Establishment Of Risk Prediction Model And Scoring Criteria For Mortality In Severe Fever With Thrombocytopenia Syndrome

Objective In order to evaluate the risk of severe fever with thrombocytopenia syndrome in time and accurately,predictive use of active treatment to change the prognosis of patients,and establish a risk prediction model and related scoring criteria for severe fever with thrombocytopenia syndrome.Methods A total of 91 patients(38 males(41.8%)and 53 females(58.2%))diagnosed as severe fever with thrombocytopenia syndrome from June 1,2020 to June 1,2022 in Chaohu Hospital of Anhui Medical University were retrospectively analyzed.SPSS 24.0 statistical software was used to test the normality of all variables and show the normal curve.The t test were performed for variables that met the normal distribution.Mann-Whitney U test was used for variables that did not meet the normal distribution.Continuous variables were expressed as median(min-max),and categorical variables were expressed as percentages.Binary logistic regression analysis was performed for variables with P < 0.05 in univariate analysis and a risk prediction model was established.The goodness of fit and prediction accuracy of the model were assessed by ROC curve,and finally the scoring criteria of the mortality risk prediction model were constructed according to the weight coefficients of each risk factor in the model.Results From June 1,2020 to April 13,2022,91 patients(38 males(41.8%)and 53females(58.2%))with confirmed fever and thrombocytopenia syndrome were included.There was a significant difference in age between the survival and death groups(P = 0.028).Among all observed subjects,there were a total of 20 deaths,with a case fatality rate(CFR)of 22.0%.The incidence of SFTS in Anhui Province was mainly concentrated in May - June,followed by October,showing a double peak.In terms of early laboratory parameters in the survival group and the death group: WBC,NE,LYM,HL,PCT,cTN-i,AMY,LPS,PLT,APTT,PT,CK,CK-MB,LDH,ALT,AST,GLOB,BUN,Cr,there was no statistical significance,and in vital signs:mean arterial pressure,respiration,and heart rate were also not statistically significant.Patients with SFTSV infection can be divided into three stages: in the first stage,no indicator shows that it can be closely associated with neobunyavirus infection and death;in the second stage,high levels of aspartate aminotransferase and low levels of albumin are factors associated with neobunyavirus infection death;high levels of LDH in the late stage of the disease are risk factors for death;and gamma globulin and ribavirin alone have no clinical effect in the treatment of SFTSV infection.Three independent risk factors for in-hospital mortality in patients with severe fever and thrombocytopenia syndrome were found,and the corresponding scores for each risk factor were given according to the risk prediction regression equation,which were: age > 60(OR: 1.197,95% CI: 1.012 -1.416,1 point),moderate fever(day)(OR: 1.671,95% CI: 1.188 - 2.351,2 points),and high fever(day)(OR: 2.954,95% CI: 1.638 - 5.327,2 points,and the patients obtained scores between 0 and 10 points,of which 0 - 3 points were low risk,4 -6 points were moderate risk,and 7 - 10 points were high risk.Homser-Lemeshow test P value > 0.05,area under ROC curve 0.948,95% confidence interval(0.891 -1.000),model goodness of fit and prediction accuracy were good.Conclusion Age > 60,moderate fever(days)and high fever(days)are independent risk factors for early judgment of death in patients with severe fever and thrombocytopenia syndrome after admission,while the first serological test after admission has little correlation with clinical practice.At the same time,the risk of death during hospitalization can be better predicted according to the scoring criteria specified by the model.A clinical predictive SFTS mortality prediction model based on age > 60 years(OR: 1.197,95% CI: 1.012-1.416),moderate fever(days)(OR: 1.671,95% CI: 1.188-2.351),and high fever(days)(OR: 2.954,95% CI: 1.638-5.327)was developed and associated scores were developed to explore a linear model for different fever time,fever,and risk of death in patients...