Friday, December 12, 2008

Dear FDA, Remove the Boxed Warning from Advair and Symbicort

If you heard the news, it seems I guessed right in my post on The Lowdown on LABA's . The FDA's advisers have recommended to withdraw the asthma indications for both LABA's, but reaffirmed the safety of ICS/LABA combinations such as Advair and Symbicort. However, all the media headlines make it sound like something bad has happened, and it is clear that some patients are already confused. Furthermore, given all the press that has surrounded this issue, the only way this will ever be put to rest is for the FDA to take the next step and remove the boxed warning on the ICS/LABA combination products.

Thus, here is a draft letter to the FDA that I will soon send to request such a move. Your comments are appreciated. Below the letter is the boxed warning for Symbicort for your reference.

Dear FDA,

I am thrilled that despite a very vocal minority that seems to wave the safety banner to promote their own agendas, your advisory panel has decided to remove the asthma indication for both formterol (Foradil) and salmeterol (Serevent), and feels strongly about the efficacy and safety of ICS/LABA combinations (Advair, Symbicort) for the treatment of asthma. The problem is that all of the news headlines which state "F.D.A. Panel Votes to Ban Asthma Drugs" and "FDA Advisers: Restrict Some Asthma Drugs" still continue to scare patients. I have one patient whose previously poorly controlled asthma was improved with Advair, but stopped because of an article she read in Men's Health Magazine that implied no doctor would ever prescribe this.

Almost all the controversy regarding LABA's has come from the SMART study, which showed a problem with salmeterol taken alone, but showed no issue when taken with an inhaled corticosteroid. At the time, you decided to give a boxed warning to all products containing LABA's. However in light of the pages and pages of data you reviwed before the recent committee meeting as well as your advisor's recommendations, it seems that now would be the time to reconsider this. The only way to definitely answer this question is a randomized trial of patients on ICS, with half on LABA and half on placebo or increased ICS does, looking at serious exacerbation or asthma death. The problem is that since these events are so rare, and since ICS/LABA does not seem to cause a problem, it would likely take a study of more than 200,000 patients to show what we already know; that ICS/LABA combination is safe and effective for asthma. In addition, no drug company or even the NIH would ever fund such a study.

Though asthma mortality has gone down since the mid-90's, despite an increasing prevalence (and probably helped with the introduction of LABA's), asthma morbidity continues to be a problem. Over 5000 ER visits and 11 deaths a day are due to asthma. ICS/LABA is the most effective treatment for asthma, but some patients and even physicians are concerned about taking them because of the LABA controversy caused primarily by the SMART study. The only way reassure doctors, patients, critics and the media that Advair and Symbicort are safe and effective is to remove the boxed warning from their labels. I truly hope you will consider this.


Dr. Matthew Mintz

Boxed Warning for Symbicort (Advair's is similar)

Important Safety InformationLong acting beta2-adrenergic agonists may increase the risk of asthma-related death. Therefore, when treating patients with asthma, SYMBICORT should only be used for patients not adequately controlled on other asthma-controller medications (e.g., low-to-medium dose inhaled corticosteroids) or whose disease severity clearly warrants initiation of treatment with two maintenance therapies. Data from a large placebo-controlled U.S. study that compared the safety of another long-acting beta2-adrenergic agonist (salmeterol) or placebo added to usual asthma therapy showed an increase in asthma-related deaths in patients receiving salmeterol. This finding with salmeterol may apply to formoterol (a long-acting beta2-adrenergic agonist), one of the active ingredients in SYMBICORT.


Unknown said...

Thanks for blogging about this. I feel much better now about being on Advair (for bronchiectasis, not asthma).

Laika said...

I would like to bring two Cochrane Systematic reviews to your attention, concluding that long-acting beta2-agonist inhalers give a higher risk of serious side events, even when used together with corticosteroids. See the last cochrane review about formoterol and my blogpost describing the podcast of this review.

The conclusion of the Cochrane authors is to keep using combinations if patients feel it helps (because there is just a small increase of risk), but to discuss the discontinuation in case of no benefit. Thus, according to their conclusions, your patient who stopped treatment with Advair should still be advised to continue taking it (as you did).

In my opinion, it is good to be aware of any possible adverse effects of medication, but these should always be balanced against the possible benefits (and vice versa) for a particular patient.

Dr. Matthew Mintz said...

Thanks for your comment.
I can't find the second Cochrane review you are referring to. The one you linked to regarding formoterol randomized patient to formoterol or placebo. Even though inhaled steroids could be used, not all the patients in all the studies were on inhaled steroids. This is exactly the same problem with the SMART study. The Cochrane authors had wanted to do a subgroup analysis (that was done in SMART) to look at the patients just taking the inhaled steroids, but this was not reported in the studies they found, so the analysis was not done. The subgroup analysis in the SMART study showed that there was no problem with patients already on an inhaled steroid. Also, the recent meta-analysis (link in orginal post) of patients on inhaled steroids with LABA added also showed no problems.
The bottom line remains the same. LABA's are potentially dangerous when used alone, but quite safe and effective when used with an inhaled steroid.
(I would be happy to look at the other Cochrane review, but couldn't easily find the one you were referring to).

Laika said...

Thanks for your response.

Here you can find the direct link to the podcasts discussing the formoterol and the salmeterol results and here is the link to the cochrane review on adverse effects of salmeterol. Both are written by the same authors.

