Search tips
Search criteria 


Logo of jrsocmedLink to Publisher's site
J R Soc Med. 2006 August; 99(8): 382–383.
PMCID: PMC1533517

Long-acting bronchodilators: time for a re-think

Asthma is a common chronic inflammatory disorder of the airways. It affects more than 5 million people, leads to around 70 000 hospital admissions and causes 1400 deaths every year in the UK alone. Despite a plethora of medical advances in this field over the last few decades the death rates show little improvement. Any further advancement in management is therefore both desirable and welcome.

Beta-agonists now form the mainstay of reliever treatment for asthma. Since their introduction there have been numerous reports showing an increase in asthma mortality and morbidity. Various explanations have been offered for these adverse effects:

  • toxicity of propellants causing bronchospasm
  • aggravation of hypoxaemia
  • cardiotoxicity secondary to hypoxaemia and hypokalaemia
  • increased amount of antigen on an unprotected airway
  • accumulation of toxic amounts of the distomer in racemic mixtures of sympathomimetics
  • tolerance/tachyphylaxis to the protective effects of beta-agonists.1

In contrast to short-acting beta-agonists, long-acting ones (LABA) allow twice daily administration, making them extremely appealing. As such, increased compliance was expected and decreased mortality and morbidity anticipated.

Originally, very potent full beta2-agonists like fenoterol were used, quickly raising concerns regarding their safety profile.2 The early 1990s witnessed the introduction of new, third-generation, highly-selective LABA that appeared to be safer than their predecessors. Salmeterol was such an agent. Shortly afterwards, and quite unexpectedly, an increase in asthma-related deaths was observed. The first randomized double-blind trial comparing symptoms control of salmeterol compared to salbutamol for 16 weeks, was soon reported with results that surprised many.3 It was suggested that patients on the salmeterol arm were almost three times as likely to die from asthma-related deaths during the trial period. However, as overall events were infrequent this trend did not reach statistical significance. A clearly worried Food and Drug Administration requested GlaxoSmithKline (GSK), salmeterol's manufacturer, to investigate this matter further. This led to the initiation of the Salmeterol Multicenter Asthma Research Trial (SMART) which was published earlier this year.4

SMART, a randomized, double-blind-controlled trial, recruiting patients older than 12 years of age, comparing the addition of salmeterol or placebo to existing asthma medication began in 1996. The primary endpoint was the outcome in combined respiratory-related deaths or life-threatening experiences and the secondary endpoint was asthma-related deaths or life-threatening experiences. Subjects were randomized to receive either salmeterol 42 mcg twice daily via a metered dose inhaler or placebo, in addition to their routine asthma medication and use short-acting beta2-agonists for symptomatic relief.

Interim analysis was scheduled when half the subjects had been recruited (26 355 patients) and provisions made to allow stopping the trial if pre-defined criteria were met.4 At this interim analysis, more patients on the salmeterol arm (37/13,176 versus 22/13,179 [confidence interval 1.01-2.89]) met the secondary endpoint criteria at a 95% level of significance. In addition, there were more overall respiratory-related deaths (24 versus 11 [95% CI 1.06-4.41]) and asthma-related deaths (13 versus 3 [95% CI 1.25-15.34]) in the total population for the salmeterol arm. Dividing the study populations into Caucasian and African-American, it appeared that African-Americans on salmeterol did worse, meeting both primary (20/2,366 versus 5/2,319 [95% CI 1.54-10.90]) and secondary endpoints (19/2,366 versus 4/2,319 [95% CI 1.68-14.45]). In view of this, the Drug Safety Monitoring Board suggested two alternatives:

  • increasing the sample by 10,000 subjects and completing the trial within 2 years or
  • stopping the trial prematurely and disseminating the results.5

GSK opted for the latter.

Using post hoc analyses, the role of inhaled corticosteroids (ICS) was also addressed, in spite of the fact that SMART was not designed to assess the effect of ICS in combination with salmeterol. Not surprisingly, those taking no ICS at baseline had a worse prognosis. If ICS were used, similar asthma-related deaths were recorded in the two arms (16/6,127 deaths in salmeterol and 13/6,138 in placebo—no statistical significance). However, when no ICS were used there was a significant difference in asthma-related deaths (21/7,049 in the salmeterol and 9/7,041 in the placebo [95% CI 1.10-5.2]) driven mainly by events in African-Americans.

