Choline is a conditionally essential nutrient required by the human body for synthesis of the neurotransmitter acetylcholine and the phospholipids phosphatidylcholine (PtdCho), lysophosphatidylcholine, sphingomyelin (SM) and plasmalogen. It also undergoes irreversible oxidation to form betaine, a major source of methyl groups. Betaine is excreted by human infants in large quantities (up to 1 mol/mol creatinine) during the first year reaching a maximum at 2-3 months of age. Since the betaine content of breast milk is very low and choline is the only known endogenous source of betaine, the supply of dietary choline to neonates may be critical particularly for premature babies whose intake may be restricted in the early days of life [ 11. Some premature babies receive expressed breast milk and we have measured the choline content, distribution and temporal variation to determine whether its supply is sufficient. Choline is present in breast milk as free choline (Cho), phosphorylcholine (PC), glycerophosphocholine (GPC), PtdCho and SM [2]. Previous studies have not considered the contributions of GPC and PC [3]. Milk samples were obtained from the Neonatal Intensive Care Unit of the Royal London Hospital. Samples were expressed by mothers and then frozen at -20°C until such time as the baby required the milk. At that time samples were defrosted and an aliquot taken for preparation for 'H nmr spectroscopy. The water soluble metabolites (Cho, PC and GPC) were measured in perchloric acid extracts of the milk, neutralised with KOH. Samples were run at pH values of 2 and 7 to distinguish the three metabolites from each other and from other metabolites with similar chemical shift values such as carnitme. Following lyophilisation the samples were dissolved in 0.6 ml 'H20 with 50 p1 lOmM fiunarate and 20 p1 20 mM 3-(trimethylsilyl)-2,2,3,3tetradeuteropropionate (TSPQ) for field locking, quantification and chemical shift referencing respectively. The samples were run at room temperature in a Jeol GSXSOO or Bruker AM400 spectrometer using a single pulse sequence (30 pulse angle, 2.73 s acquisition time and a 5 s recycling time). The choline containing phospholipids were measured in chloroform extracts. After separation the lower chloroform layer was evaporated under nitrogen, redissolved in 0.6 ml C2HCI3:C2H3O2H (2: 1) containing tetramethykhe (TMS; 0.03% v h in C2HCl3) and 0.3 mM 1,3,5-trichlorobee as the chemical shift reference and quantification standard respectively. The samples were run as descriied above. Figure 1 shows the concentration of choline metabolites in the milk expressed by one mother. This is qualitatively similar to the pattern observed for milk fiom 4 other mothers who have expressed milk over a similar time period and whose babies were born at between 28 and 37 weeks gestation. Until about one week after birth the total choline content is relatively low and is mostly present as Cho, PtdCho and SM. After that the total choline concentration rises quite markedly with the greatest increase being shown by GPC and PC which now contribute to more than half of the total choline 4 n