Treatment and Management of Segmental Neuromuscular Disorders

Clinical Signs and Symptoms

Two classic articles detail the history and examination findings in CR ( ; ). evaluated 100 patients with surgically confirmed single-level cervical radiculopathies ( Table 17.1 ). reviewed 648 cases of surgically treated single-level cervical radiculopathies. Findings in terms of pain radiation and neurologic deficits were similar to those of Yoss et al. The Murphey series did emphasize the occurrence of pain in the pectoral region in 20% of their cases; they opined that neck, periscapular, and pectoral region pain was referred from the disk itself and that arm pain was the result of nerve root compression.

In the Radhakrishnan population-based series, cervicobrachial pain was present at the onset in 98% and was radicular in 65% ( ). Paresthesias were reported by 90%, almost identical to the Yoss series. Pain on neck movement was present in 98%; paraspinal muscle spasm, in 88%; decreased reflexes, in 84% (triceps 50%, biceps or brachioradialis 34%); weakness, in 65%; and sensory loss, in 33% of participants. Employing their criteria, 45% of the patients were judged to have definite CR, 30% probable, and 25% possible. In a Cleveland Clinic series, 70% had motor and sensory symptoms, 12% had motor symptoms only, and 18% had sensory symptoms only ( ).

Localizing information in suspected CR is obtainable from the history, especially from patterns of pain radiation and paresthesias. Radiating pain on coughing, sneezing, or straining during a bowel movement is significant but seldom elicited. Increased pain on shoulder motion suggests nonradicular disease. Relief of pain by resting the hand atop the head is reportedly characteristic of CR (hand on head, Bakody, or shoulder abduction relief sign), but this phenomenon can occur with a Pancoast tumor.

Physical examination in patients with suspected CR should include an assessment of the range of motion (ROM) of the neck and arm, a search for root compression signs, detailed examination of strength and reflexes, a screening sensory examination, and probing for areas of muscle spasm or trigger points ( ). The cervical spine ROM is highly informative. Patients should be asked to put their chin to chest and to either shoulder, each ear to shoulder, and to hold the head in full extension; these maneuvers all affect the size of the intervertebral foramen. Pain produced by movements that close the foramen suggest CR. Pain on the symptomatic side on putting the ipsilateral ear to the shoulder suggests radiculopathy, but increased pain on leaning or turning away from the symptomatic side suggests a myofascial origin. Radiating pain with the head in extension and tilted slightly to the symptomatic side is highly suggestive of CR (Spurling maneuver, foraminal compression test); brief breath holding in this position will sometimes elicit the pain. The addition of axial compression does not seem to add much. Light digital compression of the jugular veins until the face is flushed and the patient is uncomfortable will sometimes elicit radicular symptoms: unilateral shoulder, arm, pectoral, or scapular pain or radiating paresthesias into the arm or hand (Viets or Naffziger sign). This is a highly specific but insensitive finding. An occasional patient has relief of pain with manual upward neck traction. Patients with a globally restricted cervical spine ROM often have extensive degenerative disease. Patients with large disk ruptures or cervical spinal stenosis may have a positive Lhermitte sign on neck flexion.

See Clinical Signs in Neurology: A Compendium for an extensive discussion and video of physical examination signs in CR ( ).

Pain or limitation of motion of any upper extremity joint should signal the possibility of nonradicular disease. The patient should be asked to put the shoulder through a full active ROM, touching the hand to the opposite shoulder and the opposite ear, then reaching behind as high between the scapulae as possible. Any pain or limitation of motion on the symptomatic side suggests bursitis, capsulitis, tendonitis, or impingement syndrome rather than CR as the cause of the patient’s pain.

