COMBINATION THERAPY
FOR CHILDHOOD ASTHMA:

INHALED CORTICOSTEROID AND LONG-ACTING BRONCHODILATORS (LABAs).

SALMETEROL AND FLUTICASONE

Greening et al. 1994 first showed that the addition of salmeterol to beclomethasone 200mcg twice per day gave better asthma control than doubling the dose of beclomethasone as judged by improved peak flows and nocturnal symptoms. The observation was confirmed by Murray et al. 1999   van Noord 1999 Woolcock 1996 Kelsen 1999 meta-analysis by Shrewsbury et al. 2000 D'Urzo et al. 2001 and  Ind 2003

All of those publications showed that the addition of a LABA to inhaled corticosteroid improved lung function ( FEV1 and peak flows), symptoms and use of rescue short-acting bronchodilator. None of them showed a decrease in exacerbations but none showed an increase in exacerbations. Initially there was concern that the addition of a LABA, because of the powerful effect on lung function and symptoms could mask deterioration in asthma (increased inflammation) when the dose of steroid was too low and therefore lead to worsening exacerbations. This did not occur. The studies with the combination of salmeterol and inhaled steroid did not result in increased rates of exacerbations compared to steroid alone but did not suggest that there was a decrease in exacerbations. Later studies with a combination of formoterol and inhaled corticosteroid (discussed below) did show a significant decrease in exacerbations compared to inhaled steroid alone. This has never been shown with salmeterol. The concern regarding masking of incomplete suppression of airway inflammation by inhaled steroid still exists and is discussed in a later section. Go to Masking  The improvement in lung function and symptoms with the addition of salmeterol to inhaled corticosteroid suggested the possibility that either salmeterol had anti-inflammatory properties or that the salmeterol was enhancing the anti-inflammatory effect of inhaled steroid.

ANTI-INFLAMMATORY PROPERTIES OF LABAs:

Studies in on cells in vitro and in animal models suggested that the addition of salmeterol might enhance the anti-inflammatory effect of steroid.  Eickelberg 1999 Adcock Review 2002 Johnson Review 2002 

Studies in human subjects reporting anti-inflammatory effects of LABAs are few in number. Dente 1999 found that treatment with salmeterol prevented an increase in airway eosinophilia following allergen challenge. However this is the only study to report an effect of salmeterol on eosinophilic inflammation.

Jeffery 2002 reported that compared with placebo, salmeterol significantly reduced the numbers of neutrophils in bronchial biopsy samples. They did not find an effect on eosinophils by salmeterol although they reported that treatment with fluticasone significantly reduced eosinophils compared to placebo.

Other studies have reported additive effects of salmeterol on suppression of different inflammatory markers in subjects being treated with inhaled corticosteroid.  Investigators from Monash University in Melbourne Li 1999 reported that after treatment with salmeterol there was a significant fall in EG1-positive eosinophils in the lamina propria (from a median 18.3 to 7.6 cells/mm, p = 0.01), which was not seen after treatment with fluticasone. The same group reported that airways of subjects with asthma had a significant increase in the number of vessels/mm2 of lamina propria compared with airways of normal subjects (524 +/- 137 vessels/mm2, n = 34 versus 425 +/- 130 vessels/mm2, n = 28; p = 0.004). There was a decrease in the density of vessels of lamina propria after treatment only in the salmeterol group compared with baseline (before, 535 +/- 153 vessels/mm2 versus after, 400 +/- 142 vessels/mm2; n = 12; p = 0.04). There was no significant change within the fluticasone (n = 11) or placebo (n = 11) treatment groups.  Orsida 2001 

In contrast to the studies cited above, the majority of studies have actually failed to show any effect of salmeterol on inflammation in the airways either as sole treatment or as add-on therapy with inhaled corticosteroid. One of the earliest studies to examine this question used bronchoalveolar lavage to examine the cell profile in asthmatics treated with beclomethasone or beclomethasone plus salmeterol. While they found no deterioration of the cell profile in the BAL, they could not find any effect of the addition of salmeterol to the beclomethasone on the markers of inflammation in BAL. They also demonstrated the improved clinical parameters with the addition of the LABA to the beclomethasone.Gardiner 1994

Kraft et al. 1997  examined the effect of salmeterol as monotherapy on night symptoms in asthma and found that salmeterol improved the number of nocturnal awakenings and decreased supplemental 24-h beta 2-agonist use in nocturnal asthma without significantly altering lung function and airway inflammation as assessed by bronchoscopy with alveolar lavage.

