Authors: Aiko Honda (Department of Pediatrics, Showa Medical University Graduate School of Medicine, Tokyo, Japan; Institute of Clinical Epidemiology (iCE), Showa Medical University, Tokyo, Japan), Yoshitaka Watanabe (Children’s Medical Center, Showa Medical University Northern Yokohama Hospital, Kanagawa, Japan), Yoshifusa Abe (Children’s Medical Center, Showa Medical University Koto Toyosu Hospital, Tokyo, Japan; Institute of Clinical Epidemiology (iCE), Showa Medical University, Tokyo, Japan), Sojiro Kusumoto (Department of Medicine, Division of Respiratory Medicine and Allergology, Showa Medical University Hospital, Tokyo, Japan), Takeshi Hasegawa (Institute of Clinical Epidemiology (iCE), Showa Medical University, Tokyo, Japan; Department of Hygiene, Public Health, and Preventive Medicine, Showa Medical University Graduate School of Medicine, Tokyo, Japan; Department of Nephrology, Showa Medical University Graduate School of Medicine, Tokyo, Japan), Hisashi Noma (Department of Interdisciplinary Statistical Mathematics, The Institute of Statistical Mathematics, Tokyo, Japan; Institute of Clinical Epidemiology (iCE), Showa Medical University, Tokyo, Japan), Erika Ota (Global Health Nursing, St.Luke’s International University, Tokyo, Japan; Institute of Clinical Epidemiology (iCE), Showa Medical University, Tokyo, Japan), Takanori Imai (Department of Pediatrics, Showa Medical University Graduate School of Medicine, Tokyo, Japan), Noyuri Yamaji (Institute of Clinical Epidemiology (iCE), Showa Medical University, Tokyo, Japan; Department of Family Nursing, Division of Health Sciences and Nursing, Graduate School of Medicine, The University of Tokyo, Tokyo, Japan), Edward Barroga (Medical English Education Center, Showa Medical University School of Medicine, Tokyo, Japan)
Source: The Cochrane Database of Systematic Reviews
Authors: Aiko Honda, Yoshitaka Watanabe, Yoshifusa Abe, Sojiro Kusumoto, Takeshi Hasegawa, Hisashi Noma, Erika Ota, Takanori Imai, Noyuri Yamaji, Edward Barroga
This is a protocol for a Cochrane Review (intervention). The objectives are as
To evaluate the clinical effectiveness and safety of high‐flow nasal cannula (HFNC) oxygen therapy compared with conventional oxygen therapy for the management of asthma exacerbations in children and adults.
We will also explore potential equity implications, including differences in access to and outcomes of HFNC oxygen therapy across diverse healthcare settings and populations.
Asthma is a chronic inflammatory disease of the airways, characterized by a variety of respiratory symptoms, such as wheezing, shortness of breath, chest tightness, and cough, and airflow obstruction that is often reversible. It is recognized as the second most prevalent chronic respiratory disease worldwide, with an estimated global burden affecting approximately 272 million people across all age groups [1]. Data from the Global Burden of Disease Study (2015) indicate that asthma is the second leading cause of mortality among chronic respiratory disorders [2]. Some people with severe exacerbation (‘asthma attack’) do not improve with medications, such as bronchodilators and systemic administration of corticosteroids, and may require ventilation support to maintain breathing [3].
Standard treatment for acute asthma exacerbations includes inhaled short‐acting beta‐agonists, systemic corticosteroids, oxygen therapy, and in severe cases, non‐invasive ventilation (NIV) or invasive mechanical ventilation. Conventional oxygen therapy is typically delivered at a low‐flow rate, by nasal cannula or face mask. NIV includes non‐invasive positive pressure ventilation (NPPV) and high‐flow nasal cannula oxygenation (HFNC).
