Content area
Objective
To evaluate the efficacy differentials among corticosteroid regimens by type, dosage and duration, we conducted a systematic review and network meta-analysis of randomised controlled trials (RCTs).
Methods
We searched four databases from their inception to March 2024 and included RCTs that evaluated oral corticosteroids for asthma exacerbations in young people aged <21 years. We compared six regimens (dexamethasone (DEXA) 0.3 mg/kg/day administered for 1 day, DEXA 0.6 mg/kg/day for 1 day, DEXA 0.6 mg/kg/day for 2 days, prednisolone (PSL) 1.0 mg/kg/day for 3 days, PSL 1.0–1.5 mg/kg/day for 5 days and PSL 2.0 mg/kg/day for 5 days). Primary outcome was relapse within 14 days, defined as unplanned visit to an emergency department or primary care physician.
Results
Eleven studies involving 2353 patients were analysed in our quantitative synthesis. There were no significant differences in the relapse rates among 15 comparisons of six regimens. As part of the results, the network estimate showed that DEXA (0.3 mg/kg/day×1 day) compared with PSL (1.0 mg/kg/day for 3 days) had a risk ratio (RR) of 0.99 (95% CI 0.56 to 1.74), and DEXA (0.6 mg/kg/day for 2 days) compared with PSL (1.0–1.5 mg/kg/day×5 days) had an RR of 1.29 (95% CI 0.84 to 1.98). The certainty of the evidence for the included comparisons was low to very low.
Conclusion
In this network meta-analysis, there were no significant differences in the efficacy of commonly used corticosteroid regimens for acute exacerbations in childhood asthma. Short-term oral DEXA may be an acceptable alternative to a longer course of PSL.
PROSPERO registration number
CRD 42023449189.
Full text
Correspondence to Dr Shunsuke Amagasa; [email protected]
WHAT IS ALREADY KNOWN ON THIS TOPIC
Short-term dexamethasone and long-term prednisolone are not significantly different in acute asthma exacerbations in children.
WHAT THIS STUDY ADDS
A network meta-analysis approach compared the type, dosage and duration of treatment with oral corticosteroids and found no significant differences in the incidence of relapse among currently used regimens.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
Short-term oral dexamethasone may be an acceptable alternative to a longer course of prednisolone.
Introduction
Children with asthma may experience exacerbations triggered by infection or other factors, often necessitating emergency medical care in a substantial subset.1–3 Treatment begins with bronchodilators for immediate relief, followed by systemic corticosteroids, which are critical for reducing inflammation and airway hyperreactivity.4 Current guidelines recommend prompt corticosteroid use for moderate-to-severe episodes and mild cases unresponsive to beta-agonists.5 Nevertheless, symptom recurrence requiring further intervention or hospitalisation occurs in approximately 5–25% of cases.6
The outpatient management of paediatric asthma exacerbations frequently involves oral corticosteroid therapy, with prescribed regimens ranging from prednisolone (1–2 mg/kg daily for 3–5 days) to dexamethasone (0.3–0.6 mg/kg daily for 1–2 days).7–11 Comparative studies, including trials and meta-analyses, have shown similar efficacy between short-term and extended prednisolone courses.7–11 However, these meta-analyses compared prednisolone regardless of dose, duration or route of administration (oral, intramuscular), and few detailed comparative studies have addressed differences in the type, dosage and duration of treatment with oral corticosteroids.
To address this gap, we conducted a systematic review and network meta-analysis (NMA) of randomised controlled trials (RCTs) involving patients treated with oral corticosteroids, aiming to evaluate the efficacy differentials among oral corticosteroid regimens in terms of type, dosage and duration.
Materials and methods
Protocols and registration
This systematic review was designed based on the Preferred Reporting Items for Systematic Reviews and the Meta-analyses (PRISMA) extension statement for reporting systematic reviews incorporating NMAs.12 The review protocol was registered in PROSPERO on 7 August 2023 (CRD 42023449189).
Studies, participants, interventions/comparators and outcomes
This review included RCTs. When the number of trials incorporating specific interventions and outcomes was limited, making NMA infeasible, they were assessed qualitatively. Conference abstracts were generally excluded from synthesis and instead evaluated qualitatively. This ensured inclusion of potentially valuable data and reduced information bias. This meta-analysis included paediatric patients aged 1–20 years with acute asthma exacerbations who were treated in an outpatient or emergency department setting, not in an inpatient setting. The NMA was used to compare multiple corticosteroid protocols for managing acute asthma exacerbations encompassing the following: 0.3 mg/kg/day for a single day, 0.6 mg/kg/day for a single day, dexamethasone 0.6 mg/kg/day for 2 days, prednisolone 1.0 mg/kg/day for 3 days, prednisolone 1.0–1.5 mg/kg/day for 5 days and prednisolone 2.0 mg/kg/day for 5 days.
The primary outcome was relapse within 14 days, defined as an unplanned visit to an emergency department or primary care physician. Secondary outcomes were hospital re-admission within 14 days and vomiting at the hospital or at home. Although quantitative integration was not possible due to insufficient numbers of studies and heterogeneity issues, we conducted a qualitative assessment of missed school days, missed parental workdays and improved scores, such as the Pediatric Respiratory Assessment Measure (PRAM).
Data sources and search details
We searched the Cochrane Central Register of Controlled Trials and MEDLINE via Ovid for eligible published clinical trials. The WHO’s International Clinical Trials Platform search portal and the ClinicalTrials.gov trials registry were also searched for ongoing trials. The final searches were conducted on 24 March 2024. Details of the search strategy and searches performed are provided in the online supplemental appendix.
