The relationship between maternal dietary patterns during pregnancy in women with gestational diabetes mellitus and infant appetitive feeding behaviour at 6 months

The study investigates whether maternal dietary patterns in late pregnancy among women with gestational diabetes mellitus (GDM) are associated with differences in infant appetitive feeding behaviours at 6 months of age. Appetite, the internal drive to seek and ingest food, is influenced by prenatal and early postnatal exposures. Flavour exposure begins in utero via amniotic fluid and continues through breastfeeding, potentially shaping later preferences. Prior human and animal studies suggest maternal diet can program offspring appetite and feeding behaviour via neurobiological pathways (homeostatic and hedonic systems) and possibly through effects on hypothalamic development and the gut microbiome. Infant appetitive traits such as food responsiveness, enjoyment of food, slowness in eating, and satiety responsiveness are associated with subsequent weight gain and adiposity. Women with GDM and their offspring are at elevated risk of later obesity and metabolic disease, but relationships between maternal dietary patterns in GDM and infant appetitive traits have not been examined. The authors hypothesised that: (1) infants of women with GDM would exhibit obesity-related appetitive traits (higher food responsiveness and enjoyment of food); (2) unhealthy maternal dietary patterns high in energy-dense, nutrient-poor foods would be associated with higher infant food responsiveness, enjoyment of food, and general appetite; and (3) healthy maternal dietary patterns high in nutrient-dense foods would be associated with higher satiety responsiveness and slowness in eating in infants.

Evidence indicates prenatal and early postnatal diet exposures shape infant flavour preferences and appetitive behaviours. Human studies link maternal macronutrient intake during pregnancy with later offspring dietary intakes, and breastfeeding exposure to greater flavour variety with broader infant acceptance of flavours relative to formula feeding. Animal studies show that maternal diet (e.g., high-fat feeding, undernutrition) can program appetite-related neural circuitry and alter hypothalamic gene expression, with downstream effects on feeding behaviour. Maternal diet may also influence early gut microbiome colonisation, producing metabolites that epigenetically affect genes regulating appetite. In infants and children, higher food responsiveness and enjoyment of food are associated with higher BMI and faster weight gain, whereas higher satiety responsiveness and slowness in eating are associated with lower BMI and slower weight gain. Despite suggestive evidence, findings are inconsistent, and the timing and persistence of altered appetitive traits are unclear. Few studies have characterised dietary patterns specifically among women with GDM; prior work in New Zealand and South Korea identified patterns resembling “Junk/Western” and “Health-conscious/vegetable-rich,” supporting the relevance of pattern-based approaches in this population.

Study design: Nested cohort secondary analysis within the TARGET Trial, a multicentre, stepped-wedge randomised trial comparing tighter versus less tight glycaemic control targets in women with GDM in New Zealand (2015–2017). Ethical approval obtained; informed consent provided. Inclusion for this study required completed FFQ at 36 weeks’ gestation, available infant sex and gestational age, pregnancy outcomes, and 6-month infant follow-up. Participants: Women with singleton pregnancies diagnosed with GDM ≥22 weeks’ gestation from 10 NZ hospitals and their infants. Maternal diet assessment: A self-administered, 1-month recall, semi-quantitative FFQ (customised from the 163-item Willett FFQ) was completed at 36 weeks’ gestation. The instrument included 65 questions; 57 food-group items were used for dietary pattern analysis covering dairy, eggs and meat, fish/seafood, breads/cereals/starches, fruits, vegetables, fast foods, beverages, sweets, baked goods, and miscellaneous. Frequency responses (ordinal) were converted to estimated weekly frequencies (e.g., never=0; 1–3/month=0.5; 1/week=1; 2–4/week=3; 5–6/week=5.5; 1/day=7; 2–3/day=17.5; 4–6/day=35). Exclusions: participants with >10 missing dietary items (assuming “never” for ≤10 missing items) and implausible energy intake (<500 or >3500 kcal/day). Dietary pattern derivation: Principal component analysis (PCA) using a polychoric correlation matrix (suitable for ordinal Likert-type data) with varimax rotation was performed in SAS (PROC FACTOR, method=principal). Data suitability was assessed via correlation matrix, Kaiser-Meyer-Olkin (KMO) measure, and Bartlett’s test of sphericity. Number of components retained was based on eigenvalues >1, scree plot inflection, and interpretability after rotation. Items with absolute loadings ≥0.3 characterised patterns. Internal consistency of items within each pattern was evaluated using Cronbach’s alpha; items with poor item-total correlations were considered for removal if improving reliability. Standardised, weighted component scores were computed and tertiles generated; higher scores indicate greater adherence to the pattern. Infant feeding practices: At 6 months postpartum, breastfeeding status (exclusive, predominantly/partially, or formula-fed) and age at introduction of solids were collected via questionnaire. Infant appetitive behaviour: At 6 months, mothers completed the Baby Eating Behaviour Questionnaire (BEBQ), with subscales for enjoyment of food (4 items), food responsiveness (6 items), slowness in eating (4 items), and satiety responsiveness (3 items), plus a general appetite item. Two items were reverse-scored; mean subscale scores were calculated and categorised as low (1–≤2.33), medium (>2.33–≤3.66), and high (>3.66–5), where higher scores indicate higher trait levels. Covariates: Considered variables included maternal age, total energy intake, BMI category, smoking status, parity, ethnicity, socioeconomic deprivation (NZDep quintiles), TARGET treatment arm, infant gestational age, birthweight, 6-month weight and weight-for-length z scores, age at solids introduction, and breastfeeding status. Variable selection for multivariable models emphasised parsimony and biological plausibility using stepwise, Max R-squared, forward, and backward selection. Statistical analysis: Descriptive statistics summarised cohort characteristics and infant appetitive traits. Maternal dietary patterns were derived via PCA; internal consistency was assessed by Cronbach’s alpha. Associations between maternal pattern scores (per SD) and infant appetitive traits were examined using general linear models with covariate adjustment per parsimonious models; sex interactions were tested. Relationships among infant appetitive traits were assessed using Pearson or Spearman correlations as appropriate. Chi-square tests compared categorical variables. Analyses were conducted in SAS 9.4.

