We read with great interest the study by Argyropoulou et al. in J Eur Acad Dermatol Venereol and commend the authors for integrating GWAS, plasma proteomics and pQTL analyses to refine the molecular landscape of hidradenitis suppurativa (HS). However, several methodological and reporting issues warrant further discussion.1 Firstly, some internal inconsistencies weaken confidence in the reported findings. The abstract refers to 60 genome-wide significant loci, whereas the Results section refers to 60 annotated SNPs reaching genome-wide significance; these terms do not appear interchangeable and may overstate the number of independent association signals. In addition, table 1 reports IHS4 categories of 71, 75 and 75, which sum to 221 rather than the full 314 cases, suggesting incomplete availability or unreported missingness. Because IHS4 is central to severity stratification, explicit denominators are needed to judge the representativeness and interpretability of the severity-related analyses. The reported top SNP coordinate on chromosome 3 also appears difficult to reconcile with the stated hg38 build and would benefit from clarification, as coordinate-level inaccuracies weaken confidence in locus annotation and downstream biological interpretation. Secondly, although SAIGE and sample-size weighted meta-analysis are reasonable analytical choices for an imbalanced case–control design, the statistical reporting remains insufficient to judge the robustness of the main signals. For the lead GWAS findings, pooled effect estimates, confidence intervals, heterogeneity statistics and key variant-level quality information are not readily interpretable from the main report, making it difficult to assess the consistency of these associations across datasets. The two highlighted HLA-DRA variants also appear to be supported mainly by the meta-analysis rather than by the Greek cohort alone and may, therefore, be better regarded as meta-analysis-supported findings than as internally replicated signals. In addition, the pQTL enrichment analysis relies on a selective comparison framework: rather than comparing all differentially expressed proteins with all non-differentially expressed proteins, the authors compared only the 15 differentially expressed proteins with the largest absolute log2 fold changes against 15 selected non-differentially expressed proteins. Because enrichment results are often sensitive to the choice of the background set and comparator group, this strategy limits interpretability unless it is more fully justified. By contrast, larger HS GWAS studies have identified relatively few independent association signals and have linked susceptibility predominantly to epidermal keratinization, keratinocyte function and Notch/Wnt-related pathways rather than to a primarily HLA-centred framework.2, 3 Thirdly, some of the interpretative claims appear stronger than the data currently support. The pathway enrichment results are driven mainly by HLA-DQB1 and HLA-DRA, suggesting that the observed immune-related signals may partly reflect the high gene density of the MHC region rather than, on their own, providing a robust basis for placing HS more firmly within the autoimmune spectrum. Likewise, although proteins such as IL-6 and IL-17A are associated with disease severity, these findings are better viewed as candidate biomarker signals rather than evidence of clinically usable biomarkers, because the present study does not provide data on discrimination, calibration or external validation. This more cautious interpretation is also supported by the broader literature: a recent proteomic study similarly linked IL-6 to HS severity, but its authors explicitly acknowledged that the association could still be influenced by treatment status and comorbidities; moreover, a systematic review concluded that no HS biomarker has yet shown sufficient clinical validity for routine use.4, 5 Addressing these issues would improve reproducibility, biological interpretation and clinical credibility. More transparent reporting and a more cautious translational framing would help position the work as a strong hypothesis-generating contribution to future multi-omic HS research. Declaration on Generative AI and AI-assisted technologies in the writing process: During the preparation of this work, the authors used ChatGPT (OpenAI) to assist with English-language editing and sentence-level rephrasing. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication. The authors were supported by the Hejiang People's Hospital-Southwest Medical University Science and Technology Strategic Cooperation Project (2023HJX-NYD04) and the Hejiang People's Hospital Special Research Project (2023HJSRY02). None declared. Not applicable. Not applicable. Data sharing is not applicable to this article as no datasets were generated or analysed during the current study.
He et al. (Wed,) studied this question.