Re: Levodopa and the Progression of Parkinson's Disease. NEJM 논문

[문형철]

레보도파를 40주(10개월)투여하고 

2주동안 레보도파를 끊으면 어떤 일이 일어나는가?

레보도파를 투여한그룹이 위약그룹보다 파킨슨병 진행도가 늦춰졌다는 논문...

그래서 건생병사 대안이 없다면, 레보도파-카르비도파를 투여하는 것이 유리

그런데..

Off period, 레보도파 유발 kyskinesia가 발생할 수밖에 없는 기간동안(최소 몇년) 이후에는?  

ORIGINAL ARTICLE

SHARE ON

•  
•  
•  
•  

Levodopa and the Progression of Parkinson's Disease

Author: The Parkinson Study Group*Author Info & Affiliations

Published December 9, 2004

N Engl J Med 2004;351:2498-2508

DOI: 10.1056/NEJMoa033447

VOL. 351 NO. 24

• • ABSTRACT
  • METHODS
  • RESULTS
  • DISCUSSION
  • NOTES
  • APPENDIX
  • REFERENCES
• • Information & Authors
  • Metrics & Citations
  • View Options
  • References
  • Media
  • Tables
  • Share

Abstract

BACKGROUND

Despite the known benefit of levodopa in reducing the symptoms of Parkinson's disease, concern has been expressed that its use might hasten neurodegeneration. This study assessed the effect of levodopa on the rate of progression of Parkinson's disease.

METHODS

In this randomized, double-blind, placebo-controlled trial, we eval‎uated 361 patients with early Parkinson's disease who were assigned to receive carbidopa–levodopa at a daily dose of 37.5 and 150 mg, 75 and 300 mg, or 150 and 600 mg, respectively, or a matching placebo for a period of 40 weeks, and then to undergo withdrawal of treatment for 2 weeks. The primary outcome was a change in scores on the Unified Parkinson's Disease Rating Scale (UPDRS) between baseline and 42 weeks. Neuroimaging studies of 142 subjects were performed at baseline and at week 40 to assess striatal dopamine-transporter density with the use of iodine-123–labeled 2-β-carboxymethoxy-3-β-(4-iodophenyl)tropane ([123I]β-CIT) uptake.

RESULTS

The severity of parkinsonism increased more in the placebo group than in all the groups receiving levodopa: the mean difference between the total score on the UPDRS at baseline and at 42 weeks was 7.8 units in the placebo group, 1.9 units in the group receiving levodopa at a dose of 150 mg daily, 1.9 in those receiving 300 mg daily, and –1.4 in those receiving 600 mg daily (P<0.001). In contrast, in a substudy of 116 patients the mean percent decline in the [123I]β-CIT uptake was significantly greater with levodopa than placebo (–6 percent among those receiving levodopa at 150 mg daily, –4 percent in those receiving it at 300 mg daily, and –7.2 percent among those receiving it at 600 mg daily, as compared with –1.4 percent among those receiving placebo; 19 patients with no dopaminergic deficits on the baseline scans were excluded from the analysis) (P=0.036). The subjects receiving the highest dose of levodopa had significantly more dyskinesia, hypertonia, infection, headache, and nausea than those receiving placebo.

CONCLUSIONS

The clinical data suggest that levodopa either slows the progression of Parkinson's disease or has a prolonged effect on the symptoms of the disease. In contrast, the neuroimaging data suggest either that levodopa accelerates the loss of nigrostriatal dopamine nerve terminals or that its pharmacologic effects modify the dopamine transporter. The potential long-term effects of levodopa on Parkinson's disease remain uncertain.

레보도파가

파킨슨병의 증상을 완화하는 데 도움이 된다고 알려져 있지만,

레보도파를 사용하면 신경 퇴화가 촉진될 수 있다는 우려가 제기되고 있습니다.

이 연구는

레보도파가

파킨슨병의 진행 속도에 미치는 영향을 평가했습니다.

연구 방법
이 무작위, 이중맹검, 위약 대조 시험에서는

초기 파킨슨병 환자 361명을 대상으로

40주 동안 카르비도파-레보도파를 각각 37.5 및 150mg, 75 및 300mg, 150 및 600mg의 일일 용량으로 투여하거나

동일한 위약을 투여한 후

2주 동안 치료를 중단하도록 배정하여 평가했습니다.

1차 결과는 기준점과 42주 사이의 통합 파킨슨병 평가 척도(UPDRS) 점수의 변화였습니다. 요오드-123 표지 2-β-카복시메톡시-3-β-(4-리오도페닐)트로판([123I]β-CIT) 흡수를 사용하여 선조체 도파민 수송체 밀도를 평가하기 위해 142명의 피험자에 대한 신경 영상 연구를 기저시점과 40주차에 실시했습니다.

결과

파킨슨병의 중증도는

위약 그룹에서 레보도파를 투여한

모든 그룹보다 더 많이 증가했습니다:

베이스라인과 42주 시점의 UPDRS 총점 평균 차이는

위약 그룹에서 7.8단위,

레보도파를 매일 150mg 투여한 그룹에서 1.9단위,

매일 300mg 투여한 그룹에서 1.9단위,

매일 600mg 투여한 그룹에서 -1.4단위였습니다(P<0.001).

반면, 116명의 환자를 대상으로 한 하위 연구에서는 위약보다 레보도파에서 [123I]β-CIT 흡수율의 평균 감소율이 유의하게 더 컸습니다(레보도파 1일 150mg 투여 환자에서 -6%, 1일 300mg 투여 환자에서 -4%, 1일 600mg 투여 환자에서 -7.2%, 위약 투여 환자에서 -1.4%; 기저 스캔에서 도파민성 결핍이 없었던 19명은 분석에서 제외)(P=0.036).

최고 용량의 레보도파를 투여받은 피험자들은

위약을 투여받은 피험자들보다

운동 이상증, 과긴장증 hypertonia , 감염, 두통 및 메스꺼움이

훨씬 더 많이 발생했습니다.

결론

임상 데이터는

레보도파가

파킨슨병의 진행을 늦추거나

파킨슨병 증상에 장기간 영향을 미친다는 것을 시사합니다.

대조적으로, 신경 영상 데이터는 레보도파가 흑질 도파민 신경 말단의 손실을 가속화하거나 약리학 적 효과가 도파민 수송체를 수정한다는 것을 시사합니다. 파킨슨병에 대한 레보도파의 잠재적인 장기적 효과는 아직 불확실합니다.

Parkinson's disease is a progressively disabling neurodegenerative disorder that is manifested clinically by bradykinesia, tremor, rigidity, flexed posture, postural instability, and freezing of gait. It is characterized pathologically by the loss of pigmented dopaminergic neurons in the substantia nigra. The course of the clinical decline parallels that of the progressive degeneration of the remaining dopaminergic neurons.1 The use of levodopa as dopamine-replacement therapy is highly effective in ameliorating the symptoms of the disease and remains the standard drug with which other therapies are compared.2,3

Because levodopa and dopamine can generate reactive oxygen species and induce the degeneration of cultured dopamine neurons, concern has been raised that levodopa could enhance oxidative stress and hasten the degeneration of residual dopamine neurons in patients with Parkinson's disease.4–6 However, levodopa is not toxic in animals and may be trophic and promote the functional recovery of damaged nigral neurons.7–10 Humans without Parkinson's disease who are exposed to levodopa do not develop nigral damage,11,12 but such persons do not have increased oxidative stress in their substantia nigra neurons.

Whether levodopa is detrimental, beneficial, or without effect on the rate of the progression of Parkinson's disease is unknown and extremely important, both scientifically and clinically. We therefore conducted a controlled clinical trial to assess the effect of levodopa on the course of Parkinson's disease.

파킨슨병은 서동증, 떨림, 경직, 구부러진 자세, 자세 불안정, 보행 정지 등이 임상적으로 나타나는 점진적으로 장애를 일으키는 신경 퇴행성 질환입니다. 병리학적으로 흑질 흑질에서 색소성 도파민 신경세포가 소실되는 것이 특징입니다. 임상적 감소 과정은 남아있는 도파민 신경세포의 점진적 퇴행과 유사합니다.1 도파민 대체 요법으로 레보도파를 사용하면 질병의 증상을 개선하는 데 매우 효과적이며 다른 치료법과 비교되는 표준 약물로 남아 있습니다.2,3

레보도파와 도파민은

활성 산소종을 생성하고

배양된 도파민 신경세포의 퇴행을 유도할 수 있기 때문에

레보도파가 파킨슨병 환자에서 산화 스트레스를 강화하고

잔류 도파민 신경세포의 퇴행을 촉진할 수 있다는 우려가 제기되어 왔습니다.4 -

그러나

레보도파는

동물에게 독성이 없으며

손상된 흑질 뉴런의 기능 회복을 촉진할 수 있습니다.7-10

레보도파에 노출된 파킨슨병이 없는 사람은

흑질 손상이 발생하지 않지만,11,12

이러한 사람은 흑질 뉴런의 산화 스트레스가 증가하지 않습니다.

레보도파가

파킨슨병의 진행 속도에 해로운지,

유익한지 또는

아무런 영향을 미치지 않는지는

과학적으로나 임상적으로 매우 중요한 문제입니다.

따라서

저희는

레보도파가 파킨슨병의 진행에 미치는 영향을 평가하기 위해

대조 임상 시험을 실시했습니다.

MethodsS

TUDY DESIGN

Our multicenter, placebo-controlled, randomized, dose-ranging, double-blind clinical trial, called the Earlier versus Later Levodopa Therapy in Parkinson Disease (ELLDOPA) study, was conceived, organized, and implemented by the Parkinson Study Group and sponsored by the National Institute of Neurological Disorders and Stroke. The Department of Defense sponsored the single-photon-emission computed-tomography (SPECT) substudy. The subjects were enrolled between September 1998 and August 2001 at 33 sites in the United States and 5 sites in Canada. The study was approved by the institutional review boards at the participating sites, and all subjects gave written informed consent. An independent safety monitoring committee monitored the data, subjects' safety, and the tolerability of the study drug. There was no prespecified formal guideline for recommending either modification or termination of the trial.

SUBJECTS

The subjects were 30 years of age or older, had received a diagnosis of Parkinson's disease within the past two years, had a rating on the modified Hoehn–Yahr scale13,14 of less than stage 3 (with stage 1 indicating unilateral disease, stage 2 mild bilateral disease, and stage 3 more advanced bilateral disease), and were considered not likely to require therapy for symptoms of the disease within the nine months after enrollment in the study. Patients were excluded if they were receiving antiparkinson medication, had been exposed to levodopa or to any dopamine agonist for more than 14 days, had an identifiable cause of parkinsonism, or had a tremor in any limb that was given a score of 3 or more on the Unified Parkinson's Disease Rating Scale (UPDRS),14 freezing of gait, loss of postural reflexes, major depression, or dementia.

Potential subjects were informed that only the assigned study drug would be permitted during the nine months of the study and that if they needed additional antiparkinson medication during this period, they would have to withdraw from the study. Subjects were randomly assigned to receive placebo or carbidopa–levodopa at a dose of 12.5 and 50 mg three times daily, 25 and 100 mg three times daily, or 50 and 200 mg three times daily, respectively. The doses were increased to the full amount over a period of nine weeks in a blinded fashion.

After 40 weeks, the subjects underwent a 3-day period of step-down withdrawal from the study drug. After two weeks without the study drug, a final assessment of the severity of the symptoms of Parkinson's disease was made. The selection of the 2-week duration for the washout period was based on reports that withdrawal from levodopa for a period up to 14 days resulted in a worsening of parkinsonism mainly within the first 7 days, with nonsignificant worsening beyond that point.15–17

대상자는 30세 이상이고,

지난 2년 이내에 파킨슨병 진단을 받았으며,

수정된 Hoehn-Yahr 척도13,14에서 3단계 미만(1단계는 편측성 질환, 2단계는 경미한 양측성 질환, 3단계는 보다 진행된 양측성 질환)으로 평가되었고,

연구 등록 후 9개월 이내에 질환의 증상에 대한 치료가 필요하지 않을 것으로 판단되는 사람으로 제한했습니다.

항파킨슨 약물을 복용 중이거나,

레보도파 또는 도파민 작용제에 14일 이상 노출된 적이 있거나,

파킨슨병의 원인이 확인된 경우,

통합 파킨슨병 평가 척도(UPDRS)에서 3점 이상의 사지 떨림이 있거나,14

보행 동결, 자세 반사 상실, 주요 우울증 또는 치매가 있는 환자는 제외되었습니다.

잠재적 피험자에게는 연구 기간 9개월 동안 지정된 연구 약물만 허용되며, 이 기간 동안 항파킨슨 약물이 추가로 필요한 경우 연구 참여를 철회해야 한다는 사실을 알렸습니다. 피험자들은 무작위로 위약 또는 카르비도파-레보도파를 각각 1일 3회 12.5mg 및 50mg, 1일 3회 25mg 및 100mg, 1일 3회 50mg 및 200mg 용량으로 투여받도록 배정되었습니다. 블라인드 방식으로 9주 동안 용량을 최대로 늘렸습니다.

40주 후, 피험자들은 3일간 시험 약물을 단계적으로 중단하는 기간을 거쳤습니다. 2주 동안 임상시험 약물을 투여하지 않은 후 파킨슨병 증상의 중증도를 최종 평가했습니다.

휴약 기간으로 2주 기간을 선택한 것은

최대 14일 동안 레보도파를 중단하면

주로 첫 7일 이내에 파킨슨병이 악화되고

그 이후에는 유의미한 악화가 나타나지 않는다는 보고에 근거한 것입니다.15-17

파킨슨병에 대한 추정 신경 보호제는 치료받지 않은 환자에서 임상 장애의 진행을 질병 진행의 지표로 사용하여 평가할 수 있습니다. 증상 치료와 관련된 혼란을 피하기 위해, 치료하지 않은 기준선부터 증상 치료 약물을 중단한 후 최종 방문까지 임상 장애의 진행을 평가하여 기저 질환의 진행을 평가할 수 있습니다. 이러한 유형의 분석에서는 증상의 영향이 제거될 수 있도록 충분한 기간의 휴약 기간을 사용하는 것이 중요합니다. 증상 효과의 시간 경과를 평가하기 위해 레보도파/카르비도파 및 브로모크립틴을 중단한 후 1일, 8일, 15일째에 31명의 환자를 평가했습니다. 평균 총 통합 파킨슨병 평가 척도 점수(+/- 표준 오차)는 1일차부터 15일차까지 7.4+/-1.5점(p <0.0001), 1일차부터 8일차까지 4.5+/-1.2점(p = 0.0009), 8일차부터 15일차까지 2.9+/-1.0점(p = 0.01)이 증가(악화)했습니다(표 1 참조). 초기 파킨슨병 환자에서 레보도파/카르비도파 및 브로모크립틴의 증상 효과를 제거하려면 최소 2주의 휴약 기간이 필요하다는 결론을 내렸습니다.

CLINICAL EVAL‎UATION

The treating investigator, who was blinded to the treatment assignment, performed a clinical eval‎uation with the use of the UPDRS14 at the screening, baseline, and interim visits (at the end of weeks 3, 9, 24, and 40) and during each of the two weeks of the washout phase. During the four interim visits, the eval‎uations were performed before the administration of the first dose of the study drug. At every visit the treating investigator inquired about adverse events. The primary rater, who was also blinded to the treatment assignment and was kept unaware of information obtained during the course of the study, performed the examination with the use of the UPDRS only at the baseline eval‎uation, which took place within four weeks after screening, and again at the final eval‎uation, two weeks after the subject had undergone complete withdrawal from the study treatment (week 42). The study coordinators and the subjects also were blinded to the treatment assignments. An emergency unblinding procedure was established but was never used.

OUTCOME

The prespecified primary outcome was the change in the severity of parkinsonism between the baseline visit and week 42, as measured with the use of the total score on the UPDRS that was obtained by the primary rater; week 42 was 14 days after the withdrawal of the study drug. The expected result was a dose-related deterioration during washout if levodopa were shown to hasten the progression of Parkinson's disease. The change in the total scores on the UPDRS measured by the treating investigator at each visit was a prespecified secondary outcome. The treating investigator assessed adverse events with the use of open-ended questioning at each visit.

SUBSTUDY

After the clinical trial was begun, we conducted a substudy with the use of SPECT to measure striatal dopamine-transporter density with the use of iodine-123–labeled 2-β-carboxymethoxy-3-β-(4-iodophenyl)tropane ([123I]β-CIT). The methods have been reported previously.18 The subjects who gave consent underwent SPECT imaging just before the baseline visit and then again before the visit at week 40. The imaging studies were performed at that point rather than after the subjects had undergone withdrawal from levodopa, at week 42, because it was thought that the subjects would not be able to tolerate the procedure (including the necessary travel) after the return of parkinsonism or after its worsening. Another reason was that there was no clear evidence of a short-term effect of levodopa on dopamine-transporter imaging.

All imaging studies were performed at Yale University or the Institute for Neurodegenerative Disorders (both in New Haven, Connecticut). The neuroimaging personnel were blinded to the treatment assignment of the subjects. The results of the SPECT studies were transferred to and analyzed by the biostatistics center of the Parkinson Study Group. The prespecified outcome was the percent change in the ratio of the specific striatal [123I]β-CIT uptake to the nondisplaceable striatal [123I]β-CIT uptake between the two images.

STATISTICAL ANALYSIS

Allowing for a dropout rate of 10 percent of the subjects enrolled, we chose a sample size of 360 subjects (i.e., 90 subjects in each of the four treatment groups) in order to provide the study with 85 percent power to detect a dose–response relationship (linear trend) between the assigned doses and the change in the subjects' total score on the UPDRS between the baseline visit and the final visit at week 42. On the basis of the results of one study,19 we anticipated a rate of worsening in the total score on the UPDRS in the placebo group of 9.5 units over the 9.5 months of the study. The study was powered to detect a linear trend that corresponded to a 4-unit difference (45 percent of 9.5 units) in the score on the UPDRS between the highest dose of levodopa (600 mg per day) and placebo — that is, a change of either 13.5 or 5.5 units from baseline in the group receiving the highest dose of the active study drug.

The primary statistical tests were two-tailed, with an alpha level of 0.05. Only subjects who completed the two-week washout phase were included in this analysis. We used the intention-to-treat principle in the analysis, even if among some subjects the dose of the study drug were to be reduced during the study. The primary analysis assessed the dose–response relationship between the assigned doses and the worsening of parkinsonism, as indicated by the changes in the total score on the UPDRS between the baseline visit and week 42. Statistical comparisons were made by analysis of covariance in a model that adjusted for differences among the investigators performing the eval‎uations and in baseline values.

Results

SUBJECTS

Of a total of 361 subjects enrolled in the study, 317 (88 percent) took the study medication for 40 weeks, and 311 (86 percent) completed the 2 weeks of washout (Figure 1). The neuroimaging substudy was begun after the enrollment of the first 108 subjects. Of the 253 subjects subsequently enrolled, 142 (56 percent) participated in the substudy and underwent the baseline SPECT. Of these, 135 (95 percent) returned for scanning at week 40. The demographic and clinical characteristics of the subjects in the treatment groups were similar at baseline, both in the entire sample and in the neuroimaging substudy (Table 1).

FIGURE 1

Random Assignments to Treatment, Completion of the Trial, and Reasons for Not Completing It.

TABLE 1

Baseline Characteristics of the Study Subjects.

CLINICAL OUTCOME

Levodopa, in a dose–response pattern, significantly (P<0.001) reduced the worsening of symptoms of Parkinson's disease as reflected in the change between the total score on the UPDRS at baseline and that at week 42 (i.e., two weeks after washout of the study medication), as compared with the change in the placebo group (Figure 2 and Table 2). The subjects in the placebo group had mild improvement at the week 3 visit, but after that their symptoms worsened steadily throughout the balance of the study period, including the two-week washout phase. A strong dose–response benefit was detected during the period in which the medication was administered beginning at week 9, when the full dose of 600 mg daily was reached in the group receiving the highest dose of levodopa, and it persisted through week 40. The scores on the UPDRS in the three levodopa groups worsened during the two-week washout period, but these groups did not deteriorate to the level observed in the placebo group, and the group receiving the highest dose of levodopa had the best result (Figure 2 and Table 2). The adverse events that were significantly more common among those receiving levodopa at 600 mg daily than in the placebo group were dyskinesias, nausea, infection, hypertonia, and headache (Table 3).

임상 결과
레보도파는 용량 반응 패턴에서 위약군의 변화와 비교했을 때 베이스라인의 UPDRS 총점과 42주차(즉, 연구 약물을 중단한 지 2주 후)의 총점 사이의 변화에 반영된 파킨슨병 증상 악화를 유의하게(P<0.001) 감소시켰습니다(그림 2 및 표 2). 위약 그룹의 피험자들은 3주차 방문 시에는 증상이 약간 개선되었지만, 그 이후에는 2주간의 휴약기를 포함하여 연구 기간 내내 증상이 꾸준히 악화되었습니다. 가장 높은 용량의 레보도파를 투여한 그룹에서 매일 600mg의 최대 용량에 도달한 9주차부터 약물을 투여하는 기간 동안 강력한 용량 반응 효과가 발견되었으며, 이러한 효과는 40주차까지 지속되었습니다. 세 레보도파 그룹의 UPDRS 점수는 2주간의 휴약 기간 동안 악화되었지만 위약 그룹에서 관찰된 수준까지 악화되지는 않았으며, 가장 높은 용량의 레보도파를 투여한 그룹이 가장 좋은 결과를 보였습니다(그림 2 및 표 2).

레보도파를

매일 600mg씩 투여한 그룹에서

위약 그룹보다 훨씬 더 흔하게 나타난 부작용은

운동 이상증, 메스꺼움, 감염, 고조증 및 두통이었습니다(표 3).

FIGURE 2

Changes in Total Scores on the Unified Parkinson's Disease Rating Scale (UPDRS) from Baseline through Eval‎uation at Week 42.

TABLE 2

Changes in the Scores on the UPDRS between Baseline and Week 42.

TABLE 3

Adverse Events.

SPECT AND [123I]Β-CIT SUBSTUDY

The mean [123I]β-CIT uptake in the striatum, caudate, and putamen at baseline was consistent with values previously reported for patients with early Parkinson's disease.20 The percent decrease in striatal [123I]β-CIT uptake over the 40 weeks of the study treatment was greater among subjects in the levodopa groups than in the placebo group, but this difference was not statistically significant (Table 4). However, 21 of the 142 subjects (14.7 percent) had a putaminal [123I]β-CIT uptake of more than 3.25 at baseline (i.e., more than 75 percent of the age-expected putaminal uptake).21 An analysis of the results of SPECT after the exclusion of the 19 subjects without a dopaminergic deficit who returned for the neuroimaging study at week 40 showed a significantly greater decrease in [123I]β-CIT uptake among those receiving levodopa than among those receiving placebo (P=0.036) (Table 4).

TABLE 4

Change in Striatal [123I]β-CIT Uptake between Baseline and Week 40.

