Age-Related Disparities in the Predictive Performance of the Shock Index for Massive Transfusion in Trauma Patients: A Retrospective Cohort Study

Abstract

Background: In trauma, the shock index (SI) is commonly used to assess the presence of significant blood loss. Prior studies have shown that the SI has a fair predictive ability for clinical outcomes such as massive transfusion (MT) or mortality in adult trauma patients. We hypothesized that the relatively lower predictive power of the SI in older adult patients compared to that of younger adult patients results in the overall fair predictive ability of the SI for clinical outcomes in adult trauma patients. Methods: This retrospective observational study analyzed adult trauma patients who presented to a single regional trauma center between 2019 and 2023, categorizing them into younger (18–64 years) and older (≥65 years) cohorts. The association between SI and MT was evaluated using simple logistic regression, while the modifying effect of age on this association was evaluated through an interaction model. The predictive performance was compared between the groups using the area under the receiver operating characteristic curve (AUC). Age-stratified AUC trends were visualized using cubic spline analysis. Results: A total of 2404 trauma patients met the inclusion criteria, including 1531 younger adults and 873 older adults. The SI was identified as an independent predictor of MT, with a stronger association in younger adults. The AUC for predicting MT was significantly higher in younger adults compared to older adults (0.801 vs. 0.666; p < 0.001), with optimal SI cut-off values of 1.18 and 0.88, respectively. Age-stratified analysis showed the highest AUC in the 41–50 age group (AUC 0.880; 95% CI, 0.836–0.916) and the lowest in the 71–80 age group (AUC 0.624; 95% CI, 0.573–0.674). Conclusions: The predictive performance of the SI for MT was influenced by age, demonstrating a lower predictive ability in older adult patients compared to younger adults.