Study Design

Patients and donors in the study provided written informed consent that was approved by the institutional review board of the University of Arkansas for Medical Sciences. The conditioning regimen for the recipients consisted of intravenous MEL 100 mg/m² over 20 minutes on day −1 for those receiving a graft from siblings, whereas recipients of unrelated donor grafts received the same dose of MEL plus 250 cGy of total-body irradiation in two fractions on day −2 and fludarabine 30 mg/m² on days −2 and −1. The outcome of the first 16 patients on the protocol has been reported previously.¹⁸

Cyclosporine was given intravenously (3 mg/kg/d) to all patients starting on day −1, with a target level of 300 ng/μL. Patients receiving unrelated donor grafts received in addition methylprednisolone 1 mg/kg/d starting on day 5 that was tapered and completely stopped by day 29 if there was no evidence of GVHD. No patient received methotrexate for graft-versus-host disease (GVHD) prophylaxis.

HLA-matched siblings were mobilized with granulocyte colony-stimulating factor (G-CSF) 10 μg/kg/d. On day 4 of G-CSF, large-volume leukapheresis (> 15 L) was performed. Mobilized peripheral-blood stem cells were also collected on days 5, 6, and 7 if necessary to attain 10 × 10⁶ CD34⁺ cells/kg of the patient’s weight. Unrelated donors had a choice between receiving G-CSF followed by leukapheresis or undergoing bone marrow harvest. On day 0, unmanipulated stem cells were infused, except in two recipients of unrelated donor grafts, where T-cell depletion was performed using CD34⁺ cell selection (Isolex; Nexell Therapeutics, Inc, Irvine, CA). The graft target was a maximum stem cell infusion of 5 × 10⁶ CD34⁺ cells/kg; the rest was stored for later use as DLI.

In the absence of GVHD, cyclosporine was tapered over 1 month starting on day 60 after mini-allograft. DLIs with escalating doses of donor CD3⁺ lymphocytes were scheduled on days 21, 42, and 112 to establish full chimeric engraftment in patients with no GVHD. Initially, patients received all scheduled DLIs. Later in the study, because of a high incidence of GVHD, cyclosporine was continued until day 120, and decisions regarding DLI were made on the basis of the chimeric and disease status of the patients. Those with mixed chimerism received additional donor cells while on cyclosporine in the absence of acute GVHD. Patients with residual or progressive disease had cyclosporine discontinued, and in the absence of GVHD, they received additional DLI with no GVHD prophylaxis. Patients who relapsed or had no response to DLI received salvage chemotherapy followed by additional DLI.

Antimicrobial prophylaxis consisted of acyclovir 400 mg every 8 hours from day 1 to 6 months after transplant, and itraconazole 400 mg/d until the CD4⁺ count exceeded 400/μL. Patients received levofloxacin 500 mg/d from day 0 to 2 years after transplantation and cotrimoxazole bid twice a week from the time of neutrophil recovery to 6 months. Patients at high risk of fungal infection and those unable to tolerate oral itraconazole received intravenous liposomal amphotericin B 1 mg/kg/d. Blood samples were checked each week for cytomegalovirus (CMV) antigenemia; those who tested positive were treated preemptively with ganciclovir. Patients received intravenous immunoglobulin 0.5 g/kg if the immunoglobulin G level was less than 500 mg/dL.

Study End Points

The end points of the study were engraftment, chimerism, toxicity, TRM, incidence of GVHD, and MM response. Chronic GVHD was evaluated in patients with at least 3 months’ follow-up. The outcome of the mini-allograft patients was compared with historical patients who received standard allotransplants at the University of Arkansas.

Engraftment was defined as neutrophil recovery to more than 0.5 × 10⁹/L for 3 consecutive days and untransfused platelet count greater than 20 × 10⁹/L. The degree of donor-recipient chimerism was assessed by polymerase chain reaction assay of short tandem repeat loci from blood and marrow according to published methods.¹⁹ Studies for chimerism were assessed monthly for 1 year. Calculating the percentage of donor DNA in comparison with the pretransplant donor and recipient genotype quantitated chimerism. The assay is able to detect chimerism if more than 5% donor or host DNA is present. Mixed chimerism was defined as the presence of more than 5% donor and host-derived cells on more than one occasion in the whole blood. Acute GVHD was assessed according to the criteria of the International Bone Marrow Transplant Registry. Other toxicities were assessed using National Cancer Institute criteria. Chronic GVHD was defined as GVHD occurring 100 days or more after bone-marrow transplant and was graded as none, limited, or extensive.

