Comparative prognostic value of CKD risk strata in HFpEF/HFmrEF: insights from TOPCAT

Abstract Chronic kidney disease (CKD) worsens prognosis in heart failure with preserved or mildly reduced ejection fraction (HFpEF/HFmrEF). Estimated glomerular filtration rate (eGFR) and urine albumin-to-creatinine ratio (UACR) capture distinct aspects of kidney disease and are independently linked to HF progression. The KDIGO framework recommended combination of both, but its incremental prognostic value in HFpEF/HFmrEF has not been systematically evaluated. We compared KDIGO risk strata against conventional renal metrics for major adverse outcomes. This post-hoc analysis of the TOPCAT trial categorized 1173 patients with HF with LVEF ≥ 45% (encompassing HFpEF and HFmrEF) into 2024 KDIGO risk strata. The associations between renal metrics (KDIGO classification, CKD stage, UACR, and continuous eGFR) and a primary composite endpoint of cardiovascular death, aborted cardiac arrest and heart failure hospitalization were assessed using Lasso‑Cox regression with 10‑fold cross‑validation, which effectively selected the most relevant variables and avoided overfitting. Model discrimination and incremental predictive value were evaluated via C-statistic, Brier score, IDI and NRI. Among 1173 patients with HFpEF/HFmrEF, the distribution across KDIGO risk categories was: Low 29.5% ( n = 346), Moderate 33.1% ( n = 388), High 23.3% ( n = 273), and Very High 14.1% ( n = 166). A graded increase in risk was observed across advancing KDIGO categories, with the Very High-risk group exhibiting a substantially elevated hazard for the primary endpoint (fully adjusted HR 1.980; 95% CI [1.260, 3.109]; p = 0.003) and HF hospitalization (HR 3.039; 95% CI [1.975, 4.652]; p < 0.001). When added to the basic clinical model, KDIGO stratification demonstrated significant incremental predictive value for the primary endpoint (IDI: 0.004, p = 0.032; NRI: 0.068, p = 0.032), with particularly consistent improvements for heart failure hospitalization. The KDIGO risk classification provided prognostic information at least comparable to individual or simply unweighted renal metrics. Although its statistical incremental gains were modest, its clinical value lies in offering a convenient, guideline-aligned, single integrated score for routine bedside use and raising more attention to isolated UACR elevation.