Background:Severe fever with thrombocytopenia syndrome (SFTS) is an emerging infectious disease known for its high mortality rate and its correlation with Cytokine Storms (CS). Timely detection of CS is crucial for improving the prognosis of the disease. The objective of this investigation was to develop a model for identifying cytokine storms in the acute phase of SFTS.
Methods:A total of 245 patients diagnosed with SFTS were included in this study between January 2020 and July 2022. Among them, 184 patients were part of the training set, while 61 patients were part of the validation set. Variables identified by LASSO were subsequently included in a multivariate logistic regression analysis to determine independent predictors. Subsequently, a nomogram was then developed to predict the likelihood of CS in SFTS patients. The predictive efficacy and clinical applicability of the nomogram model were further assessed through ROC analysis and the DCA curve.
Results:Following LASSO analysis, a total of 11 indicators were included in multivariate logistic regression analysis. The findings indicated that PLT (OR 0.865, P < 0.001), LDH (OR 1.002, P < 0.001), Na+ (OR 1.155, P = 0.005), and ALT (OR 1.019, P < 0.001) serve as independently predictors of CS in the acute phase of SFTS. Furthermore, a nomogram named the PLNA was constructed by integrating these four factors. The PLNA model exhibited favorable predictive accuracy with an AUC of 0.958. Moreover, the PLNA model exhibited excellent clinical applicability in both the training and validation sets, as evidenced by the DCA curve.
Conclusions:The PLNA model, constructed using clinical indicators, can predict the probability of cytokine storm in the acute phase of SFTS patients.
Keywords:Cytokine storm; PLNA score; Prediction nomogram; Severe fever with thrombocytopenia syndrome.