Myeloma Comparison of new patients with Bence-Jones, IgG ... - Nature

11 downloads 78 Views 130KB Size Report
Comparison of new patients with Bence-Jones, IgG and IgA myeloma receiving sequential .... Patients were admitted to hospital on the first day of HDT and the ...
Bone Marrow Transplantation (2001) 28, 29–37  2001 Nature Publishing Group All rights reserved 0268–3369/01 $15.00 www.nature.com/bmt

Myeloma Comparison of new patients with Bence-Jones, IgG and IgA myeloma receiving sequential therapy: the need to regard these immunologic subtypes as separate disease entities with specific prognostic criteria B Sirohi1, R Powles1, S Kulkarni1, C Rudin1, R Saso1, R Lal1, S Singhal1,2 J Mehta1,2, C Horton1 and J Treleaven1 1

Leukaemia and Myeloma Units, Royal Marsden NHS Trust, Sutton, Surrey, UK

Summary: Of the 61 newly diagnosed patients with Bence-Jones (BJ) myeloma presenting to our centre between May 1986 and December 1997, 53 received sequential therapy (ST) comprising infusional chemotherapy (IC) followed by highdose therapy and autotransplantation with interferon␣2b maintenance. The outcome was compared with 153 IgG and 39 IgA similarly treated myeloma patients. Response to IC and high-dose was comparable between the three subtypes but a significantly higher proportion of patients with BJ myeloma failed to receive high-dose compared to IgG (P = 0.003) and IgA (P = 0.04) myeloma. Median overall survival (OS) of patients with BJ myeloma (2.8 years) and event-free survival (EFS, 1.2 years) was significantly shorter than for patients with IgG myeloma (4.5 years, P = 0.03 and 2.1 years, P = 0.03, respectively). However, among those patients who achieved complete remission there was no difference in OS and EFS between IgG and BJ myeloma. In distinction to IgG myeloma where age and ␤2M were significant, Cox analysis on presentation features identified performance status and urine total protein as having significant impact on OS. We conclude that achieving CR is an important treatment aim in patients with BJ myeloma, conferring a similar outlook on survival as in patients with the IgG subtype, and there is a need to consider different subtypespecific staging systems when evaluating the results of published or ongoing therapeutic trials. Bone Marrow Transplantation (2001) 28, 29–37. Keywords: Bence-Jones; sequential therapy; separate entity; prognostic criteria

Since 1983, intensive treatment schedules for multiple myeloma have found increasingly widespread acceptance.

Correspondence: Prof R Powles, Head, Leukemia and Myeloma Units, Royal Marsden NHS Trust, Downs Road, Sutton, Surrey, SM2 5PT, UK 2 Division of Transplantation Medicine, South Carolina Cancer Center and Richland Memorial Hospital, Seven Richland Medical Park, Columbia, SC 29203, USA Received 26 January 2001; accepted 19 April 2001

Regimens consisting of initial infusional chemotherapy with vincristine, doxorubicin and steroids (C-VAMP/VAD) followed by high-dose melphalan with peripheral stem cell rescue and maintenance interferon-␣2b1,2 have come to be considered the benchmark for the treatment of many patients with multiple myeloma. The currently used prognostic criteria and staging systems for multiple myeloma were designed before the era of intensive sequential therapy (ST) and allogeneic transplants1,3–5 and may therefore no longer be valid. We have previously shown that in the setting of modern intensive treatments IgG myeloma is a distinct disease entity that lends itself to a new, easily reproducible staging system.6 In the present analysis we compared the response of a cohort of Bence-Jones (BJ) myeloma to ST with that of patients with IgG and IgA disease and evaluated the BJ myeloma patients for possible prognostic criteria that could lead to a similar staging system for this immunological subtype of myeloma. We chose to include only BJ, IgG and IgA myeloma in the study because they have objective endpoints for assessing disease response. Together they comprise 98% (20%, 53% and 25%, respectively) of all myeloma patients. Achieving a complete biochemical and morphological remission conferred the same survival advantage to BJ patients as to IgG patients and removed the difference in survival between the two groups. While in the context of intensive ST, ␤2M, age and Hb significantly influence the prognosis of IgG myeloma,6 we show here that the independent variables for outcome in BJ myeloma are performance status, urine total protein concentration and the addition of cyclophosphamide to the infusional chemotherapy. Thus it appears that the outlook of different immunologic subtypes of multiple myeloma is influenced by different variables at presentation and that they should not be grouped together as a single disease entity when assessing the effectiveness of treatment protocols.

