Elevated admission international normalized ratio strongly predicts mortality in victims of abusive head trauma

Abstract

Background: Victims of abusive head trauma have poor outcomes compared with other injured children. There is often a delay in diagnosis because these young patients are unable to communicate with health care providers. These critically injured patients would benefit from early identification and therapy.

Methods: We performed a retrospective review of our single hospital trauma registry from 2005 to 2014. All Level 1 pediatric (age 0-17 years) trauma patients who sustained abusive head trauma were included. Exclusion criteria included no admission coagulation studies, prehospital product transfusion, preexisting coagulation disorder, or death upon arrival. Primary outcome was mortality; secondary outcomes were early blood transfusion and neurosurgical intervention. Univariate analysis included Fisher's exact and Wilcoxon rank-sum testing; we then performed logistic regression modeling and calculated adjusted odds ratios (AORs) to control for known predictors of poor outcome including hypotension, hypothermia, acidosis, Injury Severity Score (ISS), and head Abbreviated Injury Scale (AIS) score.

Results: In 101 total subjects, 35% (n = 35) had international normalized ratio (INR) of 1.3 or greater at admission. On univariate analysis, patients with coagulation dysregulation were more likely to have hypothermia, hypotension, acidosis, high ISS, and low Glasgow Coma Scale (GCS) score (all p < 0.05). There was no difference in age, anemia, and incidence of polytrauma. Overall mortality was 24.8% (n = 25), which varied significantly based at admission INR (60% INR ≥ 1.3 vs. 6% INR > 1.3, p < 0.001). Patients with elevated INR were also more likely to have early packed red blood cell transfusion (p = 0.003) and neurosurgical intervention (p = 0.011). In logistic regression analysis, admission INR was the strongest independent predictor of mortality, with increased odds of 3.65 (p = 0.045). AOR after controlling specifically for hypotension, hypothermia, and acidosis was 6.25 (p = 0.006), and after controlling for head AIS score and admission GCS score, the AOR was 5.27 (p = 0.007).

Conclusion: Admission INR of 1.3 or greater strongly predicts mortality in abusive head trauma. These patients should be targeted for early aggressive interventions and monitoring with the goal of improving patient outcomes. Further study is warranted to investigate potential therapeutic targets in trauma-induced coagulation dysregulation.

Level of evidence: Prognostic and epidemiologic study, level III.