Serum Albumin and Inflammation–Nutrition Burden in Adult Hospitalized Atopic Dermatitis: A Case–Control Study with Complementary Genetic Evidence

Abstract

Background Atopic dermatitis (AD) is a chronic inflammatory skin disease that may involve systemic inflammation and metabolic disturbance. Serum albumin (Alb) is a routinely available negative acute-phase protein, but its association with hospitalized AD and sphingomyelin-related metabolic signals remains unclear. Methods This single-center retrospective case–control study included 814 adults (381 hospitalized patients with AD and 433 lipoma surgical inpatient controls). Alb was evaluated continuously, per standard-deviation increment, and by quartiles using multivariable logistic regression, restricted cubic splines, sensitivity analyses, and propensity-score matching. Two-sample Mendelian randomization (MR) provided complementary genetic evidence, and exploratory two-step MR examined a prespecified sphingomyelin-related phenotype. Results Alb was lower in patients with AD than in controls (38.9 ± 4.6 vs 41.2 ± 5.0 g/L; P<0.001). In the fully adjusted model, each 1-g/L increase in Alb was associated with lower odds of hospitalized AD status (OR=0.92, 95% CI: 0.89–0.96; P<0.001), with consistent findings across standardized, quartile, spline, sensitivity, and matched analyses. In the primary IVW analysis, genetically predicted higher Alb was associated with lower AD risk (OR=0.693, 95% CI: 0.519–0.926; P=0.013), whereas complementary MR estimates were not uniformly significant. Exploratory two-step MR showed associations of higher genetically predicted Alb with lower levels of the prespecified sphingomyelin-related phenotype and of higher sphingomyelin-related levels with greater AD risk. Conclusion Lower serum Alb was consistently associated with hospitalized adult AD status and may provide adjunctive information on systemic inflammation–nutrition burden in this inpatient setting. MR findings were supportive but not uniformly consistent, and the sphingomyelin analysis was exploratory rather than evidence of mediation. Prospective longitudinal studies incorporating standardized AD severity scores and serial Alb measurements are warranted.