Assessment of endothelial function in atrial fibrillation: utility of peripheral arterial tonometry

Endothelial function is an independent predictor of adverse cardiovascular outcomes. The evaluation of endothelial function via changes in vessel diameter or blood flow may be inaccurate during atrial fibrillation (AF) because of non-uniform stroke volumes. Using peripheral arterial tonometry, 50 patients with AF (25 in AF, 25 in sinus rhythm) had digital pulse amplitudes assessed at baseline and during reactive hyperaemia. Hyperaemic responses were compared over varying measurement durations (5, 10 and 15beats; 30s; and 1-10min) to determine optimal measurement duration. Endothelial responses were significantly decreased (indicating endothelial dysfunction) in patients in AF compared with patients in sinus rhythm (1.48±0.60 vs 2.05±1.13, respectively; P=0.03). Beat-to-beat pulse amplitude was highly variable during AF; although coefficients of variation (CV) for short measurement durations were large, these decreased with longer measurement durations. Bland-Altman plots revealed that limits of agreement for short measurement durations were poor. Limits of agreement became consistently narrower when measurement durations of at least 1min were used. In contrast, limits of agreement and CV for short measurement durations during sinus rhythm were significantly narrower and smaller, respectively, than during AF over similar measurement durations. Pulse amplitudes are highly variable owing to the non-uniform stroke volumes in AF. Our results suggest that methods of determining endothelial function via vessel diameters or blood flow during reactive hyperaemia should use measurement durations of at least 1min to ensure accurate and reproducible results.