Time to Remove Black Box From LABAs for Asthma

Andrew Shorr, MD, MPH


July 27, 2016

This feature requires the newest version of Flash. You can download it here.

This is Andy Shorr from Washington, DC, with a pulmonary critical care literature update. I'd like to discuss an article that was just published online ahead of print in the New England Journal of Medicine, which reported the results of the AUSTRI study.[1] This study focused on safety issues in asthma with the use of long-acting beta agonists (LABAs)—in this case, salmeterol.

As many of you know, there's a very clear black box warning from the US Food and Drug Administration (FDA) regarding the use of salmeterol in asthma, which carries a very substantial warning about mortality, hospitalizations, and intubation. This warning was based on some data from studies done nearly two decades ago, in particular the Salmeterol Multicenter Asthma Research Trial (SMART), which looked at salmeterol as monotherapy in patients with asthma.[2] The study had a clear safety concern, but it was poorly done and it certainly didn't reflect the way we use salmeterol today as part of combination therapy with an inhaled glucocorticoid in the management of moderate persistent asthma.

Because of the these warnings from the FDA, there's always been a lot of concern about the safety of salmeterol in asthma, and that's led to major pharmaceutical companies being ordered by the FDA to do large, randomized controlled trials specifically focusing on the safety of the combination of a LABA and an inhaled glucocorticoid. AUSTRI represents the first of these studies to be reported.

This was a nearly 12,000-patient study, again mandated by the FDA, conducted by GlaxoSmithKline and focusing on severe safety outcomes. In this case, the primary endpoint, which was determined by regulators, was a combination event of mortality, hospitalizations, or intubation related to asthma. The study compared the use of salmeterol plus fluticasone versus fluticasone alone, enrolling patients as young as 12 years old, so it looked at both adolescents and adults with moderate to severe persistent asthma.

In addition, this trial was unique because it required patients to have a history of at least one asthma exacerbation in the past year, but not within the past 4 weeks. Thus, the population consisted of people who were already at substantial risk for future asthma exacerbations.

One of the unique things about this trial, other than its monstrous sample size, was that it stratified people by asthma severity based on the use of medications at baseline and also by their asthma quality control based on their asthma control test (ACT) score. That's important because it wasn't just a study of people with mild asthma who were at low risk for these events. They really did try to enrich for the safety endpoint by looking at patients who had prior exacerbations and who may have had suboptimal control.

The Case for 'Noninferiority'

This was a noninferiority study, with noninferiority defined as a hazard ratio less than 2 in the salmeterol group for the safety outcome. In other words, they were willing to see up to a twofold increase in the risk for intubation, hospitalization, or death and still conclude noninferiority. That personally concerns me, because when we do noninferiority trials, it's important to choose a tight safety margin because that's usually what we're focused on. I don't think a twofold increased risk for a bad outcome really allows us to declare noninferiority. However, that's what was chosen in this trial.

Patients were randomized and were followed for 26 weeks. It was an international study and, importantly, 15% of the patients were African American. That's crucial because concerns about LABAs are increased in the African American subpopulation with asthma because they may be genotypically predisposed to different kinds of responses to medical therapy. At the end of the trial, there was really no difference in the primary safety endpoint. The hazard ratio for the event was 1.03 and the confidence interval was well below the 2.0 risk margin (0.64-1.66), so the authors concluded noninferiority for the combination therapy and monotherapy with fluticasone.

More important than answering the safety question, the authors actually looked at efficacy with an endpoint of asthma exacerbations, which were defined as a need for corticosteroids or hospitalization for asthma. What they saw was that there was over a 20% risk reduction in asthma exacerbations with combination therapy compared with monotherapy.

So, now not only have we answered the safety question, but we have also conclusively demonstrated the efficacy benefit of combination therapy in this population with asthma. Beyond that, there was also less rescue albuterol use with combination therapy, and the benefit of asthma exacerbation reduction was seen across all of the subgroups, including the African Americans who also did not have increased safety concerns.

Increased Risk for Exacerbations

This trial speaks volumes to how we answer safety questions with respiratory medications. We had a similar trial done when there were concerns about the new device for the delivery of a long-acting muscarinic agent.[3] That large, randomized controlled trial was done and the question was answered. It is the same thing here. I think that in the vast majority of patients, we now have very good data that there is no safety concern with combination therapy, and by maintaining that black box warning, the FDA is actually exposing patients to risk in terms of asthma exacerbations by not backing off and decreasing the pressure on physicians to avoid LABAs in patients with asthma.

Conversely, this trial doesn't tell us anything about the most brittle asthmatics. Again, these patients could not have had an asthma exacerbation in the month prior to enrollment. In the most brittle patients who perhaps don't comply with their medications, these data don't provide us with much new information. However, this is still nearly 12,000 patients' worth of data and it's valuable.

Another limitation of the trial, pointed out by Fernando Martinez in his editorial,[4] was that they didn't require strong diagnostic evidence of asthma. It was really just by history and report of asthma exacerbations. They didn't require spirometry or a methacholine challenge. That's clearly a limitation but also a tradeoff: If you want to enroll nearly 12,000 patients, you have to balance the cost of the study against the need to collect data. Nevertheless, with these large, simple clinical trial designs, we get lots of information across a broad population.

We will see a similar trial looking at formoterol, because it's a different agent from salmeterol.[5]

I think this study goes a long way to demonstrate the safety of combination therapy, document its efficacy, and support that when you take away LABAs from patients, you are actually exposing them to increased risk for exacerbations. Those are data that have been published previously.[6] I think the FDA really needs to go back and look at the data that it has requested and reevaluate the language in their black box warning.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: