Analysis of independent risk factors and establishment and validation of a prediction model for in-hospital mortality of multiple trauma patients
10.3760/cma.j.cn501098-20230330-00179
- VernacularTitle:多发伤患者院内死亡独立危险因素分析及预测模型的构建与验证
- Author:
Zhenjun MIAO
1
;
Dengkui ZHANG
;
Yapeng LIANG
;
Feng ZHOU
;
Zhizhen LIU
;
Huazhong CAI
Author Information
1. 江苏大学附属医院急诊中心，镇江 212001
- Keywords:
Multiple trauma;
Death;
Lactic acid;
Risk factors;
Nomograms
- From:
Chinese Journal of Trauma
2023;39(7):643-651
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To explore the independent risk factor for in-hospital mortality of patients with multiple trauma, and to construct a prediction model of risk of death and validate its efficacy.Methods:A retrospective cohort study was performed to analyze the clinical data of 1 028 patients with multiple trauma admitted to Affiliated Hospital of Jiangsu University from January 2011 to December 2021. There were 765 males and 263 females, aged 18-91 years[(53.8±12.4)years]. The injury severity score (ISS) was 16-57 points [(26.3±7.6)points]. There were 153 deaths and 875 survivals. A total of 777 patients were enrolled as the training set from January 2011 to December 2018 for building the prediction model, while another 251 patients were enrolled as validation set from January 2019 to December 2021. According to the outcomes, the training set was divided into the non-survival group (115 patients) and survival group (662 patients). The two groups were compared in terms of the gender, age, underlying disease, injury mechanism, head and neck injury, maxillofacial injury, chest injury, abdominal injury, extremity and pelvis injury, body surface injury, damage control surgery, pre-hospital time, number of injury sites, Glasgow coma score (GCS), ISS, shock index, and laboratory test results within 6 hours on admission, including blood lactate acid, white blood cell counts, neutrophil to lymphocyte ratio (NLR), platelet counts, hemoglobin, activated partial thromboplastin time (APTT), fibrinogen, D-dimer and blood glucose. Univariate analysis and multivariate Logistic regression analysis were performed to determine the independent risk factors for in-hospital mortality in patients with multiple trauma. The R software was used to establish a nomogram prediction model based on the above risk factors. Area under the receiver operating characteristic (ROC) curve (AUC), calibration curve and clinical decision curve analysis (DCA) were plotted in the training set and the validation set, and Hosmer-Lemeshow goodness-of-fit test was performed.Results:Univariate analysis showed that abdominal injury, extremity and pelvis injury, damage control surgery, GCS, ISS, shock index, blood lactic acid, white blood cell counts, NLR, platelet counts, hemoglobin, APTT, fibrinogen, D-dimer and blood glucose were correlated with in-hospital mortality in patients with multiple trauma ( P<0.05 or 0.01). Logistic regression analysis showed that GCS≤8 points ( OR=1.99, 95% CI 1.12,3.53), ISS>25 points ( OR=7.39, 95% CI 3.50, 15.61), shock index>1.0 ( OR=3.43, 95% CI 1.94,6.08), blood lactic acid>2 mmol/L ( OR=9.84, 95% CI 4.97, 19.51), fibrinogen≤1.5 g/L ( OR=2.57, 95% CI 1.39,4.74) and blood glucose>10 mmol/L ( OR=3.49, 95% CI 2.03, 5.99) were significantly correlated with their in-hospital mortality ( P<0.05 or 0.01). The ROC of the nomogram prediction model indicated that AUC of the training set was 0.91 (95% CI 0.87, 0.93) and AUC of the validation set was 0.90 (95% CI 0.84, 0.95). The calibration curve showed that the predicted probability was consistent with the actual situation in both the training set and validation set. DCA showed that the nomogram prediction model presented excellent performance in predicting in-hospital mortality. In Hosmer-Lemeshow goodness-of-fit test, χ2 value of the training set was 9.69 ( P>0.05), with validation set of 9.16 ( P>0.05). Conclusions:GCS≤8 points, ISS>25 points, shock index>1.0, blood lactic acid>2 mmol/L, fibrinogen≤1.5 g/L and blood glucose>10 mmol/L are independent risk factors for in-hospital mortality in patients with multiple trauma. The nomogram prediction model based on these 6 predictive variables shows a good predictive performance, which can help clinicians comprehensively assess the patient′s condition and identify the high-risk population.
