The effects of exercise on cardiovascular biomarkers in patients with chronic heart failure

Abstract

Background: Exercise training is recommended for chronic heart failure (HF) patients to improve functional status and reduce risk of adverse outcomes. Elevated plasma levels of amino-terminal pro-brain natriuretic peptide (NT-proBNP), high-sensitivity C-reactive protein (hs-CRP), and cardiac troponin T (cTnT) are associated with increased risk of adverse outcomes in this patient population. Whether exercise training leads to improvements in biomarkers and how such improvements relate to clinical outcomes are unclear.

Methods and results: Amino-terminal pro-brain natriuretic peptide, hs-CRP, and cTnT levels were assessed at baseline and 3 months in a cohort of 928 subjects from the HF-ACTION study, a randomized clinical trial of exercise training versus usual care in chronic HF patients with reduced left ventricular ejection fraction (<35%). Linear and logistic regressions were used to assess 3-month biomarker levels as a function of baseline value, treatment assignment (exercise training vs usual care), and volume of exercise. Linear regression and Cox proportional hazard modeling were used to evaluate the relations between changes in biomarker levels and clinical outcomes of interest that included change in peak oxygen consumption (peak VO2), hospitalizations, and mortality. Exercise training was not associated with significant changes in levels of NT-proBNP (P = .10), hs-CRP (P = .80), or detectable cTnT levels (P = .83) at 3 months. Controlling for baseline biomarker levels or volume of exercise did not alter these findings. Decreases in plasma concentrations of NT-proBNP, but not hs-CRP or cTnT, were associated with increases in peak VO2 (P < .001) at 3 months and decreased risk of hospitalizations or mortality (P ≤ .04), even after adjustment for a comprehensive set of known predictors.

Conclusions: Exercise training did not lead to meaningful changes in biomarkers of myocardial stress, inflammation, or necrosis in patients with chronic HF. Only improvements in NT-proBNP translated to reductions in peak VO2 and reduced risk of clinical events.

Figure 1

Baseline and 3-month biomarker levels according to treatment group. Scatter plots depicting the baseline (x-axis) and 3-month (y-axis) measurements of log-transformed NT-proBNP (A) and hs-CRP (B), according to treatment assignment. C, Frequency distributions of baseline (left) and 3-month (right) cTnT, according to treatment assignment.

Figure 1

Baseline and 3-month biomarker levels according to treatment group. Scatter plots depicting the baseline (x-axis) and 3-month (y-axis) measurements of log-transformed NT-proBNP (A) and hs-CRP (B), according to treatment assignment. C, Frequency distributions of baseline (left) and 3-month (right) cTnT, according to treatment assignment.

Figure 2

Baseline and 3-month biomarker levels according to amount of exercise. Median and intraquartile ranges of baseline and 3-month values for NT-proBNP and hs-CRP, according to amount of exercise. P values indicate whether levels between baseline and 3 months are statistically different.

Figure 3

Change in baseline and 3-month peak VO₂ according to biomarker improvement. Scatter plots depicting the baseline (x-axis) and 3 month (y-axis) measurements of log-transformed peak VO₂ according to levels of NT-proBNP (A), hs-CRP (B), and cTnT (C), grouped according to biomarker improvement versus no improvement at 3-month follow-up.

Figure 3

Change in baseline and 3-month peak VO₂ according to biomarker improvement. Scatter plots depicting the baseline (x-axis) and 3 month (y-axis) measurements of log-transformed peak VO₂ according to levels of NT-proBNP (A), hs-CRP (B), and cTnT (C), grouped according to biomarker improvement versus no improvement at 3-month follow-up.