Abstract
Background International asthma guidelines recommend against adrenaline administration in acute asthma unless associated with anaphylaxis or angioedema. However, administration of intra-muscular adrenaline in addition to nebulised selective β2-agonist is recommended for acute severe or life-threatening asthma in many pre-hospital guidelines. We conducted a systematic review to determine the efficacy of adrenaline in comparison to selective β2-agonist in acute asthma.
Methods We included peer-reviewed publications of randomised controlled trials (RCTs) that enrolled children or adults in any healthcare setting and compared adrenaline by any route to selective β2-agonist by any route for an acute asthma exacerbation. The primary outcome was treatment failure, as indicated by hospitalisation, stay >24hrs in emergency department, need for intubation, or death.
Results Thirty-eight of 1,140 studies were included, involving 2,275 participants. Overall quality of evidence was low. There was significant statistical heterogeneity, I2=56%. The pooled odds ratio for treatment failure with adrenaline versus selective β2-agonist was 0.99 (0.74 to 1.34), p=0.96. There was strong evidence that recruitment age-group was associated with different estimates of the risk of treatment failure; with studies recruiting adults-only having a lower risk of treatment failure with adrenaline. It was not possible to determine whether adrenaline in addition to selective β2-agonist improved outcomes.
Conclusion The limited evidence available suggests that adrenaline and selective β2-agonists have similar efficacy in acute asthma and does not support the use of adrenaline in addition to selective β2-agonists in acute asthma. There is a need for high-quality double-blind RCTs to address this issue.
PROSPERO registration number CRD42017079472
Introduction
Adrenaline, also known as epinephrine, has been used for the management of acute asthma for over 100 years.[1,2] Adrenaline is a non-selective alpha and beta-adrenergic agonist, which causes bronchodilation through stimulation of β2 adrenergic receptors leading to relaxation of bronchial smooth muscle. Alpha-agonist effects of adrenaline may also reduce airway oedema.[3,4] Historically, parenteral adrenaline by the subcutaneous or intra-muscular route was the mainstay of acute asthma treatment, [1] but inhaled selective beta2-agonists are now the primary bronchodilators of choice and international asthma guidelines do not recommend the use of adrenaline except in the context of concomitant acute asthma and anaphylaxis or angioedema.[5–8] However, the administration of intra-muscular adrenaline in addition to nebulised selective β2-agonist for acute severe or life-threatening asthma remains part of many pre-hospital ambulance guidelines, including in the UK, New Zealand and the USA[9–12] and continues to be used and advocated for in hospital by some clinicians.[4,13–17]
This systematic review and meta-analysis aimed to determine the efficacy of adrenaline in comparison with, and in addition to, selective β2-agonist for acute asthma in either children or adults with acute asthma.
Methods
Protocol and Registration
This systematic review was prospectively registered on the PROSPERO database on 28th November 2017, registration number CRD42017079472,[18] and is reported in line with the PRISMA guidelines.[19]
Eligibility Criteria
We included randomised controlled trials that enrolled children, aged 18 months to 17 years, or adults, aged 18 or over, in any healthcare setting, including pre-hospital treatment. All study participants must have had a clinical diagnosis of an acute asthma exacerbation.
Studies in non-English languages were considered for inclusion. The study must have been published in a peer-reviewed journal.
Types of intervention
The target intervention, as registered on the PROSPERO database, was administration of adrenaline by any route, compared to administration of selective beta-2-agonist by any route.
Types of outcomes
The primary outcome was treatment failure, as indicated by any one of: hospitalisation, length of stay in emergency department greater than 24 hours, the need for intubation, or death. We recognised a priori that these data may not have been reported or able to be extracted from the identified literature. Secondary outcomes of particular interest were peak expiratory flow rate, forced expiratory volume in the first second (FEV1), heart rate, length of stay, respiratory rate, oxygen saturation, pH, PCO2, and lactate; as well as the occurrence of adverse events.
Information Sources and Search
We searched the Cochrane Airways Group Register (Wiley interface, issue 10) and Scopus databases. Our search strategy is detailed in the supplementary appendix and includes search filters for the terms of (‘asthma’ OR ‘asthmatic’) AND (‘epinephrine’ OR ‘adrenaline’). Reference lists from included studies and known reviews were also hand searched. The searches were originally conducted 15th June 2018 and a repeat search on 23rd October 2020 identified no additional relevant studies.