Materials and methods Study population For this retrospective study we identified 61 newly diagnosed, consecutive patients with BJ myeloma from a pro-

Bence-Jones myeloma: a separate entity B Sirohi et al

30

spectively maintained database at the Royal Marsden Hospital. They were referred for ST to our unit between May 1986 and December 1997. According to the referral pattern at the time they represent a population-based group.1 Of these 61 patients, 53 went on to receive ST. They consented to therapy according to the care plan in place at the time. This was approved by the institutional review board. Eight patients did not receive ST. In seven this was because of a poor performance status (PS). They received treatment with i.v. low-dose cyclophosphamide (two patients), oral melphalan and prednisolone (three), cyclophosphamide, vincristine and prednisolone (one), or highdose methylprednisolone followed by VAMP (see below). A further patient received high-dose melphalan (HDM) with autologous stem cell rescue up-front. The focus of this paper is the 53 patients who proceeded to ST. They were compared with 153 newly diagnosed patients with IgG myeloma6 and 39 patients with previously untreated IgA myeloma. Sequential therapy Induction infusional chemotherapy (IC): VAMP:7 Between May 1986 and December 1988, 14 patients received vincristine 0.4 mg and doxorubicin 9 mg/m2 daily as a continuous infusion over 4 days together with methylprednisolone 1.5g i.v. or p.o. ⫻ 5 days, then tapered. C-VAMP:8 From January 1989 onwards, 34 patients received VAMP with the addition of weekly cyclophosphamide 500 mg on days 1, 8 and 15 of each cycle. V-CVAMP:9 Between January 1989 and July 1992, five patients received verapamil and CVAMP as part of a randomised trial to evaluate the role of verapamil in modifying drug resistance. Verapamil was given as a continuous infusion at a rate of 10 mg/day, starting 4 h before and ending 4 h after the VAMP infusion. All courses were repeated three weekly until maximum response, followed by an additional course. High-dose therapy (HDT): Patients were scheduled to receive HDT followed by autologous stem cell rescue 6 weeks after the first day of the last cycle of chemotherapy. Patients were admitted to hospital on the first day of HDT and the autograft was carried out in a four-bedded ward without barrier nursing. If the uncorrected glomerular filtration rate (GFR) was greater than 40 ml/min, melphalan 200 mg/m2 was given. The melphalan dose was reduced to 100 mg/m2 if the GFR was between 30 and 40 ml/min. Five patients with poor renal function received busulphan at a dose of 16 mg/kg in divided doses over 4 days. In two patients with more than 30% bone marrow infiltration by plasma cells at the time of HDT, melphalan 140 mg/m2 without stem cell rescue, with or without high-dose methylprednisolone (HDMP) was used. Until November 1992, bone marrow was harvested 6 weeks after completion of induction chemotherapy if the bone marrow aspirate at that stage showed less than 30% infiltration by plasma cells.10 Subsequently, peripheral blood stem cells (PBSC) were used following the subcutaneous administration of rhG-CSF (Amgen, Cambridge, UK) at a dose of 125 ␮g/m2 every 12 h for 7 days. Leukapheresis was commenced on day 5 for 4 consecutive

Bone Marrow Transplantation

days. This protocol was modified in June 1994 from when on G-CSF was administered at a dose of 12–16 ␮g/kg once a day for 4 days, followed by leukaphereses on 2 consecutive days starting on day 4.11 From January to December 1997, consenting patients were randomised to receive GCSF in either its glycosylated (lenograstim) or non-glycosylated (filgrastim) form at a dose of 6–11 ␮g/kg for 4 days. With this mobilisation regimen, aphereses were performed on days 4 and 5, rarely also on day 6, after a further dose of G-CSF on day 5.12 Two patients less than 55 years of age with an HLA matched sibling received an allogeneic transplant and one patient less than 50 years old had a matched unrelated donor transplant. Maintenance interferon (IFN): After haematopoeitic recovery (WBC ⬎2 × 109/l and platelets ⬎50 × 109/l) following HDT, patients were started on s.c. IFN-␣2b 3 mega units/m2 three time a week. From July 1988 to November 1991 this took place as part of a randomised trial, but later, IFN maintenance therapy was given to all patients. The IFN dose was adjusted according to toxicity.13 Estimation of Bence-Jones proteinuria (BJP) BJP was detected by electrophoresis of neat urine stained with colloidal gold. Immunological identification was carried out by immunofixation (Nigrosin stain14,15). Twentyfour hour urinary total protein quantitation was not performed routinely. Rather, disease monitoring was undertaken by the determination of the total protein in spot urine samples16 at the completion of each cycle of IC, pre-HDT, post HDT, and 3 monthly thereafter. Criteria for disease response For the assessment of response to therapy, we used the criteria described previously.17 CR was defined as no detectable BJP for a period of at least 3 months together with less than 5% plasma cells of normal morphology in the bone marrow aspirate, and normal bone marrow trephine appearances. If there was a 50% decrease in the level of BJP that was sustained for at least 1 month patients were regarded as being partial responders (PR). For patients in CR, relapse was defined as the reappearance of BJP or of a plasma cell infiltrate of more than 5% on the bone marrow. A persistent increase in BJP exceeding 25% was taken to signal relapse from PR. Patients were categorised as nonresponders (NR) if they failed to achieve either CR or PR. All patients underwent bone marrow aspirate and trephine biopsy at diagnosis, after four courses of IC, pre-HDT, post HDT and then 6 monthly for 1 year and annually thereafter. Statistical analysis Patient characteristics at presentation were compared with the Kruskal–Wallis test18 and the chi square test with Yates’ correction. Two-tailed probabilities were used throughout. Univariate analyses of the influence of each patient characteristic and treatment on overall survival (OS) and eventfree survival (EFS) were carried out according to the

