- Email: [email protected]
Hormone replacement therapy and breast cancer
THE LANCET
other events surrounding the publication. We entirely endorse the response made in the accompanying letter from the Director General Paul Nurse. Characteristically, Nurse has declined to comment on the attack you have made on him personally as Director General. Those who are acquainted with him will regard it as inconceivable that he would permit any tailoring of scientific interpretation, or of communication with the public of research results, so as to enhance fundraising at the expense of scientific evidence and the need for proper explanation of the findings. All the evidence we have at our disposal specifically negates your speculation. Furthermore, Council regards Nurse’s leadership since he took over the director generalship in September, 1996, as exemplary. We have sent your editorial and commentary, together with Nurse’s response and this letter to the Chief Charity Commissioner. *R D Cohen, T J H Clark, Nigel Althaus, John Skehel Imperial Cancer Research Fund, PO Box 123, London WC2A 3PX, UK 1 2
Horton R. ICRF: from mayhem to meltdown. Lancet 1997; 350: 1043–44. Collaborative Group on Hormonal Factors in Breast Cancer. Breast cancer and hormone replacement therapy: collaborative reanalysis of data from 51 epidemiological studies of 52 705 women with breast cancer and 108 411 women without breast cancer. Lancet 1997; 350: 1047–59.
Author’s reply In a letter to Ardi Kolah (head of communications, ICRF) dated Aug 29, 1997, Peter Selby (ICRF’s director of clinical research) wrote that the HRT overview “needs cautious interpretation if widespread panic is to be avoided with ladies stopping HRT all over the place. The absence of evidence for increased mortality and the likelihood of increased diagnosis not incidence needs more emphasis and specific cautions against over-interpretation would be wise. The importance of single vs combination HRT could be brought out more. For ICRF, Jack Cuzick should be the decider of our position and speak to the paper now and when published. That is his role” (underlinings in original). The final version of the paper—recall that the writing committee consisted of V Beral, D Bull, R Doll, T Key, R Peto, and G Reeves—was submitted to The Lancet over 2 weeks previously, on Aug 14. Selby’s letter was sent while the paper was undergoing qualitative and statistical peer review. Why was someone not responsible for either
Vol 350 • November 29, 1997
analysis of data or drafting of the report writing about the spin that the ICRF communications department should put on the HRT overview? Why, if “ICRF’s fundraisers neither seek nor are given responsibility for the interpretation of scientific data”, was this letter copied to Peter Vicary-Smith, head of ICRF fundraising? And why, given that Beral is the director of ICRF’s Cancer Epidemiology Unit, should Cuzick “be the decider of [ICRF’s] position”? The responses from Nurse and senior officers of ICRF’s Council thus raise many more perplexing questions than they answer. As Nurse points out, ICRF is justifiably anxious about any criticism it receives. In 1994, the UK Charity Commission concluded that ICRF’s “procedures for the supervision of research and control of the research results were not entirely satisfactory”. 3 years on, disconcerting problems still seem to exist, problems that should rekindle the charity commissioners’ concern. And readers might ask why, if Nurse and the members of his council are correct, no member of the overview writing committee has contacted The Lancet to support ICRF’s management of the HRT episode. Their silence seems telling. My principal charges remain unanswered. ICRF scientists do not have control over the presentation and interpretation of their data, as Selby’s letter to Kolah shows. The extraordinary attempts to induce The Lancet to collaborate with the ICRF communications department without the involvement of the HRT paper’s writing committee are dismissed too hastily by Nurse. Kolah’s first two telephone calls to me took place before the HRT paper was even accepted. How could he possibly have been calling, as Nurse claims, to make “arrangements for the news conference to take place at the time of the paper’s publication” when no decision to publish had been made? Nurse has been poorly informed about this sequence of events. Kolah’s Sept 29 letter to me (12 days after we had accepted the paper) invited a discussion about “how we will manage the publication” of the HRT overview; his concerns focused on “handling the publicity”. It is surely incredible that ICRF could allow any discussion about the publication of research conducted at one of their units without the full participation of the senior investigators. The fact that they were not invited to take part in that conversation and that Kolah first contacted us before The Lancet had made a decision to accept the overview
