To elucidate the pathophysiologic significance of red blood cell (RBC) filterability, we measured RBC rheology with our own designed nickel mesh with 3-microns pores, smaller than the previously used 5-microns pores. Vertical and cylindrical pores with no pore coincidence were regularly distributed across the filter, the pore entrances of which showed a round and rather smooth transition to the pore inside. An advantage of the nickel mesh is the repeated use (at least 100 times) of the same filter possible after ultrasonic washing. A very low concentration of RBC, i.e., 3 x 10(4) cells per cubic millimeter (hematocrit value of approximately 0.3%), was sufficient for a typical test to examine RBC filterability. The filtration of the dilute RBC suspension was not influenced by contaminating or added leukocytes up to a leukocyte count of approximately seven cells per cubic millimeter; therefore, measurements can be performed using conventionally washed RBCs. This may be practically relevant to routine use, such as in a clinical laboratory. As compared with filtration through 5-micron pores, filtration through 3-micron pores was found to be very sensitive in detecting major determinants of RBC deformability, particularly, changes in viscoelastic properties of the cell membrane, surface area/volume ratio of the cell, perturbing effects of lysophosphatidylcholine, and osmolality of the medium. The 3-micron filtration method revealed a marked impairment in the filterability of Heinz body-containing RBCs from patients with unstable hemoglobin (Hb) disease (Hb Yokohama). Thus, 3-micron-filtration measurements may contribute to several subfields of hematology.