스테로이드 주사 부작용... 디스크 시술 'epidural steroid inj' 부작용 탐구 논문

[문형철]

Local and Systemic Side Effects of Corticosteroid Injections for Musculoskeletal Indications
저널: American Journal of Roentgenology (AJR), 2024; 222(3): e2330458 (Epub 2023년 12월 20일)
저자: Sarah I. Kamel, MD, Humberto G. Rosas, MD, Tetyana Gorbachova, MD
DOI: 10.2214/AJR.23.30458
유형: Musculoskeletal Imaging Review (영상 유도 근골격계 주사 시 부작용 정리)

이 리뷰는

**근골격계 적응증(골관절염, 건병증, 활액낭염, 유착성 관절낭염, 신경염 등)**에 사용되는

코르티코스테로이드 주사의 국소(local) 및 전신(systemic) 부작용을 정리한 논문입니다.

부작용은

**즉시(immediate, 48시간 이내)**와 **지연(delayed, 48시간 이후)**으로 나누어

분류했습니다.

영상 유도 주사를 시행하는 의사(특히 영상의학과)는

환자 상담과 주사 계획 시 이 부작용을 반드시 고려해야 한다고 강조합니다.

장기 임상적 이득에 대한 근거가 제한적이므로(특히 무릎 등),

부작용 관리가 더욱 중요합니다.

1. 국소 부작용 (Local Side Effects)

주사 부위 자체에서 발생하는 문제입니다.

• Postinjection flare: 주사 후 일시적으로 통증이 더 심해지는 현상. 입자형 스테로이드에서 더 흔하며, 대개 1–2일 내 발생해 수일 내 호전됩니다.
• Skin hypopigmentation and atrophy: 피부 색이 옅어지고 얇아지는 변화. 표면(건주위·피하) 주사에서 특히 잘 나타납니다.
• Infection: 감염.
• Tendon rupture: 건 파열 또는 약화.
• Accelerated progression of osteoarthritis: 골관절염 진행 가속.
• Osseous injury: 골 손상(불충분 골절, 골괴사, 급속 관절 파괴 등).

2. 전신 부작용 (Systemic Side Effects)

전신으로 흡수되어 나타나는 영향입니다.

• Adrenal suppression or insufficiency: 부신 억제·기능 저하.
• Facial flushing: 안면 홍조.
• Hypertension: 고혈압(일시적).
• Hyperglycemia: 고혈당.
• Osteoporosis: 골다공증(반복·누적 사용 시).

3. 금기 (Contraindications)

다음 상황에서는 주사하면 안 됩니다.

• 표재성 또는 심부 감염 (Superficial or deep infection)
• 골절 (Fracture)
• 인공관절 (Prosthetic joint)

4. 특별 주의 환자군 (Populations With Special Considerations)

• 폐경 전 여성 (Premenopausal women)
• 당뇨병 환자 (Patients with diabetes) → 고혈당 특히 주의
• 운동선수 (Athletes)
• 소아 환자 (Pediatric patients)

주요 내용 구조

1. 코르티코스테로이드 제형 구분 (Table 1 요약)

• Particulate (불용성, 입자형): Triamcinolone acetonide (Kenalog), Methylprednisolone acetate (Depo-Medrol), Triamcinolone hexacetonide 등. 지속 시간이 길고 국소 작용이 강하나, 피부 위축·색소 침착, flare, 색전 위험(특히 척추)이 높음.
• Nonparticulate (가용성): Dexamethasone sodium phosphate (Decadron). 흡수·작용이 빠르고 피부 부작용이 적음.
• 표면(피하·건주위) 주사 시 nonparticulate 제제를 고려할 것을 권장.

핵심은

**Particulate(입자형·불용성)**와 **Nonparticulate(비입자형·가용성)**의 구분이며,

이에 따라 작용 지속 시간과 부작용 양상이 달라집니다.   

표 구성 설명

구분                                                                                               의미

Solubility	Particulate(입자형) vs Nonparticulate(비입자형)
Preparation(s)	시판 농도 (mg/mL)
Peritendinous or Bursa Injection	건주위·활액낭 주사 권장 용량
Small Joint Injection	소관절 주사 권장 용량
Large Joint Injection	대관절 주사 권장 용량
Equivalent Dose	참고용 등가 용량

제제별 요약

1. Particulate (입자형) 제제

• Triamcinolone acetonide (Kenalog)
  • 농도: 40, 80 mg/mL
  • 소관절 2.5–10 mg / 대관절 5–40 mg
  • 건주위는 package insert에 명시 없음
• Methylprednisolone acetate (Depo-Medrol)
  • 농도: 20, 40, 80 mg/mL
  • 건주위·활액낭 4–30 mg / 소관절 4–10 mg / 대관절 20–80 mg
• Triamcinolone hexacetonide (Aristospan)
  • 소관절 2–6 mg / 대관절 10–20 mg
• Triamcinolone acetonide extended release (Zilretta)
  • 무릎 관절만 32 mg 허가 (다른 부위는 허가되지 않음)

특징: 조직 내 잔류 시간이 길어 효과가 오래 가지만, 피부 위축·색소침착, postinjection flare, (척추 주사 시) 색전 위험이 상대적으로 높습니다.

2. Mixed / Nonparticulate 제제

• Betamethasone (Celestone Soluspan)
  • Acetate는 particulate, sodium phosphate는 nonparticulate
  • 건주위·활액낭 3 mg / 소관절 0.75–1.5 mg / 대관절 3 mg
• Dexamethasone sodium phosphate (Decadron) — 순수 Nonparticulate
  • 건초 0.4–1.0 mg / 활액낭 2–3 mg
  • 소관절 0.8–1.0 mg / 대관절 2–4 mg

특징: 흡수가 빠르고 작용 시간이 짧습니다. 피부 위축 위험이 낮아 표면(피하·건주위) 주사에 더 적합합니다.

임상적 시사점

• 표면 주사(테니스엘보, 드퀘르벵, 족저근막 등) → 피부 위축을 줄이려면 **nonparticulate(dexamethasone)**를 우선 고려.
• 대관절·깊은 부위 → particulate 제제가 흔히 사용되나, 반복 주사 시 연골·건 손상 위험을 고려해야 함.
• 표의 용량은 package insert 기반이며, 실제 임상에서는 부위·환자 상태·부작용 위험에 따라 조정됩니다.
• Zilretta는 현재 무릎 관절에만 허가된 서방형 제제입니다.

2. 국소 부작용 (Local)

시기                      주요 부작용                        설명·임상적 의미

즉시	Postinjection flare	주사 후 통증 악화. 입자형에서 더 흔하고 오래 지속(대개 1–2일 내 발생, 평균 2일 지속). 장기 효과에는 영향 적음. 결정성 관절염과 유사.
지연	피부 저색소침착·위축	표면 주사(테니스엘보, 드퀘르벵, 족저근막)에서 흔함. 수개월 내 발생, 대개 1년 내 호전. 입자형·needle track 이동과 관련.
	연조직 석회화	국소 조직 손상으로 유발. 통증 증가와 연관될 수 있음.
	감염	주사 후·수술 후 감염. 절대 금기(표재/심부 감염, 골절, 인공관절 존재 시).
	건·인대·근막 파열/약화	특히 건주위 반복 주사 시.
	골관절염 진행 가속	관절 내 반복 주사와 연관 보고.
	골 손상	불충분 골절, 골괴사, 급속 관절 파괴.

사례 내용

• 환자: 63세 여성, 내측 팔꿈치 통증으로 내원
• 검사 목적: 공통 굴근건(common flexor tendon)의 초음파 평가
• 주사 이력: 최근 6개월 이내에 영상 유도 없이 내측 팔꿈치에 triamcinolone acetonide (Kenalog) 40 mg을 2회 주사받음
• 사진 소견: 내측 팔꿈치에 **피부 변색(discoloration)과 얇아짐(thinning)**이 보임 (점선으로 표시된 부위)
• 발생 시기: 두 번째 주사 후 수주 뒤에 환자가 변화를 인지함

임상적 의미이 소견은 논문에서 설명하는 지연성 국소 부작용 중 하나인 **skin hypopigmentation and atrophy(피부 저색소침착 및 위축)**의 전형적인 예입니다.
주요 특징:

• 입자형(particulate) 스테로이드(Kenalog 등)에서 더 잘 발생
• 표면(피하·건주위) 주사에서 흔함 (특히 영상 유도 없이 시행할 경우 주사 경로를 따라 스테로이드가 퍼지면서 발생 가능)
• 대개 주사 후 수주~수개월 내에 나타나며, 대부분 1년 이내에 서서히 호전되는 경우가 많음
• 미용적으로 불편할 수 있으나 기능적 문제는 드묾

논문에서는 표면 주사 시 **nonparticulate 제제(dexamethasone 등)**를 고려하거나, 주사 후 주사 부위를 압박하여 needle track을 따라 약물이 퍼지는 것을 최소화할 것을 권장합니다.

3. 전신 부작용 (Systemic)

• 즉시: 부신 억제, 안면 홍조, 불면, 두통, 일시적 고혈압, 식후 고혈당, 정신과적 변화(의식 변화·환각), 딸꾹질, 과민반응, 안과적 문제(녹내장 악화·맥락망막병증), 수유 일시 억제.
• 지연·누적: 골다공증, 부신 기능 저하.

고혈당은 당뇨병 환자에서 특히 주의가 필요합니다. 안면 홍조는 비교적 흔히 보고됩니다.

4. 특수 환자군 주의

• 폐경 전 여성, 당뇨병 환자, 운동선수, 소아 → 추가 상담 필요.
• 절대 금기: 표재/심부 감염, 골절, 인공관절.

5. 결론 및 한계

• 빈도·지속 기간·평생 최대 용량에 대한 명확한 가이드라인이 부족합니다.
• 반복 사용의 장기 골 관련 합병증에 대한 추가 연구가 필요합니다.
• 주사 전 환자 상담과 제형·부위·횟수 선택이 핵심입니다.

스테로이드 국소부작용

스테로이드(코르티코스테로이드) 주사의 국소 부작용 기전은 입자형(particulate) 스테로이드의 결정성·지속적 존재, 섬유아세포·연골세포·건세포·골세포에 대한 직접적 세포독성·합성 억제, 기질(ECM) 분해 촉진, 면역 억제 등이 핵심입니다. 가장 체계적으로 정리한 고영향력 리뷰는 Gorbachova T 등, AJR Am J Roentgenol. 2024 (온라인 2023); 10.2214/AJR.23.30458으로, 제시하신 부작용 목록과 거의 동일하게 다룹니다.

1. Postinjection flare (주사 후 일시적 통증 악화)

입자형 스테로이드(triamcinolone acetonide, methylprednisolone acetate 등)의 **마이크로크리스탈(microcrystals)**이 활막이나 연조직에 흡수되면서 국소 결정 유도성 활막염(crystal-induced synovitis)을 일으킵니다. 이는 통풍과 유사한 무균성 염증 반응으로, 호중구 유입과 염증 매개체 방출을 유발합니다. 비입자형보다 발생률·지속 기간이 길며, 대개 1–2일 내 시작되어 수일 내 호전됩니다. 장기 예후에는 영향이 거의 없습니다.

2. Skin hypopigmentation and atrophy (피부 탈색·위축)

표재(피하·건주위) 주사 시 특히 잘 나타납니다.

• Atrophy: 섬유아세포·각질형성세포 증식 억제, 콜라겐·글리코사미노글리칸(GAG) 합성 감소, 지방세포 크기·수 감소로 진피·피하지방이 얇아집니다. 혈관 수축으로 인한 국소 저산소도 기여할 수 있습니다.
• Hypopigmentation: 멜라닌세포 자체는 대체로 보존되나 기능 저하(멜라닌 생성 감소)가 주된 기전으로 추정됩니다. 프로스타글란딘·사이토카인 억제가 관여할 수 있습니다. 선형 분포는 림프관으로의 과잉 스테로이드 확산 가설이 유력합니다.
  수개월 내 나타나 대개 1년 내 호전되나, 고농도·반복 시 지속될 수 있습니다.

이 논문은 손 외과에서 흔히 사용하는 corticosteroid injection (CI) 후 발생하는 연조직 위축(soft tissue atrophy) 과 피부 저색소증(hypopigmentation) 에 대한 문헌 고찰입니다. de Quervain 건초염, 외측 상과염(tennis elbow) 등에서 자주 쓰이는 주사인데, “비교적 안전한 방법”으로 여겨지지만 국소 부작용을 환자와 충분히 상담해야 한다는 점을 강조합니다.
발생 빈도와 자연 경과

• 체계적 고찰에서 연조직 위축은 약 1.5–40%, 저색소증은 약 1.3–4%로 보고됩니다.
• de Quervain 건초염에서는 최대 31%, 외측 상과염에서는 최대 40%까지 보고된 경우도 있습니다.
• 보통 주사 후 2–4개월에 나타나며(최대 10개월까지 지연 가능), 9–12개월 내에 자발적으로 호전되는 경우가 많습니다.

병태생리학

• 피하 조직: 지방세포(adipocyte)의 수와 크기가 감소하고, “lipophage-like” 대식세포가 침윤됩니다.
• 피부: type I·III 콜라겐 합성이 감소해 진피·표피가 얇아집니다.
• 저색소증: 멜라닌세포 기능 저하가 주요 원인으로 추정됩니다.
• 염증이나 괴사는 두드러지지 않습니다.

스테로이드 제제와 기술

• 용해도가 낮은(장기간 작용) 제제(예: triamcinolone acetonide)가 위축·저색소증 위험이 더 높습니다. 더 용해성 높은(단기간 작용) 제제가 상대적으로 안전할 수 있습니다.
• 용량·부피가 클수록, 주사 부위가 얕을수록 위험이 커집니다.
• 기술 권고: 목표 조직에 정확히 주입하고 주변 연조직으로의 확산을 최소화하세요. 너무 굵은 바늘은 약물이 새는 통로를 만들 수 있으므로 주의하고, 필요 시 초음파 유도를 고려할 수 있습니다. 최소 유효 용량을 사용하세요.

치료

1. 관찰·안심: 대부분 1년 이내 자연 호전되므로 우선 경과를 봅니다.
2. 12개월 후에도 호전 없으면: 자가 지방 이식(fat grafting)을 고려합니다. 단순한 충전 효과뿐 아니라 조직 질 개선에도 도움이 될 수 있다고 합니다.
3. 기타 보고된 방법: 생리식염수 연속 주사, 자가혈 주사, poly-L-lactic acid 등(근거는 제한적).

임상적 함의 (Hand Surgeons)

• 주사 전에 이 부작용을 환자와 반드시 상의해야 합니다. 특히 피부 노출이 많은 부위나 미용적 관심이 큰 환자에게 중요합니다.
• 얕은 구조(건초, 표층 연조직)에 주사할 때 위험이 더 큽니다.
• 예방이 최선: 적절한 제제 선택, 최소 용량, 정확한 주입 기술.

3. Infection (감염)

주사 자체로 인한 병원균 유입(피부 준비 불량·무균 술기 실패)이 가장 직접적 원인입니다. 스테로이드의 면역 억제 효과(호중구 기능 억제, 사이토카인 감소, 국소 방어 약화)가 감염 확산을 촉진합니다. 관절 내 주사의 경우 패혈성 관절염 위험이 매우 낮지만(0.01–0.05% 수준), 표재 연조직·골까지 진행할 수 있습니다. 인공관절·활동성 감염 부위에서는 절대 금기입니다.

4. Tendon rupture (건 파열·약화)

건세포(tenocyte)에 대한 직접적 독성이 핵심입니다.

• 건세포 증식·생존율 감소, 아포토시스 증가
• 콜라겐(특히 type I) 합성 억제, ECM 교체 교란(MMP 발현 변화)
• 단기적으로 생체역학적 강도(부하·강성) 저하(동물 모델에서 1–2주 내 25–40% 감소 보고)

입자형 스테로이드·건 내(intratendinous) 주사·고용량·반복 주사가 위험을 높입니다. 이미 퇴행된 건(Achilles, 슬개건, 회전근개 등)에서 더 잘 발생합니다. Wong 등(2004)의 인간 건세포 실험이 대표적 기초 근거입니다.

Wong MWN, Tang YN, Fu SC, Lee KM, Chan KM. Triamcinolone Suppresses Human Tenocyte Cellular Activity and Collagen Synthesis. Clin Orthop Relat Res. 2004;421:277-281.⁠Pubmed.ncbi.nlm.nih

이 논문은

triamcinolone acetonide(건 질환에 흔히 쓰이는 glucocorticoid)가

인간 건세포(tenocyte)에 미치는 직접적인 영향을 실험실 조건에서 조사한 연구입니다.

스테로이드 주사 후 건 파열 위험이 높아진다는

임상적 우려의 세포 수준 기전을 밝히고자 했습니다.

연구 가설과 방법

• 가설: Glucocorticoid가 인간 tenocyte의 세포 활성과 콜라겐 생성을 직접 억제한다.
• 세포: ACL 재건술 시 채취한 건강한 슬개건(patellar tendon)에서 primary tenocyte 배양을 했습니다.
• 측정:
  • 세포 생존율: MTT assay
  • 세포 증식: BrdU incorporation
  • 콜라겐 합성: ³H-proline incorporation

주요 결과

• 세포 생존율: Triamcinolone acetonide 10−910^{-9}10−9 ~ 10−410^{-4}10−4 mol/L 농도에서 용량 의존적으로 감소했습니다(대조군의 45–88%).
• 세포 증식: 모든 용량에서 약 87% ± 8%로 억제되었습니다.
• 콜라겐 합성: 1 μmol/L 처리 시 ³H-proline incorporation이 40 ± 2 cpm/1000 cells에서 27 ± 4 cpm/1000 cells로 감소했습니다.

결론과 임상적 함의

억제된 tenocyte 활성과 감소된 콜라겐 생성은 건의 구조를 손상시키고, 이후 자발적 건 파열 위험을 높일 수 있다고 해석했습니다. 이는 스테로이드 주사가 단기적으로는 염증·통증을 줄이지만, 장기적으로는 건 재생·유지에 부정적 영향을 줄 수 있음을 시사하는 in vitro 근거입니다.

같은 연구팀은 이후 논문에서 dexamethasone과 triamcinolone이 proteoglycan 생성도 억제하고, PDGF-BB가 일부 효과를 되돌릴 수 있음을 보고했습니다. 이 2004년 연구는 스테로이드 주사의 건 조직에 대한 직접 독성을 세포 수준에서 명확히 보여준 초기 핵심 논문 중 하나입니다.

5. Accelerated progression of osteoarthritis (골관절염 진행 가속)

연골세포(chondrocyte)에 대한 이화 작용이 중심입니다.

• 프로테오글리칸·type II 콜라겐 합성 억제
• MMP·ADAMTS 등 기질 분해 효소 상대적 증가
• 연골세포 생존율 감소 및 기질 품질 저하

McAlindon 등, JAMA 2017 (2년간 3개월마다 triamcinolone 40 mg vs 생리식염수, 초음파 활막염 있는 슬관절 OA): 연골 두께 감소가 스테로이드군에서 유의하게 더 컸습니다(−0.21 mm vs −0.10 mm). 통증 차이는 없었습니다.
Bharadwaj 등, Radiology 2025 (OAI 데이터, MRI WORMS): 코르티코스테로이드 주사군이 대조군·히알루론산군보다 2년 후 구조적 진행(특히 연골 퇴행)이 더 컸습니다. 고용량·빈번한 주사일수록 위험이 커지며, 드물게 쓰는 경우와의 임상적 의미는 여전히 논쟁 중입니다.

Bharadwaj UU, Lynch JA, Joseph GB, et al. Intra-articular Knee Injections and Progression of Knee Osteoarthritis: Data from the Osteoarthritis Initiative. Radiology. 2025;315(2):e233081.

이 논문은 Osteoarthritis Initiative(OAI) 데이터를 이용한 이차 분석으로, 무릎 골관절염(OA) 환자에서 관절강 내 주사(corticosteroid vs hyaluronic acid[HA])가 2년간 MRI로 평가한 구조적 진행과 통증에 미치는 영향을 비교했습니다.

연구 배경과 목적

무릎 OA 환자의 최소 10%가 corticosteroid 또는 HA 주사를 받습니다. 둘 다 증상 완화 효과가 알려져 있으나, 장기적인 구조적 진행에 대한 영향은 불분명했습니다. 이 연구는 whole-joint semiquantitative MRI 평가(WORMS)와 임상 결과로 2년간 연관성을 조사했습니다.

방법

• OAI(다기관, 종단 전향 연구) 데이터 활용.
• 총 210명 분석(평균 연령 64±7.9세, 여성 126명).
• Corticosteroid 주사군, HA 주사군, 주사를 받지 않은 대조군 비교.
• MRI를 주사 전·당시·2년 후 평가해 WORMS로 연골·골수병변·반월판 등 전체 관절 변화를 정량화.
• 통증은 WOMAC 점수로 평가.
• 통계: 반복측정 공분산분석(ANCOVA).

주요 결과

• 구조적 진행(WORMS):
  • Corticosteroid 주사는 대조군보다 WORMS 진행이 더 컸음(평균 차이 0.39, 95% CI 0.05–0.75, P=.02).
  • HA 대비로도 더 컸음(평균 차이 0.42, 95% CI 0.01–0.84, P=.04).
  • HA는 주사 시점 대비 2년 후 WORMS 진행이 감소함(평균 차이 −0.42, 95% CI −1.34 to −0.28, P=.003).
• 통증(WOMAC):
  • Corticosteroid: 주사 후 통증 감소(평균 차이 −5.20, P=.001).
  • HA: 통증 감소(평균 차이 −2.15, P=.04).

결론

Corticosteroid 주사는 HA나 대조군보다 OA 구조적 진행(특히 연골)을 가속화하는 것과 연관되었습니다. 반면 HA는 주사 후 최대 2년간 MRI상 진행을 감소시키는 경향을 보였습니다. 둘 다 단기 통증 완화 효과는 있었으나, corticosteroid는 장기 구조적 악화 위험이 더 높다는 점을 시사합니다.

6. Osseous injury (골 손상: 연골하 골절, 골괴사, 급속 관절 파괴)

골세포(osteocyte)·골모세포 아포토시스 증가가 핵심 기전입니다.

• 골세포 사멸 → 기계적 감지 네트워크 파괴, 골 강도·혈관성 저하
• 골모세포 억제 + 파골세포 상대적 활성으로 골 취약성 증가
• 연골하 불충분 골절(SIF) → 괴사·붕괴 → 급속 관절 파괴로 진행 가능

특히 고관절·슬관절 내 주사 후 보고되며, 기존 골 취약성(고령, 골다공증, 고용량·반복)이 있으면 위험이 높아집니다. Weinstein 등의 고전적 연구와 최근 ROS·산화 스트레스 매개 골세포 아포토시스 연구가 이를 뒷받침합니다. AJR 리뷰에서도 subchondral insufficiency fracture, osteonecrosis, rapid joint destruction를 별도 항목으로 정리했습니다.

임상적 함의 요약

• 입자형·고용량·표재 주사·건 내 주사·빈번한 관절 내 주사가 부작용 위험을 높입니다.
• 단기 항염증·진통 효과는 분명하나, 연골·건·골에 대한 누적 이화 작용이 장기 구조적 손상을 초래할 수 있습니다.
• 최신 가이드라인(ASRA 등)에서도 최소 유효 용량·빈도 제한을 권고합니다.

Weinstein RS, Nicholas RW, Manolagas SC. Apoptosis of Osteocytes in Glucocorticoid-Induced Osteonecrosis of the Hip. J Clin Endocrinol Metab. 2000;85(8):2907-2912.

이 논문은 glucocorticoid(스테로이드) 유발 대퇴골두 골괴사의 병태생리를 규명한 핵심 연구입니다. 전통적으로 “골괴사(osteonecrosis)”라고 불리던 현상이 실제로는 **골세포(osteocyte)의 아포토시스(apoptosis)**가 주된 병변임을 조직학적으로 밝혔습니다.

연구 배경과 목적

이전 동물·인간 연구에서 glucocorticoid가 osteoblast와 osteocyte의 apoptosis를 증가시킨다는 보고가 있었습니다. 이 연구는 그 apoptosis가 골괴사 병변과 공간적으로 어떻게 연관되는지, 그리고 다른 원인(알코올, 외상, 겸상적혈구질환)의 골괴사와 어떻게 다른지를 확인하고자 했습니다.

방법

• 인공관절 치환술을 받은 환자의 전체 대퇴골두 표본 사용:
  • 만성 glucocorticoid 치료로 인한 골괴사: 5명
  • 알코올 중독: 3명
  • 외상: 1명
• 겸상적혈구질환 환자의 대퇴경부 코어 생검: 5명
• 조직학적 검사로 apoptotic osteocyte와 골 표면 lining cell의 분포를 평가.

주요 결과

• Glucocorticoid 유발 군: 연골하 골절 crescent(subchondral fracture crescent) 바로 인접한 부위에 풍부한 apoptotic osteocyte와 cancellous bone lining cell이 관찰됨.
• 다른 원인 군: 외상이나 겸상적혈구질환 표본에서는 apoptotic bone cell이 거의 없었고, 알코올 관련 표본에서도 드물었음.
• 전형적인 괴사(necrosis)의 특징(세포 부종, 염증 등)보다는 apoptosis가 두드러짐.

결론과 해석

• “Glucocorticoid-induced osteonecrosis”는 잘못된 명칭입니다. 뼈가 괴사된 것이 아니라, cancellous lining cell과 osteocyte의 현저한 apoptosis가 주된 소견입니다.
• Osteocyte apoptosis는 누적되고 회복되기 어려운 결함으로, osteocyte network의 기계적 감각(mechanosensory) 기능을 파괴합니다.
• 이로 인해 대퇴골두의 하중 전달과 항상성 유지가 무너지면서 결국 골두 함몰로 이어지는 연쇄 반응이 시작됩니다.

이 연구는 glucocorticoid 관련 대퇴골두 괴사의 기전을 “혈관성 괴사” 중심에서 “골세포 아포토시스” 중심으로 재해석한 전환점이 된 논문입니다. 이후 많은 연구에서 이 개념이 인용되며, glucocorticoid가 뼈 강도를 감소시키는 메커니즘(양뿐 아니라 질적 손상)을 설명하는 데 중요한 근거로 활용되고 있습니다.

스테로이드 전신부작용

스테로이드 주사의 전신 부작용은 국소 주사 후 혈류로 흡수된 스테로이드가 전신적으로 작용하면서 발생합니다. 흡수는 관절 크기, 용량, 제형(특히 입자형·저용해도 제제), 주사 빈도, 염증 정도에 따라 달라지며, 경구 스테로이드와 유사한 기전을 공유합니다. 가장 체계적으로 정리한 고영향력 리뷰는 이전과 동일한 **Gorbachova T 등, AJR Am J Roentgenol. 2024 (doi: 10.2214/AJR.23.30458)**이며, 부신 억제에 대해서는 **European Society of Endocrinology & Endocrine Society 공동 가이드라인 (J Clin Endocrinol Metab. 2024)**이 중요합니다.

1. Adrenal suppression or insufficiency (부신 억제·기능 저하)

국소 주사된 스테로이드가 전신 흡수되면 시상하부-뇌하수체-부신(HPA) 축에 음성 피드백을 가합니다.

• 시상하부 CRH와 뇌하수체 ACTH 분비 억제 → 부신 피질(zona fasciculata/reticularis) 자극 감소 → 내인성 코르티솔 생성 저하 및 부신 위축.
  단일 관절 내 주사 후에도 혈중 코르티솔이 수시간 내 급감하고 1–4주 이상 억제될 수 있으며, 고용량·반복·다관절 주사 시 더 오래 지속됩니다. 회복은 개인차가 크며, 스트레스(감염, 수술, 외상) 시 부신 위기 위험이 있습니다. 메타분석에서 관절 내 주사의 생화학적 억제 발생률이 약 52%로 보고되기도 했습니다.

https://academic.oup.com/jcem/article/109/7/1657/7667842

Beuschlein F, Else T, Bancos I, et al. European Society of Endocrinology and Endocrine Society Joint Clinical Guideline: Diagnosis and Therapy of Glucocorticoid-induced Adrenal Insufficiency. J Clin Endocrinol Metab. 2024;109(7):1657-1683.

이 논문은 **European Society of Endocrinology(ESE)**와 Endocrine Society가 공동으로 발표한 첫 번째 임상 가이드라인입니다. 만성 glucocorticoid(스테로이드) 치료로 인한 **부신 기능부전(glucocorticoid-induced adrenal insufficiency)**의 진단과 치료에 대한 실질적 지침을 제공합니다.

배경

• Glucocorticoid는 항염증·면역억제 목적으로 널리 사용되며, 인구의 최소 1%가 만성적으로 복용하고 있어 부신 기능부전 위험에 노출됩니다.
• 위험은 용량, 기간, 약효, 투여 경로, 개인 감수성에 따라 달라집니다.
• HPA axis 억제 정도와 회복 속도는 개인차가 크고, 스테로이드 감량 시 나타나는 금단 증상과 부신 기능부전 증상이 겹쳐 구분이 어렵습니다.

주요 권고 사항 요약

가이드라인은 근거 수준(매우 낮음~높음)을 표시하며 권고(recommend)와 제안(suggest)으로 구분합니다.

1. 단기 치료(<3–4주)
   • 용량과 관계없이 tapering 없이 중단 가능. HPA axis 억제 위험이 낮아 검사 불필요.
2. 장기 치료 감량
   • 기저 질환이 조절된 경우에만 시도.
   • 생리적 용량 이상에서는 비교적 빠르게, 생리적 용량 근처에서는 천천히 감량.
   • 금단 증상(관절·근육통, 피로, 수면 장애, 기분 변화 등) 모니터링 필요.
3. 진단과 평가
   • 아침 코르티솔 검사로 HPA axis 회복 판단.

     • 300 nmol/L(10 μg/dL): 안전하게 중단 가능.

     • 150–300 nmol/L(5–10 μg/dL): 생리적 용량 유지 후 재검사.
   • 비경구 제형, 다중 제형 사용, 외인성 쿠싱 증후군 소견이 있는 경우 주의.
4. 환자 교육과 스트레스 시 대처
   • 스테로이드 복용 시 부신 기능부전 가능성과 응급 상황 대처법 교육 강조.
   • 생화학적 검사 없이도 최근 스테로이드 사용 환자라면 스트레스 상황에서 stress dose 커버 권고.
5. 기타
   • 저용량(prednisone 2.5–7.5 mg/day)도 심혈관 질환, 감염, 고혈압, 당뇨, 골다공증, 사망률 증가와 연관될 수 있음.
   • 개인별 회복 기간이 크게 다르므로 획일적 프로토콜보다 개별화된 접근 필요.

임상적 의미

이 가이드라인은 근거가 제한적인 영역에서도 실무 의사에게 공통된 기준을 제시한다는 점에서 의미가 큽니다. 특히 스테로이드 감량 속도와 금단 증상 vs 부신 기능부전 구분, 환자 교육의 중요성을 강조하며, 유럽과 미국 내분비학계의 공동 합의를 담았습니다.

2. Facial flushing (안면 홍조)

전신 흡수된 스테로이드에 의한 **일시적 혈관 확장(vasodilation)**이 주된 기전으로 추정됩니다. 얼굴·상흉부의 얇은 피부와 풍부한 표재 혈관망 때문에 눈에 잘 띕니다. 히스타민 매개 반응 또는 프로스타글란딘 등 혈관 활성 물질의 일시적 변화, 마스트세포 활성화 가설이 제기되나 정확한 기전은 완전히 규명되지 않았습니다. 발생률은 최대 40%까지 보고되며 여성에서 더 흔하고, 대개 수시간~수일 내 자연 소실됩니다. 재주사 시 재현되는 경우가 많습니다.

Kamel SI, Rosas HG, Gorbachova T. Local and Systemic Side Effects of Corticosteroid Injections for Musculoskeletal Indications. AJR Am J Roentgenol. 2024;222(3):e2330458.

이 논문은 근골격계 질환을 위한 corticosteroid 주사의 국소·전신 부작용을 정리한 리뷰입니다. 영상 유도 주사를 시행하는 방사선과 의사 등이 환자 상담과 주사 계획 시 알아야 할 내용을 종합했습니다.

부작용 분류

부작용은 **국소(local)**와 전신(systemic), 그리고 **즉시(≤48시간)**와 **지연(>48시간)**으로 구분됩니다.

국소 부작용

• 즉시: Postinjection flare (주사 후 통증 악화. 국소 활막염으로 추정)
• 지연:
  • 피부 저색소침착·위축
  • 연조직 석회화
  • 감염(주사 후 또는 수술 후)
  • 건·인대·근막 약화 또는 파열
  • 골관절염 진행 가속
  • 골 손상(insufficiency fracture, 골괴사, 급속 관절 파괴)

전신 부작용

• 즉시: 부신 억제, 안면 홍조, 불면, 두통, 일시적 고혈압, 식후 고혈당, 정신 증상, 딸꾹질, 과민반응, 안과 문제, 수유 억제 등
• 지연: 부신 기능부전·위기, 골다공증, 월경 불순·질 출혈 등

주요 포인트

• 금기: 표재성·심부 감염, 골절, 인공관절이 있는 경우.
• 특수 환자군: 폐경 전 여성, 당뇨병 환자, 운동선수, 소아에서는 추가 상담이 필요.
  • 당뇨 환자: 주사 후 혈당 상승 가능(1–3일 피크, 최대 500 mg/dL 보고).
  • 운동선수: 경기 직전 주사 비권고(건 약화 위험).
• 주사 빈도·기간·평생 최대 사용량에 대한 명확한 가이드라인은 아직 부족합니다.
• 지속적 사용의 장기 합병증(특히 골 관련)에 대한 추가 연구가 필요합니다.

임상적 의미

Corticosteroid 주사는 통증·염증 완화에 유용하지만, 국소(연골·건 손상, 피부 변화)와 전신(부신 억제, 고혈당, 골다공증) 위험이 있습니다. 영상 유도 주사를 하는 의사는 환자별 위험-이익을 고려해 상담하고, 특히 반복 주사 시 구조적 진행 가능성을 염두에 두어야 합니다.

3. Hypertension (고혈압, 일시적)

주로 염분·수분 저류와 혈관 수축 반응 증가로 발생합니다.

• 일부 스테로이드는 미네랄코르티코이드 수용체(MR)를 부분적으로 활성화하거나, 11β-HSD2 포화 시 코르티솔이 MR을 자극해 신장 원위세관에서 Na⁺ 재흡수 증가(ENaC, NCC 등).
• 혈관 평활근에서 혈관 수축 물질에 대한 민감도 증가.
  관절 내 주사 후 나타나는 고혈압은 대개 일시적이며, 고용량·반복 시 더 두드러질 수 있습니다. 순수 글루코코르티코이드 활성이 강한 제제일수록 미네랄코르티코이드 효과는 상대적으로 약합니다.

4. Hyperglycemia (고혈당)

글루코코르티코이드 수용체(GR)를 통한 다중 조직 작용으로 발생합니다.

• 간: 당신생(gluconeogenesis) 핵심 효소(PEPCK, G6Pase) 전사 촉진, 인슐린 억제 효과 감소.
• 골격근·지방조직: 인슐린 저항성 증가(GLUT4 전위 억제 등)로 말초 포도당 흡수 감소.
• 지방 분해·단백질 분해로 당신생 기질(글리세롤, 아미노산) 공급 증가.
• 췌장 β세포 기능 일시적 영향 가능.
  주사 후 수시간~1–3일에 정점을 이루며(식후 혈당 상승이 두드러짐), 당뇨병 환자에서 특히 문제가 됩니다. 혈당이 500 mg/dL 이상까지 오른 보고도 있습니다. Nature Reviews Endocrinology 등의 리뷰가 이 기전을 상세히 다룹니다.

Li JX, Cummins CL. Fresh insights into glucocorticoid-induced diabetes mellitus and new therapeutic directions. Nat Rev Endocrinol. 2022;18(9):540-557.

이 논문은 glucocorticoid(스테로이드) 유발 당뇨의 최신 기전과 치료 방향을 정리한 리뷰입니다. 항염증·면역억제 효과가 뛰어나지만 장기 사용 시 고혈당과 2형 당뇨를 유발하는 문제를 다루며, 대사 부작용을 줄이면서 항염증 효과를 유지하는 전략을 모색합니다.

핵심 내용배경과 역학

• Glucocorticoid는 1940년대부터 염증성 질환과 이식 거부 예방에 널리 사용되어 왔습니다.
• 일반 인구의 약 1%(고령에서는 2.5%)가 경구 제형을 복용하며, 장기 사용 시 고혈당 발생률이 높습니다(1개월 이상 사용 시 고혈당 ~32%, 당뇨 ~19% 보고).
• 내인성 쿠싱 증후군에서도 당뇨 유병률이 20–45%에 이릅니다.

기전

Glucocorticoid receptor 활성화가 여러 조직에서 일어나 inter-organ crosstalk를 통해 고혈당을 유발합니다:

• 직접 작용: 간(포도당 생성 증가), 지방·근육·뼈(말초 포도당 흡수 억제), 췌장 β세포(인슐린 분비 영향).
• 간접 작용: 장기 간 호르몬·대사산물 흐름을 통한 영향.
• 결과적으로 간 포도당 생산 증가 + 말초 포도당 이용 감소 → 고혈당·당뇨.

현재 관리와 한계

• 고혈당 관리 프로토콜이 부족하거나 제대로 활용되지 않는 경우가 많습니다.
• 스테로이드의 종류·용량·약효에 따라 비인슐린 또는 인슐린 치료를 선택합니다.
• 공복 혈당만으로는 놓칠 수 있어 식후 혈당 모니터링이 중요합니다.

새로운 치료 방향

1. Dissociated glucocorticoid receptor agonist / Selective GR modulators (SGRMs)
   • 항염증 효과는 유지하면서 대사 부작용(고혈당)을 분리하려는 시도. 일부 임상 시험 단계에 있습니다.
2. 조직 선택적 신호 차단
   • 특정 조직(예: 간)의 glucocorticoid 신호를 선택적으로 억제하는 전략.
3. 신규 타겟 발굴
   • 항염증과 대사 효과를 선택적으로 분리할 수 있는 새로운 분자 타겟 탐색.

임상적 의미

스테로이드의 강력한 항염증 효과를 포기하지 않으면서 당뇨·고혈당 위험을 줄이는 것이 장기 목표입니다. 현재는 용량·기간 최소화와 혈당 모니터링이 기본이지만, 조직 선택적 조절제나 새로운 분자 전략이 향후 대안으로 주목받고 있습니다.

5. Osteoporosis (골다공증, 반복·누적 사용 시)

반복·누적 노출 시 골 재형성 균형이 깨지면서 발생합니다.

• 초기: 파골세포 활성 증가로 골 흡수 촉진.
• 이후: 골모세포 증식·기능 억제, 골모세포·골세포 아포토시스 증가로 골 형성 감소.
• 장내 칼슘 흡수 감소, 신장 칼슘 배설 증가, 성호르몬 억제 등 간접 효과.
  결과적으로 골밀도 저하와 골절 위험이 증가합니다. 단일 주사로는 임상적으로 의미 있는 골다공증이 드물지만, 반복 주사나 전신 스테로이드와 병행 시 누적 위험이 커집니다. 고전적 기전은 다수의 고영향력 리뷰에서 일관되게 기술됩니다.

Local Immediate Side Effects

Postinjection flare is an exacerbation of pain after injection that is significantly higher than the baseline preinjection pain. This flare is theorized to result from focal synovitis secondary to intracellular incorporation of the injected corticosteroid particulate microcrystals [9]. Postinjection flare is more common and of longer duration for particulate than nonparticulate corticosteroids [10]. The patient experience of postinjection flare can be clinically similar to the experience of crystalline-induced arthropathies, such as gout. Studies including injections of the hand, wrist, and shoulder report an incidence of postinjection flare of 32.8–35.3% [11, 12]. A study assessing flare reactions after injections performed to treat de Quer-vain tenosynovitis and trigger finger found that most flares occurred 1–2 days after injection, lasted a mean of 2 days, and did not impact long-term symptom resolution [11].

Local Delayed Side Effects

Skin hypopigmentation and atrophy—Several prospective studies describe skin hypopigmentation and atrophy, particularly after superficial peritendinous injections (e.g., to treat epicondylitis or de Quervain tenosynovitis) [13, 14] (Fig. 1). Heel pad atrophy or skin depigmentation is reported in 10% of patients who received corticosteroid injection of the plantar fascia [15]. Skin atrophy and hypopigmentation manifest within a few months of injection and typically resolve within 1 year [16]. Such dermal changes likely relate specifically to the use of particulate steroids, especially if the injectate migrates along the injection pathway within the soft tissues [17]. When superficial structures are being injected, the use of nonparticulate steroids should be considered, to minimize the potential for skin atrophy. Additionally, compression of the injection site after removal of the needle may minimize the risk of steroid migration along the needle track [18].

Open in Viewer

Fig. 1 —63-year-old woman with medial elbow pain who presented for sonographic eval‎uation of common flexor tendon. Patient had received two prior injections of 40 mg of triamcinolone acetonide (Kenalog, Proficient Rx) in medial elbow without imaging guidance within past 6 months. Photograph obtained at time of presentation for imaging shows discoloration and thinning of skin at medial elbow (dashed outline). Patient noticed this change several weeks after second injection.

Soft-tissue calcification—Soft-tissue calcification is a reported iatrogenic complication of musculoskeletal corticosteroid injection [19]. Explanatory theories include local tissue damage resulting in increased pH that precipitates calcium salt or possible calcific necrosis of collagen fibers [19]. Soft-tissue and tendinous calcifications are associated with increased pain scores and may be clinically significant [19]. Radiologists should recognize this potential corticosteroid-induced complication when assessing for changes in periarticular calcifications on postinjection imaging.

