Lumbosacral radiculopathy: Pathophysiology, clinical features, and diagnosis - 수련의

[문형철]

허리디스크로 인한 radiculopathy에 대한 최신 review논문이다.

한번 읽어볼만 할듯

Lumbosacral radiculopathy. Pathophysiology, clinical featur.pdf

INTRODUCTION — Lumbosacral radiculopathy is a condition in which a disease process affects the function of one or more lumbosacral nerve roots [1]. The clinical aspects of lumbosacral radiculopathy will be reviewed here. The treatment of lumbosacral radiculopathy and other disorders of the lower spine are discussed separately. (See "Lumbosacral radiculopathy: Prognosis and treatment" and "Lumbar spinal stenosis: Pathophysiology, clinical features, and diagnosis" and "Approach to the diagnosis and eval‎uation of low back pain in adults" and "Subacute and chronic low back pain: Pharmacologic and noninterventional treatment" and "Subacute and chronic low back pain: Nonsurgical interventional treatment" and "Subacute and chronic low back pain: Surgical treatment".)

ANATOMY — The lumbar spine consists of five movable lumbar vertebral bodies, numbered L1 to L5 (figure 1). The sacrum is made up of five developmentally fused vertebral levels (S1 to S5), followed by a terminal bony prominence, the coccyx. The entire region is commonly described as the lumbosacral spine. Directly beneath each lumbar and sacral vertebrae, there is a pair of neural foramina with the same number designation, such that the L1 neural foramina are located just below the L1 vertebral body. Neural foramina are bounded superiorly and inferiorly by pedicles, anteriorly by the intervertebral disc and vertebral body, and posteriorly by facet joints (figure 1).

Through each neural foramen passes the same numbered spinal nerve root, recurrent meningeal nerves, and radicular blood vessels. On each side there are five lumbar, five sacral, and one coccygeal spinal nerve roots. All lumbar and sacral spinal nerve roots originate at the T10 to L1 vertebral level, where the spinal cord ends as the conus medullaris. A dorsal (somatic sensory) root from the posterolateral aspect of the spinal cord and a ventral (somatic motor) root from the

anterolateral aspect of the cord join in the spinal canal to form the spinal nerve root (figure 1). The roots then course down through the intraspinal canal, forming the cauda equina, until they exit at their respective neural (intervertebral) foramina. Thus, the lumbosacral nerve roots exit the spinal canal at a lower level than where they arise. A potential consequence of this arrangement is that intraspinal pathology may affect roots at higher levels than the level where the roots exit [2,3].

Cell bodies of the motor nerve fibers are located in the ventral (anterior) horns of the spinal

cord, while those of the sensory nerve fibers are in a dorsal root ganglion at each lumbar

and sacral level. Dorsal root ganglia (DRG) tend to be located within the neural foramina,

and are therefore not strictly speaking intraspinal (ie, within the lumbar canal). However, at

the low lumbar and sacral levels there is a tendency for DRG to reside proximal to the

neural foramina, within the intraspinal canal, as found in 11 to 38 percent of cases at L5

and 71 percent at S1 [4,5]. The dorsal root ganglia are attached to the vertebral body on

the transverse process [6]. Compressive radicular disease typically occurs proximal to this.

As noted above, the spinal cord normally terminates at the conus medullaris within the

lumbar intraspinal canal between the T10 and L1 vertebral levels. Exceptions include

patients with congenital spinal deformities known as spina bifida, in which the fetal conus is

tethered to ligamentous or bony structures, causing lengthening of the spinal cord during

development. In such patients, the conus medullaris can be displaced downward to the

middle or lower lumbar spine.

Rami — Just distal to the neural foramen, the nerve root divides in two, forming the dorsal

and ventral primary rami.

The small dorsal (posterior) primary ramus supplies motor innervation to the

paraspinal muscles and cutaneous innervation to the skin of the trunk and back

The large ventral (anterior) primary ramus supplies motor and sensory innervation to

the legs and trunk, including abdominal wall muscles

The dorsal rami of the spinal nerves also supply the apophyseal joints and the paraspinal

muscles. They innervate structures both above and below the level of the nerve. Clinical

eval‎uation of injury to the dorsal rami is difficult because of the overlap in areas innervated

by these nerves and because of the limited ability to clinically eval‎uate individual paraspinal

muscles. However, electromyography can be helpful in determining the distribution of

disease.

The ventral rami innervate the extremities and the trunk. These branches can be eval‎uated

by assessing the motor and sensory functions of the different myotomes and dermatomes,

respectively. However, variability to the dermatomal and myotomal distribution of

innervation exists [7,8].

Myotomes and dermatomes — The collection of muscles with significant innervation

from a single root is called a myotome. Similarly, the sensory distribution of a single root is

labeled a dermatome.

The primary manifestations of lumbosacral root disease can be broken up into dysfunction

of two distinct systems: motor and sensory. Motor dysfunction from a root lesion may

cause weakness in some or all muscles innervated by that root. However, many muscles

have innervation from multiple roots, which may result in preserved strength despite

significant involvement of a single root.

Although there are classic descriptions for the distributions of myotomes and dermatomes,

substantial variability exists in these distributions from person to person [7,8]. Sensory

fields have considerable overlap, but there are areas that are exclusively served by

individual nerves. These areas are called autonomous zones. The most important of these

in the eval‎uation of lumbosacral radiculopathies include the sole of the foot (S1), dorsum of

the foot (L5), medial calf (L4), and anterior thigh (L2 and 3) [9].

Lumbosacral myotomes are listed in the Table (table 1).

Sinuvertebral nerves — The sinuvertebral nerves are sensory nerves that innervate

various structures within the spine, such as ligamentous structures, the dura, periosteum,

and blood vessels [10]. They originate distal to the dorsal root ganglia and extend to

communicate with branches from radicular levels both above and below the level of entry,

as well as the contralateral side, making it difficult to localize pain from involvement of

these nerves.

Irritation of the sinuvertebral nerves may result in low back pain. Because they arise distal

to the nerve root, however, involvement of sinuvertebral nerves or their branches without

involvement of the rest of the nerve root is not considered to be radicular in nature.

