Evaluation of an on-line videodensitometric measurement of the red blood cell velocity in the capillaries of the human nailfold

Comparison of CapiFlow and frame by frame analysis for the assessment of capillary red blood cell velocity

CapiFlow (CF), a new fully computerized system for the measurement of capillary blood velocity (CBV) was compared to manual frame by frame analysis (a) in a model system, and (b) in finger nailfold capillaries recorded on video tape. In the model the overall agreement between the two methods was very good (figure 1), with no significant differences being noted between the two sets of results and the calculated velocities. However, when comparing frame by frame and CapiFlow directly, CapiFlow read on average 4.50 +/- 5.21% higher than frame by frame analysis (figure 2). The in vivo results obtained by the two methods showed similar dynamic changes although some differences between the overall mean CBVs were noted (capillary 1, manual 0.13 +/- 0.59 mm s-1 versus CF 0.12 +/- 0.02 mm s-1, (mean +/- SD), p = 0.354; capillary 2, manual 0.66 +/- 0.23 mm s-1 versus CF 0.47 +/- 0.09 mm s-1, p < 0.001; capillary 3, manual 2.53 +/- 0.73 mm s-1 versus CF 2.35 +/- 0.34 mm s-1, p = 0.062). Further analyses established the optimum settings of delta limit and cross correlation. Investigations into the effects of changes in window size, window distance or video settings on CBV results obtained by CapiFlow, indicated that only settings radically different from the optimum had a significant effect on the results obtained.

An evaluation of skin capillary blood flow determinations in neonates using a computerized videophotometric method

Dynamic capillaroscopy has been shown to give valid measurements of skin capillary blood flow in adults. The aim of this investigation is to evaluate the accuracy of the method when applied to studies in newborn infants. A computerized videophotometric technique was used to measure the skin capillary blood cell velocity (CBV) and the capillary erythrocyte column diameter at 64 standardized locations in 16 nailfold capillaries in five subjects. To estimate the method's reliability, we calculated the coefficient of variation (CV = 1 SD/mean x 100%) for repeated measurements at each location. The CV for CBV determinations was found to be 3.4% (0.9-13%) and the corresponding value for capillary diameter measurements was 6.7% (0.5-12%) (median and range). Although the CBV and diameter values were found to vary along the course of the capillary, the volume flow of blood cells was assumed to be the same in simultaneously studied but different sections of the capillary. The paired CBV and diameter observations were therefore used to calculate blood cell flow values at an average of four different intracapillary locations. To estimate the method's validity, the CV for these repeated flow calculations was determined and found to be 7.5% (3.2-26%). We conclude that the age-specific characteristics in neonates, mainly a high hematocrit and short, irregular skin capillary loops, do not limit the feasibility and accuracy of videophotometric microscopy for dynamic microcirculatory studies in neonates.

Capillary video red blood cell velocimetry by cross-correlation and spatial filtering

A spatial filter capillary red blood cell velocimeter was implemented in video format by means of a linear array of video photometric windows whose number could be varied from two to eight. Two windows of the same video linear window array were used to measure velocity by means of the conventional cross-correlation method. The two methods, spatial filtering and cross-correlation, were compared by means of a computer-controlled video test pattern generator designed to simulate red blood cell flow of variable hematocrit and velocity. It was found that at normal microvascular hematocrits, both methods provide essentially identical data. At low hematocrit, if the simulated red blood cell distribution is such that a cell is always present within the array, spatial filtering data processing provides a more stable measurement than does cross-correlation; however, under the same conditions spatial filtering yields measurement directly related to flux (red blood cells per unit time) rather than to velocity.

Use of digital cross-correlation for on-line determination of single-vessel blood flow in the mammalian kidney

The empirical relationship between erythrocyte velocity (Vrbc) and mean blood velocity (Vblood) was studied in quartz capillaries by television microscopy using the dual-slit technique. A newly designed desktop digital on-line cross-correlator was combined with a computer to determine Vrbc. The accuracy of the digital correlator was tested for velocities ranging from 0 to 3 mm/sec and compared with values determined using an analog tracking correlation device. There was good agreement. Small-bore glass tubes with diameters ranging from 12 to 26 micron were perfused with suspensions of erythrocytes having hematocrits between 10 and 37%. The relationship between mean blood velocity and erythrocyte velocity in these quartz tubes was found to be Vblood = 0.88 Vrbc - 0.11, and was independent of diameter and hematocrit within the range investigated. The mean ratio for Vrbc/Vblood was 1.42 +/- 0.06.

