close

Вход

Забыли?

вход по аккаунту

?

Effect of hormone therapy on risk of hip and knee joint replacement in the women's health initiative.

код для вставкиСкачать
ARTHRITIS & RHEUMATISM
Vol. 54, No. 10, October 2006, pp 3194–3204
DOI 10.1002/art.22138
© 2006, American College of Rheumatology
Effect of Hormone Therapy on Risk of
Hip and Knee Joint Replacement in
the Women’s Health Initiative
Dominic J. Cirillo,1 Robert B. Wallace,2 LieLing Wu,3 and Robert A. Yood4
arthroplasty (HR 0.84 [95% CI 0.70–1.00], P ⴝ 0.05).
However, this effect was borderline statistically significant for hip arthroplasty (HR 0.73 [95% CI 0.52–1.03],
P ⴝ 0.07), and not significant for knee arthroplasty (HR
0.87 [95% CI 0.71–1.07], P ⴝ 0.19). In the estrogen-plusprogestin trial, there was no association for total arthroplasty (HR 0.99 [95% CI 0.82–1.20], P ⴝ 0.92) or for
individual hip (HR 1.14 [95% CI 0.83–1.57], P ⴝ 0.41)
or knee (HR 0.91 [95% CI 0.72–1.15], P ⴝ 0.41)
arthroplasties.
Conclusion. These data suggest that hormone
therapy may influence joint health, but this observed
decrease in risk may be limited to unopposed estrogen
and may possibly be more important in hip than in knee
osteoarthritis.
Objective. To determine the effect of hormone
therapy on arthroplasty rates.
Methods. We examined data from the Women’s
Health Initiative placebo-controlled, double-blind, randomized trials. Community-dwelling women ages 50–79
years were enrolled at 40 US clinics. Women with prior
arthroplasty were excluded, yielding a sample size of
26,321 subjects. Women who had had hysterectomies
(n ⴝ 10,272) were randomly assigned to receive 0.625
mg/day conjugated equine estrogens (n ⴝ 5,076), or
placebo (n ⴝ 5,196), with a mean followup of 7.1 years.
Those who had not had hysterectomies (n ⴝ 16,049)
were randomly assigned to receive estrogen plus progestin (n ⴝ 8,240), given as 0.625 mg/day conjugated equine
estrogens plus 2.5 mg/day medroxyprogesterone acetate,
or placebo (n ⴝ 7,809), with a mean followup of 5.6
years. Participants reported hospitalizations, and arthroplasties were identified by procedure codes. Arthroplasties due to hip fracture were censored. Cox proportional hazards regression was used to assess hazard
ratios (HRs) and 95% confidence intervals (95% CIs)
using intent-to-treat methods and outcome of time to
first procedure.
Results. In the estrogen-alone trial, women receiving hormone therapy had significantly lower rates of any
Osteoarthritis (OA) affects approximately 1 in 8
Americans over age 25 years and is the leading chronic
medical condition among the elderly (1), reaching 80%
prevalence by age 75 years (2). Costs related to OA
include medication, direct and indirect costs of disability, and surgical procedures performed when conservative management fails to control pain and dysfunction
(1,3,4). In 1999 in the US, there were ⬃249,000 total
knee replacement surgeries for OA, and another 200,000
patients had total hip replacements (5).
The literature on hormone therapy and OA onset
and progression is complex due to varying definitions of
disease, different classifications of hormone therapy use,
joint differences, and study design effects. In addition,
many studies have not differentiated the possible effects
of various types of hormone therapy, such as unopposed
estrogen or estrogen plus progestin. Hormone therapy
has been shown to help relieve menopausal symptoms,
including joint pain (6). Some cross-sectional studies
suggest that current hormone therapy use protects
against radiographic evidence of OA (7,8). Cohort studies of radiographic OA have tended to show nonsignif-
The Women’s Health Initiative is supported by the NIH
(National Heart, Lung, and Blood Institute) and by contract N01-WH2-2110 from the Department of Health and Human Services.
1
Dominic J. Cirillo, BS: University of Iowa College of Public
Health, Iowa City; 2Robert B. Wallace, MD, MSc: University of Iowa
College of Public Health, and University of Iowa Carver College of
Medicine, Iowa City; 3LieLing Wu, MS: Fred Hutchinson Cancer
Research Center, Seattle, Washington; 4Robert A. Yood, MD: Fallon
Clinic, Worcester, Massachusetts, and University of Massachusetts
Medical School, Worcester.
Address correspondence and reprint requests to Robert B.
Wallace, MD, MSc, Department of Epidemiology C21-N GH, University of Iowa College of Public Health, 200 Hawkins Drive, Iowa City,
IA 52242. E-mail: robert-wallace@uiowa.edu.
Submitted for publication March 15, 2006; accepted in revised
form June 28, 2006.
3194
EFFECT OF ESTROGEN THERAPY ON JOINT REPLACEMENT
icant protective effects (9–11). However, since radiographic and self-reported diagnoses may not reflect
disabling disease, arthroplasty may be considered a
viable proxy outcome for severe OA (12). Studies examining risk of joint replacement due to OA have demonstrated increased risk (13), no association (14), a protective effect of hormone therapy (15), or mixed results
depending on length of use (16). These reports do not
form a cohesive picture of the potential effect of hormone therapy on OA or arthroplasty.
In this report, we present findings from the
Women’s Health Initiative (WHI), which included 2
randomized hormone trials based on participant hysterectomy status. A partial list of WHI investigators
is shown in Appendix A; a complete list is shown on
the WHI Web site (http://www.whi.org/about/
investigators.php). Women who had had a hysterectomy
were randomized to receive unopposed estrogen or
placebo, and those whose uteri were intact received a
combination of estrogen plus progestin or placebo (17).
The major findings of these trials were previously reported (18,19). Although hip and knee arthroplasty were
not primary end points, the WHI is the largest randomized, double-blind trial of postmenopausal hormone
therapy to date and provides an important opportunity
to examine the effect of hormone therapy on the need
for arthroplastic procedures among healthy, communitydwelling postmenopausal women.
PATIENTS AND METHODS
Role of the funding source. The National Heart, Lung,
and Blood Institute did not participate in the design and
conduct of the study, in the collection, analysis, and interpretation of the data, or in the preparation, review, or approval of
the manuscript.
