A reappraisal of association between ventilator-associated events and mortality among critically ill patients using marginal structural model: multicenter observational study

Abstract

Purpose: Ventilator-associated events (VAEs), developed by the US Centers for Disease Control and Prevention (CDC), are considered a new medical quality indicator associated with poor outcomes. However, previous studies failed to adjust for changes in patient severity, leaving open the possibility that VAEs are just a surrogate for severity. This study aimed to reevaluate the association between VAEs and mortality, adjusting for changes in severity over time. Methods: This multicenter observational study was conducted in 18 ICUs in Japan between May 2020 and December 2022. Patients aged ≥ 12 years who received mechanical ventilation for ≥ 3 consecutive days were eligible. VAE was diagnosed according to the CDC definition. The primary outcome was 30-day in-hospital mortality, whose association with VAE was estimated by the inverse probability weighted proportional hazards model, with temporal changes in patient severity treated as time-dependent confounders. Results: Among 1,094 subjects, 106 VAEs (9.7%) were identified, giving an incidence rate of 10.0 per 1,000 ventilation days. VAEs were significantly associated with increased 30-day hospital and ICU mortality (HR 2.00; 95% CI 1.23–3.26 and HR 1.92; 95% CI 1.03–3.57), and longer hospital and ICU length of stay (HR 0.72 95% CI 0.54–0.97 and HR 0.47; 95% CI 0.23–0.96). The population attributable risk fraction of VAE-related mortality was 8.8% for in-hospital deaths and 8.2% for ICU deaths. Conclusions: VAEs were associated with an increased risk of mortality after adjustment for severity, as a time-dependent confounder.