Screening for diabetes: does measurement of serum fructosamine help?

The diagnostic sensitivity and specificity for diabetes of serum fructosamine levels and fasting venous blood glucose concentrations were compared in 613 subjects during a diabetes community screening programme of 1049 adult Muslim Asians in Dar es Salaam, Tanzania. Using WHO (1985) criteria 228 had impaired glucose tolerance (IGT), 41 had previously been diagnosed as having diabetes while 32 had newly recognized diabetes. The mean (+/- SD) serum fructosamine levels were 20.9 +/- 3.2, 21.6 +/- 3.2, 23.9 +/- 4.9, and 30.1 +/- 7.9 (mumol g-1 albumin) in subjects with normal glucose tolerance, IGT, newly diagnosed diabetes, and previously diagnosed diabetes, respectively (p less than 0.001 for differences between groups). The specificity of values above the mean +2SD normal was 99% for abnormal glucose tolerance with a sensitivity of only 22% for diabetes. The predictive values were 44% and 97% for positive and negative results, respectively. Very little difference from normal was found for IGT subjects. Expressing fructosamine values in absolute terms or per gram albumin made little difference to sensitivity and specificity. The sensitivity was only 32% for fasting blood glucose greater than or equal to 6.7 mmol l-1, 73% for values greater than or equal to 5.5 mmol l-1, and 100% for fasting blood glucose greater than or equal to 4.5 mmol l-1. It is concluded that both serum fructosamine and fasting blood glucose are poor screening and diagnostic tests for diabetes and for IGT, and that glucose loading is required.