Participants. The WHI consisted of randomized trials
of hormone therapy as well as dietary modification and
calcium/vitamin D supplementation. Detailed discussions of
the recruitment methods and the principal findings of the WHI
trials are available elsewhere (17,19). Briefly, women who were
between the ages of 50 and 79 years at an initial screening visit
to 1 of 40 clinical sites were assessed for eligibility and
provided written informed consent to enroll in the hormone
therapy trial. The National Institutes of Health (NIH) and
Institutional Review Boards for all participating centers approved the WHI protocols and consent forms. The screening
and enrollment algorithm is shown in Figure 1.
Women were excluded from the hormone trials if they
had a terminal illness or a contraindication to hormone
therapy, or if they were considered at high risk of poor
adherence. Furthermore, in this secondary analysis, we excluded women with a history of arthroplasty at baseline.
Women who had had a hysterectomy were eligible for the
estrogen-alone trial, and women who had not had a hysterec-
3195
Figure 1. Profile of the hormone therapy trials of the Women’s
Health Initiative.
tomy were enrolled in the estrogen-plus-progestin trial. Eligible women who agreed to participate were randomly assigned
in equal proportions to receive hormone therapy or placebo
using a stratified permuted block algorithm (17,20). For
women currently receiving hormone therapy at study enrollment, a 3-month washout period was required before randomization assignment. The active study drug and placebo were
kindly supplied by Wyeth (St. Davids, PA). Women in the
estrogen-alone trial received either placebo or 0.625 mg/day
conjugated equine estrogens (Premarin; Wyeth), using blinded
dispensing. Women in the estrogen-plus-progestin trial received either placebo or 0.625 mg/day conjugated equine
estrogens plus 2.5 mg/day medroxyprogesterone acetate, administered as a single tablet (Prempro; Wyeth).
In order to reinforce adherence and to assess side
effects, followup telephone calls were conducted 6 weeks
postrandomization. If the participant developed side effects
from the treatment, the dosage or frequency could be altered
by study physicians, in most instances without unblinding the
subject or the study clinicians to the treatment assignment.
Infrequently, an unblinding officer could disclose the treatment assignment to study clinicians for symptom management
or safety concerns. Study clinicians were unblinded at rates of
2% in each group in the estrogen-alone trial and at rates of
41% and 6%, respectively, in the estrogen-plus-progestin trial
active and placebo groups; however, these clinicians were not
involved in the assessment of study outcomes.
Data collection and outcomes. Baseline measurements
were collected in a standardized manner, using extensive
medical history questionnaires, self-administered forms, interviews, and clinical examinations. Weight and height were
measured using a calibrated balance beam scale and a wallmounted stadiometer, respectively. Body mass index (BMI)
was calculated as weight (in kilograms) divided by height (in
meters squared). Race and ethnicity were self reported using
predefined categories, to assess treatment effect differences in
subgroup analyses. Other clinical, demographic, social, and
behavioral baseline characteristics were determined from selfadministered, structured questionnaires (17). Subjects were
asked about their medical history (including arthroplasty) as
3196
CIRILLO ET AL
Table 1. Baseline descriptive characteristics of the Women’s Health Initiative hormone therapy trials*
Estrogen-alone trial
Age group at screening, years
50–59
60–69
70–79
Ethnicity
White
Black
Hispanic
American Indian
Asian/Pacific Islander
Other/unspecified
Education
0–8 years
Some high school
High school/GED
School after high school
College degree or higher
Time since quitting estrogen
Nonuser
Past, ⬍5 years
Past, 5 to ⬍10 years
Past, ⱖ10 years
Current†
BMI, kg/m2
⬍18.5
18.5–24.9
ⱖ25.0
Baseline joint pain/stiffness
None
Mild
Moderate
Severe
NSAID use at baseline
Aspirin use at baseline
Estrogen-plus-progestin trial
Hormone therapy
(n ⫽ 5,076)
Placebo
(n ⫽ 5,196)
Hormone therapy
(n ⫽ 8,240)
Placebo
(n ⫽ 7,809)
1,611 (31.7)
2,287 (45.1)
1,178 (23.2)
1,650 (31.8)
2,368 (45.6)
1,178 (22.7)
2,815 (34.2)
3,734 (45.3)
1,691 (20.5)
2,652 (34.0)
3,522 (45.1)
1,635 (20.9)
3,817 (75.2)
751 (14.8)
315 (6.2)
40 (0.8)
84 (1.7)
69 (1.4)
3,889 (74.8)
802 (15.4)
322 (6.2)
34 (0.7)
76 (1.5)
73 (1.4)
6,901 (83.8)
535 (6.5)
466 (5.7)
26 (0.3)
190 (2.3)
122 (1.5)
6,532 (83.6)
563 (7.2)
411 (5.3)
29 (0.4)
169 (2.2)
105 (1.3)
176 (3.5)
330 (6.5)
1,184 (23.3)
2,179 (42.9)
1,154 (22.7)
140 (2.7)
344 (6.6)
1,136 (21.9)
2,247 (43.2)
1,284 (24.7)
195 (2.4)
356 (4.3)
1,562 (19.0)
3,250 (39.4)
2,831 (34.4)
172 (2.2)
347 (4.4)
1,547 (19.8)
2,942 (37.7)
2,743 (35.1)
2,639 (52.0)
547 (10.8)
247 (4.9)
994 (19.6)
648 (12.8)
2,636 (50.7)
564 (10.9)
263 (5.1)
1,046 (20.1)
684 (13.2)
6,073 (73.7)
714 (8.7)
325 (3.9)
584 (7.1)
540 (6.6)
5,794 (74.2)
662 (8.5)
297 (3.8)
569 (7.3)
484 (6.2)
1,080 (21.3)
1,727 (34.0)
2,241 (44.1)
1,073 (20.7)
1,840 (35.4)
2,248 (43.3)
2,526 (30.7)
2,912 (35.3)
2,768 (33.6)
2,416 (30.9)
2,759 (35.3)
2,585 (33.1)
1,178 (23.2)
2,301 (45.3)
1,151 (22.7)
386 (7.6)
1,777 (35.0)
1,044 (20.6)
1,234 (23.7)
2,343 (45.1)
1,193 (23.0)
375 (7.2)
1,846 (35.5)
1,064 (20.5)
2,437 (29.6)
3,951 (47.9)
1,428 (17.3)
360 (4.4)
2,725 (33.1)
1,668 (20.2)
2,400 (30.7)
3,663 (46.9)
1,330 (17.0)
343 (4.4)
2,650 (33.9)
1,632 (20.9)
* Values are the number (%) of patients. GED ⫽ general educational development; BMI ⫽ body mass index; NSAID ⫽ nonsteroidal
antiinflammatory drug.
