In order to find a sensitive method to evaluate adrenocortical function in premature infants, we compared low- (0.5 μg/1.73 m<sup>2</sup>) and standard-dose (250 μg/1.73 m<sup>2</sup>) adrenocorticotropin tests (LD- and SD-ACTH) in 12 very-low-birth-weight infants before and 2 days after the end of dexamethasone therapy (duration 9–14 days) for chronic lung disease. Basal serum cortisol levels were inappropriately low in several infants already before dexamethasone therapy (median 190, range 60–357 nmol/l). The 95% confidence intervals of mean serum cortisol levels at 20 min were equal in LD- and SD-ACTH, while at 60 min, the low-dose gave a clearly lower response than the standard-dose test. The LD-ACTH can be used in premature infants as in older children and adults, but the criteria for adrenocortical insufficiency need to be defined.