Azithromycin — it’s not just for chlamydia! Shockingly, the good people at Pfizer rejected my suggested slogan, but hey, their loss. Macrolide antibiotics have always occupied an odd spot in medicine, as anything that is killed by macrolides is also killed by strong language or an exceptionally stern glare. What I’m saying is, they’re not exactly the backbone of modern antimicrobial therapy.
But something weird and kind of wonderful has happens when you give them to patients with chronic obstructive pulmonary disease (COPD), and they have fewer exacerbations. Perhaps due to non-antibiotic effects, chronic daily azithro has been an approved treatment for severe COPD since a 2011 randomized controlled trial (RCT) demonstrated its ability to reduce the number of COPD exacerbations compared to placebo. While there have been a few trials of daily azithro in severe asthma previously, they have been mostly small and short-term, and have failed to demonstrate significant differences for important endpoints such as exacerbation frequency. So good on Peter Gibson and colleagues as their July 4th Lancet piece presents the results of a larger trial designed to definitively determine the benefits of azithromycin in patients with severe persistent asthma.
For my American readers who are having trouble reconciling themselves with reading a trial by am Australian investigator published in a British journal on July 4th, just open the most American thing you can find in another tab. May I make a recommendation?
If You’re Only Going to Read One Paragraph
In this double blind randomized controlled trial of azithromycin 500mg three times weekly versus matching placebo in patients with asthma not controlled on an inhaled corticosteroid (ICS) and long-acting bronchodilator (LABA), azithromycin resulted in fewer exacerbations (1.07 per patient-year for azithro compared to 1·86 per patient-year with placebo; incidence rate ratio [IRR] 0·59 [95% CI 0·47–0·74]; P < 0·0001), lower likelihood of at least one exacerbation (44% of patients in the azithromycin group versus 61% of patients in placebo, P < 0.0001), and improved asthma-related quality of life. This benefit was consistent between the eosinophilic and non-eosinophilic phenotypes. This comes at the expense of more diarrhea (34% vs 19%, P = 0.001), but similar rates of adverse events on the whole, with no significant difference in hearing loss, nausea/vomiting, or QTc prolongation (though absolute event rates of this outcome were low). While it’s difficult to make an apples-to-apples comparison of the effect size compared to monoclonal anti-IgE therapy like omalizumab due to the difference in baseline characteristics of the studied populations, it’s fair to say that azithromycin is well-tolerated over one year of therapy and offers benefit in patients who might qualify for anti-IgE monoclonals at a fraction of the cost, as well as in patients who would not qualify for these wonder drugs but still have severe disease.
A Bit of Context
If you ever want to check up on the latest expert consensus surrounding the diagnosis and management of asthma, just Google “GINA.” The Globial INitiative for Asthma, the acronym for which which could never be mistaken for crude innuendo, is the sister organization of the perhaps better known Global Initiative for Obstructive Lung Disease (GOLD — I think we all know who won the acronym lottery here). You can download their 2017 Global Strategy for Asthma Management and Prevention here.
I was initially going to write about this NEJM article, in which fancy drug designers design fancy drugs in an effort to eliminate the IgE response by any means necessary. Benralizumab is a novel humanized monoclonal antibody that binds the interleukin-5 (IL-5) receptor — IL-5 has a central role in the IgE mediated immune response, and anti-IL-5 antibodies have been previously shown to be effective in reducing the number of exacerbations in patients with an eosinophilic phenotype of asthma. The eosinophilic phenotype, by the way, refers to asthma that is principally allergic/IgE mediated, and is diagnosed via both peripheral blood eosinophil counts as well as sputum studies. Don’t thank me — thank GINA.
Shockingly, benralizumab also works for the endpoint chosen (reduction in oral glucocorticoid use — not exactly mortality, but certainly clinically important). But you know my bias. I’m a primary care doctor, which means that in addition to worrying about treating the one patient in front of me, I’m also worried about treating the 1000 patients on my panel and the 7 billion in the world. Having a population health mindset dictates that we care about costs and sustainability, and when a vial of omalizumab costs over $1000 (and that’s with a free coupon — thanks GoodRx!), you’ve gotta wonder if there’s a cheaper option.
