Tips for heading off harmful drug interactions
From the September ACP Observer, copyright © 2005 by the American College of Physicians.
By Christine Bahls
SAN FRANCISCO—In 1998, the issue of drug interactions took center stage when the Food and Drug Administration (FDA) yanked the blockbuster antihistamine terfenadine (Seldane) off the market. The agency said the drug—when taken with the antibiotics erythromycin or clarithromycin or the anti-fungal ketoconazole—could turn deadly, prolonging patients' QT interval or causing ventricular arrhythmias.
The FDA's decision to ban terfenadine came after it put a black box warning on the drug labeling and made repeated attempts to get physicians to stop co-prescribing the drug. Harmful drug interactions also caused the market to withdraw mibefradil (Posicor) and cerivastatin (Baycol).
"Seldane changed the pharmaceutical industry and the way it handled drugs in terms of drug interactions," said David A. Flockhart, ACP Member, who spoke about drug interactions at Annual Session. "And it changed the FDA. They are more careful with interactions with new drugs."
While only a few drug interactions really matter, said Dr. Flockhart, chief of clinical pharmacology at Indiana University in Indianapolis, "it's necessary to pull out the ones that do." Those include drugs that internists prescribe many times a day, such as anticoagulants, calcium channel blockers, statins, antidepressants, antiseizure medications and antibiotics.
Internists need to be especially wary, he said, of drugs that are of narrow therapeutic range, cause or augment arrhythmias, or lower the seizure threshold. And as more drugs enter the market and aging patients take multiple prescriptions, Dr. Flockhart added, the risk of adverse drug interactions will continue to rise. (See "Where to look for help.")
While the list of interactions is relatively short, Dr. Flockhart said the issue of drug interactions is remarkably complex.
Drug interactions can take place in an IV bottle, the gastrointestinal tract and, especially, the liver. They can occur between different drugs, drugs and disease states, drugs and metabolizing enzymes or organs, or drugs and foods or food supplements.
Some interactions actually boost the efficacy of the medications being combined, Dr. Flockhart noted, citing the HIV drugs saquinivir and ritonavir. "These compete for the same enzyme," he said, "so you get a much better drug because more of the drug becomes available."
But many more interactions sabotage the efficacy of one or more medications—or lead to dangerously high drug levels. The first line of defense against drug interactions is to get as comprehensive a medication history for patients as possible. That's tough, particularly when patients take dietary supplements and herbals.
"Administrative assistants are more likely to get accurate herbal information [from patients] than a nurse—who is much more likely to get that information than you are," said Dr. Flockhart,who advised physicians to enlist staff to quiz patients about their supplement use. (See "Anxious for cures, patients turn to untested remedies.")
Physicians should also find out about patients' drug allergies, history of previous drug interactions or dependence, and family history of problems with different medications.
And they should ask patients what medications they have recently stopped taking and when they stopped taking them. "Some drugs—ike [paroxetine] Paxil come 'off' very quickly," Dr. Flockhart said, "but many medications, like [fluoxetine] Prozac, can affect interactions even two weeks after patients stop taking them."
Blockers and inducers
Another key element in avoiding drug interactions is knowing which drugs block the effectiveness of other medications and which, when taken together, will boost drug levels.
"In the gastrointestinal tract, nearly all interactions with heavily positive cations—like aluminum, magnesium and iron—will bind drugs tightly," Dr. Flockhart said. "Sucralfate, for example, used to treat ulcers, is heavy in aluminum and blocks absorption of all the quinolone antibiotics fairly effectively." The same is true of all heavy magnesium compounds, including antacids and all oral iron preparations.
"We don't think of iron as a drug, and a lot of young and older women don't either," said, Dr. Flockhart, who urged physicians to ask patients about their use of iron supplements, "especially if you're giving a quinolone." Repeatedly, Dr. Flockhart said, physicians will use quinolone antibiotics—or tetracycline or azithromycin—to treat urinary tract infections when patients "don't get the quinolone" because they are taking iron supplements. "It doesn't," he added, "get better using more expensive antibiotics, like levofloxacin."
The same holds true for all medications that block acid, like proton pump inhibitors and H2 blockers, which dramatically reduce the absorption of ketoconazole, delavirdine and indinavir—important interactions to know about when treating HIV patients, Dr. Flockhart pointed out.
