No studies have formally shown analgesics, such as paracetamol, nonsteroidal anti-inflammatory drugs (NSAIDs), or opioids, to be effective for cervical radicular pain. Their use is based on reputation or the expectation that they might help. No analgesic is curative of cervical radicular pain because none addresses the cause of pain. At best, analgesics may be palliative of pain while natural resolution takes place or while the patient awaits more definitive therapy.

There is a prima facie rationale for the use of oral corticosteroids. In as much as the mechanism of cervical radicular pain might involve inflammation of the nerve roots, corticosteroids might suppress that inflammation and promote natural resolution. A typical regimen is betamethasone, 0.5 mg, four three times a day for 5 days, tapering to two three times a day for 5 days, and finally one three times a day for 5 days. No study has tested the efficacy of this or any other regimen of oral corticosteroids for cervical radicular pain.

The rationale for traction is that it decompresses spinal nerves by axially separating the vertebral bodies. While this may the case during the application of traction, there is no evidence that decompression persists once the patient resumes an upright posture. A systematic review of seven eligible studies found no evidence of efficacy for traction in the treatment of cervical radicular pain (9). A subsequent study found no value from adding traction to a program of manual therapy and exercise (10).

There is no consistent definition of what constitutes socalled conservative therapy for cervical radicular pain. It can consist of various combinations of traction, drugs, physical therapy, exercises, using a collar, bed rest, and transcutaneous electrical nerve stimulation. The few controlled trials paint a sobering picture of its efficacy for cervical radicular pain.

In one study, treatment with traction or collars was no more effective than various placebo treatments (11). Irrespective of treatment administered, about 20% of patients were cured, and at 6 months, nearly 30% were free of pain. Another study found no differences in the proportions of patients “improved” after treatment with isometric exercises or traction or no treatment (12).

In an observational study of 61 patients treated with traction, heat, and massage, 16% obtained complete relief initially, and 25% were markedly improved. Between 6 months and 5 years later, only 40% of those patients completely relieved initially remained so relieved (13).

Another observational study compared three interventions for acute cervical radicular pain: a semihard collar combined with rest for 3 to 6 weeks, twice weekly physiotherapy and home exercises for 6 weeks, and maintain activities as much as possible with no specific treatment. Each of the specific interventions achieved earlier resolution of pain (at 6 weeks), but the magnitude of difference over no specific treatment was only 12 points less on a 100-point scale (14). By 26 weeks, there were no differences between the groups. By this time, all groups had a mean pain score of 20/100, consistent with the favorable natural history of acute cervical radicular pain.

The best results for conservative care for cervical radicular pain were reported in an observational study of 28 patients, who were treated using ice, rest, a hard cervical collar, NSAIDs, manual and mechanical traction, progressive strengthening exercises of the shoulder girdle and chest and training in postural control and body mechanics, prednisone, epidural or selective nerve root injections of corticosteroids, acupuncture, and transcutaneous electrical nerve stimulation (15). An excellent outcome was defined as no pain and no limitations in activities. A good outcome was defined as minimal limitations in activity and minor neck pain only. At follow-up between 1.2 and 3.1 years, 20 of the 24 patients achieved a good or excellent outcome, and 22 returned to full work (15).

Although this study showed that surgery could be avoided by aggressive conservative therapy, many questions remain unanswered. It is not evident how representative the sample was of the general population with cervical radicular pain. It is not evident if the intensive, 3-month regimen used by the authors is cost effective and more effective than simpler measures, such as those used by others (16). It is not evident if the outcomes were due to the enthusiasm of the therapists, nor is it evident which of the components of the regimen were critical for achieving the outcome. The latter is a seminal question because whereas all of the other components of the program lack evidence of efficacy, or have been shown to be ineffective, epidural injections have been shown to be effective (see below).

Epidural injection of steroids by the interlaminar route can be performed with or without fluoroscopic guidance. The descriptive studies of this intervention have been poorly reported; they purport to announce a useful intervention but have not consistently provided meaningful data (Table 19.1). Although all reported success rates, the duration of follow-up was inconsistent and often short.

