Commonly Overlooked Drug Interactions

Potential interactions with a patient's current medications must be considered any time a new drug is prescribed. Even common interactions, however, often are not anticipated by physicians.

By Douglas S. Paauw, MD

Drug interactions are extremely common, but because they are often not suspected, their effects just as often go unnoticed. The challenge for physicians is to recognize those interactions that are clinically important from among the thousands of less important ones.

Public awareness and an interest in medical malpractice have been heightened by a recent report by the Institute of Medicine of the National Academy of Sciences indicating that an estimated 44,000 to 98,000 deaths caused by medical errors occur annually in hospitals in the United States. In fact, drug interaction errors and severe adverse effects of medication are the most common causes of iatrogenic illness. Drug-related morbidity and mortality have recently been estimated to cost more than $130 billion annually in the United States.

An adverse drug event can lead to a prolonged hospital stay and double the risk of death during hospitalization. This article will focus on the important drug interactions that are commonly overlooked.

DRUG ABSORPTION PROBLEMS

Gastric pH can drastically alter the absorptive quality of several drugs. The antifungal agents itraconazole and ketoconazole, for example, are poorly absorbed at a high gastric pH. Because such oral medications as proton pump inhibitors, H2 blockers, and antacids can increase gastric pH and impair the ability of either itraconazole or ketoconazole to be absorbed, they should not be administered when patients are undergoing antifungal therapy. In addition, the antiretroviral drug didanosine has its own antacid buffer and should not be taken with itraconazole or ketoconazole.

Fluconazole, an antifungal drug that is not affected by gastric pH, may be an appropriate alternative for patients who must continue to take pH-altering drugs. Commonly, a patient who undergoes pulse itraconazole therapy for treatment of fungal nail disease may not receive proper benefit from the drug, because he or she is also taking a proton pump inhibitor, which inhibits the absorption of itraconazole. A poor response can be particularly upsetting for those patients who must pay for the high cost of itraconazole out of their own pockets. The antibiotic cefpodoxime is also affected by gastric pH, which, if too acidic, can decrease absorption of the drug by as much as 50%.

Thyroid hormone absorption can be affected by several drugs and supplements. The iron in iron sulfate compounds, for example, can bind to thyroid hormone and thereby reduce absorption of the hormone. This interaction is variable and can lead to marked thyroxine malabsorption in some patients. To avoid such a response, the dose of each drug should be separated by at least eight hours. Aluminum hydroxide, a common antacid, can also cause a decrease in thyroid hormone absorption, according to Mersebach and colleagues, who confirmed the interaction in in vivo studies (Pharmacology and Toxicology, vol. 84, p. 107, 1999). An interaction of potentially much greater clinical significance is the recently reported effect of calcium carbonate on thyroxine absorption. Singh and colleagues have found that the impact on serum levels of thyroid-stimulating hormone is modest, but because supplemental calcium therapy is so common, the potential for a clinically major effect is great (JAMA, vol. 283, p. 2822, 2000).

These resins, such as cholestyramine and colestipol, can bind to and alter the absorption of a wide variety of drugs. The problem is especially severe when warfarin is taken: when the two drugs are administered together, as much as 95% of a warfarin dose will bind to cholestyramine. Cholesterol resins also affect the enterohepatic circulation of warfarin, a response that can decrease the half-life of the drug. These resins can affect the absorption of digoxin, thyroid hormone, thiazide, beta-blockers, and fibric acid drugs such as gemfibrozil. Because many other drugs are likely to be bound by these resins, the best solution is to ensure that the doses of resin medications and those of all prescription medications and supplements are separated by at least three hours.

Quinolone antibiotics are poorly absorbed when given with cationic compounds, particularly those containing aluminum, as in some antacids, or magnesium, which is an ingredient of some antacids and laxatives. Some antacids, such as Mylanta and Maalox, contain both aluminum hydroxide and magnesium hydroxide. The cations iron and calcium can also affect quinolone absorption, although to a lesser extent than do magnesium and aluminum. The drugs ciprofloxacin and norfloxacin appear to be more affected by cation binding than other agents in the quinolone class. The best solution is for the patient to take the quinolone antibiotic two hours before or six hours after the cationic compound.

