Both the WHO and ADA endorse HbA1c as a screening test for undiagnosed type 2 diabetes in early pregnancy. However, limited pregnancy-specific HbA1c data exist with recommended cut-points based on data from non-pregnant populations. Our aim was to determine pregnancy-specific HbA1c centiles by gestation and by ethnicity.
This is a population-based observational study. Using electronic records, laboratory data were matched to births in Christchurch during 2008-2010. Inclusion criteria were a normal glucose challenge test or 75g glucose tolerance test. Included were 6800 pregnancies, European 80% (5462), Māori 6% (415), Pacific 3% (196), and 11% (727) ‘Others’ (mostly Asian). HbA1c fell early in pregnancy reaching a nadir at 24 weeks’ gestation, thereafter increasing to baseline levels or beyond. The 97.5th centile for HbA1c by gestation in a European woman age 30 years is 5.76% (39.5mmol/mol) at 8+0 weeks’, 5.70% (38.8mmol/mol) at 16+0 weeks’, 5.66% (38.4mmol/mol) at 24+0 weeks’, and 5.99% (42.0mmol/mol) at 32+0 weeks’. Non-European women had higher mean HbA1c than Europeans, the estimated mean (SD) difference in Māori +0.13% (0.05) (1.4mmol/mol (0.5)), Pacific +0.20% (0.03) (2.2mmol/mol (0.3)), ‘Others’ +0.10% (0.03) (1.1mmol/mol (0.3)). Māori and Pacific women also had higher fasting plasma glucose levels, mean (95% confidence interval) +0.07mmol/L (0.02, 0.12) and +0.19mmol/L (0.13, 0.26) respectively, while ‘Others’ had higher plasma glucose post glucose load, +0.35mmol/L (0.24, 0.45) at 2-hours, all within the ‘normal’ range.
In conclusion, our innovative HbA1c centile chart adjusted by gestation is useful to consider when using HbA1c to screen for pre-existing hyperglycaemia in early pregnancy. Ethnicity-specific HbA1c charts may be unnecessary as glycaemia partly accounts for this observed variance. Further study could clarify whether the late pregnancy rise within the higher HbA1c centiles is due to unrecognised gestational diabetes or confounding by iron deficiency.