The level of hematocrit (Hct) is known to affect mean arterial pressure (MAP) by influencing blood viscosity. In the healthy population, an increase in Hct (and corresponding increase in viscosity) tends to raise MAP. However, data from a clinical study of type 2 diabetic patients indicate that this relationship is not universal. Instead, individuals in the lower levels of Hct range display a decrease in MAP for a given rise in Hct. After reaching a minimum, this trend is reversed, so that further increases in Hct lead to increases in MAP. We hypothesize that this anomalous behavior occurs due to changes in the circulatory autoregulation mechanism. To substantiate this hypothesis, we develop a physically based mathematical model that incorporates autoregulation mechanisms. Our model replicates the anomalous U-shaped relationship between MAP and Hct found in diabetic patients in the same range of Hct variability.