Definitions of pediatric obesity

We agree with Geoff Ball and Noreen Willows that Canadian reference data are needed to accurately estimate the current prevalence of overweight and obesity. Unfortunately, standardized reference data based on measured heights and weights of Canadian children are not available. We calculated prevalence in our study population1 using the methods of both Cole and colleagues2 (the International Obesity Task Force) and the CDC.3 To make the data more amenable for international comparisons, an editorial decision was made to report the prevalence estimates only in terms of the first method. Fig. 1 of this letter compares the prevalence rates as estimated by the 2 methods. Six points are noteworthy.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1: Prevalence of overweight (squares) and obesity (triangles) among preschool children in Newfoundland and Labrador, by sex and age group, estimated by the methods of Cole and colleagues² (the International Obesity Task Force) (A) and the Centers for Disease Control and Prevention³ (B). Error bars represent 95% confidence intervals.

First, the prevalence estimates for overweight (termed “at risk for overweight” by the CDC) and obesity (termed “overweight” by the CDC) were high regardless of which method was used. Also, the analysis indicated no significant difference between boys and girls in the prevalence of overweight or obesity as estimated by either method. Third, there were no differences between age groups with either method. Fourth, the CDC method yielded significantly higher rates of overweight and obesity (combined), although this difference appears to be accounted for more by the difference in estimates of obesity (18.0% v. 8.0%) than by the difference in estimates of overweight (18.8% v. 17.6%). Further analysis indicated that a child was more than twice as likely to be classified as obese by the CDC method than by the method of Cole and colleagues. Finally, comparisons between the 2 methods with categorical modelling procedures indicated that they classified boys and girls differently. Boys were 2.85 times more likely to be classified as obese by the CDC method than by the method of Cole and colleagues, whereas girls were 2.20 more likely to be classified as obese by the CDC method.

Clearly, care must be taken in making comparisons, particularly of rates of obesity, when different methods of classification have been used.