Correlation between self-monitored mean blood glucose and average plasma glucose estimated from glycated haemoglobin in patients attending the diabetes clinic at Dr George Mukhari Academic Hospital, Pretoria, South Africa

M K Nkoana, A A Khine


Background. Glycated haemoglobin (HbA1c) has been used for decades as a measure of chronic glycaemia. A simple linear relationship between HbA1c values and mean blood glucose (MBG) has been identified and led to conversion of HbA1c values into estimated average glucose (eAG) levels, following the findings of the A1c-Derived Average Glucose (ADAG) Study Group. The intention was to help patients with diabetes mellitus (DM) understand their glycaemic control better, as eAG is reported in the same units as self-monitored glucose levels. However, factors other than glycaemia have been found to affect the relationship between HbA1c and MBG.

Objectives. To: (i) determine the relationship between self-monitored MBG levels and HbA1c values; and (ii) evaluate the correlation between MBG levels and eAG levels calculated from HbA1c values using the regression equation derived from the ADAG Study Group in black South African patients with DM.

Methods. This was a prospective observational study of 96 diabetic patients. MBG levels were calculated using glucose measurements downloaded from the glucose meters for the previous 90 days (3 months). High-performance liquid chromatography was used for measurement of HbA1c values, collected at the end of 3 months. eAG was calculated using the regression equation from the ADAG Study Group, as follows: eAG (mmol/L) = 1.5944 × HbA1c (NGSP, %) – 2.594.

Results. A positive correlation was found between MBG and HbA1c in all participants (R2=0.69, p<0.0001). There was a wide range of MBG levels for any given HbA1c value. Clinically significant differences between MBG and eAG were found, with a ≥10% difference in 65.6% of the participants. eAG overestimated MBG in ~71.8% of the study population, with an overestimation of ≥1.6 mmol/L (28.7 mg/dL, equivalent to a 1% change in HbA1c value) in ~50% of the total study population.

Conclusions. Our findings showed an imperfect relationship between MBG levels and HbA1c values. eAG significantly overestimated MBG, and this disagreement may cause confusion among both patients and clinicians. The risk of hypoglycaemic episodes may also increase if HbA1c and eAG alone are used to intensify therapy. We recommend that the use of eAG should be validated prior to implementation in clinical practice. It would be ideal to evaluate the relationship between average glucose and HbA1c in each individual patient in order to provide more personalised diabetes care.



Authors' affiliations

M K Nkoana, Department of Chemical Pathology, School of Medicine, Sefako Makgatho Health Sciences University, Pretoria, South Africa; National Health Laboratory Service, Dr George Mukhari Academic Hospital, Pretoria, South Africa

A A Khine, Department of Chemical Pathology, Faculty of Medicine and Health Sciences, Stellenbosch University and Tygerberg Hospital, Cape Town, South Africa; National Health Laboratory Service, Cape Town, South Africa

Full Text

PDF (398KB)


Diabetes mellitus; Mean blood glucose; Estimated average glucose; Glycated haemoglobin; HbA1c

Cite this article

South African Medical Journal 2020;110(5):416-421. DOI:10.7196/SAMJ.2020.v110i5.13959

Article History

Date submitted: 2020-04-29
Date published: 2020-04-29

Article Views

Abstract views: 1485
Full text views: 774

Comments on this article

*Read our policy for posting comments here