Unreliable oral glucose tolerance test and haemoglobin A1C in beta thalassaemia major – a case for continuous glucose monitoring?
A. Choudhary,P. Giardina,Z. Antal,M. Vogiatzi
DOI: https://doi.org/10.1111/bjh.12322
2013-07-01
British Journal of Haematology
Abstract:Diabetes mellitus (DM) is a common complication in individuals with Thalassaemia Major (TM), with reported incidence rates of 14-24% (Vogiatzi et al, 2009). Although the pathophysiology is unclear, the concomitant presence of transfusion related iron overload, insulin resistance, and beta cell destruction are thought to contribute in the development of DM in this population (Cooksey et al, 2004).
The oral glucose tolerance test (OGTT) and serum hemoglobin A1C (HbA1C) levels are commonly used screening tools for diabetes, and have been shown to be valid in individuals with obesity amd family history of diabetes. However, the utility of HbA1c measurement is limited in thalassemia since both hemoglobinopathies and transfusions are known to interfere with HbA1C analysis. The results may be falsely increased or decreased depending on the proximity to transfusion, shortened erythrocyte lifespan and the assay used (Bry et al, 2001; Spencer et al, 2011; Sacks, 2003). Given the above limitations, OGTT has been proposed as the recommended screening method for glucose abnormalities in thalassaemia. However, the OGTT procedure requires patients to fast, is cumbersome and time consuming.
Continuous glucose monitoring (CGM) systems is a new technology that measures glucose concentrations in a continuous fashion, with the use of a sensor that is placed subcutaneously. CGM provides information on glucose excursion in relationship to meals and daily activities, and hence, has added a vital dimension in the diagnostic and therapeutic approach to hyperglycemia and management of subjects with diabetes or abnormal OGTT and HbA1c results. In this paper, we set out to evaluate the utility of CGM in subjects with TM, given the limitations of HbA1C and OGTT in this population. The study also documents elevated HbA1C results using high performance liquid chromatography (HPLC) in TM regardless of glycemia.
The Weill Cornell Medical College standard of care for patients with TM involves yearly screening for DM via OGTT. We invited all adults with TM who presented for their annual OGTT to wear a CGM sensor and have HbA1C measured as part of a larger study that examined relationship between CGM and OGTT results. Five subjects agreed to partake in wearing sensors and twelve had HbA1C measurements.
All underwent OGTT per protocol. CGM was performed using the Medtronic I Pro 3 day blinded glucose sensor. The sensor was inserted per manufacturer’s recommendations into the subcutaneous tissue compartment of the abdomen. The subjects were instructed to wear the sensor at all times for 72 hours. The subjects were additionally asked to check finger stick glucose (FSG) levels using a glucometer at least four times a day for calibration and comparison purposes. They were asked to keep a detailed log of meals and snacks, exercise, and any medication use. Three days later, the subjects returned for sensor removal, at which time the data was analyzed using version 1.2a of the Medtronic Solutions software. HbA1C was measured by high performance liquid chromatography (HPLC), normal range: 4 to 6%. All 12 subjects had elevated HbA1c with no clear correlation to the results of their OGTT. Mean HbA1C results were 7.4 % in 3 patients with normal glucose tolerance, 7.33% in 3 patients with pre-diabetes, and 8.16% in 6 patients with DM.
Of the five subjects who underwent CGM, three (#1, 2 and 3) met the criteria for DM based on the two hour glucose value on OGTT above 11.0 mmol/L (Table I). Two of the five subjects (#4 and 5) had normal glucose tolerance (Table I). The FSG which showed normal pre and post meal glucose readings, including in those who had an abnormal OGTT. Complete CGMS data were retrieved and downloaded for all the subjects. Results were normal in all cases. None of the subjects showed the degree of glucose variability seen during OGTT, despite reporting equal and at times greater carbohydrate intake per meal than the glucose load during OGTT. Figure 1 shows the discrepancy between the OGTT and CGM data for the three subjects with DM on OGTT.
