AIMS: To evaluate mothers' breastfeeding intention at birth and the breastfeeding rate at hospital discharge and at 3 months of life in a population of preterm infants. To study the relation between gestational age and the breastfeeding rate at birth. To analyze sociodemographic and neonatal factors that could influence breastfeeding rate. METHODS: We present a prospective study concerning preterm infants born between November 1st 2006 and April 20th 2007 at Brest University Hospital. Exclusion criteria were: death during hospitalization and neurological or digestive diseases not allowing enteral feeding. The following data were extracted from medical charts: birth weight, gestational age, medical care in the neonatal period, delivery circumstances, and feeding mode at hospital discharge. Mothers were asked their sociodemographic status and their intention to breastfeed. Parents were contacted 3 months later to determine the feeding mode at that time. Risk factors were studied for single births through univariate and multivariate analysis. RESULTS: We recruited 149 mother-infant dyads. There were 9 lost to follow-up at 3 months. Breastfeeding rates were: 69.13% (range: 61.7-76.5) at birth, 57.82% (range: 49.8-65.8) at hospital discharge, and 33.57% (range: 25.9-41.7) at 3 months. There was an inverse relation between breastfeeding rate and gestational age: OR=0.07 (0; 0.50). Other factors significantly associated with a high breastfeeding rate were: mother's high level of education: OR=3.19 (1.39; 7.33) and low birth weight: OR=0.20 (0.08; 0.53). CONCLUSION: Our breastfeeding rates are lower than recommended by WHO. Mothers of late preterm infants or with a low socioeconomic status should benefit from supportive care.