Predictors of retinal atrophy in multiple sclerosis: A longitudinal study using spectral domain optical coherence tomography with segmentation analysis

Background: Multiple sclerosis is an inflammatory demyelinating disease characterized by progressive axonal loss affecting mainly the inner retinal layers. Optical coherence tomography (OCT) provides in-vivo quantification of the retinal layers and allows measuring progressive retinal changes. Our objective was to assess the longitudinal changes in the retina using spectral domain OCT (SDOCT) and to identify independent predictors affecting retinal thinning in MS patients. Methods: A prospective study in a tertiary care MS center was conducted to study the longitudinal retinal changes in MS patients. All subjects underwent baseline and follow up OCT assessment with segmentation analysis. Regression analysis was performed to assess clinical factors (age, sex, disease duration, history of optic neuritis before baseline, non-ocular clinical relapses) and MRI disease activity during the follow-up period. Results: The study included 102 MS patients with a mean follow-up duration of 3.9 +/- SD years. At the last follow-up assessments, there were significant thinning of the average macular thickness (AMT) (p < .001), macular nerve fiber layer (MRNFL) (p < .001), ganglion cell-inner plexiform layer (GCIPL) (p < .001), and the peripapillary nerve fiber layer (PRNFL) (p < .001), compared to baseline. Early disease duration up to 10 years was associated with thinning of AMT, PRNFL, and GCIPL, while longer disease duration ( > 15 years) was associated with only GCIPL thinning. Prior optic neuritis was predictive of more thinning of PRNFL (p = < .01), while MRI activity and female gender were significantly associated with more MRNFL thinning (p = < .01). Conclusion: MS is associated with longitudinal thinning affecting AMT inner retinal layers (MRNFL, GCIPL, PRNFL). Early disease duration, female gender, MRI activity, and prior optic neuritis were predictive of faster rate of neuro-axonal loss. This may have implications in the design of future therapeutic trials.