Unmanipulated haploidentical bone marrow transplantation and post-transplant cyclophosphamide for hematologic malignanices following a myeloablative conditioning: an update.

This is a report of 148 patients with hematologic malignancies who received an unmanipulated haploidentical bone marrow transplant (BMT), followed by post-transplant high-dose cyclophosphamide (PT-CY). All patients received a myeloablative conditioning consisting of thiotepa, busulfan, fludarabine (n=92) or TBI, fludarabine (n=56). The median age was 47 years (17-74); 47 patients were in first remission (CR1), 37 in second remission (CR2) and 64 had an active disease; all patients were first grafts. The diagnosis was acute leukemia (n=75), myelodisplastic syndrome (n=24), myelofibrosis (n=16), high-grade lymphoma (n=15) and others (n=18). GVHD prophylaxis consisted in PT-CY on days +3 and +5, cyclosporine (from day 0), and mycophenolate (from day +1). The median day for neutrophil engraftment was day +18 (13-32). The cumulative incidence of grades II-IV acute GVHD was 24%, and of grades III-IV GVHD 10%. The incidence of moderate-severe chronic GVHD was 12%. With a median follow-up for the surviving patients of 313 days (100-1162), the cumulative incidence of transplant-related mortality (TRM) is 13%, and the relapse-related death is 23%. The actuarial 22 months overall survival is 77% for CR1 patients, 49% for CR2 patients and 38% for patients grafted in relapse (P<0.001). Major causes of death were relapse (22%), GVHD (2%) and infections (6%). We confirm our initial results, suggesting that a myeloablative conditioning regimen followed by unmanipulated haploidentical BMT with PT-CY, results in a low risk of acute and chronic GVHD and encouraging rates of TRM and overall survival, also for patients with active disease at the time of transplant.

DLI after haploidentical BMT with post-transplant CY.

Forty-two patients relapsing after an unmanipulated haploidentical BM transplant and post-transplant CY (PT-CY), were given 108 DLI, with median interval from transplant of 266 days (range, 67-1372). DLI were given at escalating doses, expressed as CD3+ cells/kg, without GVHD prophylaxis, and ranged from 1 × 10(3) to 1 × 10(7) cells/kg (median 5 × 10(5) cells/kg). The average number of DLI per patient was 2.6 (range, 1-6). The diagnosis was leukemias (n=32) grafted with a myeloablative regimen and Hodgkin's disease (n=10), grafted with a nonmyeloablative regimen. Leukemic patients with molecular relapse (n=20), received DLI alone (n=17) or in association with azacytidine (n=3); leukemic patients with hematologic relapse (n=12) received chemotherapy followed by DLI (n=11) or DLI alone (n=1); Hodgkin patients received DLI following 1-3 courses of chemotherapy. In these three groups the incidence of acute GVHD II-III was 15%, 17% and 10%; response rate was 45%, 33% and 70%; 2-year actuarial survival was 43%, 19% and 80% respectively. This study confirms that escalating doses of DLI can be given in the haploidentical setting with PT-CY, with a relatively low risk of acute GVHD. Response rates and survival are dependent on the underlying disease.

Risk factors for invasive aspergillosis and related mortality in recipients of allogeneic SCT from alternative donors: an analysis of 306 patients.

Invasive aspergillosis (IA) is a serious complication in patients undergoing allogeneic haematopoietic stem cell transplantation (HSCT), particularly from donors other than HLA-identical sibling. All 306 patients who underwent alternative donor HSCT between 01 January 1999 and 31 December 2006 were studied. Late IA was defined as occurring >or=40 days after HSCT. The median follow-up was 284 days (range, 1-2709). Donors were matched unrelated (n=185), mismatched related (n=69), mismatched unrelated (n=35) and unrelated cord blood (n=17). According to European Organization for Research and Treatment of Cancer/Mycoses Study Group criteria, 2 patients already had IA at HSCT, 23 had early IA and 20 had late IA (IA incidence 15%). Eight patients had proven and 37 probable IA. Multivariate analyses showed that significant predictors of IA were delayed neutrophil engraftment, extensive chronic GVHD (cGVHD), secondary neutropenia and relapse after transplant. Early IA was associated with active malignancy at HSCT, CMV reactivation and delayed lymphocyte engraftment. Late IA was predicted by cGVHD, steroid therapy, secondary neutropenia and relapse after HSCT. IA-related mortality among IA patients was 67% and was influenced by use of anti-thymocyte globulin, steroids, higher levels of creatinine, and lower levels of IgA and platelets. The outcome of IA depends on the severity of immunodeficiency and the status of the underlying disease.

