You are here

Article Text

Article menu

Download PDFPDF

Original article
High-flow nasal cannula therapy for children with bronchiolitis: a systematic review and meta-analysis
  1. Jilei Lin1,2,
  2. Yin Zhang1,2,
  3. Limei Xiong1,2,
  4. Sha Liu1,
  5. Caihui Gong1,
  6. Jihong Dai1,2
  1. 1 Respiratory Department, Children’s Hospital of Chongqing Medical University, Chongqing, China
  2. 2 Ministry of Education Key Laboratory of Child Development and Disorders, Chongqing, China
  1. Correspondence to Professor Jihong Dai, Respiratory Department, Children’s Hospital of Chongqing Medical University, Chongqing 400065, China; danieljh{at}163.com

Abstract

Objectives To review the effects and safety of high-flow nasal cannula (HFNC) for bronchiolitis.

Methods Six electronic databases including PubMed, EMBASE, Cochrane Central Register of Controlled Trials, China National Knowledge Infrastructure, CQ VIP Database and Wanfang Data were searched from their inception to 1 June 2018. Randomised controlled trials (RCTs) which investigated the effects of HFNC versus other forms of oxygen therapies for bronchiolitis were included.

Results Nine RCTs with 2121 children met the eligibility criteria. There was no significant difference in length of stay in hospital (LOS), length of oxygen supplementation (LOO), transfer to intensive care unit, incidence of intubation, respiratory rate, SpO2 and adverse events in HFNC group compared with standard oxygen therapy (SOT) and nasal continuous positive airway pressure (nCPAP) groups. A significant reduction of the incidence of treatment failure (risk ratio (RR) 0.50, 95% CI 0.40 to 0.62, p<0.01) was observed in HFNC group compared with SOT group, but there was a significant increase of the incidence of treatment failure (RR 1.61, 95% CI 1.06 to 2.42, p0.02) in HFNC group compared with nCPAP group. In subgroup analysis, LOS was significantly decreased in HFNC group compared with SOT group in low-income and middle-income countries.

Conclusion The systematic review suggests HFNC is safe as an initial respiratory management, but the evidence is still lacking to show benefits for children with bronchiolitis compared with SOT or nCPAP.

  • bronchiolitis
  • children
  • high-flow nasal cannula
  • meta-analysis

https://doi.org/10.1136/archdischild-2018-315846

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    What is already known on this topic?

    • High-flow nasal cannula (HFNC) has been paid attention to in treatment of bronchiolitis increasingly, but there is limited evidence on its effects and safety.

    • Observational studies indicated that HFNC may be more comfortable, less invasive and potentially has lower costs and fewer adverse events compared with standard oxygen therapy.

    What this study adds?

    • HFNC is safe as an initial respiratory management, but it did not significantly benefit children with bronchiolitis compared with standard oxygen therapy and nasal continuous positive airway pressure.

    • HFNC may decrease the rate of treatment failure for children with bronchiolitis compared with conventional oxygen supplementation.

    Introduction

    Bronchiolitis is a lower respiratory tract infection commonly caused by virus in children.1 The annual hospital admission rate was 46.1 per 1000 infants aged <1 year with an increasing rate having risen by an average of 1.8% per year in England, and a small number even require intensive care.2 3 Current managements of bronchiolitis for hospitalised children are limited to supplemental oxygen, respiratory support and fluid replacement.1

    In recent years, heated, humidified, high-flow nasal cannula (HFNC) oxygen therapy has emerged as a promising therapeutic option for children with bronchiolitis. It provides humidified and heated air-oxygen mixture with high flow through a nasal cannula.4 Physiologic measures of respiratory effort can be improved with generation of continuous positive airway pressure (CPAP) in bronchiolitis.5 6 HFNC may be more comfortable, less invasive and potentially has lower costs and fewer adverse events compared with other conventional non-invasive ventilation therapies.7–9 HFNC has been paid attention to in treatment of bronchiolitis increasingly, but there is limited evidence on its effects and safety. Emerging clinical evidence from observational clinical studies suggested HFNC could reduce work of breathing and may decrease need for intubation.7 10 A systematic review11 published in 2014 only included one randomised   controlled trial (RCT)12 and reported insufficient evidence by assessing the effectiveness of HFNC for treating children with bronchiolitis compared with conventional therapy (head-box oxygen). The study12 included in that systematic review was published as a letter to editor in 2012 and there was no pooled data in that systematic review, which provided a weak evidence. One review13 which evaluated CPAP and HFNC in bronchiolitis concluded CPAP and HFNC may improve physiologic and clinical outcomes associated with respiratory distress and failure due to bronchiolitis. However, quantitative support is lacking and evidence was conflicting around this result. In recent years, some RCTs14 15 did not show significant benefits of HFNC on bronchiolitis compared with standard oxygen therapy (SOT).

    Considering factors above and more RCTs have been published in recent years, we have incorporated studies till now and performed a systematic review and meta-analysis to figure out the effects and safety of HFNC for children with bronchiolitis. This review should provide more evidence on effects of HFNC for bronchiolitis.

