After about three hours of cross examination by a Merck & Co. (MRK) lawyer, a witness for the widow of a man who died after taking Vioxx reiterated Friday his belief that the drug causes heart attacks regardless of the length of use or of the size of dose.

Merck's lawyer repeatedly challenged how Wayne Ray, who studies the risk and benefits of drugs, examined data to reach his conclusion.

"No matter how you look at it, Vioxx causes higher risk every time," said Ray, who heads the Pharmaco-Epidemiology department at Vanderbilt University School of Medicine.

Jurors will be asked to decide if Vioxx contributed to the fatal heart attack suffered by Richard "Dicky" Irvin in May 2001. The 53-year old manager of a seafood distributor had been taking the drug for about a month to alleviate back pain when his colleagues found him dead at his desk.

Irvin's widow, Evelyn Irvin Plunkett, is suing Merck, which has scored a loss in Texas and a win on its home turf of New Jersey in the first two state-level Vioxx cases. The company faces about 7,000 state and federal lawsuits, and analysts have estimated its liability could reach $50 billion.

Irvin was taking 25 milligrams of Vioxx, and Merck lawyer Phil Beck focused on studies related to that amount. Patients were taking 25 milligrams in the study that led Merck to remove the drug from the market last year — because it showed a doubling of patients' risk of heart attacks and strokes after 18 months of use.

Ray testified that four out of five epidemiological studies showed that patients taking a 25 milligram dose of Vioxx carried a higher risk of heart attacks.

Beck questioned Ray' logic in reaching his determination, at one point noting that not all five studies were comparing Vioxx to the same drug. Beck also appeared to suggest Ray was selectively choosing data by highlighting Vioxx in comparison to Celebrex, a similar pain reliever made by Pfizer Inc. (PFE), in a trial that compared three drugs.

Toward the end of his cross examination, Beck said that if you "compare apples to apples," four of those epidemiological studies found that there is no greater risk of heart attack for patients using a 25 milligram dose of Vioxx.

"No, I don't think so," replied Ray. He said that "no matter how you slice the data you see an increase risk of (heart attacks and sudden death) with Vioxx at 25" milligrams.

Epidemiological studies are typically retrospective examinations of data. While they are considered useful because they examine how drugs affect the greater population, blinded, controlled clinical trials in which groups of patients are randomly given either a drug or placebo are considered the gold standard in the pharmaceutical industry.