Breast Cancer ChoicesTM
|Scrutinizing the evidence for breast
cancer procedures and treatments
Improved breast cancer survival among hormone
replacement therapy users is durable after 5 years of
Dara Christante, M.D., SuEllen Pommier, Ph.D., Jennifer Garreau, M.D.,
Patrick Muller, B.S., Brett LaFleur, B.S., Rodney Pommier, M.D.
We previously reported that breast cancer patients who used hormone
replacement therapy (HRT) had significantly lower stage tumors and higher
survival than never-users. We present an update with longer follow-up,
HRT use data, and in vitro research.
Our database of 292 postmenopausal breast cancer patients was updated
to include HRT type, duration, and disease status. In vitro effects of
estrogen (E) and/or medroxyprogesterone (MPA) on breast cancer cell
growth were measured.
Tumor prognostic factors were better and survival rates higher for both E
and combination HRT users of any duration. Use greater than 10 years
correlated with node-negative disease, mammographically detected
tumors, and 100% survival. E supported minimal proliferation; MPA induced
cell death; E+MPA results were similar to E alone.
HRT users, regardless of type or duration of HRT use, continued to have
higher survival rates. In vitro results supported the clinical finding that
outcomes for users of E and E+MPA were similar.
We previously reported that breast cancers in hormone replacement
therapy (HRT) users were smaller, lower grade, more often node-negative,
lower stage, and had significantly higher survival rates compared to those
in never-users. Recent events have raised concerns about the impact of
HRT on breast cancer. Particular concern has been raised about the use of
combinations of estrogen (E) and medroxyprogesterone acetate (MPA).
Results from the Women's Health Initiative (WHI) trial indicated that breast
cancers were more advanced among users of E and MPA (combination
HRT) than among patients receiving placebo.2 This would be expected to
result in lower survival rates among users of combination HRT. Concerns
also exist about duration of HRT use and breast cancer.
Due to these concerns, we investigated if the higher survival rate we
reported was durable after an additional 5 years of follow-up. We also
investigated whether duration or type of HRT are associated with
differences in tumor characteristics or survival rates. We have
supplemented our clinical investigation with in vitro studies in which
estrogen receptor (ER)-positive and -negative breast cancer cell lines
were treated with various concentrations of E, MPA, or combinations of E
and MPA to determine their impacts on cellular proliferation.
Our database of 292 postmenopausal women diagnosed with breast cancer
at Oregon Health & Science University between March 1994 and January
2002 from our previous publication1 was updated by review of medical
records. New data included the current disease status of each patient as
well as the type and duration of HRT used. HRT was categorized as E alone,
progestin alone, E and progestin combination, or other HRT. If E and
progestin were ever taken by a patient, either serially or concomitantly, it
was considered combination HRT. Approval for this retrospective review
was obtained from the institutional review board.
Disease-specific survival curves were constructed using the Kaplan-Meier
method5 and statistical significance between survival distributions was
determined by log-rank analysis. Significance of differences between
groups of patients was determined by chi-square analysis, Student t test,
analysis of variance (ANOVA), Mann-Whitney U test, or Kruskal-Wallis H
test. Correlations between duration of HRT use and tumor size, number of
positive lymph nodes, and stage were determined using Pearson or
Spearman coefficients of correlation. The independent effects of
prognostic factors and HRT use on survival was determined by Cox
For all laboratory experiments, 3 cell lines, T-47D, HCC1954, and HCC1937,
were obtained from the ATCC (American Type Culture Collection,
Manassas, VA) and maintained according to ATCC protocol. T-47D cells are
ER-positive, progesterone receptor (PR)-positive; HCC1954 and HCC1937
cells are ER-negative, PR-negative; both are androgen receptor (AR)-
positive.6 Cell lines with these receptor profiles were selected for study
because our group of HRT users with a higher survival rate was comprised
of a mixture of patients with ER-positive and ER-negative tumors. Cells
were plated onto 96-well plates with 104 cells per well, grown in hormone-
depleted media for 3 days, and then treated with 17-β-estradiol (E) and
medroxyprogesterone-17-acetate (MPA) (Sigma-Aldrich, St Louis, MO).
