Predictors of mortality in adult patients with ventilator-associated pneumonia: a meta-analysis.

Studies exploring predictors of mortality in patients with ventilator-associated pneumonia (VAP) produced conflicting results. The present work is a meta-analysis of studies that enrolled only patients with microbiologically confirmed VAP and reported on mortality. Potentially eligible reports were searched in PubMed, EMBASE, CINAHL, and HEALTHSTAR with no language restrictions. Twenty-six reports were included. Factors associated with mortality were malignancy (odds ratio [OR], 2.20; 95% confidence interval [CI], 1.10-4.40) at intensive care unit admission as well as inappropriate initial treatment (i.e., treatment either in vitro inactive against the causative bacteria or administered later than 24 h after diagnosis of VAP) (OR, 2.92; 95% CI, 2.01-4.22), bacteremia (OR, 2.07; 95% CI, 1.16-3.71), acute respiratory distress syndrome/acute lung injury (OR, 2.28; 95% CI, 1.24-4.21), shock (OR, 3.90; 95% CI, 2.31-6.61), sepsis (OR, 4.77; 95% CI, 2.22-10.25), disease severity, and sepsis-related organ failure score at the day of diagnosis of VAP. Isolation of nonfermenting gram-negative bacteria in general (OR, 1.71; 95% CI, 1.09-2.68) and Acinetobacter baumannii in specific (OR, 1.74; 95% CI, 1.02-2.96) was also associated with higher fatality. Intensive care unit admission caused by trauma, as opposed to other reasons, was linked to lower mortality (OR, 0.35; 95% CI, 0.22-0.57). These findings may help investigators to formulate appropriate predicting scores for patients with VAP and may further motivate clinicians to provide appropriate initial treatment and to manage sepsis and shock optimally in such patients.