Use of oral contraceptives and menopausal hormone therapy and risk of endometrial cancer stratified by parity in the aggregated sample of postmenopausal Caucasian women
Hazard ratios (HRs) and 95% confidence intervals (CI) estimated using Cox proportional hazards models with age as the underlying time scale with separate models fit to nulliparous and parous women. Models include all variables presented in Figures 2 and 3 as well as study, birth year, calendar year of entry, diabetes, and smoking status. Parous models further adjusted for number of births. P-values are presented for the heterogeneity between parous and nulliparous women based on a Wald type chi-square test. HRs for nulliparous women represented by white boxes. HRs for parous women represented black boxes.
3a. In NIH-AARP, never use includes <1 year
3b. In PLCO, duration of E + P use categorized as <6 years and ≥6 years. Analyses restricted to subset of women who reported never or current use from NIH-AARP or entry into PLCO and USRT after 1995. Follow-up censored at June 30, 2002.