Sulfato de magnesio intravenoso para el tratamiento de adultos con asma aguda en el servicio de urgencias
Resumen
Antecedentes
El asma es una afección respiratoria crónica caracterizada por la inflamación de las vías respiratorias, la constricción del músculo liso de las vías respiratorias y una alteración estructural de las vías respiratorias que es al menos parcialmente reversible. Las exacerbaciones del asma pueden ser potencialmente mortales y dan lugar a una carga significativa en los servicios de asistencia sanitaria. Se han publicado diversas guías para informar al personal que lleva a cabo el tratamiento en el contexto agudo; varias incluyen la administración de un único bolo de sulfato de magnesio intravenoso (MgSO4 IV) en los casos que no responden al tratamiento de primera línea. Sin embargo, aún no se conoce la efectividad de este enfoque, en particular en los casos menos graves.
Objetivos
Evaluar la seguridad y la eficacia del MgSO4 IV en adultos que reciben tratamiento para el asma agudo en el servicio de urgencias.
Métodos de búsqueda
Se identificaron ensayos del registro especializado del Grupo Cochrane de Vías Respiratorias (Cochrane Airways Review Group) (CAGR) hasta el 2 mayo 2014. También se hicieron búsquedas www.ClinicalTrials.gov y listas de referencias de otras revisiones, y se estableció contacto con los autores de los ensayos para solicitar información adicional.
Criterios de selección
Se incluyeron ensayos controlados aleatorios (ECA) en adultos que recibieron tratamiento para las exacerbaciones del asma en el servicio de urgencias (SU) cuando comparaban cualquier dosis de MgSO4 IV con placebo.
Obtención y análisis de los datos
Todos los autores de la revisión examinaron los títulos y resúmenes para la inclusión, y por lo menos dos autores extrajeron de forma independiente las características de los estudios, el riesgo de sesgo y datos numéricos. Los desacuerdos se resolvieron por consenso, y se estableció contacto con los investigadores de los ensayos para obtener la información que faltaba.
Los datos dicotómicos se analizaron como odds ratios mediante el uso de los participantes en estudio como la unidad de análisis, y los datos continuos se analizaron como diferencias de medias o diferencias de medias estandarizadas mediante modelos de efectos fijos. Todos los resultados se calificaron mediante GRADE y los hallazgos se presentaron en la tabla 1 de Resumen de los hallazgos.
Se realizaron análisis de subgrupos sobre el resultado primario para la gravedad inicial de las exacerbaciones y sobre si el bromuro de ipratropio se había administrado como comedicación. Los datos no publicados y los estudios en riesgo alto de sesgo en cuanto al cegamiento se extrajeron del análisis principal en los análisis de sensibilidad.
Resultados principales
Catorce estudios reunieron los criterios de inclusión, los cuales asignaron al azar a 2313 personas con asma aguda a las comparaciones de interés en esta revisión.
La mayoría de los estudios eran ensayos doble ciego que comparaban una única infusión de 1,2 g o 2 g de MgSO4 IV durante 15 a 30 minutos versus un placebo similar. Once se realizaron en un único centro, y tres fueron ensayos multicéntricos. Los participantes en casi todos los estudios ya habían recibido al menos oxígeno, beta2‐agonistas de acción breve nebulizados y corticosteroides IV en el SU; en algunos estudios, los investigadores también administraron bromuro de ipratropio. Diez estudios incluyeron sólo adultos y cuatro incluyeron adultos y niños; los mismos fueron incluidos debido a que la media de edad de los participantes fue de 18 años.
El MgSO4 intravenoso redujo los ingresos al hospital en comparación con placebo (odds ratio [OR] 0,75; intervalo de confianza [IC] del 95%: 0,60 a 0,92; I2 = 28%; valor de p = 0,18; n = 972; pruebas de alta calidad). En términos absolutos, este odds ratio se traduce en una reducción de siete ingresos al hospital por cada 100 adultos tratados con MgSO4 IV (IC del 95% dos a 13 menos). La prueba para las diferencias entre subgrupos no reveló ninguna heterogeneidad estadística entre los tres subgrupos de gravedad (I2 = 0%, valor de p 0,73) o entre los cuatro estudios que administraron bromuro de ipratropio nebulizado como comedicación y los que no lo administraron (I2 = 0%, valor de p 0,82). Los análisis de sensibilidad en los que los datos no publicados y los estudios en alto riesgo en cuanto al cegamiento se eliminaron del análisis primario no cambiaron las conclusiones.
Dentro de los resultados secundarios, las pruebas de calidad alta y moderada a través de tres índices espirométricos indican algunas mejorías en la función pulmonar con la administración de MgSO4 IV. No se encontraron diferencias entre el MgSO4 IV y el placebo para la mayoría de los resultados secundarios no espirométricos, de los cuales todos se consideraron de calidad baja o moderada (ingresos a la unidad de cuidados intensivos, duración del tratamiento en el SU, duración de la estancia hospitalaria, reingresos, tasa de respiración, presión arterial sistólica).
Los eventos adversos se informaron de forma inconsistente y no se realizó el metanálisis. Los eventos adversos citados más comúnmente en los grupos de MgSO4 IV fueron rubor, fatiga, náuseas y cefalea e hipotensión (presión arterial baja).
Conclusiones de los autores
Esta revisión aporta pruebas de que una única infusión de 1,2 g o 2 g de MgSO4 IV durante 15 a 30 minutos reduce los ingresos al hospital y mejora la función pulmonar en adultos con asma agudo que no han respondido de forma suficiente al oxígeno, los beta2‐agonistas de acción breve nebulizados y los corticosteroides IV. Las diferencias en la forma en que se realizaron los ensayos dieron lugar a dificultades para los revisores en cuanto a la evaluación de si la gravedad de la exacerbación o la comedicación adicional alteraron el efecto del tratamiento con MgSO4 IV. Se encontraron pruebas limitadas sobre otras medidas de beneficio y seguridad.
Los estudios realizados en estas poblaciones deben definir claramente los parámetros de gravedad inicial y registrar sistemáticamente los eventos adversos. Los estudios que seleccionan a participantes con exacerbaciones de gravedad variable deben considerar los resultados del subagrupamiento sobre la base de las clasificaciones aceptadas de la gravedad.
PICO
Resumen en términos sencillos
¿Las infusiones de sulfato de magnesio reducen la necesidad de ingreso al hospital en adultos con asma aguda?
¿Por qué es importante esta pregunta?
El asma es un trastorno a largo plazo que causa tos, sibilancias, disnea y opresión torácica. Cuando los síntomas empeoran significativamente, a menudo denominado ataque o “exacerbación”, este evento puede ser potencialmente mortal. El tratamiento de las exacerbaciones en el servicio de urgencias (SU) es variable, y algunas guías recomiendan la administración de sulfato de magnesio intravenoso (MgSO4 IV) cuando otros tratamientos no han ayudado. Sin embargo, no está claro si el MgSO4 IV es efectivo, en particular en los casos menos graves, y se deseaba responder a esta pregunta.
¿Cómo se respondió a la pregunta?
Se realizaron búsquedas de ensayos que comparaban MgSO4 IV versus placebo en adultos que asistían al SU con una exacerbación del asma. Las búsquedas más recientes se realizaron el 2 de mayo de 2014. El interés se centró principalmente en si el MgSO4 IV redujo el número de pacientes que necesitaban ingresar al hospital, y también se consideraron otras medidas, incluido el tiempo en el SU, la función pulmonar y las puntuaciones de los síntomas.
¿Qué se encontró?
Catorce estudios cumplieron los criterios de inclusión e incluyeron un total de 2313 personas. Estos estudios variaron con respecto a cuán graves debían ser las exacerbaciones para que los pacientes estuvieran incluidos y en cuanto a cuáles tratamientos se administraron antes del MgSO4 IV, aunque en casi todos los ensayos los participantes recibieron al menos oxígeno, fármacos de acción breve nebulizados y comprimidos o inyección de esteroides.
En términos generales, el MgSO4 IV redujo la necesidad de ingreso al hospital en comparación con placebo (siete menos por 100 tratados; intervalo de confianza del 95% dos a 13 menos). No hubo información suficiente disponible para mostrar si la reducción de los ingresos al hospital se asoció con la gravedad de la exacerbación del asma, o si se logró una diferencia al administrar otros tratamientos. Las pruebas indican que el MgSO4 IV mejoró algunos parámetros de la función pulmonar, aunque para otras medidas como la frecuencia cardíaca, la variación entre los hallazgos de los estudios redujo la confianza en los resultados. No se encontró una diferencia entre el MgSO4 IV y el placebo en la mayoría de las medidas restantes (incluido el tiempo en el SU, la frecuencia respiratoria y la presión arterial), y en general los eventos adversos se informaron de forma deficiente.
Conclusion
Esta revisión indicó que el MgSO4 IV reduce los ingresos al hospital y mejora la función pulmonar en adultos con exacerbaciones del asma en los casos en que otros fármacos de primera línea no han aliviado los síntomas agudos (es decir oxígeno, fármacos de acción breve inhalados y esteroides IV). Las pruebas para otras medidas de beneficio y seguridad fueron limitadas.
Los investigadores deben definir claramente la gravedad del asma entre los pacientes de los estudios mientras registran de forma cuidadosa los eventos adversos.
Este resumen en términos sencillos está actualizado hasta mayo de 2014.
Authors' conclusions
Summary of findings
| IV MgSO4 for treating adults with acute asthma in the ED | |||||
| Patient or population: adults with acute asthma Comparions: placebo Both intervention and placebo groups received oxygen, short‐acting beta2‐agonists and oral or intravenous steroids before the infusion. Measurements were taken between 60 and 240 minutes after the start of the infusion. | |||||
| Outcomes | Illustrative comparative risks* (95% CI) | Relative effect | No. of participants | Quality of the evidence | |
| Assumed risk | Corresponding risk | ||||
| Control | IV MgSO4 | ||||
| Hospital admissions | 569 per 1000 | 498 per 1000 | OR 0.75 | 1769 | ⊕⊕⊕⊕ |
| Intensive care unit (ICU) admissions | 14 per 1000 | 28 per 1000 | OR 2.03 | 752 | ⊕⊕⊝⊝ |
| Length of hospital stay (days) | Mean length of hospital stay in the control groups was | Mean length of hospital stay in the intervention groups was | ‐ | 949 | ⊕⊕⊝⊝ |
| ED treatment duration (minutes) | Mean duration in the placebo group was 228 minutes | Mean ED treatment duration in the intervention groups was | ‐ | 96 | ⊕⊕⊝⊝ |
| FEV1 (% predicted) | Mean FEV1 in the placebo group was 50% predicted | Mean FEV1 (% predicted) in the intervention groups was | ‐ | 523 | ⊕⊕⊕⊕ |
| PEF (L/min) | Mean PEF in the placebo group was 239 L/min | Mean PEF in the intervention groups was | ‐ | 1460 | ⊕⊕⊕⊝ |
| Respiratory rate (breaths/min) | Mean respiration rate in the placebo group was 20.7 respirations/min | Mean respiratory rate in the intervention groups was | ‐ | 1195 | ⊕⊕⊕⊝ |
| *The basis for the assumed risk (e.g. median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | |||||
| GRADE Working Group grades of evidence. | |||||
| 1One study (Green 1992) introduced risk of bias, but the rest of the studies were generally well conducted. | |||||
Background
Description of the condition
Asthma is a chronic respiratory condition characterised by airway inflammation, constriction of airway smooth muscle and structural alteration of the airways that is at least partially reversible. Common symptoms include cough, wheezing, difficulty breathing, reduced exercise tolerance and chest tightness. Common triggers include allergens, pollutants and viral infections, although endogenous factors have also been identified. The World Health Organization (WHO) recognises the global burden of asthma and estimates a worldwide prevalence of 300 million people of all ages, with 250,000 dying each year. Epidemiological data suggest that prevalence is greatest in the developed world, with prevalence amongst adults at 8.2% in the USA (CDC) and 9% to 10% in the UK (DOH 2012).
