Influence of HRT on prognostic factors for breast cancer: a systematic review after the Women's Health Initiative trial

INTRODUCTION: Mortality due to breast cancer has been reported to be the same or even lower in HRT users than in non-users. This has been attributed to earlier diagnosis and to better prognosis. Nevertheless, more advanced disease in HRT users was reported recently by the Women's Health Initiative (WHI) study. The objective of this study was to assess, using a systematic review of current literature, whether the data of the WHI study are in contradiction to observational data. METHODS: We selected 25 studies, for which we evaluated the methodology, the characteristics of the studied populations, confounding breast cancer risk factors and prognostic indicators. RESULTS: The WHI study, showing a worsening of some prognostic parameters, is in contradiction to most published observational studies. Most observational studies are retrospective, not well matched and did not consider most confounding factors. Their methodology and selection criteria varied considerably and the number of patients was often small. No differences in the distributions of histology, grade or steroid receptors were observed in the WHI trial, while this was the case in some of the observational studies. Other parameters (S phase, protein Neu, Bcl-2 gene, protein p53 and E-cadherin, cathepsin D) were not reported in the WHI trial. CONCLUSIONS: In view of these data, the current clinical message to patients should be changed: one can no longer declare that breast cancers developed while using HRT are of better prognosis.     

Issues to debate on the Women's Health Initiative (WHI) study. Hormone replacement therapy: an epidemiological dilemma?

In July 2002 the data of the prematurely stopped Estrogen plus Progestin study of the Women's Health Initiative (WHI) were presented in the Journal of the American Medical Association. The results of the Heart and Estrogen/Progestin Replacement Study (HERS/HERS II) were published in the same issue. The results of WHI for healthy postmenopausal women often are interpreted to be in analogy with the HERS/HERS II results for postmenopausal women with established coronary heart disease. As a result of HERS/HERS II and WHI, use of HRT in general became questionable. This is an unjustified judgement of HRT in general. This synoptic review and criticism of both studies will show the methodological weaknesses and their consequences and the reasons for limited generalizability of the study results from WHI and HERS/HERS II on normal HRT users.     

Issues to debate on the Women's Health Initiative (WHI) study. Epidemiology or randomized clinical trials--time out for hormone replacement therapy studies?

Over the last 40 years, there has been increasing epidemiological evidence that post-menopausal treatment with sex steroids in physiological doses may reduce the relative risk of cardiovascular disease (CVD). These findings have been supported by biological studies showing favourable changes in cardiovascular risk factors with estrogen supplementation. The impact of the so-called 'healthy user' bias has been eagerly debated, and the results of the first and only randomized long-term clinical trial of HRT for primary prevention have therefore been long awaited. The dramatic decision to halt the Women's Health Initiative (WHI) study before completion came unexpectedly as the consequence of not only an increased risk of breast cancer but also increased occurrence of cardiovascular events with HRT. Due to the superior design of the study, the results from the WHI study have had an enormous impact on the clinical recommendations of HRT to post-menopausal women, concurrent with a degradation of evidence from observational studies. It is not very likely that other long-term randomized clinical trials (RCTs) will be completed and epidemiology has certainly been disreputed-so is this 'time out' for HRT studies?     

Issues to debate on the Women's Health Initiative: failure of estrogen plus progestin therapy for prevention of breast cancer risk

Several studies on hormone replacement therapy (HRT) in the USA have been published. They revealed that the risk of breast cancer is increased with HRT more than with estrogen alone (ERT). A progestin has been given with each dose of ERT, as was the case in the Women's Health Initiative (WHI) study. The results of studies in Europe show similar trends. The increased risk of breast cancer in the WHI study was significantly higher only in women who had used HRT for several years before entering the study. The study was non-blind in 3444 cases, i.e. 40.5% of women in the estrogen plus progestin group and 6.8% in controls. If the women in the HRT group had more mammographic examinations it could change the validity of the results of the study. Estradiol-containing drugs have now been added to the list of carcinogens and the packages of these drugs have warning labels. The results of the WHI study do not support this labelling. The results of the WHI study show that the administration of HRT increases the risks of stroke and pulmonary embolism. It is reasonable to think that in the case of bleeding, at least at weekends in nursing homes (when staff levels may be low) patients were immobilized in their beds. Immobilization among women on HRT could have been more dangerous than the HRT itself. Progestins need to be delivered to the endometrium in a manner that will have the least effect on the breast. Systemic administration can be replaced by releasing progestin locally in the uterine cavity. Endometrial protection with a levonorgestrel-releasing intra-uterine system (IUS) is well tolerated. The high hepatic concentrations of estrogens given orally could be avoided by transdermal administration. New studies should be planned to reflect the situation in clinical practice. The time to start HRT in healthy menopausal women is between the ages of 45 to 55 years.     

