Скачать полный оригинал PDF с иллюстрациями
Impact of maintenance therapy on survival after autologous
stem cell transplant for multiple myeloma regardless of
response: a global analysis
by Andrew J. Cowan, Luuk Gras, Raffaella Cassano, Yoshiko Atsuta, Laurien Baaij, Francisca Bass,
Ali Bazarbachi, Mohamed Amine Bekadja, Katherine Clesham, Peter Dreger, Anita D'Souza,
Noel Estrada-Merly, Ai Sim Goh, Minako Iida, Patrick J Hayden, Koji Kawamura, Bor-Sheng Ko,
Linda Koster, Christopher Liam, Hira Mian, Arleigh McCurdy, Cristobal Augusto Frutos Ortiz, Nada Hamad,
Donal McLornan, Meng Lv, Shohei Mizuno, Aloks Srivastava, Asmaa Quessar, Eloisa Riva, Wael Saber,
John Snowden, Hiroyuki Takamatsu, Xavier Cheng-Hong Tsai, Estelle Verburgh, Ho Kim Wah,
Daniel Neumann, Anna Sureda, Mahmoud Aljurf, Mickey Boon Chai Koh, Sebastian Galeano,
Yoshihisa Kodera, Damiano Rondelli, Dietger W . Niederwieser and Laurent Garderet
Received: May 25, 2026.
Accepted: September 8, 2026.
Citation: Andrew J. Cowan, Luuk Gras, Raffaella Cassano, Yoshiko Atsuta, Laurien Baaij, Francisca Bass,
Ali Bazarbachi, Mohamed Amine Bekadja, Katherine Clesham, Peter Dreger, Anita D'Souza,
Noel Estrada-Merly, Ai Sim Goh, Minako Iida, Patrick J Hayden, Koji Kawamura, Bor-Sheng Ko,
Linda Koster, Christopher Liam, Hira Mian, Arleigh McCurdy, Cristobal Augusto Frutos Ortiz,
Nada Hamad, Donal McLornan, Meng Lv, Shohei Mizuno, Aloks Srivastava, Asmaa Quessar, Eloisa Riva,
Wael Saber, John Snowden, Hiroyuki Takamatsu, Xavier Cheng-Hong Tsai, Estelle Verburgh, Ho Kim Wah,
Daniel Neumann, Anna Sureda, Mahmoud Aljurf, Mickey Boon Chai Koh, Sebastian Galeano,
Yoshihisa Kodera, Damiano Rondelli, Dietger W . Niederwieser and Laurent Garderet.
Impact of maintenance therapy on survival after autologous stem cell transplant for multiple
myeloma regardless of response: a global analysis.
Haematologica. 2026 Sept 17. doi: 10.3324/haematol.2026.301244 [Epub ahead of print]
Publisher's Disclaimer.
E-publishing ahead of print is increasingly important for the rapid dissemination of science.
Haematologica is, therefore, E-publishing PDF files of an early version of manuscripts that have
completed a regular peer review and have been accepted for publication.
E-publishing of this PDF file has been approved by the authors.
After having E-published Ahead of Print, manuscripts will then undergo technical and English editing,
typesetting, proof correction and be presented for the authors' final approval, the final version of the
manuscript will then appear in a regular issue of the journal.
All legal disclaimers that apply to the journal also pertain to this production process.
1
Impact of maintenance therapy on survival after autologous stem cell transplant for multiple myeloma
regardless of response: a global analysis
Andrew J. Cowan 1*, Luuk Gras 2*, Raffaella Cassano3, Yoshiko Atsuta 4, Laurien Baaij 2, Francisca Bass 5, Ali
Bazarbachi6, Mohamed Amine Bekadja 7, Katherine Clesham 8, Peter Dreger 9, Anita D'Souza 10; Noel Estrada -
Merly10, Ai Sim Goh11, Minako Iida12, Patrick J. Hayden13, Koji Kawamura14, Bor-Sheng Ko15,16,17, Linda Koster,
Christopher Liam18, Hira Mian19, Arleigh McCurdy20, Cristobal Augusto Frutos Ortiz21, Nada Hamad22, Donal P.
McLornan8, Meng Lv23, Shohei Mizuno12, Aloks Srivastava24, Asmaa Quessar25, Eloisa Riva26, Wael Saber10, John
Snowden27,28, Hiroyuki Takamatsu 29, Xavier Cheng -Hong Tsai 30, Estelle Verburgh 31, Ho Kim Wah 32, Daniel
Neumann33, Anna Sureda34, Mahmoud Aljurf 35; Mickey Boon Chai Koh 36, Sebastian Galeano 37, Yoshihisa
Kodera12, Damiano Rondelli38, Dietger W. Niederwieser12,39,40* and Laurent Garderet41*
* Authors AJC and LG contributed equally to this manuscript; DWN and LG contributed equally to this