Roberts et al. 1999  studied the effect of 6 weeks of treatment with salmeterol on inflammation in subjects with stable atopic asthma in a double blind placebo-controlled parallel group study. The subjects underwent bronchoscopy with bronchoalveolar lavage (BAL) and bronchial biopsy immediately before starting treatment and again after 6 weeks. They found that regular treatment with salmeterol improved clinical indices of asthma but had no effect on the underlying inflammatory process.

Boulet et al. 2001 examined inflammatory cells in bronchial mucosa 24 hours after allergen challenge in subjects who had been treated for 7 days with salmeterol or placebo. Compared to placebo, 1 week of regular use of salmeterol was associated with an increase in bronchial inflammatory cells 24 h after allergen challenge.

Calhoun et al.  2001 used segmental allergen challenge to assess the effect of salmeterol on airway inflammation in bronchial alveolar lavage at 5 minutes and 48 hours after challenge in a double-blind placebo controlled trial. They concluded that salmeterol therapy alone did not reduce airway inflammation induced by allergen challenge, but also did not result in increased inflammation.

Lindqvist et al. 2003 compared the effect of salmeterol versus fluticasone versus cromoglycate versus placebo on the control of newly diagnosed asthma using bronchial biopsy. Salmeterol showed no anti-eosinophil effect despite giving good symptom relief. Fluticasone provided the best anti-eosinophil properties and symptom relief of the studied compounds.

These studies strongly suggested that although salmeterol was capable of giving symptom relief in asthma, it had no anti-inflammatory effect and should not be used as monotherapy in the absence of anti-inflammatory treatment.

ANTI-INFLAMMATORY EFFECT OF FORMOTEROL

Wallin et al. have compared the anti-inflammatory effect of treatment with formoterol vs. budesonide vs. placebo for 8 weeks using biopsy and BAL in a parallel group study.1999 Generally these patients, who had mild persisting asthma treated prior to the study only with a short-acting bronchodilator, had very few inflammatory cells in the biopsies and the BAL. However they found that in a subset of the patients with higher numbers of inflammatory cells in their biopsies (mast cells, eosinophils and T lymphocytes), that formoterol significantly reduced the numbers of mast cells and eosinophils but not the T cells compared to pretreatment values but not compared to the placebo. The budesonide reduced each of the cell types significantly, especially activated T cells. These data suggested that formoterol does not have a pro-inflammatory effect and might even have some anti-inflammatory effect through reduction of mast cells and eosinophils. The authors suggested that “Our finding that regular formoterol is able to reduce the baseline tissue eosinophilia might provide a mechanism for the protective effect of formoterol in the long-term management of asthma, beyond its capacity to functionally antagonize bronchoconstrictor mediators. However, with our present understanding, we wish to emphasise that long-acting b2-adrenoceptor agonists should still be used only in combination with a topical corticosteroid, as recommended in current guidelines, and should not be prescribed as a sole therapy.” More recently the authors examined the same biopsy material to determine whether the formoterol could be suppressing eosinophils by an effect on pre-formed T cell cytokines IL-4 amd IL-5 and adhesion molecules. They showed that while budesonide suppressed the IL-4 and IL-5 as well as other cytokines and adhesion molecules, formoterol did not. Wilson et al. 2001 Wallin et al. 2002  At present it is not clear how the formoterol achieved the suppression of mucosal eosinophils nor for that matter whether the observation would be reproducible in another set of subjects.

Researchers at the University of Dundee, have published two studies on small numbers of patients examining the effect or formoterol or budesonide on indirect markers of inflammation in asthma including Adenosine monophosphate (AMP) bronchial challenge, exhaled nitric oxide (NO), and serum eosinophilic cationic protein (ECP). They found that budesonide caused a reduction in each of these markers while formoterol did not. Aziz et al 2000 Lee et al. 2003  However these markers are indirect and do not clearly prove that formoterol has no anti-inflammatory activity.

At present it cannot be definitively stated whether formoterol has or does not have some anti-inflammatory effect on airway eosinophils.

