Background: The weathering hypothesis proposes that sexual minorities would have earlier onset of chronic disease like hypertension (HTN) due to chronic stress and discrimination. This hypothesis has not been fully studied as most prior work focused on prevalence analysis. Using electronic health records from a large health system in Illinois, we compared prevalence and incidence of HTN by sexual orientation. Methods: We used data from people 30-60 years old with sexual orientation data and no evidence of HIV at the time of reporting sexual orientation (index date). We then identified prevalent (before or on index date) and incident (≥1 day after index date) HTN cases using diagnosis codes, medications, and outpatient blood pressure (≥140/90 mmHg). We performed 1:4 propensity score matching of bisexual (bi) and gay/lesbian to straight adults (covariates: age, legal sex, raceðnicity, prior year utilization, year of first visit). We used logistic regression to calculate HTN prevalence ratios and Cox proportional hazards models to calculate hazard ratios. Subgroup analysis by legal sex (male/female) and age group (incidence only) was also performed. Results: The 157,334 eligible individuals (1.95% bi, 5.34% gay/lesbian) had crude HTN prevalences of bi-8.4%, gay/lesbian-11.2% and straight–10.2% (Image 1). Matching led to covariate balance. Compared to straight controls, bi (PR: 1.25, 95%CI: 1.15, 1.36) and gay/lesbian adults (PR: 1.22, 95%CI: 1.17, 1.26) had significantly higher HTN prevalence with slightly lower age at diagnosis. The prevalence remained higher in subgroup analyses by sex and using the ≥130/80 mmHg threshold (Images 2&3). Survival analysis showed bi adults had higher HTN risk (HR: 1.24, 95% CI: 1.07, 1.44) than straight controls with bi adults having lower mean diagnosis age. Stratified analyses show higher risk in males only. Sensitivity analysis with the ≥130/80 mmHg threshold showed significantly higher risk in bi adults, but subgroup analysis by sex only had significant results for females. No significant difference in HTN risk between gay/lesbian and straight controls were noted. Incident analysis by age showed higher risk in younger subgroups. (Images 2&3). Conclusion: Consistent with weathering, we found higher HTN prevalence in sexual minority adults and higher incident HTN risk in bisexual compared to matched straight adults. Confirmation with data from other health systems and community-based cohorts is needed.
Rivera et al. (Tue,) studied this question.