Tone · cochrane  ·  Domain · il23_inhibitors_psoriasis_biologics  ·  Generated 20260527_214803

Comparative Efficacy and Safety of IL-23 Inhibitors Versus Other Biologics in Adults with Moderate-to-Severe Plaque Psoriasis: A Systematic Review and Meta-Analysis

Summary of Findings

Outcome Effect estimate (95% CI) No. of participants (studies) Certainty (GRADE) Plain-language summary
PASI 90 (proportion) Pooled proportion: 0.72 (0.65 to 0.79) Not estimable (k = 43) ⊕◯◯◯ Very low IL-23 inhibitors likely achieve PASI 90 in a majority of treated patients, but the true rate may differ substantially due to high between-study variability
PASI 90 (SMD vs comparator) SMD: −0.61 (−1.11 to −0.11) Not estimable (k = 8) ⊕◯◯◯ Very low IL-23 inhibitors may be favoured over comparators for PASI 90, but confidence in this estimate is very low
PASI 90 (risk ratio vs placebo) RR: 17.98 (7.73 to 41.78) Not estimable (k = 6) ⊕◯◯◯ Very low All biologic classes are substantially more effective than placebo for PASI 90
PASI 100 (proportion) Pooled proportion: 0.49 (0.40 to 0.57) Not estimable (k = 29) ⊕◯◯◯ Very low Approximately half of patients treated with IL-23 inhibitors achieve complete skin clearance, though the true rate is very uncertain
PASI 75 (proportion) Pooled proportion: 0.82 (0.64 to 0.92) Not estimable (k = 21) ⊕◯◯◯ Very low A large majority of patients achieve PASI 75, though estimates range widely across studies
IGA 0/1 Pooled proportion: 0.78 (0.60 to 0.89) Not estimable (k = 4) ⊕◯◯◯ Very low Most patients achieve clear or almost clear skin by investigator assessment
sPGA 0/1 Pooled proportion: 0.82 (0.76 to 0.87) Not estimable (k = 5) ⊕◯◯◯ Very low Approximately four in five patients achieve clear or minimal disease
DLQI 0/1 Pooled proportion: 0.69 (0.63 to 0.74) Not estimable (k = 9) ⊕◯◯◯ Very low About two-thirds of patients report no or minimal quality-of-life impairment
Adverse events (proportion) Pooled proportion: 0.09 (0.04 to 0.18) Not estimable (k = 13) ⊕◯◯◯ Very low Adverse event rates appear low, though reporting definitions vary widely
Drug survival Pooled proportion: 0.92 (0.88 to 0.95) Not estimable (k = 7) ⊕◯◯◯ Very low Over 90% of patients persist on IL-23 inhibitor therapy, suggesting good tolerability and sustained efficacy
Treatment discontinuation Pooled proportion: 0.10 (0.05 to 0.19) Not estimable (k = 4) ⊕◯◯◯ Very low Approximately one in ten patients discontinues treatment
Infections Pooled proportion: 0.11 (0.02 to 0.37) Not estimable (k = 4) ⊕◯◯◯ Very low Infection rates are uncertain and vary markedly across studies

Overall GRADE assessment: The certainty of evidence for all pooled outcomes was rated very low, driven primarily by very serious inconsistency (I² > 80% for all outcomes) and serious risk of bias (the majority of evidence weight derived from low- or medium-credibility studies with prevalent industry sponsorship and no preregistration). Confidence in the precise magnitude of treatment effects is therefore limited, although the direction of benefit over placebo is consistent and robust across sensitivity analyses.

Background

Description of the condition

Plaque psoriasis is a chronic, immune-mediated inflammatory skin disease affecting approximately 2–3% of the global population. Moderate-to-severe disease — typically defined as body surface area (BSA) involvement ≥10%, Psoriasis Area and Severity Index (PASI) ≥10, or Dermatology Life Quality Index (DLQI) >10 — imposes a substantial burden through visible skin lesions, pruritus, and systemic comorbidities including cardiovascular disease, metabolic syndrome, and psychological distress (Kearney et al., 2025; Papp et al., 2025a).

Description of the intervention

IL-23 inhibitors (guselkumab, risankizumab, tildrakizumab) are fully human monoclonal antibodies that selectively target the p19 subunit of interleukin-23, a key cytokine in the pathogenic IL-23/IL-17 axis. These agents are compared against TNF-alpha inhibitors (adalimumab, infliximab, etanercept), IL-17 inhibitors (secukinumab, ixekizumab, brodalumab, bimekizumab), and the IL-12/23 inhibitor ustekinumab.

How the intervention might work

IL-23 acts upstream in the pathogenic cascade by promoting the survival, expansion, and effector function of Th17 cells and tissue-resident memory T cells (TRM cells). Selective IL-23p19 blockade is hypothesised to achieve durable clinical responses by depleting pathogenic TRM17 cells from lesional tissue without broadly suppressing host defence, potentially conferring a favourable long-term safety profile relative to broader immunosuppressive agents (Eyerich et al., 2024; Bernardini et al., 2022).

Why this review is important

The treatment landscape for moderate-to-severe plaque psoriasis now includes over a dozen biologic agents across four mechanistic classes. While individual randomised controlled trials (RCTs) and network meta-analyses (NMAs) have evaluated these agents, the comparative efficacy hierarchy — particularly between IL-23 inhibitors and IL-17 inhibitors — remains contested. Very few direct head-to-head trials exist between individual IL-23 inhibitors. Real-world evidence has proliferated but has not been systematically synthesised alongside trial data. This review addresses the need for a comprehensive, class-level and agent-level comparison incorporating both RCT and real-world evidence.

Methods

Search strategy

We searched Semantic Scholar, CrossRef, PubMed, Europe PMC, OpenAlex, CORE, and DOAJ from 2010 to 2026 using 17 search queries encompassing IL-23 inhibitor names, trial names (VOYAGE, UltIMMa, reSURFACE, IMMerge, IMMhance, GUIDE), comparator biologics, PASI response outcomes, and network meta-analysis terminology. A total of 1059 records were identified during discovery.

Inclusion criteria

Types of studies: Randomised controlled trials, systematic reviews, meta-analyses, network meta-analyses, and long-term extension studies.

Types of participants: Adults (≥18 years) with moderate-to-severe plaque psoriasis.

Types of interventions: At least one IL-23 inhibitor (guselkumab, risankizumab, or tildrakizumab) compared with placebo, active comparator biologics (TNF-alpha inhibitors, IL-17 inhibitors, IL-12/23 inhibitors), or small molecules.

Types of outcomes: PASI 75, PASI 90, PASI 100, IGA 0/1, DLQI scores, adverse event rates (serious adverse events, infections, major adverse cardiovascular events).

Exclusion criteria

Data collection and analysis

Effect sizes were pooled using the REML random-effects model. Heterogeneity was assessed using the I² statistic and Cochran's Q test. Publication bias was assessed using Egger's regression test where at least 10 studies were available. Sensitivity analyses included leave-one-out analysis, fixed-effect versus random-effects model comparison, and credibility-stratified pooling. Subgroup analyses by drug (guselkumab, risankizumab, tildrakizumab) and timepoint were pre-specified.

Assessment of certainty of evidence

The certainty of the evidence was assessed using the GRADE approach. Four domains were assessed mechanically: risk of bias (from study credibility tier distribution), inconsistency (from I² and Cochran's Q), imprecision (from confidence interval width and optimal information size), and publication bias (from Egger's test). Indirectness was assessed qualitatively by the review authors. Starting certainty was High for bodies of RCT evidence, with each domain potentially downgrading the certainty by one or two levels.

Results

Study selection

A total of 1059 records were identified through database searching. Of these, 97 passed initial screening as on-topic, 182 were flagged for review, and 780 were excluded as off-topic. After full-text assessment and data extraction, 195 papers contributed analysable data, 51 reported quantitative results suitable for meta-analysis, and 205 papers underwent credibility assessment.

PRISMA Flow: Paper Attrition Across Pipeline Papers identified n = 1,059 Passed relevance gate n = 279 Excluded: 780 Results extracted n = 276 Analysable (conf ≥ 0.3) n = 195 Low confidence: 81 Credibility assessed n = 205
PRISMA-style flow diagram of paper attrition.

Included studies

The evidence base comprised 52 randomised controlled trials, 38 network meta-analyses, 78 real-world observational studies, 12 long-term extension studies, 12 pairwise meta-analyses, 22 systematic reviews without meta-analysis, and 10 indirect treatment comparisons. Industry sponsorship was prevalent, with 42 studies relying on industry-sponsored RCT data.

Top 7 Methods Across the Corpus Real World Observational Study 78 Randomized Controlled Trial 52 Network Meta Analysis 38 Systematic Review Without Meta Analysis 22 Pairwise Meta Analysis 12 Long Term Extension Study 12 Indirect Treatment Comparison 10 0 25 50 75 100 Papers
Distribution of methodologies across the analysed papers.

Key phase 3 RCTs included the VOYAGE 1 and 2 trials (guselkumab, n = 1829) ("Guselkumab Efficacy by Psoriasis...", 2025), UltIMMa-1 and 2 (risankizumab, n = 997) (Al‐Janabi et al., 2019), IMMerge (risankizumab vs secukinumab, n = 327) (Warren et al., 2020a), IMMhance (risankizumab, n = 507) (Blauvelt et al., 2020), reSURFACE 1 and 2 (tildrakizumab, n = 1156) (Wytsma et al., 2024), and GUIDE (guselkumab dose-spacing, n = 822) (Eyerich et al., 2024). Long-term extension data extended to 304 weeks for risankizumab (LIMMitless, n = 897) (Papp et al., 2025a) and 252 weeks for guselkumab (VOYAGE extensions) ("Guselkumab Efficacy by Psoriasis...", 2025).

Risk of bias in included studies

D1: Randomisation D2: Deviations from interventions D3: Missing outcome data D4: Measurement of outcome D5: Selection of reported result Overall Sbidian et al. (2021) ? + + + + + Reich et al. (2020) + + ? ? Liu et al. (2025a) + ? ? ? Reich et al. (2019a) ? + + ? Guelimi et al. (2022) ? ? ? ? Shin et al. (2025) + + + ? ? Kim et al. (2023) ? ? ? Reich et al. (2019b) ? ? ? + ? ? Di Giulio et al. (2025) ? ? ? ? ? Abdullah & Shamoon (2025) + + + ? Griffiths et al. (2018a) + + ? ? Papp et al. (2017) ? + + ? ? Al‐Janabi et al. (2019) ? + ? + ? "FC12 A systematic rev... + ? ? ? ? Thaçi et al. (2021) + ? + ? ? Ismail et al. (2024) + + + ? Laws & Jardin (2023) + ? + ? "Guselkumab Efficacy b... ? + + ? Reich et al. (2022) ? ? ? Loft et al. (2019) + ? ? ? ? Youn et al. (2022) ? + + ? ? Gordon et al. (2022) + + + ? Reich et al. (2021) + + + ? Kolli et al. (2018) ? ? ? ? ? Phan et al. (2026) ? + ? ? ? Asadullah et al. (2025) + ? ? ? Trovato et al. (2026a) ? ? ? ? Carrascosa et al. (2023) ? ? ? ? Bagel et al. (2026) ? + ? ? ? Blauvelt et al. (2025) ? + + ? + Low risk ? Some concerns High risk
Risk of bias assessment across included studies, mapped to approximate RoB 2 domains from credibility assessment data. Green (+) = low risk, amber (?) = some concerns, red (−) = high risk.

Credibility Tier Distribution 199 papers High 1 (1%) Medium 114 (57%) Low 83 (42%) Uncertain 1 (1%)
Distribution of papers by credibility_assessment tier.

Only 1 study (0.5%) was rated high credibility; 114 (57%) were rated medium credibility; 83 (42%) were rated low credibility. Common methodological concerns included lack of preregistration, absent data and code availability, potential conflicts of interest, and selection bias. The mean overall credibility score was 0.44 (range 0–1).

Evidence Grounding: 4,852 of 6,976 claims verified against source text 70% 0% 25% 50% 75% 100%
Proportion of claims verified against source text.

The overwhelming predominance of medium- and low-credibility evidence — coupled with prevalent industry sponsorship and retrospective designs — means that all pooled estimates should be interpreted with caution. The direction of treatment effects is generally consistent, but their precise magnitude remains uncertain.

Effects of interventions

PASI 90 (proportion achieving ≥90% improvement)

Forty-three studies reported PASI 90 as a proportion. The pooled proportion was 0.72 (95% confidence interval 0.65 to 0.79; I² = 96.8%; τ² = 1.25; Q p < 0.001). Very low-certainty evidence suggests that approximately seven in ten patients treated with IL-23 inhibitors achieve PASI 90, though the prediction interval was wide (0.21 to 0.96), indicating substantial between-study variability.

PASI 90 (corpus-level) Study Estimate [95% CI] Weight Berry (2021) +0.75 [+0.28, +0.96] 1.4% Narcisi et al. (2023) +0.74 [+0.68, +0.79] 2.5% Blauvelt et al. (2020) +0.73 [+0.69, +0.77] 2.5% Alexis et al. (2025) +0.57 [+0.47, +0.66] 2.5% Papp et al. (2021) +0.74 [+0.65, +0.82] 2.5% Kim et al. (2023) +0.81 [+0.75, +0.86] 2.5% Pinter et al. (2023) +0.61 [+0.59, +0.64] 2.6% Campoli et al. (2024) +0.61 [+0.54, +0.68] 2.5% Thaçi et al. (2025) +0.71 [+0.69, +0.73] 2.6% Valenti et al. (2025) +0.81 [+0.78, +0.82] 2.6% Trovato et al. (2026) +0.90 [+0.85, +0.93] 2.5% Papp et al. (2025) +0.86 [+0.84, +0.88] 2.6% Amin et al. (2023) +0.85 [+0.82, +0.88] 2.5% Bagit et al. (2023) +0.70 [+0.52, +0.83] 2.3% Gordon et al. (2019) +0.86 [+0.80, +0.90] 2.5% Blauvelt et al. (2025) +0.94 [+0.56, +1.00] 1.1% Gordon et al. (2022) +0.37 [+0.33, +0.41] 2.6% Youn et al. (2022) +0.71 [+0.61, +0.80] 2.5% Griffiths et al. (2020) +0.84 [+0.82, +0.87] 2.6% Belcastro et al. (2023) +0.83 [+0.75, +0.89] 2.5% Caldarola et al. (2026) +0.28 [+0.21, +0.37] 2.5% Al‐Janabi et al. (2019) +0.75 [+0.70, +0.80] 2.5% Warren et al. (2020) +0.87 [+0.80, +0.91] 2.5% Ruggiero et al. (2021) +0.29 [+0.17, +0.47] 2.3% Mastorino et al. (2023) +0.56 [+0.53, +0.59] 2.6% Mastorino et al. (2024) +0.56 [+0.25, +0.83] 1.9% Carrascosa et al. (2021) +0.59 [+0.53, +0.65] 2.5% Bernardini et al. (2022) +0.05 [+0.02, +0.15] 2.0% Mastorino et al. (2025) +0.91 [+0.84, +0.96] 2.3% Hansel et al. (2020) +0.03 [+0.01, +0.07] 2.2% Carrascosa et al. (2023) +0.74 [+0.67, +0.80] 2.5% Asadullah et al. (2025) +0.65 [+0.59, +0.70] 2.5% Abu-Hilal et al. (2025) +0.70 [+0.58, +0.79] 2.4% Ruggiero et al. (2023) +0.76 [+0.61, +0.87] 2.3% Gönülal et al. (2023) +0.93 [+0.86, +0.96] 2.3% Fratton et al. (2025) +0.84 [+0.73, +0.91] 2.4% Griffiths et al. (2018) +0.89 [+0.85, +0.92] 2.5% Dubois (2024) +0.93 [+0.87, +0.96] 2.3% Megna et al. (2023) +0.83 [+0.75, +0.89] 2.5% Leo et al. (2024) +0.77 [+0.70, +0.83] 2.5% Licata et al. (2025) +0.86 [+0.70, +0.94] 2.2% - et al. (2024) +0.85 [+0.82, +0.88] 2.5% Ravasio et al. (2021) +0.02 [+0.00, +1.00] 0.1% Pooled (RE) +0.72 [+0.65, +0.79] I² = 96.8% τ² = 1.2538 Q = 1320.57 (df=42, p=0.000) High heterogeneity PI: +0.72 [+0.21, +0.96] ← Favours treatment Favours control →
Corpus-level forest plot for PASI 90. Pooled proportion = 0.72 [0.65, 0.79], I² = 96.8%.

Funnel plot: PASI 90 0.00 1.61 3.22 4.83 6.43 8.04 Effect estimate -16.62 -9.68 -2.74 4.19 11.13 18.07 Standard error High Medium Low Egger's test: p = 0.848 (no significant asymmetry)
Funnel plot for PASI 90 (proportion, 43 studies). Each point is one study; y-axis = precision (1/SE), x-axis = effect estimate. Dashed line = pooled estimate; shaded region = pseudo-95% confidence funnel. Egger's test: p = 0.848 (no significant asymmetry).

Subgroup analysis by drug (Q_between = 1309.3, p < 0.001):

IL-23 Inhibitor k Pooled PASI 90 (95% CI)
Guselkumab 10 0.68 (0.57 to 0.78)
Risankizumab 10 0.73 (0.65 to 0.79)
Tildrakizumab 3 0.55 (0.32 to 0.76)

PASI 90 (proportion) — by drug Subgroup (by drug) Pooled [95% CI] k I² guselkumab +0.68 [+0.57, +0.78] 10 0% risankizumab +0.73 [+0.65, +0.79] 10 0% tildrakizumab +0.55 [+0.32, +0.76] 3 62% Overall (RE) +0.72 [+0.65, +0.79] Test for subgroup differences: Q_between = 1309.34 (df=2, p=0.000) Significant subgroup differences
Subgroup forest plot for PASI 90 (proportion) stratified by drug. Each diamond = a subgroup-level pooled estimate; the bottom diamond is the overall pool. Q_between tests whether subgroups differ significantly.

An intra-class efficacy gradient was consistently observed. Risankizumab achieved the highest PASI 90 proportions, followed by guselkumab, with tildrakizumab consistently ranking lower. This pattern was replicated across NMAs (Pinter et al., 2025) and real-world cohorts (Ruggiero et al., 2021).

