Prognostic Factors Associated with Extubation Failure in Acutely Brain-Injured Patients: A Systematic Review and Meta-Analysis∗

Objective: Extubation failure in brain-injured patients is associated with increased morbidity. Our objective was to systematically review prognostic factors associated with extubation failure in acutely brain-injured adult patients receiving invasive ventilation in an ICU. Data Sources: MEDLINE, Embase, and Cochrane Central were searched from inception to January 31, 2022. Study Selection: Two reviewers independently screened citations and selected English-language cohort studies and randomized trials examining the association of prognostic factors with extubation failure. Studies were considered if they included greater than or equal to 80% adult patients with acute brain injury admitted to the ICU and mechanically ventilated for greater than or equal to 24 hours. Data Extraction: Two reviewers extracted data on population, prognostic factors, extubation outcomes, and risk of bias (using the quality in prognostic factors tool). Data Synthesis: In the primary analysis, adjusted odds ratios (aOR) for each prognostic factor were pooled using random-effects models. Certainty of evidence was assessed using Grading of Recommendations Assessment, Development and Evaluation. The search identified 7,626 citations, of which 21 studies met selection criteria. Moderate-certainty evidence suggested increased risk of extubation failure with older age (aOR, 3.0 for upper vs lower tertile; 95% CI, 1.78-5.07) and longer duration of mechanical ventilation (aOR, 3.47 for upper vs lower tertile; 95% CI, 1.68-7.19). Presence of cough (aOR, 0.40; 95% CI, 0.28-0.57) and intact swallow (aOR, 0.34; 95% CI, 0.21-0.54) probably decreased risk of extubation failure (moderate certainty). Associations of other factors with extubation failure were informed by low or very low certainty evidence. Conclusions: Patient age, duration of mechanical ventilation, and airway reflexes were associated with extubation failure in brain-injured patients with moderate certainty. Future studies are needed to determine the optimal application of these variables in clinical practice