The Cochrane Library and the treatment of croup in children: an overview of reviews




Croup is an illness characterized by barky cough, stridor, hoarseness, and occasionally respiratory distress. It is a relatively mild and short-lived illness for the majority of children. However, a small proportion of children have moderate to severe symptoms and are at risk of hospital admission, and in the most severe cases, intubation.


To synthesize the evidence currently in The Cochrane Database of Systematic Reviews (CDSR) related to the clinical effectiveness and applicability of four treatments for croup: glucocorticoids, epinephrine, heliox and humidified air.


The CDSR was searched for all systematic reviews containing the term ‘croup’ in the title. Pharmacologic and nonpharmacologic interventions for the treatment of croup in children were included. Data were extracted, compiled into tables and synthesized using qualitative and quantitative methods.

Main Results

Four reviews examining the treatment of croup were included in this overview. Children treated with epinephrine vs placebo had a significantly lower croup score at 30 min (SMD: − 0.94; 95% CI: − 1.37, − 0.51), and children treated with glucocorticoids vs placebo had significantly lower croup scores at both 6 h and 12 h (SMD: − 0.59; 95% CI: − 0.83, − 0.35 and SMD: − 0.65; 95% CI: − 1.03, − 0.27). Glucocorticoids vs placebo also significantly reduced inpatient length of stay (MD: − 10.33; 95% CI: − 17.33, − 3.36) and risk of return visits and/or re-admissions (RR: 0.49; 95% CI: 0.34, 0.71). Neither humidified air nor heliox were associated with significant reductions in croup score; however, there is a paucity of evidence examining treatment with heliox.

Author's Conclusions

Glucocorticoids are effective for children with mild croup, and both nebulized epinephrine and glucocorticoids are effective treatments for children presenting with moderate to severe croup that is associated with respiratory distress. There is insufficient evidence to determine the effectiveness of heliox, and there is sufficient evidence to suggest that humidified air is not effective in the treatment of croup. Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. The Cochrane Collaboration

Editors Note:Overviews of reviews, compiling evidence from multiple Cochrane reviews into one accessible and usable document, is a regular feature of this journal. Our aim for each overview is to focus on the treatment question, ‘which treatment should I use for this condition?’, and to highlight the Cochrane reviews and their results in doing so. It is our hope that the overview will serve as a ‘friendly front end’ to the Cochrane Library, allowing the reader a quick overview (and an exhaustive list) of Cochrane reviews relevant to the clinical decision at hand.


Description of the condition

Croup (laryngotracheobronchitis) is a common pediatric respiratory illness that affects approximately 3% of children under six years of age per year1, 2. Croup can affect infants as young as three months of age, but its incidence peaks at approximately 5% in the second year of life3. Croup can also occur more rarely in adolescents and adults, and there is a male to female preponderance of 1.4 to 14, 5. The most common etiology for croup is a viral infection, predominantly parainfluenza virus, which causes edema of the upper airway and varying degrees of upper airway obstruction6–8.

Croup is a clinical diagnosis. The illness presents with the sudden onset of a distinctive seal-like or barky cough, which may be accompanied by hoarse voice, stridor, and/or respiratory distress. Symptom onset is abrupt and often begins during the night. Non-specific symptoms often precede croup, including cough, rhinorrhea and fever. Croup symptoms are most often short-lived, with the majority of children showing spontaneous resolution of the barky cough within 48 h; however, a minority of children may continue to have symptoms for up to one week9.

Most children with croup who present for medical care may be classified as having mild croup, defined on clinical assessment as having barky cough with no stridor or significant chest-wall indrawing at rest. A smaller proportion of children have moderate croup, which is clinically defined as frequent barky cough associated with stridor and chest-wall indrawing at rest, but without other signs of distress or agitation. A very small minority of children present with symptoms of severe croup: frequent barky cough, prominent stridor, and marked indrawing of the chest-wall or sternum that is associated with agitation, lethargy, or cyanosis2, 9.

Most children with croup can be managed at home or in the outpatient setting. In population-based studies, fewer than 5% of children with croup are admitted to hospital after medical assessment2, 10. Of those who are hospitalized, 1–3% are subsequently intubated11–14. Mortality is very rare, and based upon extrapolation from various data sources, is estimated to occur in approximately 1 in 30,000 cases2, 14–20.

Description of the interventions


The symptoms of croup (barky cough, stridor, and respiratory distress) result from narrowing of the subglottic region of the upper airway. The narrowing is caused by localized inflammation and edema secondary to viral infection. Given the small diameter of children's airways, even minor swelling can increase work of breathing; this is because airway resistance increases exponentially secondary to reduction in airway diameter. Glucocorticoids are postulated to reduce upper airway edema via their anti-inflammatory properties. The most commonly administered glucocorticoid is dexamethasone, with doses ranging between 0.15 and 0.6 mg per kg21. In terms of route of administration, published trials have studied oral dexamethasone and prednisone/prednisolone, parenteral dexamethasone and methylprednisone and inhaled budesonide21.


