This study verified the diagnostic accuracy of the nutritional status classified by the international height and BMI references of the World Health Organization (WHO) (WHO/2007), International Obesity Task Force (IOTF/2012) and MULT (2023). The data pool was composed by 22 737 subjects aged five to 16 years from the Santos and Porto Alegre surveys. A correlation matrix between the z-scores of the BMI references and the skinfold measurements was calculated through the Pearson correlation coefficient (r), and the subject’s nutritional status was classified according to the international growth references. The accuracy for diagnosing obesity was performed separately by sex and using the 95th percentile of the triceps and subscapular skinfold sum, while Lin’s concordance coefficient, Bland–Altman method and the Cohen’s Kappa coefficient (Kappa) were used to verify the concordance and reliability among the BMI references. The correlation matrix showed a high positive correlation among the BMI z-scores (r ≥ 0·99) and among the skinfold measurements (r ≥ 0·86). The prevalence of stunting was higher when applying the MULT reference (3·4 %) compared with the WHO reference (2·3 %). The Bland–Altman plots showed the lowest critical difference (CD) between the height references of WHO and MULT (CD = 0·22). Among the BMI references, the WHO obesity percentile presented lower performance than MULT for boys, presenting a lower +LR value (WHO = 6·99/MULT 18 years = 10·99; 19 years = 8·99; 20 years = 8·09) for the same −LR values (0·04). Therefore, MULT reference holds promise as a valuable tool for diagnosing childhood obesity, particularly when considering sex differences. This enhances its suitability for assessing the nutritional status of Brazilian schoolchildren.

Very few large studies in Latin America have evaluated the association between waist:height ratio (W-HtR) and cardiometabolic risk in children and adolescents. Further, multivariable analyses verifying the independence of located subcutaneous fat have not been conducted so far. The aim of this study was to evaluate the associations of W-HtR and waist circumference (WC) with metabolic syndrome abnormalities and high LDL-cholesterol levels in schooled adolescents before and after adjusting for trunk skinfolds and BMI. The sample consisted of 831 boys and 841 girls aged 10–17 years. Biochemical, blood pressure and anthropometrical variables were measured. Age- and sex-specific quartiles of W-HtR and WC were used in Poisson regression models to evaluate the associations. High WC values (highest quartile v. quartiles 1–3) were associated with high TAG levels in both sexes (prevalence ratio, boys: 2·57 (95 % CI 1·91, 3·44); girls: 1·92 (95 % CI 1·49, 2·47); P<0·05), and with high blood pressure specifically in female adolescents (3·07 (95 % CI 1·58, 5·98); P<0·05), independently of trunk skinfolds or BMI (P<0·05). Associations of high WC with high fasting glucose (boys), low HDL-cholesterol and having at least two abnormalities did not remain significant in most of the adjustments for trunk skinfolds or BMI (P>0·05). High W-HtR (highest quartile v. quartiles 1–3) was only independently associated with high TAG in female adolescents (1·99 (95 % CI 1·55, 2·56); P<0·05). In conclusion, WC showed better association with cardiometabolic risk than W-HtR in the children of this study. This observation does not support W-HtR as a relevant adiposity marker for cardiovascular and metabolic risk in adolescence.