Unconditional Logistic Stepwise Regression Analysis Of Risk Factors For Sepsis In Severely Burned Children

BackgroudBurns is a common trauma. In Europe, almost0.2~2.9person need hospitalization due toburn among10,000residents annually, and nearly half of the burned patients are younger than12-year-old. Many clinical data from the burn units in our country display that, pre-schoolchildren account for21.4%~43.6%of the whole hospitalized burned patients. Because ofgrowing and immaturation of tissue and organs, children are more prone to be infected aftersevere burn. The infection can cause sepsis, which has become one of the leading cause ofdeath in paediatric burn patients.The pathogenesis of sepsis is a extremly complex process. The treatment is very difficult,if sepsis has progressed to severe sepsis and septic shock. The number of people die of severesepsis is over18million every year globally, and exceed the people died of myocardialinfarction, and still increasing in the rate of1.5%per year. Because of the great harm to humanhealth, the European Society of Intensive Care Medicine (ESICM), the Society of critical caremedicine (SCCM), the International Sepsis Forum (ISF) and other organizations collectivelysigned the “Surviving Sepsis Campaign”(SSC) in the European Critical Care Medicinemeeting, which held in Barcelona, Spain in October,2002. The joint organization alsoformulated a guidelines for treatment of sepsis (the SSC guidelines) in2004, and revised itaccording to clinical practice in2008. Nevertheless, the treatment of severe sepsis and septicshock is still very diffcult. Effective prevention is still very important to combat sepsis in thosepeople at high risk of sepsis.Since the Amerivan Association of Chest Physician（ACCP）/SCCM defined sepsis andformulated diagnostic criteria of sepsis in1991, experts have been trying to improve thediagnostic criteria, and held a joint meeting to revised the above diagnostic criteria inDecember2001. The new diagnostic criteria of sepsis is more objective and clear. After that,based on the characteristics of the patients and the diagnostic criteria of sepsis established in2001conference, various disciplines formulated their specialized diagnostic criteria of sepsis. Both burn surgery and paediatrics has formulated their own sepsis diagnostic criteria basing ondifferent patient characteristics.Because of the improvement of the diagnostic criteria, there are a lot of risk factors forsepsis had been reported, aimed at early prevention for high-risk group. The clinical studies indiffernt burn centers have demonstrated that, the patient’s age, gender, total burned surfacearea, third degree burned surface area, and inhalation injury are related to the occurrence ofsepsis. For children, experts find the age of children, respiratory infection, and underlyingdisease are closely related to the occurrence of sepsis. However, the clinical research on riskfactors for burned children with sepsis is still very rare.ObjectivesBased on clinical data collected from paediatric burn patients, valuate the influence ofpatient general imformation, injury and early treatment of injury on the occurrence ofpaediatric sepsis, try to find the risk factors of sepsis. Furthermore we study the effects of therisk factors for prognosis and treatment.Materials and methodsWe collected the paediatric burn patients in our hospital during January1,2003-December31,2011. Including criteria is as below:(1) age between1month-12years old;(2)total burned surface area≥15%TBSA (total body surface area);(3) no discharge or deadwithin3days after injury;(4) without underlying diseases;(5)without other combined injury;(6) with complete clinical data. According to the cases selection criteria, a total of679childrenwere included, and80cases were diagnosed as sepsis,10cases as severe sepsis,5cases ascomplicated with septic shock. All burned children clinical data were collected. Those includesome risk factor, for example age, gender, burn causes, total burned surface area, third degreeburned surface area, inhalation injury, early anti-shock treatment, early wound treatment, earlyenteral nutrition support; and treatment data, such as number of operations, consumption ofantibiotics, blood transfusion, and wound healing time.Grouped all cases according to sepsis occurred or not, and performed univariate andmultivariate analysis to identify risk factors of sepsis in severely burned children. Extractedsome cases from all paediatric patients, and performed case-paired study to understand the influence of risk factors which had been studied above for treatment and prognosis. All datawere processed and analyzed by SPSS18.0. Descriptive statistics included median(interquartile range) or the mean±standard deviation for continuous data and proportions forcategorical variables. Continuous variables were compared using group t-test and paired t-test,if meet the parametric test conditions,or using the Mann-Whitney U test and Wilcoxonsigned-rank test. Categorical data were compared by chi-square test or Fisher exact test. Thedifference was statistically significant when P <0.05. Unconditioned logistic regressionanalysis were used in multivariate analysis, by the stepwise forward selection method, andincluding rate of0.05and excluding rate of0.10. Calculated odd ratio (OR) and95%confidence interval (CI) of research factors.Results1. The incidence rate of sepsis, severe sepsis and septic shock were11.78%(80/679),1.47%(10/679) and0.74%(5/679) respectively. There were52male children, and28femalechildren with sepsis, and the gender rate was almost2:1(male to female). The children withsepsis were2.10(1.73-4.00) years old, total burned surface area were35.00(24.00-49.75)%TBSA, and third degree burned surface area were16.50(0.00-31.50)%TBSA.Hydrothermal burn was the leading cause, accounted for78.75%(63/80). Pseudomonasaeruginosa (81), Acinetobacter baumannii (42), Staphylococcus aureus were the top threepathogenic bacteria.2. Univariate analysis showed that there were statistically difference between the age ofchildren, total burned surface area, three degree burned surface area, and early treatment(including early anti-shock, early wound treatment, early enteral nutrition) among sepsis andnon-sepsis children. Multivariate analysis confirm that total burned surface area, three degreeburned surface area, and improper early treatment are the risk factors (OR>1) of sepsis forburned children.3. Case-paired analysis different early treatment, showed:(1) At same age, gender and extent of burn, children had high incidence rate of sepsis(58.00%vs10.00%, p=0.000), and high hospital mortality (12.0%vs0.0%, p=0.035), if theyreceived improper early treatment. (2) At same age, gender and extent of burn, comparing with the children who receivedcorrect early treatment, the children received improper early treatment need more operations(0.73±0.80vs0.48±0.75, p=0.012), and showed longer average wound healing time (25.15±14.10vs18.87±7.99, p=0.000) days,consumpted more antibiotics in unit of weight (2.91±1.79vs2.07±1.17, p=0.001) DDDs/kg, more average blood plasma (1308.33±1379.12vs1078.33±796.36, p=0.008) ml, and more total red blood cells transfusions (77.75vs38.5)unit.Conclusions1. According to our clinical data, the severely burned children with sepsis is alwaysyounger than5years old, with total burned surface area <50%TBSA and three degree burnedsurface area <30%TBSA and hydrothermal burn.2. Total burned surface area, three degree burned surface area, improper early treatmentare the independent risk factors for sepsis in severely burned children.3. Improper early treatment may increase the incidence of sepsis, hospital mortality, costof treatment in severely burned children.