CLINICAL STUDIES WITH SALMETEROL MONOTHERAPY

SMART STUDY A similar conclusion can be drawn from the cancellation of a study on the safety of Serevent (salmeterol) in the management of chronic asthma. In the SMART study, Serevent was added to a subject's ongoing asthma management; however, many of the subjects were not on maintenance inhaled steroid despite symptoms of chronic asthma. An interim analysis of the data suggested that some of those patients were at risk for severe asthma exacerbations, possibly even resulting in death. SMART CANCELLATION

STUDIES WITH FORMOTEROL AND BUDESONIDE

FACET TRIAL:

Pauwels et al. 1997 After a four-week run-in period of treatment with budesonide (800 mg twice daily), 852 patients being treated with glucocorticoids were randomly assigned to one of four treatments given twice daily by means of a dry-powder inhaler (Turbuhaler): 100mg of budesonide plus placebo, 100mg of budesonide plus 12mg of formoterol, 400mg of budesonide plus placebo, or 400mg of budesonide plus 12mg of formoterol. Terbutaline was permitted as needed. Treatment continued for one year. Outcome measures were lung function but also frequency of severe exacerbations. A severe exacerbation was defined as an increase in symptoms requiring treatment with oral steroid or a fall in peak flow readings greater than 30% from baseline on 2 consecutive days. The rates of severe and mild exacerbations were reduced by 26 percent and 40 percent, respectively, when formoterol was added to the lower dose of budesonide. The higher dose of budesonide alone reduced the rates of severe and mild exacerbations by 49 percent and 37 percent, respectively. Patients treated with formoterol and the higher dose of budesonide had the greatest reductions — 63 percent and 62 percent, respectively. Symptoms of asthma and lung function improved with both formoterol and the higher dose of budesonide, but the improvements with formoterol were greater.

This study was very important because it indicated that the addition of a LABA to inhaled corticosteroid, at least the addition of formoterol to budesonide, was capable of not only improving symptoms and lung function but also reducing the rate of asthma exacerbations. It should also be noted that the greatest improvement occurred when these patients, who were stable after a run-in of four weeks on high dose inhaled steroid were placed on 400 mcg bid of budesonide plus formoterol. When the patients were treated with low dose inhaled steroid plus formoterol, the patients improved compared to low dose inhaled steroid alone but not as much as doubling the dose of inhaled steroid alone. Since inhaled steroid is only thought to suppress eosinophilic inflammation, this suggests that the low dose inhaled steroid had failed to completely suppress the inflammation in these patients. Adding formoterol to low dose inhaled steroid, while giving improvement in asthma control would be unlikely to suppress that inflammation since from the studies cited above, at least for salmeterol, there is little evidence that the LABAs act by enhancing suppression of inflammation in the airways. That would suggest that, in these subjects, adding formoterol to low-dose inhaled steroid would improve asthma control but mask continuing airway inflammation.

As will be discussed, the authors attempted to study this possibility using the technique of induced sputum but I think that it is critical that the appropriate dose of inhaled steroid needs to be established before the addition of a LABA. The authors state in the last paragraph of the manuscript “Our results support therapeutic guidelines that recommend the addition of a long-acting inhaled b2-agonist to low doses of inhaled glucocorticoids in patients with persistent symptoms of asthma or less than optimal lung function. Increasing the maintenance dose of inhaled glucocorticoids might be a more appropriate initial therapeutic step in patients with repeated severe exacerbations of asthma.”

OPTIMA TRIAL:

This study O'Byrne 2001 - full text  was intended to examine treatment of mild asthma by assessing the effects of adding a long-acting inhaled beta-agonist, formoterol, to low doses of an inhaled corticosteroid, budesonide, for 1 yr in subjects with mild asthma, receiving no or only a small dose of inhaled corticosteroid. The 698 corticosteroid free patients (Group A) were assigned to twice daily treatment with 100mcg budesonide, 100mcg budesonide plus 4.5mcg formoterol, or placebo. The 1,272 corticosteroid-treated patients (Group B) were assigned to twice daily treatment with 100mcg budesonide, 100mcg budesonide plus 4.5mcg formoterol, 200mcg budesonide, or 200mcg budesonide plus 4.5mcg formoterol. The main outcome variables were time to the first severe asthma exacerbation and poorly controlled asthma days. In Group A, budesonide alone reduced the risk for severe exacerbations by 60% and poorly controlled days by 48%; adding formoterol increased lung function with no change in other end points. By contrast, in Group B, adding formoterol reduced the risk for the first severe exacerbation and for poorly controlled days by 43 and 30%, respectively. Thus, in corticosteroid-free patients, low dose inhaled budesonide alone reduced severe exacerbations and improved asthma control, and in patients already receiving inhaled corticosteroid, adding formoterol was more effective than doubling the corticosteroid dose.

