Untitled

ORIGINAL CONTRIBUTION Health Outcomes After Stopping Conjugated
Equine Estrogens Among Postmenopausal
Women With Prior Hysterectomy
A Randomized Controlled Trial
Andrea Z. LaCroix, PhD Context The Women's Health Initiative Estrogen-Alone Trial was stopped early af-
Rowan T. Chlebowski, MD, PhD ter a mean of 7.1 years of follow-up because of an increased risk of stroke and little JoAnn E. Manson, MD, DrPH likelihood of altering the balance of risk to benefit by the planned trial terminationdate. Postintervention health outcomes have not been reported.
Aaron K. Aragaki, MS Objective To examine health outcomes associated with randomization to treat-
Karen C. Johnson, MD, MPH ment with conjugated equine estrogens (CEE) among women with prior hysterec- tomy after a mean of 10.7 years of follow-up through August 2009.
Karen L. Margolis, MD, MPH Design, Setting, and Participants The intervention phase was a double-blind,
placebo-controlled, randomized clinical trial of 0.625 mg/d of CEE compared with pla-
Marcia L. Stefanick, PhD cebo in 10 739 US postmenopausal women aged 50 to 79 years with prior hysterec- Robert Brzyski, MD, PhD tomy. Follow-up continued after the planned trial completion date among 7645 sur- J. David Curb, MD, MPH viving participants (78%) who provided written consent.
Barbara V. Howard, PhD Main Outcome Measures The primary outcomes were coronary heart disease (CHD)
and invasive breast cancer. A global index of risks and benefits included these primary
Cora E. Lewis, MD, MSPH outcomes plus stroke, pulmonary embolism, colorectal cancer, hip fracture, and death.
Jean Wactawski-Wende, PhD Results The postintervention risk (annualized rate) for CHD among women assigned
for the WHI Investigators to CEE was 0.64% compared with 0.67% in the placebo group (hazard ratio [HR], 0.97;95% confidence interval [CI], 0.75-1.25), 0.26% vs 0.34%, respectively, for breast can- tive (WHI) Estrogen-Alone (HR, 1.00; 95% CI, 0.84-1.18). The risk of stroke was no longer elevated during the post-intervention follow-up period and was 0.36% among women receiving CEE compared Trial was a double-blind, pla- with 0.41% in the placebo group (HR, 0.89; 95% CI, 0.64-1.24), the risk of deep vein thrombosis was lower at 0.17% vs 0.27%, respectively (HR, 0.63; 95% CI, 0.41-0.98), clinical trial evaluating the effects of and the risk of hip fracture did not differ significantly and was 0.36% vs 0.28%, respec- conjugated equine estrogens (CEE) on tively (HR, 1.27; 95% CI, 0.88-1.82). Over the entire follow-up, lower breast cancer inci- chronic disease incidence among post- dence in the CEE group persisted and was 0.27% compared with 0.35% in the placebo menopausal women with prior hyster- group (HR, 0.77; 95% CI, 0.62-0.95). Health outcomes were more favorable for younger ectomy. The trial intervention was compared with older women for CHD (P=.05 for interaction), total myocardial infarction stopped 1 year early after a mean of 7.1 (P=.007 for interaction), colorectal cancer (P=.04 for interaction), total mortality (P=.04 years of follow-up because of an in- for interaction), and global index of chronic diseases (P=.009 for interaction).
creased risk of stroke and little likeli- Conclusions Among postmenopausal women with prior hysterectomy followed up
hood of altering the balance of risk to for 10.7 years, CEE use for a median of 5.9 years was not associated with an increased benefit by the planned termination date.
or decreased risk of CHD, deep vein thrombosis, stroke, hip fracture, colorectal can-cer, or total mortality. A decreased risk of breast cancer persisted.
Analyses of outcomes during the inter-vention period suggested that treat- Trial Registration clinicaltrials.gov Identifier: NCT00000611
ment effects differed by age; com- pared with older women, youngerwomen receiving CEE had a lower risk of coronary heart disease (CHD), co- Author Affiliations are listed at the end of this article.
lorectal cancer, total death, and the Corresponding Author: Andrea Z. LaCroix, PhD, Fred
global index of chronic diseases.1 How- Hutchinson Cancer Research Center, 1100 Fairview For editorial comment see p 1354.