We will focus on the comparison between HFNC and conventional oxygen therapy. Although the potential benefits of NPPV are an important clinical consideration, we chose to focus on HFNC in this review because it is a less invasive option and is expected to become more widely adopted in clinical practice. HFNC oxygen therapy is a non‐invasive respiratory management method that has been used for about 30 years. It delivers warmed, humidified, high‐flow oxygen intranasally [4]. High‐flow oxygen pushes out the dead space in the nasopharynx and stabilizes the ventilation rate. HFNC oxygen therapy has a number of advantages. It provides constant positive end‐expiratory pressure, maintains ciliary function by delivering heated and humidified oxygen, and minimizes the person's discomfort [5]. Traditionally, HFNC has been used for respiratory infections and post‐extubation complications. More recently, it has been approved for use in bronchial asthma attacks, but evidence for its use is lacking [6].
Recently, systematic reviews explored the efficacy of HFNC oxygen therapy for bronchiolitis [7, 8], and acute respiratory failure [9]. Since 2018, several randomized controlled trials (RCTs) have examined the effect of HFNC on asthma [10, 11, 12, 13]. A systematic review also explored the use of HFNC in the management of acute asthma [14].
This review will include additional recent studies not covered in previous reviews, allowing us to present updated evidence and potentially new insights into this topic. HFNC oxygen therapy may be shown to be an effective and less invasive respiratory support alternative to intubation and NPPV for acute exacerbations of asthma.
Asthma is a highly prevalent disease worldwide. Although the number of deaths due to acute exacerbations has declined with the introduction of inhaled steroids and biologic agents as long‐term controls, the asthma mortality rate was reported to be 5.96 per 100,000 (95% confidence interval 4.74 to 7.22) in 2019 [15]. Reducing asthma deaths is a priority for asthma treatment, and new treatment proposals for severe symptoms aim to reduce asthma deaths. Prevention of severe conditions can also reduce the health burden of people with asthma and their families, and may also lead to long‐term improvement of respiratory function, reduction of exacerbations, and improvement of quality of life. In addition, if non‐invasive management can be introduced to improve respiratory status and avoid tracheal intubation, ventilator‐related infections and muscle weakness caused by prolonged bed rest can be prevented, which may contribute to shorter hospital stays.
To evaluate the clinical effectiveness and safety of high‐flow nasal cannula (HFNC) oxygen therapy compared with conventional oxygen therapy for the management of asthma exacerbations in children and adults.
We will also explore potential equity implications, including differences in access to and outcomes of HFNC oxygen therapy across diverse healthcare settings and populations.
We will include individual, cluster, and cross‐over randomized controlled trials (RCTs). These designs will be included to ensure comprehensive coverage of the available evidence. We will account for potential biases specific to each design; for example, adjusting for clustering effects in cluster‐RCTs and using only first‐phase data in cross‐over trials to minimize carry‐over effects.
We will include people of all ages, both children and adults, who are being treated for an acute exacerbation of asthma in any healthcare setting, including outpatient clinics, emergency departments, or hospitals.
We will use guideline criteria, such as those outlined in the 2021 Global Strategy for Asthma Management and Prevention to define an asthma exacerbation [16], or the set of criteria predefined in the included studies. The diagnosis of asthma must be confirmed by a physician before enrollment.
If a study includes a broader population but reports data separately for the subgroup that meets our inclusion criteria (e.g. children with acute asthma), we will include and extract data only for the relevant subgroup. If subgroup data are not reported separately, we will attempt to contact the study authors. If subgroup data cannot be obtained, we will exclude the study.
We will include studies that compared high‐flow nasal cannula (HFNC) oxygen therapy with conventional oxygen therapy for acute exacerbations of asthma; the concomitant use of medication for asthma is acceptable. Conventional oxygen therapy refers to the standard methods of delivering supplemental oxygen to people using low‐flow or fixed‐flow systems. These systems provide oxygen at flow rates that are generally lower than the person’s peak inspiratory flow, and the oxygen is often not humidified or heated. For example, we will include the use of a nasal cannula, simple face mask, venturi mask, or non‐rebreather mask.
