Résumé : In an experimental model, variable and intermittent contact force (CF) resulted in a significant decrease in lesion volume. In humans, variability of CF during pulmonary vein isolation has not been characterized. Methods and Results-In 20 consecutive patients undergoing CF-guided circumferential pulmonary vein isolation, 914 radiofrequency applications (530 in sinus rhythm and 384 in atrial fibrillation) were analyzed. The variability of the 60% CF range (CF60%) was 17 ± 9.6 g. Hundred seventy-one (19%) applications were delivered with constant, 717 (78%) with variable, and 26 (3%) with intermittent CF. The mean CF and force-time integral were significantly higher during applications with variable than with intermittent or constant CF. There was no significant difference in CF variability, CF60% variability, and force-time integral between applications delivered in sinus rhythm and atrial fibrillation. The main reasons for CF variability were systolo-diastolic heart movement (29%) and respiration (27%). In 10 additional patients, during adenosine-induced atrioventricular block, the minimum CF significantly increased at 19 sites (5.3 ± 4.4 versus 13.4 ± 5.9 g; P < 0.001) and at 16 sites intermittent or variable CF became constant. At only 1 site systolo-diastolic movement remained the main reason for variable CF. Conclusions-CF during pulmonary vein isolation remains highly variable despite efforts to optimize contact. CF and CF parameters were similar during sinus rhythm and atrial fibrillation. The main reasons for CF variability are systolodiastolic heart movement and respiration. The systolo-diastolic peaks and nadirs of CF are because of ventricular contractions at the large majority of pulmonary vein isolation sites.