Discussion

We found no clinical evidence that levodopa accelerated the worsening of Parkinson's disease over the 9.5 months of observation. Rather, levodopa was associated with less worsening of parkinsonism than was placebo, consistent with the notion that it slows disease progression (Table 2 and Figure 2). In comparison with the scores on the UPDRS at baseline, the final scores, after the washout phase, had worsened by approximately 8 units in the placebo group, whereas the group receiving levodopa at 600 mg daily did not have evidence of deterioration. The groups receiving levodopa at lower doses did have deterioration, as shown by a comparison of their scores at baseline and at week 42, but the deterioration was less than in the placebo group (P<0.001).

We need to consider that a two-week washout from levodopa may have been insufficient to eliminate fully the effect of the medication on symptoms, and the results observed may be related to a profound effect of levodopa on symptoms that persists for a long time after the drug has been withdrawn. Indeed, Hauser and Holford,22 using a modeling technique, analyzed the withdrawal of levodopa in 20 patients and reported that the mean half-life of levodopa as measured by the loss of the clinical benefit was 7.9 days (95 percent confidence interval, 2.2 to 30.4 days). They suggest that a washout period of 32 days (four half-lives) may be required to eliminate 90 percent of the drug's effects on symptoms. However, we saw little deterioration after one week of washout. Near the end of the study, with the approval of the institutional review boards and of the National Institutes of Health, we asked the last 38 subjects remaining in the study to extend the washout period to four weeks. Among these subjects there was no further worsening of the scores on the UPDRS during the additional two weeks, but the small number of subjects renders this component of the study difficult to interpret.

Muenter and Tyce23 described motor responses with two types of duration, short and long, with levodopa therapy. The short-duration benefit lasts for a few hours after a single dose, and the long-duration benefit lasts several days. In our study, the long-duration benefit appears to have lasted approximately one to two weeks (Figure 2). To argue that a longer washout period might have revealed more clinical worsening than was observed, one could propose a hitherto unknown third type of duration of motor response, one that is more sustained than the so-called long-duration benefit. Such an enduring benefit could be envisioned to result from a prolonged pharmacodynamic effect, for example, on dopamine receptors.

If, however, the clinical effects observed offer evidence of neuroprotection, how can we explain the greater loss of dopamine transporter shown in the SPECT imaging studies of the subjects who received levodopa, as compared with those who received the placebo — a result that suggests the possibility of a levodopa-induced toxic effect on dopamine neurons? At the end of the study, when the SPECT neuroimaging studies were performed, the subjects were still taking levodopa, so it is possible to assume that levodopa has a pharmacologic effect on the dopamine transporter that interferes with and reduces the binding of the β-CIT ligand.

Indeed, one study found a reduction in dopamine-transporter binding with the use of positron-emission tomography (PET) and another dopamine-transporter ligand in patients with early Parkinson's disease who were treated for six weeks with levodopa at 300 mg daily.24 However, the sample size in this study, as in others that have shown no change in dopamine-transporter binding after short-term treatment with levodopa,25–27 was too small to demonstrate a statistically significant difference between levodopa and placebo. Further indirect support for the absence of a pharmacologic effect of levodopa on imaging studies of the dopamine transporter in the present study was the absence of decline in the [123I]β-CIT uptake by week 40 among the 16 subjects who received levodopa and whose SPECT scans were normal at baseline (i.e., without evidence of a dopaminergic deficit) (data not shown). However, in the absence of studies with larger samples and a longer period of treatment with levodopa, we cannot exclude the possibility that levodopa may simply down-regulate the dopamine transporter. Another consideration is that the falling dopamine concentration in the placebo group may have led to a compensatory increase in the activity of the dopamine transporter that could have increased [123I]β-CIT binding, but evidence for such an interpretation is lacking.

If levodopa has neuroprotective effects, what mechanisms could account for this property? In low concentrations and in the presence of glial cells, levodopa protects cultured dopaminergic neurons28–31 and up-regulates antioxidant and antiapoptotic proteins.29,32 Furthermore, in vivo studies suggest that levodopa can promote survival and enhance the sprouting of nigral dopamine neurons in rodents treated with the toxin 6-hydroxydopamine.9,10 The question of whether levodopa has a protective or a toxic effect in Parkinson's disease — reflecting the results of clinical examination and neuroimaging studies, respectively — cannot be answered with certainty, and future studies will be needed.

This dose–response, placebo-controlled clinical trial eval‎uating the effect of levodopa in patients with early Parkinson's disease showed a strikingly impressive dose–response clinical benefit: the higher the dose, the stronger and more lasting the benefit, and the benefit was greater even after the drug was withdrawn (Figure 2).

During withdrawal, we did not encounter the neuroleptic malignant-like effect (i.e., high fever, obtundation, and rigidity) that can occur with the sudden withdrawal of levodopa.33,34 Its absence may be related to the down-titration procedure employed; however, this complication is so rare that it might not have occurred in a study of this size even had levodopa been withdrawn suddenly, particularly in patients with early-stage disease.

Our study calls into question interpretations of the functional neuroimaging of the dopamine system. It has generally been assumed that imaging studies performed with 18F-fluorodopa PET and [123I]β-CIT SPECT can provide reliable information on the integrity of the nigrostriatal dopamine pathway. Our study and other recent clinical trials25,35 raise the possibility of a pharmacologic influence of dopaminergic therapy on these neuroimaging targets and point to the need for clarification of this issue.

Although the absence of evidence of a dopaminergic deficit in the imaging in our study and two other studies25,35 may simply represent a limitation of the sensitivity of the imaging techniques, the scans without a dopaminergic deficit raise doubt about whether the subjects had Parkinson's disease.35 In our study, the subjects with such scans at baseline had no worsening of the [123I]β-CIT uptake at week 40 (data not shown), nor did subjects with such scans in the group receiving the highest dose of levodopa have an improvement in the scores on the UPDRS (i.e., a change between the score at baseline and at week 40 of 3.38±4.25 units, as compared with a change of –4.95±10.4 units in the subjects receiving levodopa at 600 mg daily whose baseline SPECT scans showed abnormalities; P=0.002). Whether the subjects with no evidence of dopaminergic deficit on scanning do or do not have classic Parkinson's disease remains uncertain. The fact that scans of some patients with early Parkinson's disease do not have evidence of a dopaminergic deficit needs to be taken into consideration in the planning of future trials to test drugs for neuroprotective effects.

Finally, even though we cannot reconcile the clinical and imaging findings in our study, we can assure both patients with early Parkinson's disease and their physicians that, from a clinical perspective, our study did not find that levodopa hastens the progression of Parkinson's disease. On the basis of the study, we can recommend that the doses of levodopa be adjusted to fit the needs of the patient. Small doses were found to be effective, although less so than higher doses. High doses, however, were associated with a greater frequency of adverse events such as dyskinesia. For the present, until more evidence is available, we recommend customizing the dose of levodopa to the needs of the individual patient on the basis of the clinical response and the profile of adverse events.

토론

9.5개월의 관찰 기간 동안

레보도파가 파킨슨병의 악화를 가속화한다는 임상적 증거는 발견되지 않았습니다.

오히려

레보도파는

위약보다 파킨슨병의 악화를 늦춘다는 개념과 일치하는 것으로 나타났습니다(표 2 및 그림 2).

기준 시점의 UPDRS 점수와 비교했을 때, 휴약기 후 최종 점수는 위약 그룹에서 약 8점 정도 악화된 반면, 레보도파를 매일 600mg씩 투여한 그룹에서는 악화의 증거가 없었습니다. 레보도파를 더 낮은 용량으로 투여한 그룹은 베이스라인과 42주차에 점수를 비교했을 때 악화된 것으로 나타났지만, 위약 그룹에 비해 악화 정도가 덜했습니다(P<0.001).

레보도파를 2주간 중단하는 것만으로는 약물이 증상에 미치는 영향을 완전히 제거하기에 충분하지 않았을 수 있으며, 관찰된 결과는 레보도파가 약물을 중단한 후에도 오랫동안 지속되는 증상에 대한 깊은 영향과 관련이 있을 수 있다는 점을 고려해야 합니다. 실제로 모델링 기법을 사용한 하우저와홀포드22는 20명의 환자를 대상으로 레보도파의 중단을 분석한 결과, 임상적 효과 상실로 측정한 레보도파의 평균 반감기는 7.9일(95% 신뢰 구간, 2.2~30.4일)이었다고 보고했습니다.

연구진은

약물이 증상에 미치는 영향의 90%를 제거하려면

32일(반감기 4회)의 휴약 기간이 필요할 수 있다고 제안합니다.

그러나

휴약 기간 1주일 후에도

증상이 거의 악화되지 않았습니다.

연구가 끝날 무렵, 기관 심의위원회와 국립보건원의 승인을 받아 연구에 남은 마지막 38명의 피험자에게 휴약 기간을 4주까지 연장해 달라고 요청했습니다. 이 피험자들 중 추가 2주 동안 UPDRS 점수가 더 이상 악화되지 않았지만, 피험자 수가 적기 때문에 이 연구 구성 요소를 해석하기는 어려웠습니다.

Muenter와 Tyce23는 레보도파 치료의 운동 반응을 단기간과 장기간의 두 가지 유형으로 설명했습니다. 단기간 효과는 1회 투약 후 몇 시간 동안 지속되며, 장기 효과는 며칠 동안 지속됩니다. 본 연구에서 장시간 효과는 약 1~2주 동안 지속되는 것으로 나타났습니다(그림 2). 휴약 기간이 길어지면 관찰된 것보다 더 많은 임상적 악화가 나타날 수 있다고 주장하려면, 지금까지 알려지지 않은 세 번째 유형의 운동 반응 지속 기간, 즉 소위 장기 지속 효과보다 더 지속되는 효과를 제안할 수 있습니다. 이러한 지속적인 효과는 예를 들어 도파민 수용체에 대한 장기간의 약역학적 효과로 인해 발생하는 것으로 상상할 수 있습니다.

그러나 관찰된 임상 효과가 신경 보호의 증거를 제공한다면, 위약을 투여한 피험자에 비해 레보도파를 투여한 피험자의 SPECT 영상 연구에서 나타난 도파민 수송체의 더 큰 손실, 즉 레보도파가 도파민 신경세포에 독성 효과를 유발할 가능성을 시사하는 결과를 어떻게 설명할 수 있을까요? 연구가 끝날 때 SPECT 신경 영상 연구를 수행했을 때 피험자들은 여전히 레보도파를 복용하고 있었기 때문에 레보도파가 β-CIT 리간드의 결합을 방해하고 감소시키는 도파민 수송체에 약리학 적 영향을 미친다고 가정 할 수 있습니다.

실제로 한 연구에서는 양전자 방출 단층 촬영(PET)과 다른 도파민 수송체 리간드를 사용하여 6주 동안 매일 300mg의 레보도파로 치료받은 초기 파킨슨병 환자에서 도파민 수송체 결합이 감소하는 것을 발견했습니다.24 그러나 이 연구의 표본 크기는 레보도파 단기 치료 후 도파민 수송체 결합에 변화가 없는 다른 연구와 마찬가지로 25-27 너무 작아서 레보도파와 위약 간의 통계적으로 유의미한 차이를 입증하기에는 너무 적었습니다. 본 연구에서 도파민 수송체의 영상 연구에 대한 레보도파의 약리학 적 효과가 없다는 것을 간접적으로 뒷받침하는 또 다른 증거는 레보도파를 투여하고 기저시점에 SPECT 스캔이 정상 (즉, 도파민 결핍의 증거가없는) 16 명의 피험자 중 40 주까지 [123I] β-CIT 흡수가 감소하지 않았다는 것입니다 (데이터 표시되지 않음). 그러나 더 많은 샘플과 더 긴 기간 동안 레보도파를 투여한 연구가 없는 상황에서 레보도파가 단순히 도파민 수송체를 하향 조절할 가능성을 배제할 수는 없습니다. 또 다른 고려 사항은 위약 그룹의 도파민 농도 감소로 인해 도파민 수송체의 활동이 보상 적으로 증가하여 [123I] β-CIT 결합이 증가했을 수 있다는 것이지만 그러한 해석에 대한 증거는 부족합니다.

레보도파가 신경 보호 효과가 있다면 어떤 메커니즘이 이러한 특성을 설명할 수 있을까요?

낮은 농도와 신경교 세포가 있는 경우

레보도파는 배양된 도파민 신경세포를 보호하고28-31

항산화 및 항아포토시스 단백질을 상향 조절합니다.29,32

또한 생체 내 연구에 따르면 레보도파는 독소 6-하이드록시도파민으로 처리된 설치류에서 흑질 도파민 신경세포의 생존을 촉진하고 발아를 향상시킬 수 있습니다.9,10 레보도파가 파킨슨병에서 보호 효과가 있는지 독성 효과가 있는지에 대한 질문은 각각 임상 검사와 신경 영상 연구 결과를 반영하여 확실하게 대답 할 수 없으며 향후 연구가 필요합니다.

초기 파킨슨병 환자에서 레보도파의 효과를 평가한 이 용량 반응, 위약 대조 임상 시험에서는 용량이 높을수록 효과가 더 강하고 오래 지속되었으며, 약물을 중단한 후에도 효과가 더 컸다는 놀랍도록 인상적인 용량 반응 임상 효과가 나타났습니다(그림 2).

약물 중단 중에는

레보도파를 갑자기 중단할 때 발생할 수 있는

신경 이완성 악성 효과(예: 고열, 충혈, 경직)가 나타나지 않았습니다.33,34

이러한 현상이 나타나지 않은 것은 사용된 감량 절차와 관련이 있을 수 있지만, 이러한 합병증은 매우 드물어서 이 규모의 연구에서 레보도파를 갑자기 중단했더라도 특히 초기 질환 환자에서 발생하지 않았을 수도 있습니다.

이 연구는 도파민 시스템의 기능적 신경 영상에 대한 해석에 의문을 제기합니다. 일반적으로 18F- 플루오로 도파 PET 및 [123I] β-CIT SPECT로 수행 된 영상 연구는 흑질 도파민 경로의 완전성에 대한 신뢰할 수있는 정보를 제공 할 수 있다고 가정 해 왔습니다. 본 연구와 다른 최근 임상 시험25,35은 이러한 신경 영상 표적에 대한 도파민 요법의 약리학 적 영향 가능성을 제기하고이 문제에 대한 명확한 설명이 필요함을 지적합니다.

본 연구와 다른 두 연구25,35에서 영상에서 도파민성 결핍의 증거가 없다는 것은 단순히 영상 기술의 감도의 한계를 나타낼 수 있지만, 도파민성 결핍이 없는 스캔은 피험자가 파킨슨병을 앓고 있는지 여부에 대한 의문을 제기합니다.35 우리 연구에서, 베이스라인에 그러한 스캔을 한 피험자들은 40 주째에 [123I] β-CIT 흡수가 악화되지 않았으며 (데이터 표시되지 않음), 최고 용량의 레보도파를 투여 한 그룹에서 그러한 스캔을 한 피험자들은 UPDRS 점수가 개선되지 않았습니다 (즉,, 베이스라인과 40주차의 점수 변화는 3.38±4.25 단위로, 베이스라인 SPECT 스캔에서 이상이 나타난 레보도파 600mg을 매일 투여한 피험자의 변화 -4.95±10.4 단위와 비교했을 때; P=0.002). 스캔에서 도파민 결핍의 증거가 없는 피험자들이 전형적인 파킨슨병에 걸렸는지 아닌지는 아직 불확실합니다. 일부 초기 파킨슨병 환자의 스캔에서 도파민 결핍의 증거가 없다는 사실은 향후 약물의 신경 보호 효과를 테스트하기 위한 임상시험을 계획할 때 고려해야 할 사항입니다.

마지막으로, 본 연구의 임상적 결과와 영상학적 결과를 일치시킬 수는 없지만, 임상적 관점에서 볼 때 레보도파가 파킨슨병의 진행을 촉진하지 않는다는 사실을 초기 파킨슨병 환자와 의사 모두에게 확신할 수 있습니다. 이 연구를 바탕으로 환자의 필요에 따라 레보도파의 용량을 조절할 것을 권장할 수 있습니다. 고용량보다는 덜 효과적이지만 소량도 효과가 있는 것으로 밝혀졌습니다. 그러나 고용량은 운동 이상증과 같은 부작용의 빈도가 더 높은 것으로 나타났습니다. 현재로서는 더 많은 증거가 나올 때까지 임상 반응과 이상 반응의 프로파일을 바탕으로 개별 환자의 필요에 따라 레보도파 용량을 맞춤화할 것을 권장합니다.

NOTES

Supported by grants from the National Institute of Neurological Disorders and Stroke (NS34796, to Dr. Fahn), the Department of Defense (DAMD 17-99-1-9472, to Dr. Marek), and the General Clinical Research Center of the National Center for Research Resources, National Institutes of Health (MO1-RR-00044 and MO1-RR-02066). Carbidopa–levodopa tablets and the matching placebo tablets were kindly provided by Teva Pharmaceuticals (Israel).

Drs. Fahn, Oakes, Shoulson, Kieburtz, Lang, Tanner, and Marek report having served as unpaid consultants to Teva Pharmaceuticals, and Dr. Olanow reports having served as a paid consultant to Teva Pharmaceuticals. Drs. Marek and Seibyl have an equity interest in Molecular Neuroimaging (New Haven), which carried out the [123I]β-CIT SPECT imaging for this study.

We are indebted to the subjects for their participation in the study; to K. Hyland (Baylor Medical College, Houston) for the independent, blinded chemical analyses of the contents of the tablets conducted annually to ascertain the chemical stability of levodopa and carbidopa; to S. Bennett, A. Brocht, D. Graffrath, J. Janciuras, C. Orme, L. Preston, K. Rothenburgh, C. Weaver, and A. Watts of the Biostatistics and Coordination Centers at the University of Rochester, Rochester, N.Y.; to the Independent Safety Monitoring Committee (P. Tariot, chair, Monroe Community Hospital, Rochester, N.Y.; to W.J. Hall, University of Rochester; to R. Rodnitzky, University of Iowa, Iowa City); to the National Institute of Neurological Disorders and Stroke Safety Monitoring Committee (E.C. Haley, chair, University of Virginia, Charlottesville; D. Eidelberg, North Shore University Medical Center, Manhasset, N.Y.; C.A. Gatsonis, Brown University, Providence, R.I.; W. Rocca, Mayo Clinic, Rochester, Minn.); and to E.J. Oliver (administrator), National Institute of Neurological Disorders and Stroke, Bethesda, Md.

참고
국립 신경 장애 및 뇌졸중 연구소(NS34796, 판 박사), 국방부(DAMD 17-99-1-9472, 마렉 박사), 국립 보건원 국립 연구 자원 센터의 일반 임상 연구 센터(MO1-RR-00044 및 MO1-RR-02066)의 보조금을 지원받았습니다. 카비도파-레보도파 정제와 그에 상응하는 위약 정제는 Teva Pharmaceuticals(이스라엘)에서 친절하게 제공했습니다.
판, 오크스, 숄슨, 키버츠, 랭, 태너, 마렉 박사는 테바 제약의 무보수 컨설턴트로, 올라노우 박사는 테바 제약의 유급 컨설턴트로 근무했다고 보고했습니다. 마렉 박사와 세이빌 박사는 이 연구에서 [123I]β-CIT SPECT 영상을 수행한 분자 신경 이미징(뉴헤이븐)에 지분을 보유하고 있습니다.
우리는 연구에 참여한 피험자, 레보도파와 카비도파의 화학적 안정성을 확인하기 위해 매년 실시하는 정제 내용물의 독립적이고 맹검된 화학 분석을 수행한 K. Hyland(휴스턴 베일러 의과대학), S.. 베넷, A. 브로흐트, D. 그라프라스, J. 잔시우라스, C. 오메, L. 프레스턴, K. 로텐버그, C. 위버, 뉴욕 로체스터 대학 생물통계 및 조정 센터의 A. 와츠; 독립 안전성 모니터링 위원회(P.. 타리엇, 뉴욕주 로체스터 소재 먼로 커뮤니티 병원, W.J. 홀, 로체스터 대학교, R. 로드니츠키, 아이오와 시티 소재 아이오와 대학교); 국립 신경장애 및 뇌졸중 안전 모니터링 위원회(E.C. 헤일리, 샬롯츠빌 소재 버지니아 대학교, D. Eidelberg, 노스쇼어대학 메디컬센터, 맨하셋, 뉴욕; C.A. Gatsonis, 브라운대학, 프로비던스, 로드아일랜드; W. Rocca, 메이요 클리닉, 로체스터, 미네소타); 그리고 국립 신경장애 및 뇌졸중 연구소, 베데스다, 메릴랜드, E.J. 올리버(관리자)에게 보낸 편지.

