Editorials Editorials represent the opinions of the authors and T HE J OURNAL and not those of the American Medical Association. Of Designer Drugs, Magic Bullets,
and Gold Standards Evaluating the risks and benefits of estrogen and related
drugs after menopause is complex given their diverse prop-
erties. Major considerations include the effects of such agents
on cardiovascular and bone disease; breast and endometrial
cancer; the plasma lipoproteins and the blood coagulation
system; and hot flashes, vaginal bleeding, breast tenderness,
and urogenital function. Some effects may be beneficial, others
may be adverse. From these multifaceted and often disparate
properties attempts are made to construct a balance sheet to
better help select from the available agents those that provide
the most benefit and the least harm. 1,2 This approach is impor- tant in considering the appropriateness of the therapy either
for the individual patient or from the standpoint of public
health. See also p 1445. Unfortunately, no robust analysis of the benefits and risks can be made at the present time. There is a dearth of the only
kind of information that is entirely convincing, ie, that from
randomized, controlled, clinical trials with clinical outcomes.
Because these gold standard studies have not yet been done,
practice decisions have perforce been based on the vicissi-
tudes of lesser kinds of studies, resulting in an ebb and flow of
enthusiasm for postmenopausal hormone use. Currently, a
wave of enthusiasm for long-term hormone use has been gen-
erated by the undoubted benefit for preventing osteoporosis,
and by a commonly assumed but unproven benefit for pre-
venting cardiovascular disease. Based on a 1995 National
Heart, Lung, and Blood Institute survey, more than half of US
practitioners would prescribe estrogen for prevention of coro-
nary heart disease (CHD) or treatment of high blood choles-
terol (Beth Schucker, MA, oral communication, April 1998). Cardiovascular disease, including CHD, stroke, and venous thromboembolism, is the major cause of morbidity and mor-
tality in postmenopausal women, accounting for 473 569 deaths
in US women older than 55 years in 1996. 3 Therefore, the potential effects of estrogen on rates of these cardiovascular
disorders are of enormous public health importance. Numer- ous observational epidemiological studies have suggested that
use of estrogen, with or without a progestin, is associated with
a 40% to 50% lower risk of fatal and nonfatal CHD compared
with that among nonusers. 1 Stroke rates appear similar in users and nonusers of estrogen, but rates of venous thrombo-
embolism are clearly increased in estrogen users. 4 Because CHD is by far the most common outcome, estrogen effects on
CHD dominate most possible scenarios. 2 From a quantitative standpoint, a real reduction in CHD of almost half would over-
whelm any anticipated increase in risk for breast cancer. For
instance, in 1994 there were 229 628 deaths from CHD, com-
pared with 33 532 deaths from breast cancer, in women older
than 55 years. 3 Nonetheless, substantial numbers of women may be exposed to an increased risk of breast cancer. A recent
compilation of the international data concluded that estrogen
use for 5 years or more may be associated with a 35% increased
risk for breast cancer. 5 Notwithstanding the consistent, large, and statistically sig- nificant lower risk for CHD in estrogen users, these observa-
tional studies do not and cannot establish causality. Users
differ from nonusers in important respects; eg, they are thin-
ner, better educated, and more health-conscious. Substantial
differences in their CHD risk factor profiles before commenc-
ing estrogen use could readily explain most of the subsequent
differences in CHD rates. Women who use estrogen replace-
ment therapy for a number of years are by definition good
compliers, and as such may have other attributes that predict
better health. In addition, these women are or will be under
closer medical supervision and, thus, have earlier diagnosis
and treatment of health conditions, leading to a lower mortal-
ity than that among nonusers. On the other hand, women who
cease estrogen replacement therapy frequently do so because
of illness, which makes the remaining users appear even
healthier. 6 These various biases are quite strong, are impos- sible to correct for fully in observational studies, and may
account for most or even all of the apparent benefit for CHD
in observational studies. 7 Well-designed clinical trials, which are free of these biases, are needed to establish a cardioprotective effect. The available
information is scanty; in men large doses of estrogen were not
cardioprotective, 8 and a collection of small, short-term trials in women also did not support such an effect. 9 Until recently the large numbers and lengthy follow-up period that were needed
precluded adequate studies in women. However, the pressing From the National Heart, Lung, and Blood Institute, Bethesda, Md.