    Materials and methods

    Protocol and registration

    The study protocol was performed in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Statement.16 The protocol was registered with the PROSPERO International Prospective Register of Systematic Reviews (CRD42018099477).

    Search strategy

    We searched the following six electronic databases: PubMed (1966 to June 2018), EMBASE (1974 to June 2018), Cochrane Central Register of Controlled Trials (through May 2018), China National Knowledge Infrastructure (1915 to June 2018), CQ VIP Database (1989 to June 2018), Wanfang Data (1995 to June 2018) and international trial registers (ClinicalTrials.gov and ISRCTN Registry) from their inception to 1 June 2018 regardless of language. Search strategies as (bronchiolitis OR bronchopneumonia OR respiratory syncytial virus OR respiratory syncytial viruses OR RSV) AND (HFNC OR high flow nasal cannula OR humidified high-flow nasal cannula OR HHFNC OR heated humidified high-flow nasal cannula OR HHHFNC OR high flow nasal oxygen OR HFNO OR high flow oxygen OR nasal high flow) were used. The corresponding Chinese key terms were used in Chinese databases. We reviewed the reference lists of articles for other studies to supplement our search.

    Eligibility criteria

    RCTs that evaluated the effects of HFNC therapy on children with bronchiolitis were considered. The specific criteria were listed as follows: (1) children were under 24 months with a clinical diagnosis of bronchiolitis, (2) HFNC therapy was supplied in the experimental group, (3) other forms of therapy such as SOTs (nasal cannula or mask oxygen supplementation at low flow <2 L/min) or nasal CPAP (nCPAP) (the set of positive continuous pressure was not limited) with the same co-intervention as the experimental group were supplied in control group, (4) at least one of the primary or secondary outcomes was reported.

    Outcome measures

    Primary outcomes:

    1. Length of stay in hospital (LOS);

    2. Length of oxygen supplementation (LOO);

    3. Transfer to intensive care unit (ICU).

    Secondary outcomes:

    1. Treatment failure (discontinuation of current therapy and/or escalation of care due to the deterioration of disease or adverse events. Detailed definition in different studies is shown in online supplementation 1).

    2. Incidence of intubation.

    3. Length of non-invasive ventilation (length of the HFNC and other non-invasive ventilation in HFNC group, length of nCPAP and other non-invasive ventilation in control group).

    4. Paediatric intensive care unit (PICU) LOS.

    5. Length of wheezing.

    6. Respiratory rate, heart rate (HR), PaO2 of carbon dioxide (PaCO2), PaO2 of oxygen (PaO2) and pulse oxygen saturation (SpO2).

    7. Adverse events.

    Study selection and data extraction

    Two reviewers (JL and YZ) independently assessed the title and abstract of studies that were probably eligible. Then the full text was retrieved and the eligibility according to the eligibility criteria was assessed. Any disagreement was resolved through discussion with a third reviewer (CG). Data extraction was performed using a self-designed data collection form. We tried to contact the authors to obtain the original data of some included trials.

    Assessment of quality

    Two reviewers (LX and SL) independently assessed the methodological quality of all included studies using Cochrane collaboration risk of bias tool. The following domains were assessed: 1) generation of allocation sequence, 2) allocation concealment, 3) blinding of participants and researchers, 4) blinding of outcome assessors, 5) completeness of outcome data, 6) selective outcome reporting, 7) other risk of bias. Each item was marked by low, high or unclear risk of bias. Any disagreement was resolved through discussion with a third reviewer (JL).

    Synthesis methods and heterogeneity test

    Meta-analyses were conducted if data from studies were available and sufficient. Mean difference (MD) was used for continuous variables. Standard mean difference was used in continuous data analysis if the criterion or measurement for evaluating the results among different studies were different. Risk ratio (RR) was used for dichotomous variables. We calculated 95% CI for each effect size estimate. Random-effects model or fixed-effects model was used depending on the heterogeneity across studies. P<0.05 meant statistically significant. We estimated heterogeneity by the I2 statistic, which assessed the impact of heterogeneity in the meta-analysis. It demonstrated a low heterogeneity if I2 was <50%.17 To explore the causes of significant heterogeneity and identify factors modifying the effects of HFNC, we would perform subgroup analyses for primary outcomes if data were sufficient. The factors such as study location, age, sample size, mean LOS and flow rate of HFNC were prespecified for subgroup analyses. We planned to carry out sensitivity analyses for primary outcomes to investigate the effects of the trial’s quality on the outcomes by removing those studies rated as ‘high’ or ‘unclear’ risk of selection, performance or attrition bias to establish whether it was likely to impact on the findings. Begg’s test18 and Egger’s test19 were conducted to assess the potential publication bias of primary outcomes. Meta-analyses were conducted by Review Manager V.5.3 and STATA V.12.0.

    Results

    Literature search

    Our search obtained 486 records as a result of database searching. There were 321 records after 165 duplicates were removed. Two hundred eighty-five studies were removed after two reviewers (JL and YZ) went through titles and abstracts, resulting in 36 studies needing reading in full text. Seventeen studies including reviews and conference abstracts, nine observational studies and one protocol were removed in the process. None was included in manual retrieval. Consequently, nine RCTs14 15 20–26 were included in this review (figure 1). Five relevant ongoing studies were searched in the international trial registers, but no data were available from these studies.

