Background: Analysis of insulin-like growth factor I in serum (S-IGF-I) is an integral component in the diagnosis of GH-related disorders and is going to be of interest in the diagnosis and follow-up of many disorders. The objective of the present study was to develop cross-sectional reference values for S-IGF-I measured by an automated chemiluminescence immunoassay (Nichols Advantage<sup>®</sup>). Methods: The study included samples from 3,961 healthy subjects (2,201 males, 1,760 females) aged 1 month to 88 years. Six laboratories were involved in this study and the samples were analyzed by one of seven automated immunoassay systems run in these laboratories. For data analysis, polynomial age and sex-specific models were fitted after transformation of S-IGF-I values. Results: The results show the well-known age dependency of S-IGF-I levels. At ages <20, higher S-IGF-I levels were seen in girls with an estimated mean peak of 410 µg/l at age 14 and an estimated mean peak of 382 µg/l at age 16 in boys. Thereafter, a rapid decrease was seen to approximately 25 years of age, followed by a slow age-dependent decrease. In adulthood, S-IGF-I in males were slightly, but significantly higher than in females. It could be shown that the mean values of some reference sample subgroups differed significantly from the total mean. However, the multicenter approach used in this study reduces the impact of systematic population, sample handling and laboratory differences on the calculated reference mean. Conclusion: The present study establishes age- and sex-specific reference values for a fully automated immunoassay system based on a large population of healthy subjects. The established reference values may be used for this immunoassay system in different laboratories provided that the systematic difference between systems is low.