Infection—Intraarticular infection after corticosteroid injection is very rare, with an incidence thought to be less than 0.001% [20]. The incidence may be significantly less when using single-dose vials, as well as when using modern sterile technique, packaging, and cleaning methods [21]. A systematic review of case reports describes the incidences of soft-tissue infection and osteomyelitis of the underlying bone adjacent to the injection site [22]. The injection may also increase the risk of infection if the patient later undergoes surgery at the injection site. For example, studies have found that patients who received intraarticular injections of corticosteroid before a knee or hip replacement were more likely to experience a postoperative prosthetic joint infection than those who had never received an injection [23, 24]. A meta-analysis of eight studies on hip and knee joint replacements showed a 2.13 higher odds of deep infection in patients with than without a prior injection but no significant difference in superficial infection rate [25]. Another systematic review found that intraarticular injections performed within 3 months before knee arthroplasty were significantly associated with subsequent prosthetic joint infection [26]. Moreover, a study found that the use of preoperative corticosteroids was associated with a higher likelihood of superficial or deep infections in any surgical site, as well as an increased risk of respiratory, renal, and cardiac complications [27]. Given the available data, it may be advisable to delay any surgery for 1–3 months after corticosteroid injection to potentially mitigate the infection risk, particularly in patients with multiple comorbidities and chronic conditions and who are dependent on corticosteroids [28]. It is important to consider this potential infection risk before injecting patients who may be considering arthroplasty or other surgeries soon after corticosteroid injection and to counsel such patients accordingly.

Tendon, ligament, or fascial insufficiency or rupture—A rare but significant potential consequence of corticosteroid injections is tendon rupture (Figs. 2 and 3). Although its incidence is low, reported as approximately 0.1% in a meta-analysis, case reports describe such complications throughout the body [29, 30]. A review of the case reports estimated that among 95 postinjection complications, 53.7% were plantar fascia ruptures, and 26.3% were tendon ruptures involving the patellar, quadriceps, Achilles, or biceps tendon—tendons possibly already prone to increased biomechanical stress [31]. Injected corticosteroids are thought to locally decrease collagen production, which may limit potential tendon, ligament, or fascial healing and therefore increase predisposition to structural failure or rupture [32]. Even if rupture does not occur, corticosteroid use has been associated with ligament instability in the elbow, leading to refractory pain and disability [33]. A histologic study of the Achilles tendon in rabbits reported equivalent tendon weakening and loss of elasticity after either intratendinous or retrocalcaneal bursa injection, indicating that even peritendinous injections pose a risk of tendon rupture [34]. Tendon rupture may also be more strongly associated with triamcinolone acetonide than with other steroid preparations and may be associated with frequent repeated injections to a particular site [31]. Tendon rupture has been reported to occur anywhere from 2 days to 6 weeks after injection [29]. Although evidence-based consensus guidelines are lacking, some authors suggest that a maximum of three to five injections be performed at a single site, with a minimum of 2 weeks between injections, to minimize potential tendon rupture [31].

Open in Viewer

Fig. 2A —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

A, Gray-scale sonographic image obtained using 15-MHz linear probe shows intact gluteus medius insertion (dashed lines) on greater trochanter (asterisk) in long axis. Image shows thickened tendon with moderate tendinopathy but no high-grade tear. Patient underwent ultrasound-guided percutaneous tenotomy followed by injection of 20 mg of methylprednisolone acetate (Depo-Medrol, Pfizer) and 4 mL of 0.25% bupivacaine into greater trochanteric bursa. Twelve days after procedure, patient returned due to worsening pain, which was worse with weight-bearing and did not improve after course of NSAIDS and oral steroids. Patient reported no interval trauma. MRI of osseous pelvis was performed 18 days after procedure.

Open in Viewer

Fig. 2B —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

B, Coronal STIR (B) and sagittal T2-weighted (C) images, obtained 18 days after A, through greater trochanter show full-thickness tear of left gluteus medius tendon at myotendinous junction (arrow, B) with 2-cm tendon gap (arrow, C).

Open in Viewer

Fig. 2C —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

C, Coronal STIR (B) and sagittal T2-weighted (C) images, obtained 18 days after A, through greater trochanter show full-thickness tear of left gluteus medius tendon at myotendinous junction (arrow, B) with 2-cm tendon gap (arrow, C).

Open in Viewer

Fig. 3A —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

A, Gray-scale sonographic image of right hamstring tendon origin in long axis, obtained at initial presentation, shows marked thickening and hypoechoic degeneration of tendon (dashed lines) with areas of anechoic interstitial tearing (arrow). Ischial tuberosity is denoted by asterisk. Two days after third steroid injection, patient felt snap in buttock when performing household tasks.

Open in Viewer

Fig. 3B —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

B, Coronal STIR image of bony pelvis through ischial tuberosity shows near-complete rupture of right hamstring tendon origin with almost 3 cm of distal retraction (arrow).

Open in Viewer

Fig. 3C —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

C, Gray-scale sonographic image of left hamstring origin in long axis, immediately caudal to ischial tuberosity (asterisk), shows anechoic tendon gap (arrow).

Corticosteroid injections in the shoulder have been associated with a need for revision after rotator cuff repair, possibly due to underlying degeneration of the tendon quality as a result of the injection [35]. A systematic review found a nearly twofold increase in adverse outcomes after rotator cuff repair if a corticosteroid injection is administered within 6 months before surgery or if more than two injections are performed within 1 year before surgery [36]. This effect was also shown in an analysis of patients with lateral epicondylitis who underwent surgery; those receiving three or more injections were at highest risk (OR, 3.55) for requiring revision surgery [37]. Patients must be appropriately counseled regarding the potential greater need for revision surgery if a corticosteroid is administered within 6 months before surgical tendon repair.

Accelerated progression of osteoarthritis—Accelerated progression of osteoarthritis is not strictly defined but refers to an increase in joint space loss that is greater than expected to occur in a short time period. One study defined accelerated progression of osteoarthritis as a decrease in joint space by 2 mm over a 12-month period [38]. Research from the Osteoarthritis Initiative found that patients who received intraarticular knee joint injections of corticosteroid had 3.2 higher odds of showing radiographic progression of osteoarthritis than those who did not receive injections; the odds were higher at 4.67 in patients who received continuous injections [39]. Radiographic progression of osteoarthritis after injection has also been shown in the hip, with a study reporting radiographic progression of hip joint space narrowing in 44% of patients who received hip injections and were followed postprocedurally for a mean of 6 months [40].

Osseous injury—Corticosteroids induce osteocyte apoptosis, which predisposes to decreases in osseous vascularity and bone strength; these effects are particularly pronounced after intraarticular injection [41]. Indeed, subchondral insufficiency fracture, osteonecrosis, and rapid joint destruction have all been reported after intraarticular corticosteroid injection into various joints [42–44] (Figs. 4 and 5). A retrospective single-center study of 1000 patients with long-term follow-up after hip and knee corticosteroid injections showed a combined rate of osteonecrosis, insufficiency fractures, and rapid progressive osteoarthritis of 1% [45]. Such reports are controversial, and some experts believe that the reported risk of osseous injury after intraarticular injection is inflated (possibly attributed to the natural progression of osteoarthritis) and that a risk-benefit analysis would support this approach [46].

Open in Viewer

Fig. 4A —66-year-old man who presented with left groin pain.

A, Coronal STIR image through osseous pelvis shows large left hip joint effusion (asterisk).

Open in Viewer

Fig. 4B —66-year-old man who presented with left groin pain.

B, Sagittal T2-weighted image through left hip joint shows areas of full-thickness cartilage loss and subchondral marrow edema–like signal (arrows). Patient underwent same-day injection of left hip and iliopsoas bursa, receiving total of 60 mg of triamcinolone acetonide (Kenalog, Proficient Rx). Patient presented 3 months after injection with worsening hip pain.

Open in Viewer

Fig. 4C —66-year-old man who presented with left groin pain.

C, Anteroposterior radiograph of pelvis obtained 3 months after A and B shows severe joint space narrowing of left femoral articular surface with articular surface collapse and flattening (arrow).

Open in Viewer

Fig. 4D —66-year-old man who presented with left groin pain.

D, Coronal STIR (D) and coronal T1-weighted (E) images obtained same day as radiograph (C) show severe subchondral marrow edema spanning left hip and large joint effusion. Articular surface collapse (arrow, D), with areas of subchondral sclerosis (asterisk), is also present. Differential diagnosis includes rapid progression of osteoarthritis, avascular necrosis, and septic joint. Further workup identified no evidence of infection, and patient underwent left total hip arthroplasty.

Open in Viewer

Fig. 4E —66-year-old man who presented with left groin pain.

E, Coronal STIR (D) and coronal T1-weighted (E) images obtained same day as radiograph (C) show severe subchondral marrow edema spanning left hip and large joint effusion. Articular surface collapse (arrow, D), with areas of subchondral sclerosis (asterisk), is also present. Differential diagnosis includes rapid progression of osteoarthritis, avascular necrosis, and septic joint. Further workup identified no evidence of infection, and patient underwent left total hip arthroplasty.

Open in Viewer

Fig. 5A —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

A, Preinjection anteroposterior radiograph of pelvis shows severe left hip osteoarthritis with joint space narrowing, marginal spurring (arrow), and superolateral subluxation of femoral head. Intraarticular injection of 40 mg of triamcinolone acetonide (Kenalog, Proficient Rx) was performed after radiography.

Open in Viewer

Fig. 5B —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

B, Coronal STIR (B) and coronal T1-weighted (C) images of osseous pelvis obtained 2 weeks after injection show rapid destruction of femoral articular surface (asterisk) and acetabulum (arrow), with large joint effusion.

Open in Viewer

Fig. 5C —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

C, Coronal STIR (B) and coronal T1-weighted (C) images of osseous pelvis obtained 2 weeks after injection show rapid destruction of femoral articular surface (asterisk) and acetabulum (arrow), with large joint effusion.

Open in Viewer

Fig. 5D —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

D, Anteroposterior radiograph of pelvis obtained 3 weeks after injection shows progressing marked osseous resorption of femoral head, which now extends to intertrochanteric region (arrow), with cephalad migration and lateral displacement of proximal femur. No evidence of infection was found on joint aspiration or intraoperatively during total hip arthroplasty.

An expert panel review concluded that estimates of the preval‎ence of structural problems induced by intraarticular injections are highly variable and that, at present, no imaging, clinical, or laboratory markers are established for predicting subsequent osseous injury after an intraarticular injection [47]. The presence of acute joint pain that is not explained by radiographic or sonographic findings could prompt caution, with further workup obtained before potential corticosteroid injection. In this situation, MRI may help to detect osseous pathology, such as subchondral fracture, that would be a contraindication to injection; universal implementation of this approach may be cost-prohibitive given the likely low frequency of osseous injury on MRI [47]. The use of repeat radiographs before joint injection is also controversial. The panel concluded that large clinical trials are needed to determine the true incidence of these osseous complications and to formulate preintervention safety measures [47]. Although these osseous complications have a likely very low incidence, patients should be aware of the risk, and clinical judgment should be used to assess such complications' relevance for individual patients. For example, such complications may be considered more relevant in a patient with prior long-standing steroid use or a prior diagnosis of osteopenia or osteoporosis.

Systemic Immediate Side Effects

Adrenal suppression—Locally injected corticosteroid does not remain confined to the joint, and some amount is absorbed by the bloodstream. One study found that serum cortisol levels after intraarticular injection of methylprednisolone acetate were dose-dependent, peaked 2–12 hours after injection, and showed a significant downregulation for at least 1 week regardless of the injected dose [48]. A study of 10 healthy young male athletes reported adrenal suppression for up to 2 weeks after a single intra- or extraarticular injection [49]. A study found that injecting two joints simultaneously yields higher serum absorption than injecting the same total dose into one joint; this finding is theorized to be due to the exposure of a greater volume of synovium to corticosteroid, thus increasing the likelihood of systemic absorption [48]. This systemic absorption of corticosteroid accounts for many of the side effects occurring after injection, as well as the such side effects' similarity to the known side effects of exogenous steroid medication.

One study reported facial flushing (defined as erythema of the upper chest and face) after intraarticular corticosteroid injection in 40% of patients, with this side effect being unpleasant in 15% in cases, predominantly occurring in women, and being reproducible with subsequent injection [50]. Flushing is self-limiting, typically resolving within 3 days [51]. Associated systemic symptoms can include insomnia and headache.

Transient hypertension and postprandial hyperglycemia occur after intraarticular corticosteroid injection in patients both with and without diabetes and relate to suppression of the corticotropic axis; these effects may last for over 1 week [52]. These side effects' clinical implications depend on the patient's baseline condition.

Similar to the known side effects of oral or IV steroid use, psychiatric effects (e.g., altered mental status, visual hallucination) have been reported after intraarticular injection, typically occurring within 1 week after injection and occurring more commonly in older patients with decreased renal function [53, 54]. Prolonged hiccups have been reported after corticosteroid injection and are of indeterminate etiology [55].

Hypersensitivity reactions—The reported preval‎ence of hypersensitivity reactions after corticosteroid administration by any route is 0.3–0.5%; a common presentation of such reactions after musculoskeletal injection is delayed urticaria at the injection site [56, 57]. A case report described immediate anaphylaxis after intraarticular injection of triamcinolone acetonide, with confirmed hypersensitivity reaction by skin prick test [57]. Patients with frequent exposure to steroids are more likely to experience a hypersensitivity reaction. In addition, patients reporting a prior hypersensitivity reaction to corticosteroids should undergo workup for potential allergy to any alternative steroid preparation that is selected before it is administered [56]. Management of hypersensitivity reactions may include administration of oral or IV diphenhydramine hydrochloride, depending on the reaction's severity [56, 57].

Ophthalmic side effects—Central serous chorioretinopathy is a retinal detachment presenting with potentially permanent decreased vision that has been attributed to exogenously administered corticosteroids [58]. Any form of corticosteroid administration can precipitate glaucoma or increased intraocular pressure, the latter of which may lead to sudden-onset blurred vision or potentially irreversible vision loss [59]. These potential complications must be discussed before corticosteroid injection with patients who have known ophthalmic pathology and with patients reporting prior visual changes after steroid exposure.

Systemic Delayed Side Effects

As mentioned previously, intraarticular corticosteroid injection suppresses the hypothalamic-pituitary-adrenal axis, causing an at least temporary decrease in endogenous cortisol levels in 52.2% of patients [60]. This suppression can be potentially life-threatening in patients who are subject to an emotional or physical stress, such as illness or surgery, that may precipitate an adrenal crisis [6]. Most patients with adrenal insufficiency are not diagnosed until they are hospitalized for adrenal crisis, which has a reported mortality of 5% [61]. Prolonged adrenal insufficiency has been reported in some patients after repeated corticosteroid injections (i.e., iatrogenic Cushing syndrome) and may require exogenous supplementation to avoid an adrenal crisis [62, 63]. Although limited data describe prolonged adrenal suppression after musculoskeletal injection, a meta-analysis found that adrenal insufficiency may be present in 25.3% of patients 6 months after any corticosteroid administration [60]. Due to individual variation in glucocorticoid sensitivity, there is no corticosteroid dose, administration route, or treatment duration that prevents adrenal suppression [60, 64]. Furthermore, some suggest that no corticosteroids be administered for at least 4 weeks before planned surgery, to decrease the chance of adrenal crisis [28].

Osteoporosis is a known complication of long-term exogenous steroid use, attributed to steroid's negative effect on both bone quality and bone mass. A statistically significant decrease in bone mineral density of the femur has been reported after a single epidural steroid injection in postmenopausal women [65]. The literature suggests that the effect of epidural corticosteroid injection on bone mineral density is dose-dependent, whether from a single injection or multiple injections over time (i.e., cumulative dose), and may also predispose to increased vertebral fracture risk [28]. The incidence and impact of osteoporotic fractures from intra- and extraarticular corticosteroid injections for musculoskeletal indications are unknown. However, given a recognized impact of epidural steroid injections on bone mineral density, the radiologist should provide a recommendation to the primary care practitioner to consider initiating prophylactic osteoporotic treatment in postmenopausal women receiving frequent or multiple injections [28, 65]. Local Immediate Side Effects

Postinjection flare is an exacerbation of pain after injection that is significantly higher than the baseline preinjection pain. This flare is theorized to result from focal synovitis secondary to intracellular incorporation of the injected corticosteroid particulate microcrystals [9]. Postinjection flare is more common and of longer duration for particulate than nonparticulate corticosteroids [10]. The patient experience of postinjection flare can be clinically similar to the experience of crystalline-induced arthropathies, such as gout. Studies including injections of the hand, wrist, and shoulder report an incidence of postinjection flare of 32.8–35.3% [11, 12]. A study assessing flare reactions after injections performed to treat de Quer-vain tenosynovitis and trigger finger found that most flares occurred 1–2 days after injection, lasted a mean of 2 days, and did not impact long-term symptom resolution [11].

Local Delayed Side Effects

Skin hypopigmentation and atrophy—Several prospective studies describe skin hypopigmentation and atrophy, particularly after superficial peritendinous injections (e.g., to treat epicondylitis or de Quervain tenosynovitis) [13, 14] (Fig. 1). Heel pad atrophy or skin depigmentation is reported in 10% of patients who received corticosteroid injection of the plantar fascia [15]. Skin atrophy and hypopigmentation manifest within a few months of injection and typically resolve within 1 year [16]. Such dermal changes likely relate specifically to the use of particulate steroids, especially if the injectate migrates along the injection pathway within the soft tissues [17]. When superficial structures are being injected, the use of nonparticulate steroids should be considered, to minimize the potential for skin atrophy. Additionally, compression of the injection site after removal of the needle may minimize the risk of steroid migration along the needle track [18].

Open in Viewer

Fig. 1 —63-year-old woman with medial elbow pain who presented for sonographic eval‎uation of common flexor tendon. Patient had received two prior injections of 40 mg of triamcinolone acetonide (Kenalog, Proficient Rx) in medial elbow without imaging guidance within past 6 months. Photograph obtained at time of presentation for imaging shows discoloration and thinning of skin at medial elbow (dashed outline). Patient noticed this change several weeks after second injection.

Soft-tissue calcification—Soft-tissue calcification is a reported iatrogenic complication of musculoskeletal corticosteroid injection [19]. Explanatory theories include local tissue damage resulting in increased pH that precipitates calcium salt or possible calcific necrosis of collagen fibers [19]. Soft-tissue and tendinous calcifications are associated with increased pain scores and may be clinically significant [19]. Radiologists should recognize this potential corticosteroid-induced complication when assessing for changes in periarticular calcifications on postinjection imaging.

Infection—Intraarticular infection after corticosteroid injection is very rare, with an incidence thought to be less than 0.001% [20]. The incidence may be significantly less when using single-dose vials, as well as when using modern sterile technique, packaging, and cleaning methods [21]. A systematic review of case reports describes the incidences of soft-tissue infection and osteomyelitis of the underlying bone adjacent to the injection site [22]. The injection may also increase the risk of infection if the patient later undergoes surgery at the injection site. For example, studies have found that patients who received intraarticular injections of corticosteroid before a knee or hip replacement were more likely to experience a postoperative prosthetic joint infection than those who had never received an injection [23, 24]. A meta-analysis of eight studies on hip and knee joint replacements showed a 2.13 higher odds of deep infection in patients with than without a prior injection but no significant difference in superficial infection rate [25]. Another systematic review found that intraarticular injections performed within 3 months before knee arthroplasty were significantly associated with subsequent prosthetic joint infection [26]. Moreover, a study found that the use of preoperative corticosteroids was associated with a higher likelihood of superficial or deep infections in any surgical site, as well as an increased risk of respiratory, renal, and cardiac complications [27]. Given the available data, it may be advisable to delay any surgery for 1–3 months after corticosteroid injection to potentially mitigate the infection risk, particularly in patients with multiple comorbidities and chronic conditions and who are dependent on corticosteroids [28]. It is important to consider this potential infection risk before injecting patients who may be considering arthroplasty or other surgeries soon after corticosteroid injection and to counsel such patients accordingly.

Tendon, ligament, or fascial insufficiency or rupture—A rare but significant potential consequence of corticosteroid injections is tendon rupture (Figs. 2 and 3). Although its incidence is low, reported as approximately 0.1% in a meta-analysis, case reports describe such complications throughout the body [29, 30]. A review of the case reports estimated that among 95 postinjection complications, 53.7% were plantar fascia ruptures, and 26.3% were tendon ruptures involving the patellar, quadriceps, Achilles, or biceps tendon—tendons possibly already prone to increased biomechanical stress [31]. Injected corticosteroids are thought to locally decrease collagen production, which may limit potential tendon, ligament, or fascial healing and therefore increase predisposition to structural failure or rupture [32]. Even if rupture does not occur, corticosteroid use has been associated with ligament instability in the elbow, leading to refractory pain and disability [33]. A histologic study of the Achilles tendon in rabbits reported equivalent tendon weakening and loss of elasticity after either intratendinous or retrocalcaneal bursa injection, indicating that even peritendinous injections pose a risk of tendon rupture [34]. Tendon rupture may also be more strongly associated with triamcinolone acetonide than with other steroid preparations and may be associated with frequent repeated injections to a particular site [31]. Tendon rupture has been reported to occur anywhere from 2 days to 6 weeks after injection [29]. Although evidence-based consensus guidelines are lacking, some authors suggest that a maximum of three to five injections be performed at a single site, with a minimum of 2 weeks between injections, to minimize potential tendon rupture [31].

Open in Viewer

Fig. 2A —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

A, Gray-scale sonographic image obtained using 15-MHz linear probe shows intact gluteus medius insertion (dashed lines) on greater trochanter (asterisk) in long axis. Image shows thickened tendon with moderate tendinopathy but no high-grade tear. Patient underwent ultrasound-guided percutaneous tenotomy followed by injection of 20 mg of methylprednisolone acetate (Depo-Medrol, Pfizer) and 4 mL of 0.25% bupivacaine into greater trochanteric bursa. Twelve days after procedure, patient returned due to worsening pain, which was worse with weight-bearing and did not improve after course of NSAIDS and oral steroids. Patient reported no interval trauma. MRI of osseous pelvis was performed 18 days after procedure.

Open in Viewer

Fig. 2B —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

B, Coronal STIR (B) and sagittal T2-weighted (C) images, obtained 18 days after A, through greater trochanter show full-thickness tear of left gluteus medius tendon at myotendinous junction (arrow, B) with 2-cm tendon gap (arrow, C).

Open in Viewer

Fig. 2C —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

C, Coronal STIR (B) and sagittal T2-weighted (C) images, obtained 18 days after A, through greater trochanter show full-thickness tear of left gluteus medius tendon at myotendinous junction (arrow, B) with 2-cm tendon gap (arrow, C).

Open in Viewer

Fig. 3A —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

A, Gray-scale sonographic image of right hamstring tendon origin in long axis, obtained at initial presentation, shows marked thickening and hypoechoic degeneration of tendon (dashed lines) with areas of anechoic interstitial tearing (arrow). Ischial tuberosity is denoted by asterisk. Two days after third steroid injection, patient felt snap in buttock when performing household tasks.

Open in Viewer

Fig. 3B —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

B, Coronal STIR image of bony pelvis through ischial tuberosity shows near-complete rupture of right hamstring tendon origin with almost 3 cm of distal retraction (arrow).

Open in Viewer

Fig. 3C —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

C, Gray-scale sonographic image of left hamstring origin in long axis, immediately caudal to ischial tuberosity (asterisk), shows anechoic tendon gap (arrow).

Corticosteroid injections in the shoulder have been associated with a need for revision after rotator cuff repair, possibly due to underlying degeneration of the tendon quality as a result of the injection [35]. A systematic review found a nearly twofold increase in adverse outcomes after rotator cuff repair if a corticosteroid injection is administered within 6 months before surgery or if more than two injections are performed within 1 year before surgery [36]. This effect was also shown in an analysis of patients with lateral epicondylitis who underwent surgery; those receiving three or more injections were at highest risk (OR, 3.55) for requiring revision surgery [37]. Patients must be appropriately counseled regarding the potential greater need for revision surgery if a corticosteroid is administered within 6 months before surgical tendon repair.

Accelerated progression of osteoarthritis—Accelerated progression of osteoarthritis is not strictly defined but refers to an increase in joint space loss that is greater than expected to occur in a short time period. One study defined accelerated progression of osteoarthritis as a decrease in joint space by 2 mm over a 12-month period [38]. Research from the Osteoarthritis Initiative found that patients who received intraarticular knee joint injections of corticosteroid had 3.2 higher odds of showing radiographic progression of osteoarthritis than those who did not receive injections; the odds were higher at 4.67 in patients who received continuous injections [39]. Radiographic progression of osteoarthritis after injection has also been shown in the hip, with a study reporting radiographic progression of hip joint space narrowing in 44% of patients who received hip injections and were followed postprocedurally for a mean of 6 months [40].

Osseous injury—Corticosteroids induce osteocyte apoptosis, which predisposes to decreases in osseous vascularity and bone strength; these effects are particularly pronounced after intraarticular injection [41]. Indeed, subchondral insufficiency fracture, osteonecrosis, and rapid joint destruction have all been reported after intraarticular corticosteroid injection into various joints [42–44] (Figs. 4 and 5). A retrospective single-center study of 1000 patients with long-term follow-up after hip and knee corticosteroid injections showed a combined rate of osteonecrosis, insufficiency fractures, and rapid progressive osteoarthritis of 1% [45]. Such reports are controversial, and some experts believe that the reported risk of osseous injury after intraarticular injection is inflated (possibly attributed to the natural progression of osteoarthritis) and that a risk-benefit analysis would support this approach [46].

Open in Viewer

Fig. 4A —66-year-old man who presented with left groin pain.

A, Coronal STIR image through osseous pelvis shows large left hip joint effusion (asterisk).

Open in Viewer

Fig. 4B —66-year-old man who presented with left groin pain.

B, Sagittal T2-weighted image through left hip joint shows areas of full-thickness cartilage loss and subchondral marrow edema–like signal (arrows). Patient underwent same-day injection of left hip and iliopsoas bursa, receiving total of 60 mg of triamcinolone acetonide (Kenalog, Proficient Rx). Patient presented 3 months after injection with worsening hip pain.

Open in Viewer

Fig. 4C —66-year-old man who presented with left groin pain.

C, Anteroposterior radiograph of pelvis obtained 3 months after A and B shows severe joint space narrowing of left femoral articular surface with articular surface collapse and flattening (arrow).

Open in Viewer

Fig. 4D —66-year-old man who presented with left groin pain.

D, Coronal STIR (D) and coronal T1-weighted (E) images obtained same day as radiograph (C) show severe subchondral marrow edema spanning left hip and large joint effusion. Articular surface collapse (arrow, D), with areas of subchondral sclerosis (asterisk), is also present. Differential diagnosis includes rapid progression of osteoarthritis, avascular necrosis, and septic joint. Further workup identified no evidence of infection, and patient underwent left total hip arthroplasty.

Open in Viewer

Fig. 4E —66-year-old man who presented with left groin pain.

E, Coronal STIR (D) and coronal T1-weighted (E) images obtained same day as radiograph (C) show severe subchondral marrow edema spanning left hip and large joint effusion. Articular surface collapse (arrow, D), with areas of subchondral sclerosis (asterisk), is also present. Differential diagnosis includes rapid progression of osteoarthritis, avascular necrosis, and septic joint. Further workup identified no evidence of infection, and patient underwent left total hip arthroplasty.

Open in Viewer

Fig. 5A —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

A, Preinjection anteroposterior radiograph of pelvis shows severe left hip osteoarthritis with joint space narrowing, marginal spurring (arrow), and superolateral subluxation of femoral head. Intraarticular injection of 40 mg of triamcinolone acetonide (Kenalog, Proficient Rx) was performed after radiography.

Open in Viewer

Fig. 5B —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

B, Coronal STIR (B) and coronal T1-weighted (C) images of osseous pelvis obtained 2 weeks after injection show rapid destruction of femoral articular surface (asterisk) and acetabulum (arrow), with large joint effusion.

Open in Viewer

Fig. 5C —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

C, Coronal STIR (B) and coronal T1-weighted (C) images of osseous pelvis obtained 2 weeks after injection show rapid destruction of femoral articular surface (asterisk) and acetabulum (arrow), with large joint effusion.

Open in Viewer

Fig. 5D —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

D, Anteroposterior radiograph of pelvis obtained 3 weeks after injection shows progressing marked osseous resorption of femoral head, which now extends to intertrochanteric region (arrow), with cephalad migration and lateral displacement of proximal femur. No evidence of infection was found on joint aspiration or intraoperatively during total hip arthroplasty.

An expert panel review concluded that estimates of the preval‎ence of structural problems induced by intraarticular injections are highly variable and that, at present, no imaging, clinical, or laboratory markers are established for predicting subsequent osseous injury after an intraarticular injection [47]. The presence of acute joint pain that is not explained by radiographic or sonographic findings could prompt caution, with further workup obtained before potential corticosteroid injection. In this situation, MRI may help to detect osseous pathology, such as subchondral fracture, that would be a contraindication to injection; universal implementation of this approach may be cost-prohibitive given the likely low frequency of osseous injury on MRI [47]. The use of repeat radiographs before joint injection is also controversial. The panel concluded that large clinical trials are needed to determine the true incidence of these osseous complications and to formulate preintervention safety measures [47]. Although these osseous complications have a likely very low incidence, patients should be aware of the risk, and clinical judgment should be used to assess such complications' relevance for individual patients. For example, such complications may be considered more relevant in a patient with prior long-standing steroid use or a prior diagnosis of osteopenia or osteoporosis.

Systemic Immediate Side Effects

Adrenal suppression—Locally injected corticosteroid does not remain confined to the joint, and some amount is absorbed by the bloodstream. One study found that serum cortisol levels after intraarticular injection of methylprednisolone acetate were dose-dependent, peaked 2–12 hours after injection, and showed a significant downregulation for at least 1 week regardless of the injected dose [48]. A study of 10 healthy young male athletes reported adrenal suppression for up to 2 weeks after a single intra- or extraarticular injection [49]. A study found that injecting two joints simultaneously yields higher serum absorption than injecting the same total dose into one joint; this finding is theorized to be due to the exposure of a greater volume of synovium to corticosteroid, thus increasing the likelihood of systemic absorption [48]. This systemic absorption of corticosteroid accounts for many of the side effects occurring after injection, as well as the such side effects' similarity to the known side effects of exogenous steroid medication.

One study reported facial flushing (defined as erythema of the upper chest and face) after intraarticular corticosteroid injection in 40% of patients, with this side effect being unpleasant in 15% in cases, predominantly occurring in women, and being reproducible with subsequent injection [50]. Flushing is self-limiting, typically resolving within 3 days [51]. Associated systemic symptoms can include insomnia and headache.

Transient hypertension and postprandial hyperglycemia occur after intraarticular corticosteroid injection in patients both with and without diabetes and relate to suppression of the corticotropic axis; these effects may last for over 1 week [52]. These side effects' clinical implications depend on the patient's baseline condition.

Similar to the known side effects of oral or IV steroid use, psychiatric effects (e.g., altered mental status, visual hallucination) have been reported after intraarticular injection, typically occurring within 1 week after injection and occurring more commonly in older patients with decreased renal function [53, 54]. Prolonged hiccups have been reported after corticosteroid injection and are of indeterminate etiology [55].

Hypersensitivity reactions—The reported preval‎ence of hypersensitivity reactions after corticosteroid administration by any route is 0.3–0.5%; a common presentation of such reactions after musculoskeletal injection is delayed urticaria at the injection site [56, 57]. A case report described immediate anaphylaxis after intraarticular injection of triamcinolone acetonide, with confirmed hypersensitivity reaction by skin prick test [57]. Patients with frequent exposure to steroids are more likely to experience a hypersensitivity reaction. In addition, patients reporting a prior hypersensitivity reaction to corticosteroids should undergo workup for potential allergy to any alternative steroid preparation that is selected before it is administered [56]. Management of hypersensitivity reactions may include administration of oral or IV diphenhydramine hydrochloride, depending on the reaction's severity [56, 57].

Ophthalmic side effects—Central serous chorioretinopathy is a retinal detachment presenting with potentially permanent decreased vision that has been attributed to exogenously administered corticosteroids [58]. Any form of corticosteroid administration can precipitate glaucoma or increased intraocular pressure, the latter of which may lead to sudden-onset blurred vision or potentially irreversible vision loss [59]. These potential complications must be discussed before corticosteroid injection with patients who have known ophthalmic pathology and with patients reporting prior visual changes after steroid exposure.

Systemic Delayed Side Effects

As mentioned previously, intraarticular corticosteroid injection suppresses the hypothalamic-pituitary-adrenal axis, causing an at least temporary decrease in endogenous cortisol levels in 52.2% of patients [60]. This suppression can be potentially life-threatening in patients who are subject to an emotional or physical stress, such as illness or surgery, that may precipitate an adrenal crisis [6]. Most patients with adrenal insufficiency are not diagnosed until they are hospitalized for adrenal crisis, which has a reported mortality of 5% [61]. Prolonged adrenal insufficiency has been reported in some patients after repeated corticosteroid injections (i.e., iatrogenic Cushing syndrome) and may require exogenous supplementation to avoid an adrenal crisis [62, 63]. Although limited data describe prolonged adrenal suppression after musculoskeletal injection, a meta-analysis found that adrenal insufficiency may be present in 25.3% of patients 6 months after any corticosteroid administration [60]. Due to individual variation in glucocorticoid sensitivity, there is no corticosteroid dose, administration route, or treatment duration that prevents adrenal suppression [60, 64]. Furthermore, some suggest that no corticosteroids be administered for at least 4 weeks before planned surgery, to decrease the chance of adrenal crisis [28].

Osteoporosis is a known complication of long-term exogenous steroid use, attributed to steroid's negative effect on both bone quality and bone mass. A statistically significant decrease in bone mineral density of the femur has been reported after a single epidural steroid injection in postmenopausal women [65]. The literature suggests that the effect of epidural corticosteroid injection on bone mineral density is dose-dependent, whether from a single injection or multiple injections over time (i.e., cumulative dose), and may also predispose to increased vertebral fracture risk [28]. The incidence and impact of osteoporotic fractures from intra- and extraarticular corticosteroid injections for musculoskeletal indications are unknown. However, given a recognized impact of epidural steroid injections on bone mineral density, the radiologist should provide a recommendation to the primary care practitioner to consider initiating prophylactic osteoporotic treatment in postmenopausal women receiving frequent or multiple injections [28, 65]. Local Immediate Side Effects

Postinjection flare is an exacerbation of pain after injection that is significantly higher than the baseline preinjection pain. This flare is theorized to result from focal synovitis secondary to intracellular incorporation of the injected corticosteroid particulate microcrystals [9]. Postinjection flare is more common and of longer duration for particulate than nonparticulate corticosteroids [10]. The patient experience of postinjection flare can be clinically similar to the experience of crystalline-induced arthropathies, such as gout. Studies including injections of the hand, wrist, and shoulder report an incidence of postinjection flare of 32.8–35.3% [11, 12]. A study assessing flare reactions after injections performed to treat de Quer-vain tenosynovitis and trigger finger found that most flares occurred 1–2 days after injection, lasted a mean of 2 days, and did not impact long-term symptom resolution [11].

Local Delayed Side Effects

Skin hypopigmentation and atrophy—Several prospective studies describe skin hypopigmentation and atrophy, particularly after superficial peritendinous injections (e.g., to treat epicondylitis or de Quervain tenosynovitis) [13, 14] (Fig. 1). Heel pad atrophy or skin depigmentation is reported in 10% of patients who received corticosteroid injection of the plantar fascia [15]. Skin atrophy and hypopigmentation manifest within a few months of injection and typically resolve within 1 year [16]. Such dermal changes likely relate specifically to the use of particulate steroids, especially if the injectate migrates along the injection pathway within the soft tissues [17]. When superficial structures are being injected, the use of nonparticulate steroids should be considered, to minimize the potential for skin atrophy. Additionally, compression of the injection site after removal of the needle may minimize the risk of steroid migration along the needle track [18].

Open in Viewer

Fig. 1 —63-year-old woman with medial elbow pain who presented for sonographic eval‎uation of common flexor tendon. Patient had received two prior injections of 40 mg of triamcinolone acetonide (Kenalog, Proficient Rx) in medial elbow without imaging guidance within past 6 months. Photograph obtained at time of presentation for imaging shows discoloration and thinning of skin at medial elbow (dashed outline). Patient noticed this change several weeks after second injection.

Soft-tissue calcification—Soft-tissue calcification is a reported iatrogenic complication of musculoskeletal corticosteroid injection [19]. Explanatory theories include local tissue damage resulting in increased pH that precipitates calcium salt or possible calcific necrosis of collagen fibers [19]. Soft-tissue and tendinous calcifications are associated with increased pain scores and may be clinically significant [19]. Radiologists should recognize this potential corticosteroid-induced complication when assessing for changes in periarticular calcifications on postinjection imaging.

Infection—Intraarticular infection after corticosteroid injection is very rare, with an incidence thought to be less than 0.001% [20]. The incidence may be significantly less when using single-dose vials, as well as when using modern sterile technique, packaging, and cleaning methods [21]. A systematic review of case reports describes the incidences of soft-tissue infection and osteomyelitis of the underlying bone adjacent to the injection site [22]. The injection may also increase the risk of infection if the patient later undergoes surgery at the injection site. For example, studies have found that patients who received intraarticular injections of corticosteroid before a knee or hip replacement were more likely to experience a postoperative prosthetic joint infection than those who had never received an injection [23, 24]. A meta-analysis of eight studies on hip and knee joint replacements showed a 2.13 higher odds of deep infection in patients with than without a prior injection but no significant difference in superficial infection rate [25]. Another systematic review found that intraarticular injections performed within 3 months before knee arthroplasty were significantly associated with subsequent prosthetic joint infection [26]. Moreover, a study found that the use of preoperative corticosteroids was associated with a higher likelihood of superficial or deep infections in any surgical site, as well as an increased risk of respiratory, renal, and cardiac complications [27]. Given the available data, it may be advisable to delay any surgery for 1–3 months after corticosteroid injection to potentially mitigate the infection risk, particularly in patients with multiple comorbidities and chronic conditions and who are dependent on corticosteroids [28]. It is important to consider this potential infection risk before injecting patients who may be considering arthroplasty or other surgeries soon after corticosteroid injection and to counsel such patients accordingly.

Tendon, ligament, or fascial insufficiency or rupture—A rare but significant potential consequence of corticosteroid injections is tendon rupture (Figs. 2 and 3). Although its incidence is low, reported as approximately 0.1% in a meta-analysis, case reports describe such complications throughout the body [29, 30]. A review of the case reports estimated that among 95 postinjection complications, 53.7% were plantar fascia ruptures, and 26.3% were tendon ruptures involving the patellar, quadriceps, Achilles, or biceps tendon—tendons possibly already prone to increased biomechanical stress [31]. Injected corticosteroids are thought to locally decrease collagen production, which may limit potential tendon, ligament, or fascial healing and therefore increase predisposition to structural failure or rupture [32]. Even if rupture does not occur, corticosteroid use has been associated with ligament instability in the elbow, leading to refractory pain and disability [33]. A histologic study of the Achilles tendon in rabbits reported equivalent tendon weakening and loss of elasticity after either intratendinous or retrocalcaneal bursa injection, indicating that even peritendinous injections pose a risk of tendon rupture [34]. Tendon rupture may also be more strongly associated with triamcinolone acetonide than with other steroid preparations and may be associated with frequent repeated injections to a particular site [31]. Tendon rupture has been reported to occur anywhere from 2 days to 6 weeks after injection [29]. Although evidence-based consensus guidelines are lacking, some authors suggest that a maximum of three to five injections be performed at a single site, with a minimum of 2 weeks between injections, to minimize potential tendon rupture [31].

Open in Viewer

Fig. 2A —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

A, Gray-scale sonographic image obtained using 15-MHz linear probe shows intact gluteus medius insertion (dashed lines) on greater trochanter (asterisk) in long axis. Image shows thickened tendon with moderate tendinopathy but no high-grade tear. Patient underwent ultrasound-guided percutaneous tenotomy followed by injection of 20 mg of methylprednisolone acetate (Depo-Medrol, Pfizer) and 4 mL of 0.25% bupivacaine into greater trochanteric bursa. Twelve days after procedure, patient returned due to worsening pain, which was worse with weight-bearing and did not improve after course of NSAIDS and oral steroids. Patient reported no interval trauma. MRI of osseous pelvis was performed 18 days after procedure.

Open in Viewer

Fig. 2B —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

B, Coronal STIR (B) and sagittal T2-weighted (C) images, obtained 18 days after A, through greater trochanter show full-thickness tear of left gluteus medius tendon at myotendinous junction (arrow, B) with 2-cm tendon gap (arrow, C).

Open in Viewer

Fig. 2C —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

C, Coronal STIR (B) and sagittal T2-weighted (C) images, obtained 18 days after A, through greater trochanter show full-thickness tear of left gluteus medius tendon at myotendinous junction (arrow, B) with 2-cm tendon gap (arrow, C).

Open in Viewer

Fig. 3A —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

A, Gray-scale sonographic image of right hamstring tendon origin in long axis, obtained at initial presentation, shows marked thickening and hypoechoic degeneration of tendon (dashed lines) with areas of anechoic interstitial tearing (arrow). Ischial tuberosity is denoted by asterisk. Two days after third steroid injection, patient felt snap in buttock when performing household tasks.

Open in Viewer

Fig. 3B —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

B, Coronal STIR image of bony pelvis through ischial tuberosity shows near-complete rupture of right hamstring tendon origin with almost 3 cm of distal retraction (arrow).

Open in Viewer

Fig. 3C —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

C, Gray-scale sonographic image of left hamstring origin in long axis, immediately caudal to ischial tuberosity (asterisk), shows anechoic tendon gap (arrow).

