Comparison of the correlation of creatinine- and cystatin C–Based estimated GFR and their differences with new-onset heart failure in a community-based population with type 2 diabetes

Abstract

This study aimed to investigate the impact of different estimated glomerular filtration rate (eGFR) values like cystatin C-based eGFR (eGFRcys), creatinine-based eGFR (eGFRcr), and their difference (eGFRdiff; eGFRcys -eGFRcr), on the incidence of heart failure (HF) in patients with type 2 diabetes(T2D). Being a prospective cohort study, it included 7,967 patients with T2D who underwent serum creatinine and cystatin C tests as part of the Kailuan Group’s 6th annual health examination (2016). Subsequently, eGFRcys, eGFRcr, and eGFRdiff were calculated. Patients were categorized into three groups: negative (<-15 mL/min/1.73 m2), midrange (-15 to 15 mL/min/1.73 m2), and positive (> 15 mL/min/1.73 m2) eGFRdiff groups, respectively. Furthermore, the relationship between the various eGFR measurements and new-onset HF were studied using Cox proportional hazards regression, and the potential improvement in predictive capability was evaluated by adding these eGFR metrics to established HF risk models. Among 7967 participants with mean age of 60.51 years, there were 20.92% women and 79.08% men. At baseline, eGFRcys and eGFRcr values differed by more than 15 mL/min/1.73m2 in 41.3% of participants. During a median follow-up period of 3.76 years, there were 172 (2.16%) new HF cases and 517 (6.49%) all-cause deaths. The cumulative incidence of HF in the midrange, negative, and positive eGFRdiff groups was 1.74%, 4.10%, and 0.61%, respectively (p < 0.001). In multivariable adjusted models, participants in the negative eGFRdiff group had higher risk of HF compared with the midrange eGFRdiff group (HR, 2.15; 95% CI, 1.57–2.94). Conversely, participants in the positive eGFRdiff group had lower risk for HF (HR, 0.40; 95% CI, 0.17–0.93). And each 15 mL/min/ 1.73 m2 higher eGFRdiff was associated with 34% (HR, 0.66; 95% CI, 0.58 − 0.47)lower risk of incident HF. The predictive capacity for HF risk in diabetic individuals was enhanced by adding eGFRcys or eGFRdiff to established HF risk models, with eGFRcys showing more significant additional predictive value. These findings suggest that large differences between eGFRcys and eGFRcr were common in community-based population with T2D. Different eGFR metrics can independently predict HF incidence in patients with T2D. Additionally, metrics like eGFRcys and eGFRdiff provide significant predictive value for HF risks beyond traditional risk factors, with eGFRcys showing more pronounced benefits in such cases.