Survival in Immunocompromised Patients Ultimately Requiring Invasive Mechanical Ventilation: A Pooled Individual Patient Data Analysis

Rationale: Acute respiratory failure (ARF) is associated with high mortality in immunocompromised patients, particularly when invasive mechanical ventilation is needed. Therefore, noninvasive oxygenation/ventilation strategies have been developed to avoid intubation, with uncertain impact on mortality, especially when intubation is delayed. Objectives: We sought to report trends of survival over time in immunocompromised patients receiving invasive mechanical ventilation. The impact of delayed intubation after failure of noninvasive strategies was also assessed. Methods: Systematic review and meta-analysis using individual data (IPD) of studies which focused on immunocompromised adult patients with ARF requiring invasive mechanical ventilation. Studies published in English were identified through PubMed, Web of science, and Cochrane Central (2008-2018). IPD were requested to corresponding authors for all identified studies. We used mixed-effect models to estimate the effect of delayed intubation on hospital mortality and described mortality rates over time. Measurements and main results: 11087 patients were included (24 studies, 3 controlled trials and 21 cohorts), of whom 7736 (74%) were intubated within 24h of ICU admission (early intubation). Crude mortality rate was 53.2%. Adjusted survivals improved over time (from 1995 to 2017, OR for hospital mortality per year: 0.96[0.95-0.97]). For each elapsed day between ICU admission and intubation, mortality was higher (OR:1.38[1.26-1.52], p<0.001). Early intubation was significantly associated with lower mortality (OR: 0.83[0.72-0.96]), regardless of initial oxygenation strategy. These results persisted after propensity score analysis (matched OR associated with delayed intubation :1.56[1.44-1.70]). Conclusion: In immunocompromised intubated patients, survival has improved over time. Time between ICU admission and intubation is a strong predictor of mortality, suggesting a detrimental effect of late initial oxygenation failure.