manuscript
1 Andrew Cowan, MD; BC Cancer Agency and University of British Columbia;
2 EBMT Leiden Study Unit, Leiden, Netherlands;
3 Department of Medicine, University of Pisa, Italy
4 Japanese Data Center for Hematopoietic Cell Transplantation, Nagakute, Japan;
4 Aichi Medical University School of Medicine, Nagakute, Japan;
5 Intensive Hematology Unit, Hospital del Salvador, Santiago, Chile.
6 Department of Internal Medicine, American University of Beirut, Beirut, Lebanon.
7 Faculty of Medicine, Ahmed Benbella-1 University of Oran, Oran-Algeria.
8 University College London Hospital, London, United Kingdom,
9 University of Heidelberg, Heidelberg, Germany.
10 Medical College of Wisconsins, USA
11 Hospital Pulau Pinang, Malaysia
12 Aichi Medical University School of Medicine, Nagakute, Japan.
13 St James Hospital, Ireland
14 Tottori University, Tottori, Japan;
15 Department of Internal Medicine, National Taiwan University Hospital, Taipei, Taiwan
16 Department of Hematological Oncology, National Taiwan University Cancer Center, Taipei, Taiwan
17 School of Medicine, National Taiwan University College of Medicine, Taipei, Taiwan
18 Hospital Sultanah Aminah, Johor Bahru, Malaysia
19 Dept of Oncology, McMaster University, Ontario, Canada
20 Department of Medicine, Division of Hematology, The Ottawa Hospital, Ottawa, Canada
21 Instituto de Prevision Social, Asuncion, Paraguay
22 St Vincents Hospital and University of Sydney, Sydney Australia
23 Peking University People’s Hospital,
24 Christian Medical College, Vellore India
25 Abulcasis international university of health sciences, Rabat. Morocco
26 Clinical Hospital Dr. Manuel Quintela De Clinicas, Montevideo, Uruguay
27 Division of Clinical Medicine, School of Medicine and Population Health, The University of Sheffield, UK
28 Department of Haematology, Sheffield Teaching Hospitals NHS Foundation Trust, Sheffield, UK
29 Kanazawa University, Kanazawa, Japan;
30 Department of Internal Medicine, National Taiwan University, Taiwan
31 Groote Schuur Hospital, Hospital Road, Cape Town, South Africa;
32 Sunway Medical Centre, Kuala Lumpur, Malaysia
33 IMISE, University of Leipzig, Leipzig, Germany;
34 Hematology and Hematopoeitic Stemm Cell Transplant Programme, Institut Catala d’Oncologia, Barcelona, Spain
35 Oncology Center, King Faisal Specialist Hospital and Research Center, Riyadh, Saudi Arabia
36 City St Georges, University of London and St Georges University Hospitals, London, UK
37 Hospital Britanico, Montevideo, Uruguay
38 Division of Hematology/Oncology, University of Illinois at Chicago -UI Health, Chicago, USA
39 University of Leipzig, Leipzig, Germany,
40 Penn Medicine, Philadelphia, USA,
41 Sorbonne Université, Inserm, Centre de Recherche Saint-Antoine, Assistance Publique-Hôpitaux de Paris, Hôpital
Pitié Salpêtrière, Service d'Hématologie, Paris, France;
2
Corresponding author: Andrew Cowan, MD; email andrew.cowan@bccancer.bc.ca
Contributions
Andrew J. Cowan and Luuk Gras designed the research, interpreted the data, and wrote the manuscript. Luuk Gras,
Laurien Baaij, Linda Koster, and Daniel Neumann analyzed the data and performed the statistical analyses. Laurent
Garderet and Dietger W. Niederwieser conceived and supervised the study, interpreted data, and critically revised
the manuscript. Raffaella Cassano, Yoshiko Atsuta, Francisca Bass, Ali Bazarbachi, Mohamed Amine Bekadja,
Katherine Clesham, Peter Dreger, Anita D'Souza, Noel Estrada-Merly, Ai Sim Goh, Minako Iida, P.J. Hayden, Koji
Kawamura, Bor-Sheng Ko, Christopher Liam, Hira Mian, Arleigh McCurdy, Cristobal Augusto Frutos Ortiz, Nada
Hamad, Donal P. McLornan, Meng Lv, Shohei Mizuno, Asmaa Quessar, Eloisia Riva, Wael Saber, John Snowden,
Hiroyuki Takamatsu, Xavier Cheng-Hong Tsai, Estelle Verburgh, Ho Kim Wah, Anna Sureda, Mahmoud Aljurf,
Mickey Boon Chai Koh, Sebastian Galeano, Yoshihisa Kodera, and Damiano Rondelli contributed vital data and
study materials, interpreted the data, and critically reviewed the manuscript. All authors approved the final
manuscript and agreed to its submission.
Disclosures
AJC: Research Funding: JNJ, BMS, Sanofi, Regeneron, Merck, Caelum, Opna Bio, Karyopharm;
Advisory/Consulting: JNJ, BMS, Sanofi, HopeAI, Abbvie, Kite/Arcellx; Stock ownership: HopeAI
All other co-authors indicate no other relevant COI.
Data-sharing statement
The data of this study are not publicly available. These data are managed in accordance with the WBMT Research
Guidelines.
3
Abstract
The effect of different maintenance therapies on outcome after autologous hematopoietic cell transplantation
(AHCT) was analyzed in newly diagnosed multiple myeloma (NDMM) patients using global real-world data.
NDMM patients transplanted between 2013 and 2017 were selected from the Worldwide Network for Blood and
Marrow Transplantation (WBMT) registry. Overall survival (OS) was evaluated as primary and progression-free
survival (PFS), non-relapse mortality and relapse/progression incidence as secondary endpoints. Multivariable 6-
month landmark analyses were performed using Cox proportional hazards models including random effect for
country.
Information was available on 8,269 patients with lenalidomide (n=4,148; 50.2%), thalidomide (n=689; 8.3%) or
bortezomib (n=726; 8.8%) as maintenance or no maintenance (n=2,706; 32.7%). In the 6-month landmark analysis,
OS at 3 years was higher after maintenance with lenalidomide (86%), with thalidomide (80%) and with bortezomib
(83%) as compared to no maintenance (79%, p<0.001) irrespective of r esponse status at AHCT. In multivariate
analyses, maintenance with lenalidomide (HR0.57, 95%CI 0.48-0.68), thalidomide (HR0.76, 95%CI 0.58-1.00)
and bortezomib (HR0.65, 95% CI 0. 50-0.84) were associated with better OS compared to no maintenance.
Compared to no maintenance, all three medications significantly improved PFS early after HCT. The effect was
gone by 30, 6, and 12 months for lenalidomide, thalidomide, and bortezomib respectively.
Maintenance with lenalidomide, and to a lesser extend thalidomide and bortezomi b resulted in improved OS as
compared to no -maintenance in all depth s of response prior to AHCT. Weaker and shorter associations were
observed for PFS reflecting maintenance type, therapy discontinuation and survivor effects.
4
Introduction
Multiple myeloma (MM) is a plasma cell malignancy that accounts for 1% -1.8% of all cancers and is the second
most common hematological malignancy (1, 2) . Although diagnosed more frequently in high -income
sociodemographic index (SDI) countries (including Western Europe, North America, and Australia), MM remains
a global disease, with an age standardized incidence rate (ASIR) of 1.72/100,000 persons globally (2). MM was
responsible for 98,437 deaths worldwide with an age-standardized death rate (ASDR) of 1.5 per 100 ,000 persons
(95% UI, 1.3-1.7) in 2019 [3].
Currently, the state-of-the-art treatment for newly diagnosed MM (NDMM) includes induction with a proteasome
inhibitor (PI ; most commonly bortezomib ), an immunomodulatory agent (most commonly, lenalidomide or
thalidomide) and dexamethasone. In some regions of the world , induction will also typically include anti-CD38
monoclonal antibodies, either daratumumab or isatuximab (3, 4). Autologous hematopoietic cell transplantation
(AHCT) with melphalan conditioning is standard of care for eligible patients followed by maintenance therapy (3).
However, real world global data regarding maintenance agents and the impact in different response categories prior
to AHCT has not been published on a global level.
Thalidomide was the first agent explored as maintenance. In the Myeloma IX trial , the use of thalidomide
maintenance was associated with potential OS and PFS benefit in patients with favorable, but worse OS in patients
with adverse fluorescence in situ hybridization ( FISH) findings (5). In the HOVON 50 Trial, thalidomide
maintenance was compared to Interferon alpha, confirming the benefit of thalidomide at a median follow-up of 129
months (IQR 123–136), showing significantly longer event free survival in the thalidomide group compared with
the control group (6).
Larger randomized trials, and subsequent meta-analyses, have confirmed the beneficial effect of maintenance
lenalidomide after AHCT on PFS (7, 8). PFS at 3 years was 59% in the lenalidomide group, as compared with 35%
in the placebo group [hazard ratio (HR) 0.50; p<0.001]. The benefit was observed across all patient subgroups, even
those with different risk profiles based on baseline demographic or disease characteristics, including response at
the time of randomization and initial cytogenetic results. Additionally, there was an OS benefit for receipt of
5
lenalidomide maintenance (8). Furthermore, a meta-analysis including more than 1200 patients revealed a median
PFS of 52.8 months for the lenalidomide group and 23.5 months for the placebo or observation group ( HR 0.48;
95% CI, 0.41 to 0.55). In this analysis, OS was also improved with lenalidomide; the median OS was not reached
for lenalidomide, whereas it was 86.0 months for the control group (HR 0.75; 95% CI, 0.63 to 0.90; p=0.001).
Unfortunately, there was no benefit in OS in cohorts of patients with ISS-III disease or high-risk cytogenetics (7).
However, a subsequent analysis from the Myeloma XI trial demonstrated a benefit for lenalidomide maintenance
even amongst patients harboring high-risk cytogenetic abnormalities (9).
Real world data from the CIBMTR demonstrated that maintenance therapy (including lenalidomide, lenalidomide
in combination with bortezomib, or other combinations) yielded significantly improved 3 -year PFS vs no post -
transplant therapy (55% vs. 39%, p=0.0001). This benefit was most evident in patients not achieving at least
complete remission (CR) post-AHCT (10). However, the CIBMTR registry predominantly comprises patients from
high-income Socio demographic Index countries.