† A 3-month washout period was required prior to randomization.
well as current symptoms (including joint pain/stiffness, as
“mild,” “moderate,” “severe,” or “did not occur”).
Main study outcomes. Clinical outcomes were ascertained semiannually via questionnaires and annual clinic visits.
The semiannual followup included standardized questionnaires on specific symptoms and safety concerns, as well as
major health events, including the primary outcomes of heart
disease, stroke, hip fracture, and breast cancer as well as
hospitalizations and significant medical procedures. Participants provided consent to have their medical and hospital
records reviewed by study physicians. The completion rate for
clinical record acquisition and adjudication was estimated to
be 96%. Outcomes used in this analysis were obtained from
medical records and identified with International Classification of Diseases, Ninth Revision (ICD-9) codes. Total hip
replacement (ICD-9 code 81.51) and total knee replacement
(ICD-9 code 81.54) were analyzed separately, and an index of
either hip or knee arthroplasty was also considered. Hip and
femoral neck fractures, which were primary outcomes in the
hormone therapy trials, were ascertained in a similar manner,
and women were censored at the date of fracture during the
followup period (n ⫽ 340). As described elsewhere in detail
(19), the estrogen-plus-progestin trial was discontinued earlier
than scheduled after a recommendation from the Data and
Safety Monitoring Board that the overall harms outweighed
the benefits of estrogen plus progestin. Close-out of the
estrogen-alone clinical trial was originally planned for October
2004 through March 2005; early study termination (February
2004) resulted in a followup time ⬃1 year shorter than planned
(18).
Statistical analysis. All statistical analyses were conducted using the SAS System for Windows, version 9.0 (SAS
Institute, Cary, NC). The time to the first arthroplasty event
was the primary analytical outcome. Associations between
baseline variables and randomization assignment were assessed using P values from chi-square tests for categorical
variables and 2-sample t-tests for continuous variables. All
outcomes analyses used time-to-event methods and were based
EFFECT OF ESTROGEN THERAPY ON JOINT REPLACEMENT
3197
Table 2. Incidence of arthroplasty by randomization assignment in the unopposed-estrogen (estrogen-alone) trial of the Women’s
Health Initiative*
Followup time, mean ⫾ SD years
Total joint replacement‡
Total hip replacement
Total knee replacement
Censoring after nonadherence§
Total joint replacement‡
Total hip replacement
Total knee replacement
Estrogen alone
(n ⫽ 5,076)
Placebo
(n ⫽ 5,196)
HR (95% CI)†
P†
7.1 ⫾ 1.6
222 (62)
58 (16)
169 (47)
7.1 ⫾ 1.7
269 (74)
81 (22)
198 (54)
0.84 (0.70–1.00)
0.73 (0.52–1.03)
0.87 (0.71–1.07)
0.05
0.07
0.19
119 (56)
28 (13)
93 (44)
169 (78)
53 (24)
121 (56)
0.73 (0.58–0.93)
0.55 (0.35–0.88)
0.80 (0.61–1.05)
0.01
0.01
0.11
* Except where indicated otherwise, values are the number of patients undergoing arthroplasty (rate per 10,000 person-years).
† Hazard ratios (HRs), 95% confidence intervals (95% CIs), and P values from Cox proportional hazards analyses stratified by age.
‡ Events include any hip or knee arthroplasty.
§ Analyzed per-protocol, with followup time of participants censored at 6 months post-nonadherence, which was predefined as
consuming ⬍80% of assigned study pills.
on the intent-to-treat (ITT) principle. Outcome comparisons
are presented as hazard ratios (HRs) and 95% confidence
intervals (95% CIs) from Cox proportional hazards models,
stratified by age. Random assignment of subjects to calcium/
vitamin D supplementation occurred 1–2 years after randomization in the hormone trial, so the calcium/vitamin D randomization date was treated as the only time-dependent covariate
in the analysis. This adjustment did not appreciably change the
results, and therefore the findings are not presented. The
proportionality assumption was tested by including a time–
treatment interaction term in the Cox model, in addition to the
treatment assignment indicator. None of the P values was
significant at the 5% level, which indicated no evidence of
departure from proportionality. Absolute risk reduction was
calculated as the difference in incidence rates between groups.
This study was a secondary analysis of a large trial that was
designed and sufficiently powered to detect clinically relevant
changes in the primary outcomes of coronary heart disease, hip
fracture, and breast cancer, but the study was not specifically
powered to detect differences in arthroplasty rates.
Kaplan-Meier plots illustrated the risk of arthroplasty
over the followup period, and log rank test statistics were
calculated to test for differences between treatment assignment curves. Sensitivity analyses were conducted to examine
the effect of nonadherence, by repeating these analyses and
censoring followup 6 months after a woman ceased to adhere
to the assigned treatment, which was prospectively defined
either as consuming ⬍80% of study pills as determined by pill
count or as starting non–study prescribed hormone therapy.
RESULTS
Baseline subject characteristics for the entire
study cohort were presented elsewhere (19,21). The
primary reports from these trials described the results
for 5,310 women in the conjugated equine estrogens
group and 5,429 women in the placebo group for the
estrogen-alone trial (18), and for 8,506 women receiving
estrogen plus progestin and 8,102 women receiving
placebo in the estrogen-plus-progestin trial (19). For the
present analysis, the study population included these
women, but exclusions were made if the participant had
a history of arthroplasty (n ⫽ 1,026; 353 with hip
replacement, 598 with other joint arthroplasty, and 75
with both conditions).
The final sample size for this study was 26,321
(10,272 in the estrogen-alone trial and 16,049 in the
estrogen-plus-progestin trial) (Figure 1). The proportion
of women who were excluded from the analysis based on
history of arthroplasty was slightly higher for the
estrogen-alone trial compared with the estrogen-plusprogestin trial (4.3% versus 3.4%; P ⬍ 0.001), but these
subjects were balanced between treatment groups within
each trial. The treatment and control groups were also
balanced on key baseline demographic and disease risk
factors (Table 1). The only variables that were significantly different were found in the estrogen-alone trial, in
which the placebo group subjects had attained a significantly higher educational level (P ⫽ 0.02) and were
more likely to have a history of cancer (P ⬍ 0.01),
although the difference between the groups was small,
and statistical significance was driven by the large number of women enrolled in the trial. The remaining
variables were not significantly different (P ⬎ 0.10), and
no significant differences at baseline were found in the
estrogen-plus-progestin trial.
