Abstract
Background: In spite of increased vigilance of undiagnosed type 2 diabetes (DM2), the prevalence of unknown DM2 in subjects with morbid obesity is not known.
Aim: To assess the prevalence of undiagnosed DM2 and compare the performance of glycated A1c (HbA1c) and fasting glucose (FG) for the diagnosis of DM2 and prediabetes (preDM) in patients with morbid obesity.
Patients and methods: We measured fasting glucose and HbA1c in 537 consecutive patients with morbid obesity without previously known DM2.
Results: A total of 49 (9%) patients with morbid obesity had unknown DM2 out of which 16 (33%) fulfilled both the criteria for HbA1c and FG. Out of 284 (53%) subjects with preDM, 133 (47%) fulfilled both the criteria for HbA1c and FG. Measurements of agreement for FG and HbA1c were moderate for DM2 (κ = 0.461, p < .001) and fair for preDM (κ = 0.317, p < .001). Areas under the curve for FG and HbA1c in predicting unknown DM2 were 0.970 (95% CI 0.942, 0.998) and 0.894 (95% CI 0.837, 0.951) respectively. The optimal thresholds to identify unknown DM2 were FG ≥6.6 mmol/L and HbA1c ≥ 6.1% (43 mmol/mol).
Conclusions: The prevalence of DM2 remains high and both FG and HbA1c identify patients with unknown DM2. FG was slightly superior to HbA1c in predicting and separating patients with unknown DM2 from patients without DM2. We suggest that an FG ≥6.6 mmol/L or an HbA1c ≥6.1% (43 mmol/mol) may be used as primary cut points for the identification of unknown DM2 among patients with morbid obesity.
Acknowledgements
We acknowledge the contribution of Study coordinator Iren Pedersen at the Section for Morbid Obesity at Akershus University Hospital HF for the persistent effort of data collection. Thanks are also due to Fredrik Andreas Dahl for valuable advice with the statistical analysis.
Disclosure statement
No potential conflict of interest was reported by the authors.