The relation between baseline variables and the incidence of hard coronary events (or angina pectoris) was assessed using Cox’s proportional hazards regression analysis. Adjustment was made for the following factors known to be predictive of ischaemic heart disease: age; years of education; level of physical activity; waist circumference; cumulative tobacco consumption (pack years); systolic blood pressure; apolipoprotein A-1 concentration; apolipoprotein B concentration; and treatment for diabetes, dyslipidaemia, and hypertension. The validity of the proportional hazards ratios assumption for the covariates was verified. We tested the proportionality assumption using cumulative sums of martingale based residuals.29 Analyses were conducted first for each country separately and then for the whole cohort, after testing the homogeneity of relations between the two samples. All statistical analyses were carried out using the SAS statistical software package (SAS Institute, Cary, NC). All tests were considered significant at P=0.05.
