Long-term outcome results of the first tandem autotransplant trial for multiple myeloma

Authors


Dr Bart Barlogie, 4301 W. Markham, #816, Little Rock, AR 72205, USA.
E-mail: BarlogieBart@uams.edu

Summary

Total Therapy 1, the first tandem autotransplant trial for newly diagnosed patients with multiple myeloma, was designed to increase the frequency of complete response (CR) and thereby extend survival. With a median follow-up of 12 years, 62 of 231 initially enrolled patients are alive (17% at 15 years); 31 remain event free (7% at 15 years) including 16 of 94 (41%) that initially achieved CR. Currently alive patients less frequently had cytogenetic abnormalities (CAs) at baseline (P = 0·002), postenrolment (P < 0·001) and at relapse (P = 0·004); elevations of serum C-reactive protein (CRP) (P = 0·003) and lactate dehydrogenase (P = 0·029), anaemia (P = 0·029) and they more often completed two transplants within 12 months (P = 0·019). Postenrolment overall survival (OS) and event-free survival (EFS) were superior in the absence of CA of the hypodiploidy or deletion 13 variety (P < 0·001 and 0·037 respectively) and in the presence of low CRP at baseline (P = 0·001 and 0·017 respectively). Postrelapse survival was longer in the absence of CA at relapse (P < 0·001), IgA isotype (P = 0·002), International Staging System stage 3 (P = 0·014), and when patients had two protocol transplants prior to relapse (P = 0·038). Ten-year EFS and OS could be accomplished in 15% and 33% of patients, respectively, when all agents available in 1989, especially high-dose melphalan, were applied together upfront for the management of myeloma.

Melphalan (MEL) was introduced in 1962 as the first agent with significant palliative potential for patients with multiple myeloma (MM) (Bergsagel et al, 1962; Alexanian et al, 1969). Numerous large prospective randomised trials have since been conducted, collectively indicating that multi-agent combination chemotherapy failed to prolong the median survival of 30–36 months achieved with standard MEL–prednisone (MP) (Gregory et al, 1992). Dose escalation of glucocorticoids with dexamethasone in the VAD regimen (vincristine, doxorubicin, dexamethasone) (Barlogie et al, 1984) and subsequently of MEL provided meaningful salvage therapy for patients relapsing and refractory to MP (McElwain & Powles, 1983). The introduction of autografts, especially chemotherapy- or growth factor-mobilised peripheral blood stem cells (PBSCs), permitted the safe administration, also in patients over the age of 65 years, of myeloablative doses of MEL alone or combined with total body irradiation (TBI), effecting complete responses (CRs) in 10–20% of patients with advanced (Barlogie et al, 1987) and in more than 20% with newly diagnosed MM (Barlogie et al, 1986; Gianni et al, 1989; Fermand et al, 1993; Harousseau et al, 1995). We now report the long-term outcome of initially 231 newly diagnosed patients enrolled into Total Therapy 1 (TT1), the first tandem transplant trial for MM (Barlogie et al, 1997).

Patients and methods

Details of the treatment regimen have been reported previously, together with eligibility criteria, response and relapse definitions and statistical methods applied (Barlogie et al, 1997). Briefly, induction therapy consisted of VAD for three cycles, high-dose cyclophosphamide 6 g/m2 plus granulocyte–macrophage colony-stimulating factor for PBSC collection and EDAP (etoposide, dexamethasone, cytosine arabinoside, cisplatin). This was followed by two cycles of high-dose therapy with MEL 200 mg/m2, spaced 3–6 months apart; in cases that achieved less than a partial response (PR) with the first transplant, TBI (850 cGy/m2) was added to MEL 140 mg/m2 with the second transplant.

With the present median follow-up of 12 years and continuing annual re-evaluation in the currently 62 surviving patients, this report seeks to provide a long-term outcome analysis. We have addressed the baseline features of currently event-free and alive patients in the context of the whole population. Outcome was also evaluated in terms of the availability of thalidomide, introduced in March 1998 for salvage therapy (Singhal et al, 1999). The Kaplan–Meier method was used in estimating event-free and overall survival (OS) (Kaplan & Meier, 1958), while group comparisons were made using the log-rank test (Mantel, 1966). Multivariate models of prognostic factors were carried out using Cox regression (Cox et al, 1972). P-values <0·05 were considered significant

Results

Table I depicts patient characteristics at baseline. Table II documents the progression of patients through the four phases of induction, first and second transplant and post-transplant maintenance therapy with interferon. Off-study reasons are also indicated. Eventually, 82% achieved a stringently defined partial response (PR, ≥75% M-protein reduction in serum and ≥90% M-protein reduction in urinary light chain excretion); 41% achieved CR (immunofixation negative and normal bone marrow aspirate and biopsy). First and second transplants were completed in 84% and 71% of patients within medians of 7 and 12 months from the start of protocol therapy; interferon maintenance was initiated in 55% within a median of 17 months from enrolment (Fig 1). The proportions of patients surviving and remaining event free, together with the subset in continuous CR, are portrayed as Kaplan–Meier plots in Fig 2: at 10 and 15 years, respectively, 33% and 17% are alive; 15% and 7% are event free; and 18% and 12% of those achieving CR remain in continuous CR. Of the 231 study patients, 169 have died (76 from myeloma, 93 from other causes) and 200 have suffered an event (disease recurrence in 158, death in 42 without receiving salvage therapy). Of the 62 patients alive as of 1 March 2006, 31 have remained continuously relapse free.