WARFARIN INTERACTIONS

Drug interactions involving warfarin are among the most serious, because they are associated with a high risk of fatal bleeding. A few drugs decrease the effectiveness of warfarin and increase thrombotic risk. The agents that increase metabolism of warfarin are generally antiseizure drugs, such as phenytoin, carbamazepine, and phenobarbital. Rifampin also increases warfarin metabolism, and in doing so it can also lower the prothrombin time. Effective monitoring of prothrombin time can allow appropriate dose adjustments of warfarin to compensate for the increased metabolism.

More common than the agents that increase warfarin metabolism are drugs that decrease its metabolism and increase prothrombin time (see table, below). The most important example is trimethoprim-sulfamethoxazole, a widely used antibacterial agent that is probably the most common cause of prothrombin time elevation in patients undergoing warfarin therapy, who should avoid the drug. The effect on prothrombin time can appear within 2 to 3 days, although it usually peaks at 7 to 10 days. If treatment options are limited only to drugs that alter warfarin metabolism, then close monitoring of prothrombin time every two or three days during the regimen is necessary.

Warfarin can interact with drugs in other ways that increase the risk of bleeding. Nonsteroidal antiinflammatory drugs (NSAIDs) reversibly inhibit platelet function, thereby increasing bleeding time. Warfarin increases the risk of bleeding by increasing prothrombin time. The combination of the two mechanisms increases overall bleeding risk. The NSAIDs also cause gastric erosions, which in turn become sites for bleeding. Bleeding from gastric erosions caused by NSAID use is much more severe when a patient is also taking warfarin.

Celecoxib and rofecoxib, two of the new cyclooxygenase-2 (COX-2) inhibitor agents that have been recently introduced, do not inhibit platelet function or increase the risk of gastric erosions. In patients taking warfarin, rofecoxib increases prothrombin time by 8% to 11%, whereas celecoxib produces no change in prothrombin time. Bleeding has been noted in elderly patients undergoing warfarin therapy who were also given both COX-2 inhibitors. Although both agents appear to be safer antiinflammatory options than NSAIDs for patients taking warfarin, close monitoring is advised.

Acetaminophen is usually considered a safe pain reliever for patients undergoing anticoagulant therapy. A recent study by Hylek and colleagues showed that patients taking warfarin who use more than 1.5 grams (three extra-strength tablets) of acetaminophen daily are 10 times more likely to have an international normalization ratio (INR) greater than 6 (JAMA, vol. 279, p. 657, 1998).

NATURAL SUBSTANCES AND PRODUCTS CONTAINING HERBAL DERIVATIVES

Many types of interactions between pharmaceutical agents and natural substances are possible. Among the potentially serious interactions are those that occur between warfarin and the many over-the-counter medicinal products that contain herbal derivatives, which can potentiate the effect of warfarin. The absolute risk for bleeding that is generated by such interactions is not known, but it is real.

Garlic, for example, has been associated with decreased platelet aggregation, and several case reports have investigated severe bleeding in patients taking high doses of the herb. Gingko biloba contains gingkolide B, a potent inhibitor of platelet activating factor. Two case reports have described spontaneous subdural hematoma occurring in patients taking gingko. Ginger may be a thromboxane synthetase inhibitor and may decrease platelet aggregation. Dong quai contains several natural warfarin derivatives as well as ferulic acid, which produces antithrombotic activity. Patients undergoing warfarin therapy may have a sudden increase in their INR after taking dong quai. Feverfew is also associated with a high risk of bleeding in patients taking warfarin.