    Figure 1
    Figure 1

    Study flow diagram.

    Characteristics of included studies

    Nine open-label RCTs were included in different area across the world including China, India, Australia and Europe. Details of these studies are provided in table 1. Two thousand one hundred twenty-one participants under 24 months of age diagnosed with bronchiolitis were involved in this review. Participants received oxygen delivered by HFNC therapy (oxygen flow >2 L/min) in experimental groups, and received oxygen delivered by SOT (nasal cannula or mask oxygen supplementation at low flow <2 L/min) or nCPAP (positive continuous pressure was set from 4 to 8 cmH2O) in control groups. Four studies14 15 22 25 compared the effect of HFNC with SOT (nasal cannula or mask oxygen supplementation at low -flow) as an oxygen support and three21 23 24 compared the effect of HFNC with nCPAP. One study26 set two groups (SOT and nCPAP) as control groups. One study20 investigated HFNC versus standard oxygen supplementation plus hypertonic saline solution (SOT+HSS) in improving respiratory distress in children with bronchiolitis. Outcomes including LOS, LOO, transfer to ICU, treatment failure, the incidence of intubation, length of non-invasive ventilation, PICU LOS, length of wheezing, respiratory rate, HR, PaCO2, PaO2, SpO2 and adverse events were reported. Three studies21 23 24 reported the length of non-invasive ventilation. HFNC was also regarded as a kind of non-invasive ventilation in these studies. So the length of non-invasive ventilation meant the length of applying HFNC and other non-invasive ventilation in HFNC group, while the length of nCPAP and other non-invasive ventilation in nCPAP group. We tried to contact the authors to obtain the original data of some included trials, but we failed to get the reply.

    View this table:
    Table 1

    Characteristic of included studies

    Quality of individual studies

    Risk of bias of nine included studies was assessed according to the Cochrane Handbook. Only one study25 used inappropriate random sequence generation. Three studies14 21 23 reported the adequate method of allocation concealment. None seemed to use blinding of participants and personnel for the characteristics of interventions. The blinding of outcome assessment was applied in one study,14 while other included studies remained unclear in this item. In terms of incomplete outcome data, all nine studies were at low risk of bias for no data were missing. Four included studies14 20 21 23 had low risk in selective reporting, while the others had unclear risk. Other risk including potential source of bias was not found in all included studies (figure 2).

    Figure 2
    Figure 2

    Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

    Meta-analysis of primary outcomes

    Length of stay in hospital

    Five studies reported LOS.14 15 22 25 26 However, one study15 was removed from meta-analysis because the researchers used median and IQR to describe LOS (p>0.05) and the original data could not be obtained. The pooled data suggested that LOS was not significantly reduced in HFNC group (MD −1.53, 95% CI −3.33 to 0.27, p =0.10) (day) with significant heterogeneity (I2=94%) compared with SOT. One study26 suggested that LOS was not significantly reduced in HFNC group compared with nCPAP group (p 0.40) (figure 3).

    Figure 3
    Figure 3

    Comparison of high-flow nasal cannula (HFNC) vs control outcome: length of stay. nCPAP, nasal continuous positive airway pressure; SOT, standard oxygen therapy. 

    Length of oxygen supplementation

    Three included studies14 15 26 reported LOO of patients in HFNC group compared with SOT group. One study15 used median and IQR to describe the LOO (p>0.05), but we could not obtain the original data. The pooled data suggested that LOO was not significantly reduced in HFNC group (MD −0.59, 95% CI −1.70 to 0.53, p =0.30) (day) with significant heterogeneity (I2=91%). One study26 suggested there was no significant reduction on LOO in HFNC group compared with nCPAP group (p 0.54) (figure 4).

    Figure 4
    Figure 4

    Comparison of high-flow nasal cannula (HFNC) vs control outcome: length of oxygen. nCPAP, nasal continuous positive airway pressure;  SOT, standard oxygen therapy.

    Transfer to ICU

    Three included studies14 15 20 reported the incidence of transfer to ICU of patients. The pooled data suggested there was no significant reduction on transfer to ICU in HFNC group (RR 1.30, 95% CI 0.98 to 1.72, p=0.06) with insignificant heterogeneity (I2=0%). No significant difference was obtained by comparing transfer to ICU in HFNC and SOT+HSS group in one study20 (p=0.64) (figure 5).

    Figure 5
    Figure 5

    Comparison of high-flow nasal cannula (HFNC) vs control outcome: transfer to intensive care unit. HSS, hypertonic saline solution; SOT, standard oxygen therapy.

    Meta-analysis of secondary outcomes

    Treatment failure

    Four included studies14 15 23 24 reported the incidence of treatment failure of patients. The pooled data suggested there was a significant reduction in the incidence of treatment failure in HFNC group compared with SOT group (RR 0.50, 95% CI 0.40 to 0.62, p<0.01) with insignificant heterogeneity (I2=0%). The pooled data suggested there was a significant increase in the incidence of treatment failure in HFNC group compared with nCPAP group (RR 1.61, 95% CI 1.06 to 2.42, p=0.02) with insignificant heterogeneity (I2=0%) (figure 6).

