Pre-randomization NT-proBNP trajectory and outcomes in the VICTOR trial

Background In the VICTOR trial, enrollment was enriched using screening N-terminal pro B-type natriuretic peptide (NT-proBNP) thresholds of 600-6000 ng/L (900-6000 ng/L if in atrial fibrillation). NTproBNP levels, however, may have fluctuated between screening and randomization. We evaluated the impact of different NT-proBNP trajectories between screening and randomization on outcomes in the VICTOR trial. Methods The 6,105 participants in the VICTOR trial were categorized by screening-to-randomization NT-proBNP trajectory into three groups: "Stable" (staying within 600-6000 ng/L), "Drop" (falling below 600 or 900 ng/L in AFib), or "Rise" (rising above 6000 ng/L). Primary, secondary, and exploratory endpoints were assessed using Cox proportional hazard models with covariates of stratification factors defined by baseline NYHA Class, treatment, NT-proBNP trajectory subgroup, and treatment-by-subgroup interaction. A trend test was performed using a separate Cox proportional hazard model in which NT-proBNP subgroup was coded as a numeric covariate. Results By the time of randomization, NT-proBNP levels remained stable in 4,888 (80%) participants, dropped below the eligibility threshold in 832 (13.6%), and rose above 6000 ng/L in 172 (2.8%) participants. For the primary composite endpoint of cardiovascular (CV) death or hospitalization for heart failure (HHF), or for CV death or all-cause death, no significant differences were observed across the “Drop,” “Stable,” and “Rise” subgroups (hazard ratio [HR] 0.92, 95% confidence interval [CI] 0.57-1.50; HR 0.95 [0.83-1.08]; and HR 0.63 [0.39-1.01]; p-trend=0.26 for composite endpoint; HR 0.75 [0.40-1.41]; HR 0.86 [0.72-1.02]; and HR 0.48, [0.25-0.94]; p-trend=0.36 for CV death; and HR 0.78 [0.45-1.33]; HR 0.85 [0.73-0.99]; and HR 0.65 [0.38-1.13]; p-trend=0.66 for all-cause death, respectively). No differences across the three subgroups were observed for both the composite of all-cause death or any HF event (i.e., HFH, urgent HF visit, or oral diuretic intensification) and the composite of CV death or any HF event (Figure). Conclusion In this post hoc analysis of the VICTOR trial, no differences on outcomes were observed across three subgroups of NT-proBNP trajectory between screening and randomization. The estimates for the “Rise” subgroup are based on very limited data.