Other herbal medicines, such as ginseng, can decrease the effect of warfarin. In one case report, a man who had undergone effective anticoagulation therapy with warfarin after receiving a mechanical valve began taking ginseng after which his INR dropped from 3.0 to 1.5. The INR returned to 3.0, however when the patient stopped taking ginseng.

St. John's wort is a popular herbal treatment for depression and can interact with a number of drugs. In women taking oral contraceptives and St. John's wort, menstrual irregularities with breakthrough bleeding can occur. St. John's wort can lower serum levels of the protease inhibitor indinavir, thereby producing subtherapeutic concentrations of the drug and subsequent treatment failure. Serum levels of cyclosporin can also be decreased by St. John's wort. Two case reports have described acute cardiac transplant rejection occurring in patients who have taken St. John's wort while receiving cyclosporin. Persons undergoing digoxin therapy have also been reported to have much lower levels of the drug in serum when they take St. John's wort. The substance may also increase the metabolism of warfarin, resulting in a lower prothrombin time.

Recently, attention has been directed at the many potential interactions that can occur between various drugs and grapefruit juice (see table, below). Grapefruit juice inhibits the cytochrome P-450 CYP3A4 subunit in the gastrointestinal tract. This enzyme inhibition increases the rate at which several drugs are absorbed. Several HMG-CoA reductase inhibitors are affected, especially lovastatin and simvastatin. In one study by Kantola and colleagues, the area under the curve (AUC) for lovastatin was increased 15 times when patients drank 200 mL of double-strength grapefruit juice before taking 80 mg of the drug (Clinical Pharmacology and Therapeutics, vol. 63, p. 397, 1998).

In a follow-up study of patients drinking single-strength grapefruit juice and taking 40 mg of lovastatin, the AUC increased by only 40%. When simvastatin was studied in a similar way, with double-strength grapefruit juice and 60 mg of the drug, the AUC was increased by 1500%. Atorvastatin can be affected by grapefruit juice but to a lesser extent than are lovastatin and simvastatin. Pravastatin and fluvastatin do not rely on the CYP3A4 subunit for metabolism and are not affected by grapefruit juice.

INTERACTIONS THAT CAUSE CARDIAC EMERGENCIES

Several drug interactions can prolong the QT interval and increase the risk of torsade de pointes arrhythmia. The most serious of these interactions have been eliminated by the removal of several of the offending drugs from the market, including cisapride, terfenadine, and astemizole. When coadministered with drugs that inhibit the CYP3A4 subunit of the P-450 system, such as erythromycin, ketoconazole, and itraconazole the discontinued agents all had the potential to inhibit metabolism of the coadministered drug, thereby increasing their blood concentration. There are still several drugs on the market that can prolong the QT interval and potentially incite arrhythmia, especially when combined with other drugs that prolong the QT interval. These drugs include the quinolones sparfloxacin, moxifloxacin, and gatifloxacin. In September 2000, a boxed warning appeared in the product disclosure sheet that accompanied the antipsychotic drug mesoridazine. All of the above-mentioned drugs should not be given to patients taking class 1A (quinidine or procainamide) or class 3 (amiodarone or sotalol) antiarrhythmic agents.

Several drug interactions can cause hyperkalemia, a potential cardiac emergency. The most common interaction that produces hyperkalemia occurs between angiotensin-converting enzyme (ACE) inhibitors and potassium-sparing diuretic agents, such as spironolactone, amiloride, or triamterene. Such a combination is very popular now because these drugs lower the risk of death in patients who have congestive heart failure.

Administration of NSAIDs can decrease the antihypertensive effect of ACE inhibitors and lead to sodium and water retention and to hyperkalemia. Patients who ingest a salt substitute, such as potassium chloride, while taking ACE inhibitors or spironolactone also can become hyperkalemic. Trimethoprim can inhibit potassium excretion and, when combined with other drugs that inhibit potassium excretion, can produce severe hyperkalemia. Other drugs involved in interactions that produce hyperkalemia include angiotensin-receptor blockers and cyclosporine.