The dose-dependent effects of MPA were initially tested by treating ER-
positive (T-47D) and ER-negative (HCC1937) breast cancer cells with a
range of MPA (.01, .1, 1, 10, 100, and 250 nmol/L). In all subsequent
experiments, T-47D and HCC1954 cells were treated with E alone at
concentrations of 1 and 10 nmol/L or MPA alone at concentrations of 1, 10,
100, and 250 nmol/L. Cells were also treated with 1 or 10 nmol/L E in
combination with 1, 10, 100, or 250 nmol/L MPA. For all experiments, cells
grown in hormone-depleted media served as an untreated control. Cell
proliferation was measured by 3-(4,5-dimethylthiazol-2-yl)-2,5-
diphenyltetrazolium bromide (MTT) assay and expressed as the relative
percent change between treated and untreated cultures.7 All experiments
were performed in triplicate wells and repeated a minimum of 3 times.
Statistical significance of differences in proliferation of cell cultures was
determined by ANOVA and independent t tests.
Of the 292 patients, 144 reported HRT use at the time of diagnosis or
anytime in the past. Seventy-three patients (51%) used E alone, 54 (38%)
used combination HRT, 3 (2%) used progestins alone, and 14 (9%) used
another type HRT (Table 1). There were no significant differences in the
types of HRT used between patients with ductal carcinoma in situ (DCIS) or
invasive cancer. There were no significant differences in the incidence of
ER-positive or PR-positive tumors between HRT users and never-users.
Data on HER2 status was not available for many patients whose diagnosis
preceded the advent of traztuzumab therapy.
Comparison of prognostic factors and survival rates between never-users
and HRT users by HRT type and duration
[editors note: see full text article for omitted charts]
Among patients with invasive cancers, mean tumor size, nodal status, ER
and PR status, the presence of distant metastases, stage, and the method
of detection did not vary significantly by types of HRT. Tumors of patients
using either E alone or combination HRT were smaller than those of never-
users (P = .022 and .003, respectively). There were significantly more T1 (≤2.
0 cm) tumors among patients using E (P = .036) and combination HRT (P = .
001) compared with never-users. Negative lymph nodes were significantly
more common in users of E (P = .045) and combination HRT (P = .023) than
in never-users. The mean number of positive lymph nodes was also
significantly lower in users of E (P = .03) and combination HRT (P = .02)
compared with never-users.
There were no significant differences in the distribution of stage based on
type of HRT use. Compared to never-users, users of E or combination HRT
had significantly more cancers lower than stage II (P = .01 for both). The
incidence of distant metastases did not differ between never-users and
HRT users, whether compared collectively or by type of HRT.
Data were available on the duration of HRT use for 133 of 144 users (92%)
and the median duration was 7 years. Forty-seven patients (33%) used HRT
for less than 5 years, 36 patients (25%) used HRT from 5 to 10 years, and 50
patients (35%) used HRT for more than 10 years. DCIS was evenly
distributed between groups of HRT duration. Tumor size, number of
positive lymph nodes, and stage did not increase with increasing
durations of HRT use, regardless of type. Lymph node status was more
often negative in patients with greater than 10 years of HRT use compared
to never-users (P = .023). E and combination HRT users were more often
diagnosed by mammography than palpation compared to never-users (P = .
036 and .01, respectively). Patients with any type of HRT use for more than
10 years were also significantly more likely to have their tumors detected
by mammography than by palpation compared to never-users (P = .01).
After an additional 5-years of follow-up, the disease-specific 5-year
survival rate of HRT users was 92% compared with 84% for never-users
(Figure 1, P = .02). No patient with a diagnosis of DCIS had died of breast
cancer. Among patients with invasive cancers, the 5-year survival rate of
HRT users was 91% compared with 83% for never-users (P = .04). There
were no significant differences in survival between patients grouped by
HRT type. The 5-year survival rates for users of E and combination HRT
were 89% and 95%, respectively. The difference in survival rates between
combination HRT users and never-users was statistically significant (P = .
04), but the difference in survival rates between users of E and never-
users was not significant. Survival curves by type of HRT are shown in
Figure 2 and compared with the curve for never-users.
Figure 1. Survival curves for breast cancer patients who were HRT users
(circles) compared with never-users (vertical dashes). The difference in
survival was statistically significant (P = .02).
Figure 2. Survival curves for HRT users of E (circles) and combination
therapy (triangles) compared with never-users (vertical dashes). The
difference in survival between patients who used combination therapy and
never-users was statistically significant (P = .04).
