SUMMARY Erythrocyte deformability was formerly measured by its contribution to whole blood viscosity. It is now more commonly measured by filtration of erythrocytes through, or aspiration into, pores of 3-5 ,um diameter and by the measurement of shear induced erythrocyte elongation using laser diffractometry. Recent improvements in the technology for erythrocyte filtration have included the removal of acute phase reactants from test erythrocyte suspensions, ultrasonic cleaning and reuse of filter membranes, awareness of the importance of mean cell volume as a determinant of flow through 3 gm diameter pores, and the ability to detect subpopulations of less deformable erythrocytes. Measurements of erythrocyte elongation by laser diffractometry, using the Ektacytometer, are also influenced by cell size and need to be corrected for mean cell volume. These advances have greatly improved the sensitivity and specificity of rheological methods for measuring the deformability of erythrocytes and for investigating the mode of action of rheologically active drugs.An earlier review of blood rheology (the science of the deformation and flow of blood) was mainly concerned with the clinical value of measuring whole blood viscosity.' Five years later, rheological measurements in clinical studies are more often concerned with the individual components of blood (erythrocytes, leucocytes, platelets, and plasma) that collectively influence whole blood viscosity. It is therefore appropriate to review the residual role of measurements of whole blood viscosity while concentrating on new developments in the study of erythrocyte rheology.
Measurement of whole blood viscosityThis measurement still has a role as a global test, but its relatively poor sensitivity and specificity, and overdependence on the packed cell volume, have limited its clinical usefulness. Within the normal range there is a linear relation between packed cell volume and the logarithm of whole blood viscosity, but attempts to correct viscosity to a standard packed cell volume of 0-45, using regression lines, are prone to error.2 Regression lines derived from normal blood may not apply to patients' samples,