Abstract
Introduction: This systematic review and network meta-analysis aimed to compare and evaluate the efficacy and safety of five medications, dupilumab, tralokinumab, upadacitinib, baricitinib, and abrocitinib, for the treatment of adolescent atopic dermatitis (AD), in order to provide decision support to support clinical decision-making by developing more scientifically grounded and effective treatment strategies. Methods: A comprehensive search was conducted in PubMed, Embase, Web of Science (WoS), and the Cochrane database to collect randomized controlled trials (RCTs) and Phase 3 clinical trials up to April 13, 2024. Supplementary data were retrieved from trial registries, and researchers contacted study authors and pharmaceutical companies when necessary to obtain complete data. Inclusion criteria comprised treatment studies for moderate to severe AD in adolescents aged 12 and above, with outcome measures including efficacy and safety assessments. Data extraction and risk bias assessment were independently performed by two researchers, using Excel for data extraction and the netmeta package in R software for network meta-analysis. Sensitivity analysis and bias risk assessment were conducted to validate the robustness and credibility of the results. Our research protocol was registered in PROSPERO (CRD42023480597) and did not require approval from an Institutional Review Board or written informed consent. Results: In the primary efficacy outcome measures, upadacitinib 30 mg/day, upadacitinib 15 mg/day, and dupilumab 300 mg/2 weeks demonstrated excellent efficacy in EASI75 compared to placebo, significantly outperforming other medications and placebo. Dupilumab 300 mg/2 weeks, upadacitinib 30 mg/day, and upadacitinib 15 mg/day showed excellent treatment effects in IGA0/1. Among the outcome measures for improvement in itch severity rating PP-NRS4, dupilumab 300 mg/2 weeks and tralokinumab 300 mg/2 weeks showed the highest efficacy values. Compared to these medications, baricitinib 1 mg/day exhibited weaker performance across all three indicators, particularly in EASI75 and IGA0/1, with effects approaching no significant difference. Due to limited sample sizes, estimates for treatment-emergent adverse events, serious adverse events (SAEs), and drug-induced adverse events safety indicators were unstable, preventing strong conclusions on safety outcomes. There are significant differences in the incidence rates of adverse reactions such as nasopharyngitis, acne, and AD among various medications. Conclusion: Upadacitinib and dupilumab demonstrate strong efficacy and symptom improvement in the treatment of moderate to severe AD in adolescents, particularly in reducing the severity of skin lesions and itchiness. Therefore, these medications should be considered as primary treatment options for adolescents with AD. However, further studies with long-term follow-up and larger sample sizes are necessary to thoroughly investigate the safety profiles of these medications in adolescents. This underscores the importance of closely monitoring the side effects of different drugs during clinical treatment to tailor optimal therapeutic strategies based on individual patient needs.
Introduction
Atopic dermatitis (AD) is one of the most common inflammatory skin diseases, typically onset in childhood and may persist into adulthood, with onset in middle age or late adulthood being uncommon [1]. Due to its characteristics of chronic, inflammatory, recurrent nature and itchiness, AD can have a significant impact on the quality of life and mental health of adolescent patients, in particular. Currently, there are numerous medications available for the treatment of AD, such as the biologics dupilumab, tralokinumab, nemolizumab, lebrikizumab, mepolizumab, etc., and JAK inhibitors including upadacitinib, abrocitinib, baricitinib. Among the various biologics targeting AD, this study specifically compares the biologics dupilumab and tralokinumab with JAK inhibitors.
Dupilumab acts by binding to and blocking the interleukin-4 (IL-4) receptor alpha chain (IL-4Rα), thereby inhibiting the signaling pathways of both IL-4 and interleukin-13 (IL-13). It is a biologic drug approved by the FDA for use in adults and children aged 6 months and above with moderate to severe AD [2, 3]. Numerous clinical trials [4‒10] have widely recognized the efficacy and safety of dupilumab. Tralokinumab functions by inhibiting the interaction of the protein IL-13 with its cell surface receptors IL-13R α1 and IL-13R α2. It was approved by the FDA in December 2023 for adolescents aged 12 and above [11]. Upadacitinib [12] and abrocitinib [13] are oral JAK1 inhibitors. As of January 2022, the FDA has approved them for adolescents (aged 12 and above) and adults with refractory moderate to severe AD, where systemic therapies (including biologics) have not provided adequate control or are not recommended. Baricitinib is an oral reversible selective janus kinase 1/2 (JAK1/JAK2) inhibitor approved in Europe for the treatment of moderate to severe AD in adult patients [14, 15].
