The aim of the present study was to test the practicability of sequential cortisol determinations in saliva of low birth weight neonates and to evaluate the impact of systemic and inhaled glucocorticoid therapy on saliva concentrations of cortisol in preterm neonates with bronchopulmonary dysplasia (BPD). Salivary cortisol levels were measured by RIA in saliva samples from 10 full-term and 10 preterm healthy neonates and from 20 preterm neonates with BPD during systemic [dexamethasone (DEX); n = 10] or topical steroid therapy [budesonide (BUD); n = 10]. Saliva samples of each individual were collected on 3 consecutive days at 06.00, 12.00, 18.00 and 24.00 h. Cortisol levels in saliva ranged from 0.8 to 60.6 nmol/l (median 6.5 nmol/l) in full-term neonates, from 0.6 to 52.1 nmol/l (median 5.5 nmol/l) in preterm neonates, from 0.4 to 14.0 nmol/l (median 1.0 nmol/l) in preterm neonates treated with DEX and from 0.4 to 15.2 nmol/l (median 2.5 nmol/l) in preterm neonates treated with BUD. Autocorrelation analysis revealed a distinct endogenous cortisol rhythm in 2 of the 10 healthy full-term neonates and in 3 of the 10 healthy preterm neonates with a wavelength of 12–30 h. Salivary cortisol levels in preterm neonates treated with DEX or BUD were significantly lower than those measured in healthy preterm neonates. These results demonstrate that the measurement of salivary cortisol levels is a reliable and practicable way of assessing adrenal function in full-term and preterm neonates. This study also shows for the first time that some neonates display an endogenous cortisol rhythm which is not coupled to the exogenous day/night cycle. Furthermore, systemic and nebulized glucocorticoids suppress adrenal function in low-birth-weight neonates. After treatment these children should be closely monitored for potential adrenal insufficiency.