Reproducibility of retinal vessel hemoglobin oxygenation saturation in healthy subjects

Abstract

Abstract Purpose Only recently systems for the measurement of hemoglobin oxygenation saturation in retinal vessels became commercially available. In the present study a fundus camera equipped with a dual wavelength transmission filter and a CCD camera was used (Imedos, Jena, Germany). Two monochromatic fundus images at 548 and 610 nm were recorded simultaneously and the oxygen saturation (SaO2) was calculated based on the optical density ratio. Methods A study on test/retest, short‐term and day‐to‐day reproducibility of the technique was done in 14 healthy subjects. To assess test/retest variability fundus images were taken at each measurement cycle. To assess short‐term reproducibility 2 such cycles of 3 fundus images were done within 15 minutes. Finally, this cycle of measurements was done on the next day to assess day‐to‐day reproducibility. Data were evaluated separately for retinal branch arteries and branch veins. The reproducibility was calculated using the standard deviation (SD) of the measurements. Results In retinal branch veins the SaO2 was 49±9% (range 32‐69%, 38 vessels) and in retinal branch arteries 91±8% (range 77‐112%, 40 vessels). The test/retest SD was 5.4±3.7% in retinal veins and 3.0±2.3% in retinal arteries. The short term SD was 4.7±3.7% in retinal veins and 2.9±2.1% in retinal arteries. The day‐to‐day SD was 6.2±4.1% in retinal veins and 4.0±2.8% in retinal arteries. Conclusion Our data indicate that the test/retest, short‐term and day‐to‐day reproducibility of the technique is acceptable. Hence, the system may be adequate to follow oxygen saturation during disease processes. The data obtained in retinal branch arteries indicates, however, that the validity of measurements requires further studies.