GRADE assessment for PASI 90 (proportion):

Domain Judgement Downgrade Rationale
Risk of bias Serious −1 47% of evidence weight from low-credibility studies; 0% from high-credibility studies
Inconsistency Very serious −2 I² = 96.8%, Q p < 0.001; wide prediction interval spanning 0.21 to 0.96
Imprecision No concern 0 CI does not cross clinically meaningful null threshold
Publication bias No concern 0 Egger's p = 0.848
Indirectness No concern 0 Studies directly addressed the population, intervention, and outcome of interest

Certainty: Very low (starting High, downgraded 3 levels)

Sensitivity analyses:
- Fixed-effect (0.70) vs random-effects (0.72): concordant
- Leave-one-out: most influential study was Hansel et al. (2020) (shift 0.020); conclusion robust
- Credibility-stratified: medium-credibility studies (0.79, k = 19) vs low-credibility (0.66, k = 20): concordant in direction

PASI 90 (comparative effect vs other biologics)

Eight studies provided comparative PASI 90 data amenable to standardised mean difference (SMD) analysis. The pooled SMD was −0.61 (95% CI −1.11 to −0.11; I² = 98.0%; τ² = 0.50; Q p < 0.001), favouring IL-23 inhibitors over comparators.

PASI 90 (corpus-level) Study Estimate [95% CI] Weight Hjort et al. (2024) -0.01 [-0.01, +0.00] 12.8% Shi et al. (2025) -2.46 [-2.91, -2.01] 11.6% doi_10.1016/j.jaad.2018… -0.48 [-0.62, -0.34] 12.7% Diels et al. (2019) -0.57 [-0.69, -0.44] 12.7% Lv et al. (2018) -0.43 [-0.76, -0.11] 12.1% Alzghool et al. (2026) -0.11 [-0.28, +0.05] 12.6% Abdullah et al. (2025) -0.95 [-1.13, -0.77] 12.6% Lee et al. (2026) -0.03 [-0.05, -0.00] 12.8% Pooled (RE) -0.61 [-1.11, -0.11] I² = 98.0% τ² = 0.5011 Q = 357.50 (df=7, p=0.000) High heterogeneity PI: -0.61 [-2.45, +1.23] ← Favours treatment Favours control →
Corpus-level forest plot for PASI 90. Pooled smd = -0.61 [-1.11, -0.11], I² = 98.0%.

Funnel plot: PASI 90 0.00 0.05 0.11 0.16 0.21 0.26 Effect estimate -2.58 -2.04 -1.50 -0.96 -0.42 0.12 Standard error High Medium Low Egger's test: p = 0.006 (significant asymmetry)
Funnel plot for PASI 90 (smd, 8 studies). Each point is one study; y-axis = precision (1/SE), x-axis = effect estimate. Dashed line = pooled estimate; shaded region = pseudo-95% confidence funnel. Egger's test: p = 0.006 (significant asymmetry).

GRADE assessment for PASI 90 (SMD):

Domain Judgement Downgrade Rationale
Risk of bias Serious −1 38% of evidence weight from low-credibility studies
Inconsistency Very serious −2 I² = 98.0%, extreme heterogeneity
Imprecision No concern 0 CI excludes null
Publication bias Serious −1 Egger's p = 0.006, indicating funnel plot asymmetry
Indirectness Serious −1 Mixed comparators (placebo, active controls); some studies used indirect comparison methods with inherent transitivity assumptions

Certainty: Very low (starting High, downgraded 5 levels, capped at Very low)

PASI 100 (complete skin clearance)

Twenty-nine studies reported PASI 100. The pooled proportion was 0.49 (95% CI 0.40 to 0.57; I² = 97.8%; τ² = 0.88; Q p < 0.001). Very low-certainty evidence indicates that approximately half of patients achieve complete clearance, though with marked variation.

PASI 100 (corpus-level) Study Estimate [95% CI] Weight Berry (2021) +0.36 [+0.20, +0.55] 3.2% Narcisi et al. (2023) +0.59 [+0.52, +0.65] 3.7% Alexis et al. (2025) +0.30 [+0.22, +0.39] 3.6% Papp et al. (2021) +0.54 [+0.44, +0.63] 3.6% Kim et al. (2023) +0.27 [+0.21, +0.33] 3.6% Pinter et al. (2023) +0.18 [+0.16, +0.20] 3.7% Campoli et al. (2024) +0.33 [+0.26, +0.40] 3.6% Thaçi et al. (2025) +0.57 [+0.55, +0.58] 3.7% Reguiai et al. (2026) +0.62 [+0.54, +0.69] 3.6% Trovato et al. (2026) +0.82 [+0.77, +0.86] 3.6% Papp et al. (2025) +0.54 [+0.51, +0.57] 3.7% doi_10.1016_j.ad.2024.0… +0.67 [+0.63, +0.71] 3.7% Amin et al. (2023) +0.52 [+0.49, +0.56] 3.7% Blauvelt et al. (2025) +0.11 [+0.02, +0.40] 2.0% Liu et al. (2025) +0.39 [+0.19, +0.62] 2.9% Belcastro et al. (2023) +0.63 [+0.55, +0.71] 3.6% Ruggiero et al. (2021) +0.09 [+0.03, +0.24] 2.6% Carrascosa et al. (2021) +0.49 [+0.43, +0.55] 3.7% Mastorino et al. (2025) +0.95 [+0.86, +0.98] 2.7% Hansel et al. (2020) +0.49 [+0.36, +0.62] 3.5% Abu-Hilal et al. (2025) +0.35 [+0.25, +0.46] 3.5% Ruggiero et al. (2023) +0.33 [+0.21, +0.49] 3.3% Gönülal et al. (2023) +0.83 [+0.76, +0.89] 3.5% Fratton et al. (2025) +0.71 [+0.59, +0.81] 3.5% Megna et al. (2023) +0.43 [+0.36, +0.50] 3.6% Leo et al. (2024) +0.24 [+0.18, +0.31] 3.6% Licata et al. (2025) +0.34 [+0.23, +0.48] 3.4% - et al. (2024) +0.76 [+0.72, +0.80] 3.7% Reich et al. (2022) +0.41 [+0.37, +0.45] 3.7% Pooled (RE) +0.49 [+0.40, +0.57] I² = 97.8% τ² = 0.8763 Q = 1270.71 (df=28, p=0.000) High heterogeneity PI: +0.49 [+0.12, +0.87] ← Favours treatment Favours control →
Corpus-level forest plot for PASI 100. Pooled proportion = 0.49 [0.40, 0.57], I² = 97.8%.

Funnel plot: PASI 100 0.00 0.20 0.40 0.59 0.79 0.99 Effect estimate -1.65 -0.79 0.06 0.91 1.77 2.62 Standard error High Medium Low Egger's test: p = 0.321 (no significant asymmetry)
Funnel plot for PASI 100 (proportion, 29 studies). Each point is one study; y-axis = precision (1/SE), x-axis = effect estimate. Dashed line = pooled estimate; shaded region = pseudo-95% confidence funnel. Egger's test: p = 0.321 (no significant asymmetry).

Subgroup analysis by drug (Q_between = 1264.3, p < 0.001):

IL-23 Inhibitor k Pooled PASI 100 (95% CI)
Guselkumab 5 0.42 (0.28 to 0.58)
Risankizumab 7 0.56 (0.50 to 0.63)

Tildrakizumab PASI 100, pooled from the tildrakizumab-specific claim group, was 0.34 (95% CI 0.28 to 0.42; I² = 0%, k = 3), notably lower than both guselkumab and risankizumab.

PASI 100 (proportion) — by drug Subgroup (by drug) Pooled [95% CI] k I² guselkumab +0.42 [+0.28, +0.58] 5 0% risankizumab +0.56 [+0.50, +0.63] 7 0% Overall (RE) +0.49 [+0.40, +0.57] Test for subgroup differences: Q_between = 1264.30 (df=1, p=0.000) Significant subgroup differences
Subgroup forest plot for PASI 100 (proportion) stratified by drug. Each diamond = a subgroup-level pooled estimate; the bottom diamond is the overall pool. Q_between tests whether subgroups differ significantly.

Subgroup analysis by timepoint:

Timepoint k Pooled PASI 100 (95% CI)
Week 16 6 0.40 (0.22 to 0.60)
Week 52 3 0.66 (0.51 to 0.78)

PASI 100 (proportion) — by timepoint Subgroup (by timepoint) Pooled [95% CI] k I² 16 +0.40 [+0.22, +0.60] 6 0% 52 +0.66 [+0.51, +0.78] 3 0% Overall (RE) +0.49 [+0.40, +0.57] Test for subgroup differences: Q_between = 1269.32 (df=1, p=0.000) Significant subgroup differences
Subgroup forest plot for PASI 100 (proportion) stratified by timepoint. Each diamond = a subgroup-level pooled estimate; the bottom diamond is the overall pool. Q_between tests whether subgroups differ significantly.

This temporal pattern supports the observation that IL-23 inhibitors demonstrate incrementally improving response rates over time, consistent with long-term extension data showing that risankizumab maintains PASI 90 at approximately 82–86% and PASI 100 at 54–60% through 3–6 years of continuous therapy (Papp et al., 2025a).

GRADE assessment for PASI 100:

Domain Judgement Downgrade Rationale
Risk of bias Very serious −2 66% of evidence weight from low-credibility studies
Inconsistency Very serious −2 I² = 97.8%; prediction interval 0.12 to 0.87
Imprecision No concern 0 CI excludes null
Publication bias No concern 0 Egger's p = 0.321
Indirectness No concern 0 Outcome and population directly match PICO

Certainty: Very low (starting High, downgraded 4 levels)

PASI 75

Twenty-one studies reported PASI 75. The pooled proportion was 0.82 (95% CI 0.64 to 0.92; I² = 97.9%; τ² = 4.57; Q p < 0.001).

PASI 75 (corpus-level) Study Estimate [95% CI] Weight Narcisi et al. (2023) +0.91 [+0.87, +0.94] 4.9% Papp et al. (2021) +0.92 [+0.84, +0.96] 4.8% Campoli et al. (2024) +0.88 [+0.83, +0.92] 4.9% Thaçi et al. (2025) +0.84 [+0.83, +0.86] 5.0% Bagit et al. (2023) +0.73 [+0.55, +0.85] 4.8% Blauvelt et al. (2025) +0.94 [+0.56, +1.00] 3.6% Youn et al. (2022) +0.94 [+0.87, +0.97] 4.8% Belcastro et al. (2023) +0.98 [+0.93, +0.99] 4.6% Yu et al. (2024) +0.64 [+0.58, +0.69] 5.0% Igarashi et al. (2021) +0.55 [+0.45, +0.64] 4.9% Thaçi et al. (2021) +0.93 [+0.88, +0.95] 4.9% Maul et al. (2021) +0.72 [+0.71, +0.73] 5.0% Bernardini et al. (2022) +0.12 [+0.06, +0.23] 4.8% Hansel et al. (2020) +0.00 [+0.00, +0.01] 4.9% Carrascosa et al. (2023) +0.93 [+0.87, +0.96] 4.9% Kolli et al. (2018) +0.33 [+0.19, +0.51] 4.8% Abu-Hilal et al. (2025) +0.96 [+0.88, +0.99] 4.7% Gönülal et al. (2023) +0.97 [+0.92, +0.99] 4.7% Dubois (2024) +0.93 [+0.87, +0.96] 4.8% Licata et al. (2025) +0.97 [+0.82, +1.00] 4.1% - et al. (2024) +0.95 [+0.93, +0.96] 4.9% Pooled (RE) +0.82 [+0.64, +0.92] I² = 97.9% τ² = 4.5694 Q = 966.14 (df=20, p=0.000) High heterogeneity PI: +0.82 [+0.04, +1.00] ← Favours treatment Favours control →
Corpus-level forest plot for PASI 75. Pooled proportion = 0.82 [0.64, 0.92], I² = 97.9%.

Funnel plot: PASI 75 0.00 0.30 0.60 0.91 1.21 1.51 Effect estimate -2.44 -1.13 0.17 1.47 2.77 4.07 Standard error High Medium Low Egger's test: p = 0.000 (significant asymmetry)
Funnel plot for PASI 75 (proportion, 21 studies). Each point is one study; y-axis = precision (1/SE), x-axis = effect estimate. Dashed line = pooled estimate; shaded region = pseudo-95% confidence funnel. Egger's test: p = 0.000 (significant asymmetry).

Subgroup by drug:

IL-23 Inhibitor k Pooled PASI 75 (95% CI)
Guselkumab 3 0.89 (0.10 to 1.00)
Risankizumab 3 0.84 (0.71 to 0.92)
Tildrakizumab 5 0.61 (0.41 to 0.78)

PASI 75 (proportion) — by drug Subgroup (by drug) Pooled [95% CI] k I² guselkumab +0.89 [+0.10, +1.00] 3 0% risankizumab +0.84 [+0.71, +0.92] 3 0% tildrakizumab +0.61 [+0.41, +0.78] 5 0% Overall (RE) +0.82 [+0.64, +0.92] Test for subgroup differences: Q_between = 964.67 (df=2, p=0.000) Significant subgroup differences
Subgroup forest plot for PASI 75 (proportion) stratified by drug. Each diamond = a subgroup-level pooled estimate; the bottom diamond is the overall pool. Q_between tests whether subgroups differ significantly.

PASI 75 (proportion) — by timepoint Subgroup (by timepoint) Pooled [95% CI] k I² 12 +0.62 [+0.40, +0.79] 3 0% 16 +0.73 [+0.06, +0.99] 4 0% Overall (RE) +0.82 [+0.64, +0.92] Test for subgroup differences: Q_between = 965.94 (df=1, p=0.000) Significant subgroup differences
Subgroup forest plot for PASI 75 (proportion) stratified by timepoint. Each diamond = a subgroup-level pooled estimate; the bottom diamond is the overall pool. Q_between tests whether subgroups differ significantly.

GRADE assessment for PASI 75:

Domain Judgement Downgrade Rationale
Risk of bias Serious −1 43% of evidence weight from low-credibility studies
Inconsistency Very serious −2 I² = 97.9%
Imprecision No concern 0 CI excludes null
Publication bias Serious −1 Egger's p < 0.001, significant asymmetry
Indirectness No concern 0 Directly relevant to PICO

Certainty: Very low (starting High, downgraded 4 levels)

IGA 0/1

Four studies reported IGA 0/1. The pooled proportion was 0.78 (95% CI 0.60 to 0.89; I² = 97.0%; τ² = 0.70; Q p < 0.001). Very low-certainty evidence suggests that approximately eight in ten patients achieve clear or almost clear investigator assessment.

IGA 0/1 (corpus-level) Study Estimate [95% CI] Weight Alexis et al. (2025) +0.74 [+0.65, +0.82] 25.2% Kim et al. (2023) +0.87 [+0.82, +0.91] 25.3% Youn et al. (2022) +0.91 [+0.83, +0.95] 22.9% Strober et al. (2025) +0.51 [+0.46, +0.55] 26.6% Pooled (RE) +0.78 [+0.60, +0.89] I² = 97.0% τ² = 0.7026 Q = 98.56 (df=3, p=0.000) High heterogeneity PI: +0.78 [+0.06, +1.00] ← Favours treatment Favours control →
Corpus-level forest plot for IGA 0/1. Pooled proportion = 0.78 [0.60, 0.89], I² = 97.0%.

Funnel plot: IGA 0/1 0.00 0.08 0.16 0.24 0.33 0.41 Effect estimate -0.10 0.25 0.61 0.96 1.31 1.66 Standard error High Medium Low Egger's test: p = 0.815 (no significant asymmetry)
Funnel plot for IGA 0/1 (proportion, 4 studies). Each point is one study; y-axis = precision (1/SE), x-axis = effect estimate. Dashed line = pooled estimate; shaded region = pseudo-95% confidence funnel. Egger's test: p = 0.815 (no significant asymmetry).

GRADE assessment for IGA 0/1:

Domain Judgement Downgrade Rationale
Risk of bias Serious −1 27% from low-credibility studies
Inconsistency Very serious −2 I² = 97.0%
Imprecision No concern 0 CI excludes null
Publication bias No concern 0 Egger's p = 0.815
Indirectness No concern 0 Directly relevant

Certainty: Very low (starting High, downgraded 3 levels)

sPGA 0/1

Five studies reported sPGA 0/1. The pooled proportion was 0.82 (95% CI 0.76 to 0.87; I² = 82.3%; τ² = 0.12; Q p < 0.001). Very low-certainty evidence supports that most patients achieve minimal disease by physician global assessment.

spga 0/1 (corpus-level) Study Estimate [95% CI] Weight Papp et al. (2021) +0.68 [+0.58, +0.76] 19.9% Thaçi et al. (2025) +0.83 [+0.81, +0.84] 27.2% Papp et al. (2025) +0.85 [+0.82, +0.87] 25.9% Amin et al. (2023) +0.86 [+0.83, +0.88] 25.3% Blauvelt et al. (2025) +0.94 [+0.56, +1.00] 1.8% Pooled (RE) +0.82 [+0.76, +0.87] I² = 82.3% τ² = 0.1170 Q = 22.57 (df=4, p=0.000) High heterogeneity PI: +0.82 [+0.57, +0.94] ← Favours treatment Favours control →
Corpus-level forest plot for spga 0/1. Pooled proportion = 0.82 [0.76, 0.87], I² = 82.3%.

Funnel plot: sPGA 0/1 0.00 0.30 0.60 0.91 1.21 1.51 Effect estimate -2.43 -1.13 0.17 1.47 2.77 4.07 Standard error High Medium Low Egger's test: p = 0.931 (no significant asymmetry)
Funnel plot for sPGA 0/1 (proportion, 5 studies). Each point is one study; y-axis = precision (1/SE), x-axis = effect estimate. Dashed line = pooled estimate; shaded region = pseudo-95% confidence funnel. Egger's test: p = 0.931 (no significant asymmetry).

GRADE assessment for sPGA 0/1:

Domain Judgement Downgrade Rationale
Risk of bias Very serious −2 52% from low-credibility studies
Inconsistency Very serious −2 I² = 82.3%
Imprecision No concern 0 CI excludes null
Publication bias No concern 0 Egger's p = 0.931
Indirectness No concern 0 Directly relevant

Certainty: Very low (starting High, downgraded 4 levels)

DLQI 0/1 (no or minimal quality-of-life impairment)

Nine studies reported DLQI 0/1. The pooled proportion was 0.69 (95% CI 0.63 to 0.74; I² = 96.8%; τ² = 0.13; Q p < 0.001).

"More than three-quarters of patients reported little to no impact on quality of life as measured by DLQI 0/1 through 304 weeks of treatment." (Papp et al., 2025a)

dlqi 0/1 (corpus-level) Study Estimate [95% CI] Weight Kim et al. (2023) +0.61 [+0.54, +0.68] 11.5% Thaçi et al. (2025) +0.70 [+0.68, +0.72] 13.2% Tsianakas et al. (2026) +0.51 [+0.49, +0.54] 13.1% Papp et al. (2025) +0.76 [+0.73, +0.79] 12.7% Amin et al. (2023) +0.76 [+0.73, +0.79] 12.6% Blauvelt et al. (2025) +0.89 [+0.60, +0.98] 2.0% Carrascosa et al. (2023) +0.70 [+0.63, +0.77] 11.0% Asadullah et al. (2025) +0.64 [+0.59, +0.70] 12.0% Lee et al. (2026) +0.72 [+0.67, +0.77] 12.0% Pooled (RE) +0.69 [+0.63, +0.74] I² = 96.8% τ² = 0.1293 Q = 247.37 (df=8, p=0.000) High heterogeneity PI: +0.69 [+0.47, +0.84] ← Favours treatment Favours control →
Corpus-level forest plot for dlqi 0/1. Pooled proportion = 0.69 [0.63, 0.74], I² = 96.8%.

