Prospective association of serum adipocyte fatty acid‐binding protein with heart failure hospitalization in diabetes

AIMS: Adipocyte fatty acid‐binding protein (AFABP) is associated with cardiovascular diseases in type 2 diabetes. Whether circulating AFABP levels are associated with the risk of heart failure (HF) in type 2 diabetes remains undefined. We investigated the prospective association of circulating AFABP levels with incident HF hospitalization in type 2 diabetes, and its relationship to the use of sodium glucose co‐transporter 2 inhibitors (SGLT2i) which reduce HF risk. METHODS AND RESULTS: Baseline serum AFABP level was measured in 3322 Chinese participants without known history of cardiovascular diseases or hospitalization for HF, recruited from the Hong Kong West Diabetes Registry. Its association with incident HF hospitalization was evaluated using multivariable Cox regression analysis. Use of SGLT2i was included as a time‐dependent covariate. Among these 3322 participants (52.9% men; mean age 60.0 ± 12.6), 176 (5.3%) developed HF hospitalization over a median follow‐up of 8 years. Seven hundred and thirty‐one (22%) were started on SGLT2i during the study period (empagliflozin 55.1%, dapagliflozin 44.2%, canagliflozin 0.4%, and ertugliflozin 0.3%). Serum AFABP levels were significantly higher in participants who developed HF hospitalization than those who did not (men: 14.8 vs. 8.3 ng/mL; women: 21.5 vs. 14.6 ng/mL; all: 18.6 vs. 10.9 ng/mL, P < 0.001). In multivariable Cox regression analysis, baseline serum AFABP level was significantly associated with incident HF hospitalization [hazard ratio (HR) 1.38, 95% confidence interval (CI) 1.06–1.80, P = 0.019] independent of the use of SGLT2i, in a model also consisting of age; sex; body mass index; smoking status; duration of diabetes; hypertension, dyslipidaemia; atrial fibrillation; presence of chronic kidney disease and albuminuria; glycated haemoglobin and high‐sensitivity C‐reactive protein levels; and use of metformin, insulin, aspirin, furosemide, and beta‐blockers at baseline. High cumulative defined daily dose (cDDD) of SGLT2i was protective of incident HF hospitalization (HR 0.10, 95% CI 0.01–0.68, P = 0.019). The addition of circulating AFABP level to a clinical model of conventional HF risk factors provided significant improvement in the category‐free net reclassification index (11.5%, 95% CI 1.6–22.1, P = 0.02) and integrated discrimination improvement (0.3%, 95% CI 0.1–1.7, P = 0.04). A dose‐dependent reduction in cumulative incidence of HF hospitalization in response to SGLT2i, based on cDDD, was more clearly observed in participants with a higher baseline AFABP level above the sex‐specific median (P for trend <0.01). CONCLUSIONS: Circulating AFABP level is independently associated with incident HF hospitalization in type 2 diabetes and is potentially helpful in risk stratification for the prevention of HF hospitalization.