High body mass index (BMI) and large waist circumference are separately associated with increased coronary heart disease (CHD) risk but these measures are highly correlated. Their separate associations with incident CHD, cross-classifying one variable by the other, are less investigated in large-scale studies. We examined these associations in a large UK cohort (the Million Women Study), which is a prospective population-based study. We followed 496,225 women (mean age 60 years) with both waist circumference and BMI measurements who had no vascular disease or cancer. Adjusted relative risk and 20-year cumulative CHD incidence (first coronary hospitalization or death) from age 55 to 74 years were calculated using Cox regression. Plasma apolipoproteins were assayed in 6295 randomly selected participants. There were 10,998 incident coronary events after mean follow up of 5.1 years. Within each BMI category (<25, 25-29.9, ≥30 kg/m(2)), CHD risk increased with increasing waist circumference; within each waist circumference category (<70, 70-79.9, ≥79 cm), CHD risk increased with increasing BMI. The cumulative CHD incidence was lowest in women with BMI <25 kg/m(2) and waist circumference <70 cm, with 1 in 14 (95% confidence interval 1 in 12 to 16) women developing CHD in the 20 years from age 55 to 74 years, and highest in women with BMI ≥30 kg/m(2) and waist circumference ≥80 cm, with 1 in 8 (95% confidence interval 1 in 7 to 9) women developing CHD over the same period. Similar associations for apolipoprotein B to A1 ratio across adiposity categories were observed, particularly in non-obese women. Our conclusions were that both waist circumference and BMI are independently associated with incident CHD.