Translating the HbA1c assay into estimated average glucose values in children and adolescents with type 1 diabetes mellitus.

OBJECTIVE: The A1c assay, expressed as the percent of hemoglobin that is glycated, measures chronic glycemia and is widely used to judge the adequacy of diabetes treatment and adjust therapy. Day-to-day management is guided by self-monitoring of capillary glucose concentrations (milligrams per decilitre or millimoles per litter) as well as by using continuous glucose monitoring systems (CGMS). We found a mathematical relationship between A1c and average glucose (AG) levels measured by CGMS over 5 days and determined the correlation between the variable CGMS parameters and HbA1c in 50 children with type 1 diabetes mellitus (DM-1) on MDI therapy. RESEARCH DESIGN AND METHODS: A total of 50 diabetic children randomly selected from a cohort of children with DM-1 were included in the analyses. A1c levels obtained at the end of 3 months and measured in a central laboratory were compared with the AG levels during the previous 5 days recorded by CGMS. AG was calculated by combining weighted results from 5 days of continuous glucose monitoring performed before measuring HbA1c, with 3-5 point daily self-monitoring of capillary (fingerstick) glucose.
RESULTS: Linear regression analysis between the A1c and AG values provided the tightest correlations HbA1c=0.0494 MG- 2E-14, R2=0.90, P<0.0001), allowing calculation of an estimated average glucose (eAG) for A1c values.
CONCLUSIONS: Our study showed a linear relationship between HbA1C and AG values measured by CGMS for 5 days before HbA1c measurement. The AG can be easily calculated using a formula derived from linear regression analysis of HbA1c data obtained in our diabetic children.

Introduction

Clinical trials have demonstrated the association between HbA1c and both microvascular and macrovascular complications in type 1 diabetes mellitus (DM-1) (1). HbA1c estimates glucose level over the previous 2-3 months, while the continuous glucose monitoring (CGM) devices measure continuous glycemic profile over a few days and provide many information including patterns, trends and time of glucose changing. In meta-analysis studies real-time CGM appears more effective than self-monitoring of blood glucose (SMBG) in type 1 diabetes (2, 3).

The relationship between the monitoring of blood glucose (MBG) level and HbA1c has been examined in several studies, most of the studies either emphasis on infrequent capillary glucose measurements (4-6). We recorded and analyzed the level of HbA1c in relation to different glucose parameters over 5 days and measured the 24 h mean blood glucose (MBG) from in 50 children with type 1 diabetes mellitus (DM-1). Correlation studies were performed between glucose parameters measured by CGMS and HbA1c level.

Patients and Methods

Fifty randomly selected children with type 1 DM (aged between 3 and 15 years) were included in this study. They had the onset of DM-1 for more than 6 months and were able to perform finger stick glucose testing four times daily.

All children had normal thyroid function and had no other systemic illness or syndrome.

The Medtronic (Northridge, CA) MiniMed CGMS® Gold sensor was used as the CGMS in all children for 5 days prior to measuring HbA1c. Children and their parents were instructed to enter their daily blood glucose finger sticks (morning, lunch, dinner, and before bedtime) into the device for calibration, and children were blinded to the sensor reading. All participants completed CGMS for 5 consecutive days before testing their HbA1c levels. The 24 h mean blood glucose (MBG) and glucose standard deviation values (GSD), BG concentrations before and 2 h after breakfast, lunch and dinner, and the number of high (>250 mg/dL) and low (<60 mg/dL) excursions were recorded.

The study was approved by the IRB committee of Hamad Medical Centre before performing the study.

Spearman correlations and linear regression analyses were applied to quantify the relationship between HbA1C and glucose markers.

Results

The glycemic parameters measured by CGMS 5 days before measuring HbA1c in our 50 children with DM-1 with variable glycemic control are reported in table 1. In particular, 24 h MBG was positively correlated with HbA1c (r=0.90, P<0.001) in all children with DM-1. In addition, the HbA1c was correlated significantly with BG standard deviation score (SDS), BG before and after breakfast and BG after lunch (Table 2 and Figures 1-3).

Table 1.

Glucose parameters during the 5 days and HbA1c levels

Table 2.