    Figure 6
    Figure 6

    Comparison of high-flow nasal cannula (HFNC) vs control outcome: treatment failure. nCPAP, nasal continuous positive airway pressure;  SOT, standard oxygen therapy.

    Incidence of intubation

    Five included studies14 20 23 24 26 reported the incidence of intubation of patients. The pooled data suggested there was no significant difference in the incidence of intubation between HFNC and SOT groups (RR 1.98, 95% CI 0.6 to 6.56, p=0.26). The pooled data suggested there was no significant difference in the incidence of intubation between HFNC and nCPAP groups (RR 0.96, 95% CI 0.35 to 2.61, p=0.93) with insignificant heterogeneity (I2=0%) (figure 7).

    Figure 7
    Figure 7

    Comparison of high-flow nasal cannula (HFNC) vs control outcome: intubation. nCPAP, nasal continuous positive airway pressure; SOT, standard oxygen therapy.

    Length of non-invasive ventilation

    Three included studies21 23 24 reported the length of non-invasive ventilation of patients. The pooled data suggested there was no significant difference in the length of non-invasive ventilation between HFNC and nCPAP groups (MD 0.08, 95% CI −10.86 to 11.03, p=0.99) (hour) with significant heterogeneity (I2=58%) (figure 8).

    Figure 8
    Figure 8

    Comparison of high-flow nasal cannula (HFNC) vs control outcome: length of non-invasive ventilation. nCPAP, nasal continuous positive airway pressure.

    PICU LOS

    Three included studies14 23 24 reported the length in PICU of patients. One study14 suggested no significant difference in PICU LOS was found between HFNC and SOT groups (p=0.79). The pooled data of other studies suggested there was no significant difference in PICU LOS between HFNC and nCPAP groups (MD −0.15, 95% CI −1.27 to 0.98, p=0.80) (day) with insignificant heterogeneity (I2=0%) (figure 9).

    Figure 9
    Figure 9

    Comparison of high-flow nasal cannula (HFNC) vs control outcome: paediatric intensive care unit (PICU) length of stay (LOS). nCPAP, nasal continuous positive airway pressure; SOT, standard oxygen therapy.

    Length of wheezing

    Two included studies22 26 reported the length of wheezing time of patients between HFNC and SOT groups. The pooled data suggested the length of wheezing was reduced significantly in the HFNC group (MD −2.29, 95% CI −2.85 to −1.72, p<0.01) (day) with insignificant heterogeneity across studies (I2=0%). One study26 suggested there was no significant difference in the length of wheezing between HFNC group and nCPAP group (p=0.16) (figure 10).

    Figure 10
    Figure 10

    Comparison of high-flow nasal cannula (HFNC) vs control outcome: length of wheezing. nCPAP, nasal continuous positive airway pressure; SOT, standard oxygen therapy.

    Respiratory rate

    Seven included studies14 15 20–23 25 reported respiratory rate of patients. The pooled data suggested there was no significant difference in respiratory rate between HFNC group and SOT group (MD −1.65, 95% CI −10.60 to 7.31, p=0.72) with significant heterogeneity (I2=96%). There was no significant difference in respiratory rate of between HFNC group and nCPAP group (MD 1.88, 95% CI −3.38 to 7.13, p=0.48) with significant heterogeneity (I2=89%) (figure 11).

    Figure 11
    Figure 11

    Comparison of high-flow nasal cannula (HFNC) vs control outcome: respiratory rate.  nCPAP, nasal continuous positive airway pressure; SOT, standard oxygen therapy.

    Heart rate

    Four included studies14 15 21 24 reported HR of patients. The pooled data suggested there was no significant difference in HR of children participants between HFNC group and SOT group (MD −1.44, 95% CI −3.42 to 0.54, p=0.15) with insignificant heterogeneity (I2=0%). HR of children in the HFNC group was significantly reduced than those in nCPAP group (MD −7.83, 95% CI −13.46 to −2.20, p=0.006) with significant heterogeneity (I2=85%) (figure 12).

    Figure 12
    Figure 12

    Comparison of high-flow nasal cannula (HFNC) vs control outcome: heart rate. nCPAP, nasal continuous positive airway pressure; SOT, standard oxygen therapy.

    PaCO2

    Three included studies21 24 26 reported PaCO2 of patients. One study26 suggested significant reduction in PaCO2 was found in HFNC group compared with SOT group (p=0.04). The pooled data suggested PaCO2 was not reduced significantly in the HFNC group compared with those in nCPAP group (MD −0.17, 95% CI −4.91 to 4.56, p=0.94) (mm Hg) with significant heterogeneity (I2=84%) (figure 13).

    Figure 13
    Figure 13

    Comparison of high-flow nasal cannula (HFNC) vs control outcome: PaCO2. nCPAP, nasal continuous positive airway pressure; SOT, standard oxygen therapy.

