- Email: [email protected]
Treatment for elderly patients with multiple myeloma – Authors' reply
Correspondence
5
Tosi P, Zamagni E, Ronconi S, et al. Safety of autologous hematopoietic stem cell transplantation in patients with multiple myeloma and chronic renal failure. Leukemia 2000; 14: 1310–13.
Authors’ reply Jayesh Mehta, Masaharu Tsubokura and Masahiro Kami, and Daniele Focosi question second-line treatment and the representativeness of our study population. Recognising the efficacy of thalidomide as second-line therapy, Mehta comments that half of the patients who withdrew from our trial in the MP and MEL100 groups did not receive second-line therapy with thalidomide, and that this could have been the sole factor responsible for the survival advantage that we saw with MPT. Before responding, we need to clarify that data presented in The Lancet article come from the intention-to-treat (ITT) analyses done on Oct 8, 2005. However, as requested by The Lancet editors and noted in the article, data for overall survival, progression-free survival, and survival after progression were presented using both that date and a more current updated ITT analysis from Jan 8, 2007. Data on second-line therapy (as with all other data presented in the article except survival), however, are taken only from the initial analysis of Oct 8, 2005. We should also clarify that second-line therapy in our trial was neither prescribed nor mandated, but left to investigators’ discretion. Using data on second-line and other rescue treatments now available from the January, 2007, analysis, the proportion of patients who still did not receive any second-line therapy was similar in the MP (21 of 193 [11%]) and MEL100 (15 of 122 [12%]) groups, but higher in the MPT group (29 of 124 [23%]). Among patients receiving therapy after their initially randomised treatment, 74 of 145 (51%) and 48 of 89 (54%) received thalidomide as second-line treatment in the MP and MEL100 groups, respectively. These proportions rise to 104 of 145 (72%) 984
and 58 of 89 (65%) when including thalidomide treatment at any time after withdrawal from the trial, and to 113 of 145 (78%) and 67 of 89 (75%) when thalidomide, bortezomib, or lenalidomide are included as postwithdrawal treatment. Consequently, only 32 (17%) and 22 (18%) of the 193 and 122 patients who started treatment as randomised in the MP and MEL100 groups, respectively, did not receive one of these three very active anti-myeloma treatments during follow-up to Jan 8, 2007—ie, less than one patient out of five. Thus, it seems very unlikely that rescue treatment is the sole explanation for the large survival disadvantage seen in the MP and MEL100 groups compared with the MPT groups. Regarding Mehta’s other questions, per protocol, all patients recruited in a small centre and assigned to the MEL100 group were referred for transplantation to an associated large centre. When analysing overall survival according to centre recruitment (it was considered large with an enrolment of 10 patients or more into the trial), no significant effect was seen on overall survival in the MP, MPT, and MEL100 groups (p=0·36, p=0·06, and p=0·24, respectively). When defining large centre as five patients or more, this absence of effect was even more striking, especially in the MEL100 group (p=0·93, p=0·12, and p=0·94 in the three groups, respectively). We agree that some imbalance in biological variable distributions at randomisation was seen between the three groups, especially for albumin concentration. There was no imbalance for haemoglobin, white blood cell count, platelet count, and bone marrow plasmacytosis. However, it should be noted that among the variables mentioned by Mehta, albumin, haemoglobin, white blood cell count, and bone marrow plasmacytosis were all included in our survival prognostic model and that, after adjustment for prognostic factors, differences in overall survival
between the MP or MEL100 group and the MPT group remained highly significant. Additionally, when testing for the effect of C-reactive protein, serum calcium, platelet count, creatinine, and LDH concentrations, in addition to the factors used in our prognostic model, all were clearly not significant (p=0·95, p=0·26, p=0·59, p=0·34, and p=0·44, respectively). Consequently, it is highly unlikely that some imbalance in the distribution of factors, assessed at inclusion between the trial groups, could explain the observed differences in overall survival. Finally, Focosi suggests that patients enrolled in the IFM 99-06 trial were not representative of the overall elderly myeloma population. We respectfully disagree with this criticism. We excluded patients with overt, clinically significant light-chain amyloidosis and myeloma, as other similar trials have done.1 In our experience, these patients do not represent 12–30% of myeloma patients in the age-group studied. The exclusion of patients with peripheral neuropathy was logical in the late 1990s when this trial was started, and it should be noted that grade 2 neuropathy or worse is still an exclusion criterion in virtually all myeloma trials.2,3 Overall, we are convinced that our exclusion criteria were similar to those used in other comparable trials. We declare that we have no conflict of interest.