Correlation between HbA1c and CGMS glycemic data

Figure 1.

Regression of mean glucose (MG) (mg/dL) and HbA1c level

Figure 3.

Regression of GSD (glucose standard deviation values, mg/dL) and HbA1c level

Regression of mean glucose (G) (mg/dL) before breakfast and HbA1c level

Discussion

Along with its role in diagnosing diabetes, the A1c test is performed between 2 and 4 times per year to estimate average blood sugar levels over the previous 3 months. This test is used to monitor the effectiveness of diabetes treatment and to determine if overall blood sugar goals are being met. The American Diabetes Association recommends a target A1c below 7.5% in children with diabetes, which is an average blood glucose concentration (eAG) below 170 mg/dl (7-11). In the DCCT study, retrospective analysis of data derived from SMBG measurements identified a linear correlation between HbA1c and eA. However, the correlation was based on only on fingerstick glucose measurements (7, 8). The ADAG study defined a mathematical equation between HbA1c and the eAG level (eAG mg/dL=28.7×HbA1c–46.7), which has been widely used in the clinical practice and the equation was recommended by the ADA’s calculation of the eAG. In ADAG study, participants underwent CGM for 48 hours at baseline and monthly for the duration of the study, as well as the SMBG measurement 7 times per day for at least 3 days per week. Over the course of the 12-week study, approximately 2.700 glucose measurements were performed on each participant (8).

In our study, the linear equation showed that HbA1c=0.0494 MG- 2E-14. Accordingly, the predicted HbA1c based on this equation is reported in table 3.

Table 3.

These data are slightly different than that reported by the ADAG study on adults with DM. However, glucose levels comparable to those recorded by ADA were associated with higher HbA1c in our children. The mean increase of 24 h MBG per 1% increase in HbA1c found in our present study (1.1 mmol/L, 20 mg/dL) was lower than that found in either the DCCT study (1.98 mmol/L, 36 mg/dL) (12) or the ADAG study (1.59 mmol/L, 29 mg/dL) (8).

The difference may be attributed to several distinctive characteristics of our population. We included only Arab children with DM-1 with Eastern traditional diet and lifestyle which differ in many aspects compared to the Western traditions. It is well known that racial disparities exist among HbA1c values (13). In addition, there is evidence of wide fluctuations in HbA1c between individuals that are unrelated to glycemic status, suggesting the existence of high and low glycators. High glycators have consistently higher HbA1c than expected for their MBG, whereas low glycators have lower HbA1c than their MBG would suggest (14, 15). In support for this theory, an epidemiologic study found that, when matched for fasting plasma glucose(FPG), African Americans had higher HbA1c than Caucasians (16). This variation in the glycation rate may be attributed to the variations in erythrocyte survival and some yet unknown genetic elements (17-20). Moreover, we used only 5 days glucose data, not all 2-3 months samples, before HbA1c measurement. In support of our data, another study using CGMS over some but not all the 3 months, prior to HbA1c measurement, showed a strong correlation in children with type 1 diabetes, who typically had higher glucose variability (16, 20, 21).

Furthermore, HbA1c levels have been shown to be positively associated with age in nondiabetic populations even after exclusion of subjects with impaired fasting glucose (IFG) and/or impaired glucose tolerance test (IGT). Therefore, HbA1C level in children may be lower than in adults with the same MBG (22-25).

Our children with DM-1 had mean GSD=57.6 mg/dl which may increase their risk of developing diabetic complications and may influence on their HbA1c level. CGMS allows better identification of marked fluctuations in blood glucose, and therefore can improve glycemic control.

Because of the relatively small sample size of our study, further studies validating the findings from the present study in children would be required before any implementation of our results could be considered.

Conclusions

Our study showed a linear relationship between A1C and AG values measured by CGMS for 5 days before HbA1c measurement. The AG can be easily calculated using a formula derived from linear regression analysis of HbA1c data obtained in our diabetic children. Fluctuation of blood glucose can evidently affect HbA1c concentration. The proper use of CGMS enables monitoring glucose variability and can help controlling glucose fluctuations. Further studies are needed to determine whether age-specific diagnostic and treatment criteria would be appropriate