Date:  May 22nd, 2014 

Guest Skeptics: Dr. Salim Rezaie is an Assistant Professor in the Department of Emergency Medicine and Internal Medicine at the University of Texas at San Antonio. You may better know him from his website REBEL EM or twitter handle @srrezaie.

Case Scenario: 56yo man presents to the emergency department and is diagnosed with community acquired pneumonia. He has a history of osteoarthritis, which he takes ibuprofen on a PRN basis. He does not have a history of cardiac disease but does have an allergy to penicillin. You are about to discharge him home on azithromycin when the resident raises concern about an FDA warning with z-packs and arrhythmias.

Question: Does use of azithromycin lead to increased cardiovascular death?

Background: Erythromycin and clarithromycin have been shown to have an increase risk of serious cardiac arrhythmias. It was thought that Azithromycin did not share these cardio toxic effects.

Some recent literature has called into question on whether or not azithromycin can lead to serious cardiac arrhythmias and/or death.

Article: Ray et al. Azithromycin and the Risk of Cardiovascular Death. NEJM 2012.

  • Population: Adults 30-74yo enrolled in Medicaid for >1yr and regular medical care use, no life-threatening non-cardiovascular illnesses, non-nursing home resident, no prior drug abuse, no hospitalization past 30days.  Data extracted from Tennessee Medicaid program 1992-2006.
  • Intervention: Azithromycin script (5 day course)
  • Comparison: Frequency matched cohorts with placebo or other antibiotics (amoxicillin, ciprofloxacin, levofloxacin)
  • Outcomes: Primary outcome was cardiovascular death and all cause mortality.

Authors Conclusions: “During 5 days of azithromycin therapy, there was a small absolute increase in cardiovascular deaths, which was most pronounced among patients with a high baseline risk of cardiovascular disease.”

checklist-cartoonPrognostic Study Quality Checklist:

  1. The study population included or focused on those in the ED? Unsure
    • No specific comment on where patients received prescriptions
  2. The patients were representative of those with the problem? Yes
  3. The patients were sufficiently homogeneous with respect to prognostic risk? Yes
  4. Objective and unbiased outcome criteria were used? Yes
  5. The follow-up was sufficiently long and complete? Yes
    • Patients were followed up to 10 days of antibiotics
  6. The effect was large enough and precise enough to be clinically significant? Yes

Key Results: 

  • Patients taking no antibiotics = 1,391,180
  • Patients taking amoxicillin = 1,348,672 prescriptions
  • Patients taking azithromycin = 347,795 prescriptions
  • Patients taking ciprofloxacin 264,626 prescriptions
  • Patients taking levofloxacin 193,906 prescriptions

Majority of azithromycin and amoxicillin prescriptions were for upper respiratory infections (URI) or lower respiratory infections (LRI). Ciprofloxacin was mainly for urinary tract infections (URI). Levofloxacin was used for URI/LRI and UTI.

Statistical Analysis: Propensity Score Model (PMS) = statistical analysis of observational study that tries to remove confounding factors to remove bias from the analysis by comparing a treatment arm to a non-treatment arm. Check out this good review on the topic of PMS  in observational studies from the University of Manitoba.

  • Azithromycin vs. no antibiotics had increased risk of cardiovascular death HR 2.88 (CI 1.79 – 4.63) and death from any cause HR 1.85 (CI1.25 – 2.75)
  • Azithromycin vs. amoxicillin had increased risk of cardiovascular death HR 2.49 (1.38 – 4.50) and death from any cause HR 2.02 (CI 1.24 – 3.30)
  • Cardiovascular death was 29.8 per million with no antibiotics, 85.2 per million with azithromycin, and 31.5 per million with amoxicillin
  • Risk of death was highest in patients with the highest cardiovascular risks (245 additional deaths per million 5-day courses of antibiotics)
Dr. Salim Rezaie

Dr. Salim Rezaie

Comments: A challenge of observational studies is the blurring of associations with causality.  The authors here show an association of increased risk of cardiovascular/all-cause death with azithromycin 5 day prescriptions compared to placebo or amoxicillin/ciprofloxacin/levofloxacin matched controls in otherwise reasonably balanced cohorts. The event rates seem rather small (approximately 100 deaths per million scripts; 0.01%), yet the hazard ratios (HR) are likely statistically significant on the basis of huge numbers within the comparison cohorts.

An example of correlation not equaling causation was beautifully demonstrated in a recent blog called Spurious Correlations. It illustrated multiple examples of very strange things that seem to have a correlation.

  • Per capita consumption of mozzarella cheese vs. civil engineering doctorates awarded (correlation = 0.958648)
  • People who drowned after falling out of a fishing boat vs. marriage rates in Kentucky (correlation = 0.952407)
  • Honey producing bee colonies vs. juvenile arrests for possession of marijuana (correlation = -0.933389)

Screen Shot 2014-05-21 at 2.31.46 PM

In rebuttal commentaries to NEJM, various company and non-conflicted reviewers caution that prior randomized controlled trials with azithromycin have not shown increased cardiovascular deaths, and that observational study results need to be interpreted with caution.  There are possible differential factors that may be more associated/causal with cardiovascular death beyond azithromycin use (eg. Chlamydia Pneumonia infection) that are not captured in these data sets.

The Food and Drug Administration did issue a safety announcement after this study was published. It issues a warning that azithromycin “can cause abnormal changes in the electrical activity of the heart that may lead to potentially fatal irregular heart rhythm”.

