The 3Mg Trial: Randomised Controlled Trial of Intravenous Or Nebulised Magnesium Sulphate

The 3Mg Trial: Randomised controlled trial of intravenous or nebulised magnesium sulphate or standard therapy for severe acute asthma

Steve Goodacre PhD1, Judith Cohen PhD1, Mike Bradburn MSc1, Alasdair Gray MD2, Jonathan Benger MD3, Timothy Coats MD4 on behalf of the 3Mg Research Team

1School of Health and Related Research (ScHARR), University of Sheffield, 2Emergency Department, Royal Infirmary of Edinburgh, 3Faculty of Health and Life Sciences, University of the West of England, Bristol, 4Emergency Department, Leicester Royal Infirmary

Corresponding author

Steve Goodacre, Professor of Emergency Medicine

University of Sheffield, Regent Court, 30 Regent Street, Sheffield, S1 4DA

Email:

Tel: +44 114 222 0842

This is the accepted version of an article that has been published in the Lancet Respiratory Medicine:

Registration:

Lancet protocol 08PRT/503:

Abstract

Background: Previous studies suggest that intravenous (IV) or nebulised magnesium sulphate may improve respiratory function in acute asthma. We aimed to determine whether IV or nebulised magnesium sulphate improve symptoms of breathlessness and reduce the need for hospital admission in adults with severe acute asthma.

Methods: In a double-blind placebo-controlled trial undertaken in the emergency departments of 34 hospitals we randomised 1109 adults with severe acute asthma to receive either IV magnesium sulphate (2g over 20 minutes) or nebulised magnesium sulphate (3 x 500mg over one hour) alongside standard therapy including salbutamol, or standard therapy alone. Consented participants were allocated to numbered treatment packs using a telephone or internet randomisation system. A simple randomisation sequence was used in 20 hospitals participating at the outset, but switched to blocked randomisation, stratified by hospital, for subsequent hospitals. Each treatment pack contained an IV infusion and three nebuliser solutions, either of which could be active treatment or placebo. The primary outcome measures were the proportion of patients admitted to hospital (either after emergency department treatment or at any time over the subsequent seven days) and breathlessness measured on a 100mm visual analogue scale (VAS) over two hours after initiation of treatment.Participants were analysed in the groups to which they were allocated, regardless of whether they actually received or completed the allocated treatment. Recruitment continued towards a target of 1200 participants until funding expired and the trial closed. ISRCTN04417063

Findings: Hospital admission was recorded for 1084 patients (394 IV magnesium sulphate, 332 nebulised magnesium sulphate, 358 placebo, mean age 36.1 years, 763/1084 (70%) female) and VAS breathlessness for 976. IV magnesium sulphate was associated with an odds ratio of 0.73 (95% confidence interval 0.51 to 1.04, p=0.083) for hospital admission, an improvement in VAS breathlessness that was 2.6mm (-1.6 to 6.8mm, p=0.231) greater than placebo, and an improvement in percentage predicted peak expiratory flow rate (PEFR) that was 0.4% (-2.3 to 3.0%, p=0.786) less than placebo. Nebulised magnesium sulphate was associated with an odds ratio of 0.96 (0.65 to 1.40, p=0.819) for hospital admission, an improvement in VAS breathlessness that was 2.6mm (-1.8mm to 7.0mm, p=0.253) less than placebo, and an improvement in percentage predicted PEFR that was 0.6% (-2.1 to 3.4%, p=0.652) less than placebo.

Interpretation:These findings suggest that there is no role for nebulised magnesium sulphate in the management of severe acute asthma in adults and at best only a limited role for IV magnesium sulphate.

Funding: National Institute for Health Research Health Technology Assessment Programme (HTA06/01/02)

Background

Acute asthma is responsible for around 60,000 hospital admissions per year in England1. Current guidelines2,3 advise a stepwise approach to the management of exacerbations. Initially all patients should receive oxygen, nebulised 2-agonists, nebulised anticholinergic agent and corticosteroids. However, bronchodilators act within minutes whereas corticosteroids require hours to take effect. This suggests a potential role for magnesium sulphate as an additional treatment option in the therapeutic gap between nebulised bronchodilators and corticosteroids.

Magnesium sulphate has been evaluated in both the intravenous (IV) and nebulised form. The nebulised route offers the potential advantage of a quick onset of action and lower incidence of side effects. Its disadvantages include a lower dose of drug delivered and the patient requiring some respiratory effort to maximise its effectiveness. The IV route provides direct access to the venous system, allowing the delivery of high drug concentrations. The disadvantages include the requirement for intravenous access and the drug being administered by infusion over 20 minutes.