Corticosteroid injections in the shoulder have been associated with a need for revision after rotator cuff repair, possibly due to underlying degeneration of the tendon quality as a result of the injection [35]. A systematic review found a nearly twofold increase in adverse outcomes after rotator cuff repair if a corticosteroid injection is administered within 6 months before surgery or if more than two injections are performed within 1 year before surgery [36]. This effect was also shown in an analysis of patients with lateral epicondylitis who underwent surgery; those receiving three or more injections were at highest risk (OR, 3.55) for requiring revision surgery [37]. Patients must be appropriately counseled regarding the potential greater need for revision surgery if a corticosteroid is administered within 6 months before surgical tendon repair.

Accelerated progression of osteoarthritis—Accelerated progression of osteoarthritis is not strictly defined but refers to an increase in joint space loss that is greater than expected to occur in a short time period. One study defined accelerated progression of osteoarthritis as a decrease in joint space by 2 mm over a 12-month period [38]. Research from the Osteoarthritis Initiative found that patients who received intraarticular knee joint injections of corticosteroid had 3.2 higher odds of showing radiographic progression of osteoarthritis than those who did not receive injections; the odds were higher at 4.67 in patients who received continuous injections [39]. Radiographic progression of osteoarthritis after injection has also been shown in the hip, with a study reporting radiographic progression of hip joint space narrowing in 44% of patients who received hip injections and were followed postprocedurally for a mean of 6 months [40].

Osseous injury—Corticosteroids induce osteocyte apoptosis, which predisposes to decreases in osseous vascularity and bone strength; these effects are particularly pronounced after intraarticular injection [41]. Indeed, subchondral insufficiency fracture, osteonecrosis, and rapid joint destruction have all been reported after intraarticular corticosteroid injection into various joints [42–44] (Figs. 4 and 5). A retrospective single-center study of 1000 patients with long-term follow-up after hip and knee corticosteroid injections showed a combined rate of osteonecrosis, insufficiency fractures, and rapid progressive osteoarthritis of 1% [45]. Such reports are controversial, and some experts believe that the reported risk of osseous injury after intraarticular injection is inflated (possibly attributed to the natural progression of osteoarthritis) and that a risk-benefit analysis would support this approach [46].

Open in Viewer

Fig. 4A —66-year-old man who presented with left groin pain.

A, Coronal STIR image through osseous pelvis shows large left hip joint effusion (asterisk).

Open in Viewer

Fig. 4B —66-year-old man who presented with left groin pain.

B, Sagittal T2-weighted image through left hip joint shows areas of full-thickness cartilage loss and subchondral marrow edema–like signal (arrows). Patient underwent same-day injection of left hip and iliopsoas bursa, receiving total of 60 mg of triamcinolone acetonide (Kenalog, Proficient Rx). Patient presented 3 months after injection with worsening hip pain.

Open in Viewer

Fig. 4C —66-year-old man who presented with left groin pain.

C, Anteroposterior radiograph of pelvis obtained 3 months after A and B shows severe joint space narrowing of left femoral articular surface with articular surface collapse and flattening (arrow).

Open in Viewer

Fig. 4D —66-year-old man who presented with left groin pain.

D, Coronal STIR (D) and coronal T1-weighted (E) images obtained same day as radiograph (C) show severe subchondral marrow edema spanning left hip and large joint effusion. Articular surface collapse (arrow, D), with areas of subchondral sclerosis (asterisk), is also present. Differential diagnosis includes rapid progression of osteoarthritis, avascular necrosis, and septic joint. Further workup identified no evidence of infection, and patient underwent left total hip arthroplasty.

Open in Viewer

Fig. 4E —66-year-old man who presented with left groin pain.

E, Coronal STIR (D) and coronal T1-weighted (E) images obtained same day as radiograph (C) show severe subchondral marrow edema spanning left hip and large joint effusion. Articular surface collapse (arrow, D), with areas of subchondral sclerosis (asterisk), is also present. Differential diagnosis includes rapid progression of osteoarthritis, avascular necrosis, and septic joint. Further workup identified no evidence of infection, and patient underwent left total hip arthroplasty.

Open in Viewer

Fig. 5A —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

A, Preinjection anteroposterior radiograph of pelvis shows severe left hip osteoarthritis with joint space narrowing, marginal spurring (arrow), and superolateral subluxation of femoral head. Intraarticular injection of 40 mg of triamcinolone acetonide (Kenalog, Proficient Rx) was performed after radiography.

Open in Viewer

Fig. 5B —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

B, Coronal STIR (B) and coronal T1-weighted (C) images of osseous pelvis obtained 2 weeks after injection show rapid destruction of femoral articular surface (asterisk) and acetabulum (arrow), with large joint effusion.

Open in Viewer

Fig. 5C —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

C, Coronal STIR (B) and coronal T1-weighted (C) images of osseous pelvis obtained 2 weeks after injection show rapid destruction of femoral articular surface (asterisk) and acetabulum (arrow), with large joint effusion.

Open in Viewer

Fig. 5D —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

D, Anteroposterior radiograph of pelvis obtained 3 weeks after injection shows progressing marked osseous resorption of femoral head, which now extends to intertrochanteric region (arrow), with cephalad migration and lateral displacement of proximal femur. No evidence of infection was found on joint aspiration or intraoperatively during total hip arthroplasty.

An expert panel review concluded that estimates of the preval‎ence of structural problems induced by intraarticular injections are highly variable and that, at present, no imaging, clinical, or laboratory markers are established for predicting subsequent osseous injury after an intraarticular injection [47]. The presence of acute joint pain that is not explained by radiographic or sonographic findings could prompt caution, with further workup obtained before potential corticosteroid injection. In this situation, MRI may help to detect osseous pathology, such as subchondral fracture, that would be a contraindication to injection; universal implementation of this approach may be cost-prohibitive given the likely low frequency of osseous injury on MRI [47]. The use of repeat radiographs before joint injection is also controversial. The panel concluded that large clinical trials are needed to determine the true incidence of these osseous complications and to formulate preintervention safety measures [47]. Although these osseous complications have a likely very low incidence, patients should be aware of the risk, and clinical judgment should be used to assess such complications' relevance for individual patients. For example, such complications may be considered more relevant in a patient with prior long-standing steroid use or a prior diagnosis of osteopenia or osteoporosis.

Systemic Immediate Side Effects

Adrenal suppression—Locally injected corticosteroid does not remain confined to the joint, and some amount is absorbed by the bloodstream. One study found that serum cortisol levels after intraarticular injection of methylprednisolone acetate were dose-dependent, peaked 2–12 hours after injection, and showed a significant downregulation for at least 1 week regardless of the injected dose [48]. A study of 10 healthy young male athletes reported adrenal suppression for up to 2 weeks after a single intra- or extraarticular injection [49]. A study found that injecting two joints simultaneously yields higher serum absorption than injecting the same total dose into one joint; this finding is theorized to be due to the exposure of a greater volume of synovium to corticosteroid, thus increasing the likelihood of systemic absorption [48]. This systemic absorption of corticosteroid accounts for many of the side effects occurring after injection, as well as the such side effects' similarity to the known side effects of exogenous steroid medication.

One study reported facial flushing (defined as erythema of the upper chest and face) after intraarticular corticosteroid injection in 40% of patients, with this side effect being unpleasant in 15% in cases, predominantly occurring in women, and being reproducible with subsequent injection [50]. Flushing is self-limiting, typically resolving within 3 days [51]. Associated systemic symptoms can include insomnia and headache.

Transient hypertension and postprandial hyperglycemia occur after intraarticular corticosteroid injection in patients both with and without diabetes and relate to suppression of the corticotropic axis; these effects may last for over 1 week [52]. These side effects' clinical implications depend on the patient's baseline condition.

Similar to the known side effects of oral or IV steroid use, psychiatric effects (e.g., altered mental status, visual hallucination) have been reported after intraarticular injection, typically occurring within 1 week after injection and occurring more commonly in older patients with decreased renal function [53, 54]. Prolonged hiccups have been reported after corticosteroid injection and are of indeterminate etiology [55].

Hypersensitivity reactions—The reported preval‎ence of hypersensitivity reactions after corticosteroid administration by any route is 0.3–0.5%; a common presentation of such reactions after musculoskeletal injection is delayed urticaria at the injection site [56, 57]. A case report described immediate anaphylaxis after intraarticular injection of triamcinolone acetonide, with confirmed hypersensitivity reaction by skin prick test [57]. Patients with frequent exposure to steroids are more likely to experience a hypersensitivity reaction. In addition, patients reporting a prior hypersensitivity reaction to corticosteroids should undergo workup for potential allergy to any alternative steroid preparation that is selected before it is administered [56]. Management of hypersensitivity reactions may include administration of oral or IV diphenhydramine hydrochloride, depending on the reaction's severity [56, 57].

Ophthalmic side effects—Central serous chorioretinopathy is a retinal detachment presenting with potentially permanent decreased vision that has been attributed to exogenously administered corticosteroids [58]. Any form of corticosteroid administration can precipitate glaucoma or increased intraocular pressure, the latter of which may lead to sudden-onset blurred vision or potentially irreversible vision loss [59]. These potential complications must be discussed before corticosteroid injection with patients who have known ophthalmic pathology and with patients reporting prior visual changes after steroid exposure.

Systemic Delayed Side Effects

As mentioned previously, intraarticular corticosteroid injection suppresses the hypothalamic-pituitary-adrenal axis, causing an at least temporary decrease in endogenous cortisol levels in 52.2% of patients [60]. This suppression can be potentially life-threatening in patients who are subject to an emotional or physical stress, such as illness or surgery, that may precipitate an adrenal crisis [6]. Most patients with adrenal insufficiency are not diagnosed until they are hospitalized for adrenal crisis, which has a reported mortality of 5% [61]. Prolonged adrenal insufficiency has been reported in some patients after repeated corticosteroid injections (i.e., iatrogenic Cushing syndrome) and may require exogenous supplementation to avoid an adrenal crisis [62, 63]. Although limited data describe prolonged adrenal suppression after musculoskeletal injection, a meta-analysis found that adrenal insufficiency may be present in 25.3% of patients 6 months after any corticosteroid administration [60]. Due to individual variation in glucocorticoid sensitivity, there is no corticosteroid dose, administration route, or treatment duration that prevents adrenal suppression [60, 64]. Furthermore, some suggest that no corticosteroids be administered for at least 4 weeks before planned surgery, to decrease the chance of adrenal crisis [28].

Osteoporosis is a known complication of long-term exogenous steroid use, attributed to steroid's negative effect on both bone quality and bone mass. A statistically significant decrease in bone mineral density of the femur has been reported after a single epidural steroid injection in postmenopausal women [65]. The literature suggests that the effect of epidural corticosteroid injection on bone mineral density is dose-dependent, whether from a single injection or multiple injections over time (i.e., cumulative dose), and may also predispose to increased vertebral fracture risk [28]. The incidence and impact of osteoporotic fractures from intra- and extraarticular corticosteroid injections for musculoskeletal indications are unknown. However, given a recognized impact of epidural steroid injections on bone mineral density, the radiologist should provide a recommendation to the primary care practitioner to consider initiating prophylactic osteoporotic treatment in postmenopausal women receiving frequent or multiple injections [28, 65]. Local Immediate Side Effects

Postinjection flare is an exacerbation of pain after injection that is significantly higher than the baseline preinjection pain. This flare is theorized to result from focal synovitis secondary to intracellular incorporation of the injected corticosteroid particulate microcrystals [9]. Postinjection flare is more common and of longer duration for particulate than nonparticulate corticosteroids [10]. The patient experience of postinjection flare can be clinically similar to the experience of crystalline-induced arthropathies, such as gout. Studies including injections of the hand, wrist, and shoulder report an incidence of postinjection flare of 32.8–35.3% [11, 12]. A study assessing flare reactions after injections performed to treat de Quer-vain tenosynovitis and trigger finger found that most flares occurred 1–2 days after injection, lasted a mean of 2 days, and did not impact long-term symptom resolution [11].

Local Delayed Side Effects

Skin hypopigmentation and atrophy—Several prospective studies describe skin hypopigmentation and atrophy, particularly after superficial peritendinous injections (e.g., to treat epicondylitis or de Quervain tenosynovitis) [13, 14] (Fig. 1). Heel pad atrophy or skin depigmentation is reported in 10% of patients who received corticosteroid injection of the plantar fascia [15]. Skin atrophy and hypopigmentation manifest within a few months of injection and typically resolve within 1 year [16]. Such dermal changes likely relate specifically to the use of particulate steroids, especially if the injectate migrates along the injection pathway within the soft tissues [17]. When superficial structures are being injected, the use of nonparticulate steroids should be considered, to minimize the potential for skin atrophy. Additionally, compression of the injection site after removal of the needle may minimize the risk of steroid migration along the needle track [18].

Open in Viewer

Fig. 1 —63-year-old woman with medial elbow pain who presented for sonographic eval‎uation of common flexor tendon. Patient had received two prior injections of 40 mg of triamcinolone acetonide (Kenalog, Proficient Rx) in medial elbow without imaging guidance within past 6 months. Photograph obtained at time of presentation for imaging shows discoloration and thinning of skin at medial elbow (dashed outline). Patient noticed this change several weeks after second injection.

Soft-tissue calcification—Soft-tissue calcification is a reported iatrogenic complication of musculoskeletal corticosteroid injection [19]. Explanatory theories include local tissue damage resulting in increased pH that precipitates calcium salt or possible calcific necrosis of collagen fibers [19]. Soft-tissue and tendinous calcifications are associated with increased pain scores and may be clinically significant [19]. Radiologists should recognize this potential corticosteroid-induced complication when assessing for changes in periarticular calcifications on postinjection imaging.

Infection—Intraarticular infection after corticosteroid injection is very rare, with an incidence thought to be less than 0.001% [20]. The incidence may be significantly less when using single-dose vials, as well as when using modern sterile technique, packaging, and cleaning methods [21]. A systematic review of case reports describes the incidences of soft-tissue infection and osteomyelitis of the underlying bone adjacent to the injection site [22]. The injection may also increase the risk of infection if the patient later undergoes surgery at the injection site. For example, studies have found that patients who received intraarticular injections of corticosteroid before a knee or hip replacement were more likely to experience a postoperative prosthetic joint infection than those who had never received an injection [23, 24]. A meta-analysis of eight studies on hip and knee joint replacements showed a 2.13 higher odds of deep infection in patients with than without a prior injection but no significant difference in superficial infection rate [25]. Another systematic review found that intraarticular injections performed within 3 months before knee arthroplasty were significantly associated with subsequent prosthetic joint infection [26]. Moreover, a study found that the use of preoperative corticosteroids was associated with a higher likelihood of superficial or deep infections in any surgical site, as well as an increased risk of respiratory, renal, and cardiac complications [27]. Given the available data, it may be advisable to delay any surgery for 1–3 months after corticosteroid injection to potentially mitigate the infection risk, particularly in patients with multiple comorbidities and chronic conditions and who are dependent on corticosteroids [28]. It is important to consider this potential infection risk before injecting patients who may be considering arthroplasty or other surgeries soon after corticosteroid injection and to counsel such patients accordingly.

Tendon, ligament, or fascial insufficiency or rupture—A rare but significant potential consequence of corticosteroid injections is tendon rupture (Figs. 2 and 3). Although its incidence is low, reported as approximately 0.1% in a meta-analysis, case reports describe such complications throughout the body [29, 30]. A review of the case reports estimated that among 95 postinjection complications, 53.7% were plantar fascia ruptures, and 26.3% were tendon ruptures involving the patellar, quadriceps, Achilles, or biceps tendon—tendons possibly already prone to increased biomechanical stress [31]. Injected corticosteroids are thought to locally decrease collagen production, which may limit potential tendon, ligament, or fascial healing and therefore increase predisposition to structural failure or rupture [32]. Even if rupture does not occur, corticosteroid use has been associated with ligament instability in the elbow, leading to refractory pain and disability [33]. A histologic study of the Achilles tendon in rabbits reported equivalent tendon weakening and loss of elasticity after either intratendinous or retrocalcaneal bursa injection, indicating that even peritendinous injections pose a risk of tendon rupture [34]. Tendon rupture may also be more strongly associated with triamcinolone acetonide than with other steroid preparations and may be associated with frequent repeated injections to a particular site [31]. Tendon rupture has been reported to occur anywhere from 2 days to 6 weeks after injection [29]. Although evidence-based consensus guidelines are lacking, some authors suggest that a maximum of three to five injections be performed at a single site, with a minimum of 2 weeks between injections, to minimize potential tendon rupture [31].

Open in Viewer

Fig. 2A —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

A, Gray-scale sonographic image obtained using 15-MHz linear probe shows intact gluteus medius insertion (dashed lines) on greater trochanter (asterisk) in long axis. Image shows thickened tendon with moderate tendinopathy but no high-grade tear. Patient underwent ultrasound-guided percutaneous tenotomy followed by injection of 20 mg of methylprednisolone acetate (Depo-Medrol, Pfizer) and 4 mL of 0.25% bupivacaine into greater trochanteric bursa. Twelve days after procedure, patient returned due to worsening pain, which was worse with weight-bearing and did not improve after course of NSAIDS and oral steroids. Patient reported no interval trauma. MRI of osseous pelvis was performed 18 days after procedure.

Open in Viewer

Fig. 2B —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

B, Coronal STIR (B) and sagittal T2-weighted (C) images, obtained 18 days after A, through greater trochanter show full-thickness tear of left gluteus medius tendon at myotendinous junction (arrow, B) with 2-cm tendon gap (arrow, C).

Open in Viewer

Fig. 2C —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

C, Coronal STIR (B) and sagittal T2-weighted (C) images, obtained 18 days after A, through greater trochanter show full-thickness tear of left gluteus medius tendon at myotendinous junction (arrow, B) with 2-cm tendon gap (arrow, C).

Open in Viewer

Fig. 3A —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

A, Gray-scale sonographic image of right hamstring tendon origin in long axis, obtained at initial presentation, shows marked thickening and hypoechoic degeneration of tendon (dashed lines) with areas of anechoic interstitial tearing (arrow). Ischial tuberosity is denoted by asterisk. Two days after third steroid injection, patient felt snap in buttock when performing household tasks.

Open in Viewer

Fig. 3B —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

B, Coronal STIR image of bony pelvis through ischial tuberosity shows near-complete rupture of right hamstring tendon origin with almost 3 cm of distal retraction (arrow).

Open in Viewer

Fig. 3C —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

C, Gray-scale sonographic image of left hamstring origin in long axis, immediately caudal to ischial tuberosity (asterisk), shows anechoic tendon gap (arrow).

Corticosteroid injections in the shoulder have been associated with a need for revision after rotator cuff repair, possibly due to underlying degeneration of the tendon quality as a result of the injection [35]. A systematic review found a nearly twofold increase in adverse outcomes after rotator cuff repair if a corticosteroid injection is administered within 6 months before surgery or if more than two injections are performed within 1 year before surgery [36]. This effect was also shown in an analysis of patients with lateral epicondylitis who underwent surgery; those receiving three or more injections were at highest risk (OR, 3.55) for requiring revision surgery [37]. Patients must be appropriately counseled regarding the potential greater need for revision surgery if a corticosteroid is administered within 6 months before surgical tendon repair.

Accelerated progression of osteoarthritis—Accelerated progression of osteoarthritis is not strictly defined but refers to an increase in joint space loss that is greater than expected to occur in a short time period. One study defined accelerated progression of osteoarthritis as a decrease in joint space by 2 mm over a 12-month period [38]. Research from the Osteoarthritis Initiative found that patients who received intraarticular knee joint injections of corticosteroid had 3.2 higher odds of showing radiographic progression of osteoarthritis than those who did not receive injections; the odds were higher at 4.67 in patients who received continuous injections [39]. Radiographic progression of osteoarthritis after injection has also been shown in the hip, with a study reporting radiographic progression of hip joint space narrowing in 44% of patients who received hip injections and were followed postprocedurally for a mean of 6 months [40].

Osseous injury—Corticosteroids induce osteocyte apoptosis, which predisposes to decreases in osseous vascularity and bone strength; these effects are particularly pronounced after intraarticular injection [41]. Indeed, subchondral insufficiency fracture, osteonecrosis, and rapid joint destruction have all been reported after intraarticular corticosteroid injection into various joints [42–44] (Figs. 4 and 5). A retrospective single-center study of 1000 patients with long-term follow-up after hip and knee corticosteroid injections showed a combined rate of osteonecrosis, insufficiency fractures, and rapid progressive osteoarthritis of 1% [45]. Such reports are controversial, and some experts believe that the reported risk of osseous injury after intraarticular injection is inflated (possibly attributed to the natural progression of osteoarthritis) and that a risk-benefit analysis would support this approach [46].

Open in Viewer

Fig. 4A —66-year-old man who presented with left groin pain.

A, Coronal STIR image through osseous pelvis shows large left hip joint effusion (asterisk).

Open in Viewer

Fig. 4B —66-year-old man who presented with left groin pain.

B, Sagittal T2-weighted image through left hip joint shows areas of full-thickness cartilage loss and subchondral marrow edema–like signal (arrows). Patient underwent same-day injection of left hip and iliopsoas bursa, receiving total of 60 mg of triamcinolone acetonide (Kenalog, Proficient Rx). Patient presented 3 months after injection with worsening hip pain.

Open in Viewer

Fig. 4C —66-year-old man who presented with left groin pain.

C, Anteroposterior radiograph of pelvis obtained 3 months after A and B shows severe joint space narrowing of left femoral articular surface with articular surface collapse and flattening (arrow).

Open in Viewer

Fig. 4D —66-year-old man who presented with left groin pain.

D, Coronal STIR (D) and coronal T1-weighted (E) images obtained same day as radiograph (C) show severe subchondral marrow edema spanning left hip and large joint effusion. Articular surface collapse (arrow, D), with areas of subchondral sclerosis (asterisk), is also present. Differential diagnosis includes rapid progression of osteoarthritis, avascular necrosis, and septic joint. Further workup identified no evidence of infection, and patient underwent left total hip arthroplasty.

Open in Viewer

Fig. 4E —66-year-old man who presented with left groin pain.

E, Coronal STIR (D) and coronal T1-weighted (E) images obtained same day as radiograph (C) show severe subchondral marrow edema spanning left hip and large joint effusion. Articular surface collapse (arrow, D), with areas of subchondral sclerosis (asterisk), is also present. Differential diagnosis includes rapid progression of osteoarthritis, avascular necrosis, and septic joint. Further workup identified no evidence of infection, and patient underwent left total hip arthroplasty.

Open in Viewer

Fig. 5A —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

A, Preinjection anteroposterior radiograph of pelvis shows severe left hip osteoarthritis with joint space narrowing, marginal spurring (arrow), and superolateral subluxation of femoral head. Intraarticular injection of 40 mg of triamcinolone acetonide (Kenalog, Proficient Rx) was performed after radiography.

Open in Viewer

Fig. 5B —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

B, Coronal STIR (B) and coronal T1-weighted (C) images of osseous pelvis obtained 2 weeks after injection show rapid destruction of femoral articular surface (asterisk) and acetabulum (arrow), with large joint effusion.

Open in Viewer

Fig. 5C —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

C, Coronal STIR (B) and coronal T1-weighted (C) images of osseous pelvis obtained 2 weeks after injection show rapid destruction of femoral articular surface (asterisk) and acetabulum (arrow), with large joint effusion.

Open in Viewer

Fig. 5D —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

D, Anteroposterior radiograph of pelvis obtained 3 weeks after injection shows progressing marked osseous resorption of femoral head, which now extends to intertrochanteric region (arrow), with cephalad migration and lateral displacement of proximal femur. No evidence of infection was found on joint aspiration or intraoperatively during total hip arthroplasty.

An expert panel review concluded that estimates of the preval‎ence of structural problems induced by intraarticular injections are highly variable and that, at present, no imaging, clinical, or laboratory markers are established for predicting subsequent osseous injury after an intraarticular injection [47]. The presence of acute joint pain that is not explained by radiographic or sonographic findings could prompt caution, with further workup obtained before potential corticosteroid injection. In this situation, MRI may help to detect osseous pathology, such as subchondral fracture, that would be a contraindication to injection; universal implementation of this approach may be cost-prohibitive given the likely low frequency of osseous injury on MRI [47]. The use of repeat radiographs before joint injection is also controversial. The panel concluded that large clinical trials are needed to determine the true incidence of these osseous complications and to formulate preintervention safety measures [47]. Although these osseous complications have a likely very low incidence, patients should be aware of the risk, and clinical judgment should be used to assess such complications' relevance for individual patients. For example, such complications may be considered more relevant in a patient with prior long-standing steroid use or a prior diagnosis of osteopenia or osteoporosis.

Systemic Immediate Side Effects

Adrenal suppression—Locally injected corticosteroid does not remain confined to the joint, and some amount is absorbed by the bloodstream. One study found that serum cortisol levels after intraarticular injection of methylprednisolone acetate were dose-dependent, peaked 2–12 hours after injection, and showed a significant downregulation for at least 1 week regardless of the injected dose [48]. A study of 10 healthy young male athletes reported adrenal suppression for up to 2 weeks after a single intra- or extraarticular injection [49]. A study found that injecting two joints simultaneously yields higher serum absorption than injecting the same total dose into one joint; this finding is theorized to be due to the exposure of a greater volume of synovium to corticosteroid, thus increasing the likelihood of systemic absorption [48]. This systemic absorption of corticosteroid accounts for many of the side effects occurring after injection, as well as the such side effects' similarity to the known side effects of exogenous steroid medication.

One study reported facial flushing (defined as erythema of the upper chest and face) after intraarticular corticosteroid injection in 40% of patients, with this side effect being unpleasant in 15% in cases, predominantly occurring in women, and being reproducible with subsequent injection [50]. Flushing is self-limiting, typically resolving within 3 days [51]. Associated systemic symptoms can include insomnia and headache.

Transient hypertension and postprandial hyperglycemia occur after intraarticular corticosteroid injection in patients both with and without diabetes and relate to suppression of the corticotropic axis; these effects may last for over 1 week [52]. These side effects' clinical implications depend on the patient's baseline condition.

Similar to the known side effects of oral or IV steroid use, psychiatric effects (e.g., altered mental status, visual hallucination) have been reported after intraarticular injection, typically occurring within 1 week after injection and occurring more commonly in older patients with decreased renal function [53, 54]. Prolonged hiccups have been reported after corticosteroid injection and are of indeterminate etiology [55].

Hypersensitivity reactions—The reported preval‎ence of hypersensitivity reactions after corticosteroid administration by any route is 0.3–0.5%; a common presentation of such reactions after musculoskeletal injection is delayed urticaria at the injection site [56, 57]. A case report described immediate anaphylaxis after intraarticular injection of triamcinolone acetonide, with confirmed hypersensitivity reaction by skin prick test [57]. Patients with frequent exposure to steroids are more likely to experience a hypersensitivity reaction. In addition, patients reporting a prior hypersensitivity reaction to corticosteroids should undergo workup for potential allergy to any alternative steroid preparation that is selected before it is administered [56]. Management of hypersensitivity reactions may include administration of oral or IV diphenhydramine hydrochloride, depending on the reaction's severity [56, 57].

Ophthalmic side effects—Central serous chorioretinopathy is a retinal detachment presenting with potentially permanent decreased vision that has been attributed to exogenously administered corticosteroids [58]. Any form of corticosteroid administration can precipitate glaucoma or increased intraocular pressure, the latter of which may lead to sudden-onset blurred vision or potentially irreversible vision loss [59]. These potential complications must be discussed before corticosteroid injection with patients who have known ophthalmic pathology and with patients reporting prior visual changes after steroid exposure.

Systemic Delayed Side Effects

As mentioned previously, intraarticular corticosteroid injection suppresses the hypothalamic-pituitary-adrenal axis, causing an at least temporary decrease in endogenous cortisol levels in 52.2% of patients [60]. This suppression can be potentially life-threatening in patients who are subject to an emotional or physical stress, such as illness or surgery, that may precipitate an adrenal crisis [6]. Most patients with adrenal insufficiency are not diagnosed until they are hospitalized for adrenal crisis, which has a reported mortality of 5% [61]. Prolonged adrenal insufficiency has been reported in some patients after repeated corticosteroid injections (i.e., iatrogenic Cushing syndrome) and may require exogenous supplementation to avoid an adrenal crisis [62, 63]. Although limited data describe prolonged adrenal suppression after musculoskeletal injection, a meta-analysis found that adrenal insufficiency may be present in 25.3% of patients 6 months after any corticosteroid administration [60]. Due to individual variation in glucocorticoid sensitivity, there is no corticosteroid dose, administration route, or treatment duration that prevents adrenal suppression [60, 64]. Furthermore, some suggest that no corticosteroids be administered for at least 4 weeks before planned surgery, to decrease the chance of adrenal crisis [28].

Osteoporosis is a known complication of long-term exogenous steroid use, attributed to steroid's negative effect on both bone quality and bone mass. A statistically significant decrease in bone mineral density of the femur has been reported after a single epidural steroid injection in postmenopausal women [65]. The literature suggests that the effect of epidural corticosteroid injection on bone mineral density is dose-dependent, whether from a single injection or multiple injections over time (i.e., cumulative dose), and may also predispose to increased vertebral fracture risk [28]. The incidence and impact of osteoporotic fractures from intra- and extraarticular corticosteroid injections for musculoskeletal indications are unknown. However, given a recognized impact of epidural steroid injections on bone mineral density, the radiologist should provide a recommendation to the primary care practitioner to consider initiating prophylactic osteoporotic treatment in postmenopausal women receiving frequent or multiple injections [28, 65]. Local Immediate Side Effects

Postinjection flare is an exacerbation of pain after injection that is significantly higher than the baseline preinjection pain. This flare is theorized to result from focal synovitis secondary to intracellular incorporation of the injected corticosteroid particulate microcrystals [9]. Postinjection flare is more common and of longer duration for particulate than nonparticulate corticosteroids [10]. The patient experience of postinjection flare can be clinically similar to the experience of crystalline-induced arthropathies, such as gout. Studies including injections of the hand, wrist, and shoulder report an incidence of postinjection flare of 32.8–35.3% [11, 12]. A study assessing flare reactions after injections performed to treat de Quer-vain tenosynovitis and trigger finger found that most flares occurred 1–2 days after injection, lasted a mean of 2 days, and did not impact long-term symptom resolution [11].

Local Delayed Side Effects

Skin hypopigmentation and atrophy—Several prospective studies describe skin hypopigmentation and atrophy, particularly after superficial peritendinous injections (e.g., to treat epicondylitis or de Quervain tenosynovitis) [13, 14] (Fig. 1). Heel pad atrophy or skin depigmentation is reported in 10% of patients who received corticosteroid injection of the plantar fascia [15]. Skin atrophy and hypopigmentation manifest within a few months of injection and typically resolve within 1 year [16]. Such dermal changes likely relate specifically to the use of particulate steroids, especially if the injectate migrates along the injection pathway within the soft tissues [17]. When superficial structures are being injected, the use of nonparticulate steroids should be considered, to minimize the potential for skin atrophy. Additionally, compression of the injection site after removal of the needle may minimize the risk of steroid migration along the needle track [18].

Open in Viewer

Fig. 1 —63-year-old woman with medial elbow pain who presented for sonographic eval‎uation of common flexor tendon. Patient had received two prior injections of 40 mg of triamcinolone acetonide (Kenalog, Proficient Rx) in medial elbow without imaging guidance within past 6 months. Photograph obtained at time of presentation for imaging shows discoloration and thinning of skin at medial elbow (dashed outline). Patient noticed this change several weeks after second injection.

Soft-tissue calcification—Soft-tissue calcification is a reported iatrogenic complication of musculoskeletal corticosteroid injection [19]. Explanatory theories include local tissue damage resulting in increased pH that precipitates calcium salt or possible calcific necrosis of collagen fibers [19]. Soft-tissue and tendinous calcifications are associated with increased pain scores and may be clinically significant [19]. Radiologists should recognize this potential corticosteroid-induced complication when assessing for changes in periarticular calcifications on postinjection imaging.

Infection—Intraarticular infection after corticosteroid injection is very rare, with an incidence thought to be less than 0.001% [20]. The incidence may be significantly less when using single-dose vials, as well as when using modern sterile technique, packaging, and cleaning methods [21]. A systematic review of case reports describes the incidences of soft-tissue infection and osteomyelitis of the underlying bone adjacent to the injection site [22]. The injection may also increase the risk of infection if the patient later undergoes surgery at the injection site. For example, studies have found that patients who received intraarticular injections of corticosteroid before a knee or hip replacement were more likely to experience a postoperative prosthetic joint infection than those who had never received an injection [23, 24]. A meta-analysis of eight studies on hip and knee joint replacements showed a 2.13 higher odds of deep infection in patients with than without a prior injection but no significant difference in superficial infection rate [25]. Another systematic review found that intraarticular injections performed within 3 months before knee arthroplasty were significantly associated with subsequent prosthetic joint infection [26]. Moreover, a study found that the use of preoperative corticosteroids was associated with a higher likelihood of superficial or deep infections in any surgical site, as well as an increased risk of respiratory, renal, and cardiac complications [27]. Given the available data, it may be advisable to delay any surgery for 1–3 months after corticosteroid injection to potentially mitigate the infection risk, particularly in patients with multiple comorbidities and chronic conditions and who are dependent on corticosteroids [28]. It is important to consider this potential infection risk before injecting patients who may be considering arthroplasty or other surgeries soon after corticosteroid injection and to counsel such patients accordingly.

Tendon, ligament, or fascial insufficiency or rupture—A rare but significant potential consequence of corticosteroid injections is tendon rupture (Figs. 2 and 3). Although its incidence is low, reported as approximately 0.1% in a meta-analysis, case reports describe such complications throughout the body [29, 30]. A review of the case reports estimated that among 95 postinjection complications, 53.7% were plantar fascia ruptures, and 26.3% were tendon ruptures involving the patellar, quadriceps, Achilles, or biceps tendon—tendons possibly already prone to increased biomechanical stress [31]. Injected corticosteroids are thought to locally decrease collagen production, which may limit potential tendon, ligament, or fascial healing and therefore increase predisposition to structural failure or rupture [32]. Even if rupture does not occur, corticosteroid use has been associated with ligament instability in the elbow, leading to refractory pain and disability [33]. A histologic study of the Achilles tendon in rabbits reported equivalent tendon weakening and loss of elasticity after either intratendinous or retrocalcaneal bursa injection, indicating that even peritendinous injections pose a risk of tendon rupture [34]. Tendon rupture may also be more strongly associated with triamcinolone acetonide than with other steroid preparations and may be associated with frequent repeated injections to a particular site [31]. Tendon rupture has been reported to occur anywhere from 2 days to 6 weeks after injection [29]. Although evidence-based consensus guidelines are lacking, some authors suggest that a maximum of three to five injections be performed at a single site, with a minimum of 2 weeks between injections, to minimize potential tendon rupture [31].

Open in Viewer

Fig. 2A —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

A, Gray-scale sonographic image obtained using 15-MHz linear probe shows intact gluteus medius insertion (dashed lines) on greater trochanter (asterisk) in long axis. Image shows thickened tendon with moderate tendinopathy but no high-grade tear. Patient underwent ultrasound-guided percutaneous tenotomy followed by injection of 20 mg of methylprednisolone acetate (Depo-Medrol, Pfizer) and 4 mL of 0.25% bupivacaine into greater trochanteric bursa. Twelve days after procedure, patient returned due to worsening pain, which was worse with weight-bearing and did not improve after course of NSAIDS and oral steroids. Patient reported no interval trauma. MRI of osseous pelvis was performed 18 days after procedure.

Open in Viewer

Fig. 2B —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

B, Coronal STIR (B) and sagittal T2-weighted (C) images, obtained 18 days after A, through greater trochanter show full-thickness tear of left gluteus medius tendon at myotendinous junction (arrow, B) with 2-cm tendon gap (arrow, C).

Open in Viewer

Fig. 2C —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

C, Coronal STIR (B) and sagittal T2-weighted (C) images, obtained 18 days after A, through greater trochanter show full-thickness tear of left gluteus medius tendon at myotendinous junction (arrow, B) with 2-cm tendon gap (arrow, C).

Open in Viewer

Fig. 3A —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

A, Gray-scale sonographic image of right hamstring tendon origin in long axis, obtained at initial presentation, shows marked thickening and hypoechoic degeneration of tendon (dashed lines) with areas of anechoic interstitial tearing (arrow). Ischial tuberosity is denoted by asterisk. Two days after third steroid injection, patient felt snap in buttock when performing household tasks.

Open in Viewer

Fig. 3B —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

B, Coronal STIR image of bony pelvis through ischial tuberosity shows near-complete rupture of right hamstring tendon origin with almost 3 cm of distal retraction (arrow).

Open in Viewer

Fig. 3C —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

C, Gray-scale sonographic image of left hamstring origin in long axis, immediately caudal to ischial tuberosity (asterisk), shows anechoic tendon gap (arrow).

Corticosteroid injections in the shoulder have been associated with a need for revision after rotator cuff repair, possibly due to underlying degeneration of the tendon quality as a result of the injection [35]. A systematic review found a nearly twofold increase in adverse outcomes after rotator cuff repair if a corticosteroid injection is administered within 6 months before surgery or if more than two injections are performed within 1 year before surgery [36]. This effect was also shown in an analysis of patients with lateral epicondylitis who underwent surgery; those receiving three or more injections were at highest risk (OR, 3.55) for requiring revision surgery [37]. Patients must be appropriately counseled regarding the potential greater need for revision surgery if a corticosteroid is administered within 6 months before surgical tendon repair.

Accelerated progression of osteoarthritis—Accelerated progression of osteoarthritis is not strictly defined but refers to an increase in joint space loss that is greater than expected to occur in a short time period. One study defined accelerated progression of osteoarthritis as a decrease in joint space by 2 mm over a 12-month period [38]. Research from the Osteoarthritis Initiative found that patients who received intraarticular knee joint injections of corticosteroid had 3.2 higher odds of showing radiographic progression of osteoarthritis than those who did not receive injections; the odds were higher at 4.67 in patients who received continuous injections [39]. Radiographic progression of osteoarthritis after injection has also been shown in the hip, with a study reporting radiographic progression of hip joint space narrowing in 44% of patients who received hip injections and were followed postprocedurally for a mean of 6 months [40].

Osseous injury—Corticosteroids induce osteocyte apoptosis, which predisposes to decreases in osseous vascularity and bone strength; these effects are particularly pronounced after intraarticular injection [41]. Indeed, subchondral insufficiency fracture, osteonecrosis, and rapid joint destruction have all been reported after intraarticular corticosteroid injection into various joints [42–44] (Figs. 4 and 5). A retrospective single-center study of 1000 patients with long-term follow-up after hip and knee corticosteroid injections showed a combined rate of osteonecrosis, insufficiency fractures, and rapid progressive osteoarthritis of 1% [45]. Such reports are controversial, and some experts believe that the reported risk of osseous injury after intraarticular injection is inflated (possibly attributed to the natural progression of osteoarthritis) and that a risk-benefit analysis would support this approach [46].

Open in Viewer

Fig. 4A —66-year-old man who presented with left groin pain.

A, Coronal STIR image through osseous pelvis shows large left hip joint effusion (asterisk).

Open in Viewer

Fig. 4B —66-year-old man who presented with left groin pain.

B, Sagittal T2-weighted image through left hip joint shows areas of full-thickness cartilage loss and subchondral marrow edema–like signal (arrows). Patient underwent same-day injection of left hip and iliopsoas bursa, receiving total of 60 mg of triamcinolone acetonide (Kenalog, Proficient Rx). Patient presented 3 months after injection with worsening hip pain.

Open in Viewer

Fig. 4C —66-year-old man who presented with left groin pain.

C, Anteroposterior radiograph of pelvis obtained 3 months after A and B shows severe joint space narrowing of left femoral articular surface with articular surface collapse and flattening (arrow).

Open in Viewer

Fig. 4D —66-year-old man who presented with left groin pain.

D, Coronal STIR (D) and coronal T1-weighted (E) images obtained same day as radiograph (C) show severe subchondral marrow edema spanning left hip and large joint effusion. Articular surface collapse (arrow, D), with areas of subchondral sclerosis (asterisk), is also present. Differential diagnosis includes rapid progression of osteoarthritis, avascular necrosis, and septic joint. Further workup identified no evidence of infection, and patient underwent left total hip arthroplasty.

Open in Viewer

Fig. 4E —66-year-old man who presented with left groin pain.

E, Coronal STIR (D) and coronal T1-weighted (E) images obtained same day as radiograph (C) show severe subchondral marrow edema spanning left hip and large joint effusion. Articular surface collapse (arrow, D), with areas of subchondral sclerosis (asterisk), is also present. Differential diagnosis includes rapid progression of osteoarthritis, avascular necrosis, and septic joint. Further workup identified no evidence of infection, and patient underwent left total hip arthroplasty.

Open in Viewer

Fig. 5A —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

A, Preinjection anteroposterior radiograph of pelvis shows severe left hip osteoarthritis with joint space narrowing, marginal spurring (arrow), and superolateral subluxation of femoral head. Intraarticular injection of 40 mg of triamcinolone acetonide (Kenalog, Proficient Rx) was performed after radiography.

Open in Viewer

Fig. 5B —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

B, Coronal STIR (B) and coronal T1-weighted (C) images of osseous pelvis obtained 2 weeks after injection show rapid destruction of femoral articular surface (asterisk) and acetabulum (arrow), with large joint effusion.

Open in Viewer

Fig. 5C —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

C, Coronal STIR (B) and coronal T1-weighted (C) images of osseous pelvis obtained 2 weeks after injection show rapid destruction of femoral articular surface (asterisk) and acetabulum (arrow), with large joint effusion.

Open in Viewer

Fig. 5D —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

D, Anteroposterior radiograph of pelvis obtained 3 weeks after injection shows progressing marked osseous resorption of femoral head, which now extends to intertrochanteric region (arrow), with cephalad migration and lateral displacement of proximal femur. No evidence of infection was found on joint aspiration or intraoperatively during total hip arthroplasty.