Table I.   Patient baseline characteristics (n = 231).
Factorn/N (%)
  1. n, number with factor; N, number with valid data for factor; ISS, International Staging System.

Age ≥ 65 years21/231 (9)
IgA isotype41/231 (18)
Male gender143/231 (62)
C-reactive protein ≥ 4·0 mg/l104/223 (47)
B-2-microglobulin ≥ 4·0 mg/l69/229 (30)
Bone marrow plasmacytosis ≥ 30%135/231 (58)
Albumin < 35 g/l62/231 (27)
Haemoglobin < 10 g/dl78/231 (34)
Creatinine ≥ 176·8 μmol/l22/231 (10)
Lactate dehydrogenase ≥ 190 U/l49/230 (21)
ISS stage 1110/229 (48)
ISS stage 276/229 (33)
ISS stage 343/229 (19)
Without 1 prior cycle of therapy155/231 (67)
Any cytogenetic abnormality74/222 (33)
Hypodiploidy or deletion 1335/222 (16)
Completed first transplant195/231 (84)
Completed second transplant165/231 (71)
Table II.   Progression of patients through study phases and accounting for off study reasons.
PhaseN (%)Number coming off study
TotalToxicityRelapseDeath (TRM)Other%PR%CR
  1. NA, not available; TRM, transplant-related mortality; PR, partial response; CR, complete response.

Enrolment231 (100)NA0000NANA
Induction231 (100)36868 (6)146312
First transplant195 (84)30992 (2)107626
Second transplant165 (71)371626 (3)138136
Maintenance128 (55)NA08315 (2)218241
Figure 1.

 Time to completion of treatment steps: 84% completed first and 71% completed second transplant within median times of 7 and 12 months; 55% started interferon maintenance.

Figure 2.

 Clinical outcome of patients: Kaplan–Meier plots of overall survival (left panel), event-free survival (middle panel) and complete remission duration (right panel).

The details of further therapy could be identified for 96 of the 158 relapsing patients. Non-transplant therapies (applied at any time, not necessarily as primary salvage treatment) included thalidomide alone or in combination in 46%, combination chemotherapy in 23% and bortezomib alone or in combination in 14%. Altogether, 71 patients (74% of the 96 with known salvage therapy) received further transplants, including 30 who eventually had three and five who had four transplants. Further transplants represented the initial salvage modality in 49 patients. The median postrelapse survival among the 158 patients was 33 months, similar to the median event-free survival (EFS) of 31 months from the initiation of therapy; 12% were still alive 12 years later (Fig 3A). The 25 patients who had their first relapse in the thalidomide era had a superior 5-year postrelapse survival of 59% vs. 31% in the remainder (Fig 3B, P = 0·004). Such patients had several more favourable baseline features, such as a lower frequency of hypodiploidy or deletion 13-type cytogenetic abnormalities (CAs); they were less frequently anaemic, had lower levels of beta-2-microglobulin (B2M) and C-reactive protein (CRP), and fewer had International Staging System (ISS) stage 3.

Figure 3.

  Kaplan–Meier plot of postrelapse survival: (A) all patients – the median time of 33 months from treatment failure to death matches the median time of 31 months from study enrolment to treatment failure; (B) according to timing of availability of thalidomide for salvage therapy: superior outcome was noted among the patients who first relapsed after March 1998 (when thalidomide became available as salvage therapy).

Multivariate analysis was performed of parameters associated with OS and EFS from protocol enrolment and with survival from relapse (Table III). CA of the hypodiploidy or deletion 13-variety and high CRP were both associated with poor OS and EFS. ISS stage 3 was an additional high-risk feature for OS, as was high B2M and lactate dehydrogenase (LDH) for EFS. A time-dependent covariate for time to CR was associated with longer EFS. Postrelapse survival was shorter in cases with IgA isotype, ISS stage 3, and when CA was identified within 6 months of relapse; completion of two transplants prior to relapse was associated with longer postrelapse survival. Relapse in the thalidomide era was only of borderline significance after accounting for the better prognostic features noted in such patients (see above).

