BACKGROUND: Heart failure with mildly reduced or preserved ejection fraction (hereafter referred to as HFpEF) is the most common type of heart failure and is associated with a high risk of hospitalisation and death, especially in patients with overweight, obesity, or type 2 diabetes. In the STEP-HFpEF and STEP-HFpEF DM trials, semaglutide improved heart failure-related symptoms and physical limitations in participants with HFpEF. Whether semaglutide also reduces clinical heart failure events in this group remains to be established. METHODS: We conducted a post-hoc pooled, participant-level analysis of four randomised, placebo-controlled trials (SELECT, FLOW, STEP-HFpEF, and STEP-HFpEF DM) to examine the effects of once-weekly subcutaneous semaglutide (2·4 mg in SELECT, STEP-HFpEF, and STEP-HFpEF DM; 1·0 mg in FLOW) on heart failure events. The STEP-HFpEF and STEP-HFpF DM trials enrolled participants with obesity-related HFpEF, the SELECT trial enrolled participants with atherosclerotic cardiovascular disease and overweight or obesity, and the FLOW trial enrolled participants with type 2 diabetes and chronic kidney disease. Hence, for this analysis, we include all participants from the STEP-HFpEF trials and those with an investigator-reported history of HFpEF from SELECT and FLOW. The main outcomes for this analysis were the composite endpoint of time to cardiovascular death or first worsening heart failure event (defined as hospitalisation or urgent visit due to heart failure), time to first worsening heart failure event, and time to cardiovascular death. Efficacy and safety endpoints were analysed with the full analysis set (ie, all participants randomly assigned to treatment, according to the intention-to-treat principle). The SELECT, FLOW, STEP-HFpEF, and STEP-HFpEF DM trials are registered at ClinicalTrials.gov, NCT03574597, NCT03819153, NCT04788511, and NCT04916470, respectively, and all are complete. FINDINGS: Across the four trials, 3743 (16·8%) of 22 282 participants had a history of HFpEF (1914 assigned to semaglutide and 1829 assigned to placebo). In this group of participants with HFpEF, semaglutide reduced the risk of the combined endpoint of cardiovascular death or heart failure events (103 [5·4%] of 1914 in the semaglutide group had events vs 138 [7·5%] of 1829 in the placebo group; hazard ratio [HR] 0·69 [95% CI 0·53-0·89]; p=0·0045). Semaglutide also reduced the risk of worsening heart failure events (54 [2·8%] vs 86 [4·7%]; HR 0·59 [0·41-0·82]; p=0·0019). No significant effect on cardiovascular death alone was seen (59 [3·1%] vs 67 [3·7%]; HR 0·82 [0·57-1·16]; p=0·25). A lower proportion of patients treated with semaglutide had serious adverse events than did those who were treated with placebo (572 [29·9%] vs 708 [38·7%]). INTERPRETATION: In patients with HFpEF, semaglutide reduced the risk of the combined endpoint of cardiovascular death or worsening heart failure events, and worsening heart failure events alone, whereas its effect on cardiovascular death alone was not significant. These data support the use of semaglutide as an efficacious therapy to reduce the risk of clinical heart failure events in patients with HFpEF, for whom few treatment options are currently available. FUNDING: Novo Nordisk.