BACKGROUND: Air pollution may increase the incidence of ventricular cardiac arrhythmias. We investigated this in patients with implantable cardioverter defibrillators attending London clinics. METHOD: We explored associations between dates of activation of defibrillators and daily concentrations of various metrics of particulate matter and of pollutant gases at lags from 0 to 5 days, using a fixed-stratum case-crossover analysis controlling for confounding factors. RESULTS: Over an average of 1200 days of observation, 705 patients experienced 5462 activation days. Of 11 pollutants considered, we found positive associations with particle sulfate, particulate matter with aerodynamic diameter less than 10 microm and less than 2.5 microm, ozone, and sulfur dioxide. Only the association for particle sulfate was not easily explainable by chance (for 1 microg/m, lag 0-1 day, odds ratio = 1.025 [95% confidence interval = 1.003 to 1.047]). There was little or no evidence of associations with markers of primary vehicle emissions (particle number concentration, black smoke, nitrogen oxides, and carbon monoxide). There was little evidence of interactions with clinical factors such as ischemic heart disease, frequency of activation, or cardiac drugs. CONCLUSION: Overall there was little evidence of an association between air pollution and activation of implantable cardioverter defibrillators. The pollutants with positive associations tended to be those of secondary origin with a regional distribution, rather than primary pollutants emitted from transport sources.