Maternal pre-pregnancy overweight and gestational diabetes and dietary intakes among young adult offspring

Obesity affects billions worldwide and increases risks for diabetes, cardiovascular disease, and cancer. Its etiology is multifactorial, involving genetic, lifestyle, and environmental influences, with 40–70% of variability attributed to genetic factors. Prenatal environment also contributes: maternal overweight/obesity and gestational diabetes (GDM) can alter fetal growth and metabolic regulation, potentially inducing epigenetic changes that predispose to later overweight/obesity. Prior work links high maternal pre-pregnancy BMI and GDM to unfavorable offspring body composition from infancy through adulthood. Conversely, maternal healthy lifestyle pre-pregnancy reduces offspring obesity risk. Diet is a key modifiable risk factor for non-communicable diseases globally. The study hypothesized that exposure to maternal pre-pregnancy overweight/obesity or GDM programs dietary habits in offspring, leading to less healthy adult dietary intake patterns.

The paper summarizes evidence that: (1) maternal pre-pregnancy BMI and GDM are associated with higher adiposity and metabolic risks in offspring from infancy to adulthood; (2) maternal healthy lifestyle greatly lowers offspring obesity risk; (3) diet-related epigenetic changes may contribute to obesity and diabetes pathogenesis; and (4) substantial global mortality and morbidity are attributable to dietary risks. It also cites studies linking early-life conditions (e.g., birth size, preterm birth) with later dietary preferences and macronutrient composition, suggesting prenatal and perinatal factors can shape long-term dietary habits.

Design and cohorts: Combined analysis from two prospective Finnish birth cohorts: ESTER Maternal Pregnancy Disorders Study (participants born 1985–1989 in Northern Finland) and Arvo Ylppö Longitudinal Study (AYLS; births in 1985–1986 in Uusimaa). Clinical examinations and questionnaires were conducted in 2009–2011 (ESTER) and 2009–2012 (AYLS). Inclusion and grouping: Participants were categorized into three groups based on maternal pregnancy characteristics: (1) OGDM: offspring of mothers with gestational diabetes (n=190), regardless of maternal BMI; (2) ONO: offspring of normoglycaemic mothers with pre-pregnancy BMI ≥25 kg/m² and no GDM (n=155); (3) Controls: offspring of normoglycaemic mothers with pre-pregnancy BMI <25 kg/m² and no GDM (n=537). Exclusions: Offspring of mothers with type 1 (n=28) or type 2 diabetes (n=1), participants who were pregnant at examination (n=9), or reported cerebral palsy (n=8), mental disability (n=11), or severe physical disability (n=5). Final analytic sample: 882 participants (51% women), mean age 24.2 years (SD 1.3). Exposure ascertainment: Maternal GDM diagnoses were confirmed from records using 1980s Finnish criteria (venous glucose cutoffs after 75-g OGTT: fasting >5.5 mmol/L, 1-h >11.0 mmol/L, 2-h >8.0 mmol/L; ≥1 abnormal value required). Perinatal data included gestational age, maternal hypertension/pre-eclampsia, smoking during pregnancy, birthweight, etc. Anthropometry and covariates: Participant height, weight, and waist circumference were measured using standardized protocols; BMI calculated as kg/m². Questionnaires captured smoking, medications, and health status. Parental education (as proxy for childhood SES) categorized into four levels. Dietary assessment: Habitual diet assessed by a validated, semi-quantitative FFQ (131 items) covering the previous 12 months, completed on-site and reviewed by a trained nurse. Portion sizes were sex-specific. Nutrient intakes were calculated using the Finnish Fineli database; total energy in kJ/day; macronutrient intakes as percentage of total energy (E%). Diet quality: Recommended Finnish Diet Index (RDI) computed based on Finnish nutrition recommendations, incorporating: fruits/berries, vegetables, rye, salt, sucrose, alcohol, ratio of white to red/processed meat, and ratio of polyunsaturated to saturated plus trans-fatty acids. Components were scored by quartiles; alcohol scored 0/1. Maximum score 22 (including alcohol) or 21 (excluding alcohol); higher indicates healthier diet. Statistical analysis: Linear regression compared energy, macronutrient intakes, RDI, and RDI components between exposure groups and controls, stratified by sex. Three models: Model 1 adjusted for age and cohort (plus total energy for RDI analyses); Model 2 additionally adjusted for parental education, birth weight SD score, gestational age, maternal smoking during pregnancy, maternal hypertension, and pre-eclampsia; Model 3 additionally adjusted for participant BMI, smoking, and living at parental home. Non-normal variables (rye, fruits, vegetables) were log-transformed [ln(x+1)]; results presented as back-transformed percentages. Sensitivity analyses: underreporting assessed via Goldberg cut-off (reported energy intake/basal metabolic rate ≤1.14). Analyses were rerun excluding underreporters and then including all participants with additional adjustment for underreporting. Twins were excluded in a sensitivity check (n=12), and additional adjustment for maternal gestational weight gain was tested; results were unchanged. Ethics: Conducted per the Declaration of Helsinki with approvals from relevant ethics committees; written informed consent obtained.

• Sample characteristics: Mothers with GDM or pre-pregnancy overweight/obesity had higher pre-pregnancy BMI and more hypertensive disorders; their pregnancies were shorter; offspring were more often large for gestational age (GDM) or preterm/small for gestational age (ONO). As adults, ONO and OGDM participants had higher BMI and waist circumference than controls; OGDM participants more often lived with parents.
• ONO vs. controls (primary analyses):
  • Men: Total energy and macronutrient intakes were broadly similar, but after full adjustment (Model 3), daily carbohydrate intake was higher by 2.2 E% (95% CI 0.4, 4.0). Other macronutrients (fat, protein, sucrose, alcohol) showed no significant differences.
  • Women: Macronutrient shares similar; total energy intake appeared lower in Model 1 (-587.2 kJ/day; 95% CI -1192.0, 4.4) but attenuated with adjustment (Models 2–3).
  • Diet quality: RDI did not differ meaningfully (men mean difference 0.40; 95% CI -0.38, 1.18; women 0.25; 95% CI -0.50, 1.00).
• OGDM vs. controls: Total energy intake, macronutrient composition, RDI, and its components were similar for both men and women across models.
• Underreporting: Based on Goldberg cut-off, underreporting prevalence was 49.7% (ONO), 41.6% (OGDM), and 35.4% (controls) (n=346 underreporters).
  • Excluding underreporters: OGDM-men had higher energy intake in Model 1 (1342 kJ/day; 95% CI 125, 2559; p=0.031), which attenuated after further adjustment. ONO-men and ONO-women showed higher carbohydrate intake (men +2.1 E% [0.0, 4.1], p=0.047; women +3.2 E% [1.0, 5.4], p=0.006) and lower fat intake (men -1.6 E% [-3.2, 0.1], p=0.058; women -3.0 E% [-4.8, -1.2], p=0.001) vs controls; sucrose, protein, alcohol similar. ONO participants reported higher fruit intake (men +44.4% [2.5, 103.4], p=0.036; women +42.8% [6.1, 92.1], p=0.019); in men this attenuated with further adjustment. ONO-women had higher vegetable intake after adjustment (Model 2: +28.8% [2.8, 61.3], p=0.028; Model 3: +27.1% [1.0, 59.9], p=0.041). OGDM vs controls remained largely similar across components.
  • Adjusting for underreporting in all participants: ONO-men had higher carbohydrate intake (Model 1: +1.6 E% [0.0, 3.3], p=0.049; Model 3: +2.3 E% [0.5, 4.1], p=0.011). Other findings remained similar.
• Overall: The most consistent difference was higher carbohydrate intake among male offspring exposed to maternal pre-pregnancy overweight/obesity. Exposure to GDM did not associate with adult dietary intake or diet quality metrics.

The study tested whether prenatal exposure to maternal pre-pregnancy overweight/obesity or GDM is linked to adult offspring dietary intake and quality. Findings indicate largely similar diets between exposed and unexposed groups, with the notable exception of higher carbohydrate intake among men exposed to maternal pre-pregnancy overweight/obesity. Accounting for underreporting suggested a similar pattern in women (higher carbohydrates, lower fat) and higher fruit/vegetable consumption in ONO-women, indicating that increased carbohydrate intake may partly reflect higher consumption of complex carbohydrates from fruits and vegetables rather than simple sugars. The absence of differences in OGDM suggests that intrauterine hyperglycemia per se did not program adult diet in this sample. The observed patterns may reflect a complex interplay between prenatal programming, shared familial lifestyle/genetics, and current individual factors (BMI, smoking, living situation), as statistical significance in men emerged after adjusting for current characteristics. The results contribute to life-course perspectives on obesity by suggesting that while prenatal exposures influence adiposity, their translation into adult dietary behavior appears limited, with only modest, sex-specific differences detected.

Young adults exposed in utero to maternal pre-pregnancy overweight/obesity or GDM generally exhibited similar overall dietary intake and adherence to a recommended dietary pattern as controls. A consistent exception was higher carbohydrate intake among male offspring of mothers with pre-pregnancy overweight/obesity, with suggestive evidence of higher carbohydrate and lower fat intake among women when accounting for underreporting. These findings imply that prenatal exposure to maternal overweight/obesity may be associated with subtle, sex-specific differences in adult macronutrient composition, whereas exposure to GDM was not associated with adult dietary habits. Future research should explore underlying mechanisms (e.g., epigenetic pathways), refine dietary assessment with objective measures, examine food sources of carbohydrates and fats in more detail, and assess longitudinal changes across the life course to clarify causality and clinical relevance.

• Dietary assessment relied on a self-reported, though validated, FFQ, susceptible to misreporting, particularly underreporting of unhealthy foods and overreporting of healthy foods.
• Underreporting varied by group (highest in ONO); although addressed via exclusion and adjustment, residual bias is possible.
• Some potentially influential factors (e.g., eating disorders, physical activity, sedentary behavior) were unavailable and thus unadjusted.
• While powered to detect moderate-to-large differences, small between-group differences cannot be excluded.
• Generalizability may be limited to similar populations and birth cohorts; data sharing is restricted by participant consent.