BACKGROUND: Almost all human immunodeficiency virus (HIV) transmission via blood or tissues that has occurred since anti-HIV screening was implemented in 1985 is traceable to blood given after infection but before antibody seroconversion, a time that is referred to as the window period. In this study, the performance of newer assays designed to detect viral and serologic markers soon after infection is assessed, and the reduction in the window period achieved by these assays is estimated. STUDY DESIGN AND METHODS: Three cohort studies of persons at high risk for acquiring HIV infection were identified. These studies included well-controlled HIV type 1 (HIV-1) polymerase chain reaction (PCR) analyses of serial preseroconversion specimens from HIV-1- seroconverting homosexual men or intravenous drug users. Of 81 enrollees with anti-HIV-1 seroconversion documented by a viral lysate anti-HIV-1 enzyme immunosorbent assay (EIA) available in 1989, 13 (16%) had PCR-positive preseroconversion specimens. In the present study, sera from these 13 PCR-positive samples were further tested for anti- HIV by 10 contemporary EIAs and 6 supplemental assays, as well as being tested for plasma p24 antigen and HIV-1 RNA. Preseroconversion sera from 38 HIV-1 DNA PCR-negative cohort participants were also tested by selected anti-HIV EIAs and tested for p24 antigen and HIV-1 RNA. On the basis of these laboratory data and the intervals between blood drawing in all 81 men, the reduction in the preseroconversion window period achieved by these new assays was estimated with a mathematical model developed to analyze seroconversion data. RESULTS: Nine (69%) of the 13 preseroconversion PCR-positive samples had anti-HIV that was detectable by one or more contemporary anti-HIV-1 or anti-HIV type 2 EIA. Supplemental antibody assays were negative on all four EIA-nonreactive preseroconversion samples and negative or indeterminate on a high proportion of the nine EIA-reactive PCR-positive samples. Eight (61%) of the 13 samples were p24 antigen-positive, and 11 (85%) were HIV-1 RNA-positive. The estimated reductions in the window period (relative to the index viral lysate-based anti-HIV EIA) were as follows: contemporary anti-HIV-1/2 EIAs, 20.3 days (95% Cl, 8.0–32.5); p24 antigen and DNA PCR, 26.4 days (95% Cl, 12.6–38.7); and RNA PCR, 31.0 days (95% Cl, 16.7–45.3). CONCLUSION: Recent improvement in the sensitivity of anti-HIV assays has resulted in significant shortening of the preseroconversion window period. Consequently, the incremental reduction in the window period that could be achieved by implementing direct virus-detection assays has diminished significantly.