Predicting extubation failure in blunt trauma patients with pulmonary contusion.

BACKGROUND: The need for reintubation after weaning from mechanical ventilation (extubation failure) is associated with increased morbidity and mortality. In blunt trauma patients with pulmonary contusion, factors predicting successful weaning have not been reliably defined. The purpose of this study was to identify criteria predicting successful extubation in these patients.
METHODS: Retrospective review during a 10-year period at a Level 1 trauma center was performed. A total of 173 extubations in 163 blunt trauma patients with pulmonary contusion requiring mechanical ventilation. Exclusion criteria include Glasgow Coma Scale (GCS) score of less than 9T before extubation, successful use of noninvasive positive-pressure ventilation after extubation, quadriplegia, and preextubation FIO2 of greater than 0.5. Data included age, Injury Severity Score (ISS), ventilator days, as well as GCS score, FIO2, the ratio of arterial oxygen tension to FIO2 (P/F ratio), and alveolar-arterial oxygen (A-a) difference at the time of extubation. Failure was defined as reintubation within 72 hours (excluding stridor or acute decline in GCS score). Mann-Whitney U-test, χ2 analysis, and logistic regression analysis determined variables associated with extubation failure. Odds ratios were used to compare P/F and A-a values associated with failed extubation.
RESULTS: A total of 147 extubations (85%) were successful; 26 required reintubation. Patients did not differ by ISS, chest Abbreviated Injury Scale (AIS) score, presence of sternal or rib fractures, and admission pneumothorax or hemothorax. Increased age, A-a difference (≥ 120 mm Hg), and decreased P/F (<280) were associated with reintubation (p < 0.0001). By logistic regression analysis, P/F and A-a were independent variables for failed extubation; both remained independent risk factors when adjusted for age, ventilator days, GCS score, and preextubation FIO2. Using receiver operating characteristic curve inflection points for both P/F and A-a difference (area under the curve of 0.8 for both), patients with a P/F ratio less than 290 and an A-a difference of 100 mm Hg or greater were more likely to fail extubation (odds ratio, 9.2 and 8.7, respectively, p < 0.001).
CONCLUSION: Blunt trauma patients with pulmonary contusion who are likely to fail extubation can be reliably identified using the readily available criteria of P/F ratio less than 290 and A-a difference of 100 mm Hg or greater.