PATHOPHYSIOLOGY AND ETIOLOGY — The most common etiology of lumbosacral

radiculopathy is nerve root compression caused by a disc herniation or spondylosis (ie,

spinal stenosis due to degenerative arthritis affecting the spine). Additional etiologies

include nonskeletal causes of nerve root compression and noncompressive mechanisms

such as infection, inflammation, neoplasm, and vascular disease.

Pain generators in the lumbosacral spine — A number of tissues making up the low

back contain nerve fibers with pain receptors. These are listed in the Table (table 2). Of

note, intervertebral disc material does not contain significant numbers of pain fibers.

Classic radiating pain from spinal nerve injury is mediated through proximal spinal

nerves. Compression of a spinal nerve root by disc or arthritic spur leads to local

edema, ischemia, and inflammation. These factors contribute to production of pain

impulses through the spinal nerves. The pain from acute disc herniation or

spondylotic spinal nerve entrapment reflects a combination of pain generation within

the nerve root itself, as well as the pain from neighboring tissues whose pain fibers

are activated by the effect of disc herniation on dura, ligaments, and surrounding

vasculature.

Localized lumbosacral pain is thought to arise from intraspinal structures. These pain

impulses arise from the blood vessels, dura mater, and longitudinal ligaments, and

travel in the sinuvertebral nerves through the neural foramina, connecting via rami

communicantes with the extraspinal sympathetic chain.

Nonlocalized, nonradiating pain is thought to arise from muscle, bone, and ligament

outside the spinal canal. Interconnected ventral and dorsal nerve plexuses surround

the vertebral column [11]. The ventral nerve plexus serves the anterior longitudinal

ligament and has bilateral innervation. Many branches from the sympathetic trunk,

rami communicantes, and perivascular nerve plexuses join to form the ventral nerve

plexus. The dorsal nerve plexus arises from the sinuvertebral nerves and serves the

posterior longitudinal ligament.

Referred spine pain may arise from the abdominal viscera that share the same spinal

level of innervation. Organs that can potentially refer pain to the spine include the

aorta, pancreas, duodenum, colon, rectum, kidney, ureter, bladder, and pelvic organs.

Systemic illness can refer pain to the bony spine or can produce bony disease that

generates spine pain. The major categories and clues to their diagnosis are listed in

the Table (table 3).

Degenerative changes — Damage to spinal nerve roots occurs as the result of

degenerative change involving three main structures, which are:

The intervertebral discs

The uncovertebral joints

The zygapophyseal (facet) joints

Resulting bony overgrowth (osteophytes) or disc herniation at these points may directly

impinge on spinal nerve roots or the spinal cord, or their effect may be primarily to produce

instability and misalignment of the spine (ie, degenerative spondylolisthesis) that in turn

produces pain and neurologic deficits. It is not known whether changes in these different

structures are causally related or occur independently.

Degenerative spondylotic changes are common with aging and usually do not result in

radiculopathy. Low back pain associated with degenerative changes is distinct from

radicular pain and is much more frequent. Given the inability to ascertain the exact cause,

it is commonly referred to as nonspecific low back pain.

One school of thought suggests that degenerative spondylotic change is led by age-related

change in the nucleus pulposus of the disc. With age there is gradual narrowing of the disc

space coincident with changes in disc proteoglycan composition. Later, cracks develop in

the disc, and deposits of gas and calcification may form. Eventually the disc material

becomes desiccated and friable. Age-related changes also occur in the annulus fibrosus,

which becomes more fibrotic and less elastic. Fissures develop, and calcium is deposited.

As the disc shrinks and the intervertebral disc space narrows, the annulus tends to buckle

out.

Accompanying this disc degeneration are changes at the vertebral body endplates adjacent

to the disc. The marrow undergoes fibrovascular change or fatty marrow replacement.

Finally, endplate sclerosis develops. Osteophyte formation occurs at the margins of the

vertebral bodies. What triggers osteophyte formation is unclear, although spinal movement

at ligamentous attachment sites and loss of buffering tissues between bony surfaces likely

play roles. Osteophyte production appears to slow as advancing spondylosis leads to

decreasing spinal movement.

Uncovertebral joints are not true joints, but they may represent a slit in the intervertebral

disc at the point where the uncinate process makes contact with the disc and vertebral

body above. As disc substance decreases, there is more contact between the uncinate

process and adjacent bone, leading to osteophyte formation. Dorsal protrusion of these

osteophytes narrows the adjacent neural foramen.

Facet joint degeneration may not be directly related to spondylotic changes, but often

coexists. Disc degeneration is likely to put additional weight-bearing strains on facet joints,

which are not weight-bearing structures. With unnatural movement of the spine, the

synovial joint bears more structural burden, degenerates, and develops osteophytes. These

osteophytes grow into the posterior aspect of the neural foramen (picture 1).

Disc protrusion and level of injury — The relationship between disc disease and

lumbosacral radiculopathy was first identified in 1934 [12]. Disc protrusion can give rise to

different anatomic levels of nerve root compression depending on the orientation of spinal

nerve roots as they exit from the spinal cord. As noted earlier, all lumbar and sacral spinal

nerve roots are constituted at the T12-L1 vertebral level, where the spinal cord ends as the

conus medullaris. The roots then course down the canal as the cauda equina, until they exit

at their respective neural foramina.

Depending upon the nature and location of intraspinal compression, roots may be injured

at any disc level, from the L1-2 level to the level of their exit into their neural foramina. For

example, the L5 root can be compressed by a central disc protrusion at the L2-3 or L3-4

level, a lateral disc protrusion at the L4-5 level (picture 2), or disc protrusion into the

foramen at the L5-S1 level (picture 3). Because of the presence of multiple spinal nerve

roots in the cauda equina, there is increased likelihood of multiple, bilateral simultaneous

nerve root compressions.

Nerve root injury may be complete or only partial and thus may involve all or only a subset

of root fibers. With partial injury, incomplete myotomal involvement may result, making the

distinction between a radiculopathy and a peripheral nerve injury more challenging.