Reprints: Basil M. Rifkind, MD, Two Rockledge Centre, 8701 Rockledge Dr, Suite 10193, Bethesda, MD 20892. JAMA, May 13, 1998Vol 279, No. 18 Editorials 1483 nature of these questions has led to the launch of 3 such studies
in recent years. The first, the Heart and Estrogen/Progestin
Replacement Study (HERS), is a secondary prevention trial
of 2763 women with documented CHD and is scheduled to
report its results in 1998. 4 The Womens Health Initiative (WHI), which includes primary prevention trials of estrogen
with and without progestin, has completed its recruitment for
27 500 women and will have results after 2005. 10 A similar primary prevention trial in the United Kingdom, the Womens
International Study of Long-Duration Oestrogen After the
Menopause (WISDOM), is scheduled to start recruitment this
summer (Madge Vickers, PhD, oral communication, March
1998). An optimum hormone replacement therapy would have tis- sue selectivity to retain the apparent benefits of estrogens
while being free of their adverse features. This magic bullet
would reduce CHD as much as it is hoped conventional estro-
gen will do, and would offer a similar degree of protection
against osteoporosis. 11 It should not produce undesirable changes in breast or endometrial tissue. For favorable effects
on CHD, such an optimum therapy would be expected to pro-
duce corresponding reductions in atherogenic low-density
lipoprotein cholesterol (LDL-C) and increases in high-density
lipoprotein cholesterol (HDL-C) and favorably affect the co-
agulation process. 12 The level of HDL-C is a strong inverse predictor of CHD risk in women, and it is estimated that up to
half of the cardiovascular benefit could be mediated by the
higher levels induced by treatment with estrogen. 13 Drugs collectively known as selective estrogen receptor modulators are designed to have tissue selectivity and are now
being introduced. The prototypical such designer drug is the
antiestrogen tamoxifen, which has been used for the treat-
ment of breast cancer for 2 decades 14 and has recently been shown to prevent breast cancer in high-risk women as re-
ported in preliminary results from the large Breast Cancer
Prevention Trial. 15 While having antiestrogenic effects in sup- pressing breast cancer growth and in increasing hot flashes,
tamoxifen retains estrogenic stimulatory effects on the endo-
metrium, maintains bone density, and decreases LDL-C.
Recently a second selective estrogen receptor modulator,
raloxifene, was introduced. 11 Compared with placebo, raloxi- fene increases bone mineral density but is only half as effective
as estrogen in this regard. Animal studies and limited studies
in humans indicate that raloxifene does not stimulate the en-
dometrium and may not stimulate breast tissue. The Food and
Drug Administration has approved the labeling of raloxifene
for postmenopausal osteoporosis. In this issue of JAMA, Walsh and colleagues 16 report the effects of raloxifene on lipids and coagulation factors com-
pared with the effects of estrogen plus progestin. The study
did not include comparison with estrogen alone, which may
have quantitatively different effects on both lipids and coagu-
lation factors. Whereas both tested regimens produced simi-
lar reductions in LDL-C, the effects on other blood lipids were
likely to be less cardioprotective. Raloxifene did decrease
lipoprotein(a) (an atherogenic plasma lipoprotein closely as-
sociated with LDL-C but that has levels that are generally
resistant to change) and increased high-density lipoprotein-2
cholesterol. The changes in lipoprotein(a) and HDL-C were at
best only half as much as those observed with estrogen plus
progestin. Of special concern was the absence of an increase in
HDL-C levels by the approximate 10% that occurred with estrogen plus progestin; similar findings have been reported
elsewhere. 17 Another concern is the inability to interpret the data on coagulation factors. For instance, of what relevance is
a reduction in fibrinogen if the actual incidence of venous
thromboembolism is increased in studies of raloxifene? More
globally, of what relevance are comparisons with estrogen
when it is uncertain whether estrogen is cardioprotective? The tamoxifen experience may be relevant to raloxifene. As the authors acknowledge, tamoxifen induces changes in lipids
and coagulation factors similar to those of raloxifene, yet only
1 of 3 smaller breast cancer treatment trials has shown a sig-
nificant reduction in CHD event rates. 18 The Breast Cancer Prevention Trial has reported no effect of tamoxifen on CHD
event rates, but showed a reduction in its primary outcome,
breast cancer; other results were a reduction in fractures and
increases in endometrial cancer, pulmonary embolism, and
deep vein thrombosis. Although none of these trials was de-
signed to find answers for cardiovascular disease, and the num-
bers of CHD events are small, the tamoxifen findings to date
do not indicate that the observed lipid changes will necessarily
translate into clinical benefit. Other questions about the future clinical relevance of raloxi- fene remain. For example, what are the effects of raloxifene on
the brain? Estrogen affects the brain, as evidenced by reduction
