Early pregnancy HbA1c as the first screening test for gestational diabetes: results from three prospective cohorts

Abstract

Background: More than 90% of gestational diabetes cases are estimated to occur in low-income and middle-income countries (LMICs). Most current guidelines recommend an oral glucose tolerance test (OGTT) at 24-28 weeks of gestation. The OGTT is burdensome, especially in LMICs, resulting in a high proportion of women not being screened. We aimed to develop a simple and effective screening strategy for gestational diabetes. Methods: STRiDE, a prospective cohort study, was set up in seven centres in south India and seven centres in western Kenya, and included pregnant women aged 18-50 years of age and at less than 16 weeks of gestation (<20 weeks in Kenya), confirmed by dating ultrasound. We assessed the efficacy of early pregnancy HbA1c (venous and capillary point-of-care), either alone or as part of a composite risk score with age, BMI, and family history of diabetes, in predicting gestational diabetes at 24-28 weeks of gestation, in two LMICs (India and Kenya) and in a UK multi-ethnic population from the PRiDE study. A key secondary outcome was to assess whether an early pregnancy composite risk score can reduce the need for OGTTs. Gestational diabetes was diagnosed using current WHO criteria. Findings: Between Feb 15, 2016, Dec 13, 2019, we enrolled 3070 participants in India and 4104 in Kenya. 4320 participants were included from the PRiDE cohort. Gestational diabetes prevalence by OGTT at 24-28 weeks was 19·2% in India, 3·0% in Kenya, and 14·5% in the UK. Early pregnancy HbA1c was independently associated with incidence of gestational diabetes at 24-28 weeks of gestation. Adjusted risk ratios were 1·60 (95% CI 1·19-2·16) in India, 3·49 (2·8-4·34) in Kenya, and 4·72 (3·82-5·82) in the UK. Composite risk score models that combined venous or point-of-care HbA1c with age, BMI, and family history of diabetes best predicted testing positive for gestational diabetes. A population-specific, two-threshold screening strategy of rule-in and rule-out gestational diabetes using early pregnancy composite risk score could reduce the requirement of OGTTs by 50-64%. For the HbA1c-alone model, the thresholds were 5·4% (rule in) and 4·9% (rule out) in India, 6·0% (rule in) and 5·2% (rule out) in Kenya, and 5·6% (rule in) and 5·2% (rule out) in the UK. Interpretation: Early pregnancy HbA1c offers a simple screening test for gestational diabetes, allowing those at highest risk to receive early intervention and greatly reduce the need for OGTTs. This can also be carried out using point-of-care HbA1c in LMICs.