Ave N, M3-A410, PO 19024, Seattle, WA 98109 ever, the tests for interaction of age with 2011 American Medical Association. All rights reserved.
(Reprinted) JAMA, April 6, 2011—Vol 305, No. 13 1305
Downloaded From: http://jama.jamanetwork.com/ on 07/19/2013
HEALTH OUTCOMES AFTER STOPPING ESTROGEN THERAPY treatment were only statistically sig- fied suggestions of age-specific differ- survival. Women were excluded if they nificant for colorectal cancer.1 ences in effects of CEE on health out- had prior breast cancer or other can- All previous reports of this trial were comes persisted after stopping the cer within 10 years (except non– limited to outcomes occurring during melanoma skin cancer), or prior ve- the intervention phase. Herein, we re- nous thromboembolism (if screened port data on postintervention out- after 1997). The study protocol was ap- comes through a mean of 10.7 years of proved by institutional review boards follow-up. This preplanned analysis had Details of the WHI Estrogen-Alone Trial at the participating institutions and all 3 objectives: (1) to assess the long- have been published.1,2 Briefly, post- participants provided written in- term effects of the CEE intervention on menopausal women aged 50 to 79 years formed consent.
health outcomes; (2) to determine were recruited at 40 US clinical cen- A total of 10 739 women were ran- whether effects of CEE on health out- ters between 1993 and 1998. Women domly assigned to receive orally either comes differed between the interven- were eligible if they had a prior hyster- 0.625 mg/d of CEE (Premarin, Wyeth tion and postintervention periods; and ectomy, were not taking hormone Ayerst, Philadelphia, Pennsylvania) or (3) to determine if previously identi- therapy, and had an anticipated 3-year matching placebo. Randomization wasimplemented at the WHI Clinical Co-ordinating Center using a permuted Figure 1. Women's Health Initiative Hormone Therapy Estrogen-Alone Trial Through
block algorithm, stratified by clinical Extended Follow-up center and age group.1 The clinical trialtarget size of 12 375 was calculated to 373 092 Women initiated screening provide 81% power to detect a 21% re-duction in CHD at 9 years of follow- 11 941 Provided consent and reported having had a hysterectomy up. With the actual randomized samplesize, the power estimate was 72% for a 10 739 Randomized 21% reduction in CHD.
INTERVENTION PHASEDecember 1, 1993, to February 29, 2004 When the intervention phase ended after a mean of 7.1 years on February 5310 Assigned to receive 5429 Assigned to receive conjugated equine 29, 2004, vital status was known for 95% of participants, of whom 5.4% were POSTINTERVENTION PHASE deceased. By this time, 54% of partici- March 1, 2004, to March 31, 2005 pants had stopped taking their studymedication. Median time receiving 4794 Had any postintervention 4877 Had any postintervention treatment was 5.9 years in the CEE group vs 5.8 years in the placebo group (interquartile range, 2.5-7.3 years). The EXTENSION PHASEApril 1, 2005, to August 14, 2009 median adherent time receiving treat-ment (ingestion of ⬎80% of study pills) 459 Not eligible for extension 494 Not eligible for extension was 3.5 years in both groups (inter- quartile range, 1.5-6.5 years).
Clinical outcomes were collected 4851 Eligible to participate in 4935 Eligible to participate in through semiannual mailed question- naires and annual clinic visits. Out- 1073 Did not consent to extension 1068 Did not consent to extension comes were verified3 initially by trained phase participation phase participation physician adjudicators at the local clini- cal centers by medical record review, 105 Not approached 112 Not approached followed by final adjudication at theWHI Clinical Coordinating Center. All 3778 Consented to participate 3867 Consented to participate adjudicators were blinded to treat- in extension phase in extension phase ment assignment.
