Because of a persistently elevated caffeine half-life observed in a breast-fed infant during caffeine maintenance therapy, we conducted this prospective longitudinal study in two groups of infants (five exclusively breast-fed and 12 formula-fed). After 46 weeks' postconceptional age, all five breast-fed infants had a marked delay in caffeine elimination, compared with one infant in the formula-fed group. Four breast-fed infants had measurements of significantly longer caffeine half-lives compared with 12 formula-fed infants (76 +/- 13 hours v 21 +/- 28 hours and 54 +/- 9 hours v 16 +/- 13 hours at 47 to 50 weeks and 51 to 54 weeks postconceptional age, respectively), as well as significantly higher trough blood levels (three- to five-fold) after 46 weeks' postconceptional age. The fifth breast-fed infant accumulated caffeine secondary to a steep increase in caffeine half-life from 102 hours at 44 weeks to 372 hours at 51 weeks. The elevated blood caffeine levels in breast-fed infants was not related to higher daily dosage of caffeine citrate (4.4 mg/kg compared with 8.3 mg/kg in the formula-fed group at 56 weeks' postconceptional age). Daily consumption of caffeine was low or nonexistent in four nursing mothers, and transfer of caffeine to the infant was considered to be trivial. The findings from this study suggest, as does breast milk jaundice due to inhibition of glycuronyl transferase, that some components of human milk (free fatty acid, lipase activity, or other factors) inhibit or repress the postnatal normal maturation process of caffeine metabolism by hepatic cytochrome P-450.