Abnormalities in whole blood viscosity (WBV) have been implicated in hypertension. This study analyzes relations between WBV and blood pressure in the Strong Heart Study population of American Indians. We examined 676 participants (489 women, age 62+/-7 years) without prevalent cardiovascular disease or use of antihypertensive medications, digoxin, or aspirin. WBV was calculated from hematocrit and plasma protein concentration, at a shear rate of 208 seconds(-1), by a validated equation. Forty eight percent of participants were obese, 43% had diabetes, 19% had hypertension, and 30% were current smokers. WBV was higher in men, smokers, and participants with central obesity, but it was not associated with hypertension or diabetes, even accounting for confounders. After adjusting for gender, age, center, smoking, obesity, diabetes, and plasma creatinine, WBV was negatively related to pulse pressure (beta=-0.13; P<0.001) and systolic pressure (beta=-0.09; P<0.02), mainly because of negative relations with hematocrit (beta=-0.11 and -0.10). Among hypertensive individuals, pulse pressure was positively related to age, diabetes, and female gender but not to WBV (multiple R=0.63; P<0.0001); in contrast, in normotensive individuals, pulse pressure was related negatively to WBV or hematocrit, independent of body mass index, without relation to diabetes (R=0.42; P<0.0001). Thus, under normal physiological conditions, in vivo WBV is negatively related pulse pressure. In contrast, the presence of arterial hypertension makes this relation less evident.