High incidences of human herpesvirus (HHV)-6 encephalitis have recently been reported from several Japanese SCT centers. To evaluate the effect of low-dose foscarnet (PFA) in preventing HHV-6 infection among recipients of unrelated BM or cord blood (CB), we examined consecutive cohorts without prophylaxis against HHV-6 (cohort 1, n=51) and with PFA prophylaxis (cohort 2, PFA 50 mg/kg/day for 10 days after engraftment, n=67). Plasma real-time PCR assay was performed weekly. High-level reactivation defined as HHV-6 DNA > or =10(4) copies/mL by day 70 was the primary endpoint. No significant reduction of high-level reactivation was seen in cohort 2 (19.4%) compared with cohort 1 (33.8%, P=0.095). A trend was identified toward fewer high-level HHV-6 reactivations in cohort 2 among recipients of unrelated BM (P=0.067), but no difference in incidence was observed among CB recipients (P=0.75). Breakthrough HHV-6 encephalitis occurred following PFA prophylaxis in three patients, and incidence of HHV-6 encephalitis did not differ between cohort 1 (9.9%) and cohort 2 (4.5%, P=0.24). In conclusion, 50 mg/kg/day of PFA does not effectively suppress HHV-6 reactivation and cannot prevent all cases of HHV-6 encephalitis. To effectively prevent HHV-6 encephalitis, alternative approaches based on the pathogenesis of HHV-6 encephalitis will probably be required.