You're correct that not all patients were on corticosteroids. In fact concomitant use of inhaled corticosteroids was allowed, as long as this was not part of the randomized treatment regime.(the main aim being to determine the adverse effect of the LABA)
My understanding is that the subgroup analyses of the corticosteroid-using group didn't show a clear adverse effect, but that the confidence interval for patients taking inhaled corticosteroids was too wide to prove that corticosteroids abolished the risks. Therefore it is not (yet) sure that ICS/LABA is safe. The podcasts and SR-summaries also refer to data submitted to the FDA.I find the podcasts quite interesting to listen to.

So your conclusion that "The only way to definitely answer this question is another randomized trial to determine the effect of ICS/LABA on serious exacerbation or asthma death" is correct

Although the adverse effects remain serious, the risk of getting them remain very low. Whether this warrants a box warning, well, I wouldn't think so (but I'm not an expert in the field). There are no boxwarnings on aspirin because it can cause anaphylaxis in some hypersensitive individuals, are there?

Kind regards, Jacqueline

p.s. If you like, I can send you the full Cochrane Reviews this Monday.

Dr. Matthew Mintz said...

Thanks, here is the exact statement you are referring to:

For the subgroup taking inhaled corticosteroids at baseline the increase in asthma mortality is smaller and not statistically significant OR 1.52 [0.51, 4.49]. However the test for interaction between inhaled corticosteroid use and asthma related deaths is not significant (Chi² = 2.43, df = 1, (P = 0.12)), which means that the difference in the effect on asthma mortality between the subgroups given inhaled corticosteroids or not at baseline may be due to the play of chance. Moreover it should be pointed out that the confidence interval for those on inhaled corticosteroids at baseline includes the overall three-fold increase in mortality, so this finding cannot be interpreted as meaning that corticosteroids abolish any increased risk of asthma mortality from regular salmeterol."

1. Please note that almost has the patients analyzed were from the SMART study, which I have discussed in previous posts
2. The key words are "not statistically significant." Everything else that follows is an opionin. There analysis showed no increased risk in patients on ICS.
3. Their point is that statistically, patient on ICS that subsequently took salmeterol could have either cut their risk of dying by half or increased their risk of dying almost 5 fold (this is what they mean by wide confidence intervals). They state because their is a chance possibility that it could increase risk, that safety can not be perfectly established (again, commentary)
4. The only way to know the answer with 100% certaintly would be to do a randomized trial which would take over 100,000 patients. This would be to probitively difficult and expensive to do. Therefore, you need to look at all the data (including their data) which states that for patients on ICS, adding a LABA has not shown to increase the risk of death.

Laika said...

I really appreciate that you're willing to discuss this subject in depth with me.

I agree with most you say, but not with your conclusions.

To put it very simple:
1. The meta-analyses were done to find evidence for adverse effects of LABA's. These were definitively demonstrated both for formoterol and salmeterol. No discussion about that.
2. Indeed it is a matter of statistical significance but also of direction. These data do not reliably show that corticosteroids are protective. This also means that the data don't exclude that there is NO extra risk in patients treated with the ICS/LABA (being just a subgroup). Statistically the two groups (LABA only and ICS/LABA) aren't any different.
3. Another point is that using ICSI at baseline does not say people use ICSI all the time.

However, the risks are extremely low. Thus for a patient doing well on ICS/LABA there is no urgent reason for discontinuation.

The Cochrane Reviewers refer a couple of times to the FDA. Is it possible that the FDA has more (postsurveillance?) data at its disposal? Or is it just that the FDA reasons the same way as I do? Did you get any response from them?

But apart from the evidence or the lack of it, I understand your reluctance not to give a treatment which you consider effective, just because there is a very low and (still) putative risk of adverse events.

Dr. Matthew Mintz said...

I similarly appreciate the good dialogue.

You are correct that
"the data don't exclude that there is NO extra risk in patients treated with the ICS/LABA (being just a subgroup)."

However since "the meta-analyses were done to find evidence for adverse effects of LABA's", "these data do not reliably show that corticosteroids are protective" because they weren't set up to do this.
All you can say is that in the subgroup analysis of patients taking ICS, there was no evidence of harm when adding a LABA. You can not say definitely that ICS are protecitve, though it does seem to be the case. Though the statistical possibility of increasing the risk of death fivefold sounds like a concern, these are analytic numbers, and given the extremely small event rates, should not be interpreted as a a major concern in the way the Cochrane folks interpret this.

The issue I have is context. ICS plus LABA is the best therapy we have for asthma. There is really no data that refute this. This is why the 2002 guidelines recommended ICS plus LABA as the preferred therapy for patient not controlled on ICS and in patients with moderate persistent asthma. The problem is that all the controversy, mainly generated from the SMART study has gotten some physicians and patients concerned about ICS/LABA therapy. This has not led to increase use of ICS monotherapy, but rather use of leukotriene modifiers (which are not preferred therapy) or patients stopping their medications altogether.
Thus, though there is no way to 100% prove that ICS are protective, the preponderance of the evidence is that they are, and that all the scary LABA stuff can be pinned on LABA monotherapy, which the FDA has now recommended be banned. While we can scientifically argue over a theoretic possibility which will never be answered becasue the study will never be done, patients need to be reassured that the medications they are taking or that their doctor has recommended are safe. That is my intention of all my LABA related posts.