Extracorporeal photopheresis for the treatment of steroid refractory acute GVHD.

Extracorporeal photopheresis (ECP) was given to 23 patients with steroid-refractory acute GVHD (aGVHD, grade II (n=10), III (n=7) or IV (n=6)). The median duration of ECP was 7 months (1-33) and the median number of ECP cycles in each patient was 10. Twelve patients (52%) had complete responses. Eleven patients (48%) survived and 12 died, 10 of GVHD with or without infections and two of leukaemia relapse. The average grade of GVHD was reduced from 2.8 (on the first day of ECP) to 1.4 (on day +90 from ECP) (P=0.08), and the average dose of i.v. methylprednisolone from 2.17 to 0.2 mg/kg/d (P=0.004). Complete responses were obtained in 70, 42 and 0% of patients, respectively, with grades II, III and IV aGVHD; complete responses in the skin, liver and gut were 66, 27 and 40%. Patients treated within 35 days from onset of aGVHD had higher responses (83 vs 47%; P=0.1). A trend for improved survival was seen in grade III-IV aGVHD treated with ECP as compared to matched controls (38 vs 16%; P 0.08). ECP is a treatment option for patients with steroid refractory aGVHD and should be considered early in the course of the disease.

Allogeneic hemopoietic stem cell transplants for patients with relapsed acute leukemia: long-term outcome.

We assessed the long-term outcome of patients with relapsed acute myeloid (n=86) or acute lymphoid leukemia (n=66), undergoing an allogeneic hemopoietic stem cell transplantation in our unit. The median blast count in the marrow was 30%. Conditioning regimen included total body irradiation (TBI) (10-12 Gy) in 115 patients. The donor was a matched donor (n=132) or a family mismatched donor (n=20). Twenty-two patients (15%) survive disease free, with a median follow-up of 14 years: 18 are off medications. The cumulative incidence of transplant related mortality is 40% and the cumulative incidence of relapse related death (RRD) is 45%. In multivariate analysis of survival, favorable predictors were chronic graft-versus-host disease (GvHD) (P=0.0003), donor other than family mismatched (P=0.02), donor age less than 34 years (P=0.02) and blast count less than 30% (P=0.07). Patients with all four favorable predictors had a 54% survival. In multivariate analysis of relapse, protective variables were the use of TBI (P=0.005) and cGvHD (P=0.01). This study confirms that a fraction of relapsed leukemias is cured with an allogeneic transplant: selection of patients with a blast count <30%, identification of young, human leukocyte antigen-matched donors and the use of total body radiation may significantly improve the outcome.

Factors predicting response and graft-versus-host disease after donor lymphocyte infusions: a study on 593 infusions.

In the present study, we analyze factors predicting graft-versus-host disease (GvHD) and response after donor lymphocyte infusions (DLI). A total of 100 patients received 593 DLI between June 1990 and December 2000 in a bulk dose (n=14) or in escalating dose infusions (n=86). Patients were analyzed after stratification for type of relapse: (1). molecular relapse (n=6), (2). cytogenetic relapse (n=20), (3). chronic phase of chronic myeloid leukemia (CML) or complete remission of other disease post chemotherapy (n=24), (4). CML in accelerated/blastic phase (n=14), (5). resistant disease not responding to chemotherapy (n=36). The proportion of responders to DLI in these five groups was 100, 90, 75, 36 and 0% (P<0.0001). Factors predicting response by multivariate analysis were type of relapse (P<0.0001), post-DLI GvHD (P=0.005), pancytopenia (P=0.008), and a diagnosis of CML (P=0.04). Acute GvHD (grades II-IV) occurred in 21 patients (21%), and correlated in multivariate analysis with pancytopenia and less than four DLI. Other predictors of GvHD were the number of CD3+cells/infusion and serum levels of gamma-glutamyl transferase (gammaGT). The actuarial probability of treatment-related mortality was 9% for HLA identical siblings and 44% for alternative donor transplants (P=0.006). Response to DLI is predicted by tumor burden and is associated with GvHD and pancytopenia.

Allogeneic bone marrow transplantation (BMT) for adults with acute lymphoblastic leukemia (ALL): predictive role of minimal residual disease monitoring on relapse.

