Abstract
BACKGROUND: Acute respiratory distress syndrome (ARDS) patients with different lung morphology have distinct pulmonary mechanical dysfunction and outcomes. Whether lung morphology impacts the association between ventilatory variables and mortality remains unclear. Moreover, the impact of a novel combined ventilator variable [(4×DP) + RR] on morality in ARDS patients needs external validation. METHODS: We obtained data from the Chinese Database in Intensive Care (CDIC), which included adult ARDS patients who received invasive mechanical ventilation for at least 24 h. Patients were further classified into two groups based on lung morphology (focal and non-focal). Ventilatory variables were collected longitudinally within the first four days of ventilation. The primary outcome was 28-day mortality. Extended Cox regression models were employed to explore the interaction between lung morphology and longitudinal ventilatory variables on mortality. FINDINGS: We included 396 ARDS patients with different lung morphology (64.1% non-focal). The overall 28-day mortality was 34.4%. Patients with non-focal lung morphology have more severe and persistent pulmonary mechanical dysfunction and higher mortality than those with focal lung morphology. Time-varying driving pressure (DP) was more significantly associated with 28-day mortality in patients with non-focal lung morphology compared to focal lung morphology patients (P for interaction = 0.0039). The impact of DP on mortality was more significant than that of respiratory rate (RR) only in patients with non-focal lung morphology. The hazard ratio (HR) of mortality for [(4×DP) + RR] was significant in patients with non-focal lung morphology (HR 1.036, 95% CI 1.027-1.045), not in patients with focal lung morphology (HR 1.019, 95% CI 0.999-1.039). INTERPRETATION: The association between ventilator variables and mortality varied among patients with different lung morphology. [(4×DP) + RR] was only associated with mortality in patients with non-focal lung morphology. Further validation is needed.