Aging does not affect the relationship between OCT and clinical or MRI outcome measures

Background: Neurodegeneration in multiple sclerosis (MS) is intertwined with normal aging. The standard approach to address this overlap is statistical correction for the confounding effects of age, but it remains unclear to what extent this successfully isolates disease-specific processes from those related to aging. Objectives: We aimed to investigate how age influences cross-sectional associations between retinal measures—retinal nerve fiber layer (RNFL) and ganglion cell-inner plexiform layer (GCIPL)—and both MRI and clinical outcomes in MS. Methods: We compared two MS cohorts: (1) a cohort selected on birth year (Project Y) and (2) an age-heterogeneous MS cohort (PrograMS). Multilevel linear regression models were used to assess cross-sectional associations. Models were adjusted for sex, history of optic neuritis, and disease duration. In the heterogeneous cohort, age was included as an additional covariate. Results: In both the birth-year cohort (N = 378) and the heterogeneous cohort (N = 206), we observed significant associations between RNFL and GCIPL thickness and several MRI measures including whole brain volume, white matter volume, lesion volume, and thalamus volume, as well as clinical outcome measures, such as Expanded Disability Status Scale (EDSS), Symbol Digit Modality Test (SDMT). While effect sizes varied, the direction remained consistent. Including age as a covariate did not substantially change associations. For example, both RNFL and GCIPL thickness were associated with EDSS in the birth-year cohort (RNFL: β = -0.020, p < 0.001; GCIPL: β = -0.017, p = 0.005) and in the age-heterogeneous cohort (RNFL: β = -0.024, p = 0.010; GCIPL: β = -0.033, p < 0.001), and showed consistent directionality in their associations with MRI-derived brain volumes. Discussion: Our findings reveal significant cross-sectional associations between RNFL and GCIPL thickness and several MRI and clinical measures, with consistent directionality across both cohorts, and largely independent of age.