*Thierry Facon, Jean-Yves Mary, Cyrille Hulin, Lotfi Benboubker, Michel Attal, on behalf of the Intergroupe Francophone du Myélome [email protected] Service des Maladies du Sang, Hôpital Claude Huriez, Centre Hospitalier Universitaire de Lille, rue Michel Polonovski, 59037 Lille, France (TF); U717 INSERM, Université Paris 7, Hôpital Saint-Louis, Paris, France (JYM); Service d’Hématologie, CHU, Nancy, France (CH); Service d’Hématologie, CHU, Tours, France (LB); and Service d’Hématologie, CHU, Toulouse France (MA) 1
Palumbo A, Falco P, Corradini P, et al. Melphalan, prednisone, and lenalidomide treatment for newly diagnosed myeloma: a report from the GIMEMA—Italian Multiple Myeloma Network. J Clin Oncol 2007; 25: 4459–65.
www.thelancet.com Vol 371 March 22, 2008
Correspondence
3
Mateos MV, Hernandez JM, Hernandez MT, et al. Bortezomib plus melphalan and prednisone in elderly untreated patients with multiple myeloma: results of a multicenter phase 1/2 study. Blood 2006; 108: 2165–72. Rajkumar SV, Blood E, Vesole D, Fonseca R, Greipp PR. Phase III clinical trial of thalidomide plus dexamethasone compared with dexamethasone alone in newly diagnosed multiple myeloma: a clinical trial coordinated by the Eastern Cooperative Oncology Group. J Clin Oncol 2006; 24: 431–36.
Homoeopathy in the UK For a publication that bills itself as the world’s leading medical journal, it is surprising to see Udani Samarasekera’s Special Report on homoeopathy in Britain (Nov 17, p 1677),1 which merely rehashes time-worn arguments without presenting any new information. As you know, we provided you with a six-page set of responses to questions posed in connection with this report. We find it fascinating that only one of those responses was actually printed. To better inform your readers, we would like to point out that the Society of Homeopaths is the largest professional organisation registering homoeopaths in Britain, representing more than 2300 members overall. We are committed to fostering an integrated, patient-centred approach to health and wellness, treating each person’s symptoms as unique and each person’s care as an individual programme. Society-registered homoeopaths have satisfied the Society’s educational and professional requirements and agreed to practise in accordance with the Society of Homeopaths’ Code of Ethics and Practice, the Core Criteria for Homeopathic Practice, and the National Occupational Standards for Homeopathy. We concur with our colleagues across the medical profession that proper regulation is essential to delivering integrated, patient-centred care and we welcome increased regulation of the homoeopathy profession. In fact, it is something that we, as the leading professional www.thelancet.com Vol 371 March 22, 2008
organisation, have been advocating for years. I declare that I have no conflict of interest.
Paula Ross [email protected] Society of Homeopaths, Northampton NN3 6WL, UK 1
Samarasekera U. Pressure grows against homoeopathy in the UK. Lancet 2007; 370: 1677–78.
Meta-analyses of homoeopathy trials The opening statement of Ben Goldacre’s Comment on homoeopathy (Nov 17, p 1672)1 is false. He states that “Five large meta-analyses of homoeopathy trials have been done. All have had the same result: after excluding methodologically inadequate trials and accounting for publication bias, homoeopathy produced no statistically significant benefit over placebo.” In fact only one of the large meta-analyses of clinical trials of homoeopathy included correction for both publication bias and trial quality, and it showed that the effects of homoeopathy remained significantly greater than placebo when these, and other, corrections were applied singly or in combination. Remarkably, Goldacre fails even to cite this metaanalysis, although it was published in The Lancet.2 Goldacre’s statement is, without question, false. It would not have survived peer-review had this been a scientific paper. This is not a matter of scientific debate, it is plain inaccurate reporting. I declare that I have no conflict of interest.
