The Journal of Clinical Endocrinology & Metabolism Vol. 84, No. 6 1799-1802
Copyright © 1999 by The Endocrine Society
Strategies for Individualizing Patient Decisions About Hormone Therapy
Nananda F. Col,
Mark H. Eckman and
John B. Wong
and Stephen G. Pauker
Division of Clinical Decision Making, Informatics, and Telemedicine
Department of Medicine
New England Medical Center and Tufts University School
of Medicine
Boston, Massachusetts 02111
 |
Introduction
|
|---|
F UELED by hotly debated, highly publicized, and
often politicized new studies, the controversy over which post- and
perimenopausal women, if any, should receive hormone replacement
therapy (HRT) rages on. Proponents of HRT emphasize its
cardioprotective and bone preserving effects. They argue that the
benefits outweigh the risks because many more women die from coronary
heart disease (CHD) or hip fracture than from breast cancer. Opponents
of HRT highlight the increased risks of breast cancer associated with
long-term use while casting doubt upon the strength of evidence that
links HRT to cardioprotection. But both opinions derive predominantly
from the same epidemiological studies and should therefore be subjected
to the same degree of skepticism. Of course, both positions ignore the
individual risk factors that distinguish an individual woman from the
average.
Other controversies about HRT include its effectiveness in averting or
relieving menopausal symptoms and its tendency to cause undesirable
side-effects. Emerging evidence on the beneficial effects of HRT on
memory and in averting Alzheimer’s disease are often uncritically
accepted although these studies are particularly prone to selection
bias and confounding. Because these latter issues are based on less
well established data or simply on opinion, we have excluded them from
our analyses.
|
Analytic approach
|
|---|
To provide patient-specific information on the risks and benefits
of HRT, we synthesized the best currently available evidence, using a
single scale of life expectancy to measure the outcomes of treatment
according to an individual’s risk profile for CHD, breast cancer, and
osteoporosis.
We developed a decision analytic Markov state transition model (1) that
simulates the natural history of hypothetical cohorts of postmenopausal
women, one receiving HRT and the other not (2). We applied this model
to pairs of identical cohorts with differing risk factors for CHD,
breast cancer, and hip fracture. We used published regression models to
link these risk factors to disease incidence, drawing upon results from
the Framingham Heart Study and Framingham Offspring Cohorts (3, 4),
the Breast Cancer Detection Demonstration Project (5), and the Study of
Osteoporotic Fractures Research Group (6). An individual’s level of
risk for each disease was based on the composite contribution of
multiple risk factors. We drew our estimates on the impact of HRT in
primary prevention from studies that examined current users of HRT and
that incorporated duration of treatment. We assumed that HRT decreases
the risk of hip fracture by 54% (with peak impact observed after 10
years of treatment) (7), decreases the risk of CHD by 40% (8), and
increases the risk of breast cancer by 46% after 5 yr of therapy (9).
Although unopposed estrogen increases the risk for endometrial cancer
(10, 11), adding progestin to estrogen reduces this risk to that
observed in the general population (11, 12, 13). Therefore we did not model
an increased risk for endometrial cancer. Mortality rates after the
development of CHD, breast cancer, hip fracture, or endometrial cancer
were calculated from survival tables (14, 15). CHD mortality rates
were adjusted for the presence of selected prognostic factors, using
regression equations developed from the Framingham Heart Study
(16, 17). We assumed that HRT had no effect on the secondary
prevention of cardiac events among women with established CHD (18).
|
Results
|
|---|
Based on this decision model, we examined the effect of HRT on
life expectancy and disease incidence for women at differing levels of
risk for CHD, breast cancer, and hip fracture. Most 50-yr-old women are
predicted to accrue a survival benefit, and an average 50-yr-old woman
would gain about 6 months in life expectancy from HRT. There were
significant individual variations, ranging from gains of up to 41
months to losses of up to 8 months, depending on a woman’s level of
risk for CHD, breast cancer, and hip fracture. The only women not
expected to accrue any survival benefit from HRT are those at highest
risk for breast cancer (corresponding to those women with two or more
first degree relatives with breast cancer) who are also at
lowest risk for CHD (corresponding to having a favorable profile for
CHD with no risk factors). These findings demonstrate the importance of
assessing individual risk factors in estimating the impact of HRT.