However, the current comparative efficacy and safety data among these medications are unclear, especially the lack of head-to-head comparison studies of these drugs in the adolescent population, making the efficacy and safety of these five medications in adolescents inconclusive. Therefore, this systematic review and network meta-analysis aim to compare and evaluate the efficacy and safety of dupilumab, tralokinumab, upadacitinib, baricitinib, and abrocitinib in the treatment of adolescent AD, providing decision support for clinicians to develop more scientific and effective treatment strategies.
Methods
Search Strategy and Selection Criteria
In designing the search strategy and selection criteria for this network meta-analysis, we comprehensively searched the PubMed, Embase, Web of Science (WoS), and Cochrane databases. The target study types included randomized controlled trials (RCTs) and phase 3 clinical trials from the inception of the databases to April 13, 2024, to ensure a thorough collection of relevant studies. To supplement information from published clinical trials, we searched for additional results in trial registries. Additionally, we also contacted study authors and pharmaceutical companies as needed to obtain more complete data, in order to enhance the comprehensiveness and accuracy of our analysis. Refer to online supplementary material 1 for the complete search strategy (for all online suppl. material, see https://doi.org/10.1159/000543397).
Eligibility Criteria
The inclusion criteria consist of studies with treatment duration of at least 12 weeks targeting adolescents with moderate to severe AD. The study types include randomized controlled trials and phase 3 clinical trials, and involve comparing two of five drugs with each other or comparing one drug with a placebo. Outcome measures include efficacy and safety. The assessment of efficacy is based on changes in clinical signs reported by researchers, with Eczema Area and Severity Index (EASI) scores, Investigator Global Assessment (IGA) scores, and Peak Pruritus Numerical Rating Scale (PP-NRS) scores as the main parameters for assessing efficacy. The evaluation of safety should include assessing the incidence of adverse events and serious adverse event (SAEs), as well as drug-related discontinuation.
Screening and Abstraction Process
Using Endnote 20 to screen titles, abstracts, and full texts, data extraction and bias risk assessment are independently conducted by two individuals. In case of discrepancies, detailed discussions are carried out between the two individuals to review the data and evaluation methods, seeking possible solutions. If disagreements persist, consultation with supervisors and relevant experts is sought for resolution.
Data extraction is performed using an Excel spreadsheet, including the following: (1) Study basic information: study title, first author, publication date, study drugs; (2) Intervention information: intervention duration, intervention drugs and dosages, number of subjects and gender in each intervention group; (3) Baseline data: duration of the disease, extent of skin involvement, disease severity scores (EASI, IGA, NRS, etc.); (4) Outcome indicators: efficacy scores (EASI, IGA, NRS, etc.), safety indicators (incidence of adverse reactions and serious adverse reactions, drug-related discontinuation). In cases of missing data, communication with the corresponding author and drug companies is conducted to retrieve the missing data.
The Cochrane Risk of Bias Tool (RoB2) is utilized to assess the quality and bias risk of studies. The main efficacy outcome indicators include: IGA0/1 used to assess the degree of improvement in disease status, indicating patients reaching a clinical symptom-free or nearly symptom-free state. EASI used to assess the severity and extent of AD, involving the evaluation of the area and severity of skin lesions. EASI75 indicates at least 75% improvement in the subjective assessment of AD area and severity compared to baseline. PP-NRS4 refers to an improvement of at least 4 points in the PP-NRS, which measures itch severity at its worst moment over the past 24 h. The parameters measured include pruritus severity, mean pruritus score, and peak pruritus score. For safety indicator data, the number and percentage of patients experiencing any adverse events, SAEs, and drug-related discontinuation are recorded.
Data Analysis
This study aims to systematically evaluate the effectiveness of different drug treatments to provide decision support for future clinical practices. By analyzing data from multiple clinical trials, the study explores the effects of five drug treatments for adolescent AD compared to a placebo. Data processing and analysis are conducted using the netmeta package in the R software for network meta-analysis. The study calculated and compared the log odds ratios for each treatment group relative to their corresponding placebo group and presented the overall effects of each treatment using forest plots.
