Obesity duration and the risk of myocardial infarction: insights from the National Institutes of Health All of Us Research Program

Background While obesity is clearly implicated in the development of cardiometabolic diseases, there is considerable heterogeneity between individuals with comparable degrees of adiposity. Few studies have assessed the potential contribution of the duration of exposure to excess adiposity ("obesity duration") to this heterogeneity and have done so using largely historical and racially homogeneous cohorts. Atherosclerotic cardiovascular disease (ASCVD) has been inconsistently associated with obesity, possibly due to limitation of cross-sectional measure of obesity severity as the criterion for obesity in many studies. Purpose We aimed to use linked-electronic health record (EHR) data to identify patients at obesity onset [body mass index (BMI) ≥ 30kg/m²] and determine obesity duration. We then evaluated whether cumulative obesity exposure was independently associated with incident myocardial infarction (MI). Methods We used linked-EHR data from All of Us Research Program (AoU) participants identifying all individuals aged 18-65 years at first encounter with ≥ 1 subsequent BMI measurement over follow-up. We extracted BMI measurements and excluded any individuals with conditions that reduce BMI specificity (e.g., pregnancy, bariatric surgery). We then identified participants who first became obese following first EHR contact and followed them to incident MI (using established ICD-code methods in AoU) or last EHR encounter. We estimated each patient’s obesity duration and developed an EHR-specific formulation of obesity duration to account for varying follow-up time among participants – estimated cumulative proportion of time spent obese. To assess a relationship between cumulative obesity exposure and the hazard of MI, we used extended Cox regression models, adjusting for age at cohort entry, sex, race/ethnicity, income, and educational attainment level, with BMI and age at obesity onset as time-dependent covariates. Results Of 184,689 AoU participants, the onset of obesity was able to be identified specifically in 12,416 (6.7%) [71.2% female, median age at entry 38 (27, 51) yrs; Table 1]. As a proportion of follow-up time in the EHR, newly obese participants spent an estimated 50% of follow-up time obese. Among these AoU participants, 267 participants suffered an MI. A 10% increase in the estimated cumulative proportion of follow-up time spent obese was associated with significantly higher risk of MI with scant attenuation with full adjustment (HR = 1.15, 95% CI: 1.10, 1.21; Figure 1). Conclusion In this contemporary, diverse patient population, we demonstrate that cumulative obesity exposure is an independent risk factor for incident MI, supporting that obesity duration is an under-appreciated factor in obesity-associated ASCVD risk. This observation may motivate earlier intervention to mitigate cardiometabolic complications of obesity.