    PaO2

    Three included studies21 24 26 reported PaO2 of patients. One study26 suggested significant increase in PaCO2 was found in HFNC group compared with SOT group (p =0.005). The pooled data suggested PaO2 was not increased significantly in HFNC group compared with those in nCPAP group (MD 1.22, 95% CI −5.55 to 7.99, p=0.72) (mm Hg) with significant heterogeneity (I2=81%) (figure 14).

    Figure 14
    Figure 14

    Comparison of high-flow nasal cannula (HFNC) vs control outcome: PaO2. nCPAP, nasal continuous positive airway pressure; SOT, standard oxygen therapy.

    SpO2

    Four included studies14 15 21 24 reported SpO2 of patients. The pooled data suggested there was no significant difference in SpO2 between HFNC group and SOT group (MD 4.16, 95% CI −4.83 to 13.15, p=0.36) with significant heterogeneity (I2=93%). The pooled data suggested there was no significant difference in SpO2 in HFNC group compared with nCPAP group (MD 0.89, 95% CI −0.08 to 1.86, p=0.07) with insignificant heterogeneity (I2=56%) (figure 15).

    Figure 15
    Figure 15

    Comparison of high-flow nasal cannula (HFNC) vs control outcome: SpO2. nCPAP, nasal continuous positive airway pressure; SOT, standard oxygen therapy.

    Adverse events

    Seven included studies14 15 20 21 23–25 reported the incidence of adverse events of patients. The pooled data suggested there was no significant difference in adverse events between HFNC group and SOT group (RR 0.99, 95% CI 0.32 to 3.07, p=0.99) with insignificant heterogeneity (I2=0%). Two studies21 23 showed there was a significant reduction of adverse events in HFNC group compared with nCPAP group (RR 0.36, 95% CI 0.17 to 0.74, p=0.005) with insignificant heterogeneity (I2=0%). One study24 showed there was no significant difference between HFNC group and nCPAP group, but we could not pool the data due to no detailed data. Another study20 showed that no adverse events happened in both HFNC and SOT+HSS groups (figure 16).

    Figure 16
    Figure 16

    Comparison of high-flow nasal cannula (HFNC) vs control outcome: adverse events. HSS, hypertonic saline solution; nCPAP, nasal continuous positive airway pressure; SOT, standard oxygen therapy.

    Subgroup analyses

    We conducted subgroup analyses to investigate whether effects of HFNC on LOS differed according to different countries, age, sample size, different flow of HFNC and whether mean LOS≥4 days. LOS was significantly decreased in HFNC group compared with SOT group 1) in study located in low-income and middle-income countries (MD −2.32, 95% CI −3.12 to −1.52, p<0.001) (day), 2) in study that only included children under 6 months (MD −3, 95% CI −4.33 to −1.67, p<0.001) (day), 3) in study with a small sample size (MD −2.32, 95% CI −3.12 to −1.52, p<0.001) (day), 4) when LOS≥4 days (MD −2.32, 95% CI −3.12 to −1.52, p<0.001) (day) with insignificant heterogeneity (I2=24%). 5) The subgroup analysis could not be performed based on flow of HFNC due to the insufficient data and the great difference of flow in studies. We did not conduct subgroup analyses for LOS in HFNC versus nCPAP group and LOO in HFNC versus SOT group or HFNC versus nCPAP group because of the limited number of studies (table 2).

    View this table:
    Table 2

    Subgroup analyses of effects of HFNC on LOS

    Sensitivity analysis

    Sensitivity analyses were conducted for the primary outcomes by removing studies with high or unclear risk of selection, performance or attrition bias. Only one study14 had low risk of bias across these parameters. Removing studies with high or unclear risk of bias did not change the results of LOS and LOO (figure 17).

    Figure 17
    Figure 17

    Sensitivity analyses of (A) length of stay (LOS) and (B) length of oxygen (LOO) between high-flow nasal cannula (HFNC) and standard oxygen therapy (SOT) groups.

    Publication bias

    The results of the Egger’s test (p=0.11) and Begger’s test (p=1.0) suggested that there was no statistically significant publication bias of LOS of children in analysis.

    Discussion

    We conducted a systematic review of RCTs to evaluate the effects of HFNC for children with bronchiolitis. To our knowledge, it is the first meta-analysis comparing the effects and safety of HFNC in children with bronchiolitis with other therapies including SOT, nCPAP and HSS.

