Although several systematic reviews and meta-analyses have demonstrated the benefits of exercise interventions in older adults with frailty, the potential harm associated with these interventions has not been systematically synthesised. This systematic review aims to examine the adverse events reported in exercise intervention trials involving older adults with frailty and to compare the risk of adverse events between the intervention and control groups.

Searches will be performed in four electronic databases (PubMed, Cochrane Library, Web of Science and SPORTDiscus) for published trials. Eligible studies will be randomised controlled trials of exercise interventions, including older adults with frailty aged ≥60 years, with frailty identified using a validated method. Five reviewers and three referees, all with expertise in exercise interventions, will be assigned to three independent review teams to ensure efficient screening. Reviewers will independently screen titles, abstracts and full texts using Rayyan, and then extract trial and adverse event data into an Excel spreadsheet. The risk of bias in eligible trials will be assessed using the Cochrane Risk of Bias 2 (RoB-2) tool. The referees will resolve any disagreements between the two reviewers throughout the screening, data extraction and risk-of-bias assessment processes. The primary outcome is adverse events, defined as any unfavourable, unintended signs, symptoms or disease that occurred during the study period. An independent biostatistician will perform a random-effects meta-analysis using a generalised linear mixed model with a binomial likelihood and a logit link to estimate the pooled risk ratios (RRs) for adverse events in the intervention group relative to the control group. Publication bias will be evaluated using funnel plots and Egger’s regression test. Depending on the number of available studies, subgroup analyses will be conducted to examine differences in RRs according to the study quality, duration of intervention, exercise frequency, setting and supervision.

Ethical approval was not required because we did not use specific patient data. The findings of the systematic review and meta-analysis will be disseminated through publication in a peer-reviewed journal and presentation at appropriate conferences.

CRD420251180645.