Measurement of retinal oxygen saturation in patients with chronic obstructive pulmonary disease

Purpose. There is growing evidence that disturbances in retinal oxygenation may trigger ocular diseases. New instruments allow for the noninvasive measurement of retinal oxygen saturation in humans. The present study was designed to investigate the retinal oxygen saturation in patients with chronic obstructive pulmonary disease (COPD). This was also done in an effort to test the validity of retinal oxygenation measurements with a retinal vessel analyzer. Methods. In all, 16 patients with severe COPD grade 4 who were on long-term oxygen treatment were included in the study. For each patient two identical study days were scheduled. Measurements of retinal arterial and venous oxygen saturation were done using a commercially available instrument for retinal oxygen analysis. Peripheral arterial oxygen saturation values were analyzed with pulse oximetry and via a capillary blood sample drawn from the earlobe. Measurements were performed during oxygen treatment and during a period without oxygen supplementation. Analysis of all images for retinal oxygen saturation quantification was done by a masked investigator. Analysis was done using Pearson's correlation and a multivariate regression model. Results. Arterial and venous retinal oxygen saturation decreased significantly after the cessation of the oxygen therapy. The arteriovenous oxygen difference was unchanged while breathing ambient air or pure oxygen-enriched air. With both Pearson's correlation and the multivariate model, we found significant positive correlation coefficients between retinal arterial and peripheral arterial oxygen saturation as assessed with pulse oximetry as well as between retinal arterial and peripheral arterial oxygen saturation measured in blood samples. The change of oxygen saturation after discontinuation of oxygen supplementation showed a good correlation between retinal arterial oxygen saturation and peripheral arterial oxygen saturation (r=0.53, P < 0.05). Reproducibility on the two study days was high. DISCUSSION. The present study shows a good correlation between retinal arterial and peripheral arterial oxygen saturation indicating good validity of the technique.