An expert panel review concluded that estimates of the preval‎ence of structural problems induced by intraarticular injections are highly variable and that, at present, no imaging, clinical, or laboratory markers are established for predicting subsequent osseous injury after an intraarticular injection [47]. The presence of acute joint pain that is not explained by radiographic or sonographic findings could prompt caution, with further workup obtained before potential corticosteroid injection. In this situation, MRI may help to detect osseous pathology, such as subchondral fracture, that would be a contraindication to injection; universal implementation of this approach may be cost-prohibitive given the likely low frequency of osseous injury on MRI [47]. The use of repeat radiographs before joint injection is also controversial. The panel concluded that large clinical trials are needed to determine the true incidence of these osseous complications and to formulate preintervention safety measures [47]. Although these osseous complications have a likely very low incidence, patients should be aware of the risk, and clinical judgment should be used to assess such complications' relevance for individual patients. For example, such complications may be considered more relevant in a patient with prior long-standing steroid use or a prior diagnosis of osteopenia or osteoporosis.

Systemic Immediate Side Effects

Adrenal suppression—Locally injected corticosteroid does not remain confined to the joint, and some amount is absorbed by the bloodstream. One study found that serum cortisol levels after intraarticular injection of methylprednisolone acetate were dose-dependent, peaked 2–12 hours after injection, and showed a significant downregulation for at least 1 week regardless of the injected dose [48]. A study of 10 healthy young male athletes reported adrenal suppression for up to 2 weeks after a single intra- or extraarticular injection [49]. A study found that injecting two joints simultaneously yields higher serum absorption than injecting the same total dose into one joint; this finding is theorized to be due to the exposure of a greater volume of synovium to corticosteroid, thus increasing the likelihood of systemic absorption [48]. This systemic absorption of corticosteroid accounts for many of the side effects occurring after injection, as well as the such side effects' similarity to the known side effects of exogenous steroid medication.

One study reported facial flushing (defined as erythema of the upper chest and face) after intraarticular corticosteroid injection in 40% of patients, with this side effect being unpleasant in 15% in cases, predominantly occurring in women, and being reproducible with subsequent injection [50]. Flushing is self-limiting, typically resolving within 3 days [51]. Associated systemic symptoms can include insomnia and headache.

Transient hypertension and postprandial hyperglycemia occur after intraarticular corticosteroid injection in patients both with and without diabetes and relate to suppression of the corticotropic axis; these effects may last for over 1 week [52]. These side effects' clinical implications depend on the patient's baseline condition.

Similar to the known side effects of oral or IV steroid use, psychiatric effects (e.g., altered mental status, visual hallucination) have been reported after intraarticular injection, typically occurring within 1 week after injection and occurring more commonly in older patients with decreased renal function [53, 54]. Prolonged hiccups have been reported after corticosteroid injection and are of indeterminate etiology [55].

Hypersensitivity reactions—The reported preval‎ence of hypersensitivity reactions after corticosteroid administration by any route is 0.3–0.5%; a common presentation of such reactions after musculoskeletal injection is delayed urticaria at the injection site [56, 57]. A case report described immediate anaphylaxis after intraarticular injection of triamcinolone acetonide, with confirmed hypersensitivity reaction by skin prick test [57]. Patients with frequent exposure to steroids are more likely to experience a hypersensitivity reaction. In addition, patients reporting a prior hypersensitivity reaction to corticosteroids should undergo workup for potential allergy to any alternative steroid preparation that is selected before it is administered [56]. Management of hypersensitivity reactions may include administration of oral or IV diphenhydramine hydrochloride, depending on the reaction's severity [56, 57].

Ophthalmic side effects—Central serous chorioretinopathy is a retinal detachment presenting with potentially permanent decreased vision that has been attributed to exogenously administered corticosteroids [58]. Any form of corticosteroid administration can precipitate glaucoma or increased intraocular pressure, the latter of which may lead to sudden-onset blurred vision or potentially irreversible vision loss [59]. These potential complications must be discussed before corticosteroid injection with patients who have known ophthalmic pathology and with patients reporting prior visual changes after steroid exposure.

Systemic Delayed Side Effects

As mentioned previously, intraarticular corticosteroid injection suppresses the hypothalamic-pituitary-adrenal axis, causing an at least temporary decrease in endogenous cortisol levels in 52.2% of patients [60]. This suppression can be potentially life-threatening in patients who are subject to an emotional or physical stress, such as illness or surgery, that may precipitate an adrenal crisis [6]. Most patients with adrenal insufficiency are not diagnosed until they are hospitalized for adrenal crisis, which has a reported mortality of 5% [61]. Prolonged adrenal insufficiency has been reported in some patients after repeated corticosteroid injections (i.e., iatrogenic Cushing syndrome) and may require exogenous supplementation to avoid an adrenal crisis [62, 63]. Although limited data describe prolonged adrenal suppression after musculoskeletal injection, a meta-analysis found that adrenal insufficiency may be present in 25.3% of patients 6 months after any corticosteroid administration [60]. Due to individual variation in glucocorticoid sensitivity, there is no corticosteroid dose, administration route, or treatment duration that prevents adrenal suppression [60, 64]. Furthermore, some suggest that no corticosteroids be administered for at least 4 weeks before planned surgery, to decrease the chance of adrenal crisis [28].

Osteoporosis is a known complication of long-term exogenous steroid use, attributed to steroid's negative effect on both bone quality and bone mass. A statistically significant decrease in bone mineral density of the femur has been reported after a single epidural steroid injection in postmenopausal women [65]. The literature suggests that the effect of epidural corticosteroid injection on bone mineral density is dose-dependent, whether from a single injection or multiple injections over time (i.e., cumulative dose), and may also predispose to increased vertebral fracture risk [28]. The incidence and impact of osteoporotic fractures from intra- and extraarticular corticosteroid injections for musculoskeletal indications are unknown. However, given a recognized impact of epidural steroid injections on bone mineral density, the radiologist should provide a recommendation to the primary care practitioner to consider initiating prophylactic osteoporotic treatment in postmenopausal women receiving frequent or multiple injections [28, 65]. Local Immediate Side Effects

Postinjection flare is an exacerbation of pain after injection that is significantly higher than the baseline preinjection pain. This flare is theorized to result from focal synovitis secondary to intracellular incorporation of the injected corticosteroid particulate microcrystals [9]. Postinjection flare is more common and of longer duration for particulate than nonparticulate corticosteroids [10]. The patient experience of postinjection flare can be clinically similar to the experience of crystalline-induced arthropathies, such as gout. Studies including injections of the hand, wrist, and shoulder report an incidence of postinjection flare of 32.8–35.3% [11, 12]. A study assessing flare reactions after injections performed to treat de Quer-vain tenosynovitis and trigger finger found that most flares occurred 1–2 days after injection, lasted a mean of 2 days, and did not impact long-term symptom resolution [11].

Local Delayed Side Effects

Skin hypopigmentation and atrophy—Several prospective studies describe skin hypopigmentation and atrophy, particularly after superficial peritendinous injections (e.g., to treat epicondylitis or de Quervain tenosynovitis) [13, 14] (Fig. 1). Heel pad atrophy or skin depigmentation is reported in 10% of patients who received corticosteroid injection of the plantar fascia [15]. Skin atrophy and hypopigmentation manifest within a few months of injection and typically resolve within 1 year [16]. Such dermal changes likely relate specifically to the use of particulate steroids, especially if the injectate migrates along the injection pathway within the soft tissues [17]. When superficial structures are being injected, the use of nonparticulate steroids should be considered, to minimize the potential for skin atrophy. Additionally, compression of the injection site after removal of the needle may minimize the risk of steroid migration along the needle track [18].

Open in Viewer

Fig. 1 —63-year-old woman with medial elbow pain who presented for sonographic eval‎uation of common flexor tendon. Patient had received two prior injections of 40 mg of triamcinolone acetonide (Kenalog, Proficient Rx) in medial elbow without imaging guidance within past 6 months. Photograph obtained at time of presentation for imaging shows discoloration and thinning of skin at medial elbow (dashed outline). Patient noticed this change several weeks after second injection.

Soft-tissue calcification—Soft-tissue calcification is a reported iatrogenic complication of musculoskeletal corticosteroid injection [19]. Explanatory theories include local tissue damage resulting in increased pH that precipitates calcium salt or possible calcific necrosis of collagen fibers [19]. Soft-tissue and tendinous calcifications are associated with increased pain scores and may be clinically significant [19]. Radiologists should recognize this potential corticosteroid-induced complication when assessing for changes in periarticular calcifications on postinjection imaging.

Infection—Intraarticular infection after corticosteroid injection is very rare, with an incidence thought to be less than 0.001% [20]. The incidence may be significantly less when using single-dose vials, as well as when using modern sterile technique, packaging, and cleaning methods [21]. A systematic review of case reports describes the incidences of soft-tissue infection and osteomyelitis of the underlying bone adjacent to the injection site [22]. The injection may also increase the risk of infection if the patient later undergoes surgery at the injection site. For example, studies have found that patients who received intraarticular injections of corticosteroid before a knee or hip replacement were more likely to experience a postoperative prosthetic joint infection than those who had never received an injection [23, 24]. A meta-analysis of eight studies on hip and knee joint replacements showed a 2.13 higher odds of deep infection in patients with than without a prior injection but no significant difference in superficial infection rate [25]. Another systematic review found that intraarticular injections performed within 3 months before knee arthroplasty were significantly associated with subsequent prosthetic joint infection [26]. Moreover, a study found that the use of preoperative corticosteroids was associated with a higher likelihood of superficial or deep infections in any surgical site, as well as an increased risk of respiratory, renal, and cardiac complications [27]. Given the available data, it may be advisable to delay any surgery for 1–3 months after corticosteroid injection to potentially mitigate the infection risk, particularly in patients with multiple comorbidities and chronic conditions and who are dependent on corticosteroids [28]. It is important to consider this potential infection risk before injecting patients who may be considering arthroplasty or other surgeries soon after corticosteroid injection and to counsel such patients accordingly.

Tendon, ligament, or fascial insufficiency or rupture—A rare but significant potential consequence of corticosteroid injections is tendon rupture (Figs. 2 and 3). Although its incidence is low, reported as approximately 0.1% in a meta-analysis, case reports describe such complications throughout the body [29, 30]. A review of the case reports estimated that among 95 postinjection complications, 53.7% were plantar fascia ruptures, and 26.3% were tendon ruptures involving the patellar, quadriceps, Achilles, or biceps tendon—tendons possibly already prone to increased biomechanical stress [31]. Injected corticosteroids are thought to locally decrease collagen production, which may limit potential tendon, ligament, or fascial healing and therefore increase predisposition to structural failure or rupture [32]. Even if rupture does not occur, corticosteroid use has been associated with ligament instability in the elbow, leading to refractory pain and disability [33]. A histologic study of the Achilles tendon in rabbits reported equivalent tendon weakening and loss of elasticity after either intratendinous or retrocalcaneal bursa injection, indicating that even peritendinous injections pose a risk of tendon rupture [34]. Tendon rupture may also be more strongly associated with triamcinolone acetonide than with other steroid preparations and may be associated with frequent repeated injections to a particular site [31]. Tendon rupture has been reported to occur anywhere from 2 days to 6 weeks after injection [29]. Although evidence-based consensus guidelines are lacking, some authors suggest that a maximum of three to five injections be performed at a single site, with a minimum of 2 weeks between injections, to minimize potential tendon rupture [31].

Open in Viewer

Fig. 2A —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

A, Gray-scale sonographic image obtained using 15-MHz linear probe shows intact gluteus medius insertion (dashed lines) on greater trochanter (asterisk) in long axis. Image shows thickened tendon with moderate tendinopathy but no high-grade tear. Patient underwent ultrasound-guided percutaneous tenotomy followed by injection of 20 mg of methylprednisolone acetate (Depo-Medrol, Pfizer) and 4 mL of 0.25% bupivacaine into greater trochanteric bursa. Twelve days after procedure, patient returned due to worsening pain, which was worse with weight-bearing and did not improve after course of NSAIDS and oral steroids. Patient reported no interval trauma. MRI of osseous pelvis was performed 18 days after procedure.

Open in Viewer

Fig. 2B —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

B, Coronal STIR (B) and sagittal T2-weighted (C) images, obtained 18 days after A, through greater trochanter show full-thickness tear of left gluteus medius tendon at myotendinous junction (arrow, B) with 2-cm tendon gap (arrow, C).

Open in Viewer

Fig. 2C —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

C, Coronal STIR (B) and sagittal T2-weighted (C) images, obtained 18 days after A, through greater trochanter show full-thickness tear of left gluteus medius tendon at myotendinous junction (arrow, B) with 2-cm tendon gap (arrow, C).

Open in Viewer

Fig. 3A —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

A, Gray-scale sonographic image of right hamstring tendon origin in long axis, obtained at initial presentation, shows marked thickening and hypoechoic degeneration of tendon (dashed lines) with areas of anechoic interstitial tearing (arrow). Ischial tuberosity is denoted by asterisk. Two days after third steroid injection, patient felt snap in buttock when performing household tasks.

Open in Viewer

Fig. 3B —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

B, Coronal STIR image of bony pelvis through ischial tuberosity shows near-complete rupture of right hamstring tendon origin with almost 3 cm of distal retraction (arrow).

Open in Viewer

Fig. 3C —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

C, Gray-scale sonographic image of left hamstring origin in long axis, immediately caudal to ischial tuberosity (asterisk), shows anechoic tendon gap (arrow).

Corticosteroid injections in the shoulder have been associated with a need for revision after rotator cuff repair, possibly due to underlying degeneration of the tendon quality as a result of the injection [35]. A systematic review found a nearly twofold increase in adverse outcomes after rotator cuff repair if a corticosteroid injection is administered within 6 months before surgery or if more than two injections are performed within 1 year before surgery [36]. This effect was also shown in an analysis of patients with lateral epicondylitis who underwent surgery; those receiving three or more injections were at highest risk (OR, 3.55) for requiring revision surgery [37]. Patients must be appropriately counseled regarding the potential greater need for revision surgery if a corticosteroid is administered within 6 months before surgical tendon repair.

Accelerated progression of osteoarthritis—Accelerated progression of osteoarthritis is not strictly defined but refers to an increase in joint space loss that is greater than expected to occur in a short time period. One study defined accelerated progression of osteoarthritis as a decrease in joint space by 2 mm over a 12-month period [38]. Research from the Osteoarthritis Initiative found that patients who received intraarticular knee joint injections of corticosteroid had 3.2 higher odds of showing radiographic progression of osteoarthritis than those who did not receive injections; the odds were higher at 4.67 in patients who received continuous injections [39]. Radiographic progression of osteoarthritis after injection has also been shown in the hip, with a study reporting radiographic progression of hip joint space narrowing in 44% of patients who received hip injections and were followed postprocedurally for a mean of 6 months [40].

Osseous injury—Corticosteroids induce osteocyte apoptosis, which predisposes to decreases in osseous vascularity and bone strength; these effects are particularly pronounced after intraarticular injection [41]. Indeed, subchondral insufficiency fracture, osteonecrosis, and rapid joint destruction have all been reported after intraarticular corticosteroid injection into various joints [42–44] (Figs. 4 and 5). A retrospective single-center study of 1000 patients with long-term follow-up after hip and knee corticosteroid injections showed a combined rate of osteonecrosis, insufficiency fractures, and rapid progressive osteoarthritis of 1% [45]. Such reports are controversial, and some experts believe that the reported risk of osseous injury after intraarticular injection is inflated (possibly attributed to the natural progression of osteoarthritis) and that a risk-benefit analysis would support this approach [46].

Open in Viewer

Fig. 4A —66-year-old man who presented with left groin pain.

A, Coronal STIR image through osseous pelvis shows large left hip joint effusion (asterisk).

Open in Viewer

Fig. 4B —66-year-old man who presented with left groin pain.

B, Sagittal T2-weighted image through left hip joint shows areas of full-thickness cartilage loss and subchondral marrow edema–like signal (arrows). Patient underwent same-day injection of left hip and iliopsoas bursa, receiving total of 60 mg of triamcinolone acetonide (Kenalog, Proficient Rx). Patient presented 3 months after injection with worsening hip pain.

Open in Viewer

Fig. 4C —66-year-old man who presented with left groin pain.

C, Anteroposterior radiograph of pelvis obtained 3 months after A and B shows severe joint space narrowing of left femoral articular surface with articular surface collapse and flattening (arrow).

Open in Viewer

Fig. 4D —66-year-old man who presented with left groin pain.

D, Coronal STIR (D) and coronal T1-weighted (E) images obtained same day as radiograph (C) show severe subchondral marrow edema spanning left hip and large joint effusion. Articular surface collapse (arrow, D), with areas of subchondral sclerosis (asterisk), is also present. Differential diagnosis includes rapid progression of osteoarthritis, avascular necrosis, and septic joint. Further workup identified no evidence of infection, and patient underwent left total hip arthroplasty.

Open in Viewer

Fig. 4E —66-year-old man who presented with left groin pain.

E, Coronal STIR (D) and coronal T1-weighted (E) images obtained same day as radiograph (C) show severe subchondral marrow edema spanning left hip and large joint effusion. Articular surface collapse (arrow, D), with areas of subchondral sclerosis (asterisk), is also present. Differential diagnosis includes rapid progression of osteoarthritis, avascular necrosis, and septic joint. Further workup identified no evidence of infection, and patient underwent left total hip arthroplasty.

Open in Viewer

Fig. 5A —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

A, Preinjection anteroposterior radiograph of pelvis shows severe left hip osteoarthritis with joint space narrowing, marginal spurring (arrow), and superolateral subluxation of femoral head. Intraarticular injection of 40 mg of triamcinolone acetonide (Kenalog, Proficient Rx) was performed after radiography.

Open in Viewer

Fig. 5B —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

B, Coronal STIR (B) and coronal T1-weighted (C) images of osseous pelvis obtained 2 weeks after injection show rapid destruction of femoral articular surface (asterisk) and acetabulum (arrow), with large joint effusion.

Open in Viewer

Fig. 5C —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

C, Coronal STIR (B) and coronal T1-weighted (C) images of osseous pelvis obtained 2 weeks after injection show rapid destruction of femoral articular surface (asterisk) and acetabulum (arrow), with large joint effusion.

Open in Viewer

Fig. 5D —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

D, Anteroposterior radiograph of pelvis obtained 3 weeks after injection shows progressing marked osseous resorption of femoral head, which now extends to intertrochanteric region (arrow), with cephalad migration and lateral displacement of proximal femur. No evidence of infection was found on joint aspiration or intraoperatively during total hip arthroplasty.

An expert panel review concluded that estimates of the preval‎ence of structural problems induced by intraarticular injections are highly variable and that, at present, no imaging, clinical, or laboratory markers are established for predicting subsequent osseous injury after an intraarticular injection [47]. The presence of acute joint pain that is not explained by radiographic or sonographic findings could prompt caution, with further workup obtained before potential corticosteroid injection. In this situation, MRI may help to detect osseous pathology, such as subchondral fracture, that would be a contraindication to injection; universal implementation of this approach may be cost-prohibitive given the likely low frequency of osseous injury on MRI [47]. The use of repeat radiographs before joint injection is also controversial. The panel concluded that large clinical trials are needed to determine the true incidence of these osseous complications and to formulate preintervention safety measures [47]. Although these osseous complications have a likely very low incidence, patients should be aware of the risk, and clinical judgment should be used to assess such complications' relevance for individual patients. For example, such complications may be considered more relevant in a patient with prior long-standing steroid use or a prior diagnosis of osteopenia or osteoporosis.

Systemic Immediate Side Effects

Adrenal suppression—Locally injected corticosteroid does not remain confined to the joint, and some amount is absorbed by the bloodstream. One study found that serum cortisol levels after intraarticular injection of methylprednisolone acetate were dose-dependent, peaked 2–12 hours after injection, and showed a significant downregulation for at least 1 week regardless of the injected dose [48]. A study of 10 healthy young male athletes reported adrenal suppression for up to 2 weeks after a single intra- or extraarticular injection [49]. A study found that injecting two joints simultaneously yields higher serum absorption than injecting the same total dose into one joint; this finding is theorized to be due to the exposure of a greater volume of synovium to corticosteroid, thus increasing the likelihood of systemic absorption [48]. This systemic absorption of corticosteroid accounts for many of the side effects occurring after injection, as well as the such side effects' similarity to the known side effects of exogenous steroid medication.

One study reported facial flushing (defined as erythema of the upper chest and face) after intraarticular corticosteroid injection in 40% of patients, with this side effect being unpleasant in 15% in cases, predominantly occurring in women, and being reproducible with subsequent injection [50]. Flushing is self-limiting, typically resolving within 3 days [51]. Associated systemic symptoms can include insomnia and headache.

Transient hypertension and postprandial hyperglycemia occur after intraarticular corticosteroid injection in patients both with and without diabetes and relate to suppression of the corticotropic axis; these effects may last for over 1 week [52]. These side effects' clinical implications depend on the patient's baseline condition.

Similar to the known side effects of oral or IV steroid use, psychiatric effects (e.g., altered mental status, visual hallucination) have been reported after intraarticular injection, typically occurring within 1 week after injection and occurring more commonly in older patients with decreased renal function [53, 54]. Prolonged hiccups have been reported after corticosteroid injection and are of indeterminate etiology [55].

Hypersensitivity reactions—The reported preval‎ence of hypersensitivity reactions after corticosteroid administration by any route is 0.3–0.5%; a common presentation of such reactions after musculoskeletal injection is delayed urticaria at the injection site [56, 57]. A case report described immediate anaphylaxis after intraarticular injection of triamcinolone acetonide, with confirmed hypersensitivity reaction by skin prick test [57]. Patients with frequent exposure to steroids are more likely to experience a hypersensitivity reaction. In addition, patients reporting a prior hypersensitivity reaction to corticosteroids should undergo workup for potential allergy to any alternative steroid preparation that is selected before it is administered [56]. Management of hypersensitivity reactions may include administration of oral or IV diphenhydramine hydrochloride, depending on the reaction's severity [56, 57].

Ophthalmic side effects—Central serous chorioretinopathy is a retinal detachment presenting with potentially permanent decreased vision that has been attributed to exogenously administered corticosteroids [58]. Any form of corticosteroid administration can precipitate glaucoma or increased intraocular pressure, the latter of which may lead to sudden-onset blurred vision or potentially irreversible vision loss [59]. These potential complications must be discussed before corticosteroid injection with patients who have known ophthalmic pathology and with patients reporting prior visual changes after steroid exposure.

Systemic Delayed Side Effects

As mentioned previously, intraarticular corticosteroid injection suppresses the hypothalamic-pituitary-adrenal axis, causing an at least temporary decrease in endogenous cortisol levels in 52.2% of patients [60]. This suppression can be potentially life-threatening in patients who are subject to an emotional or physical stress, such as illness or surgery, that may precipitate an adrenal crisis [6]. Most patients with adrenal insufficiency are not diagnosed until they are hospitalized for adrenal crisis, which has a reported mortality of 5% [61]. Prolonged adrenal insufficiency has been reported in some patients after repeated corticosteroid injections (i.e., iatrogenic Cushing syndrome) and may require exogenous supplementation to avoid an adrenal crisis [62, 63]. Although limited data describe prolonged adrenal suppression after musculoskeletal injection, a meta-analysis found that adrenal insufficiency may be present in 25.3% of patients 6 months after any corticosteroid administration [60]. Due to individual variation in glucocorticoid sensitivity, there is no corticosteroid dose, administration route, or treatment duration that prevents adrenal suppression [60, 64]. Furthermore, some suggest that no corticosteroids be administered for at least 4 weeks before planned surgery, to decrease the chance of adrenal crisis [28].

Osteoporosis is a known complication of long-term exogenous steroid use, attributed to steroid's negative effect on both bone quality and bone mass. A statistically significant decrease in bone mineral density of the femur has been reported after a single epidural steroid injection in postmenopausal women [65]. The literature suggests that the effect of epidural corticosteroid injection on bone mineral density is dose-dependent, whether from a single injection or multiple injections over time (i.e., cumulative dose), and may also predispose to increased vertebral fracture risk [28]. The incidence and impact of osteoporotic fractures from intra- and extraarticular corticosteroid injections for musculoskeletal indications are unknown. However, given a recognized impact of epidural steroid injections on bone mineral density, the radiologist should provide a recommendation to the primary care practitioner to consider initiating prophylactic osteoporotic treatment in postmenopausal women receiving frequent or multiple injections [28, 65]. Local Immediate Side Effects

Postinjection flare is an exacerbation of pain after injection that is significantly higher than the baseline preinjection pain. This flare is theorized to result from focal synovitis secondary to intracellular incorporation of the injected corticosteroid particulate microcrystals [9]. Postinjection flare is more common and of longer duration for particulate than nonparticulate corticosteroids [10]. The patient experience of postinjection flare can be clinically similar to the experience of crystalline-induced arthropathies, such as gout. Studies including injections of the hand, wrist, and shoulder report an incidence of postinjection flare of 32.8–35.3% [11, 12]. A study assessing flare reactions after injections performed to treat de Quer-vain tenosynovitis and trigger finger found that most flares occurred 1–2 days after injection, lasted a mean of 2 days, and did not impact long-term symptom resolution [11].

Local Delayed Side Effects

Skin hypopigmentation and atrophy—Several prospective studies describe skin hypopigmentation and atrophy, particularly after superficial peritendinous injections (e.g., to treat epicondylitis or de Quervain tenosynovitis) [13, 14] (Fig. 1). Heel pad atrophy or skin depigmentation is reported in 10% of patients who received corticosteroid injection of the plantar fascia [15]. Skin atrophy and hypopigmentation manifest within a few months of injection and typically resolve within 1 year [16]. Such dermal changes likely relate specifically to the use of particulate steroids, especially if the injectate migrates along the injection pathway within the soft tissues [17]. When superficial structures are being injected, the use of nonparticulate steroids should be considered, to minimize the potential for skin atrophy. Additionally, compression of the injection site after removal of the needle may minimize the risk of steroid migration along the needle track [18].

Open in Viewer

Fig. 1 —63-year-old woman with medial elbow pain who presented for sonographic eval‎uation of common flexor tendon. Patient had received two prior injections of 40 mg of triamcinolone acetonide (Kenalog, Proficient Rx) in medial elbow without imaging guidance within past 6 months. Photograph obtained at time of presentation for imaging shows discoloration and thinning of skin at medial elbow (dashed outline). Patient noticed this change several weeks after second injection.

Soft-tissue calcification—Soft-tissue calcification is a reported iatrogenic complication of musculoskeletal corticosteroid injection [19]. Explanatory theories include local tissue damage resulting in increased pH that precipitates calcium salt or possible calcific necrosis of collagen fibers [19]. Soft-tissue and tendinous calcifications are associated with increased pain scores and may be clinically significant [19]. Radiologists should recognize this potential corticosteroid-induced complication when assessing for changes in periarticular calcifications on postinjection imaging.

Infection—Intraarticular infection after corticosteroid injection is very rare, with an incidence thought to be less than 0.001% [20]. The incidence may be significantly less when using single-dose vials, as well as when using modern sterile technique, packaging, and cleaning methods [21]. A systematic review of case reports describes the incidences of soft-tissue infection and osteomyelitis of the underlying bone adjacent to the injection site [22]. The injection may also increase the risk of infection if the patient later undergoes surgery at the injection site. For example, studies have found that patients who received intraarticular injections of corticosteroid before a knee or hip replacement were more likely to experience a postoperative prosthetic joint infection than those who had never received an injection [23, 24]. A meta-analysis of eight studies on hip and knee joint replacements showed a 2.13 higher odds of deep infection in patients with than without a prior injection but no significant difference in superficial infection rate [25]. Another systematic review found that intraarticular injections performed within 3 months before knee arthroplasty were significantly associated with subsequent prosthetic joint infection [26]. Moreover, a study found that the use of preoperative corticosteroids was associated with a higher likelihood of superficial or deep infections in any surgical site, as well as an increased risk of respiratory, renal, and cardiac complications [27]. Given the available data, it may be advisable to delay any surgery for 1–3 months after corticosteroid injection to potentially mitigate the infection risk, particularly in patients with multiple comorbidities and chronic conditions and who are dependent on corticosteroids [28]. It is important to consider this potential infection risk before injecting patients who may be considering arthroplasty or other surgeries soon after corticosteroid injection and to counsel such patients accordingly.

Tendon, ligament, or fascial insufficiency or rupture—A rare but significant potential consequence of corticosteroid injections is tendon rupture (Figs. 2 and 3). Although its incidence is low, reported as approximately 0.1% in a meta-analysis, case reports describe such complications throughout the body [29, 30]. A review of the case reports estimated that among 95 postinjection complications, 53.7% were plantar fascia ruptures, and 26.3% were tendon ruptures involving the patellar, quadriceps, Achilles, or biceps tendon—tendons possibly already prone to increased biomechanical stress [31]. Injected corticosteroids are thought to locally decrease collagen production, which may limit potential tendon, ligament, or fascial healing and therefore increase predisposition to structural failure or rupture [32]. Even if rupture does not occur, corticosteroid use has been associated with ligament instability in the elbow, leading to refractory pain and disability [33]. A histologic study of the Achilles tendon in rabbits reported equivalent tendon weakening and loss of elasticity after either intratendinous or retrocalcaneal bursa injection, indicating that even peritendinous injections pose a risk of tendon rupture [34]. Tendon rupture may also be more strongly associated with triamcinolone acetonide than with other steroid preparations and may be associated with frequent repeated injections to a particular site [31]. Tendon rupture has been reported to occur anywhere from 2 days to 6 weeks after injection [29]. Although evidence-based consensus guidelines are lacking, some authors suggest that a maximum of three to five injections be performed at a single site, with a minimum of 2 weeks between injections, to minimize potential tendon rupture [31].

Open in Viewer

Fig. 2A —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

A, Gray-scale sonographic image obtained using 15-MHz linear probe shows intact gluteus medius insertion (dashed lines) on greater trochanter (asterisk) in long axis. Image shows thickened tendon with moderate tendinopathy but no high-grade tear. Patient underwent ultrasound-guided percutaneous tenotomy followed by injection of 20 mg of methylprednisolone acetate (Depo-Medrol, Pfizer) and 4 mL of 0.25% bupivacaine into greater trochanteric bursa. Twelve days after procedure, patient returned due to worsening pain, which was worse with weight-bearing and did not improve after course of NSAIDS and oral steroids. Patient reported no interval trauma. MRI of osseous pelvis was performed 18 days after procedure.

Open in Viewer

Fig. 2B —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

B, Coronal STIR (B) and sagittal T2-weighted (C) images, obtained 18 days after A, through greater trochanter show full-thickness tear of left gluteus medius tendon at myotendinous junction (arrow, B) with 2-cm tendon gap (arrow, C).

Open in Viewer

Fig. 2C —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

C, Coronal STIR (B) and sagittal T2-weighted (C) images, obtained 18 days after A, through greater trochanter show full-thickness tear of left gluteus medius tendon at myotendinous junction (arrow, B) with 2-cm tendon gap (arrow, C).

Open in Viewer

Fig. 3A —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

A, Gray-scale sonographic image of right hamstring tendon origin in long axis, obtained at initial presentation, shows marked thickening and hypoechoic degeneration of tendon (dashed lines) with areas of anechoic interstitial tearing (arrow). Ischial tuberosity is denoted by asterisk. Two days after third steroid injection, patient felt snap in buttock when performing household tasks.

Open in Viewer

Fig. 3B —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

B, Coronal STIR image of bony pelvis through ischial tuberosity shows near-complete rupture of right hamstring tendon origin with almost 3 cm of distal retraction (arrow).

Open in Viewer

Fig. 3C —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

C, Gray-scale sonographic image of left hamstring origin in long axis, immediately caudal to ischial tuberosity (asterisk), shows anechoic tendon gap (arrow).

Corticosteroid injections in the shoulder have been associated with a need for revision after rotator cuff repair, possibly due to underlying degeneration of the tendon quality as a result of the injection [35]. A systematic review found a nearly twofold increase in adverse outcomes after rotator cuff repair if a corticosteroid injection is administered within 6 months before surgery or if more than two injections are performed within 1 year before surgery [36]. This effect was also shown in an analysis of patients with lateral epicondylitis who underwent surgery; those receiving three or more injections were at highest risk (OR, 3.55) for requiring revision surgery [37]. Patients must be appropriately counseled regarding the potential greater need for revision surgery if a corticosteroid is administered within 6 months before surgical tendon repair.

Accelerated progression of osteoarthritis—Accelerated progression of osteoarthritis is not strictly defined but refers to an increase in joint space loss that is greater than expected to occur in a short time period. One study defined accelerated progression of osteoarthritis as a decrease in joint space by 2 mm over a 12-month period [38]. Research from the Osteoarthritis Initiative found that patients who received intraarticular knee joint injections of corticosteroid had 3.2 higher odds of showing radiographic progression of osteoarthritis than those who did not receive injections; the odds were higher at 4.67 in patients who received continuous injections [39]. Radiographic progression of osteoarthritis after injection has also been shown in the hip, with a study reporting radiographic progression of hip joint space narrowing in 44% of patients who received hip injections and were followed postprocedurally for a mean of 6 months [40].

Osseous injury—Corticosteroids induce osteocyte apoptosis, which predisposes to decreases in osseous vascularity and bone strength; these effects are particularly pronounced after intraarticular injection [41]. Indeed, subchondral insufficiency fracture, osteonecrosis, and rapid joint destruction have all been reported after intraarticular corticosteroid injection into various joints [42–44] (Figs. 4 and 5). A retrospective single-center study of 1000 patients with long-term follow-up after hip and knee corticosteroid injections showed a combined rate of osteonecrosis, insufficiency fractures, and rapid progressive osteoarthritis of 1% [45]. Such reports are controversial, and some experts believe that the reported risk of osseous injury after intraarticular injection is inflated (possibly attributed to the natural progression of osteoarthritis) and that a risk-benefit analysis would support this approach [46].

Open in Viewer

Fig. 4A —66-year-old man who presented with left groin pain.

A, Coronal STIR image through osseous pelvis shows large left hip joint effusion (asterisk).

Open in Viewer

Fig. 4B —66-year-old man who presented with left groin pain.

B, Sagittal T2-weighted image through left hip joint shows areas of full-thickness cartilage loss and subchondral marrow edema–like signal (arrows). Patient underwent same-day injection of left hip and iliopsoas bursa, receiving total of 60 mg of triamcinolone acetonide (Kenalog, Proficient Rx). Patient presented 3 months after injection with worsening hip pain.

Open in Viewer

Fig. 4C —66-year-old man who presented with left groin pain.

C, Anteroposterior radiograph of pelvis obtained 3 months after A and B shows severe joint space narrowing of left femoral articular surface with articular surface collapse and flattening (arrow).

Open in Viewer

Fig. 4D —66-year-old man who presented with left groin pain.

D, Coronal STIR (D) and coronal T1-weighted (E) images obtained same day as radiograph (C) show severe subchondral marrow edema spanning left hip and large joint effusion. Articular surface collapse (arrow, D), with areas of subchondral sclerosis (asterisk), is also present. Differential diagnosis includes rapid progression of osteoarthritis, avascular necrosis, and septic joint. Further workup identified no evidence of infection, and patient underwent left total hip arthroplasty.

Open in Viewer

Fig. 4E —66-year-old man who presented with left groin pain.

E, Coronal STIR (D) and coronal T1-weighted (E) images obtained same day as radiograph (C) show severe subchondral marrow edema spanning left hip and large joint effusion. Articular surface collapse (arrow, D), with areas of subchondral sclerosis (asterisk), is also present. Differential diagnosis includes rapid progression of osteoarthritis, avascular necrosis, and septic joint. Further workup identified no evidence of infection, and patient underwent left total hip arthroplasty.

Open in Viewer

Fig. 5A —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

A, Preinjection anteroposterior radiograph of pelvis shows severe left hip osteoarthritis with joint space narrowing, marginal spurring (arrow), and superolateral subluxation of femoral head. Intraarticular injection of 40 mg of triamcinolone acetonide (Kenalog, Proficient Rx) was performed after radiography.

Open in Viewer

Fig. 5B —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

B, Coronal STIR (B) and coronal T1-weighted (C) images of osseous pelvis obtained 2 weeks after injection show rapid destruction of femoral articular surface (asterisk) and acetabulum (arrow), with large joint effusion.

Open in Viewer

Fig. 5C —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

C, Coronal STIR (B) and coronal T1-weighted (C) images of osseous pelvis obtained 2 weeks after injection show rapid destruction of femoral articular surface (asterisk) and acetabulum (arrow), with large joint effusion.

Open in Viewer

Fig. 5D —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

D, Anteroposterior radiograph of pelvis obtained 3 weeks after injection shows progressing marked osseous resorption of femoral head, which now extends to intertrochanteric region (arrow), with cephalad migration and lateral displacement of proximal femur. No evidence of infection was found on joint aspiration or intraoperatively during total hip arthroplasty.

An expert panel review concluded that estimates of the preval‎ence of structural problems induced by intraarticular injections are highly variable and that, at present, no imaging, clinical, or laboratory markers are established for predicting subsequent osseous injury after an intraarticular injection [47]. The presence of acute joint pain that is not explained by radiographic or sonographic findings could prompt caution, with further workup obtained before potential corticosteroid injection. In this situation, MRI may help to detect osseous pathology, such as subchondral fracture, that would be a contraindication to injection; universal implementation of this approach may be cost-prohibitive given the likely low frequency of osseous injury on MRI [47]. The use of repeat radiographs before joint injection is also controversial. The panel concluded that large clinical trials are needed to determine the true incidence of these osseous complications and to formulate preintervention safety measures [47]. Although these osseous complications have a likely very low incidence, patients should be aware of the risk, and clinical judgment should be used to assess such complications' relevance for individual patients. For example, such complications may be considered more relevant in a patient with prior long-standing steroid use or a prior diagnosis of osteopenia or osteoporosis.

Systemic Immediate Side Effects

Adrenal suppression—Locally injected corticosteroid does not remain confined to the joint, and some amount is absorbed by the bloodstream. One study found that serum cortisol levels after intraarticular injection of methylprednisolone acetate were dose-dependent, peaked 2–12 hours after injection, and showed a significant downregulation for at least 1 week regardless of the injected dose [48]. A study of 10 healthy young male athletes reported adrenal suppression for up to 2 weeks after a single intra- or extraarticular injection [49]. A study found that injecting two joints simultaneously yields higher serum absorption than injecting the same total dose into one joint; this finding is theorized to be due to the exposure of a greater volume of synovium to corticosteroid, thus increasing the likelihood of systemic absorption [48]. This systemic absorption of corticosteroid accounts for many of the side effects occurring after injection, as well as the such side effects' similarity to the known side effects of exogenous steroid medication.

One study reported facial flushing (defined as erythema of the upper chest and face) after intraarticular corticosteroid injection in 40% of patients, with this side effect being unpleasant in 15% in cases, predominantly occurring in women, and being reproducible with subsequent injection [50]. Flushing is self-limiting, typically resolving within 3 days [51]. Associated systemic symptoms can include insomnia and headache.

Transient hypertension and postprandial hyperglycemia occur after intraarticular corticosteroid injection in patients both with and without diabetes and relate to suppression of the corticotropic axis; these effects may last for over 1 week [52]. These side effects' clinical implications depend on the patient's baseline condition.

Similar to the known side effects of oral or IV steroid use, psychiatric effects (e.g., altered mental status, visual hallucination) have been reported after intraarticular injection, typically occurring within 1 week after injection and occurring more commonly in older patients with decreased renal function [53, 54]. Prolonged hiccups have been reported after corticosteroid injection and are of indeterminate etiology [55].

Hypersensitivity reactions—The reported preval‎ence of hypersensitivity reactions after corticosteroid administration by any route is 0.3–0.5%; a common presentation of such reactions after musculoskeletal injection is delayed urticaria at the injection site [56, 57]. A case report described immediate anaphylaxis after intraarticular injection of triamcinolone acetonide, with confirmed hypersensitivity reaction by skin prick test [57]. Patients with frequent exposure to steroids are more likely to experience a hypersensitivity reaction. In addition, patients reporting a prior hypersensitivity reaction to corticosteroids should undergo workup for potential allergy to any alternative steroid preparation that is selected before it is administered [56]. Management of hypersensitivity reactions may include administration of oral or IV diphenhydramine hydrochloride, depending on the reaction's severity [56, 57].

Ophthalmic side effects—Central serous chorioretinopathy is a retinal detachment presenting with potentially permanent decreased vision that has been attributed to exogenously administered corticosteroids [58]. Any form of corticosteroid administration can precipitate glaucoma or increased intraocular pressure, the latter of which may lead to sudden-onset blurred vision or potentially irreversible vision loss [59]. These potential complications must be discussed before corticosteroid injection with patients who have known ophthalmic pathology and with patients reporting prior visual changes after steroid exposure.

Systemic Delayed Side Effects

As mentioned previously, intraarticular corticosteroid injection suppresses the hypothalamic-pituitary-adrenal axis, causing an at least temporary decrease in endogenous cortisol levels in 52.2% of patients [60]. This suppression can be potentially life-threatening in patients who are subject to an emotional or physical stress, such as illness or surgery, that may precipitate an adrenal crisis [6]. Most patients with adrenal insufficiency are not diagnosed until they are hospitalized for adrenal crisis, which has a reported mortality of 5% [61]. Prolonged adrenal insufficiency has been reported in some patients after repeated corticosteroid injections (i.e., iatrogenic Cushing syndrome) and may require exogenous supplementation to avoid an adrenal crisis [62, 63]. Although limited data describe prolonged adrenal suppression after musculoskeletal injection, a meta-analysis found that adrenal insufficiency may be present in 25.3% of patients 6 months after any corticosteroid administration [60]. Due to individual variation in glucocorticoid sensitivity, there is no corticosteroid dose, administration route, or treatment duration that prevents adrenal suppression [60, 64]. Furthermore, some suggest that no corticosteroids be administered for at least 4 weeks before planned surgery, to decrease the chance of adrenal crisis [28].

