Treating sciatica in the face of poor evidence

BMJ 2012; 344 doi: http://dx.doi.org/10.1136/bmj.e487 (Published 13 February 2012) Cite this as: BMJ 2012;344:e487
  1. Roger Chou, associate professor of medicine
  1. 1Department of Medicine and Department of Medical Informatics and Clinical Epidemiology, Oregon Health and Science University, Portland, OR 97225, USA
  1. chour{at}ohsu.edu

It may be necessary to extrapolate from evidence on treatment of other neuropathic pain syndromes

The term “sciatica” is often applied to any presentation of low back and leg pain, although lumbosacral radiculopathy is a more specific term for the condition experienced by patients with low back pain who have impingement of lumbosacral nerve roots as they emerge from the spinal canal. This results in pain and sensory deficits in a dermatomal distribution and sometimes motor weakness in the corresponding myotomal distribution. Because the most commonly affected nerve roots are L4/L5 and L5/S1, pain typically radiates below the knee, and leg pain (elicited by performing the straight leg raise test) may be more pronounced than back pain. The most common cause of lumbosacral radiculopathy is intervertebral disc herniation, which occurs in about 3% of patients with acute low back pain.1 Other causes include spondylolisthesis and foraminal stenosis owing to degenerative osteophytes.

The linked systematic review and meta-analysis by Pinto and colleagues (doi:10.1136/bmj.e497) finds little direct evidence to support the use of any drug for this condition.2 Drugs commonly used to treat lumbosacral radiculopathy include analgesics such as paracetamol, non-steroidal anti-inflammatory drugs (NSAIDs), and opioids; drugs for neuropathic pain such as anticonvulsants and certain antidepressants; skeletal muscle relaxants; and systemic corticosteroids, whose anti-inflammatory effects may relieve pressure from herniated intervertebral discs on lumbosacral nerve roots.

Pinto and colleagues found few randomised trials of the effectiveness of such drugs, and, when meta-analysis was possible, pooled estimates generally showed no benefits over placebo for either pain or function. Furthermore, available trials typically had small sample sizes, did not perform long term follow-up, and had methodological shortcomings. The authors therefore concluded that for most drugs the quality of evidence was very low to low. Systemic corticosteroids were the one potential exception, with a meta-analysis of two small (n=78 and n=60) trials showing positive effects on short-term (more than two to three months) pain compared with placebo, although the benefit was small (about 10 points on a 100 point pain scale).

A shortcoming of the systematic review is that it focused on mean improvements in pain or disability scores and did not evaluate the likelihood of experiencing a clinically meaningful benefit,2 which is often considered a better measure of treatment effects. Several trials of systematic corticosteroids in the review reported this information, with mixed results. The systematic review also did not include some relevant older negative randomised trials of systemic corticosteroids, which reported dichotomous outcomes.3 4 5 Thus, the available evidence is insufficient to support a treatment recommendation for systemic corticosteroids, although a trial currently in progress has a target sample (n=270) substantially larger than any previous study and should help clarify their role (http://clinicaltrials.gov/show/NCT00668434).

Given the paucity of evidence available from drug trials, how should clinicians select a treatment for this common, usually painful, problem? Clinicians still need to make treatment decisions even when evidence is suboptimal. In such situations it is necessary to use “indirect” evidence by extrapolating from the findings of trials evaluating drugs for other conditions and making assumptions about generalisability.6 In this case, it is reasonable to assume that true lumbosacral radiculopathy should respond to drugs in a similar way to other types of neuropathic pain. Therefore, for patients with symptoms and signs typical of lumbosacral radiculopathy, clinicians may consider drugs that are effective for other types of neuropathic pain, such as pregabalin or gabapentin, and certain selective serotonin-noradrenaline (norepinephrine) reuptake inhibitors and tricyclic antidepressants.7 However, more research is needed to confirm that lumbosacral radiculopathy responds to drugs similarly to other types of neuropathic pain. Pinto and colleagues found that NSAIDs had small but unclear benefits, which is consistent with the perception that NSAIDs are generally not effective for neuropathic pain, although the evidence is limited.8 Opioids, although effective for neuropathic pain,9 are not a first line drug owing to the potential for misuse and overdosing. Opioids should be reserved for severe or intractable lumbosacral radiculopathy, in appropriately selected and monitored patients. Skeletal muscle relaxants and benzodiazepines are not recommended as a first line drug because they have not been well studied for neuropathic pain and can have sedative effects.

Regrettably, no evidence is available to guide drug choices in patients with back and leg pain with features inconsistent with lumbosacral radiculopathy (for example, the pain is in a non-dermatomal distribution), although it is probably reasonable to follow general guidelines on the management of low back pain.


Cite this as: BMJ 2012;344:e487


  • Research, doi:10.1136/bmj.e497
  • Competing interests: The author has completed the ICMJE uniform disclosure form at www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and declares: receipt of research funding from the American Pain Society to develop guidelines on low back pain evaluation and treatment; RC is a consultant for Wellpoint, Blue Cross Blue Shield Association, and Palladian Health (all of which deliver or manage healthcare and do not manufacture drugs); no other relationships or activities that could appear to have influenced the submitted work.

  • Provenance and peer review: Commissioned; not externally peer reviewed.