Six RCTs met the inclusion criteria (442 patients, range 26 to 140). Two RCTs used sham acupuncture as the comparator and four used active controls. Methodological quality was considered to be generally low; only two RCTs reported using an appropriate randomisation method, two trials blinded patients, one trial blinded the outcome assessor and two trials described drop-outs; none of the trials reported allocation concealment methods. All four RCTs from China reported positive results; the trials from Germany and USA did not.
Compared to steroid nerve blocks, response rate was significantly better with acupuncture (RR 1.28, 95% CI 1.08 to 1.52; Ι² 10%; two RCTs, 144 patients). Compared to sham acupuncture (using penetrating irrelevant meridian points) no significant benefit of acupuncture were observed (one RCT). Compared to steroid tablets, the only significant benefit of acupuncture was seen in distal motor latency in terms of nerve conduction study (mean difference -0.63, 95% CI -1.09 to -0.17). Compared to tuina alone, acupuncture plus tuina showed a significant benefit in terms of nerve conduction study of the median nerve (mean difference 1.05, 95% CI 0.51 to 1.59) but not the ulna nerve. Compared to sham laser acupuncture, laser acupuncture showed no benefit in terms of relieving night pain.
Adverse events (two RCTs): One RCT reported that 56 out of 173 adverse events were related to needle acupuncture. The second RCT reported an adverse event rate of 5% in the needle acupuncture group. No serious adverse events were reported in either trial.