Gestational diabetes treatment: Associations with maternal characteristics and offspring outcomes in the Born in Bradford study

Background: Gestational diabetes mellitus (GDM) is characterised by an impaired tolerance to glucose during pregnancy. If left untreated, GDM is associated with macrosomia and delivery complications and a higher risk of obesity and diabetes in childhood. GDM treatment consists of initial lifestyle changes advice which, if insufficient to restore euglycaemia, is followed by supplemental pharmaceutical treatment, using insulin or metformin. To date, the differences in maternal characteristics between supplemental pharmaceutical treatment and lifestyle changes advice groups remain largely uncertain in the UK. In addition, the extent to which each GDM treatment type can offset the adverse effects of exposure to GDM both at birth and in early childhood is unknown. Lastly, the associations between foetal exposure to metformin, compared to insulin, and birth anthropometry and childhood growth remain unclear.

Objectives: This thesis had three objectives: (1) to identify the maternal characteristics associated with GDM pharmaceutical treatment, (2) to examine the relationships between maternal GDM treatment type and neonatal anthropometry and (3) to investigate the associations between maternal GDM treatment type and childhood growth trajectories from birth to five years of age.

Methods: All analyses were conducted using data from the Born in Bradford (BiB) cohort.

Objective 1 Logistic and multinomial regression models were used to estimate the associations between maternal characteristics (e.g., glucose concentrations at diagnosis, age, ethnicity) and GDM pharmaceutical treatment, relative to lifestyle changes advice and metformin, relative to insulin.

Objective 2 In linear regression models, compared to infants not exposed to GDM, the relationships between foetal exposure to GDM treatment and birth weight, head, mid-arm and abdominal circumference and triceps and subscapular skinfold thickness were evaluated.Differences in neonatal anthropometry between metformin- and insulin-exposed infants were also quantified using linear regression.

Objective 3 A multilevel linear spline analysis was conducted to compare the weight-, length- and BMI-for-age z-scores trajectories of offspring exposed to maternal GDM treatment and those born to mothers without GDM.

Results: Objective 1 Higher severity of hyperglycaemia, obesity, smoking and White British ethnicity were maternal characteristics associated with supplemental GDM pharmaceutical treatment rather than lifestyle changes advice alone. Compared to insulin treatment, obese mothers had a higher risk of metformin treatment (relative risk ratio 3.2 (95% Confidence Interval (CI) 1.3, 7.8) and women with higher fasting glucose concentrations had a lower risk of metformin treatment (relative risk ratio 0.3 (95% CI 0.2, 0.6)).

Objective 2 Infants exposed to GDM treatment of any kind were predicted to be lighter and less adipose and have a larger head circumference at birth than infants not exposed to GDM. Foetal exposure to metformin was associated with smaller mid-arm circumference at birth (-0.3cm (95% CI -0.6, -0.07)) compared to exposure to insulin.

Objective 3 After birth, infants exposed to insulin or metformin, and not those exposed to lifestyle changes advice alone, were predicted to have more rapid gains in weight, length and BMI than infants not exposed to GDM. By five years of age, higher weight and height z-scores were observed for both insulin-exposed children (weight: 0.20 (95% CI 0.050, 0.35), height: 0.22 (95% CI 0.075, 0.36) and metformin-exposed children (weight: 0.14 (95% CI -0.19, 0.47), height: 0.35 (95% CI 0.027, 0.68), compared to children born to mothers without GDM. Metformin-exposed infants were taller and lighter than insulin-exposed infants by five years of age.

Conclusions: In addition to a higher severity of hyperglycaemia, maternal risk factors and health behaviours, which seemed to be underpinned by ethnic differences, could have contributed to the clinical decision to prescribe supplemental pharmaceutical treatment rather than lifestyle changes advice alone. GDM treatment of any kind was associated with not only an attenuation but a change in the direction of the positive associations that are normally described between foetal exposure to GDM and higher weight and adiposity at birth. After birth, faster growth in the offspring exposed to pharmaceutical treatment compared to those born to mothers without GDM may reflect postnatal influences such as infant feeding methods, child physical activity levels and dietary intake rather than treatment itself. Metformin exposure was not associated with adverse neonatal anthropometry and childhood growth patterns when compared to insulin.