个体化预测重症急性胰腺炎患者并发急性肾损伤风险列线图模型的建立及验证

Establishment and validation of an individualized column chart model for predicting the risk of acute kidney injury in patients with severe acute pancreatitis

目的构建预测重症急性胰腺炎患者并发急性肾损伤风险的列线图模型并进行验证。方法:回顾性选取本院于2021年1月~2023年6月期间收治的355例重症急性胰腺炎患者为研究对象,照7:3分为建模组(n=249)及验证组(n=106),另根据建模组患者有无急性肾损伤分为无急性肾损伤组(n=185)和急性肾损伤组(n=64)。对无急性肾损伤组和急性肾损伤组临床资料进行比较,将其中差异具有统计学意义的指标纳入多因素Logistic回归模型,确定影响重症急性胰腺炎患者发生急性肾损伤的危险因素,构建预测重症急性胰腺炎患者并发急性肾损伤的列线图模型。结果:与无急性肾损伤组比较,急性肾损伤组血肌酐(sCr)、血清尿素氮(BUN)、C反应蛋白(CRP)、中性粒细胞与淋巴细胞比值(NLR)、胸水和APACHE Ⅱ评分≥10分的比例显著升高(t/×^(2)=10.411、8.897、13.893、10.065、13.151、14.987,P<0.05);多因素Logistic回归分析结果显示,sCr、BUN、CRP、NLR水平和APACHEI评分是影响重症急性胰腺炎患者并发急性肾损伤独立危险因素(OR=1.039、2.078、1.688、2.095、4.700,P<0.05);R0C曲线显示,建模组曲线下面积为0.892(95%CI为0.853~0.931),验证组曲线下面积为0.904(95%CI:0.867~0.940);建模组和验证组校准曲线斜率均接近于1,且Hosmer~Lemeshow拟合优度检验显示,建模组×^(2)=6.860,P=0.552,验证组×^(2)=7.893,P=0.444。结论:sCr、BUN、CRP、NLR水平和APACHEⅡ评分是影响重症胰腺炎患者并发急性肾损伤独立危险因素,基于sCr、BUN、CRP、NLR水平和APACHEⅡ评分构建的列线图模型能够指导重症胰腺炎患者并发急性肾损伤的个体化预测。

Objective To construct a column chart model for predicting the risk of acute kidney injury in patients with severe acute pancreatitis and validate it.Methods A retrospective study was conducted on 355 patients with severe acute pancreatitis admit-ted to our hospital from January 2021 to June 2023.They were grouped into a modeling group(n=249)and a validation group(n=106)at 7:3.The modeling group was also separated into a non acute kidney injury group(n=185)and an acute kidney injury group(n=64)based on whether the patients had acute kidney injury.The clinical data were compared between the non acute kidney injury group and the acute kidney injury group,the indicators with statistically obvious differences were included in a multivariate logistic re-gression model,risk factors affecting the occurrence of acute kidney injury in patients with severe acute pancreatitis were identified,and a column chart model was constructed to predict the occurrence of acute kidney injury in patients with severe acute pancreatitis.Results Compared with the non acute kidney injury group,the serum creatinine(sCr),serum urea nitrogen(BUN),C-reactive pro-tein(CRP),neutrophil to lymphocyte ratio(NLR),pleural effusion,and proportion of APACHE Ⅱscore≥10 points in the acute kidney injury group obviously increased(t/×^(2)=10.411,8.897,13.893,10.065,13.151,14.987,P<0.05);the results of multi-variate Logistic regression analysis showed that sCr,BUN,CRP,NLR levels,and APACHE II score were independent risk factors for severe acute pancreatitis patients complicated with acute kidney injury(0R=1.039,2.078,1.688,2.095,4.700,P<0.05);the ROC curve showed that the area under curve of the modeling group was 0.892(95%CI:0.853~0.931),while the area under curve of the validation group was 0.904(95%Cl:0.867~0.940);the slopes of the calibration curves of both the modeling and validation groups were close to 1,and the Hosmer Limeshow goodness of fit test showed that the modeling group ×^(2)=6.860,P=0.552,and the validation group ×^(2)=7.893,P=0.444.Conclusion SCr,BUN,CRP,NLR levels,and APACHE Ⅱ score are independent risk fac-tors for acute kidney injury in patients with severe pancreatitis.A column chart model constructed based on sCr,BUN,CRP,NLR lev-els,and APACHE Ⅱ score can guide individualized prediction of acute kidney injury in patients wth severe pancreatitis.