5310 Included in analysis 5429 Included in analysis Demographic characteristics and medical history were collected by self- The intervention phase ran from December 1, 1993, to February 29, 2004. The postintervention phase began report using standardized question- on March 1, 2004, the day participants were instructed to stop study medication use (conjugated equine es-trogens or placebo) and continued through the original trial completion date (March 31, 2005). The postinter- naires. Race/ethnicity was reported by vention extension phase began on April 1, 2005, and includes follow-up for participants who provided addi- participants within predefined catego- tional consent (78% of those eligible) through August 14, 2009.
ries matching the US Census. This in- 1306 JAMA, April 6, 2011—Vol 305, No. 13 (Reprinted)
2011 American Medical Association. All rights reserved.
Downloaded From: http://jama.jamanetwork.com/ on 07/19/2013
HEALTH OUTCOMES AFTER STOPPING ESTROGEN THERAPY formation was required by the fund- quintiles of intended duration of inter- of their first relevant clinical event, or ing agency to monitor nonwhite vention (ie, elapsed time from ran- the date of death or withdrawal from representation in the trial.
domization until the intervention the study (whichever came first). For- ended on February 29, 2004). The mal tests of the differences between hazard ratios (HRs) were estimated the HRs in the intervention compared using Cox proportional hazards mod- with the postintervention phase were The postintervention period began on els5 stratified by age, prior disease (if calculated by inclusion of a binary March 1, 2004, when participants were appropriate), and randomization sta- instructed to discontinue taking the tus in the WHI Dietary Modification dependent variable as described.4 study pills. The current report reflects Trial.6 Models were constructed for Absolute rates and attributable risks a mean (SD) postintervention fol- each clinical end point in which (rate differences between CEE and low-up duration of 47.2 (20.7) months women contributed follow-up time placebo groups) also were calculated.
through August 14, 2009. After the pro- until the end of the interval, the date All statistical tests were 2-sided.
tocol-specified termination date ofMarch 31, 2005, subsequent partici- Table. Baseline Characteristicsa
pant follow-up required additional writ- No. (%) of Participants
ten consent, which was obtained from77.9% of surviving participants in the (n = 3778)
(n = 3867)
CEE group (n=3778) and 78.4% of sur- Age group at screening, y viving participants in the placebo group (n = 3867). The outcomes identified from the annual mailed question- naires were verified by medical record review as described.3 Annual mammo- grams were encouraged and tracked by annual mammography report review.
During the postintervention period, 3.6% to 4.7% of women from the CEE Asian/Pacific Islander group and 2.7% to 3.0% of women from the placebo group reported estrogen- Hormone therapy use alone use (any route of administra- tion) on annual questionnaires.
Duration of hormone therapy use, y The primary analyses included all ran- domized participants using time-to- event methods and were based on the intention-to-treat principle as de- scribed previously.4 Thus, all partici- pants were included in the analyses ac- cording to their randomized group assignment until they provided their last follow-up information (FIGURE 1).
Baseline characteristics for women who provided additional consent were com- Never pregnant (no term pregnancy) pared by randomization group using ␹2 ⱖ1 term pregnancy and t test statistics.
Age at first birth, y Annualized rates of clinical events were estimated for the intervention period, the postintervention period, Hysterectomy age group, y and the entire follow-up period by dividing the number of events by the corresponding person-time in each phase. Cumulative incidence curves were drawn for each trial phase with 2011 American Medical Association. All rights reserved.
(Reprinted) JAMA, April 6, 2011—Vol 305, No. 13 1307
Downloaded From: http://jama.jamanetwork.com/ on 07/19/2013
HEALTH OUTCOMES AFTER STOPPING ESTROGEN THERAPY tion, and overall follow-up periods are Table. Baseline Characteristicsa (continued)
summarized in FIGURE 2 and eTable 2
No. (%) of Participants
at http://www.jama.com. The HRs for CHD during the postintervention fol- (n = 3778)
(n = 3867)
low-up period were close to unity and similar to those observed during the in- Bilateral oophorectomy tervention (Figure 2 and FIGURE 3). The
Treated diabetes (pills or injections) increased stroke risk seen during the in- or high blood pressure) tervention phase was not present dur- High cholesterol (requiring pills) ing the postintervention phase (0.36% [n = 66] in the CEE group vs 0.41% Aspirin use ⱖ80 mg for ⱖ30 d [n = 77] in the placebo group; HR, 0.89 [95% confidence interval {CI}, 0.64- 1.24]; P = .05 for difference). Simi- larly, the increase in deep vein throm- bosis and pulmonary embolism with Fracture and age ⱖ55 y CEE use compared with placebo dur- No. of times fell in last 12 mo ing the intervention phase was not maintained during the postinterven- tion phase (0.28% [n = 52] vs 0.39% [n=74], respectively; HR, 0.72 [95% CI, 0.51-1.03]). For all cardiovascular Abbreviations: BMI, body mass index; CABG, coronary artery bypass graft; CEE, conjugated equine estrogens; DVT, deep vein thrombosis; PE, pulmonary embolism; PTCA, percutaneous transluminal coronary angioplasty.