Asthma can present with varying degrees of severity, and in the most severe cases, it can cause daily chronic symptoms and frequent exacerbations (defined as acute worsening of asthma symptoms). Overarching principles of treatment focus on controlling daily symptoms and preventing exacerbations through good education and appropriate use of inhalers. Short‐acting bronchodilators are given to relieve bronchospasm, and corticosteroids for the underlying inflammation; both are usually delivered via inhalers. Depending on the persistence of symptoms, inhalers can be taken regularly (maintenance therapy) or on an as‐needed basis (reliever therapy) (BTS/SIGN 2012; GINA 2011). Treatment guidelines recommend preventative management in the community and prompt interventions during exacerbations to reduce mortality and other negative outcomes (such as intubation and hospital admissions).
Description of the intervention
In severe exacerbations of asthma, which can be life threatening, most guidelines recommend the use of oxygen, nebulised or intravenous beta2‐agonists, nebulised antimuscarinics and intravenous or oral corticosteroids as first‐line treatment (BTS/SIGN 2012; GINA 2011; NACA 2006; NAEPP 2007). Beta2‐agonists are recognised as most effective in relieving bronchospasm (Teoh 2012); however, anticholinergic inhalers have also been shown to be effective in the treatment of acute asthma (Griffiths 2013). When patients show poor response to these, or when they present with a severe or life‐threatening exacerbation, a single dose of intravenous (IV) or nebulised magnesium sulfate (MgSO4) can be considered. Nebulised MgSO4 is the subject of a separate review (Powell 2012). The recommended dosage of IV MgSO4 in the UK is 1.2 g to 2 g, delivered by infusion over 20 minutes (BTS/SIGN 2012), but guidelines differ regarding how and when IV MgSO4 should be administered (Table 1),
| BTS/SIGN | GINA | NACA | NAEPP | |
| Oxygen | ✓ | ✓ | ✓ | ✓ |
| Inhaled beta2‐agonist | ✓ | ✓ | ✓ | ✓ |
| Inhaled antimuscarinic | ✓ | ✓ | ✓ | ✓ |
| Sytemic corticosteroids | ✓ | ✓ | ✓ | ✓ |
| IV beta2‐agonist | (✓) if nebulised form cannot be used reliably | x | ✓ if no response to inhaled form | x |
| IV MgSO4 | ✓ | ✓ | ✓ IV or nebulised | ✓ |
| Heliox | x | x | x | ✓ |
| IV aminophylline/theophylline | (✓) limited evidence, only after senior consultation | (✓) if inhaled beta2‐agonist unavailable | (✓) as an alternative to IV beta2‐agonist | x |
BTS/SIGN: British Thoracic Society and Scottish Intercollegiate Guidelines Network joint guideline; GINA: Global Initiative for Asthma; IV: Intravenous; NACA: National Asthma Council Australia; NAEPP: National Asthma Education and Prevention Program; ✓: Recommended; x: Not recommended; (✓): Recommended with conditions.
National guidelines also vary with respect to definitions of asthma severity and use of additional interventions. Table 1 offers a summary of treatment strategies recommended by some of these guidelines for the management of acute asthma.
How the intervention might work
Magnesium is an important intracellular and extracellular cation that plays a key role in intracellular enzymatic reactions. Its mechanism of action in the context of an exacerbation of asthma is not fully understood, but several theories have been proposed (Rowe 2013). It is believed to play a role in bronchial smooth muscle relaxation via its ability to prevent calcium ion movement into smooth muscle cells by blocking the voltage‐dependent calcium channels (Gourgoulianis 2001; Spivey 1990). Furthermore, some evidence suggests that it may reduce the neutrophilic burst seen with the inflammatory response (Cairns 1996), and that it may be involved in acetylcholine release from cholinergic nerve terminals and histamine release from mast cells (Dominguez 1998). The combination of these properties contributes to relief of airflow obstruction and provides the theoretical basis for the effectiveness of magnesium.
Why it is important to do this review
Acute asthma presentations represent a significant burden on emergency departments (EDs) and carry a substantial mortality risk, with 1143 deaths from asthma reported in the UK in 2010 (Asthma UK) and an estimated mortality rate of 1.1 deaths per 100,000 in the USA (CDC). In the UK, it is thought that "75% of hospital admissions for asthma are avoidable and as many as 90% of the deaths from asthma are preventable" (Asthma UK).
The financial burden is also significant, with a cost to the National Health Service (NHS) of £1 billion a year, 80% of which is spent on the 20% of people with the most severe disease (DOH 2012).
Current guidelines advocate the use of IV MgSO4 in the treatment of acute severe asthma, but evidence in the literature remains inconclusive (Rowe 2009). New evidence from randomised controlled trials published since the last version of this review may alter the conclusions.
Objectives
To assess the safety and efficacy of IV MgSO4 in adults treated for acute asthma in the emergency department.
Methods
Criteria for considering studies for this review
Types of studies
We included randomised controlled trials (RCTs) of any follow‐up duration reported as full text, those published as abstract only and unpublished data.
Types of participants
We included studies of adults (defined as over 18 years of age) treated in the ED for acute asthma. If studies recruited both adults and children, we contacted the study authors to try to obtain separate data from adults.
Types of interventions
We included trials comparing any dose of IV MgSO4 versus placebo. People with acute asthma often require multiple medications; therefore we included studies that allowed other treatments (for maintenance, for exacerbation itself or for other co‐morbidities), provided they were not part of the randomly assigned treatment.
Types of outcome measures
Primary outcomes
-
Hospital admissions.
Secondary outcomes
-
ED treatment duration.
-
Intensive care unit admissions.
-
Vital signs (heart rate, respiratory rate, blood pressure, oxygen saturation).
-
Spirometry (peak expiratory flow (PEF), forced expiratory volume within one second (FEV1)).
-
Validated symptom scores.
-
Adverse events.
Reporting in the trial of one or more of the outcomes listed here was not an inclusion criterion for the review.
Search methods for identification of studies
Electronic searches
We identified trials from the Cochrane Airways Review Group Specialised Register (CAGR), which is maintained by the Trials Search Co‐ordinator for the Group. The Register contains trial reports identified through systematic searches of bibliographic databases including the Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, EMBASE, CINAHL, AMED and PsycINFO, and by handsearching of respiratory journals and meeting abstracts (see Appendix 1 for further details). We searched all records in the CAGR using the search strategy described in Appendix 2.
We also conducted a search of ClinicalTrials.gov (www.ClinicalTrials.gov) and the WHO trials portal (www.who.int/ictrp/en/). We searched all databases from their inception to the present, and we imposed no restriction on language of publication.
Searching other resources
We checked reference lists of all relevant primary studies and review articles for additional references. We also searched for errata or retractions from included studies published in full text on PubMed (www.ncbi.nlm.nih.gov/pubmed) and reported within the review the date this was done.
Data collection and analysis
Selection of studies
Three review authors (KK, LK, CM) independently screened titles and abstracts for inclusion of all citations identified by the search and coded them as 'retrieve' (eligible or potentially eligible/unclear) or 'do not retrieve.' We retrieved the full‐text study reports/publications, and the review authors independently screened the full‐text documents and identified studies for inclusion. We identified and recorded reasons for exclusion of ineligible studies. We resolved disagreements through discussion, or, if required, we consulted a fourth person. We identified and excluded duplicates and collated multiple reports of the same study, so that each study rather than each report was the unit of interest in the review. We recorded the selection process in sufficient detail to complete a PRISMA flow diagram and a Characteristics of excluded studies table.
Data extraction and management
To record study characteristics and outcome data, we used a data collection form that had been piloted on at least one study in the review. All review authors (KK, LK, CM) extracted study characteristics from included studies, and all review authors independently extracted outcome data. We extracted the following study characteristics.
-
Methods: study design, duration of observation and follow‐up, details of any 'run‐in' period, number of study centres and locations, withdrawals, dates of study.
-
Participants: N, mean age, age range, gender, asthma severity*, diagnostic criteria, co‐morbidities, co‐medications, baseline lung function, inclusion and exclusion criteria.
-
Interventions: intervention, dose, comparison, concomitant and failed treatments, excluded medications.
-
Outcomes: primary and secondary outcomes specified and collected, time points reported.
-
Notes: funding for trial, notable conflicts of interest of trial authors.
We noted in the Characteristics of included studies table if outcome data were not reported in a usable way. We resolved disagreements by reaching consensus or by involving a fourth person. One review author transferred data into the Review Manager (RevMan) (version 5.2) file. We double‐checked that data were entered correctly by comparing data presented in the systematic review versus data provided in the study reports. A second review author (LK or CM) spot‐checked study characteristics for accuracy against the trial report.
Assessment of risk of bias in included studies
All review authors independently assessed risk of bias for each study using the criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011), resolving disagreements by discussion. We assessed the risk of bias according to the following domains.
-
Random sequence generation.
-
Allocation concealment.
-
Blinding of participants and personnel.
-
Blinding of outcome assessment.
-
Incomplete outcome data.
-
Selective outcome reporting.
-
Other bias.
We graded each potential source of bias as high, low or unclear and provided a quote from the study report together with a justification for our judgement in the Risk of bias in included studies table. We summarised risk of bias judgements across different studies for each of the domains listed. We considered blinding separately for different key outcomes when necessary (e.g. for unblinded outcome assessment, risk of bias for hospital admission may be very different than for a patient‐reported scale). When information on risk of bias was related to unpublished data or correspondence with a trial author, we noted this in the Risk of bias in included studies table.
When considering treatment effects, we took into account the risk of bias for all studies that contributed to that outcome.
Assessment of bias in conducting the systematic review
We conducted the review according to this published protocol and reported deviations from it in the Differences between protocol and review section of the systematic review.
Measures of treatment effect
We analysed dichotomous data as odds ratios (ORs) and continuous data as mean differences (MDs) or standardised mean differences (SMDs) with 95% confidence intervals (CIs). If studies reported several validated symptom measures, or if different scales were reported across studies, we analysed the data as SMDs in one analysis to reduce measurement error and to increase precision. We entered data presented as a scale with a consistent direction of effect. We narratively described skewed data reported as medians and interquartile ranges.
We undertook meta‐analyses only when this was meaningful (i.e. when treatments, participants and the underlying clinical question were similar enough for pooling to make sense).