A focused but detailed strength examination should at least assess the power in the deltoids, spinatus, biceps, triceps, pronators, wrist extensors, abductor pollicis brevis, and interossei. The sensory examination should concentrate on the hand, and particularly assess touch, since the large, myelinated fibers conveying light touch are more vulnerable to pressure injury than the smaller fibers carrying pain and temperature. Because of the extensive overlap of dermatomes, sensory loss in CR is seldom conspicuous. The reflex examination should include not only the standard upper extremity reflexes but the knee and ankle jerks and plantar reflexes as well. Increased lower extremity reflexes and extensor plantar responses suggest myelopathy complicating the radiculopathy.

Based on the foregoing, Box 17.1 outlines the clinical data that favor the diagnosis of CR.

Evaluation and Diagnosis

Several different clinical syndromes may ensue from degenerative spine disease, including simple, single-level radiculopathy; multilevel radiculopathy; cervical myelopathy; cervical radiculomyelopathy; and occasionally a central cord or Brown-Séquard syndrome. Rarely, radiculopathy results from other processes, such as tumor (e.g., neurofibroma, meningioma, metastasis), infection (e.g., Lyme disease, zoster, cytomegalovirus, syphilis, epidural abscess), infiltration (e.g., meningeal neoplasia, sarcoidosis), or ischemia ( ). Diabetes commonly causes a lumbosacral radiculoplexopathy syndrome but can also frequently produce a painful thoracic radiculopathy often confused with herpes zoster and rarely a cervicobrachial radiculopathy ( ). With acute trauma, roots are sometimes injured along with the cervical spine or the BP and in fact may be completely avulsed from the spinal cord in a severe BP injury. Acute and chronic inflammatory radiculoneuropathies commonly involve the roots. Other rare causes include arachnoiditis, root sleeve cyst, epidural lipomatosis, and spinal arteriovenous malformation.

A number of clinical conditions may be confused with CR, primarily brachial plexopathies, entrapment neuropathies, and non-neuropathic mimickers. The more common musculoskeletal conditions causing confusion include shoulder pathology (bursitis, tendinitis, impingement syndrome), lateral epicondylitis, and DeQuervain tenosynovitis. Cervical myofascial pain, facet joint disease, and cervical vertebral body disease can cause neck pain with referred pain to the arm. Patients with cervical strain due to whiplash rarely have radiculopathy. Lyme disease can cause meningitis and radiculitis, producing neck and arm pain. A rare patient with subarachnoid hemorrhage may present with neck pain without headache. Cervical epidural abscess or hematoma may present with neck pain and various neurologic signs. Pain can be referred to the neck, arm, or shoulder from the heart, lungs, esophagus, or upper abdomen.

Clues to the potential presence of an unusual cause of radiculopathy include a history of systemic symptoms such as fever, chills or weight loss, or a history of cancer, immunosuppression or intravenous drug use.

The Quality Assurance committee of the American Association of Electrodiagnostic Medicine has recently reviewed the utility of EMG in the evaluation of patients with CR and formulated a practice parameter addressing this problem ( ). The sensitivity of EMG for detecting abnormalities was in the range of 60% to 70%. The sensitivity is significantly higher in patients with motor involvement than in those with only pain or sensory abnormalities, and the yield increases with increasing severity of disease. Because EMG findings are rarely abnormal in normal subjects, the test is highly specific. There is generally a 65% to 85% correlation of EMG abnormalities with imaging and surgical findings. There is unanimity that C7 root lesions are the most common (±60%), C6 as the next most common (±20%), with C5 and C8 lesions making up about equal proportions of the remainder ( ; ; ; ; ).

The examiner should tailor EMG to probe the muscles most likely to be involved. In the absence of clear clinical direction, one must resort to a generic root screen. Choosing which muscles to study, and how many, has heretofore been a matter of personal opinion and preference. Recent investigations have provided some guidance. concluded that a screen of six limb muscles plus the paraspinals had a yield of 93% to 98%, and that examining more muscles made no significant additional contribution. studied 50 patients with surgically proven single-level cervical radiculopathies. They found stereotyped patterns with C5, C7, and C8 lesions, but a variable pattern with C6 lesions, which could resemble C5 or C7. Patients with a clinical and electrodiagnostic picture of C8 radiculopathy are particularly challenging, as more rostral pathology can mimic C8 radiculopathy and many do not actually have lesions involving the C8 root. In one study of 31 patients, only 16% had C8 root compression ( ).

Plain cervical spine radiographs may show degenerative changes but are of little use in the diagnosis of a specific radiculopathy. Many asymptomatic patients have degenerative changes. Plain films, especially with flexion and extension views, may be useful in patients with more advanced disease to rule out any instability. Computed tomography (CT) is much more useful for visualizing bony detail but does not visualize neural structures unless contrast agent has been instilled (CT myelography). Plain myelography is rarely used. Magnetic resonance imaging (MRI) is the mainstay of diagnosis, although disk herniations are commonly observed on MRI in patients who have no symptoms, particularly in older patients. Several MRI studies have documented that asymptomatic DDD is common. In a study of 1211 healthy volunteers, disk bulging on cervical spine MRI was frequently observed in asymptomatic subjects, including patients in their 20s ( ). Even spinal cord compression and intraparenchymal signal changes were seen in older subjects. Careful clinical and electrodiagnostic correlation therefore remains essential ( ).

Treatment and Management

Treatment relies on three approaches: mechanical, medicinal, and surgical ( Box 17.2 ). Nerve roots lying in the foramen normally enjoy freedom of movement through a small range. The size of the intervertebral foramen and the lateral recess changes dynamically with neck movement. The mainstay of conservative treatment is to reduce neck movement and increase the size of the foramen.

A soft cervical collar is usually helpful. For compressive CR, the collar should be worn “backward,” with the high side posterior to maintain the neck in slight flexion and open the foramina. Hard collars cannot be turned around in this fashion and are not as useful for a radiculopathy syndrome as for myelopathy. The soft collar should be worn at night if tolerated; otherwise, the patient should use a cervical pillow. Prolonged use of a cervical collar may weaken neck muscles, so this method of treatment should be limited to a few weeks at most.

Cervical traction for 15 to 30 minutes three times a day is often very helpful; it distracts the spine, opens the foramina, and gives the involved root respite. An over-the-door home traction unit is adequate; referral to a physical therapist is unnecessary. The patient should start with a low weight (5–8 lb) and advance as tolerated to 12 to 15 lb. Too rapid a weight increase may cause neck or jaw soreness and limit compliance. It is theoretically best for the patient to face the door with the neck slightly flexed; the combination of flexion and distraction is more effective in opening the foramen. However, it is difficult to do anything but stare at the door during such treatment; if facing away from the door appears equally efficacious for the individual patient, it is permissible and may improve compliance. Neuroimaging should be done before traction is employed.

Modality physical therapy, or local heat or ice, may provide some relief of the axial pain component, but the effects seldom persist much beyond the individual treatment session. Cervical ROM exercises are of no proven benefit. Nonsteroidal anti-inflammatory drugs (NSAIDs) may decrease the radicular inflammatory component and relieve pain. Other analgesics are often necessary in addition, including occasional narcotics. When the neurologic deficit is moderate (Medical Research Council [MRC] grade 4+/5 in the most involved muscles), a course of oral steroids is reasonable (e.g., prednisone 60–100 mg/day for 7–10 days, tapering over the next 7–10 days) ( ; ). Gabapentinoids are sometimes used in both cervical radiculopathy and LSR, but there is very little evidence to support their effectiveness. Intensive conservative therapy is usually continued for 3 to 6 weeks, and if the patient is no better after that interval, surgical referral should be considered. Muscle relaxants add little and cause side effects of sedation and depression.

Epidural steroid injections have been used for the treatment of radiculopathy, first lumbosacral and later cervical, since the 1950s. Despite this long history, there is a paucity of quality clinical evidence supporting their efficacy. Cervical injections are done via either an interlaminar or transforaminal route. One prospective, randomized trail using transforaminal injections found no difference in outcome between treatment with steroids/local anesthetic or saline/local anesthetic at 3 weeks ( ). A 2015 systematic review of the literature between 1966 and 2014 found only a single high-quality study regarding epidural steroids in cervical disk herniations. A total of 120 subjects received interlaminar injections of either local anesthetic (group I) or local anesthetic and steroid (group II). Significant improvement was seen in 77% in Group I and 80% in Group II ( ).

