Human herpesvirus-6 and bone marrow transplantation.

The polymerase chain reaction (PCR) technology offers many benefits to the study of infectious disease, but its use must be restrained by the limits of interpretation of the data generated. Wilborn et al' recently described the results of a PCR-based prospective study of human herpesvirus-6 (HHV-6) infections in bone marrow transplant (BMT) recipients. Their basic conclusions were that HHV6 infections correlate with graft-versus-host disease and that no other clinical correlations could be drawn, ie, "No specific clinical condition could be associated with HHV-6." Such a conclusion is unwarranted and does a disservice to the field of HHV-6 research. PCR is an inappropriate technology for such a study. Various aspects of the use of the PCR technique to study HHV6 infections are summarized in Fig 1. It is generally agreed that reactivations of HHV-6 infection occur frequently after BMT.* Such a reactivation event and its aftermath is illustrated in the top of the figure. Reactivation of the latent virus occurs (panel 2), and the newly established productive infection expands (panel 3) until it is cleared by antiviral therapy andor the patient's immunologic responses. However, during the productive infection a whole new set of latently infected cells have been established (panel 4). Although such techniques as virus isolation, immunohistochemical staining, and serum PCR can distinguish between latent and productive infections, qualitative PCR such as that used by Wilbom et all cannot. Thus, the appearance of a new positive PCR reaction with