FDA “Critical Path” May Lead To Changes In Dosing, Active-Control Trials

FDA's "Critical Path" drug development initiative will explore relaxing the agency's interpretation of active-control trials, Office of Medical Policy Director Robert Temple, MD, said.

"Probably the biggest problem is how to interpret active-control trials," Temple said during a Dec. 3 videoconference on the Critical Path sponsored by the Drug Information Association and FDA.

"We tend to make relatively conservative assumptions in interpreting those trials - maybe they're overly conservative."

"One of the things we're going to be thinking about is whether the past history of trials with related drugs could be used to be somewhat more aggressive in studying the standards for the new trial," he added.

Industry and FDA agree that mining the agency's vast historical trial database could foster the development of new trial techniques.

FDA has received numerous suggestions on "data mining and just pulling the data together and drawing conclusions...[from] this treasure trove of information we have at the FDA," Acting Deputy Commissioner for Operations Janet Woodcock, MD, said.

FDA established a docket to collect comments on the Critical Path initiative in April (¹ (Also see "PhRMA Seeks Route To Prevention Therapies Along FDA’s “Critical Path”" - Pink Sheet, 16 Aug, 2004.), p. 6).

"I think there is a wealth of information that can actually be shared that we have typically not shared in the past - everything that is pre-competitive or non-competitive," Novartis Senior VP-Drug Regulatory Affairs Mathias Hukkelhoven, PhD, said.

There should be a standardized method of submitting the data as well, "so that they can be analyzed in a big meta-analysis," he added.

FDA also discussed the Critical Path's effect on: (1) prioritizing biomarker development; (2) multiple dosing during Phase III trials; (3) revising the last observation carried forward measurement; (4) seeking broad labeling from enriched trials; and (5) approving drugs only for subpopulations.

Biomarker development is one of FDA's highest Critical Path priorities. The agency indicated in November it would be forming a working group by year-end to consider the development of a guidance on biomarkers (² (Also see "Biomarker Working Group To Explore Development Of Guidance" - Pink Sheet, 15 Nov, 2004.), p. 28).

"The one [biomarker] that would be most helpful would be a good predictor for hepatotoxicity:" the cause of "something like half of all the drugs that have ever been withdrawn," Temple said.

"We have very little clue from animal data....If there were some better way to predict it mechanistically, that would just make an enormous difference."

Though the Critical Path will serve to speed drug development in the future, Temple elaborated on Phase III dosing improvements industry can undertake today.

"One of the things that we've been advocating actually is more willingness to study multiple doses in Phase III," Temple said.

Despite the large costs and difficulty in examining multiple doses in large trials, the idea that a sponsor can discover all relevant dosing data in Phase II trials "is just fundamentally not correct," he noted.

Temple acknowledged that sponsors can study one or two doses in one study and two doses in another study. "I don't think you have to replicate every conceivable finding; you're supposed to bring it all together," he added.

One such example is anti-hypertensive drug trials where "every single trial is a multiple-dose trial."

"That's partly because it's easy to do. You can't do that in an outcome trial, I know, but there's an awful lot of symptomatic conditions where that can be done and I think there's a lot to be learned from it," Temple said.

"You get evidence of slope, evidence of effectiveness. You don't really lose anything, but you learn a lot about safety I think from having multiple doses."

The agency is also looking at revising its often criticized last observation carried forward measurement.

"We currently use a relatively crude measurement called last observation carried forward," Temple said.

"We don't really know that it would make much more difference to do the alternative ways, which is to sort of estimate what each person's outcome would be based on the data that you have for that person and the rest of the people."

One alternative is experimenting with ways of modeling outcomes. FDA plans to make greater use of mathematical and statistical models to guide drug development and approval decisions (³ (Also see "FDA Wants To Be “Model” Agency: Preparing For Clinical Trial Simulations" - Pink Sheet, 18 Oct, 2004.), p. 28).

Temple discussed the possibility that enriching a clinical trial by selectively enrolling likely responders may nevertheless allow a sponsor to seek a broad label for that agent.

"I don't want to try to give the definitive answer because I think it's going to be somewhat case by case," he said.

"The individual cases of whether you will apply it more broadly or not is going to have to do with how toxic the drug is and a bunch of things like that. Our history, however, is that we generally label for the enriched population."

Temple further discussed whether a drug with unknown overall efficacy, but substantial efficacy in a subpopulation, should be approved only for that smaller group.

"Suppose the fraction of people who respond is so small that you can't then do a trial to show that it works in the overall population," Temple said.

"This isn't settled yet," he noted. "I think there's a case for labeling the drug like that if it's very benign and if what it does for the people who respond to it is very valuable because you may just not be able to show it in the whole population."

As part of the Critical Path, FDA and its various centers are developing lists of top-priority opportunities. Release of the agency-wide list is being delayed several months due to more pressing safety matters (see ⁴ (Also see "Detour On FDA’s “Critical Path:” Safety Issues Divert Agency’s Attention" - Pink Sheet, 13 Dec, 2004.) ).

The Center for Drug Evaluation and Research's list will address oral contraceptive dosing and pediatric trial data.

"The urologic community was very interested in looking at the doses of the various kinds of hormones to be used in oral contraceptives and could we look back at all of the available data that the FDA had of all of the trials that we'd ever done," CDER Acting Deputy Director Douglas Throckmorton, MD, said.

"Similar sorts of things were found in pediatrics. There was a clear interest in looking at all of the available pediatric trial data figuring out when we could extrapolate from what we knew in the adults down to what we might use as far as safety and efficacy for children."

Throckmorton also took the opportunity to question industry on its "reluctance" to implement FDA suggestions.

"One of the things I've been struck with has been the reluctance on the part of industry, I'll call it reluctance, to implement some of these suggestions when they've been made to them [such as] the artificiality of the [clinical development] phases," Throckmorton said.

"Maybe it has to do with reluctance about compliance....You don't have to do this and we won't worry about it if you don't. Data collection is a similar thing," he added.

"Every large organization is inherently conservative and what has worked in the past is usually repeated," Hukkelhoven responded.

"I think it has been recognized that the GMP regulations have stifled innovation and the quality level in the pharmaceutical industry may not be comparable to the quality level in other areas."

In the last few years FDA has been trying to create more flexible regulations under initiatives part of FDA's "Pharmaceutical cGMPs for the 21st Century: A Risk Based Approach" program (⁵ (Also see "FDA Quality Systems Guidance To Be Modeled On Consent Decrees" - Pink Sheet, 27 Sep, 2004.), p. 28).

"We stick with things that we believe need to be done and cannot be changed because each time you change something the FDA may come and not like what you have done," Hukkelhoven added. "I think we should get out of that mood and if you can help us with guidance there or with open discussions. Maybe we don't need a 300 page [case report form], but nobody has the courage to actually bring it down to 15 pages."