Intravenous or nebulised magnesium sulphate versus standard therapy for severe acute asthma (3Mg trial): a double-blind, randomised controlled trial

Summary Background Previous studies suggested intravenous or nebulised magnesium sulphate (MgSO 4 ) might improve respiratory function in patients with acute asthma. We aimed to determine whether intravenous or nebulised MgSO 4 improve symptoms of breathlessness and reduce the need for hospital admission in adults with severe acute asthma . Methods In our double-blind, placebo-controlled trial, we enrolled adults (aged ≥ 16 years) with severe acute asthma at emergency departments of 34 hospitals in the UK. We excluded patients with life-threatening features or contraindication to study drugs. We used a central randomisation system to allocate participants to intravenous MgSO 4 (2 g in 20 min) or nebulised MgSO 4 (three 500 mg doses in 1 h) alongside standard therapy including salbutamol, or placebo control plus standard therapy alone. We assessed two primary outcome measures in all eligible participants who started treatment, according to assigned treatment group: the proportion of patients admitted to hospital within 7 days and breathlessness measured on a 100 mm visual analogue scale (VAS) in the 2 h after initiation of treatment. We adjusted for multiple testing using Simes’s method. The trial stopped before recruitment was completed because funding expired. This study is registered, number ISRCTN04417063. Findings Between July 30, 2008, and June 30, 2012, we recruited 1109 (92%) of 1200 patients proposed by the power calculation. 261 (79%) of 332 patients allocated nebulised MgSO 4 were admitted to hospital before 7 days, as were 285 (72%) of 394 patients allocated intravenous MgSO 4 and 281 (78%) of 358 controls. Breathlessness was assessed in 296 (89%) patients allocated nebulised MgSO 4 , 357 (91%) patients allocated intravenous MgSO 4 , and 323 (90%) controls. Rates of hospital admission did not diﬀ er between patients treated with either form of MgSO 4 compared with controls or between those treated with nebulised MgSO 4 and intravenous MgSO 4 . Change in VAS breathlessness did not diﬀ er between active treatments and control, but change in VAS was greater for patients in the intravenous MgSO 4 group than it was in the nebulised MgSO 4 group (5·1 mm, 0·8 to 9·4; p=0·019). Intravenous or nebulised MgSO 4 did not signiﬁ cantly decrease rates of hospital admission and breathlessness compared with placebo: intravenous MgSO 4 was associated with an odds ratio of 0·73 (95% CI 0·51 to 1·04; p=0·083) for hospital admission and a change in VAS breathlessness of 2·6 mm (–1·6 to 6·8; p=0·231) compared with placebo; nebulised MgSO 4 was associated with an odds ratio of 0·96 (0·65 to 1·40; p=0·819) for hospital admission and a change in VAS breathlessness of –2·6 mm (–7·0 to 1·8; p=0·253) compared with placebo. ﬁ ndings suggest nebulised MgSO 4 has no role in the management of severe acute asthma in adults and at best suggest only a limited role for intravenous MgSO 4 in this setting.


