Metabolic changes in the visual cortex are linked to retinal nerve fiber layer thinning in multiple sclerosis

Abstract

Objective: To investigate the damage to the retinal nerve fiber layer as part of the anterior visual pathway as well as an impairment of the neuronal and axonal integrity in the visual cortex as part of the posterior visual pathway with complementary neuroimaging techniques, and to correlate our results to patients' clinical symptoms concerning the visual pathway.

Design, subjects and methods: Survey of 86 patients with relapsing-remitting multiple sclerosis that were subjected to retinal nerve fiber layer thickness (RNFLT) measurement by optical coherence tomography, to a routine MRI scan including the calculation of the brain parenchymal fraction (BPF), and to magnetic resonance spectroscopy at 3 tesla, quantifying N-acetyl aspartate (NAA) concentrations in the visual cortex and normal-appearing white matter.

Results: RNFLT correlated significantly with BPF and visual cortex NAA, but not with normal-appearing white matter NAA. This was connected with the patients' history of a previous optic neuritis. In a combined model, both BPF and visual cortex NAA were independently associated with RNFLT.

Conclusions: Our data suggest the existence of functional pathway-specific damage patterns exceeding global neurodegeneration. They suggest a strong interrelationship between damage to the anterior and the posterior visual pathway.

Figure 1. ¹H-MRS voxel placement.

Visual representation of typical voxel placement for MR spectroscopy. In each patient, NAA concentrations were measured in a visual cortex voxel (VC) and two normal-appearing white matter voxels (NAWM).

Figure 2. Correlation of RNFLT with BPF and ¹H-MRS parameters.

a) Depicted is the average RNFLT, every symbol representing a single eye examined together with the corresponding BPF values. The symbols represent the patient's previous history of optic neuritis (open circles – no previous optic neuritis, grey squares - unilateral optic neuritis, black triangles – bilateral optic neuritis) A linear correlation function was calculated by a Generalised Linear Model to account for intra-individual inter-eye relationships (p = 0.001). b) Mean BPF was calculated for three groups that were defined based on their previous history of optic neuritis (white bar– no previous optic neuritis, grey bar - unilateral optic neuritis, black bar – bilateral optic neuritis). The (-) symbol indicates a trend, but a missing significant correlation of group differences as calculated by ANOVA (p = 0.055). Error bars represent 2× standard error of the mean (SEM). c) RNFLT averages are shown in relation to corresponding NAA concentrations in the visual cortex (VC). The symbols are coded as in a). The correlation is significant (p = 0.047). d) Mean visual cortex voxel (VC) NAA and the significance of group differences was calculated for optic neuritis groups as in b). The asterisk indicates statistically significant (p = 0.046) group differences. Error bars represent 2× standard error of the mean (SEM). e) RNFLT averages are shown in relation to corresponding NAA concentrations in normal-appearing white matter (NAWM). The symbols are coded as in a). No significant correlation was found (p = 0.531). f) Mean NAA in normal-appearing white matter (NAWM) and the significance of group differences was calculated for optic neuritis groups as in b) (p = 0.429). Error bars represent 2× standard error of the mean (SEM).