In our retrospective study of real-world data, we aimed to explore the worldwide use of maintenance type and
outcome. This is of considerable importance in view of diverse access to maintenance drugs worldwide. In the light
of results published so far on effect of disease response status prior to AHCT, we conducted an analysis of the
interaction between response status and maintenance treatment on outcomes after AHCT.
Methods
Study Design, Data Sources
This is a retrospective analysis of registry data, conducted through the Worldwide Network for Blood and Marrow
Transplantation ( WBMT) utilizing data from its member societies, or international/regional registries . A full
description of the registry has been published previously (11). Data obtained from patients with NDMM, aged ≥18
years at time of AHCT between 2013 and 2017 and data available on maintenance (lenalidomide, bor tezomib,
thalidomide or no maintenance /unknown) w as selected. As no individual patient data were used, no ethics
committee approval was mandated. The need for additional informed consent from patients was waived because
6
the study was performed by the secondary use of registry data, and no personal information was transferred. The
primary endpoint was OS and the secondary endpoint PFS , non-relapse mortality (NRM) and relapse/progression
incidence (RPI). Outcome data were acquired through regional registries (see supplement Table S1)
Statistical Analysis
Baseline clinical, demographic, and transplantation -related characteristics were grouped according to the
maintenance groups (lenalidomide, thalidomide, bortezomib, and no maintenance), and were reported as median,
range, and interquartile ranges for continuous variables. Maintenance was started according to general agreement
during the first months after AHCT and given for a duration of 2 to 3 years or until progression (Prog) or toxicity.
Differences between groups were assessed using p-values obtained with the χ2 test for categorical variables and the
Kruskal-Wallis test for continuous data.
To account for the variable timing of maintenance initiation and data on start and stop of maintenance not being
available, we performed a 6 -month landmark analysis. For those who had survived beyond 6 months without
relapse, OS was calculated as the time from the 6-month landmark to death from any cause and PFS as the time
from the 6-months landmark to death or disease Prog. NRM was defined as death without progression or relapse
(Rel). In analyses of OS , PFS, Rel/Prog and NRM events occurring more than three years after the 6-month
landmark were artificially censored. Median follow-up after the 6-month landmark and 95% confidence intervals
(CI) were calculated using the reverse Kaplan-Meier method. The probability of OS and PFS was estimated based
on the Kaplan-Meier method and differences were analyzed using the log-rank test. The crude cumulative incidence
estimator and Gray’s test were used for competing events ( Rel/Prog together with NRM). Multivariable analyses
(MVA) were performed using Cox (cause -specific) proportional hazards models including a random effect for
country. Variables included in MVA, a part from maintenance and response status at AHCT according to IMWP
[12] are listed in the Supplement.
We included interaction terms between maintenance treatment and disease stage at AHCT in OS and PFS models
to check whether the association of maintenance and outcome was similar across all disease stage. We analyzed
whether the association between the maintenance and outcome was constant over time ( for proportional hazards
7
assumption, see details in s upplement). Missing values were modeled using a separate missing category. All
statistical tests were 2 -sided, and significance was defined as p <0.05. All analyses were performed in R version
4.4.2 (12) using ‘survival’, ‘cmprsk,’, ‘prodlim’ packages. No adjustment for multiple comparisons were made.
Results
From 2013 to 2017, 61,802 patients were reported to the WBMT registry. Of these, 8,269 (13.4%) had information
available on maintenance, which included lenalidomide in 4,148 (50.2%), thalidomide in 689 (8.3%), bortezomib
in 726 (8.8%) and no maintenance in 2,706 (32.7%) patients. In addition, 746 patients were reported to have
received no further defined maintenance, 536 an IMiD + proteasome inhibitor combination, and 66 carfilzomib as
maintenance. These latter groups were excluded from this evaluation due to low numbers or incomplete clinical
information. Demographics of the 8,269 patients treated with lenalidomide, thalidomide, bortezomib or without
maintenance are shown in Table 1. Of those, 7,645 were alive at 6 months post-AHCT and included in the landmark
analysis (51.8% with lenalidomide, 7.6% thalidomide, 8.7% bortezomib and 31.9% without maintenance; Table 1).
Data obtained from p atients were entered by the following registries: Japan (34.7%), EBMT (32.6%), CIBMTR
(22.3%), LABMT (3.2%), Canada (2.0%), Malaysia (2.0%), Taiwan (1.9%), China (0.7%) and EMBMT ( 0.7%)
(Table S1). Maintenance differed across world regions. While in Europe , North America, Japan and EMBMT
lenalidomide was the most frequently used drug, thalidomide was the most utilized drug in Taiwan, Latin America
and China. The majority were ISS I (36.2%), were male (56.8%) and IgG myeloma (57.2%). The median time from
diagnosis to AHCT was 7.1 months (Interquartile range [IQR], 5.6 – 10.0), and the median age at AHCT was 60.0
years (range 20.2-82.4 years). Response status at AHCT was CR in 19.8%, VGPR in 36.2%, PR in 37.8%, SD/MR
in 4.4% and Rel/Prog in 1.7%. The majority had melphalan 200 mg/m2 conditioning (74.2%) and a single AHCT
(91.2%). The cytogenetic risk score was high in 22.2%, KPS ≤90% in 65.9% and high HCT-CI in 17.2% of patients,
as with the response status at HCT , were unevenly distributed between the four maintenance groups. Except for
sex, patient characteristics were not equally distributed between the groups , with a higher percentage of
lenalidomide patients with AHCT in more recent years. Thalidomide patients were at a younger age at AHCT, a
8
higher percentage were in CR, had a KPS of 100, a low HCT-CI and had standard cytogenetic risk compared to
lenalidomide.
Univariable analysis
The OS and PFS of the whole population with maintenance information (n=8266) starting from AHCT is shown in
Supplemental Figure S1. At 6 months OS was 99.0% ( 95% CI 98.7-99.2%) and PFS was 95.3% ( 95% CI 94.8-
95.9%). All following analyses include only those patients who were alive and relapse-free at 6 months post AHCT
(n=7,645). Median follow-up (IQR) after the 6-month landmark was 45.0 months (29.0-61.2 months).
A deeper response prior to AHCT and type of maintenance were associated with better OS and PFS after the 6-
month landmark analysis (Figure1, all p<0.001). OS at 3 years was 87% (95% CI 85-89) for patients in CR, 84%
(95% CI 83-86) in VGPR, 82% (95% CI 80-83) in PR and 74% (95% CI 69-78) in SD/MR/Prog (Figure 1a and
Table S2). OS at 3 years of the patients who received maintenance lenalidomide was 86% (95% CI 85-87),
thalidomide 80% (95% CI 76-84), bortezomib 83% (95% CI 80-86) and no maintenance 79% (95% CI 77-80)
(p<0.001; Figure 1b). PFS resembled the patterns of OS with respect to response status (Figure 1c and Table S2).
PFS at 3 years was 63% (95% CI 61-65%) with lenalidomide, 53% (95% CI 48-57%) with thalidomide and 52%
(95% CI 47-56%) with bortezomib maintenance as compared to 60% (95%CI 58-63%) without maintenance
(p<0.001 Figure 1d and Table S2).