in hot flashes, improvements in cognitive function, and possibly
a reduced risk for Alzheimer disease. 19 Does the fact that hot flashes were increased in the raloxifene group imply that a ben-
efit for cognitive function or dementia should not be expected,
in contrast with the possible benefit from estrogen? The one
distinct advantage of raloxifene is that, in contrast with estro-
gen (and with tamoxifen), it does not stimulate the endome-
trium. It may also be that in contrast with estrogen, raloxifene
does not stimulate the breast; however, long-term clinical out-
come data are needed to clarify these points. The uncertainties about the cardioprotective effect of es- trogens apply doubly to raloxifene. Even if estrogens were
proven to be cardioprotective, a rush to replace estrogen with
raloxifene for this indication still may be inappropriate. Leav-
ing the endometrium unstimulated may benefit only a small
number of women (and is of no benefit to those who have had
a hysterectomy) compared with the potentially much larger
numbers of women who may be put at risk if raloxifene offers
a lesser degree of cardioprotection. Should the Womens
Health Initiative and other trials show estrogen to be cardio-
protective, the next step would be to assess whether the lesser
lipid effects of raloxifene translate into less cardioprotection. 20 Meanwhile, raloxifene may be viewed as a designer drug, not
yet suitable for everyday wear. Or, ultimately, it may prove to
be a magic bullet in search of its target. Basil M. Rifkind, MD
Jacques E. Rossouw, MD 1. Grady D, Rubin SM, Petitti DB, et al. Hormone therapy to prevent disease and
prolong life in postmenopausal women. Ann Intern Med. 1992;117:1016-1037. 2. Col NF, Eckmann MH, Karas RH, et al. Patient-specific decisions about hormone
replacement therapy in postmenopausal women. JAMA. 1997;277:1140-1147. 3. National Center for Health Statistics. Vital Statistics of the United States, 1994. Vol
2, pt A. Washington, DC: Public Health Service; 1997. 4. Grady D, Hulley SB, Furberg C. Venous thromboembolic events associated with
hormone replacement therapy. JAMA. 1997;278:477. 5. Collaborative Group on Hormonal Factors in Breast Cancer. Breast cancer and
hormone replacement therapy: collaborative reanalysis of data from 51 epidemiologi-
cal studies of 52,705 women with breast cancer and 108,411 without breast cancer.
Lancet. 1997;350:1047-1059. 6. Sotelo MM, Johnson SR. The effects of hormone replacement therapy on coronary
heart disease. Endocrinol Metab Clin North Am. 1997;26:313-328. 7. Postuma WFM, Westendorp RG, Vanderbroucke JP. Cardioprotective effect of 1484 JAMA, May 13, 1998Vol 279, No. 18 Editorials hormone replacement therapy in postmenopausal women: is the evidence biased?
BMJ. 1994;308:1268-1269. 8. The Coronary Drug Project Research Group. The Coronary Drug Project: findings
leading to discontinuation of the 2.5 mg/day estrogen group. JAMA. 1973;226:652-657. 9. Hemminki E, McPherson K. Impact of postmenopausal hormone therapy on cardio-
vascular events and cancer: pooled data from clinical trials. BMJ. 1997;315:149-153. 10. The Womens Health Initiative Study Group. Design of the Womens Health Ini-
tiative Clinical Trial and Observational Study. Control Clin Trials. 1998;19:61-109. 11. Gustafsson J. Raloxifene: magic bullet for heart and bone? Nat Med. 1998;4:152-
153. 12. The Writing Group for the PEPI Trial. Effects of estrogen or estrogen/progestin
regimens on heart disease risk factors in postmenopausal women. JAMA. 1995;273:
199-208. 13. Bush TL, Barrett-Connor E, Cowan LD, et al. Cardiovascular mortality and non-
contraceptive use of estrogen in women: results from the Lipid Research Clinics Pro-
gram Follow-up Study. Circulation. 1997;75:1102-1109. 14. Jordan VC. Tamoxifen: the herald of a new era of therapeutics. J Natl Cancer Inst.
1997;89:747-749. 15. Breast Cancer Prevention Trial shows major benefit, some risk [press release].
Bethesda, Md: National Cancer Institute; April 6, 1998. 16. Walsh BW, Kuller LH, Wild RA, et al. Effects of raloxifene on serum lipids and co-
agulation factors in healthy postmenopausal women. JAMA. 1998;279:1445-1451. 17. Delmas PD, Bjarnason NH, Mitlak BH, et al. Effects of raloxifene on bone mineral
density, serum cholesterol concentrations and uterine endometrium in postmeno-
pausal women. N Engl J Med. 1997;337:1641-1647. 18. Constantino JP, Kuller LH, Ives DG, Fisher B, Dignam J. Coronary heart disease
mortality and adjuvant tamoxifen therapy. J Natl Cancer Inst. 1997;89:776-782. 19. Yaffe K, Sawaya G, Lieberburg I, et al. Estrogen therapy in postmenopausal
women: effect on cognitive function and dementia. JAMA. 1998;279:688-695. 20. Clarkson TB, Anthony MS, Jerome CP. Lack of effect of raloxifene on coronary
artery atherosclerosis of postmenopausal monkeys. J Clin Endocrinol Metab. 1998;
83:721-726. Health Care at Mass Gatherings As the number of spectators attending the Olympic Games has
steadily increased, the scope of Olympic medicine has ex-
panded beyond just caring for athletes. In this issue of JAMA,
2 articles focus on the measures taken to ensure the health of
the estimated 2.2 million athletes, staff, and spectators who
attended the 1996 Centennial Olympic Games in Atlanta, Ga.
