Abstract A121: The association of menstrual and reproductive factors with the risk of upper gastrointestinal tract cancers in the NIH-AARP cohort

Abstract

Abstract A121 Upper gastrointestinal tract cancers (UGI), including cancers of the esophagus, head and neck, and stomach have high worldwide incidence and mortality. Incidence rates are consistently higher among men than women; the male/female ratios are 2:1 for cancers of the stomach and 7:1 for cancers of the larynx. Strong environmental risk factors have been identified for these cancer sites, but do not appear to explain differences in incidence rates between men and women. Differences in sex hormones in men vs. women and among women of different ages might affect UGI risk. In nearly 10 studies of menstrual and reproductive factors with stomach cancer risk in women, significant inverse associations with older age at menopause or use of menopausal hormone therapy (MHT) were observed in some, but not all studies. Only a few studies have investigated the association of these factors with squamous cancers of the esophagus and head and neck. We prospectively investigated the association of menstrual and reproductive factors with UGI cancer risk in 201,506 women of the NIH-AARP Diet and Health cohort study. We used Cox proportional hazard models adjusted for age, alcohol intake, cigarette smoking, body mass index, physical activity, education, and total energy intake and report hazard ratios (HRs) and 95% confidence intervals. During 1,463,551 person years of follow-up from 1995/1996-2003, 162 women were diagnosed with adenocarcinomas (ACs; esophagus and stomach) and 353 women were diagnosed with squamous cell carcinomas (SCCs; oral cavity, pharynx, larynx, and esophagus). We found no significant associations between age at menopause, age at menarche, parity, or age at first birth and AC risk. Ever- versus never-use of MHT showed a borderline non-significant inverse association with adenocarcinoma (HR=0.74, 95% CI 0.54-1.02) in age adjusted models. After multivariate adjustment, the association was attenuated (HR=0.81, 95% CI, 0.59-1.12). For SCCs, we found a significant inverse association with older age at menopause but not with age at menarche, parity, or age at first birth. The HR for age at menopause of >55 years (vs. <45 years) was 0.53 (95% CI 0.28-1.01; p-value for trend=0.019). We also observed an inverse association with ever- versus never-use of MHT (HR=0.77, 95% CI 0.62-0.96). We further examined relations with the type of MHT stratified by hysterectomy status, using data from a subsequent questionnaire (1996-1997) completed by 127,385 women. In 51,515 women with a reported hysterectomy at baseline, the HRs and 95% CIs for <5 years of estrogen-plus-progestin MHT, relative to never MHT users, were 0.74 (0.43-1.27) for ACs (N=38) and 0.51 (0.18-1.46) for SCCs (N=80), respectively; for ≥5 years of use, the HRs and 95% CIs were 0.47 (0.19-1.16) and 0.28 (0.13-0.58), respectively. In 74,372 women with intact uteri, we found no associations between estrogen MHT and ACs (N=49) or SCCs (N=130). Our results suggest that estrogen-plus-progestin therapy is associated with reduced risk of UGI tract cancers, supporting the involvement of sex hormones in the etiology of these cancers. Future studies are needed to replicate these results. Citation Information: Cancer Prev Res 2008;1(7 Suppl):A121.