APPENDIX

The following members of the Parkinson Study Group participated in the Earlier versus Later Levodopa Therapy in Parkinson Disease (ELLDOPA) study and contributed to this report. Steering committee: S. Fahn, principal investigator, Columbia University, New York; D. Oakes, chief biostatistician; I. Shoulson, coprincipal investigator; K. Kieburtz, director, Clinical Trials Coordination Center; A. Rudolph, senior project coordinator, University of Rochester, Rochester, N.Y.; K. Marek, neuroimager; J. Seibyl, neuroimager, Institute for Neurodegenerative Disorders, New Haven, Conn.; A. Lang, Toronto Western Hospital, Toronto; C.W. Olanow, Mount Sinai School of Medicine, New York; C. Tanner, The Parkinson's Institute, Sunnyvale, Calif.; G. Schifitto, medical monitor, University of Rochester, Rochester, N.Y.; H. Zhao, biostatistician, University of Rochester, Rochester, N.Y.; L. Reyes, administrator, Columbia University, New York; A. Shinaman, administrator, University of Rochester, Rochester, N.Y.; Participating primary raters, treating investigators, and coordinators: Rush–Presbyterian–St. Luke's Medical Center, Chicago: C. Comella, C. Goetz, L. Blasucci; Barrow Neurological Institute, Phoenix, Ariz.: J. Samanta, M. Stacy, K. Williamson, M. Harrigan; Columbia University, New York: P. Greene, B. Ford, C. Moskowitz; Parkinson's and Movement Disorder Institute, Fountain Valley, Calif.: D. Truong, M. Pathak; Baylor College of Medicine, Houston: J. Jankovic, W. Ondo, F. Atassi, C. Hunter; Brown University, Providence, R.I.: C. Jacques, J.H. Friedman, M. Lannon; Institute for Neurodegenerative Disorders, New Haven, Conn.: D.S. Russell, D. Jennings, B. Fussell; Massachusetts General Hospital, Boston: D. Standaert, M.A. Schwarzschild, J. Growdon, M. Tennis; McGill Centre for Studies in Aging, Verdun, Que., Canada: S. Gauthier, M. Panisset, J. Hall; Oregon Health and Science University, Portland: S. Gancher, J. Hammerstad, C. Stone, B. Alexander-Brown; Albany Medical College, Albany, N.Y.: S. Factor, E. Molho, D. Brown (deceased), S. Evans; Scott and White Hospital–Texas A&M University, Temple: J. Clark, B. Manyam, P. Simpson, B. Wulbrecht, J. Whetteckey; University of Alberta, Edmonton, Canada: W. Martin, T. Roberts, P. King; University of South Florida, Tampa: R. Hauser, T. Zesiewicz, L. Gauger; University of Virginia, Charlottesville: J. Trugman, G.F. Wooten, E. Rost-Ruffner; Washington University, St. Louis: J. Perlmutter, B. Racette; University of Calgary, Alta., Canada: O. Suchowersky, R. Ranawaya, S. Wood, C. Pantella; University of Rochester, Rochester, N.Y.: R. Kurlan, I. Richard, N. Pearson; Mayo Clinic, Scottsdale, Ariz.: J. Caviness, C. Adler, M. Lind; University of Pennsylvania, Philadelphia: T. Simuni, A. Siderowf, A. Colcher, M. Lloyd; University of Miami, Miami: W. Weiner, L. Shulman, W. Koller, K. Lyons; Boston University, Boston: R. Feldman (deceased), M.-H. St.-Hilaire, S. Ellias, C.-A. Thomas; Emory University, Atlanta: J. Juncos, R. Watts, A. Partlow; The Parkinson's Institute, Sunnyvale, Calif.: J. Tetrud, D.M. Togasaki, M. Welsh, T. Stewart; University of Medicine and Dentistry of New Jersey–Robert Wood Johnson, New Brunswick: M.H. Mark, J.I. Sage, D. Caputo; Louisiana State University, New Orleans: H. Gould, J. Rao, A. McKendrick; Mount Sinai School of Medicine, New York: M. Brin, F. Danisi, R. Benabou; Ohio State University, Columbus: J. Hubble, G. Paulson, C. Reider; Toronto Western Hospital, Toronto: A. Birnbaum, J. Miyasaki, L. Johnston, J. So; University of Kansas, Kansas City: R. Pahwa, R. Dubinsky; Mayo Clinic, Jacksonville, Fla.: Z. Wszolek, R. Uitti, M. Turk; Minneapolis Veterans Affairs Hospital, Minneapolis: P. Tuite, D. Rottenberg, J. Hansen; University of Puerto Rico, San Juan: C. Serrano Ramos; University of Southern California, Los Angeles: C. Waters, M. Lew, M. Welsh, C. Kawai; Colorado Neurological Institute, Englewood: C. O'Brien, R. Kumar, L. Seeberger, D. Judd; Ottawa Civic Hospital, Ottawa: T. Mendis, C.L. Barclay, D.A. Grimes, L. Sutherland; Johns Hopkins University, Baltimore: T. Dawson, S. Reich, R. Dunlop; University of Michigan, Ann Arbor: R. Albin, K. Frey, K. Wernette.

ORIGINAL ARTICLE

SHARE ON

•  
•  
•  
•  

Levodopa and the Progression of Parkinson's Disease

Author: The Parkinson Study Group*Author Info & Affiliations

Published December 9, 2004

N Engl J Med 2004;351:2498-2508

DOI: 10.1056/NEJMoa033447

VOL. 351 NO. 24

• • ABSTRACT
  • METHODS
  • RESULTS
  • DISCUSSION
  • NOTES
  • APPENDIX
  • REFERENCES
• • Information & Authors
  • Metrics & Citations
  • View Options
  • References
  • Media
  • Tables
  • Share

AbstractBACKGROUND

Despite the known benefit of levodopa in reducing the symptoms of Parkinson's disease, concern has been expressed that its use might hasten neurodegeneration. This study assessed the effect of levodopa on the rate of progression of Parkinson's disease.

METHODS

In this randomized, double-blind, placebo-controlled trial, we eval‎uated 361 patients with early Parkinson's disease who were assigned to receive carbidopa–levodopa at a daily dose of 37.5 and 150 mg, 75 and 300 mg, or 150 and 600 mg, respectively, or a matching placebo for a period of 40 weeks, and then to undergo withdrawal of treatment for 2 weeks. The primary outcome was a change in scores on the Unified Parkinson's Disease Rating Scale (UPDRS) between baseline and 42 weeks. Neuroimaging studies of 142 subjects were performed at baseline and at week 40 to assess striatal dopamine-transporter density with the use of iodine-123–labeled 2-β-carboxymethoxy-3-β-(4-iodophenyl)tropane ([123I]β-CIT) uptake.

RESULTS

The severity of parkinsonism increased more in the placebo group than in all the groups receiving levodopa: the mean difference between the total score on the UPDRS at baseline and at 42 weeks was 7.8 units in the placebo group, 1.9 units in the group receiving levodopa at a dose of 150 mg daily, 1.9 in those receiving 300 mg daily, and –1.4 in those receiving 600 mg daily (P<0.001). In contrast, in a substudy of 116 patients the mean percent decline in the [123I]β-CIT uptake was significantly greater with levodopa than placebo (–6 percent among those receiving levodopa at 150 mg daily, –4 percent in those receiving it at 300 mg daily, and –7.2 percent among those receiving it at 600 mg daily, as compared with –1.4 percent among those receiving placebo; 19 patients with no dopaminergic deficits on the baseline scans were excluded from the analysis) (P=0.036). The subjects receiving the highest dose of levodopa had significantly more dyskinesia, hypertonia, infection, headache, and nausea than those receiving placebo.

CONCLUSIONS

The clinical data suggest that levodopa either slows the progression of Parkinson's disease or has a prolonged effect on the symptoms of the disease. In contrast, the neuroimaging data suggest either that levodopa accelerates the loss of nigrostriatal dopamine nerve terminals or that its pharmacologic effects modify the dopamine transporter. The potential long-term effects of levodopa on Parkinson's disease remain uncertain.

Parkinson's disease is a progressively disabling neurodegenerative disorder that is manifested clinically by bradykinesia, tremor, rigidity, flexed posture, postural instability, and freezing of gait. It is characterized pathologically by the loss of pigmented dopaminergic neurons in the substantia nigra. The course of the clinical decline parallels that of the progressive degeneration of the remaining dopaminergic neurons.1 The use of levodopa as dopamine-replacement therapy is highly effective in ameliorating the symptoms of the disease and remains the standard drug with which other therapies are compared.2,3

Because levodopa and dopamine can generate reactive oxygen species and induce the degeneration of cultured dopamine neurons, concern has been raised that levodopa could enhance oxidative stress and hasten the degeneration of residual dopamine neurons in patients with Parkinson's disease.4–6 However, levodopa is not toxic in animals and may be trophic and promote the functional recovery of damaged nigral neurons.7–10 Humans without Parkinson's disease who are exposed to levodopa do not develop nigral damage,11,12 but such persons do not have increased oxidative stress in their substantia nigra neurons.

Whether levodopa is detrimental, beneficial, or without effect on the rate of the progression of Parkinson's disease is unknown and extremely important, both scientifically and clinically. We therefore conducted a controlled clinical trial to assess the effect of levodopa on the course of Parkinson's disease.

MethodsSTUDY DESIGN

Our multicenter, placebo-controlled, randomized, dose-ranging, double-blind clinical trial, called the Earlier versus Later Levodopa Therapy in Parkinson Disease (ELLDOPA) study, was conceived, organized, and implemented by the Parkinson Study Group and sponsored by the National Institute of Neurological Disorders and Stroke. The Department of Defense sponsored the single-photon-emission computed-tomography (SPECT) substudy. The subjects were enrolled between September 1998 and August 2001 at 33 sites in the United States and 5 sites in Canada. The study was approved by the institutional review boards at the participating sites, and all subjects gave written informed consent. An independent safety monitoring committee monitored the data, subjects' safety, and the tolerability of the study drug. There was no prespecified formal guideline for recommending either modification or termination of the trial.

SUBJECTS

The subjects were 30 years of age or older, had received a diagnosis of Parkinson's disease within the past two years, had a rating on the modified Hoehn–Yahr scale13,14 of less than stage 3 (with stage 1 indicating unilateral disease, stage 2 mild bilateral disease, and stage 3 more advanced bilateral disease), and were considered not likely to require therapy for symptoms of the disease within the nine months after enrollment in the study. Patients were excluded if they were receiving antiparkinson medication, had been exposed to levodopa or to any dopamine agonist for more than 14 days, had an identifiable cause of parkinsonism, or had a tremor in any limb that was given a score of 3 or more on the Unified Parkinson's Disease Rating Scale (UPDRS),14 freezing of gait, loss of postural reflexes, major depression, or dementia.

Potential subjects were informed that only the assigned study drug would be permitted during the nine months of the study and that if they needed additional antiparkinson medication during this period, they would have to withdraw from the study. Subjects were randomly assigned to receive placebo or carbidopa–levodopa at a dose of 12.5 and 50 mg three times daily, 25 and 100 mg three times daily, or 50 and 200 mg three times daily, respectively. The doses were increased to the full amount over a period of nine weeks in a blinded fashion.

After 40 weeks, the subjects underwent a 3-day period of step-down withdrawal from the study drug. After two weeks without the study drug, a final assessment of the severity of the symptoms of Parkinson's disease was made. The selection of the 2-week duration for the washout period was based on reports that withdrawal from levodopa for a period up to 14 days resulted in a worsening of parkinsonism mainly within the first 7 days, with nonsignificant worsening beyond that point.15–17

CLINICAL EVAL‎UATION

The treating investigator, who was blinded to the treatment assignment, performed a clinical eval‎uation with the use of the UPDRS14 at the screening, baseline, and interim visits (at the end of weeks 3, 9, 24, and 40) and during each of the two weeks of the washout phase. During the four interim visits, the eval‎uations were performed before the administration of the first dose of the study drug. At every visit the treating investigator inquired about adverse events. The primary rater, who was also blinded to the treatment assignment and was kept unaware of information obtained during the course of the study, performed the examination with the use of the UPDRS only at the baseline eval‎uation, which took place within four weeks after screening, and again at the final eval‎uation, two weeks after the subject had undergone complete withdrawal from the study treatment (week 42). The study coordinators and the subjects also were blinded to the treatment assignments. An emergency unblinding procedure was established but was never used.

OUTCOME

The prespecified primary outcome was the change in the severity of parkinsonism between the baseline visit and week 42, as measured with the use of the total score on the UPDRS that was obtained by the primary rater; week 42 was 14 days after the withdrawal of the study drug. The expected result was a dose-related deterioration during washout if levodopa were shown to hasten the progression of Parkinson's disease. The change in the total scores on the UPDRS measured by the treating investigator at each visit was a prespecified secondary outcome. The treating investigator assessed adverse events with the use of open-ended questioning at each visit.

SUBSTUDY

After the clinical trial was begun, we conducted a substudy with the use of SPECT to measure striatal dopamine-transporter density with the use of iodine-123–labeled 2-β-carboxymethoxy-3-β-(4-iodophenyl)tropane ([123I]β-CIT). The methods have been reported previously.18 The subjects who gave consent underwent SPECT imaging just before the baseline visit and then again before the visit at week 40. The imaging studies were performed at that point rather than after the subjects had undergone withdrawal from levodopa, at week 42, because it was thought that the subjects would not be able to tolerate the procedure (including the necessary travel) after the return of parkinsonism or after its worsening. Another reason was that there was no clear evidence of a short-term effect of levodopa on dopamine-transporter imaging.

All imaging studies were performed at Yale University or the Institute for Neurodegenerative Disorders (both in New Haven, Connecticut). The neuroimaging personnel were blinded to the treatment assignment of the subjects. The results of the SPECT studies were transferred to and analyzed by the biostatistics center of the Parkinson Study Group. The prespecified outcome was the percent change in the ratio of the specific striatal [123I]β-CIT uptake to the nondisplaceable striatal [123I]β-CIT uptake between the two images.

STATISTICAL ANALYSIS

Allowing for a dropout rate of 10 percent of the subjects enrolled, we chose a sample size of 360 subjects (i.e., 90 subjects in each of the four treatment groups) in order to provide the study with 85 percent power to detect a dose–response relationship (linear trend) between the assigned doses and the change in the subjects' total score on the UPDRS between the baseline visit and the final visit at week 42. On the basis of the results of one study,19 we anticipated a rate of worsening in the total score on the UPDRS in the placebo group of 9.5 units over the 9.5 months of the study. The study was powered to detect a linear trend that corresponded to a 4-unit difference (45 percent of 9.5 units) in the score on the UPDRS between the highest dose of levodopa (600 mg per day) and placebo — that is, a change of either 13.5 or 5.5 units from baseline in the group receiving the highest dose of the active study drug.

The primary statistical tests were two-tailed, with an alpha level of 0.05. Only subjects who completed the two-week washout phase were included in this analysis. We used the intention-to-treat principle in the analysis, even if among some subjects the dose of the study drug were to be reduced during the study. The primary analysis assessed the dose–response relationship between the assigned doses and the worsening of parkinsonism, as indicated by the changes in the total score on the UPDRS between the baseline visit and week 42. Statistical comparisons were made by analysis of covariance in a model that adjusted for differences among the investigators performing the eval‎uations and in baseline values.

ResultsSUBJECTS

Of a total of 361 subjects enrolled in the study, 317 (88 percent) took the study medication for 40 weeks, and 311 (86 percent) completed the 2 weeks of washout (Figure 1). The neuroimaging substudy was begun after the enrollment of the first 108 subjects. Of the 253 subjects subsequently enrolled, 142 (56 percent) participated in the substudy and underwent the baseline SPECT. Of these, 135 (95 percent) returned for scanning at week 40. The demographic and clinical characteristics of the subjects in the treatment groups were similar at baseline, both in the entire sample and in the neuroimaging substudy (Table 1).

FIGURE 1

Random Assignments to Treatment, Completion of the Trial, and Reasons for Not Completing It.

TABLE 1

Baseline Characteristics of the Study Subjects.

CLINICAL OUTCOME

Levodopa, in a dose–response pattern, significantly (P<0.001) reduced the worsening of symptoms of Parkinson's disease as reflected in the change between the total score on the UPDRS at baseline and that at week 42 (i.e., two weeks after washout of the study medication), as compared with the change in the placebo group (Figure 2 and Table 2). The subjects in the placebo group had mild improvement at the week 3 visit, but after that their symptoms worsened steadily throughout the balance of the study period, including the two-week washout phase. A strong dose–response benefit was detected during the period in which the medication was administered beginning at week 9, when the full dose of 600 mg daily was reached in the group receiving the highest dose of levodopa, and it persisted through week 40. The scores on the UPDRS in the three levodopa groups worsened during the two-week washout period, but these groups did not deteriorate to the level observed in the placebo group, and the group receiving the highest dose of levodopa had the best result (Figure 2 and Table 2). The adverse events that were significantly more common among those receiving levodopa at 600 mg daily than in the placebo group were dyskinesias, nausea, infection, hypertonia, and headache (Table 3).

FIGURE 2

Changes in Total Scores on the Unified Parkinson's Disease Rating Scale (UPDRS) from Baseline through Eval‎uation at Week 42.

TABLE 2

Changes in the Scores on the UPDRS between Baseline and Week 42.

TABLE 3

Adverse Events.

SPECT AND [123I]Β-CIT SUBSTUDY

The mean [123I]β-CIT uptake in the striatum, caudate, and putamen at baseline was consistent with values previously reported for patients with early Parkinson's disease.20 The percent decrease in striatal [123I]β-CIT uptake over the 40 weeks of the study treatment was greater among subjects in the levodopa groups than in the placebo group, but this difference was not statistically significant (Table 4). However, 21 of the 142 subjects (14.7 percent) had a putaminal [123I]β-CIT uptake of more than 3.25 at baseline (i.e., more than 75 percent of the age-expected putaminal uptake).21 An analysis of the results of SPECT after the exclusion of the 19 subjects without a dopaminergic deficit who returned for the neuroimaging study at week 40 showed a significantly greater decrease in [123I]β-CIT uptake among those receiving levodopa than among those receiving placebo (P=0.036) (Table 4).

TABLE 4

Change in Striatal [123I]β-CIT Uptake between Baseline and Week 40.

Discussion

We found no clinical evidence that levodopa accelerated the worsening of Parkinson's disease over the 9.5 months of observation. Rather, levodopa was associated with less worsening of parkinsonism than was placebo, consistent with the notion that it slows disease progression (Table 2 and Figure 2). In comparison with the scores on the UPDRS at baseline, the final scores, after the washout phase, had worsened by approximately 8 units in the placebo group, whereas the group receiving levodopa at 600 mg daily did not have evidence of deterioration. The groups receiving levodopa at lower doses did have deterioration, as shown by a comparison of their scores at baseline and at week 42, but the deterioration was less than in the placebo group (P<0.001).

We need to consider that a two-week washout from levodopa may have been insufficient to eliminate fully the effect of the medication on symptoms, and the results observed may be related to a profound effect of levodopa on symptoms that persists for a long time after the drug has been withdrawn. Indeed, Hauser and Holford,22 using a modeling technique, analyzed the withdrawal of levodopa in 20 patients and reported that the mean half-life of levodopa as measured by the loss of the clinical benefit was 7.9 days (95 percent confidence interval, 2.2 to 30.4 days). They suggest that a washout period of 32 days (four half-lives) may be required to eliminate 90 percent of the drug's effects on symptoms. However, we saw little deterioration after one week of washout. Near the end of the study, with the approval of the institutional review boards and of the National Institutes of Health, we asked the last 38 subjects remaining in the study to extend the washout period to four weeks. Among these subjects there was no further worsening of the scores on the UPDRS during the additional two weeks, but the small number of subjects renders this component of the study difficult to interpret.

Muenter and Tyce23 described motor responses with two types of duration, short and long, with levodopa therapy. The short-duration benefit lasts for a few hours after a single dose, and the long-duration benefit lasts several days. In our study, the long-duration benefit appears to have lasted approximately one to two weeks (Figure 2). To argue that a longer washout period might have revealed more clinical worsening than was observed, one could propose a hitherto unknown third type of duration of motor response, one that is more sustained than the so-called long-duration benefit. Such an enduring benefit could be envisioned to result from a prolonged pharmacodynamic effect, for example, on dopamine receptors.

If, however, the clinical effects observed offer evidence of neuroprotection, how can we explain the greater loss of dopamine transporter shown in the SPECT imaging studies of the subjects who received levodopa, as compared with those who received the placebo — a result that suggests the possibility of a levodopa-induced toxic effect on dopamine neurons? At the end of the study, when the SPECT neuroimaging studies were performed, the subjects were still taking levodopa, so it is possible to assume that levodopa has a pharmacologic effect on the dopamine transporter that interferes with and reduces the binding of the β-CIT ligand.

Indeed, one study found a reduction in dopamine-transporter binding with the use of positron-emission tomography (PET) and another dopamine-transporter ligand in patients with early Parkinson's disease who were treated for six weeks with levodopa at 300 mg daily.24 However, the sample size in this study, as in others that have shown no change in dopamine-transporter binding after short-term treatment with levodopa,25–27 was too small to demonstrate a statistically significant difference between levodopa and placebo. Further indirect support for the absence of a pharmacologic effect of levodopa on imaging studies of the dopamine transporter in the present study was the absence of decline in the [123I]β-CIT uptake by week 40 among the 16 subjects who received levodopa and whose SPECT scans were normal at baseline (i.e., without evidence of a dopaminergic deficit) (data not shown). However, in the absence of studies with larger samples and a longer period of treatment with levodopa, we cannot exclude the possibility that levodopa may simply down-regulate the dopamine transporter. Another consideration is that the falling dopamine concentration in the placebo group may have led to a compensatory increase in the activity of the dopamine transporter that could have increased [123I]β-CIT binding, but evidence for such an interpretation is lacking.

If levodopa has neuroprotective effects, what mechanisms could account for this property? In low concentrations and in the presence of glial cells, levodopa protects cultured dopaminergic neurons28–31 and up-regulates antioxidant and antiapoptotic proteins.29,32 Furthermore, in vivo studies suggest that levodopa can promote survival and enhance the sprouting of nigral dopamine neurons in rodents treated with the toxin 6-hydroxydopamine.9,10 The question of whether levodopa has a protective or a toxic effect in Parkinson's disease — reflecting the results of clinical examination and neuroimaging studies, respectively — cannot be answered with certainty, and future studies will be needed.

This dose–response, placebo-controlled clinical trial eval‎uating the effect of levodopa in patients with early Parkinson's disease showed a strikingly impressive dose–response clinical benefit: the higher the dose, the stronger and more lasting the benefit, and the benefit was greater even after the drug was withdrawn (Figure 2).

During withdrawal, we did not encounter the neuroleptic malignant-like effect (i.e., high fever, obtundation, and rigidity) that can occur with the sudden withdrawal of levodopa.33,34 Its absence may be related to the down-titration procedure employed; however, this complication is so rare that it might not have occurred in a study of this size even had levodopa been withdrawn suddenly, particularly in patients with early-stage disease.

Our study calls into question interpretations of the functional neuroimaging of the dopamine system. It has generally been assumed that imaging studies performed with 18F-fluorodopa PET and [123I]β-CIT SPECT can provide reliable information on the integrity of the nigrostriatal dopamine pathway. Our study and other recent clinical trials25,35 raise the possibility of a pharmacologic influence of dopaminergic therapy on these neuroimaging targets and point to the need for clarification of this issue.

Although the absence of evidence of a dopaminergic deficit in the imaging in our study and two other studies25,35 may simply represent a limitation of the sensitivity of the imaging techniques, the scans without a dopaminergic deficit raise doubt about whether the subjects had Parkinson's disease.35 In our study, the subjects with such scans at baseline had no worsening of the [123I]β-CIT uptake at week 40 (data not shown), nor did subjects with such scans in the group receiving the highest dose of levodopa have an improvement in the scores on the UPDRS (i.e., a change between the score at baseline and at week 40 of 3.38±4.25 units, as compared with a change of –4.95±10.4 units in the subjects receiving levodopa at 600 mg daily whose baseline SPECT scans showed abnormalities; P=0.002). Whether the subjects with no evidence of dopaminergic deficit on scanning do or do not have classic Parkinson's disease remains uncertain. The fact that scans of some patients with early Parkinson's disease do not have evidence of a dopaminergic deficit needs to be taken into consideration in the planning of future trials to test drugs for neuroprotective effects.

Finally, even though we cannot reconcile the clinical and imaging findings in our study, we can assure both patients with early Parkinson's disease and their physicians that, from a clinical perspective, our study did not find that levodopa hastens the progression of Parkinson's disease. On the basis of the study, we can recommend that the doses of levodopa be adjusted to fit the needs of the patient. Small doses were found to be effective, although less so than higher doses. High doses, however, were associated with a greater frequency of adverse events such as dyskinesia. For the present, until more evidence is available, we recommend customizing the dose of levodopa to the needs of the individual patient on the basis of the clinical response and the profile of adverse events.