When multiple trial arms were reported in a single trial, we included only the relevant arms. When two relevant comparisons from a single study were combined in the same meta‐analysis, we halved the control group to avoid double‐counting.
Unit of analysis issues
For dichotomous outcomes, we used participants rather than events as the unit of analysis (i.e. number of adults admitted to hospital rather than number of admissions per adult).
Dealing with missing data
We contacted investigators or study sponsors to verify key study characteristics and to obtain missing numerical outcome data when possible (e.g. when a study was identified as abstract only). When this was not possible, and when missing data were thought to introduce serious bias, we conducted a sensitivity analysis to explore the impact of including such studies in the overall assessment of results.
Assessment of heterogeneity
We used the I² statistic to measure heterogeneity among the trials in each analysis. When substantial heterogeneity was identified, we explored possible causes by conducting prespecified subgroup analyses.
Assessment of reporting biases
We created and examined a funnel plot to explore possible small‐study and publication biases. We considered the impact of unpublished trials in the GRADE ratings for each outcome.
Data synthesis
We used a fixed‐effect model and performed a sensitivity analysis with random effects when significant heterogeneity was observed (I² > 30%).
Summary of findings table
We created summary of findings Table for the main comparison for seven of the prespecified outcomes. We used the five GRADE considerations (study limitations, consistency of effect, imprecision, indirectness and publication bias) to assess the quality of a body of evidence as it relates to the studies that contributed data to meta‐analyses for the prespecified outcomes (http://www.gradeworkinggroup.org/). We applied methods and recommendations described in Section 8.5 and Chapter 12 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011) using GRADEpro software. We justified all decisions to downgrade or upgrade the quality of studies by using footnotes, and we made comments to aid readers' understanding of the review when necessary.
Subgroup analysis and investigation of heterogeneity
We carried out the following subgroup analyses for the primary outcome, using the formal test for subgroup differences in Review Manager (version 5.2) (Review Manager (RevMan)).
-
Baseline severity (moderate, severe and life‐threatening exacerbations*).
-
Mean age (≤ and > 65 years).
-
Co‐medications (with or without ipratropium bromide**).
*Since there is no single accepted metric for assessment of asthma severity, we extracted baseline data relevant to the severity criteria, as stated in the British Thoracic Society (BTS) guidelines (BTS/SIGN 2012), that is,
-
Clinical features (e.g. ability to complete sentences, respiratory effort, conscious level, signs of exhaustion);
-
Previous intensive care unit admissions;
-
Pulse;
-
Blood pressure;
-
Respiratory rate;
-
Pulse oximetry;
-
Pulsed expiratory flow (PEF); and
-
Arterial blood gas.
Exacerbations of the study populations were labelled as moderate, severe or life threatening on the basis of available data, as judged by an independent assessor who was not involved in the review process and had no other details or results of the trials. Consistent with British Thoracic Society (BTS)/Scottish Intercollegiate Guidelines Network (SIGN) criteria (BTS/SIGN 2012), for which the percentage predicted PEF was available, mean values less than 33% were judged to be life threatening, 33% to 50% severe and over 50% moderate. When this measure was not available, or when the value was close to a cutoff, other criteria were consulted, and the value was then standardised across trials using studies reporting several indices. The decision to perform a subgroup analysis by severity was informed by conclusions drawn in the previous Cochrane review (Rowe 2009) that the intervention may be more effective in cases of severe or life‐threatening asthma.
**For co‐medications, we grouped studies by whether investigators gave ipratropium bromide in addition to other treatments (i.e. short‐acting beta2‐agonists (SABAs) via a nebuliser or spacer, oral or intravenous corticosteroids). Ipratropium bromide is included in most guidelines, but it is unclear whether this treatment is adopted in all EDs. Griffiths 2013 has demonstrated that it is an effective adjunct to SABAs in children with asthma exacerbation in the acute setting.
Sensitivity analysis
We plan to carry out the following sensitivity analyses.
-
Studies at high risk of bias for blinding.
-
Unpublished data.
Reaching conclusions
We have based our conclusions only on findings from the quantitative or narrative synthesis of included studies for this review. We have avoided making recommendations for practice, and our implications for research suggest priorities for future research and outline remaining uncertainties in this area.
Results
Description of studies
Full details of the conduct and characteristics of each included study can be found in Characteristics of included studies, and reasons for exclusion when full texts had to be viewed are given in Characteristics of excluded studies.
Results of the search
119 references were identified by electronic searches, and 27 additional records were identified by a search of clinicaltrials.gov. Most were excluded upon screening of titles and abstracts (n = 117). Full texts were consulted for the remaining 29 references, and 10 were excluded at this stage, primarily because the study was not conducted in an emergency setting (n = 7). Other reasons for exclusion at this stage were 'study population did not have asthma' (n = 2) and 'no placebo comparison' (n = 1). Several unsuccessful efforts were made to find a trial publication for one additional study (Abd El Kader 1997), which is awaiting classification. The remaining 18 citations related to 14 studies, which were included in this review. Trial flow is presented in Figure 1.
Study flow diagram.
Included studies
Fourteen studies met the inclusion criteria, randomly assigning 2313 people with acute asthma to the comparisons of interest in this review. Goodacre 2013 contributed the largest sample size to the analyses, with 1109 participants randomly assigned to the two intervention groups; in contrast, Del Castillo Rueda 1991 had the smallest sample size, with 16 participants randomly assigned to the intervention groups. Mean sample size across the included studies was 165. Summary characteristics of the included trials are presented in Table 2, and full details of each included study are given in Characteristics of included studies.
| Study ID | Country (centres) | Total N | Study design | Age range (years) | Dose (infusion) | Co‐medications |
| Iran | 81 | R, DB, PC | 12–85 | 25 mg/kg (30 minutes) | Nebulised SABA, IV xanthine, IV corticosteroid, O2 | |
| Turkey | 81 | R, SB, PC | 6–65 | 1 g or 2 g (unclear) | O2 (if PaO2 was < 60 mmHg) | |
| USA (2) | 149 | R, DB, PC | 18‐65 | 2 g (20 minutes) | Nebulised SABA, IV corticosteroid | |
| Thailand (1) | 34 | R, DB, PC | 15‐65 | 2 g (unclear) | Nebulised SABA, IV corticosteroid, O2 if necessary | |
| Scotland (1) | 129 | R, DB, PC | 16+ | 1.2 g (15 minutes) | Nebulised SABA, nebulised LAMA, IV corticosteroid, O2 | |
| Spain (1) | 16 | R, DB, PC | ? | 1.5 g (15 minutes) | Nebulised SABA, IV corticosteroid | |
| UK (34) | 1109 | R, DB, PC | 16+ | 2 g (20 minutes) | Nebulised SABA and LAMA, oral corticosteroid, O2 | |
| USA (1) | 137 | ? | 18‐65 | 2 g (20 minutes) | Nebulised SABA, IV corticosteroid (others at physician's discretion), O2 | |
| UK (1) | 131 | R | Adults | 1.2 g (15 minutes) | Nebulised SABA and LAMA, O2, IV corticosteroid (discretionary xanthine) | |
| USA (1) | 42 | R, DB, PC | 18‐55 | 2 g (unclear) | Nebulised SABA, IV corticosteroid, O2 | |
| USA (8) | 248 | R, DB, PC | 18‐60 | 2 g (15 minutes) | Nebulised SABA, IV corticosteroid, O2 | |
| India (1) | 70 | R, SB, PC | 18‐60 | 2 g (20 minutes) | Nebulised SABA, nebulised LAMA, IV corticosteroid, O2 | |
| USA (1) | 38 | R, DB, PC | 18‐70 | 1.2 g (20 minutes) | Nebulised SABA, IV metaproterenol, IV xanthine | |
| USA (1) | 48 | R, DB, PC | 18‐60 | 2 g (20 minutes) | Nebulised SABA, IV corticosteroid, SABA aerosol, IV xanthine |
DB: Double‐blind; IV: Intravenous; LAMA: Long‐acting muscarinic antagonist; O2: Oxygen; PaO2: Partial pressure of oxygen in arterial blood; PC: Placebo‐controlled; R: Randomised; SABA: Short‐acting beta2‐agonist; SB: Single‐blind.
Bilaceroglu 2001 included adults and children, but only 10 participants were younger than 18 years of age; mean age was 36 (± 13.4) years.
Design and duration
Most of the studies included in this review were randomised, double‐blinded, placebo‐controlled trials. Of those that were not, two were randomised, single‐blinded, placebo‐controlled trials (Bilaceroglu 2001; Singh 2008), one was unblinded with the control group receiving no placebo (Green 1992) and for two trials, the study design was unclear from the information provided (Del Castillo Rueda 1991; Matusiewicz 1994). For these two studies, the former commented on randomisation but not blinding, and the latter commented on neither randomisation nor blinding, although both studies appeared to include treatment and control groups.
The duration of the studies ranged from 45 minutes (Skobeloff 1989) to 260 minutes (Tiffany 1993). Most trials reported outcome data at the end of study treatment periods, but further follow‐up provided in five studies ranged from six hours to one month (Bijani 2001; Bilaceroglu 2001; Bloch 1995; Goodacre 2013; Silverman 2002). Most trials were conducted at a single centre, occurring within one ED, except for Bloch 1995, which was done across two EDs in the USA; Goodacre 2013, which took place across 34 EDs in the UK and Silverman 2002, which was completed across eight EDs in the USA.
Participant inclusion and exclusion criteria
All studies included participants with an exacerbation of asthma. However differences between studies included the measures used to define an exacerbation, with some using PEF and others using FEV1, as well as the time at which these measurements were taken (e.g. on arrival, after initial treatment).
PEF was used in seven studies (Bijani 2001; Bilaceroglu 2001; Bradshaw 2007; Goodacre 2013; Matusiewicz 1994; Porter 2001; Skobeloff 1989). Two studies (Bijani 2001; Skobeloff 1989) used PEF < 200 L/min, Porter 2001 used PEF < 100 or < 25% predicted and Matusiewicz 1994 specified PEF < 250 L/min or < 50% predicted as the cutoff to indicate an exacerbation. Both Bilaceroglu 2001 and Goodacre 2013 specified PEF < 50% predicted as a cutoff. FEV1 was used as a criterion for inclusion in four studies (Bilaceroglu 2001; Bloch 1995; Silverman 2002; Singh 2008), with the cutoff being FEV1 < 75% predicted (Bilaceroglu 2001; Bloch 1995) or FEV1 < 30% (Silverman 2002; Skobeloff 1989). Boonyavorakul 2000 used a severity score > 4 (Fischl Index, which is a composite of vital signs, PEF and clinical features). Three studies (Del Castillo Rueda 1991; Green 1992; Tiffany 1993) did not define the criteria used for an exacerbation.
Three studies did not define any exclusion criteria (Bijani 2001; Del Castillo Rueda 1991; Matusiewicz 1994). For the remaining studies (n = 11), exclusion criteria were quite consistent and included diabetes mellitus, congestive cardiac disease, hypertension, chronic renal failure, temperature > 38 ºC, pneumonia, pregnancy, participants requiring ventilation and those who did not provide consent.
