目的探讨列线图个体化危重症孕产妇发生急性肾损伤(AKI)风险模型的建立。方法将320例危重症孕产妇患者按照是否发生AKI分为发生AKI组和未发生AKI组,调取所有患者的基本信息及临床资料,采用Logistic回归分析筛选危重症孕产妇发生AKI的独立危险因素。然后将筛选出的独立危险因素使用R软件建立列线图预测模型,并对模型的预测性及准确度进行验证。结果通过对两组患者一般临床资料做Logistic回归分析可知,高血压(OR=2.509,95%CI:1.412～4.459)、肝功能损伤(OR=2.741,95%CI:1.542～4.870)、乳酸升高(OR=2.920,95%CI:1.650～5.166)、休克(OR=7.171,95%CI:2.431～21.152)、APACHEⅡ评分(OR=2.352,95%CI:1.274～4.341)、SOFA评分(OR=5.189,95%CI:2.812～9.576)为危重症孕产妇并发AKI的独立危险因素,差异具有统计学意义(P<0.05),均与危重症孕产妇并发AKI高度相关。基于高血压、肝功能损伤、乳酸升高、休克、APACHEⅡ评分及SOFA评分6项危重症孕产妇并发AKI的独立危险因素,建立预测危重症孕产妇并发AKI风险的列线图模型,并对该模型进行验证,预测值同实测值基本一致,说明本研究的列线图预测模型具有较好的预测能力。同时,本研究使用Bootstrap内部验证法对危重症孕产妇并发AKI的列线图模型进行验证,C-index指数高达0.829(95%CI:0.801～0.857)。说明本研究列线图模型具有良好的精准度和区分度。结论对危重症孕产妇及时的考虑高血压、肝功能损伤、休克、乳酸升高、APACHEⅡ评分及SOFA评分等因素综合评估危重症孕产妇并发AKI的发生率,能够提高对危重症孕产妇继发AKI的诊断效能,具有较高的临床应用价值,值得进一步推广使用。

Objective To explore the establishment of a risk-predicting model with nomogram for acute kidney injury(AKI) in critically ill pregnant women.Method 320 critically ill pregnant women were divided into the AKI group and non-AKI group, and the basic information and clinical data of all patients were retrieved. Logistic regression analysis was used to analyze the independent risk factors of AKI in critically ill pregnant women. The selected independent risk factors were used to establish a nomogram prediction model with R software, and the predictability and accuracy of the model will be verified.Result Logistic regression analysis of the general clinical data of the two groups showed that hypertension(OR= 2. 509,95%CI:1. 412-4. 459), liver function damage(OR=2. 741,95%CI:1. 542-4. 870), increased lactic acid(OR= 2. 920,95%CI:1. 650-5. 166), shock(OR= 7. 171,95%CI:2. 431-21. 152), APACHE Ⅱ score(OR= 2. 352,95%CI:1. 274-4. 341), SOFA score(OR= 5. 189,95%CI:2. 812-9. 576) ere independent risk factors for AKI in critically ill pregnant women, with statistical difference(P< 0. 05), all of which were highly correlated with critical maternal AKI. Based on these six independent risk factors for AKI( hypertension, liver function damage,shock, elevated lactate, APACHE II score and SOFA score), a nomogram model for predicting the risk of AKI in critically ill pregnant women was established an verified. The predicted values are basically consistent with the measured values, indicating that the nomogram model of this study has better predictive ability. Meanwhile, Bootstrap internal verification method was used to verify AKI in critically ill pregnant women, the C-index index is as high as 0. 829(95%CI: 0. 801-0. 857), indicating that the nomogram model of this study has good quality precision and discrimination.Conclusion Full consideration of factors as hypertension, liver function damage, shock, elevated lactate, APACHE II score and SOFA score in critically ill pregnant women can assess the incidence of AKI in this population comprehensively;it can improve the diagnostic efficiency of AKI in critically ill pregnant women. The model can be applicated actively in clinical medicine and deserves to be promoted.