Bence-Jones myeloma: a separate entity B Sirohi et al

Kaplan–Meier method.19 The analysis was undertaken in March 2000. Comparisons between groups were carried out using the log rank test.20 The categories chosen to represent the univariate influence of continuous variables were based on partition values derived from the medians. Multivariate analyses of the relationship between outcome and several variables were performed using forward stepwise regression by the Cox method.21,22 The forward stepwise method allows all variables to be considered for inclusion in the model but, in each step, selects the single variable that produces the largest, significant improvement of fit of the model.This procedure is preferable to the pre-selection of eligible variables on the basis of univariate P values because it allows for interrelationships between variables that may change their significance as the model develops. The following clinical and laboratory parameters were selected on the basis of findings in the literature and evaluated for prognostic relevance at presentation: age, sex, haemoglobin, white blood cell count, platelets, serum creatinine, urea, calcium, and albumin, light chain type (␭/␬), serum ␤2-microglobulin (␤2M), WHO performance status (PS), pain grade (WHO scale), number of lytic bone lesions (BL), stage of disease,23 bone marrow plasma cell percentage, urine total protein (UTP) concentration, and the presence or absence of soft tissue lesions at diagnosis. The impact on survival of the addition of cyclophosphamide to the VAMP regimen was also considered.

(range, 57–1463). Out of these 53 patients, 44 (83%) had stage III disease and 16 (30%) had soft tissue lesions at diagnosis. Three patients had AL amyloidosis, involving the heart, soft tissues and kidneys in one case each.

31

Sequential therapy Induction therapy: A median of five courses of IC was administered to the 53 BJ patients receiving ST. One patient did not complete his first cycle of IC and died on the second day of the infusion because of progressive disease and renal failure. Response to IC is shown in Table 2. High-dose treatment (HDT): Twenty-eight out of 53 (53%) patients received HDT. Nineteen patients received highdose melphalan (HDM) 200 mg/m2 with either autologous bone marrow (n = 8) or peripheral stem cell rescue (n = 11). Four patients received high-dose busulphan conditioning, three of them with autologous peripheral stem cells and one with autologous bone marrow. Two patients received HDM alone without rescue, and three patients had an allograft. Maintenance interferon: Twenty-eight out of 53 (53%) patients received interferon ␣-2b as maintenance therapy. Response to ST

Results Demography Fifty-three out of 61 patients (87%) with BJ myeloma who were referred during the study period entered the ST programme. Their characteristics are shown in Table 1. The median age was 55 years (range, 31–84), the median ␤2M was 5 ␮g/ml (range, 1.1–36.6, missing values in three patients) and the median serum creatinine was 118 ␮mol/l Table 1

The response to IC and HDT is shown in Table 2. In addition to the five patients in CR after IC – four of whom went on to receive HDT – a further 12 patients attained CR after HDT. Another five patients attained CR during IFN treatment. Median time to CR was 204 days. In the whole group, 22/53 (41.5%) patients achieved CR. Ten (19%) patients died during or after IC: five (9%) suffered treatment-related mortality (TRM; septicaemia in four and gastrointestinal haemorrhage in one) and five patients (9%) died of progressive disease (PD). Eight out

Comparison of characteristics in patients with BJ myeloma, IgG and IgA myeloma at presentation

Variable

Age (years) Creatinine ␮mol/l Urea ␮mol/l Calcium mmol/l ␤2M mg/l Albumin mmol/l Hb g/dl Platelets × 109/l PS No. of IC M:F Chain ␬/␭:both Stage I/II:III BL0–2:⬎2 ST lesions

BJ n = 53

IgG n = 153

IgA n = 39

P value IgG/BJ

P value IgA/BJ

P value IgG/IgA

55 (30–67) 118 (53–1463) 7.3 (3.6–41.4) 2.5 (2.06–4.38) 5.0 (1.3–40.5) 41 (20–51) 10.5 (7.4–15.8) 238 (63–504) 1 (0–4) 5 (1–8) 34:18 29:24:0 9:44 11:42 16 (30%)

52 (30–74) 90 (45–1644) 5.5 (1.5–34.8) 2.34 (1.75–4.81) 3.5 (0.7–33) 34 (18–56) 10.9 (7–15.6) 249 (26–517) 1 (0–4) 5 (1–12) 94:59 109:43:1 39:114 49:104 24 (16%)

54 (38–73) 97 (47–434) 6.3 (3–29.7) 2.42 (2.01–3.31) 5.3 (1.4–48.6) 33.5 (16–45) 10.2 (6.1–14.7) 199 (32–347) 0 (0–4) 5 (1–11) 25:14 27:11:1 12:27 16:23 3 (8%)

0.009 0.001 0.0002 ⬍0.0001 0.012 ⬍0.0001 0.54 0.25 0.9 0.24 0.55 0.05 0.0002 0.71 0.04

0.64 0.12 0.14 0.06 0.95 ⬍0.0001 0.08 0.04 0.4 0.09 0.98 0.17 0.25 0.27 0.02

0.04 0.16 0.06 0.04 0.009 0.4 0.08 0.0003 0.25 0.33 0.79 0.89 0.39 0.67 0.33

Median (range). ␤2M = beta 2 microglobulin; Hb = haemoglobin; PS = performance status; IC = infusional chemotherapy; M = male; F = female; ␬ = kappa; ␭ = lambda; BL = bone lesions; ST = soft tissue. Bone Marrow Transplantation

