Exercise-induced NT-proBNP by heart failure with preserved ejection fraction score

Background Heart failure with preserved ejection fraction (HFpEF) is a common condition and expected to increase in prevalence due to demographic changes. Age, hypertension, obesity, and atrial fibrillation are some of the known risk factors. The ability of resting plasma levels of N-terminal pro-B-natriuretic peptide (NT-proBNP) to predict HFpEF has been questioned, especially in patients with obesity or comorbidities like kidney failure. Several diagnostic scores have been developed, including the HFA-PEFF score based on NT-proBNP levels and echocardiographic parameters. Stress testing is recommended in patients at intermediate risk of HFpEF, and exercise has also shown to increase levels of NT-proBNP in plasma. However, whether exercise-induced NT-proBNP responses may have a role in diagnosing HFpEF is unclear. Purpose We sought to evaluate whether the change in NT-proBNP between rest and post exercise is associated with the HFA-PEFF score in a population at high risk for HFpEF. Methods We included 127 participants aged 27 to 79 years from The Norwegian Exercise in Atrial Fibrillation Trial (NEXAF), a multicenter study evaluating the impact of exercise in patients with atrial fibrillation. All participants had an insertable cardiac monitor implanted, underwent cardiopulmonary exercise testing (CPET) and a resting transthoracic echocardiogram. Blood samples were drawn before, as well as five minutes and three hours after CPET. HFpEF was defined as an HFA-PEFF score ≥5. Associations between change in NT-proBNP, HFpEF and HFA-PEFF score were evaluated by logistic and linear regression analyses, adjusting for known confounders. Results NT-proBNP was available in 120, 123 and 110 subjects at baseline, five minutes and three hours after CPET, respectively. Mean (95% CI) resting NT-proBNP was 142 (113-171) ng/L, rising to 165 (133-197) ng/L five minutes and 170 (131-209) ng/L three hours after CPET. Individuals with HFpEF had an increase in NT-proBNP levels of 15 (7-33) ng/L higher than those without HFpEF five minutes after CPET (p<0.001). Three hours after CPET the average difference had decreased to 5 (2-18) ng/L (p=0.007). A one-point increase in the HFA-PEFF score was associated with an increase in plasma levels of NT-proBNP of 2.2 (1.7-2.9) ng/L (p<0.001) at five minutes (Figure 1) and 1.7 (1.2-2.6) ng/L (p=0.008) at three hours after CPET. These associations persisted after exclusion of NT-proBNP from the HFA-PEFF score, except for the three-hour value (p=0.087). The percentage change from baseline to three hours after CPET showed an inverse relation to HFA-PEFF score (5.8% lower relative increase per point increment in HFA-PEFF score, 95% CI 2.0-9.6, p=0.003), but not at five minutes (0.6%, 95% CI -2.3-1,1, p=0.46). Conclusion The absolute increase in NT-proBNP five minutes and three hours after CPET is significantly correlated with increaseing HFA-PEFF scores. However, the differences were modest indicating minimal clinical importance.