Limitations of A1c Interpretation

Jessica G. Shepard, PharmD; Anita Airee, PharmD, BCPS; Andrew W. Dake, MD; M. Shawn McFarland, PharmD; Amit Vora, MD


South Med J. 2015;108(12):724-729. 

In This Article

Conditions That Affect the A1c

Erythrocyte Production

Certain conditions can shorten the lifespan of red cells, leading to a falsely low A1c. The average age of the red cells is reduced so cells have less exposure to the glucose in the bloodstream. These conditions include hemolytic anemia, treatment with erythropoietin, hemodialysis, and pregnancy. In hemolytic anemia red cells are destroyed faster than they can be produced, which could shorten the lifespan of cells from 120 days to 60 days.[10,11] In chronic kidney disease (CKD), hemodialysis can result in a lower A1c because of a 50% reduction in the lifespan of an erythrocyte.[16,17] When patients are treated with erythropoietin, erythropoiesis is accelerated, leading to a decrease in the average age of the circulating erythrocytes and a falsely low A1c.[18] One small study found a 0.7% decrease in A1c in patients with CKD treated with erythropoietin.[19] In pregnancy, there is a likely decrease in erythrocyte lifespan as well as lower fasting glucose levels. Occasionally, however, there is an increase in A1c because of possible hemoglobin alteration, making it easier to glycate.[20]

Blood loss and transfusion also can have an effect on A1c values.[21] Active bleeding or major blood loss causes increased reticulocyte formation, leading to lower A1c values resulting from a decrease in the average age of circulating red cells.[22] In patients who have recently received a transfusion, A1c values may be skewed depending on the A1c of the donor. Patient populations in which transfusions are commonly required include surgical patients, intensive care patients, and patients with CKD.[16,1916,19] In the CKD population, use of transfusions has declined since the approval of erythropoietin; however, erythropoietin also can affect A1c values, as discussed above.

Other conditions may prolong the red cell lifespan, causing an increase in the average age of the circulating red cell, which can lead to a falsely high A1c. Conditions that may result in a falsely high A1c include iron-, folate-, or vitamin B12-deficiency anemia; aplastic anemia; or splenectomy. With anemia, the body produces fewer red cells, which results in an increase in the average age of the circulating red cells.[23,24] Splenectomy causes slower red blood cell clearance, which also increases the average age of the red cells.[25]

Ethnicity and Hemoglobin Variants

Considering that approximately 15% of the African American population and 7.5% of the Asian population have diabetes mellitus, the reliability of accurate A1c is paramount.[2] People of African, Asian, and Mediterranean descent are at increased risk for hemoglobin variants. The most common hemoglobin variants are HbS, E, C, and D. Prevalence of hemoglobin C may be 2.3% in African Americans, and the prevalence of hemoglobin E may be 30% in Asian Americans. In addition, 12% of African Americans may have hemoglobin S and 1 in every 12 has sickle cell trait in which both hemoglobin S or C may be present. African Americans, those from eastern India, the Mediterranean, and the Middle East may have hemoglobin SC.[26–28] Hemoglobin D Punjab is most commonly found in the Punjab region of India and in the United States.[29] Hemoglobin F also may be present in up to 12% of patients. Healthy People 2020 has as a new goal in this decade to increase the proportion of hemoglobinopathy carriers who know their own carrier status.[30]

The presence of some variants can affect the net charge of the hemoglobin and/or the recognition of the glycated N-terminus by antibodies resulting in erroneous values for A1c. These hemoglobinopathies can alter the normal process of glycation; cause an abnormal peak on chromatography, making estimation unreliable; and make the red blood cell more prone to hemolysis, thereby decreasing the time in which glycation can occur.[31,32] Both HbS and HbC have a single mutation that occurs close to the N-terminus of the beta chain of the hemoglobin molecule. Because this is in close proximity to the position where glycation occurs, hemoglobin variants can interfere with some immunoassays.[32]

