Leptospirosis: Prognostic Model for Patient Mortality in the Transcarpathian Region, Ukraine.

Background: Leptospirosis is a bacterial zoonosis of worldwide distribution with a wide spectrum of clinical presentations that range from subclinical or mild to severe and fatal outcomes. Identifying clinical predictors for the severe form of the disease is critical to reduce disease complications and death. As a result, we conducted a retrospective case-control study to identify clinical markers of mortality in leptospirosis patients from the Transcarpathian region. Materials and Methods: The study used 102 medical records of patients with leptospirosis in the period from 2009 to 2019. There were 26 fatal cases and 76 survivors. Predictors were examined using univariate and bivariate statistics. Results: Fatal and nonfatal groups did not differ in age or gender composition ( p  > 0.5), nor did they differ in signs or symptoms, except that oliguria occurred significantly more often in fatal cases ( p  < 0.001). Laboratory diagnostic tests, however, differed between outcomes in 7/9 recorded variables; primarily associated with liver and kidney function as well thrombocytopenia and elevated white blood cell counts for fatal cases ( p  < 0.01 for all variables). Conclusions: Logistic regression analysis indicated that a combination of creatinine levels and direct bilirubin levels were the best predictors of patient outcome. The specificity of the model was 90.9% and the area under the receiver operator characteristic function as 93.6%. This model can be used when a patient is admitted to a hospital to better characterize patient risk.