Compendia disagree on drug-drug interactions

Researchers develop their own list of serious DDIs

Pharmacists may turn to drug interaction compendia for detailed information on serious drug-drug interactions (DDIs). New research, however, shows that the compendia seldom agree on which DDIs are considered to be of the highest clinical importance.

In the study, researchers found that of the 406 DDIs identified in one or more of the references as having the greatest clinical severity or importance, only nine were identified as major in all four compendia.

This study was the first part of a partnership devised to reduce clinically important DDIs. The partners include the Centers for Disease Control and Prevention in Atlanta, which provided financial support; the AdvancePCS Center for Healthier Aging in Hunt Valley, MD; the University of Arizona in Tucson, the University of Washington in Seattle, and the Albert Einstein College of Medicine in New York City. Their findings were published in the March/April 2004 issue of the Journal of the American Pharmacists Association.

To evaluate which DDIs are considered to be of the highest clinical importance, the researchers selected four compendia during the fall of 2001 that they thought health care professionals used the most to find out more about drug interactions. These were Evaluations of Drug Interactions (EDI), Drug Interaction Facts (DIF), Drug Interactions: Analysis and Management (DIAM), and MicroMedex (Drug-REAX) computer program. The researchers used specific criteria for each compendium to develop the clinically important drug interactions since the compendia used slightly different rating systems to classify the drug interactions.

The researchers also focused on the DDIs likely to be captured by computerized prospective drug utilization review systems in community and ambulatory pharmacies. Even so, health system or hospital pharmacists will see these combinations as well, says Daniel C. Malone, PhD, RPh, associate professor at the University of Arizona Colleges of Pharmacy and Public Health, and director of the Division of Pharmaceutical Policy within the Center for Health Outcomes and PharmacoEconomic Research.

Malone and the other researchers identified a total of 406 DDIs — listed as being at the highest level of severity in at least one of the four compendia. Thirty-five of these DDIs were listed in three compendia, 71 in two compendia, and 291 in only one compendium. Only nine of the DDIs were listed as being the most severe in all four compendia.

The researchers hypothesized several reasons for the lack of consistency across the compendia, Malone says. First, the different rating systems used in the compendia guarantee some variation. For example, EDI rated drug interactions using a four-item summary measure, such as Code 1: highly clinically significant. MicroMedex, on the other hand, used a five-item rating scale (major, moderate, minor, none, and not specified)."We tried to minimize [the discrepancy] by trying to be more inclusive, if possible."

An editor or a reviewer also rates the drug interactions, so they are somewhat subjective, Malone says. "There is no gold standard for comparison."

In addition, the evidence for most of these interactions is usually circumstantial, he says. The circumstantial evidence might include a few occasional cases reports, a pharmacodynamic study done in an animal model, or possibly a pharmacokinetic study done in an animal or human model.

"Usually, the amount of evidence to support the interaction is quite limited. Therefore, people are trying to make judgments about the seriousness of these interactions based upon little or sometimes no information. It is a theoretical basis for the interaction," Malone says.

In the study, the researchers included a table of the unique DDIs with the highest severity rating in three or four drug interaction compendia. However, an outside reviewer says he found a "disconnect from clinical reality" in the table.

"Where are the DDIs that experience tells us bring so many patients [those lucky enough not to die suddenly] to medical attention? Sulfonylureas and CYP 2C9 inhibitors, digoxin and P-glycoprotein inhibitors, angiotensin-converting enzyme inhibitors and potassium-sparing diuretics — the severity of these DDIs is undisputable, and countless patients are exposed to them each year," says David N. Juurlink, BPhm, MD, PhD, FRCPC, attending physician for the Divisions of General Internal Medicine, Clinical Pharmacology and Toxicology, and the Clinical Epidemiology Unit at Sunnybrook and Women’s College Health Sciences Centre in Toronto. He also is an assistant professor in the Departments of Medicine and Health Policy, Management, and Evaluation at the University of Toronto.

Identifying their own serious DDIs

After evaluating the compendia, the researchers then went on to develop their own list of clinically important DDIs, which they defined as those interactions most likely to cause harm if not detected. The list also was published in the same issue of the journal.

They used a three-step process to devise their list. First, they selected drug interactions from a review of drug interaction compendia followed by a systematic literature review. The researchers also developed an evidence report for each candidate DDI. These evidence reports then were evaluated and rated by an expert panel using a modified Delphi process.

For each DDI, the researchers tried to define the object and precipitant drug. The precipitant drug is the medication that affects the pharmacologic actions of pharmacokinetics of the object drug.

They identified a total of 25 clinically important DDIs after excluding products no longer available in the United States, products not likely to be used in an outpatient setting, and combinations that can be used for therapeutic benefit.

The panel did a better job of choosing interactions that occur in clinical reality, Malone reports.

"The paradox of these sort of ratings systems is that the interaction may exist in a theoretical sense, but when it is translated into day-to-day patient care, it means a lot less," he explains.

The researchers would have included other interactions had the focus been on the hospital setting, Malone notes. "Those interactions are certainly important, but all of the ones that we have identified here can and probably do occur in the hospital setting," he says. "Pharmacists in the hospital setting need to be aware of them, too."

The list may contain 25 interactions, but they can be logically grouped into several categories. "If you think of it in terms of three major groups: anticoagulants, monoamine oxidase inhibitors, and anti-infectives, you have captured 20 or 21 of the interactions identified. Then there is a small number of other interactions that fall outside that realm. It’s an easy way of putting the list into one’s memory," he says.

The downside to making such a list is that other interactions are important, too. "This is a first cut," Malone says. "I don’t want people to stop thinking about interactions after they have learned the 25. There are certainly other ones that are important."

The original goal was to identify 10-15 clinically important DDIs. That was expanded to 25, and Malone says that number could have been higher. "Phil Hansten [PharmD, a professor at the School of Pharmacy at the University of Washington], one of our expert panel members, constantly reminds me and everyone else that these are important interactions and ones that we should be aware of. There are others that are important, too, and the list is dynamic. It will change as new product and information comes out. It’s not something that one can learn and think [he or she] has a handle on it.

"You have to stay up on it. I think most pharmacists and physicians realize that."