A network meta-analysis was generated for each analysis. For the three efficacy outcome indicators (IGA0/1, EASI75, and PP-NRS4), the data imported into the R software include the number of responders and sample sizes for each treatment group and control group in each study. The dplyr package is used to group each study and calculate the proportion of responders and sample sizes for each treatment group relative to the corresponding placebo group. The original ratios (ORs) and relative risks are calculated, and the original ratios are transformed into log odds ratios (TE) while computing their standard errors (seTE). The netmeta package is utilized for network meta-analysis. A data frame containing treatment group, placebo group, TE, and seTE is created for each comparison in each study with a unique label. The netmeta function is then used to perform network meta-analysis, set as a random effects model, with the placebo group as the reference group. Finally, the forest function is employed to generate forest plots for the network meta-analysis, showing the effect estimates and confidence intervals of the log odds ratios for each treatment group.
Subgroup and Sensitivity Analyses
Due to the small sample size or homogeneity of the study population, it is not appropriate to conduct subgroup analysis in this study. In order to assess and validate the robustness of the study results and ensure that the reliability of the research conclusions is not influenced by any single treatment group, this study employs an exclusion method for sensitivity analysis. This analysis method focuses on evaluating the stability and consistency of various treatment effects under conditions of high heterogeneity in the data. By excluding specific treatments or data groups, we can observe the impact of these exclusions on the overall study results. This helps us assess the dependency of key conclusions and identify potential vulnerabilities in the study findings.
Risk of Bias and Certainty of Evidence
In this study, we utilized the Cochrane Risk of Bias tool to assess the risk of bias in each study. We also applied the Grading of Recommendations Assessment, Development and Evaluation (GRADE) criteria within Egger network meta-analysis, to evaluate the overall quality of evidence for comparisons between clinically relevant medications. However, as the GRADE grading evaluation has not yet been conducted, we are unable to provide specific results and grades at this time.
Protocol and Updates
Our protocol has been registered on PROSPERO (CRD42023480597). As this study involves the summary analysis and processing of secondary data, Institutional Review Board approval and written informed consent are not required.
Results
The study included a total of 9 articles [16‒24] shown in Table 1, comprising 8 phase III clinical trials and 1 post hoc analysis of a phase III clinical trial, encompassing 3 phase III clinical trials of abrocitinib, 2 phase III clinical trials of upadacitinib, and 1 phase III clinical trial each of baricitinib, dupilumab, and tralokinumab. In total, 1,801 patients were involved. The risk of bias is low for all assessed factors shown in Table 2. Detailed study characteristics and extracted result data are provided in online supplementary material 2. Insufficient data on the long-term outcomes of the five drugs for adolescents precluded a network meta-analysis. The network plot of adolescent treatment lasting 16 weeks exhibited a radial distribution, with doses of each drug connected to the placebo and no direct connections between nodes of different active drugs shown in Figure 1.
Summary of characteristics and outcomes of included studies
Study . | Patient characteristics . | Duration, weeks . | Intervention . | Dosages applied . | Main outcomes . | ||||
---|---|---|---|---|---|---|---|---|---|
age group . | number . | age, years, mean (SD) . | females (%) . | efficacy . | safety . | ||||
LIBERTY AD ADOL | Adolescents | 251 | 14.5 (1.7) | 103 (41) | 16 | Placebo | \ | IGA0/1 | TEAEs |
