Abstract
Introduction: HbA1c is an abstract concept that is poorly understood by children and young people (CYP). Average blood glucose (aBG) is readily available and is often used as a proxy for HbA1c when discussing glycemic control. Published models enable conversion of aBG to HbA1c, derived from adult cohorts in study conditions. There are a lack of studies validating these models with CYP in a clinical setting. This study aimed to compare estimated HbA1c (eHbA1c) obtained from aBG, with actual HbA1c in CYP with T1D attending a pediatric diabetes clinic.
Methods: HbA1c data were retrospectively extracted for all patients attending NHS Tayside pediatric diabetes clinic over 12 months (Jan18–Jan19). aBG was calculated from historical data for 3 months prior to each HbA1c measurement. aBG was converted to eHbA1c using Nathan et al’s equation. eHbA1c and actual HbA1c were compared using Pearson’s correlation. The difference between eHbA1c and actual HbA1c was analysed with respect to demographic predictors (age, diabetes duration, no. of BG checks/day) using multivariable linear regression.
Results: 165 patients had HbA1c measured during the study period. Of these, 60 patients had historical BG data. Mean age was 12.1 years (SD3.2), mean diabetes duration 5.7 years (SD3.2). Mean no. of BG checks/d was 3.8 (SD2.5,range 0.1-11.1). Mean aBG was 11.9mmol/l (SD3.2), mean actual HbA1c was 73mmol/mol (SD19). eHbA1c was highly correlated with actual HbA1c (r=0.677,p<0.001). Median difference between eHbA1c and actual HbA1c was 7 mmol/mol (IQR14). The optimum number of BG checks was 4-5/d - a lower number of BG checks/d was highly predictive of larger differences between eHbA1c and actual HbA1c (B-2.4,95%CI -3.6 to -1.2,p<0.001).
Conclusion: Conversion of aBG to eHbA1c would appear to be an accurate proxy for actual HbA1c in this pediatric cohort. Caution should be applied when interpreting eHbA1c if average number of BG checks is <4-5/d.
Methods: HbA1c data were retrospectively extracted for all patients attending NHS Tayside pediatric diabetes clinic over 12 months (Jan18–Jan19). aBG was calculated from historical data for 3 months prior to each HbA1c measurement. aBG was converted to eHbA1c using Nathan et al’s equation. eHbA1c and actual HbA1c were compared using Pearson’s correlation. The difference between eHbA1c and actual HbA1c was analysed with respect to demographic predictors (age, diabetes duration, no. of BG checks/day) using multivariable linear regression.
Results: 165 patients had HbA1c measured during the study period. Of these, 60 patients had historical BG data. Mean age was 12.1 years (SD3.2), mean diabetes duration 5.7 years (SD3.2). Mean no. of BG checks/d was 3.8 (SD2.5,range 0.1-11.1). Mean aBG was 11.9mmol/l (SD3.2), mean actual HbA1c was 73mmol/mol (SD19). eHbA1c was highly correlated with actual HbA1c (r=0.677,p<0.001). Median difference between eHbA1c and actual HbA1c was 7 mmol/mol (IQR14). The optimum number of BG checks was 4-5/d - a lower number of BG checks/d was highly predictive of larger differences between eHbA1c and actual HbA1c (B-2.4,95%CI -3.6 to -1.2,p<0.001).
Conclusion: Conversion of aBG to eHbA1c would appear to be an accurate proxy for actual HbA1c in this pediatric cohort. Caution should be applied when interpreting eHbA1c if average number of BG checks is <4-5/d.
| Original language | English |
|---|---|
| Publication status | Published - 30 Oct 2019 |
| Event | International Society for Pediatric and Adolescent Diabetes Conference 2019 - Boston, United States Duration: 30 Oct 2019 → 1 Nov 2019 https://2019.ispad.org/ |
Conference
| Conference | International Society for Pediatric and Adolescent Diabetes Conference 2019 |
|---|---|
| Abbreviated title | ISPAD 2019 |
| Country/Territory | United States |
| City | Boston |
| Period | 30/10/19 → 1/11/19 |
| Internet address |
