Escore preditor de mortalidade e biomarcadores de gravidade na leptospirose

• Ano de defesa: 2023
• Autor(a) principal: Galdino, Gabriela de Studart
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Palavras-chave em Português: Leptospirose; Biomarcadores; Mortalidade; Medicina
• Link de acesso: http://repositorio.ufc.br/handle/riufc/74674

Leptospirosis is a neglected disease that persists with relevant morbidity and mortality rates despite advances in identification and management. Early detection of severity factors in these patients is crucial to point out the necessary care. The main objectives of this work were to develop a predictor score of death in severe leptospirosis and to investigate the correlations between this and the new biomarkers of renal and endothelial damage. Method: Data from 295 hospitalized patients for leptospirosis were retrospectively analyzed to build the mortality predictor score which was eventually created in a derivation cohort by using machine learning models. For the study of biomarkers and its associations with the new score, the study was prospective and evaluated consecutive cases of leptospirosis admitted to the 3 tertiary hospitals in Fortaleza from February 2017 to April 2023. Blood and urine samples were collected upon hospital admission to quantify the biomarkers of severity, renal (MCP-1, serum and urinary NGAL and FGF-23) and endothelial (ICAM-1, VCAM-1, Angiopoietin-1, Angiopoietin-2, Syndecan-1 and vWF-A2). Results: To construct the score, Lasso was selected for regression analysis, as it demonstrated better accuracy in predicting mortality in the sample [area under the curve (AUC-ROC) = 0.776]. A prediction score based on the Lasso coefficients was performed and named LeptoScore. To simplify the initial predictive model, a new score was constructed by assigning points to the parameters with greater values of importance. The simplified score, called QuickLepto, has five variables (age > 40 years; lethargy; pulmonary alterations; mean arterial pressure < 80 mmHg and hematocrit < 30%) and showed good predictive value (AUC-ROC = 0.788). In the prospective sample, 44 patients were included, 81.8% were men, the mean age was 40.8 ± 16.9 years and 9.1% died. The correlations between the new QuickLepto score and the severity biomarkers measured at hospital admission have shown that the entire endothelial profile studied with VCAM-1 (r= 0.507, p <0.001), ICAM-1 (r= 0.311, p= 0.040), syndecan-1 (r= 0.331, p= 0.028) and Ang-2 (r= 0.442, p= 0.003) showed significant and positive correlation, except for Ang-1. In the analysis of coagulation-related parameters, the vWF-A2 biomarker was the only one that showed a significant correlation with the predictor score (r= 0.417, p=0.005). Platelets, TAP and APTT showed no significant association. In the renal profile, serum NGAL also showed a positive correlation with statistical significance (r= 0.388, p=0.009). No association was observed with creatinine, serum FGF-23 and MCP-1 and urinary NGAL. Conclusion: The new scoring system, QuickLepto, is a simple and useful tool to predict mortality in patients with leptospirosis at hospital admission. The serum NGAL renal biomarker and the endothelial biomarkers ICAM-1, VCAM-1, syndecan-1, Ang-2 and vWF-A2 may be helpful to predict death related to severe leptospirosis as well as to guide early clinical interventions aimed at reducing this outcome.