Development And Clinical Application Assessment Of A Prognostic Model For Adverse Outcomes In Geriatric Trauma Patients

1.Predictive value of GTOS,RTS and APACHE II scores for adverseoutcomes in geriatric trauma patients Background: The incidence of adverse outcomes is higher in geriatric trauma patients than in young patients,which imposes a huge burden on society and families.Currently,the prognosis assessment of trauma patients mainly relies on various trauma scores,among which the Revised Trauma Score(RTS),Geriatric Trauma Outcome Score(GTOS)and Acute Physiology and Chronic Health Evaluation II(APACHE II)are widely used,but their predictive value for geriatric trauma population hasn’t been validated.Aim: Our aim was to assess the predictive value of GTOS,RTS and APACHE II scores for adverse outcomes in geriatric trauma patients.Methods: Data of patients aged 65 years and above admitted to the department of trauma surgery,Tongji Hospital,from June 1,2016 to May 31,2020 were retrospectively analyzed.Patient data included basic characteristics,trauma information and scores,and occurrence of adverse outcomes.Multivariate logistic analysis was performed to determine whether trauma scores were independently associated with adverse outcomes.Receiver Operating Characteristic(ROC)curves and calibration curves were used to determine the accuracy and calibration of GTOS,RTS and APACHE II scores in predicting adverse outcomes.Hosmer-Lemeshow goodness of fit test was performed for trauma scores.The Delong method tested whether the difference in predictive accuracy between trauma scores was statistically significant.Results: A total of 485 patients were enrolled,307(63.3%)had adverse outcomes included in the case group,and the remaining 178(36.7%)were included in the control group.The GTOS,RTS and APACHE II scores of patients in the case group were127.2±21.1,6.27±1.44 and 20.2±5.4,respectively,while the GTOS,RTS and APACHE II scores of the control group were 103.6±16.1,7.39±0.81 and 13.8±5.2,respectively.The trauma scores were statistically different between the two groups.GTOS(OR 1.04,95%CI 1.03-1.06,P<0.001),RTS(OR 0.62,95% CI 0.47-0.80,P<0.001)and APACHE II score(OR 1.17,95% CI 1.11-1.24,P<0.001)were all independently associated with adverse outcomes.The AUC of GTOS score was 0.82(0.78-0.86),with a Hosmer-Lemeshow goodness-of-fit test of P<0.05.The AUC of RTS score was 0.76(0.72-0.80),with a Hosmer-Lemeshow goodness-of-fit test of P=0.441.The AUC of RTS score was 0.80(0.76-0.84),with a Hosmer-Lemeshow goodness-of-fit test of P=0.067.Conclusion: GTOS,RTS,and APACHE II scores are all valid tools for predicting adverse outcomes in geriatric trauma patients,with only moderate predictive accuracy.2.Development and clinical application assessment of a prognostic model based on frailty syndrome Background: Although GTOS,RTS and APACHE II scores are valid tools for predicting adverse outcomes in elderly trauma patients,the predictive accuracy is only moderate.Frailty is independently associated with adverse outcomes in the course of multiple diseases in the elderly population,including trauma.We hypothesized that developing a prognostic model based on frailty could improve the predictive accuracy of adverse outcomes.In addition,there is a lack of rapid and valid tools to diagnose frailty in geriatric trauma patients.Mid-Arm Muscle Circumference(MAMC),a simple and objectively measurable index,has been associated with multimorbidity adverse prognosis in geriatric patients.However,whether MAMC can be used to screen for frailty remains unexplored.Aim:1)to assess the applicability of MAMC to identify frailty status and determine the optimal gender-specific cut-off value;2)to validate frailty as an independent risk factor for adverse outcomes during hospitalization in geriatric trauma patients;and 3)to develop a prognostic model for geriatric trauma patients and assess its clinical application.Methods: Elderly patients admitted to the department of trauma surgery,Tongji Hospital,were selected from June 2020 to September 2021.Patient data included basic characteristics,trauma information and scores,anthropometric parameters,laboratory test indexes,adverse outcomes,length of hospitalization and physiological function at discharge.Frailty status was assessed by the Trauma Specific Frailty Index(TSFI)and MAMC,respectively.Pearson coefficients were calculated to test the correlation between TSFI and MAMC.ROC curve analysis was performed to determine the best cut-off value of MAMC for the diagnosis of frailty and then grouping.Baseline characteristics and prognostic differences between the frailty and non-frailty groups were compared.Multivariate logistic regression was used to analyze whether frailty was independently associated with adverse outcomes.Patients were randomized into training and test sets by3:1.Univariate logistic(P≤0.2 variables were entered into multivariate logistic analysis)and multivariate logistic regression were used to analyze the independent correlates of adverse outcomes.Prognostic model was constructed in the training set.ROC and calibration curves were used to determine the accuracy and calibration of the prognostic model.Hosmer-lemeshow goodness-of-fit test for the model.The Delong method was used to test whether the difference in prediction accuracy between the prognostic model and trauma scores was statistically significant.Clinical decision curves to assess the net clinical benefit of the prognostic model.Results: A total of 209 patients were enrolled in the study.MAMC was negatively correlated with TSFI(r=-0.68,P<0.001 for men;r=-0.73,P<0.001 for women)with an optimal cut-off value of 24.1 cm for men and 22.7 cm for women.82 patients were included in the frailty group and the remaining 127 patients were included in the non-frailty group.Except for unplanned operation(P=0.681),frail patients had a higher risk of adverse outcomes,more length of stay(24.5±11.2 days vs.19.5±11.8 days,P=0.005),and poorer physiological function at discharge than non-frailty patients(GOS-E,4(2-6)vs.7(6-7),P<0.001).Except for gastrointestinal bleeding(OR 2.09,95% CI0.52-9.21,P=0.302),unplanned operation(OR 1.07,95% CI 0.13-7.54,P=0.943),and unplanned transfer to ICU(OR 7.08,95% CI 0.91-148.01,P=0.097),frailty was an independent risk factor for the adverse outcomes.The independent predictors of adverse events were female(OR 0.16;95% CI 0.05-0.47),a CCI(OR 4.55;95% CI 2.30-9.86),frailty(OR 4.19;95% CI 1.24-15.25),GCS(OR 0.82;95% CI 0.67-0.99),ISS(OR 1.13;95% CI 1.00-1.28),blood transfusion unit(OR 1.54;95% CI 1.15-2.11),albumin(OR0.83;95% CI 0.71-0.96),and lactate(OR 2.68;95% CI 1.67-4.66).The AUC of the model was 0.92(0.84-0.99),with a Hosmer-Lemeshow goodness-of-fit test of P=0.193 and a Brier score of 0.12(0.07-0.17),outperformed GTOS,RTS and APACHE II scores.The prognostic model provided higher net benefit at high-risk thresholds of 0.4 to 0.8.Conclusion: MAMC is a simple and valid indicator to identify the frailty status of geriatric trauma patients.Frailty is an independent risk factor for adverse outcomes.The prognostic model constructed in this study has high calibration and accuracy,which is beneficial for early identification of high-risk patients and early intervention to improve their prognosis.