Bence-Jones myeloma: a separate entity B Sirohi et al

32

Table 2

Comparison of treatment received and response to treatment

Variable

Received VAMP:C/V-VAMP

BJ n = 53

IgG n = 153

IgA n = 39

P value IgG/BJ

P value BJ/IgA

P value IgG/IgA

14:39

38:115

8:31

0.9

0.7

0.7

0.13

0.8

1

29 (74.4)

0.001

0.06

0.8

17 10 1 1

0.85

0.98

1

Response to IC CR PR NR Died

5 23 15 10

Received HDT

28 (52.8)

Response to HDT CRa PR NR Died

16 7 2 3

(9.4) (43.4) (28.3) (18.9)

(57.2) (25.0) (7.1) (10.7)

21 90 31 11

(13.7) (58.8) (20.3) (7.2)

119 (77.8) 54 51 14 9

(45.4) (41.2) (11.8) (1.6)

6 20 9 4

(15.4) (51.3) (23.1) (10.2)

(58.7) (34.5) (3.4) (3.4)

Figures in brackets refer to percentages. a Five more patients achieved CR after starting interferon-␣2b maintenance.

The median OS of the 53 patients with BJ myeloma who entered the ST programme was 2.8 years (range, 0.01– 10.7). Their median EFS was 1.2 years (range, 0.01–6.3). The median OS of the 22 patients who achieved CR was 6.3 years (range, 2.7–10.7). The median EFS in patients who achieved CR was 3.2 years (range, 0.3–6.3). The median duration of CR was 2.3 years (range, 0.8–3.5). Of the 17 (32%) patients alive in March 2000, eight continued in CR and nine were in stable PR. Because CR is a time-dependent event, we used a landmark analysis to compare the impact of achieving CR on OS and EFS. By the time all 22 patients who were going to achieve CR had done so (2.45 years after the start of IC), only 22 of 53 patients remained alive (17 in the CR and five in no CR group) and 14 in remission (12 and two, respectively). Since at such a late landmark too small a group of patients was left for analysis, OS and EFS from the beginning of IC plus median time to CR was studied. Both OS and EFS were significantly longer for patients who achieved CR than for those who did not (P ⬍ 0.001, Figures 1 and 2).

Probability (%)

Survival

100 CR (9/22 deaths)

80

P < 0.001

60 No CR (14/17 deaths)

40 20 0

0

1

2

3

4

5

6

7

8

9

10

11

Time since start of IC + median days to CR (years) Figure 1 Log rank comparison of overall survival of patients with BJ myeloma who achieved CR (n = 22) vs those who did not (n = 17).

100

Probability (%)

of 10 patients had a serum creatinine ⬎140 ␮mol/l at diagnosis. Three (5.6%) patients died during HDT (cerebrovascular haemorrhage in one, septicaemia in two).

80

P < 0.001

CR (15/21 events)

60 40

No CR (15/15 events)

20 0

0

1

2

3

4

5

6

7

8

9

10

11

Time since start of IC + median days to CR (years)

Relapse Thirty-two out of 53 patients (60%) relapsed at a median of 2.2 years (range, 1.2–3.1). Twelve patients relapsed with extramedullary plasmacytomas (EMP, paravertebral in three, soft tissue or skin in four, pleural mass or effusion in three, pericardial effusion in one, and breast in one). Three patients relapsed with a serum IgG paraprotein. This disappeared with interferon treatment in two patients whereas the third patient was treated with C-VAMP because he also had progressive bone disease and BJP. One patient relapsed with plasma cell leukaemia. Bone Marrow Transplantation

Figure 2 Log rank comparison of event-free survival of patients with BJ myeloma who achieved CR (n = 21) vs those who did not (n = 15).

Second malignancy Two patients developed a second malignancy: one patient who developed an abdominal large cell non-Hodgkin’s lymphoma was given EPIC chemotherapy but died 4 months later. The second patient was found to have developed myelodysplasia with multiple chromosomal abnormalities [del(5), del(4), del(12) and del(18)] 6 years

Bence-Jones myeloma: a separate entity B Sirohi et al

Table 3

33

Patient characteristics and univariate analysis on the 53 BJ myelomas

Variable

Value

No. of patients

Median OS (years)

P value

Median EFS (years)

P value

PS

⬎1 ⭐1 ⬍55 ⭓55

12 41 26 27

0.4 3.4 3.3 2.2

0.003

0.2 1.9 1.3 1.2

0.003

M F I and II III

35 18 9 44

2.2 4.2 ⬎10 2.8

Creatinine

⭐170 ⬎170

30 23

3.2 1.9

0.23

1.3 0.99

0.57

Bone lesions

⭐2 ⬎2 ⭐1 ⬎1

11 42 28 25

⬎10 2.8 4.5 2.2

0.22

1.2 1.24 1.9 1.05

0.68

⭐1 ⬎1 ⬍2.5 ⭓2.5

35 18 26 27

4.2 1.9 3.4 2.7

␤2M

⬍5 ⭓5

24 29

3.4 2.2

Urea

⬍7.3 ⭓7.3 ⬍42 ⭓42

26 27 26 27

3.2 2.1 1.2 4.4

Kappa Lambda

29 24

2.7 2.8

1

1.2 1.6

0.38

Marrow plasma cells (%)

