Determinants of cord blood adipokines and association with neonatal abdominal adipose tissue distribution

Adipokines, including leptin and adiponectin, are adipocyte-derived hormones involved in energy homeostasis and metabolism. Leptin reflects adipose triglyceride stores and regulates central energy balance; it is generally elevated in obesity and varies by ethnicity. Numerous studies report positive associations between cord leptin and birth weight, and maternal characteristics such as obesity, smoking, and glycemia associate with higher cord leptin. Adiponectin mediates communication between adipose tissue and metabolic organs, promoting insulin sensitivity and suppressing gluconeogenesis in adults; lower adult adiponectin relates to adverse metabolic profiles and varies by ethnicity. Prenatal exposures (maternal obesity, GDM) have been linked to higher cord adiponectin; cord adiponectin correlates positively with leptin, fetal growth, and birthweight, and is higher in fetuses than adults. Increased abdominal adiposity (AA), particularly visceral compartments, is an independent risk factor for adverse cardiometabolic outcomes; South Asians exhibit greater abdominal obesity and insulin resistance. Prior work in the GUSTO cohort demonstrated ethnic differences in neonatal abdominal adipose tissue (AAT) distribution, with higher abdominal subcutaneous fat among Indian neonates despite lower birthweight. While cord leptin and adiponectin have been associated with overall newborn adiposity by anthropometry, few studies have assessed MRI-quantified AA in early infancy. Research questions: (1) What maternal and fetal factors determine cord blood leptin and adiponectin concentrations in Asian neonates? (2) How are cord leptin and adiponectin associated with neonatal adiposity and abdominal fat distribution measured by MRI in early infancy? Understanding these determinants and associations may illuminate early-life pathways influencing abdominal fat accumulation and later metabolic risk.

Prior literature indicates: (a) Cord leptin positively associates with birth size and predicts infancy weight gain; maternal obesity, smoking, and higher glycemia elevate cord leptin. (b) Adult adiponectin is inversely related to adiposity and insulin resistance, with ethnic differences, yet fetal/cord adiponectin is higher than adult levels and positively correlates with fetal growth, birthweight, and cord leptin. (c) Maternal adiponectin appears inversely associated with offspring birth weight and adiposity, while high fetal adiponectin may enhance insulin’s growth-promoting effects; animal studies show fetal adiponectin enhances fetal fat deposition, particularly in maternal obesity. (d) Increased abdominal adiposity, especially visceral fat, predicts cardiometabolic risk in youth and adults; South Asians show greater abdominal obesity at similar BMI. (e) Few studies have examined neonatal abdominal fat compartments by MRI in relation to cord adipokines, representing a gap this study addresses.

Design and cohort: Prospective mother-offspring birth cohort (GUSTO) in Singapore. Pregnant women ≥18 years were recruited at <14 weeks’ gestation from KKH and NUH (June 2009–September 2010). Ethical approvals obtained; informed consent provided (ClinicalTrials.gov NCT01174875). Maternal data: Demographics, lifestyle, obstetric/medical history from questionnaires and records. Pre-pregnancy BMI (ppBMI) calculated from self-reported weight and measured height; categorized by WHO Asian cutoffs. At 26–28 weeks, 75 g OGTT performed; glucose by hexokinase methods. GDM defined by 1999 WHO criteria (fasting ≥7.0 mmol/L or 2 h ≥7.8 mmol/L). Gestational weight gain categorized by 1990 IOM guidelines. Infant measurements: Birth anthropometrics from records. Triceps and subscapular skinfolds measured in triplicate (Holtain calipers). Fat mass predicted using validated GUSTO equation; in a subset, body composition measured by PEA POD (air displacement plethysmography). After exclusions (no consent, BW <2.5 kg, PEA POD %fat <5%, missing adipokines), 259 neonates remained for PEA POD–adipokine analyses. Cord blood adipokines: Umbilical venous EDTA plasma leptin measured by Procarta-5-plex-DropArray Luminex (CV 17.3%); total adiponectin by Abcam ELISA (CV 13.3%). Plate effects adjusted by median centering. MRI and AAT quantification: 333 neonates (≥34 weeks, BW >2000 g) had abdominal MRI within 2 weeks post-birth; adipokine data available for 271. Non-sedated neonates scanned (GE Signa HDxt 1.5T) from diaphragm to symphysis pubis with T1-weighted water-suppressed (WS) and non-WS axial sequences. AAT segmented into superficial subcutaneous (sSAT), deep subcutaneous (dSAT), and intra-abdominal (IAT) compartments using an in-house semi-automated MATLAB algorithm; analyses by trained physician and MR analyst. Monitoring included pulse and oxygen saturation with neonatologist present. Statistical analysis: Group comparisons by t-tests or chi-square; adipokines non-normal, compared with Mann–Whitney U. Multivariable linear regressions assessed determinants of cord adipokines (outcomes standardized) mutually adjusted for ethnicity, maternal education, tobacco exposure (cotinine-categorized), parity, maternal age, ppBMI categories, GWG categories, GDM status, gestational age at delivery, and infant sex. Associations of cord adipokines (standardized exposures) with birthweight, skinfolds, fat mass (predicted and PEA POD), and MRI AAT compartments (standardized outcomes) adjusted for ethnicity, ppBMI, GDM, gestational age, infant sex, birth length; MRI models additionally adjusted for age on MRI day. Multiple testing controlled by Benjamini–Hochberg FDR 0.05. Analyses performed in SPSS v23.