events, the cumulative HR associated a This table contains data from Women's Health Initiative participants who consented to extended follow-up after en- rollment in the Estrogen-Alone Trial (April 2005).
with CEE use was 1.06 (95% CI, 0.98- b Test of association.
1.15) (2.26% in the CEE group Calculated as weight in kilograms divided by height in meters squared.
[n = 1146] vs 2.12% in the placebogroup [n = 1113]; Figure 2).
Nominal P values are reported with- (SAS Institute Inc, Cary, North Caro- During the postintervention phase, out adjustment for multiple outcomes lina) and R software version 2.11 (R 81.2% of women in the CEE group and or sequential looks during the clinical Foundation for Statistical Computing, 81.3% of women in the placebo group trial follow-up period. Age-stratified had at least 1 mammogram. The HRs subgroup analyses are reported for 10 comparing rates of invasive breast can- outcomes. At the .05 level of signifi- cer in women randomized to CEE vs cance, 0 to 1 interaction P values could placebo were similar during the inter- be statistically significant based on Participant movement through the vention (HR, 0.79; 95% CI, 0.61-1.02) chance alone.
study is outlined in Figure 1. Among and postintervention phases (HR, 0.75; To determine whether not provid- the women who provided additional 95% CI, 0.51-1.09) (Figure 2 and ing consent to postintervention fol- consent, baseline characteristics re- Figure 3). Consequently, a statisti- low-up influenced risk estimates, in- mained similar to those previously pub- cally significant lower cumulative breast verse-probability weighting analyses lished1 and were evenly distributed by cancer incidence of 0.27% was seen in were conducted using the methods de- randomized treatment assignment the CEE group (n=151) compared with scribed.4 Adherence sensitivity analy- (TABLE). Small differences were ob-
0.35% in the placebo group (n = 199) ses also were conducted by censoring served for parity and bilateral oopho- (HR, 0.77 [95% CI, 0.62-0.95]; P=.02).
follow-up at 6 months after partici- rectomy between randomization Colorectal cancer incidence did not dif- pants became nonadherent (ingestion groups. A comparison of the percent- fer between the women in the CEE of ⬍80% of study pills or starting non- age of trial participants who con- group and the placebo group during the protocol hormone therapy). For these sented to additional follow-up by treat- intervention or postintervention peri- analyses, participants who provided ad- ment group is provided in eTable 1 at ods (Figure 2 and FIGURE 4).
ditional consent or were adherent were The reduced hip fracture risk seen included in analyses that used the in- during the intervention phase with CEE verse of the participant's estimated re- Comparison of Intervention
was not maintained in the postinter- consent or adherence probability as a and Postintervention Findings
vention phase (0.36% in the CEE group weighting factor.
Incident clinical events by randomiza- [n = 66] vs 0.28% in the placebo group All statistical analyses were con- tion assignment and corresponding [n = 53]) (HR, 1.27 [95% CI, 0.88- ducted using SAS software version 9.2 HRs for the intervention, postinterven- 1.82]; P = .01 for difference; Figure 2) 1308 JAMA, April 6, 2011—Vol 305, No. 13 (Reprinted)