Several systematic reviews and meta-analyses have evaluated the role of IV or nebulised magnesium sulphate in acute asthma4-10. The most recent10 showed that IV treatment appeared to be effective in children but was unable to draw clear conclusions about treatment in adults. Both IV treatment (10 trials, 955 adults) and nebulised treatment (7 trials, 430 adults) were associated with weak evidence of improved respiratory function compared to control. No trials directly compared IV to nebulised magnesium sulphate. The standardised mean difference (SMD) for IV treatment was 0.25 (95% CI -0.01 to 0.51, p=0.06) and for nebulised treatment was 0.17 (95% CI -0.02 to 0.36, p=0.09). Meta-analysis showed that IV treatment was associated with weak evidence of an effect on hospital admission (relative risk (RR) 0.68, 95% CI 0.46 to 1.02, p=0.06) while nebulised treatment was associated with no significant effect (RR 0.87, 95% CI 0.70 to 1.08, p=0.22). One further trial of IV magnesium sulphate in adults11 has since been published. Inclusion of this trial in the meta-analysis12 resulted in the effect upon respiratory function being slightly larger and statistically significant (SMD=0.35, 95% CI 0.06 to 0.64, p=0.02) but the effect on hospital admission remained non-significant (RR=0.85, 95% CI 0.68 to 1.06, p=0.14). It is not clear whether changes in measures of respiratory function are associated with important changes in patient management or a clinically meaningful improvement in symptoms.

Uncertainty in the evidence is reflected in treatment recommendations. Current guidelines in the United Kingdom2 and the United States3 suggest that IV magnesium sulphate should be considered in adults with life-threatening features or severe acute asthma that has not responded to inhaled bronchodilator therapy. No recommendations are made regarding nebulised magnesium sulphate.

We measured the effectiveness of IV and nebulised magnesium sulphate in adults with severe acute asthma. We specifically aimed to determine whether IV or nebulised magnesium sulphate, used alongside standard treatment including salbutamol, reduces the proportion of patients requiring hospital admission at initial presentation or during the following seven days, and whether IV or nebulised magnesium sulphate improves patient assessment of breathlessness over two hours after initiation of treatment.

Methods

Study design and patients

We undertook a multi-centre, double blind, placebo controlled, three-arm, randomised trial in 34 emergency departments in the United Kingdom. The trial protocol has been published13. Adults (age>16) attending the emergency department with severe acute asthma were eligible for recruitment (i.e. acute asthma with either PEFR < 50% of best or predicted, respiratory rate > 25/min, heart rate > 110/min, or inability to complete sentences in one breath). We excluded patients who had life threatening features (oxygen saturation < 92%, silent chest, cyanosis, poor respiratory effort, bradycardia, arrhythmia, hypotension, exhaustion, coma or confusion), those with a contraindication to either nebulised or intravenous magnesium sulphate (pregnancy, hepatic or renal failure, heart block or known hypermagnesaemia), those unable to provide written or verbal consent, and previous participants in the 3Mg trial. We amended the protocol during the trial to also exclude those who had received magnesium sulphate in the 24 hours prior to recruitment. Written or verbal consent was sought from all participants. Those initially providing verbal consent were asked for written consent as soon as their condition permitted.

Randomisation and masking

Consented participants were randomised through a telephone or internet randomisation system managed by the Sheffield Clinical Trials Research Unit (CTRU). After being entered into the trial participants were allocated to numbered treatment packs kept in the emergency department. A simple randomisation sequence was used in the 20 hospitals participating at the outset, but switched to blocked randomisation (block sizes of four or six), stratified by hospital, for subsequent hospitals to safeguard against new centres recruiting too few in any trial arm. Each treatment pack contained an IV infusion and three nebuliser solutions, either of which could be active treatment or placebo. Participants, hospital staff and research staff were blinded to the allocated treatment.

Interventions and concurrent treatments

The three treatment arms were as follows:

1. IV magnesium sulphate, 8 mmol (2g) in 100ml normal saline given over 20 minutes, and three 7.5ml vials of 0.9% saline nebulised at 20 minutes intervals
2. IV normal saline, 100ml given over 20 minutes, and three 7.5ml vials of 2 mmol (500mg) magnesium sulphate nebulised at 20 minutes intervals
3. IV normal saline, 100ml given over 20 minutes, and three 7.5ml vials of 0.9% saline nebulised at 20 minutes intervals

Standard therapy was provided in accordance with guidelines2 from the British Thoracic Society (BTS) and Scottish Intercollegiate Guidelines Network (SIGN) and consisted of oxygen, nebulised salbutamol (5mg), nebulised ipratropium (500mcg) and oral prednisolone administered during recruitment, followed by up to 5mg salbutamol added to each trial nebuliser. Other treatments were given at the discretion of the clinician. Patients were managed in the emergency department and data collected until two hours after randomisation. At this point, if not already undertaken, a final disposition decision was made (hospital admission or discharge) and initial data collection completed.