Then there are drugs that inhibit cytochrome P450 enzymes, a family of enzymes (better known as CYP) in the GI tract and liver that metabolizes a great many medications, he said. Drugs metabolized by these enzymes include all calcium channel blockers, benzodiazepines except valium, protease inhibitors, nonsedating antihistamines—and every drug ever shown to prolong the QT interval.
The list of inhibitors include ketoconazole, fluconazole at high doses, erythromycin, clarithromycin and grapefruit juice (as well as juice made from Seville oranges.) When patients ingest some inhibitors, like grapefruit juice, it can take the CYP system nearly a week to recover, Dr. Flockhart said—and the half-life of the drug doesn't matter.
Enzyme inducers, on the other hand, include carbamazepine, felbamate, phenytoin, rifampin, rifabutin and topiramate.
The P450 enzyme family is also, he said, "the reason why so many drug interactions occur with warfarin and phenytoin." (See "The role of pharmacogenomics.")
Other key factors to keep in mind to avoid drug interactions include:
Narrow therapeutic range: Metabolic dysfunctions of enzymes in individual patients can turn a drug that normally does not have a narrow therapeutic range—such as valium—dangerous. Dr. Flockhart recalled one patient participating in a clinical trial of valium who received 10 mg of valium via IV at 8 a.m. one morning—and did not wake up until 3 p.m. the next afternoon.
But many of the drugs that have a narrow therapeutic range are well-known to physicians, including warfarin—which can be inhibited by certain antibiotics—and phenytoin. Other drugs of narrow therapeutic range that need to be monitored closely include cyclosporine, digoxin, tricyclic antidepressants and theophylline.
Cardiac problems. Another dangerous type of interaction—which doomed terfenadine—prolongs patients' QT interval. Dr. Flockhart cited a study in the Sept. 9, 2004, issue of the New England Journal of Medicine that found the QT interval prolonged when erythromycin was mixed with strong inhibitors of CYP3A, such as the channel blockers verapamil and diltiazem. (The study is online.)
In the study, the adjusted rate of sudden death was twice as high for patients using both erythromycin and a calcium channel blocker, as opposed to patients who used none of the antibiotics in the study. Instead, Dr. Flockhart advised prescribing penicillin or cephalosporin with calcium channel blockers.
Reduced seizure threshold. Antibiotics—particularly those in the quinolone class, such as ciprofloxacin and levafloxocin—can again be the culprit in interactions that lower the seizure threshold, when interacting with neuroleptics. Give quinolones and neuroleptics together, Dr. Flockhart said, either by IV or mouth, and you're "asking for trouble."
Other drugs when, mixed with neuroleptics, can reduce a patient's seizure threshold include the antibiotic imipenem-cilastatin and the opioid normeperidine.
Muscle pain/rhabdomyolysis. Interactions that can lead to rhabdomyolysis become an issue in cholesterol treatment, as physicians combine statins and fibrates to bring cholesterol levels down. Drug interactions with cerivastatin, for example, resulted in the accumulation of the parent drug by inhibiting its metabolism, leading to a greater rate of rhabdomyolysis and the drug's removal from the market.
'You can't know all of the drugs, but you should be well acquainted with a small number so you are your own personal expert.'
—David A. Flockhart, ACP Member
Then there is the important issue of timing and interactions. Patients taking ibuprofen at the same time as aspirin, for instance, will find that ibuprofen "gets in the way of COX-1 effects of aspirin and reduces its efficacy," Dr. Flockhart said. Instead, patients should wait three to four hours between taking aspirin and ibuprofen.
And to help protect patients, Dr. Flockhart advised physicians to expand their own base of knowledge. Physicians should know 10 drugs "cold," he said. "You can't know all of the drugs, but you should be well acquainted with a small number so you are your own personal expert. You need to put drugs through a process to get onto your formulary, including a careful review of the literature."
The information included herein should never be used as a substitute for clinical judgment and does not represent an official position of ACP.
