Intended for healthcare professionals

Rapid response to:


Establishing a standard definition for child overweight and obesity worldwide: international survey

BMJ 2000; 320 doi: (Published 06 May 2000) Cite this as: BMJ 2000;320:1240

Rapid Response:

BMI centile charts

Is the proposed standard definition for child overweight and obesity
internationally acceptable?

The issue of internationally acceptable definitions of overweight and
obesity is discussed by Cole et. al. (1). The article is a much welcome
contribution to the ongoing debate. However, it also raises some

1) Why did the authors use for constructing the average percentile
curves based on 6 data sets only two European studies (Great Britain and
the Netherlands)? What selection criteria were used that disqualified
other existing data sets?

2) The assumption that in a population of 18-year olds, there are 10
per cent of individuals with BMI values above 25 and 3 per cent with BMI
values above 30 can be also disputed. If this were the case, we would have
to conclude that in the European population of 18 year old individuals
only 0.9 per cent of overweight and 0.3 per cent of obese can be found,
whereas in populations of the United States and Hong Kong these values
would be 3.3 and 3.1 per cent, respectively.

The results of the 5th Nationwide Anthropological Survey of children
and adolescents from 0-18-years of age, carried out in the Czech Republic
in 1991, seem to strongly indicate that the proposed definitions are too
tolerant for the European population. The survey collected information on
basic anthropometric characteristics of 86 846 individuals from 0 - 18
years old (2, 3). Empirical percentiles was smoothed (using moving
average) and resulting curves were plotted. Using the cut off points
suggested by Cole (1), the percentage of Czech children who would fall
into a category of overweight or obese was calculated.

Results clearly shows that values of the 90th percentile do not vary
significantly from the values proposed by Cole. However, values for the
97th percentile are significantly lower for the Czech population. While
the values of the 97th percentile reach 27.7 in boys and 28.2 in girls,
respectively, the value suggested by Cole, to ensure a smooth link with
values for the adult population, is BMI 30. The above mentioned results
and comparisons imply that the proposed cut off points underestimate
considerably the actual prevalence of obesity. Instead of the expected 3
per cent prevalence of obesity, we find only 1 per cent of obese

In conclusion, we believe that a starting point for developing values
for the 90th and 97th percentile in 18-year olds should be based on
analysis of growth studies from more that two European countries.

This research is supported by Grant Agency of MofH CR, grant no. 5158-3


1 Cole TJ, Belizzi MC, Flegal KM, Dietz WH: Establishing a standard
definition for child overweight and obesity worldwide: international
survey. BMJ 2000, 320:1240-3.

2 Vignerová J., Bláha P.: The Growth of the Czech Child during the past 40
Years. In: Bodzár, B.É., and Suzanne, C. (Eds) Secular Growth Changes in
Europe, Eotwos Univ. Press, Budapest, 1998. pp. 93-107.

3 Vignerová J., Lhotská L., Bláha P., Roth Z.: Growth of Czech Child
Population 0-18 Years Compared to the World Health Organisation Growth
Reference, American Journal of Human Biology, 9:459-468, 1997.

Competing interests: No competing interests

07 October 2000
J Vignerova
National Institute of Public Health, Prague, Czech Republic