While it is useful to estimate HRT’s impact on a woman’s survival,
there are other important outcome measures to consider as well. HRT
increases survival when its benefit in terms of preventing CHD or
osteoporosis outweighs its risk of inducing breast cancer. Recognizing
that women may view certain outcomes such as breast cancer as worse
than others such as CHD, we developed several simple approaches to help
women and their clinicians estimate the impact of HRT on an individual
woman’s survival as well as her chances of developing CHD, breast
cancer, and hip fracture (2, 19). These approaches link the presence
or absence of specific risk factors to the impact of HRT, using simple
tables and graphic displays (e.g. flow diagrams, bar charts,
and/or graphs). They do not require direct access to the computer model
but do require some basic clinical information, including blood
pressure, total and HDL cholesterol, family history (breast cancer and
fracture), and lifestyle factors (e.g. tobacco use, activity
level).
|
Strategic approaches to HRT
|
|---|
Some investigators have recently advocated that all newly
menopausal women delay the initiation of HRT for 10 yr when their
chances of developing heart disease or hip fracture are higher. This
targeted, "strategic" approach to HRT assumes that HRT provides
bone-preservation regardless of age of initiation, and that women are
unlikely to succumb to CHD (or hip fracture) during the interim years
when HRT is withheld. There is little evidence to support either of
these assumptions.
The first assumption rests largely upon a single recent study, which
found that women who started HRT later in life achieved a similar
degree of bone preservation as those who started at the time of
menopause (20). However, there were only 29 "late-initiators" in
this study, and the only endpoint measured was bone mineral density
rather than fracture rate. A much larger study that followed fracture
rates came to the opposite conclusion—"no effect [on fracture
reduction] was noted if estrogen was initiated late, even among
long-term users" (21). Annual loses in bone mass of 3–5% are not
uncommon in the years just after menopause (22), loses that can be
avoided by HRT.
The second assumption is based on the widely held misperception that
women are largely protected from heart disease until they reach very
advanced ages. For example, one physician states that "in women
younger than age 75, there are actually three times as many deaths from
breast cancer as there are from heart disease" (23). Although widely
promulgated, this conclusion is incorrect; the figures should be
reversed. At the age when most women approach the decision about
whether or not to use HRT (50 yr), the number of deaths from CHD
outweighs that from breast cancer. Furthermore, deaths from heart
disease outnumber deaths from breast cancer among women for each
subsequent decade (Table 1
).
While it is true that women at low risk for heart disease are unlikely
to develop premature heart disease, many perimenopausal American women
have one or more risk factors for heart disease, removing them from the
lowest risk group. One in 9 American women aged 45–64 has some form of
cardiovascular disease, and 20,800 women under age 65 die from
myocardial infarction each year (24). Because these data reflect the
combined experience of all American women, they do not necessarily
pertain to an individual, whose risk factor profile may differ
substantially from an average woman.
Even more importantly, the strategy of delaying HRT until a decade
after menopause uses age as the sole marker for risk of developing hip
fracture, CHD, and breast cancer, ignoring other risk factors that may
be even better predictors of these diseases. Although a woman’s
chances of developing CHD and osteoporosis increase as she ages, age is
but one of many predictors of risk. Compared to a 65-yr-old woman with
no CHD risk factors, a 50-yr-old woman with two risk factors for CHD
(e.g. a smoker with mild hypercholesterolemia) has a higher
risk for developing CHD over the next 10 years. Making treatment
recommendations for HRT based on age alone would result in
overtreatment of women at low risk for CHD and undertreatment of women
at high risk. If the goal is to target preventive therapies to
individuals who are expected to reap the greatest benefit by virtue of
being at the greatest risk for developing a disease, then
"strategic" use should not arbitrarily limit its focus to one of
many known risk factors, i.e. age, but should use a more
complete approach to risk stratification that includes other markers of
risk.
