Intra‐operative blood glucose measurements

Variations in haematocrit are known to affect the accuracy of reagent strip tests for glucose. We have investigated 10 patients during cardiopulmonary bypass, where intra-operative decreases in haematocrit occur. Whole blood glucose concentrations were measured on five occasions at 30 min intervals during the procedure using the Glucometer II, One Touch II and Reflolux II meters as well as a reference instrument (YSI Model 23 AM). Haematocrits were recorded simultaneously. Overall, for every 10% fall in haematocrit, Glucometer II measurements rose by 22% (r = 0.74, p < 0.00001), One Touch II measurements fell by 3% (r = 0.44, p < 0.002) and the Reflolux II measurements showed no significant variation. The One Touch II showed closer agreement to the reference (mean bias 0.3 mmol.l-1 (95% between +0.86 and -0.26)) than the Reflolux II (bias 1.58 (+3.40 to -0.24)) or the Glucometer II (bias 3.25 (+6.18 to 0.32)). Thus, depending on the meter used, spuriously large intraoperative changes in blood glucose may seem to arise where patient haematocrit varies.