Efficacy and safety of biological agents for the treatment of pediatric patients with psoriasis: A bayesian analysis of six high-quality randomized controlled trials

Background Biological agents have been used with extreme caution in children because of their possible adverse effects. Objectives This study used high-quality randomized controlled trials (RCTs) to provide high-level evidence to assess the effectiveness and safety of biological agents for treating children with psoriasis. Methods We searched PubMed, Embase, Cochrane, and Web of Science databases through October 31, 2021. We included trials reporting at least one adverse event after treatment with biological agents of patients less than 18-year-old diagnosed with psoriasis. RevMan 5.3 and Stata 15.0 software were used for meta and Bayesian analyses. Results Six trials with 864 participants were included in the analysis. The results showed a 2.37-fold higher response rate in all biologics groups than in the control group for psoriasis area and severity index 75 (PASI75) (RR= 2.37, P-value < 0.01, 95% confidence interval [CI] [1.22, 4.62]). Compared with placebo, the PASI75 response rates of etanercept (RR= 2.82, 95% [CI] [1.10, 7.21]), ustekinumab low dose (RR= 7.45, 95%[CI] [1.25, 44.58]), and ustekinumab high dose (RR= 7.25, 95%[CI] [1.21, 43.41]) were superior. Additionally, the incidence of total adverse reactions was 1.05 times higher for biologics than for controls, indicating a good safety profile (RR= 1.05, P-value = 0.53, 95%[CI] [0.92, 1.19]). Overall, these six high-quality randomized controlled trials suggest that biologics are effective and safe for pediatric patients with psoriasis. Limitations Inclusion of few relevant, high-quality RCTs. Conclusion The results of this study indicate that biologics can be used to treat children with moderate-to-severe psoriasis without the risk of adverse effects. Ustekinumab showed the best efficacy and the fewest adverse effects.


Introduction
Psoriasis is a common immune-mediated condition primarily characterized by skin lesions, affecting approximately 3% of the population worldwide (1), and nearly one-third of the patients present with disease symptoms before adulthood (2). Pediatric psoriasis is often highly visible and uncomfortable, and approximately 20% of children present with moderate to severe disease, which may require effective systemic therapy (3). Biologics are an important therapeutic option for moderate-tosevere psoriasis when other treatments are contraindicated or ineffective (4). Evidence confirms that the correlation between biologics and serious infections is small in the general population; however, biologics have been used in special populations, such as children, pregnant women, and the elderly, with extreme caution owing to the risk of adverse effects (5).
In 2002, etanercept was approved by the US Food and Drug Administration (FDA) to treat psoriasis. Twelve years later, in 2014, it was expanded to the treatment of pediatric patients (4-17 years old) with chronic moderate-to-severe plaque psoriasis (6). In 2021, secukinumab was approved by the FDA for the treatment of patients aged 6-18 years with moderate-to-severe plaque psoriasis. Currently, ixekizumab and ustekinumab are also approved to treat patients over six years of age with mild-tomoderate plaque psoriasis. However, relevant evaluations for utilizing biologics in childhood and adolescent psoriasis are lacking.
Biological agents are engineered monoclonal antibodies and fusion proteins capable of therapeutic action by blocking specific cytokines or cytokine receptors critical to psoriatic inflammation (7). It is speculated that T H 1 and T H 17 mediate psoriasis, and inhibition of T H 17 is a known therapeutic strategy. Drugs targeting tumor necrosis factor(TNF)-a, interleukin(IL)-23, and IL-17 are effective for the clinical treatment of psoriasis (8). TNF-a inhibitors dominate the therapeutic market for psoriasis treatment. Furthermore, IL-12/IL-23 and IL-17 inhibitors could reduce the expansion and production of T H 17, thereby suppressing the production of proinflammatory factors with relatively few adverse reactions (9). Accordingly, biologics are being progressively used to rapidly and effectively decrease psoriasis severity, although side effects have been noted, along with relapses after drug withdrawal.
Several studies have demonstrated the relative safety of biologics in special populations, particularly pediatric and pregnant patients (10-12). However, individual reports have limited power to characterize the risk factors for AEs in children and adolescents, and clinicians and patients are eager to receive corresponding guidance. Therefore, this review collates information from currently available studies documenting the apparent side effects of biologics, which can direct future clinical trials.

