RESUMO
INTRODUCTION: We aimed to analyze intensive care unit (ICU)-acquired pneumonia according to 7 definitions, estimating associated hospital mortality. METHODS: This cohort study was nested within an international randomized trial, evaluating the effect of probiotics on ICU-acquired pneumonia in 2650 mechanically ventilated adults. Each clinically suspected pneumonia was adjudicated by two physicians blinded to allocation and center. The primary outcome was ventilator-associated pneumonia (VAP) informed by ventilation for ≥2 days, new, progressive or persistent infiltrate plus 2 of: temperature > 38 °C or < 36 °C; leukopenia (<3 × 10(Fernando et al., 20206)/L) or leukocytosis (>10 × 10(Fernando et al., 20206)/L); and purulent sputum. We also used 6 other definitions estimating the risk of hospital mortality. RESULTS: The frequency of ICU-acquired pneumonia varied by definition: the trial primary outcome VAP (21.6%), Clinical Pulmonary Infection Score (CPIS) (24.9%), American College Chest Physicians (ACCP) (25.0%), International Sepsis Forum (ISF) (24.4%), Reducing Oxidative Stress Study (REDOXS) (17.6%), Centers for Disease Control (CDC) (7.8%), and invasively microbiologically confirmed (1.9%). The trial primary outcome VAP (HR 1.31 [1.08, 1.60]), ISF (HR 1.32 [1.09,1.60]), CPIS (HR 1.30 [1.08,1.58]) and ACCP definitions (HR 1.22 [1.00,1.47]) were associated with hospital mortality. CONCLUSIONS: Rates of ICU-acquired pneumonia vary by definition and are associated with differential increased risk of death.