FDA might not always demand that a new oncologic be compared to the current standard of care if its trial was conducted before that standard was established, Office of Oncology Drug Products Director Richard Pazdur said at an Oct. 5 advisory committee meeting.

The study under discussion was for GlaxoSmithKline's advanced renal cell carcinoma agent Votrient (pazopanib). The trial was positive, but was conducted against a placebo control outside the U.S., right about the time that five new kidney cancer agents were emerging on the U.S. market.

"Obviously ... everybody would like to see a comparative trial," Pazdur told the Oncologic Drugs Advisory Committee. "However, the sponsor was going to do this internationally, and all the patients were enrolled outside of the United States, where these drugs were not available. So, did they demonstrate efficacy? Yes, they did meet their primary endpoint, and they had secondary endpoints that corroborated that effect. Is it the most optimal trial design, now in 2009? Probably not. But it wasn't designed in 2009."

FDA has not been as lenient in other therapeutic areas. Antibiotics sponsors in particular have been held back by trial designs that are deemed not acceptable by contemporary standards, as in the case with Advanced Life Science's Restanza (cethromycin) (¹ Also see "Advanced Life Sciences Restanza Blocked By New FDA Efficacy Criteria" - Pink Sheet, 8 June, 2009.).

Pazdur outlined the appropriate approach for assessing trials with obsolete comparator arms. "The first issue that has to be addressed is has the sponsor demonstrated safety and efficacy," he said. "And then the question is, are the results of that study referable to the U.S. population," or are there differences in supportive care, or in subsequent or prior treatments. "For example, if one was designing a drug in a first-line metastatic breast cancer population and none of the patients got adjuvant therapy, that would be problematic."

Evaluating pazopanib based on the trial data, the expert panel voted unanimously in favor of pazopanib, impressed by its five-month improvement in progression-free survival and a trend, though not statistically significant, toward an overall survival benefit (² Also see "Despite FDA Reservations, ODAC Votes Yes On Schering's PegIntron And GSK's Votrient" - Pink Sheet, 5 October, 2009.).

Cross-Over Design Dooms OS Benefit

But the placebo-control design led to another issue with the trial design: the panel noted that this lack of statistical significance for the survival benefit was likely because patients in the placebo arm were permitted to cross over to active therapy at the trial's conclusion.

One committee member asked Pazdur why FDA approves such protocols knowing that subsequent therapies could confound and in fact obliterate a survival difference by treating patients in all arms with active, and perhaps quite effective, therapy.

Pazdur concurred with the critique, but noted that in the end the patients benefit. "In fact, one of the ethical discussions that we've had internally about taking trials to ex-U.S. countries where drugs are not available, and having a placebo arm, is if there was cross-over everybody benefits - you're really not experimenting on people here, and basically you have a guarantee that all the participants will be getting active therapy."

The FDA official also noted that "if you have a really effective drug, even with cross-over therapy, you can see a survival advantage potentially." He pointed out that 50 percent of the patients in the pazopanib trial got cross-over therapy, "yet we're still seeing a really good [OS] trend here. Granted it doesn't [reach statistical significance], but those curves do separate."

In the end, the trend was enough for the panel.

"This drug is effective in terms of progression-free survival and probably would have shown significance in overall survival if not for the cross-over, so in my mind it's a good demonstration of efficacy," committee member Mikkael Sekeres, Cleveland Clinic, said.

ODAC Says Liver Tox Signal Is Weak

On the safety side of the pazopanib equation, ODAC members did not agree with FDA's primary concern: that the drug might have potential to cause liver injury.

Both FDA and GSK determined that there were four cases in the trial that fulfilled Hy's Law, a set of criteria FDA uses to distinguish drugs with high and low potential for severe liver injury. Three of those four cases in the pazopanib trial resulted in death.

