In 2010, the available literature was insufficient evidence for the American Gastroenterological Association to make recommendations for or against the use of thiopurines as potential chemopreventive agents.36 Avasimibe However, recent clinical studies have provided sufficient evidence to reconsider

the potential for 6-MP and AZA to reduce the risk of colitis-associated dysplasia and CRC in patients with IBD. Two large population-based cohorts, similar to prior studies, had different results. In a Dutch cohort of 2578 patients with IBD, van Schaik and colleagues33 reported that 28 patients (1%) developed HGD or CRC during 16,289 person-years of follow-up. Two of 28 patients (7%) were on thiopurines alone and 1 patient (of 28, 4%) was on a thiopurine plus 5-ASA. Thiopurine use was associated with a significantly decreased risk of developing HGD or CRC with an adjusted hazard DNA Damage inhibitor ratio (HR) of 0.10 (95% CI 0.01–0.75). However, Pasternak and colleagues37 found no protective benefit in a Danish cohort of 45,986 IBD patients, of which 11% were on AZA (adjusted relative

risk [RR] = 1.00; 95% CI 0.61–1.63). In 2013, the first prospective study of the epidemiology of colorectal HGD and cancer in IBD in the thiopurine era was published by Beaugerie and colleagues.38 The results of the CESAME (Cancers Et Surrisque Associé aux Maladies Inflammatoires Intestinales Chlormezanone En France) trial, a French nationwide observational cohort of 19,486 patients with

IBD designed in the early 2000s to assess the risks of any cancer or HGD in IBD patients, found that 57 (0.3%) patients developed HGD or CRC during the follow-up period (37 CRC, 20 colorectal HGD). In patients with long-standing, extensive colitis, defined as disease duration of at least 10 years and extent of at least 50% of the colon, the multivariate adjusted HR for colorectal HGD and CRC was 0.28 for those who received thiopurines (95% CI 0.1–0.9; P = .03). In the study of inflammation risk by Rubin and colleagues,5 multivariate analysis identified thiopurine exposure as a significant predictive factor (adjusted OR 0.25; 95% CI 0.08–0.74). This finding, after controlling for degree of inflammation, was one of the strongest lines of evidence to date. A meta-analysis pooling of 19 studies (9 case-control and 10 cohort studies), while acknowledging high heterogeneity among studies (I2 = 68.0%, P<.001), reported that the use of thiopurine was associated with a statistically significant decreased incidence of CRC or dysplasia (HGD and LGD) with a pooled RR of 0.71 (95% CI 0.54–0.94; P = .017), even after adjustment for duration and extent of the disease. 39 In the thiopurine-treated patients, the RR of HGD and CRC was 0.72 (95% CI 0.50–1.03; P = .070) and 0.70 for CRC (95% CI 0.46–1.09; P = .111).