Hba1c Is Not Always Reliable in Estimating Glycaemic Control

Case report

HbA1c is not always reliable in estimating glycaemic control

Fahd Jan Mian1 Abstract MRCP UK, ST5 Endocrinology and Diabetes Glycated haemoglobin (HbA1c) is considered as a cornerstone for assessing long-term glycaemic control. Several studies have established a strong correlation between HbA1c and Koshy Jacob1 diabetic comorbidities. However, in certain conditions it can lead to a false interpretation of FRCP UK, Consultant, Endocrinology and Diabetes diabetes control which may result in serious errors in patient management. It is essential 1Pilgrim Hospital, United Lincolnshire Hospitals NHS clinicians are aware of the conditions that can affect glycated haemoglobin and use Trust, Boston, UK alternative methods of glycaemic control if there is a discrepancy between blood glucose readings and HbA1c. We report a case of a falsely low HbA1c which led to a wrong interpretation of glycaemic Correspondence to: control resulting in a major comorbidity. Copyright © 2016 John Wiley & Sons. Dr Fahd Jan Mian, MRCP UK, ST5 Endocrinology and Practical Diabetes 2016; 33(9): 317–319 Diabetes, Pilgrim Hospital, United Lincolnshire Hospitals NHS Trust, Sibsey Road, Boston PE21 9QS, Key words UK; email: [email protected] glycated haemoglobin; HbA1c Received: 22 June 2016 Accepted in revised form: 20 July 2016 Introduction revealed blood glucose in the range First described by Rahbar et al. in of 15–20mmol/mol. The patient 1969, HbA1c is an invaluable tool did not report any hypoglycaemic in monitoring long-term diabetes episodes; however, in view of the control. Several studies have shown HbA1c results, the gliclazide dose that the level of HbA1c correlates was initially reduced and later com- well with the glycaemic control over pletely stopped. As the blood glu- a period of two to three months. cose readings did not correspond The Diabetes Control and to the HbA1c results, continuous Complications Trial (DCCT) and glucose monitoring (CGM) was the UK Prospective Diabetes Study organised on two occasions which (UKPDS) have shown a strong confirmed the discrepancy between correlation between HbA1c and blood glucose and HbA1c. HbA1c diabetic comorbidities.1,2 It is a well was tested using two different assays established practice to measure at different hospitals with no differ- HbA1c in all patients with diabetes ent results. and is considered a cornerstone Review of medications did not for diabetes management. It results reveal any drugs that could have from an irreversible non-enzymatic accounted for the inappropriately reaction between glucose and the low HbA1c. Vitamin B12 and folate main haemoglobin in adults, HbA, levels were within the normal refer- and occurs during the life span of ence range (B12 264ng/L, folate the erythrocyte.3 7.3mg/L) and though the initial results showed macrocytosis subse- Case history quently it resolved spontaneously. A 52-year-old male with a seven-year Haemoglobin electrophoresis, hap- history of type 2 diabetes, managed toglobin and auto-antibody screen in primary care, was referred to the were all normal. Fructosamine levels diabetes clinic for painful periph- were also within the normal range. eral neuropathy. Apart from hyper- A red cell study was organised which tension there was no other signifi- indicated a reduced red cell survival cant past medical history. At the of approximately 22.5 days. A CT time of presentation the patient was scan of the abdomen was organised on metformin, gliclazide and rami- and showed splenomegaly, but no pril. Review of previous results cause for the splenomegaly could be showed an HbA1c ranging between identified. Considering the results 20–40mmol/mol, GFR 73ml/min, of investigations a diagnosis of Hb 139g/L, MCV 107fl and protein- falsely low HbA1c secondary to uria of 2.65g/24 hours/L. A review reduced red cell survival due to of finger prick blood glucose results hypersplensim was established.

