Background We investigated whether the greater increased risk of ische
mic heart disease mortality associated with diabetes among women compa
red with men could be explained by their more pronounced lipoprotein a
bnormalities. Methods and Results Seventy-six men and 45 women with di
abetes and 327 men and 496 women without diabetes were followed for an
average of 16 years in a population-based study. Cox proportional haz
ards models were used to determine the relative hazard of ischemic hea
rt disease mortality for changes in lipoprotein subfractions after adj
ustment for age, hypertension, obesity, smoking, exercise, alcohol con
sumption, and estrogen use (among women). The relative hazard of ische
mic heart disease mortality among diabetic women was 1.76 (P=.10) for
a 10-mg/dL decrement in high-density lipoprotein cholesterol (HDL-C) a
nd 3.13 (P=.01) for a 1-U increment in log, very-low-density lipoprote
in cholesterol (VLDL-C). The risk of ischemic heart disease mortality
among diabetic women relative to nondiabetic women for an HDL-C level
of 50 mg/dL and a log, VLDL-C of 3 (about 20 mg/dL) were 4.1 and 3.4,
respectively (P<.05). These lipoprotein changes were not associated wi
th ischemic heart disease mortality among men or among nondiabetic wom
en. Conclusions Excess ischemic heart disease mortality among diabetic
women is partially explained by deleterious levels of HDL-C and VLDL-
C. HDL-C levels of less than or equal to 50 mg/dL and VLDL-C levels of
greater than or equal to 20 mg/dL appear to predict ischemic heart di
sease mortality among these women and may help identify women who woul
d benefit most from intervention.