Study selection, data collection process and data elements
Rayyan software was used for systematic review.13 Two authors (SA and SU) individually screened titles and abstracts retrieved via the search and additional sources to determine eligibility based on the selection criteria. Full texts of potentially eligible studies were independently assessed by the same authors, with disagreements resolved through discussion with the third reviewer (KO). A standardised pre-pilot form was used to extract data on study quality and synthesis. Data included baseline characteristics, intervention and control details, methods, outcomes, time points and risk of bias. Two authors (SA and SU) extracted data independently; discrepancies were discussed with KO. Variables, such as age, severity and outcomes, were extracted. Age was recorded as a continuous variable, severity categorised by scores or qualitatively, and outcomes extracted as binary, including the number analysed and outcome occurrences per intervention.
Assessment of risk of bias within trials
The methods for assessment of risk of bias within trials are provided in the Supplementary Methods of the online supplemental file.
Statistical analysis
This study used NMA to compare the effects of multiple interventions simultaneously.14 15 Network plots were constructed to identify the number of studies and patients included in the meta-analysis. The R package ‘netmeta 2.5-0’ (V.4.1.2) and Confidence in Network Meta-Analysis (CINeMA), a web application that uses R, were used to perform NMA. A frequentist approach with multivariate meta-analysis was used, using a random-effects model to account for variability between studies. Effect sizes were expressed as relative risks with 95% CIs. The network meta-analysis in this study used individual arm data. Indirect and direct estimates were calculated using a node-split model. Analyses were performed using the number of patients analysed (not the number of patients randomised) for each trial. The certainty of network estimates was assessed using the CINeMA framework,16 based on six domains from the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) framework: bias within studies, bias across studies, indirectness, imprecision, heterogeneity and inconsistency.17
The transitivity assumption was assessed by comparing the distribution of clinical and methodological variables that may act as effect modifiers across treatment comparisons. Risk of bias between studies was assessed by considering pairwise meta-analyses. Conditions associated with ‘suspected’ and ‘unrecognised’ bias across studies were determined by the presence of publication bias as indicated by direct comparison. Each study within the network was assessed for indirectness in relation to the study topic, taking into account the study population, intervention, outcome and study setting. Indirectness was then categorised as low, moderate or high. A contribution matrix was then created using the study level judgements. Methods to address imprecision included contrasting the range of treatment effects covered by the 95% CI with the corresponding range. The heterogeneity of treatment effects relative to clinically important risk ratios (<0.75 or >1.33) in the CI was assessed. We compared the estimated heterogeneity variance with its predicted distribution derived from the model.18 The importance of heterogeneity with and without heterogeneity capture was assessed using CI-based assessments and agreement between predicted intervals. A prediction interval is calculated from the variance of the distribution of treatment effects (τ2). Heterogeneity of treatment effect was assessed when the clinically significant risk ratio was between 0.75 and 1.33 within the prediction interval. Inconsistency was evaluated using node-splitting analysis. Per-design interaction tests were used to statistically assess consistency, and network model discrepancies were estimated from discrepancy factors and their uncertainty.19 In our analysis using CINeMA, we assessed inconsistency within the network by comparing direct and indirect evidence.16 17 This involved assessing the agreement between effect estimates from different sources. Although statistical tests, such as treatment interaction tests, were used to identify potential discrepancies, the primary focus was on a qualitative assessment of the coherence of the evidence. Rather than relying solely on p values, CINeMA considers the context and potential clinical relevance of observed differences. This holistic approach includes consideration of whether differences between direct and indirect evidence are meaningful and significant in the context of the overall evidence base.
The assessment of publication bias and the ranking analysis are described in the Supplementary Methods in the online supplemental file.
Sensitivity analysis
To assess the robustness of the results, an NMA/MA with different assumptions about drug/dosage/duration was performed as a sensitivity analysis, with relapse, hospital re-admission and vomiting as outcomes. The first approach pools all dosages and durations of the same drug, assuming that neither dosage nor duration alters the treatment effect. The type of drug (dexamethasone vs prednisolone) is compared in this manner. The second approach pools different dosages of the same drug into a single treatment group, while differentiating between treatment nodes based on the duration of the drug’s use, focusing on the potential impact of treatment duration. The third approach takes the opposite stance, combining treatment nodes based on the duration of the drug’s use, but distinguishing between different dosages, thus focusing on the potential impact of varying dosages.
Results
Study selection and study characteristics
A literature search yielded 3287 articles (figure 1). After removing duplicates, 2380 publications were screened, and 46 underwent full-text review. The quantitative (network meta-analysis) analyses included 11 studies.6 20–29 One trial was included despite being a conference abstract, as detailed data were available on ClinicalTrials.gov.26 Four studies, excluded from the quantitative analysis due to differing outcomes or other factors, were assessed qualitatively.30–33 Inclusion and exclusion criteria, outcomes and follow-up periods for each study are shown in online supplemental table 1.
This network meta-analysis incorporated data from 2353 patients. The relevant studies, published between 2001 and 2022, recruited patients from 1999 to 2018, as indicated in table 1. The analysis evaluated six variants of oral corticosteroid therapy, differentiated by type, dosage and duration, in relation to the primary outcome. These variants comprised dexamethasone 0.3 mg/kg/day administered for 1 day (n=142, 6.0%), dexamethasone 0.6 mg/kg/day for 1 day (n=294, 12.5%), dexamethasone 0.6 mg/kg/day for 2 days (n=779, 33.1%), prednisolone 1.0 mg/kg/day for 3 days (n=221, 9.4%), prednisolone 1.0–1.5 mg/kg/day for 5 days (n=736, 31.3%) and prednisolone 2.0 mg/kg/day for 5 days (n=161, 6.8%).
