Despite the development of biological therapy, surgery is still needed in patients with refractory or complicated Crohn's disease. Thirty years ago, Rutgeerts et al. showed that postoperative clinical recurrence occurred early, affecting already one in five patients at 1 year, with endoscopic lesions preceding symptoms.1 Prevention and treatment of postoperative recurrence [POR] are important to avoid multiple resections and bowel destruction. Both thiopurines and antibodies targeting tumour necrosis factor [TNF] proved to prevent the development of postoperative endoscopic recurrence. 2,3 However, data on thiopurines in this setting are scarce and a Cochrane meta-analysis, published in 2014, could not conclude on a superiority of thiopurines over placebo. 4 Although indirect evidence suggested a better outcome with anti-TNF therapy, no head-to-head trials existed in order to choose between these two options.In that regard, the APPRECIA study─published in this issue of Journal of Crohn's and Colitis─is an interesting work addressing the head-to-head comparison of azathioprine [AZA] vs adalimumab [ADM] in the prevention of postoperative endoscopic recurrence.
5In this randomised controlled open-label study, 84 patients were randomised to receive either AZA 2.5 mg/kg body weight or ADM [160/80 mg induction and 40 mg fortnightly thereafter] within 2 weeks after the surgery. The primary endpoint was postoperative endoscopic recurrence at 1 year, defined as a modified Rutgeerts' score of at least i2b─more than five aphthous ulcers or larger lesions in the neoterminal ileum with normal mucosa in between [with or without anastomotic lesions]. 6 The superiority of ADM could not be proven: endoscopic recurrence rate in the AZA group was 59% [23/39] vs 42% [19/45] in the ADM group [p = 0.12].Although the study had a good design, it seems difficult to draw strong conclusions from these probably underpowered data. For their sample size calculation, the authors expected a 35% difference in POR between ADM and AZA groups. This was based on the tremendous effect of infliximab observed in the pilot study by Regueiro et al., with only 9% of POR at 1 year in the infliximab group vs 85% in the placebo group [delta 76%, p = 0.0006].3 Although the recently published PREVENT study also showed a significant difference between infliximab and placebo, the difference was much smaller, with 22% of POR in the infliximab group and 51% in the placebo group [delta 29%, p < 0.001].7 In both trials, POR was defined as a Rutgeerts' score of at least i2. Furthermore, in APPRECIA, the estimated drop-out rate of 10% was clearly underestimated [26% in PREVENT; 15% in POCER]. 7,8 Second, Lopez-Sanroman et al. chose to include all operated patients, without taking into account risk factors for POR like active smoking, previous resection, or penetrating disease. Although these risk factors have not been validated in a prospective setting, retrospective studies repeatedly showed their implication in POR.9 In POCER, the benefit of endoscopy-driven tr...