Baseline Arterial Blood Saturation is Predictive of Venous Response to Hyperoxia

Abstract

Abstract: : Purpose: To assess the relationship between baseline arterial blood oxygen saturation and retinal venous response to hyperoxia. Methods: Thirteen healthy subjects, with a mean age of 26.54 ± 3.82 years, were evaluated before and after 100% oxygen breathing. Three fundus oximetry images focusing on the major temporal retinal vessels adjacent to the optic disc, were obtained while breathing room air. After five minutes of 100% oxygen breathing, two additional images were taken. Oxygen breathing was maintained during imaging. Retinal vessel blood oxygen saturation (O₂sat) was measured by a newly developed digital retinal oximeter. The system takes a single fundus image and divides it optically into four identical images. Each image is then filtered to isolate a different narrow band wavelength (∼ 5 nm): one image with a wavelength at which the light absorption of oxyhemoglobin and deoxyhemoglobin is equal (589 nm), and the other three with wavelengths at which the relative absorption of the 2 species of hemoglobin differ by the greatest ratio (542, 558, and 605 nm). Each isolated wavelength image is digitized to determine light intensity. Images were analyzed using average reflectance of tissue surrounding blood vessels in the calculation of optical density (OD). OD, therefore, becomes ln(tissue/blood vessel). The ratio between an oxyhemoglobin sensitive OD and isobestic wavelength OD is calculated. This ratio is the optical density ratio (ODR), and is proportional to oxygen saturation. Baseline arterial oxygen saturation was compared to the change in venous ODR by linear regression. Results: Venous ODR was reduced by 100% oxygen breathing by an average of 57.2% (57.2% ± 17.9% SE). Baseline arterial blood saturation was statistically significantly correlated to the change in venous ODR (p = 0.0074, r² = 0.453). Conclusions: Room air levels of retinal arterial blood saturation predict venous response to hyperoxia.

Keywords: imaging/image analysis: clinical • metabolism • image processing