Reporting characteristics and risk of bias in randomised controlled trials of acupuncture analgesia published in PubMed-listed journals

Background Acupuncture analgesia has been evaluated by a number of randomised controlled trials (RCTs); however, a systematic summary of reporting quality of RCTs in this specific field is lacking. Objective To examine the reporting characteristics and risk of bias of RCTs of acupuncture analgesia indexed in the PubMed database. Methods A PubMed search of RCTs of acupuncture analgesia was conducted through November 2015. The Cochrane Collaboration Risk of Bias Tool was used to assess the risk of bias of each trial. Results 206 articles were identified across 59 journals (impact factor 0.4-20), of which 56% of articles and 86% of journals were Science Citation Index (SCI)-indexed. Nearly half of the articles were published in China. The next most represented countries of origin were the UK (22%) and USA (21%). Of the included trials, postoperative pain was the most prevalent phenotype, and manual acupuncture was the most frequently applied type of stimulation (46%). A total of 12% of articles reported on analgesic mechanisms. The most frequently used acupuncture points were LI4, ST36, PC6, SP6 and Shenmen. The overwhelming majority of trials were considered to be at high risk of bias (84%). Furthermore, 79% of trials enrolled <50 participants per treatment arm. Conclusions RCTs of acupuncture analgesia indexed in PubMed journals generally exhibited poor reporting of methodological and treatment details. Future studies should provide more information regarding clinical trial registration, blinding of participants (including sham procedures where applicable) and outcome assessors, as well as the training and qualification of acupuncturists.