Survival curves of patients grouped by durations of HRT use of <5, 5–10,
and >10 years are compared to the survival curve of never-users in Figure
3. Patients with more than 10 years of use had a 100% survival rate
compared to 84% among never-users (P = .01).
Figure 3. Survival curves for breast cancer patients based on duration of
HRT use. Patients with varying durations of use <5 years (circles), 5–10
years (triangles), or >10 years (boxes) are compared to never-users
(vertical dashes). The difference between patients with >10 years use and
never-users was statistically significant (P = .01).
Among patients with tumors detected by palpation, there was no
significant difference in 5-year survival rates between never-users and
any group of HRT users (Table 1). Among patients with invasive tumors
detected by mammography, the 5-year survival rate for HRT users was
100% compared with 90% for never-users. (Figure 4, P = .03). To estimate
the magnitude of difference in survival that could possibly be attributed to
differences in adjuvant therapies between these 2 groups, data for each
never-user with a mammographically detected tumor were entered into
Adjuvant! Online, standard version 8.0.8 The maximum potential benefit of
adjuvant therapies was recorded for each patient. The calculated median
and mean decreases in survival had maximum adjuvant therapies been
completely withheld from all never-users with mammographically detected
tumors were 3.0% and 5.6%, respectively, at 10 years.
Figure 4. Survival curves for patients with breast cancers detected by
mammography comparing HRT users (circles) to never-users (vertical
dashes). The difference in survival between HRT users and never-users
was statistically significant (P = .02).
Cox regression analysis was performed using variables of tumor size,
tumor stage, nodal status, stage, mode of detection, receptor status, and
HRT use. The only variable that was a significant independent predictor of
survival was stage.
In vitro responses to treatment with E and MPA
Exposure of the T-47D cells to 1 nmol/L E demonstrated 23% and 25%
proliferation on days 5 and 8, respectively, compared to untreated cells.
Treatment with 10 nmol/L E was inhibitory; cell growth decreased by 5% and
15% by days 5 and 8, respectively. MPA treatment alone did not have a
proliferative effect at any concentration. Treatment with MPA and 1 nmol/L
E resulted in 15% and 36% growth on days 5 and 8, respectively, which was
not statistically different from the proliferation observed with 1 nmol/L E
alone (P = .13 and .43, respectively). Treatment with MPA plus 10 nmol/L E
resulted in significant cell death by day 8, an effect that increased with
higher concentrations of MPA. Growth inhibition reached 50% and 42% by
day 8 at concentrations of 100 nmol/L and 250 nmol/L MPA, respectively.
In HCC1954 cells, in contrast to what was observed in T-47D cells,
exposure to both concentrations of E was cytotoxic. However, similar to
what was observed with T-47D cells, MPA treatment alone did not induce
proliferation at any concentration. When cells were treated with E and MPA
in combination, some variation in response was observed. The addition of
MPA, at all concentrations, blunted the cytotoxic effect observed with 1
nmol/L E treatment alone but did not induce significant proliferation.
Treatment with MPA and 10 nmol/L E did not significantly alter the
cytotoxicity observed with 10 nmol/L E alone. The mean responses of both
T-47D and HCC1954 cells to the various hormonal treatments are
summarized in Figure 5. There were no significant differences between
the results seen with E alone or E+MPA.
Figure 5. Mean percentage change in proliferation of HCC1954 (black) and
T-47D (gray) cells compared to untreated cells after 8 days of treatment
with 17-β-estradiol (E), or medroxyprogesterone-17-acetate (MPA), or
E+MPA. *Statistically significant changes compared to the untreated group.
Differences between cells treated with E alone and E+MPA were not
Dose-dependent cell death was observed in the ER-negative and ER-
positive cell lines with increasing concentrations of MPA. By day 8, the
percentages of cell death were −48%, −57%, −69%, −70%, and −73% for ER-
negative cell line and −9%, −17%, −23%, −18%, and −16% for ER-positive cell
lines at MPA concentrations of .01, .1, 1, 10, and 100 nmol/L, respectively.