Nebulized epinephrine, such as racemic or alternatively L-epinephrine, has been used widely for several decades in the treatment of croup. Racemic epinephrine is composed of equal proportions of D- and L-isomers of epinephrine, and the mechanism of action is postulated to be local vasoconstriction to decrease upper airway edema. The role of epinephrine treatment in the small proportion of children who have severe symptoms is to provide rapid reduction of respiratory distress while waiting for the effects of glucocorticoids to occur. The clinical effect of nebulized epinephrine is sustained for at least 1 h22–28 but is effectively gone within 2 h following administration26. There does not appear to be any ‘rebound’ effect: following epinephrine treatment; croup symptoms, on average, return to baseline and do not appear to worsen following administration26, 29.

Humidified Air (Mist)

Humidified air has been used for the treatment of croup since the 1800's30. The postulated mechanism of action involves cooling the airway mucosa and decreasing edema and secretion viscosity31. However, potential risks exist with humidified air administration: hot, humidified air can cause scald injuries32, moulds and fungi can be dispersed into the air from mist tents if they are not properly cleaned33, and perhaps most importantly, mist tents are cold, wet, and separate the child from the caregiver. This usually increases the child's agitation and worsens their symptoms34.


Helium is a low-density, inert gas that possesses no inherent biological or pharmacological effects. When helium is mixed with oxygen, it is known as heliox. The postulated mechanism of action of heliox relates to the low density of helium gas as compared with the nitrogen in room air. The presence of low-density gas decreases turbulent airflow through the narrowed airway, thus reducing airway resistance and respiratory distress. Heliox is not widely used and is usually limited to specialized settings, such as intensive care units. There can be practical limitations in the use of heliox in severe croup, the most important being the limited fractional inspired oxygen concentration (FiO2 of 0.3 or 30%) for a child with significant hypoxia.

Why this overview is important

Croup is a common respiratory illness of childhood. The incidence of croup peaks in the second year of life, affecting approximately 5% of children, and then gradually declines with increasing age35. Although the majority of children have mild and short-lived symptoms, croup is sometimes more severe and can require hospitalization, intubation, and mechanical ventilation. In rare cases, it leads to death. We aim to provide a concise summary of the evidence for four treatments for croup: glucocorticoids, epinephrine, heliox, and humidified air. For glucocorticoid treatment, the included review had a sub-analysis of children with mild croup, which will be included as a separate analysis in this overview.


The objectives of this overview are to synthesize the evidence currently in The Cochrane Database of Systematic Reviews related to the following questions:

  • (1)In children with croup, does treatment with glucocorticoids, inhaled epinephrine, humidified air or heliox reduce any of the following:
    • Severity of respiratory distress (as measured by clinical croup score or other measures of clinical improvement);

    • Hospital admissions and/or re-admissions;

    • Risk of intubation?

  • (2)In the subset of children with clinically mild croup, does treatment with glucocorticoids reduce return visits and/or re-admissions? This objective was chosen as these represent the majority of children treated for croup.
  • (3)In children with croup, is treatment with glucocorticoids, inhaled epinephrine, humidified air or heliox associated with significant adverse events?


Search strategy

The Cochrane Database of Systematic Reviews was searched for all systematic reviews assessing treatment options for croup. The term ‘croup’ was entered and restricted to the review title. This resulted in three systematic reviews21, 36, 37 and one protocol38. Three of the protocol authors were also authors of this overview and were able to inform us that the manuscript was fully completed and peer-reviewed; therefore the finished manuscript was included in our overview. Furthermore, two overview authors were authors of one review21 that was currently being updated. The updated manuscript was used in place of the original review in order to incorporate the most recent data into this overview. Therefore, this overview draws on data from four Cochrane reviews21, 36–38. All four included reviews identified relevant trials, contained data on results and examined a different intervention. The four interventions examined were glucocorticoids, epinephrine, heliox and humidified air.

Data extraction and synthesis

One reviewer (KR) extracted the following information from the included reviews: inclusion criteria (including population, intervention, comparison(s), and outcomes), methodological quality assessment and numeric results. Numeric data were checked for completeness and accuracy by a second reviewer (MF), and any discrepancies were resolved through discussion.

All dichotomous data were presented as relative risks (RR) with 95% confidence intervals (CIs). The RR describes the probability of the event in the treatment group compared to the probability of the event in the control group and was interpreted as statistically significant if the CI did not cross one. If there were no events for the outcome of interest, a risk difference was calculated instead of relative risk.

