HHV6 PCR was negative in all but one patient in the first days after transplantation, and became positive in 18 of the 30 patients after a median of 18 days after transplantation. This delay was in accordance with other studies in which HHV6 replication was also found early after transplantation. Most of the patients in whom HHV6 was detected by culture had positive PCR soon before, or concomitantly to, culture. Finally, HHV6 PCR remained positive for a limited period of time (4335 days). For all these reasons, we think that the method used in our study did not detect HHV6 in a latent stage. We agree with Gentile et al. (1) that detection of HHV-6 DNA in cell free samples has been associated with active viral replication. However, controversial results have been obtained using HHV6 DNA detection in serum or plasma. Using this technique, Osiowy et al. (2) found HHV6 in 21% of their control sera obtained in healthy adults, suggesting that the presence of HHV6 DNA in serum could not be systemat- ically associated with active infection. Seccherio et al. (3) detected HHV6 in plasma of 23% of bone marrow transplant recipients and 22% of HIV patients, but no clinical correla- tion was attempted for HIV patients. Better definition of the contribution of HHV6 infection to posttransplant complications can probably be obtained using quantitative techniques like HHV6 antigenemia (4) or stan- dardized quantitative PCR (5, 6). This would allow the cor- relation of viral load with symptomatic disease, to better determine the relationship between cytomegalovirus (CMV) and HHV6 infections. In our study, all patients at high risk for CMV received antiviral prophylaxis during the first 3 or 6 months after transplantation. Only five patients developed CMV infection, a median of 69 days after transplantation (range, 34 – 89 days). These infections occurred in three patients with HHV6 infection, and in two patients without HHV6 infection. All three patients with HHV6 infection developed severe CMV disease; of the 2 patients without HHV6 infection, one pa- tient developed asymptomatic CMV infection and one devel- oped severe CMV disease. Because of the low incidence of CMV infection due to effective prophylaxis, no relationship between CMV and HHV6 was established. In Gentile et al. (1), patients did not receive any antiCMV prophylaxis, but were given preemptive therapy when antigenemia became positive. The authors found a strong correlation between HHV6 DNA-positive patients and CMV infection detected by pp65 antigenemia. However, because of effective preemptive therapy, no association was found with CMV disease. The differences in populations studied, methods chosen for HHV6 detection, and strategies used to prevent CMV disease can certainly explain the contradictory results observed in both studies. FR ´ ED ´ ERIQUE JACOBS CORINNE LIESNARD Infectious Diseases Clinic and Virology Laboratory Erasme University Hospital Brussels, Belgium DOI: 10.1097/01.TP.0000131224.11605.C1 Address correspondence to: Fre ´de ´rique Jacobs, M.D., Ph.D., Infec- tious Diseases Clinic and Virology Laboratory, Erasme University Hospital, Route de Lennik 808, 1070 Brussels, Belgium. E-mail: maladies.infectieuses.erasme@ulb.ac.be. Received 5 November 2003. Accepted 11 November 2003. REFERENCES 1. Gentile G, Capobianchi A, Ferraironi M, et al. Relationship of serum human herpesvirus 6 DNA with cytomegalovirus pp65 antigenemia in allogenic bone marrow transplant recipients. Transplantation 2004; 77: 1907. 2. Osiowy C, Prud’homme I, Monette M, et al. Detection of human herpesvirus 6 DNA in serum by a microplate PCR-hybridization assay. J Clin Microbiol 1998; 36: 68. 3. Secchiero P, Carrigan DR, Asano Y, et al. Detection of human herpesvirus 6 in plasma of children with primary infection and immunosuppressed patients by polymerase chain reaction. J Infect Dis 1995; 171: 273. 4. Lautenschlager I, Linnavuori K, and Hockerstedt K. Human herpesvirus-6 antigenemia after liver transplantation. Transplantation 2000; 69: 2561. 5. Ljungman P, Wang FZ, Clark DA, et al. High levels of human herpesvirus 6 DNA in peripheral blood leucocytes are correlated to platelet engraft- ment and disease in allogeneic stem cell transplant patients. Br J Haematol 2000; 111: 774. 6. Locatelli G, Santoro F, Veglia F, et al. Real-time quantitative PCR for human herpesvirus 6 DNA. J Clin Microbiol 2000; 38: 4042. LATE-ONSET SEVERE ACUTE CELLULAR REJECTION AFTER ADULT ABO-INCOMPATIBLE LIVER TRANSPLANTATION: A CASE REPORT A 49-year-old female underwent ABO-incompatible (donor, A+; recipient, B+) living donor liver transplantation for end- stage liver disease of primary biliary cirrhosis. On the day of transplantation, the isoagglutinin titers were considerably high at 1:64 of IgM and 1:128 of IgG, despite five sessions of multiple plasmapheresis using 0.8 unit/kg/session of AB+ fresh frozen plasma during the 3 days prior to transplant. However, the deterioration of the patient’s condition did not allow us to postpone the transplant. The immunosuppressive protocol the patient received basically followed the method introduced by Tanabe et al. (1), including intraportal infusion therapy consist- ing of 0.01 g/kg/min of prostaglandin E1, 125 mg/day of meth- ylprednisolone, and 1000 mg/day of gabexate mesilate. How- ever, the isoagglutinin titers elevate further to 1:256 of IgM and 1:512 of IgG on the first postoperative day (POD), and daily plasmapheresis was started in order to decrease them. Despite this treatment, the titers increased further on POD2 with 1:1024 of both IgM and IgG, and thereafter sus- tained these values. Moreover, a sudden stoppage of bile output resulting in rapid elevation of serum total bilirubin level from 4 mg/dl to 9 mg/dl over an 8-hour period, and a rapid decrease in platelet count which fell to below 30000/uL from 110000/uL over an 8-hour period, were observed on POD2. We considered this episode indicative of antibody-mediated re- jection (AMR). Steroid pulse therapy was then started, and plasmapheresis continued daily until POD5. However, the stop- page of bile output, and sustained low platelet count below 30000/uL despite platelet transfusion, continued up to approx- imately 48 hours after onset. Fortunately, the AMR was reme- died by gammaglobulin bolus infusion at 0.4g/kg/day for 5 days started at POD4, in addition to the above protocol including intraportal infusion, as we previously reported (2). LETTERS TO THE EDITOR June 27, 2004 1909