Study Selection
Selection and data extraction utilised the Covidence online platform.[20] In the screening phase, authors (CB, DS, JH, JF or JS) reviewed the title of the full list of articles identified by the search strategy. After the initial screening, abstracts were examined and articles which appeared to meet the inclusion criteria were reviewed in full by two authors, alongside any articles where there was doubt regarding eligibility. Disagreements were resolved by consensus or arbitration by a third author.
Data Collection Process and Data items
Data for eligible studies was extracted independently by two authors (CB, JH or JS). Disagreements were resolved by consensus or arbitration by a third author (JF).
Risk of Bias
We did not plan to examine for publication bias. We assessed reporting bias within individual trials using the risk of bias framework recommended by the Cochrane Collaboration.[21]
Statistical Analysis
For the primary outcome variable of probability of treatment failure Peto’s method was planned to be used to pool the reported occurrence, to take account of the likely zero cell counts in at least one of the arms. Continuous variables were to be pooled by inverse variance weighting of the mean difference based on reported means, standard deviations, and numbers of participants. Peto’s method was also planned to be used for the occurrence of an adverse event. We planned to evaluate statistical heterogeneity using the I-squared statistic and should it be present to attempt meta-regression based on the following study-level covariates: Route of administration of the intervention, route of administration of the control, predominant age group for the study; and sub-group according to the possible combinations of route of intervention, route of control and whether adrenaline was given in addition to beta-agonist. The method of partitioning the Chi-square statistic was used for the meta-regressions. Both fixed and random effects estimates are shown in the reported pooled analyses.
SAS version 9.4 was used.
Results
Study selection
A flow diagram of study selection is shown in Figure 1. A total of 1,140 studies were screened, with 163 undergoing full text review. After full-text review 38 studies were included, involving 2,275 participants. Table 1 provides characteristics of included studies. The route of adrenaline administration was intramuscular for one study, nebulised in six and subcutaneous in the remaining 31. Two studies contained an arm in which adrenaline was administered in addition to selective β2-agonist, and in a further 5 studies some participants crossed over to receive adrenaline having previously received selective β2-agonist. In the remainder the group receiving adrenaline did not receive selective β2-agonist.
Risk of bias within studies
Risk of bias assessments are shown in Figure 2. The most recent study was conducted 15 years before this review and 26/38 were conducted more than 30 years prior. Manuscript style, methodology and reporting reflect the norms applicable at the time of publication. As a result, details of study methodology and primary outcome variables were unclear in many cases and 27/38 studies were adjudged to be at high risk of bias in at least one aspect.
Synthesis of results
Data on treatment failure was available for 17/38 studies. The individual and pooled study estimates are shown in Table 2 and illustrated in the Forest plot in Figure 3. There was important statistical heterogeneity with a Chi-square statistic of 36.4, (16 DF), P=0.003;, I2 (95% CI) 56% (24.4 to 74.5). Neither the fixed or random effects estimates favoured one or other treatment: Peto’s odds ratio pooled fixed effect 0.99 (0.74 to 1.34), p=0.96; and random effects 1.07 (0.65 to 1.76), p=0.79.
For the four sets of study-level covariates there was no evidence that heterogeneity was explained by the intervention route of administration (Table 3). There was weak evidence (P=0.06) that some of the heterogeneity was explained by the control route of administration with the parenteral route at higher risk of treatment failure compared to the inhaled route; although for the latter there was still important heterogeneity within that sub-group. There was weak evidence (P=0.09) that some of the heterogeneity was explained by the sub-groups with sub-groups 2 (IM adrenaline versus nebulised selective β2-agonist) and 4 (SC adrenaline versus parenteral selective β2-agonist) having a higher risk of treatment failure than the other three sub-groups; although there was still important heterogeneity within the sub-groups. Finally there was strong evidence that the recruitment age-group was associated with different estimates of the risk of treatment failure (P=0.006); with the studies recruiting adults-only having a lower risk of treatment failure with parenteral adrenaline and children-only a higher risk of treatment failure. There was still unexplained heterogeneity within studies that recruited a mixture of adults and adolescents.
There was also marked heterogeneity in design, country, population, intervention and control between studies. It was usually not possible to extract secondary outcome data. As a result combining secondary outcome data for meta-analysis was inappropriate and a narrative summary of these data was performed. For this summary, because of the heterogeneity of study design, the studies have been grouped according to the nature of the intervention and comparator. Summary details of key outcomes and adverse events are given in Tables S1 and S2 respectively in the supplementary appendix.
