The pathogenetic potential and the true extent of human T leukemia/lymphotropic virus type I (HTLV-I) and type II (HTLV-II) infection are unknown. To find out more about HTLV-I/II seroepidemiology and the risks of iatrogenic transmission, we performed a serological study, screening 4086 healthy blood donors. A surprisingly high percentage of serum reactivity to HTLV-I/II antigens was observed by commercial ELISA (2.08%) and immunoblotting (IB) (0.85%) analysis, although none of the samples satisfied the (IB) criteria for positivity based on detection of gag protein p24 and at least one env gene product, either gp46 or gp61/68. To clarify these inconclusive results, we performed polymerase chain reaction (PCR) analysis for HTLV-I and HTLV-II provirus detection in peripheral blood lymphocytes, obtained from individuals with an apparent pattern of seropositivity. The data obtained by PCR failed to reveal evidence of HTLV-I/II provirus integration in peripheral blood cells, ruling out the possibility of a viral infection in these cases, and pinpointing the limitations of both serological methods used. Our observations suggest that serological assays alone are not a reliable tool for blood donor screening of HTLV-I/II infection and raise the important question of interpreting inconclusive results.