(1409) Utility of Delta Shock Index in Patients With Upper Gastrointestinal Bleeding

Introduction: Shock Index (SI), defined as HR divided by systolic BP (SBP) has been associated with increased mortality and need for intervention in trauma. While current scoring systems do not reliably predict adverse outcomes in patients presenting to the emergency department (ED) with an upper gastrointestinal bleed (UGIB), SI has been shown to be a good predictor of positive extravasation on angiography and the requirement for embolization in patients with acute GIB. However, some conditions may blunt the usual tachycardic response of hemorrhage including extremes of age and medications. The Delta Shock index (DSI) may overcome these shortcomings by using changes in patients’ own hemodynamics as a marker. We aimed to investigate the utility of DSI vs ED SI in patients with GIBs.

Methods: We performed a single-center retrospective analysis of patients who arrived via EMS and found to have a GIB identified by ED ICD 10 codes from December 2017 to December 2022. DSI was defined as the change between initial SI by EMS and SI on arrival to the ED. The primary outcome was in-hospital mortality; secondary outcomes included admission, blood transfusion, and ICU admission. DSI and ED SI were compared using ROC curves and statistics from a confusion matrix. DeLong’s test for two correlated ROC curves was used to test for a difference in the AUC of the ROC curves.

Results: Over the study period, a total of 277 participants were included. The average age was 67 years old, 54% were women, average hemoglobin on arrival to the ED was 9.7 g/dL., 65% had a qSOFA score of zero and 69% with an emergency severity index of two. The AUC of the ROC curves did not differ significantly between ED SI and DSI for hospital mortality (p = 0.304), or blood transfusion (p = 0.109). ED SI AUC was significantly higher than the DSI AUC for ICU admission (p-value = 0.025) and admission from ED (p-value = 0.031).

Conclusions: DSI did not offer predictive value in for mortality, ICU admission or need for transfusion. Interestingly, ED SI was associated with the need for ICU admission in patients with GIB. We were limited by sample size, and shorter transport times in our hospital’s urban setting leading to less time for SI to deteriorate.