Dupilumab | 300 mg/4 weeks | EASI75 | SAEs | ||||||
Dupilumab | 300 mg/2 weeks | PP-NRS4 | DIAEs | ||||||
ECZTRA 6 | Adolescents | 289 | 15 (2.2) | 140 (48) | 16 | Tralokinumab | 150 mg/2 weeks | IGA0/1 | TEAEs |
Tralokinumab | 300 mg/2 weeks | EASI75 | SAEs | ||||||
Placebo | \ | PP-NRS4 | DIAEs | ||||||
Measure up 1 | Adolescents | 189 | 15.4 (1.8) | 103 (54) | 16 | Upadacitinib | 15 mg/day | IGA0/1 | TEAEs |
Upadacitinib | 30 mg/day | EASI75 | SAEs | ||||||
Placebo | \ | PP-NRS4 | DIAEs | ||||||
Measure up 2 | Adolescents | 180 | 15.5 (1.7) | 97 (54) | 16 | Upadacitinib | 15 mg/day | IGA0/1 | TEAEs |
Upadacitinib | 30 mg/day | EASI75 | SAEs | ||||||
Placebo | \ | PP-NRS4 | DIAEs | ||||||
JADE MONO-1 and JADE MONO-2 | Adolescents | 124 | 15 (1.9) | 56 (45) | 12 | Placebo | \ | IGA0/1 | TEAEs |
Abrocitinib | 100 mg/day | EASI75 | SAEs | ||||||
Abrocitinib | 200 mg/day | PP-NRS4 | DIAEs | ||||||
JADE TEEN | Adolescents | 285 | 15 (3.0) | 140 (49) | 12 | Placebo | \ | IGA0/1 | TEAEs |
Abrocitinib | 100 mg/day | EASI75 | SAEs | ||||||
Abrocitinib | 200 mg/day | PP-NRS4 | DIAEs | ||||||
BREEZE-AD PEDS | 2 to < 18 years | 440 | 12 (4.0) | 242 (55) | 16 | Baricitinib | 1 mg/day | IGA0/1 | TEAEs |
Baricitinib | 2 mg/day | EASI75 | SAEs | ||||||
Baricitinib | 4 mg/day | PP-NRS4 | DIAEs | ||||||
Placebo | \ |
Study . | Patient characteristics . | Duration, weeks . | Intervention . | Dosages applied . | Main outcomes . | ||||
---|---|---|---|---|---|---|---|---|---|
age group . | number . | age, years, mean (SD) . | females (%) . | efficacy . | safety . | ||||
LIBERTY AD ADOL | Adolescents | 251 | 14.5 (1.7) | 103 (41) | 16 | Placebo | \ | IGA0/1 | TEAEs |
Dupilumab | 300 mg/4 weeks | EASI75 | SAEs | ||||||
Dupilumab | 300 mg/2 weeks | PP-NRS4 | DIAEs | ||||||
ECZTRA 6 | Adolescents | 289 | 15 (2.2) | 140 (48) | 16 | Tralokinumab | 150 mg/2 weeks | IGA0/1 | TEAEs |
Tralokinumab | 300 mg/2 weeks | EASI75 | SAEs | ||||||
Placebo | \ | PP-NRS4 | DIAEs | ||||||
Measure up 1 | Adolescents | 189 | 15.4 (1.8) | 103 (54) | 16 | Upadacitinib | 15 mg/day | IGA0/1 | TEAEs |
Upadacitinib | 30 mg/day | EASI75 | SAEs | ||||||
Placebo | \ | PP-NRS4 | DIAEs | ||||||
Measure up 2 | Adolescents | 180 | 15.5 (1.7) | 97 (54) | 16 | Upadacitinib | 15 mg/day | IGA0/1 | TEAEs |
Upadacitinib | 30 mg/day | EASI75 | SAEs | ||||||
Placebo | \ | PP-NRS4 | DIAEs | ||||||
JADE MONO-1 and JADE MONO-2 | Adolescents | 124 | 15 (1.9) | 56 (45) | 12 | Placebo | \ | IGA0/1 | TEAEs |
Abrocitinib | 100 mg/day | EASI75 | SAEs | ||||||
Abrocitinib | 200 mg/day | PP-NRS4 | DIAEs | ||||||
JADE TEEN | Adolescents | 285 | 15 (3.0) | 140 (49) | 12 | Placebo | \ | IGA0/1 | TEAEs |
Abrocitinib | 100 mg/day | EASI75 | SAEs | ||||||
Abrocitinib | 200 mg/day | PP-NRS4 | DIAEs | ||||||
BREEZE-AD PEDS | 2 to < 18 years | 440 | 12 (4.0) | 242 (55) | 16 | Baricitinib | 1 mg/day | IGA0/1 | TEAEs |
Baricitinib | 2 mg/day | EASI75 | SAEs | ||||||
Baricitinib | 4 mg/day | PP-NRS4 | DIAEs | ||||||
Placebo | \ |
Network graphs of studies included in the analysis of adolescents treated between 12 and 16 weeks.
Network graphs of studies included in the analysis of adolescents treated between 12 and 16 weeks.
Efficacy Analysis
A random-effects model was used to assess the effectiveness of the drugs. The EASI was used to evaluate the severity and extent of AD in patients, including the assessment of the area and severity of skin lesions. In the outcome measure of EASI75 shown in Figure 2a, the results indicated that compared to placebo, the efficacy of upadacitinib 30 mg/day (effect size 1.82, 95% confidence interval: [1.33; 2.30]) and upadacitinib 15 mg/day (effect size 1.75, 95% confidence interval: [1.26; 2.24]) was most significant. Additionally, dupilumab 300 mg/2 weeks also demonstrated high efficacy (effect size 1.62, 95% confidence interval: [0.86; 2.37]). In contrast, the effect size of baricitinib 1 mg/day was close to zero (0.01), and its 95% confidence interval (−0.36; 0.37) included 0, suggesting that its effect may not be significantly different.
