Transmission of cytomegalovirus (CMV) infection frommothers to preterm infants during breastfeeding may be symptomatic and long term consequences are unknown. This study evaluated the kinetics of CMV load in breast milk and the rate of postnatal CMV transmission via breastmilk from mothers to their preterm infants. Methods: Prospective study of mother child pairs after preterm delivery before 33 weeks. Exclusion of donor breast milk and of CMV-seropositive blood products.Material used was maternal CMV serostatus, ear swab of the infant at birth, weekly screened breast milk and children’s urine by rapid viral culture. Results: During a 5-month period 28 mother-infantpairs with 34 preterm infants were studied.Eighteen women (64.3%)were CMV-seronegative at birth;breast milk samples and the infants’ urine remained CMV-negative.Eight of the 10 seropositive mothers, who had 11 preterm infants, excreted CMV into breast milk (80%). CMV excretion into breast milk was detected during the first week after delivery in 66%cases and was at its peaked between 3 to 5 weeks after delivery.Out of the 7 CMV-exposed infants, CMV transmission was confirmed in only one asymptomatic case. Total quantity of breast milk intake did not seem discriminative for CMV transmission.Conclusion: In CMV-seropositive mothers of preterm infants a high incidence of CMV excretion into breast milk was detected. Despite this high rate, symptomatic infection did not occur. However, potential risk and severity of infection may be difficult to establish. Because breastfeeding is beneficial, new procedures for gentle virus inactivation of seropositive breast milk should be assessed. Transmission of cytomegalovirus (CMV) infection from mothers to preterm infants during breastfeeding may be symptomatic and long term consequences are unknown. This study evaluated the kinetics of CMV load in breast milk and the rate of postnatal CMV transmission via breastmilk from mothers to their preterm infants. Methods: Prospective study of mother child pairs after preterm delivery before 33 weeks. Exclusion of donor breast milk and of CMV-seropositive blood products. Material used was maternal CMV serostatus, ear swab of the infant at birth, weekly screened breast milk and children’s urine by rapid viral culture. Results: During a 5-month period 28 mother-infant pairs with 34 preterm infants were studied. Eighteen women (64.3%) were CMV-seronegative at birth; breast milk samples and the infants’ urine remained CMV-negative. Eight of the 10 seropositive mothers, who had 11 preterm infants, excreted CMV into breast milk (80%). CMV excretion into breast milk was detected during the first w eek after delivery in 66% cases and was at its peaked between 3 to 5 weeks after delivery. Out of the 7 CMV-exposed infants, CMV transmission was confirmed in only one asymptomatic case. Total quantity of breast milk intake did not seem discriminative for CMV transmission. Conclusion: In CMV-seropositive mothers of preterm infants a high incidence of CMV excretion into breast milk was detected. Despite this high rate, symptomatic infection did not occur. However, potential risk and severity of infection may be difficult to establish. Because breastfeeding is beneficial, new procedures for gentle virus inactivation of seropositive breast milk should be assessed.