Funnel plot: DLQI 0/1 0.00 0.20 0.40 0.59 0.79 0.99 Effect estimate -1.45 -0.59 0.26 1.11 1.97 2.82 Standard error High Medium Low Egger's test: p = 0.828 (no significant asymmetry)
Funnel plot for DLQI 0/1 (proportion, 9 studies). Each point is one study; y-axis = precision (1/SE), x-axis = effect estimate. Dashed line = pooled estimate; shaded region = pseudo-95% confidence funnel. Egger's test: p = 0.828 (no significant asymmetry).

Certainty: Very low (starting High, downgraded 4 levels: very serious risk of bias, very serious inconsistency)

Adverse events

Thirteen studies reported adverse event proportions. The pooled proportion was 0.09 (95% CI 0.04 to 0.18; I² = 98.8%; τ² = 2.15; Q p < 0.001), reflecting extreme heterogeneity in adverse event definitions and ascertainment across studies.

adverse events (corpus-level) Study Estimate [95% CI] Weight Berry (2021) +0.05 [+0.03, +0.10] 8.0% Reich et al. (2015) +0.01 [+0.01, +0.02] 8.3% Reguiai et al. (2026) +0.05 [+0.02, +0.10] 7.9% Papp et al. (2017) +0.08 [+0.03, +0.21] 7.3% - et al. (2024) +0.03 [+0.02, +0.05] 8.3% Izu-Belloso et al. (202… +0.06 [+0.03, +0.10] 8.1% Crowley et al. (2019) +0.56 [+0.50, +0.61] 8.4% Snast et al. (2026) +0.01 [+0.00, +0.05] 5.7% Bagit et al. (2023) +0.04 [+0.01, +0.24] 5.7% Liu et al. (2025) +0.05 [+0.02, +0.13] 7.6% Ruggiero et al. (2021) +0.29 [+0.17, +0.47] 8.0% Loft et al. (2019) +0.57 [+0.55, +0.59] 8.5% Fratton et al. (2025) +0.17 [+0.10, +0.28] 8.1% Pooled (RE) +0.09 [+0.04, +0.18] I² = 98.8% τ² = 2.1511 Q = 990.95 (df=12, p=0.000) High heterogeneity PI: +0.09 [+0.00, +0.73] ← Favours treatment Favours control →
Corpus-level forest plot for adverse events. Pooled proportion = 0.09 [0.04, 0.18], I² = 98.8%.

Funnel plot: adverse events 0.00 0.24 0.47 0.71 0.95 1.18 Effect estimate -2.46 -1.44 -0.42 0.60 1.62 2.64 Standard error High Medium Low Egger's test: p = 0.215 (no significant asymmetry)
Funnel plot for adverse events (proportion, 13 studies). Each point is one study; y-axis = precision (1/SE), x-axis = effect estimate. Dashed line = pooled estimate; shaded region = pseudo-95% confidence funnel. Egger's test: p = 0.215 (no significant asymmetry).

The most frequently reported treatment-emergent adverse events with risankizumab were nasopharyngitis (13.7 E/100 PY), upper respiratory tract infection (8.0 E/100 PY), and arthralgia (4.3 E/100 PY) (Amin et al., 2023). The standardised mortality ratio was 0.31 (95% CI 0.13 to 0.61) (Amin et al., 2023). Serious hypersensitivity reactions were uncommon and not considered drug-related (Amin et al., 2023).

infections (corpus-level) Study Estimate [95% CI] Weight Reich et al. (2015) +0.33 [+0.16, +0.55] 26.6% Dubois (2024) +0.02 [+0.00, +0.07] 24.0% Crowley et al. (2019) +0.35 [+0.30, +0.41] 28.9% Hansel et al. (2020) +0.02 [+0.00, +0.11] 20.5% Pooled (RE) +0.11 [+0.02, +0.37] I² = 90.9% τ² = 2.3559 Q = 32.83 (df=3, p=0.000) High heterogeneity PI: +0.11 [+0.00, +1.00] ← Favours treatment Favours control →
Corpus-level forest plot for infections. Pooled proportion = 0.11 [0.02, 0.37], I² = 90.9%.

Funnel plot: infections 0.00 0.23 0.46 0.69 0.92 1.15 Effect estimate -2.37 -1.38 -0.39 0.60 1.59 2.58 Standard error High Medium Low Egger's test: p = 0.335 (no significant asymmetry)
Funnel plot for infections (proportion, 4 studies). Each point is one study; y-axis = precision (1/SE), x-axis = effect estimate. Dashed line = pooled estimate; shaded region = pseudo-95% confidence funnel. Egger's test: p = 0.335 (no significant asymmetry).

Infection rates were pooled across 4 studies (proportion: 0.11, 95% CI 0.02 to 0.37; I² = 90.9%), with very low certainty.

Across four sequential Cochrane NMAs, no significant difference in serious adverse events was found between biologic interventions and placebo (Al‐Janabi et al., 2019). IL-17 inhibitors showed higher rates of candidiasis (12% vs 3% for other classes) (Dubois, 2024), while IL-23 inhibitors had lower discontinuation rates attributable to adverse events than TNF-alpha inhibitors and IL-17 inhibitors.

Certainty: Very low (starting High, downgraded 4 levels: very serious risk of bias, very serious inconsistency)

Drug survival

Seven studies reported drug survival for IL-23 inhibitors. The pooled proportion was 0.92 (95% CI 0.88 to 0.95; I² = 86.0%; τ² = 0.34; Q p < 0.001), with a prediction interval of 0.70 to 0.98.

drug survival (corpus-level) Study Estimate [95% CI] Weight Reguiai et al. (2026) +0.95 [+0.89, +0.97] 14.8% Belcastro et al. (2023) +0.96 [+0.91, +0.98] 12.7% Mastorino et al. (2024) +0.75 [+0.22, +0.97] 3.9% Warren et al. (2020) +0.94 [+0.85, +0.97] 12.2% Mortato et al. (2025) +0.96 [+0.94, +0.97] 19.3% Balato et al. (2023) +0.92 [+0.90, +0.93] 19.9% Hu et al. (2026) +0.79 [+0.69, +0.86] 17.3% Pooled (RE) +0.92 [+0.88, +0.95] I² = 86.0% τ² = 0.3443 Q = 42.84 (df=6, p=0.000) High heterogeneity PI: +0.92 [+0.70, +0.98] ← Favours treatment Favours control →
Corpus-level forest plot for drug survival. Pooled proportion = 0.92 [0.88, 0.95], I² = 86.0%.

Funnel plot: drug survival 0.00 0.28 0.56 0.84 1.12 1.40 Effect estimate -2.09 -0.88 0.32 1.53 2.73 3.94 Standard error High Medium Low Egger's test: p = 0.274 (no significant asymmetry)
Funnel plot for drug survival (proportion, 7 studies). Each point is one study; y-axis = precision (1/SE), x-axis = effect estimate. Dashed line = pooled estimate; shaded region = pseudo-95% confidence funnel. Egger's test: p = 0.274 (no significant asymmetry).

Drug survival was notably high; approximately 86% of risankizumab-treated patients remained on treatment after 5 years (Mortato et al., 2025), and guselkumab demonstrated lower discontinuation rates than other study biologics (Tskhvarashvili et al., 2025).

Certainty: Very low (starting High, downgraded 4 levels: very serious risk of bias, very serious inconsistency)

Treatment discontinuation

treatment discontinuation (corpus-level) Study Estimate [95% CI] Weight Izu-Belloso et al. (202… +0.19 [+0.15, +0.25] 35.6% Ruggiero et al. (2021) +0.12 [+0.04, +0.27] 22.0% Fratton et al. (2025) +0.09 [+0.04, +0.19] 26.0% Mastorino et al. (2025) +0.02 [+0.01, +0.08] 16.4% Pooled (RE) +0.10 [+0.05, +0.19] I² = 77.9% τ² = 0.3824 Q = 13.60 (df=3, p=0.004) High heterogeneity PI: +0.10 [+0.00, +0.72] ← Favours treatment Favours control →
Corpus-level forest plot for treatment discontinuation. Pooled proportion = 0.10 [0.05, 0.19], I² = 77.9%.

Funnel plot: treatment discontinuation 0.00 0.16 0.33 0.49 0.66 0.82 Effect estimate -1.67 -0.96 -0.25 0.46 1.16 1.87 Standard error High Medium Low Egger's test: p = 0.368 (no significant asymmetry)
Funnel plot for treatment discontinuation (proportion, 4 studies). Each point is one study; y-axis = precision (1/SE), x-axis = effect estimate. Dashed line = pooled estimate; shaded region = pseudo-95% confidence funnel. Egger's test: p = 0.368 (no significant asymmetry).

Four studies reported treatment discontinuation (pooled proportion: 0.10, 95% CI 0.05 to 0.19; I² = 77.9%).

Additional outcomes

iga 0 (corpus-level) Study Estimate [95% CI] Weight Alexis et al. (2025) +0.33 [+0.24, +0.42] 32.3% Kim et al. (2023) +0.27 [+0.21, +0.33] 35.1% Youn et al. (2022) +0.52 [+0.42, +0.62] 32.6% Pooled (RE) +0.36 [+0.25, +0.49] I² = 89.1% τ² = 0.1829 Q = 18.33 (df=2, p=0.000) High heterogeneity PI: +0.36 [+0.00, +1.00] ← Favours treatment Favours control →
Corpus-level forest plot for iga 0. Pooled proportion = 0.36 [0.25, 0.49], I² = 89.1%.

Funnel plot: IGA 0 0.00 0.05 0.10 0.15 0.19 0.24 Effect estimate -0.16 0.05 0.26 0.47 0.68 0.89 Standard error High Medium Low Egger's test: p = 0.351 (no significant asymmetry)
Funnel plot for IGA 0 (proportion, 3 studies). Each point is one study; y-axis = precision (1/SE), x-axis = effect estimate. Dashed line = pooled estimate; shaded region = pseudo-95% confidence funnel. Egger's test: p = 0.351 (no significant asymmetry).

IGA 0 (complete clearance) was reported in 3 studies (pooled: 0.36, 95% CI 0.25 to 0.49; I² = 89.1%; very low certainty).

pasi <3 (corpus-level) Study Estimate [95% CI] Weight Eyerich et al. (2024) +0.83 [+0.79, +0.86] 34.0% Tsianakas et al. (2026) +0.68 [+0.65, +0.70] 34.2% Bernardini et al. (2022) +0.10 [+0.05, +0.21] 31.8% Pooled (RE) +0.52 [+0.16, +0.87] I² = 97.9% τ² = 2.4057 Q = 93.28 (df=2, p=0.000) High heterogeneity PI: +0.52 [+0.00, +1.00] ← Favours treatment Favours control →
Corpus-level forest plot for pasi <3. Pooled proportion = 0.52 [0.16, 0.87], I² = 97.9%.

pga 0/1 (corpus-level) Study Estimate [95% CI] Weight Tsianakas et al. (2026) +0.64 [+0.61, +0.66] 36.0% Igarashi et al. (2021) +0.55 [+0.45, +0.64] 34.5% Abu-Hilal et al. (2025) +0.93 [+0.85, +0.97] 29.4% Pooled (RE) +0.74 [+0.48, +0.89] I² = 91.4% τ² = 0.8946 Q = 23.35 (df=2, p=0.000) High heterogeneity PI: +0.74 [+0.00, +1.00] ← Favours treatment Favours control →
Corpus-level forest plot for pga 0/1. Pooled proportion = 0.74 [0.48, 0.89], I² = 91.4%.

pasi ≤ 3 (corpus-level) Study Estimate [95% CI] Weight Liu et al. (2025) +0.89 [+0.80, +0.94] 28.0% Mastorino et al. (2023) +0.70 [+0.67, +0.73] 38.8% Carrascosa et al. (2023) +0.88 [+0.83, +0.92] 33.2% Pooled (RE) +0.83 [+0.71, +0.91] I² = 93.8% τ² = 0.3059 Q = 32.51 (df=2, p=0.000) High heterogeneity PI: +0.83 [+0.00, +1.00] ← Favours treatment Favours control →
Corpus-level forest plot for pasi ≤ 3. Pooled proportion = 0.83 [0.71, 0.91], I² = 93.8%.

pasi (corpus-level) Study Estimate [95% CI] Weight Gottlieb et al. (2026) +1.62 [+1.03, +2.21] 12.9% Liu et al. (2025) -8.30 [-13.24, -3.36] 9.8% Phan et al. (2026) +4.70 [+4.00, +5.40] 12.9% Carrascosa et al. (2021) -9.30 [-10.20, -8.40] 12.8% Ak et al. (2024) +0.30 [+0.12, +0.48] 13.0% Ak et al. (2025) +1.40 [+0.57, +2.23] 12.8% Megna et al. (2023) +0.70 [+0.35, +1.05] 12.9% Licata et al. (2025) +1.20 [+0.60, +1.80] 12.9% Pooled (RE) -0.72 [-3.84, +2.39] I² = 98.9% τ² = 19.4711 Q = 635.39 (df=7, p=0.000) High heterogeneity PI: -0.72 [-12.20, +10.75] ← Favours treatment Favours control →
Corpus-level forest plot for pasi. Pooled mean_difference = -0.72 [-3.84, 2.39], I² = 98.9%.

pasi (corpus-level) Study Estimate [95% CI] Weight Papp et al. (2017) +0.77 [+0.69, +0.84] 20.4% Bernardini et al. (2022) +0.80 [+0.68, +0.88] 19.7% Blauvelt et al. (2019) +0.14 [+0.12, +0.17] 20.8% Lazar et al. (2026) +0.49 [+0.38, +0.60] 20.3% doi_10.36849_jdd.8217 +0.87 [+0.74, +0.94] 18.8% Pooled (RE) +0.62 [+0.34, +0.84] I² = 98.4% τ² = 1.7100 Q = 246.84 (df=4, p=0.000) High heterogeneity PI: +0.62 [+0.02, +0.99] ← Favours treatment Favours control →
Corpus-level forest plot for pasi. Pooled proportion = 0.62 [0.34, 0.84], I² = 98.4%.

dlqi (corpus-level) Study Estimate [95% CI] Weight Gottlieb et al. (2026) +1.83 [+0.84, +2.82] 13.5% Bhatia et al. (2024) -7.40 [-11.81, -2.99] 10.3% Liu et al. (2025) -6.80 [-10.85, -2.75] 10.7% Phan et al. (2026) +5.60 [+4.20, +7.00] 13.3% Gracia-Cazaña et al. (2… -0.80 [-1.28, -0.32] 13.7% Ruiter et al. (2022) +5.00 [+1.20, +8.80] 11.0% Ak et al. (2025) +0.50 [+0.00, +1.00] 13.7% Licata et al. (2025) +0.90 [+0.45, +1.35] 13.7% Pooled (RE) +0.14 [-2.67, +2.94] I² = 94.1% τ² = 14.8763 Q = 119.65 (df=7, p=0.000) High heterogeneity PI: +0.14 [-9.93, +10.20] ← Favours treatment Favours control →
Corpus-level forest plot for dlqi. Pooled mean_difference = 0.14 [-2.67, 2.94], I² = 94.1%.

dlqi (corpus-level) Study Estimate [95% CI] Weight Snast et al. (2026) +0.05 [+0.02, +0.10] 33.1% Gracia-Cazaña et al. (2… +0.59 [+0.56, +0.62] 33.6% Cuniberti et al. (2026) +0.00 [+0.00, +0.00] 33.3% Pooled (RE) +0.04 [+0.00, +0.56] I² = 99.7% τ² = 8.8171 Q = 599.45 (df=2, p=0.000) High heterogeneity PI: +0.04 [+0.00, +1.00] ← Favours treatment Favours control →
Corpus-level forest plot for dlqi. Pooled proportion = 0.04 [0.00, 0.56], I² = 99.7%.

BSA (corpus-level) Study Estimate [95% CI] Weight Gottlieb et al. (2026) +3.27 [+1.26, +5.28] 22.9% Ak et al. (2025) +0.90 [+0.15, +1.65] 28.4% Megna et al. (2023) +1.90 [+0.94, +2.86] 27.7% Lee et al. (2026) -3.10 [-5.48, -0.72] 21.0% Pooled (RE) +0.88 [-1.16, +2.92] I² = 84.4% τ² = 3.6576 Q = 19.24 (df=3, p=0.000) High heterogeneity PI: +0.88 [-8.48, +10.24] ← Favours treatment Favours control →
Corpus-level forest plot for BSA. Pooled mean_difference = 0.88 [-1.16, 2.92], I² = 84.4%.

PASI 90 (corpus-level) Study Estimate [95% CI] Weight Sbidian et al. (2020) +23.97 [+23.79, +24.15] 16.7% Sbidian et al. (2021) +25.79 [+25.61, +25.97] 16.7% Sbidian et al. (2022) +30.27 [+30.10, +30.44] 16.7% Sbidian et al. (2023) +27.35 [+27.18, +27.52] 16.7% Ismail et al. (2024) +37.81 [+37.53, +38.09] 16.7% Erichsen et al. (2019) +1.73 [+1.54, +1.95] 16.6% Pooled (RE) +17.98 [+7.73, +41.78] I² = 100.0% τ² = 1.1102 Q = 11195.69 (df=5, p=0.000) High heterogeneity PI: +17.98 [+0.76, +423.68] ← Favours treatment Favours control →
Corpus-level forest plot for PASI 90. Pooled risk_ratio = 17.98 [7.73, 41.78], I² = 100.0%.

Funnel plot: PASI 90 0.00 0.01 0.03 0.04 0.06 0.07 Effect estimate -0.07 7.86 15.80 23.74 31.68 39.61 Standard error High Medium Low Egger's test: p = 0.954 (no significant asymmetry)
Funnel plot for PASI 90 (risk_ratio, 6 studies). Each point is one study; y-axis = precision (1/SE), x-axis = effect estimate. Dashed line = pooled estimate; shaded region = pseudo-95% confidence funnel. Egger's test: p = 0.954 (no significant asymmetry).

