Abstract Aims The International Study of Comparative Health Effectiveness with Medical and Invasive Approaches (ISCHEMIA) trial prespecified an analysis to determine whether accounting for recurrent cardiovascular events in addition to first events modified understanding of the treatment effects. Methods and results Patients with stable coronary artery disease (CAD) and moderate or severe ischaemia on stress testing were randomized to either initial invasive (INV) or initial conservative (CON) management. The primary outcome was a composite of cardiovascular death, myocardial infarction (MI), and hospitalization for unstable angina, heart failure, or cardiac arrest. The Ghosh–Lin method was used to estimate mean cumulative incidence of total events with death as a competing risk. The 5179 ISCHEMIA patients experienced 670 index events (318 INV, 352 CON) and 203 recurrent events (102 INV, 101 CON). A single primary event was observed in 9.8% of INV and 10.8% of CON patients while ≥2 primary events were observed in 2.5% and 2.8%, respectively. Patients with recurrent events were older; had more frequent hypertension, diabetes, prior MI, or cerebrovascular disease; and had more multivessel CAD. The average number of primary endpoint events per 100 patients over 4 years was 18.2 in INV [95% confidence interval (CI) 15.8–20.9] and 19.7 in CON (95% CI 17.5–22.2), difference −1.5 (95% CI −5.0 to 2.0, P = 0.398). Comparable results were obtained when all-cause death was substituted for cardiovascular death and when stroke was added as an event. Conclusions In stable CAD patients with moderate or severe myocardial ischaemia enrolled in ISCHEMIA, an initial INV treatment strategy did not prevent either net recurrent events or net total events more effectively than an initial CON strategy. Clinical trial registration ISCHEMIA ClinicalTrials.gov number, NCT01471522, https://clinicaltrials.gov/ct2/show/NCT01471522.