Abstract
<jats:p>Background/Objectives: Psoriasis and atopic dermatitis have historically been conceived as opposite immunological poles (Th17 versus Th2), yet molecular overlap and increasing clinical coexistence challenge this dichotomy. We aimed to adjudicate, using a multivariate approach, among three nosological architectures: a shared inflammatory–metabolic spectrum with additive coexistence (H1), distinct phenotypes with synergy (H2) and predominant overlap (H0). Methods: Cross-sectional analysis of a selected UK Biobank subsample (160,785 participants), classified by L20/L40 codes as control, isolated atopic dermatitis (n = 12,859), isolated psoriasis (n = 15,146) and coexistence (n = 1,068). Haemato-inflammatory indices, a biochemical–metabolomic panel (including GlycA) and Olink inflammatory proteins were evaluated using Cliff&#039;s delta, quantile regression with an interaction term, MANOVA/PERMANOVA, supervised classifiers, Gaussian mixture modelling, dominant-factor ordination with the Jonckheere–Terpstra test and equivalence testing (TOST; margin |δ| = 0.147). Restricted and broad control definitions were adopted. Results: Effect sizes were largely negligible (46/48 contrasts with |δ| &lt; 0.147 under the restricted control; 38/48 formally equivalent). No psoriasis × atopic dermatitis interaction survived false-discovery-rate control (0/16), and the global multivariate interaction, although significant, was trivial (Pillai&#039;s trace = 0.0005). The dominant inflammatory factor was monotonically ordered (control &lt; atopic dermatitis &lt; psoriasis ≈ coexistence; Jonckheere |z| = 8.6; p &lt; 0.001). Multivariate separation was detectable yet negligible (PERMANOVA R² = 0.6%; multiclass AUC 0.53), and latent classes did not coincide with clinical labels (adjusted Rand index = 0.004). Conclusions: The evidence favours a shared inflammatory–metabolic spectrum with approximately additive coexistence, superimposed on predominant between-group indistinguishability. Coexistence lies at the severe extreme of a common axis rather than constituting a third, emergent phenotype.</jats:p>