Long-term performance of the fructosamine assay.

We have investigated the long-term performance of the fructosamine assay based on secondary glycated protein standards and attempted to define the interpretation of varying degrees of increase in fructosamine concentration in comparison to haemoglobin A1 (HbA1) values both in insulin dependent (IDDM) and non-insulin dependent (NIDDM) diabetic patients. Between-batch imprecision of fructosamine over 5 months was (CV) 2.5% at 2.09 mmol/L, 2.8% at 3.52 mmol/L and 3.6% at 4.14 mmol/L. Variation of fructosamine concentration in vivo in stable diabetic patients monitored over 8-18 weeks was 2.3% to 7.1%. Fructosamine correlated with HbA1 both in IDDM (n = 110, r = 0.701, P less than 0.001) and NIDDM (n = 71, r = 0.764, P less than 0.001). Specificity and sensitivity of fructosamine for the prediction of degree of control assessed on the basis of HbA1 level (cut-off point for good vs. poor control, HbA1 = 10%) was determined. In NIDDM, specificity above 90% was achieved at a fructosamine concentration of 3.4 mmol/L with a corresponding sensitivity of 64.1%. 22.5% of patients were classified differently on the basis of fructosamine as compared to HbA1. In IDDM, specificity over 90% was achieved at 3.8% mmol/L fructosamine with a sensitivity of 35%. Discordancy rate between HbA1 and fructosamine based assessment of control was 31.8%. The assessment of diabetic control based on fructosamine may be different from that based on HbA1, particularly in IDDM. Fructosamine and HbA1 should be used as complementary rather than alternative tests.