Obesity among U.S. adolescents has reached epidemic proportions, particularly for African Americans and Latinos (1). In some socioeconomically disadvantaged communities, rates of overweight approach 50% among elementary school–aged children (2). Obesity-related chronic diseases previously found in adults increasingly affect youth. For example, nearly one-third of type 2 diabetes mellitus has onset during in the second decade of life, primarily in youth of color, usually in association with obesity (3,4). A 32-year follow-up study of Norwegian teens yielded death rates 80% to 100% higher for those with body mass index (BMI) at or above the 95th percentile at baseline (5). In the Bogalusa Heart Study, overweight was highly predictive of cardiovascular disease (CVD) risk factor clustering in children as young as 7 years (6) and thus had potential as an effective screening variable for CVD risk factors.

Recognized CVD risk factors in children include high blood pressure (BP), elevated serum cholesterol, and elevated glycated hemoglobin (HbA1c) (7). Early intervention during childhood can reduce target organ damage associated with uncontrolled hypertension and other risk factors for CVD (8). However, without government-funded pediatric screening for CVD risk factors, most elevated markers for risk go undetected. Because 30% of overweight children have elevated BP, body composition assessment could identify children at high risk for CVD. Pediatric standards of care include assessment of body composition; such assessment — for example, the Fitnessgram requirement in California (9) — increasingly is required of schools. The potential for using school-based body composition assessment to identify middle-school children at risk for hypertension is particularly important for low-income racial/ethnic minority children living in low-income urban areas who have poor access to medical care.

Despite growing recognition of the pervasiveness of obesity in youth of color, few intervention studies have demonstrated efficacy in these populations (10). Hinkle, documenting the paucity of this research, described efforts to address the gap through pilot projects in community-based organizations (11). Some findings show promise in regard to diet (i.e., decreasing dietary fat, increasing fruit and vegetable consumption), physical activity (i.e., increasing physical activity levels, decreasing sedentary behavior), and body composition (i.e., decreasing body weight, slowing weight gain, reducing body-fat percentage) (12-16), but best practices are difficult to identify because of limitations imposed by small or unrepresentative samples, poorly controlled studies, and reliance on self-reported outcomes.

We present baseline data from the University of California, Los Angeles (UCLA), Community Steps to Minority Youth Fitness (CSMYF) project, a physical education (PE) replacement intervention funded jointly by the National Heart, Lung, and Blood Institute; National Institute of Diabetes and Digestive and Kidney Diseases; National Cancer Institute; and National Institute of Child Health and Human Development to address the worsening obesity epidemic among socioeconomically disadvantaged racial/ethnic minorities. CSMYF involved a partnership of the Los Angeles Unified School District, the Division of Chronic Disease Prevention and Health Promotion in the Los Angeles County Department of Public Health, and researchers from the UCLA School of Public Health. Here we report only the analysis of the baseline data to identify emerging correlates of three CVD risk factors: elevated BP, elevated serum cholesterol, and elevated HbA1c among low-income racial/ethnic minority sixth-graders who live in an urban area.

Despite socioeconomic comparability, the two schools differed on several baseline measures (Table 1). Of the 1627 students enrolled at the intervention school, 1027 (63%) were African American, 590 (36%) were Latino, and 10 (1%) were from other racial/ethnic groups. At the control school, of 2108 students enrolled, 630 (30%) were African American, 1458 (69%) were Latino, and 20 (1%) were from other racial/ethnic groups (data not shown). Students at the intervention school had higher diastolic BP values (t ₁₈₄ = 3.6; P = .001) and higher HbA1c scores (t ₁₆₈ = 2.4; P = .005) than students at the control school.

Of the 199 sixth-graders in our study, 47.4% were overweight or at risk for overweight (Table 2). Boys and girls differed only slightly on any of the measured variables (Table 3). African American students had higher diastolic BP, after correction for differences in age and height (t ₁₆₈ = −2.04; P = .04); Latinos had higher average BMI percentile (t ₁₇₅ = 2.60; P = .01).

When we controlled for student age, measures were similar between the two schools, except for slower mile run/walk times at the intervention school (t ₁₅₂ = 6.96: P < .001) and higher diastolic BP adjusted for height and age at the intervention school (t ₁₇₆ = –2.36; P = .02).

The percentage of sixth-graders at risk for overweight ranged from 15.2% among African American boys to 41.5% among Latino boys (Table 2). Cross-tabulations of sex by weight status within racial/ethnic groups showed no sex difference among African Americans or among Latinos. Weight did not differ significantly by sex or by race/ethnicity.

Similarly, neither systolic BP-defined hypertension nor diastolic BP-defined hypertension differed significantly by student sex or race/ethnicity (Table 4).

We compared the means for cholesterol, systolic and diastolic BP, and HbA1c as a function of BMI-for-age risk status (desirable weight, at risk for overweight, overweight, and underweight). One-way unconditional analysis of variance indicated a significant association between BMI-for-age and serum cholesterol (F _3,155 = 4.92; P = .003), systolic BP (F _3,166 = 22.43; P < .001), diastolic BP (F _3,166 = 10.35; P < .001), and 1-mile run/walk time (F _3,143 = 7.29; P < .001). We found no relation between weight and HbA1c level. Serum cholesterol differed significantly between students of desirable weight and overweight; serum cholesterol in students at risk for overweight did not differ significantly from that of underweight students (Table 4 and Table 5). For systolic BP, students of desirable weight and at risk for overweight did not differ from one another, but both differed significantly from overweight students. Overall, the average CVD risk factor profile for overweight students was a nonfasting total cholesterol of 189 mg/dL, BP values of 121/78 mm Hg, and a mean HbA1c value of 4.7%.

