Aims: The influence of permanent atrial fibrillation on exercise tolerance and cardio-respiratory function during exercise in heart failure (HF) is unknown.
Methods and results: We retrospectively compared the results of 942 cardiopulmonary exercise tests, performed consecutively at seven Italian laboratories, in HF patients with atrial fibrillation (n = 180) and sinus rhythm (n = 762). By multivariable logistic regression analysis, peak VO(2) (OR 0.376, 95% CI 0.240-0.588, P < 0.0001), O(2)pulse (VO(2)/heart rate, HR) (OR 0.236, 95% CI 0.152-0.366, P < 0.0001), VCO(2) (OR 3.97, 95% CI 2.163-7.287, P < 0.0001), and ventilation (OR 1.38, 95% CI 1.045-1.821, P = 0.0231) were independently associated with atrial fibrillation. Anaerobic threshold (AT) was identified in 132 of 180 (73%) atrial fibrillation and in 649 of 762 (85%) sinus rhythm patients (P = 0.0002). By multivariable logistic regression analysis, only peak VO(2) (OR 0.214, 95% CI 0.155-0.296, P < 0.0001) was independently associated with unidentified AT. At AT, atrial fibrillation HF patients had higher HR (P < 0.0001) and higher VO(2) (P < 0.001) compared with sinus rhythm HF patients. Among AT variables, by multivariable logistic regression analysis, only HR was an independent predictor of atrial fibrillation.
Conclusion: In HF patients with permanent atrial fibrillation, exercise performance is reduced as reflected by reduced peak VO(2). The finding of unidentified AT is associated with a poor performance. In atrial fibrillation patients, VO(2) is higher at AT whereas lower at peak. This last observation raises uncertainties about the use of AT data to define performance and prognosis of HF patients with atrial fibrillation.