Since nationwide surveillance for pneumococcal bacteremia in the United States is not done, community-based studies are useful alternative methods to monitor trends in this disease. Data on the incidence of pneumococcal bacteremia in Charleston County, South Carolina, from 1974 to 1976, have been used to support cost-effective pneumococcal vaccine programs for the elderly. We reevaluated the incidence of pneumococcal bacteremia in Charleston County in 1986 and 1987 to assess whether earlier estimates remained valid given changes in medical practice. During 1986 and 1987, overall annual incidence of pneumococcal bacteremia in Charleston County was 18.7 per 100,000 (95% confidence limits, 13.7 to 23.6 per 100,000), which represents a 2.3-fold increase over the earlier rate. The increase coincided with a 2.2-fold increase in the annual number of blood cultures processed at four Charleston County hospital laboratories from 1975 to 1987 despite only a 1.1-fold increase in the number of patients discharged from these hospitals. Annualized rates increased 2.3-fold for adults more than or equal to 65 years old to 53 per 100,000 and 4.6-fold for children less than 2 years old to 162 per 100,000. The case-fatality rate of bacteremic patients was 18%, compared with 21% in the earlier study. The case-fatality rate for adults more than or equal to 65 years of age was 44%. Ninety-one percent of adults 19 to 64 years old with bacteremia had underlying medical conditions for which pneumococcal vaccine is recommended; all persons 55 to 64 years old had at least one underlying condition. The marked increases in pneumococcal bacteremia rates detected are likely due to more routine culturing of blood from symptomatic patients with pneumococcal disease. These findings emphasize the need for effective programs promoting use of pneumococcal vaccine in high-risk groups, particularly those more than or equal to 65 years old, and the development of a more immunogenic vaccine for children less than 2 years old.