Azithromycin for COPD

Motivation: During the first couple of months of internship, I have often taken care of patients with COPD, and in the morning presentations, the conversation inevitably turns towards the benefits of adding an antibiotic to the steroid regimen.  Macrolide antibiotics are thought to have additional anti-inflammatory benefits beyond the standard anti-microbial effects.  Recently, this hypothesis was extended further to test for beneficial effects of azithromycin in COPD patients as a standing drug in-between flares.

Paper: "Azithromycin for Prevention of Exacerbations of COPD", Albert, R.K. et. al. NEJM (2011) 365: 689-698.

Methods: A multi-centered, randomized controlled trial.  The trial included patients over 40 with a history of COPD (>10 pack year of smoking, FEV1/FVC<70%, and FEV1<80% of predicted value).  To the included in the study, the patients had to have either used systemic steroids or required ED/hospital care in the past 12 months.  Patients could also be included if using home oxygen.  The exclusion criteria were tachycardia, prolonged QTc, and hearing impairment.  The intervention was daily dose of 250 mg of azithromycin.  The primary outcome was the time to first acute exacerbation of COPD.  The follow-up period was one year.

Results: 
Patient Recruitment: Overall, 558 patients were in the treatment arm, and 559 patients were in the placebo group.  The completion rate was 89% in the treatment arm and 90% in the placebo group.  The treatment and placebo group did not differ significantly in terms of age, gender, race, disease severity, or home treatment regimens.

COPD Exacerbation: The median time to acute exacerbation was 266 days in the azithromycin group vs. 174 days in the placebo group (p<0.001).  A total of 741 exacerbations among 558 patients occurred in the azithromycin group while 900 exacerbations occurred in the placebo group (p=0.01).  The number needed to treat to prevent one acute exacerbation of COPD was 2.86.

Secondary Outcomes: There was not a statistically significant difference between the azithromycin and placebo groups in hospitalizations related to COPD (156 vs. 200, p =0.15), exacerbations requiring intubations (11 vs. 16, p=0.56), or hospitalization for any cause (323 vs. 329, p=0.52).  There was a trend towards decreased ED or urgent care visit (199 vs 257, p=0.09).

Adverse Effects: The rate of death was 3% in the azithromycin group and 4% in the placebo group (difference not significant).  The only adverse effect that occurred more frequently in the azithromycin group was audiogram confirmed hearing decrease (25% vs 20%, p=0.04).  At the end of the one year study, patients receiving azithromycin were also more likely to be colonized in the nasopharynx with azithromycin resistant organisms.

Discussion: The trial demonstrates that a daily standing regimen of azithromycin results in decreased number of total exacerbations.  While there is evidence that increased number of exacerbations results in increased morbidity and mortality, the trial was not sufficiently powered and did not have a long-enough follow-up time to demonstrate mortality benefits or decreased hospitalizations.  I think that showing a mortality benefit or showing decreased hospital stays would have made this trial much more significant.  There was an interesting side effect of increased hearing loss in a sizable minority of patients.  The data presented in the paper though is unclear about the magnitude of hearing loss.  A little bit might be tolerable while total deafness is likely not.

In summary, if a patient keeps coming back with repeated exacerbations of COPD, a standing regimen of azithromycin might be the next medicine for them!

Magnesium for Torsades

Motivation: The standard ACLS catechism is that: "If you see torsades de pointes, give magnesium."  During the ACLS protocol, torsades is, I think, the only point where magnesium enters into the algorithm.  While supplementing someone with mag the other day, I wondered what the evidence was for use of magnesium in torsades and other rhythms.  I was a bit surprised to learn that the ACLS guideline for torsades is based mostly on two (quoted from the official ACLS publication) observational studies.  The major study in Circulation appeared in 1988 and is reviewed here:

Paper: "Treatment of torsade de pointes with magnesium sulfate" Tzivoni, D. et. al. Circulation (1988), 77: 392-397.

Methods: An unblinded, uncontrolled study in which 12 consecutive patients who developed torsades de pointes (TdP) with QT prolongation were treated with magnesium (2g IV bolus over 1 to 2 minutes followed in most patients with a continuous infusion).  Five additional patients with polymorphic ventricular tachycardia with normal QT interval were also treated with magnesium.

Results:
Patients: Of 12 patients with TdP (10 men and 2 women) and prolonged QT, six had ischemic heart disease, two had valvular rheumatic heart disease, two had atrial arrhythmia, and two did not have known heart disease.  Nine of the twelve patients were also on antiarrhythmic therapy.  The mean corrected QTc was 640 msec.  The five patients with polymorphic VT with normal QT interval who received magnesium all had chronic ischemic heart disease.