Continuous data were summarized using standardized mean differences (SMD) or mean differences (MD), both with 95% CIs. If the outcome was measured using a variety of scales (for example, a variety of croup scores), SMD was calculated. For comparison purposes, the Westley croup score was also summarized using SMD. MD and SMD were interpreted as statistically significant if the 95% CI did not cross zero.

In the event that a pooled effect estimate measure had to be calculated, random effects modelling was conducted in order to provide the most conservative estimate. When there were two or more studies included in a comparison, the accompanying I2 value was reported and represents the degree of statistical heterogeneity between the studies. An I2 value close to 0% indicates minimal or no heterogeneity of trials, whereas an I2 of approximately 50% or greater represents substantial heterogeneity39. All pooled effect estimates mentioned in the Results section have accompanying I2 values reported.


Description of included reviews

The four included reviews contained a total of 4,719 participants and are described in Table 1. All reviews included trials conducted in any setting, and two reviews21, 38 identified studies published in languages other than English. Since croup is a condition that occurs predominantly in children, all of the included reviews contained only pediatric data. However, no reviews specified age ranges.

Table I. Description of included reviews
TitleAuthorDate of most substantive updateNumber of studiesStudy sample size (Range)PopulationDefinition of croupInterventionControlOutcomes
Heliox for croup in childrenVorwerk C, Coats TJune 2009244 (15–29)ChildrenDiagnosis of croup or laryngotracheobronchitis. Children with other conditions of upper airway obstruction were excludedHelioxPlacebo or any active treatmentPrimary: change in croup score
         Secondary: change in respiratory rate, change in oxygen requirements, change in heart rate, hospital admission rate, need for intubation and adverse events
Humidified air inhalation for treating croupMoore M, Little PMay 20063145 (16–71)ChildrenClinical diagnosis of croupWarm or cool humidified air delivered by steam or humidified tentNo treatmentMortality, admission to hospital, ventilation/intensive care treatment, symptom severity or symptom scores, duration of symptoms or inpatient treatment, number of ED/outpatient visits for croup in the week following treatment and time lost from school
Glucocorticoids for croupRussell K, Liang Y, O'Gorman K, Johnson DW, Klassen TPMay 2010414299 (17–720)ChildrenCroup (hoarseness, barking cough and stridor), pseudo croup, or laryngotracheitis. Children with bronchiolitis and acute stridor were excludedGlucocorticoid (oral, parenteral or nebulized)Placebo or any active treatmentPrimary: change in croup score and return visits and/or (re)admissions
         Secondary: length of stay, patient improvement and use of additional treatment
Epinephrine for croupBjornson C, Russell K, Vandermeer B, Durec T, Klassen TP, Johnson DWApril 20108301 (14–78)ChildrenClinical diagnosis of croup (acute onset of a barky cough and stridor)Nebulized racemic or L-epinephrine, delivered with or without IPPBPlacebo, L-epinephrine or nebulized epinephrine without IPPBPrimary: change in croup score
         Secondary: intubation, hospitalization and length of stay, return visits due to ongoing croup symptoms, parental anxiety, side effects and myocardial injury or cardiac arrhythmias

The definitions of ‘croup’ were fairly homogeneous: two reviews included children with a clinical diagnosis of croup36, 38, one included children with a clinical diagnosis of croup or laryngotracheobronchitis and excluded other causes of upper airway obstruction37, and the last review used a range of alternative terms commonly used to describe croup, including croup, pseudo-croup and laryngotracheitis21. Only one review conducted a subgroup analysis based on croup severity and separately analyzed studies on children with mild croup21.

Two interventions were pharmacological21, 38 while the other two were inert or non-pharmacological36, 37. With the exception of humidified air, the remaining reviews compared the intervention to placebo or active treatment. Specifically, the following interventions and comparisons were included:

  • Glucocorticoids vs placebo or active treatment, administered via any route.

  • Racemic or L-epinephrine (with or without intermittent positive pressure breathing (IPPB) administration) vs placebo, or comparisons of different types of epinephrine and/or routes of administration.

  • Heliox vs placebo or active treatment, with similar routes of administration.

  • Warm or cool humidified air vs no treatment.

When reported, the primary outcomes were change in croup score21, 37, 38 and return visits and/or (re)admissions21. The most common outcomes were change in croup score, return visits, length of stay, and improvement (where improvement was defined in a variety of ways, including a minimal increment improvement in a croup score, overall relief of croup symptoms, and by dichotomization of croup score). To date, the majority of the research has examined the role of glucocorticoids for the treatment of croup.

