Effectiveness of inhaled corticosteroids in the treatment of acute asthma in children in the emergency department: A meta-analysis

Abstract

Objectives. This meta-analysis aimed to compare the treatment of an acute asthma attack in children in the emergency department (ED) with inhaled corticosteroids (ICS) versus placebo or oral systemic corticosteroids (SC) as assessed by the hospital admission rates.

Methods. After searching Medline, Cochrane, EMBASE, and Google Scholar, we identified ten articles that described randomized controlled trials of ICS versus placebo or oral SC for treating children with asthma attacks in the ED. Primary outcome was the hospital admission rate as defined as inpatient admission or admission into intensive care unit.

Results. Across the studies, a range of drugs and doses were used. For ICS, six studies administered budesonide (dose range: 0.4–2 mg), and three studies gave fluticasone/flunisolide (dose range: 0.5–2 mg). Six studies administered oral prednisone (dose range: 1–2 mg/kg/day), and four studies gave placebo. The rate of hospital admissions in patients treated with ICS was not significantly higher than those treated with oral SC. The rate of hospital admissions in patients treated with ICS was significantly lower than those treated with placebo.

Conclusion. ICS treatment of children with acute asthma exacerbations showed a similar rate of hospitalization as those treated by SC.

Key words::

• Asthma
• children
• emergency department
• inhaled corticosteroids
• meta-analysis

Key messages

• Inhaled corticosteroid treatment of pediatric patients with acute asthma exacerbations showed a similar rate of hospitalization as those treated by oral systemic corticosteroids.

Introduction

An asthma attack occurs in 4.3 million children yearly in the US. It triggers a hospital outpatient visit in 11.8% of asthmatic children, an emergency room visit in 10.7%, and hospitalization in 2.1% (1). Inhaled corticosteroids (ICS) are used routinely in the management of children with asthma (2). Daily long- term treatment with ICS decreases the occurrence and severity of asthma attacks and can reduce the need for oral corticosteroids on a routine basis (3). Systemic corticosteroids (SC) (oral, intravenous, or intramuscular) often are administered to mitigate acute asthma attacks in a hospital setting and may produce some improvements in patients admitted to hospital with acute asthma exacerbations (2). Benefits may include earlier discharge and fewer relapses.

It is common to use ICS for chronic management of asthma, but there is some evidence that it may also be beneficial in the acute setting (4). Systemic corticosteroids require 2–4 hours to reach peak effect (5), while the effect of ICS may be more rapid with similar efficacy as SC (6,7). Because acute asthma exacerbations occur in some children every three to four weeks in certain seasons, some primary care physicians prefer to use ICS rather than SC due to possible side effects (8). In some instances, children with an acute unresponsive asthma attack may be successfully treated with short-term high-dose budesonide therapy (4), possibly due to its local anti-inflammatory effect on the pulmonary tissues (9,10). However, this practice remains controversial (6,7,11,12).

A recent meta-analysis investigated the treatment of acute asthma in the emergency department with early use of ICS (3). Although they noted that the hospital admission rate in pediatric asthma patients was similar to that of adults with asthma, they did not report the effect of ICS administration in the emergency room on hospital admission rates of children with asthma (3). Because most studies of pediatric asthma patients have enrolled small groups, they may not have sufficient statistical power to detect a significant difference between treatments (6,8,13). Thus, the aim of this meta-analysis was to compare the effectiveness of ICS in treatment of an acute asthma exacerbation in children with placebo or SC in the emergency department by comparison of the hospital admission rates.

Methods

Literature search strategy

The Medline, Cochrane, EMBASE, and Google Scholar databases were searched using combinations of the key words, ‘asthma, acute asthma, inhaled corticosteroids, nebulized corticosteroid, systemic corticosteroid, emergency department, children’ by two independent reviewers. Titles were screened for relevancy, and abstracts of relevant articles published online as of 30 November 2012 were retrieved. The reference lists of relevant studies were hand-searched. Inclusion criteria were prospective, randomized controlled trial; subjects of less than 18 years of age; acute asthma exacerbation; comparison between ICS versus placebo or oral SC; and presentation to the emergency department. Exclusion criteria included a control group that used neither placebo nor oral SC, no numerical data for the primary outcome, non-English language publication, and other types of clinical trials (e.g. retrospective, cohort, non-randomized). If eligibility of an article was uncertain, a third reviewer was used to resolve any disagreements. We used the PRISMA criteria for reporting the review process.

