Subscribe to RSS
DOI: 10.1055/s-0033-1341681
Relationship between glycated hemoglobin and average blood glucose values – validation of the results from the ADAG study
The ADAG study, which results were subsequently recommended by ADA guidelines for a calculation of HbA1c from average glucose values, was performed in only 159 patients with type 2 diabetes. The aim of this study was to validate those results in a real life settings in a large group of patients with type 2 diabetes.
The analysis was prespecified in the protocol of PROGENS HbA1c, an observational multicenter study that assessed effectiveness and safety of human recombinant insulin produced by BIOTON, Poland, in which 4257 patients with type 2 diabetes were included (mean age 63.7 ± 9.4 years, mean BMI 30.3 ± 4.5 kg/m2, time from diagnosis 9 ± 5.5 years). At visits 2 and 3 a parallel assessment of HbA1c (using a multitest HbA1c analyser A1cNow+, Bayer HealthCare) and average blood glucose (AvG) from the preceding 90 days, calculated automatically by the glucometer, was performed. The regression formula (data at the 2nd visit) was HbA1c = 5.28 + 0.01487 x AvG 90 days. The results of the PROGENS HbA1c study differ from those from ADAG study by different slope of the regression line (and therefore different glucose values assigned to HbA1c) and broader distribution of the results by lower HbA1c values (see below).
Comparison of ADAG and PROGENS HbA1c studies: HbA1c – average glucose PROGENS HbA1c (95% CI) vs. average glucose ADAG (95% CI)
6% – 134,4 (93 – 175,7) vs. 126 (100 – 152)
7% – 149,3 (107,9 – 190,6) vs. 154 (123 – 185)
8% – 164,1 (122,8 – 205,5) vs. 183 (147 – 217)
9% – 179,0 (137,6 – 220,4) vs. 212 (170 – 249)
10%-193,9 (152,5 – 235,3) vs. 240 (193 – 282)
11%-208,8 (182,1 – 265,2) vs. 269 (217 – 314)
12%-223,7 (195,4 – 278,9) vs. 298 (240 – 347)
Conclusion: Relatively broad distribution and different glucose values assigned to the HbA1c value in both studies increase a probability of wrong calculation. Therefore, it seems that assessment of HbA1c is necessary for the proper assessment of diabetes control, and average blood glucose estimation as a tool that may replace HbA1c should not be recommended.