Methodological quality of the studies included in the reviews

Two of the reviews used the Risk of Bias tool to assess methodological quality37, 38. All trials assessing heliox had adequate sequence generation, compared to only half (4/8) of the epinephrine trials. While all trials in both reviews were appropriately blinded, only half of the trials in both reviews reported complete outcome data. The majority of the studies (6/8) assessing epinephrine treatment were deemed to be low risk of bias.

One review on glucocorticoids21 assessed methodological quality using the five-point Jadad Scale that examines randomization, double-blinding, and reporting of withdrawals. The median Jadad score was 3 (inter-quartile range 2, 4). Allocation concealment was also assessed, and almost half (19/41) reported adequate allocation concealment.

The review examining humidified air36 assessed allocation concealment and used an unspecified methodological quality assessment tool. All three included trials reported baseline comparisons of the experimental groups and two studies reported explicit diagnostic criteria and complete follow-up. Blinding and allocation concealment were reported as adequate in only one study (1/3).

Effects of interventions

Croup score and improvement

Table 2 shows croup scores for time points ranging from 20 min to 12 h, and Table 3 shows clinical improvement at time points ranging from 6–24 h. Children treated with glucocorticoids vs placebo showed significant reductions in croup score at both 6 h (SMD: − 0.59; 95% CI: − 0.83, − 0.35; I2: 67%) and 12 h (SMD: − 0.65; 95% CI: − 1.03, − 0.27; I2: 74%). At 24 h, however, these reductions in croup score were no longer significant. In addition, there were significantly more children who improved at 6 and 12 h when receiving glucocorticoid treatment (RR: 1.42; 95% CI: 1.14, 1.77; I2: 26% and RR: 1.32; 95% CI: 1.08, 1.61; I2: 52%) vs placebo; however, this difference was also not significant at 24 h.

Table II. Croup scores
ComparisonClinical scaleTime periodNumber of subjects (studies)Measure of effect (95% CI)I2Comparison favoured
  • *

    Statistically significant.

    CI: confidence interval; MD: mean difference; SMD: standardized mean difference.

    Notes: Westley scale is a 17-point scale including assessment of level of consciousness, cyanosis, stridor, air entry, and chest wall retractions. Non-Westley scales include a variety of clinical assessment tools which assess combinations of clinical observations, commonly including combinations of cough, stridor, retractions, breath sounds, and cyanosis.

  • **

    ‘All clinical scales’ include Westley and non-Westley croup scores

  • 1

    Nebulized: 6/9 oral: 1/9, inhaled: 1/9, IM: 1/9.

  • 2

    IM: 3/5, nebulized: 1/5, oral: 1/5.

  • 3

    Nebulized: 7/14, IM: 4/14, oral: 2/14, inhaled: 1/14.

  • 4

    IM: 3/8, nebulized: 3/8, oral: 2/8.

  • 5

    IM: 4/8, nebulized: 2/8, oral: 1/8, inhaled: 1/8.

  • 6

    Nebulized: 6/6.

  • 7

    IM: 4/7, oral: 2/7, nebulized: 1/7.

  • 8

    Inhaled fluticasone: 1/1.

  • 9

    Inhaled budesonide and IM dexamethasone: 1/2, nebulized budesonide and oral dexamethasone: 1/2.

  • 10

    Nebulized budesonide and oral dexamethasone: 3/3.

  • 11

    Nebulized budesonide and oral dexamethasone: 1/1.

  • 12

    IM dexamethasone and oral betamethasone: 1/1.

  • 13

    Both drugs—oral: 1/1.

  • 14

    Both drugs—oral: 1/1.

  • 15

    Both drugs—oral: 1/1.

  • 16

    Both drugs—oral: 2/2.

  • 17

    Gluocorticoids—IM: 2/3, nebulized: 1/3.

