The simmering controversy over data from the 2004 clinical trial that caused Merck to stop selling Vioxx bubbled up again yesterday, as The New England Journal of Medicine retracted part of a paper it had published about the test.

In a correction posted online, the journal said that the paper's abstract — or summary — should no longer claim that Vioxx causes heart damage only when it is used for more than 18 months. That claim, published as one of the 2004 article's conclusions, has been at the center of Merck's defense against lawsuits filed by people who say that Vioxx, the brand name for the drug rofecoxib, caused their heart attacks and strokes.

As of March 31, Merck faced more than 11,000 lawsuits, covering 23,000 plaintiffs. The company pulled the drug off the market in September 2004.

In a separate article that the journal also published online yesterday, Stephen W. Lagakos, a statistician from Harvard who consults for the journal, also disputed Merck's claim that the drug's dangers were not evident among patients who took it for less than 18 months.

Last month, Merck acknowledged that it had previously misreported the results of a crucial statistical test the company had said proved its 18-month theory. In fact, the test does not prove that theory, the company acknowledged. But Merck argued at the time and continues to argue that other data support its claim, which many independent scientists have viewed skeptically.

The New England Journal of Medicine removed the 18-month claim from the abstract of the paper. But in an apparent compromise with the authors of the paper, the 18-month claim remains in a different part of the paper, the results section.

"It's still in both the results and the conclusion of the paper, though with some slightly different wording," said Sandra Jacobs, a spokeswoman for the journal.

Dr. John A. Baron, a professor at the Dartmouth medical school and one author of the original journal article, said that he believed that the correction did not undermine the 18-month conclusion of the original paper.

"On purely scientific terms, I think this is a very minor issue, and the data are what they are, the observations remain unchanged," Dr. Baron said. "The controversy is about the conclusions drawn from those observations. People are going to some extent draw the conclusions that they want to conclude based on a wide range of motivations."