背景及目的

克罗恩病（Crohn’s disease, CD）是一种可累及全消化道的慢性炎症性疾病,超过80%的患者自然病程中须经历手术治疗。研究显示即使患者经历手术，约1/3的患者仍然有再手术的风险。而CD再手术的危险因素以及药物对再手术影响尚有待提供更多数据，以利于未来更好地控制和预防疾病进展。因此本研究目的是分别探究CD术后临床、内镜、影像学和再手术的累积复发率，以及以再手术为结局探究独立危险因素并尝试建立预测模型；同时进一步以预测模型及文献报道危险因素对患者进行分层，探讨术后药物治疗对再手术率的影响，以期为降低CD手术患者的再手术率提供临床借鉴。

研究方法

第一部分：回顾性分析2000年1月1日至2021年3月29日北京协和医院就诊、病程中有肠道切除手术史且同时满足病理诊断的CD患者，最终入组159例。收集患者一般人口学资料、临床资料、内镜资料、影像学资料及手术相关资料，采用Kaplan-Meier法绘制患者的生存曲线，得出累积术后复发率；采用非参数检验、卡方检验或Fisher精确检验进行单因素分析，Logistic回归行多因素分析得出独立危险因素。基于Logistic回归模型建立再手术风险预测模型，采用弃一交叉法对模型进行验证。

第二部分：回顾性分析第一部分中的159例CD手术患者术后治疗药物情况（包括5-氨基水杨酸、免疫抑制剂以及抗TNF-α制剂），分别按照第一部分得出的预测模型及文献报道危险因素将患者分为低危组及高危组，比较术后不同治疗药物对低危组及高危组CD再手术率的影响。采用Kaplan-Meier法计算术后是否加用维持治疗对不同危险分层患者累积再手术率的影响，以Log-Rank检验比较两组患者生存曲线，P＜0.05认为有统计学意义。采用Fisher精确检验比较不同药物对再手术率的影响。

研究结果

第一部分：CD患者术后1年、3年、5年及10年累积临床复发率分别为23.3%、52.9%、73.6%及88.5%；术后1年及3年累积内镜复发率分别为22.6%及73.6%；术后1年、3年、5年及10年累积再手术率分别为5.5%、10.8%、20.4%及42.0%；术后1年、3年及5年累积影像复发率分别为12.9%、46.4%及73.2%。吸烟、穿通型病变（B3）、出现消化道症状到初次手术时间短以及术后未加用维持治疗为CD再手术的独立危险因素。最终以吸烟史、穿通型病变（B3）以及出现消化道症状到初次手术时间三个变量建立预测模型，得出预测模型总分大于1.5分再手术风险高，预测模型ROC曲线下面积（95%CI）为0.774（0.690-0.859）。弃一交叉验证法所得模型验证ROC曲线下面积（95%CI）为0.656（0.548-0.764）。

第二部分：无论基于本研究预测模型亦或文献报道危险因素，均得出低危组患者中术后加用药物治疗与未加用药物治疗相比累积再手术率差异无统计学差异；而高危组患者中术后未加用药物治疗患者累积再手术率较加用药物治疗的患者高[（基于本研究）累积1年再手术率45.5% vs.1.9%，累积5年再手术率54.5% vs.27.2%，累积10年再手术率90.9% vs.44.6%（P＜0.001）；（基于文献）累积1年再手术率23.8% vs. 1.7%，累积5年再手术率50.6% vs.20.5%,累积10年再手术率90.1% vs.34.3%（P＜0.001）]。高危组患者中术后未加用治疗药物较术后加用免疫抑制剂或抗TNF-α制剂治疗的患者再手术率高；术后仅加用氨基水杨酸制剂较加用免疫抑制剂或抗TNF-α制剂治疗的患者再手术率高；术后加用免疫抑制剂治疗与加用抗TNF-α制剂治疗的患者再手术率差异无统计学意义。

研究结论：CD术后复发率较高，部分小肠型CD虽临床症状复发不典型，但内镜和影像已有一定复发。吸烟史、穿通型病变（B3）、出现消化道症状到初次手术时间短为CD再手术独立危险因素。CD再手术高危组患者术后积极加用药物治疗有助于降低累积再手术率，其中免疫抑制剂及抗TNF-α制剂预防再手术效果，氨基水杨酸制剂预防再手术作用有限。

Background and Objective

Crohn's disease (CD) is a chronic, disabling and immune-related disease, and over 80% of patients undergo surgical treatment during their natural course. Numerous studies showed that approximately 35% patients are still at risk for reoperation 10 years after the primary surgery. The purpose of this study was to investigate the cumulative clinical, endoscopic, imaging recurrence rate and cumulative reoperation rate respectively, and to explore independent risk factors for reoperation as well as establish a prediction model in order to provide a reference model for clinical evaluation of reoperation risk. Meanwhile, we try to stratify the patients based on the risk factors in our model and those reported by other studies, and explore the influence of postoperative medication on the reoperation rate, in order to provide reference for reducing reoperation in patients with Crohn’s disease. and reduced the occurrence of reoperation.