It is clear from this study that recently diagnosed asthma that is steroid naive, does not benefit from the addition of a LABA. It is not mentioned in the publication but there was a mean of 8 years difference in the length of time that Group B subjects had asthma (the group that benefited from the addition of LABA) compared to Group A (the group that showed no benefit). This suggests that newly diagnosed mild asthma should be treated with inhaled steroid only and not started on the combination therapy at the outset. The results also suggested that the addition of a LABA did not influence inflammation since it is hard to imagine why the inflammation would be different in mild asthma untreated versus mild asthma treated with inhaled steroid. In fact the difference in the response of the two groups suggests that by virtue of the length of time that they had asthma, Group B subjects had undergone airway remodeling. After they were treated with inhaled corticosteroid they likely had continuing symptoms because of airway hyperreactivity and as a result showed benefit from the addition of the formoterol (LABA). The question remains, if Group A takes the inhaled steroid faithfully, will that prevent the airway remodeling and therefore the need for a LABA over time?

MASKING:

It has been suggested that regular use of long-acting beta-agonists could delay recognition of (“mask”) increasing airway inflammation. The possibility of masking of inflammation by fomoterol was examined in a separate study of a smaller group of patients. Kips 2000 This was a difficult study that claimed there was no evidence for masking of inflammation by the addition of formoterol to budesonide. The inflammation was assessed in 70 asthmatic patients using the technique of induced sputum. Initially the patients were treated with 1600 mcg of budesonide daily for a one month run-in. This treatment significantly reduced the median sputum eosinophils from 4.5 to 0.68%. The subjects were then treated for 1 year with either budesonide 200 mcg plus formoterol or budesonide 800 mcg without formoterol. There was no significant change in the median percent eosinophils in either group from the < 1% at entry. However a closer reading of the study suggests that it may not have had the power to detect a difference. When the subjects were randomized to the two treatment groups the median percent eosinophils in the budesonide 800 mcg was 6.45% compared to 1.88% in the budesonide plus formoterol group. These were not significantly different because of the spread of individual sputum results but there were clearly more patients with a normal sputum eosinophil count i.e. < 2% in the budesonide formoterol arm. At entry after the run in on high dose budesonide, the medians were 0.88 vs. 0.6% respectively. During the treatment period, the sputum median % eosinophils in the budesonide 800 mcg group was 1.74% while the budesonide plus formoterol rose non-significantly to 3.41%. Although these numbers were said to be non-significant, in the absence of a power calculation it is not possible to determine whether the number of patients studied was sufficient to demonstrate a significant difference. In fact at face value, the treatment arm with low dose budesonide 200 mcg and formoterol may have allowed the median percent eosinophils to rise from the initial 1.88% prior to run-in and 0.6% at entry to treatment to 3.41% which with a larger number of subjects might have been significant. This study does not answer the question surrounding the possibility of masking with a low dose inhaled steroid plus a LABA.

A better study examining masking with salmeterol was published by  Mcivor et al. 1998  They studied a small number of asthmatic patients requiring maintenance high dose inhaled steroid. Exacerbations were induced by step-wise reduction in the dose of inhaled steroid. This is a technique described previously by these authors to induce a mild exacerbation. Gibson et al. 1992 It was found that the patients who were treated with salmeterol became symptomatic during reduction of inhaled corticosteroid with higher levels of eosinophils in their sputum compared to patients treated with placebo. Mean corticosteroid dose was reduced by 87% during salmeterol treatment, versus 69% with placebo (p = 0.04) before onset of symptoms. Sputum eosinophils increased before exacerbation despite stable symptoms, FEV1, and PEF. In the week before clinical exacerbation, sputum eosinophil counts were higher in the salmeterol-treatment arm (19.9 +/- 29.8% [mean +/- SD], versus placebo 9.3 +/- 17.6%; p = 0.006). Five subjects showed > 10% sputum eosinophilia before exacerbation during salmeterol treatment, as compared with two receiving placebo. The investigators concluded that “ the bronchodilating and symptom-relieving effects of salmeterol can mask increasing inflammation and delay awareness of worsening asthma.” The study has been criticised because of the small number of patients involved but it clearly demonstrates that masking can occur although it does not address the question of the importance of this phenomenon in clincal practice.