Macrolides are weird. Yes, the’re pretty simple ribosomal subunit inhibitors, but they also have an array of anti-inflammatory effects. They have been shown, for example, to reduce levels of inflammatory markers and neutrophils in patients with refractory asthma. We already talked a little bit about azithromycin for COPD, but it’s worth noting that they also seem to work in cystic fibrosis, reducing the total number of exacerbations and improving PFTs. Oh, and the COPD effect is in spite of the fact that > 50% of patients were colonized with bacteria resistant to macrolides. So we have a drug with a mechanism of action we know and understand, but also with a benefit that seems to be mediated by some other mechanism entirely. If this sounds familiar to you, it’s probably ringing bells of the non-LDL statin hypothesis — macrolides are rapidly becoming the statin of the lungs (hey, that’s a good title).
But what about the role of macrolides in asthma? This Cochrane reivew, last updated in 2015, does a pretty decent job summarizing the evidence for macrolides in asthma thus far. Trials have mostly been small (the largest with 210 patients) with short follow up periods (a few year-long trials here and there, but mostly 8 to 16 weeks in duration). So the good folks at Cochrane conclude that while pooling of the existing data did not support a benefit for clinical outcomes, the absence of evidence was not evidence of absence. There was room for a larger, longer, better-designed trial to address the issue once and for all.
Patients, Intervention, Comparator, and Outcomes
This multicenter, randomized, double-blind RCT recruited patients with history of asthma and consistent PFTs (obstruction with post-bronchodilator reversibility of at least 12% or a few reasonable surrogate criteria) who remained symptomatic on a standarized asthma control scale (the ACQ6) despite ICS and long-acting bronchodilator use (practically speaking, almost all patients were on both ICS and a long-acting beta agonist [LABA]). Patients were recruited without regard to eosinophilic or non-eosinophilic phenotype, but a prespecified subgroup analysis was planned. The one phenotype they did care about was COPD overlap. To exclude patients with concomitant COPD, they excluded anyone with more than a 10 pack-year smoking history if their diffusing capacity for carbon monoxide was less than 70%, as well as excluding all current smokers (verified via exhaled carbon monoxide). So, asthma patients still smoking aren’t represented here, but no huge loss — these patients don’t need azithromycin, they desperately need help to quit smoking. Other exclusion criteria included baseline QTc prolongation or hearing loss, so as to minimize the risk of adverse effects from azithromycin.
Following a two week run-in period during which they were screened for adherence to their baseline regimen (via diaries and inhaler dose counts), patients with stable ACQ6 scores and at least 80% adherence were randomized to receive azithromycin 500mg by mouth three times per week or matched placebo. This is a bit of a departure from the COPD protocol, in which patients received 250mg daily, but given the weirdly long half-life of azithro, it’s not unreasonable to think that either strategy could have similar outcomes. Still, I wonder if adverse events like diarrhea might have occurred less frequently with a lower dose.
It’s unclear how many patients were excluded due to poor adherence. 162 were lost between screening and randomization, but 108 those are under the broad umbrella of “inclusion criteria not met.” There was in fact a fair amount of attrition throughout the study, with 86 of the patients who were randomized subsequently dropping out (mostly due to being lost to follow up or to adverse events), though it happened in both groups and the patients were appropriately included in the intention-to-treat analysis using the last observation carried forward method. It’s worth noting that this may have biased the study if patients dropped out non-randomly, but there’s no evidence to suggest that this happened.
The co-primary outcomes were total number of moderate or severe asthma exacerbations (moderate defined as increase in ICS or antibiotics in conjunction with an ACQ6 score change of at least 0.5 or any increase in beta agonist use sustained for at least 2 days) over 48 weeks of follow up and asthma quality of life. Secondary outcomes included ACQ6 scoring, sputum cell counts, PFT changes, and assessment of sputum microbiota.
