Abstract

Randomization in clinical trials involving comparison of patients receiving an allogeneic HSCT versus other treatments is challenged by limitated donor availability. Biologic assignment relies on the availability of suitable donors to determine assignment to allogeneic HSCT. BMT CTN 0102 is a phase III multicenter clinical trial that compares tandem autologous HSCT (auto-auto) to an autologous HSCT followed by a nonmyeloablative allogeneic HSCT (auto-allo) for the treatment of multiple myeloma after initial systemic therapy. Transplant arm assignment was determined by the availability of an HLA-matched sibling donor. Patients were stratified into high and standard risk groups. Absence of chromosome 13 deletion by standard karyotyping and beta-2 microglobulin (B2M) ≤ 4mg/L were defined as standard-risk disease (SRD). The trial was powered to detect differences in 3-year progression-free survival in patients with SRD. A total of 710 patients from 43 centers in the U.S. were enrolled from December 2001 to March 2007. Data from 596 patients with SRD were available for this analysis (sibling donor availability for other 28 SRD patients is not yet determined); 30% (n=178) were assigned to the auto-allo arm. The median ages were 55 and 52 years for the auto-auto and auto-allo arms, respectively (p=0.01). The proportions of African-American patients were 17% and 9% in the auto-auto and auto-allo arms, respectively (p=0.006). Most patients were Durie Salmon Stage III (66% and 70%, p=0.6), in partial remission (86% and 83%, p=0.6) and had Karnofsky Performance Scores ≥ 90 (76% and 77%, p=0.5) at time of study entry. Treatment arms were balanced for B2M, baseline cardiac and hepatic functions and CMV serological status. Median times from diagnosis, initiation of systemic therapy and stem cell mobilization to first transplant were 213, 191 and 27 days, respectively and the same in both arms. The interval from the first to second transplant was 98 days for the auto-auto arm and 105 days for the auto-allo arm (p=0.02). We conclude that biologic assignment resulted in two reasonably comparable treatment groups, although statistically significant differences in age, race and interval between first and second transplant were identified. The differences in age between the groups is related to donor availability, since patients younger than 55 years had on average significantly more living siblings than older patients. Importantly, full-sibling donor availability did not account for differences in race distribution. Because the clinical significance of all detectable differences is unknown, the planned statistical analysis will adjust for these discrepancies. These results confirm that biologic assignment is feasible and might be considered when standard randomization is prohibitive, as in allogeneic HSCT clinical trials.

Author notes

Disclosure:Ownership Interests:; One of co-authors reported owning stock on Amgen and Celgene Corporation. Honoraria Information: One of the co-authors reported receiving speaker’s honoraria from Celgene Corporation. Membership Information: Some of the co-authors reported being in the speaker’s bureau of the following companies: Millenium Pharmaceutical, Genentech and Celgene Corporation.