HbA1c as a diagnostic tool for diabetes and pre-diabetes: the Bangladesh experience



This article is corrected by:

  1. Errata: Corrigenda Volume 33, Issue 2, 271, Article first published online: 18 January 2016

Correspondence to: Akhtar Hussain. E-mail: akhtar.hussain@medisin.uio.no



To evaluate HbA1c as a tool for the diagnosis of diabetes and pre-diabetes (impaired glucose tolerance and/or impaired fasting glucose) and to identify the optimal cut-off values suitable for a Bangladeshi population.


In this cross-sectional survey in a rural community, 2293 randomly selected individuals aged ≥ 20 years without prior history of diabetes were included. HbA1c and other clinical covariates necessary for the diagnosis of diabetes were recorded. Diabetes and pre-diabetes were defined according to the World Health Organization 1999 criteria. The receiver operating characteristic curve was used to determine the performance of HbA1c.


The prevalences of diabetes and pre-diabetes were 7.9 and 8.6%, respectively. Based on receiver operating characteristic curve analysis, an HbA1c cut-off value of ≥ 42 mmol/mol (≥ 6.0%) gave an optimal sensitivity of 86.2% and specificity of 93.3%, with an area under the curve of 0.949 to predict diabetes using the oral glucose tolerance test as the gold standard; a cut-off value of ≥ 38 mmol/mol (≥ 5.6%) gave an optimal sensitivity of 68.0% and specificity of 66.4%, with an area under the curve of 0.714 to predict pre-diabetes. In subjects at high risk of diabetes, HbA1c ≥ 42 mmol/mol (≥ 6.0%) showed higher sensitivity than fasting plasma glucose ≥ 7.0 mmol/l, 2-h plasma glucose ≥ 11.1 mmol/l and HbA1c ≥ 48 mmol/mol (≥ 6.5%).


An HbA1c cut-off value of ≥ 42 mmol/mol (≥ 6.0%) was highly sensitive and specific in diagnosing diabetes mellitus. This optimal cut-off level may be suitable as a diagnostic criterion for diabetes in a Bangladeshi population.