Table 1Study characteristics II
| Study | Years of recruitment | Country of recruitment | Age | Intervention | Comparator | Asthma severity | Number of patients | ||
| Intervention | Comparator | Intervention | Comparator | ||||||
| Studies included in quantitative synthesis | |||||||||
| Qureshi et al 20016 | 1998–1999 | USA | 2–18 years | DEXA (0.6 mg/kg/day×2 days) | PSL (1.0 mg/kg/day×5 days) | Mild 21.7%, moderate 56.5%, severe 22.1%* | Mild 23.0%, moderate 55.9%, severe 21.1%* | 272 | 261 |
| Kayani et al 200221 | NA | USA | 2–16 years | PSL (1.0 mg/kg/day×5 days) | PSL (2.0 mg/kg/day×5 days) | NA | 43 | 43 | |
| Altamimi et al 200622 | 2001–2003 | Canada | 2–16 years | DEXA (0.6 mg/kg/day×1 day) | PSL (1.0 mg/kg/day×5 days) | Mild to moderate† | 56 | 54 | |
| Greenberg et al 200823 | 2004–2005 | USA | 2–18 years | DEXA (0.6 mg/kg/day×2 days) | PSL (1.0 mg/kg/day×5 days) | PAS 8 (range 5–14) | PAS 6 (range 5–12) | 51 | 38 |
| Chang et al 200824 | 2004–2007 | Australia | 2–15 years | PSL (1.0 mg/kg/day×3 days) | PSL (1.0 mg/kg/day×5 days) | ASS 4.6 (SD 1.8) | ASS 4.5 (SD 1.8) | 101 | 100 |
| Cronin et al 201625 | 2011–2012 | Ireland | 2–16 years | DEXA (0.3 mg/kg/day×1 day) | PSL (1.0 mg/kg/day×3 days) | PRAM 4.38 (SD 2.53) | PRAM 4.51 (SD 2.35) | 123 | 122 |
| Paniagua et al 201720 | 2014–2015 | Spain | 1–14 years | DEXA (0.6 mg/kg/day×2 days) | PSL (1.5 mg/kg/day×1 day, followed by 1 mg/kg/day×4 days) | NA | 281 | 276 | |
| Jara-Almonte et al 201826 | NA | USA | 18 months–20 years | DEXA (0.6 mg/kg/day×1 day) | DEXA (0.6 mg/kg/day×2 days) | PIS 4 (SD 2) | PIS 5 (SD 2.85) | 37 | 39 |
| Ullah et al 201927 | 2015–2016 | Pakistan | 2–12 years | DEXA (0.6 mg/kg/day×1 day) | PSL (2.0 mg/kg/day×5 days) | Severe or life-threatening asthma were excluded† | 80 | 80 | |
| Elkharwili et al 202028 | 2016–2017 | Egypt | 2–11 years | DEXA (0.6 mg/kg/day×2 days) | DEXA (0.3 mg/kg/day×1 day)/ PSL (1.5 mg/kg/day×5 days) | NA | 20 | 20/20 | |
| Martin et al 202229 | 2015–2018 | USA | 2–20 years | DEXA (0.6 mg/kg/day×1 day) | DEXA (0.6 mg/kg/day×2 days) | PAS 7.2 (SD 1.8) | PAS 6.9 (SD 1.5) | 116 | 116 |
| Studies not being included in quantitative synthesis but included in qualitative analysis | |||||||||
| Ghafouri et al 201030 | NA | USA | 2–17 years | DEXA (0.6 mg/kg/day×1 day) | DEXA (0.6 mg/kg/day×2 days | Mild to moderate† | 52 | 48 | |
| Prasannavenkatesh 2018 | 2015–2016 | India | 2–12 years | DEXA (0.3 mg/kg/day×1 day) | PSL (1.0 mg/kg/day×3 days) | PRAM 5.83 (SD 0.83) | PRAM 6.13 (SD 0.82) | 30 | 30 |
| Banoth et al 202232 | NA | India | 2–14 years | DEXA (0.3 mg/kg/day×2 days) | PSL (2.0 mg/kg/day×5 days) | PRAM 5.4 (SD 1.06) | PRAM 5.28 (SD 1.13) | 88 | 87 |
| Tayyab et al 202233 | 2018–2019 | Pakistan | 2–12 years | DEXA (0.3 mg/kg/day×1 day) | PSL (1.0 mg/kg/day×3 days) | PRAM <6 140.7%, PRAM ≥6 850.3% | PRAM <6 120.8%, PRAM ≥6 870.2% | 156 | 156 |
*Patients were classified as ‘mild’ if their PEFR was greater than 70% of predicted or their asthma score was 5–7, ‘moderate’ if their PEFR was 50% to 70% of predicted or their asthma score was 8–11, and ‘severe’ if their PEFR was less than 50% of predicted or their asthma score was 12–15.
†Inclusion or exclusion criteria include asthma severity, but no actual number is given.
ASS, Asthma Severity Scale; DEXA, dexamethasone; PAS, Pediatric Asthma Score; PIS, Pulmonary Index Score; PRAM, Pediatric Respiratory Assessment Measure; PSL, prednisolone.
The number of patients and occurrence of outcomes for each study included in the quantitative synthesis are shown in online supplemental table 2. The pharmaceutical formulations of each study included in the quantitative synthesis are shown in online supplemental table 3.
Risk of bias within individual studies
The results for assessment of risk of bias within trials are provided in the Supplementary Results and online supplemental tables 4-6 of the online supplemental file.
