The Research On Risk Assessment Of Extensively Burned Patients And Urine Output Automatic Monitoring

Background:Acute burn injury is one of the most severe injuries and has the highest mortality.Therefore,the risk of death in burn patients should be assessed at an early stage and accoding to the risk,strategy of treatment should be modified as early as possible.MODS caused by shock is the main cause of death since 1980 s.Shock is not only an important complication but also an important factor inducing other complications.Although burn shock occurs in the early stage,it has a great impact on the whole treatment process.Fluid resuscitation is the one of key treatments for extensively burned patients.The effects of fluid resuscitation and correction of tissue cells ischaemia and hypoxia damage will directly affect the prognosis.Automated fluid resuscitation can avoid insufficient or excessive transfusion volume during shock,which will help extensively burned patients pass the shock period smoothly.Objective:To identify the risk factors of death and establish death risk nomogram model,providing warning for the severity of extensively burned patients and explore the impact of shock state on prognosis,providing reference for clinical shock state judgement;to compare the accuracy of measuring urine output in hours manually with automaticly for foundation of burn shock automatic fluid resusitation.Methods:1 The research on shock risk assessment of extensively burned patients1.1 The data source of a nomogram model for predicting the death risk of extensively burned patients231 extensively burned patients admitted to Burn Research Institute of First Affiliated Hospital of Army Medical University from January 2010 to October 2018 were investigated retrospectively.Demographic data(age,sex),injury related data(total burn surface area,inhalation injury,full-thickness burns,calculated SI,ABSI,Baux score and BI),shock related data(pre-hospital fluid infusion(yes or no),heart rate,systolic and diastolic blood pressure at admission,fluid and urine volume of the first 24 H and the second 24 H after admission,SI,BE and HCT at admission),prognosis,organs damage and treatment related data(using MV and CRRT or not,prognosis: cured or died)were collected.1.2 The data source of cluster analysis of shock in extensively burned patients118 extensively burned patients without missing indicators and discharged automatically within 48 hours admitted to Burn Research Institute of First Affiliated Hospital of Army Medical University from January 2010 to October 2018 were investigated retrospectively.Demographic data(age,sex),injury related data(total burn surface area,inhalation injury,superficial partial-thickness burn,deep partial-thickness burn,full-thickness burn,calculated BI),shock related data(shock diagnosis(yes or no),time until presentation,heart rate and systolic blood pressure at admission,HB,LAC,BE,HCT at admission,calculated MAP,SI),prognosis,organs damage and treatment related data(using MV and CRRT or not,complicated with SPESIS and MODS or not,prognosis: cured or died)were collected.2 Urine output automatic monitoring2.1 Accuracy of urine output dynamic monitoring and its clinical applicabilityA total of 68 burned patients admitted to Institute of Burns of the First Affiliated Hospital from September 2017 to July 2018 were prospective recruited in this study.All these patients were required for hourly monitoring of urine output under doctor’s orders.According to the time of admission,they were randomly assigned into manual monitoring group(control group,n=34)and device monitoring group(trial group,n=34).Their data of hourly urine output were collected for 8 h,and those from the control group were conducted by a nurse visually reading the scale on the disposable urine bag and then re-measuring the volume with a measuring cylinder.The volume of the trial group were dynamically recorded by the monitor,and also re-measured by a measuring cylinder.The hourly urine volume at each time point was compared between the 2 groups,and with the standard volume measured by the cylinder.Statistical analysis:SPSS statistical software was used for data description and analysis.T-test,chi-square test or Wilcoxon rank sum test,Kolmogorov-Smirnov test,Mann-Whitney test and linear regression analysis were performed on the data.Logistic regression model was used to determine the risk factors of death in patients with severe burns.A nomogram model was established by R software.Bootstrap method was used to validate the model internally.The NbClust package of R3.5.2 software is used to determine the optimal cluster number,and the K-means cluster method is used to classify the classes according to the determined number of classes.Results:1 The research on shock risk assessment of extensively burned patients1.1 A nomogram model for predicting the death risk of extensively burned patientsUnivariate analysis showed that the degree of inhalation injury,TBSA,24-hour fluid infusion after injury,full-thickness burns,mechanical ventilation,CRRT,BI,ABSI,Baux score were significantly correlated with the extensively burned patients’ death(P < 0.05),and other inclusion indicators had no significant difference.The results of single factor analysis with statistical significance(P<0.05)were included in multivariate logistic regression analysis,suggesting mechanical ventilation(regression coefficient: 1.187,standard error: 0.57,Wald value: 4.342,OR: 3.277,95% confidence interval: 