The lumbosacral spine is susceptible to disc herniations because of its mobility from flexion,

extension, and torsion. Seventy-five percent of flexion and extension occurs at the

lumbosacral joint [13]. This level, on the other hand, has limited torsion. Twenty percent of

flexion and extension occurs at L4-L5. The remaining 5 percent occurs between L1 and L3

[13]. As the L4-L5 and L5-S1 levels are most susceptible to injuries from routine

movements of the spine, about 90 to 95 percent of compressive radiculopathies occur at

these levels [14]. The incidence of radiculopathies is split somewhat evenly between L4-L5

and L5-S1, as the lack of torsion at L5-S1 helps to increase its stability despite its higher

degree of flexion and extension [15]. Next in frequency is L4. Other levels are uncommon.

Far lateral herniations are seen more often at the L2-4 levels. They may affect the rostral

root. This is rare, as only 10 percent of far lateral herniations will result in nerve root

compression [5]. Pain from the far lateral disc herniations may be more severe due to

compression of the dorsal root ganglia [16].

Other skeletal causes — Congenital abnormalities of the bony spinal column or its

contents occur in 25 to 50 percent of the general population [17]. Congenital narrowing of

the canal is not uncommon as a substrate upon which spondylosis may result in neurologic

disease; this is especially true of younger adults. Other developmental abnormalities

include tethered cord or diastematomyelia and spina bifida. They may lead to radicular

dysfunction due to injuries resulting from traction of the root.

Root avulsion is a rare cause of lumbosacral radiculopathy that may occur with fractures of

the sacroiliac joint or with diastasis of the symphysis pubis or the pubic rami [18].

Nonskeletal causes — Infection, inflammation, neoplasm, and vascular disease are less

common causes of radiculopathy and polyradiculopathy (table 4) [19-22]. Specific

etiologies include the following:

Diabetes mellitus, a common cause of peripheral neuropathy and polyneuropathy

Infectious diseases

Borrelia burgdorferi

Cytomegalovirus

Epidural abscess

Epstein-Barr virus

Herpes simplex virus

Human immunodeficiency virus (HIV)

Mycobacterium

Mycoplasma

Syphilis

Varicella zoster virus (Herpes zoster or shingles)

Inflammatory conditions

Acute inflammatory demyelinating polyradiculoneuropathy (AIDP; Guillain-Barre

syndrome)

Chronic inflammatory demyelinating polyradiculoneuropathy (CIDP)

Arachnoiditis related to postsurgical changes

Chemical radiculitis

Sarcoidosis

Mass lesion or malignancy

Metastasis (most common)

Epidural abscess

Intradural tumor, particularly meningioma, neurofibroma, and ependymoma [19]

Lymphoma

Myeloma

Root sleeve cyst, such as a Tarlov or perineural cyst

Vascular

Arteriovenous malformation

Vasculitis (nerve root infarction)

Radiation-induced vascular occlusion

There is generally complete involvement of the root with radiculitis caused by

leptomeningeal carcinomatosis, cytomegalovirus, or herpes zoster.

With leptomeningeal carcinomatosis, the source may be metastases from breast cancer,

lung cancer and melanoma, non-Hodgkin lymphoma, or leukemia [21]. Nerve roots may be

encased by tumor cells causing compression and ischemia.

Cytomegalovirus radiculitis occurs in immunosuppressed patients with very low CD4

counts. It is characterized by severe paresthesia, and sensory loss and weakness. Both

small and large nerve fibers are involved. There may also be bowel and bladder dysfunction

from diminished sphincter tone. Cerebrospinal fluid analysis may reveal increased protein,

decreased glucose, and a polymorphonuclear pleocytosis. Polymerase chain reaction may

be positive for cytomegalovirus [20].

Herpes zoster colonizes the dorsal root ganglion and may remain latent for years.

Reactivation is associated with a hemorrhagic lymphocytic infiltration of the ventral roots.

Patients present with an erythematous vesicular maculopapular rash in the dermatome of

the affected root that lasts for three to five days. Sensory changes characterized by marked

allodynia often follow [23].

Root infarction may occur from either large vessel or small vessel disease. It is most

commonly seen in diabetics where the microvasculature is affected. Such a process may be

one of the causes of the clinical syndrome known as diabetic amyotrophy [24]. Diabetic

amyotrophy involves the L2, L3, and L4 roots and presents with weakness, pain, and

dysesthesia primarily in the proximal leg that evolves over days to weeks. This condition is

discussed separately. (See "Epidemiology and classification of diabetic neuropathy", section

on 'Diabetic amyotrophy (lumbar polyradiculopathy)'.)

Radiculopathy may also arise in the absence of obvious compression, leading to the

hypothesis that the injury may be due to a chemical radiculitis, perhaps caused by the

rupture of the annulus fibrosus with subsequent release of inflammatory mediators tracking

along the nerve root sheath [25-27]. Electromyography may be abnormal even though

compression is not observed on imaging studies.

Fiber size — Like most peripheral nerves, lumbosacral nerve roots are composed of both

large and small fibers. Different fiber types play different roles by virtue of the information

they carry. Injury to these fibers may result in focal demyelination as seen initially with

compression, or axonal damage as seen with infarction or more severe compressive injury.

Large fibers carry motor efferent and sensory information involved with vibratory sense and

conscious and unconscious proprioception. The typical clinical findings of large fiber

involvement in radiculopathies are weakness and reflex abnormalities. Deep tendon

reflexes may be diminished or lost. Clinical testing for proprioception and vibratory sense

tends to extend across different dermatomes and thus is often not as helpful in the

eval‎uation of lumbosacral radiculopathies as small fiber modalities, as discussed below.

The smaller myelinated A-delta fibers and unmyelinated C fibers carry pain and

temperature information. A-delta fibers transmit cold sensation, while C fibers subserve

warm sensation. Isolated injury to these fibers is not common in lumbosacral

radiculopathies. However, their dermatomal distributions tend to be more specific than

those of the large fibers, increasing their importance in the clinical eval‎uation of

radiculopathies.

In addition, dysfunction of A-delta and C fibers even in the absence of overt nerve root

compression may play a significant role in the symptomatology of radiculopathies. A study

of warm and cold sensory thresholds in 32 patients with unilateral L5 or S1 radiculopathy

and disc herniation found that warm thresholds (C fiber function) were impaired to a

greater extent than cold thresholds (A-delta fibers) [28]. Since it is thought that

unmyelinated axons, such as C fibers, are less affected by compression than myelinated

ones, this finding suggests that the generation of lumbar radicular pain is mediated by

inflammation more than by nerve root compression [28].

