Comparison of the HemoCue beta-glucose photometer and reflotron for open heart surgery.

The HemoCue beta-glucose photometer (Angelholm, Sweden) was evaluated for use in monitoring blood glucose in both diabetic and nondiabetic patients undergoing open heart surgery. Because occasional discrepancies were noted in patients with low total proteins when the Reflotron (Boehringer Mannheim, Indianapolis, IN) was used for this purpose, the effects of protein and hematocrit on glucose results from both instruments were investigated and compared with plasma values from a Paramax 720 ZX (Baxter Healthcare, Irvine, CA). Linear-regression analysis of the HemoCue results (y) versus Paramax (x) yielded y = 0.956x + 0.35, r2 = 0.980, with Sy/x = 0.57 mmol/L (10.3 mg/dL). Results from the Reflotron (y) versus Paramax (x) yielded y = 1.075x - 0.10, r2 = 0.964, with Sy/x = 0.99 mmol/L (17.8 mg/dL). Bias plots of (HemoCue-Paramax) or (Reflotron-Paramax) versus glucose, hematocrit, or protein showed no effect of glucose on the results from either instrument and no effect of protein or hematocrit on the HemoCue findings. The Reflotron, however, showed a positive bias of up to 3.5 mmol/L (63 mg/dL) at protein concentrations between 30-40 g/L (3.0-4.0 g/dL) and a possible positive bias at low hematocrit levels.