Abstract LB-142: Marked efficacy of difluoromethylornithine plus sulindac in reducing recurrent colorectal adenomas in patients with prior adenomas: Results of a randomized, placebo-controlled, double-blind Phase III trial.

Abstract Introduction: Epidemiologic and experimental data suggests that colon cancer should be largely preventable through dietary or chemical intervention. Polyamines are actively involved in proliferation and differentiation of many tissues. Lowering of intracellular levels via inhibition of the initial step (ornithine decarboxylase) in the polyamine synthesis pathway by the specific enzyme-activated inhibitor DFMO results in a decrease of adenomas and cancers in many preclinical organ-site cancer models, most notably in the colon. Methods: From 1989-1997 we conducted a series of pilot and Phase IIa (short term) and Phase IIb (long term) trials in which we determined the lowest dose of DFMO at which polyamine content of the colonic flat mucosa would be lowered, with the goal of finding a non-toxic dose. In 1997 we debated whether to do a definitive trial with DFMO alone or to combine it with another chemoprevention compound. We settled on combining DFMO with the NSAID sulindac at low dose for a variety of mechanistic, experimental, and clinical reasons and began a Phase IIb trial in which the recurrence of adenomas would be correlated with polyamine changes in the flat mucosa. Several years later we expanded the trial to a full Phase III trial in which recurrence of adenomas became the primary endpoint. Results: Based on the recommendation of the Data Safety and Monitoring Board an accelerated second analysis was done after 70% of patients had completed their 3 year or off-study colonoscopy and the study was stopped for efficacy as the study goals had been met. Overall, a highly significant reduction (70%) of all adenomas, advanced adenomas (92% reduction), and multiple adenomas (95% reduction) was seen (all P<0.001), an effect comparable to that seen in animal models. Both adjacent rectal mucosal biopsy putrescine and spermidine levels were also markedly reduced in the active intervention arm after both 12 and 36 months, and there was a significant correlation between reduction in polyamines and adenoma recurrence. PGE2 levels were unaffected, although there was a trend in the correlation of reduction in prostaglandin E2 and adenoma recurrence which most likely is due to an independent action of sulindac. An exploratory analysis of the effect of the active regimen on the different segments of the colon suggested no preference in effectiveness. Toxicity was carefully monitored and all toxicities (adverse events of any nature) as well as hematologic, gastrointestinal, and cardiovascular toxicities were analyzed individually and in a number of ways; clinical toxicities were low and comparable in the two arms. Careful serial audiograms showed subclinical changes that were more frequent in the active intervention arm and were largely reversible after six months off drugs. Conclusions: These results confirm the value of de-escalation studies prior to launching a full Phase III trial in identifying a dose and regimen for the therapeutic prevention of adenomas that is both efficacious and non-toxic; this strategy may be generally applicable in the design of future large chemoprevention trials, particularly those in which combinations would be used. Larger trials will be needed to determine the absolute risk of cardiovascular events and audiologic changes from this regimen, and longer-term studies will be needed to determine whether the absolute endpoint of colorectal cancer incidence can be reduced in either patients with low stage prior colorectal cancers or in very high risk individuals.