⬍45 ⭓45

25 28

2.7 2.97

0.45

1.2 1.5

0.4

NOC

⬍5 ⭓5

24 29

0.99 3.4

0.07

0.36 1.9

0.22

14 39

1.8 3.2

0.18

0.55 1.9

0.01

Age Sex Stage

Urine total protein Pain grade Calcium

Albumin Light chain

Type of VAMP

VAMP C/VC-VAMP

after the original diagnosis and died of septicaemia a fortnight later. Prognostic factors Tables 3 and 4 summarise the variables with prognostic significance on univariate and Cox multivariate analysis. Table 4 Multivariate analysis of outcome and receiving high-dose therapy in 53 patients with light chain myeloma, using prognostic factors present at diagnosis Event

Favourable covariant

Improvement in P value

Risk factor

95% CI

OS

Good PS UTP ⭐1 g/l

0.006 0.01

2.98 2.49

1.4–6.3 1.2–5.2

EFS

Good PS UTP ⭐1 g/l IC with CTX

0.005 0.012 0.011

4.39 3.64 3.42

2.0–9.6 1.8–7.5 1.6–7.2

HDT

NOC ⬍5

0.017

3.33

0.91–8.43

OS = overall survival; PS = performance status; UTP = urine total protein; IC with CTX = IC containing cyclophosphamide; EFS = event-free survival; HDT = high-dose therapy; NOC = number of courses of infusional chemotherapy.

0.04 0.33 0.22

0.07 0.05

1.05 2.5 1.22 1.24

0.44 0.13 0.9

0.1

1.86 0.85 1.3 1.1

0.43

0.08

1.9 0.99

0.09

0.39

1.3 1.1 0.99 1.6

0.92

0.79

0.01

0.63

0.39

Because patients presented in this series were in part enrolled in the late 1980s and early 1990s, cytogenetic data were not included. The univariate analysis identified PS, age and serum albumin as significantly influencing OS and EFS, while cyclophosphamide as part of the IC had a marked effect on the duration of EFS without significantly affecting OS. The multivariate analysis showed that patients with a good PS (P = 0.006) and UTP of ⭐1 g/l (P = 0.01) had a significantly better OS, and that patients with a good PS (P = 0.005), UTP of ⭐1 g/l (P = 0.01), those with ⭐2 bone lesions, and those who received IC containing cyclophosphamide (P = 0.01) had a significantly longer EFS. Pre-treatment ␤2M and serum creatinine as well as extent of bone destruction at diagnosis had no statistically significant influence on outcome. A significantly smaller proportion of patients with BJ myeloma received HDT compared to patients with IgG and IgA myeloma (Table 2). The Cox multivariate analysis carried out with HDT as the endpoint for various presentation variables identified only the number of courses of IC as significantly influencing the probability of receiving HDT. Patients who received less than five courses of IC had a higher probability of receiving HDT (Table 4). Bone Marrow Transplantation

Bence-Jones myeloma: a separate entity B Sirohi et al

Comparison of BJ with IgG and IgA myelomas

100

Demography: Table 1 compares patient characteristics between the three types of myeloma. Patients with BJ myeloma were significantly older at diagnosis (P = 0.009), had a higher creatinine (P = 0.001), urea (P = 0.0002), calcium (P ⬍ 0.0001), ␤2M (P = 0.012), albumin (P ⬍ 0.0001), lambda light chain disease (P = 0.05) and more stage III (P = 0.0002) disease compared to IgG myelomas. When compared to IgA myeloma, BJ myeloma patients had a significantly higher albumin (P ⬍ 0.0001) and lower platelets (P = 0.04). Response to IC (Table 2): Five out of 53 BJ myeloma patients (9.4%) achieved CR to IC compared to 21 out of 153 (13.7%) with IgG myeloma and six out of 39 (15%) with IgA myeloma (P = NS). Ten out of 53 BJ myeloma patients (19%) died while receiving IC, compared to 11 of 153 (7%) of those with IgG myeloma and four of 39 (10%) with IgA myeloma (P = NS). Patients receiving HDT: Out of 53 patients with BJ myeloma, 28 (53%) patients received HDT, which was a significantly smaller proportion than among IgG (119/153; P = 0.001) or IgA myeloma patients (29/39; P = 0.06). Twenty-five out of 53 (47%) BJ patients did not receive HDT because of a poor PS (n = 6), death during IC (n = 10), and refusal (n = 9), even though two out of the latter nine had HDT after their first relapse 18 and 20 months later. The reasons for not receiving high-dose in the three groups are compared in Table 4. Response to HDT: Of the 28 patients who received HDT, 16 (57%) attained CR, compared to 54 of 119 (45.3%) IgG myelomas (P = NS) and 17 of 29 (58.6%) IgA myelomas (P = NS). Survival and relapse The median OS of patients with BJ myeloma (2.8 years) was significantly poorer (P = 0.03, Figure 3) than that of those with IgG myeloma (4.5 years) but was not significantly different from that of IgA myeloma patients (3.9 100

