Guo 2018.

Study characteristics
Methods	Observational study (retrospective cohort) conducted in China (1 centre) Mean follow‐up: 50.6 ± 23.8 months
Participants	Adults undergoing intestinal resection with ileocolic anastomosis and who also received azathioprine as maintenance therapy from 2006 to 2015. Exclusion criteria: previous intestinal resections, other segmental resections, stricturoplasty or repair at the time of the index resection, more than one bowel anastomosis during a single operation and incomplete medical information. Participants with at least one of the following risk factors were considered to be at high risk of recurrence: Age at surgery ≤ 30 years Received steroids within 1 month of surgery or anti‐TNF drugs in the 12 weeks prior to surgery C‐reactive protein > 10 mg/L the day before surgery Perforating disease Active smoking. Control group (Symptom‐driven) 33 participants (27 high risk, 6 low risk) Intervention group 1 (Endoscopy‐driven) 42 participants (39 high risk, 3 low risk) Intervention group 2 (Immediate) 91 participants (69 high risk, 22 low risk)
Interventions	This study had three groups. Control group (Symptom‐driven) Symptom‐driven therapy was defined as azathioprine initiated only in the presence of clinical recurrence. Intervention group 1 (Endoscopy‐driven) Endoscopy‐driven therapy was defined as azathioprine initiated only when asymptomatic endoscopic recurrence occurred. The endoscopic follow‐up occurred from 3 to 18 months after surgery. Intervention group 2 (Immediate) Immediate prophylactic therapy was defined as azathioprine initiated within 2 to 4 weeks after resection. The dosage of azathioprine was 1.5to 2.5 mg/kg/day.
Outcomes	Surgical recurrence (intestinal resection for complications of Crohn disease or symptoms refractory to medical treatments after the index surgery)
Notes	The three management strategies were adopted during different time periods: Management guided by clinical recurrence (symptom‐driven) was the usual strategy adopted in this centre before 2009; endoscopy‐driven management was a strategy adopted between 2011 and 2013 and immediate preventative therapy was the primary strategy adopted after 2009. The follow‐up of the control group was longer than both intervention groups (participants treated before 2009 versus participants treated after 2009). Clinical recurrence was defined as CDAI > 150 points and C‐reactive protein > 10 mg/L Endoscopic recurrence was defined as Rutgeerts' score ≥ i2 in the neoterminal ileum and/or at the anastomosis. The analyses were adjusted by potential confounders (multivariable Cox proportional hazard model). Potential confounders included into the model were selected by statistical criteria (P value <0.1 in the univariate analysis). Despite this, behaviour of the disease according to the Montreal classification and smoking status were not considered in this analysis. The authors did not declare conflict of interest. The authors declared the following funding sources: National Natural Science Foundation of China and Nanjing University of Science and Technology.

anti‐TNFa: anti‐tumour necrosis factor antibodies; CD: Crohn's disease; CDAI: Crohn's Disease Activity Index; ITT: intention‐to‐treat.