main or three-vessel coronary artery disease, angina at rest and severe ischaemia on stress testing. The primary end point was a composite of death, myocardial re-infarction or NYHA class IV heart failure. The 4-year cumulative primary event rate was 17.2% in the PCI group and 15.6% in the medical treatment group (hazard ratio (HR) for death, reinfarction or heart failure in the PCI group as compared with the medical treatment group 1.16, 95% CI 0.92 to 1.45, p = 0.20). Rates of myocardial re-infarction (both fatal and non-fatal) were 7.0% and 5.3%, respectively (HR 1.36, 95% CI 0.92 to 2.00, p = 0.13). Only 6 re-infarctions (0.6%) were related to the assigned PCI procedure. Rates of NYHA class IV heart failure (4.4% vs 4.5%) and death (9.1% vs 9.4%) were similar. There was no interaction between treatment effect and any of the subgroup variables (age, sex, race, infarct-related artery, ejection fraction, diabetes mellitus, Killip class and the time from myocardial infarction to randomisation). Although this study demonstrated a high procedural success rate with PCI and sustained vessel patency, it failed to demonstrate clinical benefit over an approximately 3 year follow-up period with respect to death, reinfarction or heart failure. In other words, a patient whose artery was opened mechanically .12 h after an infarct did not have an improved outcome when compared with the individual treated with medical therapy alone. Statistical bias may influence outcome of MI trials c When examining prognosis in post-myocardial infarction (MI) trials, what is the significance of the statistical analytical method used? Stukel and colleagues from Canada reanalysed data from a cohort of 122 124 elderly patients aged 65-84 years, all of whom were hospitalised between 1994 and 1995 and were candidates for cardiac catheterisation. Patients were then followed up for 7 years to assess the association between long-term survival and cardiac catheterisation within 30 days of hospital admission. Four main analytical methods were compared: multivariable model risk adjustment, propensity score risk adjustment, propensity-based matching and instrumental variable analysis. After adjustment for prognostic factors using standard statistical adjustment methods, cardiac catheterisation was associated with a 50% relative decrease in mortality, giving an adjusted relative risk (RR) of 0.51. By contrast, instrumental variable analysis-which controls for hidden as well as overt bias-suggested a 16% relative decrease in mortality (RR 0.84). Therefore, in this study, estimates of the observational association of cardiac catheterisation with long-term mortality from MI were highly sensitive to the analytical method used. Instrumental variable analysis may produce less biased estimates of treatment effects, but is more suited to answering policy questions rather than specific clinical questions. Ventricular non-compaction and double-outlet right ventricle of the tetralogy of Fallot type with doubly committed subarterial VSD are rare conditions in iso...