Outcome measures

Two primary outcomes were specified:

1. The health service primary outcome was the proportion of patients admitted to hospital, either after emergency department treatment or at any time over the subsequent seven days.
2. The patient-centred primary outcome was the patient’s visual analogue scale (VAS) for breathlessness over two hours after initiation of treatment. VAS breathlessness has been used to measure breathlessness during exercise14 and has been shown to correlate with respiratory function and symptomatic change in cohorts with acute asthma15,16.

Secondary outcomes included mortality, adverse events, use of ventilation or respiratory support, length of hospital stay, admission to a high dependency unit (HDU) or intensive care unit (ICU), change in PEFR and physiological variables (oxygen saturation, heart rate, respiratory rate, blood pressure) over two hours, quality of life at baseline and one month, number of unscheduled health care contacts over the subsequent month, and satisfaction with care.

Adverse events and side effects occurring during emergency department treatment were routinely recorded on the Case Report Form by the treating clinician. Key events (cardiac arrest, respiratory arrest, emergency intubation, non-invasive ventilation, pneumothorax and arrhythmia) and common side effects (flushing, nausea, vomiting and hypotension (systolic <100mmHg)) were specifically sought and recorded. Other events were recorded on a general adverse event reporting form. Patient notes were reviewed by a research nurse who recorded any side effects identified during treatment or adverse events occurring up to 30 days after treatment. Adverse events were identified and reported according to Good Clinical Practice (GCP) guidance.

Sample size

We planned to recruit 1200 participants divided equally between the three trial arms to provide the following statistical power:

1. Assuming that 80% of patients with severe acuteasthma are admitted after emergency department management and hospital admission is recorded for all participants, the study would have 90% power to detect a 10% absolute reduction in the proportion admitted (i.e. to 70%) for any pair of treatment groups compared (two-sided alpha=0.05).
2. Assuming that 80% of participants have their VAS measured then the study would have 90% power to detect a 8mm difference in a 100mm VAS at two hours after treatment initiation (two-sided alpha=0.05). Previous data have established that the standard deviation on a 100mm VAS is 30mm, and that 22mm represents a minimum clinically significant difference15.

Statistical analysis

Participants were analysed in the groups to which they were allocated, regardless of whether they actually received or completed the allocated treatment. Logistic regression was used for analysis of admission rates. For length of stay, means (medians) were compared using censored Normal (log-Normal) regression to account for interval censoring in discharged patients (for whom no time of discharge was recorded) and also admissions which were ongoing at 30 days. Number of days on ICU/HDU were compared using Mann-Whitney U test. Analysis of covariance was used for all other outcomes. The primary analysis was adjusted for hospital and is presented for observed data (complete case); further analyses using different imputation strategies were used as confirmatory analyses. A secondary explanatory analysis was undertaken limited to those who completed the treatment as per protocol. We used Simes’s (1986) method17, which is a modification of the Bonferroni method but has better power, to adjust for multiplicity arising from having two primary outcomes. The two pre-planned comparisons between the three groups were (1) active treatment (IV or nebulised) versus placebo and (2) IV versus nebulised magnesium sulphate. We also present comparisons of IV magnesium sulphate versus placebo and nebulised magnesium sulphate versus placebo for completeness. We undertook three pre-planned subgroup analyses stratified by age (above or below 50 years), baseline PEFR (above or below median) and whether the patient had received treatment with salbutamol before the trial treatments.

An independent Data Monitoring Committee (DMC) reviewed trial data at regular intervals and reported recommendations to the Trial Steering Committee in accordance with the DMC charter. The Trial was approved by the Scotland A Research Ethics Committee. The trial sponsor was Sheffield Teaching Hospitals NHS Foundation Trust.

Role of the funding source

The study funders had no role in study design; in the collection, analysis, and interpretation of data; in the writing of the report; and in the decision to submit the paper for publication. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.

Results

Patients were recruited from 34 hospitals between 30/7/2008 and 30/6/2012. Recruitment was slower than anticipated and ended when the trial funding expired. The flow of patients through the trial is shown in figure 1. Of the 1109 patients recruited, 25 withdrew without commencing trial medication, were recruited in error (protocol violations) or could not be allocated to a treatment pack, so 1084 were included in the analysis. Table 1 shows the baseline characteristics of the recruited patients. Age and sex characteristics were balanced across the groups, but there were more white ethnicity patients in the IV magnesium group and more patients who had never smoked in the nebulised magnesium group. Supplemental tables 1 and 2 show the trial and concurrent medications received by the three patient groups. There was a high degree of adherence to the trial protocol: the mean total dose of nebulised solution was 21.3ml, with 85% receiving the full dose of 22.5ml, while the mean total dose of IV infusion was 97.1ml with 89% receiving the full IV infusion.

