HbA1c of 6.5% to diagnose diabetes mellitus -- does it work for us? -- the Bellville South Africa study.

Abstract

BackgroundHbA1c has been the gold standard for glycaemic control follow-up for decades. In 2009, a level of 6.5% (48 mmol/mol) was proposed as diagnostic for diabetes. We test this cut-off in our community.MethodsParticipants (946) from a community-based study were screened for diabetes using either a fasting blood glucose or oral glucose tolerance test (OFTT). The HbA1c cut-off of 6.5% was tested for each group. A receiver operator characteristic (ROC) curve for both groups was generated to establish an optimal cut-off.ResultsOur study included 224 (23.7%) males and 722 (76.3%) females. Using fasting blood glucose alone, 117 (14%) were diagnosed with diabetes--50% had an HbA1c value of ≥6.5% (48 mmol/mol). Using an OGTT, 147 (18%) were diagnosed with diabetes--46% had an HbA1c value of ≥6.5% (48 mmol/mol). ROC curves found a level of 6.1% (43 mmol/mol) to be optimal in both groups (AUC 0.85 and 0.82 respectively). The sensitivities were 80% and 75% and the specificities 77% and 78% respectively.ConclusionsA cut off of 6.5% (48 mmol/mol) is a good diagnostic tool with its high specificity; however the low sensitivity limits its use. We found a level of 6.1% (43 mmol/mol) to be optimal. This emphasizes the need for evidenced based values to be established in various population groups.