Merck & Co has altered one of the statistical tools used to analyse data from the APPROVe study of Vioxx – which revealed that it doubled the risk of heart attack and stroke in patients – but claimed yesterday the findings were unaffected.

In a conference call, Merck said errors had been made in the statistical methods used in the study – specifically those used to assess whether the increased risk with Vioxx (rofecoxib) changed over time compared to placebo, or was constant.

The onset and duration of elevated cardiovascular risk with Vioxx is at the heart of many of the lawsuits brought claiming injury from the product, with Merck repeatedly relying on the premise that unless a plaintiff had long-term exposure to the drug, their claims are without merit.

Although conceding that the statistical method used to test this in APPROVe was flawed, the company insisted that additional analyses came to the same conclusions. However, it saod it plans toretain an independent statistical specialist to review the data and analysis.

The company said it was informing the study authors of the move, as well as the New England Journal of Medicine, which published the findings last year, and regulatory authorities.

The findings of the APPROVe study have been a cause for much debate in recent weeks. Merck first revealed the results of an extension phase of the study earlier this month, claiming that it showed that once a person stopped taking the drug, the increased cardiovascular risk tailed off.

This would mean that people claiming injury by the drug, but who suffered a thrombotic event after the discontinued therapy, could find it harder to prove Vioxx was responsible.

But some eminent cardiologists, including Steve Nissen of the Cleveland Clinic in the USA, promptly disputed that conclusion, pointing to a non-significant, 67% increase in cardiovascular risk compared to placebo after treatment stopped that, it was suggested, might have become significant if the dataset were larger.

Nissen and other commentators also suggested that the APPROVe data indicate that the risks with Vioxx increase after only four to six months’ treatment, although Merck vehemently denies that interpretation. Also, the statistical correction does not apply to the extension phase of the study, according to the firm.

Merck is still facing more than 11,500 lawsuit brought by people claiming to have been injured by Vioxx, with the last two cases resulting in multimillion damages against the firm.