    Although results of previous observational studies27 28 have proved that HFNC could shorten LOS for children with bronchiolitis compared with SOT, the pooled data of RCTs indicated that HFNC did not significantly shorten LOS and LOO. Significant heterogeneity existed when analysing LOS and LOO, so the subgroup analysis was performed for LOS to explore the potential factors influencing the result of LOS. From subgroup analyses, the study conducted in low-income and middle-income countries (eg, China) showed that HFNC could significantly shorten LOS with insignificant heterogeneity. The data of two studies14 15 which were conducted in high-income countries indicated HFNC could not effectively shorten LOS. There was significant difference between HFNC and SOT group on LOS in the subgroup with longer LOS. The levels of medical care may result in the inconsistent results.29–31 Such inconsistent results may also be related to small sample sizes and the quality of Chinese studies incorporated. The pooled data indicated that HFNC may effectively reduce the incidence of treatment failure compared with SOT, but the incidence of treatment failure was increased compared with nCPAP. Furthermore, HFNC did not effectively reduce the incidence of transfer to ICU and intubation of children with bronchiolitis compared with SOT or SOT+HSS. Although discontinuation of current therapy and/or escalation of care due to the deterioration of disease or adverse events was defined as treatment failure, definitions of treatment failure varied among studies. Clinicians stopped the planed therapy and escalated care in 34% of the infants who did not meet prespecified clinical criteria in one study.14 It was reported in another study23 that discomfort was the leading cause of treatment failure in nCPAP group, but worsening of respiratory distress signs was more frequent in the HFNC group. So it is difficult to conclude a real effect of HFNC on treatment failure by pooling data with obvious heterogeneity. The data analysis suggested that HFNC could not significantly improve PaCO2, PaO2 and SpO2 in children with bronchiolitis.

    Several observational studies32–34 suggested that nCPAP could benefit children with bronchiolitis such as reducing invasive care, PICU LOS, hospital LOS and economic burden in children with bronchiolitis. CPAP has been treated as the standardised therapy for infants with moderate-to-severe bronchiolitis in Denmark.35 Therefore, we collected evidence from RCTs to compare effects of HFNC with nCPAP on children with bronchiolitis. No statistically significant differences of LOS, LOO, the incidence of intubation, length of non-invasive ventilation, PICU LOS, PaO2 and PaCO2 were obtained in both arms, which is consistent with results of a retrospective study.36 However, the incidence of adverse events in children with bronchiolitis was lower in HFNC group.

    Most adverse events were not serious in HFCN group, only one child was reported to be suffering from pneumothorax in one study.14 Pooled data showed that there was no increase in adverse events in children treated by HFNC compared with SOT and nCPAP. Two studies23 24 showed skin sores were less frequent in HFNC group than in nCPAP group. No child died in HFNC group.

    There were several strengths in this review. It is the first meta-analysis evaluating effects and safety of HFNC on children with bronchiolitis compared with standard therapy, nCPAP. The value of HFNC was better evaluated in clinical practice. Besides, all included studies are RCTs, thus the high quality of included studies strengthens the reliability of this meta-analysis.

    Several potential limitations of the analysis need to be noted. The characteristics of the intervention impeded researchers to use appropriate blinding. The sample size of some included studies was small. The flow rate of the HFNC ranged widely thus it was difficult to make a subgroup analysis to explore the most appropriate flow rate to treat children with bronchiolitis. Significant heterogeneity of some outcomes made it hard to conclude the effects of HFNC on bronchiolitis.

    Conclusion

    The systematic review and meta-analysis suggested that HFNC is safe as an initial respiratory management for bronchiolitis but the evidence is still lacking to show significant benefit for bronchiolitis compared with SOT and nCPAP. However, HFNC may decrease the rate of treatment failure for children with bronchiolitis compared with conventional oxygen supplementation. More well-designed RCTs with larger sample sizes need to be conducted to evaluate the effects of HFNC in children with bronchiolitis in the future, especially in China.

