Analyses were performed of cumulative doses per protocol step (induction, peritransplantation, consolidation, and maintenance). The multivariate model uses stepwise selection with entry level of 0.1 and variable remains if it meets the 0.05 level. Variables considered univariately were: age ≥ 65 years, female sex, white race, albumin < 3.5 g/dL, B2M ≥ 3.5 mg/L, B2M > 5.5 mg/L, creatinine ≥ 2 mg/dL, CRP ≥ 8 mg/L, Hb < 10 g/dL, LDH ≥ 190 U/L, cytogenetic abnormalities (anytime before enrollment), GEP high-risk, GEP CD-1 subgroup, GEP CD-2 subgroup, GEP hyperdiploidy subgroup, GEP low bone disease subgroup, GEP MAF/MAFB subgroup, GEP MMSET/FGFR3 subgroup, GEP myeloid subgroup, GEP proliferation subgroup, GEP TP53 deletion, GEP NR3C1 bottom tertile, GEP NR3C1 top tertile, completed transplantation 2*, completed consolidation 1*, completed consolidation 2*, bortezomib during induction*, bortezomib during consolidation*, bortezomib during maintenance*, dexamethasone during induction*, dexamethasone during transplantation*, dexamethasone during consolidation*, dexamethasone during maintenance*, thalidomide during induction*, thalidomide during transplantation*, thalidomide during consolidation*, thalidomide during maintenance*, premature bortezomib discontinuation*, premature dexamethasone discontinuation*, premature thalidomide discontinuation*, achieved CR*, and achieved CR before Tx2*. Variables considered for the multivariate model were: B2M > 5.5 mg/L, creatinine ≥ 2 mg/dL, Hb < 10 g/dL, LDH ≥ 190 U/L, cytogenetic abnormalities (anytime before enrollment), GEP high-risk, GEP proliferation subgroup, dexamethasone during induction*, dexamethasone during transplant*, premature bortezomib discontinuation*, premature dexamethasone discontinuation*, premature thalidomide discontinuation* and achieved CR*.

B2M indicates B2-microglobulin; CRP, C-reactive protein; Hb, hemoglobin; and —, not applicable.

Variable was treated as a time-dependent variable.