Central pain sensitization may result from nerve root compression. Supporting evidence

comes from a small study of patients with L5 or S1 radiculopathy that demonstrated a

lowered pain threshold in the contralateral root, perhaps mediated by pre- and

postsynaptic modulation of opioidergic interneurons [29].

EPIDEMIOLOGY — Lumbosacral radiculopathy is one of the most common problems seen

in neurologic consultation. Although data are limited, the estimated lifetime preval‎ence is

approximately 3 to 5 percent for adults, with equal rates among men and women [1].

CLINICAL PRESENTATIONS — The clinical presentations of lumbosacral radiculopathy

vary according the level of nerve root or roots involved. The most frequent are the L5 and

S1 radiculopathies. All lumbosacral nerve roots exit the spinal canal at the neural foramina

below their respective vertebrae. As an example, the L5 nerve roots exit via the neural

foramina at the L5/S1 disc space level. Thus, posterolateral disc herniation of the L4/L5

disc usually compresses the L5 nerve root, while posterolateral disc herniation of the L5/S1

disc typically compresses the S1 nerve root.

L1 radiculopathy — Lumbar disc herniation at the L1 level is rare, and thus L1

radiculopathy is uncommon. Symptoms on presentation generally involve pain, paresthesia,

and sensory loss in the inguinal region [1]. Rarely, minor hip flexion weakness is present.

L2/L3/L4 radiculopathy — There is marked overlap of the L2, L3, and L4 innervation of

the anterior thigh muscles, making it difficult to differentiate these spinal nerve root levels

based on symptoms, neurologic examination, or electrodiagnostic testing. Thus, these

radiculopathies are generally considered as a group. These nerve roots are most commonly

involved in older patients with symptoms of spinal stenosis.

Acute back pain is the most common presenting complaint, often radiating around the

anterior aspect of the leg down into the knee and occasionally down the medial aspect of

the lower leg as far as the arch of the foot. On examination, there may be weakness of hip

flexion, knee extension, and hip adduction. Higher lesions may result in greater weakness

of the hip flexors. Sensation may be reduced over the anterior thigh down to the medial

aspect of the lower leg. A reduced knee reflex is common in the presence of moderate

weakness.

Electromyography and nerve conduction studies may reveal abnormalities confined to

muscles of the affected root(s), including the quadriceps, leg adductors, and iliopsoas, with

associated paraspinal abnormalities. Saphenous sensory response remains normal even if

sensory loss is prominent in the distal leg.

L5 radiculopathy — L5 radiculopathy is the most common radiculopathy affecting the

lumbosacral spine. It often presents with back pain that radiates down the lateral aspect of

the leg into the foot. On examination, strength can be reduced in foot dorsiflexion, toe

extension, foot inversion, and foot eversion. Weakness of leg abduction may also be

evident in severe cases due to involvement of gluteus minimus and medius. Atrophy may

be present in the extensor digitorum brevis muscle of the foot and the tibialis anterior

muscle of the lower leg. In severe cases, there may be "tibial ridging", a condition in which

the normal convexity of the anterior compartment of the leg is lost because of atrophy,

leaving a prominent sharp contour of the medial aspect of the tibial bone.

Sensory loss is confined to the lateral aspect of the lower leg and dorsum of the foot, but

may be obvious only when testing sharp sensation in the web space between the first and

second digits. Reflexes are generally normal, although the internal hamstring reflex may be

diminished on the symptomatic side.

Electromyography and nerve conduction studies typically show abnormalities in the L5

muscles, including the gluteus medius and minimus, tensor fascia lata, tibialis anterior,

tibialis posterior, short head of the biceps femoris, and the L5 paraspinals. Sensory studies

(sural and superficial peroneal responses) are normal since the lesion is almost always

proximal to the dorsal root ganglion.

S1 radiculopathy — In S1 radiculopathy, pain radiates down the posterior aspect of the

leg into the foot from the back. On examination, weakness of plantar flexion

(gastrocnemius muscle) is specific. There may also be weakness of leg extension (gluteus

maximus) and toe flexion. Sensation is generally reduced on the posterior aspect of the leg

and the lateral edge of the foot. Ankle reflex loss is typical.

Electromyography and nerve conduction studies reveal abnormalities in S1 innervated

muscles, including the gluteus maximus, long head of the biceps femoris, gastrocnemius

and soleus muscles, and the S1 paraspinals, with intact sensory responses (sural generally

tested). Soleus H reflex testing also can be performed bilaterally to identify amplitude loss

or latency prolongation on the affected side. (See "Nerve conduction studies: Late

responses", section on 'H reflex'.)

S2/S3/S4 radiculopathy — Structural radiculopathies at these lower levels are less

common than other lumbosacral radiculopathies, but may be caused by a large central disc

compressing the nerve roots intrathecally at a higher level (eg, L5). Patients can present

with sacral or buttock pain that radiates down the posterior aspect of the leg or into the

perineum. Weakness may be minimal, with urinary and fecal incontinence as well as sexual

dysfunction.

The utility of electromyography and nerve conduction studies of the legs for the eval‎uation

of S2, S3, and S4 radiculopathy is limited. However, electromyographic investigation of the

low paraspinal muscles may be helpful. In addition, abnormalities may be identified in

gluteus maximus and gastrocnemius. Sensory studies are normal. Prolongations in latency

when performing electrical bulbocavernosus reflex testing may indicate a lesion in the

region, although this is not specific for radiculopathy.

Severity — Acute lumbosacral radiculopathy can be separated into three general

categories from least to most severe:

Pure sensory/painful radicular pattern, characterized by radicular pain and a

segmental pattern of sensory dysfunction but no other neurologic deficits

Mild motor deficit pattern, characterized by radicular pain, sensory dysfunction, and

mild nonprogressive segmental motor weakness and/or reflex change

Marked motor deficit pattern, characterized by radicular pain and sensory dysfunction

with severe or worsening motor deficits

DIAGNOSIS — The diagnosis of a lumbosacral radiculopathy is clinical, and can usually be

made based upon compatible symptoms and examination findings.