    References

    1. ↵
      2. Ralston SL ,
      3. Lieberthal AS ,
      4. Meissner HC , et al
      . Clinical practice guideline: the diagnosis, management, and prevention of bronchiolitis. Pediatrics 2014;134:e1474–e1502.doi:10.1542/peds.2014-2742
      OpenUrlAbstract/FREE Full Text
    2. ↵
      2. Green CA ,
      3. Yeates D ,
      4. Goldacre A , et al
      . Admission to hospital for bronchiolitis in England: trends over five decades, geographical variation and association with perinatal characteristics and subsequent asthma. Arch Dis Child 2016;101:140–6.doi:10.1136/archdischild-2015-308723
      OpenUrlAbstract/FREE Full Text
    3. ↵
      2. Schlapbach LJ ,
      3. Straney L ,
      4. Gelbart B , et al
      . Burden of disease and change in practice in critically ill infants with bronchiolitis. Eur Respir J 2017;49:1601648.doi:10.1183/13993003.01648-2016
      OpenUrlAbstract/FREE Full Text
    4. ↵
      2. Lee JH ,
      3. Rehder KJ ,
      4. Williford L , et al
      . Use of high flow nasal cannula in critically ill infants, children, and adults: a critical review of the literature. Intensive Care Med 2013;39:247–57.doi:10.1007/s00134-012-2743-5
      OpenUrlCrossRefPubMedWeb of Science
    5. ↵
      2. Arora B ,
      3. Mahajan P ,
      4. Zidan MA , et al
      . Nasopharyngeal airway pressures in bronchiolitis patients treated with high-flow nasal cannula oxygen therapy. Pediatr Emerg Care 2012;28:1179–84.doi:10.1097/PEC.0b013e318271a671
      OpenUrlCrossRefPubMedWeb of Science
    6. ↵
      2. Urbano J ,
      3. del Castillo J ,
      4. López-Herce J , et al
      . High-flow oxygen therapy: pressure analysis in a pediatric airway model. Respir Care 2012;57:721–6.doi:10.4187/respcare.01386
      OpenUrlAbstract/FREE Full Text
    7. ↵
      2. Schibler A ,
      3. Pham TM ,
      4. Dunster KR , et al
      . Reduced intubation rates for infants after introduction of high-flow nasal prong oxygen delivery. Intensive Care Med 2011;37:847–52.doi:10.1007/s00134-011-2177-5
      OpenUrlCrossRefPubMedWeb of Science
    8. ↵
      2. González Martínez F ,
      3. González Sánchez MI ,
      4. Rodríguez Fernández R
      . [Clinical impact of introducing ventilation with high flow oxygen in the treatment of bronchiolitis in a paediatric ward]. An Pediatr 2013;78:210–5.doi:10.1016/j.anpedi.2012.11.024
      OpenUrl
    9. ↵
      2. Davison M ,
      3. Watson M ,
      4. Wockner L , et al
      . Paediatric high-flow nasal cannula therapy in children with bronchiolitis: A retrospective safety and efficacy study in a non-tertiary environment. Emerg Med Australas 2017;29:198–203.doi:10.1111/1742-6723.12741
      OpenUrlCrossRef
    10. ↵
      2. Goh CT ,
      3. Kirby LJ ,
      4. Schell DN , et al
      . Humidified high-flow nasal cannula oxygen in bronchiolitis reduces need for invasive ventilation but not intensive care admission. J Paediatr Child Health 2017;53:897–902.doi:10.1111/jpc.13564
      OpenUrl
    11. ↵
      2. Beggs S ,
      3. Wong ZH ,
      4. Kaul S , et al
      . High-flow nasal cannula therapy for infants with bronchiolitis. Cochrane Database Syst Rev 2014;1:Cd009609.
      OpenUrl
    12. ↵
      2. Hilliard TN ,
      3. Archer N ,
      4. Laura H , et al
      . Pilot study of vapotherm oxygen delivery in moderately severe bronchiolitis. Arch Dis Child 2012;97:182.2–3.doi:10.1136/archdischild-2011-301151
      OpenUrl
    13. ↵
      2. Sinha IP ,
      3. McBride AKS ,
      4. Smith R , et al
      . CPAP and high-flow nasal cannula oxygen in bronchiolitis. Chest 2015;148:810–23.doi:10.1378/chest.14-1589
      OpenUrl
    14. ↵
      2. Franklin D ,
      3. Babl FE ,
      4. Schlapbach LJ , et al
      . A randomized trial of high-flow oxygen therapy in infants with bronchiolitis. N Engl J Med 2018;378:1121–31.doi:10.1056/NEJMoa1714855
      OpenUrlCrossRefPubMed
    15. ↵
      2. Kepreotes E ,
      3. Whitehead B ,
      4. Attia J , et al
      . High-flow warm humidified oxygen versus standard low-flow nasal cannula oxygen for moderate bronchiolitis (HFWHO RCT): an open, phase 4, randomised controlled trial. Lancet 2017;389:930–9.doi:10.1016/S0140-6736(17)30061-2
      OpenUrlCrossRefPubMed
    16. ↵
      2. Moher D ,
      3. Liberati A ,
      4. Tetzlaff J , et al
      . Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. BMJ 2009;339:b2535.doi:10.1136/bmj.b2535
      OpenUrlFREE Full Text
    17. ↵
      2. Higgins JP ,
      3. Thompson SG ,
      4. Deeks JJ , et al
      . Measuring inconsistency in meta-analyses. BMJ 2003;327:557–60.doi:10.1136/bmj.327.7414.557
      OpenUrlFREE Full Text
    18. ↵
      2. Begg CB ,
      3. Mazumdar M
      . Operating characteristics of a rank correlation test for publication bias. Biometrics 1994;50:1088–101.doi:10.2307/2533446
      OpenUrlCrossRefPubMedWeb of Science
    19. ↵
      2. Egger M ,
      3. Davey Smith G ,
      4. Schneider M , et al
      . Bias in meta-analysis detected by a simple, graphical test. BMJ 1997;315:629–34.doi:10.1136/bmj.315.7109.629
      OpenUrlAbstract/FREE Full Text
    20. ↵
      2. Bueno Campaña M ,
      3. Olivares Ortiz J ,
      4. Notario Muñoz C , et al
      . High flow therapy versus hypertonic saline in bronchiolitis: randomised controlled trial. Arch Dis Child 2014;99:511–5.doi:10.1136/archdischild-2013-305443
