Continuous glucose monitoring by intravenous microdialysis: influence of membrane length and dialysis flow rate

Background The benefit of tight glucose control in the intensive care unit is controversial. Part of the debate is around the frequency of glucose measurements, and therefore, a continuous glucose monitoring system is needed. Previously, we have shown that intravenous microdialysis has the potential for this purpose but that the accuracy must be improved. The aim of this study was to investigate the effects of the microdialysis membrane length and the perfusion rate on improving the accuracy. Methods Two volunteer studies were performed, one comparing intravenous microdialysis catheters with different lengths (10 and 20?mm) and one comparing different perfusion rates (0.5, 1 and 2?mu l/min) with plasma glucose reference levels. Median values of seven samples taken over 70-min periods were compared using BlandAltman plots. Results When microdialysis membranes of 10 and 20?mm perfused at a rate of 1?mu l/min were used, the differences with measured plasma glucose levels were 30%?+/-?21% and 14%?+/-?13%. In comparison, plasma glucose measured in two different veins gave a difference of 3%?+/-?3%. In the second study, the differences between measured plasma glucose and that estimated with a microdialysis membrane of 30?mm perfused at 0.5, 1 and 2?mu l/min were 8%?+/-?7%, 25%?+/-?19% and 39%?+/-?28%. BlandAltman analyses gave the best line of equality (-0.11?mM) and the lowest limits of agreement (1.13 and -1.35?mM) when using the 30-mm membrane perfused with 0.5?mu l/min. Conclusion The agreement of the intravenous microdialysis with plasma glucose levels improved significantly when increasing the microdialysis membrane length, and thereby the membrane area, and decreasing the perfusion rate.