Forest plots. a Forest plot of EASI75. b Forest plot of IGA0/1. c Forest plot of PP-NRS4.
Forest plots. a Forest plot of EASI75. b Forest plot of IGA0/1. c Forest plot of PP-NRS4.
The IGA ranges from 0 to 4, where a score of 0 indicates clear skin and a score of 1 indicates almost clear skin. The IGA scores of 0–1 are often used as an outcome measure in clinical trials to assess treatment efficacy. In the outcome measure of IGA0/1 shown in Figure 2b, the results demonstrate that dupilumab 300 mg/2 weeks, upadacitinib 30 mg/day, and upadacitinib 15 mg/day show excellent therapeutic effects, with log odds ratios of 2.34, 2.37, and 2.05, respectively. Abrocitinib 200 mg/day and upadacitinib 15 mg/day also exhibit relatively high efficacy, with log odds ratios of 0.68 (95% confidence interval: [0.28; 1.08]) and 2.05 (confidence interval: [1.31; 2.80]), indicating that these drugs are significantly superior to placebo. In comparison, the effect of baricitinib 1 mg/day is relatively weak, with a log odds ratio of only 0.10, and its confidence interval includes 0, suggesting a possibly insignificant effect.
PP-NRS4 refers to at least a 4-point improvement in itch severity as measured on PP-NRS. By analyzing the improvement of each treatment drug compared to placebo in PP-NRS4 shown in Figure 2c, it is evident that all drugs demonstrate varying degrees of efficacy improvement compared to placebo. In particular, dupilumab 300 mg/2 weeks (effect size 2.05, 95% confidence interval: [1.05; 3.05]) and tralokinumab 300 mg/2 weeks (effect size 2.05, 95% confidence interval: [0.78; 3.12]) exhibit the highest effect sizes. The effect sizes for upadacitinib 30 mg/day and 15 mg/day are 2.11 ([1.43; 2.80]) and 1.84 ([1.14; 2.55]), respectively, showing similarly excellent performance. Statistically, abrocitinib 200 mg/day and 100 mg/day have effect sizes of 0.79 (95% confidence interval: [0.42; 1.16]) and 0.52 (95% confidence interval: [0.13; 0.92]), indicating a certain treatment potential. Additionally, the effect size of baricitinib 4 mg/day is 0.91 (95% confidence interval: [0.18; 1.64]), which falls within the effective range. However, the effect size of baricitinib 1 mg/day is 0.21, and its confidence interval is (−0.64; 1.05), suggesting that its effect may be less apparent.
Heterogeneity Testing and Management
This study analyzed the heterogeneity of three efficacy indicators in the treatment of AD in adolescents using a random-effects model. The overall effect estimates for EASI75, IGA0/1, and PP-NRS4 were 1.5139 (95% CI: 0.9252–2.1026), 1.5214 (95% CI: 0.9166–2.1262), and 1.5801 (95% CI: 1.0530–2.1071), respectively, all showing significant features of heterogeneity.
Given the limitations of the study sample and characteristics, subgroup analysis was not performed. Instead, a sensitivity analysis using exclusion methods was employed to explore heterogeneity and validate the study’s conclusions. After excluding the influence of the placebo and other factors, the heterogeneity of the data was effectively controlled, resulting in a more consistent and credible trend. The analysis revealed that excluding specific drugs from various treatment regimens led to varying degrees of overall efficacy changes.
For the efficacy of IGA0/1: the exclusion of the placebo significantly increased the overall average efficacy of IGA0/1, indicating that other drugs have stronger efficacy compared to the placebo. Excluding certain drugs such as baricitinib 4 mg/day and upadacitinib 30 mg/day resulted in a decrease in overall efficacy, highlighting their importance in enhancing treatment outcomes. Other drugs, such as abrocitinib 100 mg/day and abrocitinib 200 mg/day, had relatively minor effects on overall efficacy after exclusion.
