A clinically actionable stratification of bullous pemphigoid using a disease activity score reveals a targetable inflammatory endotype

BACKGROUND: Bullous pemphigoid (BP) exhibits significant clinical heterogeneity. The molecular drivers underlying these distinct phenotypes remain poorly understood, hindering the development of precision medicine approaches. OBJECTIVES: To correlate clinical phenotypes, defined by the Bullous Pemphigoid Disease Area Index (BPDAI), with molecular endotypes using serum proteomic profiling. METHODS: This cohort study enrolled 143 patients with BP hospitalized at a tertiary centre between January 2024 and June 2025. Patients were stratified as having inflammatory BP (I-BP; BPDAI urticaria/erythema subscore ≥ 5) and pauci-inflammatory BP (PI-BP; BPDAI urticaria/erythema subscore < 5) based on the first quartile of the subscore. Clinical features, laboratory parameters and treatment outcomes were compared. Serum proteomic profiling (Olink Target 96 Inflammation) was performed in a representative subset (n = 37). RESULTS: Of the 143 enrolled patients [mean (SD) age 73.9 (11.2) years; 84 (58.7%) men], 108 (75.5%) were classified as having I-BP and 35 (24.5%) as having PI-BP. Compared with those with PI-BP, patients with I-BP were younger, had more severe pruritus and had significantly elevated peripheral eosinophil counts (0.6 vs. 0.2 × 109 cells L-1; P = 0.001) and serum interleukin (IL)-5 levels (19.8 vs. 3.9 pg mL-1; P = 0.007). Despite receiving more intensive therapy, patients with I-BP required a longer median hospital stay (8 vs. 7 days; P = 0.03). Proteomic analysis identified eight differentially expressed proteins. IL-13, thymic stromal lymphopoietin, oncostatin M and monocyte chemoattractant protein-4 were upregulated in I-BP and showed positive correlation with the urticaria/erythema subscore (P < 0.05) but not with the erosions/blisters subscore. IL-13 demonstrated the highest diagnostic accuracy for the I-BP endotype (area under the curve = 0.87). Functional enrichment analysis confirmed differential activation of type 2 inflammation and the Janus kinase/signal transducers and activators of transcription (JAK/STAT) signalling pathway in the I-BP group. CONCLUSIONS: The BPDAI urticaria/erythema subscore serves as a reliable clinical surrogate for a T helper 2 (Th2)-driven molecular endotype in BP. The identification of a specific Th2 ligand-JAK/STAT signalling circuit in the inflammatory phenotype supports a stratified treatment approach, suggesting that patients with a high inflammatory burden may benefit from targeted type 2 immunotherapies or JAK inhibitors, which warrants further prospective validation.