Combined Prognostic Value of Follow‐Up Ejection Fraction and Natriuretic Peptide Measurements in Heart Failure With Reduced Ejection Fraction Following Initiation of Pharmacotherapy

Figures

• 

  Figure 1. Strengthening the Reporting of Observational Studies in Epidemiology diagram for the overall study design. HFrEF indicates heart failure with reduced ejection fraction; ICD‐9, International Classification of Diseases, Ninth Revision; ICD‐10, International Classification of Diseases, Tenth Revision; I/E, inclusion/exclusion; LVEF, left ventricular ejection fraction; and NT‐proBNP, N‐terminal pro‐B‐type natriuretic peptide.

• 

  Figure 2. Distribution of NT‐proBNP levels based on LVEF strata. LVEF indicates left ventricular ejection fraction; and NT‐proBNP, N‐terminal pro‐B‐type natriuretic peptide.

• 

  Figure 3. Freedom from the composite outcome of HF hospitalization or all‐cause death based on NT‐proBNP levels. Freedom from the composite outcome based on NT‐proBNP levels is shown for the overall cohort (A), the LVEF ≤40% subcohort (B), the LVEF 41% to 49% subcohort (C), and the LVEF ≥50% subcohort (D). P<0.005 by log‐rank test for all plots. HF, heart failure; HFrEF, heart failure with reduced ejection fraction; LVEF indicates left ventricular ejection fraction; NT‐proBNP, and N‐terminal pro‐B‐type natriuretic peptide.

• 

  Figure 4. Adjusted 1‐year risk estimates of the composite outcome of HF hospitalization or all‐cause death based on different strata of LVEF and NT‐proBNP measurements. Risk estimates were computed using a multivariable Cox regression model that was adjusted for relevant covariates using a complete case approach (n=5811). The following covariates were adjusted for in the model: age, sex, relevant comorbidities (atrial fibrillation, anemia, coronary artery bypass graft, coronary artery disease, chronic obstructive pulmonary disease, chronic kidney disease, diabetes, hypertension, hyperlipidemia, left bundle‐branch block, myocardial infarction, obesity, peripheral vascular disease, stroke, supraventricular tachycardia, valvular heart disease, and ventricular arrythmia), BMI, mean arterial pressure, eGFR, and GDMT medications (beta blockers, ACE inhibitors, ARBs, ARNi, MRAs, and SGLT2 inhibitors). Risk estimates were generally computed at mean values for numeric covariates and mode values for categorical covariates, with the exception that full use of contemporary GDMT was assumed (ie, use of a beta blocker, ARNi, MRA, and SGLT2 inhibitor), despite few patients in the study cohort actually…

• 

  Figure 5. Adjusted 1‐year risk estimate of the composite outcome of HF hospitalization or all‐cause death based on continuous values for LVEF and NT‐proBNP measurements. LVEF and log‐transformed NT‐proBNP measurements were modeled as continuous variables using a generalized additive model incorporating thin plate regression splines. The model was adjusted for the following covariates using a complete‐case approach (n=5811): age, sex, relevant comorbidities (atrial fibrillation, anemia, coronary artery bypass graft, coronary artery disease, chronic obstructive pulmonary disease, chronic kidney disease, diabetes, hypertension, hyperlipidemia, left bundle‐branch block, myocardial infarction, obesity, peripheral vascular disease, stroke, supraventricular tachycardia, valvular heart disease, and ventricular arrythmia), BMI, mean arterial pressure, eGFR, and GDMT medications (beta blockers, ACE inhibitors, ARBs, ARNi, MRAs, and SGLT2 inhibitors). Risk estimates were generally computed at mean values for numeric covariates and mode values for categorical covariates, with the exception that full use of contemporary GDMT was assumed (ie, use of a beta blocker, ARNi, MRA, and SGLT2 inhibitor),…