Long-term, postmenopausal use of hormone replacement therapy (HRT) appears to increase breast cancer risk. Whether the effect of HRT use on risk of breast cancer varies among histological types of invasive carcinoma is unknown. To determine associations between HRT use and risk of ductal carcinoma in situ (DCIS), invasive carcinoma with favorable histology, and invasive ductal or lobular carcinoma. Prospective cohort study whose participants were followed up continuously for 11 years from January 1986 to December 1996. Iowa Women's Health Study, a population-based random sample of postmenopausal women aged 55 to 69 years in 1986. A total of 1520 incident breast cancer cases occurred in the at-risk cohort of 37 105 women. Multivariate-adjusted relative risk (RR) of tumor-specific breast cancers associated with duration of ever, current, or past HRT use. Duration of ever HRT use was associated with risk of invasive carcinoma with a favorable histology, with an RR of 1.81 (95% confidence interval [CI], 1.07-3.07) for those who used HRT 5 or fewer years vs an RR of 2.65 (95% CI, 1.34-5.23) for those who used HRT for more than 5 years (P for trend = .005) after adjustment for age and other breast cancer risk factors. There was no association between ever HRT use and the incidence of DCIS or invasive ductal or lobular carcinoma. Among current hormone users, after adjusting for age and other breast cancer risk factors, the RRs (95% CIs) of invasive carcinoma with a favorable histology were 4.42 (2.00-9.76) and 2.63 (1.18-5.89) for 5 or fewer years of use and for more than 5 years of use, respectively. Risk of invasive ductal or lobular carcinoma was associated with current use (< or =5 years) of HRT with an RR of 1.38 (95% CI, 1.03-1.85). Exposure to HRT was associated most strongly with an increased risk of invasive breast cancer with a favorable prognosis. These data add important clinical information for assessing the risks and benefits of HRT use.