Disc herniation and foraminal stenosis due to spondylotic degeneration are the most

common etiologies for lumbosacral radiculopathy, and clinical symptoms are self-limited in

most cases. Thus, immediate diagnostic testing is not necessary for patients with

suspected radiculopathy who are neurologically intact and at low risk for neoplastic,

infectious, or inflammatory etiologies. Nevertheless, testing is suggested to confirm‎ the

diagnosis and etiology for patients who have persistent symptoms that are not adequately

controlled with conservative therapy and who are candidates for invasive treatment

options.

We recommend urgent neuroimaging in the initial assessment - followed by

electromyography if imaging is negative or equivocal - for patients with any of the following

conditions [1]:

Acute radiculopathy with rapidly progressive neurologic deficits

Radiculopathy with urinary retention, saddle anesthesia, or bilateral neurologic

symptoms or signs

Suspected neoplasm

Suspected epidural abscess

In addition, we recommend a lumbar puncture and cerebrospinal fluid analysis for patients

with a suspected inflammatory or infectious cause of lumbosacral radiculopathy.

The most useful modalities in the eval‎uation of lumbosacral radiculopathy are MRI, CT,

electromyography (EMG) and nerve conduction studies (NCS). A MRI scan of the

lumbosacral spine will identify most pathologic states that may benefit from surgical

intervention. CT or MRI myelography can assess the anatomy of the root sleeve. The

sensitivities of CT and MRI are similar for compressive radiculopathies [30]. EMG and NCS

have a high diagnostic accuracy for radiculopathy when neurologic weakness is present for

at least three weeks.

History — Pain and sensory symptoms such as paresthesia, dysesthesia, hyperesthesia, or

anaesthesia that involve a specific lumbosacral dermatome are suggestive of a radicular

process. Paresthesia occurs in 63 to 72 percent of cases, radiating pain in about 35

percent, and numbness in approximately 27 percent [31,32].

Similarly, weakness confined to the muscles in a particular lumbosacral myotome should

raise suspicion for radiculopathy, though few people spontaneously report such specific

complaints. Inability to get up from a chair suggests iliopsoas or quadriceps weakness,

while buckling of the knee is consistent with quadriceps weakness, and dragging of the toe

points to tibialis anterior weakness.

Elucidation of triggering and alleviating factors may also be helpful. Radicular pain that

worsens with Valsalva or improves while lying down suggests a compressive etiology.

Conversely, radicular pain that worsens with lying down suggests an inflammatory or

neoplastic etiology. However, such symptoms have not been shown to be sensitive or

specific for these conditions.

The acute onset of symptoms with bending, lifting, or trauma may herald a radiculopathy.

However, none of these factors is specific for radiculopathy. Bowel/bladder symptoms,

particularly new urinary incontinence, suggest a cauda equina syndrome.

Physical examination — Eval‎uation for lumbosacral radiculopathy requires a careful

neurologic examination. A summary of findings for specific nerve root levels of involvement

is found in the Table (table 5).

Additional procedures (eg, straight leg raising, reverse straight leg raising) may be useful

but their specificity, sensitivity, and reproducibility are variably limited. Vertebral

tenderness is suggestive of infection but is not specific enough to be clinically useful [14].

The spine should be eval‎uated for any cutaneous abnormalities, such as tufts of hair, nevi,

or pores, that may be suggestive of a neural tube developmental abnormality.

Deep tendon reflexes should be assessed at both the quadriceps (L2/L3/L4) and Achilles

(S1). Unfortunately, there is no reliable reflex to assess L5, the most common root

compressed in mechanical lumbosacral radicular disease. An experienced clinician may try

to elicit an L5-mediated deep tendon reflex by tapping the tendon of the tibialis posterior

adjacent to the medial malleolus, or the tendon of the internal hamstring, which is served

by the peroneal division of the biceps femoris muscle (lateral). However, the specificity of

these tests and their reliability is low. Side to side comparisons are sometimes useful.

The sensory examination is more subjective than other parts of the neurologic examination

and is further confounded by dermatomal overlap and variability. However, it can provide

important information for localization. Thus, we suggest testing the relevant lumbosacral

dermatomes for light touch, pain, and temperature. Vibratory sensation and proprioception

are usually not useful for eval‎uating radiculopathies.

Maneuvers — Specific maneuvers can be helpful in determining whether symptoms are

radicular in nature. These include the straight leg raise (Lasegue's sign), the contralateral

straight leg raise, and the reverse straight leg raise (also referred to as femoral stretch).

The straight leg raise test is done with the patient supine. The examiner raises the

patient's extended leg on the symptomatic side with the foot dorsiflexed, being

careful that the patient is not actively "helping" in lifting the leg. Straight leg raising

results in an increase in dural tension in the low lumbar and high sacral levels.

Lasegue's sign is the presence or worsening of radicular pain (not just low back

pain or hamstring pain) with the straight leg maneuver (ie, hip flexion with the

leg extended at the knee). In contrast, radicular pain is absent or unchanged

when the hip is flexed with the leg in flexion at the knee. A positive Lasegue's

sign usually occurs when hip flexion is between 30 and 60 degrees, though

positive tests can also occur at smaller and larger degrees of hip flexion.

The bowstring sign refers to the relief of radicular pain when the knee is flexed

during a positive straight leg raise.

The straight leg raise test is more sensitive but less specific than the contralateral

straight leg raise for the diagnosis of radiculopathy due to disc herniation [33]. It

is most helpful in the eval‎uation of radiculopathy at the L5 and S1 levels. In a

prospective study of 100 patients, Lasegue's sign was positive in 83 percent [34].

One study of patients with acute low back pain found that the combination of

radiating pain into the leg, sensory loss in the foot, and a positive straight leg

raise test was predictive of a herniated disc with nerve root compression [35].

The contralateral (crossed) straight leg raise test refers to passive elevation of the

unaffected leg by the examiner. The test is positive when lifting the unaffected leg

reproduces radicular pain in the affected leg. It is relatively specific for radiculopathy

due to disc herniation, but has poor sensitivity [33,35].

The reverse straight leg raise (femoral stretch) test is accomplished by placing the

patient prone on the table and passively extending the hip and leg straight up off the

plane of the table. This maneuver is most useful for eval‎uating the L2, L3, and L4

roots. However, the value of this test is limited by inadequate information on its

sensitivity and specificity.

