Nine RCTs (n=1,749) were included.
In terms of study quality, only 4 RCTs described the method of randomisation and three reported adequate detail of the intervention. Allocation concealment was not reported in adequate detail in any of the included trials. Loss to follow-up exceeded 20% in 5 trials, with mean loss to follow-up across all trials at 50%.
Acupuncture did not significantly affect cocaine abstinence (OR 0.76, 95% CI: 0.45, 1.27, P=0.3; 7 RCTs). Acupuncture was also not found to significantly effect cocaine abstinence when considering all drop-outs as treatment failures (OR 0.76, 95% CI: 0.54, 1.08, P=0.12).
Allocation concealment, use of intention-to-treat analyses and large loss to follow-up (more than 50%) did not explain the observed heterogeneity.
Reported adverse events included pain and fear of needles.