Hematological inflammatory markers as predictors of atherosclerotic progression in patients with psoriasis: A retrospective case–control study from the SH-YIMED database

Abstract

Chronic inflammation drives cardiovascular comorbidity in psoriasis. Clinical-only risk models have limited utility, whereas the predictive value of blood inflammatory markers for psoriatic atherosclerosis is unconfirmed. This hospital-based case–control study enrolled 636 patients with psoriasis who underwent vascular ultrasonography, and 134 matched pairs were obtained through 1:1 propensity score matching. We analyzed primary and exploratory hematological inflammatory markers, examined their links to atherosclerosis using logistic regression and supplementary stratified and nonlinear analyses, and constructed risk prediction models that were assessed by area under the curve, decision curve analysis, net reclassification improvement, and integrated discrimination improvement. Neutrophil-to-lymphocyte ratio (NLR) (OR = 16.33, 95% confidence interval = 7.54–35.35, P < .001) and systemic immune-inflammation index (SII) (OR = 20.78, 95% confidence interval = 7.92–54.53, P < .001) were stronger independent predictors of atherosclerosis than C-reactive protein and erythrocyte sedimentation rate, with significant age interactions for NLR (interaction OR = 7.45, P = .018) and SII (interaction OR = 28.03, P = .008). Models incorporating NLR (area under the curve = 0 .894, integrated discrimination improvement = 0.439, net reclassification improvement = 1.339) or SII (area under the curve = 0.891, integrated discrimination improvement = 0.413, net reclassification improvement = 1.281) exhibited favorable calibration and a higher clinical net benefit across decision curve analysis thresholds of 0.05–0.75. The combined C-reactive protein + NLR + SII model attained the highest area under the curve of 0.906. NLR and SII improved cardiovascular risk prediction for patients with psoriasis with atherosclerosis, enabling effective risk stratification.