Relapse
The primary outcome of relapse was reported in 11 studies involving 2353 patients.6 20–29 The network plot for relapse is shown in figure 2A, and network estimates are in table 2A and figure 3A. Forest plots of direct, indirect and network estimates for relapse are in online supplemental figure 1. There was no evidence that any specific type, dosage or duration of corticosteroid reduced the rate of relapse. There was no significant publication bias (online supplemental figure 2). A summary of confidence in network estimates for the primary outcome is in online supplemental table 7, with certainty of evidence rated low to very low. The P-scores for relapse are in online supplemental table 8.
Figure 2. Network plots for relapse, admission and vomiting. The network plot is a visual representation of direct and indirect comparisons between different treatments and interventions. The nodes represent specific interventions, with the size of each node proportional to the number of patients who received that intervention. The edges between nodes indicate direct comparisons made in trials, with the thickness of the edges reflecting the number of trials that compared the linked interventions. DEXA, dexamethasone; PSL, prednisolone; RCT, randomised controlled trial.
(A) Direct estimates, indirect estimates and network estimates for relapse; (B) direct estimates, indirect estimates and network estimates for admission; (C) direct estimates, indirect estimates and network estimates for vomiting
| Comparison | Direct estimate RR (95% CI) | Indirect estimate RR (95% CI) | P-value for inconsistency measures* | Network estimate RR (95% CI) | Certainty of evidence |
| (A) | |||||
| DEXA (0.3 mg/kg/day×1 day) vs DEXA (0.6 mg/kg/day×2 days) | 1.17 (0.48 to 2.86) | 1.34 (0.41 to 4.37) | 0.87 | 1.23 (0.60 to 2.51) | Very low |
| DEXA (0.3 mg/kg/day×1 day) vs PSL (1.0 mg/kg/day×3 days) | 0.98 (0.53 to 1.83) | 1.00 (0.25 to 3.97) | 0.98 | 0.99 (0.56 to 1.74) | Low |
| DEXA (0.3 mg/kg/day×1 day) vs PSL (1.0–1.5 mg/kg/day×5 days) | 1.75 (0.61 to 5.06) | 1.46 (0.57 to 3.71) | 0.8 | 1.58 (0.78 to 3.18) | Very low |
| DEXA (0.6 mg/kg/day×1 day) vs DEXA (0.6 mg/kg/day×2 days) | 1.34 (0.71 to 2.53) | 0.56 (0.18 to 1.75) | 0.19 | 1.09 (0.63 to 1.90) | Very low |
| DEXA (0.6 mg/kg/day×1 day) vs PSL (1.0–1.5 mg/kg/day×5 days) | 3.67 (0.42 to 31.87) | 1.26 (0.62 to 2.58) | 0.36 | 1.40 (0.71 to 2.77) | Very low |
| DEXA (0.6 mg/kg/day×1 day) vs PSL (2.0 mg/kg/day×5 days) | 1.34 (0.71 to 2.53) | 0.56 (0.18 to 1.75) | 0.04 | 1.09 (0.63 to 1.90) | Very low |
| DEXA (0.6 mg/kg/day×2 days) vs PSL (1.0–1.5 mg/kg/day×5 days) | 1.23 (0.78 to 1.93) | 2.13 (0.49to 9.22) | 0.48 | 1.29 (0.84 to 1.98) | Very low |
| DEXA (0.6 mg/kg/day×2 days) vs PSL (2.0 mg/kg/day×5 days) | 1.99 (0.56 to 7.00) | 0.36 (0.15 to 0.82) | 0.03 | 0.60 (0.30 to 1.22) | Very low |
| PSL (1.0 mg/kg/day×3 days) vs PSL (1.0–1.5 mg/kg/day×5 days) | 1.58 (0.54 to 4.68) | 1.62 (0.57 to 4.62) | 0.98 | 1.60 (0.75 to 3.40) | Very low |
| PSL (1.0–1.5 mg/kg/day×5 days) vs PSL (2.0 mg/kg/day×5 days) | 0.33 (0.01 to 7.96) | 0.48 (0.21 to 1.10) | 0.83 | 0.47 (0.21 to 1.04) | Very low |
| DEXA (0.3 mg/kg/day×1 day) vs DEXA (0.6 mg/kg/day×1 day) | NA | 1.13 (0.46 to 2.75) | NA | 1.13 (0.46 to 2.75) | Very low |
| DEXA (0.3 mg/kg/day×1 day) vs PSL (2.0 mg/kg/day×5 days) | NA | 0.74 (0.28 to 1.99) | NA | 0.74 (0.28 to 1.99) | Very low |
| DEXA (0.6 mg/kg/day×1 d) vs PSL (1.0 mg/kg/day×3 days) | NA | 0.88 (0.34 to 2.28) | NA | 0.88 (0.34 to 2.28) | Very low |
| DEXA (0.6 mg/kg/day×2 days) vs PSL (1.0 mg/kg/day×3 days) | NA | 0.80 (0.36 to 1.78) | NA | 0.80 (0.36 to 1.78) | Very low |
| PSL (1.0 mg/kg/day×3 days) vs PSL (2.0 mg/kg/day×5 days) | NA | 0.75 (0.26 to 2.14) | NA | 0.75 (0.26 to 2.14) | Very low |
| (B) | |||||
| DEXA (0.3 mg/kg/day×1 day) vs PSL (1.0 mg/kg/day×3 days) | 2.95 (0.31 to 27.97) | NA | NA | 2.95 (0.31 to 27.97) | Very low |