is surely proof that ICRF’s image comes before its science. Finally, although Nurse reassures us that it was not an ICRF scientist who leaked the paper, his response stops short of decisive action. ICRF’s “internal and external inquiries” need to be continued and published. Not to take this leak more seriously contradicts our shared view that the early disclosure of the paper led to an appalling mistake in the Sunday Times, London, UK with potentially damaging consequences to women. Richard Horton The Lancet, London, UK
SIR—We hugely enjoyed Richard Horton’s commentary1 about the ICRF. So have many of our concerned patients. It was beautifully composed, provided essential background information, and asked direct questions of the ICRF’s behaviour and motives. He does not comment upon the need for the Sunday Times to address the inaccuracies in Lois Roger’s original report (Sunday Times, Oct 5). Presumably, Horton thought that a correction would automatically appear. To state that the risk of developing breast cancer is 2·3 times higher, or more than double, that of non-users, rather than the relative risk being 1·023 per year of use—a hundredfold error— is so obvious that any reputable newspaper would have corrected it immediately. As far as we are aware the Sunday Times has issued no statement, correction, or apology. We have written but have had no response. The ICRF and the Sunday Times seem equally driven by a need for publicity, irrespective of accuracy, and are equally incapable of apologising when errors are made. It is difficult to understand why either organisation merits our trust or respect. *Malcolm Whitehead, Val Godfree Research Centre for Women’s Health, Amarant Centre at the Churchill Clinic, London SE1 7PW, UK 1
Horton R. ICRF: from mayhem to meltdown. Lancet 1997; 350: 1043–44.
Hormone replacement therapy and breast cancer S IR —In their meta-analysis of hormonal replacement therapy (HRT) and breast cancer the Collaborative Group on Hormonal Factors in Breast Cancer (Oct 11, p 1047)1 conclude that there was an increased cancer risk only among current or recent users, and that the tumours diagnosed in
1627
THE LANCET
women who used HRT were more likely to be localised, early-stage disease. These observations are remarkably similar to the associations of hormonal contraception and breast cancer. Several researchers report increased risk of less advanced breast cancer confined to young women with current or recent exposures to oral contraceptives2 or to Depo Provera.3 The consistency of these steroid hormonal effects, irrespective of whether the preparations contain oestrogens or progestagens, does not suggest that hormones promote breast cancer. If these preparations promoted the growth of malignant tumours, one would expect more rapid disease progression and thus more advanced stage tumours in hormone users at the time of diagnosis. By contrast, all the studies report more localised disease in hormone users.1–3 The epidemiological associations are, however, consistent with a diagnostic bias whereby tumours among current or recent hormone users tend to be diagnosed at an earlier stage than among non-users. Such a lead-time bias could occur if the hormones unmasked smaller malignant lesions. A mechanism for such unmasking is suggested by the known effects of steroids on benign breast disease. All these hormonal therapies suppress the growth of benign breast tumours, such as fibroadenomas or fibrocystic disease, and the suppressive effects diminish with the longer interval since last hormone use. 4,5 Thus, hormonal suppression of benign breast lumps is likely to facilitate detection and diagnosis of smaller, earlier-stage malignant tumours (because the latter are not inhibited by the steroids); such an unmasking effect would be manifest only among current or recent hormone users, because with longer intervals since last use, the suppression of benign lumps is diminished. Thus the association between hormone replacement, oral or injectable contraception, and increased risk of breast cancer might, in whole or in part, be due to a diagnostic bias. Ronald H Gray School of Hygiene and Public Health, Johns Hopkins University, Baltimore, MD 21205, USA 1
2
Collaborative Group on Hormonal Factors in Breast Cancer. Breast cancer and hormone replacement therapy: collaborative reanalysis of data from 51 epidemiological studies of 52 705 women with breast cancer and 108 411 women without breast cancer. Lancet 1997; 350: 1047–59. Collaborative Group on Hormonal Factors in Breast Cancer. Breast cancer and hormonal contraceptives: collaborative reanalysis of individual data on 53 297 women with breast
1628
3
4
5
cancer and 100 239 women without breast cancer from 54 epidemiological studies. Lancet 1996; 347: 1713–26. Skegg DC, Noonan EA, Paul C, Spears GFS, Merik O, Thomas DB. Depot medroxyprogesterone acetate and breast cancer. A pooled analysis of the World Health Organization and New Zealand studies. JAMA 1995; 273: 799–804. Brinton LA, Vessey MA, Flavel R, Yeates D. Risk factors for benign breast disease. Am J Epidemiol 1981; 113: 203–14. Ernster VL. The epidemiology of benign breast disease. Epidemiol Rev 1981; 3: 184–202.