We developed a PCR-based method to monitor clonogenic IgH VDJ rearrangement as a possible predictor of relapse in patients with acute B-ALL after allogeneic bone marrow transplantation (BMT). We studied 23 patients at diagnosis, before and after BMT. At the time of BMT, 13 patients were in first complete remission, eight in second complete remission and two in relapse. Four patients were PCR negative before BMT and remained PCR negative also after BMT (-/- pattern). They are still in remission after a median follow-up of 41 months. Nineteen patients were MRD-positive before BMT: three were PCR negative at first determination after BMT (+/- pattern) and maintain remission. Sixteen patients were PCR-positive at first determination after BMT (+/+ pattern): five became PCR negative (+/+/- pattern) (four with chronic graft-versus-host disease (GVHD) and two after donor lymphocyte infusions (DLI)). Nine patients remained PCR-positive (+/+/+ pattern) (four remain in remission, and six relapsed); two patients died before transplant. In conclusion, PCR negative patients before BMT remained negative post-BMT; many pre-BMT positive patients had initial MRD positivity after BMT: 37% of them achieved a molecular remission with cGVHD or DLI.

Prophylactic antithymocyte globulin reduces the risk of chronic graft-versus-host disease in alternative-donor bone marrow transplants.

We studied the impact of preparative regimens with or without antithymocyte globulin (ATG) on chronic GVHD in 160 patients undergoing marrow transplants from unrelated donors (n = 127) or partially mismatched related donors (n = 33). A conditioning regimen that included rabbit ATG, 7.5 to 15 mg/kg (Thymoglobuline; Sangstat, Lyon, France), was given to 102 patients, whereas a conditioning regimen without ATG was given to 58 patients. The median patient age was 34 years for the ATG group and 29 years for the non-ATG group (P = .002); otherwise the 2 groups were matched for disease phase, diagnosis, donor age, interval from diagnosis to transplantation, and number of cells infused at the time of transplant. Median follow-up for surviving patients was 4.5 years (range, l.5-9 years). The conditioning regimen was cyclophosphamide (CY) and total body irradiation (TBI) in 95 patients and CY-thiotepa in 65 patients; the source of stem cells was bone marrow for all patients. Acute GVHD grades II-IV and grades III-IV were reduced in patients receiving ATG compared to patients not receiving ATG (51% versus 74%, P = .004 and 14% versus 28%, P = .03, respectively). There were significantly fewer patients with chronic GVHD in the ATG group than in the non-ATG group at 6 months (14% versus 30%, P = .03), 1 year (7% versus 41%, P = .0001), 2 years (16% versus 36%, P = .02), and 4 years (5% versus 34%, P = .002) and beyond 4 years (0% in 19 patients at risk versus 29% in 24 patients at risk, P = .01). More patients in the ATG group than in the non-ATG group had a performance status (Karnowski score) greater than 90 at last follow-up (93% versus 56%, P = .01) and had discontinued cyclosporin treatment 2 years posttransplant (28% versus 3%, P = .003). Survival rates were comparable in the ATG and non-ATG groups for patients who received TBI (56% versus 59%, P = .7) and those who received thiotepa (33% versus 18%, P = .3). Transplant mortality and relapse rates were also comparable in the 2 groups for these patients. We conclude that pretransplant ATG administration reduces the risk of acute and chronic GVHD, improves quality of life, and increases the likelihood that discontinuation of immunosuppressive therapy will be possible.

Conventional hematopoietic stem cell transplants from identical or alternative donors are feasible in recipients relapsing after an autograft.

