[off-topic] Recent Cases Point To the Limitations Of Animal Drug Tests

[Guest guest]

BlankRecent Cases Point To the Limitations Of Animal Drug Tests

By ANNA WILDE MATHEWS

March 30, 2007; Page B1

The promising diabetes drug Galvus recently got turned back by the U.S. Food and

Drug Administration.

About 5,500 patients had taken the medicine in clinical trials at that point,

but the problem apparently wasn't with them. The agency was worried because some

monkeys who were given high doses of Galvus developed skin lesions. Humans who

took normal amounts of the drug for as long as two years didn't get the sores,

but the FDA refused to approve the drug until it saw more testing in people who

might be at higher risk.

The decision spotlighted an important unresolved scientific question: What do

the results of animal studies really tell us about humans? That question still

puzzles researchers even though guinea pigs, lab rats and their brethren have

long been part of experiments.

-------------- skip ----------------

Currently, the FDA is looking at results showing some anesthetic drugs led to

neurological damage in young primates.

Before regulators approve a drug, it typically has been tested on hundreds of

animals. The FDA requires initial testing in at least two species: one rodent,

one nonrodent. By the end of the process, mice, pigs, rabbits, dogs, monkeys and

other animals may have been used.

Many times, however, subtle results in animals are unclear and scientists just

don't know what to make of them. In the case of the new Novartis drug Galvus,

, the company's global head of pharmaceutical development, told

investors that Novartis researchers " do not understand -- do not know -- the

mechanism of the skin findings " in monkeys. They do know that " humans appear to

react to Galvus in a very different way. "

Another example of the confusing disparities that can arise in testing is the

case of the popular sleep drug Lunesta. It won FDA approval despite the fact

that tumors appeared when rats and mice took huge doses of a closely related

chemical cousin of the medication. Some FDA reviewers were concerned enough

initially to recommend rejection of Lunesta. After further analyses, however,

agency officials concluded the data from human testing didn't suggest a signal

for cancer in people. But you won't see the issue highlighted in the company's

ubiquitous green-moth commercials for the drug.

It can happen that a product doesn't hurt animals, but turns out to be poisonous

to patients. That occurred with the catastrophic British trial of an

experimental biotech drug called TGN1412, meant to treat leukemia and other

diseases. It didn't cause problems when given to monkeys and other species. Then

six people took it in a small initial study and had life-threatening convulsions

and organ failure. British regulators blamed an " unpredicted biological action

of the drug in humans " that wasn't foretold by the " apparently adequate "

preclinical studies.

-------------- skip ----------------

Scientists agree, there's no way to precisely and consistently translate from

animal to human results. After the withdrawals of the painkiller Vioxx and other

medicines with side effects that didn't fully reveal themselves during years of

testing, the need for such a translation looms particularly large.

Full story

http://online.wsj.com/article/SB117519602221153510.html?mod=health_home_inside_t\

oday_left_column