Biomarkers and Safety

Can biomarkers prevent another Vioxx disaster?

By Gil Y. Roth – Editor

In last month’s column, we explored recent advances in biomarkers—both scientific and regulatory—and discussed how they may lead to better drug development (and a quicker rate of failure, which is just as important to pipeline development). We cited drugs that benefited from biomarker research, including Herceptin and BiDil, which took very different paths to reach the marketplace.

This month, we look at how biomarkers may help when a drug is taken off the market (or help prevent it from being withdrawn). Vioxx marks the biggest withdrawal in our history, and its impact on the industry (and the rest of the COX-2 inhibitor class of anti-inflammatories) will play out for years. Can biomarkers help drug companies develop better safety profiles? And will we see a diagnostic test help bring Vioxx back to patients?

Glenn Miller, the general manager at Genzyme Analytical Services, remarked, “I think the major issue is a lack of understanding of what the important biomarkers might be. It goes beyond the current discussion about whether or not a biomarker is validated, although that has an important role to play here. Rather, it goes toward developing a scientific understanding of what the particular compound you’re interested in actually does.” He added that his division’s focus, offering pathology and molecular biology services to Pharma and Biopharma companies, can help identify the biomarkers that might be useful in clinical trials or as companion diagnostics for new drugs.

Total Recall

Historically, when drugs have gone down the recall path, they’re gone from the market and it’s too late to do much to save the product. However, since the COX-2 inhibitors were such a strong class of compounds, drug companies have much greater interest in finding a way to bring them back. In its last full year of sales (2003), Vioxx brought in more than $2.5 billion in revenues for Merck, while Celebrex and Bextra combined for $4.5 billion in sales for Pfizer in 2004. So there’s plenty of financial incentive for bringing the drugs back (Celebrex is still on the market, but its sales are way down).

It looks like it’ll be a while before we see a change; in September 2005, an FDA panel sent a “not approvable” letter to Pfizer for its injectable COX-2 inhibitor, which is marketed overseas as Dynastat.

“It’ll be interesting to see if something jumps out as a predictive marker for Vioxx retrospectively,” said Fred J. Pritchard, MDS Pharma Services’ vice president of drug development programs. “I think the predictive tools that are becoming available may be able to find the patient population that has the least risk from these types of drugs.”

Dr. Gordon Kapke of Covance Central Laboratory Services remarked, “It’s an interesting question of what’s going to come up in hindsight. I don’t think there’s a chemical biomarker that would’ve given us the information we’d have needed. Potentially, as we look at the evolution of cardiac biomarkers, Component T is very highly though of currently.” The problem is, would anyone have thought to look for Component T biomarkers if no one had noted the adverse event?

Mr. Miller commented, “If you take a look at Vioxx—and I’m not a cardiovascular expert, by any stretch of the imagination—but it seems to me that if you take a look at what went wrong with Vioxx and the publications around it, it’s clear that there was a lot of work going on, but that it wasn’t completely incorporated into the clinical trials. They were very focused trials, and the side effects that might have been looked at with particular tests were not included; consequently the studies would have to have been absolutely enormous to identify those particular side effects.”

He added, “I think you could possibly have had a study that, if you looked at particular biomarkers, you may’ve found a sub-population that would have turned out to be more susceptible to these side effects. That said, I think it’s a legit argument to say, ‘At the time of these studies, a lot of these issues weren’t known, and so we’re using 20/20 hindsight on this issue.’”

However, there is more and more info coming out—thanks to Human Genome Project and its offshoots—about the mutations (that’s “polymorphisms,” to you) involved, especially with the COX-2 inhibitors. Someday we may be able to segregate patients based on biomarkers, or at least go back and try to analyze the data about the adverse events and try to determine what constitutes the group that suffered these events. The question that we need to answer so that we can prevent another drug from having this effect is: Was there something common about the individuals who suffered cardiac problems after taking the drug?

This creates a dilemma that progress in biomarkers may only exacerbate. Let’s suppose a blood test is developed that shows a minority of the patient population is prone to cardiac episodes from the use of Vioxx. Is the FDA likely to approve bringing the drug back to market?

