Blood velocity measurement in human conjunctival vessels

The bulbar conjunctiva is one of the few areas in which blood flow in the peripheral vasculature can be directly and noninvasively observed in the human. Although extensive literature exists describing morphological changes which correlate with a variety of systemic diseases in this vasculature, little quantitative data is available on hemodynamics in either normal or abnormal states. The hemodynamic data available are primarily subjective assessments of "low flow." Approaches to place the subjective assessment on more quantitative grounds have usually been based on photographic techniques that have intrinsic inadequacies. The objective of the work reported here was to develop a system capable of providing sequential blood velocity data potentially useful for providing quantitative information on blood flow and its change in the microvessels of the human conjunctiva. The method that has evolved uses a standard Zeiss slit-lamp to image a subject's conjunctival vessels by using a 1-inch Newvicon TV camera with an electronic magnification of 2x. The video image is simultaneously recorded on a video tape recorder (VTR) to an overall system magnification of approximately 4 microm/raster line. The data acquisition phase requires approximately 5 minutes of patient time, whereas the actual determination of blood velocity in individual vessels is done offline through a modification of the dual-slit videodensimetric method. Two independently controllable video cursors are placed axially over the vessel image with the VTR in the still-frame mode. For each consecutive video field, the position of two reference points on the vessel and the position of each cursor relative to these and to each other are encoded into a computer to track the moving image caused by normal eye movement. The computer then determines new cursor coordinates to ensure a constant position within the vessel. The electrical signals obtained for each cursor site and for each video field are cross-correlated to yield the average blood velocity over the sampled time interval. The system has been calibrated in vitro from 0.2 to 2.5 mm/sec, evaluated in experimental animals, and used to measure blood velocity (0.3 to 1.5 mm/sec) in human conjunctival venules with diameters ranging from 20 to 50 microm. At this writing, blood velocity has been recorded during a period of about 3 months in the same vessel of several postmyocardial infarction patients. Thus, the method appears suitable for determining sequential changes in small vessel blood flow in patients over extended periods of time.