Sentinel serosurveillance for HIV infection has been carried out in Shaba province, Zaire, among consecutive pregnant women attending antenatal clinics from 1989 to 1991. There were four surveillance sites (three urban and one semiurban), at which a total of 13 surveillance studies were made of 4,205 women. Overall, 3.1% were HIV seropositive. There were no significant differences in HIV seroprevalence between surveillance sites, and HIV seroprevalence did not increase at any of the surveillance sites during the 2-year period of study. Since changes in the population studied did not occur between surveillance studies, it is believed that the observed stable trend reflects stable HIV seroprevalence rates in the general adult population of the surveillance sites. Collateral HIV seroprevalence data were available from 8,725 blood donors at 20 sites (six urban, 14 rural) in the province, who had an overall HIV seroprevalence of 4.6%. The higher HIV seroprevalence rate among blood donors was probably due to selection bias, since HIV seroprevalence rates in two blood banks, which relied nearly exclusively on replacement donors, were 2.7 and 2.8%, our best estimate for HIV seroprevalence in the three cities where blood banks exist and where no surveillance studies were carried out. The stable and relatively low HIV seroprevalence rates in Shaba province are in sharp contrast with the rapidly increasing and much higher rates in neighboring Zambia and other East African cities. Reasons for this discrepancy are unclear, and their eludication may yield critical information for HIV prevention programs.