TRENTON, N.J. — Authors of a study funded by Vioxx maker Merck & Co. failed to disclose in a report published in the New England Journal of Medicine in 2000 that three additional patients in a clinical study suffered heart attacks while using the now-withdrawn painkiller, the journal wrote in an editorial released yesterday.

The editorial, written by the journal's editor in chief, Dr. Jeffrey M. Drazen, executive editor Dr. Gregory D. Curfman and managing editor Stephen Morrissey, also alleges the study's authors deleted other relevant data before submitting their article for publication.

"Taken together, these inaccuracies and deletions call into question the integrity of the data on adverse cardiovascular events in this article," the doctors wrote. Excluding the three heart attacks "made certain calculations and conclusions in the article incorrect."

Adverse cardiovascular events include heart attacks, strokes and deaths.

The findings of what became known as the VIGOR study have been a key part of testimony in the three product liability trials to date over the withdrawn drug, including one in which a federal jury in Texas began deliberations yesterday afternoon. The research was published more than a year after the Food and Drug Administration approved Vioxx in May 1999.

The study was intended to compare whether Vioxx caused more stomach ulcers and bleeding among patients with rheumatoid arthritis than for those using the older, cheaper anti-inflammatory naproxen. Over an average nine-month period, Vioxx did score better on that count, but the study also showed there were a greater number of heart attacks among Vioxx users.

In the article, Merck explained differences found in that study by saying naproxen is cardioprotective.

Merck shares fell 61 cents, or 2 percent, to $29.68 in regular trading on the New York Stock Exchange, then fell an additional 85 cents, or 2.8 percent, to $28.85 in after-hours trading, after news of the journal editorial circulated.

In a statement issued late yesterday, Merck said the additional heart attacks "did not materially change any of the conclusions of the article." The company also said the information was not included because the heart attacks were reported after Merck's cut-off date for collecting information on the patients in the study.

"Nevertheless, these additional events were disclosed to the FDA in 2000, presented publicly at the FDA's Advisory Committee in February 2001, and included in numerous press releases subsequently issued by Merck," the company statement reads.

But Curfman said in an interview that Merck's arguments about the cut-off date "don't hold water" because journal articles are routinely updated with new data in the weeks before publication. "The health of the public, of many, many thousands of people, was at stake here," he said.

Data in yesterday's editorial show that 20 patients on Vioxx suffered heart attacks, instead of the 17 originally reported. Among patients in a comparison group taking naproxen, there were four heart attacks.

None of the three extra heart attacks was fatal, but all three of those patients were in a group at low risk of heart attack, Curfman said.

The journal editor said he learned of the extra heart attacks and deleted data — including the number of patients in the study who died — when he gave a deposition on Nov. 21 for attorneys representing Vioxx plaintiffs in the Houston trial and three other federal trials slated for early next year.

At the deposition, the plaintiff lawyers showed him documents they subpoenaed from Merck, including a July 2000 internal memo containing the deleted data.

Curfman said he and Morrissey spoke Monday with two researchers who led the study and the two, who do not work for Merck, said they would submit a correction to the journal as the editors requested.