Risk and Forecasting Factors of Liver Dysfunction in Military Persons with Gunshot Wounds

Abstract

The purpose of the study was to improve the results of treatment of post-traumatic reactive hepatitis in victims of gunshot wounds by improving the prognosis, early diagnosis and treatment. Materials and methods. A retrospective analysis of 233 medical records of inpatients of wounded military persons of 2014-2020 hospitalized in medical institutions of the Ministry of Defense of Ukraine for more than 14 days. All victims were examined in the laboratory using unified methods of laboratory tests. The value of relative risk was calculated. The probability of liver dysfunction was assessed using binary logistic regression with a comprehensive analysis of the values of individual indicators (risk factors). The informativeness of the forecast was determined by sensitivity and specificity, which were calculated on the basis of true and false (positive and negative) forecasts. Additionally, the quality of the binary logistics model was tested by ROC analysis to determine the area under the ROC-curve (AUC). Results and discussion. There was a significant difference in the frequency of detection of most of the analyzed indicators: mine-explosive type of injury; presence of abdominal trauma, including liver trauma, multiple soft tissue trauma, traumatic shock and anemia (p <0.05 according to Fisher's exact test). Skeletal trauma occurred in both subgroups with almost the same frequency, thoracic trauma more often, but not significantly (p >0.05 according to Fisher's exact test). All studied indicators had a significant risk of developing reactive hepatitis (upper and lower limits of 95 % confidence interval greater than 1.0). Anaemia had the greatest impact (relative risk greater than 5.6 with a specificity of 0.951), followed by multiple soft tissue injuries (more than 3.8 with a specificity of 0.833), traumatic shock (more than 3.7 with a specificity of 0.747). Relative risk values of the mine-explosive nature of injuries, abdominal trauma and liver trauma were in the range of 2.0-2.7. Using the method of Wald's exclusion, a model of reactive hepatitis prognosis was created with a prognostic value of a positive result of 78.9 %, a prognostic value of a negative result of 95.1 %, specificity of 87.5 %, and sensitivity of 91.1 %. The results of the ROC analysis were evaluated based on the area under the ROC-curve (AUC), which was 0.879 (95 % CI 0.810-0.929) at a significance level of <0.001. The cut-off value is 0.5. For the convenience of usage of the described model, the mobile application was created. Conclusion. The most significant risk factors for reactive hepatitis after gunshot wounds are anaemia, multiple soft tissue injuries, and traumatic shock. The model for predicting the development of reactive hepatitis after a gunshot wound, made in the form of a mobile application, in which the presence of abdominal trauma, multiple soft tissue trauma, traumatic shock and anaemia are used as predictors, has a specificity of 87.5 % and sensitivity of 91.1 %