      OpenUrlAbstract/FREE Full Text
    21. ↵
      2. Ji C ,
      3. Song A ,
      4. Zheng N , et al
      . Clinical application of heated,humidified high-flownasal cannula in the treatment of moderate and severebronchiolitis in infants(in Chinese). Chinese Journal of Applied Clinical Pediatrics 2017:1412–5.
    22. ↵
      2. Liu Z
      . Clinical observation on the treatment of infant bronchiolitis bywarming and humidifying high flow oxygen inhalation(in Chinese). China Medicine and Pharmacy 2017;7:71–3.
      OpenUrl
    23. ↵
      2. Milési C ,
      3. Essouri S ,
      4. Pouyau R , et al
      . High flow nasal cannula (HFNC) versus nasal continuous positive airway pressure (nCPAP) for the initial respiratory management of acute viral bronchiolitis in young infants: a multicenter randomized controlled trial (TRAMONTANE study). Intensive Care Med 2017;43:209–16.doi:10.1007/s00134-016-4617-8
      OpenUrlCrossRef
    24. ↵
      2. Sarkar M ,
      3. Sinha R ,
      4. Roychowdhoury S , et al
      . Comparative study between noninvasive continuous positive airway pressure and hot humidified high-flow nasal cannulae as a mode of respiratory support in infants with acute bronchiolitis in pediatric intensive care unit of a tertiary care hospital. Indian J Crit Care Med 2018;22:85–90.doi:10.4103/ijccm.IJCCM_274_17
      OpenUrl
    25. ↵
      2. Yan Z ,
      3. Gai J
      . Effects ofheated, humidified high-flownasal cannula intreatment of severe bronchiolitis in children: a randomized controlled study of 23 cases(in Chinese). Journal of Practical Medical Techniques 2016:1004–5.
    26. ↵
      2. Yang X ,
      3. Cui L ,
      4. Mi Q , et al
      . Clinical research of humidified high flow nasal cannula for bronchiolitis in children(in Chinese). Chinese Pediatric Emergency Medicine 2017:430–3.
    27. ↵
      2. Milani GP ,
      3. Plebani AM ,
      4. Arturi E , et al
      . Using a high-flow nasal cannula provided superior results to low-flow oxygen delivery in moderate to severe bronchiolitis. Acta Paediatr 2016;105:e368–e372.doi:10.1111/apa.13444
      OpenUrl
    28. ↵
      2. Mayfield S ,
      3. Bogossian F ,
      4. O’Malley L , et al
      . High-flow nasal cannula oxygen therapy for infants with bronchiolitis: pilot study. J Paediatr Child Health 2014;50:373–8.doi:10.1111/jpc.12509
      OpenUrlCrossRefPubMed
    29. ↵
      2. Christakis DA ,
      3. Cowan CA ,
      4. Garrison MM , et al
      . Variation in inpatient diagnostic testing and management of bronchiolitis. Pediatrics 2005;115:878–84.doi:10.1542/peds.2004-1299
      OpenUrlAbstract/FREE Full Text
    30. ↵
      2. Duke T
      . Pneumonia and bronchiolitis in developing countries. Arch Dis Child 2014;99:892–3.doi:10.1136/archdischild-2014-306838
      OpenUrlFREE Full Text
    31. ↵
      2. Hussey P ,
      3. Anderson G ,
      4. Berthelot JM , et al
      . Trends in socioeconomic disparities in health care quality in four countries. Int J Qual Health Care 2008;20:53–61.doi:10.1093/intqhc/mzm055
      OpenUrlCrossRefPubMedWeb of Science
    32. ↵
      2. Essouri S ,
      3. Laurent M ,
      4. Chevret L , et al
      . Improved clinical and economic outcomes in severe bronchiolitis with pre-emptive nCPAP ventilatory strategy. Intensive Care Med 2014;40:84–91.doi:10.1007/s00134-013-3129-z
      OpenUrlCrossRefPubMedWeb of Science
    33. ↵
      2. Campion A ,
      3. Huvenne H ,
      4. Leteurtre S , et al
      . [Non-invasive ventilation in infants with severe infection presumably due to respiratory syncytial virus: feasibility and failure criteria]. Arch Pediatr 2006;13:1404–9.doi:10.1016/j.arcped.2006.08.003
      OpenUrlCrossRefPubMed
    34. ↵
      2. Kallappa C ,
      3. Ninan T
      . Role of nasal continuous positive airway pressure (NCPAP) in bronchiolitis - experience from a large district general hospital in United Kingdom. Acta Paediatrica 2008;97:46.
      OpenUrlPubMed
    35. ↵
      2. Pedersen MB ,
      3. Vahlkvist S
      . Comparison of cpap and hfnc in management of bronchiolitis in infants and young children. Children 2017;4:E28.doi:10.3390/children4040028
    36. ↵
      2. Metge P ,
      3. Grimaldi C ,
      4. Hassid S , et al
      . Comparison of a high-flow humidified nasal cannula to nasal continuous positive airway pressure in children with acute bronchiolitis: experience in a pediatric intensive care unit. Eur J Pediatr 2014;173:953–8.doi:10.1007/s00431-014-2275-9
      OpenUrlCrossRefPubMed

    Footnotes

    • Contributors JL is responsible for conception, data search, inclusion and exclusion of studies, data extraction, assessment of methodological quality, data analysis and writing the manuscript. YZ and CG is responsible for inclusion and exclusion of studies, data extraction and data analysis. LX and SL is responsible for data search, inclusion and exclusion of studies, assessment of methodological quality and checking the data of the outcomes. JD is responsible for supervision, interpretation of results and checking the first and the final versions of the manuscript.

    • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

    • Competing interests None declared.

    • Patient consent Not required.

    • Provenance and peer review Not commissioned; externally peer reviewed.