Sildenafil (Viagra) Prolongs Cardiac Repolarization by Blocking the Rapid Component of the Delayed Rectifier Potassium Current
To the Editor:
Geelen and coworkers1 reported 6% to 15% increases in action potential duration in guinea pig hearts exposed to supratherapeutic concentrations (30 μmol/L) of sildenafil. They observed that high concentrations (IC50, 100 μmol/L) inhibited the rapid component of the delayed rectifier K+ current in Chinese hamster ovary and HEK293 cells transfected with human ether-a-go-go-related gene (HERG). Their interpretation that this could lead to adverse cardiovascular effects in humans is unfounded.
Sildenafil inhibits the human PDE5 enzyme in vitro by 50% at a concentration of 3.5 nmol/L,2 which is nearly 30 000 times below the level required to block HERG by 50%. Sildenafil is 96% protein-bound, with a peak free plasma concentration after the maximum recommended therapeutic dose (100 mg) of 40 nmol/L,3 which is 2500 times lower than the concentration required to inhibit HERG by 50%.
Because sildenafil is metabolized primarily via CYP3A4, the authors imply that during coadministration with CYP3A4 inhibitors (eg, erythromycin), sildenafil plasma levels could rise by 20-fold. However, sildenafil has a high oral bioavailability (≈40%), which limits the extent that coadministered drugs can raise peak sildenafil plasma levels after a single oral dose. A pharmacokinetic interaction study has shown that plasma levels of sildenafil increased only 2.6-fold in the presence of erythromycin.4 Thus, when given together with a potent CYP3A4 inhibitor, the maximum free plasma concentration of sildenafil (100 nmol/L) is still 1000 times lower than the concentration of sildenafil required to inhibit HERG by 50%.
Importantly, it has also been demonstrated that even at twice the therapeutic dose (200 mg), sildenafil caused no change in ECG parameters.4 In pharmacology and toxicology studies in dogs, sildenafil free plasma concentrations of ≈100 nmol/L and 1.8 μmol/L were achieved, respectively, and no increases in QT interval were observed.
There are >11 500 person-years of observation of patients, many with a history of heart disease, taking sildenafil in clinical trials. The incidence of serious cardiovascular adverse events is the same for patients randomized to sildenafil or placebo (Pfizer, unpublished data, 2000). In a Prescription Event Monitoring study in the United Kingdom of >5000 men with erectile dysfunction, the incidence of serious cardiovascular adverse events after 5 months was consistent with general population rates, with none reported in the first month of treatment (S. Shakir, MD, unpublished data, 2000). Furthermore, there have been no reports of torsades de pointes or QT prolongation from the clinical trial database or from postmarketing experience based on >40 million prescriptions to >13 million patients worldwide.
- Copyright © 2001 by American Heart Association
We thank Drs Butrous and Siegel for their comments. Indeed, data from clinical studies and postmarketing experiences indicate that sildenafil is a safe drug in a very large proportion of patients. However, cases of severe cardiovascular events have been observed in some patients taking sildenafil, with or without the concomitant use of nitrates.R1 R2 On the basis of a recent study, only 12% of all sildenafil-associated deaths occurred in men who were using concomitant nitrates.R3 Hemodynamic and ischemic mechanisms have been proposed to explain these severe cardiovascular events with sildenafil; in addition, we propose considering a proarrhythmic hypothesis.
Butrous and Siegel argue that the sildenafil concentrations required to block either IKr or HERG are much higher than those observed in patients. There are reasons to believe that concentrations of sildenafil may be high enough under certain circumstances in some patients to prolong cardiac repolarization. Sildenafil is mainly metabolized by CYP3As, with an average bioavailability of 40%. Thus, on average, one can predict that peak plasma concentrations will rise 2.5-fold under conditions of decreased CYP3As activity. Average data from the erythromycin-sildenafil study are in perfect agreement with this statement. However, CYP3As activity is known to exhibit wide intersubject variability. Consequently, in some patients, the magnitude of the drug interaction can be much larger than expected.R4 This statement is supported by examining individual data in drug interaction studies with sildenafil; moreover, a 10-fold increase in plasma concentrations of sildenafil was observed during combined administration with indinavir, saquinavir, or ritonavir.R5 One should also consider that plasma concentrations of sildenafil are increased 70% in elderly patients and that genetic polymorphisms have been reported for CYP3A isozymes, again pointing toward interindividual susceptibility. Finally, several CYP3As substrates such as calcium-channel blockers, hydroxymethylglutaryl coenzyme A reductase inhibitors, benzodiazepines, and antidepressants are often coadministered with sildenafil and can potentially inhibit its metabolism.
Butrous and Siegel also comment on the IC50 for block of HERG and the free fraction of sildenafil. The IC50 measured in the models used in our study does not correlate well with QT prolongation in humans if only the free fraction of drugs observed in human plasma is considered. This is a clear limitation of the models. However, drugs shown to block IKr or HERG in these models have been associated with QT prolongation in humans. One should not consider only the free fraction but rather the tissue concentration of the drug, which is a parameter that can be affected by other determinants, such as P-glycoprotein. Competitive inhibition of CYP3As activity during combined administration of drugs is often associated with concomitant inhibition of P-glycoprotein. Under these conditions, plasma concentrations and tissue concentrations are increased much more than predicted from the increase in the free fraction of the drugs. Along those lines, one should question why the doses of sildenafil required are so high (50 to 100 mg) and are associated with plasma concentrations in the micromolar range if efficacy is expected at low nanomolar concentrations (IC50 for block of PDE5 is 3.5 nmol/L).