Published: Mar 10, 2014
By Todd Neale, Senior Staff Writer, MedPage Today

Full Story

Action Points

• Azithromycin and levofloxacin were both associated with elevated risks of death and serious cardiac arrhythmias during standard lengths of prescription.
• Point out that although the risk of death was higher with azithromycin (through 5 days) and levofloxacin (through 10 days) compared with amoxicillin, the absolute risks were relatively low.

Azithromycin and levofloxacin were both associated with elevated risks of death and serious cardiac arrhythmias during standard lengths of prescription, a study of mostly male military veterans showed.

Through the first 5 days after a prescription was dispensed, the risk of all-cause death was 48% greater (hazard ratio 1.48, 95% CI 1.05-2.09) and the risk of serious arrhythmia was 77% greater (HR 1.77, 95% CI 1.20-2.62) with azithromycin, which is typically taken for 5 days, compared with amoxicillin, according to Gowtham Rao, MD, PhD, of the University of South Carolina in Columbia, and colleagues.

There were no longer any significant differences in risk 6 to 10 days after starting treatment, the researchers reported in the March/April issue of Annals of Family Medicine.

On the other hand, levofloxacin — typically given for 10 days — was associated with greater risks of death and serious arrhythmia through the first 10 days after starting treatment.

The results are consistent with a safety advisory issued by the FDA last year warning of the risks of possibly lethal heart rhythms when taking azithromycin or levofloxacin, Rao and colleagues noted.

“Risks and benefits of antibacterial therapies should be considered when making prescription decisions,” they wrote. “There are usually multiple antibiotic choices available for older patients, especially those with cardiac comorbidities; physicians may consider prescribing medications other than azithromycin and levofloxacin.”

Concerns about the cardiovascular risks of azithromycin were uncovered in a study published in 2012 in the New England Journal of Medicine, which showed that the risk ofcardiovascular death was higher with the macrolide antibiotic compared with amoxicillin or no antibiotic treatment. That study sparked the FDA review that culminated in the agency’s warning last year.

A study published last May, however, showed that azithromycin was not associated with a greater risk of cardiovascular death in younger and middle-age adults in Denmark compared with penicillin. The discrepancy between that study and the current analysis could be related to the younger average age of the participants and the inclusion of a general population sample in the Danish study.

To explore the issue further, Rao and colleagues examined data from veterans who received outpatient treatment with amoxicillin (979,380 patients), azithromycin (594,792 patients), or levofloxacin (201,798 patients) within the VA healthcare system from September 1999 to April 2012. The average age of the patients was 56.5.

The most common indications for antibiotic treatment overall were ear-nose-and-throat infections (29.3%), chronic obstructive pulmonary disease (14%), and other respiratory infections (11%).

Although the risk of death was higher with azithromycin (through 5 days) and levofloxacin (through 10 days) compared with amoxicillin, the absolute risks were relatively low. The numbers of deaths per million antibiotics dispensed at the end of 5 and 10 days of treatment were 154 and 324 for amoxicillin, 228 and 422 for azithromycin, and 384 and 714 for levofloxacin.

After accounting for potential confounders, the risk of death appeared to be greatest with levofloxacin. Compared with azithromycin, the risk of all-cause death was elevated through the first 5 days of treatment (HR 1.68, 95% CI 1.15-2.47) and 6 to 10 days after the start of treatment (HR 1.71, 95% CI 1.15-2.55). The risk of arrhythmia was similar between levofloxacin and azithromycin, however.

The authors acknowledged that the study was limited by the use of only three antibiotics, so the findings could not be used to select a potentially safer drug. In addition, the observational design left open the possibility of residual confounding, particularly if patients who received one drug were different in some unmeasured way from those who received another drug.