In a 2007 review of the risks of cervical transforaminal epidural steroids, the authors pointed out the lack of randomized controlled studies supporting their efficacy; that is still the case. They surveyed physician members of the American Pain Society and collected reports of 78 complications, including 16 vertebrobasilar brain infarcts, 12 cervical spinal cord infarcts, and 2 combined brain/spinal cord infarcts, including 13 fatalities. Evidence suggests an embolic mechanism due to inadvertent intra-arterial injection of particulate corticosteroid as the primary mechanism. This risk could potentially be minimized by using the nonparticulate glucocorticoid dexamethasone. Other potential mechanisms include vertebral artery perforation causing dissection/thrombosis and needle-induced vasospasm. The authors conclude that there is a significant risk of serious neurologic injury after cervical transforaminal steroid injections ( ).

Consider surgical referral in patients with an MRC strength of grade 4/5 or worse in any muscle or evidence of myelopathy or excruciating pain unresponsive to conservative treatment. In a population-based study, 26% required surgery. As with epidural steroids, there is scant support in the literature for the efficacy of surgery for CR. Prospective, randomized trials have either not shown a difference in outcome between anterior cervical discectomy and fusion (ACDF) and physical therapy or showed a slight difference favoring surgery early, with no difference at 12–24 months ( ; ; ). A prospective, randomized study with 5- to 8-year follow-up showed that ACDF combined with physiotherapy reduced neck disability and neck pain more effectively than physiotherapy alone ( ).

Some surgeons now prefer anterior cervical disk replacement/arthroplasty (ACDA; artificial cervical disk replacement) over ACDF. Theoretically, ACDA more closely simulates physiologic motion, preserves segmental mobility, and reduces adjacent segment degeneration ( ).

One problem investigators face in determining the effectiveness of any treatment for CR, from cervical traction to surgery, is that the natural history of the condition is toward spontaneous resolution. The typical CR patient is significantly improved by 2 to 3 months. There is a generally favorable long-term prognosis; 90% have minimal to no symptoms on prolonged follow-up. A systematic review and evidence-based clinical guideline on CR from the North American Spine society, in regard to the natural history, offered the following consensus statement: “It is likely that for most patients with cervical radiculopathy from degenerative disorders signs and symptoms will be self-limited and will resolve spontaneously over a variable length of time without specific treatment” ( ).

Clinical Signs and Symptoms

Involvement of some of these pain-sensitive structures can produce referred pain that radiates to the extremity (buttock, hip, thigh) and can simulate the radiating pain of nerve root origin ( Table 17.2 ). A study of 1293 cases of LBP concluded that referred pain to the lower limb most often originated from sacroiliac and facet joints. Referred pain to the extremity occurred nearly twice as often as true radicular pain and frequently mimicked the clinical presentation of radiculopathies ( ). Investigations have demonstrated that considerable pain can be referred to the buttock and thigh, with disease limited to the disk, the facet joint, or the sacroiliac joint ( ; ; ). A study of 92 patients with chronic LBP concluded that 39% had annular tears or other forms of internal disk disruption as the cause of their pain. No available clinical test differentiated between patients with internal disk disruptions and those with compressive radiculopathy ( ). A similar situation seems to exist with facet joint pain ( ; ). The first suggestion that a patient may have nerve root compression usually comes because of radiating pain into one or both lower extremities. Conservative therapy still usually suffices, even in patients with HNPs, and only 5% to 10% of patients ultimately need surgery ( ). Operative intervention is generally appropriate only when there is a combination of definite disk herniation shown by imaging studies, a corresponding syndrome of radicular pain, a corresponding neurologic deficit, and a failure to respond to conservative therapy.

reviewed the information that could be obtained from the history and physical examination in patients with LBP. They suggest trying to answer three basic questions: Is there a serious, underlying systemic disease present? Is there neurologic compromise that might require further evaluation? Are there psychological factors leading to pain amplification? Factors that suggest the possibility of underlying systemic disease include a history of cancer, unexplained weight loss, pain lasting longer than 1 month, pain unrelieved by bed rest, fever, focal spine tenderness, morning stiffness, improvement in pain with exercise, and failure of conservative treatment.