P values: BJ vs IgG = 0.033 BJ vs IgA = NS IgA vs IgG = NS

60 IgA (29/39)

40

IgG (121/153)

20 0

0

1

2

3

4

5

6

8

7

9

10

11

12

13

Time since start of IC (years) Figure 4 Log rank comparison of event-free survival of patients with BJ (n = 53), IgG (n = 153) and IgA (n = 39) myeloma.

years). The median EFS of patients with BJ myeloma was 1.2 years compared to 2.1 years in patients with IgG myeloma (P = 0.033, Figure 4) and 2.2 years in patients with IgA myeloma (P = NS). The probabilities of relapse were not significantly different between the three immunotypes (P = 0.44). Among the patients who attained CR there was no difference in OS and EFS between the three subtypes of myeloma, the median EFS being 2.3 years for BJ myeloma, 2.1 for IgG and 2.2 for IgA myeloma (Figure 5). Discussion In BJ myeloma there is no excess production of immunoglobulin heavy chains and the abnormal light chains have a shorter biological half-life than IgG or IgA paraprotein.24 In this study we wished to determine whether the different nature of the paraprotein in BJ myeloma results in a distinct mode of presentation and response to intensive ST of this disease in comparison to IgG or IgA myeloma. We have previously made a case for why we consider that the patient information contained in the database of our myeloma unit 100

Medians: BJ = 1014 (95% CI = 573–1599) P values: BJ vs IgG = 0.034 (days) IgG = 1663 (1342–2212) BJ vs IgA = NS IgA = 1445 (647–1933) IgG vs IgA = NS

80

Medians: BJ = 847 (95% CI = 273–1294) (days) IgG = 763 (588–1316) IgA = 841 (434–1224)

P values: BJ vs IgG = NS BJ vs IgA = NS IgA vs IgG = NS

80 IGA (25/39)

Probability (%)

Probability (%)

Medians: BJ = 454 (95% CI = 318–752) (days) IgG = 755 (596–888) IgA = 797 (386–1078)

80

Probability (%)

34

60 BJ (36/53)

40

IGG (94/153)

60 IgG (56/80)

40

20

20

0

0

IgA (14/22) BJ (16/22)

0

1

2

3

4

5

6

7

8

9

10

11

12

13

Time since start of IC (years) Figure 3 Log rank comparison of overall survival of patients with BJ (n = 53), IgG (n = 153) and IgA (n = 39) myeloma. Bone Marrow Transplantation

0

1

2

3

4

5

6

7

8

9

10

11

12

Time since CR (years) Figure 5 Log rank comparison of event-free survival of patients with BJ (n = 22), IgG (n = 80) and IgA (n = 22) myeloma who achieved CR.

Bence-Jones myeloma: a separate entity B Sirohi et al

is representative of an unselected population-based group.1 Furthermore, although there were progressive changes in the protocols for IC under which our patients received treatment, these changes did not influence outcome.8,25 In an earlier multivariate analysis involving BJ and IgG/IgA myeloma patients we have shown that the type of light chain secreted and the immunologic subtype of myeloma has no impact on disease out-look.26 We are not aware of any recent publication aimed at highlighting the distinctive presenting features of patients with BJ myeloma in comparison to patients affected by IgG or IgA myeloma in the context of intensive therapy. We found a number of statistically significant differences between these three cohorts of patients, but their magnitude was often too small to be considered clinically meaningful on their own (eg age at presentation, Hb or platelet count). The combination of hypercalcaemia with a raised urea and creatinine is commonly associated with BJP27 and our study confirms the presence of higher median values for these three parameters among BJ patients. It is nevertheless interesting to note that the median creatinine among BJ patients was well below our cut-off for significant renal impairment (⬍140 ␮mol/l). Similarly, the median serum calcium level of these patients was well within the normal range. Given that the performance status between the three cohorts of patients was comparable, the considerably higher median serum albumin level at diagnosis among the patients with BJ myeloma is striking. The lower albumin concentration among patients with an IgG or IgA paraprotein may be due to a homeostatic mechanism controlling the plasma oncotic pressure causing reciprocal changes in serum albumin and immunoglobulin levels.28 However, a low serum albumin is equally predictive of a much shorter OS in BJ myeloma as it is among patients with IgG myeloma.6 The immunological type of light chain produced by the BJ myelomas did not affect prognosis. Between the three immunological subtypes of myeloma, there was no difference in ability to administer IC. However, a significantly higher number of patients with BJ myeloma did not receive HDT. As shown in Table 5, it is difficult to identify any single reason for this. Among BJ myeloma 9/53 patients declined to undergo HDT and this may be related to the poorer performance status present in the five of these nine patients who had stage IIIA/B disease. However, Cox analysis (Table 4) demonstrated that none of the pre-treatment variables analysed had an effect on whether patients received HDT or not. It appears likely, however, that the combination of some of the adverse variables that were over-represented among the patients with BJ myeloma made them likely to be more debilitated and therefore less likely to receive HDT. Table 5