Methods
This review was conducted using the Cochrane Handbook on Systematic Review of Interventions and presented following the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) guidelines (13) (Data Sheet 1).

Data inclusion
We determined the inclusion and exclusion criteria for the present analysis before conducting a literature search. The inclusion criteria were as follows: RCTs reporting at least one type of AE after intervention with biologics, subjects diagnosed with psoriasis and aged<18 years, regardless of sex and ethnicity.

Data extraction
Two investigators (X.C. Cai and M. Zhang) independently screened the studies according to the inclusion criteria using self-designed data extraction templates for each included study. Two authors (X.C. Cai and X.Y. Sun) assessed the risk of bias, and three authors (Y.Q. Zhou, Y. Luo, and J.L. Chen) performed data analysis and interpretation.

Outcome measures
The primary efficacy measure was the PASI 75, a quantitative rating score that measures the severity of psoriatic lesions based on area coverage and plaque appearance, including scaling, infiltration, and erythema. In addition, the PASI 50, PASI 90, PASI 100, sPGA, sPGA0/1, and DLQI 0/1 indices were used to evaluate the efficacy outcomes. The main outcome indicators for safety evaluations were AEs and serious AEs. The secondary indicators of adverse reactions were the five types of adverse reactions produced by biologics (14). This classification could help better deal with the clinical features of these side effects, identify individual and general risk factors, and direct research in this novel area of medicine.

Statistical analysis
RevMan5.3 software provided by the Cochrane Collaboration, was used to synthesize the results of the metaanalysis. Risk ratios (RR) with 95% confidence intervals (CIs) were calculated for dichotomous data. The mean difference (MD) and standard mean difference (SMD) were used for continuous data. Across trials, a fixed-effects model was used for homogeneity (I 2 < 25%); otherwise, a random-effects model was used. Statistical significance was set at P-value < 0.05.
Network analysis was conducted in Stata 15.0 and using Markov Chain Monte Carlo (MCMC) simulation randomeffects modeling in a Bayesian frame. For each analysis, we used three chains and generated 100,000 iterations. The first 20,000 iterations were the burn-in periods to eliminate the effect of initial values, and the last 80,000 iterations were used for sampling. The convergence was evaluated using a trace plot. Inconsistency between direct and indirect comparisons was assessed by the "node-splitting" method. The surface under the cumulative ranking (SUCRA) (%) was calculated. If SUCRA approaches 100%, the intervention is the best among the included trials; conversely, if it is close to 0%, it represents the worst intervention.

Study selection and characteristics
We identified 172 studies after a preliminary search of four databases and 87 related articles (related articles and citations). After removing duplicate studies, 195 remained. Subsequently, the titles and abstracts of each study were screened, and 43 studies were available for a detailed evaluation. Of the remaining articles, 37 were excluded, of which 14 were not available in full text, 19 were non-RCTs, 2 did not report AEs, and 2 were excluded due to other reasons such as the articles being letters. Six full-text articles were included in the final analysis (15-20). All the trials were published in English ( Figure 1).
Finally, six trials involving 864 participants were included in the analysis: four clinical studies compared biologics with placebo, one study compared two biologics, and one compared a biologic with methotrexate (MTX) ( Table 1).

Risk of bias assessment
All the included RCTs described a specific stochastic approach. Three studies (15, 18, 20) did not report details regarding selection bias. All studies were double-blinded (participants and personnel) and were assessed in a blinded manner. We evaluated the completed data without selective reporting or other bias. In summary, the quality of the included studies was high, and the results of this meta-analysis were credible ( Figure 2).

