• Body composition;
  • children;
  • metabolic status;
  • transracial


What is already known about this subject
  • South Asian children at birth are thinner, but more adipose and more resistant to insulin than White Caucasian children.
  • South Asian adults are more adipose and more insulin resistant, but their greater adiposity does not fully explain the difference in insulin resistance.
  • South Asian children at 8 y are more insulin resistant than White Caucasian children.
What this study adds
  • The BMI of South Asian children at 6 y is distributed normally, while that of White Caucasian children is heavily skewed.
  • South Asian children at 6 y are slimmer, and the boys, but not the girls, are more adipose.
  • South Asian boys, but not girls are more insulin resistant. Both genders metabolically less healthy, but their adiposity explains only part of the difference.


The concept of the ‘thin–fat’ Indian baby is well established, but there is little comparative data in older children, and none that examines the metabolic correlates. Accordingly, we investigated the impact of body composition on the metabolic profiles of Asian Indian and white UK children.


Body mass index (BMI), waist circumference, sum of four skin-folds, % body fat (by dual-energy X-ray absorptiometry), glucose, insulin, insulin resistance (Homeostasis Model Assessment), trigylcerides, cholesterol [total, low-density lipoprotein, high-density lipoprotein {HDL}, total/HDL ratio] and blood pressure (systolic, diastolic and mean arterial) were measured in 262 white Caucasian children from Plymouth, UK (aged 6.9 ± 0.2 years, 57% male), and 626 Indian children from rural villages around Pune, India (aged 6.2 ± 0.1 years, 53% male).


Indian children had a significantly lower BMI (boys: −2.1 kg m−2, girls: −3.2 kg m−2, both P < 0.001), waist circumference (P < 0.001) and skin-fold thickness (P < 0.001) than white UK children, yet their % body fat was higher (boys +4.5%, P < 0.001, girls: +0.5%, P = 0.61). Independently of the differences in age and % body fat, the Indian children had higher fasting glucose (boys +0.52 mmol L−1, girls +0.39 mmol L−1, both P < 0.001), higher insulin (boys +1.69, girls +1.87 mU L−1, both P < 0.01) and were more insulin resistant (boys +0.25, girls +0.28 HOMA-IR units, both P < 0.001).


The ‘thin–fat’ phenotype observed in Indian babies is also apparent in pre-pubertal Indian children who have greater adiposity than white UK children despite significantly lower BMIs. Indian children are more insulin resistant than white UK children, even after adjustment for adiposity.