Many investigators have examined influence of race and gender on insulin resistance and other aspects of cardiovascular health, however, our study is unique in the attempt to examine whether these factors affect the degree to which cardiovascular fitness influences metabolic health. This cross-sectional study of healthy middle school children suggests that the metabolic benefits of fitness are influenced by both gender and race. Differences in the effect of fitness on inflammatory status and insulin sensitivity could provide insight into physiological characteristics that influence risk for metabolic disease in response to caloric excess and poor physical fitness. During times of nutrient scarcity, pro- and anti-inflammatory adipokines circulate in a balance that promotes energy efficiency to varying degrees. However, these same adipokines become detrimental when sedentary lifestyle leads to low fitness and accumulation of adipose tissue. People who evolved to be particularly adapted to expending high levels of energy in the setting of low caloric availability could, therefore, be at higher risk of inflammation and insulin resistance when activity levels are low and calories plentiful.
This study found that inflammation, measured by hsCRP, correlated strongly and inversely with CVF. This relationship was not confined to obese children; among non-overweight boys (BMI < 85th%tile); more than 50% of those with low fitness had an hsCRP level in the elevated range. The effect of higher CVF on reducing inflammation was highly significant in males and females in all racial groups; however this relationship was especially strong in the RD group and in male children. In both of these sub-groups the strong effect of fitness on inflammation reflected primarily low inflammation in fit subjects rather than elevated inflammation in children with low fitness.
This study also reports the new finding that gender and race influence the predictors of IR in middle school-aged children. While this study affirms BMI as the strongest identified predictor of IR overall, fitness was found to be a significant predictor of IR only in males, and race to be a significant predictor of IR only in females. This is consistent with previous work suggesting that males may have a stronger relationship between CVF and IR [1
]. A conceivable explanation for these findings would be that males were historically selected according to high fitness, rendering them more intolerant to the adverse effects of low fitness. In contrast, if females were in general selected by ability to reproduce in times of nutrient scarcity rather than physical fitness, this could increase susceptibility to adverse effects of over-nutrition.
While there is a strong correlation between an adult's fitness level and physical activity [32
], this correlation is much weaker in children [33
]. Reasons behind this difference are likely multi-factorial, but may include the fact that a healthy child's fitness is more genetically-determined than that of an adult, and childhood activity is often unstructured, un-sustained and of insufficient intensity to increase VO2 max [34
]. VO2 maximal exercise testing (or proxies there-of) measures a fitness level determined by a particular training stimulus that occurs with vigorous physical activity, such as competitive sports, but is much less indicative of habitual activity and leisure time physical activity, which also contribute to metabolic health. Studies repeatedly fail to show a relationship between habitual physical activity and VO2 max [35
]. Further complicating this issue is the fact that boys participate in sports at a higher rate than girls, giving them the type of fitness and experience with exercising at or near VO2 max but may not actually result in more daily activity overall. Cardiovascular fitness and habitual activity are both important for health, but the relative contribution of each may depend on an individual's particular genetic make-up and risk factors. Our data suggest that, if habitual activity is effective at preventing excessive weight gain, it may be more important than vigorous activity for overall health, particularly in girls.
Strengths and limitations
Strengths of this study include sample size, subject diversity, and setting. We sampled a large, racially diverse group of children with distribution of overweight/obese status which matched national trends. The school-based (rather than clinic- or lab-based) study design provided data that should be generalizable to other 'healthy' children. The school setting did introduce some limitations due to the "battlefield conditions" of the school setting, as described below, but overall, we feel that this is a major strength. Given the magnitude of the obesity challenge facing pediatrics today, it is important to develop assessments that can be done on a large and easily reproducible scale.
Limitations of this study include the reliability of PACER testing as a measurement of fitness, lack of pubertal stage documentation, and combining of the RD group into a single group. PACER has been shown to correlate very strongly with aerobic fitness measured via V02 max [27
] when administered by experienced staff, however, the PACER measurement of CVF depends heavily on speed, agility, and effort. Since males participate in competitive sports more than females [37
], their ability to have CVF measured accurately by the PACER test could be enhanced compared to females. The effort dependency of the PACER could be amplified when children perform the test in a peer environment, such as a co-ed physical education class at school.
Puberty hormones have known effects on body composition, markers of inflammation, and insulin sensitivity. Unfortunately, the school-based setting did not allow sufficient privacy to perform pubertal staging. While control for pubertal status would be optimal, the large sample size, distribution of varying stages of puberty amongst fitness levels, and the classification of fitness based on sex- and age-based normative data should all mitigate confounding effects of puberty on study results.
Finally, it is acknowledged that important differences in the relationships between CVF, IR, and inflammation likely exist between the various groups represented in the RD study group. We were prevented by statistical power calculations from performing separate analyses on each of these racial subgroups, and encourage more studies to investigate these potential differences in-depth.