Background: Frailty endangers the health of older adults. Furthermore, the prevalence of frailty continues to increase as the global population ageing.

Objective: To update evidence on the effectiveness of non-pharmacological interventions for frailty by conducting a network meta-analysis (NMA) of randomised controlled trials (RCTs).

Methods: Eight databases were searched from January 1, 2000, until September 24, 2021. RCTs of interventions for frailty among participants aged ≥60 years were considered eligible. The primary outcome was frailty. Pairwise meta-analysis and NMA were performed, with the pooled standardised mean difference (SMD) and 95% confidence interval (CI) being reported.

Results: A total of 69 RCTs were included after screening 16,058 retrieved citations. There were seven types of interventions (11 interventions) for frailty among the included RCTs. Physical activity (PA) (pooled SMD = 0.43, 95% CI: 0.34-0.51), multicomponent intervention (pooled SMD = 0.34, 95% CI: 0.23-0.45) and nutrition intervention (pooled SMD = 0.21, 95% CI: 0.06-0.35) were associated with reducing frailty compared to control, of which PA was the most effective type of intervention. In terms of specific types of PA, resistance training (pooled SMD = 0.58, 95% CI: 0.33-0.83), mind-body exercise (pooled SMD = 0.57, 95% CI: 0.24-0.90), mixed physical training (pooled SMD = 0.47, 95% CI: 0.37-0.57) and aerobic training (pooled SMD = 0.36, 95% CI: 0.09-0.62) were associated with a reduction in frailty compared to usual care. Resistance training was the most effective PA intervention.

Conclusion: Resistance training has the best potential to reduce frailty in older adults. This finding might be useful to clinicians in selecting interventions for older adults with frailty.