Patients and health care providers were encouraged to discuss this potential risk when considering appropriate antimicrobial therapy.

BillyIdolRebelYellWe promised the audience not one but two critical reviews on the SGEM this week.

So, with a REBEL-EM yell, “he cried more, more, more…”

Reference #2: Svanstrom H et al. Use of Azithromycin and Death from Cardiovascular Causes. NEJM 2013.

  • Population: Adult patients 18–64 years of age, living in Denmark between 1997–2010 not hospitalized or given antibiotics in the prior 30 days
  • Intervention: Use of azithromycin
  • Comparison: No antibiotics or use of penicillin during the same time period
  • Outcome:
    • Primary = Cardiovascular death
    • Secondary: All-cause mortality

checklist-cartoonPrognostic Study Quality Checklist:

  1. The study population included or focused on those in the ED? Unsure
    • No specific comments on where prescriptions were received. Azithromycin users were less likely to have visited an emergency department in the past month compared to penicillin users.
  2. The patients were representative of those with the problem? Yes
  3. The patients were sufficiently homogeneous with respect to prognostic risk? Yes
  4. Objective and unbiased outcome criteria were used? Yes
  5. The follow-up was sufficiently long and complete? Yes
  6. The effect was large enough and precise enough to be clinically significant? Yes

Key Results:

  • Patients taking no antibiotics = 7,084,184 prescriptions
  • Patients taking penicillin = 7,364,292 prescriptions
  • Patients taking azithromycin = 1,102,419 prescriptions
  • Risk of cardiovascular death from Azithromycin use (5 days of treatment) vs. no antibiotics = RR 2.85 (CI 1.13 – 7.24), but this may attributable to the increased risk of death associated with acute infection
  • Risk of cardiovascular death from Azithromycin use (5 days of treatment) vs. penicillin V = RR 0.93 (CI 0.56 – 1.55)
  • No increased cardiovascular death with recent (past 6 – 10 days) or past use (past 11 – 35 days)
  • No increase in all cause death

Screen Shot 2014-05-24 at 10.28.54 AMCommentary: This is another observational study but of the entire Danish population aged 18-64. It used similar propensity score model to remove confounding factors (bias). As discussed before observational trials showing correlation does not equal causation.

This study seems to find the opposite to the previous study by Ray et al which showed a higher risk of cardiovascular death with use of azithromycin. This effect, however, seemed to be linked to higher cardiovascular risk profiles in that US Medicaid cohort (up to age 74), who were not included in this younger Danish population study.

There were a number of limitations to this study. They did not provide the indication for the antibiotic treatment. There was no information on all cardiovascular risk factors or body mass index of patients. The number of events in the subgroup analysis (age and sex) were small. Cardiovascular causes of death were not specifically defined (arrhythmia vs. acute myocardial infarction).

An editorial by Mosholder et al in this same NEJM issue reaffirms that, despite concerns about QTc prolongation with azithromycin in patients with higher cardiovascular risks, it seems that macrolides are, in fact, safer than other antibiotics (eg. fluoroquinolones) in hospitalized or ambulatory patients with Community Acquired Pneumonia.

There may be other cardiovascular effects of macrolide antibiotics besides arrhythmias. SGEM#36 had discussed the risk of erythromycin and clarithromycin and its association with hypotension in patients taking calcium-channel blockers. This association did not seem to extend to azithromycin.

Local guidelines, pathogen isolates frequencies and resistance patterns should be coupled to patient cardiovascular risk profile to make the best choice about azithromycin usage. Consider the potential risks and potential benefit when prescribing antibiotics:

  • Risks: Increased arrhythmia/cardiovascular death in those at higher cardiovascular risk, resistance and more hypotensive events if concurrent use of calcium channel blockers.
  • Benefits: May be of greater benefit/safety compared to fluoroquinolones, but definitely not when compared to beta-lactams.

The authors conclusions about relative azithromycin safety in younger adult patients are likely appropriate, with the caveat that there should be some assessment of cardiovascular risk of arrhythmia/death and preferential use of alternative agents when possible.

azithromycinBottom Line:  Azithromycin is not associated with an increased risk of cardiovascular death in a general adult population when compared to other antibiotics, provided that there is a low baseline cardiovascular risk of arrhythmia/death.

Clinical Application: Azithromycin is not unsafe in general adult patients with a low risk of cardiovascular disease. This is tempered by local treatment guidelines and resistance patterns for the infectious disease for which azithromycin is indicated.

What do I tell my patient: Azithromycin can be a safe choice for bacterial infection treatment, as long as the adult is not >65yrs age and/or at elevated risk of cardiovascular disease.

Case Resolution: You thank the resident for raising the concern. A discussion is held with the 56 year old man. You consider him low risk from a cardiac standpoint and recognize he has an allergy to penicillin. A shared decision process is made and he is discharged home with the prescription for azithromycin.

Dr. Salim Rezaie covered this topic on his blog REBEL EM.

Keener Kontest: Last weeks winner was Chris Miller from Washington University in St. Louis. He knew that Video Killed the Radio Star by the Buggles was the first music video ever played on MTV in 1981.

If you want to play the Keener Kontest this week then listen to the podcast for the question. Email me your answer at TheSGEM@gmail.com with “keener” in the subject line. The first person to correctly answer the question will receive a cool skeptical prize.

Remember to be skeptical of anything you learn,

even if you heard it on the Skeptics’ Guide to Emergency Medicine.