HRT as secondary prevention of cardiovascular disease

Objective: To review the evidence of the efficacy of postmenopausal hormone replacement therapy (HRT) in secondary prevention of coronary artery disease or stroke. Results: Although a number of rather large and prolonged non-randomized observational studies have produced convincing and consistent evidence of the efficacy of HRT in the prevention of recurrence of cardiac events, the first randomized, placebo controlled trial (RCT) on heart disease and estrogen replacement study (HERS) reported no benefit of conjugated equine estrogen (CEE) and medroxyprogesterone acetate (MPA) in secondary prevention of cardiac events in women with established coronary artery disease. This was supported by RCT reporting no effect of CEE or CEE + MPA on the progress of coronary sclerosis. Similarly, some nonrandomized observational studies have evaluated the risk of recurrent stroke in regard to the use of HRT, and the data are conflicting reporting a reduced or increased risk of recurrence for HRT users. One RCT has shown that low-dose estrogen treatment can only slow down the progression of carotid arteriosclerosis in high-risk postmenopausal women, whereas two other RCTs have shown no benefit (or risk) of using HRT for secondary prevention of ischemic stroke or progression of carotid atherosclerosis. Conclusion: The evidence accumutaed so far shows that HRT has no place in secondary prevention of coronary or carotic artery disease. Its use in these patients must be based on solid nonvascular indications and expected benefits from these causes.     

Modulation of Thrombophilia Genes by Environmental Factors

Hormone replacement therapy (HRT) is associated with reduced risk of coronary heart disease (CHD) and stroke in observational studies; however the possibility of confounding by other risk factors requires prospective assessment of its risks and benefits in randomised controlled trials. The HERS trial of oral HRT in secondary CHD prevention observed an early increased risk of myocardial infarction (MI) and venous thromboembolism (VTE) with HRT: the latter risk has been confirmed by other prospective and case-control studies, and a past history of VTE or MI is now a contraindication to oral HRT. Other prospective randomised trials of HRT, CHD and stroke are in progress. Potential prothrombotic effects of oral HRT (but probably not transdermal HRT) include increased plasma factors VII and IX, activated protein C resistance and C-reactive protein; and decreased antithrombin, protein C and S, and tissue factor pathway inhibitor. Potential protective effects of HRT include decreased blood pressure, lipids, glucose intolerance, fibrinogen, viscosity and plasminogen activator inhibitor; and increased endothelial function. The overall balance of prothrombotic and protective effects varies with HRT preparations and individual women: and may be clarified by ongoing large randomised trials and case-control studies (and substudies of trials).     

HRT, osteoporosis and regulatory authorities Quis custodiet ipsos custodes?

HRT has been widely used for the relief of menopausal symptoms and the prevention and treatment of post-menopausal osteoporosis. However, following the publication of the Women's Health Initiative (WHI) and the Million Women Study (MWS), regulatory authorities issued an urgent safety restriction on HRT use in preventing post-menopausal osteoporosis, recommending that it now be considered a second-line treatment. Are such recommendations justified? Treatments for osteoporosis, in women with increased future risk for fractures but who have not yet developed the disease, should prevent all types of osteoporotic fractures. Of the available therapies, none other than HRT has been clearly demonstrated to prevent hip fractures in such women. Thus, HRT should be recommended as first-line treatment for osteoporosis prevention. Potential risks of HRT, such as increased development of breast cancer and increased thromboembolism, have long been known. The WHI showed risks in less than 0.3% of women studied, and the MWS appears to have overestimated the risk of breast cancer. Thus, no new safety issues have been identified, and the regulatory authorities may have misinterpreted the data from these recent studies. When given for the correct indications, HRT is of major benefit to many women.     