We then investigated whether the type of maintenance affected outcomes equally in the different response
categories. In univariate analysis, patients with lenalidomide maintenance had the highest OS at all response
categories prior AHCT (CR, VGPR, PR and SD/MR/Progr) and 3-year OS was 91% (95% CI 88-93%), 86% (95%
CI 84 -88%), 85% (95% CI 83 -87%) and 8 1% (95% CI 7 5-87%), respectively. Patients with thalidomide
maintenance had lower and decreasing OS at 3 years according to response status at AHCT reaching 78% (95% CI
70-85%), 84% (95% CI 78-90%), 81% (95% CI 74-88%), and 72% (95% CI 56-87%), respectively. Patients with
bortezomib maintenance, although inferior to thalidomide, had response dependent OS at 3 years of 89% (95% CI
84-95%), 85% (95% CI 80-90%), 77% (95% CI 70-83%) and 86% (95% CI 73-99%), respectively. The lowest OS
was observed in those patients who did not receive maintenance amounting to 84% (95% CI 81-88%), 80% (95%
9
CI 7 6-83%), 7 8% (75-81%) and 6 3% (95% CI 5 5-71%) respectively (Table S2 and supplemental Figure S2 ).
Similarly, patients treated with lenalidomide and to a lesser extend with thalidomide and bortezomib, had improved
3-years PFS across all response categories in comparison to no maintenance (Table S2). The differences in OS and
PFS were mainly due to a reduced incidence of Rel/Prog incidence. Although NRM was very low (2 -5%),
differences between the maintenance groups were detected. NRM was highest in the no maintenance group (Table
S2).
Multivariable Analysis
In the OS multivariate analysis of 7,547 patients (Table 2), lenalidomide (HR 0.57, 95% CI 0.48-0.68; p<0.0001),
bortezomib (HR 0.65, 95% CI, 0.50-0.84, p=0.001), and thalidomide (HR 0.76, 95% CI, 0.58-1.00; p=0.05) were
associated with a lower risk of death compared to no maintenance. There was a trend towards a lower risk of death
in the lenalidomide group compared to thalidomide (HR 0.79, 95% CI 0. 61-1.02, p= 0.07), and no significant
difference between lenalidomide and bortezomib (HR 0.88, 95% CI 0.69-1.11, p=0.27).
In the PFS analysis of 7,008 patients there was a significant difference in the association with lenalidomide (HR
0.65, 95% CI 0.5 7-0.73, p<0.0001), but not overall with thalidomide (HR 0.90, 95% CI 0. 75-1.09, p=0.29) or
bortezomib (HR 0.93, 95% CI 0.78-1.09, p=0.36). Lenalidomide was associated with better PFS than thalidomide
(HR 0.72, 95% CI 0.60-0.85, p<0.0001) and bortezomib (HR 0.68, 95% CI 0.59-0.79, p<0.0001).
We then tested whether these HR were constant during the entire follow-up period. Indeed, HR´s were not constant
during the follow-up period (Figure 2) and a trend of HR’s closer to 1 with longer time after the 6 month landmark
were noted (Table S3). The association of better OS with lenalidomide was strongest at the 6-month landmark (HR
0.25) and diminish ed with longer time after follow -up. HR’s were significantly lower for lenalidomide vs. no
maintenance comparison in OS up to 42 months after AHCT. A similar pattern was observed for thalidomide and
for Bortezomib compared to no maintenance up to 16 months and 30 months after AHCT for OS, respectively. In
the PFS comparison between lenalidomide vs. no maintenance we also observed diminishing HR´s with longer time
during follow-up, but a significant improved PFS (significant the first 30 months after HCT). There was evidence
10
for a significant better PFS for thalidomide maintenance the first 6 months and for Bortezomib maintenance the
first 12 months post-transplant (Figure 2 and Table S3)
Other significant risk factors for poorer OS resembled the results published previously on 61,797 patients [10]:
more advance disease stage prior to AHCT, older age, earlier calendar year of AHCT, lower KPS, higher ISS stage
and cytogenetic high-risk, but not gender and HCT -CI (Table 2). Response status prior to AHCT, female, ISS II
and III, high-risk cytogenetics, but not age, year of AHCT and HCT-CI were significant prognostic factors for PFS
(Table 2). There was no evidence for an interaction between the type of maintenance and the response status prior
to AHCT in either the analysis of OS (p=0. 06) and PFS (p=0.1 0) suggesting a similar beneficial effect of
maintenance across all disease stages.
Predicted OS and PFS probabilities during the first 3 years after the 6 -month landmark fo r a reference patient
according to response status at AHCT and receiving either lenalidomide, bortezomib, thalidomide or no
maintenance is given in Figure 3. Predictions were calculated using the results from the models presented in Table
2 and Figure 2 . Differences were more pronounced with higher risk of death in SD/MR/Prog disease stages at
AHCT. For maintenance therapies there was a predicted PFS benefit up to 36 months, but at 36 months after the 6-
month landmark predicted PFS was similar for bortezomib, thalidomide and no maintenance.
Discussion
In our analysis of a global multiple myeloma registry of AHCT recipients, we analyzed the impact of response
depth prior to AHCT and the interaction with maintenance on long-term outcomes. As we demonstrated previously,
better response status at the time of AHCT was significantly associated with better PFS and OS (11). In this analysis,
we demonstrated that lenalidomide maintenance improved OS and PFS irrespective of the response at the time of
AHCT. Bortezomib and thalidomide were superior to no maintenance with respect to OS, however, they were
inferior to lenalidomide maintenance, and they did demonstrate a significant lower risk of death/Rel, but only early
after HCT.
11
Our results have some important messages. Despite randomized trials demonstrating the survival advantage (7),
maintenance is still not used consistently globally, and when utilized, lenalidomide, bortezomib, and thalidomide
are all being used in different world regions. According to our global registry, lenalidomide was used most
frequently in Europe, USA, E MBMT region and Japan (in 73.2%, 66.2%, 68.5% and 25.7%, respectively; Table
S1). In contrast, China, Malaysia, the LABMT, and Taiwan more frequently utilize thalidomide. Despite these
differences, we have found that all agents deliver some benefit, though lenalidomide appeared to be the best.
Interestingly, high percentages of patients in many regions still receive no maintenance at all (see Table S1).
Our results showed the beneficial effect of lenalidomide in all patients, even in those with SD/MR or worse. The
association became smaller with longer follow up (HR lenalidomide vs. no maintenance became closer to 1) , and
after 36 months there was no sign ificant difference. The reasons for this effect may include disease progression,
survivor effects, and therapy discontinuation due to cumulative toxicities such as second primary malignancies
(SPM) or cytopenias (8). Unfortunately, we do not have data on start and stop dates of maintenance treatment .
Efforts to determine the ideal length of maintenance should be undertaken.
Our findings align with recent real world registry data ; however, our analysis utilizes a global registry including
some lower middle-income countries (LMIC). In a recently published analysis of 1109 MM patients from a large
US-based academic center, the outcomes of patients who achieved a suboptimal response prior to AHCT, defined
as less than a VGPR, were reported (13). A median PFS of 38.6 months was reported, and in a multivariate analysis,
the use of post-transplant maintenance was associated with a superior PFS HR when compared to no maintenance
of 0.60 (p<0.001). However, in this cohort, only 30% of patients did not receive maintenance, and the majority who
received maintenance had lenalidomide (73%). Only 22% received a non -lenalidomide maintenance regimen. In
an older study from the CIMBTR, investigators examined the impact of additional therapy prior to AHCT for
patients who achieved a PR or less to induction therapy (14). While additional therapy deepened response in 68%
of patients, there was no impact on PFS or OS. In our analysis, we only assessed response at the time of AHCT,
and included a global cohort of patients, which is different from the patient populations in the prior analyses.