Any glucocorticoid vs placeboWestley scale16 h452 (9)SMD: − 0.76 (−0.99, − 0.54)*42%Glucocorticoid
 Non-Westley scales2 579 (5)SMD: − 0.31 (−0.67, 0.04)68% 
 All clinical scales3** 1031 (14)SMD: − 0.59 (−0.83, − 0.35)*67%Glucocorticoid
 All clinical scales4**12 h532 (8)SMD: − 0.65 (−1.03, − 0.27)*74%Glucocorticoid
 All clinical scales5**24 h388 (8)SMD: − 0.55 (−1.10, 0.00)83% 
Budenoside vs placeboAll clinical scales6**6 h369 (6)SMD: − 0.74 (−0.95, 0.52)*55%Budenoside
Dexamethasone vs placeboAll clinical scales7**6 h645 (7)SMD: − 0.49 (−0.86, − 0.11)*0%Dexamethasone
Fluticasone vs placeboWestley scale86 h17 (1)MD: 0.10 (−1.36, 1.56) 
Dexamethasone vs budesonideNon-Westley scales96 h181 (2)SMD: − 0.39 (−1.07, 0.28)74% 
Dexamethasone and budesonide vs dexamethasoneNon-Westley scales106 h254 (3)SMD: − 0.08 (−0.64, 0.48)78% 
Dexamethasone and budesonide vs budesonideNon-Westley scales116 h129 (1)MD: − 0.20 (−0.55, 0.15) 
Dexamethasone vs betamethasoneWestley scale126 h52 (1)MD: 1.51 (0.61, 2.41)*Betamethasone
Dexamethasone vs prednisoloneWestley scale136 h99 (1)MD: 0.19 (−0.17, 0.55) 
Dexamethasone (0.30 mg/kg) vs dexamethasone (0.15 mg/kg)Non-Westley scale146 h60 (1)SMD: 0.17 (−0.34, 0.67) 
Dexamethasone (0.60 mg/kg) vs dexamethasone (0.30 mg/kg)Non Westley scale156 h60 (1)SMD: 0.21 (−0.30, 0.72) 
Dexamethasone (0.60 mg/kg) vs dexamethasone (0.15 mg/kg)Westley scale166 h129 (2)SMD: − 0.02 (−0.37, 0.32)0% 
Glucocorticoid vs epinephrineNon-Westley scales176 h129 (3)SMD: 0.18 (−0.17, 0.52)0% 
Epinephrine vs placeboAll clinical scales**30 min94 (3)SMD: − 0.94 (−1.37, − 0.51)*0%Epinephrine
Racemic epinephrine vs L-epinephrineNon-Westley scales30 min28 (1)MD: 0.56 (−0.66, 1.78) 
  2 h28 (1)MD: 1.55 (0.27, 2.83)*L-epinephrine
Epinephrine with IPPB vs epinephrine without IPPBNon-Westley scales30 min14 (1)MD: − 0.30 (−2.03, 1.43) 
  2 h14 (1)MD: − 1.40 (−3.04, 0.24) 
Humidified air vs no treatmentWestley scale20-60 min135 (3)SMD: − 0.14 (−0.75, 0.47)63% 
  2 h41 (2)SMD: − 0.11 (−0.73, 0.51)0% 
Heliox vs 30% oxygenWestley scale20 min15 (1)MD: 0.83 (−0.90, 2.56) 
Table III. Clinical improvement
ComparisonOutcomeNumber of subjects (studies)Measure of effect (95% CI)I2Comparison favoured
  • *

    Statistically significant.

  • CI: confidence interval; RR: risk ratio.

Glucocorticoid vs placeboClinical improvement at 6 h421 (6)RR: 1.42 (1.14, 1.77) *26%Glucocorticoid
 Clinical improvement at 12 h340 (6)RR: 1.32 (1.08, 1.61) *52%Glucocorticoid
 Clinical improvement at 24 h205 (4)RR: 1.24 (1.06, 1.45) *31%Glucocorticoid
Epinephrine vs placeboClinical improvement: unspecified time74 (2)RR: 1.46 (0.82, 2.60)17% 

Dexamethasone and budesonide were the most commonly studied glucocorticoids. There appeared to be no advantage of one over the other, nor was treatment with both superior to treatment with either alone. Furthermore, there was no significant difference in outcomes between children receiving 0.15 mg/kg, 0.30 mg/kg, or 0.60 mg/kg of dexamethasone, but this conclusion is based on only minimal data (three studies, 218 children). In comparisons of glucocorticoids with epinephrine, there were no significant differences in croup scores at 6, 12, or 24 h.

Treatment with epinephrine vs placebo significantly reduced croup score at 30 min (SMD: − 0.94; 95% CI: − 1.37, − 0.51; 0%); however, this treatment effect was no longer evident at 2 and 6 h post-treatment, and there was no significant difference in clinical improvement. Children randomized to L-epinephrine vs nebulized racemic epinephrine showed no significant difference in croup score at 30 min, but those receiving L-epinephrine had a significantly greater reduction in croup score at 2 h (MD: 1.55; 95% CI: 0.27, 2.83). There was no significant difference in croup score among children who received epinephrine with or without IPPB.

Children with an Emergency Department diagnosis of mild to moderate croup treated with heliox or 30% humidified oxygen showed no significant difference in 20 min or ‘post-treatment’ croup scores. Croup score for children treated with humidified air vs no treatment did not significantly differ at any of the measured time-points.

