People taking 5 days of azithromycin had a very small absolute increased risk of death, especially due to cardiovascular causes, compared to people taking amoxicillin, in a retrospective cohort review by Wayne Ray, Katherine Murray, and C. Michael Stein published in the May 17 New England Journal of Medicine.
Erythromycin and clarithromycin (the other macrolide antibiotics) were already known to be cause heart arrhythmias, and have been associated with sudden cardiac death. Azithromycin has been generally believed to be free of this macrolide drug class effect. However, about 20-25 case reports to FDA and in the medical literature have reported a possible association between azithromycin and QT-prolongation, ventricular tachycardia, and torsades de pointes, including several cases in which the arrhythmias were fatal.
What They Did
Authors queried the Tennessee Medicaid program's databases, and found 347,795 prescriptions for azithromycin in people aged 30-74 without "non cardiovascular life threatening illness" between 1992-2006.
They observed the incidence of death and cardiovascular death during the 5 days of azithromycin therapy.
They then compared that to control "time periods" of exposures including:
- A propensity-matched group of patients who got no antibiotics (i.e., calculated to be similar to the exposure cohort using a model with 153 covariates)
- Patients who received amoxicillin, levofloxacin, or ciprofloxacin (drugs with similar or overlapping indications to azithromycin). Authors describe patient characteristics in the azithromycin and amoxicillin cohorts to be "very similar," while the fluroquinolone-taking patients were more likely to be diabetic or poorly functional.
The comparison was between time periods, not patients. That is, a patient who took multiple different antibiotic courses over 10 years could be counted in the analysis separately each time -- in effect, compared against himself and other people multiple times, while taking antibiotics or not. (Because the drug courses couldn't overlap, and death could occur only once, there was no methodologic/statistical problem with this, authors say.)
What They Found
More cardiovascular deaths occurred during treatment courses with azithromycin:
- An absolute 29 people taking azithromycin died of cardiovascular causes during their 5 day treatment course.
- Of the absolute 29 cardiovascular deaths, 22 were sudden cardiac deaths.
- Among the cohort, this represented 85 cardiovascular deaths per million 5-day azithromycin courses, and 65 sudden cardiac deaths per million azithromycin courses.
- That's a 0.0085% observed rate of cardiovascular death, or about 1 in 12,000.
These rates were significantly higher than in the control groups:
- 30 cardiovascular deaths (24 sudden cardiac deaths) per million periods of no antibiotics in propensity-matched control patients. This could be considered the baseline risk or expected cardiovascular death rate.
- 32 cardiovascular deaths (22 sudden cardiac deaths) per million amoxicillin treatment courses.
- These observed rates of cardiovascular death were about 1 in 30,000, either during a 5 day period of no antibiotics or while taking amoxicillin.
This suggested an absolute increase or "excess risk" of cardiovascular death associated with azithromycin of about 1 in 20,000 or so. Relatively speaking, this translated to a hazard ratio of 2.88 for cardiovascular death (95% CI 1.79-4.63) associated with azithromycin. Azithromycin was also associated with an elevated risk of death from any cause (hazard ratio of 2.02), compared to the first 5 days of amoxicillin treatment.
Ciprofloxacin and levofloxacin were not associated with increases in risk of death, although there was a trend toward risk with levofloxacin.
What It Means
Examining the "excess deaths" associated with azithromycin -- above the observed rate while taking amoxicillin or nothing -- almost 60% of the observed excess deaths were concentrated in the 10% of the patients at highest cardiovascular risk. (Patients were divided into deciles of cardiovascular risk using authors' own methods based on the variables available, such as diagnoses of heart failure or cardiovascular disease, or taking medications for same.)
This high-cardiovascular risk group had a rate of 245 cardiovascular deaths per million azithromycin courses, a death rate of about 0.0025%, or 1 in 4,000.
Excluding these patients reduced the observed risk to somewhere between 9 per million courses (for the lowest-risk cardiovascular patients) and 45 per million (for those in the 6th-to-9th deciles of risk), by these authors' analysis.
Since all these figures are being calculated, in the final analysis, from 29 absolute cardiovascular deaths in an observational cohort -- inherently vulnerable to unmeasured confounding -- it would be helpful to see these provocative and potentially practice-changing findings replicated before concluding definitely about azithromycin's cardiovascular risk.
Clinical Takeaway: Given the prevalence of cardiovascular disease in chronic obstructive pulmonary disease (COPD) patients, and the standard practice of treating COPD exacerbations with azithromycin, awareness of this study is a must for pulmonologists, hospitalists and primary care physicians. Because you know, unfortunately, lawyers blog too!
Ray WA et al. Azithromycin and the Risk of Cardiovascular Death. N Engl J Med 2012;366:1881-1890.