We used population-based surveillance in New Jersey in 1986 to quantify the risk of invasive pneumococcal disease in persons with the acquired immunodeficiency syndrome (AIDS) and in those who went on to develop AIDS. Using pneumococcal surveillance, we also devised a method to estimate the size of the immunosuppressed population infected with the human immunodeficiency virus (HIV), the so-called pre-AIDS population. From rates of pneumococcal disease that occurred in areas with a low incidence of AIDS, we calculated the number of patients expected to contract pneumococcal disease in areas with a high incidence of AIDS. There were 59 more cases of pneumococcal disease observed than expected; 14 of these patients had AIDS by the time of pneumococcal infection. We attributed the remaining 45 cases to the increased risk of pneumococcal infection in pre-AIDS. The pre-AIDS pneumococcal cases and the attack rate of pneumococcal disease in pre-AIDS were used to estimate the size of the 1986 pre-AIDS New Jersey population as 8823 pre-AIDS cases (95% confidence interval, 7377 to 10 714) or 0.37% of the adult New Jersey population. Surveillance for marker diseases may provide a simple, independent method of estimating the pre-AIDS population.