We performed a retrospective study of 57 HIV-positive pregnant women and their children to look for maternal parameters related to rates of vertical transmission of HIV. Sixteen (28%) infants were HIV-infected. There was no positive correlation between maternal risk factors for HIV and vertical transmission of HIV. Multiparity seemed to be related to a higher prevalence of vertical transmission, whereas neither preterm delivery nor cesarean section resulted in different rates of vertical transmission, although they tended to have some protective effect. Maternal beta 2-microglobulin and serum neopterin levels were not related to different rates of vertical transmission of HIV. CD4+ cell counts did so only when cut-off values of 400/mm3 were taken. All the women in CDC group IV and all the HIV p24 antigen-positive transmitted the infection to their infants, whereas only 7/48 women in CDC groups II and III (P = 0.000006) and 9/48 HIV p24 antigen negative women (P = 0.00006) gave birth to infected infants. No other maternal characteristics were associated with different rates of vertical transmission in a multivariate analysis restricted to women in CDC groups II and III and with HIV p24 antigen negativity.