Aims: To test if the use of either HbA(1c) level or calculated clinical scores including two published scores and a new score (the Bondy score) could help in selecting overweight or obese women who should benefit from oral glucose tolerance test (OGTT) to detect dysglycaemia.
Methods: The French Data from an Epidemiological Study on the Insulin Resistance Syndrome (DESIR) score and the Finnish Diabetes risk score (Findrisk) were calculated, whereas the Bondy score was built in a development sample of 698 women, BMI 37±7 kg/m(2), without known diabetes. External validation was performed in a validation sample of 212 women.
Results: A dysglycaemia (according to OGTT results, WHO criteria) was diagnosed in 161 (23.1%) patients. Sensitivity of fasting plasma glucose (FPG)≥6.1 mmol/l and HbA(1c)≥6% to identify dysglycaemia were respectively 27 and 41%. Areas under Receiver Operator Curve (AROC) of HbA(1c), DESIR score and Findrisk to detect dysglycaemia were 0.630 [95% confidence interval 0.580-0.680], 0.606 [0.561-0.652] and 0.635 [0.588-0.683], respectively. The Bondy score, based on age and waist circumference, had a better AROC (0.674 [0.626-0.721]) than the DESIR score (P<0.05). These performances were confirmed in the validation sample. Performing OGTT only in subjects with a Bondy score≥4 (41% of the sample) had a sensitivity of 61% and a better net benefit (0.128) than measuring FPG in all subjects (0.069).
Conclusion: Performing OGTT in obese women selected on a simple clinical score is more sensitive to identify dysglycaemia than measuring FPG in all of them and may be cost-effective.
Copyright © 2011 Elsevier Masson SAS. All rights reserved.