Persons with ulcerative colitis are at increased risk for colorectal dysplasia and cancer; cancer occurs in up to 13% of patients within 20 years (1). Patients who have ulcerative colitis with primary sclerosing cholangitis may be at even greater risk. Such patients have accelerated rates and numbers of neoplastic events and are likely to benefit considerably from chemopreventive interventions that complement established preventive measures (screening, surveillance, polypectomy, and prophylactic surgery).

Fecal bile acids have long been associated with increased risk for colorectal cancer. Evidence supporting this association includes in vitro mutagenesis with deoxycholic acid (2), bile acid–induced tumor promotion in animal models (3), and epidemiologic investigations (4). Ursodiol, a noncytotoxic hydrophilic epimer of chenodeoxycholic acid, has shown promising chemopreventive efficacy in two carcinogen-induced rat models of colorectal carcinogenesis. Although the precise mechanisms underlying these effects are incompletely understood, ursodiol has been reported to enhance the expression of major histocompatibility antigens on premalignant tissues, modulate carcinogen-induced changes in protein kinase C, stabilize cellular membranes, reduce mucosal concentrations of phospholipase A2 at the messenger RNA and protein levels, and reduce mucosal prostanoids and aberrant crypt foci incidence (5).

In this issue, Tung and colleagues (6) present intriguing data linking ursodiol use to lower rates of colorectal dysplasia, based on observations of 59 patients with ulcerative colitis and primary sclerosing cholangitis who were participating in a colonoscopic surveillance trial. The authors used sound methods and defined their cohort with standardized histologic and cholangiographic criteria. All eligible patients were included in the analyses. Ursodiol exposure—specifically dosage (average, 8.9 to 9.9 mg/kg of body weight per day) and duration of use (average, 3.5 to 4.2 years)—was ascertained from medical records, with support from patient interviews. Well-established covariates likely to alter the occurrence of dysplasia (sex, age at onset of colitis, duration of colitis, duration of primary sclerosing cholangitis, Child–Pugh class, and use of concomitant medications) were similarly assessed. Dysplasia and cancer outcomes were measured according to an investigational colonoscopic surveillance and mucosal biopsy schedule (every 1 to 3 years, depending on the highest grade of dysplasia). Biopsy specimens were rigorously scrutinized, and the data were analyzed by experienced statisticians using established methods.

Despite the abundant strengths of this study, several limitations merit discussion. First, the study involved a small cohort followed for an unspecified period at a single institution. The rate of dysplasia and cancer in patients not using ursodiol was relatively high (that is, 72% after a mean of 21 years) compared with dysplasia and cancer rates reported elsewhere (that is, from 20% to 50% after 20 to 25 years' duration of ulcerative colitis). This disparity casts doubt on whether findings from the study by Tung and colleagues are applicable to all patients with ulcerative colitis and primary sclerosing cholangitis. In addition, compared with nonusers, patients who took ursodiol were significantly older at the onset of ulcerative colitis and had shorter durations of disease, two factors associated with reduced risk. The authors appropriately addressed these potential confounders by performing adjusted multivariate analyses; however, the groups may have differed in other ways that might account for dissimilar rates of dysplasia. For example, anatomic extent of colitis, an established risk factor for dysplasia and cancer in persons with ulcerative colitis (7), was not addressed. Furthermore, the use of other medications with significant chemopreventive potential in this and other cohorts (for example, folate or aspirin and other nonsteroidal anti-inflammatory drugs) was not discussed.

The utility of ursodiol in primary sclerosing cholangitis is somewhat controversial. While ursodiol has been associated with improvements in biochemical indices of liver function, it has yet to be shown to consistently reverse or retard progression of liver disease (8). Therefore, it is critical to know the rationale for prescribing ursodiol only for a subset of patients and at relatively low dosages (approximately 9 to 10 mg/kg per day compared with more standard dosing of 13 to 15 mg/kg per day). In addition, time to progression, as well as long-term follow-up assessing the efficacy of ursodiol against colorectal cancer or other clinically relevant events (for example, frequency of endoscopic evaluations, intensity of the biopsy protocol, or colectomy), would have been of considerable relevance. For example, if ursodiol were shown to delay and ultimately reduce the incidence of dysplasia, the case for its preventive potential would be stronger. Finally, the argument could have been strengthened further through the inclusion of biomarkers assessing ursodiol pharmacokinetics or pharmacodynamics (for example, mucosal kinetics or changes in concentrations of serum and fecal bile acids) to confirm patient adherence, agent bioavailability, and biological effects.

Several other possible chemopreventive agents have been identified from observational studies in persons with ulcerative colitis. Studies have reported a variety of protective effects associated with folate supplements, including normalization of rectal mucosal proliferation and reductions in risk for dysplasia and cancer (9-12). In 1994, Pinczowski and coworkers (13) reported the results of a case–control study involving 102 patients and 196 matched controls in which the use of sulfasalazine for at least 3 months was associated with a 62% reduction in risk for colorectal cancer. Moody and colleagues (14) subsequently confirmed that long-term use of sulfasalazine reduced the risk for colorectal cancer in a prospective study of 175 patients, but this effect was not observed in two later studies (6, 12). Randomized trials of these agents have yet to be initiated.

The hypothesis that ursodiol, a drug with relatively few side effects, may reduce colorectal neoplasia in persons with ulcerative colitis (with or without primary sclerosing cholangitis) or in those at risk for colorectal neoplasia in the general population remains intriguing, although still unproven. Several studies are under way to address this question more definitively. Eleven randomized, placebo-controlled trials of ursodiol have been completed in persons with primary biliary cirrhosis (8), and several others have evaluated the use of ursodiol in persons with primary sclerosing cholangitis (15, 16). It may be expedient to examine the incidence of colorectal neoplasia in these trials, expecting that the rate of neoplasia may be reduced during, or immediately following, drug exposure. Two large, ongoing placebo-controlled trials in persons at risk for sporadic colorectal adenomas will contribute more data on the chemopreventive efficacy of ursodiol within the next few years (17, Alberts D. Personal communication).

At present, it would be premature to offer ursodiol as a chemopreventive agent outside the context of a clinical trial. However, if the inverse association between ursodiol use and dysplasia prevalence described by Tung and colleagues is causal and true, it is a very exciting observation. The unmet medical needs of persons with ulcerative colitis and the promise of chemopreventive interventions suggest the near-term emergence of a new best-practice model. This model is based on the application of existing preventive measures but incorporates participation in randomized, controlled trials of novel chemopreventive agents, ensuring that the gap between clinical practice and prevention research will continue to narrow.