Early predictors for infection recurrence and death in patients with ventilator-associated pneumonia

Abstract

Objective: Early recognition of predictors of unfavorable evolution of ventilator-associated pneumonia (VAP) might prompt therapeutic measures that might improve outcome. The objective of this study was to describe resolution of VAP variables and to determine early predictors of VAP recurrence and death.

Design and setting: Description of the natural course of VAP resolution and multivariable analyses of predictors of VAP recurrence and death by day 28 after VAP onset based on the 401 patients included in the PNEUMA trial, a multiple-center, randomized study comparing 8 vs. 15 days of antibiotics for microbiologically proven VAP. Every patient included in that trial had received appropriate empirical antibiotics.

Measurements and main results: By day 28 after VAP onset, 27% of patients had VAP recurrence and 18% had died. On day 8 after VAP onset, predictors of VAP recurrence included intensive care unit admission Simplified Acute Physiology Score II (odds ratio [OR], 1.02), radiologic score (OR, 1.17), temperature (OR, 1.34), nonfermenting Gram-negative bacilli (OR, 2.00) or methicillin-resistant Staphylococcus aureus (OR, 2.50) as pathogens responsible for VAP, and mechanical ventilation dependency (OR, 2.08). Day 8 predictors of 28-day death were age (OR, 1.06), female sex (OR, 2.30), Sepsis-Related Organ Failure Assessment score (OR, 1.26), and nonfermenting Gram-negative bacilli (OR, 2.83) as pathogens responsible for VAP. However, the duration of antimicrobial therapy (8 vs. 15 days) was not associated with any of the studied adverse outcomes.

Conclusions: For patients benefiting from appropriate empirical antibiotics for VAP, early predictors of infection recurrence or death included demographic characteristics, such as age or female sex, disease severity at VAP onset, nonfermenting Gram-negative bacilli or methicillin-resistant S. aureus as VAP-causative pathogens, prolonged mechanical ventilation dependency, persistent fever, and severity of lung injury. Future studies should attempt to determine whether specific diagnostic or therapeutic strategies could markedly improve VAP outcomes when early criteria for treatment failure are present.