An Update of a Comparison of Nonmyeloablative Allografting with Autografting for Newly Diagnosed Myeloma.

The concept of genetic randomisation has been applied to assess clinical outcomes between patients with hematological cancers treated with allografting or other therapies. Though not universally accepted, the comparison by the intention-to-treat principle between patients with HLA-identical siblings, who can be assigned to allografting, and those without, who cannot receive an allograft, is used as a surrogate for an unbiased randomisation. We previously published the results of a study where the treatment assignment of 162 newly diagnosed patients younger than 65 years was based only on the presence/absence of an HLA-identical sibling (Bruno et al, N Engl J Med). First-line treatment plans included a cytoreductive autograft followed by a nonmyeloablative allograft (Tandem auto-allo) or a second melphalan-based autograft (Double-auto). Primary endpoints were overall (OS) and event-free (EFS) survivals by intention-to-treat analysis. The 80 patients with a sibling donor were offered a Tandem auto-allo and the 82 without a Double-auto after high (140–200 mg/m²) or intermediate dose melphalan (100 mg/m²). After a median follow up of 45 (range 21–90) months, OS and EFS were significantly longer in patients with donors: 80 versus 54 months (p=0.01) and 35 versus 29 months (p=0.02). Median OS was not reached in the 58 (out of 60 enrolled, 97%) patients who completed Tandem auto-allo and was 58 months in the 46 (out of 59 enrolled, 78%) who completed high-dose double-auto (p=0.03). Here, we report a update analysis. At a median follow up of 56 months, OS was not reached for the 80 patients with an HLA-identical sibling and was 56 months for those without (HR 0.53, CI 95% 0.33–0.86, p=0.009). EFS remained significantly longer in patients with HLA-identical siblings: 35 versus 29 months (HR: 0.61; 95% Cl: 0.42–0.88, p=0.008). Median OS was not reached in the 58 patients who completed Tandem auto-allo and was 63 months in the 46 who completed high-dose double-auto (HR 0.47, CI 95% 0.25–0.86, p=0.016). EFS was 43 and 33 months (HR 0.64, CI 95% 0.40–1.02, p=0.06). By multivariate analysis, the presence of HLA-identical siblings was significantly correlated with longer OS and EFS. We carried out a stratified analysis, on the intent-to-treat population, that defined a patient subgroup at high risk in the light of high b -2-microglobulin levels or presence of del(13). The adjusted HRs by Cox models were 0.34 for OS and 0.52 for EFS similar to those obtained in the whole series. Though this exploratory analysis has low statistical power, its results indicate that del(13) does not offset the advantage in OS and EFS for patients with an HLA-identical sibling, but do not exclude an impact of del(13) in those patients undergoing an allograft. Attal et al. (N Engl J Med) previously reported median OS and EFS of 58 and 30 months, respectively, after double autologous transplantation, consistent with our results (OS: 63 months; EFS: 33 months). However, the EFS after Tandem auto-allo also indicates that long-term disease control is an issue. Allografting and new drugs with molecular targets should not be viewed as mutually exclusive. Thus, it is imperative to thoroughly explore their roles in increasing the response rates and their duration in Tandem auto-allo.