Abstract 4365831: GP73 Associates Causally with Incident Heart Failure through Insulin Resistance in Community-Dwelling Adults: The ARIC Study

Background: High fat diet induces GP73 secretion from the liver, heart and adipose tissue. GP73 promotes metabolic dysfunction and insulin resistance in preclinical studies, but whether GP73’s metabolic effects mediate its association with incident HF remains unclear. Objective: We sought to determine whether metabolic dysfunction mediates the association between GP73 and incident HF in community-dwelling adults. Methods: We performed causal mediation analyses within the ongoing, longitudinal, observational, Atherosclerosis Risk in Communities (ARIC) study. GP73 levels were measured at ARIC Visit 2 using an aptamer-based assay validated against mass spectrometry (Pearson R = 0.84). The potential metabolic mediators were assessed at Visit 4 (median of 6 years after Visit 2) and included diabetes, Homeostasis Model Assessment of Insulin Resistance (HOMA-IR), the five components of the metabolic syndrome and urine albumin-to-creatinine ratio > 30 mg/g. Incident heart failure hospitalizations after Visit 4 were identified through previously reported ARIC event surveillance methods. Causal mediation analysis used the Imai-Keele-Tingley counterfactual framework. We tested the potential causal effects of GP73 on metabolic risk factors using Mendelian randomization. Results: The ARIC cohort mean age was 57 years (66% women and 20% Black). ARIC participants with GP73 above the median were more likely to have diabetes or the metabolic syndrome and had higher fasting plasma glucose, higher HOMA-IR and greater waist circumference. Metabolic risk factors that significantly mediated the association between GP73 and incident HF included HOMA-IR (33%; P=0.04), waist circumference (23%; P<0.001), systolic blood pressure (9%; P<0.001), urine albumin-to-creatinine ratio > 30 mg/g (7%; P<0.001) and HDL cholesterol (3%; P<0.001). Mendelian randomization analysis provided evidence of causal effects of GP73 on fasting pro-insulin levels, HbA1c levels, type 2 diabetes and heart failure. Reverse mediation analysis with GP73 as the mediator and reverse Mendelian randomization with GP73 as the outcome did not support causal effects of metabolic risk factors on GP73 levels. Conclusions: GP73 associates with the development of incident heart failure in community-dwelling adults through its effects on insulin resistance, adiposity and vascular dysfunction.