For the EASI75 efficacy response: excluding abrocitinib 100 mg/day and abrocitinib 200 mg/day led to a significant decrease in EASI75 efficacy, indicating the important role of these two drugs in promoting efficacy responses. Excluding the placebo showed a positive effect on improving the overall efficacy response. The impact of excluding upadacitinib 15 mg/day and upadacitinib 30 mg/day on the overall efficacy response was relatively low, demonstrating their stable efficacy.
For the NRS efficacy: excluding abrocitinib 100 mg/day and abrocitinib 200 mg/day resulted in a significant decrease in overall NRS efficacy, indicating the critical role of these two drugs in symptom improvement. Exclusion of the placebo led to improved NRS efficacy for most drugs, demonstrating the positive impact of actual intervention. The impact of excluding upadacitinib 15 mg/day and upadacitinib 30 mg/day on overall NRS efficacy was relatively low, indicating their stable clinical effect.
Safety Analysis
In the study, we conducted an analysis of three key safety indicators using a Bayesian multilevel logistic regression model to assess the impact of different treatments on treatment-emergent adverse events (TEAEs), SAEs, and drug-induced discontinuations (DIAEs). The model included fixed effects for each treatment and random effects for between-study variability. The following outlines our main findings.
TEAEs: The results showed that the treatments with abrocitinib 200 mg/day and upadacitinib 30 mg/day increased the odds of adverse events. The confidence intervals for some treatments included zero, indicating that their statistical significance was inconclusive. Nevertheless, the good convergence of the model suggests that these estimates are reliable.
SAEs: Compared to baseline, treatments including abrocitinib 200 mg/day, various doses of baricitinib, dupilumab 300 mg/2 weeks or 300 mg/4 weeks, and tralokinumab 150 mg/2 weeks or 300 mg/2 weeks did not significantly affect the incidence of adverse events. The effect estimate confidence intervals for most treatments included zero, suggesting that the impact of these treatments on SAEs was not statistically significant.
DIAEs: Most treatment regimens did not show a significant impact, although abrocitinib 200 mg/day displayed a slight positive effect. However, the confidence intervals also included zero, indicating a high level of uncertainty regarding this finding.
Based on the analysis of adverse reactions associated with various medications used in the treatment of adolescent AD, we draw the following conclusions (Table 3): in terms of upper respiratory infections, the highest incidence rates were observed in the abrocitinib 200 mg/day group (10.6%) and the tralokinumab 300 mg/2 weeks group (11.3%). The incidence of headaches was highest in the dupilumab 300 mg/2 weeks group (11%), followed by both the abrocitinib 200 mg/day and baricitinib 2 mg/day groups, each at 9.2%. For nasopharyngitis, the abrocitinib 100 mg/day group had the highest occurrence at 13.0%. Baricitinib 2 mg/day reported relatively lower rates of nasopharyngitis. Regarding acne, the upadacitinib 30 mg/day group exhibited the highest acne incidence at 15.1%. In terms of AD, the tralokinumab 150 mg/2 weeks group showed the highest incidence at 13.3%. The dupilumab 300 mg/4 weeks and 300 mg/2 weeks groups also reported high rates of AD, at 18.1% and 18.3%, respectively.
Summary table of adverse reactions related to drug treatment
Intervention group . | Total . | Upper respiratory infections . | Headache . | Nasopharyngitis . | |||
---|---|---|---|---|---|---|---|
n . | % . | n . | % . | n . | % . | ||
Baricitinib 1 mg | 121 | 3 | 2.5 | 7 | 5.8 | 4 | 3.3 |
Baricitinib 2 mg | 120 | 4 | 3.3 | 11 | 9.2 | 5 | 4.2 |
Baricitinib 4 mg | 120 | 5 | 4.2 | 6 | 5.0 | 5 | 4.2 |
