Cardiovascular Risk Factors Associated With Insulin Resistance in Children Persist Into Late Adolescence

RESEARCH DESIGN AND METHODS

The study cohort was the San Diego Study of Children’s Activity and Nutrition, an ongoing observational study that began when the subjects were 4 years of age (with follow-up assessments at intervals of ∼5–7 years) and consisted of non-Hispanic white and Mexican-American children recruited through state-funded daycare programs. Detailed methods and health behavior results have been previously reported (1,5–11). The current study sample includes 83 adolescents: 29 non-Hispanic white (17 males and 12 females) and 54 Mexican-American (25 males and 29 females) subjects who had completed study visits at age 11 and again at age 17 years. Based on demographic and laboratory data, there was no evidence of a bias in the drop out of subjects between study visits. All blood samples were obtained in the fasting state.

Anthropometric and blood pressure measures

Height, weight, waist circumference, and skinfold measurements were obtained using standardized approaches as previously described (1,12). At age 11 years, blood pressures were measured three times at 1-min intervals using a random zero sphygmomanometer. At age 17 years, five blood pressure measurements were taken using the Dinamap automatic recording instrument.

Laboratory analyses

Lipids, lipoproteins, apolipoprotein B-100 (ApoB), and insulin levels were measured as previously described (1). The LDL cholesterol–to–ApoB ratio was used as a marker of LDL size (13,14). Serum highly specific CRP was measured with a chemiluminescent assay (Diagnostics Products).

Statistical analyses

Variable comparisons were made using the Mann-Whitney U test. An IRS score was calculated by summing the rank score for each individual for the variables insulin, systolic blood pressure, triglycerides, HDL cholesterol, and LDL cholesterol–to–ApoB ratio as previously described (1). Stepwise regression analyses were used to model the predictors of IRS score at age 17 years, using measures from age 11 years.

RESULTS

Similar to our previous findings, Mexican Americans were more obese than non-Hispanic whites at age 11 years and remained more obese at age 17 years, as evidenced by higher BMI and skinfold measures (Table 1). Whereas blood pressure did not differ between the groups at age 11 years, Mexican Americans had significantly higher diastolic blood pressure at age 17 years. The absence of the typical modest age-related increase in diastolic blood pressure is presumably a consequence of the different methodology used for blood pressure measurement between visits (15). At age 17 years, the fasting insulin levels, although still higher in Mexican Americans, were no longer different between the groups, yet there was a trend for higher triglycerides and ApoB levels in the Mexican-American adolescents (P < 0.08), and they had significantly higher CRP levels. IRS scores trended higher in the Mexican-American adolescents (P = 0.07), and mean values were quite comparable within each ethnic group at both ages.

The relationship between insulin resistance and risk factors for cardiovascular disease was examined at age 17 years in the entire cohort by comparing the differences in levels of various metabolic risk factors across quartiles of fasting insulin levels. In general, the strong relationships between quartiles of fasting insulin and cardiovascular risk factors (HDL, triglycerides, and blood pressure) observed in this group at age 11 years (1) were also observed at age 17 years. Of note, mean CRP also increased across quartiles of insulin (0.9, 1.0, 1.9, and 2.3 mg/l, P < 0.05).

We then examined whether individual cardiovascular risk factors measured at age 11 years would predict IRS scores at age 17 years. All three measures of obesity (BMI, skinfolds, and waist circumference; r ≥ 0.57, P < 0.001), systolic and diastolic blood pressures, triglycerides, fasting insulin, total–to–HDL cholesterol ratio, and ApoB levels at age 11 years were all significantly correlated (r ≥ 0.27, P < 0.05) with IRS scores at age 17 years. In addition, HDL cholesterol (r = −0.46, P < 0.001) and the LDL cholesterol–to–ApoB ratio (r = −0.30, P < 0.01) at age 11 years were negatively correlated with IRS scores at age 17 years. In a stepwise regression analysis, including BMI, HDL cholesterol, triglycerides, insulin, ApoB, LDL cholesterol–to–ApoB ratio, and systolic and diastolic blood pressure, the significant predictors were BMI, triglyceride levels (log), and diastolic blood pressure, accounting for 46% of the variance in IRS scores at age 17 years (P < 0.001). Using waist circumference instead of BMI gave essentially identical results.

CONCLUSIONS

Only the Bogalusa, Muscatine, and Cardiovascular Risk in Young Finns Longitudinal Studies have tracked cardiovascular and metabolic risk factors from childhood to adolescence and/or adulthood, and none have included a Mexican-American cohort (16–18). Our data show that parameters frequently associated with risk for the IRS, such as obesity, lipid abnormalities, and fasting insulin, as well as the IRS score itself, track well from pre- to postpuberty in a community sample of both Mexican Americans and non-Hispanic whites. This is consistent with data from the prior longitudinal studies that have reported that BMI (19), insulin (20), serum lipid and lipoprotein (21–24), and blood pressure (25) levels track longitudinally in children and young adults. An additional finding of this study is that measurements of BMI or waist circumference, triglycerides, and diastolic blood pressure in childhood are relatively strong predictors of the IRS in late adolescence and may help identify those children who are most at risk for the development of diabetes and cardiovascular disease and most likely to benefit from intervention efforts.

As with many of the risk factors presented (Table 1), levels of CRP were higher in Mexican-American adolescents. This is partly explained by the fact that the Mexican-American adolescents were more obese, a known determinant of CRP (26,27). However, ethnicity may be a determinant of CRP in adolescents even after adjustment for BMI (28). CRP levels also increased in the entire cohort with higher insulin levels, suggesting that CRP and heightened levels of inflammation may be associated with the IRS even at this early age. Notably, half of the adolescent Mexican-American group in this study is already in the middle- to high-risk range, as defined by the Centers for Disease Control and Prevention (29). Although the clinical implications of elevated CRP in childhood have not been well studied, at least one study found CRP an independent predictor of both brachial artery reactivity and carotid artery intimal thickness in 11-year-old subjects (30). Moreover, there is increasing appreciation for the early role of inflammation in the development of both atherosclerosis and diabetes (2–4).