Background: Human milk is the optimal form of nutrition for infants, especially sick or compromised infants, yet international data suggest that breastfeeding (feeding at the breast) and the use of expressed human milk (mother’s and donor’s milk) are limited in patients cared for in the neonatal intensive care unit (NICU). Objective: The goal of this study was to examine feeding status at hospital discharge among high risk infants. Methods: We used the 1991 World Health Organization infant feeding definitions, applied to the 72 hour period preceding discharge from the NICU. The study sample consisted of all high risk infants discharged from July 1, 2005, to June 30, 2006 from 13 Italian NICUs. Data on infant feeding in the last 72 hours were collected at discharge from the medical records. Results: We recorded data from 2948 subjects with a median gestational age of 35 weeks (IQR, 32-38), a median birth weight of 2200 g (IQR, 1630-2920) and a median length of stay of 16 days (IQR, 8-33). At discharge, 28% of all infants were fed exclusively with human milk: 31%, 25%, 22% and 33% respectively in the <1500 g, 1500-2000 g, 2000-2499 g and ≥ 2500 g birth weight categories. The proportion of infants not fed with human milk varied from 6 to 82% across different centers. Conclusion: Our study found limited breastfeeding and use of human milk among the NICU infants at discharge. At discharge, infants with a birth weight 1500-2499 g were fed exclusively with human milk less than those in higher or lower birth weight categories.