Dr. Singh offers a plausible physiologic explanation why obese women might be less likely to be taking HRT. In our multivariate analysis, we used a waist-to-hip ratio of more than 0.85 as a marker of obesity because central adiposity is strongly correlated with risk for coronary heart disease in women (1) and might be an important confounder of HRT use. However, women with an increased waist-to-hip ratio were not statistically less likely to use HRT than other women.

As Dr. Singh infers, most of the diabetic women in our sample were overweight (78% had a waist-to-hip ratio >0.85 and 68% had a body mass index >27). In additional analyses, compared with nondiabetic women, diabetic women with a normal waist-to-hip ratio and those with a waist-to-hip ratio greater than 0.85 were less likely to use HRT (odds ratio, 0.14 [95% CI, 0.02 to 0.94] and 0.19 [CI, 0.05 to 0.68], respectively). The small number of diabetic patients with a normal waist-to-hip ratio in our sample, reflected in the wide CI, limits our ability to draw definite conclusions for these women as a separate group; however, our data do not support a difference in the use of HRT between lean diabetic women and obese diabetic women.

We have considered two other possible explanations for our finding of lower rates of HRT among diabetic women. First, these women may be taking several medications already, and they or their physicians may wish to avoid additional medications unless absolutely necessary. They may also have more issues to address during medical visits, and time constraints may preclude thorough discussions about the benefits and risks of HRT (2). In addition, physicians may be concerned that estrogen may aggravate glucose intolerance. In one recent study, HRT was associated with small improvements in fasting glucose levels but slightly decreased postprandial glucose tolerance (3). It remains unclear whether our finding of lower rates of HRT among diabetic women represents "undertreatment." A preliminary observational report from the Nurses' Health Study (4) suggested that diabetic women who use HRT have about half the risk for coronary heart disease compared with diabetic nonusers. However, a randomized trial, the Heart and Estrogen/progestin Replacement Study (5), found no benefit of HRT for the secondary prevention of coronary heart disease, even in a subgroup analysis of the approximately 19% of women in the cohort who were diabetic.