Evaluating SMART, it becomes clear that salmeterol is not significantly better than placebo when added to existing medication for treatment of asthma. Furthermore, the study raised serious concerns that salmeterol could, potentially, be harmful particularly in African-Americans, adding to previous concerns about LABA. This probably represents a class effect, and other LABA, e.g. formoterol, should not be considered safer options.6

This is worrying enough to warrant changes.

First, we would recommend that the current worrying tendency of LABA monotherapy should be abandoned and that no patient is started on a LABA without ICS. (It is our personal experience that an increasing number of patients referred to asthma clinics find themselves on LABA monotherapy). This therapy is neither supported by evidence nor encouraged by the current British Thoracic Society (BTS) guidelines.

Secondly, there is currently provision in the BTS guidelines for LABA to be discontinued if they are not deemed useful after a trial period. We feel that this should be reinforced. Moreover, until a consensus is reached on how current evidence is incorporated into clinical practice, the European Medicines Agency should follow the FDA's example and warn about LABA's potential to increase severe asthma episodes and death.7

Thirdly, Step 3 of the BTS guidelines suggests the addition of a LABA if low-dose ICS fail to achieve appropriate control for patients older than 5. We feel that this should be revised. Adequate, higher doses of ICS should be tried first and, if not successful, then the introduction of leukotriene receptor antagonists, theophyllines or anticholinergic agents should be considered.8 To date, there is sufficient evidence to suggest that leukotriene receptor antagonists9 and theophyllines10 are as effective as LABA, without the increased complications seen with LABA during SMART. At the same time, there is no robust evidence supporting routine use of anticholinergic agents, which are currently included in international guidelines for asthma management. However, there are suggestions that they might benefit a sub-group of asthmatic patients and are safe.11

Patients with more severe disease should be evaluated by a specialist to exclude non-compliance or the presence of co-existing/other conditions labelled as ‘asthma’. With the current evidence in mind, and until the safety of LABA with adequate dose ICS is proven in clinical trials, we feel that LABA use should be reserved until other safer options have been utilized first. Finally, we recommend that if any patient has asthma severe enough to require LABA, then they should be managed by an appropriate specialist rather than in primary care.

We have seen LABA transforming the lives of some of our patients and improving their quality of life. However, the accumulating evidence against their safety should not be ignored and urgent actions are necessary to safeguard our patients.


Competing interests None declared


1. Costello JF. Sympathomimetic enantiomers in the treatment of asthma. Nashville: Parthenon, 1995
2. Crane J, Pearce N, Flatt A, et al. Prescribed fenoterol and death from asthma in New Zealand, 1981-83: case-control study. Lancet 1989;i: 917-22 [PubMed]
3. Castle W, Fuller R, Hall J, Palmer J. Serevent nationwide surveillance study: comparison of salmeterol with salbutamol in asthmatic patients who require regular bronchodilator treatment. BMJ 1993;306: 1034-7 [PMC free article] [PubMed]
4. Nelson HS, Weiss ST, Bleecker ER, Yancey SW, Dorinsky PM; SMART Study Group. The Salmeterol Multicenter Asthma Research Trial: a comparison of usual pharmacotherapy for asthma or usual pharmacotherapy plus salmeterol. Chest 2006;1: 15-26 [PubMed]
5. Lurie P, Wolfe SM. Misleading data analyses in salmeterol (SMART) study. Lancet 2005;366: 1261-2 [PubMed]
6. Dahl R, Chuchalin A, Gor D, Yoxall S, Sharma R. EXCEL: A randomised trial comparing salmeterol/fluticasone propionate and formoterol/budesonide combinations in adults with persistent asthma. Resp Med 2006;100: 1152-62 [PubMed]
7. Bisgaard H, Szefler S. Long-acting beta2 agonists and paediatric asthma. Lancet 2006;367: 287-8 [PubMed]
8. Martinez FD. Safety of long-acting beta-agonists—an urgent need to clear the air. N Eng J Med 2006;353: 2637-9 [PubMed]
9. Wilson AM, Dempsey OJ, Sims EJ, Lipworth BJ. Evaluation of salmeterol or montelukast as second-line therapy for asthma not controlled with inhaled corticosteroids. Chest 2001;119: 1021-26 [PubMed]
10. Shah L, Wilson AJ, Gibson PG, Coughlan J. Long-acting beta2-agonists versus theophyllines for maintenance treatment of asthma. Cochrane Database System Rev 2006; Issue 2
11. Vakil DV, Ayiomamitis A, Nizami RM. Use of ipratropium aerosol in the long-term management of asthma. J Asthma 1985;22: 165-70 [PubMed]

Articles from Journal of the Royal Society of Medicine are provided here courtesy of Royal Society of Medicine Press