2011 American Medical Association. All rights reserved.
Downloaded From: http://jama.jamanetwork.com/ on 07/19/2013
HEALTH OUTCOMES AFTER STOPPING ESTROGEN THERAPY resulting in an overall HR of 0.92 (95% [n = 127]). During the postinterven- CI, 0.71-1.18]; 0.20% in the CEE group tion phase, hip fracture incidence was (Figure 4). Randomization to CEE did [n=114] vs 0.22% in the placebo group slightly higher in the CEE group com- not influence total mortality or the Figure 2. Effects of Conjugated Equine Estrogens (CEE) Compared With Placebo on Clinical Outcomes During the Intervention and
Postintervention Phases in the Women's Health Initiative Estrogen-Alone Trial
No. (%) of Events Cardiovascular outcomes Deep vein thrombosis (DVT) Pulmonary embolism (PE) All cardiovascular events Invasive breast cancer Colorectal cancer Death (all causes) The hazard ratios (HRs) are derived from proportional hazards models stratified by prior disease (for outcomes in which women were eligible for enrollment with andwithout the prevalent condition), age, and dietary modification randomization group. The P values for differences between the intervention and postintervention phaseswere calculated from models for the overall mean follow-up period that also included a time-dependent term for trial phase. For the intervention and overall phases,time to event equals 0 on date of randomization. For the postintervention phase, time to event equals 0 on February 29, 2004. CABG indicates coronary artery bypassgraft; CHD, coronary heart disease; CI, confidence interval; MI, myocardial infarction; and PTCA, percutaneous transluminal coronary angioplasty.
2011 American Medical Association. All rights reserved.
(Reprinted) JAMA, April 6, 2011—Vol 305, No. 13 1309
Downloaded From: http://jama.jamanetwork.com/ on 07/19/2013
HEALTH OUTCOMES AFTER STOPPING ESTROGEN THERAPY global index of chronic diseases either [n = 126] vs 0.48% [n = 124], respec- (P = .96 for interaction). The previ- during the intervention phase or dur- tively) for women aged 60 to 69 years; ously observed age interaction for co- ing the postintervention phase (Figure 2 and 1.23 (95% CI, 0.92-1.65; 0.82% lorectal cancer was significant through- and Figure 4).
[n = 101] vs 0.66% [n = 84], respec- out the entire follow-up period. Women tively) for women aged 70 to 79 years aged 70 to 79 years at entry experi- (P=.007 for interaction). A similar pat- enced a nearly 2-fold increased risk of The age-specific intervention results for tern was seen when time since meno- colorectal cancer in the CEE group a mean follow-up of 10.7 years are dis- pause (as previously defined7) instead (0.30% [n = 38] vs 0.16% in the pla- played in FIGURE 5. The overall HRs for
of age was examined for both coro- cebo group [n=21]) (HR, 1.83 [95% CI, CHD differed among women aged 50 nary end points (data not shown). Over- 1.08-3.12]; P = .04 for interaction).
to 59 years (HR, 0.59 [95% CI, 0.38- all, stroke risks were nonsignificantly The HRs for total mortality and the 0.90]; 0.18% [n=33] in the CEE group elevated for all age groups (P = .91 for global index of chronic diseases dif- vs 0.31% in the placebo group [n=56]) interaction). For deep vein thrombo- fered by age as previously suggested.7 compared with older women in which sis and pulmonary embolism, no age- Younger postmenopausal women (aged the HRs were near unity (P = .05 for in- specific differences emerged but the in- 50-59 years) who were randomized to teraction). For total myocardial infarc- creased risks observed during the CEE vs placebo had a lower risk of tion (MI), the HR was 0.54 (95% CI, intervention phase subsided during the death (0.35% [n=65] vs 0.48% [n=89], 0.34-0.86; 0.15% in the CEE group respectively; HR, 0.73 [95% CI, 0.53- [n = 27] vs 0.27% in the placebo group There were fewer invasive breast can- 1.00]) compared with no increased risk [n=50]) for women aged 50 to 59 years; cers in the CEE group compared with among women in their 60s (1.00% 1.05 (95% CI, 0.82-1.35; 0.51% the placebo group in all 3 age groups [n = 254] vs 0.96% [n = 253], respec- Figure 3. Cumulative Incidence of Coronary Heart Disease, Stroke, Pulmonary Embolism, and Invasive Breast Cancer
Coronary heart disease 5310 5153 4998 4809 4228 3579 1672 5310 5154 5000 4822 4236 3611 1693 Placebo 5429 5269 5101 4915 4287 3667 1765 Placebo 5429 5281 5112 4938 4337 3730 1806 Pulmonary embolism Invasive breast cancer Cumulative Hazar 0.01 5310 5176 5040 4889 4304 3666 1718 5310 5166 5007 4840 4261 3620 1696 Placebo 5429 5300 5156 4997 4393 3778 1827 Placebo 5429 5280 5106 4915 4301 3678 1771 Vertical dotted lines represent quintiles of duration of intended intervention and follow-up in the study population (elapsed time from randomization until the end ofthe intervention on February 29, 2004). CEE indicates conjugated equine estrogens.
a Includes events from randomization to August 14, 2009.
b Includes events from March 1, 2004, to August 14, 2009.
