To determine the inter-eye differences (IED) of ganglion cell-inner plexiform layer (GCIPL) thickness associated with multimodal visual dysfunction after demyelinating optic neuritis (ON).

ON affects up to 70% of people with multiple sclerosis (pwMS). Thinning of GCIPL is characteristic of demyelinating ON. Studies investigating multimodal visual dysfunction after demyelinating ON and its relationship with GCIPL IED are lacking.

In this cross-sectional single-center study, patients with a history of unilateral demyelinating ON and healthy controls underwent optical coherence tomography, best-corrected visual acuity (BCVA), 1.25% and 2.5% low-contrast letter acuity (LCLA), standard automated perimetry, and color vision testing. Subjects with bilateral ON history or subclinical optic neuropathy in the non-ON eye were excluded. IED were calculated for GCIPL thickness, BCVA, LCLA, and visual field mean deviation (VFMD). ROC curve analysis established optimal GCIPL IED associated with likelihood of persistent multimodal visual dysfunction after ON.

A total of 31 participants with ON history and either a diagnosis of MS or clinically isolated syndrome (average age: 40±13 years, 90% female), and 15 healthy controls (average age: 33±6 years, 67% female) were included. ON patients had higher GCIPL IED (median: 12%, Q1-Q3: 5%-16%) than healthy controls (median: 0.012%, Q1-Q3: 0.002%-0.015%) (p<0.001). In the ON cohort, GCIPL IED at a threshold of 27% predicted BCVA dysfunction (AUC=0.93, p=0.001). A threshold of 11% corresponded to 1.25% LCLA IED of ≥5 letters (AUC=0.78, p=0.005), while a 13% cut-off predicted 2.5% LCLA IED of ≥5 letters (AUC=0.85, p<0.005). A threshold of 15% predicted VFMD IED of >3 dB (AUC=0.87, p=0.008). 12% cut-off predicted dyschromatopsia in ON eyes (AUC=0.82, p=0.004). Protan and deutan were the most prevalent dyschromatopsias in ON eyes.

GCIPL IED is a promising biomarker of persistent multimodal visual dysfunction after demyelinating ON. LCLA and color vision deficits were associated with lower numerical thresholds for ganglion cell loss.