CR required the disappearance of monoclonal gammopathy in serum and urine on immunofixation analysis and normal bone marrow aspirate and biopsy (< 5% plasmacytosis) on two successive occasions 2 months apart. Near CR (nCR) was defined as a CR with only positive immunofixation by electrophoresis. Partial remission (PR) required a bone marrow with 5% to 10% plasmacytosis and a 75% reduction in M-protein. Early TRM included any death within 100 days after mini-allograft. Late TRM was defined as any death after day 100 attributed to complications of the allograft. Relapse was defined as recurrence of monoclonal protein or bone marrow plasmacytosis in case of CR. Progressive disease for non-CR patients implied at least a 25% increase in M-protein and/or bone marrow plasmacytosis or the development of new extramedullary disease.

Data are presented as of March 15, 2001. Statistical analyses were performed using SAS Version 8.0 (SAS Institute, Inc, Cary, NC). Kruskal-Wallis and Fisher’s exact tests were applied to discern imbalances in patient characteristics in the mini-allograft versus standard allograft and outcome. The Cochran-Mantel-Haenszel test was used to detect trends in outcome according to number of prior ATs and disease status. The Bowker test of symmetry for matched-pair data was performed to determine whether patients’ disease status showed net improvement after the allograft. The log-rank test was applied to assess differences between Kaplan-Meier survival curves.

Difference in overall survival (OS) between mini-allografts and standard allografts was adjusted for clinical covariates using multivariate Cox regression. Covariates considered for multivariate regression were age, albumin less than 3.5 g/dL, β₂-microglobulin greater than 4 mg/L, creatinine greater than 2 mg/dL, C-reactive protein greater than 1 mg/dL, disease status at allografting, hemoglobin less than 10 g/dL, lactate dehydrogenase greater than 190 U/L, number of prior ATs, sex, time from diagnosis and from last AT, and cytogenetic abnormalities (including those involving chromosomes 11 and 13). Mini-allograft and standard allograft plus the above covariates were entered into the initial multivariate model for variable selection. The final multivariate model was restricted to significant covariates for OS in the initial model plus transplant mode (mini-allograft v standard allograft).

Patients

Patients’ characteristics before mini-allografts are listed in Table 1. The population included 31 MM patients who were ineligible for conventional myeloablative conditioning because of prior AT, age, and/or comorbidity. The patients ranged in age from 38 to 69 years (median, 56 years). Median time from diagnosis to mini-allograft was 29 months (range, 8 to 164 months). All but one patient had received either one prior AT (13 patients), two ATs (14 patients), or three prior ATs (three patients). The median time from the last AT to mini-allograft was 12 months (range, 2.7 to 66 months). Twenty-five patients had been treated with salvage thalidomide after relapse. At the time of allografting, 17 patients had progressive disease (PD) and 14 patients had chemosensitive disease. Of the chemosensitive patients, six were primary responders (two were in CR, two were in nCR, and two were in PR), whereas eight had responsive disease after relapse, with three patients in PR, four in CR, and one in nCR. All patients had at least one poor prognostic factor as defined by the presence of high β₂-microglobulin, complex cytogenetics, or relapse after prior AT. A complex cytogenetic profile was present in 30 patients; partial or complete deletion of chromosome 13 was present in 21. Cytogenetic and clinical evidence of myelodysplastic syndromes (MDSs) in addition to active myeloma was observed in four patients. β₂-Microglobulin before allograft was elevated (> 4 mg/L) in seven patients. Before mini-allograft, the median number of CD4⁺ lymphocytes was 209/μL (range, 35 to 630/μL).

Characteristics of the donors for the mini-allograft patients are summarized in Table 2. Of the 31 patients, 28 received a fully HLA-compatible graft as determined serologically for class I (A, B, and C loci) and molecularly for class II (DRB1 and DQB1 loci). Three patients received grafts mismatched at one class II HLA locus (two sibling donors) or one class I HLA locus (one unrelated donor). Twenty-eight patients received G-CSF–mobilized grafts, and three received marrow from unrelated donors.

Engraftment and Chimerism

Patients received a median of 4.8 × 10⁶ CD34⁺ cells/kg (range, 1.5 to 12.5) and 1.7 × 10⁸ CD3⁺ cells/kg (range, 0.3 to 5.1). Two patients had received T-cell–depleted grafts containing 1 × 10⁵ CD3⁺ cells/kg. G-CSF was not routinely administered after transplantation. Five patients received G-CSF by day 16 because of slow white blood cell count recovery. Neutrophil recovery greater than 500/μL was reached at a median of 14 days (range, 10 to 46 days); the median time to an unsupported platelet count greater than 20 × 10⁹/L was 15 days (range, 0 to 50+ days). Five patients did not require any platelet support; the median number of platelet transfusions was eight (range, zero to 70+). The median number of units of RBCs transfused was 5 (range, 2 to 15), 13 (range, 5 to 18), and 14 (range, 6 to 35) for patients with full ABO matched (n = 21), minor (n = 3), or major mismatched (n = 7) donors, respectively.

