Estrogen replacement therapy (ERT) increases a woman's risk of developing endometrial cancer approximately 120% for each 5 years of use. ERT increases a woman's risk of developing breast cancer approximately 10% for each 5 years of use. To reduce the greatly increased endometrial cancer risk, progestins have been added to ERT (estrogen-progestin replacement therapy; EPRT) for between 5 and 15 days (usually 7 or 10 days) per month in a sequential fashion (sequential EPRT; SEPRT) or with each dose of ERT (continuous-combined EPRT; CEPRT). We conducted two large case-control studies in postmenopausal women in Los Angeles to evaluate the effects of these changes on endometrial and breast cancer risks. As expected CEPRT was not associated with any increased risk of endometrial cancer. SEPRT with the progestin being given for 10 days per month also did not increase endometrial cancer risk. SEPRT with the progestin being given for 7 days per month did increase endometrial cancer risk with only a relatively slight reduction in risk compared to ERT effectively proportional to the reduction in the number of days of unopposed estrogen. The sharp contrast between the effects of 7 days and 10 days of progestin in SEPRT suggests that the extent of endometrial sloughing or of 'terminal' differentiation at the completion of the progestin phase may play a critical role in determining endometrial cancer risk. This may provide an explanation of why endometrial cancer risk increases so sharply with age in young women even in countries where obesity-associated anovulation is very uncommon; extended periods of unopposed estrogen is not an explanation but less than 10 days of an 'adequate' progesterone level may be. EPRT significantly increased the risk of breast cancer. EPRT was associated with an approximately 24% increase in risk for each 5 years of use; the effect was some 212-fold greater than the effect of ERT, which we had previously predicted on theoretical grounds. This effect could also be predicted from the results on mammographic densities seen in the PEPI randomized trial of different forms of hormone replacement therapy (HRT). In the PEPI trial EPRT increased mammographic densities to a much greater extent than ERT. Progestins need to be given to protect the endometrium. They need to be delivered to the endometrium in a manner that will have the least effect on the breast. This can be carried out by using a vaginal or direct endometrial route of administration. The vaginal route will provide adequate endometrial progestin levels with low blood levels so that the effects of the progestin on the breast should be small; with the direct endometrial route the blood progestin levels are even lower, and the effects of the progestin on the breast will be effectively zero. If this is unacceptable to a woman, then giving progestins by mouth (or transdermally) for 10 days every 3 to 4 months should provide satisfactory protection of the endometrium when used with standard-dose conjugated estrogen (CE). This regimen has much less effect on the breast than monthly SEPRT or CEPRT. Two clinical trials of 10 mg per day of MPA for 14 days every 3 months and 0.625 mg/day of CE have been published. Both studies suggest that this approach may be satisfactory in that the extent of hyperplasia was minimal. More studies of this approach are urgently needed.