Pharmacological Management to Reduce Exacerbations in Adults With Asthma:  A Systematic Review and Meta-analysis FREE

Asthma is common, affecting 5% to 12% of the adult population.^1- 2 In the United States alone, asthma affects more than 200 000 adults and accounted for approximately 465 000 hospitalizations, 1.8 million emergency department visits, 10.4 million physician office visits, and 4487 deaths in 2000.³ For largely unknown reasons, US women have a 30% higher prevalence and 40% more asthma attacks than men, and blacks have a 10% higher prevalence and 20% more asthma attacks than whites.³ The prevalence of self-reported asthma has increased 74% over the past 2 decades,² with a doubling in the number of office visits and a 61% increase in asthma-related deaths.² The current total annual costs of asthma are $11 billion.^2- 5

Asthma is characterized clinically by repeated episodes of wheezing, breathlessness, chest tightness, and coughing, usually in the presence of variable (and reversible) airflow obstruction.⁶ Most patients also demonstrate airway hyperresponsiveness on methacholine or histamine challenge tests.⁷ At the heart of asthma pathophysiology is chronic airway inflammation,⁸ with infiltration of eosinophils, mast cells, and CD4⁺ T lymphocytes that express T helper cell type-2 cytokines such as interleukins 4, 5, and 13, although some individuals (particularly those with very severe chronic asthma) have a predominance of neutrophils.^9- 10 Airway remodeling is another characteristic feature of chronic persistent asthma, which consists of smooth muscle hypertrophy, thickening of basement membranes, increased mucus production, and denudation of airway epithelium.^8,10 Although many individuals with asthma have environmental allergies and evidence of atopy, some do not.¹¹ Thus, a history of allergies and atopy is helpful but cannot be relied on exclusively for diagnosing asthma.

Over the past 20 years, the basic understanding of asthma and its pathogenesis has rapidly evolved, leading to the development of novel pharmacological therapies. These include inhaled corticosteroids, long-acting β₂ agonists (LABAs), agents that affect the leukotriene pathway, combination products, and monoclonal anti-IgE therapies. While all of these therapies improve lung function to a certain extent, their long-term effects on exacerbations are less clear.

Exacerbations are one of the most important (if not the most important) end points for clinical trials in asthma because they represent periods in which patients have the greatest risk of emergency department visits, hospitalization, and even death.¹² Additionally, asthma exacerbations impose enormous amounts of emotional and financial stress, reduce quality of life, and impair the ability to work. From a societal perspective, exacerbations are the leading category of expenditures related to asthma, accounting for almost 50% of total costs.⁴ Moreover, patients having frequent exacerbations (who account for approximately 20% of the total pool of those with asthma) incur 80% of the total direct costs of asthma.⁴ Prevention of exacerbations is, therefore, a central aim in asthma management.⁶

We decided a priori to examine the effects of inhaled corticosteroids, LABAs, leukotriene pathway modifiers/receptor blockers, combination therapy with inhaled corticosteroids and LABAs, and anti-IgE therapies, because they are the most commonly used pharmacological agents for the management of adult asthma. For each of these therapies, we conducted a literature search by using MEDLINE, EMBASE, and Cochrane databases. We limited the search to English-language articles published from January 1, 1980, to April 30, 2004, reporting studies of adults (>19 years of age) in randomized clinical trials. We contacted experts to ascertain any studies that may have been missed in our initial search. As the primary purpose of this review was to ascertain the long-term effects of these therapies on rates of exacerbation, we excluded studies that did not report on exacerbations or that had a follow-up period of less than 3 months. We used the Jadad scale to adjudicate the methodologic quality of the studies.¹³ We restricted the analysis to randomized clinical trials that had a score of 3 or more, complete or near-complete follow-up data, and baseline characteristics that were well balanced between the treatment and control groups. Crossover trials were excluded because these studies in general had inadequate follow-up, poor ascertainment of exacerbation data, and time-treatment interactions that were difficult to evaluate.¹⁴ We also excluded studies that were published in abstract form only, because the methods and the results could not be fully analyzed.

Data were abstracted from each trial by 2 authors (J.M., W.Q.G.) independently using a prestandardized data abstraction form. Any discrepancies were resolved by iteration and consensus. Because we did not have access to the original patient records, we accepted the definitions for exacerbation as used by the investigators in the original studies. Although there was some heterogeneity in the way in which exacerbation was defined across the studies, most defined exacerbation as an episode requiring oral or parenteral corticosteroids, emergency visits, hospitalization, or decrease in morning peak-flow measurements of greater than 25% to 30% on 2 consecutive days.

