American Indian adults are not generally sampled in large enough numbers for valid statistical inference and comparison in national studies because of their small numbers in the population. American Indian children are also not adequately sampled in national studies. Many American Indians live in Oklahoma communities with other racial groups, allowing for comparisons that hold the geographic macroenvironment constant while possibly minimizing, though not eliminating, other potential confounders, such as socioeconomic status and regional cultural differences.

Many American children live in an obesogenic environment, regardless of whether they live in subsidized urban housing, suburban residences, or small farms in rural communities. The availability of fast food, large portion sizes, energy-dense food, labor-saving devices, screen time, and automobiles and the increased frequency of eating make survival convenient and sedentary and obesity endemic (1-3). Although the environmental risk factors for obesity are becoming endemic worldwide, disparities in risk associated with culture, geography, and affluence can also affect risk. Some of the risk factors of the built environment, such as a high density of fast-food restaurants and a lack of facilities or pleasant places to exercise, have been shown to be positively associated with the risk for obesity (4,5). While many schools, communities, and health departments are trying to alter this situation, the prevailing influence is to move less and eat more.

Reports from the National Health and Nutrition Examination Survey (NHANES) have indicated a possible leveling of the prevalence of overweight or obesity among children. However, evidence also suggests that the proportion of children who are very obese (≥97th percentile of the Centers for Disease Control and Prevention [CDC] 2000 Growth Charts) is increasing, particularly among African American and Hispanic children (6,7). American Indian children, like their racial/ethnic minority and nonminority counterparts, are also at risk (8-10). Environmental pressures, poverty and unemployment, and susceptibility to metabolic disease do not bode well for their future health (11). Rates of type 2 diabetes among American Indian adults are already among the highest for population groups in the United States (12,13).

The objective of this study was to assess and compare the weight status of a sample of American Indian and white students residing in the same rural environment and attending 3 elementary schools in Anadarko, Oklahoma, to ascertain differences by race, sex, or age.

We measured a similar number (average, 753; range, 706 to 795) of students each year (Table 1). Response rates averaged 93%. The mean age at measurement was 8.3 years (range, 5.1-12.4 y). We found no sex or race differences between children who participated in the study and the overall school population. Between 75% and 80% of participants were American Indian, and 21% to 25% were white. Approximately half of the participants were girls. Averaged over 5 years, 45.4% of American Indian students and 65.1% of white students were healthy weight or underweight. Few children (either American Indian or white) were underweight. American Indian students had a higher prevalence of being overweight, obese, or very obese than white students for each year of analysis (Table 1 and Table 2).

The percentage of American Indian children each year who were very obese (range, 25.5% to 28.2%) was greater than the percentage who were overweight (range, 17.6% to 22.3%) or obese (range, 6.1% to 8.7%). During the 5-year study period, 24.2% (weighted average) of all students were very obese (26.9% American Indian, 14.7% white).

When we controlled for sex and age and compared very obese children with children who had a healthy weight or were underweight, American Indian children had a significantly higher risk (RR weighted average, 2.0) for being very obese than white participants in each year (Table 3). When we controlled for race and age, sex was not significantly associated with risk for being very obese in any year. Increasing age was significantly associated with the risk for being very obese in 4 of 5 years.

When we controlled for sex and age and compared obese children with children who had a healthy weight or were underweight, American Indian children had a higher risk (RR weighted average, 1.6) for being obese than white children in each year. However, this risk was significantly higher only in 2005 (Table 3). When we controlled for race and age, sex was not significantly associated with risk for being very obese in any year. Increasing age was significantly associated with the risk for being obese only in 2008.

When we controlled for sex and age and compared overweight students with children who had a healthy weight or were underweight, American Indian children had a significantly higher risk (RR weighted average, 1.8) for being overweight than white students in each year (Table 3). When we controlled for race and age, boys had a significantly higher risk for being overweight than girls only in 2006. Age was not significantly associated with the risk for being overweight in any year.

Averaged over 5 years and for each year studied, American Indian elementary school students within 1 school district in a rural Oklahoma community had an increased risk of being overweight, obese, or very obese compared with their white peers; 24.2% (weighted average) of all students were very obese (26.9% American Indian, 14.7% white).

In 1 year, boys had a higher risk of being overweight than girls, and older age was sometimes significantly associated with an increased risk of being overweight or heavier. Higher percentages of American Indian children were very obese than were obese or overweight. Although these analyses were cross-sectional, they show that obesity has become an endemic condition for children in this community, and American Indian children are at greatest risk.

The increased prevalence of obesity among children and its effects in both childhood and adulthood have been documented (17-21). One study reported that children who have high levels of blood glucose were more likely to have a high BMI z score in childhood and adulthood and continue to have elevated levels of blood glucose (22). Another study found that BMI z score was significantly related to indicators of poor health, including higher levels of cholesterol, insulin, blood pressure, and liver enzymes (23). Among children in the Anadarko school district, 1 study showed that higher BMI is associated with higher risk for elevated blood pressure (24). Children in the highest quartile of BMI have also been shown to have an increased risk for premature death (25).

Using NHANES data from 2007-2008, Ogden et al reported the weight status of children aged 6 to 11 (6). Using their categories (BMI ≥85th percentile, BMI ≥95th percentile, and BMI ≥97th percentile), we compared their data with ours for 2007. The percentage of children in our study population was higher than the percentage in the NHANES population for the overweight or greater BMI category. In our study population, 51.7% of the children had a BMI equal to or greater than the 85th percentile, compared with 35.5% of the NHANES children; 32.0% of our children had a BMI equal to or greater than the 95th percentile, compared with 19.6% of the NHANES children; and 23.7% of our children had a BMI equal to or greater than the 97th percentile, compared with 14.5% of the NHANES children. These data show that the prevalence of obesity among our study population is greater than the national prevalence, and it is skewed toward the most extreme category of obesity.

Our study had several limitations. First, our data are cross-sectional. The data provide a snapshot of the weight status of children at 3 elementary schools for each of 5 consecutive years and therefore, do not indicate changes for each child over time. Second, we did not collect data for each student on other possible risk factors for obesity, such as poverty (eg, eligibility for free or reduced-price school lunch). But the aggregate difference in eligibility among our study population (79.4% of American Indian students and 47.5% of white students) as a surrogate for poverty may account for some of the obesity risk among the American Indian students. Data on other potential risk factors or confounders for obesity, such as neighborhood and proximity to exercise venues, were not available.

The strengths of our study were a high response rate, the collection of data over 5 years, and the ability to compare — by race — children living and attending schools in the same school district (ie, the same rural environment). These 3 factors helped eliminate the possibility that regional culture or school environment played a role in creating disparities in obesity among the study population. Our study demonstrated a consistently high percentage of children who were overweight, obese, or very obese, and American Indian children were shown to be at greater risk for obesity than white children.

Additional studies into the greater risk among American Indian children are warranted, and such studies should collect data on variables that can rule out socioeconomic and other environmental causes of obesity. This study makes clear the urgent need for interventions to improve nutrition, increase physical activity, and minimize sedentary behavior in schools and communities.