Sub-group 1: Subcutaneous adrenaline plus nebulised selective β2-agonist compared to nebulised selective β2-agonist alone
The question most closely linked to current clinical practice is whether addition of IM adrenaline to inhaled or nebulised selective β2-agonist improves outcome. No study addressed this issue. Two of the 38 studies included arms where one group received subcutaneous adrenaline in addition to selective β2-agonist and another received selective β2-agonist alone.
Kornberg et al. reported an unblinded parallel group RCT of 43 children aged 3 to 12yrs attending a paediatric emergency department for acute asthma. Patients received either 2.5mg salbutamol nebuliser repeated as required every 20 to 30 minutes or a single subcutaneous injection of Sus-Phrine (a combination of immediate- and sustained-release forms of adrenaline) 0.005ml/kg followed by 2.5mg salbutamol nebuliser repeated as required every 20 to 30 minutes. Treatment response was assessed at 20 and 120min. Treatment failure, in the form of hospitalisation, was present in 1/20 patients in the adrenaline group compared with 3/23 patients who received only selective β2-agonist. Both groups showed significant improvement in clinical score, peak flow and respiratory rate; with no difference between the groups.
Quadrel et al reported an unblinded three arm parallel group RCT of 154 adult patients receiving pre-hospital treatment for acute asthma. The authors compared nebulised metaprotenerol 2.5mg, subcutaneous adrenaline 0.3mg, or both. Peak flow improved with treatment in all three groups and there was no difference in treatment failure, peak flow, respiratory rate or heart rate between the groups.
Sub-groups 2&3: Parenteral adrenaline compared to nebulised selective β2-agonist
Seventeen studies compared parenteral (IM or SC) adrenaline with nebulised selective β2-agonist. Ten studies reported no difference in efficacy as measured by either treatment failure or lung function. Five studies suggested that selective β2-agonist was a more effective bronchodilator and two studies suggested adrenaline was more effective. There was weak and inconsistent evidence suggesting more frequent side-effects with adrenaline. Six studies reported more frequent side-effects (including tremor, agitation and headache) with adrenaline, two studies reported more frequent tremor with selective β2-agonist (both used terbutaline), five studies reported no difference in side-effect profile and four studies did not report details of side-effects.
Sub-group 4: Subcutaneous adrenaline compared to parenteral selective β2-agonist
Thirteen studies compared subcutaneous adrenaline and parenteral (subcutaneous or intravenous) selective β2-agonist. All 13 reported similar initial bronchodilation with adrenaline or selective β2-agonist although one study noted that by 90 minutes the bronchodilation caused by adrenaline was significantly less than that caused by salbutamol. Nine studies reported no difference in side effects, two reported greater side-effects (agitation, “excitement”, and headache) with adrenaline, and one reported mild tremor in two patients in the salbutamol group.
Sub-group 5: Subcutaneous adrenaline compared to oral selective β2-agonist
One study (Calvo et al. 1982) compared subcutaneous adrenaline and oral fenoterol and reported greater bronchodilation and a favourable side-effect profile with fenoterol.
Sub-group 6: Nebulised adrenaline compared to nebulised selective β2-agonist
Six studies explored this comparison and all showed equivalent efficacy. Three studies reported equivalent side-effect profile between groups, two reported increased frequency of mild side-effects with adrenaline and one did not report data on side-effects.
Unpublished studies
A search of the WHO International Clinical Trials Registry Platform identified one unpublished study of intramuscular adrenaline in addition to selective β2-agonist in 49 children aged 6 to 18 with a severe asthma exacerbation, NCT01705964. This study has been completed but not yet been reported and results are not available on the clinicaltrials.gov registry.
Discussion
This systematic review and meta-analysis of randomised controlled trials of adrenaline versus selective β2-agonist for acute asthma found no evidence of a different risk of treatment failure for adrenaline by any route compared to administration of selective β2-agonist by any route. However, the individual studies were largely at high risk of bias, there was evidence of statistical heterogeneity, and the studies were clinically heterogeneous. Meta-regression provided weak evidence that some of the heterogeneity was explained by the route of administration of intervention and control, with parenteral routes of administration associated with greater risk of treatment failure. There was strong evidence that the recruitment age-group was associated with different estimates of the risk of treatment failure; with the studies recruiting adults-only having a lower risk of treatment failure with adrenaline. It is unclear whether this represents a genuine difference in treatment response between adults and children or simply reflects underlying differences in the design, setting, and study methodology. Reporting of important clinical outcomes, here considered secondary outcomes, was very poor and precluded a pooled analysis, however a narrative review was consistent with no difference in other outcomes.