S IR —The Collaborative Group on Hormonal Factors in Breast Cancer1 present data showing an increasing risk of developing breast cancer in women who use hormone replacement therapy (HRT). In this, as well as in similar epidemiological studies, one fundamental bias is not considered: HRT is prescribed without reflection and monitoring and without individualisation for the particular patient. HRT is substitution and balancing of a deficient endocrine situation. Before and after starting HRT, some basic clinical questions should be considered, which has so far not been done. Is the patient really suffering from a hormonal deficiency? Which steroid is missing, and does HRT induce normal or supraphysiological hormone concentrations? Some physicians prescribe hormones as if they were vitamins. A free lunch of hormones for patients who do not need them stimulates many intracellular events in endocrinedependent tissues. Furthermore, is oestrogen substitution the first choice of HRT in every patient? It is well known that menopause often starts with a progesterone deficiency and an oestrogen excess. An unreflected addition of oestrogen induces side-effects, reduces patient compliance, and may also be responsible for cell transformation. The most important clinical point and the greatest bias in such studies is the uncontrolled conduct of HRT. Sexual steroids are transcriptional factors, they do not need the second messenger pathway and have an immediate access to genome and transcriptional machinery. No ethical committee today would approve a therapy with new transcriptional factors in such an uncontrolled and light-minded design as is seen with HRT. In another context, a higher risk for endometrial cancer was registered in patients with combined therapy than in patients with no therapy at all.2 Supraphysiological oestradiol concentrations were suspected to be involved in this problem. Recently, a highly significant relation between oestrogen-induced bone mass and the risk of breast cancer was
published.3 The researchers postulated that high bone density may indicate the effect of cumulative oestrogen exposure. It is the dosage that makes a pharmaceutical product into either a poison or a remedy. This fact should henceforth be recognised in clinical practice as well as in the interpretation of epidemiological papers. *D M Gruber, M O Sator, J C Huber Department of Gynecological Endocrinology and Reproductive Medicine, University of Vienna, A-1090 Wien, Austria 1
2
3
Collaborative Group on Hormonal Factors in Breast Cancer. Breast cancer and hormone replacement therapy: collaborative reanalysis of data from 51 epidemiological studies of 52 705 women with breast cancer and 108 411 women without breast cancer. Lancet 1997; 350: 1047–59. Beresford SAA, Weiss NS, Voigt LF, McKnight B. Risk of endometrial cancer in relation to use of oestrogen combined with cyclic progestagen therapy in postmenopausal women. Lancet 1997; 349: 458–61. Zhang Y, Kiel DP, Kreger BE, et al. Bone mass and the risk of breast cancer among postmenopausal women. N Engl J Med 1997; 336: 611–17.
