Deposition of intramyocellular lipids (IMCL) has been reported in childhood obesity, and is positively associated with fasting insulin, TG/HDL ratio and insulin resistance [1]. Most of the assessments of IMCL in children have been limited to school age children [2]. The goal of this study was to understand the contribution of ethnicity, gender, obesity and early developmental factors on IMCL determined by ¹H magnetic resonance spectroscopy (MRS) in Asian preschoolers.Parental consent for MR imaging was obtained for 242 children (age: 4.5 years) who were a part of the ‘Growing up In Singapore Towards Healthy Outcomes’ (GUSTO) birth cohort (116 males and 126 females, across three ethnic groups: 122 Chinese, 78 Malays, & 42 Indians). Gestational age-specific birth weight percentiles were used to classify children as small-for-gestational age (SGA) (<10th percentile) or appropriate-for-gestational age (AGA) (>= 10th percentile). Postnatal weight gain during the first two years was quantified as the difference between the weight-for-age z-score at 24 months and the gestational age-specific birth weight z-score. All scans were performed without sedation on a Siemens Magnetom Skyra 3T scanner. Following T₁-weighted axial localization, a point-resolved spectroscopy (PRESS) scan was performed on a 1×1×1 cm³ voxel placed in the soleus muscle with TR=2000ms, TE=33ms and 24 averages. Spectra were quantified using LCModel. The IMCL peak was normalized by the water peak from a water unsuppressed scan. T₂ correction of the IMCL and water peaks was performed using T₂ values reported in literature [3].

Differences in maternal BMI and fasting glucose, birth weight, postnatal weight gain, BMI-for-age, and IMCL across the ethnic groups and between the size-for-gestational age categories are shown in Table 1 & 2. IMCL at 4 years differed significantly between the three ethnic groups with Indians having the highest IMCL (Fig. 1). IMCL was higher in SGA children than in AGA children (Fig. 2).

A univariate general linear model was constructed with IMCL as the dependent variable, size-for-gestational age, ethnicity, gender and duration of any breastfeeding as fixed factors, and BMI-for-age z-score, postnatal weight gain, maternal BMI and maternal fasting glucose during the 26th week of pregnancy as covariates. In the fully adjusted model (Table 3), size-for-gestational age, ethnicity, BMI-for-age z-score and the interaction term between size-for-gestational age and ethnicity were statistically significant (P<0.05) predictors, while gender, duration of any breastfeeding, maternal BMI and fasting glucose and catch-up growth were not significant.

Prior works on IMCL in children and pre-adolescents have found IMCL to be positively associated with childhood obesity and maternal hyperglycemia during pregnancy due to gestational diabetes or type 2 diabetes [1]. In this work, we show that while children with higher BMI-for-age have higher IMCL, children born SGA also have higher IMCL at 4.5 years when compared to AGA children, despite being leaner, and despite their mothers having a lower BMI and fasting glucose during pregnancy. We also found higher IMCL in 4.5 year-old Indian children than in Chinese and Malay children, which recapitulates findings of an earlier study done in adults [4]. Our results suggest that Indians may have an inherent metabolic defect that predisposes them to increased IMCL accumulation from an early age. We found postnatal weight gain to be high in both Indian and children born SGA. However, postnatal weight gain was not significantly associated with IMCL in the fully adjusted model, suggesting that the trigger for IMCL deposition may predate accelerated weight gain.