Arterial blood > arterial capillary blood > venous capillary blood > venous blood. Plasma is the most convenient specimen for measuring blood glucose, and the results are most reliable. In general, the glucose concentration of whole blood is 10%-15% lower than that of plasma (due to the volume occupied by blood cells), and there is no difference between the measured values of capillary blood and venous blood when fasting, but the difference between the plasma glucose levels of the two can be 2.27±0.66 mmol/ L for the blood sample one hour after meal. Since the peripheral blood is whole blood, it is lower than that of plasma and serum, but many blood glucose meters nowadays correct the measurement to the level of plasma. The results are corrected to the plasma level. As for serum, it is lower than plasma because it takes time to form serum and RBCs consume blood glucose. The water content of whole blood is different from that of plasma, 81% and 93% respectively, and the water solubility of glucose is greater, so the measurement with whole blood will be lower than that of plasma by about 12% to 15%. The main difference between plasma and serum is that serum does not contain fibrinogen, and there is no significant difference in the concentration of blood glucose between the two. If plasma is used, there are strict requirements for the selection of anticoagulant, because anticoagulant often interferes with the measurement results, and some of them also have great influence on the enzyme reagents, so it is most suitable to use serum to determine blood glucose. When in the same environmental conditions and physiological conditions, the blood glucose concentration should be: serum ≥ plasma > peripheral blood.