After an additional 5 years of follow-up, the survival rates for HRT users
were still significantly higher than for never-users. Shuetz et al. also
recently reported significantly higher 5-year survival rates of 93% for HRT
users compared to 82% for non-users. The 5-year survival rates for the 2
groups reported in that study were remarkably similar to the rates
reported in this study. In their study of 1,072 women, Shuetz et al also
found a decreased incidence of metastatic disease among HRT users
compared to non-users. The presence of distant metastases did not differ
between HRT users and never-users or among users of different types of
HRT in this study, but there were relatively few stage IV patients in the
In this study, the higher survival rate was observed chiefly among patients
with mammographically detected invasive tumors whose survival rate was
still 100%, as it was in our previous report. In contrast, the 5-year survival
rate for never-users with mammographically detected invasive tumors was
significantly lower at 90%. Additional follow-up has not shown a change
from our previous report in which there was no significant difference in
survival rates based on HRT use among patients with invasive tumors
detected by palpation.
Given that the higher survival rate was found among mammographically
detected tumors and that the frequency of screening mammography
between HRT users and never-users was equal,1 the difference in survival
is unlikely to be due to better screening of HRT patients. It is also unlikely
that the higher survival rate is due to differences in adjuvant therapies.
The calculated median and mean decreases in survival had maximum
adjuvant therapies been completely withheld from all never-users with
mammographically detected tumors were 3.0% and 5.6%, respectively, at 10
years. The observed difference in survival rates for mammographically
detected tumors in our series was 10% and it occurred at less than 4 years.
Thus, the difference in survival is greater in magnitude and occurs sooner
than would be expected from differences in administration of adjuvant
Tumor size,10 nodal status,10 and stage11 are strong independent
predictors of breast cancer survival in large databases. In our series, HRT
users had smaller tumors, more node-negative tumors, and lower stage
disease than never-users. Regression analysis revealed that only stage
was an independent predictor of survival in our database. This indicates
that the higher survival rate of HRT users is chiefly attributable to lower
stage disease resulting from combinations of small tumors and node-
Several investigators have reported tumor size to be significantly smaller
among HRT users12, 13, 14, 15, 16 but did not differentiate by HRT type.
Others have reported lower incidences of positive nodes among HRT
users.13, 15, 16 It is the report by Chlebowski et al2 from the WHI trial that
has been in contradiction to most published observational studies. They
reported more advanced tumors among users of combination HRT
compared to women taking placebo (patients in the placebo group were
not required to be never-users). Combination HRT users had a mean tumor
size of 1.7 cm and a 26% incidence of positive nodes of compared with a
mean tumors size of 1.5 cm and a 16% incidence of positive nodes in the
placebo group. Although the 2-mm difference in mean tumor sizes was
statistically significant, it may be of little clinical significance because both
would correspond to T1c tumors.11 The statement that tumors are more
advanced in combination HRT users is therefore derived primarily from the
10% higher incidence of positive nodes. There is a strong relationship
between mean tumor size and the incidence of positive nodes, so this
large shift in the incidence of positive nodes with only a 2-mm difference
in tumor size is difficult to explain. The interpretation of the data has been
that the placebo group is normal and, by comparing the combination HRT
group to them, we must conclude that HRT made the tumors more
advanced for their size than they otherwise would have been. This
conclusion appears strengthened by the fact that it is derived from the
only randomized prospective data available.
The validity of this conclusion can be tested against other databases.
Based on data from 257,888 breast cancer patients in the Surveillance
Epidemiology and End Results (SEER) database, the correlation between
mean primary tumor size and the percentage of patients with positive
lymph nodes has been accurately described by the equation: percentage
of patients with positive lymph nodes = 65.6107594/1+8.336998433e−0.
1005638804(tumor size in mm).10 If one enters the mean tumor size of 17
mm for the combination HRT group into this equation, then the expected
incidence of positive lymph nodes is 26%, which is precisely what was
observed. This would indicate that the tumors in the combination HRT
group were not more advanced for their size than other breast cancers. In
contrast, if one enters the mean tumor size of 15 mm for the placebo group
into the equation, then the expected incidence of positive lymph nodes is
23%. This does not agree with the observed rate of 16%. The data point for
the placebo group falls well below the curve described by the equation. It
is not plausible that placebo made tumors less aggressive than expected,
so this indicates that the placebo group may not have been a typical
sample of breast cancer patients. In light of the SEER equation, an
alternative explanation of the WHI data is that the combination HRT group
had typical breast cancers and the placebo group, for some reason, had an
unusually low incidence of positive nodes. This had the effect of making
the combination HRT group look worse than it actually was. The observed
incidences of positive nodes for our patient groups were in close
agreement with the predictions of the SEER equation, indicating that we
had typical samples of breast cancer patients in our series.
