Lassa virus-immune plasma has been used to treat human Lassa fever patients; however, criteria for plasma selection were based arbitrarily on available serologic tools and protective efficacy was never directly assessed. To test the validity of plasma therapy for Lassa virus infections in an animal model, and to develop biologically relevant criteria for selection of protective immune plasma, inbred, strain 13 guinea pigs were infected with a lethal dose of Lassa virus and treated with various Lassa-immune plasmas obtained from guinea pigs, primates, and convalescent human patients. Neutralizing antibody titers were determined in a virus dilution, plaque reduction test, and were expressed as a loglo plaque-forming units (PFU) neutralization index (LNI). All guinea pigs treated with immune plasma 6 mlikgltreatment on days 0. 3, and 6 after virus inoculation were protected, provided the LNI exceeded 2.0. Plasmas obtained from donors in early convalescence (32-45 days) had low titers of N-antibody (LNI < 2 ) and failed to confer protection, despite high titers of Lassa antibody measurcd in the indirect fluorescent antibody (IFA) test. Higher doses of marginally titered plasma conferred increased protection. The degree of protection and suppression of viremia was closely associated with LNI and not IFA titers. Administration of low-titered plasma did not result in immune enhancement. A high dose of human plasma from Liberia (12 ml/kg/treatment) was required to confer complete protection to guinea pigs infected with a Lassa virus strain from Sierra Leone (LNI = 1.6), while a lower dose (3 mllkgitreatment) was sufficient for protection against a Liberian strain (LNI = 2.8), suggesting that a geographic matching of immune plasma and Lassa virus strain origin may increase treatment success. These studies support the concept of plasma therapy for Lassa infection and suggest that the plaque reduction neutralization test is more appropriate than the IFA test for predicting protective efficacy of passively administered plasma.