Figure 10

Subgroup Analysis. Hazard ratios (black squares), 95% CIs (horizontal lines), P values for the interaction between the treatment effect and any subgroup variable, and cumulative estimated 4-year event rates for the primary outcome (death from any cause, nonfatal reinfarction, or NYHA class IV heart failure requiring hospitalization or a stay in a short-stay unit) for PCI versus medical therapy for the specified subgroups are shown. Age, sex, race or ethnic group, the location of the infarct-related artery, the ejection fraction, and the time from the index myocardial infarction (MI) to randomization were prespecified. Race was self-reported. Diabetes and the highest Killip class during the index MI were not prespecified for the subgroup analysis. Originally, the cutoff point for age was 70 years, but early during the trial monitoring and before any analyses were performed, it was changed to 65 years because of insufficient numbers of patients older than 70. There was no significant interaction between treatment and subgroup variable as defined according to the prespecified value for interaction (P < 0.01). The use of a cutoff of 40% rather than the prespecified 50% for the ejection fraction did not alter the results. There was no interaction for the presence or absence of ST-segment elevation, Q-wave loss, or R-wave loss. LAD denotes left anterior descending artery. (Hochman et al. NEJM Dec 7, 2006 p2405)