Introduction
Acute asthma leads to about 60 000 hospital admissions per year in England. 1 Present guidelines 2,3 advise a stepwise approach to the management of exacerbations. Initially, all patients should receive oxygen, nebulised β 2 agonists, a nebulised anticholinergic drug, and corticosteroids. However, bronchodilators act within minutes whereas corticosteroids require hours to take eff ect. This diff erence suggests a potential role for magnesium sulphate (MgSO 4 ) as an additional treatment option in the therapeutic gap between nebulised bronchodilators and corticosteroids. MgSO 4 has been assessed in intravenous and nebulised forms. The nebulised route off ers the potential advantage of a quick onset of action and reduced incidence of side-eff ects. Its disadvantages include a reduced dose of drug delivered compared with the intravenous form and respiratory eff ort on the part of the patient to increase its eff ectiveness. The intravenous route provides direct access to the venous system, allowing the delivery of high drug concentrations. Disadvantages include the need for intravenous access and drug administration by infusion lasting about 20 min.
Several systematic reviews and meta-analyses have assessed the role of intravenous or nebulised MgSO 4 in acute asthma. [4][5][6][7][8][9][10] The most recent review 10 suggested that intravenous treatment seemed eff ective in children but was unable to draw clear conclusions about treatment in adults. Both intravenous treatment (assessed in ten trials, with 955 adults) and nebulised treatment (seven trials, 430 adults) were associated with weak evidence of improved respiratory function compared with control populations all treated with standard care. No trials directly compared intravenous MgSO 4 with nebulised MgSO 4 . The standardised mean diff erence (SMD) in respiratory function for intravenous 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 intravenous treatment was associated with weak evidence of an eff ect on hospital admission (relative risk [RR] 0·68, 95% CI 0·46 to 1·02; p=0·06), whereas nebulised treatment was associated with no signifi cant eff ect (0·87, 0·70 to 1·08; p=0·22) compared with standard care. One further trial of intravenous MgSO 4 in adults 11 has since been published. Inclusion of this trial in the metaanalysis 12 resulted in a slightly larger and signifi cant eff ect on respiratory function (SMD 0·35, 95% CI 0·06 to 0·64; p=0·02) but the eff ect on hospital admission remained non-signifi cant (RR 0·85, 95% CI 0·68 to 1·06; p=0·14). Whether changes in measures of respiratory function were associated with important changes in management of patients or a clinically meaningful improvement in symptoms was unclear.
Uncertainty in the evidence is shown in treatment recommendations. Current guidelines in the UK 2 and the USA 3 suggest that intravenous MgSO 4 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 MgSO 4 .
We aimed to assess the eff ectiveness of intravenous and nebulised MgSO 4 in adults with severe acute asthma. We specifi cally aimed to determine whether intravenous or nebulised MgSO 4 , used alongside standard treatment including salbutamol, reduces the proportion of patients requiring hospital admission at initial presentation or during the subsequent 7 days, and whether intravenous or nebulised MgSO 4 improves patient-assessed levels of breathlessness up to 2 h after the start of treatment.

Study design and patients
We undertook a multicentre, double-blind, placebocontrolled, three-arm, randomised trial at 34 emergency departments in the UK. The trial protocol was published previously.
Eligible patients were adults (aged ≥16 years) attending an emergency department with severe acute asthma (ie, acute asthma with either a peak expiratory fl ow rate of <50% of best or predicted, respiratory rate >25 breaths per min, heart rate >110 beats per 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 eff ort, bradycardia, arrhythmia, hypo tension, exhaustion, coma, or confusion), a contrain dication to either nebulised or intravenous MgSO 4 (pregnancy, hepatic or renal failure, heart block, or known hypermagnesaemia), individuals who were unable to provide written or verbal consent, and previous participants in the 3Mg trial. We amended the protocol during the trial to also exclude individuals who had received MgSO 4 in the 24 h before recruitment. We sought written or verbal consent from all participants. Patients who initially provided verbal consent were asked for written consent as soon as their condition permitted.
An independent data monitoring committee reviewed trial data at regular intervals and reported re commendations to the trial steering committee in accordance with the data monitoring committee charter. The trial was approved by the Scotland A Research Ethics Committee.

Randomisation and masking
We randomly allocated participants with a telephone or internet randomisation system, which was managed by 339   the Sheffi eld Clinical Trials Research Unit (CTRU). After entry into the trial, participants were allocated to numbered treatment packs kept in the emergency department. We used a simple randomisation sequence in the fi rst 20 participating hospitals, but switched to blocked randomisation (block sizes of four or six), stratifi ed by hospital, for subsequent hospitals to safeguard against new centres recruiting too few participants in any trial arm. Every treatment pack contained an intravenous infusion and three nebuliser solutions, either of which could be active treatment or placebo. Participants, hospital staff , and research staff were masked to allocated treatment.

