Abstract 4368897: Albuminuria Drives Type 2 Diabetes-Related Atrial Fibrillation: an ACCORD substudy

Background: Patients with type 2 diabetes (T2D) are at elevated risk for atrial fibrillation (AF), potentially driven by glycemic variability, oxidative stress, and inflammation-induced atrial remodeling. Albuminuria, an early marker of renal microvascular disease in T2D, may also signal increased AF risk, though this association remains underexplored. Hypotheses: We hypothesized that baseline albuminuria would identify individuals with T2D at increased risk for AF in the Action to Control Cardiovascular Risk in Type 2 Diabetes (ACCORD) trial. We also explored whether random allocation to intensive glucose, blood pressure, or lipid control in ACCORD modified any observed associations between albuminuria and AF. Methods: In this post hoc analysis of ACCORD, T2D patients at high cardiovascular risk without overt AF were stratified by baseline urine albumin-to-creatinine ratio ([ACR] <30 vs ≥30 mg/g). The primary end point was time to incident AF ascertained from biannual study electrocardiograms. Adjusted Cox models were used to test for associations between ACR ≥ 30 mg/g (vs. <30 mg/g) and spline-modelled log(ACR) and incident AF. We examined effects of intensive glucose lowering, blood pressure control, and fenofibrate therapy on incident AF, stratified by baseline ACR above or below 30 mg/g. Results: Among 9,694 eligible patients, 32.3% had ACR ≥30 mg/g. Over a median follow-up of 4.9 years, 127 (1.3%) developed AF, with incidence rates of 0.22 and 0.42 per 100 person-years for ACR <30 vs. ≥30 mg/g, respectively ( Figure 1 ). In models adjusted for age, race, sex, and randomized treatment, elevated ACR (≥30 mg/g) associated with an increased risk of AF ( Table ). This risk was substantially unchanged after adjustment for potential confounders such as hypertension and body mass index ( Table ). Baseline log(ACR) was linearly associated with AF risk ( Figure 2 ). Random allocation to intensive glucose or blood pressure control did not prevent incident AF, an effect that did not vary by baseline ACR category (each P interaction =NS ). Fenofibrate did not reduce AF risk overall, but its effect differed by ACR category with reduced AF risk when ACR <30 mg/g (HR 0.53; 95% CI, 0.27–1.02) and increased risk when ACR ≥30 mg/g (HR 1.85; 95% CI, 0.90–3.77; P interaction =0.010). Conclusion: Albuminuria independently associates with incident AF in high-risk T2D patients. Whether modifying albuminuria alters AF risk is worthy of further exploration.