Abrocitinib 100 mg | 146 | 12 | 8.2 | 9 | 6.2 | 19 | 13.0 |
Abrocitinib 200 mg | 142 | 15 | 10.6 | 13 | 9.2 | 13 | 9.2 |
Tralokinumab 150 mg/2 weeks | 98 | 8 | 8.2 | 5 | 5.1 | NA | NA |
Tralokinumab 300 mg/2 weeks | 97 | 11 | 11.3 | 6 | 6.2 | NA | NA |
Upadacitinib 15 mg | 122 | 11 | 9.0 | 8 | 6.6 | 5 | 4.1 |
Upadacitinib 30 mg | 126 | 11 | 8.7 | 9 | 7.1 | 5 | 4.0 |
Dupilumab 300 mg/4 weeks | 84 | 6 | 7.2 | 4 | 4.8 | 9 | 10.8 |
Dupilumab 300 mg/2 weeks | 82 | 10 | 12.2 | 9 | 11.0 | 3 | 3.7 |
Intervention group . | Total . | Upper respiratory infections . | Headache . | Nasopharyngitis . | |||
---|---|---|---|---|---|---|---|
n . | % . | n . | % . | n . | % . | ||
Baricitinib 1 mg | 121 | 3 | 2.5 | 7 | 5.8 | 4 | 3.3 |
Baricitinib 2 mg | 120 | 4 | 3.3 | 11 | 9.2 | 5 | 4.2 |
Baricitinib 4 mg | 120 | 5 | 4.2 | 6 | 5.0 | 5 | 4.2 |
Abrocitinib 100 mg | 146 | 12 | 8.2 | 9 | 6.2 | 19 | 13.0 |
Abrocitinib 200 mg | 142 | 15 | 10.6 | 13 | 9.2 | 13 | 9.2 |
Tralokinumab 150 mg/2 weeks | 98 | 8 | 8.2 | 5 | 5.1 | NA | NA |
Tralokinumab 300 mg/2 weeks | 97 | 11 | 11.3 | 6 | 6.2 | NA | NA |
Upadacitinib 15 mg | 122 | 11 | 9.0 | 8 | 6.6 | 5 | 4.1 |
Upadacitinib 30 mg | 126 | 11 | 8.7 | 9 | 7.1 | 5 | 4.0 |
Dupilumab 300 mg/4 weeks | 84 | 6 | 7.2 | 4 | 4.8 | 9 | 10.8 |
Dupilumab 300 mg/2 weeks | 82 | 10 | 12.2 | 9 | 11.0 | 3 | 3.7 |
Overall, while certain treatment regimens demonstrated potential increased risks of adverse events and discontinuations, the lack of statistical significance and wide confidence intervals suggest that these results need to be validated in a broader dataset. The analysis of random effects in the model indicated variability in baseline effects between studies, enhancing our understanding of the mechanisms behind treatment effect differences. The Rhat values for all indicators were close to 1, and there were sufficient effective sample sizes (Bulk_ESS and Tail_ESS), highlighting the stability and reliability of the model estimates. Based on these results, we recommend that when formulating future treatment guidelines and making clinical decisions, the potential differences in adverse effects of these treatments should be fully considered. When designing clinical trials, a comprehensive assessment combining treatment effects with potential adverse reactions should be conducted. This integrated analysis provides a solid statistical foundation for understanding complex treatment effects while ensuring the scientific rigor of clinical decisions.
Discussion
This study compared the efficacy of different drug treatments for moderate to severe AD in adolescents using three key efficacy indicators: EASI75, IGA0/1, and PP-NRS4. In this network meta-analysis comprising 8 phase 3 clinical trials, we summarized the efficacy and safety of dupilumab, tralokinumab, abrocitinib, upadacitinib, and baricitinib over a treatment duration of up to 16 weeks. The results indicated that these five drugs were associated with significant improvements in the index scores over the duration of treatment. In the primary efficacy outcome measures, upadacitinib 30 mg/day, upadacitinib 15 mg/day, and dupilumab 300 mg/2 weeks demonstrated excellent efficacy in achieving EASI75 compared to placebo, and their effectiveness was significantly better than other drugs and placebo. Dupilumab 300 mg/2 weeks, upadacitinib 30 mg/day, and upadacitinib 15 mg/day showed excellent treatment effects in achieving IGA0/1. In terms of PP-NRS4, dupilumab 300 mg/2 weeks and tralokinumab 300 mg/2 weeks exhibited the highest effect sizes. In comparison, baricitinib 1 mg/day showed relatively weak performance across all three indicators, especially in EASI75 and IGA0/1, with effects approaching insignificance. While abrocitinib showed some efficacy in certain instances, its performance in IGA0/1 and PP-NRS4 was not as strong as the aforementioned drugs.
