Heidelberg retinal flowmetry: factors affecting blood flow measurement.

Abstract

To evaluate factors affecting Heidelberg retinal flowmeter (HRF) measurements of retinal and optic nerve head blood flow in human subjects. The angle of incidence between laser beam and fundus, and camera distance from the eye, were evaluated for their effect upon measures of blood volume, velocity, and flow in a single 100 x 100 x 400 microns volume of temporal peripapillary retinal tissue in normal volunteers. Both intra and intersession reproducibility of these measures were studied. Intersession data were obtained by taking one image per week for 4 weeks. Finally, the intersession haemodynamic data were examined in the entire image (640 x 2560 x 400 microns), using histograms of pixel by pixel blood flow. Measures of blood volume, velocity, and flow from a single anatomical site were unaffected by laser beam to fundus angle of incidence (n = 12). As camera distance from the eye was increased (from 2 to 5 to 7 cm), flow measurements showed increasing individual changes, despite unaltered measured vessel lengths and constant overall mean flow (n = 14). The coefficient of variation for two intersession images of optic nerve head blood flow averaged 7% (n = 20); in contrast, the 4 week intersession coefficient of variation averaged 30% (n = 15). Intersession reproducibility was increased by using flow histograms from the entire image: the coefficients of variation averaged 16% for total flow and 17% for flow in the pixel of median flow. HRF measures of flow are independent of the laser beam to fundus angle of the incidence and dependent upon camera distance from the eye. Intersession reproducibility is best using pixel by pixel analysis of the entire image.