Temporal patterns of paroxysmal atrial fibrillation following DDDR pacemaker implantation.

Paroxysmal atrial fibrillation (AF) episodes have been reported to be randomly distributed. However, because patients are not always symptomatic, it has been difficult to study temporal patterns of AF. Newer implantable pulse generators have data-logging capabilities that permit the detection and analysis of temporal patterns of AF. This study tested the hypothesis that AF episodes occur in clusters over time and that these episodes are not randomly distributed in individual patients. The date and time of 582 episodes of AF were recorded from the data logs of 16 patients with a Medtronic Thera DR followed 6 weeks and 6 and 12 months after pulse generator implant. The probability of AF recurrence and the interevent intervals between successive episodes of AF were fitted to monoexponential and Weibull distributions. A Weibull distribution best described the nonrandom distribution of AF for 67% of follow-up visits. Temporal clustering of AF (interevent intervals <24 hours) declined during follow-up (95 +/- 10%, 90 +/- 11%, and 74 +/- 28% at the 6-week and 6- and 12-month visits, respectively; p <0.05). The average duration of an episode of AF tended to increase over time (0.31 hour, 95% confidence intervals [CI] 0.17 to 0.58 hours; 0.36 hours, 95% CI 0. 17 to 0.78 hours; 0.65 hours, 95% CI 0.29 to 1.45 hours [p = 0.07] at the 6-week and 6- and 12-month visits, respectively). Paroxysmal AF recurrence is nonrandomly distributed over the long term in many patients. The temporal patterns of AF change over time after pacemaker implantation. This has implications for the selection of study end points in AF clinical trials.