Procedures
Patients were allocated to receive one of three treatments: intravenous MgSO 4 (8 mmol [2 g] in 100 mL normal saline provided over 20 min) and three 7·5 mL vials of 0·9% saline nebulised at 20 min intervals (intravenous MgSO 4 group); intravenous normal saline (100 mL given over 20 min) and three 7·5 mL vials of 2 mmol (500 mg) MgSO 4 nebulised at 20 min intervals (nebulised MgSO 4 group); or intravenous normal saline (100 mL given over 20 min) and three 7·5 mL vials of 0·9% saline nebulised at 20 min intervals (placebo group). Patients received standard therapy in accordance with guidelines 2 from the British Thoracic Society and Scottish Intercollegiate Guidelines Network and consisted of oxygen, nebulised salbutamol (5 mg), nebulised ipratropium (500 μg), and oral prednisolone administered during recruitment, followed by up to 5 mg salbutamol added to each trial nebuliser. Other treatments were provided at the discretion of the clinician. Patients were managed in the emergency department and data were collected until 2 h after randomisation. At this point, if not already undertaken, a fi nal disposition decision was made (hospital admission or discharge) and initial data collection was completed.
We prespecifi ed two primary outcomes. The fi rst was a health service primary outcome, defi ned as the proportion of patients admitted to hospital, either after emergency department treatment or at any time in the subsequent 7 days. The second was a patient-centred primary outcome, defi ned as the patient's visual analogue scale (VAS) for breathlessness in the 2 h after start of treatment. VAS breathlessness has been used to measure breathlessness during exercise 13 and has been shown to correlate with respiratory function and symptomatic change in cohorts with acute asthma. 14,15 Secondary outcomes included mortality, adverse events, use of ventilation or respiratory support, length of hospital stay, admission to a high-dependency unit or intensive-care unit, change in peak expiratory fl ow rate and physiological variables (oxygen saturation, heart rate, respiratory rate, blood pressure) over 2 h, change in quality of life between baseline and 1 month, number of unscheduled health-care contacts over the subsequent month, and satisfaction with care (these outcomes will be reported elsewhere).
Treating clinicians routinely recorded adverse events and side-eff ects occurring during emergency department treatment on case report forms. Key events (cardiac arrest, respiratory arrest, emergency intubation, noninvasive ventilation, pneumothorax, and arrhythmia) Nebulised MgSO 4 (n=332)

Statistical analysis
We planned to recruit 1200 participants (400 patients per group). Assuming 80% of patients with severe acute asthma were 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 (ie, to 70%) for any pair of treatment groups compared (twosided α=0·05). Assuming 80% of participants had a VAS measurement, then the study would have 90% power to detect an 8 mm diff erence in a 100 mm VAS at 2 h after treatment initiation (two-sided α=0·05). Previous data have established that the standard deviation on a 100 mm VAS is 30 mm, and that 22 mm represents a minimum clinically signifi cant diff erence. 14 We analysed participants in the groups to which they were allocated, irrespective of whether they actually received or completed the allocated treatment. We used logistic regression for analysis of admission rates. For length of stay, we compared means with censored normal regression and medians with log-normal regression to account for interval censoring in discharged patients (for whom no time of discharge was recorded) and also admissions that were ongoing at 30 days. We compared the number of days spent in the intensive-care unit or high-dependency unit with the Mann-Whitney U test. We used ANCOVA for assessment of all other outcomes. We assessed the primary outcome in all eligible patients who started treatment, adjusted for hospital of admission. We also did additional analyses with diff erent imputation strategies as con fi rmatory analyses. We did a secondary explanatory analysis restricted to individuals who completed the treatment as per protocol. We used Simes's method, 16 which is a modifi cation of the Bonferroni method with increased power, to adjust for multiplicity arising from use of two primary outcomes. The two preplanned comparisons between the three groups were active treatment (intravenous or nebulised) versus placebo and intravenous MgSO 4 versus nebulised MgSO 4 . We also present comparisons of intravenous MgSO 4 versus placebo and nebulised MgSO 4 versus placebo for completeness. We undertook three preplanned subgroup analyses assessing the primary outcomes (hospital admission and VAS breathlessness) between active and placebo groups stratifi ed by age (≥50 years vs <50 years), baseline peak expiratory fl ow rate (less than median vs median or greater), and previous treatment with salbutamol before the trial treatments (yes vs no).
This study is registered, number ISRCTN04417063.