FDA found this especially concerning given the availability of other treatment options. If approved, pazopanib, a new tyrosine kinase inhibitor targeting the VEGF receptor, would be the sixth targeted treatment for RCC to hit the market since 2005. Its predecessors include Bayer's Nexavar (sorafenib), Pfizer's Sutent (sunitinib), Wyeth's Torisel (temsirolimus), Novartis' Afinitor (everolimus), and Roche/Genentech's Avastin (bevacizumab) in combination with interferon alfa.

In this situation, it appears that the composition of the panel worked in GSK's favor. FDA included two hepatologists as temporary voting members. Their assessment was that though the four deaths in the trial were possibly related to liver toxicity, the cases were somewhat atypical of Hy's Law due to drug-induced liver injury - and argued that the risk benefit ratio still tilted in favor of pazopanib.

Also working in pazopanib's favor was the lineup of expert consultants GSK brought to defend the drug's safety profile, including Memorial Sloan Kettering oncologist and Chair of the National Comprehensive Care Network Kidney Cancer Guidelines Panel Robert Motzer, as well as Paul Watkins, who chairs the National Institute of Diabetes and Digestive and Kidney Diseases' Steering Committee for the Drug-Induced Liver Injury Network.

"I believe that I've seen telling evidence about the efficacy," said hepatologist John Vierling, Baylor College, who voted in favor of the drug. "The safety issues appear in my judgment to be manageable. ... and the Hy's Law cases are atypical of those that are associated with drug-induced liver injury and fatality."

Fellow kidney specialist Leonard Seeff, NIDDK, agreed: "I do think there is some concern about effects on the liver and I think that continued monitoring will be important." But, "I'm not convinced that all these cases are hepatotoxicity," he said. "I am not convinced from what I've seen here that there is a high signal for drug-induced liver injury."

Hy's Law Not Yet Law In Oncology

Noting that FDA accepts much more general toxicity for drugs when they are to treat life-threatening diseases like cancer, the panel asked for the agency's stance on where that bar fell for cases of Hy's Law.

"I cannot give you any guidance on that, and that's the reason we're discussing it," Pazdur said. "Obviously, we realize that in oncology we accept a much different risk-benefit situation than outside of oncology. ... We really don't have any extensive experience with using Hy's Law" - which is why FDA brought the matter to the advisory committee.

The panel had no clear advice for FDA on the topic. "Even the very best of us disagree on [drug-induced liver injury] because we don't have a biomarker, and so we try to use the very best information we have," Seeff (one of the hepatologists) said. "And for me, with this information, I think there are some cases that there was injury to the liver, but by no means all four of them."

The agency published a draft guidance on drug-induced liver injury in 2007, asking sponsors to enroll patients with preexisting liver disease and to focus on sussing out the potential signals (³ Also see "Liver Toxicity Guidance Urges Sponsors To Include Risky Patients In Trials" - Pink Sheet, 29 October, 2007., p. 29). But many questions remain about DILI, including proper diagnosis, and how to safely include the high-risk patients in trials (⁴ Also see "Drug-Induced Liver Injury Workshop Offers More Questions Than Answers" - Pink Sheet, 7 April, 2008., p. 22).

Ultimately, this primary safety concern was deemed acceptable in light of the drug's efficacy and because pazopanib's safety profile is comparable to the toxicity of currently marketed agents.

"As somebody who treats the renal cell carcinoma patients, I've had experience with all these drugs and they all have their plusses and minuses," said panel member William Kelly, Yale University. "This drug does have definite clinical activity, and the spectrum of side effects is different but may not be any more severe than the ones out there now."

"The only way we'll know the real benefit of this drug as we go forward is to do randomized trials against the other agents," Kelly added.

That information should be forthcoming. GSK has submitted the application in Europe, and the EMEA has asked the firm to perform analysis to try to contrast pazopanib's safety and efficacy with existing therapies.

- Jamie Hammon ( ⁵ j.hammon@elsevier.com )

GSK's Votrient Clears ODAC Despite Outdated Trial, FDA Safety Concerns

More from Archive

More from Pink Sheet