Margination of leukocytes in blood flow through small tubes.

Leukocyte margination in the vessels of the microcirculation has been attributed to a flow-dependent interaction with red cells. To determine the extent of this effect, experiments with human blood were done in 100- to 180-micron tubes to detect changes in cell distribution as a function of hematocrit and flow rate. Using a flow visualization technique, the leukocyte concentration distribution was determined in 45% ghost cell suspensions. Migration of cells toward the wall was observed at centerline velocities greater than 1 mm sec-1 and increased with increasing flow rate. The effect was probably due to a more rapid inward migration of ghosts than leukocytes because of fluid inertia and cell density differences. Experiments were therefore carried out in whole blood at hematocrits from 20 to 60%, measuring the number concentration of leukocytes and erythrocytes within the tube, nt, and comparing it to that in the infusing reservoir, no, (Fahraeus effect). At mean tube shear rates G less than 100 sec-1, nt/no less than 1 for both leukocytes and erythrocytes showing net migration of cells away from the wall, although at nearly all hematocrits there was an enrichment of leukocytes relative to erythrocytes in the tubes. At G less than 50 sec-1, nt/no remained less than 1 for erythrocytes but increased to greater than 1 for leukocytes showing migration toward the wall, the increase being greatest at 20% hematocrit in the 100-micron tubes. The nature of the effect was revealed by cine films which showed that, as the flow rate decreased, erythrocytes formed rouleaux which migrated inward creating a core and displacing leukocytes to the periphery. In control experiments using washed blood cells in phosphate buffer-albumin, nt/no less than 1 for both leukocytes and erythrocytes at all G and hematocrits, and leukocytes were now distributed. Cine films of washed blood confirmed that, in the absence of rouleaux, no significant inward migration of erythrocytes occurred.