WARNER-LAMBERT’s COGNEX TOLERATED BY UP TO THREE-QUARTERS OF ALZHEIMER’s PATIENTS, FIRM ESTIMATES; ADVISORY COMMITTEE UNANIMOUSLY SUPPORTS APPROVAL
Warner-Lambert's Cognex can be tolerated by up to three- quarters of persons with Alzheimer's disease, Parke-Davis VP- Clinical Development Mark Pierce, MD, told the Peripheral & Central Nervous System Drugs Advisory Committee March 18. Committee member Ira Shoulson, MD, University of Rochester, asked Pierce what percentage of patients could stay on Cognex (tacrine) for six months in light of drop-out rates as high as 70% in clinical trials of the acetylcholinesterase inhibitor. "My guess would be anywhere between 60%-75% of patients would still be on drug" at six months, Pierce responded. High drop-out rates in the Cognex trials were due primarily to strict protocol criteria for handling elevated liver transaminase levels, Pierce said. The clinical trials required "forced titration" and allowed physicians "no flexibility" in dose adjustments to account for the problem. Based on Warner-Lambert's experience with the Cognex Treatment IND and a Cognex rechallenge study, Pierce concluded that under real-world conditions, "a large percentage of patients [are] able to tolerate [the] drug ultimately." The committee voted unanimously to recommend approval of Cognex for treatment of Alzheimer's. Pierce's estimate of 60%-75% tolerance of tacrine suggests the likely upper limit on the patient population. FDA estimates that there are 4 mil. Alzheimer's patients in the U.S. The number of patients who show benefit from the drug is likely to be much smaller than the number who tolerate it, several committee members noted. Shoulson calculated that 12% of patients treated with Cognex would show a four-point improvement on the Alzheimer's Disease Assessment Scale/Cognitive Subscale (ADAS COG), defined as a "clinically significant" improvement by the committee. Twenty percent of patients who completed clinical trials showed that four-point level of benefit, Shoulson noted. If 60% of patients can tolerate the drug, that means that 12% of patients started on the drug by clinicians will ultimately show that cognition benefit, he reasoned. The meeting was the third review of Cognex (NDA 20-241) by the advisory committee, which recommended against approval in March 1991 ("The Pink Sheet" March 25, 1991, p. 4) and recommended further trials in July 1991 ("The Pink Sheet" July 22, 1991, p. 6). The drug has been available under a Treatment IND since December 1991. The committee's decision to recommend approval March 18 was based on Warner-Lambert's ability to demonstrate statistically significant improvements in two Alzheimer's cognitive behavior scales chosen as standards by the committee: the ADAS COG and the Clinician Interview-Based Impression of Change (CIBIC). "I would conclude that the sponsor has demonstrated that this drug has beneficial effects. I would call them mild," committee member Sid Gilman, MD, University of Michigan, said. However, he said: "They are significant. They have passed the two criteria . . . I believe that we should approve the drug." Several members echoed Gilman's statement that Cognex' efficacy is limited. "The Lazarus-like effects" reported anecdotally with tacrine use "do not seem to happen," Office of Drug Evaluation I Director Robert Temple, MD, commented. FDA Neuropharmacological Drug Products Division Director Paul Leber, MD, indicated that FDA would not "attach an adjective" qualifying Cognex' efficacy in labeling. "We will say the drug was shown effective in studies that examined clinical global and cognitive performance tests" and then include full trial results in the clinical pharmacology section, he proposed. Temple reminded the committee that FDA "can reach agreements with the company" to ensure that "at least certain parts" of the clinical pharmacology section will always be included in advertising. The committee was presented with data from trial 970-61, a 30- week, placebo-controlled, double-blind, four-arm dose titration study in 663 patients, which began in October 1991 and concluded in January of this year. The agency received data from 970-61 in the middle of February. The trial was designed with input from FDA and was a condition of the agency granting a Treatment IND for Cognex. Medical Reviewer Randy Levin, MD, said the trial showed "statistically significant differences between Cognex and placebo" on the three primary outcome measures of the trial: the ADAS COG, CIBIC and the FCCA (Final Comprehensive Consensus Assessment). The ADAS COG consists of a battery of cognitive tests, such as word memorization and picture drawing. The CIBIC is a global assessment of an Alzheimer's patient conducted by a physician without any peripheral information, such as input from family members or other medical tests. After the completion of the trial, the FCCA is a second global assessment in which the physician has access to some other test results. Levin reported that patients receiving 160 mg per day of Cognex for 30 weeks showed "almost a five-point change" compared to placebo patients on the 70-point ADAS COG. Placebo patients worsened by over two points, while the high-dose Cognex patients improved by over two points. Several secondary endpoints also showed statistically significant advantages for the higher doses of Cognex, including the caregivers' global assessment and the Mini Mental State Exam. Warner-Lambert Clinical Development Director Steven Gracon said that in trial 970-61, 17% of 115 evaluable placebo-treated patients were rated as improved on the CIBIC, while in the 160 mg per day group, 42% of 65 evaluable patients were rated as improved. Of the groups titrated to maximum daily doses of 80 mg per day and 120 mg per day, the percentages of patients rated improved were 25% and 31%, respectively. Noting that a four-point improvement versus untreated patients on the ADAS COG has been correlated to prevention of six months of decline due to Alzheimer's disease, Gracon said the 30-week trial had demonstrated "clinically meaningful results." Neuropharm director Leber opened the meeting by calling 970-61 "the most important" study of Cognex because "it examines doses of tacrine as high as 160 mg a day, a dose equivalent to the maximum that [California physician Joseph] Summers et al. reported employing" in a 1986 study