Baseline characteristics of participants
The most common age range used across studies was 18 to 60 years (Silverman 2002; Singh 2008; Tiffany 1993). Bloch 1995 and Green 1992 used a range of 18 to 65 years, Porter 2001 18 to 55 years and Skobeloff 1989 18 to 60 years. Two studies (Bradshaw 2007; Goodacre 2013) included participants 16 years of age and older, whilst Boonyavorakul 2000 included participants aged 15 to 65 years. Two studies included children and reported age ranges of 12 to 85 years (Bijani 2001) and six to 65 years (Bilaceroglu 2001). From the studies for which we have only the abstract, Del Castillo Rueda 1991 did not specify the age of participants, and Matusiewicz 1994 described participants as 'adults.'
Most of the studies were well matched between control and intervention with respect to sex (other than Porter 2001, in which the IV MgSO4arm consisted of 50% men compared with 25% in the placebo arm).
Only four studies reported ethnicity data (Bilaceroglu 2001; Goodacre 2013; Green 1992; Silverman 2002). The percentage classified as 'white' ranged from 59% to 100% in three of these (Bilaceroglu 2001; Goodacre 2013; Green 1992), whereas Silverman 2002 had a greater preponderance of black and Hispanic participants, with only 11% to 14% of participants classified as 'white.'
Five studies distinguished smokers (Bilaceroglu 2001; Bloch 1995; Goodacre 2013; Silverman 2002; Singh 2008), and in all cases the placebo and intervention arms were well matched. The percentage of current smokers within these studies ranged from 7% to 10% in Singh 2008, to 30% to 35% in Goodacre 2013 and Silverman 2002. The remainder of the studies (Bilaceroglu 2001; Bloch 1995) combined current smokers and ex‐smokers, and their proportions ranged from 29% to 50%.
Three studies further stratified participants by severity of asthma using American (Bilaceroglu 2001; Bloch 1995) and British (Bradshaw 2007) Thoracic Society Guidelines.
As stated in the protocol, we categorised study populations on the basis of average severity to conduct a subgroup analysis. Judgements of severity were based on baseline severity characteristics presented in the trials, which are summarised in Table 3. The justification for each judgement is given in each study's characteristics table.
| Study ID | Inclusion | Category within trial | PEF | FEV1 | Other | Classification* |
| PEF < 200 after bronchodilator and corticosteroids | All participants | 31% predicted | ‐ | RR = 35 rpm | Life threatening | |
| PEF increasing < 50% or FEV1 < 75% predicted after single salbutamol | Moderate | 57% predicted | 43% predicted | PaO2 = 69 mmHg | Severe | |
| Severe | 32% predicted | 32% predicted | PaO2 = 64 mmHg | Life threatening | ||
| FEV1 < 75% predicted after single salbutamol | Moderate | ‐ | 40% predicted | ‐ | Severe | |
| Severe | ‐ | 20% predicted | ‐ | Life threatening | ||
| Composite severity score | All participants | ‐ | ‐ | RR = 33 rpm HR = 125 bpm | Life threatening | |
| PEF < 75% predicted | Moderate | 60% predicted 248 L/min | ‐ | HR = 102 bpm | Moderate | |
| Severe | 41% predicted 170 L/min | ‐ | HR = 109 bpm | Severe | ||
| Life threatening | 23% predicted 96 L/min | ‐ | HR = 116 bpm | Life threatening | ||
| ‐ | All participants | ‐ | ‐ | ‐ | Unknown (not in analysis) | |
| One or more of the following: PEF < 50% predicted; RR > 25, HR > 110 or cannot complete sentences, but not life threatening | All participants | 52% predicted 433 L/min | ‐ | ‐ | Moderate | |
| ‐ | All participants | 143 L/min | ‐ | RR = 29 rpm HR = 108 bpm | Severe | |
| PEF < 250 L/min or < 50% predicted | All participants | ‐ | ‐ | ‐ | Severe | |
| PEF < 100 L/min or < 25% predicted | All participants | 88.5 L/min | ‐ | RR = 31 rpm HR = 110 bpm | Life threatening | |
| FEV1 < 30% predicted | All participants | 27% predicted 143 L/min | 23% predicted | HR = 102 bpm | Life threatening | |
| FEV1 < 30% predicted | All participants | 22% predicted | 38% predicted | HR = 127 bpm | Life threatening | |
| PEF < 200 L/min, not doubled after beta‐agonist, IV corticosteroid, theophylline | All participants | ˜150 L/min (from graph) | ‐ | HR = ˜ 100 bpm from graph RR = ˜ 28 | Severe | |
| PEF < 200 L/min, not doubled after albuterol × 2 | All participants | 115 L/min | 0.95 L | ‐ | Life threatening |
bpm: Beats per minute; FEV1: Forced expiratory volume in 1 second; HR: Heart rate; PaO2: Partial pressure of oxygen in arterial blood; PEF: Peak expiratory flow; rpm: Respirations per minute;RR: Respiration rate..
Classification for the severity subgroup analysis was assigned by an independent clinician and was cross‐checked with study authors' own judgements. Discrepancies were resolved through discussion.
Characteristics of the interventions
IV MgSO4
In nine studies a dose of 2 g IV was used, usually in 50 to 100 mL (250 mL in Singh 2008) of 0.9% normal saline or 5% dextrose solution, and was infused over periods ranging from 15 to 30 minutes.
Bradshaw 2007, Del Castillo Rueda 1991, Matusiewicz 1994 and Skobeloff 1989 used a dose of 1.2 g IV MgSO4 in solutions akin to those mentioned above. Bijani 2001 used doses calculated by weight of 25 mg/kg; this reflects the broader age range of the participants.
Placebo group
All studies had a placebo arm except Green 1992, in which no placebo was administered to the control group. In all other cases, the same solution that was used to infuse IV MgSO4 to the treatment group was used as the control solution, in equal volume and over the same time period. Boonyavorakul 2000 added 2 mL sterile water to the control solution, and Silverman 2002 does not comment on the specific solution used for control but describes it as 'like appearing solution' of equal volume.
Co‐medications
A number of other drugs commonly used in acute asthma were co‐administered, and there was a degree of variation in the way this was done. In all trials participants received nebulised SABA (salbutamol and, in one case, metaproterenol sulfate), and most also described the use of oxygen (n = 10) and IV corticosteroids (n = 10) before IV MgSO4 was given.
Goodacre 2013 administered oral prednisolone rather than IV corticosteroids, and the form of corticosteroid administered was unclear in Bijani 2001, Del Castillo Rueda 1991 and Bilaceroglu 2001. In the latter, the decision to administer was based on the severity category to which the participant had been assigned.
Use of oxygen was described in 10 studies, although some study authors commented that this was the case only if clinically indicated (Bilaceroglu 2001; Boonyavorakul 2000). Some authors did not describe the use of oxygen, although they may not have considered this to be a drug treatment requiring mention in the treatment protocol (Bloch 1995; Del Castillo Rueda 1991; Skobeloff 1989; Tiffany 1993).
Three studies administered aminophylline or theophylline (Bijani 2001; Skobeloff 1989; Tiffany 1993), and in Skobeloff 1989, this was guided by serum theophylline levels.
Nebulised ipratropium bromide was administered in four studies (Bradshaw 2007; Goodacre 2013; Matusiewicz 1994; Singh 2008). Goodacre 2013 and Green 1992 commented that other interventions were permitted at the discretion of the treating physician, although they did not specify which ones were permitted.
Outcomes and analysis structure
Most studies reported the number of participants who required hospitalisation after treatment (n = 11), but secondary outcomes were inconsistently reported. Three studies reported length of hospital stay for those hospitalised, and only one study at high risk of bias reported the duration of ED treatment (Green 1992). Readmission was reported in Bloch 1995 and Goodacre 2013 after a week and a month, respectively.
Lung function was reported in most of the studies, although this was done in different ways (primarily percentage predicted FEV1 and PEF, and PEF in litres per minute). Absolute values or changes in FEV1 (L) were not consistently reported. Bloch 1995 did not report standard deviation for FEV1, but the study was included on the basis of variance derived from the P value reported in the paper. This resulted in an unusually large standard deviation but did not significantly change the final results.
In the PEF analysis, we combined three studies reporting mean change from baseline (Bijani 2001; Skobeloff 1989; Tiffany 1993) with five reporting absolute endpoint scores (Goodacre 2013; Green 1992; Matusiewicz 1994; Porter 2001; Silverman 2002).
Four studies reported heart rate, respiratory rate and systolic blood pressure (Bloch 1995; Goodacre 2013; Silverman 2002; Singh 2008). Bijani 2001 reported respiratory rate, but the data could not be included because no measure of variance was provided. Goodacre 2013 reported oxygen saturation for participants on and off oxygen separately, but because no other studies reported data, we did not perform a meta‐analysis. Partial pressure was reported in one study (Bilaceroglu 2001), but again this was not formally analysed. These results are summarised narratively.
Validated symptom scales generally were not reported in the studies, but four studies reported scores on the Borg Dyspnoea Scale (Bloch 1995; Porter 2001; Silverman 2002; Singh 2008). One additional study (Goodacre 2013) measured breathlessness using a visual analogue scale (VAS), which we chose not to analyse, as it was not validated. Boonyavorakul 2000 used the Fischl Index, which is a composite of vital signs, PEF and clinical features. As individual measures were not available, the data were not analysed.
A large degree of disparity was noted in the reporting of adverse events; this precluded pooling of data in the meta‐analysis. Five studies reported no information on adverse events (Bijani 2001; Del Castillo Rueda 1991; Matusiewicz 1994; Silverman 2002; Tiffany 1993), although Silverman 2002 noted that no major adverse events were reported. Boonyavorakul 2000 and Green 1992 described minor adverse events such as flushing and fatigue, but these were not quantified. Other studies quantified adverse events for the duration of the treatment period, which ranged from 60 to 240 minutes. As such, we summarised information across studies narratively in the results.
Subgroup and sensitivity analyses
We conducted subgroup analyses on the primary outcome (hospital admissions) for baseline severity and co‐medications. For severity, 15 groups were identified across the 11 studies reporting the outcome: two moderate, six severe and seven life threatening.
We performed the analysis based on whether ipratropium bromide was administered as described in the protocol. Bradshaw 2007, Goodacre 2013, Matusiewicz 1994 and Singh 2008 were the only studies in which ipratropium bromide was given; three of these are UK studies. However, as information about co‐medications was inconsistently reported (summarised above and in Table 2), and it was often unclear when infusions or nebulisers were given, we were conservative in interpretation and have summarised the limitations of the analysis in the discussion. We could not carry out a subgroup analysis based on mean age (≤ and > 65 years), as no trials solely recruited older adults.
We also conducted two sensitivity analyses excluding trials at high risk of bias for blinding and those that contributed only unpublished data. Bilaceroglu 2001, Green 1992 and Matusiewicz 1994 were removed from the prior, and only Matusiewicz 1994 from the latter. No full paper was available for Del Castillo Rueda 1991, but this study did not report hospital admissions, and although only an abstract was available in English for Bilaceroglu 2001, the full paper had been published in Turkish, from which we were able to obtain further information. None of the studies provided additional unpublished data for the primary outcome.
