Neonatal glucocorticoid (GC) treatment is used to prevent bronchopulmonary dysplasia (BPD) in prematurely born babies. In the 1990s, treatment regimens with relatively high doses of dexamethasone (DEX) were common. As an alternative, hydrocortisone (HC) was used. Earlier, we compared long-term effects of both GCs in children aged 7–10 and detected adverse effects of neonatal DEX treatment, but not of HC, on a range of outcomes. The aim of the current cohort study was to investigate whether long-term effects of neonatal DEX were maintained and whether effects of HC remained absent at adolescent age (14–17years). We compared 71 DEX-treated and 67 HC-treated adolescents. In addition, 71 adolescents who were not neonatally treated with GCs participated. All were born <32weeks of gestation. DEX-treated girls showed increased adrenocorticotropic hormone (ACTH) and cortisol responses in the Trier Social Stress Test. The cortisol awakening response was lower in HC-treated participants compared to untreated participants. Negative feedback function of the HPA-axis in the dexamethasone suppression test did not differ between groups. In contrast to our observations at the age of 7–10years, we did not observe group differences in mitogen-induced cytokine production at the age of 14–17years. DEX-treated girls showed more social problems and anxious/depressed behavior than HC-treated girls. Untreated girls showed more problem behavior as well. In conclusion, our results suggest that, especially in girls, neonatal DEX has a programming effect on the HPA-axis and on the ability to adjust to the environment. The loss of group differences on immune system measures indicate that potentially negative effects detected at a younger age subsided.