Response to letter to editor

We are pleased to respond to the thoughtful letter by Aryana et al.1 in regard to our recent publication in JCE.2 For clarity, note that although the analysis encompassed CIED implantations over a 28-month period, the outcome of interest (incidence of CIED infection) was limited to the 6-month post-implantation interval. We performed an additional analysis of infection incidence at 1-year postimplantation among patients for whom data were available. The authors are correct to point out that this report was based on retrospective, nonrandomized data, and we were careful to emphasize this throughout the paper. With regard to the implanting physicians, all were experienced and with active, high-volume practices. As we note, CIED infection rates among individual physicians in this group were not significantly different prior to the period encompassed by the study. With regard to procedural technique, as is detailed in the Methods section all implantations were performed in the same environment and using similar techniques and technologies; antibiotic practices are also discussed. In retrospective analyses such as this in which multiple variables are examined, imbalances are to be expected. With regard to those in the incidence of prior CIED infection and the use of temporary pacing, both were used as matching variables in the propensity analysis. As demonstrated by the low standardized differences between the Yes-AIGISRx and No-AIGISRx groups, the matching algorithm performed well in creating an analysis dataset comprising a list of patients with similar risk infection profiles.We also note that none of the patients in the No-AIGISRx group who suffered CIED infection had a prior history of infection, and the frequency of temporary pacing was similar among infected and uninfected patients (Table 3). As we state in the paper, no matter how carefully the data are parsed studies such as this are not capable of excluding bias, and thus should be considered as hypothesis generating