Examinations of blood flow were performed on days 1, 5, 15, and 19, on oral administration of the study medication, and at 15, 30, 45, 60, 75, 90, 105, 120, 150, and 180 minutes after oral administration of the study medication. Retinal blood flow was measured with scanning laser Doppler flowmetry (SLDF; 670 nm; Heidelberg Flowmeter [HRF]; Heidelberg Engineering, Heidelberg, Germany).
14 Acquired perfusion images were analyzed with automatic full-field evaluation of the perfusion images (SLDF-AFFPIA).
15 Principles of the measuring technique have been described elsewhere in detail.
16 17 18 Briefly, the laser Doppler frequency shift was measured by SLDF in each of 16,384 points in a retinal area of 2.7 mm × 0.7 mm within 2 seconds. Confocal optics of this device registered only the capillary blood flow of the superficial retinal layer of 300 μm. Detection of laser Doppler signals from deeper layers (choroid) was excluded because of confocal characteristics of the optics. Signals from vitreous or choroid were not detected because they were out of focus. A map of the retinal blood flow was generated, encoded by the laser Doppler shift. Spatial resolution of the device was 10 μm × 10 μm. In the retina, the coefficient of reliability of blood flow measurements was calculated to 0.7 to 0.8 with the use of a flowmeter (HRF; Heidelberg Engineering).
18 19 20 21 Capillary retinal blood flow was recorded in a juxtapapillary area localized 2 to 3 mm temporally beside the optic nerve head. These regions were chosen because subjects did not show relevant fixation problems during measurement. The HRF scanned an intensity matrix of 256 points × 64 lines × 128 times using a repetition rate of 4000 Hz. Backscattered intensities of each scanned point were obtained as a function of time, resulting in 16,384 intensity time curves. Collected intensity data of each retinal point of measurement were analyzed by discrete fast Fourier transformation calculating the frequency laser Doppler shift for each point of measurement. SLDF-AFFPIA was used to analyze the perfusion images off-line after the experiments. It calculated the laser Doppler frequency shift and the hemodynamic parameter flow of each pixel according to the theory of Bonner and Nossal.
22 For valid estimation of retinal blood flow by HRF, some assumptions must be fulfilled: adequate brightness, no artificial movement, and a Doppler shift lower than 2000 Hz. To meet these requirements, the resultant perfusion image was processed by the SLDF-AFFPIA with respect to underexposed and overexposed pixels, saccades, and retinal vessel tree. In the first step of the statistical analysis, the operator marked saccades and the location of the rim area. In the second step, capillaries and vessels of the retinal vessel tree were identified automatically by a vessel detection algorithm, based on the intensity and the perfusion image. After these procedures, retinal vessels with a diameter larger than 30 μm, underexposed or overexposed pixels, and saccades were excluded automatically from the perfusion image. These processes led to a perfusion map with vessels smaller than 30 μm in diameter, without lines caused by saccades, and without pixels of inadequate reflectivity. Based on this processed perfusion map and all flow values, SLDF-AFFPIA automatically calculated the mean flow, SD, and cumulative frequency distribution curve in the scanned retinal area.
Figure 1depicts the retinal area of one eye with the cumulative frequency distribution curve and the histogram of all flow values within this area.