In view of the CGM results while Inappropriately low Inappropriately high Variable effect on the investigations were ongoing, the HbA1c HbA1c HbA1c patient was commenced on insulin to improve glycaemic control. By • Haemolysis • Iron deficiency • Certain the time of presentation the patient • Acute blood loss • Vitamin B12 deficiency haemoglobinopathies had already developed severe pain- • Blood transfusion • Uraemia • Fetal haemoglobin ful peripheral sensory neuropathy • Certain • Hyperbilirubinaemia • Methaemoglobin and, despite preventive measures, haemoglobinopathies • Alcoholism developed diabetic foot complica- • Chronic liver disease tions (Charcot’s foot, osteomyelitis) • Hypertriglyceridaemia leading to below-knee amputation of the left foot. Table 1. Conditions causing inappropriately high or low HbA1c The patient remains under the follow up of the diabetes service, Mechanism Low HbA1c High HbA1c with glucose monitoring used as a method of diabetes control. Haemolysis Dapsone Ribavirin Discussion Antiretroviral This case illustrates the importance Trimethoprim of clinical awareness when interpret- Sulfamethoxazole ing HbA1c results. They may falsely Impaired glycation Vitamin C lead to what may appear to be a Vitamin E good glycaemic control. The following discussion focuses Altered haemoglobin Hydroxyurea on the conditions and factors which Assay interference Aspirin (large doses) can affect HbA1c. Chronic opiate use Although in general HbA1c is a robust marker of glycaemic control, Table 2. Drugs causing inappropriately low or high HbA1c in some conditions the test can be unreliable. In addition, certain as the Hb variant does not interfere fructosamine. In late pregnancy, drugs can cause a falsely low or high with the assay method used or the HbA1c is increased in both diabetic HbA1c. High performance liquid binding of glucose to haemoglobin.4 and non-diabetic individuals due to chromatography, immunoassay, cap- In beta thalassemia major, HbA1c iron deficiency.10 The described illary electrophoresis, and boronate can be unreliable as a method of mechanism is an increase in malond- affinity chromatography are some of glycaemic control. These patients ialdehyde, a marker of oxidative the techniques used to measure require regular transfusions, and stress, in patients with iron deficiency HbA1c. The newer assays to measure data from relatively old literature anaemia which enhances glycation HbA1c can eliminate some, but not suggest that a high concentration of of the haemoglobin.11,12 Iron all, of the errors.3 glucose in the red blood cell (RBC) replacement therapy has shown to storage medium promotes glycation reduce both HbA1c and fructosamine Factors affecting HbA1c of the packed RBCs and raises levels in diabetic and non-diabetic interpretation HbA1c values in transfused patients.5 patients with iron deficiency.13,14 Certain conditions can cause inap- However, these findings were not Other methods of glycaemic assess- propriately high or low HbA1c. Any confirmed in a study by Spenceret ment should be used in these condition which shortens the mean al. which showed HbA1c levels may patients, such as glucose monitoring, erythrocyte age, such as haemolysis actually reduce in diabetic patients at least until the iron deficiency has and blood loss, lowers HbA1c regard- receiving transfusion because of been successfully treated. In diabetic less of the assay method used. The dilution with RBCs containing patients on dialysis, HbA1c underesti- accuracy of the HbA1c methods can typical amounts of HbA1c.6 mates glycaemic control and measur- also be affected adversely by the There are only a few studies that ing glycated albumin is advised. presence of haemoglobin variants. have looked at the effect of beta Conditions affecting HbA1c are In patients with HbSS, HbCC and thalassemia trait and alpha thalas- summarised in Table 1. HbSC, HbA1c must be interpreted semia on HbA1c. Glycated haemo- with caution given the pathological globin measurement is not affected Drugs causing inappropriately processes. In such patients, an alter- in beta thalassemia trait7,8 and in low or high HbA1c native method of testing, such as alpha thalassemia except in patients Certain drugs have been reported to glycated albumin or glycated serum with HbH disease in which HbA1c cause inappropriately low or high protein, should be considered. levels were found to be lower than HbA1c. Different mechanisms have Patients with heterozygous variants controls in a recent study.9 More been postulated by which these are usually asymptomatic and have research is needed in this area. drugs can interfere with HbA1c. normal red cell survival. In these Iron deficiency can result in Drugs such as dapsone, ribavirin patients, HbA1c can be used as long falsely elevated HbA1c as well as and antiretrovirals cause haemolysis