The utility, or lack thereof, of various physical examination findings has been studied. The straight leg raising (SLR) test remains the mainstay in detecting radicular compression. The test is performed by slowly raising the symptomatic leg with the knee extended. Tension is transmitted to the nerve roots between about 30 and 70 degrees and pain increases. Pain at less than 30 degrees raises the question of nonorganicity, and some discomfort and tightness beyond 70 degrees are routine and insignificant. There are various degrees or levels of positivity. Ipsilateral leg tightness is the lowest level, pain in the back is more significant, and radiating pain in the leg is highly significant. When raising the good leg produces pain in the symptomatic leg (crossed SLR sign), the likelihood of a root lesion is very high.

The positivity of the SLR should be the same with the patient supine or seated. If a patient with a positive supine SLR does not complain or lean backward when the extended leg is brought up while in the seated position (e.g., under the guise of doing the plantar response), it is suggestive of nonorganicity. The SLR can be enhanced by passively dorsiflexing the patient’s foot just at the elevation angle at which the increased root tension begins to produce pain. A quick snap to the sciatic nerve in the popliteal fossa just as stretch begins to cause pain (bowstring sign, or popliteal compression test) accomplishes the same end. Patients with hip disease have pain on raising the leg whether the knee is bent or straight; those with root stretch signs have pain only when the knee is extended. Pain from hip disease is maximal when the hip is flexed, abducted, and externally rotated by putting the patient’s foot on the contralateral knee and pressing down slightly on the symptomatic knee ( fabere test). There are other procedures for checking the sacroiliac joints.

See Clinical Signs in Neurology: A Compendium for an extensive discussion and video of physical examination signs in LSR ( ).

The neurologic examination should include assessment of power in the major lower extremity muscle groups, but especially the dorsiflexors of the foot and toes, and the evertors and invertors of the foot. Plantar flexion of the foot is so powerful that manual testing rarely suffices. Having the patient do 10 toe raises with either foot is a better test. Sensation should be tested in the signature zones of the major roots. The status of knee and ankle reflexes is informative about the integrity of the L3–4 and S1 roots. There is no good reflex for the L5 root, but the hamstring reflexes are occasionally useful. The medial and lateral hamstrings are both innervated by both L5 and S1, but the medial hamstring tends to be more L5 and the lateral more S1. An occasional L5 radiculopathy produces a clear selective diminution of the medial hamstring reflex. Rarely, an L5 radiculopathy will cause a deceased peroneal reflex (see http://neurosigns.org/wiki/Peroneal_reflex ). In a patient with absent ankle jerks and a question of neuropathy, loss of the lateral hamstring reflex with preservation of the medial helps confirm radicular disease. Preservation of the lateral hamstring jerk with an absent ankle jerk suggests a length-dependent process (i.e., peripheral neuropathy).

Studies have found that only 8 of 27 physical tests investigated successfully discriminated between patients with chronic LBP and normal control subjects. The eight useful tests were pelvic flexion, total spinal flexion, total extension, lateral flexion, SLR, spinal tenderness, bilateral active SLR, and sit-up ( ). Tests for nonorganicity (Waddell signs) are very useful. Pain during simulated spinal rotation, pinning the patient’s hands to the sides while passively rotating the hips back and forth (no spine rotation occurs as shoulders and hips remain in a constant relationship), suggests nonorganicity. This is also suggestive with a discrepancy between the positivity of the SLR between the supine and seated position, pain in the back on pressing down on top of the head, widespread and excessive “tenderness,” general overreaction during testing, and nondermatomal/nonmyotomal neurologic signs. The presence of three of these signs suggests, if not nonorganicity, at least embellishment ( ).

