The Use of the New World Health Organization Growth Standards To Identify Trends and Determinants of Overweight in WIC Infants and Children

Year: 2008

Research Center: Department of Nutrition at the University of California, Davis

Investigator: Maalouf, Zeina, and Kathryn G. Dewey

Institution: University of California, Davis

Project Contact:
Zeina Maalouf
Department of Nutrition
University of California, Davis
One Shields Avenue
Davis, CA 95616
Phone: 530-752-4622


Childhood obesity is a worldwide epidemic. In the United States, 24.4 percent of children are overweight or at risk of overweight. Tools for the adequate diagnosis, treatment, and prevention of childhood obesity are needed. The Special Supplemental Nutrition Program for Women, Infants, and Children (WIC), reaching half of all U.S. pregnant women and their infants and providing followup care until the child is 5 years of age, provides an ideal avenue to effectively address the problem of childhood obesity.

A child’s obesity status is determined by calculating his/her body mass index (BMI) (weight (kilograms)/height2 (meters2)) and comparing it to an age- and sex-specific reference distribution. If a child’s BMI is above the 95th percentile of that distribution, the child is considered overweight. If the BMI is between the 85th and the 95th percentiles, the child is considered at risk of overweight.

Currently in the United States, the reference used for children’s BMI is the set of charts known as the Centers for Disease Control and Prevention (CDC) growth charts. These charts were constructed using nationally representative data collected from 1963 to 1994. Most of the children whose anthropometric measurements contributed to the construction of the charts were formula-fed as infants, which is not in line with feeding recommendations for optimal growth. Furthermore, the measurements were not done with sufficient frequency to capture the rapid changes in growth that occur in the first year of life. Because of these and other technical deficiencies in the charts, the World Health Organization (WHO) issued new growth charts in 2006. These charts were constructed with data from children who were provided with conditions that allowed for optimal growth (that is, optimal nutrition, environment, and healthcare). There was frequent followup of the children to collect growth data, and appropriate statistical techniques were used to construct the WHO standards from these data.

The objectives of this study were to (1) use data from Massachusetts WIC participants to determine if, and at what ages, the WHO growth charts identify a larger number of overweight children compared with the CDC children and (2) examine the risk factors associated with child overweight when growth status is assessed using the new WHO standards.

The data were collected from participants in the Massachusetts WIC program from September 2001 to October 2006. The anthropometric measurements were done every 6 months during the participants’ certification visits. The measurements are reported to the CDC as part of the State’s participation in the Pregnancy and Pediatric Nutrition Surveillance Systems (PNSS and PedNSS), from which the data were extracted. Information on sociodemographic characteristics and breastfeeding was also extracted. The age- and sex-specific BMI percentiles of children were based on their directly measured height and weight and the two growth charts (WHO and CDC). Data analysis was done using Statistical Analysis Software (SAS) version 9.1.3. The data were cleaned to exclude implausible values, according to PedNSS and PNSS edits. Consistency between a child’s data for consecutive visits was also checked. All of the data were used for descriptive statistics. To examine prevalence estimates and risk factors for overweight, a dataset with one randomly chosen visit per child was used.

The study sample included 143,787 children who collectively had 392,927 WIC visits between the ages of 2 and 5. The mothers had a mean age of 26. About 35 percent of the mothers had not completed high school, 46 percent had a high school diploma, and the remainder had at least some college training. Pre-pregnancy, 5 percent were underweight, 25 percent were overweight, and 22 percent were obese. About half (51 percent) of the children were males. Nine percent of the children had a low birthweight, and another 9 percent had a high birthweight. The racial/ethnic distribution was as follows: 43 percent of the children were Caucasian, 19 percent were African-American, 32 percent were Hispanic, and 6 percent were Asian/Hawaiian/Pacific Islanders. Of those with breastfeeding information (about half of the sample), 65 percent of the children were breastfed at some point, with an average duration of any breastfeeding of 17 weeks.

When the CDC charts were used, 44 percent of the children were classified as overweight or at risk of overweight at least once during the 3 years of followup in WIC, increasing to 53 percent when the WHO charts were used. This and all subsequent comparisons were statistically significant with p-values less than 0.0001. The prevalence of overweight at any given time point was 17 percent when using the CDC cutoffs and 24 percent when using the WHO cutoffs. The prevalence of overweight or at risk of overweight when using the CDC charts was 34 percent and 42 percent when using the WHO charts.

The difference between the prevalence estimates in using the two charts was larger when the children were younger. The estimate for the percentage of overweight children in the group that was 24-27 months of age doubled from 12.6 percent when using the CDC cutoffs to 25.3 percent when using the WHO cutoffs. The same was true when looking at the estimate of overweight or at risk of overweight: In the group that was 24-27 months of age, the estimate increases from 28 percent when using the CDC charts to 45 percent when using the WHO charts. On average, using the WHO charts detects overweight 1.3 months earlier than when using the CDC charts. In a logistic regression model controlling for birthweight and child age, significant predictors of child overweight (defined using the WHO cutoffs) include maternal pre-pregnancy weight (obese versus normal OR=1.81 (1.72-1.91)), Hispanic ethnicity (OR=1.50 (1.43-1.56, ref=non-Hispanic), maternal education (some college versus less than high school graduate (HS), OR=0.80, 0.75-0.85, HS versus less than HS OR=0.90, 0.86-0.94), and child sex (OR=0.89, 0.85-0.93, ref=male). Using the WHO charts will result in a higher estimated prevalence of childhood overweight and allow for earlier diagnosis, thus making earlier intervention possible.

Direct inquiries about this study to the Project Contact listed above.