BACKGROUND: IGF-I is a major regulator of growth, influenced primarily by diet in infancy and primarily by GH in childhood. Breastfed infants have lower IGF-I levels compared to formula fed and tend to be shorter. The higher protein content of infant formula has a stimulatory effect on IGF-I production. Conversely, studies suggest that later in childhood, those breastfed are taller and have higher IGF-I levels. Therefore, it has been suggested that the IGF-I axis may be programmed by diet during infancy. The association between IGF-I in infancy and later life is not known. OBJECTIVE: To examine the association between IGF-I in infancy and adolescence. DESIGN: Infants (109) from the observational Copenhagen cohort study. METHODS: Serum-IGF-I was measured during infancy (2, 6, and 9 months) and at follow-up at 17 years. Associations were examined by correlation tests and linear regression controlling for gender, breastfeeding, and other covariates. Likelihood ratio test based on residual log likelihood was applied for analysis including all measurements during infancy. RESULTS: There was an inverse association between IGF-I at 9 months and 17 years (r=-0.39, P=0.014, and n=40). A 1 ng/ml higher IGF-I concentration at 9 months corresponded to 0.95 ng/ml lower IGF-I concentration at 17 years. IGF-I levels at 2 and 6 months were not significantly associated with IGF-I at 17 years, but the estimated directions were negative. These associations were not changed when adjusted for breastfeeding and other covariates except IGF-I at 2 months which was significantly negatively associated with IGF-I at 17 years (P=0.030) corresponding to a 0.96 ng/ml lower IGF-I concentration at 17 years per ng/ml IGF-I at 2 months. Inclusion of all measurements during infancy showed a negative association with 17-year values (r=-0.26, P=0.043, and n=109). CONCLUSION: The results support the hypothesis that the IGF-I axis can be programmed early in life.