Patrick's test is a maneuver where the hip is externally rotated with the ipsilateral

knee flexed at 90 degrees and placed on the opposite knee. The test is positive if it

elicits hip or buttock pain. A positive test raises suspicion for hip or sacroiliac disease.

However, it is nonspecific for a radicular process. (See "Eval‎uation of the adult with

hip pain", section on 'Patrick (Fabere) test'.)

Neuroimaging — For imaging of the lumbar spine, MRI, CT, and CT myelography (CT scan

after intrathecal administration of contrast media) are equally sensitive for the diagnosis of

disc herniation [30]. For routine initial assessment, an MRI is more informative than CT

because it can also identify other intraspinal pathologies, including inflammatory,

malignant, and vascular disorders. In addition, MRI is not associated with ionizing radiation

and is less invasive than CT myelography.

However, there is a high preval‎ence of abnormal neuroimaging findings in asymptomatic

individuals, including some who have what appears to be frank nerve root compression by

MRI [36]. As an example, one study of 98 people without back pain found MRI evidence of

disc herniation in 27 percent [37]. Furthermore, lumbar spine abnormalities on MRI in

asymptomatic patients do not appear to be predictive for the future development or

duration of low back pain [38].

Although rarely indicated, CT myelography can visualize spinal nerve roots and their

trajectory though the neural foramina. It is useful for patients with intolerance of or

contraindications to MRI (eg, implanted electrical devices such as cardiac pacemakers or

defibrillators) when standard CT fails to define the anatomic correlates of the clinical

presentation. In addition, CT myelography is preferred for patients who have surgically

placed spinal hardware that produces magnetic artifacts.

A CT scan can assess osseous structures better than either plain radiography or MRI and is

therefore helpful in assessing for bony disease. However, CT alone is unable to visualize

nerve roots, so it is not helpful in the direct imaging of a radicular process.

Plain radiographs of the lumbar spine are of limited value in the eval‎uation of lumbosacral

radiculopathy. They can visualize bony structures but do not detect herniated discs. Thus,

plain radiographs are not recommended in the work-up of lumbosacral radiculopathy unless

there is a need to eval‎uate for infection, fracture, malignancy, spondylolisthesis,

degenerative changes, disc space narrowing, or prior surgery. However, normal plain films

do not exclude malignancy or infection in patients with a suspicious history. (See

"Diagnostic testing for low back pain".)

Electrodiagnosis — The primary electrodiagnostic procedures for lumbosacral

radiculopathy are electromyography (EMG) and nerve conduction studies (NCS). In

combination, the information provided gives insights into the integrity of spinal nerve roots

and their connection with the muscles they innervate. These tests are most commonly

considered in patients with persistent disabling symptoms of radiculopathy in whom

neuroimaging findings are not consistent with the clinical presentation.

In a systematic review of evidence published through mid-2006, the American Association

of Neuromuscular and Electrodiagnostic Medicine (AANEM) assessed the utility of

electrodiagnostic testing for patients with lumbosacral radiculopathy [39]. The AANEM

noted that the available data are limited by the lack of a universally accepted “gold

standard” definition for the diagnosis of lumbosacral radiculopathy, thus preventing

comparison of sensitivities and specificities reported by the included studies. With this

caveat in mind, the AANEM concluded that the following tests probably aid the clinical

diagnosis:

Peripheral limb EMG

H reflex in S1 radiculopathy (see "Nerve conduction studies: Late responses", section

on 'H reflex')

The AANEM observed that the available evidence suggests a low sensitivity for peroneal

and posterior tibial F waves [39].

Nerve conduction studies and electromyography — NCS and EMG have a high

diagnostic utility for radiculopathy when neurologic weakness is present for at least three

weeks, which is why these studies are most often considered for those with persistent

unexplained symptoms. The yield is lower in patients with only pain or sensory loss as the

manifestation of radiculopathy. For patients with nonspecific spine pain, EMG can help to

distinguish pain-related reduced muscular effort from true neurogenic weakness.

In patients with weakness due to radiculopathy, NCS and EMG together can localize the

specific spinal nerve root that is damaged, distinguish between old and new axon loss

nerve damage, and provide indirect support for the presence of demyelinating conduction

block at the root level [8,40]. Finally, electrodiagnostic testing can identify conditions that

mimic radiculopathy, such as mononeuropathies of the leg or lumbosacral plexopathy.

From studies of the relative value of electrodiagnostic testing in the diagnosis of

lumbosacral radiculopathy, the following observations have been made:

EMG and imaging studies have a comparable diagnostic sensitivity, varying between

50 to 85 percent, depending on the patient population [41-43]. In a retrospective

comparison of 47 patients who had a clinical history suggestive of either cervical or

lumbosacral radiculopathy, there was congruence between EMG and MRI findings in

60 percent of patients [41]. Agreement between EMG and MRI was highest for

patients with clearly abnormal examination findings consistent with radiculopathy.

MRI and EMG provide unique anatomic and physiologic information, respectively

[41,44]

Thus, these tests have overlapping diagnostic sensitivity but provide different kinds of

information. When precision is needed to decide whether the neuroimaging findings are

etiologically related to specific neurologic deficits, EMG is useful to support or refute this

relationship. In particular, EMG may provide objective evidence of denervation when there

is no motor deficit or uncertain motor deficit. In addition, EMG findings may help determine

the timing of the denervation (eg, distant past versus ongoing). This is particularly useful in

patients with past surgery and residual pain.

In acute radiculopathy (ie, the first three weeks), EMG and NCS provide limited but

potentially important information. The needle EMG examination is more sensitive and

provides better localizing information than NCS, but is not likely to show prominent

features of acute axon loss until three or more weeks after symptom onset. This is due to a

two to three week delay in the development of fibrillation potentials after acute motor axon

loss. In the face of marked weakness from acute radiculopathy, NCS can show loss of

amplitude of compound muscle action potentials by day eight after injury.