The major radicular syndromes include HNP, lateral recess stenosis, and spinal stenosis with cauda equina compression. Virtually all patients with radiculopathy have sciatica. The odds of a patient without sciatica having radiculopathy have been estimated at 1:1000 ( ). The details of the sciatica are noteworthy, including the exact pattern of radiation, influence of body position and movement, and presence or absence of neurologic symptoms.

With HNP or lateral recess stenosis, leg pain usually predominates over back pain. With HNP, the pain is typically worse when sitting, better when standing, better still when lying down, and generally worse in flexed than extended postures—all reflecting the known changes in intradiscal pressure that occur in these positions. With lateral recess stenosis, the pain is worse with standing or walking and relieved by sitting with the torso flexed or by lying down. Patients with HNP tend to have a positive SLR, but those with recess stenosis do not. The essence of the recess stenosis picture, then, is pain on standing, lack of pain on sitting, and a negative SLR. The essence of the HNP picture is pain worse on sitting, lessened with standing, and a positive SLR. Patients with HNP are usually in the 30 to 55 years age range; those with lateral recess stenosis are a bit older. As with CR, pain may exacerbate with cough, sneeze, or Valsalva maneuver.

Evaluation and Diagnosis

As patients mature into the seventh decade and beyond, the liability to disk rupture decreases, but degenerative spine disease attacks in a different form. Osteophytic spurs and bars; bulging disks; thickened laminae and pedicles; arthritic, hypertrophied facets; and thickened spinal ligaments all combine to narrow the spinal canal and produce the syndrome of spinal stenosis. An extension posture contributes to spinal stenosis by causing narrowing of the foramina and dorsal quadrants and buckling of the ligamentum flavum. Narrowing of the canal compresses neural and possibly vascular structures. Flexing the spine, as by leaning forward, stooping over, or sitting down, opens the canal and decreases the symptoms.

Patients with spinal stenosis and neurogenic claudication experience pain, weakness, numbness, and paresthesias/dysesthesias when standing or walking. Symptoms decrease with sitting or bending forward. An occasional patient will have a bizarre symptom, such as spontaneous erections or fecal incontinence brought on by walking. Differentiation from vascular claudication is made by the wide distribution of symptoms, the neurologic accompaniments, and the necessity to sit down for relief. Vascular claudication tends to produce focal, intense, crampy pain in one or both calves, and the pain subsides if the patient just stops and stands. Patients with vascular claudication have even more symptoms walking uphill, because of the increased leg work. Neurogenic claudication may decrease when walking uphill because of the increased spinal flexion in forward leaning. Patients with vascular claudication have as much trouble riding a bicycle as walking because of the leg work involved, whereas forward flexion on the bicycle opens up the spinal canal, allowing patients with neurogenic claudication to ride a bike with greater ease than they can walk.

An investigation of 68 patients with lumbar spinal stenosis found that pseudoclaudication, or neurogenic claudication, was the most common symptom, producing pain, numbness, or weakness on walking, frequently bilaterally and usually relieved by flexing the spine. Mild neurologic abnormalities occurred in a minority of patients. EMG showed one or more involved roots in 90% of patients ( ).

The vast majority of LSRs are due to degenerative spine disease. Other disease processes can occur and may require exclusion. As with CR, patients may develop back pain and radiculopathy with epidural abscess or hematoma, diffuse meningeal neoplasia, diabetes, and other conditions ( ).

Plain lumbosacral spine films are rarely helpful except in the setting of acute trauma or suspected infection or malignancy. Plain CT may be useful in patients who are unable to undergo MRI; when combined with intrathecal contrast, CT is highly accurate for detecting nerve root compression (CT myelography). Plain myelography is sometimes useful, particularly in patients unable to undergo MRI because of metallic fixation and in those too obese to fit in an MRI scanner. MRI is the imaging procedure of choice in LSR, again with the caveat that many asymptomatic patients have disk disease on MRI that is not clinically relevant ( ).