NOTES

Supported by grants from the National Institute of Neurological Disorders and Stroke (NS34796, to Dr. Fahn), the Department of Defense (DAMD 17-99-1-9472, to Dr. Marek), and the General Clinical Research Center of the National Center for Research Resources, National Institutes of Health (MO1-RR-00044 and MO1-RR-02066). Carbidopa–levodopa tablets and the matching placebo tablets were kindly provided by Teva Pharmaceuticals (Israel).

Drs. Fahn, Oakes, Shoulson, Kieburtz, Lang, Tanner, and Marek report having served as unpaid consultants to Teva Pharmaceuticals, and Dr. Olanow reports having served as a paid consultant to Teva Pharmaceuticals. Drs. Marek and Seibyl have an equity interest in Molecular Neuroimaging (New Haven), which carried out the [123I]β-CIT SPECT imaging for this study.

We are indebted to the subjects for their participation in the study; to K. Hyland (Baylor Medical College, Houston) for the independent, blinded chemical analyses of the contents of the tablets conducted annually to ascertain the chemical stability of levodopa and carbidopa; to S. Bennett, A. Brocht, D. Graffrath, J. Janciuras, C. Orme, L. Preston, K. Rothenburgh, C. Weaver, and A. Watts of the Biostatistics and Coordination Centers at the University of Rochester, Rochester, N.Y.; to the Independent Safety Monitoring Committee (P. Tariot, chair, Monroe Community Hospital, Rochester, N.Y.; to W.J. Hall, University of Rochester; to R. Rodnitzky, University of Iowa, Iowa City); to the National Institute of Neurological Disorders and Stroke Safety Monitoring Committee (E.C. Haley, chair, University of Virginia, Charlottesville; D. Eidelberg, North Shore University Medical Center, Manhasset, N.Y.; C.A. Gatsonis, Brown University, Providence, R.I.; W. Rocca, Mayo Clinic, Rochester, Minn.); and to E.J. Oliver (administrator), National Institute of Neurological Disorders and Stroke, Bethesda, Md.

APPENDIX

The following members of the Parkinson Study Group participated in the Earlier versus Later Levodopa Therapy in Parkinson Disease (ELLDOPA) study and contributed to this report. Steering committee: S. Fahn, principal investigator, Columbia University, New York; D. Oakes, chief biostatistician; I. Shoulson, coprincipal investigator; K. Kieburtz, director, Clinical Trials Coordination Center; A. Rudolph, senior project coordinator, University of Rochester, Rochester, N.Y.; K. Marek, neuroimager; J. Seibyl, neuroimager, Institute for Neurodegenerative Disorders, New Haven, Conn.; A. Lang, Toronto Western Hospital, Toronto; C.W. Olanow, Mount Sinai School of Medicine, New York; C. Tanner, The Parkinson's Institute, Sunnyvale, Calif.; G. Schifitto, medical monitor, University of Rochester, Rochester, N.Y.; H. Zhao, biostatistician, University of Rochester, Rochester, N.Y.; L. Reyes, administrator, Columbia University, New York; A. Shinaman, administrator, University of Rochester, Rochester, N.Y.; Participating primary raters, treating investigators, and coordinators: Rush–Presbyterian–St. Luke's Medical Center, Chicago: C. Comella, C. Goetz, L. Blasucci; Barrow Neurological Institute, Phoenix, Ariz.: J. Samanta, M. Stacy, K. Williamson, M. Harrigan; Columbia University, New York: P. Greene, B. Ford, C. Moskowitz; Parkinson's and Movement Disorder Institute, Fountain Valley, Calif.: D. Truong, M. Pathak; Baylor College of Medicine, Houston: J. Jankovic, W. Ondo, F. Atassi, C. Hunter; Brown University, Providence, R.I.: C. Jacques, J.H. Friedman, M. Lannon; Institute for Neurodegenerative Disorders, New Haven, Conn.: D.S. Russell, D. Jennings, B. Fussell; Massachusetts General Hospital, Boston: D. Standaert, M.A. Schwarzschild, J. Growdon, M. Tennis; McGill Centre for Studies in Aging, Verdun, Que., Canada: S. Gauthier, M. Panisset, J. Hall; Oregon Health and Science University, Portland: S. Gancher, J. Hammerstad, C. Stone, B. Alexander-Brown; Albany Medical College, Albany, N.Y.: S. Factor, E. Molho, D. Brown (deceased), S. Evans; Scott and White Hospital–Texas A&M University, Temple: J. Clark, B. Manyam, P. Simpson, B. Wulbrecht, J. Whetteckey; University of Alberta, Edmonton, Canada: W. Martin, T. Roberts, P. King; University of South Florida, Tampa: R. Hauser, T. Zesiewicz, L. Gauger; University of Virginia, Charlottesville: J. Trugman, G.F. Wooten, E. Rost-Ruffner; Washington University, St. Louis: J. Perlmutter, B. Racette; University of Calgary, Alta., Canada: O. Suchowersky, R. Ranawaya, S. Wood, C. Pantella; University of Rochester, Rochester, N.Y.: R. Kurlan, I. Richard, N. Pearson; Mayo Clinic, Scottsdale, Ariz.: J. Caviness, C. Adler, M. Lind; University of Pennsylvania, Philadelphia: T. Simuni, A. Siderowf, A. Colcher, M. Lloyd; University of Miami, Miami: W. Weiner, L. Shulman, W. Koller, K. Lyons; Boston University, Boston: R. Feldman (deceased), M.-H. St.-Hilaire, S. Ellias, C.-A. Thomas; Emory University, Atlanta: J. Juncos, R. Watts, A. Partlow; The Parkinson's Institute, Sunnyvale, Calif.: J. Tetrud, D.M. Togasaki, M. Welsh, T. Stewart; University of Medicine and Dentistry of New Jersey–Robert Wood Johnson, New Brunswick: M.H. Mark, J.I. Sage, D. Caputo; Louisiana State University, New Orleans: H. Gould, J. Rao, A. McKendrick; Mount Sinai School of Medicine, New York: M. Brin, F. Danisi, R. Benabou; Ohio State University, Columbus: J. Hubble, G. Paulson, C. Reider; Toronto Western Hospital, Toronto: A. Birnbaum, J. Miyasaki, L. Johnston, J. So; University of Kansas, Kansas City: R. Pahwa, R. Dubinsky; Mayo Clinic, Jacksonville, Fla.: Z. Wszolek, R. Uitti, M. Turk; Minneapolis Veterans Affairs Hospital, Minneapolis: P. Tuite, D. Rottenberg, J. Hansen; University of Puerto Rico, San Juan: C. Serrano Ramos; University of Southern California, Los Angeles: C. Waters, M. Lew, M. Welsh, C. Kawai; Colorado Neurological Institute, Englewood: C. O'Brien, R. Kumar, L. Seeberger, D. Judd; Ottawa Civic Hospital, Ottawa: T. Mendis, C.L. Barclay, D.A. Grimes, L. Sutherland; Johns Hopkins University, Baltimore: T. Dawson, S. Reich, R. Dunlop; University of Michigan, Ann Arbor: R. Albin, K. Frey, K. Wernette.

ORIGINAL ARTICLE

SHARE ON

•  
•  
•  
•  

Levodopa and the Progression of Parkinson's Disease

Author: The Parkinson Study Group*Author Info & Affiliations

Published December 9, 2004

N Engl J Med 2004;351:2498-2508

DOI: 10.1056/NEJMoa033447

VOL. 351 NO. 24

• • ABSTRACT
  • METHODS
  • RESULTS
  • DISCUSSION
  • NOTES
  • APPENDIX
  • REFERENCES
• • Information & Authors
  • Metrics & Citations
  • View Options
  • References
  • Media
  • Tables
  • Share

AbstractBACKGROUND

Despite the known benefit of levodopa in reducing the symptoms of Parkinson's disease, concern has been expressed that its use might hasten neurodegeneration. This study assessed the effect of levodopa on the rate of progression of Parkinson's disease.

METHODS

In this randomized, double-blind, placebo-controlled trial, we eval‎uated 361 patients with early Parkinson's disease who were assigned to receive carbidopa–levodopa at a daily dose of 37.5 and 150 mg, 75 and 300 mg, or 150 and 600 mg, respectively, or a matching placebo for a period of 40 weeks, and then to undergo withdrawal of treatment for 2 weeks. The primary outcome was a change in scores on the Unified Parkinson's Disease Rating Scale (UPDRS) between baseline and 42 weeks. Neuroimaging studies of 142 subjects were performed at baseline and at week 40 to assess striatal dopamine-transporter density with the use of iodine-123–labeled 2-β-carboxymethoxy-3-β-(4-iodophenyl)tropane ([123I]β-CIT) uptake.

RESULTS

The severity of parkinsonism increased more in the placebo group than in all the groups receiving levodopa: the mean difference between the total score on the UPDRS at baseline and at 42 weeks was 7.8 units in the placebo group, 1.9 units in the group receiving levodopa at a dose of 150 mg daily, 1.9 in those receiving 300 mg daily, and –1.4 in those receiving 600 mg daily (P<0.001). In contrast, in a substudy of 116 patients the mean percent decline in the [123I]β-CIT uptake was significantly greater with levodopa than placebo (–6 percent among those receiving levodopa at 150 mg daily, –4 percent in those receiving it at 300 mg daily, and –7.2 percent among those receiving it at 600 mg daily, as compared with –1.4 percent among those receiving placebo; 19 patients with no dopaminergic deficits on the baseline scans were excluded from the analysis) (P=0.036). The subjects receiving the highest dose of levodopa had significantly more dyskinesia, hypertonia, infection, headache, and nausea than those receiving placebo.

CONCLUSIONS

The clinical data suggest that levodopa either slows the progression of Parkinson's disease or has a prolonged effect on the symptoms of the disease. In contrast, the neuroimaging data suggest either that levodopa accelerates the loss of nigrostriatal dopamine nerve terminals or that its pharmacologic effects modify the dopamine transporter. The potential long-term effects of levodopa on Parkinson's disease remain uncertain.

Parkinson's disease is a progressively disabling neurodegenerative disorder that is manifested clinically by bradykinesia, tremor, rigidity, flexed posture, postural instability, and freezing of gait. It is characterized pathologically by the loss of pigmented dopaminergic neurons in the substantia nigra. The course of the clinical decline parallels that of the progressive degeneration of the remaining dopaminergic neurons.1 The use of levodopa as dopamine-replacement therapy is highly effective in ameliorating the symptoms of the disease and remains the standard drug with which other therapies are compared.2,3

Because levodopa and dopamine can generate reactive oxygen species and induce the degeneration of cultured dopamine neurons, concern has been raised that levodopa could enhance oxidative stress and hasten the degeneration of residual dopamine neurons in patients with Parkinson's disease.4–6 However, levodopa is not toxic in animals and may be trophic and promote the functional recovery of damaged nigral neurons.7–10 Humans without Parkinson's disease who are exposed to levodopa do not develop nigral damage,11,12 but such persons do not have increased oxidative stress in their substantia nigra neurons.

Whether levodopa is detrimental, beneficial, or without effect on the rate of the progression of Parkinson's disease is unknown and extremely important, both scientifically and clinically. We therefore conducted a controlled clinical trial to assess the effect of levodopa on the course of Parkinson's disease.

MethodsSTUDY DESIGN

Our multicenter, placebo-controlled, randomized, dose-ranging, double-blind clinical trial, called the Earlier versus Later Levodopa Therapy in Parkinson Disease (ELLDOPA) study, was conceived, organized, and implemented by the Parkinson Study Group and sponsored by the National Institute of Neurological Disorders and Stroke. The Department of Defense sponsored the single-photon-emission computed-tomography (SPECT) substudy. The subjects were enrolled between September 1998 and August 2001 at 33 sites in the United States and 5 sites in Canada. The study was approved by the institutional review boards at the participating sites, and all subjects gave written informed consent. An independent safety monitoring committee monitored the data, subjects' safety, and the tolerability of the study drug. There was no prespecified formal guideline for recommending either modification or termination of the trial.

SUBJECTS

The subjects were 30 years of age or older, had received a diagnosis of Parkinson's disease within the past two years, had a rating on the modified Hoehn–Yahr scale13,14 of less than stage 3 (with stage 1 indicating unilateral disease, stage 2 mild bilateral disease, and stage 3 more advanced bilateral disease), and were considered not likely to require therapy for symptoms of the disease within the nine months after enrollment in the study. Patients were excluded if they were receiving antiparkinson medication, had been exposed to levodopa or to any dopamine agonist for more than 14 days, had an identifiable cause of parkinsonism, or had a tremor in any limb that was given a score of 3 or more on the Unified Parkinson's Disease Rating Scale (UPDRS),14 freezing of gait, loss of postural reflexes, major depression, or dementia.

Potential subjects were informed that only the assigned study drug would be permitted during the nine months of the study and that if they needed additional antiparkinson medication during this period, they would have to withdraw from the study. Subjects were randomly assigned to receive placebo or carbidopa–levodopa at a dose of 12.5 and 50 mg three times daily, 25 and 100 mg three times daily, or 50 and 200 mg three times daily, respectively. The doses were increased to the full amount over a period of nine weeks in a blinded fashion.

After 40 weeks, the subjects underwent a 3-day period of step-down withdrawal from the study drug. After two weeks without the study drug, a final assessment of the severity of the symptoms of Parkinson's disease was made. The selection of the 2-week duration for the washout period was based on reports that withdrawal from levodopa for a period up to 14 days resulted in a worsening of parkinsonism mainly within the first 7 days, with nonsignificant worsening beyond that point.15–17

CLINICAL EVAL‎UATION

The treating investigator, who was blinded to the treatment assignment, performed a clinical eval‎uation with the use of the UPDRS14 at the screening, baseline, and interim visits (at the end of weeks 3, 9, 24, and 40) and during each of the two weeks of the washout phase. During the four interim visits, the eval‎uations were performed before the administration of the first dose of the study drug. At every visit the treating investigator inquired about adverse events. The primary rater, who was also blinded to the treatment assignment and was kept unaware of information obtained during the course of the study, performed the examination with the use of the UPDRS only at the baseline eval‎uation, which took place within four weeks after screening, and again at the final eval‎uation, two weeks after the subject had undergone complete withdrawal from the study treatment (week 42). The study coordinators and the subjects also were blinded to the treatment assignments. An emergency unblinding procedure was established but was never used.

OUTCOME

The prespecified primary outcome was the change in the severity of parkinsonism between the baseline visit and week 42, as measured with the use of the total score on the UPDRS that was obtained by the primary rater; week 42 was 14 days after the withdrawal of the study drug. The expected result was a dose-related deterioration during washout if levodopa were shown to hasten the progression of Parkinson's disease. The change in the total scores on the UPDRS measured by the treating investigator at each visit was a prespecified secondary outcome. The treating investigator assessed adverse events with the use of open-ended questioning at each visit.

SUBSTUDY

After the clinical trial was begun, we conducted a substudy with the use of SPECT to measure striatal dopamine-transporter density with the use of iodine-123–labeled 2-β-carboxymethoxy-3-β-(4-iodophenyl)tropane ([123I]β-CIT). The methods have been reported previously.18 The subjects who gave consent underwent SPECT imaging just before the baseline visit and then again before the visit at week 40. The imaging studies were performed at that point rather than after the subjects had undergone withdrawal from levodopa, at week 42, because it was thought that the subjects would not be able to tolerate the procedure (including the necessary travel) after the return of parkinsonism or after its worsening. Another reason was that there was no clear evidence of a short-term effect of levodopa on dopamine-transporter imaging.

All imaging studies were performed at Yale University or the Institute for Neurodegenerative Disorders (both in New Haven, Connecticut). The neuroimaging personnel were blinded to the treatment assignment of the subjects. The results of the SPECT studies were transferred to and analyzed by the biostatistics center of the Parkinson Study Group. The prespecified outcome was the percent change in the ratio of the specific striatal [123I]β-CIT uptake to the nondisplaceable striatal [123I]β-CIT uptake between the two images.

STATISTICAL ANALYSIS

Allowing for a dropout rate of 10 percent of the subjects enrolled, we chose a sample size of 360 subjects (i.e., 90 subjects in each of the four treatment groups) in order to provide the study with 85 percent power to detect a dose–response relationship (linear trend) between the assigned doses and the change in the subjects' total score on the UPDRS between the baseline visit and the final visit at week 42. On the basis of the results of one study,19 we anticipated a rate of worsening in the total score on the UPDRS in the placebo group of 9.5 units over the 9.5 months of the study. The study was powered to detect a linear trend that corresponded to a 4-unit difference (45 percent of 9.5 units) in the score on the UPDRS between the highest dose of levodopa (600 mg per day) and placebo — that is, a change of either 13.5 or 5.5 units from baseline in the group receiving the highest dose of the active study drug.

The primary statistical tests were two-tailed, with an alpha level of 0.05. Only subjects who completed the two-week washout phase were included in this analysis. We used the intention-to-treat principle in the analysis, even if among some subjects the dose of the study drug were to be reduced during the study. The primary analysis assessed the dose–response relationship between the assigned doses and the worsening of parkinsonism, as indicated by the changes in the total score on the UPDRS between the baseline visit and week 42. Statistical comparisons were made by analysis of covariance in a model that adjusted for differences among the investigators performing the eval‎uations and in baseline values.

ResultsSUBJECTS

Of a total of 361 subjects enrolled in the study, 317 (88 percent) took the study medication for 40 weeks, and 311 (86 percent) completed the 2 weeks of washout (Figure 1). The neuroimaging substudy was begun after the enrollment of the first 108 subjects. Of the 253 subjects subsequently enrolled, 142 (56 percent) participated in the substudy and underwent the baseline SPECT. Of these, 135 (95 percent) returned for scanning at week 40. The demographic and clinical characteristics of the subjects in the treatment groups were similar at baseline, both in the entire sample and in the neuroimaging substudy (Table 1).

FIGURE 1

Random Assignments to Treatment, Completion of the Trial, and Reasons for Not Completing It.

TABLE 1

Baseline Characteristics of the Study Subjects.

CLINICAL OUTCOME

Levodopa, in a dose–response pattern, significantly (P<0.001) reduced the worsening of symptoms of Parkinson's disease as reflected in the change between the total score on the UPDRS at baseline and that at week 42 (i.e., two weeks after washout of the study medication), as compared with the change in the placebo group (Figure 2 and Table 2). The subjects in the placebo group had mild improvement at the week 3 visit, but after that their symptoms worsened steadily throughout the balance of the study period, including the two-week washout phase. A strong dose–response benefit was detected during the period in which the medication was administered beginning at week 9, when the full dose of 600 mg daily was reached in the group receiving the highest dose of levodopa, and it persisted through week 40. The scores on the UPDRS in the three levodopa groups worsened during the two-week washout period, but these groups did not deteriorate to the level observed in the placebo group, and the group receiving the highest dose of levodopa had the best result (Figure 2 and Table 2). The adverse events that were significantly more common among those receiving levodopa at 600 mg daily than in the placebo group were dyskinesias, nausea, infection, hypertonia, and headache (Table 3).

FIGURE 2

Changes in Total Scores on the Unified Parkinson's Disease Rating Scale (UPDRS) from Baseline through Eval‎uation at Week 42.

TABLE 2

Changes in the Scores on the UPDRS between Baseline and Week 42.

TABLE 3

Adverse Events.

SPECT AND [123I]Β-CIT SUBSTUDY

The mean [123I]β-CIT uptake in the striatum, caudate, and putamen at baseline was consistent with values previously reported for patients with early Parkinson's disease.20 The percent decrease in striatal [123I]β-CIT uptake over the 40 weeks of the study treatment was greater among subjects in the levodopa groups than in the placebo group, but this difference was not statistically significant (Table 4). However, 21 of the 142 subjects (14.7 percent) had a putaminal [123I]β-CIT uptake of more than 3.25 at baseline (i.e., more than 75 percent of the age-expected putaminal uptake).21 An analysis of the results of SPECT after the exclusion of the 19 subjects without a dopaminergic deficit who returned for the neuroimaging study at week 40 showed a significantly greater decrease in [123I]β-CIT uptake among those receiving levodopa than among those receiving placebo (P=0.036) (Table 4).

TABLE 4

Change in Striatal [123I]β-CIT Uptake between Baseline and Week 40.

Discussion

We found no clinical evidence that levodopa accelerated the worsening of Parkinson's disease over the 9.5 months of observation. Rather, levodopa was associated with less worsening of parkinsonism than was placebo, consistent with the notion that it slows disease progression (Table 2 and Figure 2). In comparison with the scores on the UPDRS at baseline, the final scores, after the washout phase, had worsened by approximately 8 units in the placebo group, whereas the group receiving levodopa at 600 mg daily did not have evidence of deterioration. The groups receiving levodopa at lower doses did have deterioration, as shown by a comparison of their scores at baseline and at week 42, but the deterioration was less than in the placebo group (P<0.001).

We need to consider that a two-week washout from levodopa may have been insufficient to eliminate fully the effect of the medication on symptoms, and the results observed may be related to a profound effect of levodopa on symptoms that persists for a long time after the drug has been withdrawn. Indeed, Hauser and Holford,22 using a modeling technique, analyzed the withdrawal of levodopa in 20 patients and reported that the mean half-life of levodopa as measured by the loss of the clinical benefit was 7.9 days (95 percent confidence interval, 2.2 to 30.4 days). They suggest that a washout period of 32 days (four half-lives) may be required to eliminate 90 percent of the drug's effects on symptoms. However, we saw little deterioration after one week of washout. Near the end of the study, with the approval of the institutional review boards and of the National Institutes of Health, we asked the last 38 subjects remaining in the study to extend the washout period to four weeks. Among these subjects there was no further worsening of the scores on the UPDRS during the additional two weeks, but the small number of subjects renders this component of the study difficult to interpret.

Muenter and Tyce23 described motor responses with two types of duration, short and long, with levodopa therapy. The short-duration benefit lasts for a few hours after a single dose, and the long-duration benefit lasts several days. In our study, the long-duration benefit appears to have lasted approximately one to two weeks (Figure 2). To argue that a longer washout period might have revealed more clinical worsening than was observed, one could propose a hitherto unknown third type of duration of motor response, one that is more sustained than the so-called long-duration benefit. Such an enduring benefit could be envisioned to result from a prolonged pharmacodynamic effect, for example, on dopamine receptors.

If, however, the clinical effects observed offer evidence of neuroprotection, how can we explain the greater loss of dopamine transporter shown in the SPECT imaging studies of the subjects who received levodopa, as compared with those who received the placebo — a result that suggests the possibility of a levodopa-induced toxic effect on dopamine neurons? At the end of the study, when the SPECT neuroimaging studies were performed, the subjects were still taking levodopa, so it is possible to assume that levodopa has a pharmacologic effect on the dopamine transporter that interferes with and reduces the binding of the β-CIT ligand.

