High-dose therapy with autologous peripheral blood stem cell (PBSC) rescue is widely used for the treatment of malignant disease. CD34 selection of PBSC has been applied as a means of reducing contamination of the graft. Although CD34 selection results in a 2 to 3 log reduction in contaminating tumor cells without significantly delaying engraftment, many other types of cells are depleted from the CD34-enriched grafts and immune reconstitution may be impaired. In the present study, 31 cytomegalovirus (CMV)-seropositive patients who received myeloablative therapy followed by the infusion of CD34-selected autologous PBSC were assessed for the development of CMV disease in the first 100 days posttransplant. Seven patients (22.6%) developed CMV disease and 4 patients (12.9%) died from complications of their infection. In a contemporaneous group of 237 CMV-seropositive patients receiving unselected, autologous PBSC, only 10 patients (4.2%) developed CMV disease, with 5 deaths (2.1%). In a multivariate logistic regression analysis, the use of CD34-selected autologous PBSC after high-dose therapy was associated with a marked increase in the incidence of CMV disease and CMV-associated deaths.

CD34 selection of peripheral hematopoietic blood stem cell products has been applied to reduce the risk of relapse after an autologous transplantation. However, CD34 selection is also associated with a significant reduction in T-cells, natural killer cells, and monocytes, and these reductions may influence immune reconstitution and thus increase the risk for infections. An increased incidence of cytomegalovirus (CMV) disease in patients receiving CD34-selected transplants has been reported. In this study, the incidence rate of infections other than CMV is reported in 32 patients who underwent myeloablative therapy followed by the infusion of CD34-selected autologous peripheral blood stem cells (PBSC) and compared to the rate in a contemporaneous group of 273 patients who received unselected autologous PBSC during the same time period. Infection surveillance and prevention strategies were identical between the 2 groups. More non-CMV infections occurred in the recipients of CD34-selected PBSC than in recipients of unselected PBSC (78% versus 30%, P < .0001). The differences in the rates of viral infections were mainly due to dermatomal and disseminated varicella-zoster virus (VZV) (any VZV, 26% versus 4%, P = .002; disseminated VZV, 11% versus 0.3%, P = .03) and parainfluenza 3 virus infections (13% versus 3%, P = .04). Bacterial infections were also more common among CD34-selected PBSC transplant recipients (34% versus 16%, P = .01), whereas fungal infections were not significantly different between the groups. In multivariable logistic regression models, the effect of CD34 selection on infection risk remained significant for viral infections and overall non-CMV infections. Infection-related mortality was not significantly different between the groups. In conclusion, the incidence of viral and bacterial infections appears to be increased in recipients of CD34-selected autologous PBSC transplants. Because the risk for infections approaches that seen in allogeneic transplant recipients, infection surveillance, diagnostic work-up, and prevention strategies similar to those used in allogeneic recipients are warranted.