We excluded studies that defined exacerbations exclusively as episodes requiring increased use of short-acting β₂ agonists, because LABAs and other bronchodilators may decrease the need for short-acting β₂ agonists (which is expected since β₂ agonists are themselves bronchodilators) without attenuating requirements for systemic corticosteroids or emergency visits/hospitalizations. We did not include studies in which exacerbations were reported exclusively as part of the "withdrawal" data, because the definition of an exacerbation was usually not prespecified or explicitly stated. For the inhaled corticosteroid analysis, higher dose was defined as doses greater than 500 µg/d of beclomethasone equivalent and at least 2-fold higher than the inhaled steroid dose contained in the comparator therapy (eg, combination of inhaled corticosteroids and a LABA).

Where possible, for each end point, we combined the results from individual studies to produce summary effect estimates. We checked for the heterogeneity of data across individual studies using the Cochran Q test. If significant heterogeneity was observed (P≤.10), the DerSimonian and Laird random-effects model was used to pool the results together. In the absence of significant heterogeneity (P>.10), a fixed-effects model was used. To accommodate for differences in laboratory techniques of measuring values of forced expiratory volume in 1 second (FEV₁), we converted the absolute levels of FEV₁ for each study into a common unit by calculating standardized effect sizes. Standardized effect sizes were derived by dividing the mean differences in FEV₁ (from baseline to the end of the follow-up period) between those assigned to the investigational medication and those assigned to placebo for each study by the standard deviations (of the mean differences) from the studies.¹⁵ As a sensitivity analysis, a weighted mean-difference technique was also used. In all cases, the data from standardized and weighted mean-difference methods produced very similar results. All analyses were conducted using Review Manager (RevMan) version 4.2 (The Cochrane Collaboration, Oxford, England).

Randomized controlled trials have confirmed the initial observations that inhaled corticosteroids improve lung function and ameliorate patient symptoms (Figure 1).¹⁶ Overall, compared with placebo or a short-acting β₂ agonists, inhaled corticosteroids reduced clinically relevant exacerbations by nearly 55% (relative risk [RR], 0.46; 95% confidence interval [CI], 0.34-0.62; P<.001 for heterogeneity).^17- 32 Risk reduction for exacerbations was greatest in short-term studies (12 weeks' duration)^{19,22,26- 27} (RR, 0.34; 95% CI, 0.25-0.44), followed by medium-term studies (13-51 weeks' duration)^23- 24 (RR, 0.48; 95% CI, 0.17-1.38), and least in the long-term studies (≥52 weeks' duration)^{17- 18,20- 21,25} (RR, 0.55; 95% CI, 0.38-0.80). The size of the study, on the other hand, made little difference in the results. We divided the studies into tertiles based on the sample size. Large studies (>450 participants)^21,25,27 had an RR of 0.47 (95% CI, 0.31-0.73); medium-sized studies (170-449 participants),^19,22,24,26 an RR of 0.45 (95% CI, 0.25-0.81); and small studies (<170 participants),^{17- 18,20,23} an RR of 0.38 (95% CI, 0.15-0.97).

We evaluated the potential modifying effects of disease severity by dividing the studies into tertiles of mean FEV₁ values at baseline. The RR reduction of exacerbations was similar across the FEV₁tertiles (lowest FEV₁ tertile^22,26- 27: RR, 0.39; 95% CI, 0.27-0.57; middle tertile^19,21,24: RR, 0.51; 95% CI, 0.28-0.93; highest tertile^{17- 18,23,25}: RR, 0.43; 95% CI, 0.24-0.75). Several studies compared higher-dose therapy (defined as doses >500 µg/d of beclomethasone equivalent³³ and at least 2-fold higher than the comparator dose) with lower-dose steroid therapy. In these head-to-head comparisons, the use of higher-dose therapy was associated with fewer exacerbations compared with the lower dose (RR, 0.77; 95% CI, 0.67-0.89).^{18,25,28- 32}

Inhaled corticosteroids have salutary effects on FEV₁. Compared with placebo, they improved FEV₁ by approximately 330 mL (95% CI, 260-400) in the first 3 to 4 months of therapy (mean standardized estimate, 0.56 units; 95% CI, 0.45-0.66 units in favor of inhaled corticosteroids over placebo) (Figure 2).^{19,22,26- 27,34- 48} There was little improvement in FEV₁ relative to that achieved with placebo beyond the first 3 months of therapy. In the trials that had at least 6 months of follow-up, the overall improvement in FEV₁ (compared with placebo) was approximately 150 mL (95% CI, 70-23; mean standardized estimate, 0.19 units; 95% CI, 0.14-0.25)^{20- 21,24- 25,49} (Figure 2). These data suggest that the principal salutary effects of inhaled corticosteroids on FEV₁ occur within the first 3 to 4 months of initiation of therapy.

We found 13 studies (N = 3888)^50- 62 that evaluated the effects of LABAs⁶³ on health outcomes in asthma. Overall, compared with placebo, the use of LABAs was associated with a 25% reduction in exacerbations (RR, 0.75; 95% CI, 0.64-0.88; P = .43 for heterogeneity) (Figure 3). Compared with regular use of short-acting β₂agonists, the reduction in exacerbation was smaller (RR, 0.83; 95% CI, 0.67-1.05). As expected, LABAs increased FEV₁ compared with placebo (mean standardized estimate, 0.33 units; 95% CI, 0.24-0.42).

When LABAs were added to inhaled corticosteroids, exacerbations were further reduced in those who had persistent symptoms while taking low-dose corticosteroids (Figure 4). Compared with steroid monotherapy, the combination therapy with inhaled corticosteroids and LABAs was associated with a 26% reduction in exacerbations (RR, 0.74; 95% CI, 0.61-0.91; P = .07 for heterogeneity).^{25,28- 29,51,64- 72} The addition of LABAs to inhaled corticosteroids was associated with a lower exacerbation rate than was increasing (usually doubling) the dose of inhaled corticosteroids (RR, 0.86; 95% CI, 0.76-0.96; P = .65 for heterogeneity). Combination therapy did not confer an incremental benefit beyond that achieved with steroid monotherapy in patients who were previously naive to corticosteroids.²⁵ Thus, in general, LABAs should be reserved for patients who continue to have symptoms despite low-dose steroid monotherapy.

These salutary effects of LABAs, however, have to be balanced against their potential long-term adverse effects.^73- 74 Certain groups, such as blacks with asthma and individuals not taking regular anti-inflammatory therapies, may be particularly vulnerable.⁷⁵ Until more data are available, regular monotherapy with LABAs (in the absence of regular anti-inflammatory therapy) cannot be recommended at this time for most patients.

Leukotriene pathway modifiers/receptor antagonists are also effective in reducing exacerbation rates (Figure 5). Compared with placebo, leukotriene modifiers/receptor antagonists lowered exacerbation rates by 41% (RR, 0.59; 95% CI, 0.49-0.71; P = .44 for heterogeneity).^{26- 27,42,76- 82} They also had a salutary effect on morning trough FEV₁ (standardized mean difference, 0.25 units; 95% CI, 0.12-0.38). However, they were inferior to inhaled corticosteroids in reducing exacerbations (RR, 1.72; 95% CI, 1.28-2.31; P = .91 for heterogeneity for leukotriene pathway modifiers/receptor antagonists vs inhaled corticorticosteroids) and in enhancing FEV₁ (mean standardized difference of 0.56 units [95% CI, 0.45-0.66] and 0.25 units [95% CI, 0.12-0.38] for leukotriene modifiers/receptor antagonists compared with placebo). Notably, the mean age of the participants of leukotriene trials was lower than that of participants in trials of LABAs and inhaled corticosteroids. We identified 3 studies comparing the efficacy of leukotriene modifiers/antagonists and LABAs as adjunctive therapy for those already taking inhaled corticosteroids (n = 2662). There was some heterogeneity of data (P = .045). Overall, there was a nonsignificant trend toward improved efficacy of LABAs in reducing exacerbations (RR, 0.85; 95% CI, 0.70-1.03).^83- 85

