To the Editor.
—In an article in the November 2,1994, issue of The Journal, Dr Krumholz and colleagues1 prospectively followed 997 elderly subjects and reported no significant relationship between total cholesterol or high-density lipoprotein cholesterol (HDL-C) and cardiac events. Because the wide confidence intervals (CIs) included values greater than 1.0, this cannot be considered a negative study. In fact, the upper limit of their CI for adjusted CHD mortality for men and women (1.98) is a similar risk ratio to that reported previously in middle-aged men.2Part of the reason for the wide CIs is likely the size of the study and short duration of follow-up. Studies establishing the relationship between hypercholesterolemia and CHD in middle-aged individuals evaluated many more subjects for greater lengths of time. For example, the Multiple Risk Factor Intervention Trial3 (MRFIT) followed 356222 subjects over a 6-year period, and the Lipid Research Clinics