BackgroundGender-specific differences in hypothalamus–pituitary–adrenal (HPA) axis activity have been postulated to emerge during puberty. We conducted a systematic review and meta-analysis to test the hypothesis that gender-specific differences in HPA axis activity are already present in childhood.MethodsFrom inception to January 2016, PubMed and EMBASE.com were searched for studies that assessed non-stimulated cortisol in serum or saliva or cortisol in 24-h urine in healthy males and females aged ≤18 years. Studies that conform with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) statement were reported. Standardized mean differences (95% CIs) were calculated and analyzed using fixed-effect meta-analysis stratified for age: <8 years (prepubertal) and 8–18 years (peri-/postpubertal). For comparison, we ran the same analyses using random-effects models.ResultsTwo independent assessors selected 413 out of 6158 records (7%) for full-text screening, of which 79 articles were included. Of these, 58 (with data on 16,551 subjects) were included in the meta-analysis. Gender differences in cortisol metabolism differed per age group. Boys aged <8 years had 0.18 (0.06; 0.30) nmol/L higher serum and 0.21 (0.05; 0.37) nmol/L higher salivary cortisol levels, while between 8 and 18 years, boys had 0.34 (0.28; 0.40) nmol/L lower serum and 0.42 (0.38; 0.47) nmol/L lower salivary cortisol levels. In 24-h urine, cortisol was consistently higher in boys, being 0.34 (0.05; 0.64) and 0.32 (0.17; 0.47) μg/24 h higher in the <8- and 8–18-year groups, respectively. However, gender-differences in serum cortisol <8 years and between 8 and 18 years were absent when using random-effects models.ConclusionsGender differences in cortisol metabolism are already present in childhood, with higher salivary cortisol in boys aged <8 years compared to girls. This pattern was reversed after the age of 8 years. In contrast, the gender-specific difference in cortisol production as assessed through 24-h urine did not change with age. Although differences were small, and analyses of gender differences in serum cortisol were inconclusive, they might contribute to gender-specific origins of health and disease.Electronic supplementary materialThe online version of this article (doi:10.1186/s13293-016-0123-5) contains supplementary material, which is available to authorized users.