The critical and important outcomes in this systematic review include all core outcomes for asthma exacerbations and mechanical ventilation (https://www.comet-initiative.org/). All will be measured at the longest follow‐up. If cross‐over trials are found, we will include only the short‐term outcomes, such as severity score and change in dyspnea.
Mortality rateIntubation rateChanges in severity scores (e.g. Pulmonary Index Score [17], Modified Wood’s Clinical Asthma Score [mWCAS] [18], Childhood Asthma Severity Score [CAS score] [19])Changes in dyspnea (using the Borg scale [20], the Numerical Rating Scale [NRS] [21], or the Visual Analogue Scale [VAS] [22])All (serious and non‐serious) adverse events (e.g. respiratory infection, nasal mucosa, or skin trauma)
Respiratory rateDuration of hospital stay
We will search the following
Cochrane Central Register of Controlled Trials (CENTRAL), through the Cochrane Register of Studies Online (crso.cochrane.org; all years to date);MEDLINE PubMed (1946 to date);Embase Ovid SP (1974 to date);PsycINFO Ovid SP (1967 to date);CINAHL EBSCO (Cumulative Index to Nursing and Allied Health Literature; 1937 to date);AMED EBSCO (Allied and Complementary Medicine; 1995 to date).
We will use the search strategies proposed in Supplementary material 1. An Information Specialist (Asako Baraki) drafted the search strategy for MEDLINE, which will be translated using appropriate controlled vocabulary and syntax for other databases.
We will conduct additional searches in the following clinical trial registry
US National Institutes of Health Ongoing Trials Register ClinicalTrials.gov (www.clinicaltrials.gov);WHO International Clinical Trials Registry Platform (www.who.int/trialsearch).
We will conduct additional searches of trial registries and grey literature databases to identify articles that might not have appeared in the main electronic database searches (see Supplementary material 1). We will also check the reference lists of retrieved articles and reviews and ask field experts if they know of any relevant ongoing or unpublished trials. As part of this process, we will check whether the included or potentially eligible studies have any post‐publication amendments, such as expressions of concern, errata, corrigenda, or retractions. We will search the Retraction Watch Database and review notices on journal websites and PubMed (https://pubmed.ncbi.nlm.nih.gov/) to identify such amendments.
We will collect all titles and abstracts from the electronic search and download them into Rayyan [23], a web‐based platform for screening systematic review records, and Covidence [24], a literature management tool designed to support systematic review workflows. We will remove duplicates, and five review authors (AH, YW, YA, SK, NY) will independently screen the titles and abstracts. We will retrieve full‐text study reports for all references coded as ‘include’ by the review authors. The same five review authors will independently screen the full‐text reports and record reasons for exclusion for the excluded studies. We will resolve disagreements through discussion, or if required, by consulting one of the clinical authors (TH).
We will screen all included studies for retractions and expressions of concern throughout the review process. We will exclude retracted studies, and consider those with expressions of concern in sensitivity analyses, following guidance in the Cochrane Handbook for Systematic Reviews of Interventions, chapter 7 [25].
Five review authors (AH, YW, YA, SK, NY) will extract the following study data from the included studies.
Methods: study design, total duration of study, details of any 'run‐in' period, number of study centers and location, study setting, withdrawals, and date of studyParticipants: number, mean age, age range, gender, severity of condition, diagnostic criteria, baseline lung function, smoking history, inclusion criteria, and exclusion criteriaInterventions: intervention (including HFNC flow rate and duration of use), comparison, concomitant medications, and excluded medicationsOutcomes: critical and important outcomes specified and collected, and time points reportedNotes: funding for studies and notable conflicts of interest of trial authors
We will resolve disagreements by reaching a consensus and consulting with one of the clinical authors (TH) when required. One review author (AH) will transfer the study characteristics and risk of bias judgments into Review Manager (RevMan) [26], and five review authors (AH, YW, YA, SK, NY) will check and transfer the study data into the analyses.
