Retinal layers prognosticate cognitive progression independent of relapse activity in multiple sclerosis.

BACKGROUND: Optical coherence tomography (OCT) quantifies retinal neuroaxonal loss related to neurodegeneration in people with multiple sclerosis (pwMS) and is associated with physical and cognitive disability. However, no data exist on the relationship between OCT metrics and cognitive progression independent of relapse activity (cognitive PIRA). OBJECTIVE: To assess if OCT can prognosticate cognitive PIRA events in pwMS. METHODS: Prospective study with baseline OCT (peripapillary retinal nerve fiber- (pRNFL), macular ganglion cell-inner plexiform- (mGCIPL) and inner nuclear layers (mINL)) and brief international cognitive assessment for MS (BICAMS) at yearly visits. Cognitive PIRA in each test was defined as more than 10% decrease compared to previous visit (reference) and confirmed 12 months later (confirmation visit), without relapses 90 days before or 30 days after the event and confirmation visits. We employed Cox regression models adjusting for age, baseline cognitive test, education, and treatment. RESULTS: 98 pwMS followed for a median of 5 years (63% female, 81% relapsing-remitting, age: 50.5 ± 11.6 years, median EDSS: 3.0). Each 1 µm increase in pRNFL thickness was associated with a 6% lower risk of Symbol Digit Modalities Test (SDMT) PIRA events (Hazard ratio (HR) = 0.94, p = 0.027). Each 1 µm increase in mGCIPL thickness was associated with a 6% lower risk of Verbal Learning and Memory Test (VLMT) PIRA (HR = 0.94, p = 0.040). Each 1 μm increase in mINL thickness was associated with 20% lower risk of VLMT-PIRA (HR = 0.80, p = 0.030) and 50% higher risk of SDMT-PIRA events (HR = 1.50, p = 0.049). CONCLUSION: Retinal layers can be sensitive, patient-friendly prognostic markers of cognitive PIRA in pwMS.