In this study we determined the influence of improving aerobic power (VO(2max)) on basal plasma levels of insulin and glucose of 11- to 14-year-old children, while accounting for body fat, gender, pubertal status, and leisure-time physical activity (LTPA) levels. Blood samples were obtained from 349 children after an overnight fast and analyzed for plasma insulin and glucose. Height, mass, body mass index (BMI), and sum of skinfolds (Sigma triceps + subscapular sites) were measured. LTPA levels and pubertal status were estimated from questionnaires, and VO(2max) was predicted from a cycle ergometry test. Regardless of gender, insulin levels were significantly correlated (P = 0.0001) to BMI, skinfolds, pubertal stage, and predicted VO(2max), but were not related to LTPA levels. Fasting glucose levels were not correlated to measures of adiposity or exercise (LTPA score, VO(2max)) for females; however, BMI and skinfolds were correlated for males (P < 0.006). The children then took part in an 8-week aerobic exercise program. The 60 children whose VO(2max) improved (>/=3 ml x kg(-1) x min(-1)) had a greater reduction in circulating insulin than the 204 children whose VO(2max) did not increase -16 (41) vs -1 (63) pmol x l(-1); P = 0.028. The greatest change occurred in those children with the highest initial resting insulin levels. Plasma glucose levels were slightly reduced only in those children with the highest insulin levels whose VO(2max) improved (P < 0.0506). The results of this study indicate that in children, adiposity has the most significant influence on fasting insulin levels; however, increasing VO(2max) via exercise can lower insulin levels in those children with initially high levels of the hormone. In addition, LTPA does not appear to be associated with fasting insulin status, unless it is sufficient to increase VO(2max).