Does acute pyelonephritis skew outcomes in registrational trials of novel antibiotics for complicated urinary tract infections? A systematic review and meta-analysis.

Abstract

Acute pyelonephritis (AP) and complicated urinary tract infections (cUTIs) are pooled in registrational RCTs of new antibiotics. We assessed whether outcomes differ between subtypes and which component of the composite endpoint drives any difference. Across ten registrational RCTs we estimated the AP-versus-cUTI treatment effect on the composite primary endpoint (clinical cure plus microbiologic eradication at test-of-cure) using random-effects risk ratios, with risk-difference, odds-ratio and leave-one-out sensitivity analyses. Between-arm differences in the composite and its components were correlated across trials; operative definitions and stratified-reporting practices were extracted. From 5,971 patients (3,393 AP; 2,578 other cUTI), AP showed higher overall cure at TOC (pooled RR=1.12; 95% CI 1.03-1.23; I2=82%; 95% prediction interval 0.86-1.46). Between-arm differences in the composite correlated strongly with those in microbiologic eradication (r=0.98) but only weakly with those in clinical cure (r=0.30). In the single trial reporting components stratified by infection subtype (ADAPT-PO), the AP-versus-cUTI gap was larger in microbiologic eradication (+15.6 pp) than in clinical response (+3.5 pp); whether this pattern generalizes to the remaining nine trials could not be publicly tested. Risk of bias was overall some concerns in nine of ten trials, mostly for missing outcome data and selection of the reported result. AP achieved higher cure rates at TOC than other cUTIs. Heterogeneity in operative AP/cUTI definitions and limited stratified reporting of component outcomes reduce interstudy comparability. Future trials should harmonize definitions, prespecify stratification by infection subtype, and disclose clinical and microbiologic outcomes separately.