12
Given the findings of our analysis and prior meta-analyses and large randomized phase 3 trials demonstrating a PFS
and OS benefit of maintenance our findings from a real-world global registry argue strongly for routine use of
maintenance (7, 8). Although lenalidomide demonstrated the strongest HR in favor of PFS and OS at all response
levels, both thalidomide and bortezomib were also utilized globally and demonstrated a favorable impact on OS
and PFS – notably, prior phase 3 studies have shown that bortezomib demonstrates superior outcomes over
thalidomide (15). In summary, access to maintenance therapy is critical, regardless of whether lenalidomide or
bortezomib/thalidomide are used, even in cases where the response prior to AHCT was suboptimal.
Our results are strengthened by including cohorts from large regis tries across the globe with international
representation. Multiple myeloma is a global disease (1, 2), and it is critical to understand the impact of therapies
across all persons diagnosed with this cancer to deepen understanding and report outcomes equitably. We also
describe the impact of non-lenalidomide maintenance therapy at a global perspective.
However, there are some weaknesses to our analysis. Our analysis covered the years 2013-2017, and maintenance
agents utilized during this era have been supplanted by newer, sometimes expensive maintenance treatment
strategies, and thus, may not reflect current best practices . For example, daratumumab is increasingly being
incorporated into maintenance therapy, with improved outcomes – based on the results of the PERSEUS and
CASSIOPEIA trials - and we do not have data on this agent in the WAUSTIM dataset at this time (4, 16). We did
not have detailed toxicity or secondary primary malignancy data which is critical to understanding the long-term
toxicities of lenalidomide. Indeed, SPM detection may have an impact on early compared to late OS benefit; as has
been shown in previous trials, wherein the OS benefit for lenalidomide maintenance outweighed any impact related
to SPM incidence over time (7, 8). Our dataset also suffered from missing data (for instance, the EBMT registry
did not have data on no maintenance), on timing of maintenance, duration of therapy, and treatment adherence,
which are also relevant when assessing a therapy that is given over a long duration. Although our cohort attempted
to include a broader spectrum of global regions, most of our data is from Japanese, North American, and European
regions, highlighting the need to create inclusive global disease-based registries and treatment outcomes. China,
one of the most populous regions globally, contributed only a small proportion of patients to our registry. Improving
registry participation and more inclusive international collaborations are needed with support from health
13
authorities and policymakers. Finally, we lack data on induction regimens; as has been demonstrated in recent large
trials, such as CASSIOPEIA, agents included in induction and in maintenance may have an impact on PFS (17).
Taken together , our analysis of a global MM transplant registry sheds important light on the significance of
maintenance therapy after AHCT, and the ability of maintenance to improve outcomes even in the setting of a less-
than-optimal response prior to AHCT. Furthermore, although lenalidomide in our analysis resulted in superior
outcomes for PFS and OS in all response categories, both bortezomib and thalidomide, which in some regions may
be more broadly available (11), also provided an OS benefit over no receipt of maintenance therapy in multivariate
analysis. Our findings also support pragmatic trials in low resource settings to assess the feasibility and outcomes
of different maintenance strategies.
14
References
1. Cowan AJ, Allen C, Barac A, et al. Global burden of multiple myeloma: a systematic analysis
for the global burden of disease study 2016. JAMA Oncol. 2018;4(9):1221-1227.
2. Zhou L, Yu Q, Wei G, et al. Measuring the global, regional, and national burden of multiple
myeloma from 1990 to 2019. BMC Cancer. 2021;21(1):606.
3. Dimopoulos MA, Moreau P, Terpos E, et al. Multiple myeloma: EHA-ESMO clinical practice
guidelines for diagnosis, treatment and follow-up(†). Ann Oncol. 2021;32(3):309-322.
4. Moreau P, Attal M, Hulin C, et al. Bortezomib, thalidomide, and dexamethasone with or without
daratumumab before and after autologous stem-cell transplantation for newly diagnosed multiple
myeloma (CASSIOPEIA): a randomised, open-label, phase 3 study. Lancet. 2019;394(10192):29-38.
5. Morgan GJ, Gregory WM, Davies FE, et al. The role of maintenance thalidomide therapy in
multiple myeloma: MRC Myeloma IX results and meta-analysis. Blood. 2012;119(1):7-15.
6. van de Donk NW, van der Holt B, Minnema MC, et al. Thalidomide before and after autologous
stem cell transplantation in recently diagnosed multiple myeloma (HOVON-50): long-term results from
the phase 3, randomised controlled trial. Lancet Haematol. 2018;5(10):e479-e492.
7. McCarthy PL, Holstein SA, Petrucci MT, et al. Lenalidomide maintenance after autologous
stem-cell transplantation in newly diagnosed multiple myeloma: a meta-analysis. J Clin Oncol.
2017;35(29):3279-3289.
8. McCarthy PL, Owzar K, Hofmeister CC, et al. Lenalidomide after stem-cell transplantation for
multiple myeloma. N Engl J Med. 2012;366(19):1770-1781.
9. Jackson GH, Davies FE, Pawlyn C, et al. Lenalidomide maintenance versus observation for
patients with newly diagnosed multiple myeloma (Myeloma XI): a multicentre, open-label, randomised,
phase 3 trial. Lancet Oncol. 2019;20(1):57-73.
10. Cornell RF, D'Souza A, Kassim AA, et al. Maintenance versus induction therapy choice on
outcomes after autologous transplantation for multiple myeloma. Biol Blood Marrow Transplant.
2017;23(2):269-277.
11. Garderet L, Gras L, Koster L, et al. Global characteristics and outcomes of autologous
hematopoietic stem cell transplantation for newly diagnosed multiple myeloma: A study of the
worldwide network for blood and marrow transplantation (WBMT). Am J Hematol. 2024;99(11):2084-
2095.
12. R Core Team. R: A Language and Environment for Statistical Computing. 4.4.2. Vienna,
Austria: R Foundation for Statistical Computing, 2024. Available from: https://www.R-project.org/.
Accessed on 2026 Feb 01.
13. Pasvolsky O, Abid MB, Milton DR, et al. Outcomes of patients with multiple myeloma
undergoing autologous transplant with suboptimal pretransplant response. Br J Haematol.
2025;206(6):1730-1736.
14. Vij R, Kumar S, Zhang MJ, et al. Impact of pretransplant therapy and depth of disease response
before autologous transplantation for multiple myeloma. Biol Blood Marrow Transplant.
2015;21(2):335-341.
15. Sonneveld P, Schmidt-Wolf IG, van der Holt B, et al. Bortezomib induction and maintenance
treatment in patients with newly diagnosed multiple myeloma: results of the randomized phase III
HOVON-65/ GMMG-HD4 trial. J Clin Oncol. 2012;30(24):2946-2955.
16. Sonneveld P, Dimopoulos MA, Boccadoro M, et al. Daratumumab, bortezomib, lenalidomide,
and dexamethasone for multiple myeloma. N Engl J Med. 2024;390(4):301-313.
17. Corre J, Vincent L, Moreau P, et al. Daratumumab-bortezomib-thalidomide-dexamethasone for
newly diagnosed myeloma: CASSIOPEIA minimal residual disease results. Blood. 2025;146(6):679-
692.
15
Table 1. Demographic, Disease and Patient Characteristics.
