Simulations based on a branded drug’s FDA-submitted data and limited use of a generic are adequate and can help shorten generic development timelines, FDA researchers said during a recent AAPS webinar.
By David Pittman
Modeling can be an approach to test how well complex generic drug-device combinations compare to the performance of the products they are mimicking.
Without such simulations, the development of generic alternatives can take years, possibly enough to cause drugmakers to abandon work.
CASE STUDY
Food and Drug Administration (FDA) researchers used an implantable contraceptive device as a way to prove modeling can shorten generics’ development. They outlined that work during a recent American Association of Pharmaceutical Scientists (AAPS) webinar.
The agency’s Office of Generic Drugs used data from FDA submissions for levonor-gestrel to determine what would be adequate bioequivalence (BE) for a one-year test period.
FDA approved the drug’s reference in 2000, but it still does not have a generic alternative in the U.S. despite the product’s patent expiring in 2015.
“A one-year in vivo BE study can significantly shorten the product development time and could also potentially encourage competition,” Satish Sharan, Ph.D., from FDA’s Division of Quantitative Methods and Modeling within its Center for Drug Evaluation and Research (CDER), said on the webinar Bioequivalence Evaluation of Generic Drug-Device Combination Products: “Implants.”
The 32mm body is implanted in a woman’s uterus and slowly delivers a total of 52 mg of a contraception drug over a five-year period. Because of the local delivery of the product and a five-year testing period to see if a generic works as intended, it is “practically infeasible” to use traditional pharmacokinetic-based BE testing, Sharan said.
So FDA researchers put their heads together to come up with a different way.
If they could remove an implant after a year to see how much levonorgestrel remained in the drug-carrying body, they would know how much product was delivered. And using a 90 percent confidence interval, FDA could model how well a generic product needed to perform after one year to fall within FDA’s standard 80 percent-to-125 percent BE standard for generic applications.
If a generic is between 95 percent and 105 percent of a reference product’s bioequivalence after 12 months, it will hit the agency’s standard, Sharan said.
FDA looked at data from levonorgestrel’s new drug application from 2000 to learn how much residual drug was left in the drug-carrying device and used it to create a formula to learn how fast or slow drug is released on which generics could be judged.
That formula was tested in a simulation that was repeated 1,000 times to conclude the rate at which a drug could be released in a woman’s uterus at a 90 percent confidence interval to be considered working well after five years. The FDA researchers’ method was also compared to products approved in Europe and similar products to see if it held up.
The conclusion: “Modeling and simulation can be used to assess the potential BE and statistical criteria for a five-year levonorgestrel intrauterine system,” Sharan said.
The FDA found that testing the product on 20 clinical test subjects would produce an adequate number for the agency, but that was for this particular drug, Sharan said. “We have done different subject sizes as well during our analysis,” he added.
Modeling can be a way around traditional in vivo BE testing, which can either be expensive for some or impossible for other types of drugs, said Lanyan “Lucy” Fang, Ph.D., from the division of quantitative methods and modeling at FDA’s Office of Generic Drugs within the agency’s CDER.
The AAPS webinar on levonorgestrel is an example of FDA’s work on using simulation to help expedite generic development. Lessons learned can be applied to other complex generic products, Fang said. “We are actively, intensively using these quantitative approaches,” she said. “That’s something I want to share with folks.”
FDA SUPPORTS MODELING IN DRUG DEVELOPMENT
In October 2017, FDA held a two-day workshop on advancing the use of modeling and simulation to modernize generic drug development and regulation. The goals of the workshop were to share the current state of knowledge, identify opportunities for modeling in drug development, and discuss where the science is headed.
“I know that there may be different challenges to using those approaches, and the agency is willing to work with, collaborate with different stakeholders in the field to promote the use of quantitative approaches in generic drug development and regulatory activities,” Fang said.
FDA representatives iterated during the AAPS webinar that they wanted to limit questions and discussion to the science and not future potential policy or regulatory steps. But they were interested in hearing from the scientific community about its work on modeling and simulation.
New FDA Commissioner Scott Gottlieb, who came to the agency in May 2017, has prioritized getting more generic drugs on the market, recognizing the role FDA regulations have in limiting development of complex, new generics.
Through an effort called the Drug Competition Action Plan, in October FDA released draft guidance to help generic drug companies request meetings with FDA before filing applications for complex generics with the agency. If sponsors better know what FDA needs to approve new drug applications, regulators can give better feedback in early, presubmission meetings and expedite the chances of approval.
More announcements are expected throughout 2018 on different ways to develop more tools and efficient alternatives to clinical endpoint testing for complex generic drugs, where feasible, Gottlieb said last year.
Listen to the webinar in full, and to learn about upcoming webinars or to submit a proposal.
David Pittman is a science and medical journalist based in Washington, D.C.