Indeed, one study found a reduction in dopamine-transporter binding with the use of positron-emission tomography (PET) and another dopamine-transporter ligand in patients with early Parkinson's disease who were treated for six weeks with levodopa at 300 mg daily.24 However, the sample size in this study, as in others that have shown no change in dopamine-transporter binding after short-term treatment with levodopa,25–27 was too small to demonstrate a statistically significant difference between levodopa and placebo. Further indirect support for the absence of a pharmacologic effect of levodopa on imaging studies of the dopamine transporter in the present study was the absence of decline in the [123I]β-CIT uptake by week 40 among the 16 subjects who received levodopa and whose SPECT scans were normal at baseline (i.e., without evidence of a dopaminergic deficit) (data not shown). However, in the absence of studies with larger samples and a longer period of treatment with levodopa, we cannot exclude the possibility that levodopa may simply down-regulate the dopamine transporter. Another consideration is that the falling dopamine concentration in the placebo group may have led to a compensatory increase in the activity of the dopamine transporter that could have increased [123I]β-CIT binding, but evidence for such an interpretation is lacking.

If levodopa has neuroprotective effects, what mechanisms could account for this property? In low concentrations and in the presence of glial cells, levodopa protects cultured dopaminergic neurons28–31 and up-regulates antioxidant and antiapoptotic proteins.29,32 Furthermore, in vivo studies suggest that levodopa can promote survival and enhance the sprouting of nigral dopamine neurons in rodents treated with the toxin 6-hydroxydopamine.9,10 The question of whether levodopa has a protective or a toxic effect in Parkinson's disease — reflecting the results of clinical examination and neuroimaging studies, respectively — cannot be answered with certainty, and future studies will be needed.

This dose–response, placebo-controlled clinical trial eval‎uating the effect of levodopa in patients with early Parkinson's disease showed a strikingly impressive dose–response clinical benefit: the higher the dose, the stronger and more lasting the benefit, and the benefit was greater even after the drug was withdrawn (Figure 2).

During withdrawal, we did not encounter the neuroleptic malignant-like effect (i.e., high fever, obtundation, and rigidity) that can occur with the sudden withdrawal of levodopa.33,34 Its absence may be related to the down-titration procedure employed; however, this complication is so rare that it might not have occurred in a study of this size even had levodopa been withdrawn suddenly, particularly in patients with early-stage disease.

Our study calls into question interpretations of the functional neuroimaging of the dopamine system. It has generally been assumed that imaging studies performed with 18F-fluorodopa PET and [123I]β-CIT SPECT can provide reliable information on the integrity of the nigrostriatal dopamine pathway. Our study and other recent clinical trials25,35 raise the possibility of a pharmacologic influence of dopaminergic therapy on these neuroimaging targets and point to the need for clarification of this issue.

Although the absence of evidence of a dopaminergic deficit in the imaging in our study and two other studies25,35 may simply represent a limitation of the sensitivity of the imaging techniques, the scans without a dopaminergic deficit raise doubt about whether the subjects had Parkinson's disease.35 In our study, the subjects with such scans at baseline had no worsening of the [123I]β-CIT uptake at week 40 (data not shown), nor did subjects with such scans in the group receiving the highest dose of levodopa have an improvement in the scores on the UPDRS (i.e., a change between the score at baseline and at week 40 of 3.38±4.25 units, as compared with a change of –4.95±10.4 units in the subjects receiving levodopa at 600 mg daily whose baseline SPECT scans showed abnormalities; P=0.002). Whether the subjects with no evidence of dopaminergic deficit on scanning do or do not have classic Parkinson's disease remains uncertain. The fact that scans of some patients with early Parkinson's disease do not have evidence of a dopaminergic deficit needs to be taken into consideration in the planning of future trials to test drugs for neuroprotective effects.

Finally, even though we cannot reconcile the clinical and imaging findings in our study, we can assure both patients with early Parkinson's disease and their physicians that, from a clinical perspective, our study did not find that levodopa hastens the progression of Parkinson's disease. On the basis of the study, we can recommend that the doses of levodopa be adjusted to fit the needs of the patient. Small doses were found to be effective, although less so than higher doses. High doses, however, were associated with a greater frequency of adverse events such as dyskinesia. For the present, until more evidence is available, we recommend customizing the dose of levodopa to the needs of the individual patient on the basis of the clinical response and the profile of adverse events.

NOTES

Supported by grants from the National Institute of Neurological Disorders and Stroke (NS34796, to Dr. Fahn), the Department of Defense (DAMD 17-99-1-9472, to Dr. Marek), and the General Clinical Research Center of the National Center for Research Resources, National Institutes of Health (MO1-RR-00044 and MO1-RR-02066). Carbidopa–levodopa tablets and the matching placebo tablets were kindly provided by Teva Pharmaceuticals (Israel).

Drs. Fahn, Oakes, Shoulson, Kieburtz, Lang, Tanner, and Marek report having served as unpaid consultants to Teva Pharmaceuticals, and Dr. Olanow reports having served as a paid consultant to Teva Pharmaceuticals. Drs. Marek and Seibyl have an equity interest in Molecular Neuroimaging (New Haven), which carried out the [123I]β-CIT SPECT imaging for this study.

We are indebted to the subjects for their participation in the study; to K. Hyland (Baylor Medical College, Houston) for the independent, blinded chemical analyses of the contents of the tablets conducted annually to ascertain the chemical stability of levodopa and carbidopa; to S. Bennett, A. Brocht, D. Graffrath, J. Janciuras, C. Orme, L. Preston, K. Rothenburgh, C. Weaver, and A. Watts of the Biostatistics and Coordination Centers at the University of Rochester, Rochester, N.Y.; to the Independent Safety Monitoring Committee (P. Tariot, chair, Monroe Community Hospital, Rochester, N.Y.; to W.J. Hall, University of Rochester; to R. Rodnitzky, University of Iowa, Iowa City); to the National Institute of Neurological Disorders and Stroke Safety Monitoring Committee (E.C. Haley, chair, University of Virginia, Charlottesville; D. Eidelberg, North Shore University Medical Center, Manhasset, N.Y.; C.A. Gatsonis, Brown University, Providence, R.I.; W. Rocca, Mayo Clinic, Rochester, Minn.); and to E.J. Oliver (administrator), National Institute of Neurological Disorders and Stroke, Bethesda, Md.

APPENDIX

The following members of the Parkinson Study Group participated in the Earlier versus Later Levodopa Therapy in Parkinson Disease (ELLDOPA) study and contributed to this report. Steering committee: S. Fahn, principal investigator, Columbia University, New York; D. Oakes, chief biostatistician; I. Shoulson, coprincipal investigator; K. Kieburtz, director, Clinical Trials Coordination Center; A. Rudolph, senior project coordinator, University of Rochester, Rochester, N.Y.; K. Marek, neuroimager; J. Seibyl, neuroimager, Institute for Neurodegenerative Disorders, New Haven, Conn.; A. Lang, Toronto Western Hospital, Toronto; C.W. Olanow, Mount Sinai School of Medicine, New York; C. Tanner, The Parkinson's Institute, Sunnyvale, Calif.; G. Schifitto, medical monitor, University of Rochester, Rochester, N.Y.; H. Zhao, biostatistician, University of Rochester, Rochester, N.Y.; L. Reyes, administrator, Columbia University, New York; A. Shinaman, administrator, University of Rochester, Rochester, N.Y.; Participating primary raters, treating investigators, and coordinators: Rush–Presbyterian–St. Luke's Medical Center, Chicago: C. Comella, C. Goetz, L. Blasucci; Barrow Neurological Institute, Phoenix, Ariz.: J. Samanta, M. Stacy, K. Williamson, M. Harrigan; Columbia University, New York: P. Greene, B. Ford, C. Moskowitz; Parkinson's and Movement Disorder Institute, Fountain Valley, Calif.: D. Truong, M. Pathak; Baylor College of Medicine, Houston: J. Jankovic, W. Ondo, F. Atassi, C. Hunter; Brown University, Providence, R.I.: C. Jacques, J.H. Friedman, M. Lannon; Institute for Neurodegenerative Disorders, New Haven, Conn.: D.S. Russell, D. Jennings, B. Fussell; Massachusetts General Hospital, Boston: D. Standaert, M.A. Schwarzschild, J. Growdon, M. Tennis; McGill Centre for Studies in Aging, Verdun, Que., Canada: S. Gauthier, M. Panisset, J. Hall; Oregon Health and Science University, Portland: S. Gancher, J. Hammerstad, C. Stone, B. Alexander-Brown; Albany Medical College, Albany, N.Y.: S. Factor, E. Molho, D. Brown (deceased), S. Evans; Scott and White Hospital–Texas A&M University, Temple: J. Clark, B. Manyam, P. Simpson, B. Wulbrecht, J. Whetteckey; University of Alberta, Edmonton, Canada: W. Martin, T. Roberts, P. King; University of South Florida, Tampa: R. Hauser, T. Zesiewicz, L. Gauger; University of Virginia, Charlottesville: J. Trugman, G.F. Wooten, E. Rost-Ruffner; Washington University, St. Louis: J. Perlmutter, B. Racette; University of Calgary, Alta., Canada: O. Suchowersky, R. Ranawaya, S. Wood, C. Pantella; University of Rochester, Rochester, N.Y.: R. Kurlan, I. Richard, N. Pearson; Mayo Clinic, Scottsdale, Ariz.: J. Caviness, C. Adler, M. Lind; University of Pennsylvania, Philadelphia: T. Simuni, A. Siderowf, A. Colcher, M. Lloyd; University of Miami, Miami: W. Weiner, L. Shulman, W. Koller, K. Lyons; Boston University, Boston: R. Feldman (deceased), M.-H. St.-Hilaire, S. Ellias, C.-A. Thomas; Emory University, Atlanta: J. Juncos, R. Watts, A. Partlow; The Parkinson's Institute, Sunnyvale, Calif.: J. Tetrud, D.M. Togasaki, M. Welsh, T. Stewart; University of Medicine and Dentistry of New Jersey–Robert Wood Johnson, New Brunswick: M.H. Mark, J.I. Sage, D. Caputo; Louisiana State University, New Orleans: H. Gould, J. Rao, A. McKendrick; Mount Sinai School of Medicine, New York: M. Brin, F. Danisi, R. Benabou; Ohio State University, Columbus: J. Hubble, G. Paulson, C. Reider; Toronto Western Hospital, Toronto: A. Birnbaum, J. Miyasaki, L. Johnston, J. So; University of Kansas, Kansas City: R. Pahwa, R. Dubinsky; Mayo Clinic, Jacksonville, Fla.: Z. Wszolek, R. Uitti, M. Turk; Minneapolis Veterans Affairs Hospital, Minneapolis: P. Tuite, D. Rottenberg, J. Hansen; University of Puerto Rico, San Juan: C. Serrano Ramos; University of Southern California, Los Angeles: C. Waters, M. Lew, M. Welsh, C. Kawai; Colorado Neurological Institute, Englewood: C. O'Brien, R. Kumar, L. Seeberger, D. Judd; Ottawa Civic Hospital, Ottawa: T. Mendis, C.L. Barclay, D.A. Grimes, L. Sutherland; Johns Hopkins University, Baltimore: T. Dawson, S. Reich, R. Dunlop; University of Michigan, Ann Arbor: R. Albin, K. Frey, K. Wernette.

ORIGINAL ARTICLE

SHARE ON

•  
•  
•  
•  

Levodopa and the Progression of Parkinson's Disease

Author: The Parkinson Study Group*Author Info & Affiliations

Published December 9, 2004

N Engl J Med 2004;351:2498-2508

DOI: 10.1056/NEJMoa033447

VOL. 351 NO. 24

• • ABSTRACT
  • METHODS
  • RESULTS
  • DISCUSSION
  • NOTES
  • APPENDIX
  • REFERENCES
• • Information & Authors
  • Metrics & Citations
  • View Options
  • References
  • Media
  • Tables
  • Share

AbstractBACKGROUND

Despite the known benefit of levodopa in reducing the symptoms of Parkinson's disease, concern has been expressed that its use might hasten neurodegeneration. This study assessed the effect of levodopa on the rate of progression of Parkinson's disease.

METHODS

In this randomized, double-blind, placebo-controlled trial, we eval‎uated 361 patients with early Parkinson's disease who were assigned to receive carbidopa–levodopa at a daily dose of 37.5 and 150 mg, 75 and 300 mg, or 150 and 600 mg, respectively, or a matching placebo for a period of 40 weeks, and then to undergo withdrawal of treatment for 2 weeks. The primary outcome was a change in scores on the Unified Parkinson's Disease Rating Scale (UPDRS) between baseline and 42 weeks. Neuroimaging studies of 142 subjects were performed at baseline and at week 40 to assess striatal dopamine-transporter density with the use of iodine-123–labeled 2-β-carboxymethoxy-3-β-(4-iodophenyl)tropane ([123I]β-CIT) uptake.

RESULTS

The severity of parkinsonism increased more in the placebo group than in all the groups receiving levodopa: the mean difference between the total score on the UPDRS at baseline and at 42 weeks was 7.8 units in the placebo group, 1.9 units in the group receiving levodopa at a dose of 150 mg daily, 1.9 in those receiving 300 mg daily, and –1.4 in those receiving 600 mg daily (P<0.001). In contrast, in a substudy of 116 patients the mean percent decline in the [123I]β-CIT uptake was significantly greater with levodopa than placebo (–6 percent among those receiving levodopa at 150 mg daily, –4 percent in those receiving it at 300 mg daily, and –7.2 percent among those receiving it at 600 mg daily, as compared with –1.4 percent among those receiving placebo; 19 patients with no dopaminergic deficits on the baseline scans were excluded from the analysis) (P=0.036). The subjects receiving the highest dose of levodopa had significantly more dyskinesia, hypertonia, infection, headache, and nausea than those receiving placebo.

CONCLUSIONS

The clinical data suggest that levodopa either slows the progression of Parkinson's disease or has a prolonged effect on the symptoms of the disease. In contrast, the neuroimaging data suggest either that levodopa accelerates the loss of nigrostriatal dopamine nerve terminals or that its pharmacologic effects modify the dopamine transporter. The potential long-term effects of levodopa on Parkinson's disease remain uncertain.

Parkinson's disease is a progressively disabling neurodegenerative disorder that is manifested clinically by bradykinesia, tremor, rigidity, flexed posture, postural instability, and freezing of gait. It is characterized pathologically by the loss of pigmented dopaminergic neurons in the substantia nigra. The course of the clinical decline parallels that of the progressive degeneration of the remaining dopaminergic neurons.1 The use of levodopa as dopamine-replacement therapy is highly effective in ameliorating the symptoms of the disease and remains the standard drug with which other therapies are compared.2,3

Because levodopa and dopamine can generate reactive oxygen species and induce the degeneration of cultured dopamine neurons, concern has been raised that levodopa could enhance oxidative stress and hasten the degeneration of residual dopamine neurons in patients with Parkinson's disease.4–6 However, levodopa is not toxic in animals and may be trophic and promote the functional recovery of damaged nigral neurons.7–10 Humans without Parkinson's disease who are exposed to levodopa do not develop nigral damage,11,12 but such persons do not have increased oxidative stress in their substantia nigra neurons.

Whether levodopa is detrimental, beneficial, or without effect on the rate of the progression of Parkinson's disease is unknown and extremely important, both scientifically and clinically. We therefore conducted a controlled clinical trial to assess the effect of levodopa on the course of Parkinson's disease.

MethodsSTUDY DESIGN

Our multicenter, placebo-controlled, randomized, dose-ranging, double-blind clinical trial, called the Earlier versus Later Levodopa Therapy in Parkinson Disease (ELLDOPA) study, was conceived, organized, and implemented by the Parkinson Study Group and sponsored by the National Institute of Neurological Disorders and Stroke. The Department of Defense sponsored the single-photon-emission computed-tomography (SPECT) substudy. The subjects were enrolled between September 1998 and August 2001 at 33 sites in the United States and 5 sites in Canada. The study was approved by the institutional review boards at the participating sites, and all subjects gave written informed consent. An independent safety monitoring committee monitored the data, subjects' safety, and the tolerability of the study drug. There was no prespecified formal guideline for recommending either modification or termination of the trial.

SUBJECTS

The subjects were 30 years of age or older, had received a diagnosis of Parkinson's disease within the past two years, had a rating on the modified Hoehn–Yahr scale13,14 of less than stage 3 (with stage 1 indicating unilateral disease, stage 2 mild bilateral disease, and stage 3 more advanced bilateral disease), and were considered not likely to require therapy for symptoms of the disease within the nine months after enrollment in the study. Patients were excluded if they were receiving antiparkinson medication, had been exposed to levodopa or to any dopamine agonist for more than 14 days, had an identifiable cause of parkinsonism, or had a tremor in any limb that was given a score of 3 or more on the Unified Parkinson's Disease Rating Scale (UPDRS),14 freezing of gait, loss of postural reflexes, major depression, or dementia.

Potential subjects were informed that only the assigned study drug would be permitted during the nine months of the study and that if they needed additional antiparkinson medication during this period, they would have to withdraw from the study. Subjects were randomly assigned to receive placebo or carbidopa–levodopa at a dose of 12.5 and 50 mg three times daily, 25 and 100 mg three times daily, or 50 and 200 mg three times daily, respectively. The doses were increased to the full amount over a period of nine weeks in a blinded fashion.

After 40 weeks, the subjects underwent a 3-day period of step-down withdrawal from the study drug. After two weeks without the study drug, a final assessment of the severity of the symptoms of Parkinson's disease was made. The selection of the 2-week duration for the washout period was based on reports that withdrawal from levodopa for a period up to 14 days resulted in a worsening of parkinsonism mainly within the first 7 days, with nonsignificant worsening beyond that point.15–17

CLINICAL EVAL‎UATION

The treating investigator, who was blinded to the treatment assignment, performed a clinical eval‎uation with the use of the UPDRS14 at the screening, baseline, and interim visits (at the end of weeks 3, 9, 24, and 40) and during each of the two weeks of the washout phase. During the four interim visits, the eval‎uations were performed before the administration of the first dose of the study drug. At every visit the treating investigator inquired about adverse events. The primary rater, who was also blinded to the treatment assignment and was kept unaware of information obtained during the course of the study, performed the examination with the use of the UPDRS only at the baseline eval‎uation, which took place within four weeks after screening, and again at the final eval‎uation, two weeks after the subject had undergone complete withdrawal from the study treatment (week 42). The study coordinators and the subjects also were blinded to the treatment assignments. An emergency unblinding procedure was established but was never used.

OUTCOME

The prespecified primary outcome was the change in the severity of parkinsonism between the baseline visit and week 42, as measured with the use of the total score on the UPDRS that was obtained by the primary rater; week 42 was 14 days after the withdrawal of the study drug. The expected result was a dose-related deterioration during washout if levodopa were shown to hasten the progression of Parkinson's disease. The change in the total scores on the UPDRS measured by the treating investigator at each visit was a prespecified secondary outcome. The treating investigator assessed adverse events with the use of open-ended questioning at each visit.

SUBSTUDY

After the clinical trial was begun, we conducted a substudy with the use of SPECT to measure striatal dopamine-transporter density with the use of iodine-123–labeled 2-β-carboxymethoxy-3-β-(4-iodophenyl)tropane ([123I]β-CIT). The methods have been reported previously.18 The subjects who gave consent underwent SPECT imaging just before the baseline visit and then again before the visit at week 40. The imaging studies were performed at that point rather than after the subjects had undergone withdrawal from levodopa, at week 42, because it was thought that the subjects would not be able to tolerate the procedure (including the necessary travel) after the return of parkinsonism or after its worsening. Another reason was that there was no clear evidence of a short-term effect of levodopa on dopamine-transporter imaging.

All imaging studies were performed at Yale University or the Institute for Neurodegenerative Disorders (both in New Haven, Connecticut). The neuroimaging personnel were blinded to the treatment assignment of the subjects. The results of the SPECT studies were transferred to and analyzed by the biostatistics center of the Parkinson Study Group. The prespecified outcome was the percent change in the ratio of the specific striatal [123I]β-CIT uptake to the nondisplaceable striatal [123I]β-CIT uptake between the two images.

STATISTICAL ANALYSIS

Allowing for a dropout rate of 10 percent of the subjects enrolled, we chose a sample size of 360 subjects (i.e., 90 subjects in each of the four treatment groups) in order to provide the study with 85 percent power to detect a dose–response relationship (linear trend) between the assigned doses and the change in the subjects' total score on the UPDRS between the baseline visit and the final visit at week 42. On the basis of the results of one study,19 we anticipated a rate of worsening in the total score on the UPDRS in the placebo group of 9.5 units over the 9.5 months of the study. The study was powered to detect a linear trend that corresponded to a 4-unit difference (45 percent of 9.5 units) in the score on the UPDRS between the highest dose of levodopa (600 mg per day) and placebo — that is, a change of either 13.5 or 5.5 units from baseline in the group receiving the highest dose of the active study drug.

The primary statistical tests were two-tailed, with an alpha level of 0.05. Only subjects who completed the two-week washout phase were included in this analysis. We used the intention-to-treat principle in the analysis, even if among some subjects the dose of the study drug were to be reduced during the study. The primary analysis assessed the dose–response relationship between the assigned doses and the worsening of parkinsonism, as indicated by the changes in the total score on the UPDRS between the baseline visit and week 42. Statistical comparisons were made by analysis of covariance in a model that adjusted for differences among the investigators performing the eval‎uations and in baseline values.

ResultsSUBJECTS

Of a total of 361 subjects enrolled in the study, 317 (88 percent) took the study medication for 40 weeks, and 311 (86 percent) completed the 2 weeks of washout (Figure 1). The neuroimaging substudy was begun after the enrollment of the first 108 subjects. Of the 253 subjects subsequently enrolled, 142 (56 percent) participated in the substudy and underwent the baseline SPECT. Of these, 135 (95 percent) returned for scanning at week 40. The demographic and clinical characteristics of the subjects in the treatment groups were similar at baseline, both in the entire sample and in the neuroimaging substudy (Table 1).

FIGURE 1

Random Assignments to Treatment, Completion of the Trial, and Reasons for Not Completing It.

TABLE 1

Baseline Characteristics of the Study Subjects.

CLINICAL OUTCOME

Levodopa, in a dose–response pattern, significantly (P<0.001) reduced the worsening of symptoms of Parkinson's disease as reflected in the change between the total score on the UPDRS at baseline and that at week 42 (i.e., two weeks after washout of the study medication), as compared with the change in the placebo group (Figure 2 and Table 2). The subjects in the placebo group had mild improvement at the week 3 visit, but after that their symptoms worsened steadily throughout the balance of the study period, including the two-week washout phase. A strong dose–response benefit was detected during the period in which the medication was administered beginning at week 9, when the full dose of 600 mg daily was reached in the group receiving the highest dose of levodopa, and it persisted through week 40. The scores on the UPDRS in the three levodopa groups worsened during the two-week washout period, but these groups did not deteriorate to the level observed in the placebo group, and the group receiving the highest dose of levodopa had the best result (Figure 2 and Table 2). The adverse events that were significantly more common among those receiving levodopa at 600 mg daily than in the placebo group were dyskinesias, nausea, infection, hypertonia, and headache (Table 3).

FIGURE 2

Changes in Total Scores on the Unified Parkinson's Disease Rating Scale (UPDRS) from Baseline through Eval‎uation at Week 42.

