Investors lopped a quick $12.50 off Cephalon Inc. stock on Friday, which closed at $23.375 on 9.9 million shares traded, after the FDA waited until the last moment to inform the company that it wasn't prepared to approve a Treatment IND for Myotrophin IGF-1 to treat amyotrophic lateral sclerosis.
Shares of partner Chiron Corp. were down $5.50 to $104 on 1.6 million shares traded.
In a brief statement, CEPH and CHIR said the agency had concerns as to whether the data submitted from CEPH's European study support the results of the previous North American study.
In June, CEPH reported that in the 266-patient Phase II/III North American trial, patients on the high dose of 0.1 mg/kg/day of insulin-like growth factor-1 experienced less disease severity, slower progression of disease and better functional ability than patients on placebo (see BioCentury Extras, June 12 and 13, 1995).
The company subsequently reported positive results from its 183-patient European study (see BioCentury Extra, Nov. 1, 1995). Although none of the European results were as robust as those of the first trial, the company reported that they were within the same confidence intervals.
Survival was not an end point in either trial.
One red flag in the European study was an excess of deaths in the treatment group, which appears to contradict the North American findings. There were five deaths among the 59 patients in the placebo group (8.5 percent) versus 18 deaths among the 124 patients on drug (14.5 percent).
CEPH's post-hoc analysis of survival data in the North American study showed a mean survival difference of 164 days (5.5 months) favoring treated over untreated patients. The analysis on the 138 patients who had died by 18 months includes data on 259 of the 266 patients for whom the survival status was known. The 7 patients for whom there was no data were considered "alive" for purposes of the analysis, as that was the most conservative way of assessing them, according to company spokesperson Mary Fisher.
Neither CEPH nor CHIR had spoken to the FDA as of Friday, and as a result both companies said they did not know the specific issues being raised by the agency. A meeting with FDA officials had not been scheduled as of last week.
CEPH President and CEO Frank Baldino noted that the Treatment IND hasn't been rejected and that the FDA is not questioning the data from the first trial. "They agree that the North American study is impeccable," he said. "But they want some understanding of how we've reached our conclusions on the European study. Our job is to convince the FDA that the second study is supportive."
Because the agency wouldn't comment on its decision, it's only possible to speculate on the FDA's rationale. The possibilities fall into two basic categories: the FDA, for political or tactical reasons, wants to tighten up the Treatment IND process, or there's something wrong with the data relating to efficacy and/or safety.
Taking the bureaucratic theories first, one possibility is that the FDA has become sensitive to the potential for "misuse" of the Treatment IND, if companies come to see it as tantamount to a drug approval. An offhand comment by a Rhone-Poulenc Rorer official at the FDA advisory panel on Rilutek (riluzole) last September, which implied that the FDA had affirmed the safety and efficacy of the ALS drug by granting a Treatment IND, may have raised the agency's antennae.
As in most areas, the FDA appears to have a great deal of latitude in how it interprets its rules. The regulations (21 CFR 312.34) state that "In the case of a serious disease, a drug ordinarily may be made available for treatment use under this section during Phase 3 investigations or after all clinical trials have been completed; however, in appropriate circumstances, a drug may be made available for treatment use during Phase 2."
The regulations further state that drugs may be granted Treatment INDs if: "i) the drug is intended to treat a serious or immediately life-threatening disease; ii) there is no comparable or satisfactory alternative drug or other therapy available to treat that stage of the disease in the intended patient population."
An interim rule effective in October 1988 further stated: "If the preliminary analysis of test results appears promising, FDA may ask the sponsor to submit a treatment protocol to be reviewed under the treatment IND regulations. Such a treatment protocol, if submitted and granted, would serve as a bridge between the completion of early stages of clinical trials and final marketing approval.
"Once phase 2 testing and analysis is completed by the sponsor and a marketing application is submitted, FDA will evaluate the data utilizing a medical risk-benefit analysis. As part of this evaluation, FDA will consider whether the benefits of the drug outweigh the known and potential risks of the drug and the need to answer remaining questions about risks and benefits of the drug, taking into consideration the severity of the disease and the absence of satisfactory alternative therapy."