• Cord leptin and adiponectin distributions: Leptin 0.3–20.4 ng/mL; adiponectin 0.3–28.4 µg/mL; positive correlation r=0.103, p=0.003.
• Determinants of cord adipokines (multivariable, standardized β [95% CI]): • Ethnicity: vs Chinese, Malay higher leptin 0.229 (0.027, 0.431) and higher adiponectin 0.268 (0.050, 0.486); Indian higher leptin 0.628 (0.409, 0.847) and trend to higher adiponectin 0.213 (−0.023, 0.449). • GDM: higher leptin 0.418 (0.198, 0.637) and lower adiponectin −0.243 (−0.481, −0.005). • Maternal ppBMI: overweight (23.0–24.9 kg/m²) higher leptin 0.256 (0.009, 0.504); no significant association with adiponectin. • Gestational age: positive with leptin 0.180 (0.121, 0.238) and adiponectin 0.117 (0.054, 0.180) per week. • Sex: female higher leptin 0.379 (0.221, 0.536) and higher adiponectin 0.254 (0.084, 0.423).
• Associations with neonatal adiposity (per 1 SD higher cord adipokine; adjusted, FDR-corrected): • Leptin: higher birthweight 0.212 (approx. as reported), higher triceps 0.267 (0.186, 0.349), higher subscapular 0.319 (0.238, 0.400), higher predicted fat mass 0.281 (0.218, 0.343), higher PEA POD fat mass 0.378 (0.237, 0.519). MRI AAT: sSAT 0.258 (0.142, 0.374), dSAT 0.386 (0.254, 0.517), IAT 0.250 (0.118, 0.383). • Adiponectin: higher birthweight 0.110 (0.059, 0.162), higher triceps 0.158 (0.080, 0.235), higher subscapular 0.179 (0.102, 0.256), higher predicted fat mass 0.154 (0.094, 0.215), higher PEA POD fat mass 0.136 (0.014, 0.257). MRI AAT: sSAT 0.185 (0.096, 0.274), dSAT 0.173 (0.067, 0.278), IAT 0.092 (−0.011, 0.195; not significant).
• Effect modification by maternal BMI: Positive associations between cord leptin and neonatal adiposity across all maternal BMI categories. For adiponectin, significant positive associations with birthweight and neonatal adiposity observed only among neonates of obese mothers (BMI ≥25 kg/m²); interaction terms significant for adiponectin with predicted fat mass and sSAT (p<0.05).

The study demonstrates that fetal adipokine milieu at birth reflects maternal metabolic status and demographic factors and relates to neonatal abdominal fat distribution. Maternal GDM was associated with higher cord leptin and lower adiponectin, aligning with the hypothesis that maternal hyperglycemia and insulin resistance influence fetal adipokine production and adiposity. Ethnic and sex differences in cord adipokines parallel previously observed differences in neonatal AAT (e.g., higher dSAT among Indian and Malay infants), suggesting adipokines may contribute to ethnic variability in abdominal fat deposition. Cord leptin showed stronger associations than adiponectin with neonatal adiposity, especially with metabolically active dSAT and also IAT, indicating potential relevance for future metabolic risk. Unlike adult physiology where adiponectin is inversely related to adiposity, higher cord adiponectin associated with greater subcutaneous fat in neonates, possibly reflecting developmental roles in insulin sensitization and growth. The restriction of adiponectin–adiposity associations to offspring of obese mothers suggests maternal obesity may amplify adiponectin’s growth-promoting effects, perhaps via increased fetal insulin sensitivity, enhancing nutrient allocation and fat deposition. These findings address the research questions by identifying maternal and fetal determinants of cord adipokines and linking these biomarkers to MRI-quantified neonatal abdominal adiposity, underscoring their potential role in early-life programming of body fat distribution and metabolic risk.

In a multi-ethnic Asian birth cohort, cord blood leptin and adiponectin concentrations varied by ethnicity, sex, gestational age, maternal BMI, and GDM. Both adipokines were positively associated with neonatal abdominal adiposity measured by MRI, with leptin showing robust associations across subcutaneous and intra-abdominal compartments and adiponectin associating primarily with subcutaneous compartments and particularly among offspring of obese mothers. These results suggest fetal adipokines may participate in early programming of abdominal fat distribution. Future research should include longitudinal follow-up to determine whether cord adipokines predict trajectories of abdominal fat and metabolic outcomes in childhood, assessment of high molecular weight adiponectin, and mechanistic studies to elucidate pathways linking maternal metabolic status, fetal adipokines, and neonatal fat partitioning.

• MRI subset: Only a subset of eligible neonates (parental consent required) underwent MRI; differences between included and non-included participants may limit generalizability and introduce selection bias.
• Observational design: Residual confounding cannot be excluded despite multivariable adjustment.
• Measurement scope: Only total adiponectin was measured (not high molecular weight isoform), which may better reflect metabolic risk in children.
• Subset sizes: Smaller sample sizes for PEA POD and MRI analyses reduce precision in subgroup and interaction analyses.