Neuropathic pain may be caused by nerve damage, and is often followed by changes to the central nervous system. Uncertainty remains regarding the effectiveness and safety of acupuncture treatments for neuropathic pain, despite a number of clinical trials being undertaken.

To assess the analgesic efficacy and adverse events of acupuncture treatments for chronic neuropathic pain in adults.

We searched CENTRAL, MEDLINE, Embase, four Chinese databases, ClinicalTrials.gov and World Health Organization (WHO) International Clinical Trials Registry Platform (ICTRP) on 14 February 2017. We also cross checked the reference lists of included studies.

Randomised controlled trials (RCTs) with treatment duration of eight weeks or longer comparing acupuncture (either given alone or in combination with other therapies) with sham acupuncture, other active therapies, or treatment as usual, for neuropathic pain in adults. We searched for studies of acupuncture based on needle insertion and stimulation of somatic tissues for therapeutic purposes, and we excluded other methods of stimulating acupuncture points without needle insertion. We searched for studies of manual acupuncture, electroacupuncture or other acupuncture techniques used in clinical practice (such as warm needling, fire needling, etc).

We used the standard methodological procedures expected by Cochrane. The primary outcomes were pain intensity and pain relief. The secondary outcomes were any pain-related outcome indicating some improvement, withdrawals, participants experiencing any adverse event, serious adverse events and quality of life. For dichotomous outcomes, we calculated risk ratio (RR) with 95% confidence intervals (CI), and for continuous outcomes we calculated the mean difference (MD) with 95% CI. We also calculated number needed to treat for an additional beneficial outcome (NNTB) where possible. We combined all data using a random-effects model and assessed the quality of evidence using GRADE to generate ‘Summary of findings’ tables.

We included six studies involving 462 participants with chronic peripheral neuropathic pain (442 completers (251 male), mean ages 52 to 63 years). The included studies recruited 403 participants from China and 59 from the UK. Most studies included a small sample size (fewer than 50 participants per treatment arm) and all studies were at high risk of bias for blinding of participants and personnel. Most studies had unclear risk of bias for sequence generation (four out of six studies), allocation concealment (five out of six) and selective reporting (all included studies). All studies investigated manual acupuncture, and we did not identify any study comparing acupuncture with treatment as usual, nor any study investigating other acupuncture techniques (such as electroacupuncture, warm needling, fire needling).

One study compared acupuncture with sham acupuncture. We are uncertain if there is any difference between the two interventions on reducing pain intensity (n = 45; MD -0.4, 95% CI -1.83 to 1.03, very low-quality evidence), and neither group achieved ‘no worse than mild pain’ (visual analogue scale (VAS, 0-10) average score was 5.8 and 6.2 respectively in the acupuncture and sham acupuncture groups, where 0 = no pain). There was limited data on quality of life, which showed no clear difference between groups. Evidence was not available on pain relief, adverse events or other pre-defined secondary outcomes for this comparison.

Three studies compared acupuncture alone versus other therapies (mecobalamin combined with nimodipine, and inositol). Acupuncture may reduce the risk of ‘no clinical response’ to pain than other therapies (n = 209; RR 0.25, 95% CI 0.12 to 0.51), however, evidence was not available for pain intensity, pain relief, adverse events or any of the other secondary outcomes.

Two studies compared acupuncture combined with other active therapies (mecobalamin, and Xiaoke bitong capsule) versus other active therapies used alone. We found that the acupuncture combination group had a lower VAS score for pain intensity (n = 104; MD -1.02, 95% CI -1.09 to -0.95) and improved quality of life (n = 104; MD -2.19, 95% CI -2.39 to -1.99), than those receiving other therapy alone. However, the average VAS score of the acupuncture and control groups was 3.23 and 4.25 respectively, indicating neither group achieved ‘no worse than mild pain’. Furthermore, this evidence was from a single study with high risk of bias and a very small sample size. There was no evidence on pain relief and we identified no clear differences between groups on other parameters, including ‘no clinical response’ to pain and withdrawals. There was no evidence on adverse events.

The overall quality of evidence is very low due to study limitations (high risk of performance, detection, and attrition bias, and high risk of bias confounded by small study size) or imprecision. We have limited confidence in the effect estimate and the true effect is likely to be substantially different from the estimated effect.

Due to the limited data available, there is insufficient evidence to support or refute the use of acupuncture for neuropathic pain in general, or for any specific neuropathic pain condition when compared with sham acupuncture or other active therapies. Five studies are still ongoing and seven studies are awaiting classification due to the unclear treatment duration, and the results of these studies may influence the current findings.