Standard definition of child overweight and obesity worldwide

Editor—Different standards have been proposed to define overweight and obesity in childhood, and as a result it is difficult to compare published results. An international standard is useful for making international comparisons and monitoring the global epidemic of obesity. The new standard proposed by Cole et al (the International Obesity Task Force standard) meets this demand.1 By contrast, the World Health Organization recommended using the sex-age specific body mass index 85th centiles derived from data from the United States' first national health and nutrition examination survey collected in 1971-4 to define adolescent overweight for international use.2,3 Using data from large nationwide surveys from China (1991 China health and nutrition survey), Russia (1992 Russian longitudinal monitoring survey), and the United States (third national health and nutrition examination survey, 1988-94), we compared Cole et al's and the World Health Organization standards to assess overweight in children (aged 6-9) and adolescents (10-18). These three countries' total populations account for about a quarter of the global population. We examined the prevalence of overweight in each country by using each standard. The table shows that the two standards produced consistent estimates of the prevalence of overweight, although that proposed by Cole et al (the IOTF standard) produced a slightly lower estimate in children than did the WHO standard but a slightly higher estimate in adolescents. In general, the agreement was better in all adolescents, girls aged 6-9, and male adolescents. We calculated the κ coefficient to examine the agreement at an individual level. κ⩾0.8 suggests an excellent agreement.4 κ values also suggested an excellent agreement between these two standards, although the agreement varied by sample, age group, and sex. Results from studies in older children and adolescents using these two standards would seem to be comparable, although the standard proposed by Cole et al is better for international use. We found, however, that the prevalence of overweight is much lower in adolescents than children in the samples from the surveys in China and Russia, but the prevalence was similar in American children and adolescents. Further research is needed to understand whether our findings are mainly due to socioeconomic and behavioural factors; alternatively, both standards may have a higher sensitivity for indentifying individuals as overweight among children than adolescents in developing countries, where children and adolescents' growth, development, and maturation patterns are different from those of the reference populations.