Predictive value of HemoCue capillary whole blood glucose measurements in pregnancy

OBJECTIVE: We determined the ability of capillary whole blood glucose concentrations to predict venous plasma and whole blood glucose levels. METHODS: During a standard oral glucose tolerance test in 29 pregnant women, paired capillary and venous blood samples were collected for analysis of glucose concentrations by the HemoCue photometer and by central laboratory methods. RESULTS: Glucose concentrations determined serially in a single blood sample by the HemoCue method were highly reproducible, with a coefficient of variation of 2.3%. However, glucose levels in blood from two different fingersticks from the same patient varied on average by 3 mg/dL, with a maximum difference of 14 mg/dL. Although capillary whole blood glucose results obtained by the HemoCue method correlated well with venous plasma or whole blood glucose measurements (r = 0.98 and r = 0.97, respectively) over the range investigated (60-250 mg/dL), individual capillary whole blood glucose measurements were only a fair predictor of venous values, with 95% of measured venous levels within +/- 26 mg/dL and +/- 20 mg/dL for concentrations predicted for plasma and whole blood, respectively. CONCLUSION: Sampling factors rather than measurement accuracy limit the ability of capillary whole blood glucose measurements to predict venous concentrations.