Seasonal variation of mortality in the Antiarrhythmics Versus Implantable Defibrillators (AVID) study registry

Abstract

We postulated that the pattern of death would be nonrandom with respect to temporal variables. Previous studies have demonstrated increased sudden death is associated with periods of relative stress, and overall mortality is associated with temporal variables. In the Antiarrhythmics Versus Implantable Defibrillators (AVID) registry, vital status was obtained for 4,450 patients (who had a recent episode of sustained ventricular arrhythmias or unexplained syncope and inducible ventricular tachycardia) through the National Death Index Service as of December 31, 1997 (follow-up 25.5 +/- 13.7 months). Mortality was not associated with the day of the week or with holidays but was associated with season (P = .033). Seasonal variation was present both in northern and southern sites. Mortality was higher during the winter months compared to the remaining months (111.2% in winter vs 96.5% in other months, P = .036). In addition, increased mortality was associated with a high-risk season variable defined (prior to evaluation of treatment arm associations) as spring in the north and winter in the south (P < .001). The hazard ratio for death associated with this "high-risk season" measured 1.25 (P = .001) compared to the other seasons in each region. A test of interaction between "high-risk" season and implantable cardioverter-defibrillator (ICD) status suggested that the group with ICDs experienced reduced mortality during the "high-risk season" compared to the group without ICDs (P = .047). Mortality was higher in the winter months and in the respective regional "high-risk" seasons. Furthermore, seasonal variation in mortality may have been due to variation in sudden arrhythmic death, and associated increases in mortality were reduced by ICD therapy.