NCS are carried out by applying an electrical stimulus to the skin overlying a nerve trunk,

followed by the recording of the generated electrical response over either the nerve trunk

or muscle innervated by it, at some distance from the stimulation. EMG records the

electrical potentials generated in a muscle belly through a needle electrode inserted in the

muscle. Both methods are discussed in detail elsewhere. (See "Overview of

electromyography" and "Overview of nerve conduction studies".)

Somatosensory evoked potentials — Somatosensory evoked potentials (SEPs) are

performed by repetitively stimulating the tibial nerve at the ankle and recording the

propagated sensory potentials as they ascend up the leg, into the cauda equina, through

central sensory pathways in the spinal cord and brain, and then to the sensory cortex.

However, the available evidence is conflicting regarding the ability of SEPs to localize

specific nerve root compression [45-47]. As noted above, a systematic review found

insufficient evidence to reach a conclusion about the utility of dermatomal/segmental SEPs

of the L5 or S1 dermatomes [39]. Thus, the clinical value of SEPs for the diagnosis of

lumbosacral radiculopathy is uncertain and they are not part of the routine eval‎uation of

radiculopathy.

Cerebrospinal fluid analysis — A lumbar puncture for cerebrospinal fluid (CSF)

eval‎uation is indicated if there is suspicion for a neoplastic or infectious cause for

radiculopathy. In the presence of a known primary tumor, eval‎uation of the CSF is usually

indicated only if contrast-enhanced MRI is nondiagnostic. However, CSF sampling should be

done in cases without a known primary cancer when neuroimaging is nondiagnostic,

particularly in patients who fail to improve, have progressive neurologic deficits, and/or

have involvement of multiple roots.

CSF cytology can be diagnostic in patients with leptomeningeal carcinomatosis, but it may

be negative in 10 percent of cases [22]. In such patients, CSF testing should be repeated if

there is clinical suspicion for cancer. Other abnormalities in the CSF include a monocytic

pleocytosis, decreased glucose, and increased protein.

Radiculopathy may result from infectious causes, including Lyme disease, cytomegalovirus,

and herpes zoster. CSF sampling will provide the diagnosis in the first two cases, and may

be needed in the third when there is no characteristic rash.

Discography — Discography is a controversial technique of uncertain utility that involves

the injection of contrast under fluoroscopy into the nucleus of a disc thought to be the

cause of a patient's low back pain. The test is considered positive if it demonstrates an

annular disruption and reproduces the patient's usual low back pain symptoms. It is not

helpful in the eval‎uation of lumbosacral radiculopathy. (See "Diagnostic testing for low back

pain", section on 'Discography'.)

DIFFERENTIAL DIAGNOSIS — The most common cause of lumbosacral radiculopathy is

a herniated disc, though the differential diagnosis includes nonradicular back pain, lumbar

spinal stenosis, cauda equina syndrome, diabetic amyotrophy, lumbosacral plexopathy, and

mononeuropathies of the leg, such as femoral, sciatic, peroneal, and tibial nerve lesions.

Low back pain — Low back pain is one of the most common symptoms causing patients

to seek medical attention. Most patients with lumbosacral radiculopathy will also have

symptoms of low back pain. However, the majority of cases of low back pain are not due to

a primary neurologic cause. Disc disease proves to be the underlying etiology in <5 percent

of patients [48]. Furthermore, a specific etiology cannot be reliably established for most

patients with low back pain. Thus, while lumbosacral radiculopathy is one possible

underlying etiology, it is not the most common.

Since the treatment of radicular and nonradicular low back pain may differ, it is important

to distinguish the two.

Radicular low back pain is generally associated with pain that radiates down one or

both legs. There may be associated nerve root dysfunction resulting in sensory loss

and/or weakness that may or may not be overtly symptomatic. If the appropriate

roots are involved, there may also be a loss or diminution of reflexes.

Nonradicular low back pain is not associated with other neurologic symptoms or

signs. Symptoms are generally localized to the spine and paraspinal regions and may

or may not radiate into the leg. However, some patients have pain that radiates from

the lower back into the buttocks and upper leg in a manner that mimics radiculopathy

even though it is not due to nerve root impingement. Pain may be caused by injury of

non-neurologic structures such as muscles, tendons, ligaments, or fascia without

affecting the roots.

Lumbosacral radiculopathies are often misdiagnosed in cases of nonradicular low back pain.

This confusion is reinforced by the use of nonspecific nomenclature such as the term

"sciatica." Sciatica is often applied to conditions involving low back pain with or without

radiation into the leg regardless the cause. Because of its nonspecific use, the term sciatica

should be avoided if possible when referring to radicular pain.

Diagnostic testing for low back pain is often equivocal. In such cases the diagnosis is

typically that of nonspecific low back pain. Lumbosacral MRI may reveal findings suggestive

of root compression, though the patient may not have specific radicular signs or symptoms.

Persistence of findings from remote root injury and intercurrent processes such as diabetic

neuropathy may also be noted on EMG, confounding the eval‎uation for a more acute

problem.

Spinal stenosis — Lumbar spinal stenosis is a common entity, often asymptomatic, that

can be caused by a variety of congenital or acquired conditions. The most common etiology

is degenerative spondylosis. The major manifestation is neurogenic claudication, a

syndrome of bilateral, often asymmetric pain, sensory loss, and/or weakness affecting the

legs. The symptoms are produced or exacerbated by walking or prolonged standing in an

erect posture. These symptoms represent intermittent mechanical and/or ischemic

disruption of lumbosacral nerve root function. (See "Lumbar spinal stenosis:

Pathophysiology, clinical features, and diagnosis".)

In a few patients with lumbar spinal stenosis, more fixed nerve root injury may occur,

causing lumbosacral radiculopathy, and rarely cauda equina syndrome or conus medullaris

syndrome.

Cauda equina syndrome — The cauda equina syndrome is caused by an intraspinal

lesion caudal to the conus that injures two or more of the 18 nerve roots constituting the

cauda equina within the lumbar spinal canal. (See "Anatomy and localization of spinal cord

disorders", section on 'Cauda equina syndrome'.)

The cauda equina syndrome is typically associated with marked neurologic disability. The

clinical presentation is dominated by bilateral leg weakness in multiple root distributions

(L3-S1), and may be associated with perineal sensory symptoms as well as bowel, bladder,

and sexual dysfunction due to involvement of the S2-4 spinal nerve roots.