We added a post‐hoc sensitivity analysis using change from baseline instead of endpoint means from Goodacre 2013, as baseline imbalances were noted in this study.
Excluded studies
Studies that took place outside of an acute setting were excluded, as were those concerned with the effects of nebulised magnesium sulfate (the subject of another review (Powell 2012)).
We excluded trials that were exclusively concerned with children, defined as those younger than 18 years of age. These studies will be dealt with in a separate Cochrane review (Griffiths 2014). We included studies in which participants were both older and younger than 18. Bradshaw 2007 and Goodacre 2013 included participants 16 years of age and older, and we believe that these data are applicable to adults, as we would not expect significant physiological differences between the ages of 16 and 18. Bijani 2001, Boonyavorakul 2000 and Bilaceroglu 2001 included participants 12 to 85 years and 15 to 65 years of age, respectively; we endeavoured to obtain data for adults only but were ultimately unsuccessful. Age ranges were unclear in three studies, although the implication was that participants were adults (Bilaceroglu 2001; Del Castillo Rueda 1991; Matusiewicz 1994).
Risk of bias in included studies
For details of the risk of bias rating for each study and the reasons for each rating, see Characteristics of included studies. A summary of risk of bias judgements by study and domain (allocation generation, allocation concealment, blinding and incomplete data) can be found in Figure 2.
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Allocation
We assessed six studies to be at low risk of bias for random sequence generation and seven for allocation concealment. Both Bloch 1995 and Bradshaw 2007 used random number generation by pharmacy, with blinding of physicians to the allocation. Goodacre 2013 used telephone‐ or Internet‐generated randomisation sequencing, whilst Porter 2001 used a random number generator producing a code, and in both studies, numbered treatment packs were prepared in pharmacy before they were used by physicians. Silverman 2002 used 1:1 randomisation tables, and the pharmacy prepared vials of placebo or IV MgSO4 with identical appearances and labelled with study IDs. Singh 2008 used 1:1 randomisation tables, and study numbers were concealed in envelopes until allocation was completed.
Skobeloff 1989 did not provide sufficient details of random sequence allocation to warrant a low risk bias judgement but adequately described allocation concealment.
Two studies (Boonyavorakul 2000; Tiffany 1993) detailed adequate randomisation processes (computer‐generated lists); in Tiffany 1993, this was managed by pharmacy, but no information about allocation concealment was provided, and hence this study was assessed to be at unclear risk in this domain.
Bilaceroglu 2001, Del Castillo Rueda 1991 and Bijani 2001 commented on randomisation, although no further details were provided and no comment on allocation concealment was made; hence these studies were assessed as unclear in both areas. The same assessment was made with Matusiewicz 1994, for which no information about randomisation or allocation concealment was provided.
We considered Green 1992 to be at high risk of bias in these domains, as participants were allocated to control or treatment group according to the day of presentation to the department.
Blinding
In the domains of both performance and detection bias, we considered most (n = 8) of the included studies to be at low risk of bias (Bloch 1995; Boonyavorakul 2000; Bradshaw 2007; Goodacre 2013; Porter 2001; Silverman 2002; Skobeloff 1989; Tiffany 1993). These were described as double‐blinded placebo‐controlled trials, and investigators provided adequate detail about who was blinded and commented that their primary outcomes were non‐subjective assessor‐rated outcomes.
Singh 2008 described this study as single‐blinded; however through correspondence with the study author, we were able to ascertain that participants and assessors of spirometric and clinical outcomes were blinded, as was the chief resident who made the decision about admission. The individual administering the medication was unblinded; therefore we rated performance bias as 'unclear' and detection bias as 'low risk.'
We believe that although Bijani 2001 performed a double‐blinded study with decoding done at completion of the study, limited detail was provided about who the blinded parties were, and we considered this to be unclear.
We have no information for these domains from Del Castillo Rueda 1991 and Matusiewicz 1994 and have graded them as also having unclear risk of bias.
We assessed that both Bilaceroglu 2001 and Green 1992 are at high risk of bias in these domains. The former study was single‐blinded, and further correspondence with the study author confirmed that only participants were blinded to treatment, allowing for bias in assessment of outcome measures. In Green 1992, the physicians were unblinded to randomisation, and although neither participants nor respiratory therapists carrying out PEF measures were aware that a study was being conducted, they may have been aware of the treatment received.
Incomplete outcome data
We considered that in half of the included studies (n = 7), the risk of attrition bias was low, and in the other half, the risk was unclear.
In studies for which we considered the risk to be low (Bilaceroglu 2001; Bloch 1995; Boonyavorakul 2000; Bradshaw 2007; Goodacre 2013; Singh 2008; Skobeloff 1989), withdrawal rates were clearly documented and numbers were low, with similar rates reported in placebo and control groups.
In four studies (Bijani 2001; Del Castillo Rueda 1991; Matusiewicz 1994; Tiffany 1993), no information was provided about withdrawal rates, hence the reason for considering the risk to be unclear.
In Green 1992, 97 of 217 participants were excluded from analysis, with 80 participants repeat attenders (no comment on the groups to which they had been randomly assigned) and the medical records of 17 participants misplaced. No comment was made about whether there was intention to treat any of the participants who withdrew, although at the point of analysis, numbers in all groups were similar.
Porter 2001 reports that where repeat attendance to the department was documented, data from only the first presentation were used, but no further commentary was made about withdrawals.
Silverman 2002 provides a very detailed report of participants with protocol violations who were retained in the intention‐to‐treat data set and gives reasons for these inclusions. However, attrition rates were quite high and were not provided for each arm. As such it was unclear whether attrition was balanced between groups, and the study was rated as 'unclear.'
Selective reporting
We considered that only five studies demonstrated low risk of bias in reporting of outcome data: Bijani 2001; Bloch 1995; Goodacre 2013; Silverman 2002; and Singh 2008. Although Bijani 2001 did not report on arterial blood gas (ABG) results as was planned, other data were well reported, and we believe that the ABG measure was not critical to the study. Both Goodacre 2013 and Singh 2008 provided further raw data when directly contacted by the review authors, and this completed the outcomes planned for assessment. The published report of Bloch 1995 provided data at only one of the prespecified time points and FEV1 was provided graphically, but the study author provided additional data to the review authors upon request.
We considered the following studies to be unclear for risk of reporting bias: Bilaceroglu 2001; Del Castillo Rueda 1991; Green 1992; Matusiewicz 1994; and Porter 2001. We have only the abstract for both Del Castillo Rueda 1991 and Matusiewicz 1994; The former provided no outcome data, just a written description of investigator conclusions, and the latter provided outcomes at only one of the recorded time points. Bilaceroglu 2001 did provide further raw data to the review authors on request, but this still did not include all time points laid out in the methodology. Green 1992 provided outcome data, but the methodology did not indicate the primary outcome measures selected when the study was designed. Porter 2001 provided all primary outcome data at the prespecified time point; however data were also collected at other time points, and this was not reported.
We considered that four studies demonstrated high risk of reporting bias. Boonyavorakul 2000 provided only raw admission data, and severity scores were provided only in terms of 'variance.' Bradshaw 2007, Skobeloff 1989 and Tiffany 1993 provided raw data for only a subset of outcomes or time points, with remaining results presented graphically or without variance and with no reporting of raw data.
Other potential sources of bias
No additional sources of bias were identified.
Effects of interventions
Primary outcomes
Hospital admissions
Combining 11 studies (n = 972) revealed a significant reduction in hospital admissions compared with placebo (OR 0.75, 95% CI 0.60 to 0.92; high‐quality evidence; Analysis 1.1). Some heterogeneity that was not statistically significant was observed (I2 = 28%; P value 0.18). In absolute terms, this odds ratio translates to a reduction of seven hospital admissions for every 100 adults (95% CI two to 13 fewer) treated with IV MgSO4 (Figure 3). There was no reason to downgrade for any of the five domains in GRADE (risk of bias, inconsistency, indirectness, imprecision, publication bias). Specifically, risk of bias was generally low or unclear across trials, heterogeneity was not significant, trials matched the research question well, confidence intervals were relatively narrow and almost all studies contributed data to the analysis.
In the control group, 57 of 100 people were admitted to hospital, compared with 50 (95% CI 45 to 55) of 100 for the IV MgSO4 group.
Secondary outcomes
Intensive care admissions
Evidence from one study (Goodacre 2013; n = 752) showed no significant difference in admission rates between IV MgSO4 and placebo (OR 2.03, 95% CI 0.70 to 5.89; moderate‐quality evidence; Analysis 1.2). The same study reported the number of participants admitted to the high dependency unit and showed no significant difference between the two arms (OR 1.05, 95% CI 0.57 to 1.94; moderate‐quality evidence; Analysis 1.3). Both outcomes contained few events from only one study, so they were downgraded twice for imprecision, and the quality of evidence was rated as 'low.'
ED treatment duration
Only one study (Green 1992; n = 452) reported ED treatment duration and found no significant difference between IV MgSO4 and placebo (MD ‐4.00, 95% CI ‐37.02 to 29.02; low‐quality evidence; Analysis 1.4). The outcome was downgraded for risk of bias and imprecision.
Length of hospital stay (days)
Combining three studies reporting the outcome (n = 949) revealed no significant difference in time spent in hospital between the IV MgSO4 and placebo groups (MD ‐0.03, 95% CI ‐0.33 to 0.27; low‐quality evidence; Analysis 1.5). The evidence was downgraded for risk of bias and inconsistency (I2 = 53%; P value 0.10). As I2 was over the 30% defined in the protocol, we performed a sensitivity analysis using random effects, which did not change the conclusions (MD ‐0.16, 95% CI ‐0.68 to 0.37).
Readmission
Too few events were described in only two studies to indicate whether IV MgSO4 had an effect on readmission to hospital compared with placebo (OR 2.30, 95% CI 0.66 to 7.99; moderate‐quality evidence; Analysis 1.6). No statistical heterogeneity was noted between the studies (I2 = 0%; P value 0.34), but the outcome was downgraded for imprecision.
Vital signs
Heart rate
Combining four studies (n = 1195) showed a small significant reduction in heart rate with IV MgSO4 compared with placebo (MD ‐2.37, 95% CI ‐4.13 to ‐0.61; moderate‐quality evidence; Analysis 1.7). However a high degree of heterogeneity was observed, which was statistically significant and warranted downgrading (I2 = 78%; P value 0.004). A sensitivity analysis using random effects decreased precision significantly, with confidence intervals including both significant benefit and potential harm of IV MgSO4 (MD ‐2.61, 95% CI ‐6.58 to 1.35).
Respiratory rate
When five studies were combined (n = 1276), IV MgSO4 did not show a significant reduction in respiratory rate compared with placebo (MD ‐0.28, 95% CI ‐0.77 to 0.20; moderate‐quality evidence; Analysis 1.8). Heterogeneity was not significant (I2 = 1%; P value 0.39), but the evidence was downgraded for imprecision because confidence intervals included significant benefit and potential harm of the treatment.
