Fewer new drug OKs mean boost for new technology of pharmacogenomics

Street Smarts. By Scott Gottlieb, MD, amednews contributor. Feb. 11, 2002.

STREET SMARTS

A column looking at the corporate side of health care

SEE ARCHIVES

The recent withdrawal of Bayer Corp.'s Baycol, suspected to cause rhabdomyolisis and linked to 31 deaths in the United States, from the market was the 14th such recall since 1997. Each has been the result of safety concerns that cropped up after the drugs were approved.

In response, the Food and Drug Administration is taking longer to OK new drugs, and that is forcing drug makers to reset earnings expectations. That's one reason Nasdaq's Biotechnology Index is down 13% this year, to a close of 788.95 as of Jan. 26.

Biotechnology firms, more so than pharmaceutical giants, are especially vulnerable to these delays because they're usually dependent on a single product. In 2001, new molecular entities, the drugs with ingredients never marketed before in the United States, spent an average of 15.7 months being reviewed by regulators before approval -- past the high end of the old assumption of 12 to 15 months for approval. Drugs on a special "six-month" fast track for critical new medicines now take nine months or more to be approved.

The 24 new drugs approved in 2001 mark the fewest cleared by the FDA since 1994, when 23 passed agency muster. For most of the 1990s, there was a concerted effort by the FDA to increase the number of new drugs approved. The result was a doubling of the average rate of new drugs approved, to 40 a year at the end of the decade from 20 a year at the start.

A rising number of rare side effects were inevitable, as more new drugs with narrow indications are administered to heterogeneous populations far larger and sicker than any reasonable clinical trial can replicate. Even if a drug company could afford to sponsor trials large enough to uncover remote risks, it would not be practical to recruit enough eligible patients.

Despite all of its actuarial elegance, this statistical certainty provides little comfort to the FDA, which is under harsh criticism. Now we're back where we started, with regulators clamping down on new drug approvals. What has the FDA most worried is post-marketing surveillance, the so-called phase IV of clinical trials.

This is where FDA regulators rely on physicians to spot and report serious side effects caused by newly approved drugs. Regulators believe physicians have done a lousy job reporting these rare side effects. So, they've held up new approvals and withdrawn drugs whenever there's been a doubt about safety.

What can the FDA do if even the best doctors and exhaustive clinical trials can't be relied upon to uncover rare but recall-prompting side effects? New technology -- pharmacogenomics -- can mitigate their efforts. Scientists are learning how to read genes to tell if a patient will suffer a bad reaction before ever popping a pill.

Like computer code, it takes millions of different chemical signals -- the base pairs -- to make up each of the 40,000 genes distributed among chromosomes. When one of these signals is out of sequence, it is a called a single nucleotide polymorphism (known as "snips"). These snips contain the genetic variations that determine everything from the color of one's eyes to whether a drug will trigger dangerous side effects.

Among the firms specializing in uncovering the associations are Genaissance Pharmaceuticals, New Haven, Conn.; Variagenics, Cambridge, Mass.; and Sequenom, San Diego.

Like many biotech firms, their stocks have run hot and cold. All had strong IPOs in mid-2001, and all have come down significantly. Genaissance peaked at more than $30 but as of Jan. 26 traded for $4; Variagenics peaked at the same level and has dropped to $2.98; and Sequenom roared to $100 but has sunk to $7.45. The latter two companies, as often happens when a stock drops quickly, are the target of shareholder class-action lawsuits.

Despite the rocky nature of pharmacogenomic company stocks, there's still promise in the research underway. Pharmacogenomics can be used today to bolster post-marketing surveillance. All the thousands of patients prescribed a new medication could have blood spots taken and stored in an approved location.

Whenever serious adverse events are documented and characterized, DNA from patients who experienced the problem could be extracted and compared with DNA from the control patients who received the drug but did just fine. This would enable the creation of snip profiles that then could be used to screen those who are most at risk from being prescribed the new medication.

Genaissance is doing this with the class of drugs known as cholesterol-lowering statins, such as Baycol.

Also, the biotechnology industry is developing the infrastructure to create snip profiles, even as the FDA stands on the sidelines. Privately held firms such as Ardais, Lexington, Mass.; Genomics Collaborative, Cambridge, Mass.; and Lifespan Biosciences, Seattle; have sprung up for the sole purpose of banking tissue samples and correlating them with information about the patients they come from.

Publicly traded Impath, New York, says its similar effort, Predictive Oncology, is the fastest growing portion of the company. Like other pharmacogenomics companies, its stock is well off its mid-2001 peak; for Impath that has been a comparatively small drop -- to $37.50 from more than $60.

The FDA says it recognizes the potential of this new technology, but the agency hasn't issued guidelines outlining how a company might receive approval of a new drug based on a pharmacogenomic profile.

The wider use of more new drugs has generated record profits at the drug companies, but it is also exposing them to risks that jeopardize their business. More people taking more new pills are resulting in more adverse reactions. It is a statistical certainty that makes the drug companies victims of their own success.

Dr. Gottlieb is a resident in internal medicine at Mount Sinai in New York and a former analyst for the Wall Street firm Alex. Brown & Sons.