No early warning system

Harlan Krumholz ’80, a doctor, usually stays away from litigation. But when lawyers asked him to testify several years ago in lawsuits over the pain drug Vioxx, the School of Medicine cardiologist saw an opportunity.

The suits were part of a wave of litigation by patients who had taken Vioxx and suffered heart attacks. Vioxx’s manufacturer, Merck & Co., had pulled the drug from the market in 2004, after it was found to increase heart attack risk. Krumholz wanted to know: could the risk have been identified earlier? He told the lawyers he wanted raw data from Merck’s clinical trials. But more than that, he wanted “to generate knowledge that would improve drug safety going forward.” So he asked that the data be made public.

In his testimony, Krumholz explained how Merck’s clinical studies showed increased risk. After the court trials were over, he and his colleagues went to work on the data, combining the results of numerous small clinical trials. Standing alone, each trial might have shown only one or two unexpected heart attacks, Krumholz says. “They need to be pooled so you can see the real picture”: that in 2001, two years after Vioxx went on the market, patients taking it demonstrated a 35 percent greater risk of heart attack than those taking placebos. (The work appears in the November 23 Archives of Internal Medicine.)

Krumholz’s larger point: there is no requirement for investigators to release raw data—only summary results, and only within a year. He’d like a requirement that complete data be posted on the Internet within six months of each clinical trial; in the case of Vioxx, he says it might have prevented thousands of heart attacks. “The purpose” of his studies, Krumholz adds, “was not to kick Merck, but to learn lessons.”