High Risk of Chronic Graft-Versus-Host Disease in Unmanipulated Allogeneic Peripheral Blood Stem Cell Transplantation

To the Editor:

The recent report of Storek et al1 on the high risk of chronic graft-versus-host disease (cGVHD) associated with allogeneic peripheral blood stem cell transplantation (PBSCT) has prompted us to present data that support their findings and add to their observations on acute graft-versus-host disease (aGVHD) and late infections.

Between March 1993 and May 1996, 31 patients received allogeneic bone marrow transplantation (BMT), whereas 25 patients received PBSCT between June 1996 and November 1997, as primary and unmanipulated transplantations. Except for 2 patients (1 with aplastic anemia and 1 with paroxysmal nocturnal hemoglobinuria), all of the recipients were suffering from hematological malignancies. All donors were HLA-identical siblings. Demographic and clinical characteristics of the patients and the donors are given in Table1. The BMT and PBSCT groups were comparable with respect to patient age and sex, donor age and sex, diagnosis, stage of the disease at transplantation, and cytomegalovirus (CMV) serology. The percentage of patients with advanced malignancy was relatively low in both groups. As shown in Table 1, the preparative regimens were also quite similar. All of the patients received the same peritransplant infection prophylaxis. In both groups, GVHD prophylaxis consisted of cyclosporin A and short-term methotrexate. The diagnosis and grading of aGVHD were based on clinical evaluation and histologic confirmation.2 For grade II to IV aGVHD, methylprednisolone (MPR) was administered at a starting dose of 2 mg/kg/d and then was increased (up to 10 mg/kg/d) or decreased, depending on the clinical response. Patients who did not respond to high-dose MPR were considered steroid-resistant and received monoclonal antibody treatment of OKT-3 at a dose of 5 mg/d for 10 days. Intravenous Ig (IVIg) was not used on a routine basis, but patients who received OKT-3 or gancyclovir and who had cGVHD and hypogammaglobulinemia received IVIg at recommended doses. Patients who manifest CMV antigenemia during follow-up received gancyclovir as a preemptive treatment. Routine antibiotic prophylaxis was performed for 6 months after transplantation based on standard clinical practice. Clinical cGVHD was diagnosed and graded according to Seattle criteria.3 PBSCs were collected from healthy donors with filgrastim (16 μg/kg/d subcutaneously for 4 or 5 days). The recipients of PBSCT received a median of 5.13 × 10⁶(range, 1.1 to 18.28 × 10⁶). CD34⁺ cells/kg and a median of 1.5 × 10⁸ (range, 0.57 to 5.46 × 10⁸) CD3⁺ cells/kg. The median follow-up for patients receiving PBSCT and BMT was 297 days (range, 30 to 644 days) and 730 days (range, 30 to 1,815 days), respectively.

PBSCT recipients had relatively earlier neutrophil and platelet engraftment (median, 14.5 and 15 days, respectively) than BMT recipients (median, 25 and 41 days, respectively) (P < .05). Twelve patients (48%) in the PBSCT group and 19 patients (63%) in the BMT group developed aGVHD (P > .05). Eight patients in the PBSCT group and 7 patients in the BMT group had grade II to IV aGVHD, whereas the distribution of aGVHD severity was similar among the groups (P > .05). None of the BMT recipients with aGVHD showed steroid resistance, whereas 6 PBSCT recipients did (P < .01). In cases who survived for more than 40 days, CMV infection was documented at similar rates in both groups (6/22 for PBSCT and 7/29 for BMT; P > .05).

Clinical cGVHD was evaluated among the patients who were alive and relapse-free at day 100. Four of 27 (14.8%) marrow recipients versus 12 of 19 (63.16%) PBSCT recipients demonstrated clinical cGVHD (P < .01). The cumulative incidence of clinical cGVHD at the end of 1 year among the recipients surviving beyond 100 days was 65.7% for PBSCT and 7.4% for BMT. Except for the source of stem cells, none of the potential predictors of cGVHD,4(patient’s age, pretransplant recipient and donor CMV serology, and multiparity of the donor) were found to be associated with clinical cGVHD risk in a multivariable Cox regression model. In the univariable proportional hazards regression model, patients in the PBSCT group had a 2.3 times higher relative risk of developing clinical cGVHD by 1 year compared with the BMT recipients (95% confidence interval, 2.17 to 2.51; P < .001). The evaluation of the other risk factors in the univariable model showed that patients with multiparous donors also had a somewhat higher risk of experiencing cGVHD (P < .03). Clinically extensive cGVHD was observed in 2 marrow and 12 PBSC recipients. All cGVHD patients in the BMT group also experienced prior aGVHD, whereas 7 of 12 cases in the PBSCT group developed de novo cGVHD. Distribution of the organ involvement for cGVHD was fairly comparable in both groups, except for visceral organ involvement, which was higher in the PBSCT group.

Upon evaluation of the late infections occurring between 100 days and 1 year posttransplantation, the incidence (total number of infections per 1,000 patient days) of late infections among the 17 evaluable PBSCT recipients was 3.5 times higher than that in 27 evaluable BMT patients. Interestingly, the influence of cGVHD on the incidence of the late infections was apparent only in the PBSCT group (P < .01).

During immunosuppressive treatment for cGVHD, 1 of 2 BMT and 2 of 11 PBSCT recipients with extensive cGVHD died of infections. One patient who received PBSCT succumbed to liver failure due to cGVHD. Upon response to immunosuppressive drugs, treatment of 4 of 8 surviving PBSCT recipients with extensive cGvHD was stopped, whereas the remaining 4 were still under immunosuppressive therapy.

Briefly, our observations also support the recent Seattle data with an increased incidence of cGVHD in PBSCT.1 The rather low incidence of cGVHD in our BMT group compared with the generally reported figures1,4 7 may be attributed to the younger age of the recipients (median age, 31 years) as well as to that of the donors (median age, 25 years). The fact that we had fewer multiparous donors might also be a contributing factor. However, what we consider worth noting is that, despite the similarity of the two transplant groups with respect to the characteristics just mentioned above or to the others presented in Table 1, the incidence of cGVHD after PBSCT is significantly higher.

As it has been previously reported,5 6 our PBSCT recipients achieved rapid hematologic recovery without increased risk of aGVHD. However, in our series, the PBSCT patients seemed to have steroid-resistant aGVHD as well as de novo cGVHD more frequently, with the latter always being extensive. Prolonged immunosuppression either related to aGVHD treatment by OKT-3 and/or cGVHD might have contributed to the increased incidence of late infections in PBSCT recipients. Further follow-up is required to determine whether the increased incidence and relative severity of cGVHD will affect disease-free and/or overall survival.