Open AccessPredictors of Acute Mortality After Open Pelvic Fracture: Experience From 37 Patients From A Level I Trauma Center
Author(s) -
Tseng IChuan,
Chen IJung,
Chou YingChao,
Hsu YungHeng,
Yu YiHsun
Publication year - 2020
Publication title -
world journal of surgery
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.115
H-Index - 148
eISSN - 1432-2323
pISSN - 0364-2313
DOI - 10.1007/s00268-020-05675-z
Subject(s) - medicine , pelvic fracture , trauma center , logistic regression , mortality rate , confidence interval , injury severity score , youden's j statistic , univariate analysis , incidence (geometry) , surgery , multivariate analysis , receiver operating characteristic , retrospective cohort study , pelvis , poison control , emergency medicine , injury prevention , physics , optics
Background Open pelvic fractures are caused by high‐energy traumas and are accompanied by organ injuries. Despite improvements in pre‐hospital care, the acute mortality rate following open pelvic fractures remains high. This study aimed to report experiences in managing open pelvic fractures, identify potential independent predictors that contribute to acute mortality in such patients, and generate a scoring formula to predict mortality rate. Methods Open pelvic fracture patients managed during a 42‐month period were retrospectively studied. Logistic regression analysis was used to determine predictors of acute mortality. Using the Youden index, threshold values of predictors were selected. Significant predictors were weighted to create a scoring formula. The area under the curve (AUC) was tested in this specific group. Results The incidence of open pelvic fractures in all pelvic fractures was 4.9% (37/772), and the overall mortality rate was 21.6% (8/37). All the successfully resuscitated patients entered the reconstruction stage survived and underwent the complete treatment course. Univariate and multivariate logistic regression analyses revealed that the revised trauma score (RTS) was the single independent predictor of acute mortality. A scoring formula was generated following the statistical analysis. The probability of mortality was 0% and 100% when the score was above and below −2, respectively. This model predicted mortality with an AUC of 0.948 (95% confidence interval 0.881–1.000, P < 0.01). Conclusion The RTS may be a potential predictor of acute mortality in open pelvic fracture patients. Further work would be required to validate the clinical efficacy of the generated scoring formula.