Table III.   Multivariate analysis of parameters associated with overall survival and event-free survival from the start of therapy and with survival from relapse.
Model*n (%)HR (95% CI)P-value
  1. HR, hazard ratio; 95% CI, 95% confidence interval; ISS, International Staging System; B2M, beta-2-microglobulin; CRP, C-reactive protein; LDH, lactate dehydrogenase; CA, cytogenetic abnormalities; P-value from Wald chi-square test in Cox regression.*Variables selected by stepwise regression are shown. Factors considered for overall and event-free survival analysis: Complete response (CR), two transplants (as time-dependent covariates) and binary indicators of IgA isotype, age, CRP, B2M, bone marrow plasmacytosis, albumin, haemoglobin, creatinine, LDH, ISS stage 3, prior treatment, baseline cytogenetic abnormalities (any, hypodiploidy/deletion 13), with cut-off values for these indicators as listed in Table 1. The analysis of overall survival from progression/relapse included the above variables in addition to time to CR (not time dependent), indicators of 2 transplants, CR prior to first or second transplant, cytogenetic abnormalities (any, hypodiploidy/deletion 13) postenrolment and at the time of progression/relapse.

Overall survival (N = 215 of 231 with all data)
 Hypodiploidy or deletion 13 at baseline35 (16)2·62 (1·76–3·90)<0·001
 CRP ≥ 4·0 mg/l103 (48)1·70 (1·23–2·36)0·001
 ISS Stage 341 (19)1·58 (1·06–2·35)0·023
Event-free survival (N = 214 of 231 with all data)
 B2M ≥ 4·0 mg/l64 (30)1·59 (1·15–2·21)0·005
 CRP ≥ 4·0 mg/l103 (48)1·45 (1·07–1·96)0·017
 Complete response (time dependent)92 (43)0·70 (0·51–0·96)0·027
 LDH ≥ 190 U/l46 (21)1·50 (1·04–2·17)0·030
 Hypodiploidy/deletion 13 at baseline35 (16)1·52 (1·03–2·26)0·037
Overall survival from progression/relapse (N = 135 of 158 with all data)
 CA at relapse/progression52 (39)3·45 (2·39–5·18)<0·001
 IgA isotype26 (19)2·14 (1·33–3·43)0·002
 Completed second transplant104 (77)0·62 (0·40–0·97)0·038
 Relapse after thalidomide available22 (15)0·51 (0·25–1·04)0·063
 ISS stage 331 (23)1·78 (1·12–2·81)0·014

Distinguishing baseline features for patients currently alive, event free or in continuous CR are summarised in Table IV. The absence of CA at baseline and at any time thereafter, lower frequencies of anaemia, of elevated serum levels of CRP and LDH, as well as application of two transplants within 12 months of each other were characteristics of long-term survivors; low B2M levels, but not anaemia or elevated LDH, were also significant in patients remaining event free. None of these parameters significantly distinguished patients with continuous CR. The laboratory parameters obtained in the past 12 months in the majority of patients revealed normal platelet and uninvolved immunoglobulin levels in two-thirds and a normal bone marrow aspirate and biopsy in nearly 80% of survivors; focal lesions on magnetic resonance imaging studies were absent in one-half of survivors and three-quarters of those still event free (data not shown). The far-reaching implications of CA at baseline are portrayed for OS in Fig 4: in the absence of CA (67%), 15-year projections were 20%; nil in the presence of hypodiploidy or deletion 13 (16%); with intermediate outcome for those with ‘other’ CA (18%).

Table IV.   Distinguishing features of long-term survivors.
FactorAliveDeadP-valueNo eventEventP-value
  1. Values in parenthesis are expressed in percentage.

  2. n/N (%): n, number with factor; N, number with valid data for factor; CRP, C-reactive protein; ISS, International Staging System; B2M, beta-2-microglobulin; CR, complete response; LDH, lactate dehydrogenase.

  3. *Fisher exact test, otherwise chi-square test.

