The aim of the present study was to investigate the effects of manipulating hematocrit by different methods (acute exercise, training or isovolumic hemodilution) on blood viscosity in high-level aerobic endurance athletes. We hypothesized than increasing hematocrit does not always cause a rise in blood viscosity.Sixteen endurance athletes underwent maximal exercise before and after 4 weeks of training with (LHTL; n = 10) or without (placebo; n = 6) Live High-Train Low modalities. Total hemoglobin mass was measured before and after training by a carbon monoxide rebreathing technique. After training, subjects performed two maximal exercise bouts separated by isovolumic hemodilution (phlebotomy and/or plasma volume expander) to readjust red blood cell volume and plasma volume to baseline values. Blood samples were obtained before and after exercise to assess hematocrit and blood and plasma viscosity.Training session (LHTL and placebo) increased hematocrit (Hct) in all subjects but without any significant change in blood viscosity. The decrease in plasma viscosity in all groups may explain this result. Isovolumic hemodilution caused a drop of Hct without any significant change in blood viscosity at rest. Maximal exercise increased Hct, blood and plasma viscosities in both groups, regardless of isovolumic hemodilution. However, peak hemorheological values after exercise were lower after isovolumic hemodilution.In conclusion, while acute increase in Hct during exercise caused an increase of blood viscosity, the chronic increase of Hct induced by training session did not result in a rise in blood viscosity. The lowering of plasma viscosity during training may compensate for the increase of Hct, hence limiting its impact on blood viscosity.