It is increasingly recognised that the use of BMI cut-off points for diagnosing obesity (OB) and proxy measures for body fatness in a given population needs to take into account the potential impact of ethnicity on the BMI–fat % relationship in order to avoid adiposity status misclassification. This relationship was studied here in 377 Mauritian schoolchildren (200 boys and 177 girls, aged 7–13 years) belonging to the two main ethnic groups: Indian (South Asian descent) and Creole (African/Malagasy descent), with body composition assessed using an isotopic 2H dilution technique as reference. The results indicate that for the same BMI, Indians have more body fat (and less lean mass) than Creoles among both boys and girls: linear regression analysis revealed significantly higher body fat % by 4–5 units (P < 0·001) in Indians than in Creoles across a wide range of BMI (11·6–34·2 kg/m2) and body fat % (5–52 %). By applying Deurenberg’s Caucasian-based equation to predict body fat % from WHO-defined BMI thresholds for overweight (OW) and OB, and by recalculating the equivalent BMI values using a Mauritian-specific equation, it is shown that the WHO BMI cut-offs for OB and OW would need to be lowered by 4·6–5·9 units in Indian and 2·0–3·7 units in Creole children in the 7–13-year-old age group. These results have major implications for ethnic-based population research towards improving the early diagnosis of excess adiposity in this multi-ethnic population known to be at high risk for later development of type 2 diabetes and CVD.

BMI z (BMIz) score based on the Centers for Disease Control and Prevention growth charts is widely used, but it is inaccurate above the 97th percentile. We explored the performance of alternative metrics based on the absolute distance or % distance of a child’s BMI from the median BMI for sex and age. We used longitudinal data from 5628 children who were first examined <12 years to compare the tracking of three BMI metrics: distance from median, % distance from median and % distance from median on a log scale. We also explored the effects of adjusting these metrics for age differences in the distribution of BMI. The intraclass correlation coefficient (ICC) was used to compare tracking of the metrics. Metrics based on % distance (whether on the original or log scale) yielded higher ICCs compared with distance from median. The ICCs of the age-adjusted metrics were higher than that of the unadjusted metrics, particularly among children who were (1) overweight or had obesity, (2) younger and (3) followed for >3 years. The ICCs of the age-adjusted metrics were also higher compared with that of BMIz among children who were overweight or obese. Unlike BMIz, these alternative metrics do not have an upper limit and can be used for assessing BMI in all children, even those with very high BMIs. The age-adjusted % from median (on a log or linear scale) works well for all ages, while unadjusted % from median is better limited to older children or short follow-up periods.

Early adiposity rebound (EAR) predicts paediatric overweight/obesity, but current approaches do not consider both the starting point of EAR and the BMI trajectory. We compared the clinical characteristics at birth, age 3–5 and 6–8 years of children, according to the EAR and to its type (type A/type B-EAR). We assessed the childrenʼs odds of being classified as overweight/obese at age 6–8 years, according to the type of EAR as defined at age 3–5 years. As part of this two-wave observational study, 1055 children were recruited and examined at age 3–5 years. Antenatal and postnatal information was collected through interviews with parents, and weight and height from the health records. Type A and type B-EAR were defined in wave 1 according to the BMI nadir and the variation of BMI z-score between the starting point of the adiposity rebound and the last point on the curve. At 6–8 years (wave 2), 867 children were followed up; 426 (40·4 %) children demonstrated EAR. Among them, 172 had type A-EAR, higher rates of parental obesity (P < 0·05) and greater birth weight compared with other children (P < 0·001). Odds for overweight/obesity at 6–8 years, when adjusting for antenatal and postnatal factors, was 21·35 (95 % CI 10·94, 41·66) in type A-EAR children and not significant in type B-EAR children (OR 1·76; 95 % CI 0·84, 3·68) compared with children without EAR. Classification of EAR into two subtypes provides physicians with a reliable approach to identify children at risk for overweight/obesity before the age of 5 years.

Body weight is a major risk determinant of frailty, but the effect of obesity on frailty is controversial. The present study aimed to confirm the hypothesis that the risk of frailty is positively associated with obesity (BMI ≥ 30 kg/m2), but the association is mediated by the waist:height ratio (WHtR) in older women and men. A total of 2862 community-dwelling older individuals aged 70–84 years were assessed for frailty using the Korean version of Fatigue, Resistance, Ambulation, Illnesses, and Loss of weight index. Obesity (BMI ≥ 30 kg/m2) was associated with a higher risk of frailty compared with BMI 18·5–<23 kg/m2 in all the older individuals (OR 1·88; 95 % CI 1·11, 3·17; P = 0·018) and in older women (OR 1·86; 95 % CI 1·01, 3·42; P = 0·047) before adjusting for WHtR but was not associated with BMI after adjusting for WHtR. Additionally, obesity was not significantly associated with the risk of frailty before and after adjusting for WHtR in older men. Mediation analysis revealed that the association between BMI and frailty score was mediated by WHtR. Moreover, the mediating effect of WHtR on frailty score was positive in both women and men, but the frailty score was associated with BMI positively in women and negatively in men. The present study suggests that the risk of frailty is higher in obese women, which is mediated by WHtR, but not in obese men.