Preconceptional maternal immunity is known to lessen the severity of congenital CMV infection; however, it is not clear whether maternal immunity to CMV will significantly decrease the risk of congenital infection in future pregnancies. In order to determine whether preconceptional maternal immunity decreases risk of congenital CMV infection, we compared the rate of congenital infection among offspring of women who were immune at a previous pregnancy with the rate among infants whose mothers seroconverted in the interval between pregnancies. All newborns at University Hospital are screened for congenital CMV infection by viral isolation, and cord sera are saved when available. Between January 1993 and December 1995, congenital CMV infection was found in 36 newborns of 3305 (1.1%) multigravida mothers. The study population was comprised of 1700 women with a previous cord serum, including the mothers of 30 newborns with congenital CMV infection. The study population was predominantly black (83%), low income (81%) and CMV seroimmune (82%); their mean age at most recent delivery was 24 ± 5 years, and they had a mean of 2 ± 1 previous pregnancies. Among women who seroconverted between pregnancies, 12/47 (25%) had offspring with congenital CMV infection. The time interval between pregnancies was 36 ± 15 months for mothers of newborns with congenital CMV infection compared with 46 ± 33 months for mothers who seroconverted but had uninfected babies. Among women who were seroimmune at prior pregnancy, only 14/1047 (1.3%) had offspring with congenital CMV infection. The time interval between pregnancies was 36± 24 months for those with infected babies compared with 50 ± 34 months for those with uninfected newborns. The risk of congenital CMV infection is 25 times greater among offspring of women who seroconvert between pregnancies compared with those who are immune at a previous delivery. In both the seroimmune and seroconversion groups, those women with infected offspring had a slightly shorter time interval between pregnancies than women of uninfected newborns. These findings suggest that a CMV vaccine given postpartum could significantly reduce the rate of congenital CMV infection.