Patients had a mean age of ~60.5 years and were ~60% female. 39% were former smokers, with a mean history of 7.5 pack-years. Race was not reported, and as this was an Australian trial it is reasonable to assume that there was a paucity of black patients represented. 100% of patients were using an ICS, with 85% on a high-dose, and 99% were using a LABA. 4% used leukotriene modifiers, and ~17% used long-acting muscarinic medications despite not-great evidence for use of these drugs in asthma. 34% of patients had an eosinophilic phenotype.
So how’d it go?
And there you go. The azithromycin group saw fewer exacerbations (1.07 per patient-year for azithro compared to 1·86 per patient-year with placebo; incidence rate ratio [IRR] 0·59 [95% CI 0·47–0·74]; P < 0·0001) and lower likelihood of at least one exacerbation over the study period (44% of patients in the azithromycin group versus 61% of patients in placebo, P < 0.0001). This effect was consistently seen across all subgroups, including non-eosinophilic asthma and patients with or without bacteria isolated from induced sputum samples. The azithromycin group also experienced a 0.36 point improvement on the Asthma Quality of Life Questionaire (AQLQ), driven mostly by improvements in reported symptoms.
But azithromycin is a drug with a well-known adverse effect profile, including the ever-testable catastrophe of QT prolongation as well as the much more common events of diarrhea, nausea, and vomiting. Did you know that about a third of patients will vomit if you give them a gram of azithromycin? And if they do keep it down, about half of them will have diarrhea. So it should be no surprise that more patients receiving azithromycin rather than placebo complained of diarrhea, 72 [34%] vs 39 [19%] (P = 0·001). However, we can take some comfort in the fact that adverse events as a whole, as well as rates of drug discontinuation due to these events, were similar between the two groups. You can hold out some concern that the more than doubled rate of QTc prolongation in the azithro group might prove to be significant in a larger study, and certainly it should be a target for long-term observational surveillance. But in this reasonable old population, hearing loss occurred more frequently with azithromycin than placebo.
There is some reasonable concern for azithromycin resistance being induced by this kind of chronic exposure. This is not the be all and end all of antimicrobial resistance, but would certainly be inconvenient — azithromycin is a reasonable empiric outpatient therapy for community acquired pneumonia in the outpatient setting, especially given FDA warnings that fluoroquinolones should be reserved only for patients without other reasonable treatment options. Of the 180 patients providing a sputum sample at the end of the study, 37 samples grew pathogenic bacteria (similar between the two groups). Rates of azithromycin resistance among these isolates were statistically similar between groups, but just barely; 19 [49%] of 39 intervention group isolates were resistant compared to 12 [29%] of 42; P = 0.062. So, with a reasonable mechanism and a concerning trend, I think it is reasonable to assume that this kind of chronic therapy is certainly at risk for inducing azithromycin resistance. The clinical and public health significance of this of course remains to be seen.
Interestingly, there have been a few studies with non-antibiotic macrolides, moieties that have had their structure altered to have less antimicrobial activity but appear to maintain their anti-inflammatory effects. Obviously the antibiotic effect may be part of azithromycin’s mechanism of action in asthma, but it will be interesting to see whether non-antibiotic macrolides can also decrease exacerbations without inducing higher rates of macrolide resistance.
At the end of the day, as a doctor who takes care of primarily poor Medicare and Medicaid patients, this is huge for my uncontrolled asthmatics. Getting patients to attend an allergy appointment, which may have a huge copay, get diagnosed with an eosinophilic phenotype, and then navigate the incredibly convoluted process of getting these expensive monoclonal antibodies approved has been a huge barrier to bringing the benefits of novel asthma therapies to my patients (i.e. I’m 0 for 3 lifetime). Azithromycin is cheap (though chronic use will probably still require a prior auth), and I think this represents a huge opportunity to improve healthcare utilization and quality of life.
As we come up with ever more intricate and scientifically impressive therapies to treat old diseases, it’s important that we circle back on the old drugs as well. Using azithromycin for something more than just ineffectively treating viral URIs in urgent care is a boon for medicine.
Read the original study here.