Potential etiologies include developmental abnormalities such as neural tube defects,

infectious or inflammatory conditions, or mass lesions such as tumors. (See "Clinical

features and diagnosis of neoplastic epidural spinal cord compression, including cauda

equina syndrome".)

The cauda equina syndrome is a rare complication of lumbar spinal stenosis. (See "Lumbar

spinal stenosis: Pathophysiology, clinical features, and diagnosis".)

Diabetic amyotrophy — Diabetic amyotrophy involves the L2, L3, and L4 roots and

presents with weakness, pain, and dysesthesia primarily in the proximal leg that evolves

over days to weeks. This condition is discussed separately. (See "Epidemiology and

classification of diabetic neuropathy", section on 'Diabetic amyotrophy (lumbar

polyradiculopathy)'.)

SUMMARY AND RECOMMENDATIONS

Five movable lumbar vertebral bodies (L1 to L5) comprise the lumbar spine, and five

developmentally fused vertebral levels (S1 to S5) make up the sacrum (figure 1). The

entire region is commonly described as the lumbosacral spine. Directly beneath each

lumbar and sacral vertebrae there is a pair of neural foramina with the same number

designation, such that the L1 neural foramina are located just below the L1 vertebral

body. Through each neural foramen passes the same numbered spinal nerve root. On

each side there are five lumbar, five sacral, and one coccygeal spinal nerve roots.

(See 'Anatomy' above.)

All lumbar and sacral spinal nerve roots originate at the T10 to L1 vertebral level,

where the spinal cord ends as the conus medullaris. A dorsal (somatic sensory) root

from the posterolateral aspect of the spinal cord and a ventral (somatic motor) root

from the anterolateral aspect of the cord join in the spinal canal to form the spinal

nerve root (figure 1). The roots then course down through the intraspinal canal,

forming the cauda equina, until they exit at their respective neural (intervertebral)

foramina. (See 'Anatomy' above.)

The most common etiology of lumbosacral radiculopathy is nerve root compression

caused by a disc herniation or spinal stenosis due to degenerative arthritis

(spondylosis) affecting the spine. Congenital abnormalities of the bony spinal column

or its contents include tethered cord or diastematomyelia and spina bifida. They may

cause radiculopathy via traction of the root. Additional etiologies of lumbosacral

radiculopathy include nonskeletal conditions such as infection, inflammation,

neoplasm and vascular disease. (See 'Pathophysiology and etiology' above.)

Lumbar radiculopathy most commonly involves either the L5 or S1 root.

L1 radiculopathy is rare. Symptoms involve pain, paresthesia, and sensory loss in

the inguinal region. (See 'L1 radiculopathy' above.)

L2, L3, and L4 radiculopathies are most often seen in older patients with spinal

stenosis. They are generally considered as a group because of marked overlap of

innervation of the anterior thigh muscles. Acute back pain is the most common

presenting complaint, often radiating around the anterior aspect of the leg down

into the knee. (See 'L2/L3/L4 radiculopathy' above.)

L5 radiculopathy is the most common radiculopathy affecting the lumbosacral

spine. It often presents with back pain that radiates down the lateral aspect of

the leg into the foot. On examination, strength can be reduced in foot

dorsiflexion, toe extension, foot inversion, and foot eversion. Reflexes are

generally normal. (See 'L5 radiculopathy' above.)

In S1 radiculopathy, pain radiates down the posterior aspect of the leg into the

foot from the back. On examination, strength may be reduced in leg extension

(gluteus maximus), plantar flexion, and toe flexion. Sensation is generally

reduced on the posterior aspect of the leg and the lateral foot. Ankle reflex loss is

typical. (See 'S1 radiculopathy' above.)

S2, S3, and/or S4 radiculopathies due to structural etiologies are rare, but may

be caused by a large central disc that compresses the nerve roots intrathecally at

a higher level (eg, L5). Patients can present with sacral or buttock pain that

radiates down the posterior aspect of the leg or into the perineum. Weakness may

be minimal, with urinary and fecal incontinence as well as sexual dysfunction.

(See 'S2/S3/S4 radiculopathy' above.)

The diagnosis of a lumbosacral radiculopathy is clinical, and can usually be made

based upon compatible symptoms and examination findings. Eval‎uation requires a

careful neurologic examination. Immediate diagnostic testing is not necessary for

patients with suspected radiculopathy who are neurologically intact and not

suspected of having underlying neoplasm, infection, or inflammation. Nevertheless,

testing is suggested to confirm‎ the diagnosis and etiology for patients who have

persistent symptoms that are not adequately controlled with conservative therapy

and for whom invasive treatment options are an option. (See 'Diagnosis' above.)

We recommend neuroimaging as part of the initial eval‎uation for patients with any of

the following conditions (see 'Neuroimaging' above):

Acute radiculopathy with rapidly progressive neurologic deficits

Radiculopathy with urinary retention, saddle anesthesia, or bilateral neurologic

symptoms or signs

High suspicion for neoplasm or epidural abscess

For patients with persistent or severe findings in whom the etiology is not confirmed

on neuroimaging, we suggest electromyography and nerve conduction studies (See

'Electrodiagnosis' above.).

When there is suspicion for an inflammatory or infectious cause of lumbosacral

radiculopathy, we recommend a lumbar puncture for cerebrospinal fluid analysis.

(See 'Cerebrospinal fluid analysis' above.)

In addition to a herniated disc, the differential diagnosis of lumbosacral radiculopathy

includes lumbar spinal stenosis, cauda equina syndrome, diabetic amyotrophy,

lumbosacral plexopathy, and mononeuropathies of the leg, such as femoral, sciatic,

peroneal, and tibial nerve lesions. (See 'Differential diagnosis' above.)

[문형철]

(See "Lumbosacral radiculopathy: Prognosis and treatment" and "Lumbar spinal stenosis: Pathophysiology, clinical features, and diagnosis" and "Approach to the diagnosis and eval!uation of low back pain in adults" and "Subacute and chronic low back pain: Pharmacologic and noninterventional treatment" and "Subacute and chronic low back pain: Nonsurgical interventional treatment" and "Subacute and chronic low back pain: Surgical treatment".) 공재철