Although AD has a significant impact on the adolescent population, there has been relatively less focus on drug treatment research for adolescents compared to adults and children. Consequently, there is a substantial lack of head-to-head comparative studies on the treatment of severe AD in adolescents using these five drugs. In the adult population, systematic reviews and network meta-analyses [25] have compared the aforementioned drugs: abrocitinib 200 mg/day and upadacitinib 30 mg/day were found to be more effective than dupilumab, while upadacitinib 15 mg/day showed similar scores to dupilumab. Abrocitinib 100 mg/day, baricitinib 2 mg/day, baricitinib 4 mg/day, and tralokinumab 300 mg/2 weeks were found to have slightly lower scores compared to dupilumab. The conclusion that abrocitinib 200 mg/day has better effects than dupilumab contradicts our research findings. This disparity might be attributed to potential differences in treatment response among different populations, possibly related to differences in demographic characteristics and sample size. Despite the differences in research outcomes, this does not imply that abrocitinib lacks value in the treatment of AD. This study might provide a direction worth further exploration to find treatment regimens more suitable for individual needs.
Compared with placebo, treatment with abrocitinib 200 mg/day and upadacitinib 30 mg/day showed an increasing trend in TEAEs. For SAEs, the effect estimate confidence intervals for most treatments included zero, indicating that the impact of these treatments on serious adverse events was not statistically significant. Regarding DIAEs, most treatment regimens did not show a significant impact. Although abrocitinib 200 mg/day exhibited a slight positive effect, there remains uncertainty about its statistical significance. Due to the relatively small sample size, the estimates for the three safety indicators – TEAEs, SAEs, and DIAEs – are unstable, and robust conclusions about safety outcomes cannot be drawn. Therefore, further research such as long-term follow-ups and studies with larger sample sizes is still needed to explore the safety of these five drugs in the adolescent population. There are significant differences in the incidence rates of adverse reactions such as nasopharyngitis, acne, and AD among various medications. This highlights the importance of closely monitoring the side effects of different drugs during clinical treatment to tailor optimal therapeutic strategies based on individual patient needs.
Limitations
The phase 3 clinical trial of baricitinib (BREEZE-AD-PEDS) has some limitations. There is only one phase 3 clinical study of baricitinib specifically focusing on adolescents, and the age range included in this trial is 2–18 years, with adolescents accounting for 70%. Therefore, we ultimately included this study in our analysis. However, due to the unavailability of age-specific stratified data, we had to use the data from the entire trial range, which could potentially impact the accuracy of the conclusions regarding the adolescent population. In adolescents, physiological and biological characteristics may differ from younger or older patients, leading to different effects on drug efficacy and safety. Therefore, we also considered potential differences among different age groups when interpreting the results. This age limitation is an issue that requires further investigation, and future studies should strive to ensure the acquisition and utilization of independent data specifically targeting the adolescent population to more accurately assess the performance of baricitinib in this age group.
Conclusion
Upadacitinib and dupilumab have demonstrated strong efficacy and symptom improvement in the treatment of moderate to severe AD in adolescents, particularly in reducing the severity of skin lesions and itchiness. Therefore, these medications should be considered as main treatment options for adolescent AD. Meanwhile, less effective drugs, such as baricitinib 1 mg/day, may require further evaluation with a larger number of cases or longer study durations to assess their role in AD treatment. By systematically comparing the performance of multiple drugs across three different treatment indicators, this study aims to provide clinical evidence for selecting effective treatments for AD. Future research should focus more on adolescents of different races, age groups, and disease severities to obtain more comprehensive treatment efficacy data and evaluations.
Statement of Ethics
It is not applicable because this study is based exclusively on published literature.
Conflict of Interest Statement
The authors have no conflicts of interest to declare.
Funding Sources
This study was supported by research funding from Novartis, Pfizer, Astellas, Galderma, Janssen, GSK, BAYER, LEO, MEDA Pharma, and ALK Pharma. Dr. Zhao serves as a speaker/advisor for these companies and has received funding for the study design, execution, and analysis.
Author Contributions
Zuotao Zhao, Chengyue Peng, and Lijuan Liu: writing – original draft, conceptualization, and project administration. Chengyue Peng and Yaqi Zheng: data curation, formal analysis, methodology, and visualization. Yen Tan, Xiaoting Song, and Peixin Zhang: data curation, formal analysis, and investigation. Zuotao Zhao, Xiaojie Huang, and Litao Zhang: supervision, and writing – review and editing.
Additional Information
Zuotao Zhao, Chengyue Peng, and Lijuan Liu contributed equally to this work.Edited by: H.-U. Simon, Bern.
Data Availability Statement
The data used in this meta-analysis primarily came from published literature, and all referenced studies can be accessed through publicly available databases and journals. The data that support the findings of this study are openly available at https://kdocs.cn/l/ccJOw1cps4gE.