Interactions between melatonin and estrogen may regulate cerebrovascular function in women: clinical implications for the effective use of HRT during menopause and aging

A number of clinical trials associated with the Women's Health Initiative (WHI) have assessed the potential benefits of hormone replacement therapy (HRT) for protection against the development of cardiovascular disease and memory loss in menopausal women. The results of the WHI Memory Study suggest that HRT increases the risk of stroke and dementia in menopausal women. This finding has called into question the results of hundreds of basic science studies that have suggested that estrogen could protect brain cells from damage and improve cognition. A number of researchers have argued that inappropriate formulation, improper dosing, a limited study population, and poor timing of administration likely contributed to the reported findings from the clinical trial. Regarding appropriate formulation, it has been suggested that interactions between estrogen and other hormones should be considered for further investigation. A review of the literature has led us to conclude that a thorough investigation into such hormonal interactions is warranted. We hypothesize that the increased risk of cerebrovascular disease observed in menopausal women may, in part, be due to changes in the circulating levels of melatonin and estrogen and their modulatory affects on many relevant endothelial cell biological activities, such as regulation of vascular tone, adhesion to leukocytes, and angiogenesis, among others. Our hypothesis is supported by numerous studies demonstrating the reciprocal inhibitory effects of melatonin and estrogen on vascular tone, neuroprotection, and receptor expression. We believe that a thorough analysis of the distribution, localization, expression, quantification, and characterization of hormonal receptor subtypes, as well as changes in structural morphology in diseased and normal, healthy cerebrovascular tissue, will substantially aid in our understanding of the effects of HRT on the cerebrovascular circulation. The application of new molecular biological techniques such as tissue microarray analysis, gene and protein arrays, and multi-photon confocal microscopy may be of tremendous benefit in this regard.     

Concurrent administration of sustained-release bezafibrate may counteract the increased thrombotic risk associated with oral estrogen therapy.

Although hormone replacement therapy (HRT) can have many favorable effects on serum lipids and on vascular endothelium that presumably mediate the decreased risk for heart attack and stroke associated with HRT in observational epidemiology, oral estrogen also has various pro-coagulant effects: increases in serum triglycerides and factor VII activity, decreases in serum antithrombin III and protein S. This may explain the increased risk for venous thromboembolism observed with HRT and oral contraceptives, as well as the temporary increase in coronary risk noted when women with preexisting coronary disease initiate HRT. The well-tolerated hypolipidemic agent bezafibrate has anticoagulant actions that are diametrically opposed to the procoagulant effects of oral estrogen: namely, reductions in serum triglycerides and factor VII activity, and an increase in antithrombin III. However, bezafibrate could be expected to complement the protective effects of oral estrogen on serum lipids and on serum IGF-I activity. Thus, there is reason to believe that concurrent bezafibrate administration would minimize any thrombotic risk associated with HRT or oral contraception, while amplifying the health benefits of oral estrogen, and would make it more feasible to administer these therapies in women at increased vascular risk. These predictions require confirmation in controlled clinical studies. Certain natural hypolipidemic agents may also have potential as adjuvants to oral estrogen, but their effects on hemostasis require further investigation.     

HRT and gynaecologic cancer after WHI: old stuff or new doubts?

Previous studies had shown that there is little concern about endometrial cancer risk with hormone replacement therapy (HRT), whereas the risk of ovarian cancer is still debated. A new paper based on the women's health initiative (WHI) study devoted to gynaecologic cancers has been published in the October 2003 issue of JAMA, leading to the conclusion that also for endometrial and ovarian cancer risk, as the first WHI publication did for breast cancer risk, the previous large studies have been confirmed. About follow-up the authors conclude that, "The increased burden of endometrial biopsies required to assess vaginal bleeding further limits the acceptability of this regimen." These conclusions deserve a thorough analysis and must be read in the light of the burden of available epidemiological and clinical data and of recommended clinical practice. In the WHI trial the use of continuous combined HRT has resulted in a considerable increase of both imaging and invasive exams, to investigate women with bleeding or spotting, but the findings of these exams have been invariably negative. As far as, all published studies agree on the fact that continuous combined HRT either has no effect or significantly reduces the risk of both endometrial hyperplasia and cancer, there should be no reason to close-watching endometrium of women using this regimen with a different strategy from that employed on any healthy woman.     

The incidence of breast cancer and changes in the use of hormone replacement therapy: A review of the evidence

Even though a link between hormone replacement therapy (HRT) and breast cancer has been well documented in the epidemiological literature since the 1980s, it was not until publication of the results of the Women's Health Initiative (WHI) study in 2002 and the Million Women Study in 2003 that women and doctors started reconsidering the use of HRT and sales of HRT started to drop. This paper evaluates the impact of the publication of these two landmark studies on the expected and observed changes in the incidence of breast cancer.Between 2001–2002 and 2005–2006, sharp and significant reductions in the incidence of breast cancer of up to 22% were reported in many US and European populations, temporally consistent with the drop in usage of HRT. Declines in the rates of breast cancer were strongest for 50–60-year-old women (those most likely to be current users of HRT), affected mainly ER+ and PR+ cancers (those most strongly associated with HRT use), and were largest among women with the highest pre-decline prevalence of HRT use and the sharpest decline in its use.A considerable amount of scientific evidence supports the hypothesis that the decline in the incidence of breast cancer is in large part attributable to the sudden drop in HRT use following publication of the WHI and Million Women studies. Nevertheless, the problem of how to advise women contemplating HRT use today remains. Medical relief will remain necessary for many women with menopausal complaints, and so new therapeutic options need to be explored.     

Lessons to be learned from animal studies on hormones and the breast

The relation of hormone use by postmenopausal women to breast cancer risk has been controversial and unclear. A recent large randomized trial, the Women's Health Initiative (WHI) and a large observational study (Million Women Study) provided somewhat conflicting answers. The WHI found an increased incidence of breast cancer among women given hormone therapy (conjugated equine estrogen plus medroxyprogesterone acetate) but no increase in those given estrogen only therapy (conjugated equine estrogen alone). Whereas, the Million Women Study found an increased breast cancer risk among the estrogen plus progestin and the estrogen only users. This review brings comparative perspective to the issue of the effects of estrogen plus progestin versus estrogen only effects on breast cancer and is focused particularly on nonhuman primates. Although data from rodents is mixed, studies of monkeys suggest that estrogen only treatment has little or no effect on breast cell proliferation, and by inference, on breast cancer risk. On the other hand, data from both mouse and monkey studies strongly support the conclusion that the co-administration of a progestogen with an estrogen markedly increases breast cell proliferation and the potential for breast cancer promotion.     

HRT and the primary prevention of cardiovascular disease

Observational studies have consistently shown a benefit of hormone replacement therapy (HRT) on coronary heart disease (CHD), but some randomised studies have not shown any significant effect. Thus questions still remains as to whether HRT is beneficial for CHD, and in whom this benefit might be achieved. The biological effects of oestrogen on the cardiovascular system have been extensively studied, and beneficial effects on metabolic CHD risk factors, as well as on arterial function and on surrogate clinical markers of CHD, have been demonstrated. Thus it seems implausible that HRT should not benefit CHD in postmenopausal women. Most randomised trials using clinical outcomes have studied just one dose of one HRT regimen, a dose inappropriately high with the average starting age of the participants being in their mid-60s. The observational population studies largely comprise women starting on HRT at appropriate dose around the age of menopause, i.e. early 50s. In fact, it was the older women in the randomised trials that failed to show benefit, whereas there was a trend to benefit in the younger ones for whom the starting dose of HRT was appropriate. Furthermore, a pilot study of lower dose HRT in older women did not show any cardiovascular harm. Inappropriately high doses of oestrogen could cause cardiovascular harm due to transient disturbances in thrombogenesis and vascular remodelling. Whilst the greatest CHD benefit may be seen by starting HRT in the early postmenopause, this does not exclude benefit in older women given appropriate low dose therapy.     

Selection bias in semen studies due to self-selection of volunteers

BACKGROUND: Reports of a secular decrease in semen quality remain controversial, particularly due to the possibility of selection bias. We aimed to describe the potential bias due to self-selection of volunteers in semen studies involving fecund men. METHODS: Using data from the French multicentre study REPRHOM, we compared the characteristics of the partners of pregnant women for three levels of participation: completion of a refusal questionnaire (n = 698), agreement to complete the study questionnaires only (n = 676) and agreement to complete the study questionnaires and give a semen sample (n = 331, 13% of the subjects approached). RESULTS: Poorly educated men refused more often to participate than highly educated men. Semen providers were more likely to have experienced unfavourable pregnancy outcomes (odds ratio 1.68, 95% confidence interval 1.14-2.49) compared with participants completing the questionnaires only. Time to pregnancy was similar for all participants. CONCLUSIONS: This study demonstrates the existence of selection bias in semen studies associated with fertility and socio-demographic characteristics of men. The results of semen analysis for this population sample cannot be extrapolated to the whole population from which the volunteers originate. More information is required on who participates, and participation rates should be reported in semen studies to make it possible to interpret the results correctly.     

