For many patients, the use of intermittent short-acting bronchodilators is the easiest and usually an adequate strategy. Some patients with more frequent symptoms will require some form of maintenance therapy – in most cases, this is a low-dose inhaled corticosteroid.
In recent years we have seen the emergence of new strategies to manage mild asthma. Specifically, the use of long-acting B-agonists has become trendy. So, which is the best strategy? The National Asthma Council still recommends a stepped approach to limit the use of unnecessary medications and side effects. [Fun fact: This pyramidal diagram is called a ziggurat, according to an unreliable source, after the ancient Mesopotamian temples!]
That is how most of us practice. However, a new study published in the New England Journal last month claims that another strategy: using “as required” combination budesonide-efometerol (Symbicort) may be superior to our usual plans. The Novel START trial compared 3 strategies – so let us take a dive into this paper and see if the claims are justified.
POPULATION: Patients were all adults 18 – 75 years from Australia, New Zealand, Italy or the UK. 675 patients were randomised. 668 patients were included in the analyses. Patients had mild asthma, only 7.3% had reported a severe exacerbation in the previous 12 months, and 54% reported using a SABA on two or fewer occasions per week in the previous 4 weeks. Patients were excluded if they had been admitted in the year prior or had any sort of long term smoking history – i.e. they tried to exclude patients with COPD rather than primary asthma diagnosis.
INTERVENTIONS: There were 3 strategies that patients were randomised to follow for their asthma over a 52-week period.
albuterol (Ventolin), 100 μg, with two inhalations from a metered-dose inhaler as needed for symptom relief.
budesonide (Pulmicort), 200 μg, one inhalation twice daily, plus albuterol (Ventolin), 100 μg, two inhalations from as needed for symptom relief.
budesonide–formoterol (Symbicort), 200 μg of budesonide+ 6 μg of formoterol, one inhalation as needed for symptom relief. No regular preventer used.
Patients were given an Asthma Action Plan to follow and their usage of the medication was monitored using an electronic monitor. They were required to keep a log of visits for urgent / emergent care and any other steroid use eg. prednisolone for exacerbations.
OUTCOME: The primary outcome measured was the annualized rate of asthma exacerbations per patient. This was defined as worsening asthma that resulted in one or more of the following:
an urgent medical care consultation (e.g., a primary care visit, an emergency department [ED] visit, or hospital admission)
a prescription of systemic glucocorticoids for any duration
an episode of high β2-agonist use, (more than 16 doses of albuterol or 8 actuations of budesonide–formoterol over the course of 24 hours)
This seems like a sensible, patient-oriented outcome. It is also quite objective and easy to measure accurately.
There were a pile of secondary outcomes…. but those are tiger territory: there can be only one outcome in the power calculations.
So what happened? Which strategy reigned supreme?
Well, before we get to that there are a few technical factors and confounders to consider with this trial.
Firstly, it was sponsored by Astra-Zeneca, the makers of Symbicort. So this is a pharma-paper. Having said that there does seem to be a good separation of “church ‘n state” when it comes to the actual conduct of the trial. It was run by New Zealand’s Medical Research Institute independently.
Second. This is an open-label trial. The patients knew what they were taking and that may induce bias. The researchers went to some length to discuss the rationale and the practical considerations behind this design choice. Indeed to it is very tricky to do a double-blind, double-dummy placebo trial with three arms. However, bias remains inescapable in this design.
If you are a suspicious, old sceptic like my friend Dr Michael Tam then you will recall that there was another trial looking at long-acting beta-agonists (salmeterol) very ‘similarly’ named “SMART”… it was published in Chest 2006. That trial was remarkable in that there was an alarming number of severe exacerbations and deaths in the LABA group. Now, it could be a coincidence, however, if I were trying to wallpaper over some unpleasant evidence it would make sense to use a similar title to occupy the Google searches!
Lastly, the power calculation required 225 patients in each group to make the numbers work… but they didn’t quite make it. 13 were lost to follow up. So it was not mathematically pleasing to the stats nerds out there.
Ok, enough nerdiness. Onto the bottom line. what worked? Here is the table that tells the tale:
OK. So glancing at the tables as published it looks like strategies 2 and 3 are significantly better than the “albuterol as required” group for the PRIMARY OUTCOME. The write-up shows a 50% relative risk reduction – which sounds very impressive… until you look at the scale on the vertical axis.
In absolute terms, we are seeing well under a half of an “exacerbation per patient, per year”. So another way to state that outcome is that folk receiving PRN albuterol had an exacerbation every ~ 2.5 years. The folk in the other groups had one every 5 years roughly. These exacerbations were largely on the mild end of the spectrum. Overall these are very well people and really did not need much treatment at all!
I am not sure if I would want to take either a regular inhaled steroid or a PRN combination agent for that long to prevent a single exacerbation. It just seems like overkill in my head.
Whenever I read a drug company sponsored trial I have a simple method to assess the validity of the data in the real world. I think to myself: “what is the common side effect of this drug?” and “how often would I expect to see it in this population?” Then I scroll to the side-effects / adverse events section and check to see if it seems to be in the ballpark. For this trial I thought to myself: inhaled steroids can cause oral thrush. The company’s own PI leaflet says it occurs in 1 – 10% of patients. However, in this trial, the rate of oral thrush was not reported. This means it was either not seen or not counted. For me, this makes me less confident in the external validity of this data.
So for me, the BOTTOM line for this data is: patients with very mild asthma probably will do great with whatever strategy you like. In my mind, I would keep it simple and use a single agent without many side effects. Ventolin PRN would be what I would use. If there is evidence of ongoing inflammation, then add in a steroid to treat the underlying inflammation.
BEWARE the use of relative risk numbers to make a drug look good. Always look at the absolute figures and ask yourself: is it enough bang for the buck?
I am a GP working in Broome, NW of Western Australia. I work as a hospital DMO (District Med Officer) doing Emergency, Anaesthestics, some Obstetrics and a lot of miscellaneous primary care. Also on the web as @broomedocs | + Casey Parker | Contact