Data extraction

The following information and data were extracted from each study that met the inclusion criteria: study design, number of participants in each treatment group, participants’ age and gender, drug and dosage for study/control group, results, adverse effects, hospital admission rate, and reference citation. The primary outcome was deemed the rate of hospital admission.

Quality assessment and publication bias

We utilized the Delphi list to assess the quality of the included studies (14), while a funnel plot and the fail-safe N (which indicates whether the observed significance is spurious or not) were used to assess possible publication bias for the primary outcome.

Statistical analysis

The hospital admission rate was used to evaluate treatment efficacy. Hospital admission was defined as a patient being admitted as inpatient or to the intensive care unit (ICU). Odds ratios (OR) with 95% confidence intervals (CI) were calculated for binary outcome and were compared between patients with ICS and placebo or patients with ICS and SC. A chi-square-based test of homogeneity was performed, and the inconsistency index (I²) statistic was determined. If I² was < 25%, the studies were considered to be homogeneous, and fixed-effects models were calculated. If I² was > 50% or > 75%, the trials were considered to be heterogeneous or highly heterogeneous, respectively, and a random-effects model was calculated. Pooled summary statistics for ORs of the individual studies were reported. A P value of less than 0.05 was chosen for significance. All analyses were performed using Comprehensive Meta-Analysis statistical software, version 2.0 (Biostat, Englewood, NJ, USA).

Results

Of the 8260 titles initially examined from the four databases, 20 full-length publications were retrieved and evaluated. After full review of the text of the articles, we excluded ten studies for the following reasons: not randomized controlled trial (n = 1) (15); control group used an intravenous SC administration rather than an oral formulation (n = 2) (16,17); participants were adults (n = 5) (9,17–20); and primary outcome was not expressed as a numerical value (n = 3) (4,21,22). The randomized controlled trials (RCT) of ten publications met all of our inclusion criteria (Figure 1).

Figure 1. Flow chart for the selection of the ten included studies.

The ten studies examined the responses of 829 children with an asthma attack who visited the emergency department: 420 children who received at least ICS and 409 patients who received placebo or systemic steroids but no ICS (6–8,11– 13,23–26). The characteristics of each RCT in the ten publications (6–8,11–13,23–26) are summarized in Table I. Six studies in the primary analysis had compared ICS versus placebo: four studies (8,13,23,24) compared ICS versus placebo with both groups with no SC, and two studies compared ICS versus placebo with SC in both treatment groups (25,26). Four studies compared ICS versus SC (6,7,11,12). The severity of initial presentation of asthma ranged from mild to moderate in one study (7), moderate degree in four studies (10,11,21,22), moderate to severe in three studies (6,23,24), and severe degree in one study (9). Additional characteristics of the RCT and its patients are listed in Table II.

Table I. Summary of characteristics of studies included in meta-analysis.

First author	Year	Study type	Comparison	Number of cases	Age (years)	Male (%)	Admission rate (%)
Singhi S	1999	RCT	ICS versus Placebo	30 versus 30	8 versus 8	63 versus 67	0 versus 23.3
Milani GK	2004	RCT	ICS versus Placebo	17 versus 15	4 versus 4	60 versus 47	0 versus 6.7
Sekerel BE	2005	RCT	ICS versus Placebo	33 versus 34	10 versus 10	73 versus 62	0 versus 2.9
Estrada-Reyes E	2005	RCT	ICS versus Placebo	50 versus 50	10 versus 10	50 versus 64	0 versus 0
Sung L	1998	RCT	ICS + SC versus SC	24 versus 20	4 versus 4	79 versus 80	8.3 versus 25
Upham BD	2011	RCT	ICS + SC versus SC	91 versus 89	6 versus 6	70 versus 61	61.5 versus 61.8
Scarfone RJ	1995	RCT	ICS versus SC	56 versus 55	5 versus 5	57 versus 62	37.5 versus 56.4
Devidayal	1999	RCT	ICS versus SC	41 versus 39	6 versus 7	76 versus 74	12.2 versus 2.6
Schuh S	2000	RCT	ICS versus SC	51 versus 49	9 versus 10	49 versus 69	31.4 versus 10.2
Nakanishi AK	2003	RCT	ICS versus SC	27 versus 28	10 versus 11	63 versus 64	0 versus 3.6

ICS = inhaled corticosteroids; RCT = randomized controlled trial; SC = systemic corticosteroids.

Table II. Drug information and adverse events in selected studies.

First author	Year	Severity of acute exacerbation	Initial treatment	Drug and dosage of ICS	Drug and dosage of SC	Adverse events
Singhi S	1999	Moderate	Nebulized salbutamol (0.15 mg/kg/dose, max. 5 mg) for three doses	Budesonide (400 μg/dose) for three doses	Placebo	N/A
Milani GK	2004	Moderate	Nebulized salbutamol (0.15 mg/kg/dose, max. 5 mg) for three doses	Budesonide (2 mg)	Placebo	• Coughing (n = 2 in ICS, n = 1 in placebo) • Vomiting (n = 1 in placebo)
Sekerel BE	2005	Mild to moderate	Nebulized salbutamol (0.15 mg/kg/dose, max. 5 mg) for three doses	Budesonide (1 mg/dose) for three doses	Placebo	N/A
Estrada-Reyes E	2005	Moderate	Nebulized salbutamol (30 μL/kg/dose) for three doses	Fluticasone (500 μg/dose)	Placebo	No adverse effects noted in either group
Sung L	1998	Moderate to severe	Nebulized salbutamol (0.15 mg/kg/dose, max. 5 mg) for three doses	Budesonide (a single dose of 2000 μg, 4 mL) + oral prednisone (1 mg/kg, max. 50 mg)	Oral prednisone (1 mg/kg, max. 50 mg)	No adverse effects noted in either group
Upham BD	2011	Moderate to severe	Nebulized albuterol (3.75 mg/dose for patients weighing 10–20 kg, 5 mg/dose for those > 20 kg) for three doses + nebulized ipratropium bromide (500 μg/dose) for two doses	Budesonide (2 mg) + oral prednisolone, prednisone, or methylprednisolone (2 mg/kg, max. 60 mg)	Oral prednisolone, prednisone, or methylprednisolone (2 mg/kg, max. 60 mg)	• The most commonly reported AEs were rhinorrhea, headache, diarrhea, sore throat, and cough; these were equally distributed between treatment groups • Transient hyperglycemia (n = 2 in ICS) • Serious AEs (n = 3 in ICS+ SC, n = 2 in SC)
Scarfone RJ	1995	Moderate	Nebulized albuterol (0.15 mg/kg/dose, max. 5 mg) for three doses	Dexamethasone (1.5 mg/kg, max. 45 mg)	Oral prednisolone (2 mg/kg, max. 60 mg)	Vomiting (0% in ICS, 15% in SC)
Devidayal	1999	Moderate to severe	Nebulized salbutamol (0.15 mg/kg/dose) for three doses	Budesonide (800 μg)	Oral prednisolone (2 mg/kg/day rounded off to the nearest 5 mg)	N/A
Schuh S	2000	Severe	Nebulized albuterol (0.15 mg/kg/dose) for three doses + ipratropium bromide (250 μg/dose) for three doses	Fluticasone (2 mg)	Oral prednisone (2 mg/kg, max. 60 mg)	• Vomiting (n = 24 in ICS, n = 8 in SC) • Tremor (n = 17 in ICS, n = 6 in SC) • Intravenous hydration required (n = 9 in ICS, n = 0 in SC)
Nakanishi AK	2003	N/A	Nebulized albuterol (0.15 mg/kg/dose, max. 5 mg) for three doses + ipratropium bromide (0.25 mg/dose) for three doses	Flunisolide (1 mg/dose) b.i.d. for 7 days	Oral prednisone (2 mg/kg, max. 60 mg/day) for 7 days	Reported side effects were minimal

AE = adverse events; ICS = inhaled corticosteroids; SC = systemic corticosteroids; N/A = not available.

The quality of the methodology of the ten randomized trials was assessed with the Delphi list of criteria (Table III) and was considered high. The assessor, provider, and patients were blinded to the intervention in all included trials, and seven of ten studies presented measures of variability. However, only two trials included the analysis on intention to treat.

Table III. Quality assessment of trials included in the meta-analysis.

First author	Year	Was a method of randomization used?	Were the groups similar at baseline regarding the most important prognostic indicators?	Were the eligibility criteria specified?	Was the outcome assessor blinded?	Was the care provider blinded?	Was the patient blinded?	Were point estimates and measures of variability presented for the primary outcome measures?	Did the analysis include an intention-to-treat analysis?
Singhi S	1999	Yes	Yes	Yes	Yes	Yes	Yes	Yes	No
Milani GK	2004	Yes	Yes	Yes	Yes	Yes	Yes	No	No
Sekerel BE	2005	Yes	Yes	Yes	Yes	Yes	Yes	Yes	No
Estrada-Reyes E	2005	Yes	Yes	Yes	Yes	Yes	Yes	No	No
Sung L	1998	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Upham BD	2011	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Scarfone RJ	1995	Yes	Yes	Yes	Yes	Yes	Yes	No	No
Devidayal	1999	Yes	Yes	Yes	Yes	Yes	Yes	Yes	No
Schuh S	2000	Yes	Yes	Yes	Yes	Yes	Yes	Yes	No
Nakanishi AK	2003	Yes	Yes	No	Yes	Yes	Yes	Yes	No

Study characteristics and clinical outcome

The primary outcome was hospital admission rate which varied from 0% to 61.5% in the ICS group and from 0% to 61.8% in the placebo/SC group. The most recent trial (26) reported higher rates of hospital admission compared to eight of the nine other trials.

The ORs of the five studies that evaluated the effects of patients treated with ICS on admission rate ranged from 0.05 to 0.99 (Figure 2A). The homogeneity in the combined OR among the five studies (Q = 5.925, I² = 32.49%, P = 0.205) indicated the use of a fixed-effects model of analysis. The combined OR revealed that the reduction in hospital admissions in patients treated with ICS was not significant (OR = 0.74, 95% CI = 0.13–1.27, P = 0.277). Analysis of the ICS versus placebo subgroup indicated homogeneity in combined OR among the three studies (Q = 0.887, I² = 0.00%, P = 0.64); and a fixed-effects model of analysis was used. The result indicated that ICS use significantly reduced hospital admission rate (OR = 0.15, 95% CI = 0.03–0.93, P = 0.042) (Figure 2B). Analysis of the ICS plus SC versus SC subgroup indicated no significant difference in hospital admissions between patients treated with ICS plus SC and SC (OR = 0.87, 95% CI = 0.49– 1.53, P = 0.618) (Figure 2C).

Figure 2. Forest plot of comparison: inhaled corticosteroids versus placebo. A: total; B: inhaled corticosteroids versus placebo; C: inhaled corticosteroids + systemic corticosteroids versus systemic corticosteroids; and D: inhaled corticosteroids versus systemic corticosteroids.

The four trials that compared ICS versus SC (6,7,11,12) showed heterogeneity in the combined OR (Q = 13.728, I² = 78.15%, P = 0.003). The random-effects model of analysis (Figure 2D) of the combined OR showed that the rate of hospital admissions in patients treated with ICS was not significantly higher than in those treated with oral systemic corticosteroid (OR = 1.54, 95% CI = 0.30–7.83, P = 0.604) (Figure 2D).

The hospital admission rates varied between 0% and 61.5% in these ten trials: the rates ranged from 0% to 37.5% in pediatric patients treated with ICS alone (6–8,11–13,23,24) and from 8.3% to 61.5% in pediatric patients treated with ICS plus SC (25,26). In order to determine whether this wide variation in hospital admission rates was associated with the severity of the presenting asthma exacerbation, the initial treatment of the asthma exacerbation, or the drug or dose of the administered ICS during the trials, we compiled these factors for each trial in Table II. No trial solely treated patients with mild exacerbations. One trial treated patients with mild to moderate exacerbations, and its hospital admission rate was 0% for the ICS-treated group and 2.9% for the placebo group (8). Four trials treated patients with moderate asthma exacerbations, and the hospital admission rates of the ICS groups were 0%, 0%, 0%, and 37.5% (12,13,23,24), whereas the three trials that treated patients with moderate to severe asthma exacerbations reported hospital admission rates in the ICS group of 8.3%, 12.2%, and 61.5% (7,25,26). The single trial that treated patients with severe asthma exacerbations reported hospital admission rates in the ICS group of 31.4% (11). No linear relationship between the ICS groups’ severity of asthma exacerbations with hospital admission rates was detected; however, we cannot rule out that further analysis of individual patient data from the trials may detect a relationship.

The initial treatment schedules before study enrollment were very similar among the ten trials (Table II) and did not appear to correlate directly with the trials’ hospital admission rates. The relationship between drugs and hospital admission rates was complicated by the range of doses used (Table II). For example, the hospital admission rates of trials using budesonide for the ICS ranged from 0% to 61.5% (7,8,23–26). Thus, the variables that caused or contributed to the wide variation in hospital admission rates among the ten studies remain to be determined.

Quality assessment and publication bias

The quality of the methodology of the ten randomized trials was assessed with the Delphi list of criteria (Table III) and was considered high. The assessor, provider, and patients were blinded to the intervention in all included trials, and seven of ten studies presented measures of variability. However, only two trials included the analysis on intention to treat. The funnel plot for publication bias demonstrated marked evidence of asymmetry (Figure 3), indicating the presence of publication bias.

Figure 3. Funnel plot of comparison: inhaled corticosteroids versus placebo.

Discussion

This study is the first meta-analysis that compared the hospital admission rates of pediatric patients with an acute asthma attack who were treated with ICS, placebo, or SC in the emergency room. We found that the rate of hospital admissions of patients treated with ICS was not significantly higher than those treated with oral systemic corticosteroid; these results suggest that ICS treatment may provide similar benefits to most pediatric patients. Secondly, the hospital admission rate of patients treated with ICS was significantly lower than those treated with a placebo; these results indicate that ICS treatment greatly improves control of acute asthma exacerbations in children in the emergency room. The acute asthma exacerbations of the children ranged from mild to severe in the ten included studies.

A recent meta-analysis compared the hospital admission rate of adult and pediatric asthma patients treated with ICS in the emergency department with those treated with SC (3). Although they noted that the hospital admission rate in pediatric asthma patients was similar to that of adults with asthma, they did not report the effect of ICS administration in the emergency room on hospital admission rates of children with asthma (3). Their meta-analysis of the pediatric subgroup included seven studies and only compared ICS with placebo (3). In contrast, we included ten studies that compared the hospital admission rate of pediatric patients treated with the three strategies: ICS versus placebo, ICS+ SC versus SC+ placebo, and ICS versus SC.

Limitations of the meta-analysis include the publication bias, which was similar to that of a recent meta-analysis of ICS in asthmatics (3). Further analysis indicated that only the studies that compared ICS with placebo were homogeneous (I² = 0.00%). Hence, it is important to interpret the findings with caution, particularly those that compared ICS with SC. A second limitation was the inconsistent reporting of several markers for lung function. For example, different studies reported the forced expiratory volume in 1 second (FEV₁) value, whereas others reported the percent of predicted value: these distinct expressions for the same marker thwarted robust statistical analysis of these parameters. In addition, there were many other factors that differed across studies, including atopic status, different age group, dosage, severity of disease, different asthma medication prior enrollment, which confounded the analysis. A third limitation was the wide variation in the hospital admission rates of pediatric patients with acute asthma exacerbations treated at the different emergency departments: The rates reported in the ten included studies varied from 0% to 61% (6–8,11–13,23–26). Because we did not have access to the primary data, we were unable to determine the extent that the pre-enrollment treatment or the decision points for hospital admission contributed to this variation. Identifying therapeutic regimens that can reduce hospital admission rates without increasing the risk of lasting effects on growth, weight distribution, and hypothalamic-pituitary-adrenal axis (HPA) would likely benefit the pediatric patients with asthma, especially those with recurring exacerbations that are non-responsive to home therapy.

Our findings should be interpreted with caution due to the heterogeneity of the studies included in this meta-analysis. The high heterogeneity reflects the diversity in study design across these clinical studies and indicates the need for studies that use similar study design such as patient populations, methods of evaluating pulmonary function, drug dose, clinical assessments, etc. Only when such studies are available will it be possible to perform quantitative meta-analyses that can be used to draw robust conclusions regarding this important medical question.

In conclusion, this meta-analysis indicates that ICS appeared to show similar efficacy for the treatment of an acute asthma attack of mild to severe intensity in children in the emergency department as oral SC, as assessed by hospital admission rates. For those pediatric patients with recurring asthma attacks, ICS treatment in the emergency department rather than oral SC may reduce their risk to the side effects of SC.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.