Bronchodilators (β₂ agonists) and steroids are well established effective treatments for all severities of asthma.^1–5 However, there are still children with moderate to severe acute asthmatic attacks that may have insufficient improvement, leading to hospitalisation or intensive care unit (ICU) admission, and in the worst cases, severe morbidity and even mortality. Anticholinergics and methylxanthines have controversial efficacies.^6–8 There is an ongoing need to study new effective bronchodilating agents as additional therapies for relieving moderate to severe asthmatic attacks.

Magnesium sulphate is one of the agents being intensively studied. Although a systematic review published by the Cochrane Collaboration showed that magnesium sulphate is effective in the treatment of severe acute asthma, it included mainly adult studies.⁹ Whether intravenous magnesium sulphate is effective in treating acute asthma in children is still controversial.

We performed a meta-analysis to evaluate the effectiveness of intravenous magnesium sulphate in preventing hospitalisation or ICU admission in children with acute asthmatic attacks, either used alone or in addition to standard therapies.

METHODS

The search of relevant studies, assessment of trial qualities, extraction of data, synthesis of data by meta-analysis, examination of heterogeneity, and drawing of conclusions were performed by two independent reviewers; discrepancies were resolved by consensus.

Selection criteria

Inclusion criteria were controlled clinical trials of intravenous magnesium sulphate of any therapeutic dose in children below 18 years of age using outcome measures of hospitalisation, ICU admission, clinical symptom scores, or pulmonary function tests. Both published and unpublished studies were included. There was no language restriction. The exclusion criterion was multiple publications.

Search strategies

Six databases (Medline, Pubmed, Embase, Cochrane Library, Cochrane Central Register of Controlled Trials, and China Journet Net) were searched under the following terms in the keyword search: “magnesium” and “asthma” and “children”. Reference lists and citations of identified studies were located by hand search to explore any published studies missed. Personal contacts with authors and specialists were made as far as possible to obtain unpublished data.

Quality assessment

The trial quality was assessed using the standard Jadad score¹⁰ based on the adequacy of randomisation, blinding, and follow up, with the maximum score of 5 points.

Data extraction

Baseline characteristics of treatment and control groups including age, gender, baseline SpO₂, symptom severity, symptom scores, and results of pulmonary function tests (PEFR, FEV₁, and FVC) were extracted. The primary outcome measure chosen was rate of hospitalisation or ICU admission, expressed as odds ratio (OR) in treatment compared to control group. Secondary outcome measures were rate of persistent severe bronchoconstriction as indicated by PEFR <60% predicted, between group difference in symptom scores and percentage change of pulmonary function tests at the end of the trial. Stratification of outcomes by treatment dose, age, and severity class (mild, moderate or severe) for subgroup analyses would be performed if data were available.

Data analyses

Comprehensive meta-analysis software (available at http://www.meta-analysis.com) was used for all statistical analyses. Statistical heterogeneity in the outcome measures was assessed using the Cochrane Q test, with p < 0.1 considered as significant. The random effect model was used for pooling results if significant statistical heterogeneity was present. Otherwise the fixed effect model was used. Clinical heterogeneity among studies was explored in terms of study design, patients recruited, baseline characteristics, intervention, co-therapies, and outcome measures. Sensitivity analysis was performed by reanalysis, excluding studies with poor quality (Jadad score <3).

RESULTS

Under the specified keyword search strategy, 35, 70, 31, 7, 9, and 0 articles were identified in Medline, Pubmed, Embase, Cochrane Library, Cochrane Central Register of Controlled Trials, and China Journet Net respectively. Five independent randomised placebo controlled trials satisfying inclusion criteria were identified.^11–15 No additional published or unpublished article satisfying inclusion criteria was found on hand search of the citations of these studies and personal contacts.

All trials were carried out at emergency departments comparing intravenous magnesium sulphate to placebo in treating paediatric patients with acute moderate to severe asthmatic attacks, with co-therapies of inhaled β₂ agonists and systemic steroids. All patients who entered the trials had similar severity of asthma as indicated by inadequate response to first line treatment with three doses of β₂ agonists. All the studies were assessed to be of good quality (Jadad score ⩾4).

Table 1 presents the main results of individual studies and the Jadad scores. Since no study mentioned rate of ICU admission, we could only evaluate primary outcome of hospitalisation. Table 2 shows the pooled effect size estimates of primary and secondary outcomes and tests of heterogeneity. The pooled OR estimate for the primary outcome is illustrated by the Forrest plot in fig 1.

Four of the studies concluded that magnesium sulphate was effective in paediatric acute asthma while one study found it to be ineffective. After pooling the results together, intravenous magnesium sulphate was effective in avoiding hospitalisation. The absolute risk reduction (ARR) for hospitalisation is 0.257 (95% CI 0.124 to 0.389, p = 0.0001). The number needed to treat (NNT) in avoiding hospitalisation is 4 (95% CI 3 to 8).