Forty-nine percent of women with hysterectomies had been previous or current users of hormone
therapy at study enrollment, whereas only 26% of
women in the estrogen-plus-progestin study used hormones prior to the start of the study. Among the past
users of hormone therapy in the estrogen-alone trial,
41% had not received hormone therapy in at least 10
3198
CIRILLO ET AL
Table 3. Incidence of arthroplasty by randomization assignment in the estrogen-plus-progestin trial of the Women’s Health
Initiative*
Followup time, mean ⫾ SD years
Total joint replacement‡
Total hip replacement
Total knee replacement
Censoring after nonadherence§
Total joint replacement‡
Total hip replacement
Total knee replacement
Estrogen plus
progestin
(n ⫽ 8,240)
Placebo
(n ⫽ 7,809)
HR (95% CI)†
P†
5.7 ⫾ 1.3
222 (48)
87 (19)
138 (30)
5.6 ⫾ 1.3
206 (48)
70 (16)
140 (32)
0.99 (0.82–1.20)
1.14 (0.83–1.57)
0.91 (0.72–1.15)
0.92
0.41
0.41
136 (43)
49 (16)
87 (28)
138 (44)
48 (15)
94 (30)
1.02 (0.81–1.29)
1.08 (0.72–1.61)
0.95 (0.71–1.27)
0.77
0.71
0.72
* Except where indicated otherwise, values are the number of patients undergoing arthroplasty (rate per 10,000 person-years).
† Hazard ratios (HRs), 95% confidence intervals (95% CIs), and P values from Cox proportional hazards analyses stratified by age.
‡ Events include any hip or knee arthroplasty.
§ Analyzed per-protocol, with followup time of participants censored at 6 months post-nonadherence, which was predefined as
consuming ⬍80% of assigned study pills.
years, and 27% had been currently receiving hormone
therapy prior to randomization. In the estrogen-alone
trial, 44% of participants were overweight (BMI ⬎25),
compared with 33% in the estrogen-plus-progestin trial.
The estrogen-alone trial was stopped on February 29, 2004, with a mean followup period of 7.1 years
and a maximum followup period of 10.2 years (18), with
comparable adherence between the estrogen and placebo groups (53.8% for both treatments at study termination). Since the estrogen-plus-progestin trial was
halted early (on July 7, 2002), the mean followup time in
this analysis was 5.6 years, with a maximum followup
period of 8.6 years at the time of trial closure. The mean
followup time in this report was 5.6 years for those who
received placebo and 5.7 years for those who received
estrogen plus progestin. Forty-two percent of subjects
randomly assigned to receive estrogen plus progestin
and 38% of subjects randomly assigned to receive placebo were nonadherent during the study.
In the estrogen-alone trial, the incidence rate for
total hip replacement events was 16 per 10,000 personyears in the estrogen-alone group compared with 22 per
10,000 person-years in the placebo group (Table 2). For
total knee replacement events, the rates were 47 and 54
per 10,000 person-years in the estrogen-alone and placebo groups, respectively. A total of 222 women had at
least 1 total joint replacement in the estrogen-alone
group, compared with a total of 269 women in the
placebo group, which yielded incidence rates for any
total joint replacement of 62 per 10,000 person-years in
the estrogen-alone group and 74 per 10,000 person-years
in the placebo group. In Cox proportional hazards
modeling, estrogen alone was found to be protective
against any joint replacement compared with placebo
(HR 0.84 [95% CI 0.70–1.00], P ⫽ 0.05). When hip and
knee joints were considered separately, the protective
effect approached statistical significance in predicting
total hip replacement (HR 0.73 [95% CI 0.52–1.03], P ⫽
0.07), but not in predicting total knee replacement (HR
0.87 [95% CI 0.71–1.07], P ⫽ 0.19).
In the estrogen-plus-progestin trial, the estimated
rate of any arthroplasty was 48 per 10,000 person-years
in both the estrogen-plus-progestin and placebo groups
(Table 3). For total hip replacement, the rates were 19
and 16 per 10,000 person-years in the estrogen-plusprogestin and placebo groups, respectively. For total
knee replacement, the rates were 30 and 32 per 10,000
person-years, respectively. The HRs were not statistically significant for total hip replacement (HR 1.14 [95%
CI 0.83–1.57], P ⫽ 0.41), total knee replacement (HR
0.91 [95% CI 0.72–1.15], P ⫽ 0.41), or for the index of
any joint replacement (HR 0.99 [95% CI 0.82–1.20], P ⫽
0.92).
When women were censored 6 months after they
became nonadherent, the protective effect of estrogen
alone was strengthened (Table 2). The HRs became 0.73
for any arthroplasty (95% CI 0.58–0.93, P ⫽ 0.01), 0.55
for total hip replacement (95% CI 0.35–0.88, P ⫽ 0.01),
and 0.80 for total knee replacement (95% CI 0.61–1.05,
P ⫽ 0.11). However, censoring nonadherent participants
had little effect on the estimates or significance in the
estrogen-plus-progestin trial (Table 3). Adjustment for
educational attainment did not appreciably change the
findings (data not shown).
EFFECT OF ESTROGEN THERAPY ON JOINT REPLACEMENT
3199
Figure 2. Kaplan-Meier estimate of cumulative hazards for arthroplasty in the unopposed-estrogen trial. Risks of hip (A) and knee (B) joint
replacement are shown separately among women receiving conjugated equine estrogens (CEE) alone or placebo. Total joint replacement (C)
includes time to either outcome. A further analysis was conducted with censoring occurring at 6 months after nonadherence to randomization
assignment (D). HR ⫽ hazard ratio; 95% CI ⫽ 95% confidence interval.
The possibility of differential hormone therapy
effects in some subgroups, including subjects grouped
according to age, ethnicity, past hormone therapy, and
tobacco and alcohol consumption, was considered for
each joint separately. When estrogen alone was considered as a predictor of total hip replacement (crude HR
0.73), there was no evidence of effect modification by
these variables. However, when estrogen alone was
considered with total knee replacement (crude HR
0.87), significantly different hormone therapy effects
among race/ethnic groups were observed (P for interaction ⫽ 0.04). The protective effect was limited to whites
(HR 0.79 [95% CI 0.63–0.99]). Hormone therapy
showed no effect in blacks (HR 1.02 [95% CI 0.60–
1.73]). Among Hispanics, hormone therapy appeared to
elevate risk (HR 3.51 [95% CI 0.97–12.76]). However, it
3200
CIRILLO ET AL
should be noted that these findings are based on a small
number of minority participants, and the probability of
at least 1 interaction test being significant by chance
alone was ⬎0.60 based on the number of tests conducted.