1310 JAMA, April 6, 2011—Vol 305, No. 13 (Reprinted)
2011 American Medical Association. All rights reserved.
Downloaded From: http://jama.jamanetwork.com/ on 07/19/2013
HEALTH OUTCOMES AFTER STOPPING ESTROGEN THERAPY tively; HR, 1.04 [95% CI, 0.88-1.24]), women aged 50 to 59 years who re- tually identical to those described and a slight increased risk of death ceived CEE compared with women who herein with P values for interaction re- among women in their 70s (2.02% received placebo had 12 fewer acute flecting some loss of precision with the [n = 258] vs 1.83% [n = 239], respec- MIs, 13 fewer deaths, and 18 fewer ad- inverse-probability weights: CHD tively; HR, 1.12 [95% CI, 0.94-1.33]; verse events in the global index of (P = .23); total MI (P = .01); colorectal P = .04 for interaction). A similar pat- chronic diseases. In contrast, women cancer (P = .09); death (P = .13); and tern was observed by age for women aged 70 to 79 years who received CEE global index of chronic diseases randomized to CEE vs placebo for the compared with women who received (P = .02). In each case, women in their global index of chronic diseases with placebo had 16 excess MIs, 19 excess 50s had more favorable HRs than older a possible overall benefit among deaths, and 48 excess adverse events in women (aged 70-79 years).
younger women (aged 50-59 years: the global index of chronic diseases.
The results also were similar when 1.04% [n = 184] vs 1.22% [n = 217], re- women were censored 6 months after spectively; HR, 0.85 [95% CI, 0.70- becoming nonadherent to study medi- 1.03]) and possible harm among the The results were similar when using in- cation during the intervention period.
oldest women (aged 70-79 years: 4.04% verse-probability weighting to ac- Adherence-adjusted HRs for the over- [n = 466] vs 3.56% [n = 423], respec- count for censoring due to those not all follow-up period using inverse- tively; HR, 1.15 [95% CI, 1.01-1.32]; providing consent for postinterven- probability weighting showed an in- P = .009 for interaction).
tion follow-up. The HR for breast creased risk of stroke with CEE use Expressed as absolute rates per cancer for the cumulative follow-up (HR, 1.50; 95% CI, 1.11-2.05) and a 10 000 women annualized over the av- period became 0.81 (95% CI, 0.64- lower risk of breast cancer (HR, 0.68; erage follow-up period of 10.7 years, 1.01). Age-stratified results were vir- 95% CI, 0.49-0.95). No significant age Figure 4. Cumulative Incidence of Colorectal Cancer, Hip Fracture, Death, and Global Index of Chronic Diseases
Colorectal cancer 5310 5169 5028 4882 4309 3672 1722 5310 5180 5043 4888 4307 3656 1697 Placebo 5429 5287 5136 4983 4380 3766 1826 Placebo 5429 5295 5141 4978 4363 3749 1819 5310 5186 5056 4916 4345 3708 1744 5310 5071 4846 4577 3950 3297 1505 Placebo 5429 5305 5167 5018 4420 3809 1848 Placebo 5429 5191 4931 4662 3996 3356 1599 Vertical dotted lines represent quintiles of duration of intended intervention and follow-up in the study population (elapsed time from randomization until the end ofthe intervention on February 29, 2004). CEE indicates conjugated equine estrogens.
a Includes events from randomization to August 14, 2009.
b Includes events from March 1, 2004, to August 14, 2009.
2011 American Medical Association. All rights reserved.
(Reprinted) JAMA, April 6, 2011—Vol 305, No. 13 1311
Downloaded From: http://jama.jamanetwork.com/ on 07/19/2013
HEALTH OUTCOMES AFTER STOPPING ESTROGEN THERAPY interactions emerged for any outcome and venous thromboembolism seen tial harm among older women, were ob- in the adherence-adjusted analyses; among women randomized to CEE dur- served for CHD, total MI, colorectal however, power was limited due to sub- ing the intervention period rapidly dis- cancer, total mortality, and the global stantial censoring.
sipated during the postintervention pe- index of chronic diseases.
riod as did the protective effect on risk The statistically significant reduc- of hip fracture. The lower incidence of tion in breast cancer incidence seen breast cancer seen among women ran- with CEE use continued a trend that Among postmenopausal women with domized to CEE during the interven- emerged during the intervention prior hysterectomy who stopped tak- tion period became statistically signifi- period.8,9 This finding differs from the ing CEE after a median of 5.9 years of preponderance10-12 but not all13,14 use, several patterns of health risks and Considering the entire follow-up pe- observational studies that suggest benefits seen during the intervention riod, rates of total mortality and the CEE use, especially in lean women15,16 period were not maintained during the global index of chronic diseases were and after long duration of exposure,17 postintervention period, while other essentially the same in the CEE and pla- increases breast cancer incidence. We trends persisted. For CHD (a primary cebo groups. Statistically significant age previously reported no significant dif- trial end point), the HRs remained null interactions for CEE use, suggesting ferences by body mass index for CEE after stopping the intervention and greater safety and possible benefit effects on breast cancer incidence overall. The increases in risk of stroke among women in their 50s and poten- among participants in this trial.8 Figure 5. Cumulative Annualized Incidence Rates for Clinical Outcomes in the Women's Health Initiative Estrogen-Alone Trial According to
10-Year Age Groups at Enrollment
Hazard ratio (HR) No. (%) of Events Overall follow-up Event by Age Group, y Intervention phase Deep vein thrombosis Pulmonary embolism Invasive breast cancer Colorectal cancer Death (all causes) Annualized incidence rates were estimated for the overall follow-up period by dividing the number of events by the corresponding person-time for participants in eachage stratum. The black squares indicate the HRs for the overall follow-up period. For comparison, the HRs for the intervention phase are shown as blue bars. CEEindicates conjugated equine estrogen; CHD, coronary heart disease; CI, confidence interval; MI, myocardial infarction.
1312 JAMA, April 6, 2011—Vol 305, No. 13 (Reprinted)
2011 American Medical Association. All rights reserved.
Downloaded From: http://jama.jamanetwork.com/ on 07/19/2013
HEALTH OUTCOMES AFTER STOPPING ESTROGEN THERAPY ported by preclinical34 and clinical Science Center, University of Texas, San Antonio (DrBrzyski); John A. Burns School of Medicine, Univer- Women Study have suggested,18 based data35-37 but are not applicable to older sity of Hawaii and Pacific Health Research Institute, on recent findings,9,19,20 that time from women. In a subset of WHI partici- Honolulu (Dr Curb); MedStar Research Institute, Hy- menopause (longer in the WHI vs pants aged 50 to 59 years at study en- attsville, Maryland (Dr Howard); Division of Preven-tive Medicine, University of Alabama, Birmingham shorter in usual clinical practice and try, coronary artery calcium measure- (Dr Lewis); and Department of Social and Preventive observational study cohorts) may Medicine, State University of New York, Buffalo (DrWactawski-Wende).
account for some of the differences in atherosclerotic plaque burden, were Author Contributions: Dr LaCroix had full access to
risk estimates from various studies.
lower following trial completion among all of the data in the study and takes responsibility forthe integrity of the data and the accuracy of the data Alternatively, confounding by differ- women randomized to CEE vs pla- ential mammogram use in the obser- cebo.35 Other support derives from non- Study concept and design: Manson, Margolis, vational studies (higher in estrogen human primate models36 and observa- Stefanick, Wactawski-Wende.
Acquisition of data: LaCroix, Chlebowski, Manson, users) may explain the finding of tional studies.38-40 An important caveat Johnson, Margolis, Stefanick, Brzyski, Curb, Howard, higher breast cancer incidence among is that study participants took unop- Wactawski-Wende.
Analysis and interpretation of data: LaCroix, hormone therapy users.20 Future sub- posed estrogen for a median duration Chlebowski, Manson, Aragaki, