[PP.18.27] DISCORDANCE BETWEEN ORAL GLUCOSE TOLERANCE TEST AND GLYCATED HAEMOGLOBIN FOR DYSGLYCAEMIA DIAGNOSIS IN MIXED-ANCESTRY SOUTH AFRICAN: THE INFLUENCE OF GLYCATION GAP AND ANAEMIA. (September 2017)

• Record Type: Journal Article
• Authors: Kengne, A.; Erasmus, R.T.; Matsha, T.E.
• Is Part Of: Journal of hypertension. Volume 35(2017)Supplement 2
• Publication Date: 2017-09
• Subjects: Hypertension -- Periodicals; Hypertension -- Periodicals; 616.132005
• DOI: 10.1097/01.hjh.0000523690.19227.85
• Languages: English
• ISSNs: 1473-5598

Abstract

Abstract : Objective: Variations in absolute hemoglobin levels and glycabilities among individuals are major non-glycemic drivers of the discordance between HbA1c and glucose-based tests in classifying glucose tolerance statuses. We investigated whether low hemoglobin levels and glycation gaps (the difference between measured and predicted HbA1c) account for the disagreement between HbA1c and glucose-based tests in the diagnosis of abnormal glucose tolerance in African populations. Figure. No caption available. Design and method: A total of 791 (men 21.4%) mixed-ancestry South Africans with no prior diabetes were recruited from the Bellville-South Township in Cape Town. Internationally advocated cut-offs were used to classify glucose tolerance status based on oral glucose tolerance test (OGTT) results and HbA1c levels. Correlation/regressions methods were used to predict HbA1c from simultaneously measured fructosamine levels. Results: OGTT-based glucose tolerance status distribution was normotolerance (72.1%), prediabetes (19.7%) and diabetes (8.2%). HbA1c equivalents were 39.3%, 48.4% and 12.2%. Predicted HbA1c and absolute glycation gap varied significantly (and in the same direction) across OGTT and HbA1c defined glucose tolerance statuses (all p < 0.0002). Total haemoglobin decreased with prevalent anaemia increased (both p < 0.0001) with worsening of HbA1c-defined glucose tolerance. For instance the prevalence of anaemia was 14.6%, 20.6% and 26.8% in normotolerant, prediabetics and diabetics respectively. Mean corpuscular volume, mean concentration of haemoglobin, ion levels decreased, while red cell distribution width increased with worsening of HbA1c-based glucose tolerance (all p < 0.0001). Total haemoglobin and all red cell related parameters were similar across OGTT-defined glucose tolerance status. The agreement between OGTT and measured HbA1c in classifying glucose tolerance status was kappa 0.23 (95%CI: 0.17–0.28) overall, 0.10 (0.01–0.20) in anaemic and 0.26 (0.20–0.32) in non-anaemic participants. Equivalent figures after correction for glycation gap using measured fructosamien levels were 0.20 (0.15–0.26), 0.28 (0.15–0.41) and 0.18 (0.12–0.24). Conclusions: The low agreement between OGTT and HbA1c in classifying glucose tolerance status in this population was partially explained by glycation gap particularly in participant with anaemia. These findings challenge the uncritical application of HbA1c measurements to rank glucose tolerance status against the background of high prevalence of anaemia in African populations.