The Kaplan-Meier estimates of cumulative hazards of total hip replacement, total knee replacement,
and any joint replacement in the estrogen-alone trial
showed the effects of hormone therapy since the time of
study enrollment (Figures 2A–C). The estrogen-plusprogestin trial curves did not show a divergence in risk
between the treatment and placebo groups based on
Kaplan-Meier plots and log-likelihood test (data not
shown). In the estrogen-alone trial, the risk for hip
replacements appeared to diverge after 1 year of followup, whereas the apparent divergence in knee replacement risk began at 2 years. When the analysis was
limited by censoring subjects 6 months after they became
nonadherent to the assigned therapy (Figure 2D), the
estrogen-alone trial showed more separation between
the curves but the timing of hormone therapy effects was
not substantially altered.
DISCUSSION
This study included 2 separate populations, that
is, more than 26,000 healthy, community-dwelling postmenopausal women with and without hysterectomy.
Those who had had a hysterectomy were randomly
assigned to receive a common oral estrogen therapy, and
those who had not had a hysterectomy were randomly
assigned to receive estrogen plus progestin. We studied
the rates, during followup, of hip and knee arthroplasties, which are important clinical end points of severely
symptomatic OA. A statistically significant decrease in
arthroplasty risk was seen in the estrogen-alone trial.
Estrogen alone also appeared to be stronger in preventing hip arthroplasty (27% decrease) than in preventing
knee arthroplasty (13% decrease), although the individual joint findings were not statistically significant in the
ITT analysis. The estimated absolute risk reduction for
total joint replacement was 12 per 10,000 person-years,
and, similarly, was 6 per 10,000 person-years for total hip
replacement and 7 per 10,000 person-years for total
knee replacement. When we censored women whose
adherence to their assigned therapy dropped below
80%, the protective effect was more striking. However,
the estrogen-plus-progestin trial still showed no relationship between hormone use and arthroplasty risk. To our
knowledge, this is the first randomized trial to show
evidence of a protective effect attributed to estrogen-
alone therapy, while demonstrating no effect of estrogen
in the presence of progestin.
Observational studies evaluating clinical symptoms, signs, and physician diagnosis have indicated that
hormone therapy increases the risk of OA. Results of
the cross-sectional Rancho Bernardo study suggested
that hip and hand OA (but not knee OA), as defined by
pain questionnaires and clinical examination, were more
prevalent among hormone therapy users, but the investigators also found a nonsignificant protective effect of
recent hormone therapy initiation among current users
(22). Cohort studies have also tended to reveal an
increased OA incidence among hormone therapy users.
In the National Health and Nutrition Examination Survey Epidemiological Follow-up Study, investigators
found a dose-dependent association of duration of
hormone therapy use with risk of self-reported incident
OA (23). A cohort study among Canadian women, the
Population Health Survey, showed a 2-fold relative risk
for self-reported incident OA over a 2-year period
among current, long-term hormone therapy users (24).
Oliveria and colleagues used data from a nested case–
control study to show that new hormone therapy use was
associated with new clinical diagnosis of arthritis (25).
The investigators in that study found that OA diagnosis
was associated with frequency of physician visits.
Since OA may have an insidious onset, detection
of incident disease may be subject to surveillance bias.
Health care utilization may be a consequence of joint
symptoms, but frequent visits for other reasons, such as
the treatment of perimenopausal symptoms, could result in coincidental diagnosis of OA. Therefore, this
bias may cause positive associations among users of
hormone therapy due to study design artifacts, especially
when study outcomes are new diagnosis or medical
claims of OA.
Unlike the studies of clinical manifestations,
most radiographic studies suggested a protective effect
of hormone therapy against radiographic OA incidence
or progression (7–11). Nevitt and colleagues found
radiographic evidence of protection against hip OA
among hormone therapy users in the cross-sectional
Study of Osteoporotic Fractures, and the protective
effect was greater among long-term hormone therapy
users (7). Furthermore, the protective effect appeared to
be limited to unopposed estrogen (7). The Chingford
initial cross-sectional study also showed a protective
association with radiographic knee OA (8), although the
later followup study showed a nonsignificant protective
effect against incident knee osteophytosis (10). The
Framingham Osteoarthritis Study initially showed a
EFFECT OF ESTROGEN THERAPY ON JOINT REPLACEMENT
modest, nonsignificant protective effect among women
with longer hormone therapy use (9). After 14 years of
followup, the Framingham Study showed a nonsignificant protective effect against progression of radiographic knee OA among postmenopausal women currently using hormone therapy (11). Previous findings of
protective effects could also be affected by confounding
effects, since estrogen users have been shown to have
healthier behaviors and attitudes (26), but our randomized trial limits the likelihood of confounding.
The radiographic findings suggest that hormone
therapy may influence the detectable progression of OA,
but it is not clear if this relates to symptom severity.
Requirement for arthroplasty has been suggested as a
sensitive and specific proxy outcome for OA progression, particularly in clinical trials (27). In the Nurses’
Health Study cohort, investigators found no association
between hormone therapy and hip replacement due to
arthritis (14). A Swedish case–control study showed a
significant 80% increase in risk of knee arthroplasty due
to OA among postmenopausal women (13). However, a
second Swedish case–control study showed that hormone therapy provided a significant protective effect
against hip replacement due to OA (15). A British
case–control study of hip replacement showed a protective effect among long-term users of hormone therapy
and a harmful effect among new users (16). Some
difficulties may arise in generalizing findings on arthroplasty across health care systems, and reports of these
studies did not distinguish hormone therapy formulation. Arthroplasty may also depend on access to and
utilization of health services, but these factors would be
expected to be balanced in this randomized study.
Only 1 randomized trial, the Heart and Estrogen/
Progestin Replacement Study (HERS), evaluated
estrogen-plus-progestin therapy and OA and found no
effect of estrogen plus progestin on knee pain severity or
disability (28). Our methods differed from those of the
HERS, since we did not have a standardized measure of
joint symptoms. However, the HERS cannot provide a
full picture of the relationship between estrogen and OA
because the HERS investigators did not examine OA of
the hip, arthroplasty outcomes, women without cardiovascular disease, and women using unopposed estrogen.