At 21 to 30 days after transplantation, 25 of 31 patients were full donor chimeras (Table 3). Two patients (including one who received an unrelated T-cell–depleted graft) failed to engraft even after an additional allogeneic stem cell infusion. Eventually, both patients received an autologous stem cell rescue; one achieved and is maintaining an nCR more than 1 year after the mini-allograft, whereas the other had progressive disease. Interestingly, four of seven patients with major ABO mismatches were mixed chimeras at day 30, including the two patients who rejected their allografts. By day 100, 25 (89%) of 28 assessable patients were full donor chimera.

Disease Response

The data on clinical response after mini-allograft are listed in Table 4. In these 31 poor-prognosis patients, 19 (61%) had an excellent response after mini-allograft, with a stringently defined CR achieved in 12 patients and nCR in seven (only immunofixation-positive). Three other patients achieved a PR.

None of the four patients with MM and MDS before mini-allograft had cytogenetic or clinical evidence of MDS or MM after allografting. Three are alive and disease free (at 9+, 15+, and 36+ months), and one died of pneumonia on day 400 in CR.

Not unexpectedly, myeloma responses in refractory patients were only seen after the development of acute GVHD. Status at mini-allograft significantly affected the postallograft responses: patients who underwent transplantation in CR/nCR remained in CR/nCR, and none had relapsed at last follow-up, four of five patients who received allograft in PR had achieved nCR/CR, and only 10 of 17 patients with PD before the allograft had a response (nCR/CR in seven patients and PR in three). Overall, 14 patients improved after allograft and one worsened (P = .007). A total of five patients had evidence of PD after mini-allograft; three of them had received two prior ATs and had nonresponsive disease before mini-allograft.

Donor Lymphocyte Infusions and Subsequent Therapy

As stable allografts were established after a single infusion of allogeneic stem cells with a high incidence of GVHD noted in the first phase of the study, the routine use of preemptive DLI was limited. DLIs were only given to establish full chimerism or to patients with evidence of extensive residual/or progressive disease. For the whole group, 18 patients received DLI: five patients received three DLIs, five patients received two DLIs, and eight patients received only one DLI. Six patients received DLIs to attain full donor chimerism and 12 to eradicate residual disease.

Three patients had a persistent large tumor burden. They received dexamethasone, cyclophosphamide, etoposide, and cisplatin chemotherapy before DLI on day 60. All three patients developed clinical grade 2/3 GVHD; two attained CR and one a PR.

Three patients with evidence of progressive disease after day 100 of the mini-allograft received salvage therapy with dexamethasone, thalidomide, cisplatin, doxorubicin, cyclophosphamide, and etoposide and DLI. One patient who had an extramedullary relapse with a large renal mass and CNS involvement achieved and sustained a CR, but eventually died of GVHD complications at day 514. One is in stable PR at day 400 and one died with PD at day 320. CR attained after salvage chemotherapy and DLI were not included in the overall responses (19 in CR/nCR and three in PR) reported after mini-allograft.

Toxicity and Mortality

In this cohort of heavily pretreated, high-risk MM patients, this mini-allograft protocol was reasonably tolerated in both related and unrelated settings in the early posttransplant phase. The major regimen-related toxicities observed are listed in Table 5.

Eighteen patients developed acute GVHD (> grade 2 in 16 and skin grade 1 in two), and 12 developed GVHD for the first time after donor lymphocyte infusions. Grade 2 GVHD responded well to methylprednisolone and cyclosporine. Ten patients developed chronic GVHD; six had extensive chronic GVHD. The incidence of GVHD was not affected by age (P = .7), the number of prior ATs (0.2), or the disease status at mini-allograft (P = .9) (Table 4).

Pneumonitis developed in eight patients, in association with engraftment in four; three of eight patients required ventilator support. One patient developed bronchiolitis obliterans and eventually died of bacterial pneumonia on day 400. Veno-occlusive disease was not observed. Most patients had more than one episode of fever. Eight patients had coagulase-negative Staphylococcus bacteremia. There was a high incidence of CMV antigenemia in the first 16 patients. All were CMV antibody-positive before mini-allograft and had a CMV-positive donor. This complication was successfully treated with ganciclovir. No CMV disease developed in any of the patients. The last 15 patients received preconditioning ganciclovir, and only three patients had evidence of CMV activation during the first 100 days. Invasive aspergillosis developed in three patients after steroid therapy for GVHD. One patient had seizures secondary to high cyclosporine level. Pure red cell aplasia developed in one patient who received bone marrow from an unrelated donor with major ABO mismatch.

