Background and aims:
The Hyperglycaemia and Adverse Outcomes in Pregnancy Study group clearly demonstrated a continuum between glucose and proportion of infants born large for gestational age (LGA: ≥90th centile). We aimed to determine if this varied according to ethnicity.

Materials and Methods:
Pregnant women who underwent a 24-28 week 75g oral glucose tolerance test in an inner-city healthcare centre were retrospectively analysed (n=4562). Baseline maternal demographics, glycaemia and fetal birthweight/ birthweight centile were compared across five ethnic groups: White Caucasian (n=1379), Black African-Caribbean (n=591), South Asian (n=392), Mixed ethnicity/ any other Asian ethnicity (n=1499) and Other/ Ethnicity unknown (n=701). Variations in the proportion of LGA infants by category of glycaemia and body mass index (BMI) were examined. 

Results:
Significant differences were observed in maternal age, early pregnancy BMI, fasting plasma glucose (FPG), 120-minute value glucose, fetal birthweight and birthweight centile across the groups. The proportion of infants born LGA increased with each increment in FPG category in four of the ethnic groups but not in the Other/ Unknown group: variations in LGA incidence by category of glycaemia were significant within each ethnic group. In contrast, no clear pattern emerged between 120-minute glucose values and LGA incidence: variations in LGA incidence were only significant in Black African-Caribbean and the Other/ Unknown ethnic groups.

An overall trend towards an increasing proportion of LGA infants with each 5.0kg/m2 increment in BMI was observed: variations in LGA incidence were only significant in women of White and Mixed/ Other Asian ethnicity.

Regression analyses were then used to determine if an interaction between glucose, early pregnancy BMI and fetal birthweight existed that was dependent on ethnicity. Following adjustment for gestational age and GDM diagnosis, a 1mmol/L increase in FPG increased birthweight by 157g (p<0.001): this interaction was ethnic group dependent (effect size significantly smaller in women of Mixed ethnicity/ other Asian or Other/ Unknown group). Interactions between 120-minute glucose value and fetal birthweight were also dependent on ethnicity with black African-Caribbean women exhibiting the largest effect size (27.6g increase in birthweight for every 1mmol/L increase in glucose). These interactions persisted following adjustment for early pregnancy BMI. 

The interaction between early pregnancy BMI and fetal birthweight was not dependent on ethnicity. 

Conclusions:
These data suggest that the impact of glucose and BMI on fetal overgrowth varies according to ethnicity. This study has demonstrated an ethnic group dependent effect on the interaction between glucose and fetal birthweight. However ethnicity did not impact on the interaction with early pregnancy BMI and fetal birthweight. These data suggest that ethnicity specific glucose thresholds for the diagnosis of GDM may be warranted.

Ethnic variations in glucose, maternal weight and fetal overgrowth in a multi-ethnic inner city antenatal cohort