BACKGROUND AND OBJECTIVES: The risk of relapse after autologous bone marrow transplantation (ASCT) is high and is related to the type of malignancy and phase of the disease. The outcome for the patient who relapses after an autologous transplant is poor. Some of these patients achieve a remission with conventional chemotherapy, but it is usually short-lasting. Most of them succumb to the original disease. One further therapeutic possibility is an allogeneic transplant which would confer the potential advantage of a graft-versus-leukemia effect in addition to the lack of tumor contamination of the graft and to a high-dose intensity conditioning regimen. DESIGN AND METHODS: We have studied the outcome of 31 patients with hematologic malignancies who underwent an allogeneic hematopoietic stem cell transplant (HSCT) after failing an autologous transplant because of relapse (n=29) or persistent aplasia (n=2). The median age at allograft was 36 years (18-55) and the interval from autograft to allograft was 21 months (3-141). The source of stem-cells was unmanipulated bone marrow (n=26) or growth-factor-mobilized peripheral blood (n=5). The donor was an HLA-identical sibling (n=7), or an alternative donor (n=24) (family mismatched n=11, or matched unrelated n=13). The conditioning regimen was cyclophosphamide and thiotepa (n=22), or cyclophosphamide and total body irradiation (n=9) Graft-versus-host disease (GvHD) prophylaxis consisted of cyclosporine (CyA) + methotrexate (MTX). RESULTS: Acute GvHD was scored as 0-I, II, or III-IV in 39%, 48%, and 13% of the patients, respectively. Sixteen patients died of transplant-related complications and one of progressive disease. With a median follow-up of 220 days (9-2104) the actuarial 2-year transplant-related mortality (TRM) was 51%, the actuarial relapse risk 37%, the actuarial survival 46%. Fifteen patients (48%) are alive in complete remission, with a median follow-up of 32 months (range 2-71). INTERPRETATION AND CONCLUSIONS: These data suggest that patients relapsing after an autotransplant should be screened for potential related or unrelated donors: although TRM remains high there is a definite chance of long-term disease-free survival if these patients are allografted.

Factors influencing haematological recovery after allogeneic haemopoietic stem cell transplants: graft-versus-host disease, donor type, cytomegalovirus infections and cell dose.

Platelet recovery after allogeneic haemopoietic stem cell transplant (HSCT) and predictive factors were analysed in 342 patients with haematological malignancies. All patients were prepared with cyclophosphamide plus total body irradiation, and received an unmanipulated HSCT from an HLA-identical sibling (n = 270), a matched unrelated donor (n = 67) or an identical twin (n = 5). The source of stem cells was peripheral blood (n = 15) or bone marrow (n = 327). Graft-vs.-host disease (GvHD) prophylaxis consisted of cyclosporin A with or without methotrexate. The proportion of patients with < 50 x 10(9)/l platelets on d +50, d +100, d +200 and d +365 after HSCT was 26%, 27%, 14% and 11% respectively. Thrombocytopenia was independent of the degree of complete donor chimaerism. Four variables were predictive of platelet recovery: donor type, acute GvHD, cytomegalovirus (CMV) infection and number of cells infused at transplant. Recipients of an unrelated graft had lower platelet counts (49 x 10(9)/l) on d +50 than identical sibling grafts (10(8) x 10(9)/l) (P < 0.001) and twin grafts (149 x 10(9)/l) (P < 0.001). Patients with GvHD grades 0, I, II, III and IV had significantly different platelet counts on d +50 (153 x 10(9)/l, 102 x 10(9)/l, 85 x 10(9)/l, 32 x 10(9)/l and 22 x 10(9)/l; P < 0.001) and thereafter. Thrombocytopenia was more frequent in patients with high-level CMV antigenaemia (> four positive cells/2 x 105) (P < 0.0001) and in patients who received a low cell dose at transplant (< or = 4.1 x 10(8)/kg) (P = 0.009). Platelet counts predicted transplant-related mortality (TRM) and were higher at all time intervals in patients surviving the transplant. Patients with grade II GvHD and > 50 x 10(9)/l platelets had a lower TRM than patients with grade II GvHD and < or = 50 x 10(9)/l platelets (14% vs. 40%, P < 0.0001). In conclusion, (i) a significant proportion of allogeneic HSCT recipients are thrombocytopenic long-term, irrespective of complete donor chimaerism, (ii) thrombocytopenia identifies patients at greater risk of lethal complications, and (iii) platelet recovery is influenced by GvHD, donor type, CMV infections and cell dose, not by stem cell source or other patient-disease-related variables.

Pre-emptive therapy of acute graft-versus-host disease: a pilot study with antithymocyte globulin (ATG).

We have previously shown that patients at high risk of graft-versus-host disease (GVHD) and transplant-related mortality (TRM) can be identified on day +7 following an allogeneic bone marrow transplant (BMT), based on serum bilirubin and blood urea nitrogen levels. One possible approach to reduce the risk of GVHD and TRM, is pre-emptive treatment with T cell antibodies. We report a pilot study testing the feasibility of this approach in 18 high risk patients, with a median age of 41, 83% of whom had advanced disease, undergoing an alternative donor BMT (family mismatched in five and unrelated in 13). The patients received three doses of rabbit antithymocyte globulin (ATG) (Thymoglobuline; Sangstat) 1.25 mg/kg on alternate days, starting at a median interval of 11 days (range 7-13) after BMT. Controls were 20 historical unrelated donor transplants (median age 35, 63% with advanced disease), with a high score from our original publication in 1999. The actuarial 1 year TRM of the ATG-treated patients was 40% compared to 60% for untreated controls (P = 0.06). Severe grade III-IV aGVHD developed in 27% of the ATG-treated patients, and in 55% of the controls (P = 0.08). This study indicates that early pre-emptive treatment of aGVHD in day +7 high risk patients is feasible and may lead to a reduction of aGVHD and TRM. This approach is being tested in a prospective randomized trial.

