Background Elevated N-terminal pro-B-type natriuretic peptide (NT-proBNP) levels are common in patients with chronic kidney disease (CKD), but uniform thresholds derived from the general population may not accurately predict heart failure (HF) risk across stages of kidney function. This study aimed to determine whether thresholds specific to kidney function categories improve HF risk prediction in CKD.

Methods This retrospective cohort study used data from the China Renal Data System, including 18 261 patients with CKD without prior HF. Kidney function-specific thresholds for NT-proBNP were established based on estimated glomerular filtration rate (eGFR) categories, and associations with HF risk were assessed using multivariable Cox proportional hazard models. The predictive value of these thresholds was compared with a uniform threshold of 125 pg/mL using Net Reclassification Improvement (NRI).

Results Elevated NT-proBNP was observed in 67% of patients using the uniform threshold compared with 23% when using eGFR-specific thresholds. Optimal NT-proBNP thresholds increased with declining kidney function, reaching the highest level in stage 5 CKD (eGFR <15 mL/min/1.73 m²). eGFR-specific thresholds significantly improved HF risk prediction, with NRI gains of 19% to 55% across stages 1 to 5, while the uniform threshold added no predictive value for patients with stage 5 CKD.

Conclusions In patients with CKD, NT-proBNP levels must be interpreted in the context of kidney function, as eGFR-specific thresholds provide superior HF risk stratification. These findings support adopting kidney function-adjusted thresholds rather than a uniform cut-off to improve HF risk prediction.