Development of a preliminary risk index to identify trauma patients at risk for an unplanned intubation

The development of respiratory failure requiring an emergent unplanned intubation (UI) is a potentially preventable complication associated with increased morbidity and mortality. The objective of this study was to develop a clinical risk index for UI based on readily available clinical data to assist in the identification of trauma patients at risk for this complication. We also sought to determine the impact of UI on patient outcomes.

This is a 3-year retrospective analysis of our Level 1 trauma center registry to identify all patients requiring a UI. Patients who required a UI were compared with patients who were never intubated. An additive risk index consisting of 10 clinical variables was created using the final significant variables from a stepwise logistic regression model. The sensitivity and specificity of every possible index score were calculated and added together to calculate the "gain in certainty" values.

During the 3-year period, 7,552 patients were admitted, of whom 967 (12.8%) required intubation. Of these, 55 (5.7%) underwent a UI. The final risk index consisted of 10 variables as follows: age 55 years to 64 years, age 65 years or older, male sex, Glasgow Coma Scale (GCS) score of 9 to 13, seizures, chronic obstructive pulmonary disease, traumatic brain injury, four or more rib fractures, spine fractures, and long-bone fractures. Gain in certainty was maximized at an index score of 4, with the highest combined sensitivity and specificity of 86.0% and 74.9%, respectively. The probability of UI increased from 0.9% at a score of 1 to 2.9% at 4 and 43% at 9. UI was associated with increased overall complications, length of stay, and mortality (p < 0.001).

UI is a potentially preventable adverse event associated with poor outcomes. Identification of patients at risk for this complication may be possible through the development of an additive risk index. Prospective validation of the risk index is potentially warranted.

Diagnostic study, level III.