SMITHKLINE BEECHAM’s KYTRIL POTENTIAL CARCINOGENICITY AND CARDIOVASCULAR EFFECTS DERAIL APPROVAL RECOMMENDATION; SECOND-LINE USE WITH WARNINGS SUGGESTED

SmithKline Beecham's anti-emetic Kytril (granisetron) should not be approved at this time because of questions regarding potential carcinogenicity and cardiovascular effects, FDA's Gastrointestinal Drugs Advisory Committee concluded at its March 11 meeting. "This committee would like emotionally to see granisetron receive approval...but that is not a recommendation given the safety profile," committee consultant reviewer Thomas Burks, MD, University of Texas Health Science Center-Houston, summarized. "The concern engendered by the carcinogenic effects of granisetron outweighs at this time the benefit of the compound," Burks said. The committee agreed unanimously that Kytril appears to be effective in preventing nausea and vomiting associated with chemotherapy. However, it decided that the antiemetic cannot be approved at this time without further evaluation of the importance of animal studies showing liver tumor formation in rats and mice treated with granisetron. The committee also wanted to be certain that cardiovascular events reported in Kytril studies were not unique to the compound. SmithKline Beecham faces an unexpectedly difficult problem in trying to gain approval of its 5HT[3] receptor antagonist in the U.S. Granisetron is approved in approximately 20 countries and has been administered to over 300,000 patients, the company said. In a statement issued after the March 11 meeting, SB said that the committee "felt unable to recommend approval of Kytril at this time because of two specific concerns that fell outside of their area of expertise." The firm added that "it was confident it can work with the FDA to resolve these issues." The committee and its consultants acknowledged that the existence of an already marketed 5HT[3] receptor antagonist for chemotherapy-induced emesis -- Glaxo's Zofran (ondansetron) -- was a major factor in their decision not to recommend approval of granisetron. "While this drug application must be reviewed on its intrinsic merits, we cannot ignore the fact that another 5HT[3] antagonist of similar efficacy in cisplatin-induced emesis, but without carcinogenic activity, is already available for use," Burks noted. "I feel there is a drug on the market already [that is] excellent and does not have this nuance or more than nuance, definitely a major problem in terms of tumors," committee member Helen Shields, MD, Beth Israel Hospital (Boston), said. SmithKline Beecham VP-Safety Assessment Gwyn Morgan, MD, presented data from two-year rat studies that show hepatocellular adenoma and carcinoma formation starting at granisetron doses of 5 mg/kg per day, which is over 100 times the human dose used in clinical studies. Similar tumors appeared in mice treated with granisetron for 104 weeks at doses of 50 mg/kg per day. At the meeting, SB representatives argued that while the carcinogenicity data in animals is undisputed, there is no evidence that granisetron is genotoxic. They maintained that although granisetron is a "promoter" that may cause tumor cell proliferation, the drug is not an "initiator" because in vitro tests show that it does not contribute to gene mutations or chromosomal damage. Morgan likened granisetron to phenobarbitol, a drug with carcinogenic effects in animals but no genotoxic potential in humans. FDA Gastrointestinal & Coagulation Drug Products Division Supervisory Pharmacologist Jasti Choudary, PhD, emphasized that granisetron's carcinogenic and genotoxic potential should not be minimized. He asserted that the tumors found in the animal studies are "far in excess of what one can expect in this strain of rat." In addition, Choudary maintained that in at least one mutagenicity test granisetron was found to "increase unscheduled DNA synthesis." He also cited the fact that tumor formation was dose related as evidence that there is a potential carcinogenic risk associated with granisetron. Some committee members cited granisetron's lingering presence in body tissues, for more than 48 hours in some cases, as a further indication of the drug's potential toxicity. "I am really concerned about the fact that this seems to hang around in tissues for a long period of time...in the eye and skin and other places," committee member Peter Banks, MD, Brigham and Women's Hospital, said. Committee members urged FDA and the company to resolve the issue of whether Kytril is genotoxic. "I think there is a clear difference of opinion about genotoxicity," committee member Arvery Rogers, MD, University of Miami School of Medicine, said. "I think an effort should be made to resolve those differences." The committee voted eight to two that if granisetron is shown to be non-genotoxic, the drug could be approved by FDA for limited use. If approved, committee Chairman Rosemarie Fisher, MD, Yale University, said Kytril labeling should bear a "black box warning" and be used "as a second-line drug." Committee consultant Burks agreed that "if approved granisetron should carry adequate warnings of carcinogenic potential and its use should be restricted to patients who have failed to respond to ondansetron or metoclopromide." Burks recommended that the company conduct two additional carcinogenicity studies in animals. One would examine "the relative carcinogenic activity of metabolite D predominant in humans and metabolite E predominant in rats." Burks noted that "if carcinogenicity is associated primarily or exclusively with metabolite E, the parent drug may be less of a hazard in humans than the present rodent data would suggest." The other study would examine "the carcinogenic activity of granisetron in combination with appropriate anticancer drugs or other initiator drugs to determine whether there is additivity of tumor-inducing effect." The committee suggested that Kytril may have a place as a first-line therapy if the drug were to serve a niche population. Burks said that "if [Kytril] is far better than ondansetron or if it covers a population of patients...who are resistant to ondansetron, that would be a powerful argument for favorable consideration." SB currently is conducting two studies that are comparing Kytril and Zofran. The committee considered the evidence for Kytril's efficacy in preventing nausea and vomiting in patients receiving chemotherapy to be convincing. "There is no question that the sponsor has provided us with two clinical studies that show [Kytril] is efficacious," committee reviewer Vay Go, MD, University of California at Los Angeles, commented. Studies presented by SmithKline Beecham in over 1,500 patients showed that Kytril produced complete responses (no nausea or vomiting) in 60-90% of patients receiving cisplatin chemotherapy compared to under 10% for placebo. The main question raised by the committee about the efficacy studies involved SB's selection of Kytril dose. The company admitted that it had chosen a trial dose of 40 mcg/kg per day "rather arbitrarily." Burks maintained that since there appears to be no therapeutic advantage for the 40 mcg/kg dose over the 10 mcg/kg dose and that "the carcinogenic effect in rats and mice seems to follow a steep dose response curve," the recommended dose should be 10 mcg/kg. SB argued against the 10 mcg/kg dose because of a concern about breakthrough emesis. However, a majority of the committee members voted that the dose should be 10 mcg/kg. Several committee members recommended that the dose be adjusted from 10 mcg/kg to 40 mcg/kg depending upon the emetogenic effect of the chemotherapy. While some committee members were not concerned by cardiovascular adverse events -- such as asymptomatic arrhythmias, atrial fibrillation and large changes in blood pressure -- reported in the Kytril studies, other committee members maintained that the data on cardiac side effects is insufficient. Some committee members suggested that the cardiovascular events may be a class effect. "If there are adverse cardiac effects they are well likely to be 5HT[3] receptor-mediated and therefore a study of all the drugs in this class would be of benefit," Burks said. Holter monitoring studies pre- or post-approval were recommended by the committee.