Length of stay and return visits and/or re-admissions

Table 4 presents data on length of stay for children with croup who were admitted to the hospital or treated in the Emergency Department, and Table 5 presents data on hospital re-admissions. Length of stay for children receiving glucocorticoids vs placebo was significantly shorter for both children admitted to the hospital (MD: − 10.33; 95% CI: − 17.33, − 3.36; I2: 61%) and those treated in the Emergency Department (MD: − 21.00; 95% CI: − 31.55, − 6.45). Compared to placebo, epinephrine significantly reduced the length of hospital stay for children admitted with croup (MD: − 32.00; 95% CI: − 59.14, − 4.86) but did not affect the length of stay for those treated in the Emergency Department. For children treated with epinephrine, assessments of length of stay were each based on one small trial; only one study reported return visits and noted that no child in the epinephrine or placebo group returned to the hospital.

Table IV. Length of stay (h)
ComparisonOutcomeNumber of subjects (studies)Measure of effect (95% CI)I2Comparison favoured
  • *

    Statistically significant

  • CI: confidence interval; MD: mean difference

Glucocorticoid vs placeboLength of stay—inpatients795 (8)MD: − 10.33 (−17.33, − 3.36) *61%Glucocorticoid
 Length of stay—outpatients30 (1)MD: − 21.00 (−31.55, − 6.45) *Glucocorticoid
Epinephrine vs placeboLength of stay—inpatients37(1)MD: − 32.00 (−59.14, − 4.86)*Epinephrine
 Length of stay—outpatients54 (1)MD: − 1.80 (−4.07, 0.47) 
Table V. Return visits and/or (re)admissions
ComparisonNumber of subjects (studies)Measure of effect (95% CI)I2Comparison favoured
  • *

    Statistically significant

  • CI: confidence interval; RD: risk difference; RR: risk ratio.

Any glucocorticoid vs placebo1679 (10)RR: 0.49 (0.34, 0.71) *49%Glucocorticoid
Dexamethasone vs placebo1453 (8)RR: 0.52 (0.34, 0.79) *51%Dexamethasone
Budenoside vs placebo228 (4)RR: 0.39 (0.17, 0.92) *30%Budenoside
Epinephrine vs placebo54 (1)RD: 0.00 (−0.07, 0.07) 
Dexamethasone vs prednisolone219 (2)RR: 0.32 (0.17, 0.60) *Dexamethasone
Dexamethasone vs budesonide183 (2)RD: 0.01 (−0.02, 0.05)0% 
Budenoside and dexamethasone vs dexamethasone252 (3)RR: 0.87 (0.41, 1.85)0% 
Oral dexamethasone vs intramuscular dexamethasone372 (2)RR: 0.80 (0.58, 1.12)0% 
Dexamethasone (0.60 mg/kg) vs dexamethasone (0.15 mg/kg)130 (2)RR: 1.04 (0.62, 1.75)0% 
Dexamethasone (0.30 mg/kg) vs dexamethasone (0.15 mg/kg)60 (1)RR: 2.81 (0.12, 66.40) 
Dexamethasone (0.60 mg/kg) vs dexamethasone (0.30 mg/kg)60 (1)RR: 1.87 (0.18, 19.55) 
Dexamethasone vs betamethasone52 (1)RR: 0.95 (0.67, 1.34) 
Any glucocorticoid vs epinephrine66 (1)RD: 0.00 (−0.06, 0.06) 

All children randomized to either heliox or 30% humidified oxygen were discharged from the hospital. Return visits and/or re-admissions were not reported in the review concerning humidified air treatment. Also, the risk of return visits and/or re-admissions was significantly less among children treated with glucocorticoids compared to placebo (RR: 0.49; 95% CI: 0.34, 0.71; I2: 49%).

Mild croup

Three studies within the glucocorticoid review40–42 included children with mild croup, and the pooled risk ratio indicated a significant reduction in the risk of return visits and/or re-admissions when treated with glucocorticoids (RR: 0.54 ; 95% CI: 0.30, 0.95; 48%).

Need for Intubation

Compared to placebo, there was no significant difference in the risk of intubation for children treated with glucocorticoids (RD: 0.00, 95% CI: − 0.01, 0.01; note that RR could not be used for this assessment since most of the studies did not report any children as having been intubated). The risk of intubation was not significantly different among children treated with L-epinephrine vs racemic epinephrine and epinephrine with or without IPPB. Lastly, no child in the heliox review required intubation (which was inferred from the authors' statement that no study reported a child requiring intensive care treatment).

Adverse events

Adverse events were not an outcome of interest in the glucocorticoid and humidified air reviews. No study in the epinephrine review reported significant adverse events, and no adverse events were associated with heliox treatment.