Systolic blood pressure
Four studies (Bloch 1995; Bradshaw 2007; Goodacre 2013; Silverman 2002; n = 1264) reporting systolic blood pressure showed no difference between IV MgSO4 and placebo (MD 0.08, 95% CI ‐1.89 to 2.05; moderate‐quality evidence; Analysis 1.9). Heterogeneity was high, and although it was not statistically significant, authors considered it large enough to warrant downgrading for inconsistency (I2 = 51%; P value 0.11). A sensitivity analysis using random effects did not change the conclusions (MD ‐0.73, 95% CI ‐4.13 to 2.67).
Oxygen saturations
One study reported outcomes separately for those receiving and those not receiving oxygen (Goodacre 2013). This outcome was not reported in other studies; therefore we were unable to meta‐analyse the data.
Spirometry
FEV1 (% predicted)
When four studies were combined (Bilaceroglu 2001; Bloch 1995; Silverman 2002; Singh 2008) (n = 523), significant improvement in percentage predicted FEV1 was seen in the IV MgSO4 group compared with the placebo group (MD 4.41, 95% CI 1.75 to 7.06; high‐quality evidence; Analysis 1.10). No significant heterogeneity was noted among studies (I2= 14%; P value 0.33).
During data analysis, reported standard deviations in Bilaceroglu 2001 were outliers and appeared to be more consistent with standard error values; the author confirmed that this was the case. In addition, Bloch 1995 reported no standard deviations; therefore the standard error of the mean was calculated from the graphs.
PEF (% predicted)
Three studies (Bradshaw 2007; Goodacre 2013; Silverman 2002; n = 1129) reported PEF (% predicted) and showed a statistically significant improvement in PEF with IV MgSO4 compared with placebo (MD 4.78, 95% CI 2.14 to 7.43; high‐quality evidence; Analysis 1.11). Heterogeneity between studies was high but was not statistically significant (I2 = 45%; P value 0.16), so the evidence was not downgraded. A sensitivity analysis with random effects did not change our conclusions (MD 5.17, 95% CI 1.15 to 9.19). On the basis of observed baseline imbalances in the largest study (Goodacre 2013), a second sensitivity analysis using change from baseline instead of endpoint means substantially reduced the effect (MD 1.57, 95% CI ‐0.55 to 3.69; I2 = 79%, P = 0.009; Analysis 2.5).
PEF (L/min)
Combining eight studies (n = 1460) revealed that IV MgSO4 improved PEF compared with placebo (MD 17.40, 95% CI 8.64 to 26.17; moderate‐quality evidence; Analysis 1.12). However statistically significant heterogeneity between the studies (I2 = 50%; P value 0.05) warranted downgrading. A sensitivity analysis with random effects did not change our conclusions (MD 18.35, 95% CI 4.12 to 32.58). As with the percentage PEF predicted analysis, a second sensitivity analysis using Goodacre 2013 change from baseline substantially reduced the magnitude of effect (MD 9.44, 95% CI 2.07 to 16.81; I2 = 68%, P = 0.003; Analysis 2.6).
Validated symptom scores
Five studies used symptom scales, all measuring breathlessness (n = 1237). The Borg Dyspnoea Scale was used by four studies (Bloch 1995; Porter 2001; Silverman 2002; Singh 2008), and Goodacre 2013 used a VAS for breathlessness. Data for the Borg Dyspnoea Scale revealed no significant change with IV MgSO4 compared with placebo (MD ‐0.22, 95% CI ‐0.55 to 0.12; high‐quality evidence; Analysis 1.13), and no significant heterogeneity between studies was noted (I2 = 0%; P value 0.82).
Similarly, Goodacre 2013 reported no significant change in VAS score with IV MgSO4 compared with placebo (MD ‐3.00, 95% CI ‐7.09 to 1.09).
Adverse events
The most commonly cited adverse events were flushing, fatigue, nausea and headache; some study authors also commented on hypotension.
Bilaceroglu 2001 reported flushing in 42% of those receiving IV MgSO4 versus no flushing in the placebo group. Although paraesthesia, vertigo and hypotension were also reported, no marked differences between treatment and placebo arms were observed.
Bloch 1995 reported that 58% of those receiving IV MgSO4 reported adverse events, including the sensation of flushing, fatigue and burning at the IV site, with one participant experiencing transient urticaria in the upper extremities.
Bradshaw 2007 reported minor adverse events in 8% of those receiving IV MgSO4 (headache, flushing, dizziness), with only one participant in the placebo arm reporting flushing (1.5%).
Goodacre 2013 reported the rate of adverse events (death, arrhythmia, cardiac arrest, non‐invasive ventilation, intubation, other) as 13% in the treatment group compared with 10% in the placebo group, although these rates fall almost entirely in the 'other' category. One death of an unspecified cause (1%) was reported in the IV MgSO4 group compared with none in the placebo group. No other trials reported deaths. Goodacre 2013 reported commonly cited adverse events as a separate category and revealed a statistically significant increase in adverse events in the IV MgSO4 group (OR 1.68, 95% CI 1.07 to 2.63; P value 0.025).
Both Porter 2001 and Singh 2008 reported that the difference between rates of adverse events (including deep tendon reflexes) among participants given IV MgSO4 versus placebo was not statistically significant.
Skobeloff 1989 reported higher rates of fatigue (32% vs 11%), warmth (26%) and lightheadedness (5%) in the IV MgSO4 group, but the numbers in this study were small.
With respect to blood pressure, Bilaceroglu 2001 reported hypotension in 5% versus 3% of participants in the treatment versus placebo groups, whilst Goodacre 2013 reported 8% versus 6%, respectively. Bradshaw 2007 reported a non‐significant trend for decreasing blood pressure at 60 minutes, and Singh 2008 reported no hypotension.
Subgroup analyses
Baseline severity (moderate, severe and life‐threatening exacerbations)
The test for subgroup differences revealed no statistical heterogeneity between the three severity subgroups (I2 = 0%; P value 0.73), and between‐trial heterogeneity was significant within all three subgroups (I2 = 50%; P value 0.01).
Mean age (≤ and > 65 years)
Most studies included participants over age 65, but all population mean ages were much lower than the cutoff. As we did not have access to individual participant data within the trials, we were unable to draw any conclusions regarding potential differential effects of IV MgSO4 due to age.
Co‐medications (with and without nebulised ipratropium bromide)
The test for subgroup differences showed no significant differences between the four studies that administered nebulised ipratropium bromide as a co‐medication and those that did not (I2 = 0%; P value 0.82). Between‐trial heterogeneity was not statistically significant within either of the two subgroups (I2 = 28%; P value 0.18).
Sensitivity analysis
Studies at high risk of bias for blinding
When three studies that were given a 'high' or 'unclear' rating for blinding were removed (Bilaceroglu 2001; Green 1992; Matusiewicz 1994), the pooled effect for hospital admissions was slightly larger in favour of IV MgSO4 (OR 0.72, 95% CI 0.57 to 0.91; Analysis 2.3). Heterogeneity was slightly larger than in the main analysis, but this difference was not statistically significant (I2 = 35%; P value 0.15).
Unpublished data
Of the two studies for which only a conference abstract was available, one reported hospital admissions (Matusiewicz 1994). When this study was removed from the primary outcome, the magnitude of the effect in favour of IV MgSO4 was slightly increased (OR 0.73, 95% CI 0.58 to 0.91), but this did not change the conclusions. Some heterogeneity that was not significant was reported (I2 = 32%; P value 0.15).
Discussion
Summary of main results
Fourteen studies met the inclusion criteria, randomly assigning 2313 people with acute asthma to the comparisons of interest in this review. A recent large study (Goodacre 2013) accounted for a large proportion of the total number of participants (n = 752).
The included studies were mostly randomised, double‐blinded trials comparing 1.2 g or 2 g IV MgSO4 versus a matching placebo infusion. All of these studies included participants who had an exacerbation of asthma, although definitions and inclusion criteria varied. Ten studies included only adults; four included adults and children and were included because the mean age was over 18 years. Inclusion criteria varied, and studies assigned a level of severity to participants, which we then verified against BTS/SIGN 2012 criteria, confirming that all studies included exacerbations of at least moderate severity.
Eleven studies could be included in the primary analysis and showed that IV MgSO4 reduced hospital admissions compared with placebo (OR 0.75, 95% CI 0.60 to 0.92; I2 = 28%; P value 0.18; n = 972; high‐quality evidence). In absolute terms, this odds ratio translates into a reduction of seven hospital admissions for every 100 adults (95% CI two to 13 fewer) treated with IV MgSO4 (Figure 3). The test for subgroup differences did not reveal statistical heterogeneity between the three severity subgroups (I2 = 0%; P value 0.73), or between the four studies that administered nebulised ipratropium bromide as a co‐medication and those that did not (I2 = 0%;, P value 0.82). Sensitivity analyses removing unpublished data and studies at high risk for blinding from the primary analysis did not change conclusions; this increased our confidence in the effect.
Within the secondary outcomes, evidence of high and moderate quality across three spirometric indices suggested some improvement in lung function with IV MgSO4; however the clinical significance of the size of these effects is uncertain, and baseline imbalances in the largest study reduced our confidence in some of the findings. Although close, the mean difference in PEF (L/min) found in this meta‐analysis did not reach the minimal clinically important difference (MCID) defined by Santanello 1999 (18.79 L/min). There are no accepted MCIDs for the percentage predicted measures reported in most of the trials. Mean FEV1 in litres, for which an MCID does exist, was reported in only two of the 14 trials.
No difference between IV MgSO4 and placebo was found for most of the non‐spirometric secondary outcomes, all of which were rated of low or moderate quality (intensive care admissions, ED treatment duration, length of hospital stay, readmission, respiration rate, systolic blood pressure).
Adverse events were inconsistently reported and were not meta‐analysed. The most commonly cited adverse events in the IV MgSO4 groups were flushing, fatigue, nausea and headache and hypotension. However we found no significant difference in blood pressure between the IV MgSO4 and placebo groups.
Overall completeness and applicability of evidence
A large degree of variation between prescribing procedures was evident in the trials, but doses used in the included studies are in accord with current BTS/SIGN 2012, GINA 2011 and NACA 2006 guidelines. However, the treatment protocols differed as to when the decision to administer IV MgSO4 was made; the dosage, frequency and form of co‐medications and the order in which the medications were administered in relation to one another. We suspect that differences between individual EDs both within and among countries were significant, and insufficient reporting in the trials themselves further complicated interpretation of the subgroup analysis for co‐medications. As such, although no evidence suggested a difference in the efficacy of IV MgSO4 delivered in settings where ipratropium bromide was prescribed, we cannot exclude the possibility that other combinations of co‐medications may significantly alter the effectiveness of IV MgSO4. Moreover, as almost all of the studies administered short‐acting beta2‐agonists, oxygen and IV corticosteroids before MgSO4, the evidence is suitably applied to situations for which these medications have already been prescribed. Doses of magnesium used and method and rate of delivery were relatively consistent across studies (1.2 g to 2 g via 15 to 30‐minute infusion), so it is not clear whether the same effect would be observed with alternative administrations (e.g. higher dose, bolus).
The definition of hospital admission may have varied between the healthcare settings in which these studies were carried out, and this was not clearly defined in the studies. We accept that variation exists in the broader health and economic environments and in health infrastructures, such as the use of clinical decision making or observation wards, and that this is likely to have influenced the decision to admit. This variation is likely to have introduced heterogeneity in the primary outcome.
