What could be done to prevent another Vioxx? This pain medication for arthritis became a blockbuster after its introduction in 1999, only to be taken off the market in 2004 when a study linked the drug to an increased risk of heart attack and strokes.
A new study published Monday in Archives of Internal Medicine offers an ambitious proposal to determine a drug’s risks sooner than they might otherwise become evident. The authors propose a system to examine widely prescribed drugs through safety analyses that would pool data as they emerge from various clinical trials of a medication and aggregate the information for a fuller picture of a drug’s harms and benefits.
As policy makers in Washington push for various forms of evidence gathering to determine the safety and effectiveness of medical treatments, the study proposes a broad model for monitoring drug safety that would consist of detailed publicly available data that independent researchers could freely analyze.
Such a database could be continually updated and aggregated with new information, as the results of new studies were published, to calculate a near real-time balance sheet of a drug’s risks and benefits.
Dr. Joseph S. Ross, the lead author of the study, said the objective of the researchers was to determine whether this kind of sequential cumulative analysis could serve as a new model for monitoring the safety of drugs after they enter the market.
“How could we be doing post-market surveillance in a better way?” said Dr. Ross, an assistant professor of geriatrics and palliative medicine at Mount Sinai School of Medicine in Manhattan. “For drugs that we are concerned about, that are high risk, this could be a blueprint going forward.”
Drug makers and researchers regularly conduct studies called meta-analyses in which they aggregate results from some previous studies of a drug, using the combined information to summarize the benefits and risks of a treatment. But the new study proposes a continuously updated model. Using data from Vioxx as their example, the researchers reported that Merck, the drug’s maker, could have known about the increased cardiovascular risk several years earlier if the company had used the methodology.
The cumulative analysis of the researchers showed that by May 2002 — more than two years before Merck pulled Vioxx off the market — study patients on the drug had a 39 percent increased risk of heart attack or death compared with people taking a placebo.
Dr. Ross and his co-authors have each worked as paid consultants to plaintiffs involved in lawsuits against Merck over Vioxx injury claims. The researchers used documents obtained through the litigation, including detailed data from Merck’s Vioxx studies, to conduct their analysis.
In a statement Monday, Merck questioned the methodology of the authors, saying that the researchers had used categories that were too broad. The researchers, for example, counted heart attacks and deaths that happened after patients had stopped taking Vioxx.
In 2000, a study financed by Merck had reported an increased cardiovascular risk in patients taking Vioxx compared to those taking a pain medication called naproxen. The Food and Drug Administration then asked Merck to add a warning to the drug’s label.
But Merck said that the company had conducted its own regularly updated analyses of clinical trial data on Vioxx and had not determined there was a problem until 2004.
“Based on our own analyses, we did not see an increase in cardiovascular events with Vioxx compared to a placebo,” Doug Watson, a cardiovascular epidemiologist at Merck, said in a phone interview Monday.
Merck voluntarily withdrew Vioxx in 2004, within a week of learning that a clinical trial had found an increased cardiovascular risk with the drug, the statement said.
Requiring the kind of constant cumulative analyses for all drugs proposed in the Ross article would require an act of Congress, said Dr. Janet Woodcock, the director of the F.D.A.’s Center for Drug Evaluation and Research.
The F.D.A. has stepped up its safety measures since reports of heart problems with Vioxx first surfaced, Dr. Woodcock said. The agency has added staff members in all divisions to monitor drug safety, has strengthened its Office of Surveillance and Epidemiology and has instituted a postmarket safety tracking system for new drugs, she said. The agency also regularly conducts its own meta-analyses of drug safety, particularly when there is a signal of a health problem, Dr. Woodcock said.
Meanwhile, in 2007, Congress passed an amendment to federal drug law that has given the agency authority to require drug makers to put in place risk management programs for certain drugs and to conduct postmarket safety studies, she said. The amendment also gives the agency the power to order changes to drug labels to reflect new safety information and requires drug makers to make public the results of clinical trials.
But some critics said Congress should give the agency more power — and more resources — to provide more detailed drug safety information to doctors and consumers. More information would be helpful, these critics said, because the studies required to gain approval for a new drug are often too short in duration or too small to identify all serious side effects. Sometimes, problems associated with a drug may not become known until it is used over time by several hundreds of thousands of people, they said.
“There is this kind of dogma in medicine that you shouldn’t use any drug for the first seven years after it’s released, because it takes that long to figure out its harms and benefits,” said Dr. Michael Steinman, an assistant professor of medicine at the medical school of the University of California, San Francisco.
Dr. Steinman listed a number of drugs, in addition to Vioxx, that makers removed from the market after reports of health problems: Baycol, a cholesterol-lowering drug taken off the market by Bayer in 2001 after reports of serious muscle problems; Zelnorm, a drug for irritable bowel syndrome, withdrawn by Novartis in 2007 after an analysis of safety data found an increased risk of heart problems; and Bextra, taken off the market by Pfizer in 2005 after reports that the drug could increase the risk of heart attacks and stroke.
Dr. Woodcock said cumulative analyses of the type the Ross article proposes could be useful to check out hypotheses about particular safety problems associated with particular drugs.
But there are so many drugs on the market that it would probably be impossible, if not prohibitively expensive, to cumulatively track all of them, said Dr. Elliott M. Antman, a professor of medicine at the Harvard Medical School. Still, Dr. Antman said it would make sense for the agency to give conditional approval to certain new drugs, requiring drug makers to file cumulative analyses of safety and efficacy data several years later in order to obtain permanent approval.