|
Data linking HRT to cardioprotection
|
|---|
Because most of the gains HRT provides in life expectancy result
from its cardioprotective effect, it is important to assess the
strength of the data upon which our findings are based. Although most
of the data suggesting a protective effect in the primary prevention of
cardiac events are drawn from nonrandomized, observational studies
(25), one small randomized trial found HRT to be associated with
reduction in CHD of 33% (although there were only four outcomes in
this study) (26). Numerous randomized trials examining intermediate
endpoints such as cholesterol (27), fibrinolysis (PAI-1) (28),
vasodilation (29), and platelet aggregation (30) support HRT having a
cardioprotective effect but do not help to determine the magnitude of
this effect on clinical events such as myocardial infarction.
The argument has been made that the lower incidence of heart disease
observed among users of HRT simply reflects selection bias: only
relatively healthy women are prescribed HRT. There is evidence that
women who elected to take HRT are, in most but not all respects,
healthier than women who do not take HRT (8). To minimize this
selection bias, we drew our estimates of the impact of HRT on CHD from
the Nurses’ Health Study (8), which carefully adjusted their risk
estimates for potential confounders, including age, body mass index,
diabetes, blood pressure, hypercholesterolemia, cigarette smoking,
parental history of myocardial infarction before the age of 60 yr, age
at menopause, and type of menopause (natural vs. surgical).
If the observed cardioprotection were substantially biased by
selection, then one would expect the observed cardioprotective effect
to largely dissipate after controlling for baseline differences between
users and nonusers of HRT. While there was some attenuation in this
risk estimate after adjustment, a 40% risk reduction remained (the
unadjusted estimate was 55%). Although complete adjustment for
selection bias is not possible, the magnitude of the risk reduction
remaining after adjustment for known risk factors seems unlikely to be
attributable to some unidentified confounder. Additionally, a similar
cardioprotective effect was observed in various subgroups defined by
the presence or absence of specific risk factors. For example, when
analysis was limited to low-risk women, the relative risk (RR) of HRT
on CHD was 0.67 among current HRT users compared with women who had
never used HRT. Similarly, women who smoked and used HRT had a RR of
CHD of 0.43 compared with smokers who did not use HRT. Subgroups
defined by hypertension, cholesterol, body mass index, and age all had
similar RR’s.
From the above observational cohort studies, our model assumes that HRT
is effective in the primary prevention of CHD (i.e. among
women who do not already have CHD). How does the recently published
HERS trial affect this assumption (18)? Among women with established
CHD, this randomized trial found that HRT did not significantly
decrease the rate of secondary cardiac events (relative
hazard 0.99, 95% CI, 0.80–1.22). However, when analyses were
restricted to women who were compliant with study medications, the
relative hazard did decrease (0.87), but was still not statistically
significantly. HRT was associated with a statistically significant time
trend, with more CHD events in the group receiving HRT than in the
placebo group in year 1, but fewer in years 4 and 5. Because of
uncertainties in the impact of HRT among women with established CHD, we
assumed that HRT has no impact on the secondary prevention of CHD.
Most of the survival benefit from HRT results from its cardioprotective
effect; its osteoporotic protective effect has a smaller impact on
survival because of the relatively late age at which fractures occur.
The higher the baseline risk for CHD, the greater the survival benefits
from HRT. Therefore, any change in risk factors that would lower the
risk of CHD for the population at large (such as a decrease in the
prevalence of smoking, diabetes, hypertension, or hypercholesterolemia)
might well dampen the expected benefits of HRT on survival. A recently
published study suggests that HMG CoA reductase inhibitors are
effective in the primary prevention of CHD (31). The use of these
"statins" would be expected to decrease the cardioprotective
benefits of HRT by decreasing the underlying risk for CHD events.