The previous version of this review (Rowe 2009) suggested the possibility of greater efficacy of treatment in more severe exacerbations; this partially informed our decision to perform a subgroup analysis based on severity. We did not find a statistically significant difference between the three severity subgroups; however, the method that we used to allocate baseline severity had limitations. We based this classification on BTS/SIGN 2012 criteria, but reporting of baseline metrics on which this guidance is based was insufficient in several studies. In studies that subdivided the population on the basis of severity (Bilaceroglu 2001; Bloch 1995; Bradshaw 2007), subgroups with a more severe condition gained greater benefit with respect to hospital admission. This suggests that within‐study subgroups may serve as a more reliable way of assessing severity as an effect modifier by controlling for differences in other variables that may exist between study protocols.
We were unable to draw conclusions regarding the potential effects of age on study outcomes, as none of the studies recruited older adults. It is possible that diagnosis in this age group would be complicated by important co‐morbidities (e.g. chronic obstructive pulmonary disease (COPD)), and that these might also affect the safety and effectiveness of IV MgSO4. As such, it is likely that the conclusions of this review are not applicable to this population, or to children younger than the age of 12.
Several outcomes showed significant statistical heterogeneity among studies that was not accounted for by subgrouping results by severity of exacerbations (heart rate (HR), systolic blood pressure (BP), PEF in L/min, length of hospital stay). For HR, systolic BP and PEF in L/min, variation may be explained in part by when and how the measurement was taken, measurement error and the influence of co‐medications. Length of hospital stay is highly dependent on local hospital guidelines and procedures.
We were unable to meta‐analyse data related to adverse events and therefore could not draw conclusions about the safety of IV MgSO4 in asthma. Some commonly cited adverse events were consistently reported among the studies; however, the methods of recording adverse events appeared unsystematic.
Quality of the evidence
We used GRADEpro software to assess the quality of all outcomes; this assessment is summarised in the text and in the summary of findings Table for the main comparison. Most outcomes were not downgraded for risk of bias, and in the two cases in which this was done, the decision was related primarily to insufficient blinding. It is unclear how this may have affected results for the primary outcome (i.e. decision to admit), but a sensitivity analysis excluding studies in which blinding was insufficient or unclear showed that this bias is unlikely to have significantly affected the pooled estimate.
Several outcomes were downgraded for inconsistency, that is, statistical heterogeneity, between studies. In these cases we performed sensitivity analyses using a random‐effects model, which did not alter conclusions. The clinical source of the statistical heterogeneity remains unclear in most cases, as planned subgroup analyses were performed only on the primary outcome. Most of the secondary outcomes for which heterogeneity was observed contained a small number of studies; therefore it is unlikely that subgrouping of results would have allowed a meaningful distinction between severity or co‐medication subgroups.
Studies included in this review were directly relevant to our review question with respect to participants recruited, interventions and comparisons provided, healthcare setting selected, and outcome measures used, so none of the evidence was downgraded for indirectness.
Four outcomes were downgraded for imprecision on the basis of their wide confidence intervals (intensive care unit (ICU) admission, ED treatment duration, PEF in L/min, respiratory rate (RR)). In each case the review authors made a clinical judgement regarding the minimal clinically important difference in relation to the confidence intervals. Moreover, with the exception of ED treatment duration, evidence from related outcomes (e.g. other spirometric measures) helped us draw conclusions when imprecision was due to a small number of participants or events.
No outcomes were downgraded for publication bias, although several of the secondary outcomes included a small number of studies. No incidences were identified in which studies stated outcomes and failed to report them, but this was generally a result of insufficient reporting of intended outcomes and the fact that the studies could not be linked to trial registrations. Most studies were conducted before adherence to trial registration or reporting standards was common practice, so in most cases we were unable to definitively judge whether the evidence was compromised as a result of deliberate or inadvertent selective reporting.
To resolve uncertainties related to risk of bias and missing data, we made an effort to contact all study authors. We received additional data from four of these authors (Bilaceroglu 2001; Bloch 1995; Goodacre 2013; Singh 2008), were unable to obtain current contact details for two (Matusiewicz 1994; Porter 2001) and received no response from the remaining eight.
Potential biases in the review process
We made every effort to adhere to Cochrane methods during the review process. All study characteristics and numerical data were extracted by at least two review authors, and discrepancies were resolved through discussion. The same was true for risk of bias ratings, and none of the review authors have conflicting interests.
We performed relatively broad searches that were screened by at least two review authors independently, and we included studies regardless of language of publication. As a result, It is unlikely that any published studies were missed during study selection. In addition, review authors attempted to contact all study authors to clarify study methodology or to obtain additional data when details were not included in the published reports. We received detailed replies and additional data from four study authors, but in most cases, it was unclear whether study authors had failed to receive the request or were simply unable to provide the information required.
The subgroup analysis based on exacerbation severity introduced the potential for internal bias, despite efforts to remove bias by consultation with an independent fourth party. Although we were transparent in the method of classification, an element of subjectivity due to reporting standards was noted in some trials; this reduced our confidence in the subgroup findings.
Agreements and disagreements with other studies or reviews
Our literature search identified five systematic reviews with meta‐analyses comparing use of IV MgSO4 versus placebo in adults with acute asthma (Alter 2000; Mohammed 2007; Rodrigo 2000; Rowe 2009; Shan 2013). One of these, Rowe 2009, was a previous Cochrane review, and Shan 2013, the most recent research synthesis, included the greatest number of trials (n = 16), 10 of which are included in this review.
The main outcomes analysed were hospital admissions and spirometric data. None of the existing reviews found a statistically significant reduction in hospital admissions across all severity subgroups. However Rowe 2009 found a significant reduction in hospital admissions within the more severe group (OR 0.10, 95% CI 0.04 to 0.27) and suggested that IV MgSO4 might play a role in these more severe exacerbations. Hospital admission data for Shan 2013 were on the border of statistical significance (P value 0.06).
Alter 2000 and Shan 2013 reported significant improvement in pooled spirometric measures for those receiving IV MgSO4, whilst Mohammed 2007 reported weak evidence to support this (SMD 0.25, 95% CI ‐0.01 to 0.51; P value 0.05). Pooled analyses in Rodrigo 2000 and Rowe 2009 showed no significant improvement in lung function for those given IV MgSO4.
In keeping with these reviews, we found evidence that IV MgSO4 improves lung function on a variety of spirometric measures. However, our findings differ in that when all data regardless of severity criteria were pooled, a statistically significant reduction in hospital admissions was seen among those treated with IV MgSO4. We did not draw firm conclusions regarding the extent to which severity of exacerbation affects the efficacy of IV MgSO4 because differences in the ways the studies were conducted made it difficult to assess the effect of exacerbation severity independent of other effect moderators.
Several reasons may account for the discrepancies between our conclusions and those of previous evidence syntheses. Unlike some previous systematic reviews (Alter 2000; Mohammed 2007; Rowe 2009; Shan 2013), our inclusion criteria did not include paediatric trials. Several additional trials have been published since the previous version of this review, and this warranted synthesising of data for adults separately from data for children (Bijani 2001; Bilaceroglu 2001; Boonyavorakul 2000; Bradshaw 2007; Del Castillo Rueda 1991; Goodacre 2013; Matusiewicz 1994; Porter 2001; Singh 2008). One of these trials, Goodacre 2013, is a recent randomised controlled trial with a large sample size; it accounted for a significant proportion of the total weight in several of our analyses. Evidence for the use of IV MgSO4 in the paediatric population will be analysed in a separate Cochrane review, which is currently in production. Some previous syntheses have included trials of nebulised magnesium sulfate (Mohammed 2007; Rodrigo 2000; Shan 2013), which we did not include, as this is the subject of an existing Cochrane review (Powell 2012).
Study flow diagram.
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
In the control group, 57 of 100 people were admitted to hospital, compared with 50 (95% CI 45 to 55) of 100 for the IV MgSO4 group.
Comparison 1 IV MgSO4 versus placebo, Outcome 1 Hospital admissions.
Comparison 1 IV MgSO4 versus placebo, Outcome 2 Intensive care unit (ICU) admissions.
Comparison 1 IV MgSO4 versus placebo, Outcome 3 High dependency unit (HDU) admissions.
Comparison 1 IV MgSO4 versus placebo, Outcome 4 ED treatment duration (minutes).
Comparison 1 IV MgSO4 versus placebo, Outcome 5 Length of hospital stay (days).
Comparison 1 IV MgSO4 versus placebo, Outcome 6 Readmission.
Comparison 1 IV MgSO4 versus placebo, Outcome 7 Heart rate (bpm).
Comparison 1 IV MgSO4 versus placebo, Outcome 8 Respiratory rate (breaths/min).
Comparison 1 IV MgSO4 versus placebo, Outcome 9 Systolic blood pressure (mmHg).
Comparison 1 IV MgSO4 versus placebo, Outcome 10 FEV1 (% predicted).
Comparison 1 IV MgSO4 versus placebo, Outcome 11 PEF (% predicted).
Comparison 1 IV MgSO4 versus placebo, Outcome 12 PEF (L/min).
Comparison 1 IV MgSO4 versus placebo, Outcome 13 Borg Dyspnoea Scale score.
Comparison 2 IV MgSO4 versus placebo (subgroup and sensitivity analyses), Outcome 1 Hospital admissions (by severity).
Comparison 2 IV MgSO4 versus placebo (subgroup and sensitivity analyses), Outcome 2 Hospital admissions (by co‐medications).
Comparison 2 IV MgSO4 versus placebo (subgroup and sensitivity analyses), Outcome 3 Hospital admissions (risk of bias sensitivity).
Comparison 2 IV MgSO4 versus placebo (subgroup and sensitivity analyses), Outcome 4 Hospital admissions (unpublished sensitivity).
Comparison 2 IV MgSO4 versus placebo (subgroup and sensitivity analyses), Outcome 5 PEF % predicted (Goodacre change score sensitivity).
