脓毒症合并急性肾损伤生存预后列线图预测模型的构建与验证

Construction and validation of nomogram prediction model for survival prognosis of sepsis combined with acute kidney injury

目的探讨脓毒症合并急性肾损伤(AKI)生存预后影响因素并建立预后列线图预测模型。方法回顾性纳入192例脓毒症合并AKI患者,根据随访截止时的预后情况将其分为死亡组88例和存活组104例,收集其临床资料和随访情况并分组进行比较。采用多因素logistic回归分析评估脓毒症合并AKI预后的影响因素。建立脓毒症合并AKI预后列线图预测模型,运用Bootstrap法对模型进行内部验证,经Hosmer-Lemeshow拟合优度检验,采用受试者工作特征(ROC)曲线、校正曲线及一致性指数(CI)评定模型的预测价值。结果死亡组AKI 3期、合并低蛋白血症、器官衰竭数量>3个、机械通气患者比例及肌酐(Cr)、辅助性T细胞(Th)17、IL-17水平、液体平衡量、ICU住院时间均高于存活组,AKI 1期患者比例及血清白蛋白(ALB)水平均低于存活组(P<0.05)。多因素logistic回归分析结果显示,AKI分期(2期、3期)、ICU住院时间、Th17、IL-17水平及液体平衡量是脓毒症合并AKI预后的独立影响因素(P<0.05)。建立脓毒症合并AKI预后的列线图预测模型,经内部验证,CI=0.864,其预测脓毒症合并AKI预后的校正曲线趋势接近理想曲线。预测模型预测脓毒症合并AKI预后的ROC曲线下面积(AUC)为0.864。经Hosmer-Lemeshow检验,该模型的风险预测值与实际值的P=0.170。结论脓毒症合并AKI的预后受多种因素影响,基于独立影响因素建立的列线图预测模型预测效能较好,具有一定临床适用性。

Objective To investigate the factors influencing the survival prognosis of sepsis combined with acute kidney injury(AKI)and establish nomogram prediction model for prognosis.Methods A total of 192 patients with sepsis combined with AKI were retrospectively included and divided into death group(88 cases)and survival group(104 cases)according to the prognosis conditions at the end of follow-up.Clinical data and follow-up information were collected and grouped for comparison.Multivariate logistic regression analysis was used to evaluate the influencing factors of prognosis of sepsis combined with AKI.Nomogram prediction model for prognosis of sepsis combined with AKI was established.Bootstrap method was used to verify the model internally.After Hosmer-Lemeshow goodness of fit test,receiver operating characteristic(ROC)curve,calibration curve and consistency index(CI)were used to evaluate the predictive value of the model.Results In the death group,the proportion of patients with AKI stage 3,combined with hypoproteinemia,number of organ failure>3,mechanical ventilation,the levels of Cr,Th17,IL-17 and fluid balance volume,time of stay in ICU were higher than those in survival group,the proportion of patients with AKI stage 1 and level of ALB were lower than those in survival group(P<0.05).Multivariate logistic regression analysis showed that AKI stage(stage 2,stage 3),time of stay in ICU,Th17,IL-17 and fluid balance volume were independent factors influencing the prognosis of sepsis combined with AKI(P<0.05).The nomogram prediction model for prognosis of sepsis combined with AKI was established.The internal verification showed that CI=0.864,and the trend of the correction curve for predicting the prognosis of sepsis combined with AKI was close to the ideal curve.The area under ROC curve(AUC)of the predictive model for the prognosis of sepsis with AKI was 0.864.According to Hosmer-Lemeshow test,the predicted risk value and the actual risk value of the model were P=0.170.Conclusion The prognosis of sepsis combined with AKI is affected by many factors.The nomogram prediction model based on the independent influencing factors has better prediction efficiency and certain clinical applicability.