LAUSR

We appreciate the interest in our review of antibiotic duration in hospitalized patients. Drs. Sikkens and van Agtmael comment that drug pharmacokinetics can alter true treatment duration.1,2 Specifically, azithromycin has a long half-life in tissues.3 We did not consider pharmacokinetics in our prespecified protocol for study inclusion, nor require that studies compare the same drug between treatment groups. This is consistent with a systematic review of antibiotic duration in community-acquired pneumonia, which included 3 of the 4 studies comparing shortcourse azithromycin to a longer course of another antibiotic.4 Similarly, in a recent pilot study of antibiotic duration in bloodstream infections, only treatment duration was prespecified.5 We agree that the differing pharmacokinetics between drugs is a limitation to our findings. To assess whether the inclusion of studies using short-course azithromycin biased our conclusions, we performed an additional meta-analysis for clinical efficacy excluding the 4 studies that compared azithromycin with another drug. This subgroup included 9 trials comprising 1270 patients. The overall risk difference was 0.3% (95% CI −2.7%, 3.3%), consistent with the primary findings that short-course antibiotic treatment is non-inferior to long-course antibiotic treatment. None of these 4 studies examined mortality; thus, the meta-analyses for shortterm and long-term mortality are unaffected.