Objective: To determine whether the shock index (SI), defined as the ratio of heart rate (HR) to systolic blood pressure (SBP), is a useful marker for significant injury in trauma patients.
Methods: A retrospective database analysis was used to relate the SI to the clinical measures: death within 24 hours, injury severity score (ISS) > or = 16, intensive care unit (ICU) stay > or = 1 day, and amount of blood transfused (BT) > or = 2 units. Consecutive trauma patients seen at one level I trauma center over a 24-month period were reviewed; excluded were patients not requiring trauma team consultation, or those with either incomplete records, severe head injury (Glasgow Coma Scale score < or = 8), or age < 14 years. The SI was calculated from ED admission vital signs. Receiver operating characteristic (ROC) curves were used to find the value of the SI that maximized the sum of sensitivity and specificity for predicting each measure, separately; a separate analysis was done to determine the optimal SI threshold for predicting any of the severity measures.
Results: 1,101 cases met study criteria. The optimal SI values (by ROC analysis) for predicting the severity measures were: 1.10 for death < 24 hours, 0.71 for ISS > or = 16, 0.77 for ICU > or = 1 day, and 0.85 for BT > or = 2 units. The optimal SI value (by ROC analysis) for any of the above measures was 0.83; use of this SI cutoff provided a sensitivity of 37% (95% CI 32-42%), a specificity of 83% (95% CI 80-87%), and a negative predictive value of 58% (95% CI 54-61%) for any measure. This SI threshold predicted between 24% fewer cases and 4% more cases of poor outcome than did the optimal thresholds HR and SBP, respectively.
Conclusion: The optimal SI threshold performed similarly to the optimal threshold HR or SBP for prediction of injury severity.