Background: The measurement of many parameters of human blood is usually performed in plasma or serum. Since lipoproteins or apolipoproteins, for example, are found almost exclusively in the plasma fraction after low-speed centrifugation, these parameters can be expected to be distributed in a different plasma volume depending on the hematocrit value. Therefore, the measured plasma levels might be relatively too low or too high in comparison to the whole blood concentrations in the case of abnormal hematocrit levels. The aim of our experiments was to evaluate the extent of differences between whole blood and plasma concentrations, taking as an example lipoprotein(a) [Lp(a)] in hemodialysis patients with documented decreased hematocrit values.
Methods: Lp(a) was measured in plasma as well as whole blood of 15 hemodialysis patients with low hematocrit values (0.29 +/- 0.02) in comparison to 11 control subjects (0.45 +/- 0.04).
Results: Plasma concentrations were 27% higher in patients than in controls (19.7 vs. 15.5 mg/dl). The relative difference was twice as high (59%) when measured in whole blood (13.5 vs. 8.5 mg/dl). Similar relative differences were observed when whole blood concentrations of 125 hemodialysis patients and 256 controls were calculated with the formula [Lp(a)plasma * (1-hematocrit)].
Conclusions: Our findings clearly demonstrate that hematocrit is a strong confounding variable of lipoprotein measurement in epidemiological studies when concentrations are measured in plasma, especially in cases of abnormal hematocrit values. Furthermore, studies investigating the longitudinal changes of lipoproteins should consider potential hematocrit changes.