Reactions to Prescribed Drugs Kill Thousands Annually, Study Finds

New York Times

Wednesday, April 15, 1998


More than 100,000 people a year die in American hospitals from adverse reactions to medication, making drug reactions one of the leading causes of death in the United States, researchers are reporting Wednesday.

The deaths, as described in the current issue of the Journal of the American Medical Association, are not due to mistakes by doctors in prescribing drugs or by patients in using them. Rather, drug reactions occur because virtually all medications can have bad side effects in some people, even when taken in proper doses.

"We want to increase awareness that drugs have a toxic component," said Dr. Bruce Pomeranz, an author of the study and a professor of neuroscience at the University of Toronto. "It's not rare."

The findings are a surprise, though, because deaths from adverse drug reactions often go unrecognized. Even when the cause is known to be a drug, that is rarely recorded on the death certificate, Pomeranz said; a certificate might list a stomach hemorrhage as the cause of death, without mentioning the drug that brought it on.

Researchers said Tuesday, however, that medications did far more good than harm and that patients should not stop taking them for fear of reactions.

The study said that the average number of hospital deaths from medications was 106,000 per year. In 1994, for example, there were 106,000 such deaths as well as about 2.2 million nonfatal adverse drug reactions. Nearly 7 percent of hospital patients were found to have had reactions, Pomeranz and his team at the university found, and 0.32 percent suffered fatal reactions. "These numbers are surprisingly high," Pomeranz said.

But the scientists reached their conclusion not in one large study but by combining the results of 39 smaller studies. This technique, called meta-analysis, can enable researchers to draw statistically significant conclusions from studies that individually are too small, but critics of the method say it can produce questionable results. The authors of the study published Wednesday emphasized that its conclusions should be viewed with caution.

Still, even researchers who said the study may have overestimated the number of drug-related deaths said it had identified a serious problem.

Dr. Lucian Leape, a professor of health policy at the Harvard School of Public Health, said that he had "problems" with the figures and that the real number of drug-related deaths might be half what the researchers estimated. Even so, he said, "it's still a big problem. This is a very useful piece of research."

No comprehensive records are kept of medication-related deaths, Pomeranz said. Doctors and hospitals rarely report such deaths. Health care officials are not required even to keep track of them, although the Health Care Financing Administration is developing regulations that would compel hospitals to monitor patients for adverse drug reactions.

"Do you want to guess how many death certificates in 1994 showed deaths due to drug reactions?" Pomeranz asked. "It was 156."

The Food and Drug Administration, which asks doctors to report adverse drug reactions, received notice of only 3,500 such deaths in 1994. And that number is not comprehensive because the reporting is voluntary.

Dr. Michael A. Friedman, lead deputy commissioner of the FDA, said that the agency's reporting system was not meant to be encyclopedic but to identify new, previously undescribed adverse drug reactions so that the agency could alert the public and the medical profession to them.

Friedman declined to comment on the large number of deaths from adverse reactions reported in the study, but he said "we recognize that every medication has some side effects," particularly the powerful medications taken to treat life-threatening conditions like cancer and severe heart disease.

Pomeranz and his colleagues based their research on 39 studies performed in American hospitals by other researchers from 1966 to 1996. These studies were designed to monitor patients for adverse drug reactions, which were considered separately from mistakes in prescribing or administering medications.

Pomeranz's team extrapolated from those studies to make national estimates. Initially, the researchers looked at data from around the world, Pomeranz said, but ultimately focused on the United States because the most studies had been done here.

Patients in the studies had had drug reactions while in the hospital, or were admitted because of such reactions. About a quarter of the problems were allergic reactions, and the rest were other side effects.

After analyzing the results of the 39 studies, Pomeranz and his colleagues estimated that there were 76,000 to 137,000 deaths a year from drug reactions. They averaged the figures to estimate that the number of deaths was mostly likely about 106,000.

Using the lowest estimate makes drug reactions the sixth-leading cause of death in the United States, higher than diabetes and pneumonia and lower than heart disease, cancer, stroke, lung disease and accidents. Calculating from the mean, drug reactions would be the fourth-leading cause of death.

Dr. David Bates, an associate professor of medicine at Harvard University who wrote an editorial in the journal about the study, said other studies had shown that certain types of drugs account for the most adverse reactions: antibiotics, heart medications, blood thinners and chemotherapy agents for cancer. He said that the patients most vulnerable to adverse reactions were those in intensive care, who are sickest and most likely to be receiving many different drugs -- as many as 20 to 40.

When people do die from a medication, he said, it is most likely due to failure of the kidneys or liver, disturbances in heart rhythm or destruction of the bone marrow.

Bates, who has conducted studies on drug reactions, said the estimates in Pomeranz's study should be "viewed with caution." He said he agreed with Leape that they might be high, because the figures came from large teaching hospitals with the sickest patients, where more drug use and higher rates of drug reactions would be expected than in smaller community hospitals. Bates said he thought the true number of deaths might be closer to 50,000 a year, which he said would still be cause for great concern.

Better awareness can help prevent some of the deaths, Pomeranz said, even if they are not caused by errors in prescribing or dosage. In one study, he said, just installing a computer system in a hospital to monitor drug reactions lowered the rate of adverse reactions.

And better record keeping can help prevent patients' being given drugs they have had allergic reactions to in the past, Pomeranz said.

Bates also said that increased awareness and monitoring could help prevent some reactions. "If you want to decrease the rate, you have to be able to measure it," he said.

Bates also said it was important for doctors to warn patients about potential side effects of medications, and for patients to report reactions to their doctors quickly.

But Leape said that people might not realize that their symptoms came from medication rather than the illness they were being treated for. "Patients come to doctors all the time with symptoms from drugs and don't know it," he said.

Even though some reactions cannot be avoided today, Bates said, they may become preventable in the future. Drugs can sometimes be improved, for instance, to eliminate unwanted effects. The widely used allergy drug Seldane is one example: after it was found to cause heart problems in some people when combined with other medications, it was reformulated to reduce the risk.

Despite their concern over drug-related deaths, both Pomeranz and Bates said that in the vast majority of cases, the benefits of medication far outweighed the risks.

Bates said that drugs probably saved millions of lives every year, and that it would be a dangerous mistake for people to abandon their prescriptions for fear of serious side effects.

"People should not be frightened of their medications," said Friedman of the Food and Drug Administration. "They should be respectful of them."

Other Places of Interest on The Web Incidence of Adverse Drug Reactions in Hospitalized Patients , an abstract from The Journal of the American Medical Association .