In Reply: Dr Oh and colleagues suggest adjusting for electrocardiographic LVH in our mortality analyses. When this variable was added as a covariate, we observed no change in the adjusted hazard ratio for mortality (1.08 per standard deviation increment in PR interval; 95% confidence interval, 1.02-1.13; P = .005). Although electrocardiographic LVH is a predictor of mortality, it was not significantly associated with PR interval in our data.
In response to Dr Doraiswamy, we note that 89% of the pacemakers were implanted for a primary indication of high-grade AV block or sinus node dysfunction. The other 11% were attributed to a variety of indications, including postoperative changes or myocardial infarction, with high-grade AV block or sinus node dysfunction often noted as a secondary indication. Analyses examining the relationship of specific conduction abnormalities with outcomes were limited by inadequate power. With regard to nodal-blocking agents, we excluded individuals using medications at baseline that affect cardiac conduction times other than β-blockers or calcium-channel blockers. Overall, the prevalence of PR interval prolongation in the setting of nodal agents was very low (0.2% of the sample), a consequence of the generally healthy study cohort. The association of baseline PR interval with mortality in a subgroup analysis limited to those with a PR interval less than 200 milliseconds showed similar results (adjusted hazard ratio, 1.05; 95% confidence interval, 0.99-1.12; P = .10), but power was attenuated by the narrower range of PR interval values in this subgroup. Because our study sample was predominantly white and of European descent, we agree that further studies are warranted in more diverse populations.