Abstract & Commentary
Synopsis: Patients who use both erythromycin and CYP3A inhibitors, particularly calcium channel blockers, had an increased risk of sudden death from cardiac causes.
Source: Ray WA, et al. N Engl J Med. 2004;351: 1089-1096.
In this paper, Ray and colleagues examined a large database of Tennessee Medicaid enrollees to see if use of the antibiotic erythromycin was associated with an increase in sudden death from a cardiac cause. Erythromycin is a commonly used macrolide antibiotic that has been marketed for many years. There have been sporadic cases of torsades de pointe in patients receiving both oral and intravenous erythromycin, but the relationship with intravenous use has been more solidly established. In vitro studies have shown that erythromycin may prolong the QT interval due to blockade of the human HERG potassium channel. Erythromycin is also extensively metabolized by cytochrome P-450 3A (CYP3A) isozymes. Other drugs with this combination of properties have been associated with QT prolongation and sudden death.
Ray et al examined sudden deaths among a cohort of Tennessee Medicaid enrollees. Subjects had been enrolled in Medicaid for 1 year and were between the ages of 15 and 84. They were not residing in a long-term facility and had no evidence of a concurrent life-threatening noncardiac illness. Data from medical encounter forms, files regarding prescriptions, outpatient visits, hospital admissions, and nursing home stays were used to identify the study cohort, to determine exposure to the drugs examined, to identify potential causes of sudden death from cardiac causes, and to classify the preexisting cardiovascular and noncardiovascular conditions among the members of the cohorts. Patients who received prescriptions for erythromycin were identified. Data from patients who received amoxicillin, an antibiotic with similar clinical indications, but no potential for QT prolongation, were compared to data from the erythromycin group. The database was also examined for drug usage of compounds that significantly inhibit metabolism of cytochrome CYP3A substrates. The drugs included all had been shown to double the area under the contraction curve for CYP3A substrates. The drugs selected were the azole antifungal drugs, diltiazem, verapamil, and troleandomycin.
The study outcome was sudden death from cardiac causes occurring in community settings. Death certificates and other records were examined to identify potential cases of sudden death. Only out-of-hospital sudden deaths were included.
The study cohort included 1.25 million person years of follow-up. The mean age among members of the cohort was 45 years, with only 25% of the subjects age 65 years or older. The subjects were 70% female, and 58% were white. In the group, there was a total of 1476 sudden deaths from cardiac causes, for a rate of 1.2 deaths per 1000 person years. There were 5305 person years of current use of erythromycin and 6846 person years of use of amoxicillin. Current and former uses of erythromycin, and current uses of amoxicillin, were very similar with regard to their demographic and clinical characteristics.
The rate of sudden death from cardiac causes was twice as high among current users of erythromycin, (incidence-rate ratio 2.1, P = 0.03), as among those who did not use the compound. Former users of erythromycin or current users of amoxicillin did not show an increase risk of sudden death. There were 10 deaths among those currently taking erythromycin. Three of these deaths were in patients who were also using a CYP3A inhibitor, either diltiazem or verapamil. The incidence-rate ratio for these 3 patients was 5.35 (P = 0.04) indicating a risk of sudden death more than 5 times as high as that among those who use neither CYP3A inhibitors nor any of the study antibiotics. There were 7 deaths among current users of erythromycin who were not on a CYP3A inhibitor. The incidence-rate ratio in this group was still increased at 1.79, but the confidence interval was now 0.85 to 3.76. No other drugs listed in the database were shown to be associated with sudden death.
Ray et al conclude that patients who use both erythromycin and CYP3A inhibitors, particularly calcium channel blockers, had an increased risk of sudden death from cardiac causes. Ray et al recommend that combinations of erythromycin and CYP3A inhibitors be avoided in clinical practice.
Comment by John P. DiMarco, MD, PhD
Drug-induced prolongation of the QT interval, with progression to polymorphic ventricular tachycardia, torsades de pointes was first described in patients receiving quinidine for treatment of atrial fibrillation. It is now recognized that the most common mechanism for drug induced torsades de pointes is blockade of the HERG potassium channel. This results in prolongation of repolarization, the development of repetitive early afterdepolarizations and polymorphic VT. Torsades de pointes is more common in women, and may be related to other factors which cause QT prolongation, eg, bradycardia, hypokalemia, severe hypomagnesemia, congestive heart failure, or ventricular hypertrophy. Often multiple factors are involved.
Many drugs produce minor changes in repolarizing currents, and can prolong the QT interval. The QT interval is also difficult to measure with precision, and the QT varies in the same individual over time and with changes in heart rate. Although correction factors for heart rate changes have been described, they are inexact. In most patients, minor effects on the HERG channel are easily compensated for and have no clinical sequelae. However, in patients who have either previously undetected forms of a long QT syndrome, and are exposed to higher than expected concentrations of a single drug, or are subject to effects from multiple drugs, torsades de pointe can be precipitated.
In this paper, Ray et al performed a careful examination of a large database to see if erythromycin, a drug with a known potential to block HERG channels, is actually associated with increased risk of sudden death. The database was a from a Medicaid population, so relatively few elderly individuals were involved. A Medicare population may have been more revealing. By examining this large Tennessee Medicaid database, Ray et al were able to demonstrate an increased risk ratio among all patients who used erythromycin, and in particular, among those in whom it would have been expected that erythromycin plasma concentrations would have been increased at least 2-fold. The absolute number of deaths is quite small. Without careful statistical analysis of this large database, the increase in risk would never have been detected clinically.
The clinical implications of these observations however, remain uncertain. Certainly, physicians should be aware if a drug has been reported to prolong the QT interval. If so, physicians should know the pharmacokinetic properties of the drug, and take care that the drug is not used either with other drugs which might raise its plasma concentration or in situations, for example, renal failure, where an excess concentration of the drug might be expected. In those cases, alternate therapy should certainly be prescribed. Keeping up-to-date on the potential for drug interactions, and a knowledge of the risk factors for developing torsades, should provide an adequate level of safety.
Dr. DiMarco, Professor of Medicine, Division of Cardiology, University of Virginia, Charlottesville, is on the Editorial Board of Clinical Cardiology Alert.