Although human immunodeficiency virus type 2 (HIV-2) infection among United States residents is considered rare, there are US populations at high risk. Few studies have surveyed these populations with a high likelihood of infection, that is, those with high percentages of persons from HIV-2-endemic areas and high prevalences of behaviors that would allow for transmission. Patients (n = 832) enrolled in a confidential HIV serosurvey at a hospital that serves a community with a relatively high percentage of West African immigrants, drug injectors, and persons who practice high-risk sexual activity were evaluated. Sera were tested for HIV type 1 (HIV-1) and HIV-2 by rapid enzyme immunoassays, standard enzyme immunoassays and Western blots. Eight of 832 patients were weakly reactive to HIV-2 on rapid assay, but none was confirmed to be infected when tested by standard immunoassay and Western blot. Five of these eight were reactive to HIV-1. Weak reactivity to HIV-2 antibody on the rapid assay is best explained by cross-reactivity with HIV-1 antibody; thus, even in this population at high risk for infection, false-positive reactions are more likely than true infections. The finding that HIV-2 is absent in this population at potentially high risk for infection corroborates the findings of other studies that HIV-2 infection is rare among US residents. These results support previous recommendations that, in settings other than blood collection facilities, HIV-2 testing should be selectively offered to persons with epidemiologic risk factors.