Malarone-donation programme See page 1624 S IR —Peter Bloland and colleagues (p 1624)1 address the important issue of parasite resistance in the control of malaria and the potential impact of inappropriate introduction of new antimalarials in endemic areas. The questions they raise about the Malarone-donation programme are ones to which we have given great thought from the programme’s inception. We agree that the programme must move forward with great care, and it is. Malaria is a serious and complex health issue. It is the world’s most lethal tropical parasite disease and causes over 2 million deaths each year. About 40% of the world’s population reside in endemic areas, mainly the poorest parts of Africa, Asia, and Latin America. Malarone, a combination of atovaquone and proguanil HCl, is an important new antimalarial, but it is expensive to produce and patients in countries where malaria is endemic would have limited access to it. Consequently, Glaxo Wellcome proposed to us that it donate up to 1 million treatment doses each year. The company feels strongly that the programme be developed and implemented appropriately and responsibly, from an international public-health perspective. From the
Vol 350 • November 29, 1997
other events surrounding the publication. We entirely endorse the response made in the accompanying letter from the Director General Paul Nurse. Characteristically, Nurse has declined to comment on the attack you have made on him personally as Director General. Those who are acquainted with him will regard it as inconceivable that he would permit any tailoring of scientific interpretation, or of communication with the public of research results, so as to enhance fundraising at the expense of scientific evidence and the need for proper explanation of the findings. All the evidence we have at our disposal specifically negates your speculation. Furthermore, Council regards Nurse’s leadership since he took over the director generalship in September, 1996, as exemplary. We have sent your editorial and commentary, together with Nurse’s response and this letter to the Chief Charity Commissioner. *R D Cohen, T J H Clark, Nigel Althaus, John Skehel Imperial Cancer Research Fund, PO Box 123, London WC2A 3PX, UK 1 2
Horton R. ICRF: from mayhem to meltdown. Lancet 1997; 350: 1043–44. Collaborative Group on Hormonal Factors in Breast Cancer. Breast cancer and hormone replacement therapy: collaborative reanalysis of data from 51 epidemiological studies of 52 705 women with breast cancer and 108 411 women without breast cancer. Lancet 1997; 350: 1047–59.
Author’s reply In a letter to Ardi Kolah (head of communications, ICRF) dated Aug 29, 1997, Peter Selby (ICRF’s director of clinical research) wrote that the HRT overview “needs cautious interpretation if widespread panic is to be avoided with ladies stopping HRT all over the place. The absence of evidence for increased mortality and the likelihood of increased diagnosis not incidence needs more emphasis and specific cautions against over-interpretation would be wise. The importance of single vs combination HRT could be brought out more. For ICRF, Jack Cuzick should be the decider of our position and speak to the paper now and when published. That is his role” (underlinings in original). The final version of the paper—recall that the writing committee consisted of V Beral, D Bull, R Doll, T Key, R Peto, and G Reeves—was submitted to The Lancet over 2 weeks previously, on Aug 14. Selby’s letter was sent while the paper was undergoing qualitative and statistical peer review. Why was someone not responsible for either
Vol 350 • November 29, 1997
analysis of data or drafting of the report writing about the spin that the ICRF communications department should put on the HRT overview? Why, if “ICRF’s fundraisers neither seek nor are given responsibility for the interpretation of scientific data”, was this letter copied to Peter Vicary-Smith, head of ICRF fundraising? And why, given that Beral is the director of ICRF’s Cancer Epidemiology Unit, should Cuzick “be the decider of [ICRF’s] position”? The responses from Nurse and senior officers of ICRF’s Council thus raise many more perplexing questions than they answer. As Nurse points out, ICRF is justifiably anxious about any criticism it receives. In 1994, the UK Charity Commission concluded that ICRF’s “procedures for the supervision of research and control of the research results were not entirely satisfactory”. 3 years on, disconcerting problems still seem to exist, problems that should rekindle the charity commissioners’ concern. And readers might ask why, if Nurse and the members of his council are correct, no member of the overview writing committee has contacted The Lancet to support ICRF’s management of the HRT episode. Their silence seems telling. My principal charges remain unanswered. ICRF scientists do not have control over the presentation and interpretation of their data, as Selby’s letter to Kolah shows. The extraordinary attempts to induce The Lancet to collaborate with the ICRF communications department without the involvement of the HRT paper’s writing committee are dismissed too hastily by Nurse. Kolah’s first two telephone calls to me took place before the HRT paper was even accepted. How could he possibly have been calling, as Nurse claims, to make “arrangements for the news conference to take place at the time of the paper’s publication” when no decision to publish had been made? Nurse has been poorly informed about this sequence of events. Kolah’s Sept 29 letter to me (12 days after we had accepted the paper) invited a discussion about “how we will manage the publication” of the HRT overview; his concerns focused on “handling the publicity”. It is surely incredible that ICRF could allow any discussion about the publication of research conducted at one of their units without the full participation of the senior investigators. The fact that they were not invited to take part in that conversation and that Kolah first contacted us before The Lancet had made a decision to accept the overview