The article by Meehan et al 1 describes the planning and imple- mentation of a public health surveillance and response system.
The article by Wetterhall et al 2 examines the 10 715 physician- patient encounters recorded during the 17 days of the Olympic
Games. These articles provide valuable information for plan-
ning future Olympic Games, and present useful lessons that
can be applied to medical care for other mass gatherings and
special events. See also pp 1463 and 1469. Application of a scientific approach to the planning and analysis of medical care for special events is a relatively new
occurrence. Until recently, little more than individual subjec-
tive reports have appeared in the medical literature, reflect-
ing a haphazard approach. 3 However, increasing attention is being focused on the unique issues involved in providing care
during large gatherings. Although the phrase mass gathering
medical care frequently is used to describe these activities,
this phrase also can apply to the care of temporarily displaced
populations, including disaster victims and refugees, whose
care requires distinct resources and a vastly different ap-
proach. To avoid confusion, mass gathering medical care
should have more inclusive usage, whereas special event medi-
cal care should be reserved as a more specific phrase, charac-
terized by Baker et al 4 as the provision of preventive, or definitive primary medical care or hospital referral to well
persons attending or participating in major sports, recre-
ational, or political events. As this ambiguous terminology illustrates, the lack of stan- dardized definitions, indicators, and methods represents the
greatest barrier to advancement of this field. Among 3 pub- lished reports from recent Olympic Games held in North
America, meaningful comparison or trend analysis is difficult
because of differing definitions of medical facility usage rates
and approaches to analysis. 1,4,5 Wetterhall and colleagues 2 are the first to use accreditation status (eg, athletes, staff, spec-
tators) to evaluate medical usage patterns, but do not make
use of patient acuity measures. Such a measure was developed
and recommended following the 1988 Calgary Winter Games. 4 This lack of longitudinal standardization in analysis of Olympic
medical care may be an inevitable result of the decentralized
nature of Olympic planning, but it precludes a better under-
standing of future special event medical care needs. Even with a more standardized approach, there remains considerable variation in the nature of these events and in the
populations attending them. Patterns of system use are af-
fected by many variables, often in ways that are not intuitive.
For example, a complex relationship exists between crowd
size and medical service use, with variables such as event
location, duration, and logistics frequently overshadowing the
effect of population size. Spectators attending longer-duration
outdoor events, especially those that allow spectator ambula-
tion (ie, canoeing and kayaking) may be considerably more
likely to require medical care compared with persons attend-
ing much larger, seated stadium events (eg, basketball). En-
vironmental factors, 6 population demographics, and the base- line health status of participants also may have a dramatic
effect on usage patterns of medical care. 7 A better understand- ing of the effects and interaction of these variables is needed
to develop models capable of predicting future resource needs. Meehan and colleagues 1 describe a systematic, coordinated process in planning Atlantas public health response. Several
points deserve emphasis. Public health officials must be in-
volved from the time of initial proposal development. Ad-
equate funding for public health components must be allo-
cated. Prediction of resource requirements should be based on
a comprehensive evaluation of anticipated medical care usage
rates and assessment of public health risks. This analysis
should then guide development of specific prevention strate-
gies, monitoring systems, and contingency planning. The value
of legislation and policy modification to reduce identifiable
public health risks has been reported in previous Olympics 8 and is now underscored. Meehan et al 1 also demonstrate the need for health officials to remain aggressive and for vocal
advocates to guard against compromises to the health of the
public from potential political and commercial conflicts (such From the Department of Emergency Medicine, The Johns Hopkins University School of Medicine (Dr Green), and the Department of International Health, The Johns
Hopkins University School of Hygiene and Public Health (Dr Burnham), Baltimore, Md. Reprints: Gary B. Green, MD, MPH, Department of Emergency Medicine, Johns Hopkins University School of Medicine, 600 N Wolfe St, Marburg B-186, Baltimore, MD
21287-2080. JAMA, May 13, 1998Vol 279, No. 18 Editorials 1485