The relationship between contrast sensitivity and retinal nerve fiber layer thickness in patients with glaucoma.

PURPOSE: To evaluate the correlation between contrast sensitivity (CS) and retinal nerve fiber layer (RNFL) in different areas of vision.

METHODS: This report is a sub-analysis of a prospective, observational cohort study investigating changes in performance-based assessment, vision-related quality of life, and clinical measures in patients with moderate to advanced glaucoma. The study included 161 participants with at least a 2-year history of glaucoma who underwent annual testing for 4 years. Contrast sensitivity was measured using the Spaeth/Richman contrast sensitivity (SPARCS) test, while RNFL thickness (RNFLT) was measured using Cirrus optical coherence tomography (OCT). Statistical analyses were performed to determine correlations between CS and RNFLT; the correlations were calculated for each annual visit, totaling four correlation coefficients for each patient over the course of 4 years.

RESULTS: The SPARCS score in the left upper area of vision correlated the most strongly with the RNFLT of the inferior quadrant for both eyes at each annual visit, specifically in the seven o'clock sector for the left eye and the six o'clock sector for the right eye (p < 0.05). There were no discernible trends for the correlations between the other areas of CS and RNFL quadrants or clock hours over the 4 years of the study. Linear regression between the SPARCS total score and average RNFLT showed a significant direct correlation at each visit (p < 0.01).

CONCLUSIONS: Contrast sensitivity in the left upper area of vision for both eyes correlated most strongly with the thickness of the inferior quadrant of the RNFL. These fibers project to the temporal portion of the right occipital lobe, implying a potential center for contrast perception in this area. The longitudinal nature of the study suggests that CS may be a predictive tool for changes in RNFL in patients with glaucoma. Despite this finding, retinal damage and its relationship to CS was diffuse. In addition, SPARCS was shown to predict RNFLT. Further research is warranted to understand how CS can be used as a tool in the clinical setting.