| DEXA (0.6 mg/kg/day×1 day) vs PSL (1.0–1.5 mg/kg/day×5 days) | 2.89 (0.31 to 26.96) | NA | NA | 2.89 (0.31 to 26.96) | Very low |
| DEXA (0.6 mg/kg/day×2 day) vs PSL (1.0–1.5 mg/kg/day×5 days) | 0.98 (0.27–3.47) | NA | NA | 0.98 (0.27–3.47) | Very low |
| PSL (1.0 mg/kg/day×3 days) vs PSL (1.0–1.5 mg/kg/day×5 days) | 2.97 (0.12to 72.05) | NA | NA | 2.97 (0.12 to 72.05) | Very low |
| DEXA (0.3 mg/kg/day×1 day) vs DEXA (0.6 mg/kg/day×1 day) | NA | 3.03 (0.03 to 271.48) | NA | 3.03 (0.03 to 271.48) | Very low |
| DEXA (0.3 mg/kg/day×1 day) vs DEXA (0.6 mg/kg/day×2 days) | NA | 8.94 (0.15 to 540.50) | NA | 8.94 (0.15 to 540.50) | Very low |
| DEXA (0.3 mg/kg/day×1 day) vs PSL (1.0–1.5 mg/kg/day×5 days) | NA | 8.77 (0.18 to 433.94) | NA | 8.77 (0.18 to 433.94) | Very low |
| DEXA (0.6 mg/kg/day×1 day) vs DEXA (0.6 mg/kg/day×2 days) | NA | 2.95 (0.23 to 38.37) | NA | 2.95 (0.23 to 38.37) | Very low |
| DEXA (0.6 mg/kg/day×1 day) vs PSL (1.0 mg/kg/day×3 days) | NA | 0.97 (0.02 to 47.74) | NA | 0.97 (0.02 to 47.74) | Very low |
| DEXA (0.6 mg/kg/day×2 days) vs PSL (1.0 mg/kg/day×3 days) | NA | 0.33 (0.01 to 10.19) | NA | 0.33 (0.01 to 10.19) | Very low |
| (C) | |||||
| DEXA (0.3 mg/kg/day×1 day) vs DEXA (0.6 mg/kg/day×2 days) | 0.75 (0.19 to 2.93) | 1.14 (0.08 to 17.27) | 0.79 | 0.82 (0.24 to 2.76) | Very low |
| DEXA (0.3 mg/kg/day×1 day) vs PSL (1.0–1.5 mg/kg/day×5 days) | 0.43 (0.13 to 1.43) | 0.21 (0.001 to 29.72) | 0.79 | 0.41 (0.13to 1.33) | Very low |
| DEXA (0.6 mg/kg/day×1 day) vs DEXA (0.6 mg/kg/day×2 days) | 2.50 (0.81 to 7.74) | 0.61 (0.02 to 15.49) | 0.42 | 2.14 (0.74 to 6.23) | Very low |
| DEXA (0.6 mg/kg/day×1 day) vs PSL (1.0–1.5 mg/kg/day×5 days) | 0.32 (0.01 to 7.72) | 1.31 (0.37 to 4.67) | 0.42 | 1.08 (0.33 to 3.51) | Very low |
| DEXA (0.6 mg/kg/day×2 days) vs PSL (1.0–1.5 mg/kg/day×5 days) | 0.53 (0.30 to 0.94) | 0.13 (0.004 to 3.76) | 0.42 | 0.50 (0.29 to 0.89) | Very low |
| DEXA (0.6 mg/kg/day×2 days) vs PSL (2.0 mg/kg/day×5 days) | 0.53 (0.18 to 1.55) | NA | NA | 0.53 (0.18 to 1.55) | Very low |
| DEXA (0.3 mg/kg/day×1 day) vs DEXA (0.6 mg/kg/day×1 day) | NA | 0.38 (0.08to 1.89) | NA | 0.38 (0.08 to 1.89) | Very low |
| DEXA (0.3 mg/kg/day×1 day) vs PSL (2.0 mg/kg/day×5 days) | NA | 0.43 (0.09 to 2.20) | NA | 0.43 (0.09 to 2.20) | Very low |
| DEXA (0.6 mg/kg/day×1 day) vs PSL (2.0 mg/kg/day×5 days) | NA | 1.14 (0.25 to 5.17) | NA | 1.14 (0.25 to 5.17) | Very low |
| PSL (1.0–1.5 mg/kg/day×5 days) vs PSL (2.0 mg/kg/day×5 days) | NA | 1.06 (0.32 to 3.54) | NA | 1.06 (0.32 to 3.54) | Very low |
The RR represents the risk of outcome occurrence in the group on the left side of the comparison relative to the group on the right side. RR >1 indicates a higher risk in the left-side group, while RR <1 indicates a higher risk in the right-side group (eg, in the top comparison, the network estimate shows that DEXA (0.3 mg/kg/day×1 day) is associated with a 1.23 times higher risk of relapse compared to DEXA (0.6 mg/kg/day×2 days)).
RR >1 indicates a higher risk in the group on the left side of the comparison relative to the group on the right side (eg, in the top comparison, the network estimate shows that DEXA (0.3 mg/kg/day×1 day) is associated with a 2.95 times higher risk of admission compared to PSL (1.0 mg/kg/day×3 days)).
RR >1 indicates a higher risk in the group on the left side of the comparison relative to the group on the right side (eg, in the top comparison, the network estimate shows that DEXA (0.3 mg/kg/day×1 day) is associated with a 20% lower risk of vomiting compared to DEXA (0.6 mg/kg/day×2 days)).
*Inconsistency measures are the difference between direct and indirect estimates
DEXA, dexamethasone; PSL, prednisolone; RR, risk ratio.
Figure 3. Forest plots of network estimates for (A) relapse, (B) admission and (C) vomiting. DEXA, dexamethasone; PSL, prednisolone.
