摘要: |
目的 探讨儿童急性非单纯性阑尾炎发病的危险因素,并构建预测模型。方法 回顾性分析2018年1月至2019年12月于我院行腹腔镜阑尾切除术并经病理确诊的582例急性阑尾炎患儿的临床资料,按术中所见和病理类型将所有患儿分为非单纯性阑尾炎组(382例,包括化脓、坏疽和穿孔性阑尾炎)和单纯性阑尾炎组(200例)。比较两组患儿的临床资料,取差异有统计学意义的变量进行多因素logistic回归分析,筛选急性非单纯性阑尾炎发病的危险因素,并建立logistic回归预测模型。绘制ROC曲线评价预测模型对儿童急性非单纯性阑尾炎的诊断价值。收集2020年1月至12月于我院行阑尾切除手术的急性阑尾炎病例160例对预测模型的诊断价值进行前瞻性验证。结果 582例患儿中,382(65.6%)例为非单纯性阑尾炎。多因素logistic回归分析显示,白细胞计数(WBC;OR=1.305,95% CI 1.189~1.432,P<0.001)、中性粒细胞比例(NP;OR=1.046,95% CI 1.012~1.080,P=0.007)、CRP(OR=1.046,95% CI 1.010~1.082,P=0.011)、白蛋白(ALB;OR=0.973,95% CI 0.952~0.995,P=0.014)为非单纯性阑尾炎患儿发病的独立危险因素。基于这4个因素的logistic回归方程为logit(P)=—5.614+0.263×WBC+0.045×NP+0.045×CRP—0.027×ALB。概率模型为P=1/[1+Exp(—5.614+0.263×WBC+0.045×NP+0.045×CRP—0.027×ALB)],Hosmer-Lemeshow检验提示该预测模型有较好的拟合度(P=0.622)。预测模型ROC曲线的AUC为0.931(95% CI 0.896~0.957,P<0.001),灵敏度为0.830,特异度为0.905。前瞻性验证预测模型的灵敏度和特异度分别为0.882、0.820。结论 白细胞计数、中性粒细胞比例和CRP水平升高及低白蛋白血症是儿童急性非单纯性阑尾炎发病的独立危险因素,根据上述因素构建的预测模型具有良好的诊断价值,有助于急性单纯性阑尾炎与非单纯性阑尾炎的鉴别。 |
关键词: 急性阑尾炎 儿童 危险因素 诊断 预测模型 |
DOI:10.16781/j.CN31-2187/R.20210683 |
投稿时间:2021-07-10修订日期:2021-08-26 |
基金项目:武汉市卫生局科研项目(WG13B11). |
|
Risk factors and prediction model for acute non-simple appendicitis in children |
GUO Hongxi,YANG Jun*,BIAN Hongqiang,YANG Hu,GUO Qin,SUN Beibei |
(Department of General Surgery, Wuhan Children's Hospital (Wuhan Maternal and Child Healthcare Hospital), Tongji Medical College, Huazhong University of Science & Technology, Wuhan 430014, Hubei, China *Corresponding author) |
Abstract: |
Objective To explore the risk factors of acute non-simple appendicitis in children, and to establish a prediction model. Methods Retrospective analysis was performed with the clinical data of 582 children with acute appendicitis who underwent laparoscopic appendectomy and were diagnosed by pathology in our hospital from Jan. 2018 to Dec. 2019. According to the intraoperative findings and pathological types, the patients were divided into non-simple appendicitis group (382 cases, including suppurative, gangrene and perforation) and simple appendicitis group (200 cases). The clinical data of the 2 groups were compared and statistically significant variables were used for multivariate logistic regression analysis to identify risk factors of acute non-simple appendicitis, and finally to establish a logistic regression prediction model. The performance of the prediction model was assessed by receiver operating characteristic (ROC) curve. Prospective validation of the prediction model was conducted with 160 patients with acute appendicitis undergoing appendectomy in our hospital from Jan. to Dec. 2020. Results Of the 582 patients, 382 (65.6%) had non-simple appendicitis. Multivariate logistic regression analysis demonstrated that white blood cell (WBC) (odds ratio[OR]=1.305, 95% confidence interval[CI]1.189-1.432, P<0.001), neutrophil percentage (NP) (OR=1.046, 95% CI 1.012-1.080, P=0.007), C reactive protein (CRP) (OR=1.046, 95% CI 1.010-1.082, P=0.011), and albumin (ALB) (OR=0.973, 95% CI 0.952-0.995, P=0.014) were independent risk factors for non-simple appendicitis in children. The logistic regression equation based on the 4 factors was logit(P)= —5.614+0.263×WBC+0.045×NP+0.045×CRP—0.027×ALB. The probability model was P=1/(1+Exp[—5.614+0.263×WBC+0.045×NP+0.045×CRP—0.027×ALB]). The Hosmer-Lemeshow test indicates that this model has high fitting degree (P=0.622). The area under curve of the ROC curve for the prediction model was 0.931 (95% CI 0.896-0.957, P<0.001), with a sensitivity of 0.830 and a specificity of 0.905. Prospective validation showed that the sensitivity and specificity of the prediction model were 0.882 and 0.820, respectively. Conclusion Increased white blood cell count, neutrophil percentage and CRP level, and hypoalbuminemia are independent risk factors for acute non-simple appendicitis in children. The prediction model based on the above factors has good diagnostic value and can help differentiate between acute simple and non-simple appendicitis. |
Key words: acute appendicitis children risk factors diagnosis prediction model |