Background  Debate exists regarding to whether thiopurine therapy is as effective in ulcerative colitis (UC) as it is in Crohn's disease.

Aim  To review systematically the efficacy of azathioprine (AZA) and mercaptopurine (MP) in UC, and to conduct a meta-analysis of randomized clinical trials evaluating the efficacy of AZA/MP for the induction or maintenance of UC clinical remission.

Methods  Selection of studies: Evaluating AZA/MP for induction and/or maintenance of clinical remission of UC. Randomized-controlled-trials comparing AZA/MP with placebo/5-aminosalicylates were included in the meta-analysis. Search strategy: Electronic and manual. Study quality: Independently assessed by two reviewers. Data synthesis: By ‘intention-to-treat’.

Results  Thirty noncontrolled studies (1632 patients) were included in the systematic review. Mean efficacy of AZA/MP was 65% for induction and 76% for maintenance of the remission. Seven controlled studies were included in the meta-analysis. (i) Induction of remission: four studies (89 AZA/MP-treated patients) showed mean efficacy of 73% vs. 64% in controls (OR = 1.59; 95% CI = 0.59–4.29). (ii) Maintenance of remission: six studies (124 AZA/MP-treated patients) showed mean efficacy of 60% vs. 37% in controls (OR = 2.56; 95% CI = 1.51–4.34). When only studies comparing AZA/MP vs. placebo were considered, OR was 2.59 (95% CI = 1.26–5.3), absolute risk reduction was 23% and number-needed-to-treat (NNT) to prevent one recurrence was 5.

Conclusion  Thiopurine drugs (AZA/MP) are more effective than placebo for the prevention of relapse in UC, with an NNT of 5 and an absolute risk reduction of 23%.