Introduction: Gestational diabetes, the most common medical complication of pregnancy, is diagnosed using an oral glucose tolerance test (OGTT), which has limited accuracy, reproducibility and practicality. Pre-analytical processing of glucose influences diagnosis rates, but the effect of this on a population level is unknown. We assessed how many cases of gestational diabetes were missed through suboptimal specimen processing, and if replacing the OGTT with a 28-week HbA1c was a suitable alternative for identification of hyperglycemia pregnancies.
Methods: We used data from a prospective cohort, the observational study of pregnancy hyperglycemia, endocrine causes, lipids, insulin and autoimmunity (OPHELIA), which included pregnant participants from nine UK centres recruited at the time of a 75g antenatal OGTT (criteria of the National Institute of Health and Care Excellence). Standard glucose processing (local procedures) was compared to enhanced glucose processing (storage on ice, rapid centrifugation, aliquoting and freezing <2.5 hrs).
Results: We recruited 1308 participants of mean age 31.5 years (SD 5.0) and BMI 33.0kg/m2 (6.8) of 82.5% white ethnicity, representative of the UK population. Enhanced glucose processing resulted in glucose levels ~0.6 mmol/L (10.9 mg/dL) higher than standard processing, associated with an increase in gestational diabetes diagnosis from 9% to 22%. The 13% of women who had a missed diagnosis due to standard specimen processing were not representative of the whole cohort: they were more likely to be of white ethnicity with BMI >30 kg/m2. HbA1c was not associated with most relevant pregnancy outcomes and receiver operator curves (ROC) demonstrated that the HbA1c was not a strong predictor of gestational diabetes diagnosis (AUROC 0.74; 95%CI 0.68 to 0.79).
Conclusions: Neither an oral glucose tolerance test with standard glucose processing nor 28-week HbA1c offers consistently accurate diagnosis of gestational diabetes.
L.C. Kusinski: None. D. Jones: None. P. Barker: None. K. Burling: None. I. Halsall: None. C.L. Meek: Research Support; Dexcom, Inc.
NIHR Cambridge BRC; Diabetes UK Harry Keen Intermediate Clinical Fellowship (17/0005712); EFSD-Novo Nordisk Foundation Future Leader's Award (NNF19SA058974)