You might call it a made-for-TV drug. Approved for human use in the U.S. but not marketed that way, an arthritis medicine called Rimadyl languished for nearly 10 years in developmental limbo, then emerged in a surprising new form: Instead of a human drug, it was now a drug for arthritic dogs. And it became a hit.

With the aid of slick commercials featuring once-lame dogs bounding happily about, Rimadyl changed the way veterinarians treated dogs. "Clients would walk in and say, 'What about this Rimadyl?' " says George Siemering, who practices in Springfield, Va.

Today, those TV spots are gone. The reason has to do with dogs like Montana.

A six-year-old Siberian husky with stiff back legs, Montana hobbled out of a vet's office in Brooklyn, N.Y., six months ago accompanied by his human, Angela Giglio, and a supply of Rimadyl pills. At first, the drug appeared to work. But then Montana lost his appetite. He went limp, wobbling instead of walking. Finally he didn't walk at all. He ate leaves, vomited, had seizures and, eventually, was put to sleep. An autopsy showed the sort of liver damage associated with a bad drug reaction.

Pet drugs are big business -- an estimated $3 billion world-wide -- and Rimadyl is one of the bestsellers. It has been given to more than four million dogs in the U.S. and more abroad, brought Pfizer Inc. tens of millions of dollars in sales, and pleased many veterinarians and dog owners. But the drug has also stirred a controversy, with other pet owners complaining that nobody warned them of its risks.

Montana's owner, Ms. Giglio, is among them. After she informed Pfizer and the Food and Drug administration of her relatively youthful dog's death, Pfizer offered her $440 "as a gesture of good will" and to cover part of the medical costs. Insulted by the offer and a stipulation that she agree to tell no one
about the payment except her tax preparer, she refused to sign and didn't take the money. "There's just no way in my conscience or heart I can release them from blame," she says.

After reports of bad reactions and deaths started streaming in to the FDA, the agency suggested that Pfizer mention "death" as a possible side effect in a warning letter to vets, on labels and in TV ads. Pfizer eventually did use the word with vets and on labels, but when given an ultimatum about the commercials -- mention "death" in the audio or end the ads -- Pfizer chose to drop them.

Pfizer's director of animal-products technical services, Edward W. Kanara, says that when reports started coming in, "we acted extremely promptly based on the information we had." Pfizer points out that reported adverse events involve less than 1% of treated dogs.

Since Rimadyl's 1997 launch, the FDA has received reports of about 1,000 dogs that died or were put to sleep and 7,000 more that had bad reactions after taking the drug, records and official estimates indicate. The FDA says such events are significantly underreported.

While the numbers include cases "possibly" related to Rimadyl, it is hard to be sure. Many dogs given the arthritis drug are older, and few are autopsied after they die. Pfizer says it analyzed cases of Rimadyl treated dogs that died in 1998 and found a link to Rimadyl to be "likely" in 12% of cases and "not likely" in 22%; it says there was too little information for a judgment about the others.

Still Approved

Despite these problems, the FDA says Rimadyl deserves to be on the market, provided vets take the proper precautions. These include advising dog owners what bad reactions to watch for and periodically doing liver-function or other lab tests.

Within a few weeks, Pfizer will begin affixing a safety sheet directly to packages of Rimadyl pills. It is the first time either FDA officials or Pfizer can recall such a step being taken in the world of animal drugs.

Rimadyl -- generically carprofen -- is an anti-inflammatory medicine. Developer Roche Laboratories expected to market it for people in 1988 and received FDA approval, but shelved the plan after concluding the market for such drugs was too crowded. In addition, some outside experts expressed concerns; a commentary in a pharmaceutical journal noted unusual liver-function readings in 14% to 20% of test subjects and opined that "until additional data on carprofen are available, older compounds should probably be tried initially."

The idea of switching the product to the animal-drug track soon arose. A couple of corporate transactions later, it ended up in the hands of Pfizer's animal-drug unit.

There, it was treated to the kind of sophisticated marketing Pfizer does well. A survey of 885 dog owners was done. Besides shedding light on favorite dog names (Jake, Ginger, Lady), the poll revealed that one-fifth of dog owners would be willing to spend "whatever it took" to buy an aging dog an extra year or two of life. No fewer than 53% agreed that "my dog is a better companion than other members of my family."

The FDA requires safety and efficacy testing for animal drugs just as for human ones, but animal-drug tests are smaller. Pfizer says about 500 dogs got Rimadyl in various trials, which is no more than a fifth of the number of subjects in comparable human-drug trials. Some dogs showed unusual liver-function readings and one young beagle on a high dose died, but for the most part, the FDA and Pfizer didn't find side effects alarming. The drug was approved for an early-1997 launch.

That same year, the FDA made it easier to market drugs directly to consumers on TV. Soon, Pfizer was running commercials in which a once-stiff yellow Labrador retriever named Lady bounded over a fallen tree as she fetched tennis balls beside a lake. In another ad, a dog leapt through a window and slid down a banister.

There were also full-page magazine ads and a public-relations campaign, whose results, the PR firm later said, included 1,785 print stories, 856 radio reports and 245 TV news reports "generating 25.5 million positive impressions on the product."