1.073-10.008,P<0.05),CRRT(regression coefficient: 1.72,standard error: 0.435,Wald value: 15.672,OR: 5.587,95% confidence interval: 2.384-13.093,P < 0.001)and burn index(regression coefficient: 0.064,standard error: 0.014,Wald value: 22.043,OR:1.067,95% confidence interval: 1.038-1.096,P < 0.001)were independent risk factors for extensively burned patients’ death.There were significant differences in mortality,burn severity,incidence of inhalation injury and severity of inhalation injury between the two groups treated with MV and without MV(P < 0.001).There were significant differences in mortality and burn severity between the two groups treated with CRRT and without CRRT(P < 0.001).The prediction model of extensively burned patients’ death risk was drawn based on the three independent risk factors.Through the Bootstrap self-sampling method,the resolution of the nomogram model was verified 1000 times.The initial C-index was 0.90,and the corrected C-index was 0.89.According to the nomogram model,231 extensively burned patients were scored and the ROC curve was drawn.The optimum threshold of ROC was 0.23.The sensitivity and specificity of this point were 0.86 and 0.80 respectively.The area under ROC curve is 0.90(95% confidence interval 0.86-0.942).1.2 Cluster analysis of shock in extensively burned patientsBI,full-thickness burns,TBSA,SI and HB at admission played an important role in grouping 118 cases of extensively burned patients.The indexes in group 1 were BI(72.93 ±10.70),SI(1.00±0.26),HB(186.16±25.52),CRRT(58.1%),MODS(39.5%)and mortality rate(51.2%).The indexes of group 2 were BI(40.07±9.30),SI(0.88±0.26),HB(175.17±23.92),CRRT(18.7%),MODS(8%)and mortality rate(12%).There were significant differences in the incidence of inhalation injury and MV,prognostic indicators such as mortality,MODS incidence and CRRT treatment rate(P < 0.001).But there was no difference in the number of the days of seriously or critically ill and the incidence of SPESIS between the two groups.Group 1 had high SI,95% confidence interval was 0.92-1.07,while group 2 had low SI,95% confidence interval was 0.82-0.94.0.95 was recommended as an auxiliary diagnostic value for burn shock.Group 2 was at low level of HB,95% confidence interval was 169.67-180.68,while group 1 was at high level for the first time,95% confidence interval was 178.31-194.02.The low value of group 1 was 178g/L and the high value of group 2 was 180g/L.It is suggested that HB(≥179g/L)be used as an auxiliary diagnostic criterion for shock in extensively burned patients.75.4 2% of the extensively burned patients were diagnosed as shock with SI(≥0.95)or HB(≥179 g/L),32.20% of the extensively burned patients were diagnosed as shock by doctors,and there were statistical differences in the consistency of shock diagnosis(P < 0.05).The proportion of extensively burned patients diagnosed as shock with MAP < 65 mmHg was 2.54%,32.20% of the extensively burned patients were diagnosed as shock by doctors,and there were not statistical differences in the consistency of shock diagnosis(P>0.05).2 Urine output automatic monitoring2.1 Accuracy of urine output dynamic monitoring and its clinical applicabilityThe median of measurement deviation was 15.00 mL and the measurement deviation rate was 17.75% in the control group,which was significantly higher than those of the experimental group(2.00 mL and 1.71%,P<0.001).The trial group had a better resolvable coefficient(R2= 0.999),and its slope of the regression equation(0.988)was closer to 1,when compared with the control group(R2=0.860,0.890).The percentage within the acceptable error range was significantly higher in the trial group than the control group(93.75% vs 14.00%,P<0.001).Among the 272 time points of monitoring,punctual measuring was only 34 times(12.50%)in the control group,and the maximal time error was 20(median 10).But for the trial group,the punctual measuring was obtained at every time point.Chi-square test showed that statistical difference was found in the punctuality ratio of measurement time between the 2 groups(P<0.001).Conclusions:1.The data of 231 extensively burned patients aged 18-60 admitted within 48 hours after burn to the First Affiliated Hospital of Military Medical University from January 2010 to October 2018 was retrospectivly studied.The death risk factors of the patients were identified and a nomogram model for predicting the death risk was established based on the death risk factors.According to the nomogram model,the corresponding score of death risk of burned patients could be obtained which would provide a reference for clinical treatment adjustment.2.The study also retrospectively analyzed the data of 118 extensively burned patients without missing indicators and discharged automatically within 48 hours.It is proved that BI,SI and HB at admission determined the different shock states and prognosis of two groups.The division of SI and HB at admission is 95% confidence interval upper limit of non-shock group(SI ≥0.95 or HB ≥179g/L)which has certain clinical value in screening extensively burned shock patients.3.The urine output dynamic monitor has much higher accuracy(the measurement deviation rate was 1.71%)than manual monitoring(the measurement median deviation rate was 17.75%).The urine output dynamic monitor can measure the urine volume on time,while the punctual measuring of the manual monitoring control group is only 12.5%.The urine output dynamic monitor has better clinical applicability and is worthy of promotion in clinical practice.