Use of Acupuncture for Adult Health Conditions, 2013 to 2021: A Systematic Review

This study represents a comprehensive mapping of the scientific evidence on acupuncture for adult health conditions, covering the period from 2013 to 2021. The investigators performed a systematic review of systematic reviews — a methodological approach known as an 'overview' — to assess the certainty or quality of available evidence on the effects of acupuncture.

The methodology involved a computerized search of five major databases: PubMed, Allied and Complementary Medicine Database, Cochrane Database of Systematic Reviews, Web of Science, and Database of Abstracts of Reviews of Effects. The fundamental inclusion criterion was that systematic reviews must use formal methods for assessing evidence certainty, such as the GRADE (Grading of Recommendations, Assessment, Development and Evaluations) system.

Of the 1,617 titles initially identified, 434 systematic reviews were considered eligible. Of these, 127 used formal methods to assess evidence certainty and 82 were included in the final mapping, covering 56 different health conditions. The conditions were categorized into three main groups: pain (including chronic pain, migraine, fibromyalgia, arthritis), mental health conditions (depression, anxiety, insomnia, substance use disorders), and other conditions (stroke, fertility, gastrointestinal and respiratory problems).

The results reveal a striking finding: despite the vast number of randomized controlled trials and systematic reviews on acupuncture, only a very small minority of conclusions were rated as having high or moderate evidence certainty. Specifically, only 4 conclusions with high-certainty evidence and 31 conclusions with moderate-certainty evidence were identified. The vast majority of conclusions (more than 60) were classified as having low or very low evidence certainty.

Looking at these higher-certainty conclusions more closely, approximately 75% of the moderate- or high-certainty evidence involved comparisons of acupuncture with sham acupuncture or no treatment, rather than comparisons with other established active therapies. In addition, about 10% of the conclusions classified as moderate certainty actually indicated that acupuncture was not superior to the comparator treatment.

Conditions with high-certainty evidence included: shoulder pain relief (compared with sham acupuncture), improvement in pain, fatigue, and sleep quality in fibromyalgia, lack of benefit in assisted reproductive therapy, and greater efficacy of electroacupuncture in stroke when combined with conventional treatment. Conditions with moderate-certainty evidence included relief of chronic musculoskeletal pain, reduction of preoperative anxiety, improvement of migraine and tension-type headache, and benefits in certain gastrointestinal conditions.

As for adverse events, 19 reviews formally assessed the safety of acupuncture. Most reported adverse events similar to or lower than in the control groups, with low- to high-certainty evidence that acupuncture is at least as safe as usual care.

The clinical implications of this study are significant. For clinicians, patients, and policy makers, the results suggest that, although acupuncture has been extensively studied, confidence in the available evidence remains limited for most conditions. This does not imply that acupuncture is ineffective, but rather that the methodological quality of the studies or the magnitude of the effects may not be sufficient to generate high-certainty conclusions.

The study also reveals a possible imbalance in the allocation of research resources: during the 9-year period analyzed, roughly as many systematic reviews were produced as new randomized controlled trials. The authors suggest that the field would benefit more from investments in new high-quality clinical trials than in additional reviews that frequently arrive at low-certainty conclusions.

Limitations include the possibility of not having identified all relevant reviews, especially those in languages not indexed in the databases searched, and reliance on the assessments made by the original review authors. An important conceptual limitation relates to the controversy over what constitutes adequate 'sham' acupuncture, since different forms of sham may have their own physiological effects, complicating interpretation of results.