Deceptive HbA1c in a patient with pure red cell aplasia

In September, 2008, a 59-year-old Japanese woman with pure red cell aplasia was diagnosed with diabetes mellitus by a haematologist, on the basis of a high concentration of HbA1c (8·2%). At the time of diagnosis, her fasting plasma glucose was 5·8 mmol/L. She immediately began taking acarbose, 300 mg daily. About 3 years later, she began sitagliptin, a dipeptidyl peptidase-4 inhibitor, at a daily dose of 50 mg. Her blood haemoglobin concentration remained low (50–70 g/L). Despite additional treatment with longterm blood transfusions (but not recombinant human erythropoietin), her HbA1c remained high (8·0%). Therefore, the haematologist deemed her glycaemic control to be a more serious problem than her HbA1c concentration. In May, 2012, this patient was referred to our department for help to control her diabetes mellitus. High-performance liquid chromatography (HPLC) and immunoassay both indicated similar HbA1c con centrations, and the HPLC chromatogram contained no abnormal haemoglobin peaks. Despite her high HbA1c level, fasting plasma glucose (6·5 mmol/L), 1-h postprandial glucose (6·9 mmol/L), glycated albumin (14·5%), and 1,5-anhydroglucitol (67·0 μmol/L) showed only mild glucose intolerance. Thus, we judged this patient to have near-normal glucose tolerance and recommended that she stop taking oral hypoglycaemic agents. After stopping these agents, in her fi nal follow-up in November, 2012, fasting plasma glucose, glycated albumin, and 1,5-anhydroglucitol were still stable (6·6 mmol/L, 16·0%, and 112·0 μmol/L, respec tively), but her HbA1c was inappropriately high (7·8%). HbA1c tests are more stable and convenient than are tests for fasting plasma glucose and oral glucose tolerance. For this and other reasons, in 2009 an International aExpert Committee recommended that these tests be used to diagnose diabetes. Moreover, HbA1c tests are widely used for glycaemic management, since HbA1c is a pivotal biomarker whose high concentrations strongly predict complications of diabetes. However, some physicians mistakenly believe that a patient’s HbA1c concentration always reveals that patient’s diabetic status. Certain types of anaemia tend to make an HbA1c test inaccurate. For example, anaemia with abnormal erythrocyte turn over, such as haemolytic anaemia, nephrogenic anaemia treated with e rythropoietin, blood transfusions, and secondary anaemia due to liver cirrhosis all lower HbA1c, whereas iron defi ciency anaemia increases it (panel). In addition, in people with abnormal haemo globin, a condition sometimes associated with anaemia, the HbA1c con centration diff ers dependent on the method used to measure it. Thus, use of HbA1c concentrations to diagnose diabetes in patients with anaemia can lead to overdiagnosis or underdiagnosis in such patients. To the best of our knowledge, our report is the fi rst to show that pure red cell aplasia, a form of aplastic anaemia, can raise HbA1c to an inappropriate concentration. High HbA1c in patients with pure red cell aplasia might be due to slow erythrocyte turnover such as iron defi ciency anaemia that increases the mean age of circulating erythrocytes in the setting of decreased erythropoiesis and raises a proportion of highly glycated haemoglobin, although further studies are needed. In any case, tests measuring serum glycated albumin and 1,5-anhydroglucital, which are usually unaff ected by pure red cell aplasia, should be used instead of HbA1c tests to monitor glycaemic control in patients with pure red cell aplasia.