Preterm babies are exposed to multiple stressors and this may have long- term effects. In particular, high levels of endogenous cortisol might have a programming effect on the hypothalamicpituitary- adrenal axis as may administered glucocorticoids. In this study, we aimed to test the hypothesis that the level of endogenous and exogenous glucocorticoid exposure during the neonatal period predicts the saliva cortisol response to immunization at 4 mo of age. We followed 45 babies born below 32 wk gestation. We showed that their concentration of plasma cortisol during the first 4 wk was 358, 314, 231, and 195 nmol/L cortisol, respectively (geometric mean). This is four to seven times higher than fetal levels at the same gestational age range. We used routine immunization at 4 mo and 12 mo as a stressor and measured the change in saliva cortisol as the stress response. Mean circulating cortisol in the first 4 wk predicted the cortisol response at 4 but not at 12 mo. Path analysis showed that birthweight for gestational age, thera- peutic antenatal steroids, and therapeutic postnatal steroids also contributed to the magnitude of the saliva cortisol response at 4 mo. This provides evidence that the magnitude of glucocorticoid exposure, both endogenous and exogenous, may have an effect on later stress responses. Preterm babies are exposed to multiple stressors and this may have long-term effects. In particular, high levels of endogenous cortisol might have a programming effect on the hypothalamic pituitary-adrenal axis as may administered glucocorticoids. In this study, we aimed to test the hypothesis that the level of endogenous and exogenous glucocorticoid exposure during the neonatal period predicts the saliva cortisol response to immunization at 4 mo of age. We were 45 babies born below 32 wk gestation. We showed that their concentration of plasma cortisol during the first 4 wk was 358, 314, 231, and 195 nmol / L cortisol, respectively (geometric mean). This is four to seven times higher than fetal levels at the same gestational age range. We used routine immunization at 4 mo and 12 mo as a stressor and measured the change in saliva cortisol as the stress response. Mean circulating cortisol in the first 4 wk predicted the cortisol response at 4 but not at 12 mo. Path analysis showed tha t birthweight for gestational age, thera-peutic antenatal steroids, and therapeutic postnatal steroids also contributed to the magnitude of the saliva cortisol response at 4 mo. This provides evidence that the magnitude of glucocorticoid exposure, both endogenous and exogenous, may have an effect on later stress responses.