Erythromycin, QT-Interval Prolongation, and the Risk of Sudden Death

Abstract & Commentary

Source: Ray WA, et al. Oral erythromycin and the risk of sudden death from cardiac causes. N Engl J Med 2004;351: 1089-1096.

Drug-drug interactions are a potentially hazardous pitfall of practicing medicine; physicians are confronted with new drugs and must become knowledgeable of their adverse effects as well as the potential for interactions with other medicines. One of the more important adverse effects is the rather common potential for certain and varied agents to prolong ventricular repolarization, represented by the QT interval on the electrocardiogram (ECG). Prolongation of the QT interval can predispose the development of torsades de pointes—a type of polymorphic ventricular tachycardia that (if self-terminated) may cause symptoms such as near- or true syncope, or may lead to sudden death.

Erythromycin is a common antibiotic known for its propensity to prolong the QT interval.

In this pharmacoepidemiologic study, the authors used the Tennessee Medicaid database to study the association between oral erythromycin and the risk of sudden death from cardiac causes to find the clinical relevance of the known association between erythromycin and QT-interval prolongation. Eligible patients between 1998 and 1994 included those who were 15-84 years of age, not residing in a long-term facility for medical reasons, and without evidence of a life-threatening cardiac illness. The authors examined whether oral erythromycin use was associated with an increased risk of sudden death when prescribed with and without other agents known to potently inhibit the hepatic cytochrome P450 enzyme responsible for its metabolism, CYP3A. These drugs, determined a priori based upon a known effect to at least double the area under the time-plasma concentration curve of a recognized CYP3A substrate, included the azole antifungals (ketoconazole, itraconazole, and fluconazole), diltiazem, verapamil, and troleandomycin. Protease inhibitors were excluded because patients with HIV were excluded. Patients who were former users of erythromycin, as well as those who were current users of amoxicillin (no effect on CYP3A, but with a similar antimicrobial spectrum to erythromycin) also were analyzed to uncover confounding effects. Attempts were made to define the main outcome measure, sudden death, as stringently as possible. Insufficient medical documentation for a variety of reasons, or death from another cause, resulted in only 1476 cases deemed to be sudden death from a cardiac cause from the 4404 potential qualifying deaths reviewed.

The study ultimately found 10 deaths within 5305 person-years of current erythromycin users, eight deaths among 6846 person-years of current amoxicillin users, and 100 deaths among 111,779 person-years of former erythromycin users. Those numbers yielded an incidence-rate ratio of 2.01 for the current erythromycin group (95% CI, 1.08-3.75). A group with no use of study antibiotics (1358 deaths in more than 1.1 million person-years) served as the reference group. Three of the eight deaths in the current erythromycin group were patients with concurrent use of a CYP3A inhibitor (one on diltiazem, two on verapamil), yielding an incidence-rate ratio of 5.35 (95% CI, 1.72-16.64)—this was the only group with an increased likelihood of sudden cardiac death. Subsequent analyses found no evidence of concurrent use of other agents known to prolong the QT interval (e.g., cisapride, terfenadine, astemizole, type Ia antidysrhythmics). The authors conclude that the rate of sudden death from cardiac causes was twice as high among current oral erythromycin users and more than five times as high in patients using oral erythromycin and a study drug known to increase its concentration.

Commentary by Richard Harrigan, MD

This study made for interesting reading, as did the accompanying editorial.1 The editorial did not laud or assail the study. In fact, it made only passing mention of it. Instead, the editorial stresses the key take-home points about prolongation of the QT interval, and includes a good table of drugs that prolong the QT interval as well as those that inhibit cytochrome P450 metabolism.

Our attention should be focused on five classes of drugs that do the latter: antifungal agents, antiretroviral drugs, calcium channel antagonists, selective serotonin-reuptake inhibitors, and certain antibiotics (some macrolides and fluoroquinolones). Those drugs are potent cytochrome P450 inhibitors, and when added to a drug regimen that contains an agent known to prolong the QT interval, the result may be a significant pharmacokinetic effect on the QT-prolonging drug. Therefore, we must be cognizant not only of drugs that prolong the QT (avoiding adding them to others that prolong the QT, or using them in situations—congenital long QT, hypokalemia—predisposing to prolonged ventricular repolarization), but of agents that prolong the metabolism of these drugs.

The editorial offers two noteworthy web-based resources on this topic: and

Regarding this study, the point is well taken, even if the numbers are not compelling. Despite a best-case data scenario—a huge patient cohort—the incidence of the main outcome measure is still quite low, and the confidence intervals reflect the uncertainty in the strength of the association. Moreover, there remains a leap in logic to accept that these cases of cardiac sudden death were indeed mediated by torsades de pointes, allegedly secondary to the study agents. Yet, when added to the case report data that have provided the basis for concern thus far, it seems prudent to remain wary of this phenomenon and careful in our prescription practices.

Dr. Harrigan, Associate Professor of Emergency Medicine, Temple University Hospital and School of Medicine, Philadelphia, PA, is Editor of Emergency Medicine Alert.


1. Liu BA, et al. Drugs and the QT interval-caveat doctor. N Engl J Med 2004;351:1053-1056.