Low-inflammatory phenotype predicts steroid resistance in optic neuritis: a cohort study and risk stratification model

High-dose glucocorticoids remain the standard acute-phase therapy for optic neuritis (ON), yet a proportion of patients respond poorly to this first-line treatment. This study aims to evaluate the association between baseline hematological indices and steroid sensitivity in patients with first-episode ON, and to develop a practical prediction model to identify patients at risk of steroid resistance. A retrospective cohort study was conducted on 64 patients treated with intravenous glucocorticoid therapy. Patients were stratified into steroid-sensitive or steroid-resistant groups based on the change in best corrected visual acuity (ΔLogMAR). Baseline clinical and hematological parameters were analyzed using logistic regression and receiver operating characteristic curves to determine optimal predictive thresholds. The steroid-sensitive group demonstrated higher baseline neutrophil (NEU) and white blood cell (WBC) counts compared to the resistant group (both P < 0.05). Paradoxically, contrary to their established role as poor prognostic markers for disease severity, elevated baseline NEU and WBC correlated with better visual improvements ( P < 0.01). Univariate logistic regression analysis identified NEU, WBC, lymphocyte (LYM) and symptom-to-treatment interval as significant predictors of steroid sensitivity (all P < 0.05). Further multivariate analysis confirmed WBC as an independent predictor (OR = 1.60, P = 0.015). The derived three-tiered prediction model revealed that patients with a low-inflammatory phenotype (Tier C) exhibited a steroid response rate of only 47.8%, significantly lower than the 86.4% in the high-inflammatory group (Tier A, P = 0.002). High systemic inflammation serves as a predictor of steroid sensitivity in ON. The proposed exploratory stratification model offers a preliminary framework to help identify high-risk patients, supporting heightened clinical vigilance and an earlier consideration for transition to second-line therapies such as plasma exchange for those with a low-inflammatory phenotype.