that showed dramatic improvements in Alzheimer's patients treated with tacrine. "Even more critically," Leber added, "970-61 provides information on the effects of tacrine and placebo for periods up to 30 weeks." Leber also said that the study is "notable" because it achieved statistically significant differences in favor of tacrine on the ADAS COG and the CIBIC, two measures that "virtually every expert with whom the agency has consulted has agreed can serve as a sufficient test of an antidementia drug's effectiveness." In addition, 970-61 results are "of note" because the CIBIC "has been criticized as setting too demanding a standard for antidementia drugs of the current generation to meet," Leber said. At a February 1992 meeting of the advisory committee, Leber said that "if you got movement on [the CIBIC], you would know you had a large drug effect" ("The Pink Sheet" March 2, 1992, p. 13). Trial 970-61 is the second study to demonstrate statistically significant differences between Cognex and placebo, Leber pointed out. A 468-patient study (970-26) published in the Nov. 11, 1992 issue of the Journal of the American Medical Association found statistically significant differences between tacrine and placebo on the ADAS COG and the Caregiver-Rated Clinical Global Impression of Change (CGIC) ("The Pink Sheet" Nov. 30, 1992, T&G-2). Commenting on study 970-26 and study 970-61, Levin said: "The major problem that both these studies share is the large number of drop-outs." He noted that after 30 weeks in trial 970-61, approximately 30% of placebo-treated patients had dropped out, "often due to lack of efficacy." In contrast, "over 70%" of patients in the 160 mg group dropped out. Overall, 59% of all patients in 970-61 dropped out. FDA Statistical Reviewer Nancy Smith, PhD, said that most of the drop-outs in tacrine-treated groups were due to adverse events, including liver enzyme increases and gastrointestinal side effects. Addressing the concern that the drop-outs could skew efficacy data in favor of Cognex, Smith analyzed the data with an intent- to-treat analysis and a "last observation carried forward" analysis. Both analyses, which inputted values for missing data points, yielded smaller "but still statistically significant" treatment differences between Cognex and placebo, Smith said. In summary, she said the drop-out rates "did not affect the results" of the comparisons between the 160 mg group and placebo in 970-61. "We feel that we are not losing anything...or at least what we are losing did not affect the results," she concluded. In recommending approval, committee members said they were reassured by long-term safety data showing that liver toxicity and other serious adverse events are not likely with Cognex, despite the high drop-out rates. "It appears that the drug is quite safe and with proper monitoring the sponsor has shown that we are not dealing with a drug that is toxic," Richard Penn, MD, Rush-Presbyterian-St. Luke's Medical Center, said. Warner-Lambert's Pierce presented the Cognex safety database to the committee. The database includes 437 patients from clinical pharmacology trials, 2,568 patients from efficacy studies, 433 patients from non-Parke-Davis trials, and 4,183 patients from the Treatment IND. Of those patients, 1,032 were evaluated in parallel-group, placebo-controlled trials, Pierce said. The most common adverse event was elevated liver transaminase (alanine aminotransferase or ALT) levels, seen in 25.7% of patients in controlled trials. Nausea and vomiting, diarrhea, rash and other cholinergic side effects were also seen, Pierce said. Adverse events characterized as "serious" by clinicians were observed with very low frequency, he added. Serious ALT elevations occurred in .4% of Cognex-treated patients. Severe nausea and vomiting were also seen with .4% frequency and serious rash was observed in .2% of patients. Warner-Lambert evaluated its database for cases of agranulocytosis at FDA's request. A .4% incidence of agranulocytosis was cited as one reason for recommending against approval of Hoechst-Roussel's Alzheimer's treatment Mentane (velnacrine) by the advisory committee last November ("The Pink Sheet" Nov. 9, 1992, p. 5). Warner-Lambert found four cases of agranulocytosis, Pierce said, of which three were apparently related to other drugs or conditions. "We would conclude that we have one possible case," Pierce said. ALT elevations were defined as a side effect requiring drop- out if they reached three times normal level, Pierce said. In the 970-61 trial, 29.1% of patients had such enzyme elevations. However, he added, 44 of 56 patients who entered a formal rechallenge protocol were successfully treated with Cognex after dropping out of the trial. That study and reports from the Treatment IND program led Pierce to conclude that increased transaminase levels were a manageable side effect with weekly blood monitoring. The likely dosage regimen will recommend titration at six week intervals from 40 mg up to 160 mg with weekly blood monitoring for elevated ALTs through 18 weeks, Pierce indicated. In clinical trials, physicians saw patients every six weeks, he said. "I think that the issue of how often liver function tests are needed is something we should have more information" about, Shoulson said. "Inconvenience for Alzheimer's patients in terms of going somewhere is not a trivial issue." Shoulson and other committee members frequently cited cost -- both of the drug and necessary monitoring -- as a major factor causing them to hesitate about recommending approval. FDA's Temple reminded the committee that it should not consider cost in its decision: "Cost/benefit is not what we do...There are other people in this world who do get to worry about that and should." Committee consultant Leonard Berg, MD, Washington University School of Medicine, told the committee what his advice to patients would be about Cognex: "The cost is substantial not only for family but for third-party payors and society at large," Berg said. "The benefits appear to be small, but the chance of a very serious side effect also appears to be small." While Berg recommended that the drug be made available, as a clinician "I would probably not give a recommendation" to the patient.
Sign in to continue reading.
New to Pink Sheet?
Start a free trial today!
Register for our free email digests: