Monitoring disease activity in atopic dermatitis (AD) remains challenging under biologic therapy. Clinical scores are informative but may not provide the objectivity and consistency over time required for long-term monitoring 1. With the recent expansion of systemic treatment options, there is a growing need for reliable biomarkers that capture disease activity longitudinally. Although multiple blood cytokines are known to decrease after initiation of dupilumab 2, 3, it remains unclear which biomarkers reliably reflect concurrent disease activity during therapy. We enrolled 170 Japanese patients with AD and 28 healthy controls (HCs) to explore blood biomarkers for monitoring AD severity. Of the 170 patients with AD, 24 received dupilumab and underwent longitudinal blood biomarker profiling over 6 months (Table S1). First, the levels of 17 candidate cytokines were compared between patients with AD and HCs. Interleukin (IL)-4, IL-6, IL-16, IL-17, IL-22, C-C motif chemokine ligand (CCL)5, CCL17, CCL20, CCL27, eosinophil-derived neurotoxin (EDN), tumor necrosis factor (TNF)-α, and thymic stromal lymphopoietin (TSLP) levels were significantly elevated in patients with AD (Figure S1). Meanwhile, IL-6, IL-10, IL-16, IL-17, IL-18, IL-22, CCL17, CCL20, CCL27, CXCL9, EDN, and TNF-α levels showed significant correlations with disease severity (Eczema Area and Severity Index EASI) (Figures 1 and S2). Notably, CCL17, IL-22, and CCL27 levels showed strong correlations (r > 0.5) with EASI, consistent with previous reports (Figure 1) 4, 5. The subjective pruritus score was most strongly correlated with CCL27 and IL-22 levels (Figure S3). We next assessed monthly blood cytokine levels in 24 patients receiving dupilumab (Figures 2 and S4). Clinical responses to dupilumab were favorable, with 12 (55%) and 19 (86%) of 22 patients with available EASI data achieving EASI75 and EASI50 at month 6, respectively (representing ≥ 75% and ≥ 50% improvement from baseline EASI). Of the cytokines correlated with EASI in the initial analysis, IL-22 and IL-18 consistently correlated with disease severity at six of seven time points (pretreatment, months 1–4, and month 6; r = 0.53–0.80 and r = 0.42–0.61, respectively). In contrast, the CCL17 level was significant only at pretreatment and month 1, with no correlation with EASI observed during months 2–6. These findings suggested that, during dupilumab treatment, IL-22, followed by IL-18, more consistently reflected concurrent disease activity throughout the treatment course, whereas the association between CCL17 and disease severity was limited to the early phase of treatment (Figure 2). We next analyzed longitudinal changes in EASI scores together with cytokine levels to relate biomarker dynamics to clinical response during dupilumab treatment (Figure S5). Median EASI decreased from 25.5 at baseline to 12.9 after one month and continued to decline thereafter; however, substantial inter-individual variability persisted over time. Levels of IL-22, IL-18, and CCL17 all decreased after treatment; however, CCL17 declined more steeply and exhibited a compressed range of variation, limiting its ability to capture concurrent disease activity under dupilumab treatment. In contrast, the levels of IL-22 and IL-18 retained measurable variability across time points, supporting their potential as practical biomarkers for longitudinal monitoring. Given that previous studies have proposed composite cytokine biomarkers 6, we next assessed whether IL-18 adds explanatory value beyond IL-22 during dupilumab treatment (Table S2). To do so, we compared three linear mixed-effect models using EASI as the outcome variable for months 1–6. The IL-22–only model showed a good fit (AIC = 770.47; marginal R2 = 0.359), whereas adding IL-18 resulted in only a modest improvement (AIC = 767.04; marginal R2 = 0.385). The IL-18–only model showed the poorest fit (AIC = 781.22; marginal R2 = 0.282). These results indicated that IL-18 added only modest additional information beyond IL-22 in our longitudinal models. Th2 cytokines play a central role in AD pathogenesis. Among their markers, CCL17 is widely used to assess severity 7. However, in severe AD, non-Th2 pathways, such as type 17- and type 22-related inflammation, also play critical roles 8, 9. Previous studies have shown that Asian AD populations display a stronger Th17 polarization than Caucasians 10, suggesting that ethnic differences may partly influence biomarker profiles. In our study of Japanese patients, along with Th2 biomarkers (CCL17, EDN, and CCL27), other inflammatory markers, including IL-6, IL-16, IL-18, IL-22, and TNF-α, were correlated with disease severity. In dupilumab-treated patients, CCL17 levels declined rapidly following treatment initiation, as reported in previous studies 2, 3. At later time points during treatment, the marked reduction in CCL17 levels resulted in attenuated variability, which likely limited its association with concurrent disease severity. From a mechanistic perspective, this pattern may reflect the pharmacodynamic suppression of type 2 inflammation by dupilumab. In contrast, IL-22 and IL-18, which are not limited to the Th2 axis, continued to show a positive association with concurrent EASI scores throughout the treatment course. This persistent association may reflect distinct kinetics of pathway modulation, with type 17-/22-associated signals persisting longer, whereas type 2 responses are rapidly suppressed by dupilumab, as suggested in our previous skin transcriptomic study 9. Together, these findings highlight the contribution of multiple pathways to severe AD and the need to monitor them in management. While confirmatory studies in other ethnic populations are needed to establish broader generalizability of our findings, our study suggests that IL-22 and IL-18 may serve as indicators of residual disease activity in dupilumab-treated patients with AD. A.F.-N. and H.Ka. conceived the study, designed the research, and drafted the manuscript. E.Y. conducted the statistical analysis. A.F.-N., H.Ka., S.O., Y.I., K.T., and K.Y. collected clinical data and samples. T.H. conducted laboratory assays and data management. H. Ko and M.A. supervised the project. All authors contributed to the revision and approved the final version of the manuscript. We are grateful to all the participants for their cooperation in this research. We thank Y. Yano, H. Maeo, S. Shibata, M. Tanaka, Y. Toriumi, and S. Saeki for their support in patient sampling. We thank D. Karalis and F. Zhao of Sysmex R K. Yoshida, T. Oyama, T. Kaji, and S. Hazama of Sysmex Corporation for plasma sample analysis, and H Lunding, K. Oda, and A. Nehrmann of Sysmex R&D Centre Europe GmbH for the establishment of reagents for immunoassays. This study was supported by the Japan Agency for Medical Research and Development (AMED) (JP17ek0410046, JP20ek0410079, JP22ek0410098, JP23ek0410118 and JP25ek0410135), the Japan Science and Technology Agency (JST) support program for starting up innovation hubs (JPMJIH1504), the Japan Society for the Promotion of Science (JSPS) KAKENHI (21K08338, 22K08391), and the Maruho Takagi Dermatology Foundation. This work was supported by the Japan Science and Technology Agency, the Japan Society for the Promotion of Science, and the Japan Agency for Medical Research and Development. T. Hasegawa is employed by Sysmex R&D Center Europe GmbH. All other authors declare no conflicts of interest. H. Koseki has received research funds (grants paid to his institution) from Maruho and Kao. M. Amagai has received research support and funds (grants paid to his institution) from Maruho, Ono, Torii, Sato, and Taiho. H. Kawasaki received research funds (grants paid to his institution) from Torii and Takagi. A patent application is pending in Japan for the blood cytokines associated with dupilumab treatment. The data that support the findings of this study are available from the corresponding author upon reasonable request. Supplementary Table 1 Patient demographics. Supplementary Table 2. Linear mixed-effect model results during months 1–6 of dupilumab treatment. Supplementary Figure 1. Comparison of blood cytokine levels in HCs and AD patients. Supplementary Figure 2. Nonsignificant correlations between blood cytokine levels and EASI scores in AD patients. Supplementary Figure 3. Correlations between blood cytokine levels and pruritus scores in AD patients. Supplementary Figure 4. Correlations between blood cytokine levels and EASI scores in AD patients before and during dupilumab treatment. Supplementary Figure 5. Longitudinal changes in EASI scores and blood cytokine levels in AD patients during dupilumab treatment. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.
Fukushima‐Nomura et al. (Wed,) studied this question.