Maintenance All
n (%)
Lenalidomide
n (%);
Thalidomide
n (%)
Bortezomib
n (%)
No maintenance
n (%)
Total patients 8269 (100.0) 4148 (50.2) 689 (8.3) 726 (8.8) 2706 (32.7)
Patients alive at 6 months landmark 7645(100.0) 3960 (51.8) 583 (7.6) 664 (8.7) 2438 (31.9)
ISS at diagnosis
I 2310 (36.2) 1150 (39.0) 154 (28.6) 171 (35.0) 835 (34.6)
II 2348 (36.8) 1065 (36.1) 187 (34.7) 176 (36.0) 920 (38.2)
III 1729 (27.1) 734 (24.9) 198 (36.7) 142 (29.0) 655 (27.2)
missing 1882 (22.8) 1199 (28.9) 150 (21.8) 237 (32.6) 96 (10.9)
Sex
Male 4696 (56.8) 2405 (58.0) 390 (56.6) 394 (54.3) 1507 (55.7)
Female 3570 (43.2) 1741 (42.0) 299 (43.4) 332 (45.7) 1198 (44.3)
MM classification
IgG 4658 (57.2) 2418 (59.0) 360 (53.7) 377 (53.1) 1503 (56.2)
IgA 1556 (19.1) 729 (17.8) 137 (20.4) 141 (19.9) 549 (20.5)
Light chain 1656 (20.3) 832 (20.3) 143 (21.3) 166 (23.4) 515 (19.3)
Other Ig 162 (2.0) 64 (1.6) 19 (2.8) 8 (1.1) 71 (2.7)
Non-secretory 117 (1.4) 53 (1.3) 12 (1.8) 18 (2.5) 34 (1.3)
missing 120 (1.5) 52 (1.3) 18 (2.6) 16 (2.2) 34 (1.3)
Diagnosis-AHCT Interval median (IQR) months 7.1 (5.6-10.0) 6.7 (5.3-9.2) 8.3 (6.2-12.0) 6.9 (5.4-9.6) 7.7 (5.9-10.8)
Year of AHCT
2013 1298 (15.7) 541 (13.0) 90 (13.1) 87 (12.0) 580 (21.4)
2014 1378 (16.7) 565 (13.6) 122 (17.7) 119 (16.4) 572 (21.1)
2015 1562 (18.9) 706 (17.0) 151 (21.9) 179 (24.7) 526 (19.4)
2016 1830 (22.1) 977 (23.6) 170 (24.7) 178 (24.5) 505 (18.7)
2017 2198 (26.6) 1357 (32.7) 156 (22.6) 163 (22.5) 522 (19.3)
Age at AHCT median
(IQR) years
60.0
(53.9-65.0)
60.6
(54.2-65.6)
58.6
(51.5-63.4)
60.5
(52.5-65.4)
60.0
(54.0-65.0)
(range) years (20.2-82.4) (20.2-82.4) (21-75.3) (30.3-79.3) (25-77.8)
Disease status at AHCT
CR 1620 (19.8) 671 (16.4) 228 (33.5) 165 (22.9) 556 (20.7)
VGPR 2961 (36.2) 1600 (39.2) 237 (34.9) 273 (37.9) 851 (31.6)
PR 3087 (37.8) 1602 (39.2) 170 (25.0) 247 (34.3) 1068 (39.7)
SD/MR 357 (4.4) 164 (4.0) 22 (3.2) 28 (3.9) 143 (5.3)
Rel/Prog/Prim. Refr. 135 (1.7) 44 (1.1) 23 (3.4) 6 (0.8) 62 (2.3)
Untreated 17 (0.2) 5 (0.1) 1 (0.1) 11 (0.4)
missing 92 (1.1) 62 (1.5) 9 (1.3) 6 (0.8) 15 (0.6)
Cytogenetics risk score at diagnosis
Standard 4636 (77.8) 2108 (73.2) 255 (81.0) 290 (63.3) 1983 (85.8)
High 1325 (22.2) 770 (26.8) 60 (19.0) 168 (36.7) 327 (14.2)
missing 2308 (27.9) 1270 (30.6) 374 (54.3) 268 (36.7) 396 (14.6)
Karnofsky PS at AHCT
100 2690 (34.1) 1237 (31.4) 257 (41.1) 199 (29.1) 997 (37.7)
≤90 5207 (65.9) 2708 (68.6) 368 (58.9) 485 (70.9) 1646 (62.3)
missing 372 (4.5) 203 (4.9) 64 (9.3) 42 (5.8) 63 (2.3)
HCT-CI risk group
Low (0) 4535 (59.7) 2068 (54.9) 395 (68.1) 373 (56.9) 1699 (65.6)
Intermediate (1-2) 1753 (23.1) 900 (23.9) 136 (23.4) 168 (25.6) 549 (21.2)
High (≥ 3) 1309 (17.2) 802 (21.3) 49 (8.4) 115 (17.5) 343 (13.2)
missing 672 (8.1) 378 (9.1) 109 (15.8) 70 (9.6) 115 (4.2)
Conditioning
Mel200 6119 (74.2) 3114 (75.3) 470 (68.2) 398 (55.0) 2137 (79.1)
Mel140 1132 (13.7) 591 (14.3) 98 (14.2) 134 (18.5) 309 (11.4)
Mel unknown dose 754 (9.1) 362 (8.8) 45 (6.5) 157 (21.7) 190 (7.0)
Other conditioning 243 (2.9) 68 (1.6) 76 (11.0) 34 (4.7) 65 (2.4)
Tandem protocol
No 7535 (91.2) 3793 (91.5) 558 (81.0) 652 (89.8) 2532 (93.6)
es 731 (8.8) 353 (8.5) 131 (19.0) 74 (10.2) 173 (6.4)
Abbreviations: CR, complete response; VGPR, Very good partial response; PR, partial response; SD/MR, Stable disease/minimal response; Rel/Prog/Prim. Refr., Relapse/Progression/Primary Refractory;
Mel200, Melphalan 200 mg/m2; Mel140, melphalan 140 mg/m2.
16
Table 2. Multivariable analysis of OS and PFS according to maintenance and disease stage at auto-HCT
adjusted for other baseline variables included in the model. The hazard ratio’s of maintenance therapy compared
no maintenance are shown in Figure 2.
