Analysis Of Risk Factors And Construction Of A Predictive Model For Invasive Fungal Infections After Severe Burns

Objective:To analyze the epidemiological characteristics of invasive fungal infection（IFI）after severe burns,and to explore the risk factors for IFI in patients with severe burns,to establish a logistic regression prediction model to screen out the high-risk groups that may develop IFI,so as to target the high-risk groups for early intervention and reduce the incidence and mortality of IFI.Methods:Using a retrospective nested case-control study,patients with severe burns from2012-2021 at the First Affiliated Hospital of Nanchang University Burn Plastic and Trauma Restoration Medical Center were used as the study cohort population.The patients were divided into IFI and non-IFI groups according to the IFI diagnostic criteria,and the epidemiological characteristics of the IFI group were analyzed.The106 patients with IFI occurring at age>14 years were used as the case group,and then matched with the same sex,age±3,and length of hospitalization（the length of hospitalization of the matched non-IFI patients with IFI was at least at least the interval between their admission and infection）,and 212 patients without IFI were randomly matched in a 1:2 ratio as the control group.Single-factor analysis of variance method and multi-factor logistic stepwise regression method were performed for the two groups to identify independent risk factors for the development of IFI in patients with severe burns and to establish a prediction model.In addition,the Hosmer-Lemeshow（H-L）was used to assess the calibration ability of the prediction model.The efficacy of the prediction model was assessed using the subject operating characteristic curve ROC（Receiver Operating Characteristic,ROC）.Results:1.Epidemiology:the study included 1297 patients with severe burns,of whom166 were diagnosed with IFI,and the annual incidence of IFI ranged from 7.87%to22.73%,with an overall increasing trend.Candida tropicalis was the most predominant pathogen of infection in IFI burn patients,with the proportion of Candida albicans infections decreasing and the proportion of non-Candida albicans infections trending upward.The main types of Candida including Candida subsmoothis,Candida albicans and Candida tropicalis are not less resistant to azole antifungal drugs.The incidence of IFI was higher in patients with injuries due to blast and electric shock,and IFI was more likely to occur in patients with severe burns admitted in summer and autumn,and was mostly concentrated in the second to third weeks after admission.2.Multifactorial outcome analysis showed that diabetes mellitus or stress hyperglycemia（OR=6.304,95%CI:2.691-14.767,P<0.001）,indwelling urinary catheter≥7d（OR=5.541,95%CI:1.863-16.483,P=0.002）,parenteral nutrition（PN）≥5d（OR=2.679,95%CI:1.309-5.483,P=0.007）,central venous line≥3d（OR=2.319,95%CI:1.15-4.676,P=0.019）,number of antibiotics used（OR=1.932,95%CI:1.399-2.669,P<0.001）,use of acid suppressants（OR=4.729,95%CI:1.375-16.27,P=0.014）,mechanical ventilation≥3d（OR=2.711,95%CI:1.271-5.784,P=0.01）,number of procedures（OR=1.816,95%CI:1.353-2.439,P<0.001）,and Acute Physiology and Chronic Health Evaluation Scoring SystemⅡ（APACHEII）（OR=1.205,95%CI:1.091-1.331,P<0.001）were independent risk factors for IFI,and prophylactic use of antifungal drugs（OR=0.06,95%CI:0.023-0.155,P<0.001）were independent protective factors.3.Preducation Model:logit P=-8.54+（1.841X₁）+（1.712X₂）+（0.985X₃）+（0.841X₄）+（0.659X₅）+（-2.816X₆）+（1.554X₇）+（0.186X₈）+（0.597X₉）+(0.997X₁₀).X₁ is diabetes or stress hyperglycemia,X₂ was indwelling urinary catheter≥7d,X₃ was PN≥5d,X₄was central venous placement≥3d,X₅ was number of antibiotics used,X₆ was prophylactic use of antifungal drugs,X₇ was use of acid suppressants,X₈ was APACHEII score,X₉ was number of surgeries,X₁₀ was mechanical ventilation≥3d.H-L test showed P=0.644>0.05 model The calibration ability was good.Area under the curve of the ROC=0.901,95%CI（0.878-0.941）,P<0.001,the model has good predictive ability.Conclusion:1.The incidence of IFI is increasing year by year,and the main strain composition is changing from Candida albicans to non-Candida albicans,and non-Candida albicans is resistant to azole antifungal drugs.2.Having diabetes mellitus or developing stress hyperglycemia,high APACHEII score,using more types of antibiotics,more surgeries and acid suppressant treatment,prolonged indwelling urinary catheter,PN,central venous placement,and mechanical ventilation are independent risk factors for the development of IFI in patients with severe burns.3.The prediction model for the risk of IFI in patients with severe burns constructed based on the above independent risk factors was validated,and the model showed good predictive performance,which provided guidance for clinical interventions to reduce the occurrence of IFI.