Data on the impact of the duration of use of HRT on breast cancer
prognostic factors and survival are limited, but no differences have been
reported.16 In our study, duration of use did not worsen any prognostic
factor regardless of whether patients were considered together or
grouped by type of HRT used. When patients were arranged in groups by
durations of use, those with greater than 10 years of use were significantly
more likely to have negative nodes and to have tumors that were detected
Although there were no statistically significant differences in survival
rates among HRT users based on the type of HRT used, patients who used
combination HRT had significantly higher survival rates than never-users.
Users of E HRT also had higher survival rates than never-users, but the
difference did not achieve statistical significance.
Tissue culture results showed that 1 nmol/L E alone resulted in significant
proliferation of ER-positive cells. In contrast, 10 nmol/L E alone
consistently resulted in significant decreased proliferation of ER-positive
cells. We have previously reported that concentrations between 1 nmol/L
and 10 nmol/L support proliferation of ER-positive cell lines. Interestingly,
in cultured breast cancer cells that have acquired anti-estrogen and
aromatase inhibitor resistance, unexpected cytotoxicity has been
observed in cells with normally proliferative concentrations of E, which
may explain the observed results.17, 18 As expected, significant
proliferation of ER-negative cells was not induced by treatment with E
No concentration of MPA induced a proliferative effect in any of the cell
lines. When cells were treated with combinations of MPA and E, the net
effects were no different than when they were treated with E alone. For
example, when ER-positive cells demonstrated proliferation with 1 nmol/L
E, the addition of MPA did not alter this proliferative effect. Similarly, when
E treatment induced growth inhibition in either ER-positive or ER-negative
cells, the addition of MPA did not alter this effect to the point where
significant cellular proliferation was observed. If E induced growth
inhibition in either ER-positive or ER-negative cells, the addition of MPA
never resulted in significant stimulation of proliferation. Our data indicate
that observed net effects on growth were driven by the concentration of E
used and not by MPA. Maximum serum concentrations of MPA in patients
are dose-dependent and range from 1.2 pg/mL to 4.8 pg/mL (.003–.01
nmol/L).19 In this study, the concentrations of MPA used for in vitro testing
were within and above the range measured in patients.
A limitation to these experiments is that the only effect of hormonal
treatment that was measured was cellular proliferation. The effects of
hormones on numerous other prognostic factors such as cellular
differentiation, potential to metastasize to lymph nodes or distant organs,
and interactions between tumor cells and supporting tissues were not
tested in this in vitro model. The impact of hormones on these other
factors was derived from the clinical outcomes measured in the study.
Even the finding that E can support proliferation of ER-positive cells must
be viewed in the context of an in vitro model. In this environment, cells
are viable for only a short period of time when grown under minimal
conditions that are supplemented with E only. Extending proliferation for
longer periods requires the addition of growth factors and multiple other
hormones to support the growth of breast cancer cells over time.
However, the results of our in vitro studies support our clinical findings
that the addition of MPA to E does not produce outcomes significantly
different than those observed with E alone.
In this study, we found that the higher survival rates of breast cancer
patients who used HRT were durable after an additional 5 years of follow-
up. Higher survival rates of HRT patients were found chiefly among
patients with mammographically detected tumors. Both E and combination
HRT use were associated with smaller tumors, a lower incidence of
positive lymph nodes, and lower stages of breast cancer. Greater than 10
years of HRT use was also associated with more mammographically
detected tumors and a 100% survival rate. Users of combination HRT did
not have more advanced tumors than users of E alone or never-users.
These clinical findings are supported by in vitro experiments in which the
addition of MPA to E did not alter the effects observed with E alone.
