Letters

Cardioprotective effect of hormone replacement therapy Is not due to a selection bias

In their paper they talk about "total cancer within each study," but in many instances the studies were reporting the mortality from cancer rather than the incidence of cancer. This is an important distinction because, typically, patients who die of cancer have had the disease diagnosed previously. It is well known that many physicians refrain from giving hormones to postmenopausal women who have already been diagnosed as having cancer. Hence studies of mortality from cancer will naturally tend to show a spurious protective association of postmenopausal hormones unless women with cancer at the baseline have been excluded. This spurious association has nothing to do with the selection of women according to their risk of coronary disease.

A direct and more relevant approach would have been to assess the distribution of risk factors for coronary heart disease among users and non- users of oestrogen. Most studies that have made this comparison have indeed found that users tend to have a more favourable profile of risk factors. Because this is primarily a sociological rather than a biological association, however, the differences in the profiles of risk factors will vary widely according to the population studied. In particular, one would expect that in general population surveys the differences in the profiles of risk factors would be greater than those in the relatively homogeneous populations from which much of our understanding of the epidemiology of hormones and heart disease derives. The table shows the distribution of risk factors for two large epidemiological studies that have provided data on oestrogen replacement therapy and heart disease; the differences in the profiles of risk factors are fairly modest. Moreover, adjustment for these differences has not materially altered the estimates of relative risk. Hence selection of healthier women for hormone treatment is unlikely to explain a major portion of the reduction in coronary heart disease.

Profiles of risk factors in users and non-users of oestrogen in two studies
----------------------------------------------------------
                                Current    Past    Never
                                 user*     user    used
----------------------------------------------------------
                       Nurses' health study4

Current smoker (%)                11.2     14.7     14.5
% With hypertension               23.2     25.0     21.8
% With diabetes                    2.7      3.8      3.5
% With high serum
 cholesterol                       9.9     11.2      7.6
% With parental myocardial
 infarction before age 60         10.6     10.0      9.3
% Taking vigorous physical
 activity >=once per week         48.2     43.1     42.4
% With body mass index
 >=29 kg/m2                  9.8     13.3     15.0

            Lipid Research Clinics programme5

 <High school education           16                25
% Smokers                         33                31
% Taking regular exercise         12                10
% Drinking alcohol                82                79
Mean body mass index
 (kg/m2                     24.7              25.7
Mean age (years)                  53.8              52.6
Mean systolic blood pressure
 (mm Hg)                         129.0             127.7
Mean diastolic blood
 pressure (mm Hg)                 79.9              79.5
Mean cholesterol                 234.8             235.2
----------------------------------------------------------
 * In Lipid Research Clinics programme data for current and past users were
combined.

M Stampfer, F Grodstein

Harvard School of Public Health, Boston, MA, USA Channing Laboratory, Department of Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, MA 02115, USA

1. Posthuma WFM, Westendorp RGJ, Vendenbroucke JP. Cardioprotective effect of hormone replacement therapy in postmenopausal women: is the evidence biased? BMJ 1994;308:1268-9. (14 May.)
2. Posthuma WFM, Westendorp RGJ, Vandenbroucke JP. Hormone replacement therapy. BMJ 1994;309:191-2. (16 July.) [Free Full Text]
3. Stevenson JC, Baum M. Hormone replacement therapy. BMJ 1994;309:191. (16 July.)
4. Stampfer MJ, Colditz GA, Willett WC, Manson JE, Rosner B, Speizer FE, et al. Postmenopausal estrogen therapy and cardiovascular disease: ten-year follow-up from the nurses' health study. N Engl J Med 1991;325:756-62. [Abstract]
5. Bush TL, Barrett-Connor E, Cowan LD, Criqui MH, Wallace RB, Suchindran CM, et al. Cardiovascular mortality and noncontraceptive use of estrogen in women: results from the Lipid Research Clinics program follow-up study. Circulation 1987;75:1102-9. [Abstract/Free Full Text]

Benefits women with established cardiovascular disease

EDITOR, - Ward F M Posthuma and colleagues postulate that the reported benefits of hormone replacement therapy in reducing the risk of cardiovascular disease in postmenopausal women may be due to unintended selection of relatively healthy women.¹ The implication of this is that, had sufficient women with cardiovascular disease been present at the start of the epidemiological studies, a result with a lower significance would have been obtained. It is relatively simple to analyse this hypothesis by using data that were only partially considered in their review.²

Sullivan et al presented findings from a cohort of 1822 women with cardiovascular disease documented angiographically.² Over 10 years the relative risk of death from cardiovascular disease fell by 89% among women using oestrogen compared with non-users. This implies that there is a greater benefit for postmenopausal women with established cardiovascular disease than for healthy women. It should also be remembered that premenopausal women with oestrogen deficiency resulting from a surgical or premature menopause have a risk of cardiovascular disease roughly three times that of women with normal oestrogen concentrations.

A second fundamental flaw in Posthuma and colleagues' paper is that the authors chose total cancer as being a disease that is "unlikely to be influenced by oestrogen." This assumption may be unsound. Fotsis et al showed that 2-methoxyoestradiol, an endogenous oestrogen metabolite, is a potent inhibitor of the proliferation and migration of endothelial cells and an inhibitor of angiogenesis both in vitro and in vivo.³ Angiogenesis is essential for successful tumour growth, and the antiangiogenic properties of interferon alfa have been shown to be promising in the treatment of haemangiomas.⁴ Such findings in no way confirm a protective effect against cancer in oestrogen users after the menopause but may go some way to explaining the absence of the much predicted increase in breast cancer in such women. More fundamentally, it shows that one must be careful about making assertions about the impossibility of an observed effect.

Although the degree of protection that oestrogen replacement therapy offers against cardiovascular disease may be subject to selection bias favouring healthy subjects, whether this results in an overestimate or an underestimate of the true benefit cannot be stated with certainty. It remains to be elucidated whether oestrogen replacement therapy may also reduce the risk of all cancers in postmenopausal women through an effect on angiogenesis.

D Spiers

E Merck Pharmaceuticals, Alton, Hampshire GU34 1HG

1. Posthuma WFM, Westendorp RGJ, Vandenbroucke JP. Cardioprotective effect of hormone replacement therapy in postmenopausal women: is the evidence biased? BMJ 1994;308:1268-9. (14 May.)
2. Sullivan JM, vander Zwaag, Hughes JP, Maddock V, Kroetz FW, Ramanathan KB, et al. Estrogen replacement and coronary heart disease: effect on survival in postmenopausal women. Arch Intern Med 1990;150:2557-62. [Abstract]
3. Fotsis T, Zhang Y, Pepper MS, Adlercreutz H, Montesano R, Nawroth PP, et al. The endogenous oestrogen metabolite 2-methoxyoestradiol inhibits angiogenesis and suppresses tumour growth. Nature 1994;368:237-9. [Medline]
4. White CW, Sondheimer HM, Crouch HM, Wilson EC, Fan LL. Treatment of pulmonary hemangiomatosis with recombinant interferon alfa-2a. N Eng J Med 1989;320:1197-200. [Medline]

Authors' reply

EDITOR, - For reasons other than David Spiers suggests, the paper by Sullivan et al clarifies the subject. The figure, based on their data, shows the percentage of women using oestrogens at the time of cardiac catheterisation for each year of the study. The downward trend in the late 1970s indicates that the prescription of oestrogens was omitted in patients with signs and symptoms of coronary artery disease. The authors ascribed this reluctance to prescribe oestrogens to the increased frequency of myocardial infarction observed in men receiving oestrogens for secondary prevention. Most probably, oestrogen replacement therapy was stopped in patients who developed coronary artery disease and continued only in those with a favourable course. This differential prescription may explain why the effect of use of oestrogen was not significant when the data were analysed by treatment received at the time of angiography. It also explains the lower mortality in women who subsequently used oestrogens.¹

View larger version (16K): [in this window] [in a new window]   Percentage of women receiving oestrogen replacement therapy at time of diagnostic coronary angiography (data from Sullivan et al's study1)

Even more fundamental than making assertions about the (im)possibility of an effect is weighing arguments for a plausible biological explanation.² Here we have to consider the possible beneficial effect of 2- methoxyoestradiol as an antiangiogenic compound not only for haemangiomas but, as stated in the original paper, also for rheumatoid arthritis, psoriasis, and diabetic retinopathy. As 2-methoxyoestradiol is an endogenous oestrogen metabolite the concentrations are probably higher in women. 2-Methoxyoestradiol is unlikely to have an important action because in most of these diseases neither sex predominates, whereas rheumatoid arthritis is almost exclusively a disease of women. On the other hand, recent Swedish evidence has led to the conclusion that the high oestrogen concentrations associated with pregnancy have a long term protective effect against breast cancer by inducing differentiation of normal mammary stem cells but increase the risk by stimulating the growth of cells that have undergone the early stages of malignant transformation.³

Now we know that the observational studies on cardiovascular diseases have been subject to selection of healthy subjects we should realise that their results are difficult to interpret.^4,5 They give us clues for further development of scientific ideas on the relation between sex hormones and cardiovascular disease. An inference about preventive measures is a step too far.

W F M Posthuma, R G J Westendorp, J P Vandenbroucke 

Department of Clinical Epidemiology, Leiden University Hospital, PO Box 9600, 2300 RC Leiden, Netherlands.

1. Sullivan JM, vander Zwaag R, Hughes JP, Maddock V, Kroetz FW, Ramanathan KB, et al. Estrogen replacement and coronary artery disease: effect on survival in postmenopausal women. Arch Intern Med 1990;150:2557-62.
1. Cole P. The hypothesis generating machine. Epidemiology 1993;4:271-3. [Medline]
2. Lambe M, Hsieh C, Trichopoulos D, Ekbom A, Pavia M, Adami H-O. Transient increase in the risk of breast cancer after giving birth. N Engl J Med 1994;331:5-9. [Abstract/Free Full Text]
3. Hemminki E, Sihvo S. A review of postmenopasal hormone therapy recommendations: potential for selection bias. Obstet Gynecol 1993;82:1021-8. [Abstract/Free Full Text]
4. Rosenberg L. Hormone replacement therapy: The need for reconsideration. Am J Publ Health 1993;83:1670-3. [Abstract/Free Full Text]

CiteULike    Complore    Connotea    Del.icio.us    Digg    Reddit    Technorati    What's this?

Relevant Article

Cardioprotective effect of hormone replacement therapy in postmenopausal women: is the evidence biased?

W F M Potshuma, R G J Westendorp, and J P Vandenbroucke
BMJ 1994 308: 1268-1269. [Abstract] [Full Text]