Energy concentration of human milk estimated from 24-h pools and various abbreviated sampling schemes.

The concentration of gross energy in 24-h pooled samples of milk is compared with that of single samples obtained during the same period. Samples were collected from mothers exclusively breastfeeding their infants on demand. The objective of the comparison was to evaluate simplified sampling schemes for approximating the concentration of energy of milk produced over 24 h. The lowest concentration of energy of single samples occurred in milk collected between 12 midnight and 6 a.m. The concentration of energy of milk was related negatively to the volume expressed and the time interval preceding each expression. The interval preceding each expression was correlated positively to the volume expressed. These results indicate that the pattern of milk expression is important in the design of sampling schemes. Therefore, sampling schemes may not be transferable between populations with distinct feeding patterns. The highest correlation between single samples and the 24-h value was obtained from samples collected between 12 midnight and 6 a.m. (r = 0.84, p less than 0.001). The widths of the 95% confidence and prediction intervals for the equation relating values from single samples to the 24-h pools were +/- 2.7% and +/- 10.5%, respectively, of the mean 24-h pool sample value when three single samples (collected between 12 midnight and 6 a.m., 6 a.m. and 12 noon, and 6 p.m. and 12 midnight) were included in the predictive equation.