Deviating glucose results in an international dual-center study. A root cause investigation

During a dual-center study on obese and normal weight children and adolescents, focusing on glucose metabolism, we observed a marked difference in glucose results (N = 16,840) between the two sites, Salzburg, Austria and Uppsala, Sweden (P < 0.001). After excluding differences in patient characteristics between the two populations as cause of this finding, we investigated other preanalytic influences. Finally, only the tubes used for blood collection at the two sites were left to evaluate. While the Vacuette FC-Mix tube (Greiner Bio-One, Kremsmunster, Austria) was used in Uppsala, in Salzburg blood collections were performed with a lithium heparin tube (LH-Monovette, Sarstedt, Germany). To prove our hypothesis, we collected two blood samples in either of these tubes from 51 children (Salzburg N = 27, Uppsala N = 24) and compared the measured glucose results. Indeed, we found the suspected bias and calculated a correction formula, which significantly diminished the differences of glucose results between the two sites (P = 0.023). Our finding is in line with those of other studies and although this issue should be widely known, we feel that it is widely neglected, especially when comparing glucose concentrations across Europe, using large databases without any information on preanalytic sample handling.