TABLE 2

Changes in the Scores on the UPDRS between Baseline and Week 42.

TABLE 3

Adverse Events.

SPECT AND [123I]Β-CIT SUBSTUDY

The mean [123I]β-CIT uptake in the striatum, caudate, and putamen at baseline was consistent with values previously reported for patients with early Parkinson's disease.20 The percent decrease in striatal [123I]β-CIT uptake over the 40 weeks of the study treatment was greater among subjects in the levodopa groups than in the placebo group, but this difference was not statistically significant (Table 4). However, 21 of the 142 subjects (14.7 percent) had a putaminal [123I]β-CIT uptake of more than 3.25 at baseline (i.e., more than 75 percent of the age-expected putaminal uptake).21 An analysis of the results of SPECT after the exclusion of the 19 subjects without a dopaminergic deficit who returned for the neuroimaging study at week 40 showed a significantly greater decrease in [123I]β-CIT uptake among those receiving levodopa than among those receiving placebo (P=0.036) (Table 4).

TABLE 4

Change in Striatal [123I]β-CIT Uptake between Baseline and Week 40.

Discussion

We found no clinical evidence that levodopa accelerated the worsening of Parkinson's disease over the 9.5 months of observation. Rather, levodopa was associated with less worsening of parkinsonism than was placebo, consistent with the notion that it slows disease progression (Table 2 and Figure 2). In comparison with the scores on the UPDRS at baseline, the final scores, after the washout phase, had worsened by approximately 8 units in the placebo group, whereas the group receiving levodopa at 600 mg daily did not have evidence of deterioration. The groups receiving levodopa at lower doses did have deterioration, as shown by a comparison of their scores at baseline and at week 42, but the deterioration was less than in the placebo group (P<0.001).

We need to consider that a two-week washout from levodopa may have been insufficient to eliminate fully the effect of the medication on symptoms, and the results observed may be related to a profound effect of levodopa on symptoms that persists for a long time after the drug has been withdrawn. Indeed, Hauser and Holford,22 using a modeling technique, analyzed the withdrawal of levodopa in 20 patients and reported that the mean half-life of levodopa as measured by the loss of the clinical benefit was 7.9 days (95 percent confidence interval, 2.2 to 30.4 days). They suggest that a washout period of 32 days (four half-lives) may be required to eliminate 90 percent of the drug's effects on symptoms. However, we saw little deterioration after one week of washout. Near the end of the study, with the approval of the institutional review boards and of the National Institutes of Health, we asked the last 38 subjects remaining in the study to extend the washout period to four weeks. Among these subjects there was no further worsening of the scores on the UPDRS during the additional two weeks, but the small number of subjects renders this component of the study difficult to interpret.

Muenter and Tyce23 described motor responses with two types of duration, short and long, with levodopa therapy. The short-duration benefit lasts for a few hours after a single dose, and the long-duration benefit lasts several days. In our study, the long-duration benefit appears to have lasted approximately one to two weeks (Figure 2). To argue that a longer washout period might have revealed more clinical worsening than was observed, one could propose a hitherto unknown third type of duration of motor response, one that is more sustained than the so-called long-duration benefit. Such an enduring benefit could be envisioned to result from a prolonged pharmacodynamic effect, for example, on dopamine receptors.

If, however, the clinical effects observed offer evidence of neuroprotection, how can we explain the greater loss of dopamine transporter shown in the SPECT imaging studies of the subjects who received levodopa, as compared with those who received the placebo — a result that suggests the possibility of a levodopa-induced toxic effect on dopamine neurons? At the end of the study, when the SPECT neuroimaging studies were performed, the subjects were still taking levodopa, so it is possible to assume that levodopa has a pharmacologic effect on the dopamine transporter that interferes with and reduces the binding of the β-CIT ligand.

Indeed, one study found a reduction in dopamine-transporter binding with the use of positron-emission tomography (PET) and another dopamine-transporter ligand in patients with early Parkinson's disease who were treated for six weeks with levodopa at 300 mg daily.24 However, the sample size in this study, as in others that have shown no change in dopamine-transporter binding after short-term treatment with levodopa,25–27 was too small to demonstrate a statistically significant difference between levodopa and placebo. Further indirect support for the absence of a pharmacologic effect of levodopa on imaging studies of the dopamine transporter in the present study was the absence of decline in the [123I]β-CIT uptake by week 40 among the 16 subjects who received levodopa and whose SPECT scans were normal at baseline (i.e., without evidence of a dopaminergic deficit) (data not shown). However, in the absence of studies with larger samples and a longer period of treatment with levodopa, we cannot exclude the possibility that levodopa may simply down-regulate the dopamine transporter. Another consideration is that the falling dopamine concentration in the placebo group may have led to a compensatory increase in the activity of the dopamine transporter that could have increased [123I]β-CIT binding, but evidence for such an interpretation is lacking.

If levodopa has neuroprotective effects, what mechanisms could account for this property? In low concentrations and in the presence of glial cells, levodopa protects cultured dopaminergic neurons28–31 and up-regulates antioxidant and antiapoptotic proteins.29,32 Furthermore, in vivo studies suggest that levodopa can promote survival and enhance the sprouting of nigral dopamine neurons in rodents treated with the toxin 6-hydroxydopamine.9,10 The question of whether levodopa has a protective or a toxic effect in Parkinson's disease — reflecting the results of clinical examination and neuroimaging studies, respectively — cannot be answered with certainty, and future studies will be needed.

This dose–response, placebo-controlled clinical trial eval‎uating the effect of levodopa in patients with early Parkinson's disease showed a strikingly impressive dose–response clinical benefit: the higher the dose, the stronger and more lasting the benefit, and the benefit was greater even after the drug was withdrawn (Figure 2).

During withdrawal, we did not encounter the neuroleptic malignant-like effect (i.e., high fever, obtundation, and rigidity) that can occur with the sudden withdrawal of levodopa.33,34 Its absence may be related to the down-titration procedure employed; however, this complication is so rare that it might not have occurred in a study of this size even had levodopa been withdrawn suddenly, particularly in patients with early-stage disease.

Our study calls into question interpretations of the functional neuroimaging of the dopamine system. It has generally been assumed that imaging studies performed with 18F-fluorodopa PET and [123I]β-CIT SPECT can provide reliable information on the integrity of the nigrostriatal dopamine pathway. Our study and other recent clinical trials25,35 raise the possibility of a pharmacologic influence of dopaminergic therapy on these neuroimaging targets and point to the need for clarification of this issue.

Although the absence of evidence of a dopaminergic deficit in the imaging in our study and two other studies25,35 may simply represent a limitation of the sensitivity of the imaging techniques, the scans without a dopaminergic deficit raise doubt about whether the subjects had Parkinson's disease.35 In our study, the subjects with such scans at baseline had no worsening of the [123I]β-CIT uptake at week 40 (data not shown), nor did subjects with such scans in the group receiving the highest dose of levodopa have an improvement in the scores on the UPDRS (i.e., a change between the score at baseline and at week 40 of 3.38±4.25 units, as compared with a change of –4.95±10.4 units in the subjects receiving levodopa at 600 mg daily whose baseline SPECT scans showed abnormalities; P=0.002). Whether the subjects with no evidence of dopaminergic deficit on scanning do or do not have classic Parkinson's disease remains uncertain. The fact that scans of some patients with early Parkinson's disease do not have evidence of a dopaminergic deficit needs to be taken into consideration in the planning of future trials to test drugs for neuroprotective effects.

Finally, even though we cannot reconcile the clinical and imaging findings in our study, we can assure both patients with early Parkinson's disease and their physicians that, from a clinical perspective, our study did not find that levodopa hastens the progression of Parkinson's disease. On the basis of the study, we can recommend that the doses of levodopa be adjusted to fit the needs of the patient. Small doses were found to be effective, although less so than higher doses. High doses, however, were associated with a greater frequency of adverse events such as dyskinesia. For the present, until more evidence is available, we recommend customizing the dose of levodopa to the needs of the individual patient on the basis of the clinical response and the profile of adverse events.

NOTES

Supported by grants from the National Institute of Neurological Disorders and Stroke (NS34796, to Dr. Fahn), the Department of Defense (DAMD 17-99-1-9472, to Dr. Marek), and the General Clinical Research Center of the National Center for Research Resources, National Institutes of Health (MO1-RR-00044 and MO1-RR-02066). Carbidopa–levodopa tablets and the matching placebo tablets were kindly provided by Teva Pharmaceuticals (Israel).

Drs. Fahn, Oakes, Shoulson, Kieburtz, Lang, Tanner, and Marek report having served as unpaid consultants to Teva Pharmaceuticals, and Dr. Olanow reports having served as a paid consultant to Teva Pharmaceuticals. Drs. Marek and Seibyl have an equity interest in Molecular Neuroimaging (New Haven), which carried out the [123I]β-CIT SPECT imaging for this study.

We are indebted to the subjects for their participation in the study; to K. Hyland (Baylor Medical College, Houston) for the independent, blinded chemical analyses of the contents of the tablets conducted annually to ascertain the chemical stability of levodopa and carbidopa; to S. Bennett, A. Brocht, D. Graffrath, J. Janciuras, C. Orme, L. Preston, K. Rothenburgh, C. Weaver, and A. Watts of the Biostatistics and Coordination Centers at the University of Rochester, Rochester, N.Y.; to the Independent Safety Monitoring Committee (P. Tariot, chair, Monroe Community Hospital, Rochester, N.Y.; to W.J. Hall, University of Rochester; to R. Rodnitzky, University of Iowa, Iowa City); to the National Institute of Neurological Disorders and Stroke Safety Monitoring Committee (E.C. Haley, chair, University of Virginia, Charlottesville; D. Eidelberg, North Shore University Medical Center, Manhasset, N.Y.; C.A. Gatsonis, Brown University, Providence, R.I.; W. Rocca, Mayo Clinic, Rochester, Minn.); and to E.J. Oliver (administrator), National Institute of Neurological Disorders and Stroke, Bethesda, Md.

APPENDIX

The following members of the Parkinson Study Group participated in the Earlier versus Later Levodopa Therapy in Parkinson Disease (ELLDOPA) study and contributed to this report. Steering committee: S. Fahn, principal investigator, Columbia University, New York; D. Oakes, chief biostatistician; I. Shoulson, coprincipal investigator; K. Kieburtz, director, Clinical Trials Coordination Center; A. Rudolph, senior project coordinator, University of Rochester, Rochester, N.Y.; K. Marek, neuroimager; J. Seibyl, neuroimager, Institute for Neurodegenerative Disorders, New Haven, Conn.; A. Lang, Toronto Western Hospital, Toronto; C.W. Olanow, Mount Sinai School of Medicine, New York; C. Tanner, The Parkinson's Institute, Sunnyvale, Calif.; G. Schifitto, medical monitor, University of Rochester, Rochester, N.Y.; H. Zhao, biostatistician, University of Rochester, Rochester, N.Y.; L. Reyes, administrator, Columbia University, New York; A. Shinaman, administrator, University of Rochester, Rochester, N.Y.; Participating primary raters, treating investigators, and coordinators: Rush–Presbyterian–St. Luke's Medical Center, Chicago: C. Comella, C. Goetz, L. Blasucci; Barrow Neurological Institute, Phoenix, Ariz.: J. Samanta, M. Stacy, K. Williamson, M. Harrigan; Columbia University, New York: P. Greene, B. Ford, C. Moskowitz; Parkinson's and Movement Disorder Institute, Fountain Valley, Calif.: D. Truong, M. Pathak; Baylor College of Medicine, Houston: J. Jankovic, W. Ondo, F. Atassi, C. Hunter; Brown University, Providence, R.I.: C. Jacques, J.H. Friedman, M. Lannon; Institute for Neurodegenerative Disorders, New Haven, Conn.: D.S. Russell, D. Jennings, B. Fussell; Massachusetts General Hospital, Boston: D. Standaert, M.A. Schwarzschild, J. Growdon, M. Tennis; McGill Centre for Studies in Aging, Verdun, Que., Canada: S. Gauthier, M. Panisset, J. Hall; Oregon Health and Science University, Portland: S. Gancher, J. Hammerstad, C. Stone, B. Alexander-Brown; Albany Medical College, Albany, N.Y.: S. Factor, E. Molho, D. Brown (deceased), S. Evans; Scott and White Hospital–Texas A&M University, Temple: J. Clark, B. Manyam, P. Simpson, B. Wulbrecht, J. Whetteckey; University of Alberta, Edmonton, Canada: W. Martin, T. Roberts, P. King; University of South Florida, Tampa: R. Hauser, T. Zesiewicz, L. Gauger; University of Virginia, Charlottesville: J. Trugman, G.F. Wooten, E. Rost-Ruffner; Washington University, St. Louis: J. Perlmutter, B. Racette; University of Calgary, Alta., Canada: O. Suchowersky, R. Ranawaya, S. Wood, C. Pantella; University of Rochester, Rochester, N.Y.: R. Kurlan, I. Richard, N. Pearson; Mayo Clinic, Scottsdale, Ariz.: J. Caviness, C. Adler, M. Lind; University of Pennsylvania, Philadelphia: T. Simuni, A. Siderowf, A. Colcher, M. Lloyd; University of Miami, Miami: W. Weiner, L. Shulman, W. Koller, K. Lyons; Boston University, Boston: R. Feldman (deceased), M.-H. St.-Hilaire, S. Ellias, C.-A. Thomas; Emory University, Atlanta: J. Juncos, R. Watts, A. Partlow; The Parkinson's Institute, Sunnyvale, Calif.: J. Tetrud, D.M. Togasaki, M. Welsh, T. Stewart; University of Medicine and Dentistry of New Jersey–Robert Wood Johnson, New Brunswick: M.H. Mark, J.I. Sage, D. Caputo; Louisiana State University, New Orleans: H. Gould, J. Rao, A. McKendrick; Mount Sinai School of Medicine, New York: M. Brin, F. Danisi, R. Benabou; Ohio State University, Columbus: J. Hubble, G. Paulson, C. Reider; Toronto Western Hospital, Toronto: A. Birnbaum, J. Miyasaki, L. Johnston, J. So; University of Kansas, Kansas City: R. Pahwa, R. Dubinsky; Mayo Clinic, Jacksonville, Fla.: Z. Wszolek, R. Uitti, M. Turk; Minneapolis Veterans Affairs Hospital, Minneapolis: P. Tuite, D. Rottenberg, J. Hansen; University of Puerto Rico, San Juan: C. Serrano Ramos; University of Southern California, Los Angeles: C. Waters, M. Lew, M. Welsh, C. Kawai; Colorado Neurological Institute, Englewood: C. O'Brien, R. Kumar, L. Seeberger, D. Judd; Ottawa Civic Hospital, Ottawa: T. Mendis, C.L. Barclay, D.A. Grimes, L. Sutherland; Johns Hopkins University, Baltimore: T. Dawson, S. Reich, R. Dunlop; University of Michigan, Ann Arbor: R. Albin, K. Frey, K. Wernette.

ORIGINAL ARTICLE

SHARE ON

•  
•  
•  
•  

Levodopa and the Progression of Parkinson's Disease

Author: The Parkinson Study Group*Author Info & Affiliations

Published December 9, 2004

N Engl J Med 2004;351:2498-2508

DOI: 10.1056/NEJMoa033447

VOL. 351 NO. 24

• • ABSTRACT
  • METHODS
  • RESULTS
  • DISCUSSION
  • NOTES
  • APPENDIX
  • REFERENCES
• • Information & Authors
  • Metrics & Citations
  • View Options
  • References
  • Media
  • Tables
  • Share

AbstractBACKGROUND

Despite the known benefit of levodopa in reducing the symptoms of Parkinson's disease, concern has been expressed that its use might hasten neurodegeneration. This study assessed the effect of levodopa on the rate of progression of Parkinson's disease.

METHODS

In this randomized, double-blind, placebo-controlled trial, we eval‎uated 361 patients with early Parkinson's disease who were assigned to receive carbidopa–levodopa at a daily dose of 37.5 and 150 mg, 75 and 300 mg, or 150 and 600 mg, respectively, or a matching placebo for a period of 40 weeks, and then to undergo withdrawal of treatment for 2 weeks. The primary outcome was a change in scores on the Unified Parkinson's Disease Rating Scale (UPDRS) between baseline and 42 weeks. Neuroimaging studies of 142 subjects were performed at baseline and at week 40 to assess striatal dopamine-transporter density with the use of iodine-123–labeled 2-β-carboxymethoxy-3-β-(4-iodophenyl)tropane ([123I]β-CIT) uptake.

RESULTS

The severity of parkinsonism increased more in the placebo group than in all the groups receiving levodopa: the mean difference between the total score on the UPDRS at baseline and at 42 weeks was 7.8 units in the placebo group, 1.9 units in the group receiving levodopa at a dose of 150 mg daily, 1.9 in those receiving 300 mg daily, and –1.4 in those receiving 600 mg daily (P<0.001). In contrast, in a substudy of 116 patients the mean percent decline in the [123I]β-CIT uptake was significantly greater with levodopa than placebo (–6 percent among those receiving levodopa at 150 mg daily, –4 percent in those receiving it at 300 mg daily, and –7.2 percent among those receiving it at 600 mg daily, as compared with –1.4 percent among those receiving placebo; 19 patients with no dopaminergic deficits on the baseline scans were excluded from the analysis) (P=0.036). The subjects receiving the highest dose of levodopa had significantly more dyskinesia, hypertonia, infection, headache, and nausea than those receiving placebo.

CONCLUSIONS

The clinical data suggest that levodopa either slows the progression of Parkinson's disease or has a prolonged effect on the symptoms of the disease. In contrast, the neuroimaging data suggest either that levodopa accelerates the loss of nigrostriatal dopamine nerve terminals or that its pharmacologic effects modify the dopamine transporter. The potential long-term effects of levodopa on Parkinson's disease remain uncertain.

Parkinson's disease is a progressively disabling neurodegenerative disorder that is manifested clinically by bradykinesia, tremor, rigidity, flexed posture, postural instability, and freezing of gait. It is characterized pathologically by the loss of pigmented dopaminergic neurons in the substantia nigra. The course of the clinical decline parallels that of the progressive degeneration of the remaining dopaminergic neurons.1 The use of levodopa as dopamine-replacement therapy is highly effective in ameliorating the symptoms of the disease and remains the standard drug with which other therapies are compared.2,3

Because levodopa and dopamine can generate reactive oxygen species and induce the degeneration of cultured dopamine neurons, concern has been raised that levodopa could enhance oxidative stress and hasten the degeneration of residual dopamine neurons in patients with Parkinson's disease.4–6 However, levodopa is not toxic in animals and may be trophic and promote the functional recovery of damaged nigral neurons.7–10 Humans without Parkinson's disease who are exposed to levodopa do not develop nigral damage,11,12 but such persons do not have increased oxidative stress in their substantia nigra neurons.

Whether levodopa is detrimental, beneficial, or without effect on the rate of the progression of Parkinson's disease is unknown and extremely important, both scientifically and clinically. We therefore conducted a controlled clinical trial to assess the effect of levodopa on the course of Parkinson's disease.

MethodsSTUDY DESIGN

Our multicenter, placebo-controlled, randomized, dose-ranging, double-blind clinical trial, called the Earlier versus Later Levodopa Therapy in Parkinson Disease (ELLDOPA) study, was conceived, organized, and implemented by the Parkinson Study Group and sponsored by the National Institute of Neurological Disorders and Stroke. The Department of Defense sponsored the single-photon-emission computed-tomography (SPECT) substudy. The subjects were enrolled between September 1998 and August 2001 at 33 sites in the United States and 5 sites in Canada. The study was approved by the institutional review boards at the participating sites, and all subjects gave written informed consent. An independent safety monitoring committee monitored the data, subjects' safety, and the tolerability of the study drug. There was no prespecified formal guideline for recommending either modification or termination of the trial.

SUBJECTS

The subjects were 30 years of age or older, had received a diagnosis of Parkinson's disease within the past two years, had a rating on the modified Hoehn–Yahr scale13,14 of less than stage 3 (with stage 1 indicating unilateral disease, stage 2 mild bilateral disease, and stage 3 more advanced bilateral disease), and were considered not likely to require therapy for symptoms of the disease within the nine months after enrollment in the study. Patients were excluded if they were receiving antiparkinson medication, had been exposed to levodopa or to any dopamine agonist for more than 14 days, had an identifiable cause of parkinsonism, or had a tremor in any limb that was given a score of 3 or more on the Unified Parkinson's Disease Rating Scale (UPDRS),14 freezing of gait, loss of postural reflexes, major depression, or dementia.

Potential subjects were informed that only the assigned study drug would be permitted during the nine months of the study and that if they needed additional antiparkinson medication during this period, they would have to withdraw from the study. Subjects were randomly assigned to receive placebo or carbidopa–levodopa at a dose of 12.5 and 50 mg three times daily, 25 and 100 mg three times daily, or 50 and 200 mg three times daily, respectively. The doses were increased to the full amount over a period of nine weeks in a blinded fashion.

After 40 weeks, the subjects underwent a 3-day period of step-down withdrawal from the study drug. After two weeks without the study drug, a final assessment of the severity of the symptoms of Parkinson's disease was made. The selection of the 2-week duration for the washout period was based on reports that withdrawal from levodopa for a period up to 14 days resulted in a worsening of parkinsonism mainly within the first 7 days, with nonsignificant worsening beyond that point.15–17

CLINICAL EVAL‎UATION

The treating investigator, who was blinded to the treatment assignment, performed a clinical eval‎uation with the use of the UPDRS14 at the screening, baseline, and interim visits (at the end of weeks 3, 9, 24, and 40) and during each of the two weeks of the washout phase. During the four interim visits, the eval‎uations were performed before the administration of the first dose of the study drug. At every visit the treating investigator inquired about adverse events. The primary rater, who was also blinded to the treatment assignment and was kept unaware of information obtained during the course of the study, performed the examination with the use of the UPDRS only at the baseline eval‎uation, which took place within four weeks after screening, and again at the final eval‎uation, two weeks after the subject had undergone complete withdrawal from the study treatment (week 42). The study coordinators and the subjects also were blinded to the treatment assignments. An emergency unblinding procedure was established but was never used.

OUTCOME

The prespecified primary outcome was the change in the severity of parkinsonism between the baseline visit and week 42, as measured with the use of the total score on the UPDRS that was obtained by the primary rater; week 42 was 14 days after the withdrawal of the study drug. The expected result was a dose-related deterioration during washout if levodopa were shown to hasten the progression of Parkinson's disease. The change in the total scores on the UPDRS measured by the treating investigator at each visit was a prespecified secondary outcome. The treating investigator assessed adverse events with the use of open-ended questioning at each visit.