Our findings were among healthy postmenopausal
women, and significant findings were found only in the
estrogen-alone trial and were stronger for hip, rather
than knee, arthroplasty. This large clinical trial provides
the best evidence to date of a possible protective effect
of estrogen against arthroplasty risk.
One interpretation of the differential effect of
3201
hormone therapy between the estrogen-alone and
estrogen-plus-progestin trials may involve population
differences. That is, women in the estrogen-alone trial
had hysterectomies, and therefore may have had surgically induced menopause and longer durations of “estrogen deficiency” prior to the study compared with
women with natural menopause. Women in the
estrogen-alone trial reported more hormone therapy use
and had greater BMI at baseline than women in the
estrogen-plus-progestin trial.
Since the study protocols were virtually identical
between the separate trials, the fact that these data show
a relationship only in the estrogen-alone trial and not in
the estrogen-plus-progestin trial suggests a true association. If these findings are biologically meaningful, oral
estrogen, in the absence of progestin, must affect OA
development. Based on animal and in vitro data, estrogen appears to have an influential role in the health of
cartilage and large joints. Some studies suggest that
estradiol at physiologic levels yields net beneficial effects
on cartilage metabolism, but that higher doses become
deleterious (29). Both known estrogen receptors (ERs),
ER␣ and ER␤, are expressed in chondrocytes from
human OA hip and knee cartilage (30,31), and polymorphisms in ERs have been associated with generalized
OA (32).
Estrogen is active in the cartilage and synovium.
Estrogen can have harmful effects, since estradiol increases interleukin-6 production in the synovial fluid by
chondrocytes, which leads to joint inflammation and
degeneration (33). However, estrogen also protects
against cartilage degeneration through inhibition of matrix metalloproteinases (MMPs) (34). Additionally, estrogen and progesterone alter the function of fibroblastlike synoviocytes, antagonizing the effects of MMPs
(35). Estrogen also exerts anabolic effects on cartilage
matrix via insulin-like growth factor (IGF) metabolism
by increasing IGF and IGF binding proteins (IGFBPs)
(31,34). IGFBP-2 favors the cartilage matrix by storing
IGF-1 and also increasing proteoglycan production (31).
Also, estrogen inhibits MMP-1, decreasing the breakdown of type II collagen in OA patients (34). Estrogen
therapy reduced cartilage lesions and osteophyte formation in monkeys (36). Furthermore, progestin may modulate the activity of estrogen in the articular space,
suggesting the possibility that local regulation could
contribute to the relationships observed in our study. In
a study of ovariectomized monkeys that compared unopposed estrogen or estrogen plus progestin with control treatment, the estrogen group had significantly
greater levels of IGF-1, IGF-2, IGFBP-1, and IGFBP-3
3202
in the synovial fluid, and levels in the estrogen-plusprogestin group were intermediate between those in the
estrogen and control groups (37).
It is possible that estrogen could have effects
outside of the cartilage itself, such as neurohumoral
effects on pain sensation (38), mood, or weight gain,
which could mediate symptom severity and necessity for
arthroplasty. Estrogen may provide a gross anatomic
effect on maintaining the structural integrity of the knee
joint, as was shown in a study on ovariectomized ewes
(39). In an imaging study of knee cartilage, long-term
users of estrogen had more tibial cartilage than those
who had never used estrogen (40). Estrogen also has
known antiresorptive effects on bone metabolism, and
the estrogenic effect may be analogous to those found in
recent studies of other antiresorptive drugs. Bisphosphonates were found to alter the progression of subchondral
sclerosis in animal models (41) and decreased joint
space narrowing and OA symptoms in a randomized
trial (42). Findings from these studies suggested that
bone remodeling can significantly influence OA progression. However, these alternate explanations do not
discount the observed estrogen association; instead, they
provide insight about potential mechanisms underlying
the association of oral conjugated equine estrogens with
joint replacement risk.
There are some limitations to this study. While it
is likely that the vast majority of the joint replacements
were due to OA, we could not exclude the possibility
that some were due to other causes, such as osteonecrosis or rheumatoid arthritis. However, it is likely that
these would have been distributed equally among all
study groups. Also, women with a history of arthroplasty
were excluded after randomization. These exclusions
applied to a relatively small proportion of subjects and
were balanced between the treatment groups (4.4% in
the estrogen-alone group compared with 4.3% in the
placebo group [P ⫽ 0.77] and 3.1% in the estrogen-plusprogestin group compared with 3.6% in the placebo
group [P ⫽ 0.08]). However, unmeasured confounders
could contribute to differences in the treatment groups.
Unblinding of study staff due to safety concerns was
higher in the estrogen-plus-progestin trial, but these
clinicians were not involved in the assessment of clinical
outcomes and would not have had direct influence on
orthopedic procedures. It is unlikely that the protective
effect against hip arthroplasty seen in the estrogen-alone
trial could be explained by differential surveillance in the
treatment group, because, if anything, increased surveillance in the treatment group would tend to cause an
apparent increase in risk. Finally, these trials have
CIRILLO ET AL
previously shown protective effects against osteoporotic
hip fracture risk with both types of hormone therapy
(18,19). Osteoporotic hip fracture is another significant
cause of total hip replacement, and hormone therapy
could protect against this. However, we censored women
with hip fracture, and the protective effect was limited to
one trial.
More research is needed to elucidate the direct
mechanism of the effect of hormone therapy, since this
cannot be determined from the design of the present
study. These reported findings suggest that estrogen
therapy confers a reduction in the risk of joint degeneration, but since the effect was limited to unopposed
estrogen and may be more important for hip replacement than for knee replacement, there is a need for
further understanding of estrogen’s potential impact on
joint health.
REFERENCES
1. Beaupre GS, Stevens SS, Carter DR. Mechanobiology in the
development, maintenance, and degeneration of articular cartilage. J Rehabil Res Dev 2000;37:145–51.
2. Cooper C, Campbell L, Byng P, Croft P, Coggon D. Occupational
activity and the risk of hip osteoarthritis. Ann Rheum Dis 1996;
55:680–2.
3. Elders MJ. The increasing impact of arthritis on public health.
J Rheumatol Suppl 2000;60:6–8.
4. Canale ST, editor. Campbell’s operative orthopaedics. Vol. 4. 9th
ed. St. Louis: Mosby; 1998.
5. American Academy of Orthopaedic Surgeons. Improving musculoskeletal care in America. URL: http://www3.aaos.org/research/
imca/OA_knee_TOC.htm.