Methods

Part one：Retrospective analysis was performed on 159 patients with CD who visited Peking Union Medical College Hospital during January 1, 2000 to March 29, 2021 and with a history of intestinal surgery and a pathological diagnosis. The general demographic data, clinical data, endoscopic data, imaging data and surgical data of the patients were collected, and Kaplan-Meier method was used to analyze the cumulative surgical recurrence rate. Non-parametric test, chi-square test or Fisher's exact test was used for univariate analysis. Logistic analysis was performed to  identify independent risk factors for postoperative recurrence. The risk prediction model for postoperative reoperation was established based on Logistic regression model, and the model was verified by the leave-one-out cross validation.

Part two: The effect of postoperative drug therapies (including 5 - aminosalicylic acid, immunosuppressants and biological agents) on the reoperation rate were compared respectively according to the risk stratification of reoperation. Kaplan-Meier method was used to calculate the effect of postoperative drug therapies on cumulative reoperation rate of patients with different risk stratification, and log-rank test was used to compare the survival curves of patients in the two groups, P < 0.05 was considered statistically significant. Fisher's exact test was used to compare the effects of different drugs on reoperation rate.

Results

Part one: The cumulative clinical recurrence rates of CD patients at 1, 3, 5 and 10 years were 23.3%, 52.9%, 73.6% and 88.5%, respectively. The cumulative endoscopic recurrence rates at 1 and 3 years were 22.6% and 73.6%, respectively. The cumulative reoperation rates at 1, 3 5 and 10 years were 5.5%, 10.8%, 20.4% and 42.0%, respectively. And the cumulative imaging recurrence rates were 12.9%, 46.4% and 73.2% at 1, 3 and 5 years, respectively. Smoking history, perforation-type lesions (B3), short interval from the onset of gastrointestinal symptoms to the initial operation, and the absence of postoperative drug therapies were independent risk factors for reoperation. Finally, a prediction model was established based on three variables: smoking history, perforating lesions (B3) and the time between the occurrence gastrointestinal symptoms to the first operation. Patients with the total score greater than 1.5 points were considered at high risk for reoperation. The area under the ROC curve of the prediction model (95%CI) was 0.774(0.690-0.859). And the area under the ROC curve (95%CI) was 0.656(0.548-0.764) for the model verified by the leave-one-out cross validation.

Part two: Based on both the prediction model of this study and the risk factors in the literature, the difference between reoperation rate was not statistically significant in the low risk group with or without postoperative drug therapies While in the high-risk group, the cumulative reoperation rate in patients receiving postoperative drug therapies was lower than that in patients without drug therapy[ (based on this study) the cumulative reoperation rate 1 year after the primary surgery was 1.9% vs. 23.8%，the cumulative reoperation rate 5 years after the primary surgery was 27.2% vs. 54.5%, the cumulative reoperation rate 10 years after the primary surgery was 44.6% vs. 90.0%(P < 0.001); (based on the literature) the cumulative reoperation rate 1,5 and 10years after the primary surgery were 1.7% vs. 23.8%，20.5% vs. 50.6% and 34.3% vs. 90.1%(P < 0.001)].And the rate of reoperation was higher in patients who did not receive postoperative drug therapies than those who received immunosuppressants or biologics; Patients treated with aminosalicylic acid alone after surgery had a higher rate of reoperation than those treated with immunosuppressants or biologics.There was no significant difference in the reoperation rate between the patients treated with immunosuppressants and those treated with biologic.

Conclusions

The recurrence rate of CD is high and the clinical symptomatic recurrence of some small bowel CD is atypical. The independent risk factors for reoperation of CD are smoking history, B3, the short interval between onset of gastrointestinal symptoms and initial surgery. The prediction model of this study can be used to stratify the reoperation risk for CD. In the group with high-risk of CD reoperation, the postoperative drug therapies can reduce the cumulative reoperation rate. Immunosuppressant and TNF-αinhibitors have a preventive effect on reoperation, while aminosalicylic acid have a limited preventive effect.