Background: Thiopurines are used as maintenance therapy in ulcerative colitis (UC). However, data about long-term efficacy are limited. The aim of this study was to evaluate their long-term efficacy, and to search for predictive factors of failure.
Methods: In this retrospective monocentric study, we included UC patients responding to treatment by thiopurine alone or associated with 5-aminosalicylates (5ASA) for at least 6 months. Demographic, clinical and biological data were collected at diagnosis, 6 months after thiopurine onset (considered as T0) and at the time of treatment withdrawal, or last follow-up visit.
Clinical status at T0 was classified into two groups: response (clinical improvement but persistence of symptoms or steroids) and remission (no symptoms and no steroids). Therapeutic failure was defined by: need for anti-TNF alpha, surgery, treatment withdrawal due to side effects or flare up at the date of latest news
Results: We included 130 patients (52% males), with a median follow-up of 86 months (IQR 55–130). UC was predominantly left-sided (53%) and pancolic (35%). Major indication for thiopurine was steroid-dependence (45%), and 19 patients (15%) had severe acute colitis prior to thiopurine. Median duration of treatment was 42 months (IQR 23–80), with 119 (92%) patients having azathioprine.
Eighty-seven patients (67%) maintained clinical remission on thiopurine, and 43 (33%) underwent treatment failure. Mean duration of treatment was 25months in failure group vs 52 months in remission group. Five years after initiation, 36% (n=47) of patients were still on thiopurine. Reasons of treatment withdrawal were: secondary failure (n=31), major side effect (n=9), persistent remission (n=40) and other reason (n=3). In univariate analysis, older age at diagnosis (RR 1.02; CI95% 1.0–1.04; p=0.04) and clinical response at T0 (RR 1.98; CI95% 1.03–3.82; p=0.04) were associated to treatment failure. In multivariate analysis, only clinical response at T0 vs remission was associated with failure (RR 2.07; CI95% 1.07–3.4; p=0.03).
Relapse rate after withdrawal for persistent remission was 43%, with a median time of 29 months.
Forty patients (31%) had a significant side effect, and 9 of them (7%) stopped treatment due to it. Side effects were mainly hematologic and hepatologic. Nine neoplasic and pre-neoplasic lesions were observed during follow-up: 2 colorectal high risk adenomas, 3 basal cell carcinomas, 1 in situ cervix carcinoma and 3 other cancers (prostate, breast, lung).
Conclusions: This study shows a long term efficacy of thiopurine as a maintenance treatment in UC. Risk of relapse after thiopurine withdrawal is significant. Neoplasia rate due to thiopurine seems to be acceptable in this cohort.