A potential interaction between warfarin and atovaquone

To report a case of increased international normalized ratio (INR) in a patient established on warfarin therapy who was then initiated on atovaquone therapy.

A 53-year-old African American male with HIV was prescribed warfarin 5 mg/day for 12 months after diagnosis of idiopathic deep vein thrombosis and bilateral pulmonary emboli (target INR 2.5 [range 2.0-3.0]). The patient required Pneumocystis jiroveci pneumonia prophylaxis and was prescribed atovaquone instead of trimethoprim/sulfamethoxazole therapy because of the latter drug's known interaction with warfarin. The patient's INR rose by greater than 50% (from 2.3 to 3.5) after 7 days of concomitant warfarin and atovaquone. In response, the patient's total weekly warfarin dose was decreased by 5%. Eight days later, the patient's INR was still supratherapeutic at 3.1. Approximately 4 weeks later, his INR was 4.2. One dose of warfarin was withheld and then the total weekly warfarin dosage was decreased by another 10%. Eight days later, the patient discontinued atovaquone therapy but continued on warfarin as prescribed. One day after atovaquone discontinuation, his INR decreased to 1.7. Due to this subtherapeutic INR level, 8 days later the total weekly warfarin dose was increased by 5%. Although a follow-up appointment was scheduled, no further INR values were obtained because the patient's 12-month course of anticoagulation therapy was completed and warfarin was discontinued. The patient did not report any adverse effects or signs or symptoms of hemorrhage while his INR values were supratherapeutic.

Warfarin's potential for interactions with other highly protein-bound drugs, such as atovaquone, can result in displacement from protein binding sites and increased serum concentrations of warfarin. Based on a search of MEDLINE/PubMed, International Pharmaceutical Abstracts, and the Food and Drug Administration MedWatch Adverse Event Reporting Program (all through July 31, 2010), no cases were found of an interaction between atovaquone and warfarin. The Horn Drug Interaction Probability Scale calculated this to be a probable interaction between warfarin and atovaquone.

Although current medication references do not report an interaction between atovaquone and warfarin, knowledge of their pharmacodynamic properties can enable practitioners to anticipate the consequences of a possible transient increase in warfarin serum concentration, such as that seen in our patient, when given concomitantly.