Five review authors (AH, YW, YA, SK, NY) will independently assess the included studies for risk of bias, using Cochrane's ROB 2 [27]. The assessment will focus on the main outcomes, included in the summary of findings table. Specifically, we will apply RoB 2 to the following
Mortality rateIntubation rateChanges in severity scoresChanges in dyspneaAdverse events (all serious and non‐serious)Respiratory rateDuration of hospital stay
Any disagreement will be resolved by discussion, and if necessary, involving a third author (TH). For assessment of the overall risk of bias, we will use the predefined signaling questions and algorithm tools [28]. Our primary effect of interest is the effect of assignment to the intervention at baseline, regardless of whether the intervention was received as intended (i.e. the intention‐to‐treat effect).
We will evaluate the trials for risk of bias in the following domains.
The randomization processDeviations from intended interventionsMissing outcome dataMeasurement of the outcomeSelection of the reported results
We will judge each domain using the following RoB 2 low risk of bias, some concerns, or high risk of bias, and we will provide justifications for each judgment in the risk of bias table, including quotes from the study report, when applicable. For each domain, we will summarize the bias risk judgments between different studies. We will use the following algorithm to judge the overall risk of bias for each outcome.
Low risk of all domains are judged to be at low risk of bias.Some at least one domain is judged to raise some concerns, but none are at high risk of bias.High risk of at least one domain is judged to be at high risk of bias, or there are multiple domains with some concerns that substantially lower confidence in the result.
We will use risk ratios (RRs) for dichotomous outcomes (e.g. intubation and mortality rate). For continuous outcome data (e.g. changes in asthma score and respiratory rate), we will use mean difference (MD) if the same measurement scales were used, or standardized mean difference (SMD) if different scales were used. We will report 95% confidence intervals (95% CIs) for all outcomes.
We will analyze data using participants with one or more events as the unit of analysis. We will consider variations in study designs that may impact our analyses. If study participants are not individually randomized to one of two intervention groups, we will adjust the unit of analysis issue on the studies as follows.
If we identify studies in which the assignment to interventions was in population‐level clusters, we will extract statistics to adjust effect measures as appropriate for the study design. We will extract the statistical methods necessary to account for clustering to estimate the variance of the outcome, including design effects and within‐cluster correlations. We will include any results for which the effect measures for each outcome have been adjusted for intra‐data correlations in the meta‐analysis. If the results of cluster‐randomized trials were not adjusted for the impact of the cluster design, we will adjust the sample size using the intra‐cluster correlation coefficient (ICC) estimates, described in the Cochrane Handbook for Systematic Reviews of Interventions, chapter 23 [29]. When we cannot obtain a reliable estimate of the ICC, we will use a summary measures approach and perform the analysis at the cluster level.
If we include any trials with multiple intervention arms that are compared against the same control condition in the same meta‐analysis, we will either combine groups to create a single pair‐wise comparison or select one pair of interventions and exclude the others. If the interventions in each arm are sufficiently clinically similar in the intervention delivery platform (e.g. two arms only differ by flow rate of HFNC), we will estimate the combined results for two intervention groups to one comparator group.
If we identify cross‐over trials, we will only include data from the first phase of the trial to avoid potential carry‐over effects. We will only extract short‐term outcomes measured after the first intervention period (e.g. severity scores, respiratory rate). We will not use outcomes influenced by longer‐term effects (e.g. length of stay, mortality).
We will contact investigators or study sponsors to verify key study characteristics and to obtain missing numerical outcome data, when possible (e.g. when an identified study was only an abstract). When this is not possible, and when missing data are thought to introduce serious bias, we will explore the impact of including such studies in the overall assessment of the results by performing a sensitivity analysis.
We will assess the risk of reporting bias using funnel plots to assess for the potential existence of small‐study bias if more than ten trials are included for analysis. We will also assess the existence of reporting bias, using the Egger’s test [30].
