Invisible Harms, Visible Benefits? Adverse Event Reporting in Physical Activity Interventions for Neurodevelopmental...

Clinical guidelines and practice increasingly endorse physical activity interventions for individuals with neurodevelopmental disorders. However, because monitoring and reporting of potential intervention-related harms are often incomplete, the current evidence base regarding safety remains limited.

A systematic search of PubMed, Web of Science, Scopus, the Cochrane Library, and EBSCO was conducted on November 11, 2025, followed by three rounds of supplementary snowball searching completed on November 20, 2025. Eligible studies were randomized controlled trials of physical activity interventions in populations with neurodevelopmental disorders. Following prespecified procedures recommended by the CONSORT-Harms extension and the Cochrane Handbook for Systematic Reviews of Interventions, we characterized adverse event monitoring and reporting practices across studies, summarized reasons for participant withdrawal, and considered the potential safety signals these reasons might indicate. Random-effects meta-analyses estimated the relative risk of adverse events for physical activity interventions versus non-exercise controls, and potential moderators were examined via subgroup and meta-regression analyses.

Among the 41,852 deduplicated records, 309 randomized controlled trials were included, comprising 322 study arms and 13,228 participants with neurodevelopmental disorders. None of the included trials fully complied with CONSORT-Harms recommendations for systematic harm monitoring and reporting. Only 58 trials described adverse event monitoring, mostly using passive, non-prespecified procedures; among those reporting results, approximately one quarter reported at least one adverse event or adverse effect, most commonly intervention-induced delayed-onset muscle soreness, mild pain or irritation, and related psychological discomfort. Among the 201 studies that reported reasons for participant withdrawal, 38 statements were too vague to determine the event characteristics. Sixteen studies reported adverse events (e.g., health problems and emotional discomfort); however, because none used a systematic adjudication procedure, attributing these events to the intervention was difficult. Forty-five studies met the inclusion criteria for meta-analysis. The pooled analysis indicated that physical activity interventions did not significantly increase the risk of reported adverse events (RR = 1.14; 95% CI 0.67-1.94). However, trials with prespecified monitoring procedures had a significantly higher risk estimate (RR = 4.32, 95% CI 1.36-13.68).

Existing reported results suggest that these interventions do not significantly increase the overall risk of adverse events. However, the strength of this conclusion is limited by widespread systematic underreporting in the original studies and by substantial heterogeneity in methods used to monitor, identify, and assess harms. In contrast, studies that implemented prespecified adverse event monitoring procedures were generally able to identify and report adverse events more comprehensively, rather than indicating that these trials carry a genuinely higher risk. Future research should establish and implement standardized procedures for adverse event monitoring and reporting, standardize procedures for adjudicating and attributing reasons for withdrawal, and systematically report key variables related to intervention implementation and descriptive adverse event data to strengthen the credibility of benefit-risk assessments of physical activity interventions in this population.