Table 2.

Relative heterogeneity of the included studies.

Author	Patient Group and Age	Study Design	Outcomes	Key Results	Study Weaknesses
[52] Sparrow and Geelhoed, 2006.	Children aged 3 months to 142 months old presenting to a single emergency department with mild to moderate croup.	Randomization and blinding Double blind, randomised equivalence study. Sample size and attrition 133. No attrition Medication dosages Patients received a single dose of 0.15 mg/kg dexamethasone 0.6 dexamethasone or 1 mg/kg prednisolone. Lengths of time of administration Single dose of prednisolone, 1 mg/kg, matched for potency with a single dose of dexamethasone in children with mild to moderate croup.	Primary outcome was the magnitude and rate of reduction in Westley croup score. Clinical observations at 30 minutes after administration of steroid; hourly for the next four hours and four hourly thereafter until discharge. Criteria for discharge home were minimal stridor or chest wall retractions—that is, a croup score of 1 or 0. Re-attendance at medical care within 7-10 days.	Responses to treatment: Five out of 68 (7%) children who had received dexamethasone returned for medical care versus 19/65 (29%) children who had received prednisolone. No significant difference between the three groups in magnitude or rate of Westley score reduction.	Included patients up to the age of 12 years, although very uncommon for older children to be affected. Exclusion criteria included prior administration of steroids, non-English speakers and no access to telephone. Small numbers.
[53] Fifoot and Ting, 2007	Children 6 months to 6 years presenting to a single emergency department with mild or moderate croup	Randomization and blinding Double blind, randomised trial. Sample size and attrition 99 children with 86 patients followed up by telephone (87%). Medication dosages Patients were randomised to receive 1 mg/kg prednisolone, 0.15 mg/kg dexamethasone or 0.6 mg/kg dexamethasone. Lengths of time of administration Single dose.	Primary outcome measures were the magnitude and rate of reduction in Westley croup score, rate of return for medical care with ongoing croup, and further treatment with steroids in the week following index presentation. Secondary outcome measures were the proportion of subjects requiring admission or salvage therapy, such as nebulized adrenaline, during index presentation. Follow-up by telephone interview at 7 days, Taussig score.	Responses to treatment No significant difference in admission rates, duration of symptoms or attendances. No significant difference between treatment groups.	Small sample size. Large number of eligible patients not recruited. Primary outcome over short time period.
[54] Garbutt et al, 2013	Children aged 1-8 years presenting to primary care offices with mild or moderate croup in the USA.	Randomization and blinding Double blinded randomised comparison trial Sample size and attrition Eighty-seven children randomised with 98% follow up at 11 days. Medication dosages and lengths of time of administration Prednisolone 2mg/ kg per day for 3 days versus one dose of 0.6mg dexamethasone and two placebo doses	Additional health care within 11 days of randomization assessed by self-report Telephone Out Patient score. Secondary outcomes included: duration of croup symptoms; disturbed sleep; parental stress; adverse events including sleep problems, mood changes, headache or dizziness, nausea, stomach pain, and secondary infections.	No difference in outcomes of either group for child or parent.	Small numbers. People with no telephone and lacking English spoken language were excluded.
[55] Parker et al, 2004	Children aged 4 months to 11 years, median age 2 years presenting in an emergency department with mild or moderate croup	Retrospective chart review. Randomization and blinding Not undertaken Sample size and attrition 188 eligible for analysis from 814 patients coded as croup. Medication dosages 1mg/kg prednisolone. Lengths of time of administration Not specified but implicitly single dose only	How long stridor at rest persisted after administration of prednisolone 1mg /kg. Westley and Geelhoed croup scores. To determine whether children with mild croup had a more rapid resolution of stridor at rest.	Average length of time from SAR to cessation was 6.5 hours. Children with mild croup improved quicker although this is reported as not clinically significant.	Retrospective chart reviews rely on completeness of contemporaneous data entry. Inter rater reliability reported as moderate. Lack of power calculation