Objective: Pain, stress as well as drugs may affect metabolic and endocrine measurements, especially in stressed children. The aim was to study how release of glucose and stress hormones are affected when procedural sedation with nitrous oxide or midazolam are used for establishing intravenous access in obese and lean children. Methods: In a prospective, double-blind, randomized study 90 children, 60 obese and 30 growth-retarded (GR), aged 5–18 years, with reported anxiety or difficulties connected with i.v. access, were randomized to 1 of 3 groups: oral midazolam (0.3 mg/kg, max. 15 mg), 50% nitrous oxide (N2O), and 10% N2O. In addition, all children received anesthesia cream (EMLA®) locally 1 h before i.v. access. Blood samples were drawn at 4 time points during 30 min after establishing venous access and, when feasible, after 24 h. The 24-hour sample was regarded as obtained during unstressed condition. The effect of procedural sedation was analyzed. Children’s evaluations of pain (Numeric Rating Scale) and procedure (Likert Scale) were correlated with mean values of cortisol and glucose after i.v. access. For the metabolic and hormone control measurements, 60 children aged 4–18 years (40 obese and 20 GR) served as controls. These children underwent a 24-hour blood sampling and did not receive sedation. The control samples were drawn 10–12 h after i.v. access. Results: After midazolam, significantly lower cortisol levels were found compared to both 50% N2O and 10% N2O and to unstressed controls. The growth hormone levels decreased with time in the midazolam group compared to 50 and 10% N2O, where the effect of time was reversed. Glucose levels among GR children increased from 0 to 30 min, whereas the opposite was found in obese children regardless of treatment. A post hoc analysis demonstrated significant correlations between children’s evaluations of the procedure and mean values of cortisol (r = –0.53), growth hormone (r = –0.52), and norepinephrine (r = –0.5) in children treated with a very low dose of N2O (10%). Conclusions: When sedation is insufficient during i.v. access, and blood sampling pain and stress affect hormone values, treatment with N2O or midazolam influence the glucose and stress hormone levels differently. These differences need to be accounted for when results are used for diagnosis and clinical decisions.