Osteoporosis is a known complication of long-term exogenous steroid use, attributed to steroid's negative effect on both bone quality and bone mass. A statistically significant decrease in bone mineral density of the femur has been reported after a single epidural steroid injection in postmenopausal women [65]. The literature suggests that the effect of epidural corticosteroid injection on bone mineral density is dose-dependent, whether from a single injection or multiple injections over time (i.e., cumulative dose), and may also predispose to increased vertebral fracture risk [28]. The incidence and impact of osteoporotic fractures from intra- and extraarticular corticosteroid injections for musculoskeletal indications are unknown. However, given a recognized impact of epidural steroid injections on bone mineral density, the radiologist should provide a recommendation to the primary care practitioner to consider initiating prophylactic osteoporotic treatment in postmenopausal women receiving frequent or multiple injections [28, 65]. Local Immediate Side Effects

Postinjection flare is an exacerbation of pain after injection that is significantly higher than the baseline preinjection pain. This flare is theorized to result from focal synovitis secondary to intracellular incorporation of the injected corticosteroid particulate microcrystals [9]. Postinjection flare is more common and of longer duration for particulate than nonparticulate corticosteroids [10]. The patient experience of postinjection flare can be clinically similar to the experience of crystalline-induced arthropathies, such as gout. Studies including injections of the hand, wrist, and shoulder report an incidence of postinjection flare of 32.8–35.3% [11, 12]. A study assessing flare reactions after injections performed to treat de Quer-vain tenosynovitis and trigger finger found that most flares occurred 1–2 days after injection, lasted a mean of 2 days, and did not impact long-term symptom resolution [11].

Local Delayed Side Effects

Skin hypopigmentation and atrophy—Several prospective studies describe skin hypopigmentation and atrophy, particularly after superficial peritendinous injections (e.g., to treat epicondylitis or de Quervain tenosynovitis) [13, 14] (Fig. 1). Heel pad atrophy or skin depigmentation is reported in 10% of patients who received corticosteroid injection of the plantar fascia [15]. Skin atrophy and hypopigmentation manifest within a few months of injection and typically resolve within 1 year [16]. Such dermal changes likely relate specifically to the use of particulate steroids, especially if the injectate migrates along the injection pathway within the soft tissues [17]. When superficial structures are being injected, the use of nonparticulate steroids should be considered, to minimize the potential for skin atrophy. Additionally, compression of the injection site after removal of the needle may minimize the risk of steroid migration along the needle track [18].

Open in Viewer

Fig. 1 —63-year-old woman with medial elbow pain who presented for sonographic eval‎uation of common flexor tendon. Patient had received two prior injections of 40 mg of triamcinolone acetonide (Kenalog, Proficient Rx) in medial elbow without imaging guidance within past 6 months. Photograph obtained at time of presentation for imaging shows discoloration and thinning of skin at medial elbow (dashed outline). Patient noticed this change several weeks after second injection.

Soft-tissue calcification—Soft-tissue calcification is a reported iatrogenic complication of musculoskeletal corticosteroid injection [19]. Explanatory theories include local tissue damage resulting in increased pH that precipitates calcium salt or possible calcific necrosis of collagen fibers [19]. Soft-tissue and tendinous calcifications are associated with increased pain scores and may be clinically significant [19]. Radiologists should recognize this potential corticosteroid-induced complication when assessing for changes in periarticular calcifications on postinjection imaging.

Infection—Intraarticular infection after corticosteroid injection is very rare, with an incidence thought to be less than 0.001% [20]. The incidence may be significantly less when using single-dose vials, as well as when using modern sterile technique, packaging, and cleaning methods [21]. A systematic review of case reports describes the incidences of soft-tissue infection and osteomyelitis of the underlying bone adjacent to the injection site [22]. The injection may also increase the risk of infection if the patient later undergoes surgery at the injection site. For example, studies have found that patients who received intraarticular injections of corticosteroid before a knee or hip replacement were more likely to experience a postoperative prosthetic joint infection than those who had never received an injection [23, 24]. A meta-analysis of eight studies on hip and knee joint replacements showed a 2.13 higher odds of deep infection in patients with than without a prior injection but no significant difference in superficial infection rate [25]. Another systematic review found that intraarticular injections performed within 3 months before knee arthroplasty were significantly associated with subsequent prosthetic joint infection [26]. Moreover, a study found that the use of preoperative corticosteroids was associated with a higher likelihood of superficial or deep infections in any surgical site, as well as an increased risk of respiratory, renal, and cardiac complications [27]. Given the available data, it may be advisable to delay any surgery for 1–3 months after corticosteroid injection to potentially mitigate the infection risk, particularly in patients with multiple comorbidities and chronic conditions and who are dependent on corticosteroids [28]. It is important to consider this potential infection risk before injecting patients who may be considering arthroplasty or other surgeries soon after corticosteroid injection and to counsel such patients accordingly.

Tendon, ligament, or fascial insufficiency or rupture—A rare but significant potential consequence of corticosteroid injections is tendon rupture (Figs. 2 and 3). Although its incidence is low, reported as approximately 0.1% in a meta-analysis, case reports describe such complications throughout the body [29, 30]. A review of the case reports estimated that among 95 postinjection complications, 53.7% were plantar fascia ruptures, and 26.3% were tendon ruptures involving the patellar, quadriceps, Achilles, or biceps tendon—tendons possibly already prone to increased biomechanical stress [31]. Injected corticosteroids are thought to locally decrease collagen production, which may limit potential tendon, ligament, or fascial healing and therefore increase predisposition to structural failure or rupture [32]. Even if rupture does not occur, corticosteroid use has been associated with ligament instability in the elbow, leading to refractory pain and disability [33]. A histologic study of the Achilles tendon in rabbits reported equivalent tendon weakening and loss of elasticity after either intratendinous or retrocalcaneal bursa injection, indicating that even peritendinous injections pose a risk of tendon rupture [34]. Tendon rupture may also be more strongly associated with triamcinolone acetonide than with other steroid preparations and may be associated with frequent repeated injections to a particular site [31]. Tendon rupture has been reported to occur anywhere from 2 days to 6 weeks after injection [29]. Although evidence-based consensus guidelines are lacking, some authors suggest that a maximum of three to five injections be performed at a single site, with a minimum of 2 weeks between injections, to minimize potential tendon rupture [31].

Open in Viewer

Fig. 2A —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

A, Gray-scale sonographic image obtained using 15-MHz linear probe shows intact gluteus medius insertion (dashed lines) on greater trochanter (asterisk) in long axis. Image shows thickened tendon with moderate tendinopathy but no high-grade tear. Patient underwent ultrasound-guided percutaneous tenotomy followed by injection of 20 mg of methylprednisolone acetate (Depo-Medrol, Pfizer) and 4 mL of 0.25% bupivacaine into greater trochanteric bursa. Twelve days after procedure, patient returned due to worsening pain, which was worse with weight-bearing and did not improve after course of NSAIDS and oral steroids. Patient reported no interval trauma. MRI of osseous pelvis was performed 18 days after procedure.

Open in Viewer

Fig. 2B —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

B, Coronal STIR (B) and sagittal T2-weighted (C) images, obtained 18 days after A, through greater trochanter show full-thickness tear of left gluteus medius tendon at myotendinous junction (arrow, B) with 2-cm tendon gap (arrow, C).

Open in Viewer

Fig. 2C —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

C, Coronal STIR (B) and sagittal T2-weighted (C) images, obtained 18 days after A, through greater trochanter show full-thickness tear of left gluteus medius tendon at myotendinous junction (arrow, B) with 2-cm tendon gap (arrow, C).

Open in Viewer

Fig. 3A —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

A, Gray-scale sonographic image of right hamstring tendon origin in long axis, obtained at initial presentation, shows marked thickening and hypoechoic degeneration of tendon (dashed lines) with areas of anechoic interstitial tearing (arrow). Ischial tuberosity is denoted by asterisk. Two days after third steroid injection, patient felt snap in buttock when performing household tasks.

Open in Viewer

Fig. 3B —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

B, Coronal STIR image of bony pelvis through ischial tuberosity shows near-complete rupture of right hamstring tendon origin with almost 3 cm of distal retraction (arrow).

Open in Viewer

Fig. 3C —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

C, Gray-scale sonographic image of left hamstring origin in long axis, immediately caudal to ischial tuberosity (asterisk), shows anechoic tendon gap (arrow).

Corticosteroid injections in the shoulder have been associated with a need for revision after rotator cuff repair, possibly due to underlying degeneration of the tendon quality as a result of the injection [35]. A systematic review found a nearly twofold increase in adverse outcomes after rotator cuff repair if a corticosteroid injection is administered within 6 months before surgery or if more than two injections are performed within 1 year before surgery [36]. This effect was also shown in an analysis of patients with lateral epicondylitis who underwent surgery; those receiving three or more injections were at highest risk (OR, 3.55) for requiring revision surgery [37]. Patients must be appropriately counseled regarding the potential greater need for revision surgery if a corticosteroid is administered within 6 months before surgical tendon repair.

Accelerated progression of osteoarthritis—Accelerated progression of osteoarthritis is not strictly defined but refers to an increase in joint space loss that is greater than expected to occur in a short time period. One study defined accelerated progression of osteoarthritis as a decrease in joint space by 2 mm over a 12-month period [38]. Research from the Osteoarthritis Initiative found that patients who received intraarticular knee joint injections of corticosteroid had 3.2 higher odds of showing radiographic progression of osteoarthritis than those who did not receive injections; the odds were higher at 4.67 in patients who received continuous injections [39]. Radiographic progression of osteoarthritis after injection has also been shown in the hip, with a study reporting radiographic progression of hip joint space narrowing in 44% of patients who received hip injections and were followed postprocedurally for a mean of 6 months [40].

Osseous injury—Corticosteroids induce osteocyte apoptosis, which predisposes to decreases in osseous vascularity and bone strength; these effects are particularly pronounced after intraarticular injection [41]. Indeed, subchondral insufficiency fracture, osteonecrosis, and rapid joint destruction have all been reported after intraarticular corticosteroid injection into various joints [42–44] (Figs. 4 and 5). A retrospective single-center study of 1000 patients with long-term follow-up after hip and knee corticosteroid injections showed a combined rate of osteonecrosis, insufficiency fractures, and rapid progressive osteoarthritis of 1% [45]. Such reports are controversial, and some experts believe that the reported risk of osseous injury after intraarticular injection is inflated (possibly attributed to the natural progression of osteoarthritis) and that a risk-benefit analysis would support this approach [46].

Open in Viewer

Fig. 4A —66-year-old man who presented with left groin pain.

A, Coronal STIR image through osseous pelvis shows large left hip joint effusion (asterisk).

Open in Viewer

Fig. 4B —66-year-old man who presented with left groin pain.

B, Sagittal T2-weighted image through left hip joint shows areas of full-thickness cartilage loss and subchondral marrow edema–like signal (arrows). Patient underwent same-day injection of left hip and iliopsoas bursa, receiving total of 60 mg of triamcinolone acetonide (Kenalog, Proficient Rx). Patient presented 3 months after injection with worsening hip pain.

Open in Viewer

Fig. 4C —66-year-old man who presented with left groin pain.

C, Anteroposterior radiograph of pelvis obtained 3 months after A and B shows severe joint space narrowing of left femoral articular surface with articular surface collapse and flattening (arrow).

Open in Viewer

Fig. 4D —66-year-old man who presented with left groin pain.

D, Coronal STIR (D) and coronal T1-weighted (E) images obtained same day as radiograph (C) show severe subchondral marrow edema spanning left hip and large joint effusion. Articular surface collapse (arrow, D), with areas of subchondral sclerosis (asterisk), is also present. Differential diagnosis includes rapid progression of osteoarthritis, avascular necrosis, and septic joint. Further workup identified no evidence of infection, and patient underwent left total hip arthroplasty.

Open in Viewer

Fig. 4E —66-year-old man who presented with left groin pain.

E, Coronal STIR (D) and coronal T1-weighted (E) images obtained same day as radiograph (C) show severe subchondral marrow edema spanning left hip and large joint effusion. Articular surface collapse (arrow, D), with areas of subchondral sclerosis (asterisk), is also present. Differential diagnosis includes rapid progression of osteoarthritis, avascular necrosis, and septic joint. Further workup identified no evidence of infection, and patient underwent left total hip arthroplasty.

Open in Viewer

Fig. 5A —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

A, Preinjection anteroposterior radiograph of pelvis shows severe left hip osteoarthritis with joint space narrowing, marginal spurring (arrow), and superolateral subluxation of femoral head. Intraarticular injection of 40 mg of triamcinolone acetonide (Kenalog, Proficient Rx) was performed after radiography.

Open in Viewer

Fig. 5B —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

B, Coronal STIR (B) and coronal T1-weighted (C) images of osseous pelvis obtained 2 weeks after injection show rapid destruction of femoral articular surface (asterisk) and acetabulum (arrow), with large joint effusion.

Open in Viewer

Fig. 5C —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

C, Coronal STIR (B) and coronal T1-weighted (C) images of osseous pelvis obtained 2 weeks after injection show rapid destruction of femoral articular surface (asterisk) and acetabulum (arrow), with large joint effusion.

Open in Viewer

Fig. 5D —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

D, Anteroposterior radiograph of pelvis obtained 3 weeks after injection shows progressing marked osseous resorption of femoral head, which now extends to intertrochanteric region (arrow), with cephalad migration and lateral displacement of proximal femur. No evidence of infection was found on joint aspiration or intraoperatively during total hip arthroplasty.

An expert panel review concluded that estimates of the preval‎ence of structural problems induced by intraarticular injections are highly variable and that, at present, no imaging, clinical, or laboratory markers are established for predicting subsequent osseous injury after an intraarticular injection [47]. The presence of acute joint pain that is not explained by radiographic or sonographic findings could prompt caution, with further workup obtained before potential corticosteroid injection. In this situation, MRI may help to detect osseous pathology, such as subchondral fracture, that would be a contraindication to injection; universal implementation of this approach may be cost-prohibitive given the likely low frequency of osseous injury on MRI [47]. The use of repeat radiographs before joint injection is also controversial. The panel concluded that large clinical trials are needed to determine the true incidence of these osseous complications and to formulate preintervention safety measures [47]. Although these osseous complications have a likely very low incidence, patients should be aware of the risk, and clinical judgment should be used to assess such complications' relevance for individual patients. For example, such complications may be considered more relevant in a patient with prior long-standing steroid use or a prior diagnosis of osteopenia or osteoporosis.

Systemic Immediate Side Effects

Adrenal suppression—Locally injected corticosteroid does not remain confined to the joint, and some amount is absorbed by the bloodstream. One study found that serum cortisol levels after intraarticular injection of methylprednisolone acetate were dose-dependent, peaked 2–12 hours after injection, and showed a significant downregulation for at least 1 week regardless of the injected dose [48]. A study of 10 healthy young male athletes reported adrenal suppression for up to 2 weeks after a single intra- or extraarticular injection [49]. A study found that injecting two joints simultaneously yields higher serum absorption than injecting the same total dose into one joint; this finding is theorized to be due to the exposure of a greater volume of synovium to corticosteroid, thus increasing the likelihood of systemic absorption [48]. This systemic absorption of corticosteroid accounts for many of the side effects occurring after injection, as well as the such side effects' similarity to the known side effects of exogenous steroid medication.

One study reported facial flushing (defined as erythema of the upper chest and face) after intraarticular corticosteroid injection in 40% of patients, with this side effect being unpleasant in 15% in cases, predominantly occurring in women, and being reproducible with subsequent injection [50]. Flushing is self-limiting, typically resolving within 3 days [51]. Associated systemic symptoms can include insomnia and headache.

Transient hypertension and postprandial hyperglycemia occur after intraarticular corticosteroid injection in patients both with and without diabetes and relate to suppression of the corticotropic axis; these effects may last for over 1 week [52]. These side effects' clinical implications depend on the patient's baseline condition.

Similar to the known side effects of oral or IV steroid use, psychiatric effects (e.g., altered mental status, visual hallucination) have been reported after intraarticular injection, typically occurring within 1 week after injection and occurring more commonly in older patients with decreased renal function [53, 54]. Prolonged hiccups have been reported after corticosteroid injection and are of indeterminate etiology [55].

Hypersensitivity reactions—The reported preval‎ence of hypersensitivity reactions after corticosteroid administration by any route is 0.3–0.5%; a common presentation of such reactions after musculoskeletal injection is delayed urticaria at the injection site [56, 57]. A case report described immediate anaphylaxis after intraarticular injection of triamcinolone acetonide, with confirmed hypersensitivity reaction by skin prick test [57]. Patients with frequent exposure to steroids are more likely to experience a hypersensitivity reaction. In addition, patients reporting a prior hypersensitivity reaction to corticosteroids should undergo workup for potential allergy to any alternative steroid preparation that is selected before it is administered [56]. Management of hypersensitivity reactions may include administration of oral or IV diphenhydramine hydrochloride, depending on the reaction's severity [56, 57].

Ophthalmic side effects—Central serous chorioretinopathy is a retinal detachment presenting with potentially permanent decreased vision that has been attributed to exogenously administered corticosteroids [58]. Any form of corticosteroid administration can precipitate glaucoma or increased intraocular pressure, the latter of which may lead to sudden-onset blurred vision or potentially irreversible vision loss [59]. These potential complications must be discussed before corticosteroid injection with patients who have known ophthalmic pathology and with patients reporting prior visual changes after steroid exposure.

Systemic Delayed Side Effects

As mentioned previously, intraarticular corticosteroid injection suppresses the hypothalamic-pituitary-adrenal axis, causing an at least temporary decrease in endogenous cortisol levels in 52.2% of patients [60]. This suppression can be potentially life-threatening in patients who are subject to an emotional or physical stress, such as illness or surgery, that may precipitate an adrenal crisis [6]. Most patients with adrenal insufficiency are not diagnosed until they are hospitalized for adrenal crisis, which has a reported mortality of 5% [61]. Prolonged adrenal insufficiency has been reported in some patients after repeated corticosteroid injections (i.e., iatrogenic Cushing syndrome) and may require exogenous supplementation to avoid an adrenal crisis [62, 63]. Although limited data describe prolonged adrenal suppression after musculoskeletal injection, a meta-analysis found that adrenal insufficiency may be present in 25.3% of patients 6 months after any corticosteroid administration [60]. Due to individual variation in glucocorticoid sensitivity, there is no corticosteroid dose, administration route, or treatment duration that prevents adrenal suppression [60, 64]. Furthermore, some suggest that no corticosteroids be administered for at least 4 weeks before planned surgery, to decrease the chance of adrenal crisis [28].

Osteoporosis is a known complication of long-term exogenous steroid use, attributed to steroid's negative effect on both bone quality and bone mass. A statistically significant decrease in bone mineral density of the femur has been reported after a single epidural steroid injection in postmenopausal women [65]. The literature suggests that the effect of epidural corticosteroid injection on bone mineral density is dose-dependent, whether from a single injection or multiple injections over time (i.e., cumulative dose), and may also predispose to increased vertebral fracture risk [28]. The incidence and impact of osteoporotic fractures from intra- and extraarticular corticosteroid injections for musculoskeletal indications are unknown. However, given a recognized impact of epidural steroid injections on bone mineral density, the radiologist should provide a recommendation to the primary care practitioner to consider initiating prophylactic osteoporotic treatment in postmenopausal women receiving frequent or multiple injections [28, 65]. Local Immediate Side Effects

Postinjection flare is an exacerbation of pain after injection that is significantly higher than the baseline preinjection pain. This flare is theorized to result from focal synovitis secondary to intracellular incorporation of the injected corticosteroid particulate microcrystals [9]. Postinjection flare is more common and of longer duration for particulate than nonparticulate corticosteroids [10]. The patient experience of postinjection flare can be clinically similar to the experience of crystalline-induced arthropathies, such as gout. Studies including injections of the hand, wrist, and shoulder report an incidence of postinjection flare of 32.8–35.3% [11, 12]. A study assessing flare reactions after injections performed to treat de Quer-vain tenosynovitis and trigger finger found that most flares occurred 1–2 days after injection, lasted a mean of 2 days, and did not impact long-term symptom resolution [11].

Local Delayed Side Effects

Skin hypopigmentation and atrophy—Several prospective studies describe skin hypopigmentation and atrophy, particularly after superficial peritendinous injections (e.g., to treat epicondylitis or de Quervain tenosynovitis) [13, 14] (Fig. 1). Heel pad atrophy or skin depigmentation is reported in 10% of patients who received corticosteroid injection of the plantar fascia [15]. Skin atrophy and hypopigmentation manifest within a few months of injection and typically resolve within 1 year [16]. Such dermal changes likely relate specifically to the use of particulate steroids, especially if the injectate migrates along the injection pathway within the soft tissues [17]. When superficial structures are being injected, the use of nonparticulate steroids should be considered, to minimize the potential for skin atrophy. Additionally, compression of the injection site after removal of the needle may minimize the risk of steroid migration along the needle track [18].

Open in Viewer

Fig. 1 —63-year-old woman with medial elbow pain who presented for sonographic eval‎uation of common flexor tendon. Patient had received two prior injections of 40 mg of triamcinolone acetonide (Kenalog, Proficient Rx) in medial elbow without imaging guidance within past 6 months. Photograph obtained at time of presentation for imaging shows discoloration and thinning of skin at medial elbow (dashed outline). Patient noticed this change several weeks after second injection.

Soft-tissue calcification—Soft-tissue calcification is a reported iatrogenic complication of musculoskeletal corticosteroid injection [19]. Explanatory theories include local tissue damage resulting in increased pH that precipitates calcium salt or possible calcific necrosis of collagen fibers [19]. Soft-tissue and tendinous calcifications are associated with increased pain scores and may be clinically significant [19]. Radiologists should recognize this potential corticosteroid-induced complication when assessing for changes in periarticular calcifications on postinjection imaging.

Infection—Intraarticular infection after corticosteroid injection is very rare, with an incidence thought to be less than 0.001% [20]. The incidence may be significantly less when using single-dose vials, as well as when using modern sterile technique, packaging, and cleaning methods [21]. A systematic review of case reports describes the incidences of soft-tissue infection and osteomyelitis of the underlying bone adjacent to the injection site [22]. The injection may also increase the risk of infection if the patient later undergoes surgery at the injection site. For example, studies have found that patients who received intraarticular injections of corticosteroid before a knee or hip replacement were more likely to experience a postoperative prosthetic joint infection than those who had never received an injection [23, 24]. A meta-analysis of eight studies on hip and knee joint replacements showed a 2.13 higher odds of deep infection in patients with than without a prior injection but no significant difference in superficial infection rate [25]. Another systematic review found that intraarticular injections performed within 3 months before knee arthroplasty were significantly associated with subsequent prosthetic joint infection [26]. Moreover, a study found that the use of preoperative corticosteroids was associated with a higher likelihood of superficial or deep infections in any surgical site, as well as an increased risk of respiratory, renal, and cardiac complications [27]. Given the available data, it may be advisable to delay any surgery for 1–3 months after corticosteroid injection to potentially mitigate the infection risk, particularly in patients with multiple comorbidities and chronic conditions and who are dependent on corticosteroids [28]. It is important to consider this potential infection risk before injecting patients who may be considering arthroplasty or other surgeries soon after corticosteroid injection and to counsel such patients accordingly.

Tendon, ligament, or fascial insufficiency or rupture—A rare but significant potential consequence of corticosteroid injections is tendon rupture (Figs. 2 and 3). Although its incidence is low, reported as approximately 0.1% in a meta-analysis, case reports describe such complications throughout the body [29, 30]. A review of the case reports estimated that among 95 postinjection complications, 53.7% were plantar fascia ruptures, and 26.3% were tendon ruptures involving the patellar, quadriceps, Achilles, or biceps tendon—tendons possibly already prone to increased biomechanical stress [31]. Injected corticosteroids are thought to locally decrease collagen production, which may limit potential tendon, ligament, or fascial healing and therefore increase predisposition to structural failure or rupture [32]. Even if rupture does not occur, corticosteroid use has been associated with ligament instability in the elbow, leading to refractory pain and disability [33]. A histologic study of the Achilles tendon in rabbits reported equivalent tendon weakening and loss of elasticity after either intratendinous or retrocalcaneal bursa injection, indicating that even peritendinous injections pose a risk of tendon rupture [34]. Tendon rupture may also be more strongly associated with triamcinolone acetonide than with other steroid preparations and may be associated with frequent repeated injections to a particular site [31]. Tendon rupture has been reported to occur anywhere from 2 days to 6 weeks after injection [29]. Although evidence-based consensus guidelines are lacking, some authors suggest that a maximum of three to five injections be performed at a single site, with a minimum of 2 weeks between injections, to minimize potential tendon rupture [31].

Open in Viewer

Fig. 2A —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

A, Gray-scale sonographic image obtained using 15-MHz linear probe shows intact gluteus medius insertion (dashed lines) on greater trochanter (asterisk) in long axis. Image shows thickened tendon with moderate tendinopathy but no high-grade tear. Patient underwent ultrasound-guided percutaneous tenotomy followed by injection of 20 mg of methylprednisolone acetate (Depo-Medrol, Pfizer) and 4 mL of 0.25% bupivacaine into greater trochanteric bursa. Twelve days after procedure, patient returned due to worsening pain, which was worse with weight-bearing and did not improve after course of NSAIDS and oral steroids. Patient reported no interval trauma. MRI of osseous pelvis was performed 18 days after procedure.

Open in Viewer

Fig. 2B —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

B, Coronal STIR (B) and sagittal T2-weighted (C) images, obtained 18 days after A, through greater trochanter show full-thickness tear of left gluteus medius tendon at myotendinous junction (arrow, B) with 2-cm tendon gap (arrow, C).

Open in Viewer

Fig. 2C —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

C, Coronal STIR (B) and sagittal T2-weighted (C) images, obtained 18 days after A, through greater trochanter show full-thickness tear of left gluteus medius tendon at myotendinous junction (arrow, B) with 2-cm tendon gap (arrow, C).

Open in Viewer

Fig. 3A —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

A, Gray-scale sonographic image of right hamstring tendon origin in long axis, obtained at initial presentation, shows marked thickening and hypoechoic degeneration of tendon (dashed lines) with areas of anechoic interstitial tearing (arrow). Ischial tuberosity is denoted by asterisk. Two days after third steroid injection, patient felt snap in buttock when performing household tasks.

Open in Viewer

Fig. 3B —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

B, Coronal STIR image of bony pelvis through ischial tuberosity shows near-complete rupture of right hamstring tendon origin with almost 3 cm of distal retraction (arrow).

Open in Viewer

Fig. 3C —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

C, Gray-scale sonographic image of left hamstring origin in long axis, immediately caudal to ischial tuberosity (asterisk), shows anechoic tendon gap (arrow).

Corticosteroid injections in the shoulder have been associated with a need for revision after rotator cuff repair, possibly due to underlying degeneration of the tendon quality as a result of the injection [35]. A systematic review found a nearly twofold increase in adverse outcomes after rotator cuff repair if a corticosteroid injection is administered within 6 months before surgery or if more than two injections are performed within 1 year before surgery [36]. This effect was also shown in an analysis of patients with lateral epicondylitis who underwent surgery; those receiving three or more injections were at highest risk (OR, 3.55) for requiring revision surgery [37]. Patients must be appropriately counseled regarding the potential greater need for revision surgery if a corticosteroid is administered within 6 months before surgical tendon repair.

Accelerated progression of osteoarthritis—Accelerated progression of osteoarthritis is not strictly defined but refers to an increase in joint space loss that is greater than expected to occur in a short time period. One study defined accelerated progression of osteoarthritis as a decrease in joint space by 2 mm over a 12-month period [38]. Research from the Osteoarthritis Initiative found that patients who received intraarticular knee joint injections of corticosteroid had 3.2 higher odds of showing radiographic progression of osteoarthritis than those who did not receive injections; the odds were higher at 4.67 in patients who received continuous injections [39]. Radiographic progression of osteoarthritis after injection has also been shown in the hip, with a study reporting radiographic progression of hip joint space narrowing in 44% of patients who received hip injections and were followed postprocedurally for a mean of 6 months [40].

Osseous injury—Corticosteroids induce osteocyte apoptosis, which predisposes to decreases in osseous vascularity and bone strength; these effects are particularly pronounced after intraarticular injection [41]. Indeed, subchondral insufficiency fracture, osteonecrosis, and rapid joint destruction have all been reported after intraarticular corticosteroid injection into various joints [42–44] (Figs. 4 and 5). A retrospective single-center study of 1000 patients with long-term follow-up after hip and knee corticosteroid injections showed a combined rate of osteonecrosis, insufficiency fractures, and rapid progressive osteoarthritis of 1% [45]. Such reports are controversial, and some experts believe that the reported risk of osseous injury after intraarticular injection is inflated (possibly attributed to the natural progression of osteoarthritis) and that a risk-benefit analysis would support this approach [46].

Open in Viewer

Fig. 4A —66-year-old man who presented with left groin pain.

A, Coronal STIR image through osseous pelvis shows large left hip joint effusion (asterisk).

Open in Viewer

Fig. 4B —66-year-old man who presented with left groin pain.

B, Sagittal T2-weighted image through left hip joint shows areas of full-thickness cartilage loss and subchondral marrow edema–like signal (arrows). Patient underwent same-day injection of left hip and iliopsoas bursa, receiving total of 60 mg of triamcinolone acetonide (Kenalog, Proficient Rx). Patient presented 3 months after injection with worsening hip pain.

Open in Viewer

Fig. 4C —66-year-old man who presented with left groin pain.

C, Anteroposterior radiograph of pelvis obtained 3 months after A and B shows severe joint space narrowing of left femoral articular surface with articular surface collapse and flattening (arrow).

Open in Viewer

Fig. 4D —66-year-old man who presented with left groin pain.

D, Coronal STIR (D) and coronal T1-weighted (E) images obtained same day as radiograph (C) show severe subchondral marrow edema spanning left hip and large joint effusion. Articular surface collapse (arrow, D), with areas of subchondral sclerosis (asterisk), is also present. Differential diagnosis includes rapid progression of osteoarthritis, avascular necrosis, and septic joint. Further workup identified no evidence of infection, and patient underwent left total hip arthroplasty.

Open in Viewer

Fig. 4E —66-year-old man who presented with left groin pain.

E, Coronal STIR (D) and coronal T1-weighted (E) images obtained same day as radiograph (C) show severe subchondral marrow edema spanning left hip and large joint effusion. Articular surface collapse (arrow, D), with areas of subchondral sclerosis (asterisk), is also present. Differential diagnosis includes rapid progression of osteoarthritis, avascular necrosis, and septic joint. Further workup identified no evidence of infection, and patient underwent left total hip arthroplasty.

Open in Viewer

Fig. 5A —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

A, Preinjection anteroposterior radiograph of pelvis shows severe left hip osteoarthritis with joint space narrowing, marginal spurring (arrow), and superolateral subluxation of femoral head. Intraarticular injection of 40 mg of triamcinolone acetonide (Kenalog, Proficient Rx) was performed after radiography.

Open in Viewer

Fig. 5B —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

B, Coronal STIR (B) and coronal T1-weighted (C) images of osseous pelvis obtained 2 weeks after injection show rapid destruction of femoral articular surface (asterisk) and acetabulum (arrow), with large joint effusion.

Open in Viewer

Fig. 5C —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

C, Coronal STIR (B) and coronal T1-weighted (C) images of osseous pelvis obtained 2 weeks after injection show rapid destruction of femoral articular surface (asterisk) and acetabulum (arrow), with large joint effusion.

Open in Viewer

Fig. 5D —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

D, Anteroposterior radiograph of pelvis obtained 3 weeks after injection shows progressing marked osseous resorption of femoral head, which now extends to intertrochanteric region (arrow), with cephalad migration and lateral displacement of proximal femur. No evidence of infection was found on joint aspiration or intraoperatively during total hip arthroplasty.

An expert panel review concluded that estimates of the preval‎ence of structural problems induced by intraarticular injections are highly variable and that, at present, no imaging, clinical, or laboratory markers are established for predicting subsequent osseous injury after an intraarticular injection [47]. The presence of acute joint pain that is not explained by radiographic or sonographic findings could prompt caution, with further workup obtained before potential corticosteroid injection. In this situation, MRI may help to detect osseous pathology, such as subchondral fracture, that would be a contraindication to injection; universal implementation of this approach may be cost-prohibitive given the likely low frequency of osseous injury on MRI [47]. The use of repeat radiographs before joint injection is also controversial. The panel concluded that large clinical trials are needed to determine the true incidence of these osseous complications and to formulate preintervention safety measures [47]. Although these osseous complications have a likely very low incidence, patients should be aware of the risk, and clinical judgment should be used to assess such complications' relevance for individual patients. For example, such complications may be considered more relevant in a patient with prior long-standing steroid use or a prior diagnosis of osteopenia or osteoporosis.

Systemic Immediate Side Effects

Adrenal suppression—Locally injected corticosteroid does not remain confined to the joint, and some amount is absorbed by the bloodstream. One study found that serum cortisol levels after intraarticular injection of methylprednisolone acetate were dose-dependent, peaked 2–12 hours after injection, and showed a significant downregulation for at least 1 week regardless of the injected dose [48]. A study of 10 healthy young male athletes reported adrenal suppression for up to 2 weeks after a single intra- or extraarticular injection [49]. A study found that injecting two joints simultaneously yields higher serum absorption than injecting the same total dose into one joint; this finding is theorized to be due to the exposure of a greater volume of synovium to corticosteroid, thus increasing the likelihood of systemic absorption [48]. This systemic absorption of corticosteroid accounts for many of the side effects occurring after injection, as well as the such side effects' similarity to the known side effects of exogenous steroid medication.

One study reported facial flushing (defined as erythema of the upper chest and face) after intraarticular corticosteroid injection in 40% of patients, with this side effect being unpleasant in 15% in cases, predominantly occurring in women, and being reproducible with subsequent injection [50]. Flushing is self-limiting, typically resolving within 3 days [51]. Associated systemic symptoms can include insomnia and headache.

Transient hypertension and postprandial hyperglycemia occur after intraarticular corticosteroid injection in patients both with and without diabetes and relate to suppression of the corticotropic axis; these effects may last for over 1 week [52]. These side effects' clinical implications depend on the patient's baseline condition.

Similar to the known side effects of oral or IV steroid use, psychiatric effects (e.g., altered mental status, visual hallucination) have been reported after intraarticular injection, typically occurring within 1 week after injection and occurring more commonly in older patients with decreased renal function [53, 54]. Prolonged hiccups have been reported after corticosteroid injection and are of indeterminate etiology [55].

Hypersensitivity reactions—The reported preval‎ence of hypersensitivity reactions after corticosteroid administration by any route is 0.3–0.5%; a common presentation of such reactions after musculoskeletal injection is delayed urticaria at the injection site [56, 57]. A case report described immediate anaphylaxis after intraarticular injection of triamcinolone acetonide, with confirmed hypersensitivity reaction by skin prick test [57]. Patients with frequent exposure to steroids are more likely to experience a hypersensitivity reaction. In addition, patients reporting a prior hypersensitivity reaction to corticosteroids should undergo workup for potential allergy to any alternative steroid preparation that is selected before it is administered [56]. Management of hypersensitivity reactions may include administration of oral or IV diphenhydramine hydrochloride, depending on the reaction's severity [56, 57].

Ophthalmic side effects—Central serous chorioretinopathy is a retinal detachment presenting with potentially permanent decreased vision that has been attributed to exogenously administered corticosteroids [58]. Any form of corticosteroid administration can precipitate glaucoma or increased intraocular pressure, the latter of which may lead to sudden-onset blurred vision or potentially irreversible vision loss [59]. These potential complications must be discussed before corticosteroid injection with patients who have known ophthalmic pathology and with patients reporting prior visual changes after steroid exposure.

Systemic Delayed Side Effects

As mentioned previously, intraarticular corticosteroid injection suppresses the hypothalamic-pituitary-adrenal axis, causing an at least temporary decrease in endogenous cortisol levels in 52.2% of patients [60]. This suppression can be potentially life-threatening in patients who are subject to an emotional or physical stress, such as illness or surgery, that may precipitate an adrenal crisis [6]. Most patients with adrenal insufficiency are not diagnosed until they are hospitalized for adrenal crisis, which has a reported mortality of 5% [61]. Prolonged adrenal insufficiency has been reported in some patients after repeated corticosteroid injections (i.e., iatrogenic Cushing syndrome) and may require exogenous supplementation to avoid an adrenal crisis [62, 63]. Although limited data describe prolonged adrenal suppression after musculoskeletal injection, a meta-analysis found that adrenal insufficiency may be present in 25.3% of patients 6 months after any corticosteroid administration [60]. Due to individual variation in glucocorticoid sensitivity, there is no corticosteroid dose, administration route, or treatment duration that prevents adrenal suppression [60, 64]. Furthermore, some suggest that no corticosteroids be administered for at least 4 weeks before planned surgery, to decrease the chance of adrenal crisis [28].

Osteoporosis is a known complication of long-term exogenous steroid use, attributed to steroid's negative effect on both bone quality and bone mass. A statistically significant decrease in bone mineral density of the femur has been reported after a single epidural steroid injection in postmenopausal women [65]. The literature suggests that the effect of epidural corticosteroid injection on bone mineral density is dose-dependent, whether from a single injection or multiple injections over time (i.e., cumulative dose), and may also predispose to increased vertebral fracture risk [28]. The incidence and impact of osteoporotic fractures from intra- and extraarticular corticosteroid injections for musculoskeletal indications are unknown. However, given a recognized impact of epidural steroid injections on bone mineral density, the radiologist should provide a recommendation to the primary care practitioner to consider initiating prophylactic osteoporotic treatment in postmenopausal women receiving frequent or multiple injections [28, 65]. Local Immediate Side Effects

Postinjection flare is an exacerbation of pain after injection that is significantly higher than the baseline preinjection pain. This flare is theorized to result from focal synovitis secondary to intracellular incorporation of the injected corticosteroid particulate microcrystals [9]. Postinjection flare is more common and of longer duration for particulate than nonparticulate corticosteroids [10]. The patient experience of postinjection flare can be clinically similar to the experience of crystalline-induced arthropathies, such as gout. Studies including injections of the hand, wrist, and shoulder report an incidence of postinjection flare of 32.8–35.3% [11, 12]. A study assessing flare reactions after injections performed to treat de Quer-vain tenosynovitis and trigger finger found that most flares occurred 1–2 days after injection, lasted a mean of 2 days, and did not impact long-term symptom resolution [11].

Local Delayed Side Effects

Skin hypopigmentation and atrophy—Several prospective studies describe skin hypopigmentation and atrophy, particularly after superficial peritendinous injections (e.g., to treat epicondylitis or de Quervain tenosynovitis) [13, 14] (Fig. 1). Heel pad atrophy or skin depigmentation is reported in 10% of patients who received corticosteroid injection of the plantar fascia [15]. Skin atrophy and hypopigmentation manifest within a few months of injection and typically resolve within 1 year [16]. Such dermal changes likely relate specifically to the use of particulate steroids, especially if the injectate migrates along the injection pathway within the soft tissues [17]. When superficial structures are being injected, the use of nonparticulate steroids should be considered, to minimize the potential for skin atrophy. Additionally, compression of the injection site after removal of the needle may minimize the risk of steroid migration along the needle track [18].

Open in Viewer

Fig. 1 —63-year-old woman with medial elbow pain who presented for sonographic eval‎uation of common flexor tendon. Patient had received two prior injections of 40 mg of triamcinolone acetonide (Kenalog, Proficient Rx) in medial elbow without imaging guidance within past 6 months. Photograph obtained at time of presentation for imaging shows discoloration and thinning of skin at medial elbow (dashed outline). Patient noticed this change several weeks after second injection.

Soft-tissue calcification—Soft-tissue calcification is a reported iatrogenic complication of musculoskeletal corticosteroid injection [19]. Explanatory theories include local tissue damage resulting in increased pH that precipitates calcium salt or possible calcific necrosis of collagen fibers [19]. Soft-tissue and tendinous calcifications are associated with increased pain scores and may be clinically significant [19]. Radiologists should recognize this potential corticosteroid-induced complication when assessing for changes in periarticular calcifications on postinjection imaging.

Infection—Intraarticular infection after corticosteroid injection is very rare, with an incidence thought to be less than 0.001% [20]. The incidence may be significantly less when using single-dose vials, as well as when using modern sterile technique, packaging, and cleaning methods [21]. A systematic review of case reports describes the incidences of soft-tissue infection and osteomyelitis of the underlying bone adjacent to the injection site [22]. The injection may also increase the risk of infection if the patient later undergoes surgery at the injection site. For example, studies have found that patients who received intraarticular injections of corticosteroid before a knee or hip replacement were more likely to experience a postoperative prosthetic joint infection than those who had never received an injection [23, 24]. A meta-analysis of eight studies on hip and knee joint replacements showed a 2.13 higher odds of deep infection in patients with than without a prior injection but no significant difference in superficial infection rate [25]. Another systematic review found that intraarticular injections performed within 3 months before knee arthroplasty were significantly associated with subsequent prosthetic joint infection [26]. Moreover, a study found that the use of preoperative corticosteroids was associated with a higher likelihood of superficial or deep infections in any surgical site, as well as an increased risk of respiratory, renal, and cardiac complications [27]. Given the available data, it may be advisable to delay any surgery for 1–3 months after corticosteroid injection to potentially mitigate the infection risk, particularly in patients with multiple comorbidities and chronic conditions and who are dependent on corticosteroids [28]. It is important to consider this potential infection risk before injecting patients who may be considering arthroplasty or other surgeries soon after corticosteroid injection and to counsel such patients accordingly.