Early on, vets were floored by the drug's effects. "The results in some cases have been pretty darn close to miraculous," says David Whitten of the Hilldale Veterinary Hospital in Southfield, Mich. "I'm using this drug on my own dog. It has been effective. But as with all medications, side effects are certainly a problem."

The First Complaints

Indeed, within months of the launch, vets at Colorado State University in Fort Collins noticed troubling reactions. Labrador retrievers seemed particularly affected. Since the safety studies for Rimadyl had emphasized testing on young beagles, Pfizer went back to conduct another, small test just on Labs; it says that test showed no particular problem.

Bill Keller, an FDA veterinary-medicine official, notes that "any time you take a product from the investigation and put it into actual practice, you're going to see things you didn't expect." But reports about Rimadyl came in by the hundreds. The FDA had received just over 3,000 animal-drug bad-reaction reports in 1996, the year before Rimadyl's launch; in 1998, the drug's first full year, Rimadyl alone produced more than that many.

They swamped the FDA's tiny Center for Veterinary Medicine in Rockville, Md. Pfizer was scrambling as well. "Basically, their response," says Dr. Keller, "was 'Tell us what you want us to do. We love the fact that it's selling so well, but we don't know what to do with all these adverse reactions.' "

The FDA and Pfizer discussed a "Dear Doctor" letter to be sent to vets. FDA records show the agency found parts of an early Pfizer draft "unacceptable as they are promotional in tone... ." It was revised.

The records also show Pfizer disagreed with the FDA's suggestion that the letter cite "death" as a possible side effect. To get the letter out, the FDA told Pfizer it was "agreeing to your exclusion of the 'death' syndrome from the letter at this time. However, we will revisit the 'death' syndrome issue and other potential side effects for possible inclusion in labeling at a later date." So the term didn't appear in the first warning Pfizer sent, in mid-1997.

Clear Benefits

Meanwhile, dog owners were asking for Rimadyl. "It was their advertising that sold me on the drug," says Michelle Walsh, a Phoenix woman who says her miniature schnauzer was given it and later died.

Not that vets needed much convincing. They saw clear benefits from the drug. On top of that, they could get points from Pfizer for each Rimadyl purchase they made; points were redeemable for PalmPilots, Zip drives for PCs and other equipment.

Although Pfizer's letter told vets to explain to owners the signs of a bad reaction to Rimadyl, such as vomiting, lethargy or diarrhea, it is evident that a great many didn't. The FDA's Dr. Keller says, "There are a lot of veterinarians who don't think they need to take the time, or who forget, or for whatever reason are not providing animal owners with this information."

Donna Allen, whose chow-mix, Maggie, started on Rimadyl last summer, says, "All my vet did was give me this little bag of pills, with no information." She says Maggie "didn't want to take it, but I made her."

After four weeks, Maggie began to vomit violently, Ms. Allen says. The dog vanished from their home outside Birmingham, Ala., and later was found lying in a ditch. Ms. Allen loaded her into a truck and sped 35 miles to a veterinary clinic, but the five-year-old dog died. Her vet wouldn't implicate Rimadyl in the death until Ms. Allen urged him to send the dog's internal organs to the University of Illinois vet school, where an examination showed liver toxicity.

Maggie was buried under a marker adorned with the figure of an angel. And Ms. Allen took to the streets, delivering a letter to all the vets in the area urging them to "understand that Rimadyl helps certain dogs, but it is poison to other dogs."

The D-Word

As the complaints poured in, the FDA told Pfizer it would have to revisit the label issue. Pfizer had referred to "fatal outcomes" on the label as a possible effect of the drug class to which Rimadyl belonged, but not specifically of this drug. Now the agency asked that Pfizer cite "death" prominently as a possible side effect of the drug. Describing the back and forth with Pfizer, the FDA's Dr. Keller says, "They did it. They weren't enthusiastic about it, but they have always been cooperative. And that's part of the nature of the game we play with industry."

But the FDA also wanted the word "death" in the audio of commercials. Pfizer indicated this "would be devastating to the product," FDA minutes of a February 1999 meeting show. A company spokesman says that "putting 'death' on a 30-second commercial and in proper context was something we didn't think was possible." Rather than do so, it eventually pulled the commercials.

Pfizer says it now will do traditional marketing to vets, making sure they know the proper way to use the drug. Another "Dear Doctor" letter will soon go out, and the company will start attaching a safety sheet to pill packages.

Pfizer acknowledges it has a perception problem with some dog owners; a consumer group, for instance, has mounted a campaign dubbed BARKS, for Be Aware of Rimadyl's Known Side-effects. The company is contacting dog owners who have told their stories on the Internet, and it is offering to pay medical and diagnostic expenses for some dogs who may have been harmed by Rimadyl.

But Pfizer stands firmly behind the value of the drug, of which it says sales have continued to grow. Most vets also remain strongly behind Rimadyl. Owners, too, generally say they think the drug is important -- they just want to know the risks.

Atlantan Roger Williams gave his mixed-breed terrier, William, Rimadyl for more than a year and believes it contributed to the dog's death. "But if I had to do it all over, I would give my dog Rimadyl again," he says. "The difference is I would have known what to expect. Without Rimadyl, William was miserable. And what's the point of living another three years if you're miserable?"