In reply
Thank you for the letters and the opportunity for us to respond. We agree that risk factors associated with the metabolic syndrome may have influenced our results despite adjusting for several related confounders. It would be useful to test our finding by status of metabolic syndrome. However, measures for fasting blood glucose, high-density lipoprotein cholesterol, and triglyceride values are available only in a subset of 3706 women in the Women's Health Initiative Clinical Trial Arm.1 Statistical power would have been compromised had we restricted our full analysis to this subset. As it stands, we cannot rule out that metabolic syndrome was present in many of the women taking statins and that this influenced results. We agree that this issue begs for a large, placebo-controlled, randomized clinical trial without our limitations to directly assess risk of incident diabetes with statin use in postmenopausal women and in other populations as well.