Each year, Americans spend more than $21 billion on herbals and other supplements. But according to Stephen W. Bent, ACP Member, assistant professor of medicine at the University of California, San Francisco, the Food and Drug Administration does not completely regulate these products. And only a few handfuls of the thousands of complementary products on the market have been seriously tested for efficacy and safety.
"Many companies will do a small trial with 20 patients for a week, and say 'there were no problems,' " said Dr. Bent, who does research on the safety of complementary products. "That is completely insufficient."
Because herbals and other dietary supplements can play a big role in drug interactions, it's important that physicians and patients discuss the pros and cons of using supplements.
But patients are often reluctant to talk about their personal formulary. To get patients to feel more comfortable discussing their herbal use, Dr. Bent said he tells them about the number of people nationwide who use these products—and that he is interested in interactions, so he'd like to know what they are taking.
The typical herbal or supplement user is a female with urinary tract problems, chronic pain or a psychiatric illness, said Dr. Bent. "They tend to be people who have problems for which medical treatments are not particularly effective". The herbal alternatives, however, can be dangerous. Only St. John's wort has been seriously tested, Dr. Bent said—and those studies have revealed serious drug interaction issues. The herbal can eliminate the efficacy of birth control pills, inhibit chemotherapy and protease inhibitors, and interact with some antidepressants and the blood thinner warfarin.
Other drug interaction information on herbals and supplements is based on case studies, "which are not a great form of scientific evidence," Dr. Bent said. He recently reviewed 15 case studies that showed gingko balboa could pose a bleeding risk in patients taking warfarin. The other "G" herbals—garlic and ginseng—could have the same effect, he said.
For more information on drug-supplement interactions, see PIER.
In an Annual Session presentation on drug interactions, David A. Flockhart, ACP Member, told internists where not to look for information they need on drug interactions: the "Physicians' Desk Reference," which he said lists only those drugs that are still under patent protection.
Instead, he steered internists to other resources they can use to determine what type of interaction, if any, they should expect when prescribing different medications to individual patients:
Washington Manual of Medical Therapeutics. Now in its 31st edition, the "Washington Manual," which is produced by Washington University School of Medicine, is also available in PalmOS and PocketPC formats. Available at major book retailers, $59.95.
www.drug-interactions.com: Run by Dr. Flockhart's group at Indiana University in Indianapolis, this Web site is organized by the specific enzymes belonging to the CYP3 system. Free access.
Gold Clinical Pharmacology. Produced by Gold Standard Multimedia, this resource includes CD-Rom, Internet and PDA. For pricing information, go online.
ePocrates. Available in free-standing models or software packages. For pricing information, go online.
Lexi-Comp: Has handheld and online products, including PDA downloads. For pricing information, go online.
The most widely prescribed opiate is codeine, but 7% of all patients can't metabolize the drug to produce analgesic morphine, so get no pain relief from using it.
According to David A. Flockhart, ACP Member, that group of patients includes 5% of the African-American children with sickle cell disease who have been given acetaminophen with codeine as outpatients during sickle cell crises, and been labeled in their charts as drug abusers who need high does of opiates.
"They're getting no analgesic effect from codeine," said Dr. Flockhart, chief of clinical pharmacology at Indiana University in Indianapolis, "so they should get morphine or another good opiate."
Pharmacogenomics play a major role in drug interactions for many conditions, said Dr. Flockhart, who spoke about drug interactions at Annual Session. Just recently, for instance, the FDA approved labeling that included specific pharmacogenomic testing guidelines for three oncologic drugs: six-mercaptopurine and azathioprine, which are used to treat acute lymphocytic leukemia; and irinotecan, used in colon cancer treatments.
As more genetic tests become available, physicians will play a growing role in deciphering how pharmacogenomics affect individual patients' response to medications. Cancer therapies will continue to be the focus of many pharmacogenomic advances, as will antidepressants.
"When it comes to antidepressants, the value of pharmacogenomics cannot be overemphasized," Dr. Flockhart said. "Many patients now taking these medications need to try two or three different types before they find the right one, with each drug taking at least four weeks to take effect."
Internist Archives Quick Links
Internal Medicine Meeting 2015 Digital Presentations
Choose from over 170 recorded Scientific Program Sessions and Pre-Courses. Available in a variety of packages and formats so you can choose the combination that works best for you.