Although many of the assays used to measure A1c are no longer affected by hemoglobin variants, it is advisable to suspect a hemoglobinopathy when the A1c results do not correlate with the self-monitoring blood glucose or if the A1c value changed following a change in laboratory A1c methods.[32] A comprehensive list of assays that are affected by hemoglobin variants that is available at the NGSP Web site (

Although hemoglobin variants may be largely responsible for the differences in A1c, some controlled studies have hypothesized other etiologies. Herman et al and others observed differences in African Americans, Hispanics, and Asian Americans compared with whites in a study controlled for body mass index, blood pressure, age, sex, beta-cell function, and insulin resistance.[34,3534,35] Although blood glucose levels were similar, high postprandial readings were observed, which may reflect an ethnic difference of a high carbohydrate diet. As such, the differences plausibly explained by blood glucose did not accurately reflect the 24-hour glucose profile.[36] In a study that included individuals without known diabetes mellitus, the difference in A1c between minorities and whites increased as the A1c increased, however.[35] In addition, diabetic complications also may differ as shown by a study of 2804 white and 1008 black individuals, which found that black patients experienced retinopathy at significantly lower A1c values (5.5%–5.9%) than white patients (6.0%–6.4%).[37] The above is a compelling reason for further research to determine whether the normal A1c range should be adjusted for ethnicity.

Chemically Modified Hemoglobin

The measurement of A1c can be affected by chemically modified hemoglobin such as acetylated or carbamylated hemoglobin. High-dose aspirin approximating 500 mg/day has been shown to interfere with A1c values by causing acetylation of the hemoglobin molecule leading to a change in the ionic charge.[38] This change in charge makes it difficult for the ion-exchange or electrophoresis methods to distinguish acetylated hemoglobin from glycated hemoglobin, subsequently measuring both, resulting in a falsely elevated A1c.[38] Renal failure can result in elevated urea, which dissociates into cyanate. Cyanate in its reactive form can then bind to hemoglobin to form carbamylated hemoglobin. Carbamylated hemoglobin has a charge similar to the charge of glycated hemoglobin and may interfere with ion-exchange or electrophoresis methods.[16] It has been suggested that the interference is only significant when serum urea levels exceed 84 mg/dL (30 mmol/L).[16]

Altered Rate of Glycation

Vitamins C and E also have been found to interfere with A1c values.[39,40] It is believed that they reduce glycation of hemoglobin and can affect A1c values, depending on the type of test used. In a study of 12 nondiabetic individuals taking 1 g/day of vitamin C for 3 months, a significant decrease in A1c was observed when measured using affinity chromatography, but an increase in A1c was seen when electrophoresis was used. The authors explained that vitamin C binding to hemoglobin changes the charge of the hemoglobin similarly to the binding of glucose, causing an increase in A1c when measured by electrophoresis. Vitamin C hemoglobin product has a different structure than glycated hemoglobin, causing a decrease in A1c when measured by affinity chromatography.[39] Another small study of 30 nondiabetic individuals found that vitamin C did not interfere with A1c values when measured by electrophoresis, affinity chromatography, or immunoassay, however. The individuals in this study were placed on doses as high as 1500 mg/day for 12 weeks.[41]


The effect of age in relation to the interpretation of A1c may be more a function of age-related changes in glucose regulation, particularly in elderly adults.[42–44] Pani et al found that A1c increases with age in a group of nondiabetic individuals, but other studies have found no change with increasing age.[45,46] Plausible explanations have included a decrease in physical activity, increased insulin resistance, mitochondrial dysfunction, and changes associated with insulin secretion.[42–44] Nowicka et al, however, found in a study of 1156 obese children and adolescents that cutpoints for A1c were diagnostic for prediabetes mellitus in 47% when measured by A1c and 62% when measured by oral glucose tolerance tests.[47] Although the ADA Standards of Care have acknowledged this, no age-specific values have been recommended.[44]