and can lower the HbA1c as a result 5. Choudhary A, et al. Unreliable oral glucose toler- of reduced erythrocyte lifespan.15 Key points ance test and HbA1c in beta thalassaemia major – a case for continuous glucose monitoring? Br J Similarly, hydroxyurea may falsely Haematol 2013;162:132–5. lower HbA1c by causing a shift from ● HbA1c, in general a robust marker of 6. Spencer DH, et al. Red cell transfusion decreases HbA to HbF.16 Aspirin in large doses diabetes control, can be unreliable in hemoglobin A1c in patients with diabetes. Clin Chem 2011;57:344–6. used chronically can lead to falsely certain conditions and may lead to 7. Polage C, et al. Effects of beta thalassemia minor elevated HbA1c due to acetylation of serious errors in patients’ management on results of six glycated haemoglobin methods. haemoglobin. Anti-oxidants such as ● Alternative methods of glycaemic control Clin Chim Acta 2004;350(1–2):123–8. vitamin C and E in high doses can should be used if there is a discrepancy 8. Al-Fadhli SM, et al. Effect of sickle cell trait and B-thalassemia minor on determinations of HbA1c falsely lower HbA1c by interfering between blood glucose and HbA1c by an immunoassay method. Saudi Med J 2001; with the glycation of haemoglobin.17 ● It is essential clinicians are aware of the 22(8):686–9. 9. Xu A, et al. Effects of -thalassemia on HbA1c Chronic use of opiates reportedly conditions that can affect HbA1c and α measurement. J Clin Lab Anal 2016 May 17. doi: can increase HbA1c by interfering should correlate HbA1c results with 10.1002/jcla.21983. [Epub ahead of print.] with the assay, but the exact mecha- blood glucose readings 10. NGSP. Harmonizing haemoglobin A1c testing. www. nism is unknown.18 ngsp.org/factors.asp. 11. Madhikarmi NL, Murthy KRS. Effect of iron Drugs causing inappropriately deficiency anaemia on lipid peroxidation and low or high HbA1c are summarised Declaration of interests antioxidant systems. J Coll Med Sci Nepal 2011; in Table 2. There are no conflicts of interest 7(4):34–43. declared. 12. Selvaraj N, et al. Effect of lipid peroxides and anti- oxidants on glycation of hemoglobin: an in vitro Conclusion study on human erythrocytes. Clin Chim Acta Glycated haemoglobin, though a References 2006;366:190–5. robust marker, can sometimes lead 1. Diabetes Control and Complications Trial Research 13. Coban E, et al. Effect of iron deficiency anemia Group. The effect of intensive treatment of diabetes on the levels of hemoglobin A1c in non-diabetic to the wrong interpretation of gly- on the development and progression of long-term patients. Acta Haematol 2004;112:126–8. caemic control and may result in complications in insulin-dependent diabetes melli- 14. Tarim O, et al. Effects of iron deficiency anemia on serious errors. It is essential clini- tus. N Engl J Med 1993;329:977–86. hemoglobin A1c in type 1 diabetes mellitus. Pediatr cians are aware of the conditions 2. UK Prospective Diabetes Study Group. Intensive Int 1999;41:357–62. blood-glucose control with sulphonylureas or 15. Unnikrishnan R, et al. Drugs affecting HbA1c levels. that can affect HbA1c. insulin compared with conventional treatment Indian J Endocrinol Metab 2012;16(4):528–31. In situations where there is a and risk of complications in patients with 16. Karsegard J, et al. Spurious glycohemoglobin values suspicion of a falsely low or high type 2 diabetes (UKPDS 33). Lancet 1998;352: associated with hydroxyurea treatment. Diabetes 837–53. Care 1997;20:1211–2. 1c HbA , the results should be corre- 3. Chu CH, et al. Very low HbA1c values in a patient 17. Camargo JL, et al. The effect of aspirin and vitamins lated with blood glucose readings with clinical silent hemoglobin variant (Hemoglobin C and E on HbA1c assays. Clin Chim Acta 2006; and, in the case of a discrepancy, J). A case report. J Intern Med Taiwan 2007;18: 372:206–9. alternative methods of glycaemic 45–50. 18. Ceriello A, et al. Increased glycosylated haemo- 4. Little RR, Roberts WL. A review of variant hemoglobin A1 in opiate addicts: Evidence for a hyper assessment should be used and globins interfering with hemoglobin A1c measure- glycaemic effect of morphine. Diabetologia 1982; potential causes investigated. ment. Diabetes Sci Technol 2009;3:446–51. 22:379. POEMs

Low HbA1c, BP and cholesterol in elderly patients with type 2 associated with high mortality Clinical question so forth. Approximately half the cohort were women, one- Is there an association between mortality and glycaemic third had co-existing cardiac disease, and about half had control, blood pressure levels and cholesterol levels in diabetes for at least 10 years. Approximately 10% of the patients with type 2 diabetes who are older than 80 years? patients were older than 90 years. There was a median of two years of follow up, during which about 4500 patients Reference died (17%, 105 per 1000 person-years). The authors esti- Hamada S, Gulliford MC. Mortality in individuals aged mated the mortality, adjusting for a variety of factors, 80 and older with type 2 diabetes mellitus in relation including age, sex and duration of diabetes. Similar to to glycosylated hemoglobin, blood pressure, and total other studies, the relationship between mortality and gly- cholesterol. J Am Geriatr Soc 2016;64(7):1425–31. caemic control as measured by HbA1c and blood pressure levels followed a U-shaped pattern, with the nadir for Synopsis HbA1c between 7% and 7.4% (53–57mmol/mol) and for These authors evaluated a population-based primary care blood pressure between 150/90mmHg and 155/95mmHg. database in the UK to identify nearly 26 000 patients older For each of these, the mortality progressively worsened than 80 years with type 2 diabetes. The database includes with lower and higher HbA1c or blood pressure levels. The data on physical examinations, test results, diagnoses and relationship between total cholesterol and mortality was medications. The researchers mined this database for all more curvilinear, with the highest mortality associated with they could: clinical data, comorbidities, number of clinic the lowest cholesterol levels and an asymptotic decrease as visits, classes of prescribed medication, smoking status, and cholesterol levels increase.