It was not possible to determine whether the addition of intramuscular adrenaline to nebulised selective β2-agonist improves outcomes in patients with acute asthma. There is very limited evidence from two unblinded studies of the addition of subcutaneous adrenaline to nebulised selective β2-agonist, which suggests that the addition of subcutaneous adrenaline did not reduce treatment failure or lead to greater bronchodilation than nebulised selective β2-agonist alone. The established pharmacokinetics of adrenaline versus other selective β2-agonists means that adrenaline dosing is required more frequently, typically every 20 minutes.[14]
There is inconsistent and low-quality evidence that parenteral adrenaline has a worse side-effect profile than parenteral selective β2-agonists. Side-effect data was inconsistently reported however agitation, tremor and headache appeared more frequently with adrenaline. Tachycardia was not consistently reported to be present more frequently with adrenaline and the data was insufficient to assess whether cardiovascular side-effects were more common with adrenaline.
It is appropriate to highlight that adrenaline was the reference treatment at the time many of the studies included in this review were published, with selective β2-agonist less frequently used, and so the purpose of the majority of studies was to compare the efficacy and side-effect profile of selective β2-agonists to that standard. This may explain why the clinical scenario of most relevance today, addition of intramuscular adrenaline to selective β2-agonist, was rarely addressed.
We did not limit our search by language and location although we were only able to include studies written in English, French or Spanish, and so excluded 15 studies. A limitation of this review is that the overall low quality of evidence with marked clinical heterogeneity precluded meta-analysis of secondary outcome data. A weakness of this analysis is that some of the sub-groups only had one or two studies contributing data and so estimates of heterogeneity explicable by these sub-groups and the effects within sub-groups may not be reliable. The findings of this review are limited to the doses and routes studied in the included trials. Due to the limited evidence base, we are unable to draw definitive conclusions on the efficacy of adrenaline in comparison to selective β2-agonist in the treatment of acute asthma exacerbations and there is clinical equipoise as to whether addition of adrenaline to selective β2-agonist improves outcome compared to selective β2-agonist alone.
There is a clear difference between the management recommended in national and international asthma guidelines, where adrenaline use is only recommended if there are features of anaphylaxis or angioedema, and that recommended in many pre-hospital care ambulance guidelines, where intramuscular adrenaline is recommended in severe or life-threatening asthma. There is therefore an urgent need for high quality evidence to determine whether there is a role for adrenaline in the management of severe and life-threatening asthma. If there is a benefit, asthma guidelines and clinical practice in hospital will need to be updated to include the use of parenteral adrenaline in addition to selective β2-agonist in this circumstance. Conversely, if there is no evidence of benefit updates to pre-hospital guidelines will be necessary to advise against the use of parenteral adrenaline in the absence of anaphylaxis.
Conclusions
The limited evidence available suggests that adrenaline and selective β2-agonists have similar efficacy in acute asthma but that adrenaline has a worse side-effect profile. There was no evidence of benefit from the use of adrenaline in addition to selective β2-agonists in acute asthma but the available evidence is insufficient to rule out a clinically important benefit. There is clinical equipoise as to whether addition of adrenaline improves outcomes and a need for high-quality double-blind RCTs to address this issue. The primary question that studies should aim to address is whether the addition of intramuscular adrenaline to inhaled selective β2-agonist reduces the risk of treatment failure or death in adults or children with acute asthma.
Data Availability
The data that support the findings of this study are available from the corresponding author upon reasonable request
Funding
The MRINZ is supported by Independent Research Organisation funding from the Health Research Council of New Zealand.
Contribution
JF, CB, DS, JH and RB proposed the systematic review. JF, CB, DS, JH and JS screened the studies and extracted the data. MW conducted the meta-analysis. JF wrote the first draft of the manuscript. All authors had access to the data and read and approved the final version of the manuscript. JF is guarantor of the data.
Acknowledgements
The authors would like to thank Denise Fabian and Jung Cho for assistance obtaining journal articles, Gwendoline Peter for assistance translating one of the manuscripts and Andrew Davies for clinical conversations which helped conceive this research question.
References