Based on the data in the four included Cochrane reviews, treatment with nebulized epinephrine was associated with significant short term improvement (croup score at 30 min post-treatment), and glucocorticoid treatment was effective in reducing croup scores at 6 and 12 h following treatment. While both epinephrine and glucocorticoid treatments were associated with shorter length of stay in hospital, only the finding for glucocorticoid treatments should be considered definitive, as the finding for epinephrine was based on one small trial. Treatment with glucocorticoids was also associated with a significant reduction in return visits for medical care. Published evidence suggests that treatment with humidified air is not effective in children with croup. While existing evidence does not demonstrate a benefit of heliox for the treatment of children with moderate to severe croup, insufficient evidence has been published to determine if heliox is truly effective.

Overall completeness and applicability of evidence

Overall, there were adequate numbers and sample sizes in studies comparing glucocorticoids vs placebo. Studies comparing different glucocorticoids and different dose sizes were of small size, as were studies examining heliox and humidified air. Apart from the review on glucocorticoid treatment, we were unable to analyze data according to severity of croup. Given that a major concern for the clinician is the potential for severe respiratory distress requiring hospitalization or intubation, data on the effectiveness of treatments by subgroups (mild/moderate/severe croup) would be helpful in the clinical setting, and the lack of this data is a limitation of this overview.

Quality of the evidence

Overall, the included trials were of good quality, although minor methodological limitations were identified in each of the included reviews. The glucocorticoid review included trials of good quality, and more than half of the included trials were described as having adequate allocation concealment. In the epinephrine review, most trials were evaluated to be of low risk of bias and all included appropriate blinding; however, only half of included trials reported adequate sequence generation or complete data. The trials in the heliox review had adequate sequence generation and blinding, but there were only two trials with small enrolment. The review on humidified air included only three relatively small trials, two of which had inadequate blinding and allocation concealment.

Croup score and clinical improvement

Treatment with nebulized epinephrine was associated with short-term improvement in croup score (at 30 min post treatment), an effect that did not persist at 2 and 6 h after treatment. In a child with moderate to severe croup, epinephrine is usually administered concurrently with glucocorticoids to reduce respiratory distress while awaiting the effects of the glucocorticoid treatment. We favour the use of L-epinephrine by simple nebulization over racemic epinephrine or IPPB, given the relative ease of administration of epinephrine by simple mask as opposed to IPPB, the wide availability of L-epinephrine as compared with racemic epinephrine, and the lack of evidence of superiority for the latter treatments. Though nebulized epinephrine is widely used in children with moderate to severe croup, it is unknown whether there is additional or additive benefit when it is administered concurrently with glucocorticoids.

Glucocorticoid treatment was associated with significant reduction in croup score at 6 and 12 h following treatment. The glucocorticoid effect was not noted 24 h post dosing; however, this may be because fewer numbers of children were assessed at 24 h (as many children would have been discharged from medical care by then) or because croup improves on its own over a relatively short time frame, thus potentially minimizing the differences between treatment and placebo groups.

A related outcome recorded in some studies was termed ‘improvement’. Improvement was defined in a variety of ways in different studies, but overall there was a significant difference favouring improvement with glucocorticoids as compared to placebo. This is consistent with the improvement seen in croup scores. Although both inhaled budesonide and oral and parenteral dexamethasone were associated with clinical improvement, we favour the use of oral dexamethasone given the ease of administration and lower cost.

Based upon the data analysed within this review, we did not find sufficient evidence to recommend lower doses of dexamethasone in the treatment of croup over the standard dose of 0.6 mg/kg per dose. In the literature, one meta-analysis from 198943 and four more recent randomized trials44–47 show conflicting evidence regarding the optimal dose. The 1989 meta-analysis included six studies of children admitted to hospital, and concluded that higher doses of corticosteroid were associated with better outcome. On the other hand, the four randomized trials found no significant difference in primary outcome measures between different corticosteroid doses, thus suggesting that a dose of 0.15 mg/kg dexamethasone may be adequate. However, all four randomized trials differed in study design, setting (inpatient and outpatient), and severity of croup. Furthermore, none were designed as non-inferiority studies, and all were of small sample size.

Although heliox was not associated with significant improvement in croup score at 20 min post treatment, the studies were small and possibly under-powered to detect a true clinical benefit. However, given the current lack of evidence for its benefit and the reported difficulties in administering it, we do not recommend the routine use of heliox.

Humidified air was not associated with a significant improvement in croup score at 20 min post treatment. Published trials are reasonably large and well-designed and, though not specifically designed to detect non-superiority, are likely adequate to infer that humidified air is not effective. Given the lack of current evidence and potential for difficulties associated with administration of humidified air, we do not recommend humidified air for the treatment of croup.

Return visits and/or re-admissions and length of stay

Treatment with glucocorticoid was associated with both a statistically and clinically significant reduction in return visits and/or re-admission to hospital as compared to placebo. A comparable effect was seen with dexamethasone and budesonide.