Hospital re-admission
The secondary outcome of admission was reported in five studies involving 1643 patients.6 20 22 24 25 The network plot for admission is shown in figure 2B, and network estimates are in table 2B and figure 3B. There was no evidence that any specific type, dosage or duration of corticosteroid reduced the rate of admission. A summary of confidence in network estimates for admission is in online supplemental table 9, with all values rated very low. The P-scores for admission are in online supplemental table 8.
Vomiting
The secondary outcome of vomiting was reported in five studies involving 1581 patients.6 20 22 23 28 29 The network plot for vomiting is shown in figure 2C, and network estimates are in table 2C and figure 3C. Dexamethasone (0.6 mg/kg/day×2 days) was associated with less vomiting compared with prednisolone (1.0 mg/kg/day×5 days) (risk ratio 0.50 (95% CI 0.29–0.89)). There was no evidence that another specific type, dosage or duration of corticosteroid decreased the rate of vomiting. A summary of confidence in network estimates for vomiting is shown in online supplemental table 10, with all values rated very low. The P-scores for vomiting are in online supplemental table 8.
Results of qualitative evaluation
The results of the qualitative evaluation are shown in the Supplementary Results and online supplemental tables 11, 12 of the online supplemental file.
Sensitivity analysis
Pairwise analyses for relapse, hospital re-admission and vomiting comparing all dosage and duration combinations of dexamethasone and prednisolone are shown in online supplemental figures 3-5. Network plots and the results of the NMA for recurrence, re-admission and vomiting, pooling different dosages of the same drug type into one treatment group and distinguishing nodes by duration, are presented in online supplemental tables 13-15 and figures 6-8. Similarly, network plots and NMA results pooling different durations and distinguishing nodes by dosage are shown in online supplemental tables 16-18 and figures 6-8. No significant differences were found among drug type, dosage or duration in any of the sensitivity analyses, as relapse and hospital re-admission. In a sensitivity analysis with vomiting as the outcome, dexamethasone was associated with significantly less vomiting in a pairwise analysis of dexamethasone and prednisolone, and in a NMA of dexamethasone (any dose×1 day) versus prednisolone (any dose×5 days) and dexamethasone (0.6 mg/kg/day) versus prednisolone (1.0 mg/kg/day).
Discussion
The current study assessed six commonly used corticosteroid protocols for managing acute asthma exacerbations in paediatric populations, ranging from dexamethasone (0.3–0.6 mg/kg daily for 1–2 days) to prednisolone (1–2 mg/kg daily for 3–5 days). The NMA suggested that there were no significant differences in important clinical outcomes, such as relapse rates, hospital admissions or instances of vomiting, among the varied corticosteroid types, dosages and treatment durations.
Previous meta-analyses found no significant difference between short-term dexamethasone and 3–5 days of prednisolone in paediatric asthma management.7–11 However, these studies grouped dexamethasone and prednisolone regardless of dosage or duration7 8 10 11 and did not distinguish administration routes, such as intramuscular versus oral.7 10 11 One NMA compared placebo with short-term prednisolone, dexamethasone and intramuscular corticosteroids but lacked detailed examinations of treatment types, amounts and durations.9 The current NMA addressed these gaps by employing a methodology that meticulously examined the differences in type, dosage and duration of oral corticosteroid regimens. Our analysis found no significant differences in relapse rates across regimens. Qualitative evaluations also found no substantial differences in PRAM score changes or consumer-focused outcomes, such as missed school or parental workdays. Therefore, these findings suggest that short-term oral dexamethasone offers advantages in patient compliance and familial convenience.
In some analyses, vomiting risk differed between prednisolone and dexamethasone, possibly due to formulation and taste.34 Cronin et al, for instance, used the superior but more expensive prednisolone sodium phosphate and observed little difference in vomiting rates between the two. However, as some studies lacked detailed formulation and taste data, further research is needed to clarify their impact.
Although no significant differences were observed and the certainty of the evidence is low to very low, some relapse comparisons with relatively narrow confidence intervals are noteworthy. For example, the network estimate for dexamethasone (0.3 mg/kg/day for 1 day) versus prednisolone (1.0 mg/kg/day for 3 days) showed a risk ratio of 0.99 (95% CI 0.56 to 1.74), indicating comparable efficacy. Moreover, absolute numbers in online supplemental table 2 suggest that relapse rates are similarly equivalent in the actual study population. For prednisolone (1.0–1.5 mg/kg/day for 5 days) versus prednisolone (2.0 mg/kg/day for 5 days), the network estimate showed a risk ratio of 0.47 (95% CI 0.21 to 1.04), suggesting the risk is less than half, with the upper CI close to 1. Online supplemental table 2 also does not indicate any clear advantage of higher doses of prednisolone.
The anti-inflammatory potency of corticosteroids indicates that 1 mg of prednisolone is equivalent to 0.15 mg of dexamethasone,35 implying greater daily dose potency for dexamethasone in many studies. However, our sensitivity analysis did not reveal a notable difference between doses, suggesting that the impact of this potency difference is limited.
There is ongoing debate about the diagnosis and management of asthma and viral-induced wheezing, particularly in children aged 5 years and younger. This controversy centres on the challenges of distinguishing asthma from transient viral-induced wheezing, leading to different management approaches. According to the Global Initiative for Asthma (GINA) guidelines, age-specific considerations are critical for accurate diagnosis and treatment, underscoring the need for careful clinical assessment in this population.5 However, no studies have compared the type and dosage of oral corticosteroids specifically for children under five, and the impact of age on corticosteroids efficacy remains unclear. Age may be an effect modifier; however, evidence for age-related differences is limited. Moreover, analysing infants, children and adolescents as a single group may have masked true effects. Further clinical trials focusing on specific age groups, particularly on children under five, are needed to provide clearer guidance on optimal treatment strategies.