Serum cholinesterase is an early and sensitive marker of graft-versus host-disease (GVHD) and transplant-related mortality (TRM).

Serum cholinesterase (CHE) has been reported to be a significant indicator of liver function and prognosis in patients with cirrhosis. On the other hand, liver complications are frequent following allogeneic stem cell transplantation (HSCT). We therefore tested whether CHE was predictive of graft-versus-host disease and outcome in HSCT recipients. We studied 689 patients receiving a HSCT from an HLA-identical sibling (SIB) (n = 511), an alternative donor (n = 173) or a syngeneic twin (n = 5). Acute graft-versus-host disease (GVHD) was scored as 0-I, II, III-IV in 325 (47%), 279 (41%), and 85 patients (12%) respectively; 190 (28%) patients died of transplant-related complications (TRM). On day -7 the median CHE serum level was comparable in patients who either survived or died of TRM (5900 IU/l). On day 0, serum CHE levels were respectively 2310 and 2120 IU/l (P = NS) indicating the impact of the conditioning regimen. On day +7 after HSCT, the median level for surviving patients was 2598 IU/l vs 2309 IU/l for patients who subsequently died (P = 0.0002), on day +21 CHE levels were respectively 3348 vs 2528 IU/l (P < 0.00001), on day +50, 3575 vs 2358 IU/l (P < 0.00001) and on day +100 4193 vs 2729 IU/l (P < 0.00001). CHE levels on day +50 strongly correlated with aGVHD (3803 vs 3070 vs 1933 IU/l for patients with GVHD grade 0-I, II, and III-IV, respectively (P < 0.00001) and relapse (3569 for patients relapsing vs 3115 IU/l for patients not relapsing, P = 0.0006). In conclusion, (1) serum cholinesterase is a simple and reliable marker of acute GVHD and transplant-related complications; and (2) high CHE levels on day +50 predict relapse. If confirmed, the latter patients may be eligible for early reduction of immunosuppressive therapy.

Quality of life in 244 recipients of allogeneic bone marrow transplantation.

The number of long-term survivors after allogeneic bone marrow transplantation (BMT) has been increasing over the past years, and quality of life (QOL) has become an important end-point. We studied 244 patients undergoing an allogeneic BMT to identify factors and events influencing psychosocial outcome. Patients enrolled received the Psychosocial Adjustment to Illness Scale (PAIS) questionnaire assessing psychological and social adjustment to chronic illness or its sequelae. Eighty-two per cent of patients had a haematological disease. The median age was 28 years at BMT, and the median follow-up was 61 months. The median overall PAIS score for all patients was 56 (range 22-76): 25% (n = 61) of patients were considered to have a good QOL ( 75 percentile score). Factors associated with a poor QOL in multivariate analysis were: patients' age at BMT (> 25 years, P < 0.01); presence of long-term sequelae (P < 0.01); chronic graft-versus-host disease (GVHD) (P < 0.05); and a short interval from BMT (< 5 years; P < 0.05). The QOL improved with time: 12% of patients reported a good QOL within 5 years compared with 38% after this time point and, conversely, 38% reported a poor QOL within 5 years compared with 24% after this time point (P < 0. 0001). Older patients had significantly poorer QOL compared with younger patients (< or = 25 years; P = 0.01). Females had significantly poorer scores when compared with males in the sexual (P < 0.0001) and psychological domains (P = 0.001). The data suggest that (i) one-third of patients undergoing allogeneic BMT report a poor QOL; (ii) factors associated with poor QOL are older age, presence of long-term sequelae, chronic GVHD and short follow-up; (iii) QOL is superior in long-term survivors; and (iv) BMT affects different aspects of life in males and females. A longitudinal study is ongoing to prove the effect of time on quality of life.

Reduced intensity thiotepa-cyclophosphamide conditioning for allogeneic haemopoietic stem cell transplants (HSCT) in patients up to 60 years of age.

