Implanted cardioverter defibrillators (ICDs) monitor patients for episodes of cardiac arrhythmias and can initiate a therapeutic intervention to restore normal heart rhythm. These devices also record dates, times, and electrograms of these episodes. We examined the effects of air pollution on the incidence of arrhythmias in 195 cardiac patients with ICD devices in the Boston metropolitan area between July 1995 and July 2002. Gaseous air pollutant and meteorologic data were measured on essentially all days, fine particle mass on 80% of the days, and black carbon (BC) on 61% of the days. Date and time of detected arrhythmias, intracardiac electrograms, and therapeutic interventions were downloaded during the patients' regular follow-up visits every 3 months on average. A cardiac electrophysiologist reviewed electrograms recorded before, during, and after the arrhythmias and categorized them into ventricular and supraventricular events. Risk of arrhythmias associated with air pollution was estimated using logistic regression with adjustments for season, temperature, relative humidity, day of the week, and patient. We found increased relative risks of ventricular arrhythmias (VAs) associated with an increase in 2-day mean concentrations for all air pollutants considered, although these associations were not statistically significant. The relative risks of supraventricular arrhythmias (SVAs) increased in association with 2-day mean concentrations for all air pollutants, and this association was significant only for sulfur dioxide (SO2) at 4 ppb (odds ratio [OR] = 1.33; 95% confidence interval [CI] = 1.04, 1.70). The positive associations of VAs and SVAs with particulate matter less than 2.5microm in aerodynamic diameter (PM2.5; also referred to as fine particles), carbon monoxide (CO), nitrogen dioxide (NO2), BC, and SO2 suggest a link with motor vehicle pollutants. We explored patient characteristics that may have identified subjects susceptible to the effects of air pollution. The association of air pollution with SVAs was blunted by regularly prescribed beta-blockers. We found stronger associations of air pollution with VAs for episodes within 3 days of a previous arrhythmia, suggesting that VAs were triggered by air pollution episodes in combination with other factors that raised the patient's underlying risk. Although ICDs are specifically designed to monitor and treat only VAs, these results suggest that air pollution may trigger both VAs and SVAs.