News Release

Intake Of Acetaminophen Increases Risk Of Excessive Anticoagulation In Patients Taking Warfarin

Peer-Reviewed Publication

Massachusetts General Hospital

People who take the anticoagulant drug warfarin are at increased risk of excessive anticoagulation if they also take large amounts of the pain reliever acetaminophen, according to a study from the Massachusetts General Hospital (MGH). Excessive anticoagulation is known to raise the risk of major hemorrhage. The report appears in the March 4 Journal of the American Medical Association.

"While acetaminophen generally is a very safe drug for pain and fever, continuous use -- for seven days or more -- should prompt closer monitoring of anticoagulation levels to reduce the risk of hemorrhage," says Elaine Hylek, MD, MPH, of the MGH Medical Practices Unit, the paper's first author. Acetaminophen is often the pain reliever of choice for people with potential bleeding problems because aspirin and non-steroidal anti-inflammatory drugs like ibuprofen can cause gastric bleeding.

Warfarin is a blood-thinning drug taken by millions of people with a heart condition called atrial fibrillation to prevent the development of blood clots that could cause strokes. It also is taken to prevent clots around replacement heart valves and to dissolve blood clots in the legs or lungs. Because the drug's effects can be altered by many factors -- including illness, other medications, and eating foods containing vitamin K -- warfarin treatment is a delicate balancing act between two serious risks. Too little anticoagulant effect and the patient could develop a clot that might travel to the brain and cause a stroke, too much anticoagulation and the patient could suffer a major hemorrhage.

Patients on long-term treatment usually have a monthly blood test called the INR (international normalized ratio) to measure their levels of anticoagulation. The MGH team surveyed 289 patients from the hospital's anticoagulation unit within 24 hours of a regular outpatient INR test. Among these patients, 93 had been found to have an INR greater than 6 -- a level associated with increased hemorrhage risk -- and 196 were controls with INRs in their target range of 2 to 3. Patients were contacted by phone and asked a variety of questions regarding their health, diet and medications taken in the weeks before their blood test.

The researchers found several factors associated with an increased risk of an INR greater than 6, the most important of which was ingestion of acetaminophen. Patients who took the equivalent of at least four regular strength (325 mg) tablets per day for 7 days or longer had a ten-fold increase in the risk of INR of 6 or more. The risk dropped with lower dosages, and there was no increased risk seen in patient taking 6 or fewer tablets per week.

A high-risk INR level also was associated with recently starting to take other medications known to increase anticoagulation, including a variety of common antibiotics; with having an advanced malignancy or a diarrheal illness, and with a reduction in overall food intake. Eating foods rich in vitamin K -- such as broccoli, cabbage, green peas, lettuce, liver, spinach and greens -- was associated with a decreased risk of having an INR over 6.

Although some previous research suggested that acetaminophen might increase warfarin's blood-thinning effect, Hylek notes, this is the first large demonstration of its impact in usual clinical practice.

"We don't want patients on anticoagulation therapy to avoid treating pain or fever, but if they are continuously taking a maximum over-the-counter dose of acetaminophen, they should alert their physician or anticoagulation clinic to determine whether their warfarin dose should be adjusted."

Hylek's coauthors on the study were Daniel Singer, MD; Heather Heiman; Steven Skates, PhD; and Mary Sheehan, RN, all of the MGH Department of Medicine. The study was supported in part by the National Institute on Aging and the Eliot Shoolman Fund of the MGH.

###



Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.