1. Cheek J, Lacy J, Toth-Fejel S, et al. The impact of hormone replacement therapy on the detection and stage
of breast cancer. Arch Surg. 2002;137:1015–1019. MEDLINE | CrossRef
2. Chlebowski RT, Hendrix SL, Langer RD, et al. Influence of estrogen plus progestin on breast cancer and
mammography in healthy postmenopausal women: the Women's Health Initiative randomized trial. JAMA. 2003;
3. Collaborative Group on Hormonal Factors in Breast Cancer. Breast cancer and hormone replacement
therapy: collaborative reanalysis of data from 51 epidemiological studies of 52,705 women with breast cancer
and 108,411 women without breast cancer. Lancet. 1997;350:1047–1059. Abstract | Full Text | Full-Text PDF
(425 KB) | MEDLINE | CrossRef
4. Anderson GL, Chlebowski RT, Rossouw JE, et al. Prior hormone therapy and breast cancer risk in the Women's
Health Initiative randomized trial of estrogen plus progestin. Maturitas. 2006;55:103–115. Abstract | Full Text |
Full-Text PDF (401 KB) | MEDLINE | CrossRef
5. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. Am J Statistics. 1958;53:457–
6. Hardin C, Pommier R, Calhoun K, et al. A new hormonal therapy for estrogen receptor-negative breast
cancer. World J Surg. 2007;31:1041–1046. MEDLINE | CrossRef
7. Mosmann T. Rapid colorimetric assay for cellular growth and survival: application to proliferation and
cytotoxicity assays. J Immunol Methods. 1983;65:55–63. MEDLINE | CrossRef
8. Adjuvant! Online. Accessed May 26, 2008. http://www.adjuvantonline.com.
9. Schuetz F, Diel IJ, Pueschel M, et al. Reduced incidence of distant metastases and lower mortality in 1072
patients with breast cancer with a history of hormone replacement therapy. Am J Obstet Gynecol. 2007;196:
10. In: Ries LAG, Young JL, Keel GE, et al. editor. SEER Survival Monograph: Cancer Survival Among
Andults: U.S. SEER Program, 1988–2001, Patient and Tumor Characteristics. Bethesda, MD: National Institutes
of Health; 2007;.
11. Bonnier P, Bessenay F, Sasco AJ, et al. Impact of menopausal hormone-replacement therapy on clinical
and laboratory characteristics of breast cancer. Int J Cancer. 1998;79:278–282. MEDLINE | CrossRef
12. Manjer J, Malina J, Berglund G, et al. Increased incidence of small and well-differentiated breast tumours
in post-menopausal women following hormone-replacement therapy. Int J Cancer. 2001;92:919–922. MEDLINE
13. Biglia N, Sgro L, Defabiani E, et al. The influence of hormone replacement therapy on the pathology of
breast cancer. Eur J Surg Oncol. 2005;31:467–472. Abstract | Full Text | Full-Text PDF (101 KB) | MEDLINE |
14. Sacchini V, Zurrida S, Andreoni G, et al. Pathologic and biological prognostic factors of breast cancers in
short- and long-term hormone replacement therapy users. Ann Surg Oncol. 2002;9:266–271. MEDLINE |
15. In: Greene FL, Page D, Fleming ID editor. AJCC Cancer Staging Manual. Ed 6. New York, NY: Springer;
16. Antoine C, Liebens F, Carly B, et al. Influence of HRT on prognostic factors for breast cancer: a systematic
review after the Women's Health Initiative trial. Hum Reprod. 2004;19:741–756. MEDLINE | CrossRef
17. Planas-Silva MD, Waltz PK, Kilker RL. Estrogen induces death of tamoxifen-resistant MCF-7 cells:
contrasting effect of the estrogen receptor downregulator fulvestrant. J Steroid Biochem Mol Biol. 2006;98:193–
198. MEDLINE | CrossRef
18. Lewis JS, Osipo C, Meeke K, et al. Estrogen-induced apoptosis in a breast cancer model resistant to long-
term estrogen withdrawal. J Steroid Biochem Mol Biol. 2005;94:131–141. MEDLINE | CrossRef
19. Prempro. Philadelphia, PA: Wyeth; 2007;[package insert].
Please consult your licensed health care practitioner before implementing any health strategy.
The information provided on this site is designed to support, not replace, the relationship that
exists between a patient/site visitor and his/her existing physician. This site accepts no
advertising. The contents of this site are copyrighted 2004-2013 by Breast Cancer Choices, Inc.
Contact us with comments or for reprint permission at email@example.com
Web page updated January 2013.
In this study, the
rate was still
100%, as it was in
In contrast, the
lower at 90%.
|HRT takers' survival at 5 years: