Association of NT-proBNP and BNP With Future Clinical Outcomes in Patients With ESKD: A Systematic Review and Meta-analysis

Rationale & Objective: Use of brain natriuretic peptide (BNP) and N-terminal pro-BNP (NT-proBNP) for cardiovascular (CV) risk assessment in patients with end-stage kidney disease (ESKD) remains unclear. We examined the associations between different threshold elevations of these peptide levels and clinical outcomes in patients with ESKD. Study Design: Systematic review and meta-analysis. Setting & Study Populations: We searched MEDLINE and EMBASE (through September 2019) for observational studies of adults with ESKD (estimated glomerular filtration rate <= 15 mL/min/1.73 m(2) or receiving maintenance dialysis). Selection Criteria for Studies: Studies that reported NT-proBNP or BNP levels and future CV events, CV mortality, or all-cause mortality. Data Extraction: Cohort characteristics and measures of risk associated with study-specified peptide thresholds. Analytical Approach: Hazard ratios (HRs) for clinical outcomes associated with different NT-proBNP and BNP ranges were categorized into common thresholds and pooled using random-effects meta-analysis. Results: We identified 61 studies for inclusion in our review (19,688 people). 49 provided sufficient detail for inclusion in meta-analysis. Pooled unadjusted HRs for CV mortality were progressively greater for greater thresholds of NT-proBNP, from 1.45 (95% CI, 0.91-2.32) for levels > 2,000 pg/mL to 5.95 (95% CI, 4.23-8.37) for levels > 15,000 pg/mL. Risk for all-cause mortality was significantly higher at all NT-proBNP thresholds ranging from > 1,000 to > 20,000 pg/mL (HR range, 1.53-4.00). BNP levels > 550 pg/mL were associated with increased risk for CV mortality (HR, 2.54; 95% CI, 1.49-4.33), while the risks for all-cause mortality were 2.04 (95% CI, 0.82-5.12) at BNP levels > 100 pg/mL and 2.97 (95% CI, 2.21-3.98) at BNP levels > 550 pg/mL. Adjusted analyses demonstrated similarly greater risks for CV and all-cause mortality with greater NT-proBNP concentrations. Limitations: Incomplete outcome reporting and risk for outcome reporting bias. Estimation of risk for CV events for specific thresholds of both peptides were limited by poor precision. Conclusions: ESKD-specific NT-proBNP and BNP level thresholds of elevation are associated with increased risk for CV and all-cause mortality. This information may help guide interpretation of NT-proBNP and BNP levels in patients with ESKD.