SUBSTUDY

After the clinical trial was begun, we conducted a substudy with the use of SPECT to measure striatal dopamine-transporter density with the use of iodine-123–labeled 2-β-carboxymethoxy-3-β-(4-iodophenyl)tropane ([123I]β-CIT). The methods have been reported previously.18 The subjects who gave consent underwent SPECT imaging just before the baseline visit and then again before the visit at week 40. The imaging studies were performed at that point rather than after the subjects had undergone withdrawal from levodopa, at week 42, because it was thought that the subjects would not be able to tolerate the procedure (including the necessary travel) after the return of parkinsonism or after its worsening. Another reason was that there was no clear evidence of a short-term effect of levodopa on dopamine-transporter imaging.

All imaging studies were performed at Yale University or the Institute for Neurodegenerative Disorders (both in New Haven, Connecticut). The neuroimaging personnel were blinded to the treatment assignment of the subjects. The results of the SPECT studies were transferred to and analyzed by the biostatistics center of the Parkinson Study Group. The prespecified outcome was the percent change in the ratio of the specific striatal [123I]β-CIT uptake to the nondisplaceable striatal [123I]β-CIT uptake between the two images.

STATISTICAL ANALYSIS

Allowing for a dropout rate of 10 percent of the subjects enrolled, we chose a sample size of 360 subjects (i.e., 90 subjects in each of the four treatment groups) in order to provide the study with 85 percent power to detect a dose–response relationship (linear trend) between the assigned doses and the change in the subjects' total score on the UPDRS between the baseline visit and the final visit at week 42. On the basis of the results of one study,19 we anticipated a rate of worsening in the total score on the UPDRS in the placebo group of 9.5 units over the 9.5 months of the study. The study was powered to detect a linear trend that corresponded to a 4-unit difference (45 percent of 9.5 units) in the score on the UPDRS between the highest dose of levodopa (600 mg per day) and placebo — that is, a change of either 13.5 or 5.5 units from baseline in the group receiving the highest dose of the active study drug.

The primary statistical tests were two-tailed, with an alpha level of 0.05. Only subjects who completed the two-week washout phase were included in this analysis. We used the intention-to-treat principle in the analysis, even if among some subjects the dose of the study drug were to be reduced during the study. The primary analysis assessed the dose–response relationship between the assigned doses and the worsening of parkinsonism, as indicated by the changes in the total score on the UPDRS between the baseline visit and week 42. Statistical comparisons were made by analysis of covariance in a model that adjusted for differences among the investigators performing the eval‎uations and in baseline values.

ResultsSUBJECTS

Of a total of 361 subjects enrolled in the study, 317 (88 percent) took the study medication for 40 weeks, and 311 (86 percent) completed the 2 weeks of washout (Figure 1). The neuroimaging substudy was begun after the enrollment of the first 108 subjects. Of the 253 subjects subsequently enrolled, 142 (56 percent) participated in the substudy and underwent the baseline SPECT. Of these, 135 (95 percent) returned for scanning at week 40. The demographic and clinical characteristics of the subjects in the treatment groups were similar at baseline, both in the entire sample and in the neuroimaging substudy (Table 1).

FIGURE 1

Random Assignments to Treatment, Completion of the Trial, and Reasons for Not Completing It.

TABLE 1

Baseline Characteristics of the Study Subjects.

CLINICAL OUTCOME

Levodopa, in a dose–response pattern, significantly (P<0.001) reduced the worsening of symptoms of Parkinson's disease as reflected in the change between the total score on the UPDRS at baseline and that at week 42 (i.e., two weeks after washout of the study medication), as compared with the change in the placebo group (Figure 2 and Table 2). The subjects in the placebo group had mild improvement at the week 3 visit, but after that their symptoms worsened steadily throughout the balance of the study period, including the two-week washout phase. A strong dose–response benefit was detected during the period in which the medication was administered beginning at week 9, when the full dose of 600 mg daily was reached in the group receiving the highest dose of levodopa, and it persisted through week 40. The scores on the UPDRS in the three levodopa groups worsened during the two-week washout period, but these groups did not deteriorate to the level observed in the placebo group, and the group receiving the highest dose of levodopa had the best result (Figure 2 and Table 2). The adverse events that were significantly more common among those receiving levodopa at 600 mg daily than in the placebo group were dyskinesias, nausea, infection, hypertonia, and headache (Table 3).

FIGURE 2

Changes in Total Scores on the Unified Parkinson's Disease Rating Scale (UPDRS) from Baseline through Eval‎uation at Week 42.

TABLE 2

Changes in the Scores on the UPDRS between Baseline and Week 42.

TABLE 3

Adverse Events.

SPECT AND [123I]Β-CIT SUBSTUDY

The mean [123I]β-CIT uptake in the striatum, caudate, and putamen at baseline was consistent with values previously reported for patients with early Parkinson's disease.20 The percent decrease in striatal [123I]β-CIT uptake over the 40 weeks of the study treatment was greater among subjects in the levodopa groups than in the placebo group, but this difference was not statistically significant (Table 4). However, 21 of the 142 subjects (14.7 percent) had a putaminal [123I]β-CIT uptake of more than 3.25 at baseline (i.e., more than 75 percent of the age-expected putaminal uptake).21 An analysis of the results of SPECT after the exclusion of the 19 subjects without a dopaminergic deficit who returned for the neuroimaging study at week 40 showed a significantly greater decrease in [123I]β-CIT uptake among those receiving levodopa than among those receiving placebo (P=0.036) (Table 4).

TABLE 4

Change in Striatal [123I]β-CIT Uptake between Baseline and Week 40.

Discussion

We found no clinical evidence that levodopa accelerated the worsening of Parkinson's disease over the 9.5 months of observation. Rather, levodopa was associated with less worsening of parkinsonism than was placebo, consistent with the notion that it slows disease progression (Table 2 and Figure 2). In comparison with the scores on the UPDRS at baseline, the final scores, after the washout phase, had worsened by approximately 8 units in the placebo group, whereas the group receiving levodopa at 600 mg daily did not have evidence of deterioration. The groups receiving levodopa at lower doses did have deterioration, as shown by a comparison of their scores at baseline and at week 42, but the deterioration was less than in the placebo group (P<0.001).

We need to consider that a two-week washout from levodopa may have been insufficient to eliminate fully the effect of the medication on symptoms, and the results observed may be related to a profound effect of levodopa on symptoms that persists for a long time after the drug has been withdrawn. Indeed, Hauser and Holford,22 using a modeling technique, analyzed the withdrawal of levodopa in 20 patients and reported that the mean half-life of levodopa as measured by the loss of the clinical benefit was 7.9 days (95 percent confidence interval, 2.2 to 30.4 days). They suggest that a washout period of 32 days (four half-lives) may be required to eliminate 90 percent of the drug's effects on symptoms. However, we saw little deterioration after one week of washout. Near the end of the study, with the approval of the institutional review boards and of the National Institutes of Health, we asked the last 38 subjects remaining in the study to extend the washout period to four weeks. Among these subjects there was no further worsening of the scores on the UPDRS during the additional two weeks, but the small number of subjects renders this component of the study difficult to interpret.

Muenter and Tyce23 described motor responses with two types of duration, short and long, with levodopa therapy. The short-duration benefit lasts for a few hours after a single dose, and the long-duration benefit lasts several days. In our study, the long-duration benefit appears to have lasted approximately one to two weeks (Figure 2). To argue that a longer washout period might have revealed more clinical worsening than was observed, one could propose a hitherto unknown third type of duration of motor response, one that is more sustained than the so-called long-duration benefit. Such an enduring benefit could be envisioned to result from a prolonged pharmacodynamic effect, for example, on dopamine receptors.

If, however, the clinical effects observed offer evidence of neuroprotection, how can we explain the greater loss of dopamine transporter shown in the SPECT imaging studies of the subjects who received levodopa, as compared with those who received the placebo — a result that suggests the possibility of a levodopa-induced toxic effect on dopamine neurons? At the end of the study, when the SPECT neuroimaging studies were performed, the subjects were still taking levodopa, so it is possible to assume that levodopa has a pharmacologic effect on the dopamine transporter that interferes with and reduces the binding of the β-CIT ligand.

Indeed, one study found a reduction in dopamine-transporter binding with the use of positron-emission tomography (PET) and another dopamine-transporter ligand in patients with early Parkinson's disease who were treated for six weeks with levodopa at 300 mg daily.24 However, the sample size in this study, as in others that have shown no change in dopamine-transporter binding after short-term treatment with levodopa,25–27 was too small to demonstrate a statistically significant difference between levodopa and placebo. Further indirect support for the absence of a pharmacologic effect of levodopa on imaging studies of the dopamine transporter in the present study was the absence of decline in the [123I]β-CIT uptake by week 40 among the 16 subjects who received levodopa and whose SPECT scans were normal at baseline (i.e., without evidence of a dopaminergic deficit) (data not shown). However, in the absence of studies with larger samples and a longer period of treatment with levodopa, we cannot exclude the possibility that levodopa may simply down-regulate the dopamine transporter. Another consideration is that the falling dopamine concentration in the placebo group may have led to a compensatory increase in the activity of the dopamine transporter that could have increased [123I]β-CIT binding, but evidence for such an interpretation is lacking.

If levodopa has neuroprotective effects, what mechanisms could account for this property? In low concentrations and in the presence of glial cells, levodopa protects cultured dopaminergic neurons28–31 and up-regulates antioxidant and antiapoptotic proteins.29,32 Furthermore, in vivo studies suggest that levodopa can promote survival and enhance the sprouting of nigral dopamine neurons in rodents treated with the toxin 6-hydroxydopamine.9,10 The question of whether levodopa has a protective or a toxic effect in Parkinson's disease — reflecting the results of clinical examination and neuroimaging studies, respectively — cannot be answered with certainty, and future studies will be needed.

This dose–response, placebo-controlled clinical trial eval‎uating the effect of levodopa in patients with early Parkinson's disease showed a strikingly impressive dose–response clinical benefit: the higher the dose, the stronger and more lasting the benefit, and the benefit was greater even after the drug was withdrawn (Figure 2).

During withdrawal, we did not encounter the neuroleptic malignant-like effect (i.e., high fever, obtundation, and rigidity) that can occur with the sudden withdrawal of levodopa.33,34 Its absence may be related to the down-titration procedure employed; however, this complication is so rare that it might not have occurred in a study of this size even had levodopa been withdrawn suddenly, particularly in patients with early-stage disease.

Our study calls into question interpretations of the functional neuroimaging of the dopamine system. It has generally been assumed that imaging studies performed with 18F-fluorodopa PET and [123I]β-CIT SPECT can provide reliable information on the integrity of the nigrostriatal dopamine pathway. Our study and other recent clinical trials25,35 raise the possibility of a pharmacologic influence of dopaminergic therapy on these neuroimaging targets and point to the need for clarification of this issue.

Although the absence of evidence of a dopaminergic deficit in the imaging in our study and two other studies25,35 may simply represent a limitation of the sensitivity of the imaging techniques, the scans without a dopaminergic deficit raise doubt about whether the subjects had Parkinson's disease.35 In our study, the subjects with such scans at baseline had no worsening of the [123I]β-CIT uptake at week 40 (data not shown), nor did subjects with such scans in the group receiving the highest dose of levodopa have an improvement in the scores on the UPDRS (i.e., a change between the score at baseline and at week 40 of 3.38±4.25 units, as compared with a change of –4.95±10.4 units in the subjects receiving levodopa at 600 mg daily whose baseline SPECT scans showed abnormalities; P=0.002). Whether the subjects with no evidence of dopaminergic deficit on scanning do or do not have classic Parkinson's disease remains uncertain. The fact that scans of some patients with early Parkinson's disease do not have evidence of a dopaminergic deficit needs to be taken into consideration in the planning of future trials to test drugs for neuroprotective effects.

Finally, even though we cannot reconcile the clinical and imaging findings in our study, we can assure both patients with early Parkinson's disease and their physicians that, from a clinical perspective, our study did not find that levodopa hastens the progression of Parkinson's disease. On the basis of the study, we can recommend that the doses of levodopa be adjusted to fit the needs of the patient. Small doses were found to be effective, although less so than higher doses. High doses, however, were associated with a greater frequency of adverse events such as dyskinesia. For the present, until more evidence is available, we recommend customizing the dose of levodopa to the needs of the individual patient on the basis of the clinical response and the profile of adverse events.

NOTES

Supported by grants from the National Institute of Neurological Disorders and Stroke (NS34796, to Dr. Fahn), the Department of Defense (DAMD 17-99-1-9472, to Dr. Marek), and the General Clinical Research Center of the National Center for Research Resources, National Institutes of Health (MO1-RR-00044 and MO1-RR-02066). Carbidopa–levodopa tablets and the matching placebo tablets were kindly provided by Teva Pharmaceuticals (Israel).

Drs. Fahn, Oakes, Shoulson, Kieburtz, Lang, Tanner, and Marek report having served as unpaid consultants to Teva Pharmaceuticals, and Dr. Olanow reports having served as a paid consultant to Teva Pharmaceuticals. Drs. Marek and Seibyl have an equity interest in Molecular Neuroimaging (New Haven), which carried out the [123I]β-CIT SPECT imaging for this study.

We are indebted to the subjects for their participation in the study; to K. Hyland (Baylor Medical College, Houston) for the independent, blinded chemical analyses of the contents of the tablets conducted annually to ascertain the chemical stability of levodopa and carbidopa; to S. Bennett, A. Brocht, D. Graffrath, J. Janciuras, C. Orme, L. Preston, K. Rothenburgh, C. Weaver, and A. Watts of the Biostatistics and Coordination Centers at the University of Rochester, Rochester, N.Y.; to the Independent Safety Monitoring Committee (P. Tariot, chair, Monroe Community Hospital, Rochester, N.Y.; to W.J. Hall, University of Rochester; to R. Rodnitzky, University of Iowa, Iowa City); to the National Institute of Neurological Disorders and Stroke Safety Monitoring Committee (E.C. Haley, chair, University of Virginia, Charlottesville; D. Eidelberg, North Shore University Medical Center, Manhasset, N.Y.; C.A. Gatsonis, Brown University, Providence, R.I.; W. Rocca, Mayo Clinic, Rochester, Minn.); and to E.J. Oliver (administrator), National Institute of Neurological Disorders and Stroke, Bethesda, Md.

APPENDIX

The following members of the Parkinson Study Group participated in the Earlier versus Later Levodopa Therapy in Parkinson Disease (ELLDOPA) study and contributed to this report. Steering committee: S. Fahn, principal investigator, Columbia University, New York; D. Oakes, chief biostatistician; I. Shoulson, coprincipal investigator; K. Kieburtz, director, Clinical Trials Coordination Center; A. Rudolph, senior project coordinator, University of Rochester, Rochester, N.Y.; K. Marek, neuroimager; J. Seibyl, neuroimager, Institute for Neurodegenerative Disorders, New Haven, Conn.; A. Lang, Toronto Western Hospital, Toronto; C.W. Olanow, Mount Sinai School of Medicine, New York; C. Tanner, The Parkinson's Institute, Sunnyvale, Calif.; G. Schifitto, medical monitor, University of Rochester, Rochester, N.Y.; H. Zhao, biostatistician, University of Rochester, Rochester, N.Y.; L. Reyes, administrator, Columbia University, New York; A. Shinaman, administrator, University of Rochester, Rochester, N.Y.; Participating primary raters, treating investigators, and coordinators: Rush–Presbyterian–St. Luke's Medical Center, Chicago: C. Comella, C. Goetz, L. Blasucci; Barrow Neurological Institute, Phoenix, Ariz.: J. Samanta, M. Stacy, K. Williamson, M. Harrigan; Columbia University, New York: P. Greene, B. Ford, C. Moskowitz; Parkinson's and Movement Disorder Institute, Fountain Valley, Calif.: D. Truong, M. Pathak; Baylor College of Medicine, Houston: J. Jankovic, W. Ondo, F. Atassi, C. Hunter; Brown University, Providence, R.I.: C. Jacques, J.H. Friedman, M. Lannon; Institute for Neurodegenerative Disorders, New Haven, Conn.: D.S. Russell, D. Jennings, B. Fussell; Massachusetts General Hospital, Boston: D. Standaert, M.A. Schwarzschild, J. Growdon, M. Tennis; McGill Centre for Studies in Aging, Verdun, Que., Canada: S. Gauthier, M. Panisset, J. Hall; Oregon Health and Science University, Portland: S. Gancher, J. Hammerstad, C. Stone, B. Alexander-Brown; Albany Medical College, Albany, N.Y.: S. Factor, E. Molho, D. Brown (deceased), S. Evans; Scott and White Hospital–Texas A&M University, Temple: J. Clark, B. Manyam, P. Simpson, B. Wulbrecht, J. Whetteckey; University of Alberta, Edmonton, Canada: W. Martin, T. Roberts, P. King; University of South Florida, Tampa: R. Hauser, T. Zesiewicz, L. Gauger; University of Virginia, Charlottesville: J. Trugman, G.F. Wooten, E. Rost-Ruffner; Washington University, St. Louis: J. Perlmutter, B. Racette; University of Calgary, Alta., Canada: O. Suchowersky, R. Ranawaya, S. Wood, C. Pantella; University of Rochester, Rochester, N.Y.: R. Kurlan, I. Richard, N. Pearson; Mayo Clinic, Scottsdale, Ariz.: J. Caviness, C. Adler, M. Lind; University of Pennsylvania, Philadelphia: T. Simuni, A. Siderowf, A. Colcher, M. Lloyd; University of Miami, Miami: W. Weiner, L. Shulman, W. Koller, K. Lyons; Boston University, Boston: R. Feldman (deceased), M.-H. St.-Hilaire, S. Ellias, C.-A. Thomas; Emory University, Atlanta: J. Juncos, R. Watts, A. Partlow; The Parkinson's Institute, Sunnyvale, Calif.: J. Tetrud, D.M. Togasaki, M. Welsh, T. Stewart; University of Medicine and Dentistry of New Jersey–Robert Wood Johnson, New Brunswick: M.H. Mark, J.I. Sage, D. Caputo; Louisiana State University, New Orleans: H. Gould, J. Rao, A. McKendrick; Mount Sinai School of Medicine, New York: M. Brin, F. Danisi, R. Benabou; Ohio State University, Columbus: J. Hubble, G. Paulson, C. Reider; Toronto Western Hospital, Toronto: A. Birnbaum, J. Miyasaki, L. Johnston, J. So; University of Kansas, Kansas City: R. Pahwa, R. Dubinsky; Mayo Clinic, Jacksonville, Fla.: Z. Wszolek, R. Uitti, M. Turk; Minneapolis Veterans Affairs Hospital, Minneapolis: P. Tuite, D. Rottenberg, J. Hansen; University of Puerto Rico, San Juan: C. Serrano Ramos; University of Southern California, Los Angeles: C. Waters, M. Lew, M. Welsh, C. Kawai; Colorado Neurological Institute, Englewood: C. O'Brien, R. Kumar, L. Seeberger, D. Judd; Ottawa Civic Hospital, Ottawa: T. Mendis, C.L. Barclay, D.A. Grimes, L. Sutherland; Johns Hopkins University, Baltimore: T. Dawson, S. Reich, R. Dunlop; University of Michigan, Ann Arbor: R. Albin, K. Frey, K. Wernette.

ORIGINAL ARTICLE

SHARE ON

•  
•  
•  
•  

Levodopa and the Progression of Parkinson's Disease

Author: The Parkinson Study Group*Author Info & Affiliations

Published December 9, 2004

N Engl J Med 2004;351:2498-2508

DOI: 10.1056/NEJMoa033447

VOL. 351 NO. 24

• • ABSTRACT
  • METHODS
  • RESULTS
  • DISCUSSION
  • NOTES
  • APPENDIX
  • REFERENCES
• • Information & Authors
  • Metrics & Citations
  • View Options
  • References
  • Media
  • Tables
  • Share

AbstractBACKGROUND

Despite the known benefit of levodopa in reducing the symptoms of Parkinson's disease, concern has been expressed that its use might hasten neurodegeneration. This study assessed the effect of levodopa on the rate of progression of Parkinson's disease.

METHODS

In this randomized, double-blind, placebo-controlled trial, we eval‎uated 361 patients with early Parkinson's disease who were assigned to receive carbidopa–levodopa at a daily dose of 37.5 and 150 mg, 75 and 300 mg, or 150 and 600 mg, respectively, or a matching placebo for a period of 40 weeks, and then to undergo withdrawal of treatment for 2 weeks. The primary outcome was a change in scores on the Unified Parkinson's Disease Rating Scale (UPDRS) between baseline and 42 weeks. Neuroimaging studies of 142 subjects were performed at baseline and at week 40 to assess striatal dopamine-transporter density with the use of iodine-123–labeled 2-β-carboxymethoxy-3-β-(4-iodophenyl)tropane ([123I]β-CIT) uptake.

RESULTS

The severity of parkinsonism increased more in the placebo group than in all the groups receiving levodopa: the mean difference between the total score on the UPDRS at baseline and at 42 weeks was 7.8 units in the placebo group, 1.9 units in the group receiving levodopa at a dose of 150 mg daily, 1.9 in those receiving 300 mg daily, and –1.4 in those receiving 600 mg daily (P<0.001). In contrast, in a substudy of 116 patients the mean percent decline in the [123I]β-CIT uptake was significantly greater with levodopa than placebo (–6 percent among those receiving levodopa at 150 mg daily, –4 percent in those receiving it at 300 mg daily, and –7.2 percent among those receiving it at 600 mg daily, as compared with –1.4 percent among those receiving placebo; 19 patients with no dopaminergic deficits on the baseline scans were excluded from the analysis) (P=0.036). The subjects receiving the highest dose of levodopa had significantly more dyskinesia, hypertonia, infection, headache, and nausea than those receiving placebo.

CONCLUSIONS

The clinical data suggest that levodopa either slows the progression of Parkinson's disease or has a prolonged effect on the symptoms of the disease. In contrast, the neuroimaging data suggest either that levodopa accelerates the loss of nigrostriatal dopamine nerve terminals or that its pharmacologic effects modify the dopamine transporter. The potential long-term effects of levodopa on Parkinson's disease remain uncertain.

Parkinson's disease is a progressively disabling neurodegenerative disorder that is manifested clinically by bradykinesia, tremor, rigidity, flexed posture, postural instability, and freezing of gait. It is characterized pathologically by the loss of pigmented dopaminergic neurons in the substantia nigra. The course of the clinical decline parallels that of the progressive degeneration of the remaining dopaminergic neurons.1 The use of levodopa as dopamine-replacement therapy is highly effective in ameliorating the symptoms of the disease and remains the standard drug with which other therapies are compared.2,3

Because levodopa and dopamine can generate reactive oxygen species and induce the degeneration of cultured dopamine neurons, concern has been raised that levodopa could enhance oxidative stress and hasten the degeneration of residual dopamine neurons in patients with Parkinson's disease.4–6 However, levodopa is not toxic in animals and may be trophic and promote the functional recovery of damaged nigral neurons.7–10 Humans without Parkinson's disease who are exposed to levodopa do not develop nigral damage,11,12 but such persons do not have increased oxidative stress in their substantia nigra neurons.

Whether levodopa is detrimental, beneficial, or without effect on the rate of the progression of Parkinson's disease is unknown and extremely important, both scientifically and clinically. We therefore conducted a controlled clinical trial to assess the effect of levodopa on the course of Parkinson's disease.