Tendon, ligament, or fascial insufficiency or rupture—A rare but significant potential consequence of corticosteroid injections is tendon rupture (Figs. 2 and 3). Although its incidence is low, reported as approximately 0.1% in a meta-analysis, case reports describe such complications throughout the body [29, 30]. A review of the case reports estimated that among 95 postinjection complications, 53.7% were plantar fascia ruptures, and 26.3% were tendon ruptures involving the patellar, quadriceps, Achilles, or biceps tendon—tendons possibly already prone to increased biomechanical stress [31]. Injected corticosteroids are thought to locally decrease collagen production, which may limit potential tendon, ligament, or fascial healing and therefore increase predisposition to structural failure or rupture [32]. Even if rupture does not occur, corticosteroid use has been associated with ligament instability in the elbow, leading to refractory pain and disability [33]. A histologic study of the Achilles tendon in rabbits reported equivalent tendon weakening and loss of elasticity after either intratendinous or retrocalcaneal bursa injection, indicating that even peritendinous injections pose a risk of tendon rupture [34]. Tendon rupture may also be more strongly associated with triamcinolone acetonide than with other steroid preparations and may be associated with frequent repeated injections to a particular site [31]. Tendon rupture has been reported to occur anywhere from 2 days to 6 weeks after injection [29]. Although evidence-based consensus guidelines are lacking, some authors suggest that a maximum of three to five injections be performed at a single site, with a minimum of 2 weeks between injections, to minimize potential tendon rupture [31].

Open in Viewer

Fig. 2A —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

A, Gray-scale sonographic image obtained using 15-MHz linear probe shows intact gluteus medius insertion (dashed lines) on greater trochanter (asterisk) in long axis. Image shows thickened tendon with moderate tendinopathy but no high-grade tear. Patient underwent ultrasound-guided percutaneous tenotomy followed by injection of 20 mg of methylprednisolone acetate (Depo-Medrol, Pfizer) and 4 mL of 0.25% bupivacaine into greater trochanteric bursa. Twelve days after procedure, patient returned due to worsening pain, which was worse with weight-bearing and did not improve after course of NSAIDS and oral steroids. Patient reported no interval trauma. MRI of osseous pelvis was performed 18 days after procedure.

Open in Viewer

Fig. 2B —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

B, Coronal STIR (B) and sagittal T2-weighted (C) images, obtained 18 days after A, through greater trochanter show full-thickness tear of left gluteus medius tendon at myotendinous junction (arrow, B) with 2-cm tendon gap (arrow, C).

Open in Viewer

Fig. 2C —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

C, Coronal STIR (B) and sagittal T2-weighted (C) images, obtained 18 days after A, through greater trochanter show full-thickness tear of left gluteus medius tendon at myotendinous junction (arrow, B) with 2-cm tendon gap (arrow, C).

Open in Viewer

Fig. 3A —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

A, Gray-scale sonographic image of right hamstring tendon origin in long axis, obtained at initial presentation, shows marked thickening and hypoechoic degeneration of tendon (dashed lines) with areas of anechoic interstitial tearing (arrow). Ischial tuberosity is denoted by asterisk. Two days after third steroid injection, patient felt snap in buttock when performing household tasks.

Open in Viewer

Fig. 3B —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

B, Coronal STIR image of bony pelvis through ischial tuberosity shows near-complete rupture of right hamstring tendon origin with almost 3 cm of distal retraction (arrow).

Open in Viewer

Fig. 3C —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

C, Gray-scale sonographic image of left hamstring origin in long axis, immediately caudal to ischial tuberosity (asterisk), shows anechoic tendon gap (arrow).

Corticosteroid injections in the shoulder have been associated with a need for revision after rotator cuff repair, possibly due to underlying degeneration of the tendon quality as a result of the injection [35]. A systematic review found a nearly twofold increase in adverse outcomes after rotator cuff repair if a corticosteroid injection is administered within 6 months before surgery or if more than two injections are performed within 1 year before surgery [36]. This effect was also shown in an analysis of patients with lateral epicondylitis who underwent surgery; those receiving three or more injections were at highest risk (OR, 3.55) for requiring revision surgery [37]. Patients must be appropriately counseled regarding the potential greater need for revision surgery if a corticosteroid is administered within 6 months before surgical tendon repair.

Accelerated progression of osteoarthritis—Accelerated progression of osteoarthritis is not strictly defined but refers to an increase in joint space loss that is greater than expected to occur in a short time period. One study defined accelerated progression of osteoarthritis as a decrease in joint space by 2 mm over a 12-month period [38]. Research from the Osteoarthritis Initiative found that patients who received intraarticular knee joint injections of corticosteroid had 3.2 higher odds of showing radiographic progression of osteoarthritis than those who did not receive injections; the odds were higher at 4.67 in patients who received continuous injections [39]. Radiographic progression of osteoarthritis after injection has also been shown in the hip, with a study reporting radiographic progression of hip joint space narrowing in 44% of patients who received hip injections and were followed postprocedurally for a mean of 6 months [40].

Osseous injury—Corticosteroids induce osteocyte apoptosis, which predisposes to decreases in osseous vascularity and bone strength; these effects are particularly pronounced after intraarticular injection [41]. Indeed, subchondral insufficiency fracture, osteonecrosis, and rapid joint destruction have all been reported after intraarticular corticosteroid injection into various joints [42–44] (Figs. 4 and 5). A retrospective single-center study of 1000 patients with long-term follow-up after hip and knee corticosteroid injections showed a combined rate of osteonecrosis, insufficiency fractures, and rapid progressive osteoarthritis of 1% [45]. Such reports are controversial, and some experts believe that the reported risk of osseous injury after intraarticular injection is inflated (possibly attributed to the natural progression of osteoarthritis) and that a risk-benefit analysis would support this approach [46].

Open in Viewer

Fig. 4A —66-year-old man who presented with left groin pain.

A, Coronal STIR image through osseous pelvis shows large left hip joint effusion (asterisk).

Open in Viewer

Fig. 4B —66-year-old man who presented with left groin pain.

B, Sagittal T2-weighted image through left hip joint shows areas of full-thickness cartilage loss and subchondral marrow edema–like signal (arrows). Patient underwent same-day injection of left hip and iliopsoas bursa, receiving total of 60 mg of triamcinolone acetonide (Kenalog, Proficient Rx). Patient presented 3 months after injection with worsening hip pain.

Open in Viewer

Fig. 4C —66-year-old man who presented with left groin pain.

C, Anteroposterior radiograph of pelvis obtained 3 months after A and B shows severe joint space narrowing of left femoral articular surface with articular surface collapse and flattening (arrow).

Open in Viewer

Fig. 4D —66-year-old man who presented with left groin pain.

D, Coronal STIR (D) and coronal T1-weighted (E) images obtained same day as radiograph (C) show severe subchondral marrow edema spanning left hip and large joint effusion. Articular surface collapse (arrow, D), with areas of subchondral sclerosis (asterisk), is also present. Differential diagnosis includes rapid progression of osteoarthritis, avascular necrosis, and septic joint. Further workup identified no evidence of infection, and patient underwent left total hip arthroplasty.

Open in Viewer

Fig. 4E —66-year-old man who presented with left groin pain.

E, Coronal STIR (D) and coronal T1-weighted (E) images obtained same day as radiograph (C) show severe subchondral marrow edema spanning left hip and large joint effusion. Articular surface collapse (arrow, D), with areas of subchondral sclerosis (asterisk), is also present. Differential diagnosis includes rapid progression of osteoarthritis, avascular necrosis, and septic joint. Further workup identified no evidence of infection, and patient underwent left total hip arthroplasty.

Open in Viewer

Fig. 5A —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

A, Preinjection anteroposterior radiograph of pelvis shows severe left hip osteoarthritis with joint space narrowing, marginal spurring (arrow), and superolateral subluxation of femoral head. Intraarticular injection of 40 mg of triamcinolone acetonide (Kenalog, Proficient Rx) was performed after radiography.

Open in Viewer

Fig. 5B —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

B, Coronal STIR (B) and coronal T1-weighted (C) images of osseous pelvis obtained 2 weeks after injection show rapid destruction of femoral articular surface (asterisk) and acetabulum (arrow), with large joint effusion.

Open in Viewer

Fig. 5C —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

C, Coronal STIR (B) and coronal T1-weighted (C) images of osseous pelvis obtained 2 weeks after injection show rapid destruction of femoral articular surface (asterisk) and acetabulum (arrow), with large joint effusion.

Open in Viewer

Fig. 5D —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

D, Anteroposterior radiograph of pelvis obtained 3 weeks after injection shows progressing marked osseous resorption of femoral head, which now extends to intertrochanteric region (arrow), with cephalad migration and lateral displacement of proximal femur. No evidence of infection was found on joint aspiration or intraoperatively during total hip arthroplasty.

An expert panel review concluded that estimates of the preval‎ence of structural problems induced by intraarticular injections are highly variable and that, at present, no imaging, clinical, or laboratory markers are established for predicting subsequent osseous injury after an intraarticular injection [47]. The presence of acute joint pain that is not explained by radiographic or sonographic findings could prompt caution, with further workup obtained before potential corticosteroid injection. In this situation, MRI may help to detect osseous pathology, such as subchondral fracture, that would be a contraindication to injection; universal implementation of this approach may be cost-prohibitive given the likely low frequency of osseous injury on MRI [47]. The use of repeat radiographs before joint injection is also controversial. The panel concluded that large clinical trials are needed to determine the true incidence of these osseous complications and to formulate preintervention safety measures [47]. Although these osseous complications have a likely very low incidence, patients should be aware of the risk, and clinical judgment should be used to assess such complications' relevance for individual patients. For example, such complications may be considered more relevant in a patient with prior long-standing steroid use or a prior diagnosis of osteopenia or osteoporosis.

Systemic Immediate Side Effects

Adrenal suppression—Locally injected corticosteroid does not remain confined to the joint, and some amount is absorbed by the bloodstream. One study found that serum cortisol levels after intraarticular injection of methylprednisolone acetate were dose-dependent, peaked 2–12 hours after injection, and showed a significant downregulation for at least 1 week regardless of the injected dose [48]. A study of 10 healthy young male athletes reported adrenal suppression for up to 2 weeks after a single intra- or extraarticular injection [49]. A study found that injecting two joints simultaneously yields higher serum absorption than injecting the same total dose into one joint; this finding is theorized to be due to the exposure of a greater volume of synovium to corticosteroid, thus increasing the likelihood of systemic absorption [48]. This systemic absorption of corticosteroid accounts for many of the side effects occurring after injection, as well as the such side effects' similarity to the known side effects of exogenous steroid medication.

One study reported facial flushing (defined as erythema of the upper chest and face) after intraarticular corticosteroid injection in 40% of patients, with this side effect being unpleasant in 15% in cases, predominantly occurring in women, and being reproducible with subsequent injection [50]. Flushing is self-limiting, typically resolving within 3 days [51]. Associated systemic symptoms can include insomnia and headache.

Transient hypertension and postprandial hyperglycemia occur after intraarticular corticosteroid injection in patients both with and without diabetes and relate to suppression of the corticotropic axis; these effects may last for over 1 week [52]. These side effects' clinical implications depend on the patient's baseline condition.

Similar to the known side effects of oral or IV steroid use, psychiatric effects (e.g., altered mental status, visual hallucination) have been reported after intraarticular injection, typically occurring within 1 week after injection and occurring more commonly in older patients with decreased renal function [53, 54]. Prolonged hiccups have been reported after corticosteroid injection and are of indeterminate etiology [55].

Hypersensitivity reactions—The reported preval‎ence of hypersensitivity reactions after corticosteroid administration by any route is 0.3–0.5%; a common presentation of such reactions after musculoskeletal injection is delayed urticaria at the injection site [56, 57]. A case report described immediate anaphylaxis after intraarticular injection of triamcinolone acetonide, with confirmed hypersensitivity reaction by skin prick test [57]. Patients with frequent exposure to steroids are more likely to experience a hypersensitivity reaction. In addition, patients reporting a prior hypersensitivity reaction to corticosteroids should undergo workup for potential allergy to any alternative steroid preparation that is selected before it is administered [56]. Management of hypersensitivity reactions may include administration of oral or IV diphenhydramine hydrochloride, depending on the reaction's severity [56, 57].

Ophthalmic side effects—Central serous chorioretinopathy is a retinal detachment presenting with potentially permanent decreased vision that has been attributed to exogenously administered corticosteroids [58]. Any form of corticosteroid administration can precipitate glaucoma or increased intraocular pressure, the latter of which may lead to sudden-onset blurred vision or potentially irreversible vision loss [59]. These potential complications must be discussed before corticosteroid injection with patients who have known ophthalmic pathology and with patients reporting prior visual changes after steroid exposure.

Systemic Delayed Side Effects

As mentioned previously, intraarticular corticosteroid injection suppresses the hypothalamic-pituitary-adrenal axis, causing an at least temporary decrease in endogenous cortisol levels in 52.2% of patients [60]. This suppression can be potentially life-threatening in patients who are subject to an emotional or physical stress, such as illness or surgery, that may precipitate an adrenal crisis [6]. Most patients with adrenal insufficiency are not diagnosed until they are hospitalized for adrenal crisis, which has a reported mortality of 5% [61]. Prolonged adrenal insufficiency has been reported in some patients after repeated corticosteroid injections (i.e., iatrogenic Cushing syndrome) and may require exogenous supplementation to avoid an adrenal crisis [62, 63]. Although limited data describe prolonged adrenal suppression after musculoskeletal injection, a meta-analysis found that adrenal insufficiency may be present in 25.3% of patients 6 months after any corticosteroid administration [60]. Due to individual variation in glucocorticoid sensitivity, there is no corticosteroid dose, administration route, or treatment duration that prevents adrenal suppression [60, 64]. Furthermore, some suggest that no corticosteroids be administered for at least 4 weeks before planned surgery, to decrease the chance of adrenal crisis [28].

Osteoporosis is a known complication of long-term exogenous steroid use, attributed to steroid's negative effect on both bone quality and bone mass. A statistically significant decrease in bone mineral density of the femur has been reported after a single epidural steroid injection in postmenopausal women [65]. The literature suggests that the effect of epidural corticosteroid injection on bone mineral density is dose-dependent, whether from a single injection or multiple injections over time (i.e., cumulative dose), and may also predispose to increased vertebral fracture risk [28]. The incidence and impact of osteoporotic fractures from intra- and extraarticular corticosteroid injections for musculoskeletal indications are unknown. However, given a recognized impact of epidural steroid injections on bone mineral density, the radiologist should provide a recommendation to the primary care practitioner to consider initiating prophylactic osteoporotic treatment in postmenopausal women receiving frequent or multiple injections [28, 65]. Local Immediate Side Effects

Postinjection flare is an exacerbation of pain after injection that is significantly higher than the baseline preinjection pain. This flare is theorized to result from focal synovitis secondary to intracellular incorporation of the injected corticosteroid particulate microcrystals [9]. Postinjection flare is more common and of longer duration for particulate than nonparticulate corticosteroids [10]. The patient experience of postinjection flare can be clinically similar to the experience of crystalline-induced arthropathies, such as gout. Studies including injections of the hand, wrist, and shoulder report an incidence of postinjection flare of 32.8–35.3% [11, 12]. A study assessing flare reactions after injections performed to treat de Quer-vain tenosynovitis and trigger finger found that most flares occurred 1–2 days after injection, lasted a mean of 2 days, and did not impact long-term symptom resolution [11].

Local Delayed Side Effects

Skin hypopigmentation and atrophy—Several prospective studies describe skin hypopigmentation and atrophy, particularly after superficial peritendinous injections (e.g., to treat epicondylitis or de Quervain tenosynovitis) [13, 14] (Fig. 1). Heel pad atrophy or skin depigmentation is reported in 10% of patients who received corticosteroid injection of the plantar fascia [15]. Skin atrophy and hypopigmentation manifest within a few months of injection and typically resolve within 1 year [16]. Such dermal changes likely relate specifically to the use of particulate steroids, especially if the injectate migrates along the injection pathway within the soft tissues [17]. When superficial structures are being injected, the use of nonparticulate steroids should be considered, to minimize the potential for skin atrophy. Additionally, compression of the injection site after removal of the needle may minimize the risk of steroid migration along the needle track [18].

Open in Viewer

Fig. 1 —63-year-old woman with medial elbow pain who presented for sonographic eval‎uation of common flexor tendon. Patient had received two prior injections of 40 mg of triamcinolone acetonide (Kenalog, Proficient Rx) in medial elbow without imaging guidance within past 6 months. Photograph obtained at time of presentation for imaging shows discoloration and thinning of skin at medial elbow (dashed outline). Patient noticed this change several weeks after second injection.

Soft-tissue calcification—Soft-tissue calcification is a reported iatrogenic complication of musculoskeletal corticosteroid injection [19]. Explanatory theories include local tissue damage resulting in increased pH that precipitates calcium salt or possible calcific necrosis of collagen fibers [19]. Soft-tissue and tendinous calcifications are associated with increased pain scores and may be clinically significant [19]. Radiologists should recognize this potential corticosteroid-induced complication when assessing for changes in periarticular calcifications on postinjection imaging.

Infection—Intraarticular infection after corticosteroid injection is very rare, with an incidence thought to be less than 0.001% [20]. The incidence may be significantly less when using single-dose vials, as well as when using modern sterile technique, packaging, and cleaning methods [21]. A systematic review of case reports describes the incidences of soft-tissue infection and osteomyelitis of the underlying bone adjacent to the injection site [22]. The injection may also increase the risk of infection if the patient later undergoes surgery at the injection site. For example, studies have found that patients who received intraarticular injections of corticosteroid before a knee or hip replacement were more likely to experience a postoperative prosthetic joint infection than those who had never received an injection [23, 24]. A meta-analysis of eight studies on hip and knee joint replacements showed a 2.13 higher odds of deep infection in patients with than without a prior injection but no significant difference in superficial infection rate [25]. Another systematic review found that intraarticular injections performed within 3 months before knee arthroplasty were significantly associated with subsequent prosthetic joint infection [26]. Moreover, a study found that the use of preoperative corticosteroids was associated with a higher likelihood of superficial or deep infections in any surgical site, as well as an increased risk of respiratory, renal, and cardiac complications [27]. Given the available data, it may be advisable to delay any surgery for 1–3 months after corticosteroid injection to potentially mitigate the infection risk, particularly in patients with multiple comorbidities and chronic conditions and who are dependent on corticosteroids [28]. It is important to consider this potential infection risk before injecting patients who may be considering arthroplasty or other surgeries soon after corticosteroid injection and to counsel such patients accordingly.

Tendon, ligament, or fascial insufficiency or rupture—A rare but significant potential consequence of corticosteroid injections is tendon rupture (Figs. 2 and 3). Although its incidence is low, reported as approximately 0.1% in a meta-analysis, case reports describe such complications throughout the body [29, 30]. A review of the case reports estimated that among 95 postinjection complications, 53.7% were plantar fascia ruptures, and 26.3% were tendon ruptures involving the patellar, quadriceps, Achilles, or biceps tendon—tendons possibly already prone to increased biomechanical stress [31]. Injected corticosteroids are thought to locally decrease collagen production, which may limit potential tendon, ligament, or fascial healing and therefore increase predisposition to structural failure or rupture [32]. Even if rupture does not occur, corticosteroid use has been associated with ligament instability in the elbow, leading to refractory pain and disability [33]. A histologic study of the Achilles tendon in rabbits reported equivalent tendon weakening and loss of elasticity after either intratendinous or retrocalcaneal bursa injection, indicating that even peritendinous injections pose a risk of tendon rupture [34]. Tendon rupture may also be more strongly associated with triamcinolone acetonide than with other steroid preparations and may be associated with frequent repeated injections to a particular site [31]. Tendon rupture has been reported to occur anywhere from 2 days to 6 weeks after injection [29]. Although evidence-based consensus guidelines are lacking, some authors suggest that a maximum of three to five injections be performed at a single site, with a minimum of 2 weeks between injections, to minimize potential tendon rupture [31].

Open in Viewer

Fig. 2A —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

A, Gray-scale sonographic image obtained using 15-MHz linear probe shows intact gluteus medius insertion (dashed lines) on greater trochanter (asterisk) in long axis. Image shows thickened tendon with moderate tendinopathy but no high-grade tear. Patient underwent ultrasound-guided percutaneous tenotomy followed by injection of 20 mg of methylprednisolone acetate (Depo-Medrol, Pfizer) and 4 mL of 0.25% bupivacaine into greater trochanteric bursa. Twelve days after procedure, patient returned due to worsening pain, which was worse with weight-bearing and did not improve after course of NSAIDS and oral steroids. Patient reported no interval trauma. MRI of osseous pelvis was performed 18 days after procedure.

Open in Viewer

Fig. 2B —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

B, Coronal STIR (B) and sagittal T2-weighted (C) images, obtained 18 days after A, through greater trochanter show full-thickness tear of left gluteus medius tendon at myotendinous junction (arrow, B) with 2-cm tendon gap (arrow, C).

Open in Viewer

Fig. 2C —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

C, Coronal STIR (B) and sagittal T2-weighted (C) images, obtained 18 days after A, through greater trochanter show full-thickness tear of left gluteus medius tendon at myotendinous junction (arrow, B) with 2-cm tendon gap (arrow, C).

Open in Viewer

Fig. 3A —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

A, Gray-scale sonographic image of right hamstring tendon origin in long axis, obtained at initial presentation, shows marked thickening and hypoechoic degeneration of tendon (dashed lines) with areas of anechoic interstitial tearing (arrow). Ischial tuberosity is denoted by asterisk. Two days after third steroid injection, patient felt snap in buttock when performing household tasks.

Open in Viewer

Fig. 3B —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

B, Coronal STIR image of bony pelvis through ischial tuberosity shows near-complete rupture of right hamstring tendon origin with almost 3 cm of distal retraction (arrow).

Open in Viewer

Fig. 3C —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

C, Gray-scale sonographic image of left hamstring origin in long axis, immediately caudal to ischial tuberosity (asterisk), shows anechoic tendon gap (arrow).

Corticosteroid injections in the shoulder have been associated with a need for revision after rotator cuff repair, possibly due to underlying degeneration of the tendon quality as a result of the injection [35]. A systematic review found a nearly twofold increase in adverse outcomes after rotator cuff repair if a corticosteroid injection is administered within 6 months before surgery or if more than two injections are performed within 1 year before surgery [36]. This effect was also shown in an analysis of patients with lateral epicondylitis who underwent surgery; those receiving three or more injections were at highest risk (OR, 3.55) for requiring revision surgery [37]. Patients must be appropriately counseled regarding the potential greater need for revision surgery if a corticosteroid is administered within 6 months before surgical tendon repair.

Accelerated progression of osteoarthritis—Accelerated progression of osteoarthritis is not strictly defined but refers to an increase in joint space loss that is greater than expected to occur in a short time period. One study defined accelerated progression of osteoarthritis as a decrease in joint space by 2 mm over a 12-month period [38]. Research from the Osteoarthritis Initiative found that patients who received intraarticular knee joint injections of corticosteroid had 3.2 higher odds of showing radiographic progression of osteoarthritis than those who did not receive injections; the odds were higher at 4.67 in patients who received continuous injections [39]. Radiographic progression of osteoarthritis after injection has also been shown in the hip, with a study reporting radiographic progression of hip joint space narrowing in 44% of patients who received hip injections and were followed postprocedurally for a mean of 6 months [40].

Osseous injury—Corticosteroids induce osteocyte apoptosis, which predisposes to decreases in osseous vascularity and bone strength; these effects are particularly pronounced after intraarticular injection [41]. Indeed, subchondral insufficiency fracture, osteonecrosis, and rapid joint destruction have all been reported after intraarticular corticosteroid injection into various joints [42–44] (Figs. 4 and 5). A retrospective single-center study of 1000 patients with long-term follow-up after hip and knee corticosteroid injections showed a combined rate of osteonecrosis, insufficiency fractures, and rapid progressive osteoarthritis of 1% [45]. Such reports are controversial, and some experts believe that the reported risk of osseous injury after intraarticular injection is inflated (possibly attributed to the natural progression of osteoarthritis) and that a risk-benefit analysis would support this approach [46].

Open in Viewer

Fig. 4A —66-year-old man who presented with left groin pain.

A, Coronal STIR image through osseous pelvis shows large left hip joint effusion (asterisk).

Open in Viewer

Fig. 4B —66-year-old man who presented with left groin pain.

B, Sagittal T2-weighted image through left hip joint shows areas of full-thickness cartilage loss and subchondral marrow edema–like signal (arrows). Patient underwent same-day injection of left hip and iliopsoas bursa, receiving total of 60 mg of triamcinolone acetonide (Kenalog, Proficient Rx). Patient presented 3 months after injection with worsening hip pain.

Open in Viewer

Fig. 4C —66-year-old man who presented with left groin pain.

C, Anteroposterior radiograph of pelvis obtained 3 months after A and B shows severe joint space narrowing of left femoral articular surface with articular surface collapse and flattening (arrow).

Open in Viewer

Fig. 4D —66-year-old man who presented with left groin pain.

D, Coronal STIR (D) and coronal T1-weighted (E) images obtained same day as radiograph (C) show severe subchondral marrow edema spanning left hip and large joint effusion. Articular surface collapse (arrow, D), with areas of subchondral sclerosis (asterisk), is also present. Differential diagnosis includes rapid progression of osteoarthritis, avascular necrosis, and septic joint. Further workup identified no evidence of infection, and patient underwent left total hip arthroplasty.

Open in Viewer

Fig. 4E —66-year-old man who presented with left groin pain.

E, Coronal STIR (D) and coronal T1-weighted (E) images obtained same day as radiograph (C) show severe subchondral marrow edema spanning left hip and large joint effusion. Articular surface collapse (arrow, D), with areas of subchondral sclerosis (asterisk), is also present. Differential diagnosis includes rapid progression of osteoarthritis, avascular necrosis, and septic joint. Further workup identified no evidence of infection, and patient underwent left total hip arthroplasty.

Open in Viewer

Fig. 5A —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

A, Preinjection anteroposterior radiograph of pelvis shows severe left hip osteoarthritis with joint space narrowing, marginal spurring (arrow), and superolateral subluxation of femoral head. Intraarticular injection of 40 mg of triamcinolone acetonide (Kenalog, Proficient Rx) was performed after radiography.

Open in Viewer

Fig. 5B —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

B, Coronal STIR (B) and coronal T1-weighted (C) images of osseous pelvis obtained 2 weeks after injection show rapid destruction of femoral articular surface (asterisk) and acetabulum (arrow), with large joint effusion.

Open in Viewer

Fig. 5C —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

C, Coronal STIR (B) and coronal T1-weighted (C) images of osseous pelvis obtained 2 weeks after injection show rapid destruction of femoral articular surface (asterisk) and acetabulum (arrow), with large joint effusion.

Open in Viewer

Fig. 5D —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

D, Anteroposterior radiograph of pelvis obtained 3 weeks after injection shows progressing marked osseous resorption of femoral head, which now extends to intertrochanteric region (arrow), with cephalad migration and lateral displacement of proximal femur. No evidence of infection was found on joint aspiration or intraoperatively during total hip arthroplasty.

An expert panel review concluded that estimates of the preval‎ence of structural problems induced by intraarticular injections are highly variable and that, at present, no imaging, clinical, or laboratory markers are established for predicting subsequent osseous injury after an intraarticular injection [47]. The presence of acute joint pain that is not explained by radiographic or sonographic findings could prompt caution, with further workup obtained before potential corticosteroid injection. In this situation, MRI may help to detect osseous pathology, such as subchondral fracture, that would be a contraindication to injection; universal implementation of this approach may be cost-prohibitive given the likely low frequency of osseous injury on MRI [47]. The use of repeat radiographs before joint injection is also controversial. The panel concluded that large clinical trials are needed to determine the true incidence of these osseous complications and to formulate preintervention safety measures [47]. Although these osseous complications have a likely very low incidence, patients should be aware of the risk, and clinical judgment should be used to assess such complications' relevance for individual patients. For example, such complications may be considered more relevant in a patient with prior long-standing steroid use or a prior diagnosis of osteopenia or osteoporosis.

Systemic Immediate Side Effects

Adrenal suppression—Locally injected corticosteroid does not remain confined to the joint, and some amount is absorbed by the bloodstream. One study found that serum cortisol levels after intraarticular injection of methylprednisolone acetate were dose-dependent, peaked 2–12 hours after injection, and showed a significant downregulation for at least 1 week regardless of the injected dose [48]. A study of 10 healthy young male athletes reported adrenal suppression for up to 2 weeks after a single intra- or extraarticular injection [49]. A study found that injecting two joints simultaneously yields higher serum absorption than injecting the same total dose into one joint; this finding is theorized to be due to the exposure of a greater volume of synovium to corticosteroid, thus increasing the likelihood of systemic absorption [48]. This systemic absorption of corticosteroid accounts for many of the side effects occurring after injection, as well as the such side effects' similarity to the known side effects of exogenous steroid medication.

One study reported facial flushing (defined as erythema of the upper chest and face) after intraarticular corticosteroid injection in 40% of patients, with this side effect being unpleasant in 15% in cases, predominantly occurring in women, and being reproducible with subsequent injection [50]. Flushing is self-limiting, typically resolving within 3 days [51]. Associated systemic symptoms can include insomnia and headache.

Transient hypertension and postprandial hyperglycemia occur after intraarticular corticosteroid injection in patients both with and without diabetes and relate to suppression of the corticotropic axis; these effects may last for over 1 week [52]. These side effects' clinical implications depend on the patient's baseline condition.

Similar to the known side effects of oral or IV steroid use, psychiatric effects (e.g., altered mental status, visual hallucination) have been reported after intraarticular injection, typically occurring within 1 week after injection and occurring more commonly in older patients with decreased renal function [53, 54]. Prolonged hiccups have been reported after corticosteroid injection and are of indeterminate etiology [55].

Hypersensitivity reactions—The reported preval‎ence of hypersensitivity reactions after corticosteroid administration by any route is 0.3–0.5%; a common presentation of such reactions after musculoskeletal injection is delayed urticaria at the injection site [56, 57]. A case report described immediate anaphylaxis after intraarticular injection of triamcinolone acetonide, with confirmed hypersensitivity reaction by skin prick test [57]. Patients with frequent exposure to steroids are more likely to experience a hypersensitivity reaction. In addition, patients reporting a prior hypersensitivity reaction to corticosteroids should undergo workup for potential allergy to any alternative steroid preparation that is selected before it is administered [56]. Management of hypersensitivity reactions may include administration of oral or IV diphenhydramine hydrochloride, depending on the reaction's severity [56, 57].

Ophthalmic side effects—Central serous chorioretinopathy is a retinal detachment presenting with potentially permanent decreased vision that has been attributed to exogenously administered corticosteroids [58]. Any form of corticosteroid administration can precipitate glaucoma or increased intraocular pressure, the latter of which may lead to sudden-onset blurred vision or potentially irreversible vision loss [59]. These potential complications must be discussed before corticosteroid injection with patients who have known ophthalmic pathology and with patients reporting prior visual changes after steroid exposure.

Systemic Delayed Side Effects

As mentioned previously, intraarticular corticosteroid injection suppresses the hypothalamic-pituitary-adrenal axis, causing an at least temporary decrease in endogenous cortisol levels in 52.2% of patients [60]. This suppression can be potentially life-threatening in patients who are subject to an emotional or physical stress, such as illness or surgery, that may precipitate an adrenal crisis [6]. Most patients with adrenal insufficiency are not diagnosed until they are hospitalized for adrenal crisis, which has a reported mortality of 5% [61]. Prolonged adrenal insufficiency has been reported in some patients after repeated corticosteroid injections (i.e., iatrogenic Cushing syndrome) and may require exogenous supplementation to avoid an adrenal crisis [62, 63]. Although limited data describe prolonged adrenal suppression after musculoskeletal injection, a meta-analysis found that adrenal insufficiency may be present in 25.3% of patients 6 months after any corticosteroid administration [60]. Due to individual variation in glucocorticoid sensitivity, there is no corticosteroid dose, administration route, or treatment duration that prevents adrenal suppression [60, 64]. Furthermore, some suggest that no corticosteroids be administered for at least 4 weeks before planned surgery, to decrease the chance of adrenal crisis [28].

Osteoporosis is a known complication of long-term exogenous steroid use, attributed to steroid's negative effect on both bone quality and bone mass. A statistically significant decrease in bone mineral density of the femur has been reported after a single epidural steroid injection in postmenopausal women [65]. The literature suggests that the effect of epidural corticosteroid injection on bone mineral density is dose-dependent, whether from a single injection or multiple injections over time (i.e., cumulative dose), and may also predispose to increased vertebral fracture risk [28]. The incidence and impact of osteoporotic fractures from intra- and extraarticular corticosteroid injections for musculoskeletal indications are unknown. However, given a recognized impact of epidural steroid injections on bone mineral density, the radiologist should provide a recommendation to the primary care practitioner to consider initiating prophylactic osteoporotic treatment in postmenopausal women receiving frequent or multiple injections [28, 65]. Local Immediate Side Effects

Postinjection flare is an exacerbation of pain after injection that is significantly higher than the baseline preinjection pain. This flare is theorized to result from focal synovitis secondary to intracellular incorporation of the injected corticosteroid particulate microcrystals [9]. Postinjection flare is more common and of longer duration for particulate than nonparticulate corticosteroids [10]. The patient experience of postinjection flare can be clinically similar to the experience of crystalline-induced arthropathies, such as gout. Studies including injections of the hand, wrist, and shoulder report an incidence of postinjection flare of 32.8–35.3% [11, 12]. A study assessing flare reactions after injections performed to treat de Quer-vain tenosynovitis and trigger finger found that most flares occurred 1–2 days after injection, lasted a mean of 2 days, and did not impact long-term symptom resolution [11].

Local Delayed Side Effects

Skin hypopigmentation and atrophy—Several prospective studies describe skin hypopigmentation and atrophy, particularly after superficial peritendinous injections (e.g., to treat epicondylitis or de Quervain tenosynovitis) [13, 14] (Fig. 1). Heel pad atrophy or skin depigmentation is reported in 10% of patients who received corticosteroid injection of the plantar fascia [15]. Skin atrophy and hypopigmentation manifest within a few months of injection and typically resolve within 1 year [16]. Such dermal changes likely relate specifically to the use of particulate steroids, especially if the injectate migrates along the injection pathway within the soft tissues [17]. When superficial structures are being injected, the use of nonparticulate steroids should be considered, to minimize the potential for skin atrophy. Additionally, compression of the injection site after removal of the needle may minimize the risk of steroid migration along the needle track [18].

Open in Viewer

Fig. 1 —63-year-old woman with medial elbow pain who presented for sonographic eval‎uation of common flexor tendon. Patient had received two prior injections of 40 mg of triamcinolone acetonide (Kenalog, Proficient Rx) in medial elbow without imaging guidance within past 6 months. Photograph obtained at time of presentation for imaging shows discoloration and thinning of skin at medial elbow (dashed outline). Patient noticed this change several weeks after second injection.

Soft-tissue calcification—Soft-tissue calcification is a reported iatrogenic complication of musculoskeletal corticosteroid injection [19]. Explanatory theories include local tissue damage resulting in increased pH that precipitates calcium salt or possible calcific necrosis of collagen fibers [19]. Soft-tissue and tendinous calcifications are associated with increased pain scores and may be clinically significant [19]. Radiologists should recognize this potential corticosteroid-induced complication when assessing for changes in periarticular calcifications on postinjection imaging.

Infection—Intraarticular infection after corticosteroid injection is very rare, with an incidence thought to be less than 0.001% [20]. The incidence may be significantly less when using single-dose vials, as well as when using modern sterile technique, packaging, and cleaning methods [21]. A systematic review of case reports describes the incidences of soft-tissue infection and osteomyelitis of the underlying bone adjacent to the injection site [22]. The injection may also increase the risk of infection if the patient later undergoes surgery at the injection site. For example, studies have found that patients who received intraarticular injections of corticosteroid before a knee or hip replacement were more likely to experience a postoperative prosthetic joint infection than those who had never received an injection [23, 24]. A meta-analysis of eight studies on hip and knee joint replacements showed a 2.13 higher odds of deep infection in patients with than without a prior injection but no significant difference in superficial infection rate [25]. Another systematic review found that intraarticular injections performed within 3 months before knee arthroplasty were significantly associated with subsequent prosthetic joint infection [26]. Moreover, a study found that the use of preoperative corticosteroids was associated with a higher likelihood of superficial or deep infections in any surgical site, as well as an increased risk of respiratory, renal, and cardiac complications [27]. Given the available data, it may be advisable to delay any surgery for 1–3 months after corticosteroid injection to potentially mitigate the infection risk, particularly in patients with multiple comorbidities and chronic conditions and who are dependent on corticosteroids [28]. It is important to consider this potential infection risk before injecting patients who may be considering arthroplasty or other surgeries soon after corticosteroid injection and to counsel such patients accordingly.

Tendon, ligament, or fascial insufficiency or rupture—A rare but significant potential consequence of corticosteroid injections is tendon rupture (Figs. 2 and 3). Although its incidence is low, reported as approximately 0.1% in a meta-analysis, case reports describe such complications throughout the body [29, 30]. A review of the case reports estimated that among 95 postinjection complications, 53.7% were plantar fascia ruptures, and 26.3% were tendon ruptures involving the patellar, quadriceps, Achilles, or biceps tendon—tendons possibly already prone to increased biomechanical stress [31]. Injected corticosteroids are thought to locally decrease collagen production, which may limit potential tendon, ligament, or fascial healing and therefore increase predisposition to structural failure or rupture [32]. Even if rupture does not occur, corticosteroid use has been associated with ligament instability in the elbow, leading to refractory pain and disability [33]. A histologic study of the Achilles tendon in rabbits reported equivalent tendon weakening and loss of elasticity after either intratendinous or retrocalcaneal bursa injection, indicating that even peritendinous injections pose a risk of tendon rupture [34]. Tendon rupture may also be more strongly associated with triamcinolone acetonide than with other steroid preparations and may be associated with frequent repeated injections to a particular site [31]. Tendon rupture has been reported to occur anywhere from 2 days to 6 weeks after injection [29]. Although evidence-based consensus guidelines are lacking, some authors suggest that a maximum of three to five injections be performed at a single site, with a minimum of 2 weeks between injections, to minimize potential tendon rupture [31].

Open in Viewer

Fig. 2A —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

A, Gray-scale sonographic image obtained using 15-MHz linear probe shows intact gluteus medius insertion (dashed lines) on greater trochanter (asterisk) in long axis. Image shows thickened tendon with moderate tendinopathy but no high-grade tear. Patient underwent ultrasound-guided percutaneous tenotomy followed by injection of 20 mg of methylprednisolone acetate (Depo-Medrol, Pfizer) and 4 mL of 0.25% bupivacaine into greater trochanteric bursa. Twelve days after procedure, patient returned due to worsening pain, which was worse with weight-bearing and did not improve after course of NSAIDS and oral steroids. Patient reported no interval trauma. MRI of osseous pelvis was performed 18 days after procedure.

Open in Viewer

Fig. 2B —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

B, Coronal STIR (B) and sagittal T2-weighted (C) images, obtained 18 days after A, through greater trochanter show full-thickness tear of left gluteus medius tendon at myotendinous junction (arrow, B) with 2-cm tendon gap (arrow, C).

Open in Viewer

Fig. 2C —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

C, Coronal STIR (B) and sagittal T2-weighted (C) images, obtained 18 days after A, through greater trochanter show full-thickness tear of left gluteus medius tendon at myotendinous junction (arrow, B) with 2-cm tendon gap (arrow, C).

Open in Viewer

Fig. 3A —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

A, Gray-scale sonographic image of right hamstring tendon origin in long axis, obtained at initial presentation, shows marked thickening and hypoechoic degeneration of tendon (dashed lines) with areas of anechoic interstitial tearing (arrow). Ischial tuberosity is denoted by asterisk. Two days after third steroid injection, patient felt snap in buttock when performing household tasks.

Open in Viewer

Fig. 3B —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

B, Coronal STIR image of bony pelvis through ischial tuberosity shows near-complete rupture of right hamstring tendon origin with almost 3 cm of distal retraction (arrow).

Open in Viewer

Fig. 3C —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

C, Gray-scale sonographic image of left hamstring origin in long axis, immediately caudal to ischial tuberosity (asterisk), shows anechoic tendon gap (arrow).

Corticosteroid injections in the shoulder have been associated with a need for revision after rotator cuff repair, possibly due to underlying degeneration of the tendon quality as a result of the injection [35]. A systematic review found a nearly twofold increase in adverse outcomes after rotator cuff repair if a corticosteroid injection is administered within 6 months before surgery or if more than two injections are performed within 1 year before surgery [36]. This effect was also shown in an analysis of patients with lateral epicondylitis who underwent surgery; those receiving three or more injections were at highest risk (OR, 3.55) for requiring revision surgery [37]. Patients must be appropriately counseled regarding the potential greater need for revision surgery if a corticosteroid is administered within 6 months before surgical tendon repair.

Accelerated progression of osteoarthritis—Accelerated progression of osteoarthritis is not strictly defined but refers to an increase in joint space loss that is greater than expected to occur in a short time period. One study defined accelerated progression of osteoarthritis as a decrease in joint space by 2 mm over a 12-month period [38]. Research from the Osteoarthritis Initiative found that patients who received intraarticular knee joint injections of corticosteroid had 3.2 higher odds of showing radiographic progression of osteoarthritis than those who did not receive injections; the odds were higher at 4.67 in patients who received continuous injections [39]. Radiographic progression of osteoarthritis after injection has also been shown in the hip, with a study reporting radiographic progression of hip joint space narrowing in 44% of patients who received hip injections and were followed postprocedurally for a mean of 6 months [40].

