In 12 of 13 phase 3 and 4 comparative clinical trials, all-cause mortality was higher in the tigecycline group versus the comparator group. Study-level mortality risk differences were pooled using a random-effects meta-analysis. Statistical models evaluated the association between patient-level all-cause mortality and baseline factors using logistic regression, recursive partitioning [classification and regression tree (CART) analysis] and survival techniques. The estimated risk difference (tigecycline minus comparator) in all-cause mortality from the meta-analysis was 0.6% (95% confidence interval 0.1-1.2%). Statistical modelling identified baseline bacteraemia associated with mortality only in the tigecycline group. In patients with ventilator-associated pneumonia (VAP) and baseline bacteraemia, mortality was 50.0% (9/18) for tigecycline versus 7.7% (1/13) for the comparator group. Study-level and patient-level analyses have identified that patients in the hospital-acquired pneumonia trial, particularly those with VAP with baseline bacteraemia, were at a higher risk of clinical failure and mortality.
Copyright © 2013 Elsevier B.V. and the International Society of Chemotherapy. All rights reserved.