STUDIES IN CHILDREN

Langton Hewer 1995 studied the addition of salmeterol or placebo to ongoing treatment with inhaled corticosteroid in 24 children aged 12 - 17 years. There were significant improvements in morning and evening peak flow rates and spirometry when measured on four occasions during the study period.

von Berg 1998  performed a 12 month, multicentre, double-blind, randomized, parallel study, in which 426 asthmatic children aged 5-15 years old received salmeterol 50 micrograms b.i.d. or placebo b.i.d. via the Diskhaler. They had access to salbutamol as needed. At the end of 12 months of treatment with salmeterol, the adjusted change from baseline for morning and evening peak expiratory flow rate (PEF) was 56 and 47.l min-1, respectively, and this was significantly greater than placebo (P < 0.01; P < 0.05). Exacerbation rates did not differ between groups and results were not dependent upon concurrent inhaled steroid use. Neither treatment caused a change of > or = 1 doubling dose in PC20/PD20 either during or on stopping treatment.

Verberne 1998 studied 177 children with moderate asthma being treated with 400 mcg beclomethasone per day. They were randomised to three groups and treated blindly with the addition of salmeterol 50 mcg twice daily, beclomethasone 200 mcg bid or placebo bid. They found that adding salmeterol or doubling the dose of beclomethasone gave no additional benefit over that from 200 mg beclomethasone twice daily in this selected group of children with moderate asthma, in which the compliance with medication was excellent. They hypothesised that in children the inflammatory changes in the airways might be more reversible by inhaled corticosteroid treatment than they are in adults, who typically have long-standing asthma. In fact the OPTIMA study suggests that even in adults with recent onset asthma (OPTIMA group A), the response to inhaled steroid is not enhanced by the addition of a LABA.

Akpinarli 1999 performed a small study with 32 symptomatic asthmatic children with moderate to severe asthma who continued their regular treatment with inhaled corticosteroid but half of whom received the addition of formoterol in a double-blind-placebo controlled fashion. They reported that the additon of formoterol for six weeks decreased symptom scores, PEFR variability, the number of rescue salbutamol doses while increasing morning and evening PEFR significantly.

Tal et al. 2002 reported results of a double-blind, double-dummy, randomized, parallel-group, multicenter trial. After a 2-4-week run-in period, 286 asthmatic children (177 boys, 109 girls; mean age, 11 years; mean forced expiratory volume in 1 sec (FEV(1)), 75% predicted normal), previously treated with inhaled corticosteroids (average dose 548 mcg/day), were randomized to 12 weeks' treatment with either budesonide/formoterol 80/4.5 mcg, two inhalations twice daily (n = 148), or an equivalent dose of budesonide 100 mcg, two inhalations twice daily (n = 138). They found that morning PEF (primary variable) and evening PEF increased to a significantly greater extent with budesonide/formoterol than with budesonide alone. They also found no difference in symptom scores, exacerbations and use of rescue medication between the two groups. This study raises the question as to whether the slight increase in lung function is worth the addition of the LABA to the inhaled corticosteroid in young children.

Bensch 2002 examined the effectiveness of inhaled formoterol over a period of 12 months in asthmatic children who were still symptomatic despite anti-inflammatory treatment. After a run-in period, 518 patients (5 to 12 years old) were randomized in a double-blind manner to receive 12 or 24 mcg formoterol dry powder or placebo twice daily for 12 months in addition to their anti-inflammatory treatment. The primary outcome variable was the area under the curve for forced expiratory volume in 1 second measured over 12 hours after the morning dose of study medication. The area under the curve for FEV1 after the first dose of treatment and after 3 and 12 months of treatment was significantly greater for patients receiving formoterol 12 mcg and 24 mcg than for patients receiving placebo (all P < or = 0.0062). Compared with placebo, both doses of formoterol significantly improved morning and evening premedication peak expiratory flow rate (all P < 0.001). In the group treated with formoterol 24 mcg, median symptom score and median dose of rescue medication at night were lower than during the run-in period, whereas the opposite occurred in the placebo group. HOWEVER the incidence of hospitalizations for asthma was higher in the formoterol groups than in the placebo group. These results suggest that not only was there no decrease in the rate of exacerbations but there may have been some masking of symptoms so that breakthrough symptoms were severe enough to require hospitalization.