Comparison 2 IV MgSO4 versus placebo (subgroup and sensitivity analyses), Outcome 6 PEF L/min (Goodacre change score sensitivity).
| IV MgSO4 for treating adults with acute asthma in the ED | |||||
| Patient or population: adults with acute asthma Comparions: placebo Both intervention and placebo groups received oxygen, short‐acting beta2‐agonists and oral or intravenous steroids before the infusion. Measurements were taken between 60 and 240 minutes after the start of the infusion. | |||||
| Outcomes | Illustrative comparative risks* (95% CI) | Relative effect | No. of participants | Quality of the evidence | |
| Assumed risk | Corresponding risk | ||||
| Control | IV MgSO4 | ||||
| Hospital admissions | 569 per 1000 | 498 per 1000 | OR 0.75 | 1769 | ⊕⊕⊕⊕ |
| Intensive care unit (ICU) admissions | 14 per 1000 | 28 per 1000 | OR 2.03 | 752 | ⊕⊕⊝⊝ |
| Length of hospital stay (days) | Mean length of hospital stay in the control groups was | Mean length of hospital stay in the intervention groups was | ‐ | 949 | ⊕⊕⊝⊝ |
| ED treatment duration (minutes) | Mean duration in the placebo group was 228 minutes | Mean ED treatment duration in the intervention groups was | ‐ | 96 | ⊕⊕⊝⊝ |
| FEV1 (% predicted) | Mean FEV1 in the placebo group was 50% predicted | Mean FEV1 (% predicted) in the intervention groups was | ‐ | 523 | ⊕⊕⊕⊕ |
| PEF (L/min) | Mean PEF in the placebo group was 239 L/min | Mean PEF in the intervention groups was | ‐ | 1460 | ⊕⊕⊕⊝ |
| Respiratory rate (breaths/min) | Mean respiration rate in the placebo group was 20.7 respirations/min | Mean respiratory rate in the intervention groups was | ‐ | 1195 | ⊕⊕⊕⊝ |
| *The basis for the assumed risk (e.g. median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | |||||
| GRADE Working Group grades of evidence. | |||||
| 1One study (Green 1992) introduced risk of bias, but the rest of the studies were generally well conducted. | |||||
| BTS/SIGN | GINA | NACA | NAEPP | |
| Oxygen | ✓ | ✓ | ✓ | ✓ |
| Inhaled beta2‐agonist | ✓ | ✓ | ✓ | ✓ |
| Inhaled antimuscarinic | ✓ | ✓ | ✓ | ✓ |
| Sytemic corticosteroids | ✓ | ✓ | ✓ | ✓ |
| IV beta2‐agonist | (✓) if nebulised form cannot be used reliably | x | ✓ if no response to inhaled form | x |
| IV MgSO4 | ✓ | ✓ | ✓ IV or nebulised | ✓ |
| Heliox | x | x | x | ✓ |
| IV aminophylline/theophylline | (✓) limited evidence, only after senior consultation | (✓) if inhaled beta2‐agonist unavailable | (✓) as an alternative to IV beta2‐agonist | x |
| BTS/SIGN: British Thoracic Society and Scottish Intercollegiate Guidelines Network joint guideline; GINA: Global Initiative for Asthma; IV: Intravenous; NACA: National Asthma Council Australia; NAEPP: National Asthma Education and Prevention Program; ✓: Recommended; x: Not recommended; (✓): Recommended with conditions. | ||||
| Study ID | Country (centres) | Total N | Study design | Age range (years) | Dose (infusion) | Co‐medications |
| Iran | 81 | R, DB, PC | 12–85 | 25 mg/kg (30 minutes) | Nebulised SABA, IV xanthine, IV corticosteroid, O2 | |
| Turkey | 81 | R, SB, PC | 6–65 | 1 g or 2 g (unclear) | O2 (if PaO2 was < 60 mmHg) | |
| USA (2) | 149 | R, DB, PC | 18‐65 | 2 g (20 minutes) | Nebulised SABA, IV corticosteroid | |
| Thailand (1) | 34 | R, DB, PC | 15‐65 | 2 g (unclear) | Nebulised SABA, IV corticosteroid, O2 if necessary | |
| Scotland (1) | 129 | R, DB, PC | 16+ | 1.2 g (15 minutes) | Nebulised SABA, nebulised LAMA, IV corticosteroid, O2 | |
| Spain (1) | 16 | R, DB, PC | ? | 1.5 g (15 minutes) | Nebulised SABA, IV corticosteroid | |
| UK (34) | 1109 | R, DB, PC | 16+ | 2 g (20 minutes) | Nebulised SABA and LAMA, oral corticosteroid, O2 | |
| USA (1) | 137 | ? | 18‐65 | 2 g (20 minutes) | Nebulised SABA, IV corticosteroid (others at physician's discretion), O2 | |
| UK (1) | 131 | R | Adults | 1.2 g (15 minutes) | Nebulised SABA and LAMA, O2, IV corticosteroid (discretionary xanthine) | |
| USA (1) | 42 | R, DB, PC | 18‐55 | 2 g (unclear) | Nebulised SABA, IV corticosteroid, O2 | |
| USA (8) | 248 | R, DB, PC | 18‐60 | 2 g (15 minutes) | Nebulised SABA, IV corticosteroid, O2 | |
| India (1) | 70 | R, SB, PC | 18‐60 | 2 g (20 minutes) | Nebulised SABA, nebulised LAMA, IV corticosteroid, O2 | |
| USA (1) | 38 | R, DB, PC | 18‐70 | 1.2 g (20 minutes) | Nebulised SABA, IV metaproterenol, IV xanthine | |
| USA (1) | 48 | R, DB, PC | 18‐60 | 2 g (20 minutes) | Nebulised SABA, IV corticosteroid, SABA aerosol, IV xanthine | |
| DB: Double‐blind; IV: Intravenous; LAMA: Long‐acting muscarinic antagonist; O2: Oxygen; PaO2: Partial pressure of oxygen in arterial blood; PC: Placebo‐controlled; R: Randomised; SABA: Short‐acting beta2‐agonist; SB: Single‐blind. Bilaceroglu 2001 included adults and children, but only 10 participants were younger than 18 years of age; mean age was 36 (± 13.4) years. | ||||||
| Study ID | Inclusion | Category within trial | PEF | FEV1 | Other | Classification* |
| PEF < 200 after bronchodilator and corticosteroids | All participants | 31% predicted | ‐ | RR = 35 rpm | Life threatening | |
| PEF increasing < 50% or FEV1 < 75% predicted after single salbutamol | Moderate | 57% predicted | 43% predicted | PaO2 = 69 mmHg | Severe | |
| Severe | 32% predicted | 32% predicted | PaO2 = 64 mmHg | Life threatening | ||
| FEV1 < 75% predicted after single salbutamol | Moderate | ‐ | 40% predicted | ‐ | Severe | |
| Severe | ‐ | 20% predicted | ‐ | Life threatening | ||
| Composite severity score | All participants | ‐ | ‐ | RR = 33 rpm HR = 125 bpm | Life threatening | |
| PEF < 75% predicted | Moderate | 60% predicted 248 L/min | ‐ | HR = 102 bpm | Moderate | |
| Severe | 41% predicted 170 L/min | ‐ | HR = 109 bpm | Severe | ||
| Life threatening | 23% predicted 96 L/min | ‐ | HR = 116 bpm | Life threatening | ||
| ‐ | All participants | ‐ | ‐ | ‐ | Unknown (not in analysis) | |
| One or more of the following: PEF < 50% predicted; RR > 25, HR > 110 or cannot complete sentences, but not life threatening | All participants | 52% predicted 433 L/min | ‐ | ‐ | Moderate | |
| ‐ | All participants | 143 L/min | ‐ | RR = 29 rpm HR = 108 bpm | Severe | |
| PEF < 250 L/min or < 50% predicted | All participants | ‐ | ‐ | ‐ | Severe | |
| PEF < 100 L/min or < 25% predicted | All participants | 88.5 L/min | ‐ | RR = 31 rpm HR = 110 bpm | Life threatening | |
| FEV1 < 30% predicted | All participants | 27% predicted 143 L/min | 23% predicted | HR = 102 bpm | Life threatening | |
| FEV1 < 30% predicted | All participants | 22% predicted | 38% predicted | HR = 127 bpm | Life threatening | |
| PEF < 200 L/min, not doubled after beta‐agonist, IV corticosteroid, theophylline | All participants | ˜150 L/min (from graph) | ‐ | HR = ˜ 100 bpm from graph RR = ˜ 28 | Severe | |
| PEF < 200 L/min, not doubled after albuterol × 2 | All participants | 115 L/min | 0.95 L | ‐ | Life threatening | |
| bpm: Beats per minute; FEV1: Forced expiratory volume in 1 second; HR: Heart rate; PaO2: Partial pressure of oxygen in arterial blood; PEF: Peak expiratory flow; rpm: Respirations per minute;RR: Respiration rate.. Classification for the severity subgroup analysis was assigned by an independent clinician and was cross‐checked with study authors' own judgements. Discrepancies were resolved through discussion. | ||||||
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Hospital admissions Show forest plot | 11 | 1769 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.75 [0.60, 0.92] |
| 2 Intensive care unit (ICU) admissions Show forest plot | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 3 High dependency unit (HDU) admissions Show forest plot | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 4 ED treatment duration (minutes) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 5 Length of hospital stay (days) Show forest plot | 3 | 949 | Mean Difference (IV, Fixed, 95% CI) | ‐0.03 [‐0.33, 0.27] |
| 6 Readmission Show forest plot | 2 | 887 | Odds Ratio (M‐H, Fixed, 95% CI) | 2.30 [0.66, 7.99] |
| 7 Heart rate (bpm) Show forest plot | 4 | 1195 | Mean Difference (IV, Fixed, 95% CI) | ‐2.37 [‐4.13, ‐0.61] |
| 8 Respiratory rate (breaths/min) Show forest plot | 4 | 1195 | Mean Difference (IV, Fixed, 95% CI) | ‐0.28 [‐0.77, 0.20] |
| 9 Systolic blood pressure (mmHg) Show forest plot | 4 | 1264 | Mean Difference (IV, Fixed, 95% CI) | 0.08 [‐1.89, 2.05] |
| 10 FEV1 (% predicted) Show forest plot | 4 | 523 | Mean Difference (IV, Fixed, 95% CI) | 4.41 [1.75, 7.06] |
| 11 PEF (% predicted) Show forest plot | 3 | 1129 | Mean Difference (IV, Fixed, 95% CI) | 4.78 [2.14, 7.43] |
| 12 PEF (L/min) Show forest plot | 8 | 1460 | Mean Difference (IV, Fixed, 95% CI) | 17.40 [8.64, 26.17] |
| 13 Borg Dyspnoea Scale score Show forest plot | 4 | Mean Difference (IV, Fixed, 95% CI) | Subtotals only | |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Hospital admissions (by severity) Show forest plot | 11 | 1743 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.76 [0.62, 0.95] |
| 1.1 Moderate | 2 | 791 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.76 [0.55, 1.04] |
| 1.2 Severe | 6 | 474 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.87 [0.58, 1.31] |
| 1.3 Life threatening | 7 | 478 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.69 [0.46, 1.03] |
| 2 Hospital admissions (by co‐medications) Show forest plot | 11 | 1769 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.75 [0.60, 0.92] |
| 2.1 Nebulised ipratropium bromide | 4 | 1072 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.73 [0.55, 0.96] |
| 2.2 No nebulised ipratropium bromide | 7 | 697 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.77 [0.55, 1.06] |
| 3 Hospital admissions (risk of bias sensitivity) Show forest plot | 8 | 1437 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.72 [0.57, 0.91] |
| 4 Hospital admissions (unpublished sensitivity) Show forest plot | 10 | 1638 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.73 [0.58, 0.91] |
| 5 PEF % predicted (Goodacre change score sensitivity) Show forest plot | 3 | 1129 | Mean Difference (IV, Fixed, 95% CI) | 1.57 [‐0.55, 3.69] |
| 6 PEF L/min (Goodacre change score sensitivity) Show forest plot | 8 | 1460 | Mean Difference (IV, Fixed, 95% CI) | 9.44 [2.07, 16.81] |