Osseous injury—Corticosteroids induce osteocyte apoptosis, which predisposes to decreases in osseous vascularity and bone strength; these effects are particularly pronounced after intraarticular injection [41]. Indeed, subchondral insufficiency fracture, osteonecrosis, and rapid joint destruction have all been reported after intraarticular corticosteroid injection into various joints [42–44] (Figs. 4 and 5). A retrospective single-center study of 1000 patients with long-term follow-up after hip and knee corticosteroid injections showed a combined rate of osteonecrosis, insufficiency fractures, and rapid progressive osteoarthritis of 1% [45]. Such reports are controversial, and some experts believe that the reported risk of osseous injury after intraarticular injection is inflated (possibly attributed to the natural progression of osteoarthritis) and that a risk-benefit analysis would support this approach [46].

Open in Viewer

Fig. 4A —66-year-old man who presented with left groin pain.

A, Coronal STIR image through osseous pelvis shows large left hip joint effusion (asterisk).

Open in Viewer

Fig. 4B —66-year-old man who presented with left groin pain.

B, Sagittal T2-weighted image through left hip joint shows areas of full-thickness cartilage loss and subchondral marrow edema–like signal (arrows). Patient underwent same-day injection of left hip and iliopsoas bursa, receiving total of 60 mg of triamcinolone acetonide (Kenalog, Proficient Rx). Patient presented 3 months after injection with worsening hip pain.

Open in Viewer

Fig. 4C —66-year-old man who presented with left groin pain.

C, Anteroposterior radiograph of pelvis obtained 3 months after A and B shows severe joint space narrowing of left femoral articular surface with articular surface collapse and flattening (arrow).

Open in Viewer

Fig. 4D —66-year-old man who presented with left groin pain.

D, Coronal STIR (D) and coronal T1-weighted (E) images obtained same day as radiograph (C) show severe subchondral marrow edema spanning left hip and large joint effusion. Articular surface collapse (arrow, D), with areas of subchondral sclerosis (asterisk), is also present. Differential diagnosis includes rapid progression of osteoarthritis, avascular necrosis, and septic joint. Further workup identified no evidence of infection, and patient underwent left total hip arthroplasty.

Open in Viewer

Fig. 4E —66-year-old man who presented with left groin pain.

E, Coronal STIR (D) and coronal T1-weighted (E) images obtained same day as radiograph (C) show severe subchondral marrow edema spanning left hip and large joint effusion. Articular surface collapse (arrow, D), with areas of subchondral sclerosis (asterisk), is also present. Differential diagnosis includes rapid progression of osteoarthritis, avascular necrosis, and septic joint. Further workup identified no evidence of infection, and patient underwent left total hip arthroplasty.

Open in Viewer

Fig. 5A —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

A, Preinjection anteroposterior radiograph of pelvis shows severe left hip osteoarthritis with joint space narrowing, marginal spurring (arrow), and superolateral subluxation of femoral head. Intraarticular injection of 40 mg of triamcinolone acetonide (Kenalog, Proficient Rx) was performed after radiography.

Open in Viewer

Fig. 5B —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

B, Coronal STIR (B) and coronal T1-weighted (C) images of osseous pelvis obtained 2 weeks after injection show rapid destruction of femoral articular surface (asterisk) and acetabulum (arrow), with large joint effusion.

Open in Viewer

Fig. 5C —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

C, Coronal STIR (B) and coronal T1-weighted (C) images of osseous pelvis obtained 2 weeks after injection show rapid destruction of femoral articular surface (asterisk) and acetabulum (arrow), with large joint effusion.

Open in Viewer

Fig. 5D —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

D, Anteroposterior radiograph of pelvis obtained 3 weeks after injection shows progressing marked osseous resorption of femoral head, which now extends to intertrochanteric region (arrow), with cephalad migration and lateral displacement of proximal femur. No evidence of infection was found on joint aspiration or intraoperatively during total hip arthroplasty.

An expert panel review concluded that estimates of the preval‎ence of structural problems induced by intraarticular injections are highly variable and that, at present, no imaging, clinical, or laboratory markers are established for predicting subsequent osseous injury after an intraarticular injection [47]. The presence of acute joint pain that is not explained by radiographic or sonographic findings could prompt caution, with further workup obtained before potential corticosteroid injection. In this situation, MRI may help to detect osseous pathology, such as subchondral fracture, that would be a contraindication to injection; universal implementation of this approach may be cost-prohibitive given the likely low frequency of osseous injury on MRI [47]. The use of repeat radiographs before joint injection is also controversial. The panel concluded that large clinical trials are needed to determine the true incidence of these osseous complications and to formulate preintervention safety measures [47]. Although these osseous complications have a likely very low incidence, patients should be aware of the risk, and clinical judgment should be used to assess such complications' relevance for individual patients. For example, such complications may be considered more relevant in a patient with prior long-standing steroid use or a prior diagnosis of osteopenia or osteoporosis.

Systemic Immediate Side Effects

Adrenal suppression—Locally injected corticosteroid does not remain confined to the joint, and some amount is absorbed by the bloodstream. One study found that serum cortisol levels after intraarticular injection of methylprednisolone acetate were dose-dependent, peaked 2–12 hours after injection, and showed a significant downregulation for at least 1 week regardless of the injected dose [48]. A study of 10 healthy young male athletes reported adrenal suppression for up to 2 weeks after a single intra- or extraarticular injection [49]. A study found that injecting two joints simultaneously yields higher serum absorption than injecting the same total dose into one joint; this finding is theorized to be due to the exposure of a greater volume of synovium to corticosteroid, thus increasing the likelihood of systemic absorption [48]. This systemic absorption of corticosteroid accounts for many of the side effects occurring after injection, as well as the such side effects' similarity to the known side effects of exogenous steroid medication.

One study reported facial flushing (defined as erythema of the upper chest and face) after intraarticular corticosteroid injection in 40% of patients, with this side effect being unpleasant in 15% in cases, predominantly occurring in women, and being reproducible with subsequent injection [50]. Flushing is self-limiting, typically resolving within 3 days [51]. Associated systemic symptoms can include insomnia and headache.

Transient hypertension and postprandial hyperglycemia occur after intraarticular corticosteroid injection in patients both with and without diabetes and relate to suppression of the corticotropic axis; these effects may last for over 1 week [52]. These side effects' clinical implications depend on the patient's baseline condition.

Similar to the known side effects of oral or IV steroid use, psychiatric effects (e.g., altered mental status, visual hallucination) have been reported after intraarticular injection, typically occurring within 1 week after injection and occurring more commonly in older patients with decreased renal function [53, 54]. Prolonged hiccups have been reported after corticosteroid injection and are of indeterminate etiology [55].

Hypersensitivity reactions—The reported preval‎ence of hypersensitivity reactions after corticosteroid administration by any route is 0.3–0.5%; a common presentation of such reactions after musculoskeletal injection is delayed urticaria at the injection site [56, 57]. A case report described immediate anaphylaxis after intraarticular injection of triamcinolone acetonide, with confirmed hypersensitivity reaction by skin prick test [57]. Patients with frequent exposure to steroids are more likely to experience a hypersensitivity reaction. In addition, patients reporting a prior hypersensitivity reaction to corticosteroids should undergo workup for potential allergy to any alternative steroid preparation that is selected before it is administered [56]. Management of hypersensitivity reactions may include administration of oral or IV diphenhydramine hydrochloride, depending on the reaction's severity [56, 57].

Ophthalmic side effects—Central serous chorioretinopathy is a retinal detachment presenting with potentially permanent decreased vision that has been attributed to exogenously administered corticosteroids [58]. Any form of corticosteroid administration can precipitate glaucoma or increased intraocular pressure, the latter of which may lead to sudden-onset blurred vision or potentially irreversible vision loss [59]. These potential complications must be discussed before corticosteroid injection with patients who have known ophthalmic pathology and with patients reporting prior visual changes after steroid exposure.

Systemic Delayed Side Effects

As mentioned previously, intraarticular corticosteroid injection suppresses the hypothalamic-pituitary-adrenal axis, causing an at least temporary decrease in endogenous cortisol levels in 52.2% of patients [60]. This suppression can be potentially life-threatening in patients who are subject to an emotional or physical stress, such as illness or surgery, that may precipitate an adrenal crisis [6]. Most patients with adrenal insufficiency are not diagnosed until they are hospitalized for adrenal crisis, which has a reported mortality of 5% [61]. Prolonged adrenal insufficiency has been reported in some patients after repeated corticosteroid injections (i.e., iatrogenic Cushing syndrome) and may require exogenous supplementation to avoid an adrenal crisis [62, 63]. Although limited data describe prolonged adrenal suppression after musculoskeletal injection, a meta-analysis found that adrenal insufficiency may be present in 25.3% of patients 6 months after any corticosteroid administration [60]. Due to individual variation in glucocorticoid sensitivity, there is no corticosteroid dose, administration route, or treatment duration that prevents adrenal suppression [60, 64]. Furthermore, some suggest that no corticosteroids be administered for at least 4 weeks before planned surgery, to decrease the chance of adrenal crisis [28].

Osteoporosis is a known complication of long-term exogenous steroid use, attributed to steroid's negative effect on both bone quality and bone mass. A statistically significant decrease in bone mineral density of the femur has been reported after a single epidural steroid injection in postmenopausal women [65]. The literature suggests that the effect of epidural corticosteroid injection on bone mineral density is dose-dependent, whether from a single injection or multiple injections over time (i.e., cumulative dose), and may also predispose to increased vertebral fracture risk [28]. The incidence and impact of osteoporotic fractures from intra- and extraarticular corticosteroid injections for musculoskeletal indications are unknown. However, given a recognized impact of epidural steroid injections on bone mineral density, the radiologist should provide a recommendation to the primary care practitioner to consider initiating prophylactic osteoporotic treatment in postmenopausal women receiving frequent or multiple injections [28, 65]. Local Immediate Side Effects

Postinjection flare is an exacerbation of pain after injection that is significantly higher than the baseline preinjection pain. This flare is theorized to result from focal synovitis secondary to intracellular incorporation of the injected corticosteroid particulate microcrystals [9]. Postinjection flare is more common and of longer duration for particulate than nonparticulate corticosteroids [10]. The patient experience of postinjection flare can be clinically similar to the experience of crystalline-induced arthropathies, such as gout. Studies including injections of the hand, wrist, and shoulder report an incidence of postinjection flare of 32.8–35.3% [11, 12]. A study assessing flare reactions after injections performed to treat de Quer-vain tenosynovitis and trigger finger found that most flares occurred 1–2 days after injection, lasted a mean of 2 days, and did not impact long-term symptom resolution [11].

Local Delayed Side Effects

Skin hypopigmentation and atrophy—Several prospective studies describe skin hypopigmentation and atrophy, particularly after superficial peritendinous injections (e.g., to treat epicondylitis or de Quervain tenosynovitis) [13, 14] (Fig. 1). Heel pad atrophy or skin depigmentation is reported in 10% of patients who received corticosteroid injection of the plantar fascia [15]. Skin atrophy and hypopigmentation manifest within a few months of injection and typically resolve within 1 year [16]. Such dermal changes likely relate specifically to the use of particulate steroids, especially if the injectate migrates along the injection pathway within the soft tissues [17]. When superficial structures are being injected, the use of nonparticulate steroids should be considered, to minimize the potential for skin atrophy. Additionally, compression of the injection site after removal of the needle may minimize the risk of steroid migration along the needle track [18].

Open in Viewer

Fig. 1 —63-year-old woman with medial elbow pain who presented for sonographic eval‎uation of common flexor tendon. Patient had received two prior injections of 40 mg of triamcinolone acetonide (Kenalog, Proficient Rx) in medial elbow without imaging guidance within past 6 months. Photograph obtained at time of presentation for imaging shows discoloration and thinning of skin at medial elbow (dashed outline). Patient noticed this change several weeks after second injection.

Soft-tissue calcification—Soft-tissue calcification is a reported iatrogenic complication of musculoskeletal corticosteroid injection [19]. Explanatory theories include local tissue damage resulting in increased pH that precipitates calcium salt or possible calcific necrosis of collagen fibers [19]. Soft-tissue and tendinous calcifications are associated with increased pain scores and may be clinically significant [19]. Radiologists should recognize this potential corticosteroid-induced complication when assessing for changes in periarticular calcifications on postinjection imaging.

Infection—Intraarticular infection after corticosteroid injection is very rare, with an incidence thought to be less than 0.001% [20]. The incidence may be significantly less when using single-dose vials, as well as when using modern sterile technique, packaging, and cleaning methods [21]. A systematic review of case reports describes the incidences of soft-tissue infection and osteomyelitis of the underlying bone adjacent to the injection site [22]. The injection may also increase the risk of infection if the patient later undergoes surgery at the injection site. For example, studies have found that patients who received intraarticular injections of corticosteroid before a knee or hip replacement were more likely to experience a postoperative prosthetic joint infection than those who had never received an injection [23, 24]. A meta-analysis of eight studies on hip and knee joint replacements showed a 2.13 higher odds of deep infection in patients with than without a prior injection but no significant difference in superficial infection rate [25]. Another systematic review found that intraarticular injections performed within 3 months before knee arthroplasty were significantly associated with subsequent prosthetic joint infection [26]. Moreover, a study found that the use of preoperative corticosteroids was associated with a higher likelihood of superficial or deep infections in any surgical site, as well as an increased risk of respiratory, renal, and cardiac complications [27]. Given the available data, it may be advisable to delay any surgery for 1–3 months after corticosteroid injection to potentially mitigate the infection risk, particularly in patients with multiple comorbidities and chronic conditions and who are dependent on corticosteroids [28]. It is important to consider this potential infection risk before injecting patients who may be considering arthroplasty or other surgeries soon after corticosteroid injection and to counsel such patients accordingly.

Tendon, ligament, or fascial insufficiency or rupture—A rare but significant potential consequence of corticosteroid injections is tendon rupture (Figs. 2 and 3). Although its incidence is low, reported as approximately 0.1% in a meta-analysis, case reports describe such complications throughout the body [29, 30]. A review of the case reports estimated that among 95 postinjection complications, 53.7% were plantar fascia ruptures, and 26.3% were tendon ruptures involving the patellar, quadriceps, Achilles, or biceps tendon—tendons possibly already prone to increased biomechanical stress [31]. Injected corticosteroids are thought to locally decrease collagen production, which may limit potential tendon, ligament, or fascial healing and therefore increase predisposition to structural failure or rupture [32]. Even if rupture does not occur, corticosteroid use has been associated with ligament instability in the elbow, leading to refractory pain and disability [33]. A histologic study of the Achilles tendon in rabbits reported equivalent tendon weakening and loss of elasticity after either intratendinous or retrocalcaneal bursa injection, indicating that even peritendinous injections pose a risk of tendon rupture [34]. Tendon rupture may also be more strongly associated with triamcinolone acetonide than with other steroid preparations and may be associated with frequent repeated injections to a particular site [31]. Tendon rupture has been reported to occur anywhere from 2 days to 6 weeks after injection [29]. Although evidence-based consensus guidelines are lacking, some authors suggest that a maximum of three to five injections be performed at a single site, with a minimum of 2 weeks between injections, to minimize potential tendon rupture [31].

Open in Viewer

Fig. 2A —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

A, Gray-scale sonographic image obtained using 15-MHz linear probe shows intact gluteus medius insertion (dashed lines) on greater trochanter (asterisk) in long axis. Image shows thickened tendon with moderate tendinopathy but no high-grade tear. Patient underwent ultrasound-guided percutaneous tenotomy followed by injection of 20 mg of methylprednisolone acetate (Depo-Medrol, Pfizer) and 4 mL of 0.25% bupivacaine into greater trochanteric bursa. Twelve days after procedure, patient returned due to worsening pain, which was worse with weight-bearing and did not improve after course of NSAIDS and oral steroids. Patient reported no interval trauma. MRI of osseous pelvis was performed 18 days after procedure.

Open in Viewer

Fig. 2B —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

B, Coronal STIR (B) and sagittal T2-weighted (C) images, obtained 18 days after A, through greater trochanter show full-thickness tear of left gluteus medius tendon at myotendinous junction (arrow, B) with 2-cm tendon gap (arrow, C).

Open in Viewer

Fig. 2C —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

C, Coronal STIR (B) and sagittal T2-weighted (C) images, obtained 18 days after A, through greater trochanter show full-thickness tear of left gluteus medius tendon at myotendinous junction (arrow, B) with 2-cm tendon gap (arrow, C).

Open in Viewer

Fig. 3A —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

A, Gray-scale sonographic image of right hamstring tendon origin in long axis, obtained at initial presentation, shows marked thickening and hypoechoic degeneration of tendon (dashed lines) with areas of anechoic interstitial tearing (arrow). Ischial tuberosity is denoted by asterisk. Two days after third steroid injection, patient felt snap in buttock when performing household tasks.

Open in Viewer

Fig. 3B —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

B, Coronal STIR image of bony pelvis through ischial tuberosity shows near-complete rupture of right hamstring tendon origin with almost 3 cm of distal retraction (arrow).

Open in Viewer

Fig. 3C —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

C, Gray-scale sonographic image of left hamstring origin in long axis, immediately caudal to ischial tuberosity (asterisk), shows anechoic tendon gap (arrow).

Corticosteroid injections in the shoulder have been associated with a need for revision after rotator cuff repair, possibly due to underlying degeneration of the tendon quality as a result of the injection [35]. A systematic review found a nearly twofold increase in adverse outcomes after rotator cuff repair if a corticosteroid injection is administered within 6 months before surgery or if more than two injections are performed within 1 year before surgery [36]. This effect was also shown in an analysis of patients with lateral epicondylitis who underwent surgery; those receiving three or more injections were at highest risk (OR, 3.55) for requiring revision surgery [37]. Patients must be appropriately counseled regarding the potential greater need for revision surgery if a corticosteroid is administered within 6 months before surgical tendon repair.

Accelerated progression of osteoarthritis—Accelerated progression of osteoarthritis is not strictly defined but refers to an increase in joint space loss that is greater than expected to occur in a short time period. One study defined accelerated progression of osteoarthritis as a decrease in joint space by 2 mm over a 12-month period [38]. Research from the Osteoarthritis Initiative found that patients who received intraarticular knee joint injections of corticosteroid had 3.2 higher odds of showing radiographic progression of osteoarthritis than those who did not receive injections; the odds were higher at 4.67 in patients who received continuous injections [39]. Radiographic progression of osteoarthritis after injection has also been shown in the hip, with a study reporting radiographic progression of hip joint space narrowing in 44% of patients who received hip injections and were followed postprocedurally for a mean of 6 months [40].

Osseous injury—Corticosteroids induce osteocyte apoptosis, which predisposes to decreases in osseous vascularity and bone strength; these effects are particularly pronounced after intraarticular injection [41]. Indeed, subchondral insufficiency fracture, osteonecrosis, and rapid joint destruction have all been reported after intraarticular corticosteroid injection into various joints [42–44] (Figs. 4 and 5). A retrospective single-center study of 1000 patients with long-term follow-up after hip and knee corticosteroid injections showed a combined rate of osteonecrosis, insufficiency fractures, and rapid progressive osteoarthritis of 1% [45]. Such reports are controversial, and some experts believe that the reported risk of osseous injury after intraarticular injection is inflated (possibly attributed to the natural progression of osteoarthritis) and that a risk-benefit analysis would support this approach [46].

Open in Viewer

Fig. 4A —66-year-old man who presented with left groin pain.

A, Coronal STIR image through osseous pelvis shows large left hip joint effusion (asterisk).

Open in Viewer

Fig. 4B —66-year-old man who presented with left groin pain.

B, Sagittal T2-weighted image through left hip joint shows areas of full-thickness cartilage loss and subchondral marrow edema–like signal (arrows). Patient underwent same-day injection of left hip and iliopsoas bursa, receiving total of 60 mg of triamcinolone acetonide (Kenalog, Proficient Rx). Patient presented 3 months after injection with worsening hip pain.

Open in Viewer

Fig. 4C —66-year-old man who presented with left groin pain.

C, Anteroposterior radiograph of pelvis obtained 3 months after A and B shows severe joint space narrowing of left femoral articular surface with articular surface collapse and flattening (arrow).

Open in Viewer

Fig. 4D —66-year-old man who presented with left groin pain.

D, Coronal STIR (D) and coronal T1-weighted (E) images obtained same day as radiograph (C) show severe subchondral marrow edema spanning left hip and large joint effusion. Articular surface collapse (arrow, D), with areas of subchondral sclerosis (asterisk), is also present. Differential diagnosis includes rapid progression of osteoarthritis, avascular necrosis, and septic joint. Further workup identified no evidence of infection, and patient underwent left total hip arthroplasty.

Open in Viewer

Fig. 4E —66-year-old man who presented with left groin pain.

E, Coronal STIR (D) and coronal T1-weighted (E) images obtained same day as radiograph (C) show severe subchondral marrow edema spanning left hip and large joint effusion. Articular surface collapse (arrow, D), with areas of subchondral sclerosis (asterisk), is also present. Differential diagnosis includes rapid progression of osteoarthritis, avascular necrosis, and septic joint. Further workup identified no evidence of infection, and patient underwent left total hip arthroplasty.

Open in Viewer

Fig. 5A —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

A, Preinjection anteroposterior radiograph of pelvis shows severe left hip osteoarthritis with joint space narrowing, marginal spurring (arrow), and superolateral subluxation of femoral head. Intraarticular injection of 40 mg of triamcinolone acetonide (Kenalog, Proficient Rx) was performed after radiography.

Open in Viewer

Fig. 5B —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

B, Coronal STIR (B) and coronal T1-weighted (C) images of osseous pelvis obtained 2 weeks after injection show rapid destruction of femoral articular surface (asterisk) and acetabulum (arrow), with large joint effusion.

Open in Viewer

Fig. 5C —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

C, Coronal STIR (B) and coronal T1-weighted (C) images of osseous pelvis obtained 2 weeks after injection show rapid destruction of femoral articular surface (asterisk) and acetabulum (arrow), with large joint effusion.

Open in Viewer

Fig. 5D —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

D, Anteroposterior radiograph of pelvis obtained 3 weeks after injection shows progressing marked osseous resorption of femoral head, which now extends to intertrochanteric region (arrow), with cephalad migration and lateral displacement of proximal femur. No evidence of infection was found on joint aspiration or intraoperatively during total hip arthroplasty.

An expert panel review concluded that estimates of the preval‎ence of structural problems induced by intraarticular injections are highly variable and that, at present, no imaging, clinical, or laboratory markers are established for predicting subsequent osseous injury after an intraarticular injection [47]. The presence of acute joint pain that is not explained by radiographic or sonographic findings could prompt caution, with further workup obtained before potential corticosteroid injection. In this situation, MRI may help to detect osseous pathology, such as subchondral fracture, that would be a contraindication to injection; universal implementation of this approach may be cost-prohibitive given the likely low frequency of osseous injury on MRI [47]. The use of repeat radiographs before joint injection is also controversial. The panel concluded that large clinical trials are needed to determine the true incidence of these osseous complications and to formulate preintervention safety measures [47]. Although these osseous complications have a likely very low incidence, patients should be aware of the risk, and clinical judgment should be used to assess such complications' relevance for individual patients. For example, such complications may be considered more relevant in a patient with prior long-standing steroid use or a prior diagnosis of osteopenia or osteoporosis.

Systemic Immediate Side Effects

Adrenal suppression—Locally injected corticosteroid does not remain confined to the joint, and some amount is absorbed by the bloodstream. One study found that serum cortisol levels after intraarticular injection of methylprednisolone acetate were dose-dependent, peaked 2–12 hours after injection, and showed a significant downregulation for at least 1 week regardless of the injected dose [48]. A study of 10 healthy young male athletes reported adrenal suppression for up to 2 weeks after a single intra- or extraarticular injection [49]. A study found that injecting two joints simultaneously yields higher serum absorption than injecting the same total dose into one joint; this finding is theorized to be due to the exposure of a greater volume of synovium to corticosteroid, thus increasing the likelihood of systemic absorption [48]. This systemic absorption of corticosteroid accounts for many of the side effects occurring after injection, as well as the such side effects' similarity to the known side effects of exogenous steroid medication.

One study reported facial flushing (defined as erythema of the upper chest and face) after intraarticular corticosteroid injection in 40% of patients, with this side effect being unpleasant in 15% in cases, predominantly occurring in women, and being reproducible with subsequent injection [50]. Flushing is self-limiting, typically resolving within 3 days [51]. Associated systemic symptoms can include insomnia and headache.

Transient hypertension and postprandial hyperglycemia occur after intraarticular corticosteroid injection in patients both with and without diabetes and relate to suppression of the corticotropic axis; these effects may last for over 1 week [52]. These side effects' clinical implications depend on the patient's baseline condition.

Similar to the known side effects of oral or IV steroid use, psychiatric effects (e.g., altered mental status, visual hallucination) have been reported after intraarticular injection, typically occurring within 1 week after injection and occurring more commonly in older patients with decreased renal function [53, 54]. Prolonged hiccups have been reported after corticosteroid injection and are of indeterminate etiology [55].

Hypersensitivity reactions—The reported preval‎ence of hypersensitivity reactions after corticosteroid administration by any route is 0.3–0.5%; a common presentation of such reactions after musculoskeletal injection is delayed urticaria at the injection site [56, 57]. A case report described immediate anaphylaxis after intraarticular injection of triamcinolone acetonide, with confirmed hypersensitivity reaction by skin prick test [57]. Patients with frequent exposure to steroids are more likely to experience a hypersensitivity reaction. In addition, patients reporting a prior hypersensitivity reaction to corticosteroids should undergo workup for potential allergy to any alternative steroid preparation that is selected before it is administered [56]. Management of hypersensitivity reactions may include administration of oral or IV diphenhydramine hydrochloride, depending on the reaction's severity [56, 57].

Ophthalmic side effects—Central serous chorioretinopathy is a retinal detachment presenting with potentially permanent decreased vision that has been attributed to exogenously administered corticosteroids [58]. Any form of corticosteroid administration can precipitate glaucoma or increased intraocular pressure, the latter of which may lead to sudden-onset blurred vision or potentially irreversible vision loss [59]. These potential complications must be discussed before corticosteroid injection with patients who have known ophthalmic pathology and with patients reporting prior visual changes after steroid exposure.

Systemic Delayed Side Effects

As mentioned previously, intraarticular corticosteroid injection suppresses the hypothalamic-pituitary-adrenal axis, causing an at least temporary decrease in endogenous cortisol levels in 52.2% of patients [60]. This suppression can be potentially life-threatening in patients who are subject to an emotional or physical stress, such as illness or surgery, that may precipitate an adrenal crisis [6]. Most patients with adrenal insufficiency are not diagnosed until they are hospitalized for adrenal crisis, which has a reported mortality of 5% [61]. Prolonged adrenal insufficiency has been reported in some patients after repeated corticosteroid injections (i.e., iatrogenic Cushing syndrome) and may require exogenous supplementation to avoid an adrenal crisis [62, 63]. Although limited data describe prolonged adrenal suppression after musculoskeletal injection, a meta-analysis found that adrenal insufficiency may be present in 25.3% of patients 6 months after any corticosteroid administration [60]. Due to individual variation in glucocorticoid sensitivity, there is no corticosteroid dose, administration route, or treatment duration that prevents adrenal suppression [60, 64]. Furthermore, some suggest that no corticosteroids be administered for at least 4 weeks before planned surgery, to decrease the chance of adrenal crisis [28].

Osteoporosis is a known complication of long-term exogenous steroid use, attributed to steroid's negative effect on both bone quality and bone mass. A statistically significant decrease in bone mineral density of the femur has been reported after a single epidural steroid injection in postmenopausal women [65]. The literature suggests that the effect of epidural corticosteroid injection on bone mineral density is dose-dependent, whether from a single injection or multiple injections over time (i.e., cumulative dose), and may also predispose to increased vertebral fracture risk [28]. The incidence and impact of osteoporotic fractures from intra- and extraarticular corticosteroid injections for musculoskeletal indications are unknown. However, given a recognized impact of epidural steroid injections on bone mineral density, the radiologist should provide a recommendation to the primary care practitioner to consider initiating prophylactic osteoporotic treatment in postmenopausal women receiving frequent or multiple injections [28, 65]. Local Immediate Side Effects

Postinjection flare is an exacerbation of pain after injection that is significantly higher than the baseline preinjection pain. This flare is theorized to result from focal synovitis secondary to intracellular incorporation of the injected corticosteroid particulate microcrystals [9]. Postinjection flare is more common and of longer duration for particulate than nonparticulate corticosteroids [10]. The patient experience of postinjection flare can be clinically similar to the experience of crystalline-induced arthropathies, such as gout. Studies including injections of the hand, wrist, and shoulder report an incidence of postinjection flare of 32.8–35.3% [11, 12]. A study assessing flare reactions after injections performed to treat de Quer-vain tenosynovitis and trigger finger found that most flares occurred 1–2 days after injection, lasted a mean of 2 days, and did not impact long-term symptom resolution [11].

Local Delayed Side Effects

Skin hypopigmentation and atrophy—Several prospective studies describe skin hypopigmentation and atrophy, particularly after superficial peritendinous injections (e.g., to treat epicondylitis or de Quervain tenosynovitis) [13, 14] (Fig. 1). Heel pad atrophy or skin depigmentation is reported in 10% of patients who received corticosteroid injection of the plantar fascia [15]. Skin atrophy and hypopigmentation manifest within a few months of injection and typically resolve within 1 year [16]. Such dermal changes likely relate specifically to the use of particulate steroids, especially if the injectate migrates along the injection pathway within the soft tissues [17]. When superficial structures are being injected, the use of nonparticulate steroids should be considered, to minimize the potential for skin atrophy. Additionally, compression of the injection site after removal of the needle may minimize the risk of steroid migration along the needle track [18].

Open in Viewer

Fig. 1 —63-year-old woman with medial elbow pain who presented for sonographic eval‎uation of common flexor tendon. Patient had received two prior injections of 40 mg of triamcinolone acetonide (Kenalog, Proficient Rx) in medial elbow without imaging guidance within past 6 months. Photograph obtained at time of presentation for imaging shows discoloration and thinning of skin at medial elbow (dashed outline). Patient noticed this change several weeks after second injection.

Soft-tissue calcification—Soft-tissue calcification is a reported iatrogenic complication of musculoskeletal corticosteroid injection [19]. Explanatory theories include local tissue damage resulting in increased pH that precipitates calcium salt or possible calcific necrosis of collagen fibers [19]. Soft-tissue and tendinous calcifications are associated with increased pain scores and may be clinically significant [19]. Radiologists should recognize this potential corticosteroid-induced complication when assessing for changes in periarticular calcifications on postinjection imaging.

Infection—Intraarticular infection after corticosteroid injection is very rare, with an incidence thought to be less than 0.001% [20]. The incidence may be significantly less when using single-dose vials, as well as when using modern sterile technique, packaging, and cleaning methods [21]. A systematic review of case reports describes the incidences of soft-tissue infection and osteomyelitis of the underlying bone adjacent to the injection site [22]. The injection may also increase the risk of infection if the patient later undergoes surgery at the injection site. For example, studies have found that patients who received intraarticular injections of corticosteroid before a knee or hip replacement were more likely to experience a postoperative prosthetic joint infection than those who had never received an injection [23, 24]. A meta-analysis of eight studies on hip and knee joint replacements showed a 2.13 higher odds of deep infection in patients with than without a prior injection but no significant difference in superficial infection rate [25]. Another systematic review found that intraarticular injections performed within 3 months before knee arthroplasty were significantly associated with subsequent prosthetic joint infection [26]. Moreover, a study found that the use of preoperative corticosteroids was associated with a higher likelihood of superficial or deep infections in any surgical site, as well as an increased risk of respiratory, renal, and cardiac complications [27]. Given the available data, it may be advisable to delay any surgery for 1–3 months after corticosteroid injection to potentially mitigate the infection risk, particularly in patients with multiple comorbidities and chronic conditions and who are dependent on corticosteroids [28]. It is important to consider this potential infection risk before injecting patients who may be considering arthroplasty or other surgeries soon after corticosteroid injection and to counsel such patients accordingly.

Tendon, ligament, or fascial insufficiency or rupture—A rare but significant potential consequence of corticosteroid injections is tendon rupture (Figs. 2 and 3). Although its incidence is low, reported as approximately 0.1% in a meta-analysis, case reports describe such complications throughout the body [29, 30]. A review of the case reports estimated that among 95 postinjection complications, 53.7% were plantar fascia ruptures, and 26.3% were tendon ruptures involving the patellar, quadriceps, Achilles, or biceps tendon—tendons possibly already prone to increased biomechanical stress [31]. Injected corticosteroids are thought to locally decrease collagen production, which may limit potential tendon, ligament, or fascial healing and therefore increase predisposition to structural failure or rupture [32]. Even if rupture does not occur, corticosteroid use has been associated with ligament instability in the elbow, leading to refractory pain and disability [33]. A histologic study of the Achilles tendon in rabbits reported equivalent tendon weakening and loss of elasticity after either intratendinous or retrocalcaneal bursa injection, indicating that even peritendinous injections pose a risk of tendon rupture [34]. Tendon rupture may also be more strongly associated with triamcinolone acetonide than with other steroid preparations and may be associated with frequent repeated injections to a particular site [31]. Tendon rupture has been reported to occur anywhere from 2 days to 6 weeks after injection [29]. Although evidence-based consensus guidelines are lacking, some authors suggest that a maximum of three to five injections be performed at a single site, with a minimum of 2 weeks between injections, to minimize potential tendon rupture [31].

Open in Viewer

Fig. 2A —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

A, Gray-scale sonographic image obtained using 15-MHz linear probe shows intact gluteus medius insertion (dashed lines) on greater trochanter (asterisk) in long axis. Image shows thickened tendon with moderate tendinopathy but no high-grade tear. Patient underwent ultrasound-guided percutaneous tenotomy followed by injection of 20 mg of methylprednisolone acetate (Depo-Medrol, Pfizer) and 4 mL of 0.25% bupivacaine into greater trochanteric bursa. Twelve days after procedure, patient returned due to worsening pain, which was worse with weight-bearing and did not improve after course of NSAIDS and oral steroids. Patient reported no interval trauma. MRI of osseous pelvis was performed 18 days after procedure.

Open in Viewer

Fig. 2B —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

B, Coronal STIR (B) and sagittal T2-weighted (C) images, obtained 18 days after A, through greater trochanter show full-thickness tear of left gluteus medius tendon at myotendinous junction (arrow, B) with 2-cm tendon gap (arrow, C).

Open in Viewer

Fig. 2C —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

C, Coronal STIR (B) and sagittal T2-weighted (C) images, obtained 18 days after A, through greater trochanter show full-thickness tear of left gluteus medius tendon at myotendinous junction (arrow, B) with 2-cm tendon gap (arrow, C).

Open in Viewer

Fig. 3A —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

A, Gray-scale sonographic image of right hamstring tendon origin in long axis, obtained at initial presentation, shows marked thickening and hypoechoic degeneration of tendon (dashed lines) with areas of anechoic interstitial tearing (arrow). Ischial tuberosity is denoted by asterisk. Two days after third steroid injection, patient felt snap in buttock when performing household tasks.

Open in Viewer

Fig. 3B —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

B, Coronal STIR image of bony pelvis through ischial tuberosity shows near-complete rupture of right hamstring tendon origin with almost 3 cm of distal retraction (arrow).

Open in Viewer

Fig. 3C —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

C, Gray-scale sonographic image of left hamstring origin in long axis, immediately caudal to ischial tuberosity (asterisk), shows anechoic tendon gap (arrow).

Corticosteroid injections in the shoulder have been associated with a need for revision after rotator cuff repair, possibly due to underlying degeneration of the tendon quality as a result of the injection [35]. A systematic review found a nearly twofold increase in adverse outcomes after rotator cuff repair if a corticosteroid injection is administered within 6 months before surgery or if more than two injections are performed within 1 year before surgery [36]. This effect was also shown in an analysis of patients with lateral epicondylitis who underwent surgery; those receiving three or more injections were at highest risk (OR, 3.55) for requiring revision surgery [37]. Patients must be appropriately counseled regarding the potential greater need for revision surgery if a corticosteroid is administered within 6 months before surgical tendon repair.

Accelerated progression of osteoarthritis—Accelerated progression of osteoarthritis is not strictly defined but refers to an increase in joint space loss that is greater than expected to occur in a short time period. One study defined accelerated progression of osteoarthritis as a decrease in joint space by 2 mm over a 12-month period [38]. Research from the Osteoarthritis Initiative found that patients who received intraarticular knee joint injections of corticosteroid had 3.2 higher odds of showing radiographic progression of osteoarthritis than those who did not receive injections; the odds were higher at 4.67 in patients who received continuous injections [39]. Radiographic progression of osteoarthritis after injection has also been shown in the hip, with a study reporting radiographic progression of hip joint space narrowing in 44% of patients who received hip injections and were followed postprocedurally for a mean of 6 months [40].

Osseous injury—Corticosteroids induce osteocyte apoptosis, which predisposes to decreases in osseous vascularity and bone strength; these effects are particularly pronounced after intraarticular injection [41]. Indeed, subchondral insufficiency fracture, osteonecrosis, and rapid joint destruction have all been reported after intraarticular corticosteroid injection into various joints [42–44] (Figs. 4 and 5). A retrospective single-center study of 1000 patients with long-term follow-up after hip and knee corticosteroid injections showed a combined rate of osteonecrosis, insufficiency fractures, and rapid progressive osteoarthritis of 1% [45]. Such reports are controversial, and some experts believe that the reported risk of osseous injury after intraarticular injection is inflated (possibly attributed to the natural progression of osteoarthritis) and that a risk-benefit analysis would support this approach [46].

Open in Viewer

Fig. 4A —66-year-old man who presented with left groin pain.

A, Coronal STIR image through osseous pelvis shows large left hip joint effusion (asterisk).

Open in Viewer

Fig. 4B —66-year-old man who presented with left groin pain.

B, Sagittal T2-weighted image through left hip joint shows areas of full-thickness cartilage loss and subchondral marrow edema–like signal (arrows). Patient underwent same-day injection of left hip and iliopsoas bursa, receiving total of 60 mg of triamcinolone acetonide (Kenalog, Proficient Rx). Patient presented 3 months after injection with worsening hip pain.

Open in Viewer

Fig. 4C —66-year-old man who presented with left groin pain.

C, Anteroposterior radiograph of pelvis obtained 3 months after A and B shows severe joint space narrowing of left femoral articular surface with articular surface collapse and flattening (arrow).

Open in Viewer

Fig. 4D —66-year-old man who presented with left groin pain.

D, Coronal STIR (D) and coronal T1-weighted (E) images obtained same day as radiograph (C) show severe subchondral marrow edema spanning left hip and large joint effusion. Articular surface collapse (arrow, D), with areas of subchondral sclerosis (asterisk), is also present. Differential diagnosis includes rapid progression of osteoarthritis, avascular necrosis, and septic joint. Further workup identified no evidence of infection, and patient underwent left total hip arthroplasty.

Open in Viewer

Fig. 4E —66-year-old man who presented with left groin pain.

E, Coronal STIR (D) and coronal T1-weighted (E) images obtained same day as radiograph (C) show severe subchondral marrow edema spanning left hip and large joint effusion. Articular surface collapse (arrow, D), with areas of subchondral sclerosis (asterisk), is also present. Differential diagnosis includes rapid progression of osteoarthritis, avascular necrosis, and septic joint. Further workup identified no evidence of infection, and patient underwent left total hip arthroplasty.

Open in Viewer

Fig. 5A —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

A, Preinjection anteroposterior radiograph of pelvis shows severe left hip osteoarthritis with joint space narrowing, marginal spurring (arrow), and superolateral subluxation of femoral head. Intraarticular injection of 40 mg of triamcinolone acetonide (Kenalog, Proficient Rx) was performed after radiography.

Open in Viewer

Fig. 5B —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

B, Coronal STIR (B) and coronal T1-weighted (C) images of osseous pelvis obtained 2 weeks after injection show rapid destruction of femoral articular surface (asterisk) and acetabulum (arrow), with large joint effusion.

Open in Viewer

Fig. 5C —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

C, Coronal STIR (B) and coronal T1-weighted (C) images of osseous pelvis obtained 2 weeks after injection show rapid destruction of femoral articular surface (asterisk) and acetabulum (arrow), with large joint effusion.

Open in Viewer

Fig. 5D —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

D, Anteroposterior radiograph of pelvis obtained 3 weeks after injection shows progressing marked osseous resorption of femoral head, which now extends to intertrochanteric region (arrow), with cephalad migration and lateral displacement of proximal femur. No evidence of infection was found on joint aspiration or intraoperatively during total hip arthroplasty.

An expert panel review concluded that estimates of the preval‎ence of structural problems induced by intraarticular injections are highly variable and that, at present, no imaging, clinical, or laboratory markers are established for predicting subsequent osseous injury after an intraarticular injection [47]. The presence of acute joint pain that is not explained by radiographic or sonographic findings could prompt caution, with further workup obtained before potential corticosteroid injection. In this situation, MRI may help to detect osseous pathology, such as subchondral fracture, that would be a contraindication to injection; universal implementation of this approach may be cost-prohibitive given the likely low frequency of osseous injury on MRI [47]. The use of repeat radiographs before joint injection is also controversial. The panel concluded that large clinical trials are needed to determine the true incidence of these osseous complications and to formulate preintervention safety measures [47]. Although these osseous complications have a likely very low incidence, patients should be aware of the risk, and clinical judgment should be used to assess such complications' relevance for individual patients. For example, such complications may be considered more relevant in a patient with prior long-standing steroid use or a prior diagnosis of osteopenia or osteoporosis.