OS (n=7,547) PFS (n=7,008)
HR (95% CI) (Overall)
p-value HR (95% CI) (Overall)
p-value
Maintenance (<0.0001) (<0.0001)
No maintenance 1.00 1.00
Lenalidomide 0.57 (0.48-0.68) <0.0001 0.65 (0.57-0.73) <0.0001
Thalidomide 0.76 (0.58-1.00) 0.05 0.90 (0.75-1.09) 0.29
Bortezomib 0.65 (0.50-0.84) 0.001 0.93 (0.78-1.09) 0.36
Lenalidomide vs. thalidomide 0.79 (0.61-1.02) 0.07 0.72 (0.60-0.85) <0.0001
Lenalidomide vs. bortezomib 0.88 (0.69-1.11) 0.27 0.68 (0.59-0.79) <0.0001
Disease stage at AHCT (<0.0001) (<0.0001)
CR 1.00 1.00
VGPR 1.32 (1.09-1.60) 0.004 1.23 (1.09-1.40) 0.0008
PR 1.63 (1.35-1.97) <0.0001 1.59 (1.41-1.80) <0.0001
SD/MR/Relapse/Progression 2.72 (2.10-3.50) <0.0001 1.81 (1.50-2.19) <0.0001
Age at AHCT (per 10 years older) 1.09 (1.01-1.17) 0.04 1.06 (1.01-1.11) 0.02
Year of AHCT (per year later) 1.01 (0.96-1.06) 0.71 0.97 (0.94-1.00) 0.08
Sex
Male 1.00 1.00
Female 0.88 (0.78-1.00) 0.05 0.91 (0.84-0.99) 0.02
Karnofsky PS at AHCT
100 1.00 1.00
≤90 1.23 (1.07-1.42) 0.004 1.01 (0.92-1.11) 0.79
MM classification (<0.0001) (0.007)
IgG 1.00 1.00
IgA 1.33 (1.14-1.56) 0.0003 1.19 (1.07-1.33) 0.001
Light chain 1.00 (0.85-1.19) 0.96 1.14 (1.03-1.27) 0.02
Other Ig 2.23 (1.60-3.10) <0.0001 1.41 (1.05-1.89) 0.02
Non-secretory 1.54 (0.92-2.59) 0.10 1.07 (0.74-1.54) 0.73
Conditioning dosage (0.44) (0.89)
Mel 200 1.00 1.00
Mel 140 0.99 (0.83-1.18) 0.89 0.99 (0.88-1.12) 0.92
Other conditioning 0.77 (0.48-1.23) 0.50 0.99 (0.75-1.31) 0.95
ISS at diagnosis (<0.0001) (<0.0001)
I 1.00 1.00
II 1.41 (1.18-1.70) 0.0002 1.28 (1.14-1.44) <0.0001
III 2.38 (1.99-2.85) <0.0001 1.80 (1.60-2.02) <0.0001
Cytogenetics
Normal risk 1.00 1.00
High risk 2.17 (1.86-2.54) <0.0001 1.63 (1.46-1.82) <0.0001
HCT-CI at AHCT (0.08) (0.53)
Low risk (0) 1.00 1.00
Intermediate risk (1-2) 1.01 (0.87-1.19) 0.87 1.06 (0.96-1.18) 0.23
High risk (≥3) 1.23 (1.01-1.46) 0.02 1.09 (0.97-1.23) 0.15
Models included a country specific random effect. Overall p-values were obtained using the likelihood ratio test and test whether in multicategorical
variables associations between categories and the outcome in question are all the same. Missing values were modeled using a m issing category
(not shown in this table). Abbreviations: OS, Overall survival; PFS, progression-free survival; CR, complete response; VGPR, very good partial
response; PR, partial response; SD, stable disease; MR, minor response; Prog, progression; CI, confidence interval; PS, performance score; AHCT,
autologous hematopoietic stem cell transplantation; HCT-CI, hematopoietic stem cell transplantation comorbidity index; ISS, International staging
system; Ig, Immunoglobulin
17
Legend to Figures.
Figure 1. Kaplan Meier-plots for overall survival (OS) and progression free survival (PFS): a) OS by
Response/disease status at AHCT, b) OS by maintenance, c) PFS by Response/disease status at
AHCT, and d) PFS by maintenance. Shaded areas show the 95% confidence intervals. Numbers
below the graphs show the number of patients still at risk. Abbreviations: AHCT, autologous
hematopoietic stem cell transplantation, CR, complete response; VGPR, very good partial response;
PR, partial response; SD, stable disease; MR, minor response; Prog, progression
Figure 2. Time dependent hazard ratios of a) OS and b) PFS of maintenance therapies compared to no
maintenance modelled over the follow-up period. Dashed lines show the 95% confidence intervals.
Figure 3. Predicted OS and PFS probabilities for reference patients receiving either lenalidomide,
bortezomib, thalidomide or no maintenance according to disease stage at AHCT. The predictions
were made using the results presented in Table 2. We assumed the patient was a male MM patient, with
standard cytogenetic risk score and an age of 60 years at AHCT (in 2017), KPS ≤90 at AHCT, IgG
classification, conditioning with Mel 200, ISS I, a low risk HCT-CI and a country random effect of 0.
1
Supplement:
Global Participating Registries
1. The European Society for Blood and Marrow Transplantation (EBMT; www.ebmt.org). Information on
maintenance in this group did not distinguish between missing or no maintenance given, therefore none of
the EBMT patients were included in the no maintenance category.
2. The Center for International Blood and Marrow Transplantation (CIBMTR; www.cibmtr.org), the
United States of America (USA),
3. The Canada registry using the Ottawa Blood Disease Center MM Database (OBDCMMD),
4. Latin American Blood and Marrow Transplantation group (LABMT).
5. The Asian Pacific Blood and Marrow Transplant Group (APBMT; www.apbmt.org) with reporting
registries
a. Myeloma Transplant Registry, Ministry of Health, Malaysia (MTRMOHM)
b. Japan Society for Transplantation and Cellular Therapy/ Japanese Data Center for
Hematopoietic Cell Transplantation (JSTCT/JDCHCT)
c. Taiwan Society of Blood and Marrow Transplantation (TBMT)
d. Beijing Bone Marrow Transplant registry
6. Eastern Mediterranean Blood and Marrow Transplant Group (EMBMT).
Supplementary
2
Supplement to Methods
Variables included in multivariable analyses and their definitions
Apart from maintenance and disease stage at AHCT models included patient sex, International Staging
System (ISS) at diagnosis, high-risk cytogenetic (t(4;14) and/or del 17p and/or t(14;16) and in Europe
t(4;14) and/or del 17p and/or t(14;16) t(14;20) and/or hypodiploid and/or 1q gain and/or deletion 1p),
myeloma sub-classification, year of AHCT, time from diagnosis to AHCT, Karnofsky performance status
(KPS), HCT-specific comorbidity index (HCT-CI) and melphalan conditioning dose.
Regarding response status prior to AHCT, CR, very good partial response (VGPR), partial response (PR),
minor response (MR)/stable disease (SD), and Rel / Prog were defined according to the International
Myeloma Working Group criteria [11]. Conditioning was categorized as either melphalan 140 mg/m2, 200
mg/m2, and “others” if melphalan was used in combination with additional drugs. We included interaction
terms between maintenance treatment and disease stage at AHCT in OS and PFS models to check whether
the association of maintenance and outcome was similar across all disease stage. We analyzed whether the
association between the maintenance and outcome was constant over time (the proportional hazards
assumption) by including an interaction term between maintenance and follow-up time in the models using
the tt function part of the coxph function in R. In these models the country random effect was first estimated
without the time-dependent coefficients and subsequently included in models with time dependent
coefficients as a fixed variable using the offset statement.
3
Supplemental Table S1
Registry Contribution
Maintenance All
n (%)
Lenalidomide
n (%);
Thalidomide
n (%)
Bortezomib
n (%)
No maintenance
n (%)
Total patients 8269 (100.0) 4148 (50.2) 689 (8.3) 726 (8.8) 2706 (32.7)
Patients alive at 6 months
landmark 7645(100.0) 3960 (51.8) 583 (7.6) 664 (8.7) 2438 (31.9)
Japan 2866 (34.7) 756 (18.2) 17 (2.5) 96 (13.2) 1997 (73.8)
% of registry 100.0 26.9 0.6 3.3 69.7
EBMT 2691 (32.6) 1969 (47.5) 335 (48.6) 387 (53.3)
% of registry 100.0 (73.2)* (12.4)* (14.4)*
CIBMTR 1841 (22.3) 1218 (29.4) 4 (0.6) 180 (24.8) 439 (16.2)
% of registry 100.0 66.2 0.2 9.8 23.8
EMBMT 54 (0.7) 37 (0.9) 9 (1.3) 8 (1.1)
% of registry 100.0 68.5 16.7 14.8
Taiwan 158 (1.9) 6 (0.1) 152 (22.1)
% of registry 100.0 3.8 96.2
LABMT 261 (3.2) 67 (1.6) 105 (15.2) 43 (5.9) 46 (1.7)
% of registry 100.0 25.7 40.2 16.5 17.6
Canada 168 (2.0) 75 (1.8)
93 (3.4)
% of registry 100.0 44.6 55.4
Malaysia 166 (2.0) 7 (0.2) 36 (5.2) 10 (1.4) 113 (4.2)
% of registry 100.0 4.2 21.7 6.0 68.1
China 61 (0.7) 11 (0.3) 31 (4.5) 2 (0.3) 17 (0.6)
% of registry 100.0 18.0 50.8 3.3 27.9
Abbreviations: AHCT, autologous hematopoietic cell transplantation; EBMT, European Bone Marrow Transplant; CIBMTR, Center for
International Blood and Marrow Transplant; EMBMT, Eastern Mediterranean Blood and Marrow Transplantation; LABMT, Latin American
Bone Marrow Transplantation;
4
Supplemental Table S2. Estimated OS, PFS and cumulative incidence of relapse and NRM at 3 year after 6 month landmark, according to
maintenance groups and disease stage at auto-HCT. Log-rank p-values (OS and PFS) and Gray’s test p-values (relapse and
NRM) apply to the whole length of the curves, not just to the 3-year estimate.
