Temporal Patterns of Ventricular Tachyarrhythmia Recurrences Follow a Weibull Distribution

• Published in: Journal of cardiovascular electrophysiology 2005-02, Vol.16 (2), p.181-185
• Main Authors: WOOD, MARK A., GUNDERSON, BRUCE, XIA, AMY, ZHOU, XIAOHONG, PADMANABHAN, VASANT, ELLENBOGEN, KENNETH A.
• Format: Article
• Language: English
• Subjects: Aged; Cluster Analysis; Defibrillators, Implantable; Female; Humans; Male; Models, Statistical; Recurrence; Statistical Distributions; Tachycardia, Ventricular - physiopathology; Tachycardia, Ventricular - therapy; temporal pattern; Time Factors; ventricular tachyarrhythmias; ventricular tachycardia storm; Weibull distribution
• ISSN: 1045-3873; 1540-8167
• DOI: 10.1046/j.1540-8167.2005.40121.x

Introduction: The objective of this study was to test whether the temporal patterns of ventricular tachyarrhythmia recurrences in patients with implantable cardioverter‐defibrillator (ICD) follow a random or a clustered distribution. Methods: Data analysis was conducted using the Medtronic (Minneapolis, MN) Gem DR database of 521 ICD patients. Patients with ≥3 sustained ventricular tachyarrhythmia detections that resulted in shock or antitachycardia pacing therapies were studied. The times between consecutively treated ICD detections for each patient were compared to an exponential model of random recurrences and a Weibull model for clustered recurrences. Results: Seventy‐one patients had ≥ 3VT episodes during follow‐up of 131 ± 86 days. A total of 2347 VT episodes were recorded (33 ± 65 episodes/patient, median 10 episodes/patient). Patient age was 66 ± 13 years, 78% male, 83% coronary artery disease, ejection fraction 31 ± 11%, and 63% were taking antiarrhythmic drugs. By the Kolmogorov‐Smirnov goodness‐of‐fit test, 38 of 71 patients (53.5%) showed that the pattern of detections differed from an exponential model (P < 0.01 for each patient and the proportion of patients was similar to chance at P = 0.65). In contrast, only 11 out of 71 patients (15.5%) showed that the pattern differed from the Weibull model (P < 0.01 for each patient). The proportion of patients fitting the Weibull model was significantly greater than chance and was greater that the proportion fitting the exponential model (both P < 0.001). The time interval between consecutive detections was less than 1 hour for 78% of all 2347 detections. The proportion of all 521 patients with ≥2, ≥3, ≥ 4, ≥6, ≥8, and ≥10 ICD detections in a 24‐hour period was 10.5%, 9.5%, 8.1%, 7.0%, 6.3%, and 5.2%, respectively. Conclusion: In most patients with ≥3 ICD detections, the recurrence pattern of treated ventricular tachyarrhythmia detections are clustered and can be described by a Weibull distribution. The proportion of patients with multiple detections in a 24‐hour period declines in a linear fashion as the number of events in 24 hours increases from 2 to 10 events.