Overall, with concomitant inhaled corticosteroid therapy, the use of recombinant monoclonal anti-IgE antibody was associated with a 45% reduction in exacerbations over the first 12 to 16 weeks of therapy (RR, 0.55; 95% CI, 0.45-0.66; P = .15 for heterogeneity).^86- 89 Even in the presence of inhaled corticosteroid reduction and withdrawal, the effectiveness of monoclonal anti-IgE antibody was retained over a 12-week period (RR, 0.59; 95% CI, 0.48-0.74). Because all of these studies were performed among patients with asthma who had allergy skin test results that were positive for at least 1 or 2 perennial (and common) allergens as well as elevated serum IgE levels (≥30 IU/mL), anti-IgE therapy cannot be recommended for patients with asthma who do not have these characteristics. Anti-IgE therapy also improved FEV₁, though the magnitude of the improvement was modest (weighted mean difference between anti-IgE therapy and placebo: 2.9% [95% CI, 1.3%-4.5%] of predicted FEV₁ in favor of anti-IgE therapy; or mean standardized difference of 0.18 units [95% CI, 0.07-0.29 units] in favor of anti-IgE therapy).

Currently, asthma is believed to be a disease of airway inflammation, caused by allergic sensitization of airways and accompanied by dysfunction of airway smooth muscle cells.^8,90 Corticosteroids are potent (but nonspecific) anti-inflammatory agents and, as such, appear to be the therapies most effective in controlling asthma symptoms and improving lung function.⁹¹ Since airway inflammation is present even in mild disease, inhaled corticosteroids are the first-line treatments for patients who need more than an occasional inhalation of short-acting β₂ agonists.⁶ In those with moderate to severe airflow impairment, higher-dose therapy appears to produce greater beneficial effects on the risk of exacerbations than does lower-dose therapy. However, the salutary effects on exacerbations should be balanced against their potential adverse effects. In a dose-dependent manner, inhaled corticosteroids have been associated with bone demineralization, osteoporosis, hip fractures, cataracts, glaucoma, skin bruising, and adrenal suppression,^92- 101 although some studies have not confirmed these associations.^102- 106

The clinical trials evaluated herein were too short and underpowered to determine whether inhaled corticosteroids do indeed cause these adverse effects. Because proper inhaler technique, use of a spacer, and mouth rinsing after each actuation significantly reduce the systemic absorption of corticosteroids, patients should be educated on these safeguards.¹⁰⁷ Whether inhaled corticosteroids reduce mortality in asthma is uncertain. Because, fortunately, asthma deaths are relatively rare in the western world, none of the clinical trials had sufficient power to detect this end point.¹⁰⁸ However, several observational studies have demonstrated a protective association between therapy with inhaled corticosteroids and asthma mortality in various populations and across different jurisdictions.^109- 111 It is unclear whether the provision of inhaled corticosteroids to patients who are taking oral corticosteroids after hospital or emergency department visits reduces the risk of relapses.¹¹² However, once oral corticosteroids are discontinued, such patients should receive inhaled corticosteroids.

In those patients whose disease remains out of control despite low-dose inhaled corticosteroid therapy, addition of a LABA appears reasonable. By themselves, LABAs have only a modest beneficial effect in reducing exacerbations. However, when given in combination with inhaled corticosteroids, the overall risk of exacerbation is reduced by approximately 26% compared with inhaled corticosteroid monotherapy. Combination therapy (with LABAs and low-dose inhaled corticosteroids) is slightly more effective than high-dose inhaled corticosteroid therapy. Monotherapy with LABAs, on the other hand, is in general best avoided because it is less effective than combination therapy or monotherapy with steroids.

For patients with mild airflow obstruction who are unwilling or unable to take inhaled corticosteroids, treatment with a leukotriene pathway modifier/receptor antagonist should be considered. These agents are less effective in reducing clinical exacerbations than monotherapy with inhaled corticosteroids. However, compared with placebo, they significantly reduce exacerbations by approximately 40%. Although these medications are generally safe and well tolerated, sporadic cases of Churg-Strauss syndrome have been associated with their use.^113- 117 Whether leukotriene pathway modifiers or receptor antagonists are directly responsible for this syndrome or whether these vasculitic cases resulted from withdrawal of corticosteroids (in response to the therapeutic effects of leukotriene pathway modifiers) is uncertain.¹¹⁷

The precise role of monoclonal anti-IgE antibody therapy in the management of chronic asthma is unclear. Short-term studies have demonstrated that these medications have salutary effects on exacerbations above and beyond those achieved by inhaled corticosteroids among patients with asthma who have allergy skin test results that are positive for at least 1 or 2 perennial (and common) allergens as well as elevated serum IgE levels (≥30 IU/mL). However, as the studies have been relatively short, the long-term effects of these therapies on lung function and, more importantly, on the clinical course of patients with asthma remain uncertain. As such, they cannot be routinely recommended for most patients with asthma.

