Abstract 787: Estimating the influence of sex on cancer risk: An analysis in a large US prospective cohort study

Abstract

Abstract Historically in the United States, cancer incidence has been significantly higher in men than women. This disparity is often attributed to gender-related behavioral risk factors, such as smoking and drinking. The degree to which differences in age-adjusted cancer rates are due to sex (e.g., chromosomes, hormones) is unknown. We estimated the effect of sex on cancer risk by comparing cancer incidence at 21 shared sites between men and women while adjusting for known gender-related risk factors. We analyzed data from 122,826 women and 171,274 men enrolled in the NIH AARP Diet and Health Study (1995-2011), contributing 3.5 million person-years of follow-up. Cox proportional hazards models were used to estimate male-to-female (M:F) adjusted hazard ratios (aHR) and 95% confidence intervals (CI) for each of the cancers, controlling for relevant baseline cancer-specific demographic (age, race, marital status, education), lifestyle (self-reported health status, family history of cancer, diet, body mass index, physical activity), and behavioral factors (tobacco, alcohol use). We used the Peters-Belson method to partition the observed differences between the sexes in cancer incidence into the proportion of the difference explained by the modifiable environmental and behavioral exposures (explained disparity) and the remaining proportion that cannot be explained by these exposures (unexplained disparity). After adjustment for cancer-specific risk factors, risks remained significantly higher in men for nearly all sites. The highest M:F aHR was for esophageal adenocarcinoma (aHR: 10.59, 95% CI: 7.19, 15.61) followed by gastric cardia (aHR: 4.55, 95% CI: 3.28, 6.31) and larynx (aHR: 3.73, 95% CI: 2.84, 4.89) cancers. In contrast, men had a lower adjusted incidence of thyroid (aHR: 0.57, 95% CI: 0.48, 0.68) and gallbladder (aHR: 0.55, 95% CI: 0.37, 0.82) cancers compared with women. Using the Peters-Belson method we found that, if men had the same exposure distribution as the women (under a counterfactual), the male predominance would remain for 12 cancers (bladder, skin, lung, colon, kidney, rectum, esophageal adenocarcinoma, larynx, gastric cardia, esophageal squamous cell carcinoma, pancreas, and oropharynx). For 4 cancers (liver, gastric non-cardia, oral cavity, and anus), if men had the exposure distribution of women, they would have experienced a higher cancer incidence than what was observed. This elevated risk in men compared with women, after adjustment for risk factors, indicates potential innate sex differences across nearly all shared cancer sites. Funding: Intramural Research Program of the National Institutes of Health, National Cancer Institute, Division of Cancer Epidemiology and Genetics Citation Format: Sarah S. Jackson, Morgan A. Marks, Hormuzd Katki, Barry I. Graubard, Michael D. Cook, Anil Chaturvedi. Estimating the influence of sex on cancer risk: An analysis in a large US prospective cohort study [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 787.