Line in the sand
Given the common perception that the FDA has become less arbitrary on drug approvals in the face of political pressure to move drugs to market faster, it's also possible that the agency sees Myotrophin as a good case to make the point that it won't be pushed too far.
Because RPR's Rilutek had shown a survival benefit as a primary end point, and was first to reach the market for ALS, the agency isn't under pressure to approve Myotrophin as the sole available therapy. The FDA is thus freer to focus on the possible shortcomings of the European study, particularly if there are concerns about deaths in that trial.
"The FDA is not in a position where they have to approve a second agent, because there's already one out there with a survival effect," said Benjamin Brooks, professor of neurology and director of the ALS Clinical Research Center at the University of Wisconsin, Madison. "There's also an ongoing trial of BDNF, which may have a survival benefit and an effect on breathing," he said. "So the FDA may say it has time to really evaluate the safety of IGF-1."
A Phase III study of brain-derived neurotrophic factor, being developed by Regeneron Pharmaceuticals Inc. (REGN, Tarrytown, N.Y.) and Amgen Inc. (AMGN, Thousand Oaks, Calif.), is fully enrolled, with results expected at the end of 1996.
One investigator from the Myotrophin trial, Eugene Lai of Baylor College of Medicine in Houston, also thinks the FDA may have a point to prove. "The FDA is between a rock and a hard place," he said. "There's a demand for new drugs, but they don't want to be viewed as pushovers. The approval of Rilutek was a political decision because there was no other treatment. Now the pressure is off and they can be more careful and more scientific. So other companies and drugs will be served notice that they're not easy."
Any inconsistencies between data from the first and second trials would also slow the agency down, said Brooks. "Having more data now may be confusing to the FDA. If results are similar to the old analysis, things tend to go along smoothly. But if it requires a new analysis, the clock tends to get reset."
Understanding the data
Bureaucratic maneuvering would be irrelevant if there is something wrong with the data. Assuming the FDA would like to notch another drug approval on its belt, the fact that the agency felt compelled to delay a Treatment IND for Myotrophin would imply that the FDA sees problems that it can't ignore.
"By the letter of the law, a Treatment IND is granted where there's a high probability of approval," said Merrill Lynch analyst Jeffrey Casdin. "So what they're saying right now is there's not a high probability of approval at this point in time."
Casdin, who has been one of the most thoughtful critics of the data, considers the European study to be non-confirmatory. Specifically, the primary end point in the North American trial was disease progression over nine months of therapy, using a three point slope to plot progression (p=0.01). In the European trial, the corresponding p value using a three point slope was 0.06, not significant. But the European study used as its primary end point a two-point slope, which was significant (p=0.04), but less so than the first study. Other results in the European study also were not as robust.
The excess deaths in the European treatment group also raise questions, Casdin said. "Clearly the FDA had to dig into mortality questions because of the imbalance." However, he noted that time has passed since those data were released, and it's possible that the imbalance has disappeared as more patients from both the treatment and control groups have died.
Casdin said that if IGF-1 does cause excess mortality, one plausible explanation would be based on the fact that the growth factor stimulates muscle cell growth. Adding muscle mass can increase muscle strength, but muscle cells require nerves. As a result, he said, neurons that are already compromised by the disease would be additionally stressed, which could lead to greater degeneration.
User point of view
If safety is an issue, it's not obvious to the clinicians using the drug, although their experience is with relatively small groups. Based on their own experiences, none of the Myotrophin investigators interviewed by BioCentury said there was any evidence that people on drug were dying earlier than those on placebo, and two thought they were seeing at least some mortality benefit.
Deborah Gelinas of Forbes North MDA/ALS Center in San Francisco has six or seven patients still alive after three years out of 45 in the initial study at her center. "In this disease, there should be zero alive at this point," she said. "We're especially seeing a benefit in certain patients who have just deteriorated more slowly."
"I can't think why the FDA turned it down, because it was a very good study and a very good drug," said Gelinas. "I'm totally astounded. I've been an investigator on both the Myotrophin and Rilutek trials, and the Myotrophin data are not marginal at all."
Lai also said that about one-third of the 36 patients who were in the trial at his center are still alive. He's still seeing slower progression in the patients who started the trial on Myotrophin compared to those who switched over later, "although the difference is so small that all the investigators may not recognize it until you put all the data together and analyze it - it's pretty much like all drugs for progressive diseases."