Although this review has focused on pharmacological treatment of asthma, nonpharmacological interventions are often of value in the management of chronic stable asthma. In general, treatment for asthma in adults depends on the severity of the symptoms and lung-function measurements. Either FEV₁ or peak expiratory flows should be used to assess lung function and to guide therapy. In most circumstances, FEV₁ is preferred over peak expiratory flow because the latter is more effort-dependent, demonstrates greater intrasubject and intersubject variability, and is less sensitive than FEV₁ in detecting mild airway obstruction.¹¹⁸ Regardless of the device used, the overall aim is to control symptoms, preserve lung function, and maintain good quality of life using the minimum amount of medications.⁶ Education aimed at self-management is useful for most individuals with asthma.¹¹⁹ Control of environment, particularly in eliminating known allergens, is recommended.⁶ General measures to maintain a healthy lifestyle are strongly encouraged. Physicians should counsel their patients with asthma not to smoke, and to lose weight if overweight or obese, to improve their asthma control. Weight loss has been demonstrated to reduce symptoms and improve lung function as well as health-related quality of life in obese patients with asthma.^120- 121 A proposed scheme that integrates symptoms and lung function as a guide to therapy is shown in Table 1.

There were several important limitations to this systematic review. First, we could not evaluate whether current smoking status materially modified the effects of inhaled corticosteroids. Active smoking may directly or indirectly (through creation of oxidative reactive species) neutralize the anti-inflammatory effects of corticosteroids by inhibiting the recruitment and action of histone deacetylase, which is responsible for down-regulating expression of proinflammatory cytokines from various inflammatory cells, such as alveolar macrophages.¹²² A recent study indicates that active smoking markedly attenuates the efficacy of inhaled and oral corticosteroids in stable asthma.¹²³ Indeed, in active smokers, even 2 weeks of oral corticosteroid therapy did not lead to significant improvements in lung function or symptom scores.^124- 125 Moreover, we could not determine whether obesity, race, and other risk factors can modify treatment effects of various antiasthma medications. Large comparative studies are needed in each of these subgroups to address this important issue. Second, because the clinical trials included in this review were relatively short in duration, the long-term adverse effects of antiasthma medications could not be adequately addressed. β₂ Agonists significantly increase heart rate and decrease potassium concentrations, which may predispose susceptible individuals with asthma to cardiovascular events.¹²⁵ Several epidemiologic studies have linked the use of β₂ agonists to increased cardiovascular morbidity and mortality.^126- 127 Thus, they should be used with caution in those with cardiovascular comorbid conditions. Third, we did not evaluate the usefulness of noninvasive markers for monitoring disease activity and predicting therapeutic responses in patients with asthma. Measurements of nitric oxide from exhaled gases and of inflammatory cells/cytokines from induced sputum show early promise as clinically relevant markers of airway inflammation in asthma.¹²⁸ They may be especially useful in separating asthma from other inflammatory conditions of the airway and in evaluating therapeutic responsiveness.¹²⁹ However, more work is needed to define their incremental value above and beyond conventional measurements such as symptoms and lung function in the management of asthma. Finally, because airway inflammation in asthma is heterogeneous,⁶ considerable variations in therapeutic responses can be expected among patients. Therefore, the data from the present review are not meant to replace clinical judgment or intuition; they should be used as clinical aids for practicing physicians.

In summary, there is now a wealth of evidence supporting the use of inhaled corticosteroids in low doses as first-line therapy for adult patients with asthma who require more than an occasional use of short-acting β₂ agonists for control of their disease. In those with airflow obstruction who continue to have symptoms despite low-dose steroid therapy, the addition of long-acting β₂ agonists is reasonable. Alternatively, the dose of inhaled corticosteroids may be increased, although this may be associated with increased risk of adverse effects. In relatively young patients with asthma who cannot or will not take inhaled corticosteroids, monotherapy with leukotriene pathway modifiers is effective in reducing exacerbation rates; however, these agents are less effective than monotherapy with inhaled corticosteroids.