Treatment and Management

Investigators face the same problems in determining the effectiveness of treatment for LSR as for CR, since symptoms of LSR due to disk herniation are typically self-limited. In patients with underlying spondylosis due to degenerative arthritis, radicular symptomatology may wax and wane, making the long-term prognosis less favorable than with disk herniation.

Studies have clearly shown that bed rest is not efficacious in patients with nonspecific or mechanical LBP ( ). There is less information about the role of bed rest in acute LSR, and brief, judicious bed rest may still be useful for short-term symptom control in patients who are in acute, severe pain. At the very least, if the patient remains up and active, he or she should avoid activities such as bending or lifting, which are known to increase intradiscal pressure. Patients often find particular positions that are comfortable and particular activities to avoid.

Any benefit from agents such as benzodiazepines and cyclobenzaprine are probably due more to the drowsiness induced, helping to keep the patient inactive, than to any muscle relaxant effects. NSAIDs may be useful both for pain control and to help reduce acute radicular inflammation, although trials have not substantiated their usefulness compared to placebo ( ). Opioids may be needed for pain relief in the acute situation. A short course of oral steroids is sometimes useful. Supportive evidence is marginal ( ; ). Drugs used to treat neuropathic pain, such as tricyclic antidepressants, gabapentin, and similar agents, may be useful for the management of chronic radiculopathy, but there is little evidence that they are helpful in the acute situation.

Many studies have been done regarding epidural steroids for LSR and spinal stenosis, leading in turn to a number of meta-analyses ( ; ; ; ). The conclusion in general is that any effect of epidural steroids is short lived and the treatment effect is small. The systematic review and meta-analysis by concluded that there was effectiveness for radiculopathy but the benefits were small and not sustained, and that there was no effectiveness for spinal stenosis. The 2016 systematic review and meta-analysis by concluded that lidocaine alone and lidocaine in conjunction with steroids were equally effective. In six studies with 649 patients, there was no difference in improvement comparing lidocaine alone to lidocaine with steroids at 3 or 12 months. Epidural steroids administered with sodium chloride solution were ineffective. This raises the possibility that any benefit from steroids may arise through some mechanism other than an anti-inflammatory one and that perhaps epidural injections should use local anesthetics alone, as was routinely the practice between 1901 and the early 1950s.

Physical therapy such as back exercises should probably be avoided with an acute painful radiculopathy. After the acute stage has subsided, physical therapy may certainly be useful. However, if back exercises reproduce radicular pain, they should be stopped immediately. Numerous other nonoperative treatments are occasionally used, although demonstration of effectiveness in randomized trials is generally lacking. These treatments include transcutaneous electrical nerve stimulation, acupuncture, massage, and spinal manipulation.

The vast majority of patients with acute LSR improve with time and conservative management, even when the disk herniation remains visible on MRI. Surgical referral should be prompt if the patient develops any evidence of cauda equina syndrome. Whether patients with more routine LSR benefit from surgery remains controversial. There is no evidence that early surgery, in the absence of a severe or progressive neurologic deficit, improves the outcome for radiculopathy due to lumbar disk herniation with or without spinal stenosis. Many patients with a minor motor deficit, such as a mild foot drop, recover with nonsurgical treatment.

If surgery is required, a number of techniques are currently employed, including traditional open laminectomy and discectomy, microdiscectomy with hemilaminectomy, and minimally invasive techniques such as percutaneous discectomy, laser discectomy, percutaneous endoscopic discectomy, microendoscopic discectomy, and radiofrequency nucleoplasty ( ). Studies comparing surgery to nonsurgical management for radiculopathy generally show that operated patients improve more quickly but that outcomes are similar within 2 years ( ; ). Nonspecific LBP with DDD associated with subacute or chronic LBP without radicular symptoms is another topic altogether.