Background:
Systematic reviews show poor sensitivity of HbA1c in identifying GDM1. However, elevated HbA1c is associated with adverse pregnancy outcomes2.
Aim:
Methods:
We examined de-identified prospectively collected demographic and outcome data from singleton ADIPS-2014 criteria diagnosed GDM pregnancies (Mar2016-Mar2024), with a sub-group of 812 pregnancies with capillary blood glucose levels (BGLs). All women were provided education on blood glucose monitoring and medical nutritional therapy. Insulin was commenced if treatment targets were not met (FBGL<5.3mmol/L, 1hrBGL<7.4mmol/L, 2hrBGL<7.0mmol/L). Metformin was not used. Sensitivity for GDM diagnosis was calculated across thresholds for HbA1c. Two groups were defined – High and Low HbA1c groups, according to patients above and below the mean HbA1c in each cohort, respectively. Pearson’s correlation coefficient (r) was calculated for HbA1c against plasma fasting, 1hr-OGTT, 2hr-OGTT, and capillary mean fasting/post-prandial BGLs.
We compared differences in demographic and outcome findings between the groups. Outcomes assessed included insulin use, caesarean section, early delivery (<37weeks), LGA (>90th percentile), SGA (<10th percentile), neonatal hypoglycaemia (<2.6mmol/L) and neonatal jaundice (requiring phototherapy).
Results:
There were 2990 GDM pregnancies. Mean HbA1c was 5.2±0.4%. The sensitivities for GDM diagnosis using different HbA1c thresholds decreased with increasing HbA1c-cutoffs of 4.8%, 5.1%, 5.3% and 5.6% corresponded to sensitivities of 89%, 61%, 42% and 18%, respectively. HbA1c was more strongly correlated with fasting plasma glucose (r=0.26) and fasting capillary BGLs (r=0.29) than 1hr-OGTT (r=0.24), 2hr-OGTT (r=0.23) and mean post-prandial capillary BGLs (r=0.213), all p<0.0001. Higher HbA1c (1569/2900, 54%), defined as above HbA1c≥ 5.2% was associated with higher maternal age (32.0±5.2vs31.0±5.4,p<0.0001), gravida (3.0±1.9vs2.8±1.8, p<0.0001), parity (1.4±1.4vs1.2±1.3, p<0.0001), maternal BMI (28.6±6.5vs25.6±5.4kg/m2, p<0.0001), and total gestational weight gain (9.4±6.8vs8.6±5.3kg, p<0.001). Furthermore, higher HbA1c was associated with increasing rates of insulin therapy (57.4%vs37.6%, p<0.0001), caesarean section (40.7%vs31.3%, p<0.0001), LGA (14.7%vs10.0%, p<0.001), neonatal hypoglycaemia (12.1%vs9.2%, p<0.05), and higher insulin requirements (33.0±32.3vs22.0±18.7units, p<0.0001).
Conclusions:
(1) HbA1c should not be used for GDM diagnosis - as approximately 50% of women with GDM will have a HbA1c<5.2%,
(2) HbA1c is slightly better correlated with fasting as compared to post-prandial glycaemic status.
(3) GDM women with higher HbA1c have higher risk maternal characteristics and have higher rates of adverse pregnancy outcomes.