By Turna Ray and Bernadette Toner
This article has been updated from a previous version to include additional information from Novartis and FDA officials.
Novartis plans to discuss with the US Food and Drug Administration what it needs to garner US marketing approval for Prexige, a Cox-2-inhibiting painkiller that the agency did not approve in 2007 due to hepatic safety concerns.
The company submitted lumiracoxib and a safety genetic marker for the management of osteoarthritis pain to the European regulatory authorities in December 2009. A Novartis spokesperson told Pharmacogenomics Reporter last week that the company "plans to have discussions with the FDA about the specific requirements for the approval of lumiracoxib when used in conjunction with a biomarker program."
The compound, lumiracoxib, was approved in the EU in 2006, but was withdrawn a year later from most of the 50 countries it was marketed in, including several European countries, Canada, and Australia, following reports of serious liver adverse events and deaths. The drug is currently marketed in a few Latin American and Caribbean countries with risk mitigation strategies.
Despite a report last year that Novartis was planning to refile a New Drug Application with the FDA for the drug with a companion genetic test, the firm was not ready to announce publicly its intention to do so. At the time, Novartis told Pharmacogenomics Reporter that it had plans to reintroduce Prexige with a genetic marker in several global markets, but not in the US.
Now, it is clear that last year Novartis was utilizing FDA's Voluntary Exploratory Data Submissions program to get feedback from PGx experts at the agency on the safety biomarker being used to develop a companion test.
Although the VXDS program allows sponsors to discuss PGx data with the agency without regulatory repercussions, an increasing number of these educational meetings between the agency and sponsors have resulted in NDA submissions, according to Issam Zineh, associate director for genomics at the Office of Clinical Pharmacology in FDA's Center for Drug Evaluation and Research, who spoke at a conference in Bethesda, Md., last week.
"Those who haven't considered VXDS… I suggest you do. We had extremely valuable discussion and feedback," said Kevin Carl, Novartis' director of global regulatory affairs, at the same meeting on pharmacogenomics, hosted by the Drug Information Agency. "Obviously, it's not binding, but it is useful for giving you some flavor and direction around whether or not the final marker has a future."
As part of the VXDS program, Novartis also met with FDA's Pharmacogenomics Working Group last year. The group, chaired by Lawrence Lesko, director of the Office of Clinical Pharmacology at FDA's CDER, is responsible for drafting guidance for industry on pharmacogenomics and meets with industry to discuss submission requirements for PGx data.
Michael Little, global head of diagnostics development at Novartis Molecular Diagnostics, said that the company's molecular diagnostics unit, which it launched last year, has been conducting biomarker studies for the drug and had identified a marker that can identify patients who should not receive it due to the risk of liver-related adverse events.
Little, speaking at the Molecular Medicine Tri-Conference in San Francisco last week, said that the resubmission of lumiracoxib could be the "first example" of a molecular diagnostic-based "drug rescue" in the industry.
"The genetic biomarker can identify patients at risk of certain liver-related side effects from lumiracoxib and make them ineligible for therapy," a Novartis spokesperson explained to GenomeWeb via e-mail. "A risk mitigation system would be implemented to screen out [osteoarthritis] patients with the genetic biomarker and exclude them from lumiracoxib treatment."
The PGx marker associated with hepatic safety in lumiracoxib-treated patients, DQA1*0102, was identified by researchers via retrospective analysis of the Therapeutic Arthritis Research and Gastrointestinal Event Trial, or TARGET. This study involved more than 18,244 patients, of which 9,117 were randomized to receive lumiracoxib, 4,730 received naproxen, and 4,397 received ibuprofen. More than 10,000 participants consented to DNA sampling and 4,518 lumiracoxib-treated patients were genotyped.
DQA1*0102 was shown in studies to have 83 percent sensitivity, 68.2 percent specificity, and 99.7 percent negative predictive value. Researchers estimate between 25 percent to 35 percent allele frequency in the general population. Novartis used an FDA-cleared in vitro diagnostic to genotype patient samples, Carl said.
"Overexclusion [of patients] is going to be a reality" when using the gene marker to determine who gets the drug, Carl noted. "We're okay with that since there are therapeutic alternatives. There are other options for these patients. So if we exclude inappropriately, as long as we're getting the patients at risk, that's really our goal here."
Carl noted that US and European regulators have indicated to the company that they support the utility of DQA1*0102 in predicting which patients are at risk for serious liver injury with lumiracoxib treatment.
"The biomarker will be used to foster confidence in markets where lumiracoxib was pulled," he said, adding that the gene marker will also help the company "cross the approval threshold" in countries where the drug was never approved, such as the US.
Novartis is planning to use a "widely used certified diagnostic test" in conjunction with the commercial relaunch of lumiracoxib. Genotyping will be done at accredited laboratories, according to Carl. Furthermore, the company is planning to conduct a prospective observational study to confirm the benefit of genetic testing once the Rx/Dx is broadly reintroduced on the market.