is surely proof that ICRF’s image comes before its science. Finally, although Nurse reassures us that it was not an ICRF scientist who leaked the paper, his response stops short of decisive action. ICRF’s “internal and external inquiries” need to be continued and published. Not to take this leak more seriously contradicts our shared view that the early disclosure of the paper led to an appalling mistake in the Sunday Times, London, UK with potentially damaging consequences to women. Richard Horton The Lancet, London, UK
SIR—We hugely enjoyed Richard Horton’s commentary1 about the ICRF. So have many of our concerned patients. It was beautifully composed, provided essential background information, and asked direct questions of the ICRF’s behaviour and motives. He does not comment upon the need for the Sunday Times to address the inaccuracies in Lois Roger’s original report (Sunday Times, Oct 5). Presumably, Horton thought that a correction would automatically appear. To state that the risk of developing breast cancer is 2·3 times higher, or more than double, that of non-users, rather than the relative risk being 1·023 per year of use—a hundredfold error— is so obvious that any reputable newspaper would have corrected it immediately. As far as we are aware the Sunday Times has issued no statement, correction, or apology. We have written but have had no response. The ICRF and the Sunday Times seem equally driven by a need for publicity, irrespective of accuracy, and are equally incapable of apologising when errors are made. It is difficult to understand why either organisation merits our trust or respect. *Malcolm Whitehead, Val Godfree Research Centre for Women’s Health, Amarant Centre at the Churchill Clinic, London SE1 7PW, UK 1
Horton R. ICRF: from mayhem to meltdown. Lancet 1997; 350: 1043–44.
Hormone replacement therapy and breast cancer S IR —In their meta-analysis of hormonal replacement therapy (HRT) and breast cancer the Collaborative Group on Hormonal Factors in Breast Cancer (Oct 11, p 1047)1 conclude that there was an increased cancer risk only among current or recent users, and that the tumours diagnosed in
1627
THE LANCET
women who used HRT were more likely to be localised, early-stage disease. These observations are remarkably similar to the associations of hormonal contraception and breast cancer. Several researchers report increased risk of less advanced breast cancer confined to young women with current or recent exposures to oral contraceptives2 or to Depo Provera.3 The consistency of these steroid hormonal effects, irrespective of whether the preparations contain oestrogens or progestagens, does not suggest that hormones promote breast cancer. If these preparations promoted the growth of malignant tumours, one would expect more rapid disease progression and thus more advanced stage tumours in hormone users at the time of diagnosis. By contrast, all the studies report more localised disease in hormone users.1–3 The epidemiological associations are, however, consistent with a diagnostic bias whereby tumours among current or recent hormone users tend to be diagnosed at an earlier stage than among non-users. Such a lead-time bias could occur if the hormones unmasked smaller malignant lesions. A mechanism for such unmasking is suggested by the known effects of steroids on benign breast disease. All these hormonal therapies suppress the growth of benign breast tumours, such as fibroadenomas or fibrocystic disease, and the suppressive effects diminish with the longer interval since last hormone use. 4,5 Thus, hormonal suppression of benign breast lumps is likely to facilitate detection and diagnosis of smaller, earlier-stage malignant tumours (because the latter are not inhibited by the steroids); such an unmasking effect would be manifest only among current or recent hormone users, because with longer intervals since last use, the suppression of benign lumps is diminished. Thus the association between hormone replacement, oral or injectable contraception, and increased risk of breast cancer might, in whole or in part, be due to a diagnostic bias. Ronald H Gray School of Hygiene and Public Health, Johns Hopkins University, Baltimore, MD 21205, USA 1
2
Collaborative Group on Hormonal Factors in Breast Cancer. Breast cancer and hormone replacement therapy: collaborative reanalysis of data from 51 epidemiological studies of 52 705 women with breast cancer and 108 411 women without breast cancer. Lancet 1997; 350: 1047–59. Collaborative Group on Hormonal Factors in Breast Cancer. Breast cancer and hormonal contraceptives: collaborative reanalysis of individual data on 53 297 women with breast
1628
3
4
5
cancer and 100 239 women without breast cancer from 54 epidemiological studies. Lancet 1996; 347: 1713–26. Skegg DC, Noonan EA, Paul C, Spears GFS, Merik O, Thomas DB. Depot medroxyprogesterone acetate and breast cancer. A pooled analysis of the World Health Organization and New Zealand studies. JAMA 1995; 273: 799–804. Brinton LA, Vessey MA, Flavel R, Yeates D. Risk factors for benign breast disease. Am J Epidemiol 1981; 113: 203–14. Ernster VL. The epidemiology of benign breast disease. Epidemiol Rev 1981; 3: 184–202.