Figure 1
Discrepancy between OGTT results (on the left) and their corresponding CGM (on the right). OGTT results reveal glucose concentrations above 11.0 mmol/L (200 mg/dl) at 120min, which is consistent with DM. In contrast, all glucose values by CGM are within ...
Table I
Subjects characteristics with discrepancies between OGTT and CMS, and their HbA1C, OGTT and CGMS results.
Transfusion dependent TM and DM have a unique and incompletely understood relationship. Similar to other transfusion dependent diseases, the risk of iron overload remains substantial despite improved methods of chelation, resulting in a high prevalence of DM in this population (Vogiatzi et al, 2009; Cooksey et al, 2004). Multiple large scale studies have demonstrated that patients with type 2 DM are at an increased risk for cardiovascular complications, with post-prandial hyperglycemia being a significant risk factor (King et al, 1999; Nathan et al, 2005; Hanefeld et al, 1996; Cengiz, 2009; Blake et al, 2004). Although the significance of similar carbohydrate abnormalities in the aging TM population is unclear at this time, it is reasonable to assume that they are predisposed to the same long term complications, morbidity, and mortality, making accurate and timely diagnosis of DM critical. The onset of hyperglycemia in patients with TM can be insidious and silent. Patients may go through extended periods of impaired glucose tolerance and variable degrees of glycemic excursions. The additional inflammatory stimulus caused by iron deposition, coupled with hyperglycemia, leads to the potential for significant micro and macro vascular complications (Cooksey et al, 2004).
Current ADA guidelines recommend screening “at risk populations” with either fasting glucose and insulin levels, random HbA1c levels, or a formal oral glucose tolerance test. These screens impose a burden on patients with TM who are already coming in regularly for transfusions. Glucose and insulin levels require the patients to fast, while the OGTT is invasive and time consuming. Furthermore, the use of HbA1c as a screening tool for DM or assessment of glycemic control is inappropriate in TM. Transfusions may contain hemoglobin molecules exposed to glucose concentrations that are different from those of the transfusion recipient. Data from old literature suggests that high concentration of glucose in the RBC storage medium promotes glycation of the packed RBC and raises HbA1c values in the transfused patients. These findings were not confirmed in recent studies using current blood bank procedures. Moreover, they indicate that transfusions reduce the HbA1c concentration in diabetic patients due to dilution (Spencer et al, 2011). Hemoglobinopathies including thalassemia syndromes, factors that impact red blood cell survival and age, uremia, hyperbilirubinemia, and iron deficiency may all alter HbA1c test results (Herman et al, 2012). Racial and ethnic differences, genetic variation in hemoglobin glycation and assay methodology may also influence the results (Herman et al, 2012). Specific to this report, HPLC is widely used and, in general, is considered superior to immunoassays. However, hemoglobin variants may create an abnormal peak very close to the HbA1C peak, which interferes with HbA1C reading leading to spurious results. In such case, a visual inspection of the HPLC chromatogram may give valuable information regarding hemoglobin variants (Bry et al, 2001), but this is not an established procedure in commercial laboratories. The persistently elevated HbA1C values by HPLC in TM may result in unnecessary and potential life threatening medical interventions.
CGM is a powerful tool to assess diabetes control and the effect of medical management. CGM gives a detailed recording of glucose excursions under daily activities. All three patients who were found to be diabetic by OGTT had normoglycemia by CGM, which is an unexpected finding. We acknowledge the limitations related to small sample size. We can also offer limited insight as to the mechanism. OGTT entails the ingestion of a large bolus of simple sugar. In contrast during CGM, subjects consume a variety of simple and complex carbohydrates, which may evoke different response of gut mediated hormones. Studies evaluating alpha and beta cell reserve and function in Thalaessemia suggest that the effect of iron deposition and inflammation on the pancreatic islets may result in both alpha and beta cell damage (De Sanctis et al, 1985). This type of injury can lead to disturbed insulin and glucagon release from islets, and may contribute to our findings.
Our preliminary report highlights the potential that OGTT results can be erroneous in this population, and that CGM may be more accurate for patients to define their degree of glucose control. Large scale studies may shed light on the role of CGM in TM.