Reasons for not receiving HDT

Reason

BJ 25/53

IgG 34/153

IgA 10/39

P value

Refused Poor PS Died

9 6 10

8 15 11

4 2 4

0.70 0.33 0.94

Among BJ myeloma patients. only the number of courses of IC influenced the likelihood of a patient going on to receive HDT. Given that the most effective part of ST is undoubtedly the high-dose melphalan, one should maximise the probability that patients will receive this part of the treatment. Continuing to give IC to maximum response appears to decrease the probability of proceeding to autografting because of a falling performance status. It may also increase drug resistance. Therefore, it may prove valuable to design a programme that includes HDT early in the management of BJ myeloma, such as the up-front delivery of high-dose treatment.29 With the advances in supportive care over recent years, the TRM of such a protocol should not be as high as was reported earlier.25 Since the probability of relapse was not significantly different between BJ and IgG myeloma, the excess mortality observed among BJ patients may reflect a greater impact on general health of their subtype of myeloma than that experienced by patients with IgG myeloma. It is nevertheless important to note that even though we were able to confirm the significant difference in OS and EFS between IgG and BJ myeloma,30 the OS and EFS of patients with BJ myeloma who attained CR were not significantly different from those with IgG disease. However, this finding is not based on an intention to treat analysis and only 53% of BJ myeloma patients received HDT. Nevertheless, it underlines the value of attempting to induce CR with early treatment intensification in this immunological subtype of myeloma as well. The treatment-related mortality observed in this series was 9% with induction chemotherapy and 6% with high-dose therapy which is higher than the IgG myelomas but we have argued in an earlier report that in our experience treatment adjustment according to renal failure is not warranted in myeloma patients.31 The pattern of relapse in our group of patients was unusual in as far as 12 patients relapsed with EMP. In six of them, soft tissue lesions were present at diagnosis, but in the other six the occurrence of EMP was a new event. Moreover, three other patients developed an IgG paraprotein during disease progression, suggestive of clonal evolution. In the multivariate analysis, only PS and UTP were found to be independent predictors of OS. Similarly, out of the many other factors considered, only the same two parameters together with the inclusion of cyclophosphamide in the IC regimen were predictive of EFS. Contrary to what is the case in IgG myeloma, ␤2M was not an independent predictive factor for survival in BJ myeloma. It was also surprising to note that the serum creatinine level at diagnosis did not influence the outcome of BJ myeloma. This is in contradiction to the findings of Merlini et al32 and implies that category B of the Durie–Salmon staging system23 does not correlate well with prognosis in BJ myeloma in the context of modern intensive treatments. Performance status had been incorporated in the MRC staging system and was one of the independent variables influencing OS and EFS in BJ myelomas but it has been criticised for being difficult to apply reproducibly.33 The Durie–Salmon staging system is based on the finding that the myeloma cell mass correlates with the extent of bone destruction, Hb level, serum calcium and the level of

35

Bone Marrow Transplantation

Bence-Jones myeloma: a separate entity B Sirohi et al

36

the ‘M’ component in serum and urine. Serum creatinine was also found to be a good predictor for survival, even though it does not correlate with tumour cell mass. This staging system has hitherto been the one most widely used in clinical trials, but Merlini et al32 showed that in myeloma, the value of prognostic variables depends on the type of paraprotein produced. They found that serum creatinine, calcium and percentage of bone marrow plasma cells predicted prognosis for IgG and BJ myeloma patients, whereas Hb and calcium together with the level of the ‘M’ component were predictive for IgA patients. We, too, have argued previously that in multiple myeloma reliable risk stratification for the comparisons of treatment programmes is only possible if different patient groups are staged separately in easily demarcated subgroups, such as IgG and IgA myeloma.6 We have recently published a new staging system for IgG myeloma.6 In the present study, we were able to demonstrate that BJ myeloma has a distinct clinical presentation and response to treatment. Although we were able to define prognostic criteria, there was an insufficient number of patients with BJ myeloma to devise a new staging system for this subtype of multiple myeloma as well. To this end, the findings in this paper could be incorporated into a new, multicentre database established in the era of myeloma therapy with intensive ST. Acknowledgements We thank the David Adams Leukaemia Fund, Bud Flanagan Leukaemia Fund and the Institute of Cancer Research.

References 1 Powles R, Raje N, Milan S et al. Outcome assessment of a population based group of 195 unselected myeloma patients under 70 years of age offered intensive treatment. Bone Marrow Transplant 1997; 20: 435–443. 2 Barlogie B, Jagannath S, Desikan KR et al. Total therapy with tandem transplants for newly diagnosed multiple myeloma. Blood 1999; 93: 55–65. 3 Attal M, Harousseau JL, Stoppa AM et al. A prospective, randomized trial of autologous bone marrow transplantation and chemotherapy in multiple myeloma. Intergroupe Francais du Myelome. New Engl J Med 1996; 335: 91–97. 4 Bjorkstrand BB, Ljungman P, Svensson H et al. Allogeneic bone marrow transplantation versus autologous stem cell transplantation in multiple myeloma: a retrospective case matched study from the European Group for Blood and Marrow Transplantation. Blood 1996; 88: 4711–4718. 5 Kulkarni S, Powles RL, Treleaven JG et al. Impact of previous high-dose therapy on outcome after allografting for multiple myeloma. Bone Marrow Transplant 1999; 23: 675–680. 6 Sirohi B, Powles R, Mehta J et al. Complete remission rate and outcome after intensive treatment of 177 patients under 75 years of age with IgG myeloma defining a circumscribed disease entity with a new staging system. Br J Haematol 1999; 107: 656–666. 7 Forgeson GV, Selby P, Lakhani S et al. Infused vincristine and adriamycin with high dose methylprednisolone (VAMP) in advanced previously treated multiple myeloma patients. Br J Cancer 1988; 58: 469–473. Bone Marrow Transplantation