Peter Fisher peter.fi[email protected]
Author’s reply I am grateful for the opportunity to clarify two small errors. One of my references was incorrect,1 and, as Peter Fisher correctly implies, I should have simply said: “...when these analyses were restricted to the best quality trials, homoeopathy produced no statistically significant benefit over placebo.” This trivial change makes no difference at all. Fisher cites one meta-analysis from 1997,2 which is frequently quoted by homoeopaths as proof of efficacy. However this positive finding is perhaps weak: the odds ratio for the 26 goodquality studies, corrected for publication bias, was 1·78 (95% CI 1·03–3·10). The same trials from this meta-analysis have been reanalysed extensively3 and significant doubt has been cast on the initial analysis. Most notably the same researchers, working on the same set of 89 trials4 in 1999 concluded that their own reanalysis “weakened the findings of our original meta-analysis… it seems, therefore, likely that our meta-analysis at least overestimated the effects of homeopathic treatments.” Many clinicians are clear that they can see a role for homoeopathy, even if it does perform no better than placebo. I would hope that homoeopaths might now divert their attention to performing randomised controlled (albeit unblinded) trials comparing “visiting a homoeopathy clinic” against “general practitioner’s treatment as usual”, since this might be the clinical question of more interest to patients—ie, not “do the pills work better than placebo” but “will the experience of visiting a homoeopath help me to feel better?” I write a column for the Guardian newspaper. I have just started an academic post funded by the UK National Health Service.
Ben Goldacre [email protected]
Royal London Homoeopathic Hospital, London WC1 3HR, UK
The Guardian, London EC1R 3ER, UK
1
1
2
Goldacre B. Benefits and risks of homoeopathy. Lancet 2007; 370: 1672–73. Linde K, Clausius N, Ramirez G, et al. Are the clinical effects of homoeopathy placebo effects? A meta-analysis of placebo-controlled trials. Lancet 1997; 350: 834–43.
Photolibrary
2
2
Kleijnen J, Knipschild P, ter Riet G. Clinical trials of homoeopathy. BMJ 1991; 302: 316–23. Linde K, Clausius N, Ramirez G. Are the clinical effects of homoeopathy placebo effects? A meta-analysis of placebo-controlled trials. Lancet 1997; 350: 834–43.
985
5
Tosi P, Zamagni E, Ronconi S, et al. Safety of autologous hematopoietic stem cell transplantation in patients with multiple myeloma and chronic renal failure. Leukemia 2000; 14: 1310–13.
Authors’ reply Jayesh Mehta, Masaharu Tsubokura and Masahiro Kami, and Daniele Focosi question second-line treatment and the representativeness of our study population. Recognising the efficacy of thalidomide as second-line therapy, Mehta comments that half of the patients who withdrew from our trial in the MP and MEL100 groups did not receive second-line therapy with thalidomide, and that this could have been the sole factor responsible for the survival advantage that we saw with MPT. Before responding, we need to clarify that data presented in The Lancet article come from the intention-to-treat (ITT) analyses done on Oct 8, 2005. However, as requested by The Lancet editors and noted in the article, data for overall survival, progression-free survival, and survival after progression were presented using both that date and a more current updated ITT analysis from Jan 8, 2007. Data on second-line therapy (as with all other data presented in the article except survival), however, are taken only from the initial analysis of Oct 8, 2005. We should also clarify that second-line therapy in our trial was neither prescribed nor mandated, but left to investigators’ discretion. Using data on second-line and other rescue treatments now available from the January, 2007, analysis, the proportion of patients who still did not receive any second-line therapy was similar in the MP (21 of 193 [11%]) and MEL100 (15 of 122 [12%]) groups, but higher in the MPT group (29 of 124 [23%]). Among patients receiving therapy after their initially randomised treatment, 74 of 145 (51%) and 48 of 89 (54%) received thalidomide as second-line treatment in the MP and MEL100 groups, respectively. These proportions rise to 104 of 145 (72%) 984
and 58 of 89 (65%) when including thalidomide treatment at any time after withdrawal from the trial, and to 113 of 145 (78%) and 67 of 89 (75%) when thalidomide, bortezomib, or lenalidomide are included as postwithdrawal treatment. Consequently, only 32 (17%) and 22 (18%) of the 193 and 122 patients who started treatment as randomised in the MP and MEL100 groups, respectively, did not receive one of these three very active anti-myeloma treatments during follow-up to Jan 8, 2007—ie, less than one patient out of five. Thus, it seems very unlikely that rescue treatment is the sole explanation for the large survival disadvantage seen in the MP and MEL100 groups compared with the MPT groups. Regarding Mehta’s other questions, per protocol, all patients recruited in a small centre and assigned to the MEL100 group were referred for transplantation to an associated large centre. When analysing overall survival according to centre recruitment (it was considered large with an enrolment of 10 patients or more into the trial), no significant effect was seen on overall survival in the MP, MPT, and MEL100 groups (p=0·36, p=0·06, and p=0·24, respectively). When defining large centre as five patients or more, this absence of effect was even more striking, especially in the MEL100 group (p=0·93, p=0·12, and p=0·94 in the three groups, respectively). We agree that some imbalance in biological variable distributions at randomisation was seen between the three groups, especially for albumin concentration. There was no imbalance for haemoglobin, white blood cell count, platelet count, and bone marrow plasmacytosis. However, it should be noted that among the variables mentioned by Mehta, albumin, haemoglobin, white blood cell count, and bone marrow plasmacytosis were all included in our survival prognostic model and that, after adjustment for prognostic factors, differences in overall survival
between the MP or MEL100 group and the MPT group remained highly significant. Additionally, when testing for the effect of C-reactive protein, serum calcium, platelet count, creatinine, and LDH concentrations, in addition to the factors used in our prognostic model, all were clearly not significant (p=0·95, p=0·26, p=0·59, p=0·34, and p=0·44, respectively). Consequently, it is highly unlikely that some imbalance in the distribution of factors, assessed at inclusion between the trial groups, could explain the observed differences in overall survival. Finally, Focosi suggests that patients enrolled in the IFM 99-06 trial were not representative of the overall elderly myeloma population. We respectfully disagree with this criticism. We excluded patients with overt, clinically significant light-chain amyloidosis and myeloma, as other similar trials have done.1 In our experience, these patients do not represent 12–30% of myeloma patients in the age-group studied. The exclusion of patients with peripheral neuropathy was logical in the late 1990s when this trial was started, and it should be noted that grade 2 neuropathy or worse is still an exclusion criterion in virtually all myeloma trials.2,3 Overall, we are convinced that our exclusion criteria were similar to those used in other comparable trials. We declare that we have no conflict of interest.
*Thierry Facon, Jean-Yves Mary, Cyrille Hulin, Lotfi Benboubker, Michel Attal, on behalf of the Intergroupe Francophone du Myélome [email protected] Service des Maladies du Sang, Hôpital Claude Huriez, Centre Hospitalier Universitaire de Lille, rue Michel Polonovski, 59037 Lille, France (TF); U717 INSERM, Université Paris 7, Hôpital Saint-Louis, Paris, France (JYM); Service d’Hématologie, CHU, Nancy, France (CH); Service d’Hématologie, CHU, Tours, France (LB); and Service d’Hématologie, CHU, Toulouse France (MA) 1
Palumbo A, Falco P, Corradini P, et al. Melphalan, prednisone, and lenalidomide treatment for newly diagnosed myeloma: a report from the GIMEMA—Italian Multiple Myeloma Network. J Clin Oncol 2007; 25: 4459–65.
www.thelancet.com Vol 371 March 22, 2008
Correspondence
3
Mateos MV, Hernandez JM, Hernandez MT, et al. Bortezomib plus melphalan and prednisone in elderly untreated patients with multiple myeloma: results of a multicenter phase 1/2 study. Blood 2006; 108: 2165–72. Rajkumar SV, Blood E, Vesole D, Fonseca R, Greipp PR. Phase III clinical trial of thalidomide plus dexamethasone compared with dexamethasone alone in newly diagnosed multiple myeloma: a clinical trial coordinated by the Eastern Cooperative Oncology Group. J Clin Oncol 2006; 24: 431–36.
