HOUSTON — Conventional wisdom said the first federal Vioxx trial was Merck & Co.'s to lose.

The Florida man whose 2001 death of a heart attack lay at the center of the case took the once-popular painkiller for just a month. Richard "Dicky" Irvin got the prescription from his son-in-law without a medical checkup. His arteries were clogged by more than half in some places.

Merck insisted no science showed that Vioxx could heighten the risk of heart attacks or strokes unless taken for 18 months or more. A long-term study showing that 18 months of constant use could increase the chances of a heart attack prompted the company to withdraw the painkiller from the market in September last year.

But jurors couldn't reach a unanimous verdict — as required in federal court — after more than 18 hours of deliberations, prompting U.S. District Judge Eldon Fallon to declare a mistrial yesterday. He will confer with attorneys next week to set a date for a retrial.

For now, the result leaves Merck with a win and a loss in state trials and an undecided in its first battle in the federal forum widely considered to be more disciplined and corporate-friendly than many state courts.

Irvin's widow, though disappointed, matter-of-factly accepted the outcome and nodded as her attorneys vowed to retry the case.

"I know in my heart my husband had died because of Vioxx and I had to push forward because of that," said Evelyn Irvin Plunkett, 56, surrounded by the couple's apparently stunned children. "A lot of people have been injured, hurt, died because of that drug.

"I have never backed down from the feelings I had," she said.

The mistrial also leaves Merck facing a new jury that could hear allegations that the company withheld information from the New England Journal of Medicine about a 2000 Vioxx study, so the drug would appear safer than it was. A separate, longer-term study disclosed last year led to the drug's withdrawal.

Merck shares closed down 72 cents, or nearly 2.5 percent, at $28.41 yesterday.

Two jurors who voted in favor of Merck told The Associated Press only one of the nine jurors refused to absolve the company of liability.

One of those jurors, Amanda Toungate, said she didn't believe the drug caused Irvin's death, but added that Merck should have done a better job telling patients about Vioxx's risks.

"He had too many other risk factors," she said.

Merck lead attorney Phil Beck said that he could not confirm the 8-1 jury split because the judge instructed lawyers not to contact the jurors.

"If it is accurate, it's very gratifying; 8-to-1 is good enough to win in almost every state court in the country, including Texas. But in the federal system you need to win 9-0 and hopefully next time we'll get all nine votes," he said.

About 20 million people took Vioxx before Merck withdrew the $2.5 billion seller.

"You can't lie to people like the New England Journal of Medicine and get away with it," said Plunkett's lawyer Jere Beasley. "We look forward to the next trial."

Merck lead attorney Phil Beck said last week's revelations from the medical journal would be a "non-issue" in a retrial and that Merck supplied all the updated safety data to the Food and Drug Administration before the journal article was published.

The journal reported after deliberations had begun Thursday that authors of a report about the Merck-funded study — including Merck's head of clinical trials and a key witness for Merck, Dr. Alise Reicin — omitted three patients' heart attacks in the data they submitted to the journal, making it appear that Vioxx caused four times, rather than five times, as many heart attacks as the painkiller naproxen.

Merck attributed the disparity to naproxen's heart-friendly qualities rather than to a defect in Vioxx. The FDA criticized Merck for making the assertion, for which they had no scientific evidence.