急症腹腔镜胆囊切除术中转开腹的预警模型构建与验证

Construction and verification of an early warning model for switching to laparotomy during emergency laparoscopic cholecystectomy

目的:探讨急症腹腔镜胆囊切除术中转开腹的危险因素,建立个体化列线图预警模型并进行验证。方法:回顾性选取2018年2月至2020年12月584例行腹腔镜胆囊切除术的急性胆囊炎患者,按照7∶3比例随机分为建模组(n=408)与验证组(n=176),采用Logistic回归模型分析变量,建立中转开腹列线图预警模型,采用验证组对预警模型的预测效能进行外部验证。结果:建模组与验证组仅在胆囊壁厚度、胆囊炎发作次数、术前总胆红素水平方面差异有统计学意义(P<0.05),其余指标差异均无统计学意义(P>0.05);Logistic多因素分析显示,胆囊壁厚度≥5 mm、胆囊炎发作次数≥3次、急性胆囊炎分型为坏疽型/穿孔、发病至手术时间≥72 h、术前总胆红素高是腹腔镜胆囊切除术中转开腹的独立危险因素(P<0.05);对列线图预警模型进行验证,建模组与验证组受试者工作特征曲线下面积分别为0.872、0.843,且两组校正曲线均显示预测的中转开腹风险与实际中转开腹风险的相关性良好;Hosmer-Lemeshow拟合优度检验χ^(2)=6.629,P=0.577;临床决策曲线分析显示阈概率为0.13~0.85,模型净收益率高于0。结论:列线图预警模型可准确实现对急症腹腔镜胆囊切除术中转开腹的个体化预测,能更有效地帮助医护人员将中转开腹的高危患者筛选出来,以提供更精准的临床治疗与护理方案。

Objective:To explore the risk factors of converting to laparotomy during emergency laparoscopic cholecystectomy(LC),to develop and verify an individualized nomogram early warning model.Methods:From Feb.2018 to Dec.2020,584 patients who suffered from acute cholecystitis and underwent LC were retrospectively selected and randomly divided into modeling group(n=408)and verification group(n=176)according to a 7∶3 ratio.The logistic regression model was used to analyze the variables,the nomogram early warning model of switching to laparotomy was established,and the verification group was used to externally verify the predictive effectiveness of the early warning model.Results:The general data of the modeling group and the verification group only differed in the thickness of the gallbladder wall,the number of cholecystitis episodes,and the level of total bilirubin before surgery(P<0.05),and there was no statistical difference in other indicators(P>0.05).Logistic multivariate analysis showed that the gallbladder wall thickness≥5 mm,the number of cholecystitis episodes≥3 times,the type of gangrene/perforation acute cholecystitis,time from onset to operation≥72 h and the higher preoperative total bilirubin were independent risk factors for LC conversion to laparotomy(P<0.05).In the validation of the nomogram early warning model,the area under the receiver operator characteristic curve of the modeling group and the verification group was 0.872 and 0.843,respectively.And the calibration curves of the two groups showed that the predicted risk of conversion to laparotomy correlated well with the actual risk of conversion to laparotomy.The model passed Hosmer-Lemeshow goodness of fit testχ^(2)=6.629,P=0.577;decision curve analysis showed that the threshold probability was 0.13-0.85,and the net income ratio of model was higher than 0.Conclusions:The nomogram early warning model can accurately realize the individualized prediction of emergency LC conversion to laparotomy,it can more effectively help medical staff to screen out high-risk patients of conversion to laparotomy,and provide more accurate plans of clinical treatment and nursing.