8 Raje N, Powles R, Kulkarni S et al. A comparison of vincristine and doxorubicin infusional chemotherapy with methylprednisolone (VAMP) with the addition of weekly cyclophosphamide (C-VAMP) as induction treatment followed by autografting in previously untreated myeloma. Br J Haematol 1997; 97: 153–160. 9 Gore ME, Selby PJ, Millar B et al. The use of verapamil to overcome drug resistance in myeloma. Proc Am Soc Clin Oncol 1988: 7: 228. 10 Cunningham D, Paz-Ares L, Milan S et al. High dose melphalan and autologous bone marrow transplantation as consolidation in previously untreated myeloma. J Clin Oncol 1994; 12: 759–736. 11 Raje N, Powles R, Horton C et al. Comparison of marrow vs blood derived stem cells for autografting in previously untreated multiple myeloma. Br J Cancer 1997; 75: 1684– 1689. 12 Kulkarni S, Powles R, Treleaven J et al. Randomized trial of equal doses of lenograstim and filgrastim for mobilisation of peripheral blood stem cells in patients with hematologic malignancies. Bone Marrow Transplant 2000; 25: S260. 13 Cunningham D, Powles RL, Malpas J et al. A randomised trial of maintenance interferon following high dose chemotherapy in multiple myeloma; long term follow up results. Br J Haematol 1998; 102: 495–502. 14 Kohn J, Riches PG. A cellulose acetate immuno-fixation technique. J Immunolog Meth 1978; 20: 325–331. 15 Karinkanta HH, Nieminen EJ. An improved technique for immunofixation of electropherograms. Clin Chem 1978; 24: 1639–1641. 16 Watanaba N, Kamel S, Ohkubo A et al. Urinary protein as measured with a pyrogallol red-molybdate complex, manually and in a Hitachi 726 automated analyzer. Clin Chem 1986; 328: 1551–1554. 17 Gore ME, Selby PJ, Viner C et al. Intensive treatment of multiple myeloma and criteria for complete remission. Lancet 1989; 2: 879–881. 18 Kruskal WH, Wallis WA. Use of ranks in one-criterion variance analysis, J Am Stat Assoc 1952; 47: 583–621. 19 Kaplan EL, Meier P. Non parametric estimation from incomplete observations. J Am Stat Assoc 1958; 53: 457–481. 20 Peto R, Pike MC, Armitage P et al. Design and analysis of randomized clinical trials requiring prolonged observation of each patient. II. Analysis and examples. Br J Cancer 1977; 35: 1–39. 21 Cox DR. Regression models and life tables. J Roy Stat Soc Br 1972; 24: 187–220. 22 BMDP statistical software release 7. BMDP Statistical Software Manual. University of California Press: Berkeley, 1992; 825–866. 23 Durie BGM, Salmon SE. A clinical staging system for multiple myeloma. Cancer 1975; 36: 842–854. 24 Hobbs JR. Immunoglobulins in clinical chemistry. Adv Clin Chem 1971; 14: 219–317. 25 Cunningham D, Paz-Ares L, Gore ME et al. High dose melphalan for multiple myeloma: long term follow up data. J Clin Oncol 1994; 12: 764–768. 26 Sirohi B, Powles R, Treleaven J et al. Light chain disease (LCD): a different form of myeloma compared with IgG disease. Blood 1999; 10 (Suppl. 1): 2567. 27 Smolens P, Barnes JL, Kreisberg R. Hypercalcemia can potentiate the nephrotoxicity of Bence Jones proteins. J Lab Clin Med 1987; 110: 460–465. 28 Howells DPM. Serum albumin and survival in myelomatosis. Br Med J 1971; 3: 583. 29 McElwain TJ, Powles R. High dose intravenous melphalan for

Bence-Jones myeloma: a separate entity B Sirohi et al

plasma cell leukaemia and myeloma. Lancet 1983; 2: 822– 824. 30 Jagannath S, Barlogie B, Dicke K et al. Autologous bone marrow transplantation in multiple myeloma: identification of prognostic factors. Blood 1990; 76: 1860–1866. 31 Sirohi B, Powles R, Kulkarni S et al. Glomerular filtration rate (GFR) as a surrogate marker for transplant-related toxicity and overall outcome in patients of multiple myeloma (MM)

homogenously treated with high dose melphalan (HDM) 200/m2. Bone Marrow Transplant 2000; 25 (Suppl. 1): 694. 32 Merlini G, Waldenstrom JG, Jayakar S. A new improved clinical staging system for multiple myeloma based on analysis of 123 treated patients. Blood 1980; 55: 1011–1019. 33 Medical Research Council Working Party on Leukemia in adults. Prognostic features in the third MRC myelomatosis trial. Br J Cancer 1980; 42: 831–840.

37

Bone Marrow Transplantation