In this study, the certainty of evidence was very low for most comparisons, primarily due to ‘serious concerns’ about imprecision.16 17 This stemmed from the small number of cases and events in each RCT, resulting in wide confidence intervals for effect sizes. This was also a consequence of the large number of groups compared with studying each treatment in detail. The risk of bias in each trial also reduced the certainty of the evidence. Large, high-quality RCTs are needed to provide further evidence and enable detailed comparisons of regimen effectiveness.
Strengths and limitations
In this study, we used NMA to examine the effects of oral corticosteroids on acute asthma exacerbations, focusing on type, dosage and duration of administration. This method enables simultaneous comparisons of multiple regimens using direct and indirect evidence, while reducing treatment heterogeneity compared with previous meta-analyses. We also performed three sensitivity analyses to confirm the robustness of our results under different hypotheses for drug type/dosage/duration.
However, this study has limitations. First, due to the detailed comparisons for each treatment regimen and several smaller trials, there may be a small group of cases, which may have caused problems with the precision and power of the comparisons. Second, several RCTs had concerns or a high risk of bias due to problems with the randomisation process, deviations from the intended interventions, and missing outcome data. Third, although the detailed subdivision of regimens reduced heterogeneity in treatment interventions, heterogeneity in inclusion criteria, follow-up protocols, and so forth was observed, as in previous meta-analyses.
Conclusions
In this network meta-analysis, there were no significant differences in the efficacy of commonly used corticosteroid regimens for acute exacerbations in childhood asthma. Short-term oral dexamethasone may be an acceptable alternative to a longer course of prednisolone.
Data availability statement
Data are available upon reasonable request.
Ethics statements
Patient consent for publication
Not applicable.
Ethics approval
Not applicable.
Contributors SA and S Uematsu contributed to the conception and design of the study. SA, S Utsumi and KO contributed to data acquisition, analysis and interpretation of data. SA contributed to analysis and data interpretation. SA drafted the article. S Utsumi, KO and S Uematsu critically revised the article for important intellectual content. All authors have read and approved the final version of the article. SA takes responsibility for the paper as a whole. SA is responsible for the overall content as guarantor.
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.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.
1 Scarfone RJ, Friedlaender E. Corticosteroids in acute asthma: past, present, and future. Pediatr Emerg Care 2003; 19: 355–61. doi:10.1097/01.pec.0000092585.40174.f6
2 Choi J, Lee GL. Common Pediatric Respiratory Emergencies. Emerg Med Clin North Am 2012; 30: 529–63. doi:10.1016/j.emc.2011.10.009
3 Akinbami LJ, Simon AE, Rossen LM. Changing Trends in Asthma Prevalence Among Children. Pediatrics 2016; 137: 1–7. doi:10.1542/peds.2015-2354
4 Taylor IK, Shaw RJ. The mechanism of action of corticosteroids in asthma. Respir Med 1993; 87: 261–77. doi:10.1016/0954-6111(93)90022-r
5 Global Institute for Asthma. Global Strategy for Asthma Management and Prevention, 2023. Available: https://ginasthma.org/2023-gina-main-report/
6 Qureshi F, Zaritsky A, Poirier MP. Comparative efficacy of oral dexamethasone versus oral prednisone in acute pediatric asthma. J Pediatr 2001; 139: 20–6. doi:10.1067/mpd.2001.115021
7 Keeney GE, Gray MP, Morrison AK, et al. Dexamethasone for acute asthma exacerbations in children: a meta-analysis. Pediatrics 2014; 133: 493–9. doi:10.1542/peds.2013-2273
8 Wei J, Lu Y, Han F, et al. Oral Dexamethasone vs. Oral Prednisone for Children With Acute Asthma Exacerbations: A Systematic Review and Meta-Analysis. Front Pediatr 2019; 7: 503. doi:10.3389/fped.2019.00503
9 Kirkland SW, Vandermeer B, Campbell S, et al. Evaluating the effectiveness of systemic corticosteroids to mitigate relapse in children assessed and treated for acute asthma: A network meta-analysis. J Asthma 2019; 56: 522–33. doi:10.1080/02770903.2018.1467444
10 Cai K-J, Su S-Q, Wang Y-G, et al. Dexamethasone Versus Prednisone or Prednisolone for Acute Pediatric Asthma Exacerbations in the Emergency Department: A Meta-Analysis. Pediatr Emerg Care 2021; 37: e1139–44. doi:10.1097/PEC.0000000000001926
11 Dahan E, El Ghazal N, Nakanishi H, et al. Dexamethasone versus prednisone/prednisolone in the management of pediatric patients with acute asthmatic exacerbations: a systematic review and meta-analysis. J Asthma 2023; 60: 1481–92. doi:10.1080/02770903.2022.2155189
12 Hutton B, Salanti G, Caldwell DM, et al. The PRISMA extension statement for reporting of systematic reviews incorporating network meta-analyses of health care interventions: checklist and explanations. Ann Intern Med 2015; 162: 777–84. doi:10.7326/M14-2385
