Venous Thromboembolic Risk Stratification In Pediatric Trauma: A Pediatric Trauma Society Research Committee Multicenter Analysis
Daniel F. Labuz1, *Aaron J. Cunningham1, Joseph Tobias1, *Christopher W. Marenco2, *Mauricio Escobar3, Max D. Hazeltine4, Muriel A. Cleary4, *Meera Kotagal5, *Richard A. Falcone, Jr.5, *Sara C. Fallon6, *Bindi Naik-Mathuria6, Taleen MacArthur7, *Denise B. Klinkner7, Aashka Shah8, *Artur Chernoguz8, Alexandra Dixon1, *Nicholas A. Hamilton1, *Mubeen A. Jafri9
1Oregon Health & Science University, Portland, OR; 2Madigan Army Medical Center, Tacoma, WA; 3Mary Bridge Children's Hospital, Tacoma, WA; 4University of Massachusetts Medical School, Worcester, MA; 5Cincinnati Children's Hospital Medical Center, Cincinnati, OH; 6Texas Children's Hospital, Houston, TX; 7Mayo Clinic, Rochester, MN; 8Tufts Medical Center, Boston, MA; 9Randall Children's Hospital, Portland, OR
PURPOSE: While venous thromboembolism (VTE) in injured children is rare, its consequences are significant. Little consensus on the prevention of VTE in injured children exists despite many published risk-stratification schemes. We performed a multicenter study to assess the utility of one such algorithm.
METHODS: Local trauma registries at eight institutions were queried for all patients < 18 years of age from 2009-2018. Clinical variables were applied to the prediction algorithm (Figure 1A), and VTE risk was calculated. Additional chart review was performed of all VTE cases and randomly selected non-VTE cases in a 1:2 ratio. The area under a receiver operator characteristic (AUROC) curve was calculated to test discriminatory ability. Test statistics were tabulated.
RESULTS: Review identified 42,424 patients with 66 episodes of VTE (0.16%). Prophylaxis against VTE was used in 4.7%. Application of VTE prediction algorithm classified 3.7% as moderate- or high-risk (>1% risk). AUROC was 0.897 (0.843-0.951, 95%CI) (Figure 1B). In this national population, use of the VTE risk algorithm provided an accurate (96.5%) screening tool with a negative predictive value of 99.9% (96.1%-96.8%, 95%CI) narrowing the prophylaxis cohort to 3.7% of all pediatric trauma patients (Figure 1C). Number-needed-to-screen using the VTE risk algorithm was 24.
CONCLUSION: A VTE risk algorithm using ten clinical variables can identify patients at risk for venous thromboembolic disease with strong discrimination. Prospective application should be studied.
Back to 2020 Abstracts