Videomicroscopic method for direct determination of blood flow to the papilla of the kidney

We adapted the technique of videomicroscopy for direct determination of blood flow in individual capillaries of the papilla of the kidney, the ascending vasa recta (AVR) and descending vasa recta (DVR). The papilla was exposed in anesthetized rats and positioned under a video-camera-microscope and viewed under epiillumination. The intravenous infusion of fluorescein-isothiocyanate (FITC)-labeled gamma globulin was combined with fluorescence microscopy to enhance the contrast among plasma, red blood cells and capillary walls. On the television monitor, the walls were clearly outlined, enabling the measurement of capillary diameter. The velocity of red cells (Vrbc) in individual vasa recta was measured using the dual slit technique. From the videotape recorded microscopic image of a vas rectum, two photometric signals were obtained by integrating the light intensity from two electronic "windows" positioned closely together over the same capillary. Red cell velocity was calculated by dividing the distance between the two windows by the time delay between signals. The delay was determined using analog correlation tracking or digital cross correlation techniques. Single vasa recta blood flow was calculated from capillary diameter, Vrbc, and F (Fahraeus factor), which converts Vrbc to average whole blood velocity, Vblood. In quartz capillaries the same size as vasa recta, the ratio F = Vrbc/Vblood = 1.42 +/- 0.06. Total papillary blood inflow and outflow was calculated by multiplying the total number of DVR or AVR times the mean single capillary blood flow for DVR or AVR, respectively.

Direct determination of vasa recta blood flow in the rat renal papilla

Blood flow in vasa recta capillaries of the exposed renal papilla of young antidiuretic rats (n = 18) was determined by an adaptation of the video-photometric technique of Intaglietta. The erythrocyte velocity and capillary diameter in vasa recta (n = 97) were measured at the same location by means of fluorescence video microscopy, with fluorescein-labeled bovine gamma-globulin as a plasma marker. A factor relating erythrocyte velocity to mean cross-sectional blood velocity was determined in vitro to permit the calculation of single vasa recta blood flows from the measured indices, erythrocyte velocity and capillary diameter. Mean blood flow in descending vasa recta was 8.83 +/- 0.96 (SE) nl/min, significantly greater than that in ascending vasa recta, 4.82 +/- 0.34 nl/min. The total numbers of ascending and descending vasa recta at the base of the exposed papilla were also determined. Over 1500 vasa recta were identified as ascending vasa recta or descending vasa recta in electron micrographs of three papillas. At this level in the papilla (2 mm from the tip), there were four ascending vasa recta for each descending vas rectum. From the total numbers of ascending vasa recta and descending vas rectum, single vessel blood flows were converted to total blood flow. Total blood outflow in all ascending vasa recta, 11.3 microliter/min, substantially exceeded total blood inflow in all descending vasa recta, 5.2 microliter/min. The difference between outflow and inflow (6.1 microliter/min) represents an estimate of water by the papillary microcirculation, and is more than adequate to accommodate the known rate of water reabsorption from the collecting ducts of the exposed papilla.

Analysis of network flow distribution: computational aid to minimize experimental expenditure

Several problems in microvascular research, e.g., total blood flow and red cell flux distribution, can only be solved by evaluating the flow distribution to all branches of microvessel networks. We present a method to obtain values of volumetric blood flow and red cell flux in individual networks by taking measurements in an adequately chosen number of the vessels (50-33%) in the network. Actual measurements are not mandatory in all vessels because in a major part of them computational quantification of flows and fluxes is possible. The proposed method consists of a procedure to document in matrix form the connections of a network found in a preparation. By appropriate mathematical manipulations of the matrix, a preselection of measuring sites within the network can be achieved. From the flow values in these sites the remaining vessel flows can be calculated by utilizing the law of mass conservation in all branching points and confluxes. The worst case error of the procedure can be calculated from the errors of the single measurements. The method can also be applied, whenever, for experimental reasons (e.g., poor microscopic focus, flow too rapid or too slow), certain branches do not lend themselves for measurement. A BASIC program can be made available upon request, which carries out all calculations necessary for an appropriate selection in such a short time (about 1 min) so that it can easily be performed before the start of data acquisition.