MethodsSTUDY DESIGN

Our multicenter, placebo-controlled, randomized, dose-ranging, double-blind clinical trial, called the Earlier versus Later Levodopa Therapy in Parkinson Disease (ELLDOPA) study, was conceived, organized, and implemented by the Parkinson Study Group and sponsored by the National Institute of Neurological Disorders and Stroke. The Department of Defense sponsored the single-photon-emission computed-tomography (SPECT) substudy. The subjects were enrolled between September 1998 and August 2001 at 33 sites in the United States and 5 sites in Canada. The study was approved by the institutional review boards at the participating sites, and all subjects gave written informed consent. An independent safety monitoring committee monitored the data, subjects' safety, and the tolerability of the study drug. There was no prespecified formal guideline for recommending either modification or termination of the trial.

SUBJECTS

The subjects were 30 years of age or older, had received a diagnosis of Parkinson's disease within the past two years, had a rating on the modified Hoehn–Yahr scale13,14 of less than stage 3 (with stage 1 indicating unilateral disease, stage 2 mild bilateral disease, and stage 3 more advanced bilateral disease), and were considered not likely to require therapy for symptoms of the disease within the nine months after enrollment in the study. Patients were excluded if they were receiving antiparkinson medication, had been exposed to levodopa or to any dopamine agonist for more than 14 days, had an identifiable cause of parkinsonism, or had a tremor in any limb that was given a score of 3 or more on the Unified Parkinson's Disease Rating Scale (UPDRS),14 freezing of gait, loss of postural reflexes, major depression, or dementia.

Potential subjects were informed that only the assigned study drug would be permitted during the nine months of the study and that if they needed additional antiparkinson medication during this period, they would have to withdraw from the study. Subjects were randomly assigned to receive placebo or carbidopa–levodopa at a dose of 12.5 and 50 mg three times daily, 25 and 100 mg three times daily, or 50 and 200 mg three times daily, respectively. The doses were increased to the full amount over a period of nine weeks in a blinded fashion.

After 40 weeks, the subjects underwent a 3-day period of step-down withdrawal from the study drug. After two weeks without the study drug, a final assessment of the severity of the symptoms of Parkinson's disease was made. The selection of the 2-week duration for the washout period was based on reports that withdrawal from levodopa for a period up to 14 days resulted in a worsening of parkinsonism mainly within the first 7 days, with nonsignificant worsening beyond that point.15–17

CLINICAL EVAL‎UATION

The treating investigator, who was blinded to the treatment assignment, performed a clinical eval‎uation with the use of the UPDRS14 at the screening, baseline, and interim visits (at the end of weeks 3, 9, 24, and 40) and during each of the two weeks of the washout phase. During the four interim visits, the eval‎uations were performed before the administration of the first dose of the study drug. At every visit the treating investigator inquired about adverse events. The primary rater, who was also blinded to the treatment assignment and was kept unaware of information obtained during the course of the study, performed the examination with the use of the UPDRS only at the baseline eval‎uation, which took place within four weeks after screening, and again at the final eval‎uation, two weeks after the subject had undergone complete withdrawal from the study treatment (week 42). The study coordinators and the subjects also were blinded to the treatment assignments. An emergency unblinding procedure was established but was never used.

OUTCOME

The prespecified primary outcome was the change in the severity of parkinsonism between the baseline visit and week 42, as measured with the use of the total score on the UPDRS that was obtained by the primary rater; week 42 was 14 days after the withdrawal of the study drug. The expected result was a dose-related deterioration during washout if levodopa were shown to hasten the progression of Parkinson's disease. The change in the total scores on the UPDRS measured by the treating investigator at each visit was a prespecified secondary outcome. The treating investigator assessed adverse events with the use of open-ended questioning at each visit.

SUBSTUDY

After the clinical trial was begun, we conducted a substudy with the use of SPECT to measure striatal dopamine-transporter density with the use of iodine-123–labeled 2-β-carboxymethoxy-3-β-(4-iodophenyl)tropane ([123I]β-CIT). The methods have been reported previously.18 The subjects who gave consent underwent SPECT imaging just before the baseline visit and then again before the visit at week 40. The imaging studies were performed at that point rather than after the subjects had undergone withdrawal from levodopa, at week 42, because it was thought that the subjects would not be able to tolerate the procedure (including the necessary travel) after the return of parkinsonism or after its worsening. Another reason was that there was no clear evidence of a short-term effect of levodopa on dopamine-transporter imaging.

All imaging studies were performed at Yale University or the Institute for Neurodegenerative Disorders (both in New Haven, Connecticut). The neuroimaging personnel were blinded to the treatment assignment of the subjects. The results of the SPECT studies were transferred to and analyzed by the biostatistics center of the Parkinson Study Group. The prespecified outcome was the percent change in the ratio of the specific striatal [123I]β-CIT uptake to the nondisplaceable striatal [123I]β-CIT uptake between the two images.

STATISTICAL ANALYSIS

Allowing for a dropout rate of 10 percent of the subjects enrolled, we chose a sample size of 360 subjects (i.e., 90 subjects in each of the four treatment groups) in order to provide the study with 85 percent power to detect a dose–response relationship (linear trend) between the assigned doses and the change in the subjects' total score on the UPDRS between the baseline visit and the final visit at week 42. On the basis of the results of one study,19 we anticipated a rate of worsening in the total score on the UPDRS in the placebo group of 9.5 units over the 9.5 months of the study. The study was powered to detect a linear trend that corresponded to a 4-unit difference (45 percent of 9.5 units) in the score on the UPDRS between the highest dose of levodopa (600 mg per day) and placebo — that is, a change of either 13.5 or 5.5 units from baseline in the group receiving the highest dose of the active study drug.

The primary statistical tests were two-tailed, with an alpha level of 0.05. Only subjects who completed the two-week washout phase were included in this analysis. We used the intention-to-treat principle in the analysis, even if among some subjects the dose of the study drug were to be reduced during the study. The primary analysis assessed the dose–response relationship between the assigned doses and the worsening of parkinsonism, as indicated by the changes in the total score on the UPDRS between the baseline visit and week 42. Statistical comparisons were made by analysis of covariance in a model that adjusted for differences among the investigators performing the eval‎uations and in baseline values.

ResultsSUBJECTS

Of a total of 361 subjects enrolled in the study, 317 (88 percent) took the study medication for 40 weeks, and 311 (86 percent) completed the 2 weeks of washout (Figure 1). The neuroimaging substudy was begun after the enrollment of the first 108 subjects. Of the 253 subjects subsequently enrolled, 142 (56 percent) participated in the substudy and underwent the baseline SPECT. Of these, 135 (95 percent) returned for scanning at week 40. The demographic and clinical characteristics of the subjects in the treatment groups were similar at baseline, both in the entire sample and in the neuroimaging substudy (Table 1).

FIGURE 1

Random Assignments to Treatment, Completion of the Trial, and Reasons for Not Completing It.

TABLE 1

Baseline Characteristics of the Study Subjects.

CLINICAL OUTCOME

Levodopa, in a dose–response pattern, significantly (P<0.001) reduced the worsening of symptoms of Parkinson's disease as reflected in the change between the total score on the UPDRS at baseline and that at week 42 (i.e., two weeks after washout of the study medication), as compared with the change in the placebo group (Figure 2 and Table 2). The subjects in the placebo group had mild improvement at the week 3 visit, but after that their symptoms worsened steadily throughout the balance of the study period, including the two-week washout phase. A strong dose–response benefit was detected during the period in which the medication was administered beginning at week 9, when the full dose of 600 mg daily was reached in the group receiving the highest dose of levodopa, and it persisted through week 40. The scores on the UPDRS in the three levodopa groups worsened during the two-week washout period, but these groups did not deteriorate to the level observed in the placebo group, and the group receiving the highest dose of levodopa had the best result (Figure 2 and Table 2). The adverse events that were significantly more common among those receiving levodopa at 600 mg daily than in the placebo group were dyskinesias, nausea, infection, hypertonia, and headache (Table 3).

FIGURE 2

Changes in Total Scores on the Unified Parkinson's Disease Rating Scale (UPDRS) from Baseline through Eval‎uation at Week 42.

TABLE 2

Changes in the Scores on the UPDRS between Baseline and Week 42.

TABLE 3

Adverse Events.

SPECT AND [123I]Β-CIT SUBSTUDY

The mean [123I]β-CIT uptake in the striatum, caudate, and putamen at baseline was consistent with values previously reported for patients with early Parkinson's disease.20 The percent decrease in striatal [123I]β-CIT uptake over the 40 weeks of the study treatment was greater among subjects in the levodopa groups than in the placebo group, but this difference was not statistically significant (Table 4). However, 21 of the 142 subjects (14.7 percent) had a putaminal [123I]β-CIT uptake of more than 3.25 at baseline (i.e., more than 75 percent of the age-expected putaminal uptake).21 An analysis of the results of SPECT after the exclusion of the 19 subjects without a dopaminergic deficit who returned for the neuroimaging study at week 40 showed a significantly greater decrease in [123I]β-CIT uptake among those receiving levodopa than among those receiving placebo (P=0.036) (Table 4).

TABLE 4

Change in Striatal [123I]β-CIT Uptake between Baseline and Week 40.

Discussion

We found no clinical evidence that levodopa accelerated the worsening of Parkinson's disease over the 9.5 months of observation. Rather, levodopa was associated with less worsening of parkinsonism than was placebo, consistent with the notion that it slows disease progression (Table 2 and Figure 2). In comparison with the scores on the UPDRS at baseline, the final scores, after the washout phase, had worsened by approximately 8 units in the placebo group, whereas the group receiving levodopa at 600 mg daily did not have evidence of deterioration. The groups receiving levodopa at lower doses did have deterioration, as shown by a comparison of their scores at baseline and at week 42, but the deterioration was less than in the placebo group (P<0.001).

We need to consider that a two-week washout from levodopa may have been insufficient to eliminate fully the effect of the medication on symptoms, and the results observed may be related to a profound effect of levodopa on symptoms that persists for a long time after the drug has been withdrawn. Indeed, Hauser and Holford,22 using a modeling technique, analyzed the withdrawal of levodopa in 20 patients and reported that the mean half-life of levodopa as measured by the loss of the clinical benefit was 7.9 days (95 percent confidence interval, 2.2 to 30.4 days). They suggest that a washout period of 32 days (four half-lives) may be required to eliminate 90 percent of the drug's effects on symptoms. However, we saw little deterioration after one week of washout. Near the end of the study, with the approval of the institutional review boards and of the National Institutes of Health, we asked the last 38 subjects remaining in the study to extend the washout period to four weeks. Among these subjects there was no further worsening of the scores on the UPDRS during the additional two weeks, but the small number of subjects renders this component of the study difficult to interpret.

Muenter and Tyce23 described motor responses with two types of duration, short and long, with levodopa therapy. The short-duration benefit lasts for a few hours after a single dose, and the long-duration benefit lasts several days. In our study, the long-duration benefit appears to have lasted approximately one to two weeks (Figure 2). To argue that a longer washout period might have revealed more clinical worsening than was observed, one could propose a hitherto unknown third type of duration of motor response, one that is more sustained than the so-called long-duration benefit. Such an enduring benefit could be envisioned to result from a prolonged pharmacodynamic effect, for example, on dopamine receptors.

If, however, the clinical effects observed offer evidence of neuroprotection, how can we explain the greater loss of dopamine transporter shown in the SPECT imaging studies of the subjects who received levodopa, as compared with those who received the placebo — a result that suggests the possibility of a levodopa-induced toxic effect on dopamine neurons? At the end of the study, when the SPECT neuroimaging studies were performed, the subjects were still taking levodopa, so it is possible to assume that levodopa has a pharmacologic effect on the dopamine transporter that interferes with and reduces the binding of the β-CIT ligand.

Indeed, one study found a reduction in dopamine-transporter binding with the use of positron-emission tomography (PET) and another dopamine-transporter ligand in patients with early Parkinson's disease who were treated for six weeks with levodopa at 300 mg daily.24 However, the sample size in this study, as in others that have shown no change in dopamine-transporter binding after short-term treatment with levodopa,25–27 was too small to demonstrate a statistically significant difference between levodopa and placebo. Further indirect support for the absence of a pharmacologic effect of levodopa on imaging studies of the dopamine transporter in the present study was the absence of decline in the [123I]β-CIT uptake by week 40 among the 16 subjects who received levodopa and whose SPECT scans were normal at baseline (i.e., without evidence of a dopaminergic deficit) (data not shown). However, in the absence of studies with larger samples and a longer period of treatment with levodopa, we cannot exclude the possibility that levodopa may simply down-regulate the dopamine transporter. Another consideration is that the falling dopamine concentration in the placebo group may have led to a compensatory increase in the activity of the dopamine transporter that could have increased [123I]β-CIT binding, but evidence for such an interpretation is lacking.

If levodopa has neuroprotective effects, what mechanisms could account for this property? In low concentrations and in the presence of glial cells, levodopa protects cultured dopaminergic neurons28–31 and up-regulates antioxidant and antiapoptotic proteins.29,32 Furthermore, in vivo studies suggest that levodopa can promote survival and enhance the sprouting of nigral dopamine neurons in rodents treated with the toxin 6-hydroxydopamine.9,10 The question of whether levodopa has a protective or a toxic effect in Parkinson's disease — reflecting the results of clinical examination and neuroimaging studies, respectively — cannot be answered with certainty, and future studies will be needed.

This dose–response, placebo-controlled clinical trial eval‎uating the effect of levodopa in patients with early Parkinson's disease showed a strikingly impressive dose–response clinical benefit: the higher the dose, the stronger and more lasting the benefit, and the benefit was greater even after the drug was withdrawn (Figure 2).

During withdrawal, we did not encounter the neuroleptic malignant-like effect (i.e., high fever, obtundation, and rigidity) that can occur with the sudden withdrawal of levodopa.33,34 Its absence may be related to the down-titration procedure employed; however, this complication is so rare that it might not have occurred in a study of this size even had levodopa been withdrawn suddenly, particularly in patients with early-stage disease.

Our study calls into question interpretations of the functional neuroimaging of the dopamine system. It has generally been assumed that imaging studies performed with 18F-fluorodopa PET and [123I]β-CIT SPECT can provide reliable information on the integrity of the nigrostriatal dopamine pathway. Our study and other recent clinical trials25,35 raise the possibility of a pharmacologic influence of dopaminergic therapy on these neuroimaging targets and point to the need for clarification of this issue.

Although the absence of evidence of a dopaminergic deficit in the imaging in our study and two other studies25,35 may simply represent a limitation of the sensitivity of the imaging techniques, the scans without a dopaminergic deficit raise doubt about whether the subjects had Parkinson's disease.35 In our study, the subjects with such scans at baseline had no worsening of the [123I]β-CIT uptake at week 40 (data not shown), nor did subjects with such scans in the group receiving the highest dose of levodopa have an improvement in the scores on the UPDRS (i.e., a change between the score at baseline and at week 40 of 3.38±4.25 units, as compared with a change of –4.95±10.4 units in the subjects receiving levodopa at 600 mg daily whose baseline SPECT scans showed abnormalities; P=0.002). Whether the subjects with no evidence of dopaminergic deficit on scanning do or do not have classic Parkinson's disease remains uncertain. The fact that scans of some patients with early Parkinson's disease do not have evidence of a dopaminergic deficit needs to be taken into consideration in the planning of future trials to test drugs for neuroprotective effects.

Finally, even though we cannot reconcile the clinical and imaging findings in our study, we can assure both patients with early Parkinson's disease and their physicians that, from a clinical perspective, our study did not find that levodopa hastens the progression of Parkinson's disease. On the basis of the study, we can recommend that the doses of levodopa be adjusted to fit the needs of the patient. Small doses were found to be effective, although less so than higher doses. High doses, however, were associated with a greater frequency of adverse events such as dyskinesia. For the present, until more evidence is available, we recommend customizing the dose of levodopa to the needs of the individual patient on the basis of the clinical response and the profile of adverse events.

NOTES

Supported by grants from the National Institute of Neurological Disorders and Stroke (NS34796, to Dr. Fahn), the Department of Defense (DAMD 17-99-1-9472, to Dr. Marek), and the General Clinical Research Center of the National Center for Research Resources, National Institutes of Health (MO1-RR-00044 and MO1-RR-02066). Carbidopa–levodopa tablets and the matching placebo tablets were kindly provided by Teva Pharmaceuticals (Israel).

Drs. Fahn, Oakes, Shoulson, Kieburtz, Lang, Tanner, and Marek report having served as unpaid consultants to Teva Pharmaceuticals, and Dr. Olanow reports having served as a paid consultant to Teva Pharmaceuticals. Drs. Marek and Seibyl have an equity interest in Molecular Neuroimaging (New Haven), which carried out the [123I]β-CIT SPECT imaging for this study.

We are indebted to the subjects for their participation in the study; to K. Hyland (Baylor Medical College, Houston) for the independent, blinded chemical analyses of the contents of the tablets conducted annually to ascertain the chemical stability of levodopa and carbidopa; to S. Bennett, A. Brocht, D. Graffrath, J. Janciuras, C. Orme, L. Preston, K. Rothenburgh, C. Weaver, and A. Watts of the Biostatistics and Coordination Centers at the University of Rochester, Rochester, N.Y.; to the Independent Safety Monitoring Committee (P. Tariot, chair, Monroe Community Hospital, Rochester, N.Y.; to W.J. Hall, University of Rochester; to R. Rodnitzky, University of Iowa, Iowa City); to the National Institute of Neurological Disorders and Stroke Safety Monitoring Committee (E.C. Haley, chair, University of Virginia, Charlottesville; D. Eidelberg, North Shore University Medical Center, Manhasset, N.Y.; C.A. Gatsonis, Brown University, Providence, R.I.; W. Rocca, Mayo Clinic, Rochester, Minn.); and to E.J. Oliver (administrator), National Institute of Neurological Disorders and Stroke, Bethesda, Md.

APPENDIX

The following members of the Parkinson Study Group participated in the Earlier versus Later Levodopa Therapy in Parkinson Disease (ELLDOPA) study and contributed to this report. Steering committee: S. Fahn, principal investigator, Columbia University, New York; D. Oakes, chief biostatistician; I. Shoulson, coprincipal investigator; K. Kieburtz, director, Clinical Trials Coordination Center; A. Rudolph, senior project coordinator, University of Rochester, Rochester, N.Y.; K. Marek, neuroimager; J. Seibyl, neuroimager, Institute for Neurodegenerative Disorders, New Haven, Conn.; A. Lang, Toronto Western Hospital, Toronto; C.W. Olanow, Mount Sinai School of Medicine, New York; C. Tanner, The Parkinson's Institute, Sunnyvale, Calif.; G. Schifitto, medical monitor, University of Rochester, Rochester, N.Y.; H. Zhao, biostatistician, University of Rochester, Rochester, N.Y.; L. Reyes, administrator, Columbia University, New York; A. Shinaman, administrator, University of Rochester, Rochester, N.Y.; Participating primary raters, treating investigators, and coordinators: Rush–Presbyterian–St. Luke's Medical Center, Chicago: C. Comella, C. Goetz, L. Blasucci; Barrow Neurological Institute, Phoenix, Ariz.: J. Samanta, M. Stacy, K. Williamson, M. Harrigan; Columbia University, New York: P. Greene, B. Ford, C. Moskowitz; Parkinson's and Movement Disorder Institute, Fountain Valley, Calif.: D. Truong, M. Pathak; Baylor College of Medicine, Houston: J. Jankovic, W. Ondo, F. Atassi, C. Hunter; Brown University, Providence, R.I.: C. Jacques, J.H. Friedman, M. Lannon; Institute for Neurodegenerative Disorders, New Haven, Conn.: D.S. Russell, D. Jennings, B. Fussell; Massachusetts General Hospital, Boston: D. Standaert, M.A. Schwarzschild, J. Growdon, M. Tennis; McGill Centre for Studies in Aging, Verdun, Que., Canada: S. Gauthier, M. Panisset, J. Hall; Oregon Health and Science University, Portland: S. Gancher, J. Hammerstad, C. Stone, B. Alexander-Brown; Albany Medical College, Albany, N.Y.: S. Factor, E. Molho, D. Brown (deceased), S. Evans; Scott and White Hospital–Texas A&M University, Temple: J. Clark, B. Manyam, P. Simpson, B. Wulbrecht, J. Whetteckey; University of Alberta, Edmonton, Canada: W. Martin, T. Roberts, P. King; University of South Florida, Tampa: R. Hauser, T. Zesiewicz, L. Gauger; University of Virginia, Charlottesville: J. Trugman, G.F. Wooten, E. Rost-Ruffner; Washington University, St. Louis: J. Perlmutter, B. Racette; University of Calgary, Alta., Canada: O. Suchowersky, R. Ranawaya, S. Wood, C. Pantella; University of Rochester, Rochester, N.Y.: R. Kurlan, I. Richard, N. Pearson; Mayo Clinic, Scottsdale, Ariz.: J. Caviness, C. Adler, M. Lind; University of Pennsylvania, Philadelphia: T. Simuni, A. Siderowf, A. Colcher, M. Lloyd; University of Miami, Miami: W. Weiner, L. Shulman, W. Koller, K. Lyons; Boston University, Boston: R. Feldman (deceased), M.-H. St.-Hilaire, S. Ellias, C.-A. Thomas; Emory University, Atlanta: J. Juncos, R. Watts, A. Partlow; The Parkinson's Institute, Sunnyvale, Calif.: J. Tetrud, D.M. Togasaki, M. Welsh, T. Stewart; University of Medicine and Dentistry of New Jersey–Robert Wood Johnson, New Brunswick: M.H. Mark, J.I. Sage, D. Caputo; Louisiana State University, New Orleans: H. Gould, J. Rao, A. McKendrick; Mount Sinai School of Medicine, New York: M. Brin, F. Danisi, R. Benabou; Ohio State University, Columbus: J. Hubble, G. Paulson, C. Reider; Toronto Western Hospital, Toronto: A. Birnbaum, J. Miyasaki, L. Johnston, J. So; University of Kansas, Kansas City: R. Pahwa, R. Dubinsky; Mayo Clinic, Jacksonville, Fla.: Z. Wszolek, R. Uitti, M. Turk; Minneapolis Veterans Affairs Hospital, Minneapolis: P. Tuite, D. Rottenberg, J. Hansen; University of Puerto Rico, San Juan: C. Serrano Ramos; University of Southern California, Los Angeles: C. Waters, M. Lew, M. Welsh, C. Kawai; Colorado Neurological Institute, Englewood: C. O'Brien, R. Kumar, L. Seeberger, D. Judd; Ottawa Civic Hospital, Ottawa: T. Mendis, C.L. Barclay, D.A. Grimes, L. Sutherland; Johns Hopkins University, Baltimore: T. Dawson, S. Reich, R. Dunlop; University of Michigan, Ann Arbor: R. Albin, K. Frey, K. Wernette.