Toward targeted assessment of inspiratory muscle weakness before extubation: a Pooled individual patient data analysis from multiple observational cohorts in critically ill ventilated patients

Bedside neuromuscular tests are commonly proposed to assess readiness for extubation in critically ill patients. However, when used in unselected populations, their ability to predict clinically relevant outcomes such as extubation failure has been inconsistent. A more targeted use of these tests, restricted to patients in whom neuromuscular weakness is more likely to contribute to weaning failure, may improve their clinical relevance. We conducted a pooled individual patient data analysis of multiple observational cohorts including adult ICU patients receiving invasive mechanical ventilation. The primary outcome was extubation failure, defined as reintubation or death within 48 h after extubation. Available neuromuscular assessments included maximal inspiratory pressure (MIP), maximal expiratory pressure (MEP), peak cough flow (PCF), and Medical Research Council (MRC) sum score. Prespecified thresholds were MIP ≤ 30 cmH₂O, MEP ≤ 50 cmH₂O, PCF ≤ 60 L/min, and MRC < 48/60. Associations with extubation failure were assessed in the overall population and according to prior mechanical ventilation duration. Among 494 included patients, 458 had an available extubation outcome and 72 (16%) experienced extubation failure. MIP was available in 230 patients, MEP in 124, PCF in 156, and MRC score in 252. In the overall population, MIP ≤ 30 cmH₂O showed only a borderline association with extubation failure (24% vs. 14%; OR 1.94, 95% CI 0.98 to 3.84; p = 0.06). In patients ventilated for more than 5 days, extubation failure occurred in 33% with MIP ≤ 30 cmH₂O versus 14% with MIP > 30 cmH₂O. After adjustment, MIP ≤ 30 cmH₂O remained associated with extubation failure (adjusted OR 3.52, 95% CI 1.37 to 9.03; p = 0.009). No significant adjusted association was observed for MEP, PCF, or MRC. Neuromuscular tests may have limited prognostic value when applied broadly before extubation. Low MIP appears more informative after prolonged mechanical ventilation, supporting a targeted rather than systematic use of respiratory neuromuscular assessment before extubation.