Outcomes Efficacy
Psoriasis Area and Severity Index (PASI) 75 PASI 75 indicates that the lesion score improved by 75% compared with the baseline score. In the present study, the PASI 75 was used to evaluate the primary efficacy of biologics for broader applications. We observed that the efficacy of biologics (ustekinumab, ixekizumab, and etanercept) was significantly higher than that of placebo ( Table 2, Data Sheet 2). In addition, adalimumab was more effective than MTX (Table 2; Data Sheet 2). Collectively, the biologic-treated groups presented 2.37-fold better results than the control group (

PASI 50
Two included articles employed PASI 50 as the outcome indicator, and the response rate with biologics was better than that with placebo at the endpoint (

PASI 100
The PASI 100 score after ixekizumab treatment was markedly greater (27.76-fold better) than that after placebo treatment (

Static Physician's Global Assessment (sPGA)
The sPGA score assesses the average thickness, erythema, and scaling of psoriatic lesions, with scores ranging from 0 (clear) to 4 (severe), with 0/1 indicating clear or almost clear. Two studies used sPGA 0 as the outcome indicator, and the results were significant, favoring treatment with ustekinumab and ixekizumab over placebo (

Children's Dermatology Life Quality Index (CDLQI)
Three studies used the CDLQI scores to evaluate the general curative effect as a complete response. The biologics (ustekinumab and ixekizumab) showed better performance against psoriatic pathogenesis in adolescent patients ( Table 2, Data Sheet 2). Overall, the efficacy outcomes were better in the biologic group than in the control group.

Safety AEs
We compared among the groups. Surprisingly, there was no significant difference in the frequency of overall adverse reactions Flowchart of study inclusion according to the PRISMA 2009 guidelines. RCTs: randomized control trials; AEs, adverse events.

Serious AEs
We also compared the frequency of serious adverse reactions between the biologics and placebo and MTX groups. The probability of severe AEs was higher in the biologic group than in the control group (

Specific adverse reactions
Type a AEs occur after the abundant release of inflammatory factors; these AEs are most frequent with complicated and changeable symptoms. Type b AEs are immune-mediated and are more serious than type a AEs, including immediate and delayed hypersensitivity reactions. Types g and d AEs are short-and long-term toxicities, respectively. Type ϵ AEs occur during drug withdrawal, particularly when a drug is suddenly stopped (14).

Type a
In this study, type a AEs were the most frequently observed adverse reactions. The incidence of infection was 1.06-fold higher in the biologics group than in the control group (P-value = 0.48; RR =  Figure 3).

Type b
In the present study, injection-site reactions were the most common type of b-adverse reactions, and the incidence of infection at the injection site was 2.60-fold higher in the biologics group than   (Figure 3).

Type g
The most frequently reported type g adverse reaction was a hand fracture. The probability of hand fracture was 0.75-fold higher in the biologic group than in the control group (P-value = 0.67; RR = 0.75; 95%[CI] [0.43, 1.30]) ( Table 4, Data Sheet 5). Other common type g AEs are Crohn's disease (CD) and irritable bowel disease; only one case of these two reactions, induced following ixekizumab therapy, was reported in the present study (Figure 3).

Type ϵ
Three types of adverse reactions, namely, depression, agitation, and self-injury, were documented in the biologics group ( Figure 3).

Bayesian analysis outcomes
Five RCTs involving 791 patients reported a PASI of 75. The Bayesian analysis performed was a network meta-analysis ( Figure 4). The size of the nodes represented the sample size of each study; the link between two nodes indicated direct comparisons, with the thickness of the lines representing the number of reports. No inconsistencies were detected when using the node-splitting method (P-value > 0.05) (Data Sheet 6). The scatter chart revealed a low bias ( Figure 5).
Based on the Bayesian framework, we performed a metaanalysis using an MCMC random-effects model and generated 21 pairs of comparisons, of which three showed significant differences ( Figure 6). Compared with that of A (placebo), the  Figure 7. The efficacy comparison estimated using the ladder diagram is presented in Table 5. Bolded font indicates a statistical difference.