Patients with prolonged QTc: In nine of the twelve patients, a single bolus of magnesium stopped TdP.  In the other three patients, TdP was stopped after second Mag bolus.  In eight of 12 patients, potassium levels were below 3.5, which was repleted with oral and intravenous potassium.  The mean QTc did not change after magnesium bolus.  Magnesium levels were available in eight patients, and all were normal.

Patients without prolonged QTc: In the five patients with polymorphic ventricular tachycardia, none responded to multiple boluses of magnesium.

Discussion: One of the major points of the paper for me is that sometimes amazing results trump poorly designed studies.  Most of the patients had multiple confounders, including hypokalemia and concurrent anti-arrhythmic administration, but achieving a 100% response rate really makes this paper landmark.  And, really since the publication of the paper, magnesium has become the standard therapy of choice for TdP.  There were a couple of other interesting points in the paper as well.  The first is that not every polymorphic ventricular tachycardia responds to magnesium - only polymorphic v-tach in patient with prolonged QTc can be properly called TdP.  Also, while magnesium bolus is key, the precipitating factor for TdP is not hypomagnesemia since magnesium levels were normal.  Finally, repleting potassium and other electrolytes are pretty important too! 

The Flutter-Fibrillation Spectrum

Motivation: Every time I encounter a patient with atrial flutter and hard to control ventricular rates, I wonder why the patient is on the floor and not in an electrophysiology suite. I mean, why should we spend time titrating diltiazem when the ablation procedure (ablation of cavo-tricuspid isthmus) has more than 90% success rate.  In fact, general guidelines state that in contrast to a-fib, for patients with recurrent atrial flutter, ablation should be the first-line approach rather than rate control.  Ablation, however, is not a good long-term fix if patients with atrial flutter also develop atrial fibrillation, which does not respond well to ablation.  I have at times heard of the "flutter fibrillation spectrum." But, how often do patients with atrial flutter actually develop atrial fibrillation in the future?

Paper: "Risk of Stroke in Patients with Atrial Flutter", Biblo, L.A. et. al. Am. J. of Card. (2001) 87: 346-349.  http://www.ncbi.nlm.nih.gov/pubmed?term=11165976%20

Methods: In this large retrospective study, the authors scanned Medicare inpatient files in 1984 for patients older than 65 who were diagnosed with atrial flutter or atrial fibrillation without stroke.  For control, a 5% random sample of other hospitalized patients were chosen without atrial flutter or fibrillation.  Patients were followed for 8 years.  The primary endpoints assessed were incidence of stroke and incidence of atrial fibrillation in patients with atrial flutter.

Results:
Subjects: The study followed 17,413 patients with atrial flutter, 337,428 with atrial fibrillation, and 395,147 controls.

Atrial Flutter to Fibrillation Incidence: Patients with atrial flutter developed atrial fibrillation in an almost linear fashion over the 8 year follow-up period.  By about 6.5 years of follow-up, half of the initial group of patients with atrial flutter had developed an episode of atrial fibrillation.  The top three factors which predicted which patients would develop fibrillation were rheumatic heart disease (Risk Ratio: 1.464, CI: 1.250-1.715)., systemic hypertension (RR: 1.333, CI: 1.267-1.402), and congestive heart failure (RR: 1.243, CI: 1.174-1.316).

Stroke Risk: After adjusting for known risk factors like hypertension, CHF, rheumatic heart disease, DM, and MI, the stroke risk in patients with atrial flutter was greater than controls (RR: 1.406, p < 0.0001).  The adjusted stroke risk in patients with atrial fibrillation (RR: 1.642, p < 0.0001) was greater than those with atrial flutter.  Importantly, the stroke risk for patients initially with atrial flutter with new onset atrial fibrillation was not significantly different from the stroke risk in patients with atrial fibrillation.

Discussion: I think that this paper nicely illustrates why it is hard to manage atrial flutter.  Patients with atrial flutter are at significant risk for atrial fibrillation (about half have atrial fibrillation by about 6.5 years of follow-up).  In these patients, going through ablation is questionable since a-fib requires rate-control and anti-coagulation regardless.  On the other hand, for the half of patients who do not develop a-fib by seven years, conversion to sinus rhythm through ablation would save them many years of hazards of warfarin anti-coagulation.  I think that it is still unclear exactly which subgroup benefits most from ablation.  Likely as suggested by the paper, patients with additional risk factors like HTN, CHF, and rheumatic heart disease are likely to go on to develop a-fib.  Another point underscored by the paper is that while both atrial flutter and fibrillation have elevated stroke risk, patients with atrial fibrillation are at higher risk than those with flutter.   Finally, while the paper is impressive in the large number of subjects tracked, the paper is retrospective in nature and only examines patients older than 65.   Also, I have some doubts about how accurately atrial flutter/fibrillation is coded during hospitalizations (!).