There could be between‐studies heterogeneity due to the severity of the acute asthma exacerbations or the age of the participants. Thus, we will use a random‐effects model with between‐study variance estimated by the restricted maximum‐likelihood (REML) method for our meta‐analyses, to address the clinical and methodological diversity and statistical heterogeneity expected in the included studies [31]. We will use a random‐effects model for the analysis of both dichotomous and continuous outcomes. We will use RevMan for the data analysis [26].
If meta‐analysis is not feasible because of substantial heterogeneity or insufficient data, we will synthesize findings narratively, following the Synthesis Without Meta‐analysis (SWiM) reporting guidelines, in the Cochrane Handbook for Systematic Reviews of Interventions, chapter 12 [32]. This will include a structured presentation of results in tables, grouped by outcome and intervention characteristics, with consideration of the direction and consistency of effects.
We will use the DerSimonian‐Laird random‐effects model for synthesis analyses to account for heterogeneity between studies. Statistical heterogeneity among the trials will be assessed using the heterogeneity variance Tau^2^, Higgins’ I² statistic, and the Q‐statistic (Cochrane’s test) [31].
To evaluate clinical heterogeneity due to potential effect modifiers, we will perform subgroup analyses or meta‐regression analyses. Specifically, we plan the following subgroup analyses of the critical outcomes to assess the influences of potential effect modifiers.
Age: Pediatric (< 15 years), adolescent, and adult (≥ 15 years)Income Level: high‐income countries versus low‐ and middle‐income countries, based on the World Bank Country and Lending Groups classification [33]Clinical participants receiving HFNC therapy in an intensive care unit (ICU) setting versus those treated in non‐ICU hospital settings
In this systematic review, we acknowledge the importance of equity‐related factors. We aim to explore whether the effects of HFNC oxygen therapy differ based on key characteristics such as age, income levels, and healthcare settings. To guide our assessment of these equity considerations, we will use the PROGRESS‐Plus framework (Place of residence, Race/ethnicity, Occupation, Gender, Religion, Education, Socioeconomic status, Social capital; https://www.trialforge.org/trial-diversity/pro-edi/). This approach will help to ensure that our findings are applicable across diverse populations and contexts.
We will conduct sensitivity analyses for all main outcomes (mortality, intubation rate, changes in severity scores, changes in dyspnea, adverse events, respiratory rate, and duration of hospital stay). These analyses exclude studies judged to be at high overall risk of bias based on the RoB 2 tool, and assess the impact of missing data. We will also compare the results from a fixed‐effect model with those using a random‐effects model.
We will create a summary of findings table that includes all outcomes listed in the Outcome measures section at the longest follow‐up. The selected outcomes are as
Mortality rateIntubation rateChanges in severity scoresChanges in dyspneaAdverse eventsRespiratory rateDuration of hospital stay
We will use the five GRADE considerations (overall risk of bias, consistency of effect, imprecision, indirectness, and publication bias) [34] to assess the quality of a body of evidence as it relates to studies that contributed data to meta‐analyses for prespecified
Mortality rateIntubation rateChanges in severity scoresChanges in dyspneaAdverse eventsRespiratory rateDuration of hospital stay
We will use the methods and recommendations described in the Cochrane Handbook for Systematic Reviews of Interventions, chapter 14 on GRADE methodology [35], and GRADEpro GDT software [36].
Consumers will not be involved in this review, although the review authors do use the core outcome sets for the review’s outcomes, which were developed with consumer involvement.
Supplementary materials are available with the online version of this 10.1002/14651858.CD016205.
Supplementary materials are published alongside the article and contain additional data and information that support or enhance the article. Supplementary materials may not be subject to the same editorial scrutiny as the content of the article and Cochrane has not copyedited, typeset or proofread these materials. The material in these sections has been supplied by the author(s) for publication under a Licence for Publication and the author(s) are solely responsible for the material. Cochrane accordingly gives no representations or warranties of any kind in relation to, and accepts no liability for any reliance on or use of, such material.