The following describes a patient on a stable regimen of warfarin who developed severe hypoprothrombinemia and bleeding 4 weeks after starting gemfibrozil. Despite a warning by the manufacturer, only one report of this interaction has been published in the literature. This interaction may be overlooked by clinicians, which may result in a serious bleeding risk for patients on warfarin. (CHEST 1998; 114:641-642)
Key words: gemfibrozil; hypoprothrombinemia; warfarin
Abbreviation: INR=international normalized ratio
Numerous drug interactions with warfarin have been well documented. One interaction that may be over looked is the combination of gemfibrozil and warfarin. This could be due to the lack of citations in the medical literature describing this interaction, either by anecdotes or clinical investigation. We report a case of significant hypoprothrombinemia and bleeding when gemfibrozil was added to a stable regimen of warfarin.
A 73-year-old man was admitted to the hospital after 3 days of fatigue, black stools, and a large bruise on the right arm. Five weeks prior to admission, the patient was prescribed gemfibrozil, 1,200 mg daily, for hyperlipidemia. No other changes in medications were made. The patient was taking warfarin, 5 mg alternating with 7.5 mg daily, for the past 3 years for deep vein thrombosis. Over this time period, 27 prothrombin time tests were performed in the anticoagulation clinic with a mean international normalized ratio (INR) of 2.6 (range, 1.8 to 3.8). The patient was on a stable diet and was never noted to drink alcohol. On admission, laboratory results revealed a prothrombin time of 66 s (INR 43; international sensitivity index=2.1; control prothrombin time=11 s), a hematocrit of 31% (which had been 41% 2 months previously), and stools that tested positive for blood. The patient was normotensive, but did have an orthostatic decrease in systolic BP. Warfarin and gemfibrozil were discontinued and vitamin K, 10 mg, was administered intravenously. The next day, the patient's INR was normal and the melena had resolved. On the third hospital day, warfarin was restarted at the previous dosage. The patient was discharged from the hospital on the next day with partial resolution of the large bruise and no melena. Three INRs measured over the next 8 weeks ranged from 1.9 to 3.0 without a change in dose and no evidence of bleeding.
The strong temporal relationship between starting gemfibrozil and subsequent hypoprothrombinemia implicates a gemfibrozil-warfarin drug interaction in this patient. The gemfibrozil manufacturer's package insert warns of a possible interaction between these two drugs, although there are few data in the medical literature corroborating this interaction. A search of MEDLINE through 1966 revealed only one report of this interaction. In that ease, the author reported a "higher" prothrombin time in a patient on stable doses of warfarin 2 weeks after starting gemfibrozil. Specific details were not offered in the brief report.
The lack of published eases could be due to underreporting. The gemfibrozil-warfarin interaction is listed in the gemfibrozil manufacturer's package insert and, therefore, clinicians may assume that this interaction is commonplace and not worth reporting. Another possibility is that clinicians may dismiss this interaction with the perception that it is unimportant due lack of descriptive reports.
Gemfibrozil is chemically related to clofibrate, which is well known to interact with warfarin. The mechanism appears to be unrelated to altered metabolism or protein binding of warfarin. One report suggests an interaction at the receptor site must account for clofibrate enhancing warfarin's hypoprothrombinemia. Gemfibrozil is highly protein bound and, unlike clofibrate, could displace warfarin from its binding sites resulting in an enhanced effect. It is unknown if this or possibly an effect on CYP2C9 activity (enzyme responsible for warfarin metabolism) are contributing factors for this interaction.
Few other drug interactions have been described with gemfibrozil. Gemfibrozil has been demonstrated to increase the incidence of myopathy with hydroxymethyl glutaryl CoA reductase inhibitors. It is not known if this is related to an altered disposition of these drugs. One author reported a case of hypoglycemia in a patient taking glyburide temporally related to gemfibrozil administration. Altered protein binding of warfarin was postulated in this case. The paucity of reported drug interactions with gemfibrozil suggests that this agent does not significantly interfere with the disposition or action of most drugs, including warfarin. One author classified the interaction with warfarin as "doubtful" after a systematic critical review of the literature for potential warfarin drug interactions. The lack of data may lead health-care professionals to underestimate the drug-interacting potential of gemfibrozil and warfarin, which occurred in this case.
Gemfibrozil has the potential to markedly enhance the effect of warfarin. Increased monitoring is prudent when gemfibrozil is added to a stable regimen of warfarin.
 Hansten PD, Horn JR. Drug interactions analysis and management. Vancouver, Wash: Applied Therapeutics, Inc, 1997
 Ahmad S. Gemfibrozil interaction with warfarin sodium (coumadin) [letter]. Chest 1990; 98:1041-42
 Bjornsson TD, Meffin PJ, Swezey S, et al. Clofibrate displaces warfarin from plasma proteins in man: an example of a pure displacement interaction. J Pharmacol Exp Ther 1979; 210: 316-21
 Pond SM, Graham GG, Wade DN, et al. The effect of allopurinol and clofibrate on the elimination of coumarin anticoagulants in man. Aust N Z J Med 1975; 5:324-28
 McEvoy GK, Litvak K, Welsh OH. Drug information 97. Bethesda, MD: American Society of Health-System Pharmacists, Inc, 1997; 1344-49
 Ahmand S. Gemfibrozil: interaction with glyburide. South Med J 1991; 84:102
 Wells PS, Holbrook AM, Crowther NR, et al. Interactions of warfarin with drugs and food. Ann Intern Med 1994; 121: 676-83
(*) From the Veterans Affairs Medical Center, Prescott, Ariz. Manuscript received November 20, 1997; revision accepted February 5, 1998.
Correspondence to: Joseph P. Rindone, PharmD, VA Medical Center (119), Prescoth, AZ 86313
COPYRIGHT 1998 American College of Chest Physicians
COPYRIGHT 2000 Gale Group