PASI 100 (corpus-level) Study Estimate [95% CI] Weight doi_10.1016/j.jaad.2018… -0.51 [-0.64, -0.38] 36.7% Hu et al. (2026) -1.00 [-1.75, -0.24] 27.7% Barcelos et al. (2021) -2.07 [-2.33, -1.81] 35.6% Pooled (RE) -1.20 [-1.99, -0.41] I² = 98.2% τ² = 0.4414 Q = 111.28 (df=2, p=0.000) High heterogeneity PI: -1.20 [-11.08, +8.68] ← Favours treatment Favours control →
Corpus-level forest plot for PASI 100. Pooled smd = -1.20 [-1.99, -0.41], I² = 98.2%.

Funnel plot: PASI 100 0.00 0.09 0.18 0.27 0.36 0.44 Effect estimate -2.16 -1.78 -1.39 -1.01 -0.63 -0.25 Standard error High Medium Low Egger's test: p = 0.652 (no significant asymmetry)
Funnel plot for PASI 100 (smd, 3 studies). Each point is one study; y-axis = precision (1/SE), x-axis = effect estimate. Dashed line = pooled estimate; shaded region = pseudo-95% confidence funnel. Egger's test: p = 0.652 (no significant asymmetry).

PASI 75 (corpus-level) Study Estimate [95% CI] Weight Shi et al. (2025) -2.55 [-3.14, -1.95] 23.5% doi_10.1016/j.jaad.2018… -0.42 [-0.60, -0.24] 25.5% Lv et al. (2018) -2.07 [-2.31, -1.83] 25.4% Alzghool et al. (2026) -0.17 [-0.33, -0.01] 25.6% Pooled (RE) -1.28 [-2.27, -0.28] I² = 98.6% τ² = 1.0024 Q = 212.34 (df=3, p=0.000) High heterogeneity PI: -1.28 [-6.11, +3.56] ← Favours treatment Favours control →
Corpus-level forest plot for PASI 75. Pooled smd = -1.28 [-2.27, -0.28], I² = 98.6%.

Funnel plot: PASI 75 0.00 0.07 0.14 0.21 0.28 0.35 Effect estimate -2.66 -2.14 -1.62 -1.10 -0.57 -0.05 Standard error High Medium Low Egger's test: p = 0.276 (no significant asymmetry)
Funnel plot for PASI 75 (smd, 4 studies). Each point is one study; y-axis = precision (1/SE), x-axis = effect estimate. Dashed line = pooled estimate; shaded region = pseudo-95% confidence funnel. Egger's test: p = 0.276 (no significant asymmetry).

Funnel plot: PASI 0.00 0.58 1.16 1.74 2.32 2.90 Effect estimate -10.01 -6.88 -3.74 -0.60 2.54 5.67 Standard error High Medium Low Egger's test: p = 0.799 (no significant asymmetry)
Funnel plot for PASI (mean_difference, 8 studies). Each point is one study; y-axis = precision (1/SE), x-axis = effect estimate. Dashed line = pooled estimate; shaded region = pseudo-95% confidence funnel. Egger's test: p = 0.799 (no significant asymmetry).

Funnel plot: PASI 0.00 0.10 0.20 0.30 0.40 0.50 Effect estimate -0.47 -0.03 0.40 0.84 1.27 1.71 Standard error High Medium Low Egger's test: p = 0.029 (significant asymmetry)
Funnel plot for PASI (proportion, 5 studies). Each point is one study; y-axis = precision (1/SE), x-axis = effect estimate. Dashed line = pooled estimate; shaded region = pseudo-95% confidence funnel. Egger's test: p = 0.029 (significant asymmetry).

Funnel plot: PASI <3 0.00 0.10 0.20 0.30 0.40 0.49 Effect estimate -0.54 -0.12 0.31 0.74 1.16 1.59 Standard error High Medium Low Egger's test: p = 0.494 (no significant asymmetry)
Funnel plot for PASI <3 (proportion, 3 studies). Each point is one study; y-axis = precision (1/SE), x-axis = effect estimate. Dashed line = pooled estimate; shaded region = pseudo-95% confidence funnel. Egger's test: p = 0.494 (no significant asymmetry).

Funnel plot: DLQI 0.00 0.52 1.03 1.55 2.07 2.59 Effect estimate -8.05 -5.19 -2.33 0.53 3.39 6.25 Standard error High Medium Low Egger's test: p = 0.815 (no significant asymmetry)
Funnel plot for DLQI (mean_difference, 8 studies). Each point is one study; y-axis = precision (1/SE), x-axis = effect estimate. Dashed line = pooled estimate; shaded region = pseudo-95% confidence funnel. Egger's test: p = 0.815 (no significant asymmetry).

Funnel plot: DLQI 0.00 0.09 0.18 0.27 0.36 0.45 Effect estimate -0.93 -0.54 -0.15 0.23 0.62 1.01 Standard error High Medium Low Egger's test: p = 0.454 (no significant asymmetry)
Funnel plot for DLQI (proportion, 3 studies). Each point is one study; y-axis = precision (1/SE), x-axis = effect estimate. Dashed line = pooled estimate; shaded region = pseudo-95% confidence funnel. Egger's test: p = 0.454 (no significant asymmetry).

Funnel plot: BSA 0.00 0.28 0.56 0.84 1.12 1.40 Effect estimate -3.44 -1.96 -0.48 1.00 2.48 3.96 Standard error High Medium Low Egger's test: p = 0.770 (no significant asymmetry)
Funnel plot for BSA (mean_difference, 4 studies). Each point is one study; y-axis = precision (1/SE), x-axis = effect estimate. Dashed line = pooled estimate; shaded region = pseudo-95% confidence funnel. Egger's test: p = 0.770 (no significant asymmetry).

Funnel plot: PGA 0/1 0.00 0.10 0.21 0.31 0.42 0.52 Effect estimate -0.39 0.06 0.51 0.96 1.41 1.86 Standard error High Medium Low Egger's test: p = 0.751 (no significant asymmetry)
Funnel plot for PGA 0/1 (proportion, 3 studies). Each point is one study; y-axis = precision (1/SE), x-axis = effect estimate. Dashed line = pooled estimate; shaded region = pseudo-95% confidence funnel. Egger's test: p = 0.751 (no significant asymmetry).

Funnel plot: PASI ≤ 3 0.00 0.08 0.16 0.24 0.32 0.40 Effect estimate -0.04 0.31 0.65 1.00 1.35 1.70 Standard error High Medium Low Egger's test: p = 0.266 (no significant asymmetry)
Funnel plot for PASI ≤ 3 (proportion, 3 studies). Each point is one study; y-axis = precision (1/SE), x-axis = effect estimate. Dashed line = pooled estimate; shaded region = pseudo-95% confidence funnel. Egger's test: p = 0.266 (no significant asymmetry).

adverse events leading to discontinuation — g013 Study Estimate [95% CI] Weight Reguiai et al. (2026) +0.04 [+0.02, +0.09] 31.6% Izu-Belloso et al. (202… +0.19 [+0.15, +0.25] 41.8% Ruggiero et al. (2021) +0.12 [+0.04, +0.27] 26.6% Pooled (RE) +0.11 [+0.05, +0.22] I² = 86.2% τ² = 0.4127 Q = 14.47 (df=2, p=0.001) High heterogeneity PI: +0.11 [+0.00, +1.00] ← Favours treatment Favours control →
Forest plot for adverse events leading to discontinuation. Pooled proportion = 0.11 [0.05, 0.22], I² = 86.2%.

sPGA 0 — g022 Study Estimate [95% CI] Weight Blauvelt et al. (2025) +0.94 [+0.56, +1.00] 5.8% Papp et al. (2021) +0.68 [+0.58, +0.76] 43.4% Papp et al. (2025) +0.85 [+0.82, +0.87] 50.8% Pooled (RE) +0.79 [+0.67, +0.88] I² = 89.1% τ² = 0.2124 Q = 18.33 (df=2, p=0.000) High heterogeneity PI: +0.79 [+0.00, +1.00] ← Favours treatment Favours control →
Forest plot for sPGA 0. Pooled proportion = 0.79 [0.67, 0.88], I² = 89.1%.

sGPA 0/1 — g102 Study Estimate [95% CI] Weight Blauvelt et al. (2025) +0.94 [+0.56, +1.00] 0.3% Thaçi et al. (2025) +0.83 [+0.81, +0.84] 66.8% Amin et al. (2023) +0.86 [+0.83, +0.88] 32.9% Pooled (RE) +0.84 [+0.82, +0.86] I² = 59.6% τ² = 0.0061 Q = 4.95 (df=2, p=0.084) Moderate heterogeneity PI: +0.84 [+0.56, +0.95] ← Favours treatment Favours control →
Forest plot for sGPA 0/1. Pooled proportion = 0.84 [0.82, 0.86], I² = 59.6%.

treatment_discontinuation — g149 Study Estimate [95% CI] Weight Fratton et al. (2025) +0.09 [+0.04, +0.19] 49.9% Ruggiero et al. (2021) +0.12 [+0.04, +0.27] 32.1% Mastorino et al. (2025) +0.02 [+0.01, +0.08] 17.9% Pooled (RE) +0.08 [+0.04, +0.13] I² = 55.3% τ² = 0.0000 Q = 4.47 (df=2, p=0.107) Moderate heterogeneity PI: +0.08 [+0.00, +0.80] ← Favours treatment Favours control →
Forest plot for treatment_discontinuation. Pooled proportion = 0.08 [0.04, 0.13], I² = 55.3%.

Outcomes without quantitative pooling

Serious adverse events (SAEs): Four successive Cochrane NMAs consistently found no significant difference in SAE rates between biologic interventions and placebo (Al‐Janabi et al., 2019). Quantitative pooling of SAE proportions was not performed due to heterogeneous reporting definitions across included primary studies.

Major adverse cardiovascular events (MACE): No statistically significant difference in MACE risk was identified between biologic therapies and placebo, or between biologic classes. This finding was supported by dedicated systematic reviews and meta-analyses within the corpus.

Candidiasis: IL-17 inhibitors were associated with higher rates of candidiasis (12% vs 3% for other classes) (Dubois, 2024). This outcome was not pooled quantitatively due to inconsistent reporting across studies.

Dose-spacing outcomes: Guselkumab super-responders maintained disease control with extended every-16-week dosing, with only small differences from every-8-week dosing (mean PASI 0.1 vs 0.4) (Eyerich et al., 2024). Both risankizumab and guselkumab demonstrated feasibility of dose de-escalation in real-world settings, with drug survival exceeding 89% on de-escalated regimens (Mastorino et al., 2025).

Discussion

Summary of main results

This systematic review synthesised evidence from 195 papers across multiple study designs evaluating three IL-23 inhibitors for moderate-to-severe plaque psoriasis. All three IL-23 inhibitors — guselkumab, risankizumab, and tildrakizumab — demonstrated clinically meaningful efficacy across PASI response thresholds, IGA endpoints, and quality-of-life measures, with favourable safety profiles relative to other biologic classes.

A consistent intra-class efficacy gradient was observed: risankizumab achieved the highest PASI 90 and PASI 100 response rates, followed by guselkumab, with tildrakizumab ranking lowest for high-hurdle endpoints. Tildrakizumab PASI 100 was pooled at 0.34 (95% CI 0.28 to 0.42; I² = 0%, k = 3), substantially below risankizumab estimates of approximately 0.56 (Balato et al., 2023). However, tildrakizumab may offer a uniquely favourable safety profile that partially offsets this differential (Imafuku et al., 2021; Bagel et al., 2026).

When compared against other biologic classes, evidence from multiple NMAs indicated that IL-17 inhibitors (ixekizumab, brodalumab) and bimekizumab achieved the fastest onset and highest peak PASI 90 rates at weeks 12–16, whereas IL-23 inhibitors — particularly risankizumab — demonstrated superior durability over 1–5 years (Warren et al., 2020a; Dubois, 2024). A five-year registry analysis found that 80% of IL-23 inhibitor-treated patients maintained PASI 90 at five years compared with 65% for IL-17 inhibitors and 38% for TNF inhibitors (Dubois, 2024).

"Risankizumab 150 mg every 12 weeks demonstrated a favorable safety profile and was well tolerated through 6 years of continuous therapy." (Papp et al., 2025a)

Safety profiles were reassuring across all IL-23 inhibitors. SAE rates were not significantly different from placebo across NMAs. Drug survival at 1–2 years was pooled at 92%, higher than historical data for TNF-alpha inhibitors. Infliximab and adalimumab showed higher serious infection risk compared with risankizumab and secukinumab.

Agreements and disagreements with other reviews

The findings are broadly consistent with the Cochrane series of NMAs on systemic treatments for plaque psoriasis, which identified bimekizumab, ixekizumab, and risankizumab as offering the best compromise between PASI 90 efficacy and safety acceptability (Warren et al., 2020a). The present analysis extends these findings by incorporating real-world evidence confirming the durability advantage of IL-23 inhibitors and the within-class gradient.

A key area of disagreement relates to the relative ranking of guselkumab versus IL-17 inhibitors. Some NMAs placed guselkumab below ixekizumab and brodalumab for short-term PASI 90, while PSoHO observational data showed comparable effectiveness at 12 months with guselkumab demonstrating better durability outcomes at 24 months (Pinter et al., 2023). This discrepancy likely reflects differences in timepoint selection and patient populations studied.

Strengths and limitations

Strengths include the breadth of the evidence base (195 analysable papers, encompassing RCTs, NMAs, long-term extensions, and real-world registries), pre-specified subgroup analyses by drug and timepoint, and multiple sensitivity analyses confirming robustness of conclusions.

Limitations are substantial. First, heterogeneity was very high (I² > 80%) for all pooled outcomes, reflecting differences in study populations, timepoints, outcome definitions, and geographic settings. Second, the majority of evidence derived from low- or medium-credibility studies; only one study achieved high-credibility rating. Industry sponsorship was prevalent, introducing potential reporting and publication bias. Third, very few direct head-to-head trials compared IL-23 inhibitors against one another (the IMMerge trial comparing risankizumab vs secukinumab being a notable exception) (Warren et al., 2020a), and NMA estimates rely on transitivity assumptions that may not hold across heterogeneous trial populations. Fourth, adverse event definitions and ascertainment varied markedly, precluding reliable quantitative safety comparisons. Finally, the search was limited to indexed databases and may have missed grey literature.

Implications for practice

Given that all evidence was rated very low certainty, treatment recommendations must be made cautiously. Nevertheless, the consistent direction and large magnitude of effects support the following observations:

Implications for research

Future research priorities include:

  1. Direct head-to-head RCTs between IL-23 inhibitors (guselkumab vs risankizumab vs tildrakizumab) to resolve the intra-class efficacy hierarchy with high certainty.
  2. Longer-term comparative data (≥5 years) across biologic classes, incorporating both efficacy durability and cumulative safety endpoints.
  3. Standardised adverse event reporting to enable meaningful quantitative safety comparisons.
  4. Under-represented populations: Greater inclusion of patients with diverse ethnic backgrounds, elderly patients, and those with significant comorbidities (Augustin et al., 2020; Kim et al., 2023).
  5. Biomarker-guided treatment selection to identify patients likely to achieve super-response or those at risk of treatment failure.
  6. Pragmatic dose-optimisation trials to establish evidence-based protocols for dose de-escalation in stable responders.

Authors' conclusions

Very low-certainty evidence suggests that all three IL-23 inhibitors are effective for moderate-to-severe plaque psoriasis, with a consistent intra-class gradient favouring risankizumab over guselkumab and tildrakizumab for high-hurdle efficacy endpoints. Compared with other biologic classes, IL-23 inhibitors appear to offer superior long-term durability and favourable safety profiles, while IL-17 inhibitors achieve faster onset and competitive short-term efficacy. All pooled estimates are limited by very high heterogeneity and predominantly low- to medium-credibility evidence; future direct comparative trials are needed to increase certainty and guide personalised treatment decisions.

Figures

Leave-one-out sensitivity: PASI 90 (proportion). Robust to single-study exclusion.
PASI 90 (proportion) — leave-one-out Dropped study Recomputed pool [95% CI] Berry (2021) +0.72 [+0.65, +0.79] 97% Narcisi et al. (2023) +0.72 [+0.65, +0.79] 97% Blauvelt et al. (2020) +0.72 [+0.65, +0.79] 97% Alexis et al. (2025) +0.73 [+0.65, +0.79] 97% Papp et al. (2021) +0.72 [+0.65, +0.79] 97% Kim et al. (2023) +0.72 [+0.64, +0.79] 97% Pinter et al. (2023) +0.73 [+0.65, +0.79] 97% Campoli et al. (2024) +0.73 [+0.65, +0.79] 97% Thaçi et al. (2025) +0.73 [+0.65, +0.79] 97% Valenti et al. (2025) +0.72 [+0.64, +0.79] 97% Trovato et al. (2026) +0.72 [+0.64, +0.79] 97% Papp et al. (2025) +0.72 [+0.64, +0.79] 97% Amin et al. (2023) +0.72 [+0.64, +0.79] 97% Bagit et al. (2023) +0.73 [+0.65, +0.79] 97% Gordon et al. (2019) +0.72 [+0.64, +0.79] 97% Blauvelt et al. (2025) +0.72 [+0.64, +0.79] 97% Gordon et al. (2022) +0.73 [+0.66, +0.80] 96% Youn et al. (2022) +0.73 [+0.65, +0.79] 97% Griffiths et al. (2020) +0.72 [+0.64, +0.79] 97% Belcastro et al. (2023) +0.72 [+0.64, +0.79] 97% Caldarola et al. (2026) +0.73 [+0.66, +0.80] 97% Al‐Janabi et al. (2019) +0.72 [+0.65, +0.79] 97% Warren et al. (2020) +0.72 [+0.64, +0.79] 97% Ruggiero et al. (2021) +0.73 [+0.66, +0.80] 97% Mastorino et al. (2023) +0.73 [+0.65, +0.79] 97% Mastorino et al. (2024) +0.73 [+0.65, +0.79] 97% Carrascosa et al. (2021) +0.73 [+0.65, +0.79] 97% Bernardini et al. (2022) +0.74 [+0.67, +0.80] 97% Mastorino et al. (2025) +0.72 [+0.64, +0.78] 97% Hansel et al. (2020) +0.74 [+0.68, +0.80] 97% Carrascosa et al. (2023) +0.72 [+0.65, +0.79] 97% Asadullah et al. (2025) +0.73 [+0.65, +0.79] 97% Abu-Hilal et al. (2025) +0.73 [+0.65, +0.79] 97% Ruggiero et al. (2023) +0.72 [+0.65, +0.79] 97% Gönülal et al. (2023) +0.72 [+0.64, +0.78] 97% Fratton et al. (2025) +0.72 [+0.64, +0.79] 97% Griffiths et al. (2018) +0.72 [+0.64, +0.79] 97% Dubois (2024) +0.72 [+0.64, +0.78] 97% Megna et al. (2023) +0.72 [+0.64, +0.79] 97% Leo et al. (2024) +0.72 [+0.65, +0.79] 97% Licata et al. (2025) +0.72 [+0.64, +0.79] 97% - et al. (2024) +0.72 [+0.64, +0.79] 97% Ravasio et al. (2021) +0.73 [+0.65, +0.79] 97% Full-pool reference: +0.72 [+0.65, +0.79] (dashed oxblood line) Conclusion is robust to any single study exclusion.
Leave-one-out plot for PASI 90 (proportion). Each row shows the recomputed pool when one study is dropped; the dashed line marks the full-pool estimate. Rows flagged in oxblood would flip the direction or significance of the conclusion.
PASI 90 (proportion) — cumulative meta-analysis Year (k studies) Cumulative pool [95% CI] 2019 (k=2) +0.88 [+0.85, +0.90] 9% 2019 (k=3) +0.84 [+0.76, +0.90] 90% 2020 (k=4) +0.82 [+0.74, +0.87] 92% 2020 (k=5) +0.82 [+0.76, +0.87] 91% 2020 (k=6) +0.83 [+0.78, +0.87] 89% 2020 (k=7) +0.71 [+0.41, +0.90] 96% 2021 (k=8) +0.72 [+0.44, +0.89] 96% 2021 (k=9) +0.72 [+0.48, +0.88] 95% 2021 (k=10) +0.68 [+0.45, +0.85] 95% 2021 (k=11) +0.68 [+0.46, +0.84] 96% 2021 (k=12) +0.67 [+0.46, +0.83] 95% 2022 (k=13) +0.65 [+0.44, +0.81] 98% 2022 (k=14) +0.65 [+0.47, +0.80] 98% 2022 (k=15) +0.60 [+0.39, +0.78] 98% 2023 (k=16) +0.61 [+0.41, +0.77] 98% 2023 (k=17) +0.62 [+0.44, +0.78] 97% 2023 (k=18) +0.62 [+0.45, +0.77] 97% 2023 (k=19) +0.64 [+0.48, +0.78] 98% 2023 (k=20) +0.64 [+0.49, +0.77] 97% 2023 (k=21) +0.66 [+0.51, +0.78] 97% 2023 (k=22) +0.65 [+0.51, +0.77] 97% 2023 (k=23) +0.66 [+0.52, +0.77] 97% 2023 (k=24) +0.66 [+0.54, +0.77] 97% 2023 (k=25) +0.68 [+0.55, +0.78] 97% 2023 (k=26) +0.69 [+0.57, +0.78] 97% 2024 (k=27) +0.68 [+0.57, +0.78] 97% 2024 (k=28) +0.68 [+0.57, +0.77] 97% 2024 (k=29) +0.69 [+0.59, +0.78] 97% 2024 (k=30) +0.70 [+0.59, +0.78] 97% 2024 (k=31) +0.70 [+0.60, +0.79] 97% 2025 (k=32) +0.70 [+0.60, +0.78] 97% 2025 (k=33) +0.70 [+0.61, +0.78] 97% 2025 (k=34) +0.70 [+0.61, +0.78] 97% 2025 (k=35) +0.71 [+0.62, +0.78] 97% 2025 (k=36) +0.72 [+0.63, +0.79] 97% 2025 (k=37) +0.72 [+0.64, +0.79] 97% 2025 (k=38) +0.72 [+0.64, +0.79] 97% 2025 (k=39) +0.72 [+0.64, +0.79] 97% 2025 (k=40) +0.73 [+0.65, +0.79] 97% 2025 (k=41) +0.73 [+0.65, +0.79] 97% 2026 (k=42) +0.73 [+0.66, +0.80] 97% 2026 (k=43) +0.72 [+0.65, +0.79] 97% Final pool: +0.72 [+0.65, +0.79] (dashed oxblood line) Evidence appears to have stabilised — the most recent additions do not materially shift the pooled estimate.
Cumulative meta-analysis for PASI 90 (proportion). Each row shows the pooled estimate when the first k studies (ordered by publication year) are included. The dashed line marks the final pool; stable rows near the line indicate evidence has plateaued.
Leave-one-out sensitivity: PASI 90 (smd). Robust to single-study exclusion.
PASI 90 (smd) — leave-one-out Dropped study Recomputed pool [95% CI] Hjort et al. (2024) -0.70 [-1.24, -0.16] 98% Shi et al. (2025) -0.36 [-0.60, -0.12] 98% doi_10.1016/j.jaad.2018… -0.63 [-1.20, -0.06] 98% Diels et al. (2019) -0.62 [-1.19, -0.05] 98% Lv et al. (2018) -0.64 [-1.20, -0.07] 98% Alzghool et al. (2026) -0.68 [-1.24, -0.13] 98% Abdullah et al. (2025) -0.56 [-1.12, -0.00] 98% Lee et al. (2026) -0.70 [-1.24, -0.15] 98% Full-pool reference: -0.61 [-1.11, -0.11] (dashed oxblood line) Conclusion is robust to any single study exclusion.
Leave-one-out plot for PASI 90 (smd). Each row shows the recomputed pool when one study is dropped; the dashed line marks the full-pool estimate. Rows flagged in oxblood would flip the direction or significance of the conclusion.
PASI 90 (smd) — cumulative meta-analysis Year (k studies) Cumulative pool [95% CI] 2018 (k=2) -0.48 [-0.60, -0.35] 0% 2019 (k=3) -0.52 [-0.61, -0.43] 0% 2024 (k=4) -0.36 [-0.60, -0.12] 98% 2025 (k=5) -0.77 [-1.50, -0.04] 98% 2025 (k=6) -0.80 [-1.40, -0.19] 99% 2026 (k=7) -0.70 [-1.24, -0.15] 98% 2026 (k=8) -0.61 [-1.11, -0.11] 98% Final pool: -0.61 [-1.11, -0.11] (dashed oxblood line) Evidence appears to have stabilised — the most recent additions do not materially shift the pooled estimate.
Cumulative meta-analysis for PASI 90 (smd). Each row shows the pooled estimate when the first k studies (ordered by publication year) are included. The dashed line marks the final pool; stable rows near the line indicate evidence has plateaued.
Leave-one-out sensitivity: PASI 90 (risk_ratio). Robust to single-study exclusion.
PASI 90 (risk_ratio) — leave-one-out Dropped study Recomputed pool [95% CI] Sbidian et al. (2020) +16.97 [+6.21, +46.34] 100% Sbidian et al. (2021) +16.72 [+6.15, +45.47] 100% Sbidian et al. (2022) +16.19 [+6.03, +43.46] 100% Sbidian et al. (2023) +16.52 [+6.10, +44.75] 100% Ismail et al. (2024) +15.49 [+5.93, +40.48] 100% Erichsen et al. (2019) +28.66 [+24.95, +32.93] 100% Full-pool reference: +17.98 [+7.73, +41.78] (dashed oxblood line) Conclusion is robust to any single study exclusion.
Leave-one-out plot for PASI 90 (risk_ratio). Each row shows the recomputed pool when one study is dropped; the dashed line marks the full-pool estimate. Rows flagged in oxblood would flip the direction or significance of the conclusion.
PASI 90 (risk_ratio) — cumulative meta-analysis Year (k studies) Cumulative pool [95% CI] 2020 (k=2) +6.45 [+1.04, +39.86] 100% 2021 (k=3) +10.24 [+2.47, +42.41] 100% 2022 (k=4) +13.43 [+4.21, +42.86] 100% 2023 (k=5) +15.49 [+5.93, +40.48] 100% 2024 (k=6) +17.98 [+7.73, +41.78] 100% Final pool: +17.98 [+7.73, +41.78] (dashed oxblood line) Evidence appears to have stabilised — the most recent additions do not materially shift the pooled estimate.
Cumulative meta-analysis for PASI 90 (risk_ratio). Each row shows the pooled estimate when the first k studies (ordered by publication year) are included. The dashed line marks the final pool; stable rows near the line indicate evidence has plateaued.
Leave-one-out sensitivity: PASI 100 (proportion). Robust to single-study exclusion.
PASI 100 (proportion) — leave-one-out Dropped study Recomputed pool [95% CI] Berry (2021) +0.49 [+0.40, +0.58] 98% Narcisi et al. (2023) +0.48 [+0.39, +0.57] 98% Alexis et al. (2025) +0.49 [+0.40, +0.58] 98% Papp et al. (2021) +0.48 [+0.39, +0.58] 98% Kim et al. (2023) +0.50 [+0.41, +0.58] 98% Pinter et al. (2023) +0.50 [+0.41, +0.59] 95% Campoli et al. (2024) +0.49 [+0.40, +0.58] 98% Thaçi et al. (2025) +0.48 [+0.39, +0.57] 98% Reguiai et al. (2026) +0.48 [+0.39, +0.57] 98% Trovato et al. (2026) +0.47 [+0.39, +0.56] 98% Papp et al. (2025) +0.48 [+0.39, +0.58] 98% doi_10.1016_j.ad.2024.0… +0.48 [+0.39, +0.57] 98% Amin et al. (2023) +0.48 [+0.39, +0.58] 98% Blauvelt et al. (2025) +0.50 [+0.41, +0.58] 98% Liu et al. (2025) +0.49 [+0.40, +0.58] 98% Belcastro et al. (2023) +0.48 [+0.39, +0.57] 98% Ruggiero et al. (2021) +0.50 [+0.42, +0.59] 98% Carrascosa et al. (2021) +0.49 [+0.40, +0.58] 98% Mastorino et al. (2025) +0.47 [+0.39, +0.55] 98% Hansel et al. (2020) +0.49 [+0.40, +0.58] 98% Abu-Hilal et al. (2025) +0.49 [+0.40, +0.58] 98% Ruggiero et al. (2023) +0.49 [+0.40, +0.58] 98% Gönülal et al. (2023) +0.47 [+0.39, +0.56] 98% Fratton et al. (2025) +0.48 [+0.39, +0.57] 98% Megna et al. (2023) +0.49 [+0.40, +0.58] 98% Leo et al. (2024) +0.50 [+0.41, +0.59] 98% Licata et al. (2025) +0.49 [+0.40, +0.58] 98% - et al. (2024) +0.47 [+0.39, +0.56] 98% Reich et al. (2022) +0.49 [+0.40, +0.58] 98% Full-pool reference: +0.49 [+0.40, +0.57] (dashed oxblood line) Conclusion is robust to any single study exclusion.
Leave-one-out plot for PASI 100 (proportion). Each row shows the recomputed pool when one study is dropped; the dashed line marks the full-pool estimate. Rows flagged in oxblood would flip the direction or significance of the conclusion.
Leave-one-out sensitivity: PASI 100 (smd). Robust to single-study exclusion.
PASI 100 (smd) — leave-one-out Dropped study Recomputed pool [95% CI] doi_10.1016/j.jaad.2018… -1.66 [-2.37, -0.96] 86% Hu et al. (2026) -1.28 [-2.37, -0.20] 99% Barcelos et al. (2021) -0.53 [-0.66, -0.40] 35% Full-pool reference: -1.20 [-1.99, -0.41] (dashed oxblood line) Conclusion is robust to any single study exclusion.
Leave-one-out plot for PASI 100 (smd). Each row shows the recomputed pool when one study is dropped; the dashed line marks the full-pool estimate. Rows flagged in oxblood would flip the direction or significance of the conclusion.
PASI 100 (smd) — cumulative meta-analysis Year (k studies) Cumulative pool [95% CI] 2021 (k=2) -1.28 [-2.37, -0.20] 99% 2026 (k=3) -1.20 [-1.99, -0.41] 98% Final pool: -1.20 [-1.99, -0.41] (dashed oxblood line) Evidence appears to have stabilised — the most recent additions do not materially shift the pooled estimate.
Cumulative meta-analysis for PASI 100 (smd). Each row shows the pooled estimate when the first k studies (ordered by publication year) are included. The dashed line marks the final pool; stable rows near the line indicate evidence has plateaued.
Leave-one-out sensitivity: adverse events (proportion). Robust to single-study exclusion.
adverse events (proportion) — leave-one-out Dropped study Recomputed pool [95% CI] Berry (2021) +0.09 [+0.04, +0.19] 99% Reich et al. (2015) +0.10 [+0.05, +0.21] 98% Reguiai et al. (2026) +0.09 [+0.04, +0.19] 99% Papp et al. (2017) +0.09 [+0.04, +0.19] 99% - et al. (2024) +0.09 [+0.04, +0.20] 99% Izu-Belloso et al. (202… +0.09 [+0.04, +0.19] 99% Crowley et al. (2019) +0.07 [+0.03, +0.14] 99% Snast et al. (2026) +0.10 [+0.05, +0.20] 99% Bagit et al. (2023) +0.09 [+0.04, +0.19] 99% Liu et al. (2025) +0.09 [+0.04, +0.19] 99% Ruggiero et al. (2021) +0.08 [+0.03, +0.16] 99% Loft et al. (2019) +0.07 [+0.03, +0.14] 98% Fratton et al. (2025) +0.08 [+0.03, +0.18] 99% Full-pool reference: +0.09 [+0.04, +0.18] (dashed oxblood line) Conclusion is robust to any single study exclusion.
Leave-one-out plot for adverse events (proportion). Each row shows the recomputed pool when one study is dropped; the dashed line marks the full-pool estimate. Rows flagged in oxblood would flip the direction or significance of the conclusion.
adverse events (proportion) — cumulative meta-analysis Year (k studies) Cumulative pool [95% CI] 2017 (k=2) +0.03 [+0.01, +0.09] 89% 2019 (k=3) +0.10 [+0.01, +0.50] 99% 2019 (k=4) +0.18 [+0.03, +0.59] 99% 2021 (k=5) +0.14 [+0.03, +0.45] 99% 2021 (k=6) +0.16 [+0.05, +0.43] 99% 2023 (k=7) +0.14 [+0.04, +0.37] 99% 2024 (k=8) +0.12 [+0.04, +0.30] 99% 2025 (k=9) +0.11 [+0.04, +0.26] 99% 2025 (k=10) +0.11 [+0.05, +0.25] 99% 2026 (k=11) +0.10 [+0.05, +0.22] 99% 2026 (k=12) +0.10 [+0.05, +0.20] 99% 2026 (k=13) +0.09 [+0.04, +0.18] 99% Final pool: +0.09 [+0.04, +0.18] (dashed oxblood line) Evidence appears to have stabilised — the most recent additions do not materially shift the pooled estimate.
Cumulative meta-analysis for adverse events (proportion). Each row shows the pooled estimate when the first k studies (ordered by publication year) are included. The dashed line marks the final pool; stable rows near the line indicate evidence has plateaued.
Leave-one-out sensitivity: PASI 75 (proportion). Robust to single-study exclusion.
PASI 75 (proportion) — leave-one-out Dropped study Recomputed pool [95% CI] Narcisi et al. (2023) +0.81 [+0.62, +0.92] 98% Papp et al. (2021) +0.81 [+0.62, +0.92] 98% Campoli et al. (2024) +0.81 [+0.62, +0.92] 98% Thaçi et al. (2025) +0.82 [+0.63, +0.92] 98% Bagit et al. (2023) +0.82 [+0.63, +0.93] 98% Blauvelt et al. (2025) +0.81 [+0.62, +0.92] 98% Youn et al. (2022) +0.81 [+0.62, +0.92] 98% Belcastro et al. (2023) +0.80 [+0.61, +0.91] 98% Yu et al. (2024) +0.83 [+0.64, +0.93] 98% Igarashi et al. (2021) +0.83 [+0.65, +0.93] 98% Thaçi et al. (2021) +0.81 [+0.62, +0.92] 98% Maul et al. (2021) +0.82 [+0.63, +0.93] 98% Bernardini et al. (2022) +0.84 [+0.68, +0.93] 98% Hansel et al. (2020) +0.86 [+0.77, +0.92] 96% Carrascosa et al. (2023) +0.81 [+0.62, +0.92] 98% Kolli et al. (2018) +0.83 [+0.66, +0.93] 98% Abu-Hilal et al. (2025) +0.81 [+0.61, +0.92] 98% Gönülal et al. (2023) +0.80 [+0.61, +0.91] 98% Dubois (2024) +0.81 [+0.62, +0.92] 98% Licata et al. (2025) +0.81 [+0.61, +0.92] 98% - et al. (2024) +0.81 [+0.61, +0.92] 98% Full-pool reference: +0.82 [+0.64, +0.92] (dashed oxblood line) Conclusion is robust to any single study exclusion.
Leave-one-out plot for PASI 75 (proportion). Each row shows the recomputed pool when one study is dropped; the dashed line marks the full-pool estimate. Rows flagged in oxblood would flip the direction or significance of the conclusion.
PASI 75 (proportion) — cumulative meta-analysis Year (k studies) Cumulative pool [95% CI] 2020 (k=2) +0.04 [+0.00, +0.57] 99% 2021 (k=3) +0.20 [+0.01, +0.92] 99% 2021 (k=4) +0.27 [+0.02, +0.88] 99% 2021 (k=5) +0.43 [+0.05, +0.91] 99% 2021 (k=6) +0.48 [+0.09, +0.90] 99% 2022 (k=7) +0.58 [+0.15, +0.91] 99% 2022 (k=8) +0.51 [+0.14, +0.87] 99% 2023 (k=9) +0.57 [+0.19, +0.88] 99% 2023 (k=10) +0.59 [+0.23, +0.87] 98% 2023 (k=11) +0.66 [+0.29, +0.90] 98% 2023 (k=12) +0.69 [+0.35, +0.90] 98% 2023 (k=13) +0.73 [+0.41, +0.92] 98% 2024 (k=14) +0.75 [+0.45, +0.91] 98% 2024 (k=15) +0.74 [+0.46, +0.90] 98% 2024 (k=16) +0.76 [+0.50, +0.91] 98% 2024 (k=17) +0.78 [+0.54, +0.91] 98% 2025 (k=18) +0.78 [+0.56, +0.91] 98% 2025 (k=19) +0.79 [+0.58, +0.91] 98% 2025 (k=20) +0.81 [+0.61, +0.92] 98% 2025 (k=21) +0.82 [+0.64, +0.92] 98% Final pool: +0.82 [+0.64, +0.92] (dashed oxblood line) Evidence appears to have stabilised — the most recent additions do not materially shift the pooled estimate.
Cumulative meta-analysis for PASI 75 (proportion). Each row shows the pooled estimate when the first k studies (ordered by publication year) are included. The dashed line marks the final pool; stable rows near the line indicate evidence has plateaued.
Leave-one-out sensitivity: PASI 75 (smd). Conclusion is sensitive to specific studies.
PASI 75 (smd) — leave-one-out Dropped study Recomputed pool [95% CI] ▲ Shi et al. (2025) -0.88 [-1.84, +0.07] 99% doi_10.1016/j.jaad.2018… -1.57 [-2.74, -0.40] 99% ▲ Lv et al. (2018) -1.01 [-2.18, +0.17] 96% Alzghool et al. (2026) -1.65 [-2.69, -0.61] 99% Full-pool reference: -1.28 [-2.27, -0.28] (dashed oxblood line) 2 study(s) flip the conclusion when excluded — result is sensitive to individual studies.
Leave-one-out plot for PASI 75 (smd). Each row shows the recomputed pool when one study is dropped; the dashed line marks the full-pool estimate. Rows flagged in oxblood would flip the direction or significance of the conclusion.
PASI 75 (smd) — cumulative meta-analysis Year (k studies) Cumulative pool [95% CI] 2018 (k=2) -1.24 [-2.39, -0.10] 99% 2025 (k=3) -1.65 [-2.69, -0.61] 99% 2026 (k=4) -1.28 [-2.27, -0.28] 99% Final pool: -1.28 [-2.27, -0.28] (dashed oxblood line) Evidence appears to have stabilised — the most recent additions do not materially shift the pooled estimate.
Cumulative meta-analysis for PASI 75 (smd). Each row shows the pooled estimate when the first k studies (ordered by publication year) are included. The dashed line marks the final pool; stable rows near the line indicate evidence has plateaued.
Leave-one-out sensitivity: PASI (mean_difference). Conclusion is sensitive to specific studies.
PASI (mean_difference) — leave-one-out Dropped study Recomputed pool [95% CI] Gottlieb et al. (2026) -1.09 [-4.62, +2.43] 99% ▲ Liu et al. (2025) +0.10 [-2.89, +3.09] 99% Phan et al. (2026) -1.51 [-4.68, +1.66] 99% ▲ Carrascosa et al. (2021) +0.97 [-0.73, +2.67] 96% Ak et al. (2024) -0.91 [-4.49, +2.67] 99% Ak et al. (2025) -1.06 [-4.59, +2.47] 99% Megna et al. (2023) -0.96 [-4.53, +2.60] 99% Licata et al. (2025) -1.03 [-4.58, +2.51] 99% Full-pool reference: -0.72 [-3.84, +2.39] (dashed oxblood line) 2 study(s) flip the conclusion when excluded — result is sensitive to individual studies.
Leave-one-out plot for PASI (mean_difference). Each row shows the recomputed pool when one study is dropped; the dashed line marks the full-pool estimate. Rows flagged in oxblood would flip the direction or significance of the conclusion.
PASI (mean_difference) — cumulative meta-analysis Year (k studies) Cumulative pool [95% CI] 2023 (k=2) -4.28 [-11.21, +2.65] 100% 2024 (k=3) -2.75 [-7.96, +2.47] 100% 2025 (k=4) -3.88 [-8.51, +0.75] 99% 2025 (k=5) -2.78 [-6.87, +1.32] 99% 2025 (k=6) -2.07 [-5.64, +1.50] 99% 2026 (k=7) -1.51 [-4.68, +1.66] 99% 2026 (k=8) -0.72 [-3.84, +2.39] 99% Final pool: -0.72 [-3.84, +2.39] (dashed oxblood line) Evidence appears to have stabilised — the most recent additions do not materially shift the pooled estimate.
Cumulative meta-analysis for PASI (mean_difference). Each row shows the pooled estimate when the first k studies (ordered by publication year) are included. The dashed line marks the final pool; stable rows near the line indicate evidence has plateaued.
Leave-one-out sensitivity: PASI (proportion). Conclusion is sensitive to specific studies.
PASI (proportion) — leave-one-out Dropped study Recomputed pool [95% CI] Papp et al. (2017) +0.58 [+0.25, +0.85] 98% Bernardini et al. (2022) +0.57 [+0.25, +0.84] 98% ▲ Blauvelt et al. (2019) +0.74 [+0.59, +0.85] 89% Lazar et al. (2026) +0.65 [+0.31, +0.89] 99% doi_10.36849_jdd.8217 +0.54 [+0.25, +0.80] 99% Full-pool reference: +0.62 [+0.34, +0.84] (dashed oxblood line) 1 study(s) flip the conclusion when excluded — result is sensitive to individual studies.
Leave-one-out plot for PASI (proportion). Each row shows the recomputed pool when one study is dropped; the dashed line marks the full-pool estimate. Rows flagged in oxblood would flip the direction or significance of the conclusion.
Leave-one-out sensitivity: sPGA 0/1 (proportion). Robust to single-study exclusion.
sPGA 0/1 (proportion) — leave-one-out Dropped study Recomputed pool [95% CI] Papp et al. (2021) +0.84 [+0.82, +0.86] 50% Thaçi et al. (2025) +0.82 [+0.73, +0.88] 86% Papp et al. (2025) +0.81 [+0.73, +0.87] 85% Amin et al. (2023) +0.81 [+0.73, +0.86] 84% Blauvelt et al. (2025) +0.82 [+0.76, +0.86] 86% Full-pool reference: +0.82 [+0.76, +0.87] (dashed oxblood line) Conclusion is robust to any single study exclusion.
Leave-one-out plot for sPGA 0/1 (proportion). Each row shows the recomputed pool when one study is dropped; the dashed line marks the full-pool estimate. Rows flagged in oxblood would flip the direction or significance of the conclusion.
sPGA 0/1 (proportion) — cumulative meta-analysis Year (k studies) Cumulative pool [95% CI] 2023 (k=2) +0.79 [+0.64, +0.88] 95% 2025 (k=3) +0.80 [+0.72, +0.87] 90% 2025 (k=4) +0.82 [+0.76, +0.86] 86% 2025 (k=5) +0.82 [+0.76, +0.87] 82% Final pool: +0.82 [+0.76, +0.87] (dashed oxblood line) Evidence appears to have stabilised — the most recent additions do not materially shift the pooled estimate.
Cumulative meta-analysis for sPGA 0/1 (proportion). Each row shows the pooled estimate when the first k studies (ordered by publication year) are included. The dashed line marks the final pool; stable rows near the line indicate evidence has plateaued.
Leave-one-out sensitivity: PASI <3 (proportion). Conclusion is sensitive to specific studies.
PASI <3 (proportion) — leave-one-out Dropped study Recomputed pool [95% CI] Eyerich et al. (2024) +0.34 [+0.06, +0.80] 98% Tsianakas et al. (2026) +0.44 [+0.05, +0.91] 99% ▲ Bernardini et al. (2022) +0.76 [+0.64, +0.85] 98% Full-pool reference: +0.52 [+0.16, +0.87] (dashed oxblood line) 1 study(s) flip the conclusion when excluded — result is sensitive to individual studies.
Leave-one-out plot for PASI <3 (proportion). Each row shows the recomputed pool when one study is dropped; the dashed line marks the full-pool estimate. Rows flagged in oxblood would flip the direction or significance of the conclusion.
PASI <3 (proportion) — cumulative meta-analysis Year (k studies) Cumulative pool [95% CI] 2024 (k=2) +0.44 [+0.05, +0.91] 99% 2026 (k=3) +0.52 [+0.16, +0.87] 98% Final pool: +0.52 [+0.16, +0.87] (dashed oxblood line) Evidence appears to have stabilised — the most recent additions do not materially shift the pooled estimate.
Cumulative meta-analysis for PASI <3 (proportion). Each row shows the pooled estimate when the first k studies (ordered by publication year) are included. The dashed line marks the final pool; stable rows near the line indicate evidence has plateaued.
Leave-one-out sensitivity: IGA 0/1 (proportion). Robust to single-study exclusion.
IGA 0/1 (proportion) — leave-one-out Dropped study Recomputed pool [95% CI] Alexis et al. (2025) +0.80 [+0.56, +0.92] 98% Kim et al. (2023) +0.74 [+0.51, +0.89] 96% Youn et al. (2022) +0.73 [+0.52, +0.86] 97% Strober et al. (2025) +0.84 [+0.76, +0.91] 83% Full-pool reference: +0.78 [+0.60, +0.89] (dashed oxblood line) Conclusion is robust to any single study exclusion.
Leave-one-out plot for IGA 0/1 (proportion). Each row shows the recomputed pool when one study is dropped; the dashed line marks the full-pool estimate. Rows flagged in oxblood would flip the direction or significance of the conclusion.
IGA 0/1 (proportion) — cumulative meta-analysis Year (k studies) Cumulative pool [95% CI] 2023 (k=2) +0.88 [+0.84, +0.91] 0% 2025 (k=3) +0.84 [+0.76, +0.91] 83% 2025 (k=4) +0.78 [+0.60, +0.89] 97% Final pool: +0.78 [+0.60, +0.89] (dashed oxblood line) Evidence appears to have stabilised — the most recent additions do not materially shift the pooled estimate.
Cumulative meta-analysis for IGA 0/1 (proportion). Each row shows the pooled estimate when the first k studies (ordered by publication year) are included. The dashed line marks the final pool; stable rows near the line indicate evidence has plateaued.
Leave-one-out sensitivity: IGA 0 (proportion). Conclusion is sensitive to specific studies.
IGA 0 (proportion) — leave-one-out Dropped study Recomputed pool [95% CI] ▲ Alexis et al. (2025) +0.38 [+0.23, +0.57] 94% ▲ Kim et al. (2023) +0.42 [+0.29, +0.56] 87% Youn et al. (2022) +0.29 [+0.24, +0.34] 7% Full-pool reference: +0.36 [+0.25, +0.49] (dashed oxblood line) 2 study(s) flip the conclusion when excluded — result is sensitive to individual studies.
Leave-one-out plot for IGA 0 (proportion). Each row shows the recomputed pool when one study is dropped; the dashed line marks the full-pool estimate. Rows flagged in oxblood would flip the direction or significance of the conclusion.
IGA 0 (proportion) — cumulative meta-analysis Year (k studies) Cumulative pool [95% CI] 2023 (k=2) +0.38 [+0.23, +0.57] 94% 2025 (k=3) +0.36 [+0.25, +0.49] 89% Final pool: +0.36 [+0.25, +0.49] (dashed oxblood line) Evidence appears to have stabilised — the most recent additions do not materially shift the pooled estimate.
Cumulative meta-analysis for IGA 0 (proportion). Each row shows the pooled estimate when the first k studies (ordered by publication year) are included. The dashed line marks the final pool; stable rows near the line indicate evidence has plateaued.
Leave-one-out sensitivity: DLQI (mean_difference). Conclusion is sensitive to specific studies.
DLQI (mean_difference) — leave-one-out Dropped study Recomputed pool [95% CI] ▲ Gottlieb et al. (2026) -0.17 [-3.42, +3.08] 95% Bhatia et al. (2024) +1.04 [-1.37, +3.46] 94% Liu et al. (2025) +1.02 [-1.44, +3.48] 94% ▲ Phan et al. (2026) -0.64 [-3.30, +2.02] 91% Gracia-Cazaña et al. (2… +0.24 [-3.04, +3.52] 93% ▲ Ruiter et al. (2022) -0.45 [-3.37, +2.47] 95% Ak et al. (2025) +0.03 [-3.27, +3.33] 95% ▲ Licata et al. (2025) -0.03 [-3.32, +3.27] 95% Full-pool reference: +0.14 [-2.67, +2.94] (dashed oxblood line) 4 study(s) flip the conclusion when excluded — result is sensitive to individual studies.
Leave-one-out plot for DLQI (mean_difference). Each row shows the recomputed pool when one study is dropped; the dashed line marks the full-pool estimate. Rows flagged in oxblood would flip the direction or significance of the conclusion.
DLQI (mean_difference) — cumulative meta-analysis Year (k studies) Cumulative pool [95% CI] 2023 (k=2) +1.42 [-2.38, +5.22] 89% 2024 (k=3) -0.94 [-6.27, +4.38] 89% 2025 (k=4) -2.34 [-7.10, +2.41] 88% 2025 (k=5) -1.67 [-5.43, +2.10] 90% 2025 (k=6) -1.14 [-4.23, +1.95] 91% 2026 (k=7) -0.64 [-3.30, +2.02] 91% 2026 (k=8) +0.14 [-2.67, +2.94] 94% Final pool: +0.14 [-2.67, +2.94] (dashed oxblood line) Evidence appears to have stabilised — the most recent additions do not materially shift the pooled estimate.
Cumulative meta-analysis for DLQI (mean_difference). Each row shows the pooled estimate when the first k studies (ordered by publication year) are included. The dashed line marks the final pool; stable rows near the line indicate evidence has plateaued.
Leave-one-out sensitivity: DLQI (proportion). Conclusion is sensitive to specific studies.
DLQI (proportion) — leave-one-out Dropped study Recomputed pool [95% CI] Snast et al. (2026) +0.04 [+0.00, +0.86] 100% ▲ Gracia-Cazaña et al. (2… +0.01 [+0.00, +0.10] 98% Cuniberti et al. (2026) +0.23 [+0.03, +0.74] 99% Full-pool reference: +0.04 [+0.00, +0.56] (dashed oxblood line) 1 study(s) flip the conclusion when excluded — result is sensitive to individual studies.
Leave-one-out plot for DLQI (proportion). Each row shows the recomputed pool when one study is dropped; the dashed line marks the full-pool estimate. Rows flagged in oxblood would flip the direction or significance of the conclusion.
DLQI (proportion) — cumulative meta-analysis Year (k studies) Cumulative pool [95% CI] 2026 (k=2) +0.23 [+0.03, +0.74] 99% 2026 (k=3) +0.04 [+0.00, +0.56] 100% Final pool: +0.04 [+0.00, +0.56] (dashed oxblood line) Evidence appears to have stabilised — the most recent additions do not materially shift the pooled estimate.
Cumulative meta-analysis for DLQI (proportion). Each row shows the pooled estimate when the first k studies (ordered by publication year) are included. The dashed line marks the final pool; stable rows near the line indicate evidence has plateaued.
Leave-one-out sensitivity: infections (proportion). Conclusion is sensitive to specific studies.
infections (proportion) — leave-one-out Dropped study Recomputed pool [95% CI] Reich et al. (2015) +0.06 [+0.01, +0.33] 94% ▲ Dubois (2024) +0.19 [+0.05, +0.50] 82% Crowley et al. (2019) +0.06 [+0.01, +0.29] 90% ▲ Hansel et al. (2020) +0.16 [+0.03, +0.51] 91% Full-pool reference: +0.11 [+0.02, +0.37] (dashed oxblood line) 2 study(s) flip the conclusion when excluded — result is sensitive to individual studies.
Leave-one-out plot for infections (proportion). Each row shows the recomputed pool when one study is dropped; the dashed line marks the full-pool estimate. Rows flagged in oxblood would flip the direction or significance of the conclusion.
infections (proportion) — cumulative meta-analysis Year (k studies) Cumulative pool [95% CI] 2019 (k=2) +0.35 [+0.30, +0.41] 0% 2020 (k=3) +0.19 [+0.05, +0.50] 82% 2024 (k=4) +0.11 [+0.02, +0.37] 91% Final pool: +0.11 [+0.02, +0.37] (dashed oxblood line) Evidence appears to have stabilised — the most recent additions do not materially shift the pooled estimate.
Cumulative meta-analysis for infections (proportion). Each row shows the pooled estimate when the first k studies (ordered by publication year) are included. The dashed line marks the final pool; stable rows near the line indicate evidence has plateaued.
Leave-one-out sensitivity: DLQI 0/1 (proportion). Robust to single-study exclusion.
DLQI 0/1 (proportion) — leave-one-out Dropped study Recomputed pool [95% CI] Kim et al. (2023) +0.70 [+0.64, +0.75] 97% Thaçi et al. (2025) +0.69 [+0.62, +0.75] 97% Tsianakas et al. (2026) +0.71 [+0.67, +0.75] 83% Papp et al. (2025) +0.68 [+0.61, +0.73] 97% Amin et al. (2023) +0.68 [+0.61, +0.73] 97% Blauvelt et al. (2025) +0.68 [+0.62, +0.73] 97% Carrascosa et al. (2023) +0.69 [+0.62, +0.74] 97% Asadullah et al. (2025) +0.69 [+0.63, +0.75] 97% Lee et al. (2026) +0.68 [+0.62, +0.74] 97% Full-pool reference: +0.69 [+0.63, +0.74] (dashed oxblood line) Conclusion is robust to any single study exclusion.
Leave-one-out plot for DLQI 0/1 (proportion). Each row shows the recomputed pool when one study is dropped; the dashed line marks the full-pool estimate. Rows flagged in oxblood would flip the direction or significance of the conclusion.
DLQI 0/1 (proportion) — cumulative meta-analysis Year (k studies) Cumulative pool [95% CI] 2023 (k=2) +0.70 [+0.58, +0.79] 94% 2023 (k=3) +0.70 [+0.62, +0.77] 89% 2025 (k=4) +0.70 [+0.65, +0.75] 85% 2025 (k=5) +0.72 [+0.67, +0.76] 87% 2025 (k=6) +0.72 [+0.67, +0.76] 85% 2025 (k=7) +0.71 [+0.66, +0.75] 85% 2026 (k=8) +0.68 [+0.62, +0.74] 97% 2026 (k=9) +0.69 [+0.63, +0.74] 97% Final pool: +0.69 [+0.63, +0.74] (dashed oxblood line) Evidence appears to have stabilised — the most recent additions do not materially shift the pooled estimate.
Cumulative meta-analysis for DLQI 0/1 (proportion). Each row shows the pooled estimate when the first k studies (ordered by publication year) are included. The dashed line marks the final pool; stable rows near the line indicate evidence has plateaued.
Leave-one-out sensitivity: drug survival (proportion). Robust to single-study exclusion.
drug survival (proportion) — leave-one-out Dropped study Recomputed pool [95% CI] Reguiai et al. (2026) +0.92 [+0.86, +0.95] 88% Belcastro et al. (2023) +0.92 [+0.86, +0.95] 88% Mastorino et al. (2024) +0.93 [+0.88, +0.96] 88% Warren et al. (2020) +0.92 [+0.87, +0.96] 88% Mortato et al. (2025) +0.91 [+0.86, +0.95] 78% Balato et al. (2023) +0.93 [+0.87, +0.96] 87% Hu et al. (2026) +0.94 [+0.92, +0.96] 72% Full-pool reference: +0.92 [+0.88, +0.95] (dashed oxblood line) Conclusion is robust to any single study exclusion.
Leave-one-out plot for drug survival (proportion). Each row shows the recomputed pool when one study is dropped; the dashed line marks the full-pool estimate. Rows flagged in oxblood would flip the direction or significance of the conclusion.
drug survival (proportion) — cumulative meta-analysis Year (k studies) Cumulative pool [95% CI] 2023 (k=2) +0.95 [+0.91, +0.97] 0% 2023 (k=3) +0.92 [+0.90, +0.93] 32% 2024 (k=4) +0.92 [+0.90, +0.93] 28% 2025 (k=5) +0.94 [+0.91, +0.96] 78% 2026 (k=6) +0.94 [+0.92, +0.96] 72% 2026 (k=7) +0.92 [+0.88, +0.95] 86% Final pool: +0.92 [+0.88, +0.95] (dashed oxblood line) Evidence appears to have stabilised — the most recent additions do not materially shift the pooled estimate.
Cumulative meta-analysis for drug survival (proportion). Each row shows the pooled estimate when the first k studies (ordered by publication year) are included. The dashed line marks the final pool; stable rows near the line indicate evidence has plateaued.
Leave-one-out sensitivity: treatment discontinuation (proportion). Robust to single-study exclusion.
treatment discontinuation (proportion) — leave-one-out Dropped study Recomputed pool [95% CI] Izu-Belloso et al. (202… +0.08 [+0.04, +0.13] 55% Ruggiero et al. (2021) +0.09 [+0.04, +0.22] 85% Fratton et al. (2025) +0.10 [+0.04, +0.24] 82% Mastorino et al. (2025) +0.15 [+0.11, +0.22] 53% Full-pool reference: +0.10 [+0.05, +0.19] (dashed oxblood line) Conclusion is robust to any single study exclusion.
Leave-one-out plot for treatment discontinuation (proportion). Each row shows the recomputed pool when one study is dropped; the dashed line marks the full-pool estimate. Rows flagged in oxblood would flip the direction or significance of the conclusion.
treatment discontinuation (proportion) — cumulative meta-analysis Year (k studies) Cumulative pool [95% CI] 2025 (k=2) +0.10 [+0.05, +0.18] 0% 2025 (k=3) +0.08 [+0.04, +0.13] 55% 2026 (k=4) +0.10 [+0.05, +0.19] 78% Final pool: +0.10 [+0.05, +0.19] (dashed oxblood line) Evidence appears to have stabilised — the most recent additions do not materially shift the pooled estimate.
Cumulative meta-analysis for treatment discontinuation (proportion). Each row shows the pooled estimate when the first k studies (ordered by publication year) are included. The dashed line marks the final pool; stable rows near the line indicate evidence has plateaued.
Leave-one-out sensitivity: BSA (mean_difference). Conclusion is sensitive to specific studies.
BSA (mean_difference) — leave-one-out Dropped study Recomputed pool [95% CI] Gottlieb et al. (2026) +0.21 [-1.95, +2.36] 86% Ak et al. (2025) +0.81 [-2.11, +3.74] 89% Megna et al. (2023) +0.46 [-2.31, +3.23] 88% ▲ Lee et al. (2026) +1.62 [+0.78, +2.46] 67% Full-pool reference: +0.88 [-1.16, +2.92] (dashed oxblood line) 1 study(s) flip the conclusion when excluded — result is sensitive to individual studies.
Leave-one-out plot for BSA (mean_difference). Each row shows the recomputed pool when one study is dropped; the dashed line marks the full-pool estimate. Rows flagged in oxblood would flip the direction or significance of the conclusion.
BSA (mean_difference) — cumulative meta-analysis Year (k studies) Cumulative pool [95% CI] 2025 (k=2) +1.31 [+0.64, +1.97] 61% 2026 (k=3) +1.62 [+0.78, +2.46] 67% 2026 (k=4) +0.88 [-1.16, +2.92] 84% Final pool: +0.88 [-1.16, +2.92] (dashed oxblood line) Evidence appears to have stabilised — the most recent additions do not materially shift the pooled estimate.
Cumulative meta-analysis for BSA (mean_difference). Each row shows the pooled estimate when the first k studies (ordered by publication year) are included. The dashed line marks the final pool; stable rows near the line indicate evidence has plateaued.
Leave-one-out sensitivity: PGA 0/1 (proportion). Conclusion is sensitive to specific studies.
PGA 0/1 (proportion) — leave-one-out Dropped study Recomputed pool [95% CI] Tsianakas et al. (2026) +0.79 [+0.42, +0.95] 96% ▲ Igarashi et al. (2021) +0.81 [+0.52, +0.95] 95% ▲ Abu-Hilal et al. (2025) +0.63 [+0.61, +0.65] 66% Full-pool reference: +0.74 [+0.48, +0.89] (dashed oxblood line) 2 study(s) flip the conclusion when excluded — result is sensitive to individual studies.
Leave-one-out plot for PGA 0/1 (proportion). Each row shows the recomputed pool when one study is dropped; the dashed line marks the full-pool estimate. Rows flagged in oxblood would flip the direction or significance of the conclusion.
PGA 0/1 (proportion) — cumulative meta-analysis Year (k studies) Cumulative pool [95% CI] 2025 (k=2) +0.79 [+0.42, +0.95] 96% 2026 (k=3) +0.74 [+0.48, +0.89] 91% Final pool: +0.74 [+0.48, +0.89] (dashed oxblood line) Evidence appears to have stabilised — the most recent additions do not materially shift the pooled estimate.
Cumulative meta-analysis for PGA 0/1 (proportion). Each row shows the pooled estimate when the first k studies (ordered by publication year) are included. The dashed line marks the final pool; stable rows near the line indicate evidence has plateaued.
Leave-one-out sensitivity: PASI ≤ 3 (proportion). Robust to single-study exclusion.
PASI ≤ 3 (proportion) — leave-one-out Dropped study Recomputed pool [95% CI] Liu et al. (2025) +0.80 [+0.64, +0.90] 96% Mastorino et al. (2023) +0.88 [+0.84, +0.92] 0% Carrascosa et al. (2023) +0.79 [+0.63, +0.90] 91% Full-pool reference: +0.83 [+0.71, +0.91] (dashed oxblood line) Conclusion is robust to any single study exclusion.
Leave-one-out plot for PASI ≤ 3 (proportion). Each row shows the recomputed pool when one study is dropped; the dashed line marks the full-pool estimate. Rows flagged in oxblood would flip the direction or significance of the conclusion.
PASI ≤ 3 (proportion) — cumulative meta-analysis Year (k studies) Cumulative pool [95% CI] 2023 (k=2) +0.80 [+0.64, +0.90] 96% 2025 (k=3) +0.83 [+0.71, +0.91] 94% Final pool: +0.83 [+0.71, +0.91] (dashed oxblood line) Evidence appears to have stabilised — the most recent additions do not materially shift the pooled estimate.
Cumulative meta-analysis for PASI ≤ 3 (proportion). Each row shows the pooled estimate when the first k studies (ordered by publication year) are included. The dashed line marks the final pool; stable rows near the line indicate evidence has plateaued.
Cross-Paper Agreement Across Claim Groups 200 groups Consistent 100 (50%) Partial 83 (42%) Contradictory 17 (8%)
Agreement types across semantically grouped claims.

References

*Consistency of response (PASI 90 or 100 and IGA 0 or 0/1) in patients with moderate to severe psoriasis treated with guselkumab: Results from the VOYAGE 1 and 2 trials*. (2018). Journal of the American Academy of Dermatology. https://doi.org/10.1016/j.jaad.2018.05.377

*Eficacia y seguridad del risankizumab en pacientes con psoriasis en el mundo real : Un estudio retrospectivo, multicéntrico y no intervencionista*. (2025).

*Guselkumab Efficacy by Psoriasis Disease Severity and Treatment History: VOYAGE 1 and 2 Post Hoc Analyses*. (2025).

*Tildrakizumab Real-World Effectiveness and Safety Over 64 Weeks in Patients With Moderate-to-Severe Plaque Psoriasis*. (2024).

Abdullah, Abdullah Adnan Abdullah, & Shamoon, Linda Polus Bando (2025). Comparative Effectiveness of IL-17 and IL-23 Inhibitors for Moderate-to-Severe Plaque Psoriasis: A Systematic Review and Network Meta-Analysis (Preprint). https://doi.org/10.2196/preprints.85894

Abu-Hilal, M., Cowger, J., & Bawazir, M. (2025). Real-World Effectiveness of Tildrakizumab for Moderate-to-Severe Plaque Psoriasis in Canada. *Journal of cutaneous medicine and surgery*. https://doi.org/10.1177/12034754241302827

Ak, Melike, Birkenmaier, Ion, & Boehncke, Wolf-Henning (2025). Sex Disparities of Health-related Quality of Life in Moderate to Severe Psoriasis: A Real-world Analysis from the Swiss Psoriasis Registry (SDNTT). *Society for the Publication of Acta Dermato-Venereologica*. https://doi.org/10.2340/actadv.v105.42296

Ak, Melike, Burlando, Martina, & Cerminara, Sara E (2024). Tildrakizumab Treatment for Psoriasis in Real-world Practice: An Analysis from the Swiss Registry (SDNTT). *Society for the Publication of Acta Dermato-Venereologica*. https://doi.org/10.2340/actadv.v104.40946

Alexis, A., McMichael, A., & Soung, J. (2025). Guselkumab for Moderate to Severe Psoriasis Across All Skin Tones: Cohort A of the VISIBLE Randomized Clinical Trial. *JAMA dermatology*. https://doi.org/10.1001/jamadermatol.2025.1836

Alzghool, M., Jiquan, S., & Bashir, M.B. (2026). Comparative efficacy and safety of IL-17 and IL-23 inhibitors for moderate-to-severe psoriasis in Asian populations: a systematic review and meta-analysis. *BMC immunology*. https://doi.org/10.1186/s12865-025-00789-2

Al‐Janabi, A., Jabbar‐Lopez, Z. K., & Griffiths, C. (2019). Risankizumab vs. ustekinumab for plaque psoriasis: a critical appraisal. *British Journal of Dermatology*. https://doi.org/10.1111/bjd.17624

Amin, A.Z., Azam, T., & Beissert, Stefan (2023). Long-term safety and efficacy of risankizumab for the treatment of moderate-to-severe plaque psoriasis : Interim analysis of the LIMMitless open-label extension trial up to 5 years of follow-up. https://doi.org/10.1016/j.jaad.2023.07.1024

Asadullah, Khusru, Gerdes, Christina Valeska, & Gomez, Mario (2025). Effectiveness, safety and quality‐of‐life effects of guselkumab and ustekinumab in patients with psoriasis: Week 104 results from the non‐interventional, prospective, German multicentre PERSIST study. *Wiley*. https://doi.org/10.1111/jdv.19296

Augustin, Matthias, Lambert, Jo, & Zema, Carla (2020). Effect of Risankizumab on Patient-Reported Outcomes in Moderate to Severe Psoriasis. *JAMA Dermatology*. https://doi.org/10.1001/jamadermatol.2020.3617

Bagel, Jerry, Gogineni, Ranga, & Shaikh, Asif (2026). Efficacy and Safety of Tildrakizumab for Moderate-to-Severe Plaque Psoriasis with Diabetes: Pooled Subgroup Analysis of reSURFACE 1 and reSURFACE 2. *Dermatology and Therapy*. https://doi.org/10.1007/s13555-026-01708-y

Bagit, A., Maliyar, K., & Georgakopoulos, J. R. (2023). Real-world effectiveness and safety of risankizumab in patients with plaque psoriasis in whom guselkumab failed recently: A multicenter retrospective study of switching within the interleukin-23 inhibitor class. *JAAD International*. https://doi.org/10.1016/j.jdin.2023.05.006

Balato, A., Bardazzi, F., & Burlando, M. (2023). Drug survival of IL-12/23, IL-17 and IL-23 inhibitors for moderate-to-severe plaque psoriasis: a retrospective multicenter real-world experience on 5932 treatment courses – IL PSO (Italian landscape psoriasis). *Frontiers Media SA*. https://doi.org/10.3389/fimmu.2023.1341708

Barcelos, Anabela, Filipe, Paulo, & Fonseca, João Eurico (2021). A systematic review with network meta-analysis of the available biologic therapies for psoriatic disease domains. *'Frontiers Media SA'*. https://doi.org/10.3389/fmed.2020.618163

Belcastro, A., Bianchi, L., & Galluzzo, M. (2023). Guselkumab for treatment of moderate-to-severe plaque psoriasis: real-life effectiveness and drug-survival for up to 148 weeks. https://doi.org/10.1080/14712598.2023.2194485

Bernardini, N., Calabrese, L., & Caldarola, G. (2022). Tildrakizumab in moderate-to-severe plaque psoriasis: A multicenter, retrospective, real-life study. https://doi.org/10.1111/dth.15488

Berry, Angel L (2021). Is Risankizumab Effective in Decreasing Scaling and Percentage of Body Surface Area Affected Amongst Adults with Moderate-Severe Plaque Psoriasis?.

Bhatia, Neal, Heim, Jayme, & Vasquez, J Gabriel (2024). Long-term quality of life outcomes from a phase 4 study of tildrakizumab in patients with moderate-to-severe plaque psoriasis in a real-world setting. *The Journal of dermatological treatment*. https://doi.org/10.1080/09546634.2024.2310631

Blauvelt, A., Leonardi, C. L., & Gooderham, M. (2020). Efficacy and Safety of Continuous Risankizumab Therapy vs Treatment Withdrawal in Patients With Moderate to Severe Plaque Psoriasis. *JAMA Dermatology*. https://doi.org/10.1001/jamadermatol.2020.0723

Blauvelt, A., Sofen, H., & Papp, K. (2019). Tildrakizumab efficacy and impact on quality of life up to 52 weeks in patients with moderate‐to‐severe psoriasis: a pooled analysis of two randomized controlled trials. *Journal of the European Academy of Dermatology and Venereology*. https://doi.org/10.1111/jdv.15862

Blauvelt, Andrew, Jiang, Rundong, & Shi, Linyu (2025). A randomized phase 2 clinical trial to treat moderate-to-severe plaque psoriasis patients with high-induction dosing of risankizumab. *Nature communications*. https://doi.org/10.1038/s41467-025-67475-0

Caldarola, Giacomo, De Luca, Eleonora, & Balato, Anna (2026). Intraclass switching of interleukin-23 inhibitors in psoriasis: effectiveness, patterns and predictors of response - a retrospective multicentre study in Italy. *Clinical and experimental dermatology*. https://doi.org/10.1093/ced/llaf476

Campoli, Marco, Carrera, Carlo G., & Carugno, Andrea (2024). A 3-Year Multicentric Study on Switching from Ustekinumab to Guselkumab in Partial Responders with Psoriasis—IL PSO (Italian Landscape Psoriasis). *Dermatology and Therapy*. https://doi.org/10.1007/s13555-024-01270-5

Carrascosa, José Manuel, Carretero, G., & de la Cueva, Pablo (2021). Effectiveness and safety of guselkumab for the treatment of psoriasis in real-world settings at 24 weeks : A retrospective, observational, multicentre study by the Spanish Psoriasis Group. *'Wiley'*. https://doi.org/10.1111/dth.15231

Carrascosa, José Manuel, Costanzo, A., & Cuccia, A. (2023). Tildrakizumab improves high burden skin symptoms, impaired sleep and quality of life of moderate-to-severe plaque psoriasis patients in conditions close to clinical practice. *Journal of the European Academy of Dermatology and Venereology*. https://doi.org/10.1111/jdv.19229

Crowley, J., Warren, R. B., & Cather, J. (2019). Safety of selective <scp>IL</scp>‐23p19 inhibitors for the treatment of psoriasis. *Journal of the European Academy of Dermatology and Venereology*. https://doi.org/10.1111/jdv.15653

Cuniberti, Francesco, Miniotti, Marco, & Bailon, Mariagiulia (2026). From skin clearance to psychological wellbeing: real-world outcomes of biologic therapy in psoriasis. *Frontiers in psychology*. https://doi.org/10.3389/fpsyg.2026.1735777

Diels, J., Thilakarathne, P., & Cameron, C. (2019). Adjusted treatment COMPArisons between guSelkumab and uStekinumab for treatment of moderate‐to‐severe plaque psoriasis: the COMPASS analysis. *British Journal of Dermatology*. https://doi.org/10.1111/bjd.18634

Drerup, Katharina A, Seemann, Claudia, & Gerdes, Sascha (2022). Effective and Safe Treatment of Psoriatic Disease with the Anti-IL-23p19 Biologic Tildrakizumab: Results of a Real-World Prospective Cohort Study in Nonselected Patients. *Dermatology (Basel, Switzerland)*. https://doi.org/10.1159/000519924

Dubois, Pierre (2024). Evaluation of novel biologic agents in the treatment of psoriasis: A focus on long-term efficacy. *International Journal of Dermatology Sciences*. https://doi.org/10.33545/26649772.2024.v6.i1a.59

Erichsen, C., Jensen, P. E. H., & Kofoed, K. F. (2019). Biologic therapies targeting the interleukin (<scp>IL</scp>)‐23/<scp>IL</scp>‐17 immune axis for the treatment of moderate‐to‐severe plaque psoriasis: a systematic review and meta‐analysis. *Journal of the European Academy of Dermatology and Venereology*. https://doi.org/10.1111/jdv.15879

Eyerich, K., Asadullah, K., & Pinter, A. (2024). Noninferiority of 16-Week vs 8-Week Guselkumab Dosing in Super Responders for Maintaining Control of Psoriasis. *JAMA Dermatology*. https://doi.org/10.1001/jamadermatol.2024.2463

Fratton, Z., Bighetti, S., & Bettolini, L. (2025). Real-World Experience of Guselkumab in the Elderly Population. *Psoriasis (Auckland, N.Z.)*. https://doi.org/10.2147/ptt.s549502

Gordon, K B, Reich, K, & Crowley, J J (2022). Disease activity and treatment efficacy using patient-level Psoriasis Area and Severity Index scores from tildrakizumab phase 3 clinical trials. *The Journal of dermatological treatment*. https://doi.org/10.1080/09546634.2020.1747590

Gordon, Kenneth B, Armstrong, April W, & Foley, Peter (2019). Guselkumab Efficacy after Withdrawal Is Associated with Suppression of Serum IL-23-Regulated IL-17 and IL-22 in Psoriasis: VOYAGE 2 Study. *The Journal of investigative dermatology*. https://doi.org/10.1016/j.jid.2019.05.016

Gottlieb, Scott, Farberg, Aaron S, & Bhatia, Neal (2026). Real-World Benefit of Tildrakizumab for Moderate-to-Severe Plaque Psoriasis: Findings from a Systematic Literature Review and Meta-Analysis. *Dermatology and therapy*. https://doi.org/10.1007/s13555-025-01625-6

Gracia-Cazaña, T., Bernal-Masferrer, L., & Morales-Callaghan, A.M. (2023). Risankizumab for the Treatment of Moderate to Severe Psoriasis: Impact on Health-Related Quality of Life and Psychological Wellbeing. *Clinical, cosmetic and investigational dermatology*. https://doi.org/10.2147/ccid.s296544

Griffiths, C., Papp, K., & Song, M. (2020). Continuous treatment with guselkumab maintains clinical responses through 4 years in patients with moderate-to-severe psoriasis: results from VOYAGE 1. *Journal of Dermatological Treatment*. https://doi.org/10.1080/09546634.2020.1782817

Griffiths, CEM, Papp, KA, & Song, M (2018). Maintenance of Response With Guselkumab for up to 3 Years’ Treatment in the Phase 3 VOYAGE 1 Trial of Patients With Plaque Psoriasis. *SKIN The Journal of Cutaneous Medicine*. https://doi.org/10.25251/skin.2.supp.91

Gönülal, M., Balcı, D.D., & Öztürkcan, S. (2023). Risankizumab for the Treatment of the Patients with Moderate to Severe Plaque Psoriasis During a 24-Week Period: Real-Life Experience. *Clinical, cosmetic and investigational dermatology*. https://doi.org/10.2147/ccid.s442427

Hansel, K., Zangrilli, A., & Bianchi, L. (2020). A multicenter study on effectiveness and safety of risankizumab in psoriasis: an Italian 16‐week real‐life experience during the COVID‐19 pandemic. *Journal of the European Academy of Dermatology and Venereology*. https://doi.org/10.1111/jdv.17003

Hjort, Gustav, Schwarz, Christopher Willy, & Skov, Lone (2024). Clinical Characteristics Associated With Response to Biologics in the Treatment of Psoriasis. *JAMA Dermatology*. https://doi.org/10.1001/jamadermatol.2024.1677

Hu, R., Li, Q., & Liu, X. (2026). A 76-week real-world multidimensional analysis of guselkumab in moderate to severe plaque psoriasis: a retrospective cohort study based on Chinese clinical practice standards. *Frontiers in immunology*. https://doi.org/10.3389/fimmu.2026.1684996

Igarashi, Atsuyuki, Nakagawa, Hidemi, & Morita, Akimichi (2021). Efficacy and safety of tildrakizumab in Japanese patients with moderate to severe plaque psoriasis: Results from a 64-week phase 3 study (reSURFACE 1). *The Journal of dermatology*. https://doi.org/10.1111/1346-8138.15789

Imafuku, S., Nakagawa, H., & Igarashi, A. (2021). Long‐term efficacy and safety of tildrakizumab in Japanese patients with moderate to severe plaque psoriasis: Results from a 5‐year extension of a phase 3 study (reSURFACE 1). *The Journal of Dermatology*. https://doi.org/10.1111/1346-8138.15763

Ismail, Omar, Jaber, Kamel, & Jaber, Yazan (2024). Short term efficacy of biological treatment for moderate-to-severe plaque psoriasis: a systematic review and network meta-analysis. *Archives of dermatological research*. https://doi.org/10.1007/s00403-024-03398-y

Izu-Belloso, R., Gainza-Apraiz, I., & Meruelo-Ruano, M. (2026). Real-World Effectiveness and Safety of Tildrakizumab in Moderate-to-Severe Psoriasis: A Multicenter Experience in the Basque Country. *Dermatology and therapy*. https://doi.org/10.1007/s13555-025-01607-8

Kearney, N., Gorecki, P., & Acciarri, L. (2025). Treatment of Plaque Psoriasis with Guselkumab Reduces Systemic Inflammatory Burden as Measured by Neutrophil/Lymphocyte Ratio, Platelet/Lymphocyte Ratio, and Monocyte/Lymphocyte Ratio: A post hoc Analysis of Three Randomised Clinical Trials. *Dermatology*. https://doi.org/10.1159/000545148

Kim, B., Jo, S. J., & Youn, S. W. (2023). Five-year Maintenance of Clinical Response and Consistent Safety Profile for Guselkumab in Asian patients with Psoriasis from VOYAGE 1 and VOYAGE 2. *Dermatology and Therapy*. https://doi.org/10.1007/s13555-023-01026-7

Kojanova, M., Pejrilova, D., & Fialova, J. (2026). Super-Response to Guselkumab Treatment in Patients With Moderate-to-Severe Psoriasis: Real-World Data With Up to Five Years of Follow-Up in The Czech Republic. *International journal of dermatology*. https://doi.org/10.1111/ijd.70126

Kolli, S. S., Gabros, S. D., & Pona, A. (2018). Tildrakizumab: A Review of Phase II and III Clinical Trials. *Annals of Pharmacotherapy*. https://doi.org/10.1177/1060028018809522

Lazar, Andrada-Luciana, Bolboacă, Sorana D, & Baican, Adrian-Lucian (2026). Exploratory Retrospective Assessment of Patients with Psoriasis Receiving Biological Therapy. *Medicina (Kaunas, Lithuania)*. https://doi.org/10.3390/medicina62020257

Lee, Y.B., Shin, B.S., & Kim, M. (2026). Factors Predicting Guselkumab Treatment Response in Patients with Moderate-to-Severe Plaque Psoriasis: A Post Hoc Analysis of Korean Real-World Data. *Journal of clinical medicine*. https://doi.org/10.3390/jcm15020704

Leo, F., Roccuzzo, G., & Mastorino, L. (2024). Psoriasis in Childbearing Age: A Real-Life, Retrospective, Single-Center Study on Anti-IL17 and IL-23 Agents. https://doi.org/10.3390/jcm13216401

Licata, G., Di Brizzi, E.V., & Castelli, F. (2025). Tildrakizumab and Quality of Life: Deep Dive into the Impact of Psoriasis and Treatment on Different Domains-Should Psychosocial Life Impairment Be Considered a Comorbidity?. *Journal of clinical medicine*. https://doi.org/10.3390/jcm14010223

Liu, Qianzi, Hu, Kun, & Sha, Yang (2025). Tildrakizumab in real-world Chinese psoriasis: efficacy-safety profiles from a 28-week retrospective cohort with geriatric, late-onset and metabolic syndrome stratification. *The Journal of dermatological treatment*. https://doi.org/10.1080/09546634.2025.2598183

Loft, N.D., Vaengebjerg, S., & Halling, A.‐S. (2019). Adverse events with IL‐17 and IL‐23 inhibitors for psoriasis and psoriatic arthritis: a systematic review and meta‐analysis of phase III studies. *Journal of the European Academy of Dermatology and Venereology*. https://doi.org/10.1111/jdv.16073

Lv, Jingjing, Zhou, Dongmei, & Wang, Yan (2018). Quantitative evaluation to efficacy and safety of therapies for psoriasis: A network meta-analysis. *Molecular pain*. https://doi.org/10.1177/1744806918762205

Mastorino, Luca, Dapavo, Paolo, & Ortoncelli, Michela (2025). Dose Modulation Strategies in Psoriatic Patients: Real-Life Pilot Comparison Between Risankizumab and Guselkumab up to 12 Months After Dose Spacing. *Experimental dermatology*. https://doi.org/10.1111/exd.70062

Mastorino, Luca, Dapavo, Paolo, & Susca, Sara (2023). Drug survival and clinical effectiveness of secukinumab, ixekizumab, brodalumab, guselkumab, risankizumab, tildrakizumab for psoriasis treatment. *JDDG: Journal der Deutschen Dermatologischen Gesellschaft*. https://doi.org/10.1111/ddg.15251

Mastorino, Luca, Dapavo, Paolo, & Susca, Sara (2024). <i>Drug survival</i> und klinische Wirksamkeit von Secukinumab, Ixekizumab, Brodalumab, Guselkumab, Risankizumab und Tildrakizumab in der Behandlung der Psoriasis. *JDDG: Journal der Deutschen Dermatologischen Gesellschaft*. https://doi.org/10.1111/ddg.15251_g

Maul, J., Augustin, M., & Sorbe, C. (2021). Association of sex and systemic therapy treatment outcomes in psoriasis: a two‐country, multicentre, prospective, noninterventional registry study*. *British Journal of Dermatology*. https://doi.org/10.1111/bjd.20387

Megna, Matteo, Ruggiero, Angelo, & Battista, Teresa (2023). Long-Term Efficacy and Safety of Risankizumab for Moderate to Severe Psoriasis: A 2-Year Real-Life Retrospective Study. *Journal of Clinical Medicine*. https://doi.org/10.3390/jcm12093233

Mortato, E., Talamonti, M., & Marcelli, L. (2025). Long-Term Real-World Effectiveness and Drug Survival of Guselkumab in Patients with Psoriasis: A 5-Year Retrospective Study. *Psoriasis (Auckland, N.Z.)*. https://doi.org/10.2147/ptt.s533005

Narcisi, A., Costanzo, A., & Carrera, C. G. (2023). Real-life effectiveness of tildrakizumab in chronic plaque psoriasis:

A 52-week multicentre retrospective study—IL PSO (Italian

landscape psoriasis).

Papp, K., Blauvelt, A., & Bukhalo, M. (2017). Risankizumab versus Ustekinumab for Moderate-to-Severe Plaque Psoriasis. *New England Journal of Medicine*. https://doi.org/10.1056/nejmoa1607017

Papp, K.A., Lebwohl, M.G., & Puig, L. (2025). Long-Term Safety and Efficacy of Risankizumab to Treat Moderate-to-Severe Plaque Psoriasis: Final LIMMitless Phase 3, Open-Label Extension Trial Results. *American journal of clinical dermatology*. https://doi.org/10.1007/s40257-025-00964-6

Papp, Kim A., de Vente, Saskia, & Zeng, Jiewei (2021). Long-Term Safety and Efficacy of Risankizumab in Patients with Moderate-to-Severe Chronic Plaque Psoriasis: Results from a Phase 2 Open-Label Extension Trial. *Dermatology and Therapy*. https://doi.org/10.1007/s13555-021-00490-3

Parthasarathi, A., Savariah, D. D., & Marodia, S. (2024). Comparative Efficacy and Safety of TNF-Alpha Inhibitors, IL-17 Inhibitors, and IL-23 Inhibitors in the Treatment of Moderate to Severe Psoriasis: A Systematic Review. *International Journal For Multidisciplinary Research*. https://doi.org/10.36948/ijfmr.2024.v06i06.29806

Phan, Duc Binh, Laws, Philip, & Smith, Catherine H (2026). Using biologic therapies as first-line systemic treatment for psoriasis: A cohort study following the target trial emulation framework from the British Association of Dermatologists Biologics and Immunomodulators Register (BADBIR). *The British journal of dermatology*. https://doi.org/10.1093/bjd/ljag098

Pinter, A., Costanzo, A., & Khattri, S. (2023). Comparative Effectiveness and Durability of Biologics in Clinical Practice: Month 12 Outcomes from the International, Observational Psoriasis Study of Health Outcomes (PSoHO). *Dermatology and Therapy*. https://doi.org/10.1007/s13555-023-01086-9

Pinter, Andreas, Brnabic, Alan, & Trovato, Emanuele (2025). Comparative Effectiveness and Durability of Biologics Through 24 Months for Patients with Moderate-to-Severe Psoriasis: Results from the International, Observational Psoriasis Study of Health Outcomes (PSoHO). *Dermatology and therapy*. https://doi.org/10.1007/s13555-025-01494-z

Ravasio, R., Costanzo, A., & Antonelli, S. (2021). Number needed to treat for interleukin inhibitors approved for the treatment of moderate-to-severe plaque psoriasis in Italy. *Global & Regional Health Technology Assessment*. https://doi.org/10.33393/grhta.2021.2222

Reguiai, Ziad, Ghislain, Pierre-Dominique, & Baudier, Emilie (2026). Real-World Effectiveness of Brodalumab in Challenging Psoriasis Subgroups: Insights from the PSO-TARGET Cohort. *Dermatology and therapy*. https://doi.org/10.1007/s13555-025-01629-2

Reich, K, Gordon, K B, & Strober, B (2022). Super-response to guselkumab treatment in patients with moderate-to-severe psoriasis: age, body weight, baseline Psoriasis Area and Severity Index, and baseline Investigator's Global Assessment scores predict complete skin clearance. *Journal of the European Academy of Dermatology and Venereology : JEADV*. https://doi.org/10.1111/jdv.18474

Reich, K., Mrowietz, U., & Radtke, M. A. (2015). Drug safety of systemic treatments for psoriasis: results from The German Psoriasis Registry PsoBest. *Archives of Dermatological Research*. https://doi.org/10.1007/s00403-015-1593-8

Ruggiero, A., Fabbrocicni, G., & Cacciapuoti, S. (2023). Tildrakizumab for the Treatment of Moderate-to-Severe Psoriasis: Results from 52 Weeks Real-Life Retrospective Study. *Clinical Cosmetic and Investigational Dermatology*. https://doi.org/10.2147/ccid.s402183

Ruggiero, A., Fabbrocini, G., & Cinelli, E. (2021). Anti‐interleukin‐23 for psoriasis in elderly patients: guselkumab, risankizumab and tildrakizumab in real‐world practice. *Clinical and Experimental Dermatology*. https://doi.org/10.1111/ced.14979

Ruiter, C. C. D., & Rustemeyer, T. (2022). Biologics Can Significantly Improve Dermatology Life Quality Index (DLQI) in Psoriatic Patients: A Systematic Review. *Psoriasis Targets and Therapy*. https://doi.org/10.2147/ptt.s356568

Sbidian, Emilie, Chaimani, Anna, & Garcia-Doval, Ignacio (2021). Systemic pharmacological treatments for chronic plaque psoriasis: a network meta-analysis. *The Cochrane database of systematic reviews*. https://doi.org/10.1002/14651858.CD011535.pub4

Sbidian, É., Chaimani, A., & García‐Doval, I. (2020). Systemic pharmacological treatments for chronic plaque psoriasis: a network meta-analysis. *Cochrane Database of Systematic Reviews*. https://doi.org/10.1002/14651858.cd011535.pub3

Sbidian, É., Chaimani, A., & García‐Doval, I. (2022). Systemic pharmacological treatments for chronic plaque psoriasis: a network meta-analysis. *Cochrane Database of Systematic Reviews*. https://doi.org/10.1002/14651858.cd011535.pub5

Sbidian, É., Chaimani, A., & Guelimi, R. (2023). Systemic pharmacological treatments for chronic plaque psoriasis: a network meta-analysis. *Cochrane Database of Systematic Reviews*. https://doi.org/10.1002/14651858.cd011535.pub6

Shi, A., Shu, Y., & Haddad, J.E. (2025). Cardiovascular and Kidney Outcomes After Systemic Treatment for Plaque Psoriasis: A Systematic Review and Network Meta-analysis. *Dermatology and therapy*. https://doi.org/10.1007/s13555-025-01472-5

Snast, Igor, Valdman-Grinshpoun, Yuliya, & Bathish, Naji (2026). Effectiveness, Quality of Life and Durability of Risankizumab in Patients with Moderate-to-Severe Psoriasis: Real-World Evidence from the PRIMMA Study. *Dermatology and therapy*. https://doi.org/10.1007/s13555-026-01729-7

Strober, B., Menter, A., & Leonardi, C. (2020). Efficacy of risankizumab in patients with moderate‐to‐severe plaque psoriasis by baseline demographics, disease characteristics and prior biologic therapy: an integrated analysis of the phase III UltIMMa‐1 and UltIMMa‐2 studies. *Journal of the European Academy of Dermatology and Venereology*. https://doi.org/10.1111/jdv.16521

Strober, Bruce, Armstrong, April, & Fitzgerald, Timothy (2025). Long-Term Effectiveness of Guselkumab vs. Other Biologic Therapies Among Plaque Psoriasis Patients in the CorEvitas Psoriasis Registry. *SKIN The Journal of Cutaneous Medicine*. https://doi.org/10.25251/skin.10.supp.531

Thaçi, D., Piaserico, S., & Warren, R. B. (2021). Five‐year efficacy and safety of tildrakizumab in patients with moderate‐to‐severe psoriasis who respond at week 28: pooled analyses of two randomized phase III clinical trials (reSURFACE 1 and reSURFACE 2)*. *British Journal of Dermatology*. https://doi.org/10.1111/bjd.19866

Thaçi, Diamant, Ohtsuki, Mamitaro, & Maul, Julia-Tatjana (2025). Real-World Effectiveness of Risankizumab in Patients with Moderate-to-Severe Psoriasis: Interim Analysis from the VALUE Global Prospective Post-marketing Observational Study at 25 Months. *Dermatology and therapy*. https://doi.org/10.1007/s13555-025-01342-0

Trovato, Emanuele, Cartocci, Alessandra, & Gaiani, Francesca (2026). Real-World Efficacy of IL-23 Inhibitors in Psoriasis Affecting High-Impact Areas: Indirect Comparison of Tildrakizumab 200 mg, Risankizumab, and Guselkumab-IL PSO (Italian Landscape Psoriasis). *Dermatology and therapy*. https://doi.org/10.1007/s13555-026-01662-9

Tsianakas, Athanasios, Magnolo, Nina, & Kempf, Afra (2026). Tildrakizumab Real-World Experience (T-REX) in Plaque Psoriasis: Meta-Analysis of Four Non-interventional Studies. *Dermatology and therapy*. https://doi.org/10.1007/s13555-026-01713-1

Tskhvarashvili, G, Aher, K, & Sveide, I (2025). Persistence of advanced systemic pharmacological treatment of moderate-to-severe psoriasis among bio-naïve patients-A retrospective register-based cohort study in Finland and Sweden. *Journal of the European Academy of Dermatology and Venereology : JEADV*. https://doi.org/10.1111/jdv.20198

Valenti, M., Ibba, L., & Di Giulio, S. (2025). Guselkumab Retention, Effectiveness, and Safety in Psoriasis: A 260-Week Real-World Multicenter Retrospective Study Exploring the Role of Concomitant PsA-IL PSO (Italian Landscape Psoriasis). *Dermatology and therapy*. https://doi.org/10.1007/s13555-025-01476-1

Warren, R. B., Blauvelt, A., & Poulin, Y. (2020). Efficacy and safety of risankizumab vs. secukinumab in patients with moderate‐to‐severe plaque psoriasis (IMMerge): results from a phase III, randomized, open‐label, efficacy–assessor‐blinded clinical trial*. *British Journal of Dermatology*. https://doi.org/10.1111/bjd.19341

Warren, R. B., Carrascosa, J., & Fumero, E. (2020). Time to relapse after tildrakizumab withdrawal in patients with moderate‐to‐severe psoriasis who were responders at week 28: <i>post hoc</i> analysis through 64 weeks from reSURFACE 1 trial. *Journal of the European Academy of Dermatology and Venereology*. https://doi.org/10.1111/jdv.16964

Wytsma, Jennifer, Woo, Taylor Evart, & Parsons, Laurie (2024). Long-Term Efficacy of Tildrakizumab in the Treatment of Psoriasis. *Skin therapy letter*.

Youn, Sang Woong, Yu, Dae Young, & Kim, Tae Yoon (2022). Efficacy and safety of guselkumab compared with placebo and adalimumab in Korean patients with moderate-to-severe psoriasis: post-hoc analysis from the phase III, double-blind, placebo- and active-comparator-controlled VOYAGE 1/2 trials. *The Journal of dermatological treatment*. https://doi.org/10.1080/09546634.2020.1770174

Yu, C., Geng, S., & Yang, B. (2024). Tildrakizumab for moderate-to-severe plaque psoriasis in Chinese patients: A 12-week randomized placebo-controlled phase III trial with long-term extension. *Chinese medical journal*. https://doi.org/10.1097/cm9.0000000000002873