Purpose: To study whether there exists a relationship between retinal nerve fiber layer thickness and oxygen saturation and to identify whether perivascular thickness (pRNFL) or quadrant thickness (qRNFL) has a stronger correlation.
Methods: Patients without any detectable ocular abnormality on complete ophthalmic evaluation underwent non-invasive photo-spectrometric retinal oximetry on the Oxymap T1 retinal oximeter (Oxymap hf, Reykjavik, Iceland) and spectral-domain optical coherence tomography to determine if the pRNFL (Spectralis, Heidelberg, Germany) and the qRNFL (RtVue, Optovue) correlated with the measured light intensities and the calculated retinal oxygen saturations.
Results: 30 eyes of 30 patients (120 arterioles, 120 venules) were included in the study. The mean arteriolar saturations (%) were temporal: 84.5 (95% CI 82.7-86.3) and nasal: 93.5 (95% CI 91.0-95.8); venous saturations were temporal: 54.4 (95% CI 52.9-55.9) and nasal: 59.9 (95% CI 58.3-61.4). The mean pRNFL thickness (µm) around arterioles was temporal: 153.1 (95% CI 146-159) and nasal: 109.4 (95% CI 103-115); around venules it was temporal: 147.4 (95% CI 140-154) and nasal: 101.2 (95% CI 95-107). The oxygen saturations correlated significantly with pRNFL (arteriolar r = -0.514; p < 0.001 and venous r = -0.513; p < 0.001) but did not show a significant correlation with qRNFL values.
Conclusion: This relationship between perivascular RNFL could be physiological or artifactual. If physiological, it would explain and form the basis of altered oxygen saturation in a multitude of diseases. If artifactual, it would necessitate the incorporation of a pRNFL correction. However, the results of this study enable us to expect increased saturation in areas of thinner RNFL or vice versa with the current methodology of retinal vessel oximetry as obtained by dual-wavelength photo-spectrometric fundus imaging.
Keywords: Metabolism; RNFL; oximetry; retinal nerve fiber layer; retinal oxygen saturation.