UNLABELLED: Transplant-related mortality (TRM) remains a major problem in older patients undergoing allogeneic haemopoietic stem cell transplants (HSCTs). We have therefore explored a less intensive conditioning in 33 patients with a median age of 52 years (range 43-60) transplanted from human leucocyte antigen (HLA)-identical siblings. The underlying disease was chronic myeloid leukaemia (n = 15), acute myeloid leukaemia (n = 6), myelodysplasia (n = 7) or a chronic lymphoproliferative disorder (n = 5); 15 patients (45%) had advanced disease. The regimen consisted of thiotepa (THIO; 10 mg/kg) on day -5 and cyclophosphamide (CY; 50 mg/kg) on days -3 and -2 (total dose 100 mg/kg). The source was bone marrow (BM) (n = 17) or granulocyte colony-stimulating factor (G-CSF)-mobilized peripheral blood (PB) (n = 16), which were infused without manipulation. Graft-versus-host disease (GVHD) prophylaxis consisted of cyclosporin A (CyA) and a short course of methotrexate. Mean time to achieve a neutrophil count of 0.5 x 109/l was 17 d (range 11-23) and full donor chimaerism was detected in 79% of patients by day 100. Acute GVHD grade III or IV occurred in 3% of patients. Chronic GVHD was seen in 45% of patients, with a significant difference for PB (69%) compared with BM transplants (23%) (P = 0.009). For BM grafts, the actuarial 2-year TRM was 6%, the relapse 56% and survival 87%; for PB grafts, these figures were, respectively, 27%, 33% and 68%. Twenty-five patients are alive at a median follow-up of 762 d (range 216-1615) and 20 patients (60%) remain free of disease. Thirteen patients (39%) received donor lymphocyte infusion (DLI) either for persisting or relapsing disease and six patients had complete remission. IN CONCLUSION: (i) patients up to the age of 60 years can be allografted with reduced intensity conditioning; (ii) the procedure was associated with a low transplant-related mortality, particularly for bone marrow grafts, because of a lower risk of chronic GVHD; and (iii) DLI were required after transplant in half the patients for persisting disease or relapse.

Foscarnet prophylaxis of cytomegalovirus infections in patients undergoing allogeneic bone marrow transplantation (BMT): a dose-finding study.

This is a dose-finding study using foscarnet for CMV prophylaxis after allogeneic bone marrow transplantation (BMT) in 20 high risk patients (unrelated donors, or T cell depleted, and/or advanced disease). Foscarnet was started on day +1 after BMT and continued until day +100. We explored four different dose levels, patients being entered at the lowest dose level until one patient experiences CMV-reactivation, identified as two consecutive positive CMV antigenemias (CMVAg-emia). The four dose levels expressed as mg/kg/day between days 1 and 30 (induction) and between days 31 and 100 (maintenance) were respectively: dose level I = 60/30 (n = 5); dose level II = 120/60 (n = 4); dose level III = 120/90 (n = 5) and dose level IV = 120/120 (n = 6). All patients showed engraftment: PMN > or =0.5 x 109/l at a median interval of 16, 21, 17, 15 days after BMT, and Plt > or =30x10(9)/l on days 19, 16, 17, 17 respectively. CMVAg-emia was seen in 10 patients at a median interval of 53 days post-BMT (range 33-89) with a median of 10 CMV antigen+ cells (range 1-16). There was a dose effect of foscarnet on CMVAg-emia: respectively 4/5 patients (80%), 2/4 (50%), 3/5 (60%) and 1/6 (18%) at dose levels I, II, III, IV (P = 0.1). CMV disease was seen in 3/9 (33%) at dose levels I, II and 0/11 at dose levels III, IV (P = 0. 07). The median number of CMV antigen-positive cells at diagnosis of CMV infection was different: 13 in dose levels I-II and two in dose levels III-IV (P = 0.01). Increased creatininine was seen in 15 patients with a mean of 1.8 mg% (range 1.5-5.7) and was the cause of discontinuation in nine patients (45%). Renal toxicity was reversible in all nine patients. Overall actuarial TRM at 2 years was 31%: 47% for patients at dose levels I-II and 19% for patients at dose levels III-IV. In conclusion, foscarnet exhibits a dose-dependent prophylactic effect on CMVAg-emia, CMV disease and transplant-related mortality with acceptable and reversible renal toxicity.