The results of primary outcome analysis are shown in table 2 (admission to hospital) and table 3 (VAS breathlessness). The odds ratios for admission to hospital were 0.84 (95% CI 0.61 to 1.15, p=0.276) for active treatment versus placebo, 0.76 (0.53 to 1.10, p=0.146) for IV versus nebuliser, 0.73 (0.51 to 1.04, p=0.083) for IV versus placebo, and 0.96 (0.65 to1.40, p=0.819) for nebuliser versus placebo. The mean differences in improvement in VAS were 0.0mm (95% CI -1.9 to 1.9mm, p=0.999) for active treatment versus placebo, 5.1mm (0.8 to 9.4mm, p=0.019) for IV versus nebuliser, 2.6mm (-1.6 to 6.8mm, p=0.231) for IV versus placebo, and -2.6mm (-7.0 to 1.8mm, p=0.253) for nebuliser versus placebo (a positive value indicates a greater improvement than the comparator). Further analyses were run with plausible imputations for the 108 (10%) patients with no 2-hour change in VAS recorded; these had no material impact on the findings.

Table 4 shows the analysis of PEFR as a percentage of predicted PEFR. The mean differences in improvement in % predicted PEFR were -0.5% (95% CI -2.9 to 1.9%, p=0.676) for active treatment versus placebo, 0.3% (-2.4 to 3.0%, p=0.841) for IV versus nebuliser, -0.4% (-3.0 to 2.3%, p=0.786) for IV versus placebo, and -0.6% (-3.4 to 2.1%), p=0.652) for nebuliser versus placebo (a positive value indicates a greater improvement than the comparator). There were no significant differences in any comparisons of physiological measures. Full details of physiological measures and oxygen flow rates are provided in supplemental tables 3 to 8.

Table 5 shows analysis of length of stay, admission to the ICU or HDU, and use of respiratory support. There was no significant difference between the three groups in these outcomes. Figure 2 shows the proportion of patients in hospital by treatment group as a function of time from hospital admission. Any small difference between the groups had disappeared by 24 hours.

Table 6 shows adverse events and side effects. The odds ratios for suffering any side effect were 1.68 (95% CI 1.11 to 2.52, p=0.014) for active treatment versus placebo, 1.00 (0.66 to 1.52, p=0.988) for IV versus nebuliser, 1.68 (1.07 to 2.63, p=0.025) for IV versus placebo, and 1.67 (1.05 to 2.66, p=0.031) for nebuliser versus placebo. Table 7 shows the medications prescribed to patients discharged after emergency department treatment. Most patients received prednisolone and a few received additional inhalers.

There were no significant findings on pre-planned subgroup analysis. In particular, the odds ratio for hospital admission with IV treatment versus placebo was 0.76 (95% CI 0.45 to 1.30) in patients presenting with more severe asthma (defined as PEFR<=50% of predicted) and 0.67 (95% CI 0.42 to 1.06) in those presenting with less severe asthma.

Discussion

The 3Mg trial is the largest trial of magnesium sulphate ever undertaken in acute asthma, the first powered on the basis of detecting a meaningful difference in admission to hospital and the only trial to directly compare IV to nebulised treatment. We were unable to demonstrate a clinically worthwhile benefit from either IV or nebulised magnesium sulphate compared to placebo. There was some evidence of IV magnesium sulphate having an effect upon hospital admission and the confidence interval for this estimate includes the possibility of both a worthwhile effect and no effect, but any effect on breathlessness was smaller than the minimum clinically significant difference14. There was no suggestion of an effect from nebulised magnesium sulphate in either primary outcome.

Meta-analysis of previous trials suggested evidence of benefit from both IV and nebulised magnesium sulphate10. This contrasts with our findings of no benefit from nebulised treatment and weak evidence of benefit from IV treatment. There are a number of potential reasons for this inconsistency. Meta-analysis of previous trials may be subject to publication bias if positive trials are preferentially submitted and accepted for publication. Some previous trials may have been limited by inadequate allocation concealment or blinding that inflated estimates of treatment effects. All three arms of the 3Mg trial received treatment with nebulised β-agonists which may have limited the potential for magnesium sulphate to provide additional bronchodilatation, whereas it was not always clear that all patients received optimal standard treatment in previous trials. In this respect it is worth noting that patients in the control arm showed marked improvements in peak expiratory flow rate and VAS breathlessness, and few required respiratory support, indicating a good response to standard treatment alone.