The pooled result on short term lung function suggested that there was an approximately 85% reduction in risk of persistent bronchoconstriction with PEFR <60% of predicted values, with all studies showing consistent reduction in risks, although to somewhat different extent.

The pooled result on asthma symptom score also showed significant improvement, with individual studies showing consistent results. It was noted that two different scoring systems were used in the studies, with three of the studies^12–15 using the clinical asthma symptom score and one using the pulmonary index. However, their results pointed in the same direction, favouring treatment compared to placebo with statistical significance, irrespective of the scoring system used.

No subgroup analysis was performed in different age groups because stratified outcome data were not available in any of the studies. Since all studies included only patients with moderate to severe asthma, we could not evaluate the efficacy of magnesium sulphate in mild asthma. No study compared the effectiveness of different dosages of magnesium sulphate, and therefore we could not further evaluate the dose-response relation. Sensitivity analysis was not performed because of the small number of studies and similarity of trial quality. Possible areas of heterogeneity of the trials are shown in table 3 (available on the ADC website).

DISCUSSION

This is the first meta-analysis investigating the effectiveness of intravenous magnesium sulphate in treating acute childhood asthma; it showed a favourable effect. The NNT in avoiding hospitalisation is only four, implying probable high cost effectiveness. In addition to β₂ agonists and systemic steroids, intravenous magnesium sulphate also appears effective in relieving bronchoconstriction and improving asthma symptoms.

Three different single dose regimes of intravenous magnesium were used in the five studies: 25 mg/kg in two studies, 40 mg/kg in two, and 75 mg/kg in one. Interestingly, one study using 25 mg/kg and another study using 40 mg/kg magnesium showed a significant reduction in the rate of hospitalisation, whereas the other study using 25 mg/kg and the study using 75 mg/kg magnesium reported an insignificant effect on hospitalisation. Whether this inconsistency in dose-response relation is related to intrinsic dosage related effect or confounding factors such as criteria of hospitalisation is not apparent in the current meta-analysis and requires further evaluation in future studies. However, the inconsistent dosage effect might reflect a small sample size of individual studies, resulting in inadequate statistical power to detect a significant effect, especially in view of the consistent trend of reduction in hospitalisation and statistical homogeneity of this primary outcome across studies. Nevertheless, we observed no clear evidence to suggest that higher dosage of magnesium results in better outcomes; further studies are therefore needed to identify the optimal dosage of magnesium with maximal efficacy.

Intravenous magnesium sulphate therapy was well tolerated with only minor side effects reported, such as epigastric or facial warmth, flushing, pain and numbness at infusion site, dry mouth, and malaise.^12,16 It is generally safe to be used in children with moderate to severe acute asthma, although rare adverse effects might have escaped detection in the small samples.

The mechanism of action of magnesium sulphate is not clearly defined. Magnesium ion decreases the uptake of calcium by bronchial smooth muscle cells, which in turn leads to bronchodilation.¹⁷ Magnesium may also have a role in inhibiting mast cell degranulation, thus reducing inflammatory mediators such as histamine, thromboxanes, and leukotrienes.¹⁷ In addition, magnesium inhibits the release of acetylcholine from motor nerve terminals and depresses the excitability of muscle fibre membranes.^18,19

There were several limitations in the current meta-analysis. Firstly, studies were heterogeneous with respect to patient inclusion criteria, baseline characteristics, magnesium sulphate dosage, co-therapies, and outcome measures, although there was no statistically significant heterogeneity in the primary outcome. Hence care should be taken in applying the pooled result to individual patients in clinical practice. Secondly, the total number of patients involved in all the studies combined is still relatively small. The study by Scarfone and colleagues¹⁵ comprised about 30% of patients and concluded that magnesium is ineffective. Two of the five studies reporting effectiveness of magnesium had the same first author^11,14 and involved one third of the total number of patients. Thirdly, the pooled effect estimate, though significant statistically, was not precise as only a few studies have been performed. Fourthly, subgroup analysis was not possible because stratification of outcomes or raw data were not available. Finally, publication bias might be present as suggested by funnel plot asymmetry (fig 2). Unpublished negative studies might have been missed.

Conclusion

Based on existing data, intravenous magnesium sulphate is likely to be effective in avoiding hospitalisation and improving bronchoconstriction and clinical symptoms of moderate to severe acute asthma in children, when added to standard therapies of inhaled bronchodilators and systemic steroids.

Further studies are needed to evaluate the effectiveness of intravenous magnesium sulphate in different severity classes of asthma and in patients of different age groups. Studies should also aim at further defining the indications, possible contraindications, and optimal dosage of intravenous magnesium sulphate.