Systemic Immediate Side Effects

Adrenal suppression—Locally injected corticosteroid does not remain confined to the joint, and some amount is absorbed by the bloodstream. One study found that serum cortisol levels after intraarticular injection of methylprednisolone acetate were dose-dependent, peaked 2–12 hours after injection, and showed a significant downregulation for at least 1 week regardless of the injected dose [48]. A study of 10 healthy young male athletes reported adrenal suppression for up to 2 weeks after a single intra- or extraarticular injection [49]. A study found that injecting two joints simultaneously yields higher serum absorption than injecting the same total dose into one joint; this finding is theorized to be due to the exposure of a greater volume of synovium to corticosteroid, thus increasing the likelihood of systemic absorption [48]. This systemic absorption of corticosteroid accounts for many of the side effects occurring after injection, as well as the such side effects' similarity to the known side effects of exogenous steroid medication.

One study reported facial flushing (defined as erythema of the upper chest and face) after intraarticular corticosteroid injection in 40% of patients, with this side effect being unpleasant in 15% in cases, predominantly occurring in women, and being reproducible with subsequent injection [50]. Flushing is self-limiting, typically resolving within 3 days [51]. Associated systemic symptoms can include insomnia and headache.

Transient hypertension and postprandial hyperglycemia occur after intraarticular corticosteroid injection in patients both with and without diabetes and relate to suppression of the corticotropic axis; these effects may last for over 1 week [52]. These side effects' clinical implications depend on the patient's baseline condition.

Similar to the known side effects of oral or IV steroid use, psychiatric effects (e.g., altered mental status, visual hallucination) have been reported after intraarticular injection, typically occurring within 1 week after injection and occurring more commonly in older patients with decreased renal function [53, 54]. Prolonged hiccups have been reported after corticosteroid injection and are of indeterminate etiology [55].

Hypersensitivity reactions—The reported preval‎ence of hypersensitivity reactions after corticosteroid administration by any route is 0.3–0.5%; a common presentation of such reactions after musculoskeletal injection is delayed urticaria at the injection site [56, 57]. A case report described immediate anaphylaxis after intraarticular injection of triamcinolone acetonide, with confirmed hypersensitivity reaction by skin prick test [57]. Patients with frequent exposure to steroids are more likely to experience a hypersensitivity reaction. In addition, patients reporting a prior hypersensitivity reaction to corticosteroids should undergo workup for potential allergy to any alternative steroid preparation that is selected before it is administered [56]. Management of hypersensitivity reactions may include administration of oral or IV diphenhydramine hydrochloride, depending on the reaction's severity [56, 57].

Ophthalmic side effects—Central serous chorioretinopathy is a retinal detachment presenting with potentially permanent decreased vision that has been attributed to exogenously administered corticosteroids [58]. Any form of corticosteroid administration can precipitate glaucoma or increased intraocular pressure, the latter of which may lead to sudden-onset blurred vision or potentially irreversible vision loss [59]. These potential complications must be discussed before corticosteroid injection with patients who have known ophthalmic pathology and with patients reporting prior visual changes after steroid exposure.

Systemic Delayed Side Effects

As mentioned previously, intraarticular corticosteroid injection suppresses the hypothalamic-pituitary-adrenal axis, causing an at least temporary decrease in endogenous cortisol levels in 52.2% of patients [60]. This suppression can be potentially life-threatening in patients who are subject to an emotional or physical stress, such as illness or surgery, that may precipitate an adrenal crisis [6]. Most patients with adrenal insufficiency are not diagnosed until they are hospitalized for adrenal crisis, which has a reported mortality of 5% [61]. Prolonged adrenal insufficiency has been reported in some patients after repeated corticosteroid injections (i.e., iatrogenic Cushing syndrome) and may require exogenous supplementation to avoid an adrenal crisis [62, 63]. Although limited data describe prolonged adrenal suppression after musculoskeletal injection, a meta-analysis found that adrenal insufficiency may be present in 25.3% of patients 6 months after any corticosteroid administration [60]. Due to individual variation in glucocorticoid sensitivity, there is no corticosteroid dose, administration route, or treatment duration that prevents adrenal suppression [60, 64]. Furthermore, some suggest that no corticosteroids be administered for at least 4 weeks before planned surgery, to decrease the chance of adrenal crisis [28].

Osteoporosis is a known complication of long-term exogenous steroid use, attributed to steroid's negative effect on both bone quality and bone mass. A statistically significant decrease in bone mineral density of the femur has been reported after a single epidural steroid injection in postmenopausal women [65]. The literature suggests that the effect of epidural corticosteroid injection on bone mineral density is dose-dependent, whether from a single injection or multiple injections over time (i.e., cumulative dose), and may also predispose to increased vertebral fracture risk [28]. The incidence and impact of osteoporotic fractures from intra- and extraarticular corticosteroid injections for musculoskeletal indications are unknown. However, given a recognized impact of epidural steroid injections on bone mineral density, the radiologist should provide a recommendation to the primary care practitioner to consider initiating prophylactic osteoporotic treatment in postmenopausal women receiving frequent or multiple injections [28, 65]. Local Immediate Side Effects

Postinjection flare is an exacerbation of pain after injection that is significantly higher than the baseline preinjection pain. This flare is theorized to result from focal synovitis secondary to intracellular incorporation of the injected corticosteroid particulate microcrystals [9]. Postinjection flare is more common and of longer duration for particulate than nonparticulate corticosteroids [10]. The patient experience of postinjection flare can be clinically similar to the experience of crystalline-induced arthropathies, such as gout. Studies including injections of the hand, wrist, and shoulder report an incidence of postinjection flare of 32.8–35.3% [11, 12]. A study assessing flare reactions after injections performed to treat de Quer-vain tenosynovitis and trigger finger found that most flares occurred 1–2 days after injection, lasted a mean of 2 days, and did not impact long-term symptom resolution [11].

Local Delayed Side Effects

Skin hypopigmentation and atrophy—Several prospective studies describe skin hypopigmentation and atrophy, particularly after superficial peritendinous injections (e.g., to treat epicondylitis or de Quervain tenosynovitis) [13, 14] (Fig. 1). Heel pad atrophy or skin depigmentation is reported in 10% of patients who received corticosteroid injection of the plantar fascia [15]. Skin atrophy and hypopigmentation manifest within a few months of injection and typically resolve within 1 year [16]. Such dermal changes likely relate specifically to the use of particulate steroids, especially if the injectate migrates along the injection pathway within the soft tissues [17]. When superficial structures are being injected, the use of nonparticulate steroids should be considered, to minimize the potential for skin atrophy. Additionally, compression of the injection site after removal of the needle may minimize the risk of steroid migration along the needle track [18].

Open in Viewer

Fig. 1 —63-year-old woman with medial elbow pain who presented for sonographic eval‎uation of common flexor tendon. Patient had received two prior injections of 40 mg of triamcinolone acetonide (Kenalog, Proficient Rx) in medial elbow without imaging guidance within past 6 months. Photograph obtained at time of presentation for imaging shows discoloration and thinning of skin at medial elbow (dashed outline). Patient noticed this change several weeks after second injection.

Soft-tissue calcification—Soft-tissue calcification is a reported iatrogenic complication of musculoskeletal corticosteroid injection [19]. Explanatory theories include local tissue damage resulting in increased pH that precipitates calcium salt or possible calcific necrosis of collagen fibers [19]. Soft-tissue and tendinous calcifications are associated with increased pain scores and may be clinically significant [19]. Radiologists should recognize this potential corticosteroid-induced complication when assessing for changes in periarticular calcifications on postinjection imaging.

Infection—Intraarticular infection after corticosteroid injection is very rare, with an incidence thought to be less than 0.001% [20]. The incidence may be significantly less when using single-dose vials, as well as when using modern sterile technique, packaging, and cleaning methods [21]. A systematic review of case reports describes the incidences of soft-tissue infection and osteomyelitis of the underlying bone adjacent to the injection site [22]. The injection may also increase the risk of infection if the patient later undergoes surgery at the injection site. For example, studies have found that patients who received intraarticular injections of corticosteroid before a knee or hip replacement were more likely to experience a postoperative prosthetic joint infection than those who had never received an injection [23, 24]. A meta-analysis of eight studies on hip and knee joint replacements showed a 2.13 higher odds of deep infection in patients with than without a prior injection but no significant difference in superficial infection rate [25]. Another systematic review found that intraarticular injections performed within 3 months before knee arthroplasty were significantly associated with subsequent prosthetic joint infection [26]. Moreover, a study found that the use of preoperative corticosteroids was associated with a higher likelihood of superficial or deep infections in any surgical site, as well as an increased risk of respiratory, renal, and cardiac complications [27]. Given the available data, it may be advisable to delay any surgery for 1–3 months after corticosteroid injection to potentially mitigate the infection risk, particularly in patients with multiple comorbidities and chronic conditions and who are dependent on corticosteroids [28]. It is important to consider this potential infection risk before injecting patients who may be considering arthroplasty or other surgeries soon after corticosteroid injection and to counsel such patients accordingly.

Tendon, ligament, or fascial insufficiency or rupture—A rare but significant potential consequence of corticosteroid injections is tendon rupture (Figs. 2 and 3). Although its incidence is low, reported as approximately 0.1% in a meta-analysis, case reports describe such complications throughout the body [29, 30]. A review of the case reports estimated that among 95 postinjection complications, 53.7% were plantar fascia ruptures, and 26.3% were tendon ruptures involving the patellar, quadriceps, Achilles, or biceps tendon—tendons possibly already prone to increased biomechanical stress [31]. Injected corticosteroids are thought to locally decrease collagen production, which may limit potential tendon, ligament, or fascial healing and therefore increase predisposition to structural failure or rupture [32]. Even if rupture does not occur, corticosteroid use has been associated with ligament instability in the elbow, leading to refractory pain and disability [33]. A histologic study of the Achilles tendon in rabbits reported equivalent tendon weakening and loss of elasticity after either intratendinous or retrocalcaneal bursa injection, indicating that even peritendinous injections pose a risk of tendon rupture [34]. Tendon rupture may also be more strongly associated with triamcinolone acetonide than with other steroid preparations and may be associated with frequent repeated injections to a particular site [31]. Tendon rupture has been reported to occur anywhere from 2 days to 6 weeks after injection [29]. Although evidence-based consensus guidelines are lacking, some authors suggest that a maximum of three to five injections be performed at a single site, with a minimum of 2 weeks between injections, to minimize potential tendon rupture [31].

Open in Viewer

Fig. 2A —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

A, Gray-scale sonographic image obtained using 15-MHz linear probe shows intact gluteus medius insertion (dashed lines) on greater trochanter (asterisk) in long axis. Image shows thickened tendon with moderate tendinopathy but no high-grade tear. Patient underwent ultrasound-guided percutaneous tenotomy followed by injection of 20 mg of methylprednisolone acetate (Depo-Medrol, Pfizer) and 4 mL of 0.25% bupivacaine into greater trochanteric bursa. Twelve days after procedure, patient returned due to worsening pain, which was worse with weight-bearing and did not improve after course of NSAIDS and oral steroids. Patient reported no interval trauma. MRI of osseous pelvis was performed 18 days after procedure.

Open in Viewer

Fig. 2B —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

B, Coronal STIR (B) and sagittal T2-weighted (C) images, obtained 18 days after A, through greater trochanter show full-thickness tear of left gluteus medius tendon at myotendinous junction (arrow, B) with 2-cm tendon gap (arrow, C).

Open in Viewer

Fig. 2C —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

C, Coronal STIR (B) and sagittal T2-weighted (C) images, obtained 18 days after A, through greater trochanter show full-thickness tear of left gluteus medius tendon at myotendinous junction (arrow, B) with 2-cm tendon gap (arrow, C).

Open in Viewer

Fig. 3A —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

A, Gray-scale sonographic image of right hamstring tendon origin in long axis, obtained at initial presentation, shows marked thickening and hypoechoic degeneration of tendon (dashed lines) with areas of anechoic interstitial tearing (arrow). Ischial tuberosity is denoted by asterisk. Two days after third steroid injection, patient felt snap in buttock when performing household tasks.

Open in Viewer

Fig. 3B —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

B, Coronal STIR image of bony pelvis through ischial tuberosity shows near-complete rupture of right hamstring tendon origin with almost 3 cm of distal retraction (arrow).

Open in Viewer

Fig. 3C —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

C, Gray-scale sonographic image of left hamstring origin in long axis, immediately caudal to ischial tuberosity (asterisk), shows anechoic tendon gap (arrow).

Corticosteroid injections in the shoulder have been associated with a need for revision after rotator cuff repair, possibly due to underlying degeneration of the tendon quality as a result of the injection [35]. A systematic review found a nearly twofold increase in adverse outcomes after rotator cuff repair if a corticosteroid injection is administered within 6 months before surgery or if more than two injections are performed within 1 year before surgery [36]. This effect was also shown in an analysis of patients with lateral epicondylitis who underwent surgery; those receiving three or more injections were at highest risk (OR, 3.55) for requiring revision surgery [37]. Patients must be appropriately counseled regarding the potential greater need for revision surgery if a corticosteroid is administered within 6 months before surgical tendon repair.

Accelerated progression of osteoarthritis—Accelerated progression of osteoarthritis is not strictly defined but refers to an increase in joint space loss that is greater than expected to occur in a short time period. One study defined accelerated progression of osteoarthritis as a decrease in joint space by 2 mm over a 12-month period [38]. Research from the Osteoarthritis Initiative found that patients who received intraarticular knee joint injections of corticosteroid had 3.2 higher odds of showing radiographic progression of osteoarthritis than those who did not receive injections; the odds were higher at 4.67 in patients who received continuous injections [39]. Radiographic progression of osteoarthritis after injection has also been shown in the hip, with a study reporting radiographic progression of hip joint space narrowing in 44% of patients who received hip injections and were followed postprocedurally for a mean of 6 months [40].

Osseous injury—Corticosteroids induce osteocyte apoptosis, which predisposes to decreases in osseous vascularity and bone strength; these effects are particularly pronounced after intraarticular injection [41]. Indeed, subchondral insufficiency fracture, osteonecrosis, and rapid joint destruction have all been reported after intraarticular corticosteroid injection into various joints [42–44] (Figs. 4 and 5). A retrospective single-center study of 1000 patients with long-term follow-up after hip and knee corticosteroid injections showed a combined rate of osteonecrosis, insufficiency fractures, and rapid progressive osteoarthritis of 1% [45]. Such reports are controversial, and some experts believe that the reported risk of osseous injury after intraarticular injection is inflated (possibly attributed to the natural progression of osteoarthritis) and that a risk-benefit analysis would support this approach [46].

Open in Viewer

Fig. 4A —66-year-old man who presented with left groin pain.

A, Coronal STIR image through osseous pelvis shows large left hip joint effusion (asterisk).

Open in Viewer

Fig. 4B —66-year-old man who presented with left groin pain.

B, Sagittal T2-weighted image through left hip joint shows areas of full-thickness cartilage loss and subchondral marrow edema–like signal (arrows). Patient underwent same-day injection of left hip and iliopsoas bursa, receiving total of 60 mg of triamcinolone acetonide (Kenalog, Proficient Rx). Patient presented 3 months after injection with worsening hip pain.

Open in Viewer

Fig. 4C —66-year-old man who presented with left groin pain.

C, Anteroposterior radiograph of pelvis obtained 3 months after A and B shows severe joint space narrowing of left femoral articular surface with articular surface collapse and flattening (arrow).

Open in Viewer

Fig. 4D —66-year-old man who presented with left groin pain.

D, Coronal STIR (D) and coronal T1-weighted (E) images obtained same day as radiograph (C) show severe subchondral marrow edema spanning left hip and large joint effusion. Articular surface collapse (arrow, D), with areas of subchondral sclerosis (asterisk), is also present. Differential diagnosis includes rapid progression of osteoarthritis, avascular necrosis, and septic joint. Further workup identified no evidence of infection, and patient underwent left total hip arthroplasty.

Open in Viewer

Fig. 4E —66-year-old man who presented with left groin pain.

E, Coronal STIR (D) and coronal T1-weighted (E) images obtained same day as radiograph (C) show severe subchondral marrow edema spanning left hip and large joint effusion. Articular surface collapse (arrow, D), with areas of subchondral sclerosis (asterisk), is also present. Differential diagnosis includes rapid progression of osteoarthritis, avascular necrosis, and septic joint. Further workup identified no evidence of infection, and patient underwent left total hip arthroplasty.

Open in Viewer

Fig. 5A —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

A, Preinjection anteroposterior radiograph of pelvis shows severe left hip osteoarthritis with joint space narrowing, marginal spurring (arrow), and superolateral subluxation of femoral head. Intraarticular injection of 40 mg of triamcinolone acetonide (Kenalog, Proficient Rx) was performed after radiography.

Open in Viewer

Fig. 5B —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

B, Coronal STIR (B) and coronal T1-weighted (C) images of osseous pelvis obtained 2 weeks after injection show rapid destruction of femoral articular surface (asterisk) and acetabulum (arrow), with large joint effusion.

Open in Viewer

Fig. 5C —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

C, Coronal STIR (B) and coronal T1-weighted (C) images of osseous pelvis obtained 2 weeks after injection show rapid destruction of femoral articular surface (asterisk) and acetabulum (arrow), with large joint effusion.

Open in Viewer

Fig. 5D —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

D, Anteroposterior radiograph of pelvis obtained 3 weeks after injection shows progressing marked osseous resorption of femoral head, which now extends to intertrochanteric region (arrow), with cephalad migration and lateral displacement of proximal femur. No evidence of infection was found on joint aspiration or intraoperatively during total hip arthroplasty.

An expert panel review concluded that estimates of the preval‎ence of structural problems induced by intraarticular injections are highly variable and that, at present, no imaging, clinical, or laboratory markers are established for predicting subsequent osseous injury after an intraarticular injection [47]. The presence of acute joint pain that is not explained by radiographic or sonographic findings could prompt caution, with further workup obtained before potential corticosteroid injection. In this situation, MRI may help to detect osseous pathology, such as subchondral fracture, that would be a contraindication to injection; universal implementation of this approach may be cost-prohibitive given the likely low frequency of osseous injury on MRI [47]. The use of repeat radiographs before joint injection is also controversial. The panel concluded that large clinical trials are needed to determine the true incidence of these osseous complications and to formulate preintervention safety measures [47]. Although these osseous complications have a likely very low incidence, patients should be aware of the risk, and clinical judgment should be used to assess such complications' relevance for individual patients. For example, such complications may be considered more relevant in a patient with prior long-standing steroid use or a prior diagnosis of osteopenia or osteoporosis.

Systemic Immediate Side Effects

Adrenal suppression—Locally injected corticosteroid does not remain confined to the joint, and some amount is absorbed by the bloodstream. One study found that serum cortisol levels after intraarticular injection of methylprednisolone acetate were dose-dependent, peaked 2–12 hours after injection, and showed a significant downregulation for at least 1 week regardless of the injected dose [48]. A study of 10 healthy young male athletes reported adrenal suppression for up to 2 weeks after a single intra- or extraarticular injection [49]. A study found that injecting two joints simultaneously yields higher serum absorption than injecting the same total dose into one joint; this finding is theorized to be due to the exposure of a greater volume of synovium to corticosteroid, thus increasing the likelihood of systemic absorption [48]. This systemic absorption of corticosteroid accounts for many of the side effects occurring after injection, as well as the such side effects' similarity to the known side effects of exogenous steroid medication.

One study reported facial flushing (defined as erythema of the upper chest and face) after intraarticular corticosteroid injection in 40% of patients, with this side effect being unpleasant in 15% in cases, predominantly occurring in women, and being reproducible with subsequent injection [50]. Flushing is self-limiting, typically resolving within 3 days [51]. Associated systemic symptoms can include insomnia and headache.

Transient hypertension and postprandial hyperglycemia occur after intraarticular corticosteroid injection in patients both with and without diabetes and relate to suppression of the corticotropic axis; these effects may last for over 1 week [52]. These side effects' clinical implications depend on the patient's baseline condition.

Similar to the known side effects of oral or IV steroid use, psychiatric effects (e.g., altered mental status, visual hallucination) have been reported after intraarticular injection, typically occurring within 1 week after injection and occurring more commonly in older patients with decreased renal function [53, 54]. Prolonged hiccups have been reported after corticosteroid injection and are of indeterminate etiology [55].

Hypersensitivity reactions—The reported preval‎ence of hypersensitivity reactions after corticosteroid administration by any route is 0.3–0.5%; a common presentation of such reactions after musculoskeletal injection is delayed urticaria at the injection site [56, 57]. A case report described immediate anaphylaxis after intraarticular injection of triamcinolone acetonide, with confirmed hypersensitivity reaction by skin prick test [57]. Patients with frequent exposure to steroids are more likely to experience a hypersensitivity reaction. In addition, patients reporting a prior hypersensitivity reaction to corticosteroids should undergo workup for potential allergy to any alternative steroid preparation that is selected before it is administered [56]. Management of hypersensitivity reactions may include administration of oral or IV diphenhydramine hydrochloride, depending on the reaction's severity [56, 57].

Ophthalmic side effects—Central serous chorioretinopathy is a retinal detachment presenting with potentially permanent decreased vision that has been attributed to exogenously administered corticosteroids [58]. Any form of corticosteroid administration can precipitate glaucoma or increased intraocular pressure, the latter of which may lead to sudden-onset blurred vision or potentially irreversible vision loss [59]. These potential complications must be discussed before corticosteroid injection with patients who have known ophthalmic pathology and with patients reporting prior visual changes after steroid exposure.

Systemic Delayed Side Effects

As mentioned previously, intraarticular corticosteroid injection suppresses the hypothalamic-pituitary-adrenal axis, causing an at least temporary decrease in endogenous cortisol levels in 52.2% of patients [60]. This suppression can be potentially life-threatening in patients who are subject to an emotional or physical stress, such as illness or surgery, that may precipitate an adrenal crisis [6]. Most patients with adrenal insufficiency are not diagnosed until they are hospitalized for adrenal crisis, which has a reported mortality of 5% [61]. Prolonged adrenal insufficiency has been reported in some patients after repeated corticosteroid injections (i.e., iatrogenic Cushing syndrome) and may require exogenous supplementation to avoid an adrenal crisis [62, 63]. Although limited data describe prolonged adrenal suppression after musculoskeletal injection, a meta-analysis found that adrenal insufficiency may be present in 25.3% of patients 6 months after any corticosteroid administration [60]. Due to individual variation in glucocorticoid sensitivity, there is no corticosteroid dose, administration route, or treatment duration that prevents adrenal suppression [60, 64]. Furthermore, some suggest that no corticosteroids be administered for at least 4 weeks before planned surgery, to decrease the chance of adrenal crisis [28].

Osteoporosis is a known complication of long-term exogenous steroid use, attributed to steroid's negative effect on both bone quality and bone mass. A statistically significant decrease in bone mineral density of the femur has been reported after a single epidural steroid injection in postmenopausal women [65]. The literature suggests that the effect of epidural corticosteroid injection on bone mineral density is dose-dependent, whether from a single injection or multiple injections over time (i.e., cumulative dose), and may also predispose to increased vertebral fracture risk [28]. The incidence and impact of osteoporotic fractures from intra- and extraarticular corticosteroid injections for musculoskeletal indications are unknown. However, given a recognized impact of epidural steroid injections on bone mineral density, the radiologist should provide a recommendation to the primary care practitioner to consider initiating prophylactic osteoporotic treatment in postmenopausal women receiving frequent or multiple injections [28, 65]. Local Immediate Side Effects

Postinjection flare is an exacerbation of pain after injection that is significantly higher than the baseline preinjection pain. This flare is theorized to result from focal synovitis secondary to intracellular incorporation of the injected corticosteroid particulate microcrystals [9]. Postinjection flare is more common and of longer duration for particulate than nonparticulate corticosteroids [10]. The patient experience of postinjection flare can be clinically similar to the experience of crystalline-induced arthropathies, such as gout. Studies including injections of the hand, wrist, and shoulder report an incidence of postinjection flare of 32.8–35.3% [11, 12]. A study assessing flare reactions after injections performed to treat de Quer-vain tenosynovitis and trigger finger found that most flares occurred 1–2 days after injection, lasted a mean of 2 days, and did not impact long-term symptom resolution [11].

Local Delayed Side Effects

Skin hypopigmentation and atrophy—Several prospective studies describe skin hypopigmentation and atrophy, particularly after superficial peritendinous injections (e.g., to treat epicondylitis or de Quervain tenosynovitis) [13, 14] (Fig. 1). Heel pad atrophy or skin depigmentation is reported in 10% of patients who received corticosteroid injection of the plantar fascia [15]. Skin atrophy and hypopigmentation manifest within a few months of injection and typically resolve within 1 year [16]. Such dermal changes likely relate specifically to the use of particulate steroids, especially if the injectate migrates along the injection pathway within the soft tissues [17]. When superficial structures are being injected, the use of nonparticulate steroids should be considered, to minimize the potential for skin atrophy. Additionally, compression of the injection site after removal of the needle may minimize the risk of steroid migration along the needle track [18].

Open in Viewer

Fig. 1 —63-year-old woman with medial elbow pain who presented for sonographic eval‎uation of common flexor tendon. Patient had received two prior injections of 40 mg of triamcinolone acetonide (Kenalog, Proficient Rx) in medial elbow without imaging guidance within past 6 months. Photograph obtained at time of presentation for imaging shows discoloration and thinning of skin at medial elbow (dashed outline). Patient noticed this change several weeks after second injection.

Soft-tissue calcification—Soft-tissue calcification is a reported iatrogenic complication of musculoskeletal corticosteroid injection [19]. Explanatory theories include local tissue damage resulting in increased pH that precipitates calcium salt or possible calcific necrosis of collagen fibers [19]. Soft-tissue and tendinous calcifications are associated with increased pain scores and may be clinically significant [19]. Radiologists should recognize this potential corticosteroid-induced complication when assessing for changes in periarticular calcifications on postinjection imaging.

Infection—Intraarticular infection after corticosteroid injection is very rare, with an incidence thought to be less than 0.001% [20]. The incidence may be significantly less when using single-dose vials, as well as when using modern sterile technique, packaging, and cleaning methods [21]. A systematic review of case reports describes the incidences of soft-tissue infection and osteomyelitis of the underlying bone adjacent to the injection site [22]. The injection may also increase the risk of infection if the patient later undergoes surgery at the injection site. For example, studies have found that patients who received intraarticular injections of corticosteroid before a knee or hip replacement were more likely to experience a postoperative prosthetic joint infection than those who had never received an injection [23, 24]. A meta-analysis of eight studies on hip and knee joint replacements showed a 2.13 higher odds of deep infection in patients with than without a prior injection but no significant difference in superficial infection rate [25]. Another systematic review found that intraarticular injections performed within 3 months before knee arthroplasty were significantly associated with subsequent prosthetic joint infection [26]. Moreover, a study found that the use of preoperative corticosteroids was associated with a higher likelihood of superficial or deep infections in any surgical site, as well as an increased risk of respiratory, renal, and cardiac complications [27]. Given the available data, it may be advisable to delay any surgery for 1–3 months after corticosteroid injection to potentially mitigate the infection risk, particularly in patients with multiple comorbidities and chronic conditions and who are dependent on corticosteroids [28]. It is important to consider this potential infection risk before injecting patients who may be considering arthroplasty or other surgeries soon after corticosteroid injection and to counsel such patients accordingly.

Tendon, ligament, or fascial insufficiency or rupture—A rare but significant potential consequence of corticosteroid injections is tendon rupture (Figs. 2 and 3). Although its incidence is low, reported as approximately 0.1% in a meta-analysis, case reports describe such complications throughout the body [29, 30]. A review of the case reports estimated that among 95 postinjection complications, 53.7% were plantar fascia ruptures, and 26.3% were tendon ruptures involving the patellar, quadriceps, Achilles, or biceps tendon—tendons possibly already prone to increased biomechanical stress [31]. Injected corticosteroids are thought to locally decrease collagen production, which may limit potential tendon, ligament, or fascial healing and therefore increase predisposition to structural failure or rupture [32]. Even if rupture does not occur, corticosteroid use has been associated with ligament instability in the elbow, leading to refractory pain and disability [33]. A histologic study of the Achilles tendon in rabbits reported equivalent tendon weakening and loss of elasticity after either intratendinous or retrocalcaneal bursa injection, indicating that even peritendinous injections pose a risk of tendon rupture [34]. Tendon rupture may also be more strongly associated with triamcinolone acetonide than with other steroid preparations and may be associated with frequent repeated injections to a particular site [31]. Tendon rupture has been reported to occur anywhere from 2 days to 6 weeks after injection [29]. Although evidence-based consensus guidelines are lacking, some authors suggest that a maximum of three to five injections be performed at a single site, with a minimum of 2 weeks between injections, to minimize potential tendon rupture [31].

Open in Viewer

Fig. 2A —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

A, Gray-scale sonographic image obtained using 15-MHz linear probe shows intact gluteus medius insertion (dashed lines) on greater trochanter (asterisk) in long axis. Image shows thickened tendon with moderate tendinopathy but no high-grade tear. Patient underwent ultrasound-guided percutaneous tenotomy followed by injection of 20 mg of methylprednisolone acetate (Depo-Medrol, Pfizer) and 4 mL of 0.25% bupivacaine into greater trochanteric bursa. Twelve days after procedure, patient returned due to worsening pain, which was worse with weight-bearing and did not improve after course of NSAIDS and oral steroids. Patient reported no interval trauma. MRI of osseous pelvis was performed 18 days after procedure.

Open in Viewer

Fig. 2B —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

B, Coronal STIR (B) and sagittal T2-weighted (C) images, obtained 18 days after A, through greater trochanter show full-thickness tear of left gluteus medius tendon at myotendinous junction (arrow, B) with 2-cm tendon gap (arrow, C).

Open in Viewer

Fig. 2C —50-year-old man with left lateral hip pain that was not relieved by conservative measures who was referred from physiatry service for percutaneous intervention.

C, Coronal STIR (B) and sagittal T2-weighted (C) images, obtained 18 days after A, through greater trochanter show full-thickness tear of left gluteus medius tendon at myotendinous junction (arrow, B) with 2-cm tendon gap (arrow, C).

Open in Viewer

Fig. 3A —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

A, Gray-scale sonographic image of right hamstring tendon origin in long axis, obtained at initial presentation, shows marked thickening and hypoechoic degeneration of tendon (dashed lines) with areas of anechoic interstitial tearing (arrow). Ischial tuberosity is denoted by asterisk. Two days after third steroid injection, patient felt snap in buttock when performing household tasks.

Open in Viewer

Fig. 3B —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

B, Coronal STIR image of bony pelvis through ischial tuberosity shows near-complete rupture of right hamstring tendon origin with almost 3 cm of distal retraction (arrow).

Open in Viewer

Fig. 3C —71-year-old woman who presented with chronic right buttock pain. Patient had received one injection of 10–20 mg of triamcinolone acetonide (Kenalog, Proficient Rx) per year directed to ischial bursa for previous 3 years.

C, Gray-scale sonographic image of left hamstring origin in long axis, immediately caudal to ischial tuberosity (asterisk), shows anechoic tendon gap (arrow).

Corticosteroid injections in the shoulder have been associated with a need for revision after rotator cuff repair, possibly due to underlying degeneration of the tendon quality as a result of the injection [35]. A systematic review found a nearly twofold increase in adverse outcomes after rotator cuff repair if a corticosteroid injection is administered within 6 months before surgery or if more than two injections are performed within 1 year before surgery [36]. This effect was also shown in an analysis of patients with lateral epicondylitis who underwent surgery; those receiving three or more injections were at highest risk (OR, 3.55) for requiring revision surgery [37]. Patients must be appropriately counseled regarding the potential greater need for revision surgery if a corticosteroid is administered within 6 months before surgical tendon repair.

Accelerated progression of osteoarthritis—Accelerated progression of osteoarthritis is not strictly defined but refers to an increase in joint space loss that is greater than expected to occur in a short time period. One study defined accelerated progression of osteoarthritis as a decrease in joint space by 2 mm over a 12-month period [38]. Research from the Osteoarthritis Initiative found that patients who received intraarticular knee joint injections of corticosteroid had 3.2 higher odds of showing radiographic progression of osteoarthritis than those who did not receive injections; the odds were higher at 4.67 in patients who received continuous injections [39]. Radiographic progression of osteoarthritis after injection has also been shown in the hip, with a study reporting radiographic progression of hip joint space narrowing in 44% of patients who received hip injections and were followed postprocedurally for a mean of 6 months [40].

Osseous injury—Corticosteroids induce osteocyte apoptosis, which predisposes to decreases in osseous vascularity and bone strength; these effects are particularly pronounced after intraarticular injection [41]. Indeed, subchondral insufficiency fracture, osteonecrosis, and rapid joint destruction have all been reported after intraarticular corticosteroid injection into various joints [42–44] (Figs. 4 and 5). A retrospective single-center study of 1000 patients with long-term follow-up after hip and knee corticosteroid injections showed a combined rate of osteonecrosis, insufficiency fractures, and rapid progressive osteoarthritis of 1% [45]. Such reports are controversial, and some experts believe that the reported risk of osseous injury after intraarticular injection is inflated (possibly attributed to the natural progression of osteoarthritis) and that a risk-benefit analysis would support this approach [46].

Open in Viewer

Fig. 4A —66-year-old man who presented with left groin pain.

A, Coronal STIR image through osseous pelvis shows large left hip joint effusion (asterisk).

Open in Viewer

Fig. 4B —66-year-old man who presented with left groin pain.

B, Sagittal T2-weighted image through left hip joint shows areas of full-thickness cartilage loss and subchondral marrow edema–like signal (arrows). Patient underwent same-day injection of left hip and iliopsoas bursa, receiving total of 60 mg of triamcinolone acetonide (Kenalog, Proficient Rx). Patient presented 3 months after injection with worsening hip pain.

Open in Viewer

Fig. 4C —66-year-old man who presented with left groin pain.

C, Anteroposterior radiograph of pelvis obtained 3 months after A and B shows severe joint space narrowing of left femoral articular surface with articular surface collapse and flattening (arrow).

Open in Viewer

Fig. 4D —66-year-old man who presented with left groin pain.

D, Coronal STIR (D) and coronal T1-weighted (E) images obtained same day as radiograph (C) show severe subchondral marrow edema spanning left hip and large joint effusion. Articular surface collapse (arrow, D), with areas of subchondral sclerosis (asterisk), is also present. Differential diagnosis includes rapid progression of osteoarthritis, avascular necrosis, and septic joint. Further workup identified no evidence of infection, and patient underwent left total hip arthroplasty.

Open in Viewer

Fig. 4E —66-year-old man who presented with left groin pain.

E, Coronal STIR (D) and coronal T1-weighted (E) images obtained same day as radiograph (C) show severe subchondral marrow edema spanning left hip and large joint effusion. Articular surface collapse (arrow, D), with areas of subchondral sclerosis (asterisk), is also present. Differential diagnosis includes rapid progression of osteoarthritis, avascular necrosis, and septic joint. Further workup identified no evidence of infection, and patient underwent left total hip arthroplasty.

Open in Viewer

Fig. 5A —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

A, Preinjection anteroposterior radiograph of pelvis shows severe left hip osteoarthritis with joint space narrowing, marginal spurring (arrow), and superolateral subluxation of femoral head. Intraarticular injection of 40 mg of triamcinolone acetonide (Kenalog, Proficient Rx) was performed after radiography.

Open in Viewer

Fig. 5B —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

B, Coronal STIR (B) and coronal T1-weighted (C) images of osseous pelvis obtained 2 weeks after injection show rapid destruction of femoral articular surface (asterisk) and acetabulum (arrow), with large joint effusion.

Open in Viewer

Fig. 5C —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

C, Coronal STIR (B) and coronal T1-weighted (C) images of osseous pelvis obtained 2 weeks after injection show rapid destruction of femoral articular surface (asterisk) and acetabulum (arrow), with large joint effusion.

Open in Viewer

Fig. 5D —77-year-old woman who presented for therapeutic corticosteroid injection to treat symptomatic moderate left hip osteoarthritis, with worsening left hip pain and difficulty with weight-bearing.

D, Anteroposterior radiograph of pelvis obtained 3 weeks after injection shows progressing marked osseous resorption of femoral head, which now extends to intertrochanteric region (arrow), with cephalad migration and lateral displacement of proximal femur. No evidence of infection was found on joint aspiration or intraoperatively during total hip arthroplasty.

An expert panel review concluded that estimates of the preval‎ence of structural problems induced by intraarticular injections are highly variable and that, at present, no imaging, clinical, or laboratory markers are established for predicting subsequent osseous injury after an intraarticular injection [47]. The presence of acute joint pain that is not explained by radiographic or sonographic findings could prompt caution, with further workup obtained before potential corticosteroid injection. In this situation, MRI may help to detect osseous pathology, such as subchondral fracture, that would be a contraindication to injection; universal implementation of this approach may be cost-prohibitive given the likely low frequency of osseous injury on MRI [47]. The use of repeat radiographs before joint injection is also controversial. The panel concluded that large clinical trials are needed to determine the true incidence of these osseous complications and to formulate preintervention safety measures [47]. Although these osseous complications have a likely very low incidence, patients should be aware of the risk, and clinical judgment should be used to assess such complications' relevance for individual patients. For example, such complications may be considered more relevant in a patient with prior long-standing steroid use or a prior diagnosis of osteopenia or osteoporosis.

Systemic Immediate Side Effects

Adrenal suppression—Locally injected corticosteroid does not remain confined to the joint, and some amount is absorbed by the bloodstream. One study found that serum cortisol levels after intraarticular injection of methylprednisolone acetate were dose-dependent, peaked 2–12 hours after injection, and showed a significant downregulation for at least 1 week regardless of the injected dose [48]. A study of 10 healthy young male athletes reported adrenal suppression for up to 2 weeks after a single intra- or extraarticular injection [49]. A study found that injecting two joints simultaneously yields higher serum absorption than injecting the same total dose into one joint; this finding is theorized to be due to the exposure of a greater volume of synovium to corticosteroid, thus increasing the likelihood of systemic absorption [48]. This systemic absorption of corticosteroid accounts for many of the side effects occurring after injection, as well as the such side effects' similarity to the known side effects of exogenous steroid medication.

One study reported facial flushing (defined as erythema of the upper chest and face) after intraarticular corticosteroid injection in 40% of patients, with this side effect being unpleasant in 15% in cases, predominantly occurring in women, and being reproducible with subsequent injection [50]. Flushing is self-limiting, typically resolving within 3 days [51]. Associated systemic symptoms can include insomnia and headache.

Transient hypertension and postprandial hyperglycemia occur after intraarticular corticosteroid injection in patients both with and without diabetes and relate to suppression of the corticotropic axis; these effects may last for over 1 week [52]. These side effects' clinical implications depend on the patient's baseline condition.

Similar to the known side effects of oral or IV steroid use, psychiatric effects (e.g., altered mental status, visual hallucination) have been reported after intraarticular injection, typically occurring within 1 week after injection and occurring more commonly in older patients with decreased renal function [53, 54]. Prolonged hiccups have been reported after corticosteroid injection and are of indeterminate etiology [55].

Hypersensitivity reactions—The reported preval‎ence of hypersensitivity reactions after corticosteroid administration by any route is 0.3–0.5%; a common presentation of such reactions after musculoskeletal injection is delayed urticaria at the injection site [56, 57]. A case report described immediate anaphylaxis after intraarticular injection of triamcinolone acetonide, with confirmed hypersensitivity reaction by skin prick test [57]. Patients with frequent exposure to steroids are more likely to experience a hypersensitivity reaction. In addition, patients reporting a prior hypersensitivity reaction to corticosteroids should undergo workup for potential allergy to any alternative steroid preparation that is selected before it is administered [56]. Management of hypersensitivity reactions may include administration of oral or IV diphenhydramine hydrochloride, depending on the reaction's severity [56, 57].

Ophthalmic side effects—Central serous chorioretinopathy is a retinal detachment presenting with potentially permanent decreased vision that has been attributed to exogenously administered corticosteroids [58]. Any form of corticosteroid administration can precipitate glaucoma or increased intraocular pressure, the latter of which may lead to sudden-onset blurred vision or potentially irreversible vision loss [59]. These potential complications must be discussed before corticosteroid injection with patients who have known ophthalmic pathology and with patients reporting prior visual changes after steroid exposure.

Systemic Delayed Side Effects

As mentioned previously, intraarticular corticosteroid injection suppresses the hypothalamic-pituitary-adrenal axis, causing an at least temporary decrease in endogenous cortisol levels in 52.2% of patients [60]. This suppression can be potentially life-threatening in patients who are subject to an emotional or physical stress, such as illness or surgery, that may precipitate an adrenal crisis [6]. Most patients with adrenal insufficiency are not diagnosed until they are hospitalized for adrenal crisis, which has a reported mortality of 5% [61]. Prolonged adrenal insufficiency has been reported in some patients after repeated corticosteroid injections (i.e., iatrogenic Cushing syndrome) and may require exogenous supplementation to avoid an adrenal crisis [62, 63]. Although limited data describe prolonged adrenal suppression after musculoskeletal injection, a meta-analysis found that adrenal insufficiency may be present in 25.3% of patients 6 months after any corticosteroid administration [60]. Due to individual variation in glucocorticoid sensitivity, there is no corticosteroid dose, administration route, or treatment duration that prevents adrenal suppression [60, 64]. Furthermore, some suggest that no corticosteroids be administered for at least 4 weeks before planned surgery, to decrease the chance of adrenal crisis [28].

Osteoporosis is a known complication of long-term exogenous steroid use, attributed to steroid's negative effect on both bone quality and bone mass. A statistically significant decrease in bone mineral density of the femur has been reported after a single epidural steroid injection in postmenopausal women [65]. The literature suggests that the effect of epidural corticosteroid injection on bone mineral density is dose-dependent, whether from a single injection or multiple injections over time (i.e., cumulative dose), and may also predispose to increased vertebral fracture risk [28]. The incidence and impact of osteoporotic fractures from intra- and extraarticular corticosteroid injections for musculoskeletal indications are unknown. However, given a recognized impact of epidural steroid injections on bone mineral density, the radiologist should provide a recommendation to the primary care practitioner to consider initiating prophylactic osteoporotic treatment in postmenopausal women receiving frequent or multiple injections [28, 65].