Overall CR VGPR PR SD/MR/Prog
n % (95% CI) n % (95% CI) n % (95% CI) n % (95% CI) n % (95% CI)
OS at 3 years
Total 7645 83 (82-84) 1498 87 (85-89) 2747 84 (83-86) 2882 82 (80-83) 434 74 (69-78)
Lenalidomide 3960 86 (85-87) 634 91 (88-93) 1533 86 (84-88) 1539 85 (83-87) 196 81 (75-87)
Thalidomide 583 80 (76-84) 188 78 (70-85) 188 84 (78-90) 155 81 (74-88) 44 72 (56-87)
Bortezomib 664 83 (80-86) 154 89 (84-95) 245 85 (80-90) 229 77 (70-83) 32 86 (73-99)
None 2438 79 (77-80) 522 84 (81-88) 781 80 (76-83) 959 78 (75-81) 162 63 (55-71)
p-value <0.001 <0.001 <0.001 <0.001 0.05
PFS at 3 years
Total 7098 60 (59-62) 1407 67 (64-70) 2578 62 (60-64) 2667 57 (54-59) 368 52 (47-58)
Lenalidomide 3797 63 (61-65) 612 70 (66-74) 1477 65 (62-68) 1467 59 (56-62) 186 57 (49-65)
Thalidomide 553 53 (48-57) 179 53 (44-62) 181 54 (45-62) 145 50 (42-59) 40 45 (28-63)
Bortezomib 625 52 (47-56) 150 71 (63-79) 233 50 (43-57) 209 41 (34-49) 30 48 (29-67)
None 2123 60 (58-63) 466 67 (63-72) 687 61 (57-65) 846 57 (54-61) 112 49 (39-59)
p-value <0.001 0.002 <0.001 <0.001 0.30
Relapse cumulative incidence at 3 years
Total 7098 36 (34-37) 1407 29 (26-32) 2578 35 (33-37) 2667 39 (37-41) 368 43 (38-49)
Lenalidomide 3797 34 (33-36) 612 28 (24-32) 1477 33 (30-36) 1467 38 (35-40) 186 44 (36-53)
Thalidomide 553 45 (40-49) 179 41 (32-49) 181 43 (35-52) 145 50 (41-58) 40 52 (34-70)
Bortezomib 625 45 (41-49) 150 27 (20-35) 233 45 (38-52) 209 56 (48-63) 30 57 (37-76)
None 2123 33 (31-35) 466 27 (22-31) 687 33 (29-37) 846 36 (32-39) 112 59 (45-74)
p-value <0.001 0.04 <0.001 <0.001 0.30
NRM cumulative incidence at 3 years
Total 7098 4 (3-5) 1407 4 (3-5) 2578 3 (3-4) 2667 5 (4-5) 368 4 (2-6)
Lenalidomide 3797 3 (2-3) 612 2 (1-4) 1477 2 (1-3) 1467 4 (3-5) 186 2 (0-5)
Thalidomide 553 3 (1-4) 179 7 (2-11) 181 3 (0-5) 145 0 (0-0) 40 0 (0-0)
Bortezomib 625 3 (2-5) 150 2 (0-4) 233 5 (2-7) 209 3 (0-6) 30 4 (0-11)
None 2123 6 (5-8) 466 6 (3-8) 687 6 (4-8) 846 7 (5-9) 112 9 (3-14)
p-value <0.001 0.01 <0.001 0.04 0.030
Abbreviations: CR, complete response; VGPR, very good partial response; PR, partial response; SD, stable disease; MR, minor response; Prog, progression; CI, confidence interval;
OS, overall survival; PFS, progression-free survival; NRM, non-relapse mortality.
5
Supplemental Table S3. Time dependent hazard ratio’s (HR) of maintenance therapies vs. no maintenance and 95% CI’s after the 6-month
landmark.
Lenalidomide vs. no maintenance Thalidomide vs. no maintenance Bortezomib vs. no maintenance
Time point
after 6-month
landmark
OS HR
(95% CI)
p-
value*
(<0.001)
PFS HR
(95% CI)
p-
value*
(0.02)
OS HR
(95% CI)
p-
value*
(0.04)
PFS HR
(95% CI)
p-
value*
(0.01)
OS HR
(95% CI)
p-
value*
(0.02)
PFS HR
(95% CI)
p-
value*
(0.002)
0 months 0.25
(0.17-0.37)
<0.001 0.36
(0.29-0.45)
<0.001 0.35
(0.17-0.72)
0.005 0.58
(0.40-0.84)
0.004 0.41
(0.21-0.81) 0.01
0.56
(0.40-0.79)
0.001
6 months 0.40
(0.32-0.50)
<0.001 0.53
(0.47-0.60)
<0.001 0.54
(0.37-0.81)
0.003 0.78
(0.64-0.95)
0.01 0.54
(0.37-0.77) <0.001
0.78
(0.65-0.94)
0.009
12 months 0.53
(0.46-0.62)
<0.001 0.67
(0.60-0.74)
<0.001 0.72
(0.55-0.93)
0.01 0.93
(0.79-1.09)
0.37 0.63
(0.49-0.80) <0.001
0.95
(0.82-1.11)
0.52
18 months 0.63
(0.55-0.74)
<0.001 0.77
(0.68-0.86)
<0.001 0.84
(0.66-1.09)
0.19 1.03
(0.85-1.25)
0.74 0.69
(0.54-0.89) 0.004
1.08
(0.91-1.28)
0.41
24 months 0.70
(0.60-0.84)
<0.001 0.84
(0.73-0.95)
0.008 0.93
(0.70-1.24)
0.64 1.10
(0.88-1.37)
0.39 0.74
(0.55-0.98) 0.04
1.16
(0.95-1.41)
0.15
30 months 0.75
(0.62-0.91)
0.003 0.88
(0.76-1.02)
0.08 0.99
(0.72-1.36)
0.96 1.15
(0.90-1.46)
0.27 0.76
(0.56-1.05) 0.09
1.21
(0.97-1.51)
0.09
36 months 0.78
(0.64-0.96)
0.02 0.91
(0.78-1.06)
0.22 1.03
(0.73-1.45)
0.86 1.17
(0.91-1.51)
0.22 0.78
(0.56-1.09) 0.15
1.24
(0.99-1.57)
0.07
6
Supplemental Figure S1. Overall survival (OS) and progression-free survival (PFS) after autologous
hematopoietic stem cell transplantation. Numbers below the graph indicate the number at risk. Shaded areas the
95% confidence intervals.
7
Supplemental Figure S2. Landmark analysis of OS (a) and PFS (b) for maintenance lenalidomide vs no maintenance by IMWG response (***, p<0.001;
**, p<0.05; *, p≥0.05). P-values apply to the whole length of the curve 0-36 months not just at 36 months. Numbers below the graphs show the number of
patients still at risk.