To determine the impact of overweight on BP, controlling for other variables, we regressed systolic and diastolic BP on sex, racial/ethnic group, age, time to complete a 1-mile run/walk, and BMI-for-age risk status. Time to complete the mile, ranging from 7 minutes to 20 minutes 16 seconds, was a crude measure of physical fitness. We did not compute regression analyses for HbA1c because the bivariate analyses showed no association between BMI-for-age and HbA1c.

Students' times to run/walk 1 mile fell just short of statistical significance in predicting systolic hypertension (odds ratio [OR], 1.15; 95% confidence interval [CI], 0.99–1.33; P = .06), but overweight significantly predicted systolic hypertension (OR, 3.42; 95% CI, 1.5–7.8) (Table 6). Overweight students were nearly 3.5 times more likely than students of desirable weight to have systolic BPs in the hyptertensive range. Overweight also contributed to a higher diastolic BP (OR, 8.01; 95% CI, 3.1–20.6). Overweight students were eight times more likely than students of desirable weight to have diastolic BPs in the hyptertensive range. The influence of race/ethnicity on diastolic BP was not significant (OR, 2.56; 95% CI, 1.06–6.17), although Latinos were at greater risk than were African Americans (data not shown). If a student was categorized as hypertensive on the basis of either a systolic or diastolic BP, BMI-for-age was the major predictor (OR, 7.04; P < .001).

In contrast, multivariate analysis showed no association between BMI-for-age and elevated serum cholesterol (>200 mg/dL); none of the assessed variables independently predicted elevated serum cholesterol in our student sample.

Our baseline sample of racial/ethnic minority sixth-graders in two inner-city middle schools indicated that nearly one-third (29.9%) were at risk for overweight. Sex or racial/ethnic differences in the prevalence of overweight were minimal. These weight statistics are comparable to the 31% of adolescents observed in central Mexico who were recently assessed as being overweight (26). Even with 30% of middle-school students already overweight, these findings nonetheless may underestimate the extent of obesity in inner-city middle schools because more health-conscious students may be predisposed to volunteer for a school-based nutrition and physical activity program. Because we did not collect data from students who did not participate, we could not gauge whether selection bias affected the generalizability of the data.

Our serum cholesterol and BP findings revealed that overweight, low-income racial/ethnic minority children are at higher risk for CVD than were adolescents in the Bogalusa Heart Study (6). Overweight children aged 11 and 12 years in our study had a mean nonfasting total serum cholesterol of 189 mg/dL (Bogalusa median for 11- to 17-year-olds was 157 mg/dL) and a BP of 112/70 mm Hg (Bogalusa median for children aged 11 to 17 years was 107/66 mm Hg). The mean serum cholesterol value of overweight students was significantly higher than for desirable weight sixth-graders, and both systolic and diastolic BP levels were significantly higher for overweight students than for both desirable weight and at-risk-for-overweight students. Thus, overweight could function as a crude variable for screening youth at risk for high BP but was less useful in identifying those with elevated serum cholesterol. Physical fitness, as reflected by students' 1-mile run/walk time, was not successful in multivariate prediction of any of the CVD risk factors. Of students at risk for systolic hypertension (i.e., prehypertensive, hypertensive, or severely hypertensive), 51.1% also were overweight; for diastolic hypertension, 56.4% also were overweight.

A strength of our study is its focus on Latino and African American middle school children living in a low-income area of Los Angeles. Few previous studies have focused on low-income racial/ethnic minority children in the western United States. A limitation of the findings is the cross-sectional nature of the data, particularly BP. A clinical diagnosis of pediatric hypertension requires several BP assessments distributed over multiple visits (27). The consistent relation between overweight and elevated risk for pediatric hypertension, however, cannot be explained as an artifact of the one-time measurement (25).

Both our logistic regression and the literature support the superior utility of overweight and physical fitness as markers for elevated BP rather than for elevated serum cholesterol (27,28). The measure of physical fitness was imperfect because motivational and social factors are potentially important additional influences on the students' time to complete the 1-mile run/walk. Nonetheless, overweight independently predicted both systolic and diastolic BP and weakly predicted serum cholesterol in these analyses. The odds of an overweight minority youth being classified as hypertensive, either from excessive systolic BP or from excessive diastolic BP, was about eightfold higher than that of their peers of desirable weight. No other nondemographic variable was significantly associated with observed risk for high BP, including physical fitness as measured by the 1-mile run/walk.

Analyses of these baseline data highlight the need for obesity prevention and intervention programs targeted to racial/ethnic minority youth to reduce their current and future risk for hypertension. Bringing overall student nutrition (at school and at home) into compliance with the current Dietary Guidelines for Americans (29) should help reduce risk for hypertension, especially through adherence to dietary sodium standards for African Americans (30), and should contribute to decreased risk for obesity (31). For the nearly 50% of hypertensive children whose hypertension would not have been diagnosed by their overweight status, these results also suggest that direct measurement of BP is indicated. Even though weight status in minority children aged 11 and 12 years could be a marker for other elevated risk factors for CVD (particularly high BP), research is needed to determine whether intervention-mediated changes in weight can reduce risk for pediatric high BP and reduce future risk for high BP in adulthood. The importance of this question is underscored by recent data from the National Health and Nutrition Examination Survey (NHANES) showing a reversal at the end of the last decade in the downward trend of the prevalence of high BP (32). To increase the lifespans of children in this generation, we must identify effective strategies to reverse the high levels of weight-related risk factors for CVD documented here in low-income, racial/ethnic minority children.