6. Barnabei VM, Cochrane BB, Aragaki AK, Nygaard I, Williams
RS, McGovern PG, et al. Menopausal symptoms and treatmentrelated effects of estrogen and progestin in the Women’s Health
Initiative. Obstet Gynecol 2005;105(5 Pt 1):1063–73.
7. Nevitt MC, Cummings SR, Lane NE, Hochberg MC, Scott JC,
Pressman AR, et al, Study of Osteoporotic Fractures Research
Group. Association of estrogen replacement therapy with the risk
of osteoarthritis of the hip in elderly white women. Arch Intern
Med 1996;156:2073–80.
8. Spector TD, Nandra D, Hart DJ, Doyle DV. Is hormone
replacement therapy protective for hand and knee osteoarthritis in women? The Chingford Study. Ann Rheum Dis 1997;56:
432–4.
9. Hannan MT, Felson DT, Anderson JJ, Naimark A, Kannel WB.
Estrogen use and radiographic osteoarthritis of the knee in
women: the Framingham Osteoarthritis Study. Arthritis Rheum
1990;33:525–32.
10. Hart DJ, Doyle DV, Spector TD. Incidence and risk factors for
radiographic knee osteoarthritis in middle-aged women: the
Chingford Study. Arthritis Rheum 1999;42:17–24.
11. Zhang Y, McAlindon TE, Hannan MT, Chaisson CE, Klein R,
Wilson PW, et al. Estrogen replacement therapy and worsening of
radiographic knee osteoarthritis: the Framingham Study. Arthritis
Rheum 1998;41:1867–73.
12. Dougados M, Gueguen A, Nguyen M, Berdah L, Lequesne M,
Mazieres B, et al. Requirement for total hip arthroplasty: an
EFFECT OF ESTROGEN THERAPY ON JOINT REPLACEMENT
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
outcome measure of hip osteoarthritis? J Rheumatol 1999;26:
855–61.
Sandmark H, Hogstedt C, Lewold S, Vingard E. Osteoarthrosis of
the knee in men and women in association with overweight,
smoking, and hormone therapy. Ann Rheum Dis 1999;58:151–5.
Karlson EW, Mandl LA, Aweh GN, Sangha O, Liang MH,
Grodstein F. Total hip replacement due to osteoarthritis: the
importance of age, obesity, and other modifiable risk factors. Am J
Med 2003;114:93–8.
Vingard E, Alfredsson L, Malchau H. Lifestyle factors and hip
arthrosis: a case referent study of body mass index, smoking and
hormone therapy in 503 Swedish women. Acta Orthop Scand
1997;68:216–20.
Dennison EM, Arden NK, Kellingray S, Croft P, Coggon D,
Cooper C. Hormone replacement therapy, other reproductive
variables and symptomatic hip osteoarthritis in elderly white
women: a case-control study. Br J Rheumatol 1998;37:
1198–202.
The Women’s Health Initiative Study Group. Design of the
Women’s Health Initiative clinical trial and observational study.
Control Clin Trials 1998;19:61–109.
Women’s Health Initiative Steering Committee. Effects of conjugated equine estrogen in postmenopausal women with hysterectomy: the Women’s Health Initiative randomized controlled trial.
JAMA 2004;291:1701–12.
Writing Group for the Women’s Health Initiative. Risks and
benefits of estrogen plus progestin in healthy postmenopausal
women: principal results from the Women’s Health Initiative
randomized controlled trial. JAMA 2002;288:321–33.
Anderson GL, Manson J, Wallace R, Lund B, Hall D, Davis S, et
al. Implementation of the Women’s Health Initiative study design.
Ann Epidemiol 2003;13(9 Suppl):S5–17.
Stefanick ML, Cochrane BB, Hsia J, Barad DH, Liu JH, Johnson
SR. The Women’s Health Initiative postmenopausal hormone
trials: overview and baseline characteristics of participants. Ann
Epidemiol 2003;13(9 Suppl):S78–86.
Von Muhlen D, Morton D, von Muhlen CA, Barrett-Connor E.
Postmenopausal estrogen and increased risk of clinical osteoarthritis at the hip, hand, and knee in older women. J Womens
Health Gend Based Med 2002;11:511–8.
Sahyoun NR, Brett KM, Hochberg MC, Pamuk ER. Estrogen
replacement therapy and incidence of self-reported physiciandiagnosed arthritis. Prev Med 1999;28:458–64.
Wilkins K. Hormone replacement therapy and incident arthritis.
Health Rep 1999;11:49–57(Eng); 57–66(Fre).
Oliveria SA, Felson DT, Klein RA, Reed JI, Walker AM. Estrogen
replacement therapy and the development of osteoarthritis. Epidemiology 1996;7:415–9.
Posthuma WF, Westendorp RG, Vandenbroucke JP. Cardioprotective effect of hormone replacement therapy in postmenopausal
women: is the evidence biased? BMJ 1994;308:1268–9.
Dougados M. Monitoring osteoarthritis progression and therapy.
Osteoarthritis Cartilage 2004;12 Suppl A:S55–60.
Nevitt MC, Felson DT, Williams EN, Grady D. The effect of
estrogen plus progestin on knee symptoms and related disability in
postmenopausal women: the Heart and Estrogen/Progestin Replacement Study, a randomized, double-blind, placebo-controlled
trial. Arthritis Rheum 2001;44:811–8.
Richette P, Corvol M, Bardin T. Estrogens, cartilage, and osteoarthritis. Joint Bone Spine 2003;70:257–62.
Ushiyama T, Ueyama H, Inoue K, Ohkubo I, Hukuda S. Expression of genes for estrogen receptors ␣ and ␤ in human articular
chondrocytes. Osteoarthritis Cartilage 1999;7:560–6.
Richmond RS, Carlson CS, Register TC, Shanker G, Loeser RF.
Functional estrogen receptors in adult articular cartilage: estrogen
replacement therapy increases chondrocyte synthesis of proteogly-
3203
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
cans and insulin-like growth factor binding protein 2. Arthritis
Rheum 2000;43:2081–90.
Ushiyama T, Ueyama H, Inoue K, Nishioka J, Ohkubo I, Hukuda
S. Estrogen receptor gene polymorphism and generalized osteoarthritis. J Rheumatol 1998;25:134–7.
Guerne PA, Carson DA, Lotz M. IL-6 production by human articular
chondrocytes: modulation of its synthesis by cytokines, growth factors,
and hormones in vitro. J Immunol 1990;144:499–505.
