Azithromycin – The Heart of the Matter

By Stan Deresinski, MD, FACP, FIDSA, Clinical Professor of Medicine, Stanford University, Hospital Epidemiologist, Sequoia Hospital, Redwood City, CA, is Editor for Infectious Disease Alert.

Source: Ray WA, et al. Azithromycin and the risk of cardiovascular death. N Engl J Med 2012;366:1881-90.

Ray and colleagues examined the risk of cardiovascular death among patients 30 to 74 years of age enrolled in the Tennessee Medicaid program that had been prescribed azithromycin between 1992 and 2006. Individuals with known life-threatening cardiovascular illness, drug abuse, hospitalization in the previous 30 days, or residence in a nursing home in the previous year were excluded. The target patients, who received 347,795 5-day azithromycin prescriptions, were compared to propensity-score- matched individuals who took no antibiotics (1,391,180 control periods), and to patients prescribed amoxicillin (1,348,672 prescriptions), ciprofloxacin (264,626), or levofloxacin (193,906).

Compared to "no antibiotic" controls, the hazard ratio for cardiovascular death during 5 days of prescribed azithromycin therapy was 2.88 (95% CI, 1.79 to 4.63; P<0.001) and that of death from any cause was 1.85 (95% CI, 1.25 to 2.75; P=0.002). Azithromycin therapy was also associated with an increased hazard ratio for both cardiovascular and total deaths when compared to amoxicillin therapy, which itself had no increased hazard relative to "no antibiotic" controls. Relative to amoxicillin, azithromycin was associated with an estimated 47 added cardiovascular deaths per 1 million prescriptions. The degree of hazard was, however, strongly associated with the presence of preexisting cardiac risk factors. Those in the lowest deciles of risk (1-5) and those in deciles 6-9 had 9 and 45 excess deaths per 1 million prescriptions respectively, while those in risk decile 10 had 245 excess deaths per million prescriptions. The high-risk decile 10 enrollees accounted for 59% of cardiovascular deaths while receiving azithromycin. Azithromycin recipients also had a significantly greater hazard ratio of sudden death (2.71; 95% CI, 1.58 to 4.54) when compared to amoxicillin recipients.

Azithromycin was also associated with increased hazard of death relative to ciprofloxacin but did not significantly differ from levofloxacin. The increased risk associated with azithromycin therapy did not extend beyond the 5 days for which it was prescribed.


No direct evidence demonstrating the reason for the excess cardiac deaths is provided in this study. The finding, however, that there was an increased risk of sudden cardiac deaths (in addition to other types of cardiovascular deaths) leads to a strong hypothesis that arrhythmias played a significant role. This investigation was in fact prompted, at least in part, by the knowledge that the macrolide antibiotics prolong cardiac repolarization, as measured by the QT interval, and their use has been associated with the development of serious ventricular arrhythmias, especially torsades de pointe. This concern has proven to be well founded with regard to erythromycin and clarithromycin, which also have added potential risk when they are coadministered with drugs that inhibit CYP450, the enzyme responsible for their metabolic clearance (azithromycin is not affected). A previous examination of erythromycin prescriptions in Tennessee Medicaid enrollees found that they were associated with a doubling of the risk of sudden cardiac death and that this risk increased to five-fold in those who concomitantly took medications that inhibited CYP3A.1 On the other hand, a precise understanding of the pro-arrhythmogenic risk of the azalide, azithromycin, has been lacking. QT prolongation, torsades de pointe, and polymorphic ventricular tachycardias in azithromycin recipients have been reported in only a small number of patients. In addition, the authors indicate that "at least" 20 cases of associated torsades have been reported to the FDA. These numbers must be put into the context of the fact that in 2010 azithromycin was the seventh most frequently prescribed medication in the U.S. with 52.6 million prescriptions.2

In the U.S., azithromycin has received FDA approval for use in the treatment of acute bacterial exacerbations of chronic pulmonary disease, acute bacterial sinusitis, community acquired pneumonia, pharyngitis/tonsillitis, uncomplicated skin and skin structure infections, urethritis and cervicitis, and genital ulcer disease. In March of 2012, the FDA issued the following statement:

"Prolongation of the QT interval: Cases of torsades de pointes have been spontaneously reported during postmarketing surveillance in patients receiving azithromycin. Although the absolute risk is unknown, it appears to be low with azithromycin likely due to the lack of appreciable drug interactions, and the observation that it is rarely reported as a postmarketing adverse event. However, it would be prudent to avoid use in patients with known prolongation of the QT interval, patients with ongoing proarrhythmic conditions such as uncorrected hypokalemia or hypomagnesemia, clinically significant bradycardia, and in patients receiving Class IA (quinidine, procainamide) or Class III (dofetilide, amiodarone, sotalol) antiarrhythmic agents. Elderly patients may be more susceptible to drug-associated effects on the QT interval."3

In response to the publication by Ray and colleagues, FDA has indicated that they are reviewing its results and will communicate the results of that review when it is completed. In the meantime, they provide the following advice:

"Patients taking azithromycin should not stop taking their medicine without talking to their healthcare professional. Healthcare professionals should be aware of the potential for QT interval prolongation and heart arrhythmias when prescribing or administering macrolides."4

The 47 excess cardiovascular deaths per 1 million azithromycin prescriptions in the U.S. together with the 52.6 million azithromycin prescriptions in 2010 would indicate that there were approximately 2472 deaths attributed to azithromycin that year. This may be compared to a preliminary estimate of 646,421 deaths in individuals 25 – 74 years of age in 2010 [5]. For further comparison, preliminary data for that year indicates that, for all ages, there were 8352 deaths related to HIV infection, 7554 from viral hepatitis, 7284 from Clostridium difficile infection, and 569 from tuberculosis.5

The fact that more than 50 million azithromycin prescriptions are written in the U.S. in a single year should be surprising but, unfortunately, it is not. The perceptions of azithromycin have been that it is well tolerated, safe, and has few, if any, pharmacokinetic interactions. This view, together with the ease of writing a prescription for it ("Z-Pak" – a brilliant ploy by the marketing department) – has made it easy to make it the antibiotic of choice for many patients, especially those who do not actually need an antibiotic. The report by Ray and colleagues certainly disrupts this view, although it would be quite useful to determine if their results can be duplicated elsewhere, especially since it is likely impossible to totally eliminate all confounders in such an analysis. In the meantime, perhaps the most important result of this study may be to make clinicians think twice about whether their next antibiotic prescription is truly necessary. Clinicians may also wish to examine the extraordinarily long list of other medications, most not antibiotics, that may prolong cardiac repolarization.6


  1. Ray WA, et al. Oral erythromycin and the risk of sudden death from cardiac causes. N Engl J Med. 2004 Sep 9;351(11):1089-96.
  2. IMS Institute for Healthcare Informatics. The use of medicines in the United States: Review of 2010. April 2011.
  3. Medwatch Safety Labeling.
  4. FDA. Zithromax (azithromycin): FDA statement on risk of cardiovascular death:
  5. Murphy SL, et al. Division of Vital Statistics. Deaths: Preliminary Data for 2010. National Vital Statistics Reports. Volume 60, Number 4.
  6. Center for Education and Research on Therapeutics: