Seduced by hundreds of millions of dollars in advertising, Americans spend $6 billion a year for the arthritis pain-killers Vioxx and Celebrex, said to be as good as over-the-counter drugs -- and easier on the stomach.

But the two have not lived up to their hype, according to published research and interviews with arthritis doctors and drug specialists. Vioxx, which may be better for the stomach, appears to have a far worse side effect than over-the-counter drugs: an increased risk of heart attacks.

 For reasons doctors don't yet understand, Celebrex does not seem to be linked to heart problems, but there's conflicting evidence whether it has any benefit for the stomach.

There's also no evidence that these drugs are any better at fighting pain than comparable doses of older medications like ibuprofen -- though they cost six times as much. At CVS, a month's supply of ibuprofen at the doses people with arthritis normally take costs $17.55; a comparable supply of Celebrex costs $107.99, and Vioxx, $110.99. Depending on copays, the consumer may actually end up paying the same amount or even more for over-the-counter drugs, though, because insurance doesn't defray costs for non-prescription medications.

Vioxx and Celebrex, like their lesser-known cousins Bextra and Mobic, are in a relatively new class of drugs called Cox-2 inhibitors.

Despite the saturation advertising aimed at the 22 million Americans who suffer from arthritis, "there is no proof that these Cox-2 drugs are better for arthritis pain than traditional ibuprofen, and there is also no proof, except for a certain subset of patients, that these drugs are safer," said Dr. Neil Minkoff, medical director of pharmacy at Harvard Pilgrim Health Care, which insures 800,000 people in Massachusetts, New Hampshire and Maine. Some insurers, including Harvard Pilgrim, now insist that doctors get "prior approval" before writing prescriptions for them.

Back in the late 1990s, when the hype over Vioxx and Celebrex began, the excitement seemed well-founded. Researchers had long known that prostaglandins, natural chemicals that are made in cells under the direction of enzymes called cyclooxygenases, were the root cause of pain and inflammation.

But it turned out there were actually two types of cyclooxygenases, dubbed Cox-1 and Cox-2. Cox-1 releases "good" prostaglandins that protect the stomach. Cox-2 releases "bad" prostaglandins that further drive inflammation. The new drugs were hailed for their ability to block only Cox-2 and leave the good prostaglandins intact, something that traditional over-the-counter anti-inflammatory drugs like ibuprofen could not do.

Eager to get the pain-killing benefits without stomach-killing side effects like bleeding ulcers, doctors and patients flocked to the new drugs to help dampen the pain of arthritis, as well as migraines, acute pain and menstrual cramps. Then the bad news began trickling out.

The Cox-2 inhibitors actually "don't eliminate gastrointestinal side effects, they just reduce them somewhat, and even that is more problematic than first thought," said Dr. Jerry Avorn, chief of pharmacoepidemiology and pharmacoeconomics at Brigham and Women's Hospital, and author of the forthcoming book, "Powerful Medicines" on the subject.

In September, 2000, a study on 8,000 patients showed that people taking Celebrex had fewer ulcer complications than those taking over-the-counter medications.

But the study, published in the Journal of the American Medical Association, focused only on figures from six months of drug use, "the most appropriate" time-frame, according to Dr. Gail Cawkwell, worldwide medical team leader for Pfizer, which sells Celebrex.

When a later analysis by the US Food and Drug Administration looked at 12 months' worth of data, this apparent advantage disappeared, said Dr. Michael Wolfe, chief of gastroenterology at Boston Medical Center.