In our own 12 week study of 302 children aged 6 - 11 years 2004  who continued their maintenance dose of inhaled steroid and added formoterol or placebo. Zimmerman, D'Urzo, Berube: Pediatric Pulmonology 37:122–127 (2004) we found that there was an improvment in morning and evening peak flow rate and FEV1 but no change in the days to first exacerbation or quality of life measurements. The exacerbation rate was low in all of the children even though they were entered on the basis of continuing symptoms and slightly low spirometry. We suggested that the children were not symptomatic enough and the study was too short to detect a change in exacerbations but it is also possible that young children do not require anything beyond treatment with inhaled steroid much as the OPTIMA group A adults. The improvement in lung function demonstrated by the Tal study and our own while statistically significant may not be clinically significant since the increase was less than 10% in all cases. Neither the Tal study nor our own showed a decrease in exacerbation rates but both studies were only 12 weeks in length and might have been too short to assess differences in exacerbations since the studies in adults (FACET and OPTIMA) were year-long studies. In contrast the Bensch study was year long and failed to show improvement in exacerbation rate with the addition of formoterol to inhaled corticosteroid and in fact suggested there was an increase in hospitalization rates for those patients treated with the addition of formoterol.

Bisgaard 2003 analysed the literature on the addition of LABAs to inhaled corticosteroid in children to determine if the addition of a LABA had any effect on exacerbations and hospitalizations. Eight studies were identified in which an exacerbation was defined as deterioration in a patient's asthma requiring a change in prescribed medication or not defined but reported as an asthma exacerbation or an asthma-related hospitalization. Analysis of data from the eight studies revealed no apparent protection from an asthma exacerbation among children on a LABA compared to patients on comparator treatment. He concluded that the “lack of evidence for the control of asthma exacerbations in children regularly using a LABA should bring into question its general use as add-on therapy”.

CONCLUSIONS:

CONTROVERSY RESULTING FROM THE SMART STUDY:

There has been continuing controversy over whether there is a risk for increased serious acute episodes of asthma and even a risk for death from asthma when long-acting bronchodilators are used in patient management.

There is clearly a risk for serious episodes and death when the labas are used as monotherapy but the FDA has issued a warning for increased risk even when inhaled steroids are used with the laba FDA and FDA.

This led to Martinez writing an editorial in the New England Journal of Medicine entitled "Safety of Long-Acting Beta-Agonists - An Urgent Need to Clear the Air. New Eng J Med 353:25 2637, Dec 22, 2005.In that editorial, he actually suggested that "in mild-moderate asthma, inhaled corticosteroids should be used in sufficient amounts to control chronic symptoms. If symptoms cannot be controlled in this way, some such patients may also benefit from the addition of Leukotriene antagonists or low-dose theophylline therapy. With adequate doses of inhaled corticosteroids and other treatments, long-acting beta-agonists should not usually be needed."Martinez 2005. Martinez apparently served on the FDA Advisory Committee that met in July 2005 regarding long-acting beta-agonists. He was clearly impressed by the data that he examined in committee. His recommendation in the editorial contradicts various guidelines on the management of asthma and led to a letter by Mintz, calling attention to the fact that the National Asthma Education and Prevention Program expert panel, preferred the addition of long-acting beta-agonists to Leukotriene antagonists as add-on medication when inhaled corticosteroids were insufficient to control asthma. He called attention to the fact that there was no evidence that when salmeterol is combined with inhaled steroid, there were any increase in deaths. Mintz   Martinez replied "Dr. Mintz’s arguments are based on a false premise: that the increased risk of severe side effects of long-acting beta-agonists does not occur in patients taking inhaled corticosteroids. A thorough analysis by the Food and Drug Administration concluded that the data from the SMART study do not allow any definitive conclusions with regard to this issue."

Wolfe and Lurie published a letter in the Lancet (reprinted on the web as Misleading data) claiming that GlaxoSmithKline had never published the SMART data and had submitted an analysis to the FDA that minimized the deaths and severe reactions by including post-study data. However the SMART study has now been published in the journal Chest and is freely available in its entirety. Nelson, Weiss, Bleecker et al 2006 This publication reiterated that the deaths seemed to occur primarily in African-American subjects and there are numbers of possible reasons for this outcome.

