Peripapillary choroidal thickness in glaucoma measured with optical coherence tomography

As choroidal changes have been suggested in glaucoma, we examined peripapillary choroidal thickness (CT) in patients with and without primary open-angle glaucoma (POAG) using spectral-domain optical coherence tomography (SD-OCT). We collected measurements retrospectively on 70 eyes of 70 patients consecutively undergoing SD-OCT. POAG (n = 31) and suspect eyes (n = 39) had two reliable and repeatable Humphrey 24-2 visual fields with glaucoma hemifield test outside or within normal limits, respectively. A 360-degree peripapillary scan was performed using the standard protocol for retinal nerve fiber layer (RNFL) assessment. Using provided software, two independent masked investigators manually segmented CT as the area of visible choroidal vasculature. Agreement between investigators was determined using Lin's concordance correlation coefficient (CCC). A single masked observer determined clock hours of parapapillary atrophy (PPA) and the presence of ßPPA for each optic nerve quadrant. Correlation between RNFL and CT was assessed; two-sample t-tests were used to determine differences in RNFL and CT between POAG and suspect eyes; and linear regression was used to model changes in RNFL and CT. We found that independent measurements of CT by two observers were highly correlated (Lin's CCC for global CT; ρ_c = 0.93, p < 0.001). RNFL and CT measurements were not significantly correlated for any peripapillary location (|r| ≤ 0.15, p ≥ 0.22). Global CT (ß = -1.94, 95% confidence interval [CI] -2.76, -1.13) but not RNFL thickness (ß = -0.18, 95% CI -0.58, 0.22) decreased significantly with age. Compared to suspect eyes, eyes with POAG had significantly thinner RNFL measurements at all locations (p ≤ 0.005) but CT measurements did not differ between groups for any location (p ≥ 0.13). Adjusting for glaucoma status and age, total (ß = 3.15 95% CI -0.24, 6.53) and ß clock hours of PPA (ß = 1.33, 95% CI -1.72, 4.38) were not significantly associated with global CT; the spatial distribution of PPA was not associated with underlying CT, though PPA was graded subjectively and may have been subject to spatial mismatch with a singular peripapillary eccentricity on SD-OCT. We conclude that eyes with POAG did not demonstrate reduced CT nor was there a correlation between RNFL and CT maps. This study does not support the use of CT assessment in glaucoma diagnosis or management.