Digoxin toxicity and ritonavir

Abstract

P-glycoprotein is an MDR1 gene product that functions as an ATP-dependent drug efflux pump. P-glycoprotein is expressed in the apical membrane of epithelial and endothelial cells of a number of tissues, including the liver, small intestines, kidney and the blood–brain barrier, where it results in reduced absorption, enhanced elimination and the prevention of drug entry into the central nervous system, respectively. Digoxin is a substrate for p-glycoprotein, and given its low therapeutic index its concurrent use with drugs that have a propensity to increase digoxin exposure is of significant concern. Clinically significant drug interactions with digoxin have been associated with itraconazole, quinidine, clarithromycin and verapamil that inhibit intestinal or renal digoxin excretion by inhibiting p-glycoprotein-mediated transport [1]. All protease inhibitors are substrates for p-glycoprotein [2]. The predominant effect of ritonavir on p-glycoprotein expression has been somewhat controversial [3]. A recent pharmacokinetic study in normal volunteers [4] showed that ritonavir decreased the renal clearance of digoxin. It is plausible that the mechanism of this is through the inhibition of p-glycoprotein-mediated transport in the renal proximal tubules. We describe a case of digoxin toxicity, which we believe was precipitated by the introduction of ritonavir. A 61-year-old woman presented to the emergency department with increasing nausea and vomiting 3 days after the addition of ritonavir 200 mg twice a day to lamivudine, indinavir and stavudine. Before this she had been tolerating indinavir 800 mg three times a day, lamivudine 150 mg twice a day and stavudine 40 mg twice a day for the past 3 years. She had been diagnosed as being HIV infected in 1995 after her husband died from AIDS-related complications. The past history was significant for rheumatic heart disease, with valve replacement and permanent pacemaker insertion in 1992. She had been taking a stable dose of digoxin (0.250 mg a day) since developing atrial fibrillation after open-heart surgery 8 years earlier. Other concurrent medications at the time of admission included digoxin, coumadin 5 mg alternating with 10 mg a day and aerosolized pentamidine once a month. The basic laboratory work-up on admission was normal apart from an unconjugated bilirubin level of 129 μmol/l, which was up from 30 μmol/l, probably explained by the increased exposure to indinavir after the introduction of ritonavir. The electrocardiogram showed a paced rhythm. The digoxin level on admission, measured by the CEDIA immunoassay (Roche Diagnostic Systems Inc., Branchburg, NJ, USA), approximately 5 h after her last dose was 7.2 nmol/l (N 1–2.6). Digoxin levels were repeated in duplicate using the Immono1System (Bayer Diagnostics, New York, USA) to rule out any interference with bilirubin. Levels 11, 15 and 27 h after the last dose were 5.5, 4.5 and 2.7 nmol/l. By the second day in hospital she was clinically much improved. Digoxin was permanently discontinued without any sequelae. The original antiretroviral drugs (lamivudine, indinavir and stavudine) were restarted without significant side-effects. The inhibition of p-glycoprotein-mediated transport by ritonavir may be one mechanism to explain digoxin toxicity in this woman previously on a stable dose of digoxin for several years. A more recent study suggested that ritonavir may be unique among protease inhibitors in its inhibition of p-glycoprotein, and that this may occur in particular at the level of the renal tubules. Although p-glycoprotein is often co-expressed with cytochrome P450 3A4, this case appears to represent drug interaction mediated through p-glycoprotein because digoxin is a substrate for p-glycoprotein but not cytochrome P450 3A4. Of interest is the fact that the patient described had been on a stable dose of indinavir for 3 years without any evidence of toxicity. This is in support of in-vitro data, which suggest that indinavir is not a significant inhibitor of p-glycoprotein-mediated transport. As the HIV population ages the potential for these types of drug interactions will increase with more exposure to drugs such as digoxin. Other factors such as genetic polymorphisms in MDR1 may be important in predicting the likelihood of a clinically significant drug interaction as outlined in a recent study of pharmacokinetic interactions between clarithromycin and digoxin [3,5]. Our case and pharmacokinetic data support the fact that patients who are concurrently treated with ritonavir and digoxin may be at risk of digoxin toxicity and should be closely monitored.