Breast cancer risk in the WHI study: the problem of obesity

In the climacteric, about 40% of the women have occult breast tumors the growth of which may be stimulated by hormones. Many genetic, reproductive and lifestyle factors may influence the incidence of breast cancer. Epidemiological data suggest that the increase in the relative risk (RR) of breast cancer induced by hormone replacement therapy (HRT) is comparable with that associated with early menarche, late menopause, late first birth, alcohol consumption, etc. One of the most important risk factors is obesity which exceeds the effect of HRT by far, and in overweight postmenopausal women the elevated risk of breast cancer is not further increased by HRT. As in the WHI study the majority of women was overweight or obese, this trial was unsuitable for the investigation of breast cancer risk. In the women treated with an estrogen/progestin combination, the RR of breast cancer rose only in those women who have been treated with hormones prior to the study, suggesting a selection bias. In the women not pretreated with hormones, it was not elevated. In the estrogen-only arm of the WHI study, there was no increase but a steady decrease in the RR of breast cancer during 6.8 years of estrogen therapy. This result was unexpected, as estrogens are known to facilitate the development and growth of breast tumors, and the effect is enhanced by the addition of progestins. Obese women are at high risk to develop a metabolic syndrome including insulin resistance and hyperinsulinemia. In postmenopausal women, elevated insulin levels are not only associated with an increased risk for cardiovascular disease, but also for breast cancer. This might explain the effects observed in both arms of the WHI study: HRT with relative low doses of estrogens may improve insulin resistance and, hence, reduce the elevated breast cancer risk in obese patients, whereas this beneficial estrogen effect may be antagonized by progestins. The principal options for the reduction of breast cancer risk in postmenopausal women are the prevention of overweight and obesity to avoid the development of hyperinsulinemia, the medical treatment of insulin resistance, the use of low doses of estrogens and the reduction of exposure to progestins. The latter might include long-cycles with the sequential use of appropriate progestins every 3 months for 14 days. There are large inter-individual variations in the proliferative response to estrogens of the endometrium. Control by vaginalsonography and progestin challenge tests may help to identify those women who may be candidates for low-dose estrogen-only therapy.     

Menopause in crisis post-Women's Health Initiative? A view based on personal clinical experience

Menopausal women should not consider that hormonal treatment is an obligatory long-term commitment. Estrogen-based treatments are extremely effective for vasomotor symptom relief and for vaginal atrophy. HRT also is one of several effective methods for the primary prevention of osteoporosis. If trials were done early after the menopause when the endothelium is likely still to be intact, estrogen-based treatment might be shown to prevent coronary heart disease. However, greater efficacy is to be expected from smoking cessation, proper nutrition, exercise, moderate alcohol consumption, statins, beta-blockers and angiotensin-converting enzyme inhibitors. The treatment options for a menopausal woman should include non-drug-related strategies, non-hormonal pharmaceutical therapies as well as hormonal treatments. The first objective of this contribution is to call to the attention of practising physicians the fact that the Women's Health Initiative (WHI) and Heart and Estrogen/Progestin Replacement Study (HERS) studies involved women much older than the early postmenopausal age groups for whom HRT is prescribed because of symptoms. The second objective is to emphasize that the attending physicians must not only treat the symptomatic women but also prevent the occurrence of diseases more prevalent after 60 years of age. Hormones can safely be used for the former, when not contraindicated, whereas for the latter non-pharmacological interventions and non-hormonal medications are preferable.     

Effects of the estrogen replacement therapy on different biochemical markers of endothelial reactivity in recent postmenopausal healthy women

The risk of cardiovascular disease significantly increases after menopause. Accordingly, many evidences suggest that estrogens may positively affect the production of different vasoactive factors, such as nitric oxide, prostacyclin, endothelin-1 and catecholamines and induce favourable changes in lipid profile. However, although observational data indicate that hormone replacement therapy (HRT) reduces significantly the cardiovascular risk, recent results have added controversial data to the institution of HRT in postmenopausal women. These last studies present numerous bias, related to inclusion of a single combination HRT regimen, recruitment women of older age groups who began treatment years after menopause and generally with pre-existing coronary artery disease. In fact, it is known that aging and atherosclerotic injury may induce estrogen receptors depletion, endothelial dysfunction and thrombosis, thus potentially decreasing HRT efficacy. Therefore, particular attention must be paid to the age of the woman and the complexity of atheroscleriotic lesions as determinants of the response to HRT for each patient. Moreover, the maintenance of an healthy and normal functioning endothelium after menopause emerges as a major target to retain maximal cardiovascular benefit from this treatment.     

