Laboratory assessment of a commercial kit for measuring fructosamine in serum

We have evaluated the laboratory performance and clinical usefulness of the Roche fructosamine kit. As used with an Abbott ABA 100 bichromatic analyzer, the kit response varied linearly with fructosamine concentration to 5.0 mmol/L (deoxymorpholinofructose equivalents). Interbatch precision was 4.1% and 3.6% for respective fructosamine concentrations of 3.2 and 5.0 mmol/L; intrabatch precision was 3.2% and 3.0% (fructosamine = 3.0 and 4.0 mmol/L). In 55 nondiabetic subjects all fructosamine values were less than 3.0 mmol/L, 95% were less than 2.7 mmol/L. For both fructosamine and glycated hemoglobin (Hb A1c) the 95th percentile of the reference range corresponded to approximately the 10th percentile of values observed in 108 diabetic subjects. In the latter subjects fructosamine concentrations correlated somewhat (r = 0.6504) with the Hb A1c value and for eight diabetic subjects indicated a similar degree of diabetic control over a 10-week period. From assessments of sensitivity and specificity for predicting abnormal glucose tolerance in 145 subjects, we conclude that this assay of serum fructosamine reflects diabetic control about as well as Hb A1c estimation, but neither can replace the glucose tolerance test for the diagnosis of diabetes.