Lee YJ, Lee EB, Kwon YE, Lee JJ, Cho WS, Kim HA, et al. Effect
of estrogen on the expression of matrix metalloproteinase
(MMP)-1, MMP-3, and MMP-13 and tissue inhibitor of metalloproteinase-1 in osteoarthritis chondrocytes. Rheumatol Int 2003;
23:282–8.
Khalkhali-Ellis Z, Seftor EA, Nieva DR, Handa RJ, Price RH Jr,
Kirschmann DA, et al. Estrogen and progesterone regulation of
human fibroblast-like synoviocyte function in vitro: implications in
rheumatoid arthritis. J Rheumatol 2000;27:1622–31.
Ham KD, Loeser RF, Lindgren BR, Carlson CS. Effects of
long-term estrogen replacement therapy on osteoarthritis severity
in cynomolgus monkeys. Arthritis Rheum 2002;46:1956–64.
Fernihough JK, Richmond RS, Carlson CS, Cherpes T, Holly JM,
Loeser RF. Estrogen replacement therapy modulation of the
insulin-like growth factor system in monkey knee joints. Arthritis
Rheum 1999;42:2103–11.
Blacklock AD, Johnson MS, Krizsan-Agbas D, Smith PG. Estrogen increases sensory nociceptor neuritogenesis in vitro by a
direct, nerve growth factor-independent mechanism. Eur J Neurosci 2005;21:2320–8.
Turner AS, Athanasiou KA, Zhu CF, Alvis MR, Bryant HU.
Biochemical effects of estrogen on articular cartilage in ovariectomized sheep. Osteoarthritis Cartilage 1997;5:63–9.
Wluka AE, Davis SR, Bailey M, Stuckey SL, Cicuttini FM. Users
of oestrogen replacement therapy have more knee cartilage than
non-users. Ann Rheum Dis 2001;60:332–6.
Hayami T, Pickarski M, Wesolowski GA, McLane J, Bone A,
Destefano J, et al. The role of subchondral bone remodeling in
osteoarthritis: reduction of cartilage degeneration and prevention
of osteophyte formation by alendronate in the rat anterior cruciate
ligament transection model. Arthritis Rheum 2004;50:1193–206.
Spector TD, Conaghan PG, Buckland-Wright JC, Garnero P,
Cline GA, Beary JF, et al. Effect of risedronate on joint structure
and symptoms of knee osteoarthritis: results of the BRISK randomized, controlled trial. Arthritis Res Ther 2005;7:R625–33.
APPENDIX A: WHI INVESTIGATORS
Program Office: Barbara Alving, Jacques Rossouw, Shari
Ludlam, Linda Pottern, Joan McGowan, Leslie Ford, Nancy Geller
(National Heart, Lung, and Blood Institute, Bethesda, MD).
Clinical Coordinating Centers: Ross Prentice, Garnet Anderson, Andrea LaCroix, Charles L. Kooperberg, Ruth E. Patterson,
Anne McTiernan (Fred Hutchinson Cancer Research Center, Seattle,
WA); Sally Shumaker (Wake Forest University School of Medicine,
Winston-Salem, NC); Evan Stein (Medical Research Labs, Highland
Heights, KY); Steven Cummings (University of California at San
Francisco).
Clinical Centers: Sylvia Wassertheil-Smoller (Albert Einstein
College of Medicine, Bronx, NY); Jennifer Hays (Baylor College of
Medicine, Houston, TX); JoAnn Manson (Brigham and Women’s
Hospital, Harvard Medical School, Boston, MA); Annlouise R. Assaf
(Brown University, Providence, RI); Lawrence Phillips (Emory University, Atlanta, GA); Shirley Beresford (Fred Hutchinson Cancer
Research Center, Seattle, WA); Judith Hsia (George Washington
University Medical Center, Washington, DC); Rowan Chlebowski
(Harbor–UCLA Research and Education Institute, Torrance, CA);
Evelyn Whitlock (Kaiser Permanente Center for Health Research,
3204
Portland, OR); Bette Caan (Kaiser Permanente Division of Research,
Oakland, CA); Jane Morley Kotchen (Medical College of Wisconsin,
Milwaukee); Barbara V. Howard (MedStar Research Institute/
Howard University, Washington, DC); Linda Van Horn (Northwestern University, Chicago/Evanston, IL); Henry Black (Rush Medical
Center, Chicago, IL); Marcia L. Stefanick (Stanford Prevention Research Center, Stanford, CA); Dorothy Lane (State University of New
York at Stony Brook); Rebecca Jackson (The Ohio State University,
Columbus); Cora E. Lewis (University of Alabama at Birmingham);
Tamsen Bassford (University of Arizona, Tucson/Phoenix); Jean
Wactawski-Wende (University at Buffalo, Buffalo, NY); John Robbins
(University of California at Davis, Sacramento); F. Allan Hubbell
(University of California at Irvine); Howard Judd (University of
California at Los Angeles); Robert D. Langer (University of California
at San Diego, La Jolla/Chula Vista); Margery Gass (University of
CIRILLO ET AL
Cincinnati, Cincinnati, OH); Marian Limacher (University of Florida,
Gainesville/Jacksonville); David Curb (University of Hawaii, Honolulu); Robert B. Wallace (University of Iowa, Iowa City/Davenport);
Judith Ockene (University of Massachusetts/Fallon Clinic, Worcester); Norman Lasser (University of Medicine and Dentistry of New
Jersey, Newark); Mary Jo O’Sullivan (University of Miami, Miami,
FL); Karen Margolis (University of Minnesota, Minneapolis); Robert
Brunner (University of Nevada, Reno); Gerardo Heiss (University of
North Carolina, Chapel Hill); Lewis Kuller (University of Pittsburgh,
Pittsburgh, PA); Karen C. Johnson (University of Tennessee, Memphis); Robert Brzyski (University of Texas Health Science Center, San
Antonio); Gloria E. Sarto (University of Wisconsin, Madison); Denise
Bonds (Wake Forest University School of Medicine, Winston-Salem,
NC); Susan Hendrix (Wayne State University School of Medicine/
Hutzel Hospital, Detroit, MI).
Документ
Категория
Без категории
Просмотров
1
Размер файла
116 Кб
Теги
effect, women, hip, health, joint, knee, replacement, risk, hormone, therapy, initiative
1/--страниц
Пожаловаться на содержимое документа