- Sample: Of 339 women with FFQ data at 36 weeks’ gestation, 14 were excluded (8 with ≥10 missing FFQ items; 6 implausible energy intakes), leaving 325 for dietary pattern analysis. Of these, 247 (76%) completed the 6-month infant appetitive behaviour questionnaire. - Infant feeding at 6 months: 14.2% exclusively breastfed; 61.5% predominantly/partially breastfed; 24.3% exclusively formula-fed. Mean age at solids introduction was 5.1 months (SD 0.72). - Infant appetitive traits (overall): 92.3% had high enjoyment of food (>3.66–5); 54.5% had high general appetite; 4.9% had high food responsiveness. Food responsiveness correlated with general appetite (r=0.44, p<0.0001) and slowness in eating (r=0.21, p=0.0008). Enjoyment of food correlated negatively with satiety responsiveness (r=−0.22, p=0.0006) and slowness in eating (r=−0.22, p=0.0006). - Maternal dietary patterns: PCA of 57 FFQ items (KMO=0.75; Bartlett’s p<0.0001) retained three components explaining 28.3% of total variance: (1) Junk (energy-dense, discretionary foods and drinks); (2) Mixed (vegetables, fats/oils, starchy vegetables, nuts, dairy, meats, beverages incl. tea/coffee and water); (3) Health-conscious (high-fibre cereals, brown rice, fruits, fish/seafood, legumes, eggs, low-fat cheese). Cronbach’s alpha: Junk 0.63; Mixed 0.81; Health-conscious 0.75. - Associations between maternal diet and infant appetitive traits: In multivariable models adjusting for sex, 6-month weight-for-age z score, and NZDep at study entry (parsimonious model): • Health-conscious maternal pattern was inversely associated with enjoyment of food in boys (β=−0.24, 95% CI −0.36 to −0.11, p=0.0003) but not in girls (β=−0.02, 95% CI −0.12 to 0.08, p=0.70); sex interaction p=0.004. • Health-conscious pattern was positively associated with slowness in eating (β≈0.13, 95% CI about 0.02 to 0.24; p≈0.01–0.025). • No other significant associations were observed for Junk or Mixed patterns with appetitive traits after adjustment.

The study shows that among women with GDM, a late-pregnancy Health-conscious dietary pattern is associated with infant appetitive traits at 6 months that are linked to lower obesity risk: specifically, lower enjoyment of food in boys and higher slowness in eating overall. These findings align with evidence that prenatal nutritional environments can program appetite regulation via neurobiological pathways (homeostatic and hedonic systems) and possibly via microbiome-mediated epigenetic mechanisms. The observed sex-specific association (significant in boys but not girls) may relate to known sex differences in obesity risk and dietary preferences across childhood, with boys often exhibiting higher preferences for energy-dense foods. The identification of three robust dietary patterns in GDM mirrors patterns found in other NZ cohorts, supporting generalisability within this context. Clinically, results suggest that promoting healthier dietary patterns in late pregnancy among women with GDM may improve early appetitive control in offspring, potentially moderating later obesity risk, particularly in boys. However, the persistence of these associations and their translation into differences in growth and adiposity remain to be determined.

Among women with gestational diabetes mellitus, a Health-conscious dietary pattern in the third trimester is associated with infant appetitive behaviours at 6 months indicative of better appetitive control (lower enjoyment of food in boys and higher slowness in eating). Given the elevated obesity risk in offspring of women with GDM, maternal dietary guidance during pregnancy may represent a viable intervention target to influence early appetitive traits. Future research should assess whether these associations persist into later childhood, relate to actual dietary intakes and growth trajectories, and explore mechanisms underlying sex-specific effects.

- Generalisability: Responders to the infant questionnaire differed from non-responders in ethnicity and socioeconomic status; cohort may not fully represent all women with GDM in New Zealand. - Measurement bias: Maternal diet and infant appetitive behaviour were self-reported (FFQ and questionnaire), susceptible to recall and social desirability biases compared to real-time dietary records or observational measures. - PCA subjectivity: Choice of food groupings, number of components, and interpretation can introduce subjectivity; although internal consistency was acceptable and patterns align with prior NZ work. - Variance explained: The three patterns explained 28.3% of dietary variance, leaving substantial unexplained variability in diet. - Residual confounding: Despite adjustment for multiple covariates, unmeasured or inadequately measured confounders may remain. - Short follow-up: Appetitive behaviours were assessed only at 6 months; persistence and links to later diet, growth, and adiposity are unknown.