Effect of empiric antibiotics against Pseudomonas aeruginosa on mortality in hospitalized patients: a systematic review and meta-analysis—authors’ response

We appreciate the thoughtful comments by Dr Kumar et al.¹ about our systematic review² on the effect of empirical antibiotics against Pseudomonas aeruginosa on mortality in hospitalized patients. Dr Kumar et al. suggest that the manuscript would have been improved if we had performed subgroup analyses including study risk of bias and infection severity, and if we had reported prediction intervals in addition to confidence intervals. We did, fortunately, include both subgroups in our subgroup analysis, which can be found in Table S6 of the systematic review article¹ (studies with higher risk of bias or restricted to sicker patients such as ICU patients reported somewhat greater benefit for empirical antibiotics against Pseudomonas). The 95% prediction interval for the primary outcome of adjusted OR was 0.15–1.03 compared with the 95% CI of 0.32–0.50. As Dr Kumar et al. point out, the prediction interval is wider because it accounts for the between-study variability in the analysis.

Funding

This research was supported in part by the Veterans' Affairs Clinical Science Research and Development (VA CSRD grant no. IK2 CX001981 to AC).

Transparency declarations

None to declare.

References