While Mr. Miller is optimistic that we’ll be able to form better gauges of safety through analysis of biomarkers and development of proper diagnostics, he’s not so sure it’ll lead to a return for Vioxx. “It’ll be very difficult to put a drug like Vioxx back on the market, even if there’s a safety test associated with it. The safety issue is—and should be—the primary concern of the FDA, and it’s going to be very tough to ensure that physicians actually use the tests that are out there, and don’t just prescribe the drug without it.”

After all, it’s one thing if a drug has some unpleasant side effects and a few patients fall through the cracks of safety testing. It’s another thing when the side effect is a fatal heart attack.

Indication Vs. Effect

In addition to the severity of the side effects, one scientist we spoke to contended that the “relaunch” issue will also hinge on the severity of the indication being treated.

“In oncology—given the very serious, critical natures of a lot of those treatments—there’ll be a different level of testing for polymorphisms, and maybe the results of those tests will lead to use of one protocol instead of another. But you’re talking about a very serious disease in that case,” said the source.

She added, “If you’re talking about arthritis, where it’s a long-term, chronic disease, in which there may be three or four different methods of dealing with the pain of the inflammation, then having a test that will separate out the people who could die from that treatment vs. those who will be helped by it is not so compelling an argument. I think the FDA is much more likely to say, ‘People may die from this treatment if they don’t get the test, and there are still other ways of treating it that don’t kill people, so why are we even making this decision?’”

The consensus among those we spoke to was that, if it’s a life-threatening disease and there is no other alternative, the Agency may approve a product with heavy-duty black-box warnings to test effectively. But it appears that, for chronic/non-threatening indications, the FDA will be much stricter on safety, and biomarkers will be used more for efficacy purposes.

Others agreed that the FDA’s first cause for approving a drug is its safety profile. We’d all like to see greater efficacy, of course—and that’s where targeted therapies have to come in—but drug companies have to pay attention to safety first.

Personalized Wakeup Call

Some feel that the issues with COX-2 drugs will be a wakeup call for drug companies, as far as biomarkers go. “Years ago, big drug companies were averse to the idea of personalized medicine,” said one source. “I think the last couple of years, and especially what we’re seeing with Vioxx, has brought them to the realization that they are going to have to break up their patient populations and reduce the individual markets for drugs by developing efficacy and safety diagnostics. Now I think they’re going to be developing series of drugs, or classes, so that they can reach individual patients with the right products. In total, they may have a series of targeted therapies that may have a significant market size, and they’ll have to learn to deal with smaller volumes per product. It’ll offset with a greater market penetration in each patient group.”

Another industry figure commented, “If you look at the expense of developing an appropriate patient population screening test and diagnostics and monitoring over time, and compare that to the cost of the Vioxx recalls and lawsuits, then there’s no discussion. What was that first Vioxx penalty? A quarter of a billion dollars?”

Said Mr. Pritchard, “I think the big message Vioxx brought us is that just because you get a drug to market, it doesn’t mean you’re through understanding the disease, or improving how the drug should be prescribed. The risk of drug development that companies take on doesn’t end when they start getting a return on their investment.”

The economics of personalized medicine is more complicated than that, of course. Dr. Steven Toon, senior vice president of clinical pharmacology, head of the Medeval unit of ICON, reminded us that economics are regional, and that personalized medicine may have more obstacles than science. “In the UK, the economics of the health system might dictate against personalized medicine,” said Dr. Toon. “With a socialized medicine system, there has been a resistance on the part of purchasing agencies against drugs that can be shown to have a very, very good efficacy in a specific subgroup of the population.”

Genzyme’s Mr. Miller added, “Even though the Human Genome Project is ‘finished,’ we’re only at the beginning of extracting the information out of it. Medicine is still going to be an art, for several generations. The only way you can make it a science is to completely understand the impact of particular medications or the mechanism of a particular disease. It’s going to be a long time before we’re at that level. What we’re doing in the diagnostic industry—and in personalized medicine—is helping to support physicians in the practice of their art.”