Epinephrine treatment was not associated with reduced return visits or readmissions in one small study included within this review. However, given the objective benefit in reduction in croup score in children with moderate or severe croup, we support its use in this population.

Mild croup

The three glucocorticoid trials focused on children with mild croup symptoms showed approximately a 50% reduction in the number of return visits to health care providers. Given that approximately 15% of children who are not treated with glucocorticoids return to care, we think a 50% reduction would be considered important to both parents and clinicians, and consequently recommend that glucocorticoids be administered to children with mild croup symptoms as well as those with moderate to severe symptoms.


Although there was no significant difference found in the risk of intubation among children treated with glucocorticoid vs placebo within the included review, the numbers may have been too small to show a significant difference as the outcome is very rare.

In the literature, there is data to suggest a benefit of glucocorticoid treatment in reducing the rate of intubation and the duration of intubation in croup. A meta-analysis of trials including those with pseudorandomization found a five-fold reduction in intubation rate in children with severe croup and impending respiratory failure43, and a well-designed trial in children with severe croup who were intubated found that treatment with glucocorticoid reduced the length of ventilation by one-third and the need for reintubation by seven-fold48. Finally, a publication of historical interest describes a five-fold reduction in intubation rate in children with croup admitted to a tertiary care hospital following the introduction of routine treatment with glucocorticoid49.

Adverse events

No adverse events were reported in the review of epinephrine treatment. The administration of one dose at a time of nebulized epinephrine to children has not been associated with clinically significant increases in blood pressure or heart rate, nor with any adverse effects25, 27, 29, 47, 50, 51. The most frequent side effects noted in clinical trials have been mild, and have included tachycardia and pallor52. Therefore intermittent use of nebulized epinephrine appears to be safe for children with croup-related respiratory distress. A single case report describes a previously healthy child who developed ventricular tachycardia and myocardial infarction following three doses of nebulized epinephrine administered within a one hour time period53. Therefore for children in severe respiratory distress necessitating frequent doses (e.g. more than one per h) of nebulized epinephrine, we would advocate that the clinicians consult a pediatric intensivist to discuss further treatment and/or transfer to the intensive care setting.

Although the glucocorticoid review did not report on adverse effects, there are no reports—to our knowledge—of potential adverse effects stemming from glucocorticoid treatment of children with croup. Furthermore, we would expect any adverse effects associated from glucocorticoid treatment to be rare, but would consequently acknowledge that current evidence is insufficient to be certain that rare events do not occur.


Implications for research

  • There is a need for corticosteroid dose ranging studies designed to show equivalence.

  • There is a need for studies comparing treatment of moderate and severe croup with a single dose of corticosteroid vs multiple doses.

  • Trials are needed examining the effectiveness of treatment with heliox in the acute setting for children with severe croup.

Implications for practice

  • In children presenting with moderate to severe croup symptoms, nebulized epinephrine treatment is associated with rapid reduction of respiratory distress.

  • Nebulized L-epinephrine treatment administered by simple mask may be preferable to racemic epinephrine or epinephrine administered by IPPB, given the relative ease of use and wider availability of L-epinephrine, and lack of evidence of superiority of the other treatments.

  • In children with moderate to severe croup, glucocorticoid treatment is associated with significant reduction in respiratory distress, reduced length of stay in the emergency department and reduced risk of hospital admission.

  • Dexamethasone may be preferable to other forms of oral or inhaled corticosteroid for the treatment of croup, given the lack of evidence of superiority of one particular glucocorticoid over another, and the relative ease of use and low cost of oral dexamethasone.

  • There is insufficient evidence to support the routine use of heliox in the treatment of croup; however, more studies are needed.

  • There is insufficient evidence to support the routine use of humidified air in the treatment of croup.


The authors would like to acknowledge Denise Thomson for her assistance with this overview. Kelly Russell holds an Alberta Heritage Foundation for Medical Research doctoral student award.

Contributions of Authors

All authors contributed to this overview. CB contributed to writing the Introduction, Discussion and Conclusions sections. KR extracted the data and contributed to writing the Methods and Results sections. MF verified data and contributed to writing the Results section. DWJ contributed to writing the Introduction, Discussion, and Conclusions sections. CB is the primary author of this report. All authors contributed to editing all sections of the overview and take responsibility for the manuscript.

Declarations of Interest

Dr Johnson and Ms Russell are authors on two of the included reviews (Epinephrine for Croup and Glucocorticoids for Croup) and Dr Bjornson is an author of one included review (Epinephrine for Croup). Also Dr Johnson is an author of several included trials and Dr Bjornson is an author of one included trial.