LAUSR

AIMS Accurate diabetes diagnosis is important in Africa, where rates are increasing, and the disease largely undiagnosed. The cumbersome oral glucose tolerance test (OGTT) remains the reference standard, while alternative diagnostic methods are not yet established in Africans. We assessed the ability of fasting plasma glucose (FPG), HbA1c and fructosamine, to diagnose OGTT-based abnormal glucose tolerance in mixed-ancestry South Africans. METHODS Mixed-ancestry adults, residing in Cape Town were examined between February and November 2015. OGTT values were used to classify glucose tolerance status as: screen-detected diabetes, prediabetes, dysglycaemia (combination of diabetes and prediabetes) and normal glucose tolerance. RESULTS Of the 793 participants included, 65 (8.2%) had screen-detected diabetes, 157 (19.8%) prediabetes and 571 (72.0%) normal glucose tolerance. Correlations of FPG and 2-h glucose with HbA1c (r=0.51 and 0.52) were higher than those with fructosamine (0.34 and 0.30), both p<0.0001. The highest c-statistic for the prediction of abnormal glucose tolerance was recorded with 2-h glucose [c-statistic=0.997 (screen-detected diabetes), 0.979 (prediabetes) and 0.984 (dysglycaemia)] and the lowest with fructosamine (0.865, 0.596 and 0.677). At recommended or data-specific optimal cut-offs, no combination of FPG, HbA1c and fructosamine did better than 2-h glucose, while FPG was better than HbA1c and fructosamine on a range of performance measures. CONCLUSIONS Abnormal glucose tolerance in this population is overwhelmingly expressed through 2-h glucose's abnormalities; and no combination of FPG, HbA1c and fructosamine was effective at accurately discriminating OGTT-defined abnormal glucose tolerance. Tested non-glucose based strategies are unreliable alternatives to OGTT for dysglycaemia diagnosis in this population.