In adults, leptin seems to cross the blood-brain barrier by a saturable transporter. This may contribute to the development of obesity. The present study in healthy children investigates leptin levels in plasma and cerebrospinal fluid (CSF) in relation to body constitution. This prospective study analyzed leptin levels in plasma and CSF samples (stored at -80 C) of patients without CNS infection or blood-brain barrier dysfunction. Inclusion criteria included temperature less than 38.5 C, C-reactive protein levels below 10 mg/L, CSF leukocyte levels less than 10(7)/L, no need for neurosurgical or oncological treatment, and no history of trauma. Four groups were designated according to body mass index. Sixty-five children (28 girls and 37 boys) entered the study. Plasma leptin (median) was 7.4 in girls and 2.6 ng/mL in boys., CSF leptin was 0.273 and 0.204 ng/mL, respectively, leading to CSF/plasma ratios of 0.045 and 0.071, respectively. Ratios were clearly dependent on body mass index percentiles (r = -0.484; P < 0.01, significant differences between groups by ANOVA). Median plasma leptin levels in the 4 groups (body mass index, <10th, 10th-50th, 50th-90th, and >90th percentile) were 2.0, 2.3, 4.1, and 8.8 ng/mL; CSF/plasma ratios were inversely related: 8.2%, 7.6%, 5.5% and 3.6%. In healthy children, CSF leptin levels account for approximately 5% of plasma levels. CSF/plasma ratios in girls are lower than those in boys, explaining why calorie intake and energy expenditure are not grossly different despite large differences in circulating plasma leptin concentrations. CSF/plasma ratios of lean children are higher than those in obese children. The dynamic changes in the CSF/plasma ratios are more pronounced in lean children, i.e. the nonlinear transport characteristics of the leptin system amplifies the information about changes in body energy stores in this population, indicating that leptin is part of a mechanism to protect the body from critical weight loss rather than to avoid obesity.