Homoeopathy in the UK For a publication that bills itself as the world’s leading medical journal, it is surprising to see Udani Samarasekera’s Special Report on homoeopathy in Britain (Nov 17, p 1677),1 which merely rehashes time-worn arguments without presenting any new information. As you know, we provided you with a six-page set of responses to questions posed in connection with this report. We find it fascinating that only one of those responses was actually printed. To better inform your readers, we would like to point out that the Society of Homeopaths is the largest professional organisation registering homoeopaths in Britain, representing more than 2300 members overall. We are committed to fostering an integrated, patient-centred approach to health and wellness, treating each person’s symptoms as unique and each person’s care as an individual programme. Society-registered homoeopaths have satisfied the Society’s educational and professional requirements and agreed to practise in accordance with the Society of Homeopaths’ Code of Ethics and Practice, the Core Criteria for Homeopathic Practice, and the National Occupational Standards for Homeopathy. We concur with our colleagues across the medical profession that proper regulation is essential to delivering integrated, patient-centred care and we welcome increased regulation of the homoeopathy profession. In fact, it is something that we, as the leading professional www.thelancet.com Vol 371 March 22, 2008
organisation, have been advocating for years. I declare that I have no conflict of interest.
Paula Ross [email protected] Society of Homeopaths, Northampton NN3 6WL, UK 1
Samarasekera U. Pressure grows against homoeopathy in the UK. Lancet 2007; 370: 1677–78.
Meta-analyses of homoeopathy trials The opening statement of Ben Goldacre’s Comment on homoeopathy (Nov 17, p 1672)1 is false. He states that “Five large meta-analyses of homoeopathy trials have been done. All have had the same result: after excluding methodologically inadequate trials and accounting for publication bias, homoeopathy produced no statistically significant benefit over placebo.” In fact only one of the large meta-analyses of clinical trials of homoeopathy included correction for both publication bias and trial quality, and it showed that the effects of homoeopathy remained significantly greater than placebo when these, and other, corrections were applied singly or in combination. Remarkably, Goldacre fails even to cite this metaanalysis, although it was published in The Lancet.2 Goldacre’s statement is, without question, false. It would not have survived peer-review had this been a scientific paper. This is not a matter of scientific debate, it is plain inaccurate reporting. I declare that I have no conflict of interest.
Peter Fisher peter.fi[email protected]
Author’s reply I am grateful for the opportunity to clarify two small errors. One of my references was incorrect,1 and, as Peter Fisher correctly implies, I should have simply said: “...when these analyses were restricted to the best quality trials, homoeopathy produced no statistically significant benefit over placebo.” This trivial change makes no difference at all. Fisher cites one meta-analysis from 1997,2 which is frequently quoted by homoeopaths as proof of efficacy. However this positive finding is perhaps weak: the odds ratio for the 26 goodquality studies, corrected for publication bias, was 1·78 (95% CI 1·03–3·10). The same trials from this meta-analysis have been reanalysed extensively3 and significant doubt has been cast on the initial analysis. Most notably the same researchers, working on the same set of 89 trials4 in 1999 concluded that their own reanalysis “weakened the findings of our original meta-analysis… it seems, therefore, likely that our meta-analysis at least overestimated the effects of homeopathic treatments.” Many clinicians are clear that they can see a role for homoeopathy, even if it does perform no better than placebo. I would hope that homoeopaths might now divert their attention to performing randomised controlled (albeit unblinded) trials comparing “visiting a homoeopathy clinic” against “general practitioner’s treatment as usual”, since this might be the clinical question of more interest to patients—ie, not “do the pills work better than placebo” but “will the experience of visiting a homoeopath help me to feel better?” I write a column for the Guardian newspaper. I have just started an academic post funded by the UK National Health Service.
Ben Goldacre [email protected]
Royal London Homoeopathic Hospital, London WC1 3HR, UK
The Guardian, London EC1R 3ER, UK
1
1
2
Goldacre B. Benefits and risks of homoeopathy. Lancet 2007; 370: 1672–73. Linde K, Clausius N, Ramirez G, et al. Are the clinical effects of homoeopathy placebo effects? A meta-analysis of placebo-controlled trials. Lancet 1997; 350: 834–43.
Photolibrary
2
2
Kleijnen J, Knipschild P, ter Riet G. Clinical trials of homoeopathy. BMJ 1991; 302: 316–23. Linde K, Clausius N, Ramirez G. Are the clinical effects of homoeopathy placebo effects? A meta-analysis of placebo-controlled trials. Lancet 1997; 350: 834–43.
985