Trends and quality of randomized controlled trials on acupuncture conducted in Japan by decade from the 1960s to the 2010s: a systematic review.

backgroundMore new randomized controlled trials (RCTs) on acupuncture have been published in Japan since our last updated systematic review (2010). This systematic review aimed to evaluate the quality of RCTs on acupuncture conducted in Japan and understand the decade-wise changes in the methodological characteristics of the relevant RCTs.

methodsThe literature search was performed using Ichushi Web, the Cochrane Central Register of Controlled Trials (CENTRAL), PubMed and our team's compilation of relevant papers. We included full-length papers reporting RCTs that examined the clinical effects of acupuncture on patients in Japan published in or before 2019. We assessed the risk of bias (RoB), sample size, control setting, negative trial reporting, informed consent, ethics approval, trial registration, and adverse event reporting.

resultsA total of 99 articles reporting 108 eligible RCTs were identified. The number of RCTs published in each decade was 1, 6, 9, 5, 40, and 47 in the 1960s, 1970s, 1980s, 1990s, 2000s, and 2010s, respectively. Quality assessment using the Cochrane RoB tool revealed that "sequence generation" improved in and after 1990 (73%-80% of RCTs were rated as "low") and "blinding of outcome assessors" slightly improved in and after the 2000s (40%-50% judged as "low"). However, "high" or "unclear" remained the dominant grades in other domains. Clinical trial registration and adverse events were reported only in 9% and 28% of the included RCTs even in the 2010s, respectively. A different acupuncture method or different point selection (e.g., deep vs. shallow insertion) was the most dominant control setting before 1990, while sham (or "placebo") needling and/or sham acupoints became the most dominant in the 2000s. The proportion of RCTs with positive results was 80% in the 2000s and 69% in the 2010s.

conclusionsThe quality of RCTs on acupuncture conducted in Japan did not appear to have improved over the decades except for "sequence generation." While the culture of submitting negative trial reports was prevalent in the Japanese acupuncture research milieu as late as the 1990s, the overall quality of the relevant trials needs to be further improved.