Frequency and predictors of ventilator-associated pneumonia recurrence: a meta-analysis

Abstract

Large clinical series focusing on the risk factors associated with recurrence after the onset of an initial episode of ventilator-associated pneumonia (VAP) produced inconsistent results. A meta-analysis would be helpful to shed light on the issue. Our objective was to estimate the frequency of VAP recurrence and to identify risk factors associated with it. PubMed, Scopus, Current Contents, and references of retrieved articles were searched without language restrictions. Pooled odds ratios (ORs) and 95% confidence intervals (CIs) were calculated by using both the Mantel-Haenszel fixed-effect and the DerSimonian-Laird random effects models. The overall frequency of VAP recurrence in 969 patients of the seven eligible reports was 26.8%. Among the 20 evaluated risk factors, only acute lung injury/acute respiratory distress syndrome (OR, 1.76; 95% CI, 1.12-2.75) and shock (OR, 1.55; 95% CI, 1.01-2.41) at the day of diagnosis of the first VAP episode were found to be associated with VAP recurrence. There was also evidence, albeit inconsistent, that severity of illness at intensive care unit admission was associated with VAP recurrence. Recurrence involves almost one in four cases of VAP and is associated with acute lung injury/acute respiratory distress syndrome and shock, but not with first-episode causative pathogens. Recognition of these predictors may permit the timely implementation of measures to prevent recurrence of VAP.