The mean prevalence of azathioprine (AZA) or 6-mercaptopurine (MP)-induced liver injury in patients with inflammatory bowel disease was approximately 3%, and the mean annual drug-induced liver disorder rate was only 1.4%. However, this low figure calculated from retrospective studies contrasts with a much higher incidence (>10%) reported by a prospective study. Thiopurine-induced hepatotoxicity can be grouped into three syndromes: hypersensitivity, idiosyncratic cholestatic reaction, and endothelial cell injury (with resultant raised portal pressures, veno-occlusive disease, or peliosis hepatis). A small percentage of patients present with a slight elevation of liver tests (LTs) that do not have clinical implications and LTs return to normal values during the follow-up, indicating that it is not always necessary to adjust the dose of the immunomodulator. However, when abnormalities in LTs are more marked, the dose of AZA/MP may be reduced 50%, with posterior clinical and analytical controls. With this strategy, LTs frequently normalize spontaneously, and the initial AZA/MP dose may be cautiously prescribed again. Thiopurines may induce an unusual severe cholestatic jaundice that may not regress but even progress despite thiopurine withdrawal. Therefore, these drugs should be completely withdrawn, and not only tapered, in those patients presenting clinically significant jaundice. Despite a lack of evidence that monitoring of LTs is necessary in patients receiving AZA/MP, routinely performed laboratory controls including LTs seem recommendable. However, the optimal monitoring schedule remains to be established. As long-term hepatotoxicity seems to be an unpredictable and potentially severe adverse drug reaction of 6-thioguanine, this drug should not be administered outside a clinical trial setting. (Am J Gastroenterol 2007;102:1518-527).