Background: We aimed to explore the incidence and risk of adverse events associated with manual therapies.

Method: The main health electronic databases, plus those specific to allied medicine and manual therapy professions, were searched. Our inclusion criteria for relevant studies were: manual therapies only; administered by regulated therapists; a clearly described intervention; adverse events reported. We performed a meta-analysis using incident estimates of proportions and random effects models from relevant prospective cohort studies and randomised controlled trials (RCTs) published after 1997.

Results: Eight prospective cohort studies and 31 manual therapy RCTs were identified and analysed. The incidence estimate of proportions of minor or moderate transient adverse events after manual therapy was ~41% (CI 95% 17–68%) in the cohort studies and 22% (CI 95% 11.1–36.2%) in the RCTs. The estimate for major adverse events was between 0.007 and 0.13%. No deaths or vascular accidents occurred in any studies. The pooled relative risk (RR) for experiencing adverse events with exercise, or with sham, passive or control interventions compared to manual therapy was similar, but compared to drug therapies greater (RR 0.05, CI 95% 0.01–0.20) and less when compared to general practitioner or usual care (RR 1.91, CI 95% 1.39–2.64).

Conclusions: Our data indicate a very low risk of major adverse events with manual therapy, but around half manual therapy patients may experience minor to moderate adverse events after treatment. The relative risk of adverse events appears greater with drug therapy but less with usual medical care.

Conflicts of Interest: D Carnes & T Mars

Source of Funding: National Council for Osteopathic Research