Role of the funding source
The sponsor of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all the data in the study and had fi nal responsibility for the decision to submit for publication.
We noted no signifi cant diff erences on pre-planned subgroup analysis. Rates of hospital admission with intravenous MgSO 4 did not diff er from placebo (odds ratio 0·76, 95% CI 0·45-1·30; p=0·332) in patients presenting with more severe asthma (defi ned as median peak expiratory fl ow rate or lower) or in individuals presenting with less severe asthma (0·67, 0·42-1·06; p=0·088). Data for rates of hospital admission split by age group and previous salbutamol treatment will be reported elsewhere.

Discussion
To our knowledge, the 3Mg trial is the largest trial of MgSO 4 undertaken in acute asthma, the fi rst trial powered to detect a meaningful diff erence in rates of admission to hospital, and the fi rst to directly compare intravenous treatment with nebulised treatment (panel). We did not show a clinically meaningful benefi t from either intravenous or nebulised MgSO 4 compared with placebo. Intravenous MgSO 4 might have an eff ect on rates of hospital admission and the confi dence interval for this estimate included the possibility of both a worthwhile eff ect and no eff ect, but any eff ect we noted on breathlessness was smaller than the minimum clinically signifi cant diff erence. 13 We noted no suggestion of an eff ect from nebulised MgSO 4 in either primary outcome.
Meta-analysis of previous trials suggested evidence of benefi t from intravenous and nebulised MgSO 4 . 10 This suggestion contrasts with our fi ndings of no benefi t from nebulised treatment and weak evidence of benefi t from intravenous treatment. Several factors might explain this inconsistency. Meta-analysis can be subject to publication bias if positive trials are preferentially submitted and accepted for publication. Some previous trials might have been restricted by inadequate allocation concealment or masking that infl ated estimates of treatment eff ects. Patients in all three arms of the 3Mg trial received treatment with nebulised β agonists which might have restricted the ability of MgSO 4 to provide additional bronchodilation, whereas it was not always clear that all patients received optimum standard treatment in previous trials. Notably, patients in the control group of 3Mg had improvements in peak expiratory fl ow rate and VAS breathlessness, and few required respiratory support, suggesting a good response to standard treatment alone.
One potential explanation that can probably be discounted is that the 3Mg trial treatment was inadequate, in terms of the planned dose and actual amount of drug given. The protocol-specifi ed doses of intravenous and nebulised MgSO 4 were at the top end of doses used in previous trials, and most patients received the full dose Total number of events will not equal the sum of individual events if a patient has more than one side-eff ect.  of the relevant drugs (appendix). Pragmatic trials carry a risk that trial treatment will be delivered in a suboptimal manner, but we noted no evidence of this in the 3Mg trial.
Our fi ndings for nebulised MgSO 4 contrast with those of the MAGNETIC trial, 17 which showed an improvement in asthma severity score at 60 min after treatment with nebulised MgSO 4 compared with placebo in children and adolescents with acute severe asthma. 17 Previous metaanalysis 10 suggested that intravenous MgSO 4 is more eff ective in children than adults. Findings from 3Mg and MAGNETIC suggest that the same is true of nebulised MgSO 4 .
The 3Mg trial had some limitations. The trial terminated when funding expired and had recruited 1109 patients (92%) of a target of 1200. Despite this shortfall, the trial had 84% power to detect a 10% diff erence in admission rate for nebulised treatment versus placebo and 87% power for intravenous MgSO 4 versus placebo based on the original sample-size projections. Furthermore, VAS breathlessness was recorded for 90% of the study population, as opposed to the anticipated 80% in the power calculation, so there was no loss of power to detect a diff erence in this outcome.