Most descriptive studies suggest that between 10% and 30% of patients might get complete relief of pain, but for uncertain periods; a further 20% to 40% might get at least 75% relief. Remarkable are the results of Castagnera et al. (22) who reported 21% of patients maintaining complete relief for 4 years and a further 43% maintaining 75% relief for this period. Such outcomes are unparalleled in the literature.

One study (23) reported results that do not lend themselves to inclusion in Table 19.1. In patients with symptoms for less than 180 days, the mean reduction of pain was between 70% and 86%, depending on duration, with standard deviations between 11% and 18%. In patients with chronic pain, the mean reduction was 58% following single injections but 74% following a continuous infusion.

Cervical epidural steroids have been validated by one controlled study (24). Twenty-five patients received epidural injections of steroid and lidocaine, and 17 received injections of steroid and lidocaine into the dorsal neck muscles. Patients were entitled to one to three injections depending on response. The authors reported significant differences in favor of epidural steroids, using a statistical comparison of relative risk, which is a generous statistical test. A more demanding test is a comparison of success rates, using 95% confidence intervals of a proportion. When this test is applied (Table 19.2), some but not all comparisons still favor epidural steroids. At 1 week after
treatment, the proportions of patients achieved greater than 75% relief or between 50% and 74% relief of pain after epidural steroids were not significantly greater than the corresponding proportions after intramuscular steroids, but the combined proportion of patients achieving at least 50% relief is greater for epidural steroids (Table 19.2). At 1 year after treatment, the proportions of patients achieving at least 75% relief and greater than 50% relief significantly favor epidural steroids.

These results refute nonspecific treatment effects and refute a systemic effect of steroids. They justify the epidural placement of steroids.

Transforaminal injections involve placing a needle into the intervertebral foramen that lodges the affected nerve. The procedure is performed under fluoroscopic or computerized tomography guidance.

The descriptive studies that reported outcomes in any detail achieved successful outcomes in only small proportions of patients, namely, complete relief in 14% of 51 patients (26); 29% of 32 patients with complete relief at 3 and 6 months, diminishing to 20% at 12 months (27); and five of 21 patients with pain less than 10/100 at 4 months (28). These figures portray only modest success rates, and ones compatible with placebo responses.

More impressive success rates were reported by other descriptive studies, such as 60% of 20 patients (29) and 60% of 30 patients (30). In both instances, however, the authors grouped excellent and good results, without providing separate figures for each category. In the former, a study a good result amounted to a pain score if less than 3/10; in the latter, it was at least 50% reduction in pain. Another study reported a decrease of mean pain scores from 6.5 to 2, in 19 patients, at 4 months (31).

Two controlled trials have been reported, one as a full paper (32) and one as an abstract (33). The first study randomized patients to receive either carbocaine and methylprednisolone or carbocaine and normal saline. Six out of twenty patients in each group achieved significant subjective reductions in radicular pain at 3 weeks, but the outcomes were not further defined (32). In the second study, 4 (16%) of 25 patients treated with bupivacaine alone and 2 (10%) of 20 treated with bupivacaine and betamethasone avoided surgery for 20 months (33).

Both studies attest to a modest success rate for cervical transforaminal injections of steroids, consonant with the earlier descriptive studies. Both controlled trials showed no attributable effect of adding steroids to the injection. Whether or not local anesthetic alone has a therapeutic effect or only a placebo effect has yet to be determined.

Cervical transforaminal injections of steroids have been associated with catastrophic neurologic complications
(34, 35, 36 and 37). The prevailing view is that if injected into a radicular artery, or into the vertebral artery, particulate steroids act as emboli and cause spinal cord or brainstem infarction. The incidence of such complications is not known, but at least 78 cases have been recorded in the literature (34). Some exponents of the procedure have reported no instances of neurologic complications in 1,063 patients, and only minor complications in 1.6% of cases (38). Implicitly, serious complications are rare, but they are devastating in nature. Ironic is that fact that steroids have been implicated as the offending agent, yet steroids have been shown to be redundant for the small success rate of this procedure.