13 Ouzzani M, Hammady H, Fedorowicz Z, et al. Rayyan-a web and mobile app for systematic reviews. Syst Rev 2016; 5: 210. doi:10.1186/s13643-016-0384-4
14 Watt J, Tricco AC, Straus S, et al. Research Techniques Made Simple: Network Meta-Analysis. J Invest Dermatol 2019; 139: 4–12. doi:10.1016/j.jid.2018.10.028
15 Dias S, Caldwell DM. Network meta-analysis explained. Arch Dis Child Fetal Neonatal Ed 2019; 104: F8–12. doi:10.1136/archdischild-2018-315224
16 Nikolakopoulou A, Higgins JPT, Papakonstantinou T, et al. CINeMA: An approach for assessing confidence in the results of a network meta-analysis. PLoS Med 2020; 17: e1003082. doi:10.1371/journal.pmed.1003082
17 Papakonstantinou T, Nikolakopoulou A, Higgins JPT, et al. CINeMA: Software for semiautomated assessment of the confidence in the results of network meta-analysis. Campbell Syst Rev 2020; 16: e1080. doi:10.1002/cl2.1080
18 Rhodes KM, Turner RM, Higgins JPT. Predictive distributions were developed for the extent of heterogeneity in meta-analyses of continuous outcome data. J Clin Epidemiol 2015; 68: 52–60. doi:10.1016/j.jclinepi.2014.08.012
19 Higgins JPT, Jackson D, Barrett JK, et al. Consistency and inconsistency in network meta-analysis: concepts and models for multi-arm studies. Res Synth Methods 2012; 3: 98–110. doi:10.1002/jrsm.1044
20 Paniagua N, Lopez R, Muñoz N, et al. Randomized Trial of Dexamethasone Versus Prednisone for Children with Acute Asthma Exacerbations. J Pediatr 2017; 191: 190–6. doi:10.1016/j.jpeds.2017.08.030
21 Kayani S, Shannon DC. Adverse behavioral effects of treatment for acute exacerbation of asthma in children: a comparison of two doses of oral steroids. Chest 2002; 122: 624–8. doi:10.1378/chest.122.2.624
22 Altamimi S, Robertson G, Jastaniah W, et al. Single-dose oral dexamethasone in the emergency management of children with exacerbations of mild to moderate asthma. Pediatr Emerg Care 2006; 22: 786–93. doi:10.1097/01.pec.0000248683.09895.08
23 Greenberg RA, Kerby G, Roosevelt GE. A Comparison of Oral Dexamethasone With Oral Prednisone in Pediatric Asthma Exacerbations Treated in the Emergency Department. Clin Pediatr (Phila) 2008; 47: 817–23. doi:10.1177/0009922808316988
24 Chang AB, Clark R, Sloots TP, et al. A 5- versus 3-day course of oral corticosteroids for children with asthma exacerbations who are not hospitalised: a randomised controlled trial. Med J Aust 2008; 189: 306–10. doi:10.5694/j.1326-5377.2008.tb02046.x
25 Cronin JJ, McCoy S, Kennedy U, et al. A Randomized Trial of Single-Dose Oral Dexamethasone Versus Multidose Prednisolone for Acute Exacerbations of Asthma in Children Who Attend the Emergency Department. Ann Emerg Med 2016; 67: 593–601. doi:10.1016/j.annemergmed.2015.08.001
26 Jara-Almonte G, Melville L, Kelly C. One vs Two Doses of Dexamethasone for Pediatric Asthma Exacerbation. Acad Emerg Med 2018; 25: S182.
27 Ullah I, Ul Haq S, Khan RM, et al. To Compare the Efficacy of Single Dose Oral Dexamethasone versus Multi-Dose Prednisolone in the Treatment of Acute Asthma Exacerbation. Med Forum 2019; 30: 59.
28 Elkharwili DA, Ibrahim OM, Elazab GA, et al. Two regimens of dexamethasone versus prednisolone for acute exacerbations in asthmatic Egyptian children. Eur J Hosp Pharm 2020; 27: 151–6. doi:10.1136/ejhpharm-2018-001707
29 Martin M, Penque M, Wrotniak BH, et al. Single-Dose Dexamethasone Is Not Inferior to 2 Doses in Mild to Moderate Pediatric Asthma Exacerbations in the Emergency Department. Pediatr Emerg Care 2022; 38: e1285–90. doi:10.1097/PEC.0000000000002727
30 Ghafouri N, Sharieff GQ, Rajasingham A, et al. Comparison of one-dose and two-dose regimens of oral dexamethason in the management of acute asthma exacerbations in the pediatric emergency department. Pediatr Emerg Care 2010; 26: 705–6.
31 Prasannavenkatesh K, Gunasekaran D, Pandi K, et al. Single-Dose Oral Dexamethasone Compared with Three Day Course of Oral Prednisolone in Children with Moderate Exacerbation of Asthma-A Pilot Double-Blinded Randomised Controlled Trial. J Clin Diagn Res 2008; 12.
32 Banoth B, Verma A, Bhalla K, et al. Comparative effectiveness of oral dexamethasone vs. oral prednisolone for acute exacerbation of asthma: A randomized control trial. J Family Med Prim Care 2022; 11: 1395–400.
33 Tayyab A, Asif A, Qazi S, et al. Effectiveness of single dose oral dexamethasone versus multidose prednisolone for treatment of acute exacerbations of asthma among children. Annals of Medicine & Surgery 2022; 84: 104799. doi:10.1016/j.amsu.2022.104799
34 Aljebab F, Alanazi M, Choonara I, et al. Observational study on the palatability and tolerability of oral prednisolone and oral dexamethasone in children in Saudi Arabia and the UK. Arch Dis Child 2018; 103: 83–8. doi:10.1136/archdischild-2017-312697
35 Glucocorticoid therapy Treatment summaries - BNF - NICE, Available: https://bnf.nice.org.uk/treatment-summaries/glucocorticoid-therapy/
© 2025 Author(s) (or their employer(s)) 2025. No commercial re-use. See rights and permissions. Published by BMJ Group.