Studies have shown that children born small-for-gestational age (SGA) are at an increased risk of subsequently developing obesity and diabetes [1-3]. In vitro studies have shown a higher lipid content in adipocytes differentiated from mesenchymal stem cells obtained from the umbilical cord of the SGA neonates compared to appropriate-for-gestational age (AGA) infants [4]. However, validating this in children is difficult as assessment of adipocyte hypertrophy has traditionally necessitated adipose tissue biopsies. In earlier studies on rats and humans, we have shown that the hydrolipidic ratio (HLR) determined using magnetic resonance spectroscopy (MRS) is negatively associated with adipocyte hypertrophy [5]. In this work, we examined the size-for-gestational age specific differences in abdominal subcutaneous fat and HLR in Asian preschoolers.

The study population included 1208 subjects (681 Chinese, 307 Malays and 220 Indians) aged 4.5 years from the Growing Up in Singapore Towards healthy Outcomes (GUSTO) birth cohort [6]. Anthropometric measurements of these subjects were obtained at multiple time points from birth to age 4.5 years. Gestational age-specific birth weight percentiles were used to classify the children as SGA (<10th percentile) or AGA (>10th percentile). Abdominal images with 5 mm slice thickness and in-plane resolution of 0.94 × 0.94 mm were acquired using water suppressed HASTE sequence (TR=1000 ms, TE=95 ms) on a 3T MR scanner (Siemens Skyra) from a sub-cohort of 54 subjects at age 4.5 years. The subcutaneous adipose tissue (SAT) between the diaphragm and upper sacrum was segmented using an automated algorithm [7] and manually edited for correction of mis-segmented regions. The segmented SAT volumes were used to derive a predictive equation for SAT by automatic linear modeling using SPSS 23.0 software. Forward stepwise selection using the Akaike information criterion was used to select/remove predictors. Anthropometric skinfold measures, weight, height, gender and ethnicity were entered as predictor variables. SAT was predicted using the following equation: SAT (cm³) = 794.26 + 73.66 × weight (kg) + 43.62 × subscapular skinfold (mm) – 18.61 × height (cm) + 17.09 × triceps skinfold (mm) that had a predictive power of R² = 0.88.

Volume localized ¹H MR spectrum of the SAT was acquired from a 1 × 1 × 1 cm³ voxel placed in the posterior deep SAT using PRESS sequence (TE/TR = 30/2000 ms, 8 avg) in a sub-cohort of 211 subjects (98 Chinese, 72 Malays and 41 Indians). The peak resonances were fitted and quantified using in-house developed MATLAB program. The HLR (water/(water+fat)) was determined from the concentration of water and all the lipid peaks.

Table 1 shows the adiposity measures of SGA and AGA subjects. The SGA children had a significantly lower mean birth weight (2532 g vs. 3114 g) than AGA children. Even after a faster postnatal weight gain compared to AGA during the first year (Fig. 1), the abdominal SAT did not catch-up with that of AGA subjects at age 4.5 years, after adjusting for breastfeeding duration and gender. While the SAT volumes were lower in the SGA group, the SAT HLR was not significantly different between the groups. This suggests that SGA children had a comparable degree of subcutaneous adipocyte hypertrophy even at lower levels of adiposity. Longitudinal measurements at later time points can provide a more complete picture of the changes in abdominal fat volumes and hypertrophy between the two groups.