Substantial changes in a woman’s risk profile should prompt a
reassessment of the risks and benefits of HRT.
|
Conclusion
|
|---|
Some investigators argue that it is inappropriate to guide
clinical practice in the absence of randomized, controlled trials
(RCT’s) and that any decision about HRT should be deferred until such
evidence is available. The two large RCT’s examining this question
will not report for at least another decade, while millions of women
will need to make this decision in the interim. Furthermore, arguing
that until then all women should decline HRT or that
all women should take HRT prejudges those trials.
Additionally, it is unclear whether these trials will ever be able to
clarify the extent to which HRT increases the risk of breast cancer
(which is treated as a secondary outcome in these trials) or to address
the impact of HRT among subgroups of women at elevated risk for
multiple diseases because of limited statistical power of these
long-term trials. While awaiting the results of these trials, we
propose using the best data currently available to help guide women and
their physicians with this difficult individual decision.
|
References
|
|---|
-
Beck JR, Pauker SG. 1983 The Markov process in
medical prognosis. Med Decis Making. 3:419–458.
-
Col NF, Eckman MH, Karas RH, et al. 1997 Patient-specific decisions about hormone replacement therapy in
postmenopausal women. JAMA. 277:1140–1147.[Abstract]
-
Anderson KM, Odell PM, Wilson PWF, Kannel WB. 1990 Cardiovascular disease risk profiles. Am Heart J. 121:293–298.
-
Anderson KM, Wilson PWF, Odell PM, Kannel WB. 1991 An updated coronary risk profile, A statement for health professionals. Circulation. 83:356–362.[Free Full Text]
-
Gail MK, Brinton LA, Byar DP, et al. 1989 Projecting individualized probabilities of developing breast cancer for
white females who are being examined annually. J Natl Cancer Inst. 81:1879–1886.[Abstract/Free Full Text]
-
Cummings SR, Nevitt MC, Browner WS, et al. 1995 Risk factors for hip fracture in white women. N Engl J Med. 332:767–773.[Abstract/Free Full Text]
-
Weiss NS, Ure CL, Ballard JH, Williams AR, Daling
JR. 1980 Decreased risk of fractures of the hip and lower forearm
with postmenopausal use of estrogen. N Engl J Med. 303:1195–1198.[Abstract]
-
Grodstein F, Stampfer MJ, Manson JE, et al. 1996 Postmenopausal estrogen and progestin use and the risk of
cardiovascular disease. N Engl J Med. 335:453–461.[Abstract/Free Full Text]
-
Colditz GA, Hankinson SE, Hunter DJ, et al. 1995 The use of estrogens and progestins and the risk of breast cancer in
postmenopausal women. N Engl J Med. 332:1589–1593.[Abstract/Free Full Text]
-
Grady D, Rubin SM, Petitti DB, et al. 1992 Hormone
therapy to prevent disease and prolong life in postmenopausal women. Ann Intern Med. 117:1016–1037.
-
Grady D, Gebretsadik T, Kerlikowske K, Ernster V,
Petitte D. 1995 Hormone replacement therapy and endometrial cancer
risk: A meta-analysis. Obstet Gynecol. 85:304–313.[Abstract]
-
The Writing Group for the PEPI Trial. 1996 Effects
of hormone replacement therapy on endometrial histology in
postmenopausal women. The Postmenopausal Estrogen/Progestin
Interventions (PEPI) Trial. JAMA. 275:370–375.[Abstract]
-
Speroff L, Rowan J, Symons J, Genant H, Wilborn W, for
the CHART Study Group. 1996 The comparative effect on bone
density, endometrium, and lipids of continuous hormones as replacement
therapy (CHART Study). JAMA. 276:1397–1403.[Abstract]
-
Beck JR, Pauker SG. 1982 A convenient approximation
of life expectancy (The "DEALE") I. Validation of the method. Am J Med. 73:883–888.[CrossRef][Medline]
-
Beck JR, Pauker SG. 1982 A convenient approximation
of life expectancy (The "DEALE") II. Use in medical decision
making. Am J Med. 73:889–897.[CrossRef][Medline]
-
Kannel WB, Wolf PA, Garrison RJ. 1988 The
Framingham Study: An epidemiological investigation of cardiovascular
disease. Section 35: Survival following initial cardiovascular events.