Overall, 12 patients have died. TRM in the first 100 days was 10% (n = 3). All were very-high-risk patients (one was hemodialysis dependent, and one had received three prior ATs, with a history of diffuse interstitial pneumonitis); two had received unrelated donor grafts. Three patients died of PD and six including five patients in CR/nCR died from late complications including sepsis (n = 3), aspergillosis (n = 2), and toxoplasmosis (n = 1), which were secondary to immunosuppressive therapy for GVHD. Disease status at the time of mini-allograft significantly affected overall mortality: there was one death (11%) among nine patients who underwent transplantation in CR/nCR, versus two (40%) of five patients who underwent allografting in PR and nine (64%) of 14 patients who underwent allografting with PD (P = .04) (Table 4). There was a trend for higher mortality for patients who received two or more ATs before mini-allograft (P = .06) (Table 4).

Survival

With a median follow-up of 6 months (range, 45 to 730 days) 19 patients are alive with Karnofsky performance status scores of 90 to 100. The median overall durations of event-free survival (EFS) and OS were 15 months, and the projected actuarial survival at 1 and 2 years is 71% and 31%, respectively (Fig 1). The number of prior ATs significantly affected EFS and OS (Fig 2). At 1 year, there was a significantly longer EFS (86% v 31%, P = .01) and OS (86% v 48%, P = .04) when mini-allograft was performed after one versus two or more prior ATs. In a subset analysis and with a short follow-up, patients with responsive disease (CR, nCR, and stable PR) who underwent mini-allograft after one prior AT (n = 9) had a remarkable 100% EFS and OS. The difference between this subset and all the other patients (PD and/or ≥ two prior ATs) allografted was statistically significant for EFS (P = .005) and OS (P = .01) (Fig 3).

View larger version:

• In this window
• In a new window

• PowerPoint Slide for Teaching

Fig 1. Actuarial and overall survival and event-free survival after mini-allografts.

View larger version:

• In this window
• In a new window

• PowerPoint Slide for Teaching

Fig 2. Actuarial survival after mini-allografts according to the number of prior autotransplants. Kaplan-Meier curves show significantly better EFS and OS among patients who received one AT prior to mini-allograft than those who received two or more ATs.

View larger version:

• In this window
• In a new window

• PowerPoint Slide for Teaching

Fig 3. Actuarial survival after allografts according to number of prior autotransplants and disease status. Patients with responsive disease (RD; including CR, nCR, and PR) receiving one AT before mini-allograft had significantly better EFS and OS compared with patients with progressive disease (PD) and/or ≥ two ATs.

Standard Allografts Versus Mini-Allografts

The comparison of patients’ characteristics receiving standard allografts versus mini-allografts is provided in Table 1. Patients receiving mini-allograft were significantly older than those receiving a standard conditioning regimen (median age, 56 v 46 years; P < .0001). Many mini-allograft recipients had additional high-risk features, two patients had renal failure and were dialysis dependent at time of mini-allograft, and four had additional MDS clones with relapsed MM. Ninety-seven percent (n = 30) of the mini-allograft recipients had chromosomal abnormalities v 62% (n = 52) in the standard allograft patients (P < .0001). Chromosome 13 abnormalities were detected in 68% in the mini-allograft patients v 35% in the standard allograft patients (P = .003). Time from last AT to allograft was comparable in the two groups, whereas time from diagnosis was longer in the mini-allograft patients. Conditioning regimens for the standard allotransplant patients were mostly TBI based and variable regimens for GVHD prophylaxis were used. Donors for the standard allotransplant included 14 unrelated and 79 matched sibling donors. Fifteen patients received peripheral-blood stem cells and 78 patients received bone marrow; of the marrow group, 33 had T-cell–depleted grafts.

When outcome of the mini-allograft recipients was compared with those who received standard allografts, TRM in the first 100 days was significantly lower in the mini-allograft than in the standard allotransplant patients (three of 31 [10%] v 27 of 93 [29%], P = .03, respectively). Moreover, even with the relatively short follow-up, mini-allograft patients tend to show a lower mortality during the first year (P = .09); the difference is especially pronounced (P = .05) for the subset of patients who received one AT (Fig 4). After adjusting for significant clinical variables in the multivariate model, standard allotransplant continued to be the only variable that negatively affected OS (hazard ratio = 2.85; P = .007).

View larger version:

• In this window
• In a new window

• PowerPoint Slide for Teaching

Fig 4. First-year mortality in standard allografts and mini-allografts. (A) First-year mortality in the standard allograft (n = 93) and mini-allograft (n = 31) patients. (B) First-year mortality in standard allograft (n = 73) and mini-allograft (n = 14) patients with one prior AT.