Exercise-induced troponin release in patients with atrial fibrillation: magnitude and determinants in the NEXAF trial
H Wallin, K E Johnson, A Berg Sellevold, M Flasnes, I Haagenrud Langmo, J P Loennechen, B M Nes, J M LetnesAbstract
Background
Exercise-induced elevations in cardiac troponin are frequently observed and well described in healthy individuals, but determinants of this response in patients with atrial fibrillation (AF) remain incompletely defined.
Purpose
We aimed to identify independent physiological and structural determinants of exercise-induced high-sensitivity cardiac troponin I (cTnI) release in AF patients studied in sinus rhythm.
Methods
We included 184 participants (27–79 years) from the Norwegian Exercise in Atrial Fibrillation Trial (NEXAF). All underwent symptom-limited cardiopulmonary exercise testing (CPET) and resting transthoracic echocardiography while in sinus rhythm. Blood samples were obtained before, 5 minutes, 3 hours, and 24 hours after CPET. Temporal changes were assessed using the Friedman test with post-hoc comparisons. Multivariable linear regression models were constructed with peak post-exercise cTnI as the dependent variable, adjusting for baseline cTnI, age, sex, estimated glomerular filtration rate, and AF burden during the preceding 24 hours. cTnI concentrations were natural log-transformed (ln cTnI) due to right-skewed distribution. CPET-derived measures of cardiac workload, effort, aerobic capacity, and ventilatory efficiency, as well as echocardiographic indices of atrial and ventricular structure and function, were first explored in univariable analyses and subsequently entered into multivariable models.
Results
cTnI increased significantly following exercise, with incomplete normalization at 24 hours. Median (IQR) cTnI concentrations were 4.5 (3.4–7.2) ng/L at rest, 5.1 (3.8–8.4) ng/L at 5 minutes, 9.8 (6.3–14.8) ng/L at 3 hours, and 5.7 (3.9–9.6) ng/L at 24 hours (all p<0.001). In multivariable regression analyses, post-exercise cTnI concentrations (ln cTnI) were independently associated with peak heart rate during exercise (β=0.007, p=0.01) and left ventricular mass index (β=0.007, p=0.009), adjusting for confounders. In contrast, VO2peak, peak respiratory exchange ratio, diastolic indices including E/e′, and left atrial volume index were not associated with post-exercise cTnI concentrations. VE/VCO2 measured at a fixed submaximal workload showed a borderline association in univariable analyses (p=0.05) but did not remain significant after adjustment for peak heart rate and LVMI.
Conclusions
In AF patients studied in sinus rhythm, exercise-induced cTnI release is primarily related to acute cardiac workload and underlying myocardial structural remodeling rather than maximal aerobic capacity, resting diastolic function, or effort-related indices. These findings support the concept that troponin I elevations following exercise in AF reflect the interaction between myocardial workload and structural substrate, independent of active arrhythmia.