OBJECTIVE The first human retrovirus, HTLV-I, was isolated in 1980; HTLV-II was described later. The former is endemic in southwestern Japan, the Caribbean and equatorial Africa; whereas the latter prevails in intravenous drug addicts, being also endemic in American indian populations. Both viruses are either sexually transmitted, from mother to child mainly by breast-feeding, by blood transfusion or by sharing contaminated needles. With regard to transmission, since they are intracellular viruses, it occurs only when whole blood or cellular components are transfused; this is not the case when either plasma or plasma derivatives are used. The likelihood of transmission decreases as the storage time increases. HTLV-I is associated, at least, with two diseases: adult T-cell leukaemia/lymphoma (ATLL), and the tropical spastic paraparesis (TSP) or HTLV-I-associated myelopathy (HAM). ATLL occurs after a latency period of 20 to 30 years; whereas the incubation period ranges from 3 to 5 years in the case of the neurological disease. Most individuals infected with the virus remain healthy; the risk of developing the hematological complication is 2-4% whereas it is below 1% in the case of TSP. No clear association of HTLV-II with any known disease has been reported as yet. In this study, we have assessed the prevalence of HTLV-I and HTLV-II in the sera of the blood donors who have come to our Division, with the aim of avoiding the spreading of this oncogenic virus by transfusion. This study could serve as a measure of the infection in the general population. METHODS A total of 28,897 samples were analyzed from May 1993 to January 1995. Anti-HTLV-I/II antibodies were analyzed by the method of passive agglutination of gelatin participles (PA). Samples which reacted were tested again by the same method, and those reacting for the second time were further confirmed by Western blot (WBT), a method with the ability to differentiate between antibodies anti-HTLV-I and anti-HTLV-II. RESULTS Of the 28,897 samples, 47 were repeatedly reactive by PA (0.16%). Analysis by WBT resulted in 10 reactive results with HTLV-I (0.035%), 2 reactive results with HTLV-II (0.007%); in one sample it could not be determined whether the anti-HTLV-I or anti-HTLV-II antibody was present. Of the remaining samples, 21 were non-reacting, whereas 13 were indeterminated. CONCLUSIONS Prevalence of HTLV-I and HTLV-II seropositive blood donors is low and similar to that found in other non-endemic countries. We believe that routine evaluation of anti-HTLV-I and HTLV-II antibodies in blood donors would be warranted in our country, since transmission of the viruses by transfusion of blood components has been clearly shown. It is possible that the recipients of the reactive units do not develop the disease. Nevertheless, these individuals constitute an important source of virus dissemination, both during the perinatal period and by sexual intercourse. In fact, advise to seropositive donors would prevent transmission by these routes. Lastly, it should be noticed that investigation of anti-HTLV-I/II antibodies could result in a surrogate method for detecting other viral infections transmitted by these routes.