Variability of breast cancer risk in observational studies of hormone replacement therapy: a meta-regression analysis

OBJECTIVES: A re-analysis of data from 51 epidemiological studies reported a significant 14% increase in the risk of breast cancer associated with the use of hormone replacement therapy (HRT). Unlike randomized trials, these observational studies varied in design and methods. This study was conducted to explore the impact of study design factors on the rate ratio. METHODS: We performed a meta-regression analysis of 39 epidemiological studies of HRT and breast cancer. The rate ratio of breast cancer associated with ever use of HRT was evaluated in relation to study design, study period, country, primary study objective, method of exposure measurement, age control, adjustment factors related to reproduction and menopause, and the presence of breast cancer surveillance. We used stepwise multiple regression analysis, weighted by the inverse of the variance of the logarithm of the rate ratio, to estimate ratios of rate ratios for these factors. RESULTS: Exposure measured by personal interview and/or medical record review was associated with a 14% lower rate ratio estimate as compared with telephone interview or self-administered questionnaire (P = 0.018). Among studies that did not adjust for age at menopause, the rate ratio was 12% lower if the primary objective was HRT effect than not (P = 0.016), while it was 43% higher among studies that adjusted for age at menopause (P = 0.042). An index that included as six desirable design features, breast cancer surveillance, matching of controls, more reliable exposure information, adjustment for age at menopause and reproductive risk factors, and as primary objective the effect of HRT suggests that studies with none of these properties would lead to a rate ratio estimate of 1.14 (95% CI: 1.00-1.29) while studies with all six properties would produce a rate ratio of 0.98 (95% CI: 0.83-1.15). CONCLUSIONS: Design factors of epidemiological studies could be an alternative explanation for the reported 14% increase in the risk of breast cancer associated with the use of HRT.     

Myocardial infarction risk and hormone replacement: differences between products

OBJECTIVES: To establish the risk of myocardial infarction (MI) in users of hormone replacement therapy (HRT) compared with non-users and to compare the risk between different HRT regimens. METHODS: A population-based cohort and case-control study, and a case-control study nested within a cohort of HRT users, using the UK General Practice Research Database. Differences between HRT regimen, mode of administration and duration and recency of use were examined whilst adjusting for confounding. RESULTS: In the cohort and case-control study, 4537 cases of MI were identified in 2.62 million observed women years, cases were age-matched to 27,220 controls. In both studies, current and past HRT use were associated with reduced risk estimates for MI compared with no prior use. MIs were less likely to be fatal amongst women who had used HRT than amongst never users (OR(adj) 0.58; 95% CI 0.45-0.75). No difference in risk was seen between current and past use, oral and transdermal HRT or between different regimens (p>0.44). In the nested study, no difference was found in the association with MI risk between different oestrogen-progestogen combinations or between different combinations and tibolone. Unopposed oestrogen use was not associated with a decrease in risk compared with combined HRT. CONCLUSIONS: These results are consistent with previous observational studies in supporting the hypothesis that use of postmenopausal HRT is associated with a decrease in risk of acute myocardial infarction (AMI). Case fatality differed between HRT users and non-users, suggesting a protective effect of HRT. This study does not demonstrate a difference between regimens.     

Patient selection factors in angiographic studies: A conceptual formulation and empirical test

Studies using coronary angiography patients as the study population to assess etiologic relationships are vulnerable to bias produced by patient selection factors, but the direction and magnitude of the bias in any particular study are uncertain. Partly for this reason, some authors have questioned the use of angiography data for studies of etiology. Nevertheless, this type of data continues to be used in etiologic studies. Therefore, in this paper a conceptual framework is described to quantify the degree of bias under different assumptions about the process whereby patients are selected for angiography. The direction and degree of bias depend on the pattern of patient selection, as well as the measure of association used (e.g., the risk ratio or odds ratio). One method is then described, based on assessing how physicians refer patients to angiography, which estimates the degree of bias produced by selection factors. This study is limited to the situation of a two-level risk factor (e.g., type A/B behavior) in relation to a two-level outcome measure (e.g., significant coronary artery occlusion versus no significant coronary artery occlusion). While type A/B behavior is used as an example in this paper, the concepts apply to any behavioral or nonbehavioral predictor variable.This work was funded in part by The National Institutes of Health, U.S. Department of Health and Human Services under Grants HL26235 and T32 HL07365.