Thirty year follow-up. NIH Publication No. 88-2969. Bethesda, Maryland.
National Heart, Lung, and Blood Institutes of Health.
-
The Framingham Study. 1970 An epidemiological
investigation of cardiovascular disease. Kannel WB, Gordon T, eds
Section 25. Joyce Schiffman. Survival following certain cardiovascular
events. Washington, D.C., U.S. Government Printing Office.
-
Hulley S, Grady D, Bush T, et al. 1998 Randomized
trial of estrogen plus progestin for secondary prevention of coronary
heart disease in postmenopausal women. JAMA. 280:605–613.[Abstract/Free Full Text]
-
Col NF. 1997 A woman doctor’s guide to hormone
therapy. How to choose what’s right for you. Worcester, MA: Chandler
House Press.
-
Schneider DL, Barrett-Connor EL, Morton DJ. 1997 Timing of postmenopausal estrogen for optimal bone mineral density. The
Rancho Bernardo Study. JAMA. 277:543–547.[Abstract]
-
Cauley JA, Seeley DG, Ensrud K, Ettinger B, Black D,
Cummings SR, for the Study of Osteoporotic Fractures Research
Group. 1995 Estrogen replacement therapy and fractures in older
women. Ann Intern Med. 122:9–16.[Abstract/Free Full Text]
-
Consensus Development Conference. 1993 Diagnosis,
prophylaxis, and treatment of osteoporosis. Am J Med. 94:646–650.[CrossRef][Medline]
-
Love S. 1997 Sometimes mother nature knows best.
New York Times. natl. ed. Mar 20: A39.
-
American Heart Association. 1993 Cardiovascular
disease in women. A special report. Special Writing Group: Eaker ED,
Chesebro JH, Sacks FM, Wenger NK, Whisnant JP, Winston M. American
Heart Association.
-
Grady D, Rubin SM, Petitti DB, et al. 1992 Hormone
therapy to prevent disease and prolong life in postmenopausal women. Ann Intern Med. 117:1016–1037.
-
Nachtigall LE, Nachtigall RH, Nachtigall RD, Beckman
EM. 1983 Estrogen replacement therapy II: a prospective study in
the relationship to carcinoma and cardiovascular and metabolic
problems. Obstet Gynecol. 62:435–443.[Abstract/Free Full Text]
-
The Writing Group for the PEPI Trial. 1995 Effects
of estrogen or estrogen/progestin regimens on heart disease risk
factors in postmenopausal women. The Postmenopausal Estrogen/Progestin
Interventions (PEPI) Trail. JAMA. 273:199–208.[Abstract]
-
Koh KK, Mincemoyer R, Bui MN, et al. 1997 Effects
of hormone replacement therapy on fibrinolysis in postmenopausal women. N Engl J Med. 336:683–690.[Abstract/Free Full Text]
-
Gilligan DM, Badar DM, Panza JA, Quyyumi AA, Cannon RO.1994 Acute vascular effects of estrogen in postmenopausal women. 90:786–791.
-
Bar J, Tepper R, Fuchs J, Pardo Y, Goldberger S, Ovadia
J. 1993 The effect of estrogen replacement therapy on platelet
aggregation and adenosine triphosphate release in postmenopausal women. Obstet Gynecol. 81:261–264.[Abstract/Free Full Text]
-
Downs JR, Clearfield M, Weis S, et al. 1998 Primary
prevention of acute coronary events with lovastatin in men and women
with average cholesterol levels. JAMA. 279:1615–1622.[Abstract/Free Full Text]
-
U.S. Department of Health and Human Services. 1995 Vital Statistics of the United States. Volume II-mortality, Part B.
National Center for Health Statistics 1995; Hyattsville,
Maryland.
Top
Introduction
Analytic approach