Discussion
Juvenile-onset psoriasis is typically associated with a high probability of a positive family history and more serious symptoms, necessitating prompt therapeutic intervention (21). Treatment with biologics is the most effective for psoriasis, with a low incidence of adverse reactions in adults and a certain degree of safety (22). However, caution should be exercised when treating special populations, such as children. The present analysis provides comprehensive data on the use of biologics in children and adolescents with psoriasis. This is the first report of AEs in pediatric patients with psoriasis who were systematically exposed to biologics. Studies have reported that the curative effect of biologics is considerable in pediatric patients compared to that in controls, and the AEs rates are similar to those in adults. In the present study, accumulated data suggested that anti-TNFa preparations, such as etanercept, have lower efficacy, while adalimumab seemed more effective in treating psoriasis in children and adolescents. A previous study confirmed that the efficacy of etanercept may be similar to that of MTX (23, 24), which is consistent with our findings. Our Bayesian analysis showed differences in the efficacy of etanercept TNF-a inhibitor and ustekinumab IL-23 inhibitor compared with placebo. The highest dose of ustekinumab showed the best efficacy. Based on subjective evaluation scales (sPGA and CDLQI), ixekizumab was preferred for symptom improvement by clinicians, whereas ustekinumab was preferred by patients. The frequency of AEs between the biologics group and controls was almost equivalent, whereas the symptom severity was 1.22-fold higher in the biologics group than in the control group. These results indicated that the probability of AEs following treatment with biologics was not high, but rather more severe, which may decrease their acceptability.
To characterize the severe AEs caused by biologics, we classified and analyzed the different AEs induced by the biological agents. Accordingly, type a adverse reactions appeared to occur the most frequently. Children with psoriasis are prone to infections, headache, nasopharyngitis, injection-site infections, and skin eruptions. Compared to other biologics, etanercept is more likely to cause adverse reactions. Ustekinumab caused fewer AEs in children with psoriasis and could be recommended as a therapeutic option. In summary, our analysis showed that ustekinumab conferred better clinical efficacy with a low incidence of AEs, and hence could be recommended.
Herein, we reviewed retrospective RCTs evaluating biologics to determine cautionary issues and treatment risks during therapy in adolescents. This information could potentially enhance the quality and efficiency of future clinical research. The strengths of this study are as follows: (i) the included studies were of high quality, minimizing selection bias; (ii) all included studies had corresponding control groups, monitoring the bias and generating compelling evidence; and (iii) no pharmaceutical industry was involved in the present work. However, the analysis was limited by the small sample size, which was insufficient to determine the recommended dose and treatment duration of novel biological agents for pediatric patients. Future studies should focus on applying biologics in special populations, and clinicians should encourage relevant patients to enroll in prospective pharmacovigilance registries, undoubtedly helping to tackle unresolved questions in this field. Mapping of specific adverse events between biological agents and control groups. MTX: methotrexate. IBD: inflammatory bowel disease.
Finally, because of the short duration of use of biologics in children and the small number of relevant RCTs, the sample size of this study was not large enough, and the database search only yielded studies on secukinumab, ixekizumab, adalimumab, ustekinumab, and etanercept; the other biologics were not evaluated. Furthermore, the treatment duration of the included studies was not uniform and the recurrence rate was not reported.    SUCRA rank plot of seven interventions. If the SUCRA approaches 100%, the intervention is the best among the included trials; conversely, if it is close to 0%, it represents the worst intervention.

Conclusion
In conclusion, biologics can effectively treat children with psoriasis and can greatly improve their quality of life, eliminate lesions, and improve pruritus severity. Although AEs were reported in all the included studies, biologics can still be safely used to treat pediatric patients with moderate-to-severe psoriasis. Additionally, our analysis revealed that ustekinumab conferred good clinical efficacy with a low incidence of AEs, and hence could be recommended.

Data availability statement
The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding authors.

Author contributions
XC and YR conceived the study. XC and YR designed this study. Y-qZ, YL, and JC performed literature search and data extraction. XC and MZ assessed the quality of the trials and analyzed the data. YR and CW drafted the original manuscript, and BL and XL contributed to the manuscript revision. All authors have read and approved the final manuscript.