Another Myotrophin investigator, Ruediger Kratz of Georgetown University Hospital in Washington, is more cautious about the actual clinical benefit. Of the 23 patients at his center who were in the trial, six are alive. "That's not much different than what we'd normally expect at this time period," he said. "Does it prolong people's lives? - so far only statistically. In clinical terms, three months is not something you notice. But patients on Myotrophin are definitely not dying earlier."
Internal to company
Other clinicians express doubts about the European data. "I don't know all the negatives the FDA saw," said Brooks, "but one thing the scientific community had as a concern was that all the analyses were internal to the company, and no one knew what was included and what wasn't included. In contrast, the riluzole study was opened finally and everyone knew what the data were."
Brooks also said he had concerns based on the asymmetries with respect to deaths. "I think what we all would be concerned about is safety because of that mortality imbalance, though it might just be the way things go in a small study." He also noted that if enough studies of any drug are done, there will inevitably be a negative trial.
Merit Cudkowicz, a neurologist at Massachusetts General Hospital, who wasn't a participant in the Myotrophin trials, also has interpreted the results cautiously. "The European data doesn't look as good as the North American data, and the North American data wasn't great," she said. "It's always been hard to form an opinion about IGF-1. Since no one knew what to make of the North American data, it was hard to know what to think about the European data."
Gelinas is putting her newly diagnosed patients on Rilutek if they request it, although she said the benefit is too small to see on an individual basis. "The statistics showed a benefit in 1,000 patients, but it must be a very small benefit," she said. "In the trial, some of the patients who I thought were doing well turned out to be on placebo. But the side effects aren't terrible, so I have no qualms in patients going on it."
Like Gelinas, Baylor's Lai is putting his newly diagnosed ALS patients on Rilutek if they want it. "It is scientifically proven, although I grit my teeth a little about the price," he said. "Like all such drugs, the dropout rate will tell: after paying that much money, patients will only stay on it if they benefit."
Kratz, Brooks and other clinicians all see combination therapy as the real upside. "I want both drugs approved so I can try them together," Kratz said. "You can expect more benefit if you can work on two aspects of the disease."
Kratz believes that the optimal dose for Myotrophin hasn't yet been determined. "Because the side effects are negligible, we felt we maybe didn't give enough," he said. The Georgetown clinician said he expects that Myotrophin ultimately will be approved. "I am also quite convinced that our data are not any worse than the Rilutek data, and the FDA itself has said the North American data are sound."
Brooks predicted that Myotrophin will be approved, although another study may be required.
On that score, CEPH's Baldino said a new trial is the worst case scenario. He pointed out that the company has an ongoing study in Japan, of the same size as the European trial, which will be fully accrued mid-year. End points are the same as for the North American trial, with the primary end point being change on the Appel functional rating scale. Secondary end points include quality of life and forced vital capacity (FVC), a measure of lung function.
The communications issue
Wrapped up in all these questions about the data is a communications problem between the company and members of both the clinical and Wall Street communities. Although no individuals want to go on the record with their gripes, clinicians complained immediately after the European study was presented in Dublin that the company presented the data, rather than the investigators, and about the lack of time for questions.
Several observers from both the clinical and analyst communities, none of whom wanted to be named, last week noted that the company has discussed multiple versions of the data, which is confusing at best, and at worst raises concerns about the actual data. One suggested that the company would be well-served by having an outside statistical group, preferably from a university, do an independent analysis.
In a similar vein, there are rumblings about the delay in publishing the North American study. The skeptical view is that this implies that the data can't pass muster. However,Gelinas of Forbes North MDA/ALS Center said the problem is merely one of verbosity.
"One thing that definitely has hurt us was not having a publication out," she said. "It's been delayed because of the people spearheading the publication committee. They made a tentative submission, and got comments back that they should cut it. Even I get bored reading it, so you know it's too long."
There also has been confusion about whether the data analysis should have been based on a two-point slope or a three-point slope. Several Wall Street analysts have complained privately that the company has been inconsistent in answering this question. As a result, analysts have been busily constructing their own analyses, which only adds to the confusion.
The end result is that only an FDA review will provide the ultimate answers.