Abstract 4147548: Usefulness of Baseline NT-proBNP Non-Response Score and Health Status Measures in Assessing Treatment Responses in Heart Failure with Reduced Ejection Fraction

Background: The NT-proBNP nonresponse score (NNRS) previously derived from the PROTECT and BATTLESCARRED studies was proposed (Gaggin et al, Congest Heart Fail 2013), with the goal to identify patients with chronic heart failure (HF) who would not likely to response to outpatient medical treatment. Hypothesis: The NNRS can be used with health status measures (Kansas City Cardiomyopathy Questionnaire Overall Summary Score [KCCQ-OSS]) for assessing treatment response and prognostication in patients with chronic HFrEF. Aims: To validate the prognostic value of NNRS in comparison with KCCQ-OSS. Method: Data on the GUIDE-IT trial were used to derive the NNRS based on 4 predictors and their integeric score (IS) including baseline NT-proBNP (< 1,000, 1,000-5,000, and > 5,000 pg/mL; IS = 0, 2, 3, respectively), heart rate < 60 beats per min (IS = 1), NYHA class III or IV (IS = 1), and history of atrial fibrillation (IS = 1). Multivariable logistic regression analysis was used to determine the predictive value of the NNRS and KCCQ-OSS greater than the median values for NT-proBNP response with adjustment for age, sex, body mass index, comorbidities, baseline creatinine and NT-proBNP levels. Kaplan-Meier curves and log rank test were used to assess survival from CV mortality or HF hospitalization across patients with high and low NNRS and KCCQ-OSS. Results: Among 877 patients (median NNRS=3 [IQR 2-4], and KCCQ-OSS=59 [IQR 40-75]), 252 (28.7%) patients had NT-proBNP >1,000 pg/mL at 12 months, and 321 (36.6%) patients had CV mortality or HF hospitalization during median follow up of 300 days. The NNRS and KCCQ-OSS correlated poorly ( Figure 1A ). After adjusting for covariates, higher NNRS (adjusted OR 1.99, 95% CI 1.02-3.90, P 0.044) and KCCQ-OSS (adjusted OR 2.14, 95% CI 1.33-3.45, P 0.002) were associated with a higher risk of lack of NT-proBNP response at 12 months. Both NNRS and KCCQ-OSS can be used complimentarily for prognostication (log rank P<0.001, Figure 1B ). Conclusions: The NNRS and KCCQ-OSS were associated with persistently elevated NT-proBNP 12 months independently of baseline NT-proBNP levels and can be used complimentarily for prognostication in patients with HFrEF.