严重烧伤患者并发肺动脉栓塞风险预测模型的建立与验证

OBJECTIVE: To screen the risk factors for pulmonary embolism in severe burn patients, based on which, a risk prediction model was established and validated. METHODS: This study was a retrospective case series study. The clinical data of 267 severe burn patients who met the inclusion criteria and were admitted to the Department of Burns of Wuhan Third Hospital from March 2020 to March 2023 were collected, including 159 males and 108 females, aged 18-82 years. The patients were divided into pulmonary embolism group (26 cases) and non-pulmonary embolism group (241 cases) according to whether they were complicated with pulmonary embolism. The following data of patients in the 2 groups were collected and compared, including gender, age, body mass index, bedtime during treatment, cause of burn, albumin level on admission, combination of chronic obstructive pulmonary disease (COPD), combination of diabetes mellitus, combination of hypertension, combination of inhalation injury, and the abbreviated burn severity index (ABSI) on admission. The indicators with statistically significant differences between the two groups were conducted with univariate and multivariate logistic regression analyses to identify the independent risk factors for pulmonary embolism in 267 severe burn patients. Based on these findings, a nomogram prediction model was established. The performance of the prediction model was evaluated by the receiver operating characteristic (ROC) curve, while its validation was conducted through calibration curve and clinical decision curve analysis. RESULTS: The proportions of beyond 60 years old, bedtime over 7 days during treatment, combination of COPD, and combination of diabetes mellitus (with χ(2) values of 7.75, 29.15, 29.86, and 5.94, respectively), and ABSI score on admission (t=6.01) of patients in pulmonary embolism group were significantly higher than those in non-pulmonary embolism group (P < 0.05). There were no statistically significant differences in the other indicators between the two groups of patients (P > 0.05). The univariate logistic regression analysis showed that age, bedtime during treatment, combination of COPD, combination of diabetes mellitus, and ABSI score on admission were the risk factors for pulmonary embolism in severe burn patients (with odds ratios of 3.40, 14.87, 17.78, 2.80, and 1.88, respectively, 95% confidence intervals of 1.38-8.39, 4.34-50.98, 4.63-68.22, 1.19-6.58, and 1.47-2.41, respectively, P < 0.05). The multivariate logistic regression analysis showed that bedtime over 7 days during treatment, combination of COPD, and high ABSI score on admission were the independent risk factors for pulmonary embolism in severe burn patients (with odds ratios of 11.02, 30.82, and 1.86, respectively, 95% confidence intervals of 2.76-43.98, 3.55-267.33, and 1.38-2.50, respectively, P < 0.05). Based on the three aforementioned independent risk factors, a nomogram prediction model for the risk of pulmonary embolism in severe burn patients was established. The ROC curve of prediction model showed that the area under the ROC curve was 0.91 (with 95% confidence interval of 0.82-0.99). When the optimal cut-off value of 25% was taken, the sensitivity and specificity of prediction model was 84.6% and 93.4%, respectively. The calibration curve showed that the calibration curve of prediction model was around the ideal curve, with a consistency index of 0.80 in Cox regression (with 95% confidence interval of 0.74-0.87). The clinical decision curve showed that the threshold probability value of the prediction model was in the range of 1% to 98%, with net return rate over 0. CONCLUSIONS: The independent risk factors for pulmonary embolism in severe burn patients include bedtime over 7 days during treatment, combination of COPD, and high ABSI score on admission. The nomogram prediction model established based on this has good predictive value for complicated pulmonary embolism in severe burn patients.