3Mg was designed as a pragmatic trial to determine the eff ectiveness of use of MgSO 4 alongside other treatments as part of routine emergency department practice. The study population was pragmatically defi ned by use of information routinely available to emergency department staff . Thus, fi ndings should be generalisable to typical adult patients attending hospital with acute asthma, but the design also means that the study population could have included some patients with other diagnoses. We assessed MgSO 4 alongside standard treatment rather than comparing it to elements of standard treatment. This design might have reduced the potential for MgSO 4 to make a clinically meaningful diff erence, but withholding standard treatment would have been unethical. We selected primary outcomes that measured the eff ect of treatment on symptoms (VAS breathlessness) and management (hospital admission). We also measured physiological parameters and peak expiratory fl ow rate as secondary outcomes. Other measures, such as forced expiratory volume in 1 s, might have been more sensitive to changes in respiratory function, but these are not routinely measured in the emergency department, and would not provide evidence of clinical eff ectiveness. Demonstration of clinical eff ectiveness requires a meaningful improvement in symptoms or management of patients, not just a change in respiratory parameters. Finally, we deliberately excluded patients with life-threatening asthma and were unable to power the study to detect diff erences in serious adverse outcomes (including death), so we were unable to determine whether MgSO 4 has an eff ect on serious adverse outcomes in lifethreatening asthma.
The fi ndings of our trial suggest that there is no role for nebulised MgSO 4 in the management of severe acute asthma in adults and at best a limited role for intravenous MgSO 4 in this setting. Patients receiving standard treatment had striking improvements in rates of breathlessness and peak expiratory fl ow, and few required respiratory support. Although most patients were admitted to hospital, nebulised MgSO 4 did not reduce the admission rate and we noted only weak evidence of an eff ect from intravenous MgSO 4 . The low rate of sideeff ects and adverse events (other than those related to the underlying illness) suggests a low risk of harm from intravenous administration but the corresponding evidence of benefi t is modest and uncertain.
Further clinical trials of MgSO 4 in adults with acute asthma are unlikely to be worthwhile. If intravenous treatment has an eff ect on admission rates or adverse events that was not detected by 3Mg then a much larger trial would be needed to detect such an eff ect. The logistic barriers to undertaking clinical trials in patients with a medical emergency would seem to prevent a larger trial being feasible at an acceptable cost.

Contributors
The coapplicants (see appendix) designed the trial and developed the research proposal and submitted it for funding. The Project Management Group and Local Investigators undertook the trial with independent oversight from the trial steering committee and data monitoring committee. SG wrote the fi rst draft of the report. MB did the statistical analysis. SG, MB, JC, AG, TC, and JB contributed to redrafting of the paper and approved the fi nal draft.

Confl icts of interest
We declare that we have no confl icts of interest.

Panel: Research in context
Systematic review A 2007 systematic review, 10 which was updated in 2009, 12 identifi ed 11 trials of intravenous MgSO 4 in 1018 adults and seven trials of nebulised MgSO 4 in 430 adults with acute asthma. Meta-analysis suggested that both intravenous and nebulised treatment had potentially worthwhile eff ects on respiratory function and showed non-signifi cant trends towards reduced rates of admission to hospital.

Interpretation
Our large pragmatic study failed to provide convincing evidence that intravenous or nebulised MgSO 4 produce clinically worthwhile benefi ts in adults with severe acute asthma. Although MgSO 4 is a safe treatment with few signifi cant side-eff ects, current data do not support a role for MgSO 4 in the standard treatment of adults with severe acute asthma.