S IR —The Collaborative Group on Hormonal Factors in Breast Cancer1 present data showing an increasing risk of developing breast cancer in women who use hormone replacement therapy (HRT). In this, as well as in similar epidemiological studies, one fundamental bias is not considered: HRT is prescribed without reflection and monitoring and without individualisation for the particular patient. HRT is substitution and balancing of a deficient endocrine situation. Before and after starting HRT, some basic clinical questions should be considered, which has so far not been done. Is the patient really suffering from a hormonal deficiency? Which steroid is missing, and does HRT induce normal or supraphysiological hormone concentrations? Some physicians prescribe hormones as if they were vitamins. A free lunch of hormones for patients who do not need them stimulates many intracellular events in endocrinedependent tissues. Furthermore, is oestrogen substitution the first choice of HRT in every patient? It is well known that menopause often starts with a progesterone deficiency and an oestrogen excess. An unreflected addition of oestrogen induces side-effects, reduces patient compliance, and may also be responsible for cell transformation. The most important clinical point and the greatest bias in such studies is the uncontrolled conduct of HRT. Sexual steroids are transcriptional factors, they do not need the second messenger pathway and have an immediate access to genome and transcriptional machinery. No ethical committee today would approve a therapy with new transcriptional factors in such an uncontrolled and light-minded design as is seen with HRT. In another context, a higher risk for endometrial cancer was registered in patients with combined therapy than in patients with no therapy at all.2 Supraphysiological oestradiol concentrations were suspected to be involved in this problem. Recently, a highly significant relation between oestrogen-induced bone mass and the risk of breast cancer was
published.3 The researchers postulated that high bone density may indicate the effect of cumulative oestrogen exposure. It is the dosage that makes a pharmaceutical product into either a poison or a remedy. This fact should henceforth be recognised in clinical practice as well as in the interpretation of epidemiological papers. *D M Gruber, M O Sator, J C Huber Department of Gynecological Endocrinology and Reproductive Medicine, University of Vienna, A-1090 Wien, Austria 1
2
3
Collaborative Group on Hormonal Factors in Breast Cancer. Breast cancer and hormone replacement therapy: collaborative reanalysis of data from 51 epidemiological studies of 52 705 women with breast cancer and 108 411 women without breast cancer. Lancet 1997; 350: 1047–59. Beresford SAA, Weiss NS, Voigt LF, McKnight B. Risk of endometrial cancer in relation to use of oestrogen combined with cyclic progestagen therapy in postmenopausal women. Lancet 1997; 349: 458–61. Zhang Y, Kiel DP, Kreger BE, et al. Bone mass and the risk of breast cancer among postmenopausal women. N Engl J Med 1997; 336: 611–17.
Malarone-donation programme See page 1624 S IR —Peter Bloland and colleagues (p 1624)1 address the important issue of parasite resistance in the control of malaria and the potential impact of inappropriate introduction of new antimalarials in endemic areas. The questions they raise about the Malarone-donation programme are ones to which we have given great thought from the programme’s inception. We agree that the programme must move forward with great care, and it is. Malaria is a serious and complex health issue. It is the world’s most lethal tropical parasite disease and causes over 2 million deaths each year. About 40% of the world’s population reside in endemic areas, mainly the poorest parts of Africa, Asia, and Latin America. Malarone, a combination of atovaquone and proguanil HCl, is an important new antimalarial, but it is expensive to produce and patients in countries where malaria is endemic would have limited access to it. Consequently, Glaxo Wellcome proposed to us that it donate up to 1 million treatment doses each year. The company feels strongly that the programme be developed and implemented appropriately and responsibly, from an international public-health perspective. From the
Vol 350 • November 29, 1997