Gross errors made by routine blood sampling from two sites using a tourniquet applied at different positions.

Tourniquet application in routine blood sampling procedures may induce considerable haemoconcentration and thereby augment the concentration of serum protein and protein-bound substances. To evaluate this effect serum protein, calcium and magnesium were measured in 31 healthy persons before and after 3-min venous stasis induced by a standardized tourniquet. Four sites of tourniquet and two sampling sites on the arm were compared. The concentration of measured serum constituents rose after the stasis period (5--13% rise in serum protein), the changes being similar at all tourniquet positions. However, blood sampled from a cubital arm vein showed significantly larger haemoconcentration than samples obtained simultaneously more distally on the arm. Taking three samples in a series, the serum concentration of protein, calcium and magnesium rose from tube to tube (5--10% change). The study demonstrates that neither use of a standardized tourniquet nor release of the tourniquet before blood sampling eliminates the errors inherent in routine venepuncture. Adjustment to a constant serum protein level minimizes these errors.