CRP ≥ 4·0 mg/l18/60 (30)86/163 (53)0·0036/29 (21)98/194 (51)0·003
B2M ≥ 4·0 mg/l13/61 (21)56/168 (33)0·0804/30 (13)65/199 (33)0·031
LDH ≥ 190 U/l7/61 (11)42/169 (25)0·0293/30 (10)46/200 (23)0·105
Hb < 10 g/dl14/62 (23)64/169 (38)0·0297/31 (23)71/200 (36)0·157
Bone marrow plasmacytosis  ≥ 30%30/62 (48)105/169 (62)0·06015/31 (48)120/200 (60)0·222
Age ≥ 65 years3/62 (5)18/169 (11)0·1731/31 (3)20/200 (10)0·222*
IgA isotope7/62 (11)34/169 (20)0·1204/31 (13)37/200 (19)0·448
Creatinine ≥ 176·8 μmol/l4/62 (6)18/169 (11)0·3352/31 (6)20/200 (10)0·531*
Albumin < 35 g/l17/62 (27)45/169 (27)0·9048/31 (26)54/200 (27)0·889
ISS stage 135/61 (57)75/168 (45)0·08819/30 (63)91/199 (46)0·072
ISS stage 218/61 (30)58/168 (35)0·4767/30 (23)69/199 (35)0·219
ISS stage 38/61 (13)35/168 (21)0·1864/30 (13)39/199 (20)0·413
Without 1 prior cycle of therapy42/62 (68)113/169 (67)0·90023/31 (74)132/200 (66)0·366
Second transplant within 12 months (patients with two transplants)43/51 (84)76/114 (67)0·01923/24 (96)96/141 (68)0·005
Indicator of confirmed CR29/62 (47)65/169 (380·25416/31 (52)78/200 (39)0·183
Cytogenetic abnormalities at baseline10/59 (17)64/163 (39)0·0024/29 (14)70/193 (36)0·017
Hypodiploidy/deletion 13 at baseline2/59 (3)33/163 (20)0·0022/29 (7)33/193 (17)0·160
Cytogenetic abnormalities postenrolment18/60 (30)88/160 (55)<0·0016/29 (21)100/191 (52)0·001
Cytogenetic abnormalities at relapse/progression2/25 (8)46/122 (38)0·004NA53/158 (34)NA
Figure 4.

 Kaplan–Meier survival plots from study enrolment according to the presence of certain cytogenetic abnormalities (CAs): survival was superior among patients presenting without CA, poorest in the presence of hypodiploidy or chromosome 13 deletion (CA13/Hypo) and intermediate in the presence of ‘other CA’.

Discussion

The results presented here support the notion that dose-intensified MEL, to the level of bone marrow aplasia requiring haematopoietic cellular support, can accomplish 10- and 15-year survival rates in a substantial proportion of patients, one-half of whom have not yet relapsed, including one-fifth (17%) of the initial 41% who are enjoying continuous complete remission. Both the original publication of TT1 data (Barlogie et al, 1997) and the current long-term update support the existence of myeloma sub-entities best captured by cytogenetics, historically (Tricot et al, 1995; Desikan et al, 2000), and now even better by molecular genetics (Avet-Loiseau et al, 2002; Zhan et al, 2002). Although CA was associated with inferior outcome overall, a fraction of such patients enjoyed prolonged survival.

Total Therapy 1 pilot data have since been confirmed in the InterGroupe Francophone du Myelome (IFM)94 trial, which prospectively randomised patients to single versus tandem transplants (Attal et al, 2003). Long-term data of TT1 and IFM94 trials (tandem versus single transplant; Attal et al, 2003) are depicted in Fig 5. Collectively, these results underscore the principal feasibility of 10-year survivorship in 30% of cases, setting a new standard of myeloma outcome against which new treatments need to be measured. The advent of several new classes of agents effective in the post-transplant salvage setting [immunomodulatory agents thalidomide (Singhal et al, 1999) and lenalidomide (Richardson et al, 2002), and the proteasome inhibitor bortezomib (Richardson et al, 2003)] holds great promise for further prolongation of disease control and cure when appropriately used in combination (Facon et al, 2005; Mateos et al, 2005; Palumbo et al, 2006).

Figure 5.

 Kaplan–Meier plots of survival of patients treated with Total Therapy 1 (TT1) in the context of IFM94 trial results: TT1 survival was similar to survival for the IFM94 tandem transplant arm.

Although not independently significant, the superior survival of patients relapsing after thalidomide became available is interesting in the context of our recent data from Total Therapy 2 (TT2, randomising patients to receive thalidomide or not upfront, in addition to intensive therapy including tandem transplantation) (Barlogie et al, 2006). In the TT2 study, patients relapsing on thalidomide responded more poorly to all types of salvage therapy and had a significantly shorter postrelapse survival compared with those who had not received thalidomide previously. If having had access to thalidomide (i.e. TT1 relapse after March 1998) could be considered analogous to having access to but not taking the drug for induction (TT2 patients randomised to no thalidomide), it could be argued that the best way to use thalidomide is not for induction therapy but for either maintenance therapy, as done in the recent IFM99 study (Attal et al, 2005), or for relapsed disease, as done by us. This would also suggest that the widespread use of thalidomide–dexamethasone as induction therapy could prove detrimental in the long run; a supposition underpinned by the recent Eastern Cooperative Oncology Group study, demonstrating comparable 2-year survival with thalidomide–dexamethasone and dexamethasone alone despite higher response rates with the former (Rajkumar et al, 2006).

Acknowledgements

This work was supported in part by a programme project grant from the National Cancer Institute (CA55819). The authors thank referring physicians and patients for their trust.

Ancillary