RESUMO
An association between gastroesophageal reflux disease (GERD) and diabetes mellitus (DM) has been reported. Studies have not been population-based and have failed to include a representative sample of African American subjects. The aim of the study was to determine if DM is independently associated with GERD among urban African Americans. Single-center, population-based survey utilizing a complex, stratified sampling design. To obtain a simple random sample of the entire African American community, targeted survey zones and hand-delivered invitations were identified. Participating subjects had to be self-described African American, age ≥18. Surveys were completed at a computer terminal assisted by a research coordinator. Four hundred nineteen subjects (weighted sample size of 21 264 [20 888-23 930]). GERD prevalence was 23.7% (95% confidence interval [CI] 23.2-23.9). GERD prevalence was 41.5 % in those with DM versus 20.6 % for those without (P < 0.001). Those with GERD had DM longer but had lower glycohemoglobin levels. The prevalence of ≥2 DM comorbidities was higher in those with GERD (odds ratio [OR] = 2.06; 95% CI 1.71-2.48). In the final model, age >40, DM, increasing body mass index, harmful drinking, and increasing smoking dependence were independently associated with GERD. For DM, there was significant effect modification by gender. In males, the risk was (OR = 4.63; 95% CI 3.96-5.40), while in females, the risk was markedly attenuated (OR = 1.79; 95% CI 1.61-2.00). Among urban African Americans, there is an independent association between DM and GERD that appears to be stronger in men. More information is needed to understand this association.