META-ANALYSIS BY SALPETER ET AL
A Meta-Analysis: Effect of Long-Acting Beta-Agonists on Severe Asthma Exacerbations and Asthma-Related Deaths was published by Salpeter, et al. and is available in its entirety on the web. meta-analysis 2006 These authors selected randomized, placebo-controlled trials that lasted at least 3 months and evaluated long-acting Beta-agonist use in patients with asthma. Pooled results were obtained from 19 trials with 33 826 participants and indicated that long-acting Beta-agonists increased exacerbations requiring hospitalization (OR, 2.6 [95% CI, 1.6 to 4.3]) and lifethreatening exacerbations (OR, 1.8 [CI, 1.1 to 2.9]) compared with placebo. Hospitalizations were statistically significantly increased with salmeterol (OR, 1.7 [CI, 1.1 to 2.7]) and formoterol (OR, 3.2 [CI, 1.7 to 6.0]) and in children (OR, 3.9 [CI, 1.7 to 8.8]) and adults (OR, 2.0 [CI, 1.1 to 3.9]). The absolute increase in hospitalization was 0.7% (CI, 0.1% to 1.3%) over 6 months. The risk for asthma-related deaths was increased (OR, 3.5 [CI, 1.3 to 9.3]), with a pooled risk difference of 0.07% (CI, 0.01% to 0.1%). They admitted that the analysis was limited by the small number of deaths which limited the reliability in assessing this risk. Furthermore many studies did not report information on the outcomes of interest.

In an editorial go to editorial accompanying the Saltpeter meta-analysis, Glassroth calls attention to basic research studies that found polymorphisms at several positions within the human Beta2-adrenergic receptor gene. Asthmatic patients homozygous for a variant with an arginine at the 16th amino acid position of the Beta2-adrenergic receptor (the Arg/Arg genotype) may actually experience declines in airflow and worsening asthma control when using Beta-agonists to treat their asthma Israel 2000. This genotype occurs in approximately one sixth of the U.S. population and may be more common in some racial or ethnic groups, such as African-American persons Drysdale 2000. Glassroth suggested that this phenotype may be important in the conflicting results seen in different studies. The SMART trial had 18% African-American subjects and they may have accounted for most of the adverse effects seen with the LABAS. Glassroth recommends a very conservative approach to the use of combined inhaled steroid and LABAs. "First, it is important to follow current guidelines that emphasize the use of inhaled corticosteroids as the first line of treatment for patients with mild to moderate persistent asthma symptoms. Physicians should not use LABAs as initial therapy for any asthmatic patient. Second, physicians caring for patients who do not achieve at least good control (defined by recent guidelines as minimal daily or nocturnal symptoms and infrequent exacerbations requiring systemic corticosteroids or emergency departments) ...should first escalate the dose of inhaled corticosteroids to achieve asthma control. If satisfactory control is not attained at maximal doses of inhaled corticosteroids, then a LABA should be added." However it should be noted that a recent study has tested the hypothesis that the arg/arg phenotype might influence treatment effects with salmeterol and did not find any effect on patient response to salmeterol and fluticasone. Bleecker 2006

The SALPETER META-ANALYSIS has been heavily criticized for relying too heavily on the SMART study which has been one of the few studies powered to examine risk of severe episodes and death. However the SMART study was not rigorous in the use of inhaled steroid. The subjects continued the inhaled steroid they had been taking at study entry but it is not clear whether they maintained the steroid use throughout the study. The SMART study was discontinued early because of the apparent risk for severe episodes and even death especially among the African-American subjects. Nelson and Dorinsky letter 2006 disagree with the meta-analysis conclusion that there is a risk for increased deaths and severe exacerbations even when inhaled steroids are combined with the LABAs. They base their disagreement on the fact that 80% percent of the subjects in the metaanalysis were from the SMART study where the use of steroid was not tightly controlled. Moreover other large studies have demonstrated that inhaled steroid plus LABA reduces exacerbations compared to increased inhaled steroid and the rate of death from asthma has been declining even as the sales of Combined inhaled steroid and LABA have increased.

Ernst el al writing for the Canadian Guidelines committee on management of asthma, takes issue with the Salpeter meta-analysis. Ernst 2006 for the following reasons:

1. In 12 of 19 studies selected for the analysis, there was no requirement for inhaled steroid during the study periods.
2. In the SMART study, the Salmeterol was provided but the inhaled corticosteroid was not, so that there was no incentive to maintain the use of the inhaled corticosteroid.
3. The Salpeter meta-analysis differs sharply from two Cochrane reviews examining the risk for severe exacerbations requiring hospitalization in patients taking long-acting Beta-agonists in addition to inhaled corticosteroid. In these studies, the inhaled corticosteroid was definitely maintained. An excess of hospital admissions was not observed with the addition of LABAs to ics.
4. Many "pivotal studies of concurrent therapy" were not included in the Salpeter meta-analysis because they were not placebo-controlled.

Glassroth responded to Ernst et al  Glassroth 2006 by citing their agreement that inhaled corticosteroid are the first-line management for asthma but disagreeing on the dose. He marshalled data supporting increased doses of inhaled steroid and felt that the literature did not support the contention of a significant increase in side-effects from increasing the steroid dose. He states "When all of these data are considered, I believe that the recommendation of ensuring that inhaled corticosteroids have been used to the maximum benefit before considering LABAs is reasonable. At the bottom line, however, we agree on the most important point: Asthma is a serious condition and it should be treated until controlled".

Many authors have written to journals criticising the meta-analysis by Salpeter et al. and strongly suggesting that the current guidelines which recommend early introduction of LABAS as additive medication even with moderate doses of inhaled steroid are correct and should be continued. However it is true that there has not been a study sufficiently large enough to answer the question of whether long-acting Beta-agents are safe when used in the presence of inhaled corticosteroids. Ernst et al. call for a meta-analysis to be performed with indivual patient data culled from the various trials of LABAa and ICS done to date.

STUDIES IN CHILDREN

As cited above, there is little data supporting the need for Long-acting Beta-agents in the management of childhood asthma. In fact there is data suggesting that the use of LABAS in the managment of childhood asthma might lead to an increase in severe exacerbations and hospitalizations. This has led Bisgaard to call for studies to be done specifically in children assessing exacerbations and not simply extrapolating from adults.Bisgaard 2003 and Bisgaard, Szefler, 2006.

More recently, Bisgaard has published the pediatric portion of a study designed to test the use of a combination of low dose budesonide/formoterol used as both maintenance and reliever medication in pediatric asthma. These patients were part of a larger study in adults, testing the same concept. Nevertheless, the pediatric portion of the study was designed to be analysed as a separate study. In this double-blind, randomized study for 12 months three arms were compared, a low fixed dose of budesonide/formoterol, a moderate fixed dose of budesonide alone, and a low dose of budesonide/ formoterol used as both maintenance and reliever (Symbicort, maintenance and reliever therapy - SMART). The other two arms used a short-acting bronchodilator as needed for symptom relief. The primary outcome variable was the time to first exacerbation and SMART prolonged the time to first exacerbation and significantly reduced the number of severe exacerbations compared to the other two arms. Growth in the children was 1 cm better for the SMART arm as well as the low-dose fixed combination arm compared to the higher dose inhaled steroid arm. There was an increased exacerbation rate and increased number of emergency room visits, hospitalizations and courses of oral steroid needed in the fixed low dose steroid/laba arm compared to the fixed dose of inhaled steroid alone (a higher dose than in the fixed steroid/laba arm). The dose of inhaled steroid in the fixed dose inhaled steroid/laba arm was clearly too low. In this study, to be eligible for randomization, the patient had to have used 8 or more inhalations of terbutaline in the last 10 days of run-in along with their usual inhaled steroid dose. That is practically daily use of short-acting bronchodilator which would indicate that the patients were not under control on constant inhaled steroid of 200 to 500 mcg/day (the inhaled steroid could be of any type). They were then randomized to the three arms and only the steroid arm had sufficient inhaled steroid to control the asthma and in that arm the results were not as good as in the SMART arm.

CONCLUSION:

• Generally children can be controlled with small or moderate doses of inhaled steroid. There is no need for the use of combination therapy in the majority of children.
• Although it is probably true that the combination of inhaled steroid with Laba mitigates the small threat from Laba seen in the SMART study, it has not been established formally.
• Combination therapy should never be used as initial therapy to manage childhood asthma partly because it is probably unnecessary but also because of the possible small risk.
• In the Bisgaard SMART study (an inapt choice of name given that the SMART trial in adults is the controversial study), the children appeared to be significantly out of control (using bronchodilator almost daily in the run-in) even while using regular inhaled steroid. Perhaps they were not taking inhaled steroid at all and the problem was compliance.
• If it is clear that an older child's asthma cannot be controlled on a moderate dose of inhaled steroid, and a compliance issue has been ruled out, then that child might be a candidate for add-on therapy and I would still tend to choose long-acting Beta-agents over leukotriene receptor antagonists much as recommended in the 2003 update of the Canadian Asthma Management Guidelines Summary of recommendations.