P667 Thiopurine withdrawal during sustained clinical remission in inflammatory bowel disease: relapse and recapture rates in 72 patients, a report from a district general hospital

Long-term treatment with thiopurines is associated with an increased risk of opportunistic infection, lymphoma and other malignancies. Treatment withdrawal should be considered in patients who are in deep remission. To establish the percentage of patients that relapse after withdrawal of azathioprine and time from withdrawal to relapse and recapture rate. We searched our IBD database for patients where azathioprine (AZA) was withdrawn following a deep remission (defined as faecal calprotectin (FCP) < 200 mg/kg and/ endoscopic remission defined as no ulceration or quiescent inflammation). We reviewed the length of time on AZA, the relapse rates following withdrawal and recapture rate following re-starting. 72 patients were identified (36 Crohn’s disease (CD); 36 ulcerative colitis (UC). Mean age at stopping = 61.5 years (range 21–84). 36 (50%) male; Median duration of thiopurine use prior to withdrawal = 6.5 years (range 1.0–14.8). Fifty-two (72%) had FCP measured; mean = 37 (5–161). Endoscopic remission was confirmed in 36 (50%). In 3 (18.8%) FCP was normal with mild inflammation endoscopically. A total of 28 (38.8%) patients relapsed; 17 UC (60.7%) 11 CD (39.3%), mean time until relapse =16 months (range 7 days–79 months). Relapse was confirmed endoscopically in 22 (78.6%) and by FCP in 6 (21.4%). Mean FCP 940 (range 459–2329); 20 (68.9%) re-started AZA with concomitant prednisolone. Of these 14 (70%) entered clinical remission. Six patients (50% UC) failed to enter remission with AZA. Three patients required admission (10.3%). A clinical decision was made to start 2 (6.9%) on methotrexate because of better risk profile in older age. Seven (24.1%) re-started mesalazine monotherapy. 1 (3.1%) opted for colectomy (immunosuppression was not re-started due to a diagnosis of malignant melanoma). And 1 had vedolizumab due to a diagnosis of sarcoma. In total 16 patients (57.1%) entered and maintained remission (length of follow-up = mean 29.7 months (range 2–120) without requiring surgery or biologics. Our data demonstrate that 38.8% of patients stopping thiopurine will relapse and 79.4% of these patients will relapse within 3 years. Reassuringly, 71.5% achieved remission on re-starting AZA/methotrexate/ mesalazine. Interestingly, 3 of our patients were identified to be in remission with FCP but had mild inflammation on endoscopy. These three patients subsequently relapsed. This raises the possibility of the need for diagnosis of remission endoscopically rather than using FCP alone. These data should aid discussion regarding the safety of withdrawing thiopurine in patients in a long-term remission.