Chronic Kidney Disease and Heart Failure: A Burden of Proof Meta-analysis

Abstract

Chronic kidney disease (CKD), defined by prolonged reduction in estimated glomerular filtration rate (eGFR), is an established risk factor for cardiovascular disease; however, the quantified magnitude of risk of reduced renal function on incident heart failure (HF) remains poorly characterized across the full exposure range of CKD severity. This study applies the Institute for Health Metrics and Evaluation’s (IHME) Burden of Proof (BoP) meta-analytic methodology to quantify the conservative excess risk of incident HF across the continuum of eGFR. A systematic review was conducted in PubMed (January 1980 - August 2025) following PRISMA 2020 guidelines, yielding 12 prospective cohort studies for inclusion in the primary meta-analytic pipeline from 733 records identified. Dose-response meta-analysis was performed using the BoP pipeline, which employs a Bayesian regularized spline approach to estimate a non-linear risk function. The meta-regression yielded an inverse, non-linear risk relationship between eGFR and incident HF, with the greatest change in risk occurring at low levels of eGFR and attenuation toward the null at higher levels. The exposure-averaged Burden of Proof Risk Function (BPRF) was 1.11, reflecting an average of 11% excess risk of incident HF under the most conservative interpretation of the evidence, yielding a Risk-Outcome Score of approximately 0.10 and a two-star rating. This is the first application of the BoP methodology to the CKD–incident HF risk-outcome pair, with results indicating a consistent, modest association between reduced eGFR and increased risk of incident HF, with the greatest risk elevation occurring below the clinical CKD threshold of eGFR 60. These findings suggest that HF screening efforts may benefit from consideration of renal function decline prior to formal CKD diagnosis, and further large-scale cohort studies across diverse populations are needed to refine the dose-response relationship and its implications for clinical risk stratification.