The Heart and Estrogen/progestin Replacement Study (HERS) found no overall effect of 4.1 years of therapy with estrogen plus progestin for secondary prevention of coronary heart disease in postmenopausal women. However, within the overall null effect, a 50% increase in cardiovascular events was seen in the first year, followed by fewer events after 2 years of treatment in the hormone therapy group than in the placebo group. Understanding the cause of this pattern of early increase and late reduction in risk is key to interpreting the HERS results and reconciling them with the large number of observational and other studies of the cardiovascular effects of estrogen. There are two possibilities. One is that the HERS regimen of estrogen plus progestin has no effect on risk for heart disease, and the pattern of changing risk over time is simply the result of chance or confounding. The other is that the pattern of early increase and late reduction in risk is due to real but opposing effects of this regimen. Several lines of evidence support each possibility. Attrition of a susceptible cohort of women uniquely at risk for a cardiovascular complication from hormone therapy coupled with a gradually progressive beneficial effect due to lipid lowering and other factors is a promising potential explanation. The HERS results remind us of the need for clinical trials to evaluate both the benefits and risks of new therapies. They also illustrate how much more we need to know about the cardiovascular effects of hormone replacement therapy.