Sex/race categories, social determinants, and cardiovascular events in Brazilian adults with diabetes: a longitudinal analysis from ELSA-Brasil

Social determinants of health influence both the development of diabetes and its complications. However, evidence on how combined sex/race categories relate to cardiovascular risk among individuals with diabetes remains limited, particularly in middle-income countries. To investigate whether combined categories of sex and self-reported race/skin color are associated with incident cardiovascular events in Brazilian adults with diabetes. This prospective cohort study included adults with diabetes and no prior cardiovascular disease at baseline (2008–2010) from ELSA-Brasil. Participants were followed through subsequent study waves. The primary outcome was incident major cardiovascular events. The main exposure was a cross-classified sex/race variable (White men, Black men, White women, and Black women). Cox proportional hazards models were fitted using a hierarchical strategy, from minimally adjusted models to models additionally including socioeconomic and clinical variables. Variables such as HbA1c, hypertension, obesity, diabetes duration, chronic kidney disease, and albuminuria were considered potential downstream factors and were therefore interpreted cautiously. During follow-up, incident cardiovascular events were observed. In minimally adjusted analyses, differences across combined sex/race groups were observed but were attenuated after adjustment for socioeconomic and clinical factors. Lower educational attainment, longer diabetes duration, higher HbA1c, hypertension, and kidney disease markers were more consistently associated with cardiovascular events than the sex/race categories themselves. In this ELSA-Brasil cohort of adults with diabetes, combined sex/race categories were not statistically associated with incident cardiovascular events after accounting for measured socioeconomic and clinical variables. These findings should not be interpreted as evidence that intersectional or structural inequalities are unimportant; rather, they suggest that measurable socioeconomic and cardiometabolic pathways may partly explain the observed disparities. Limited event numbers, possible overadjustment, and the absence of direct measures of racism, discrimination, and healthcare barriers should be considered when interpreting the results.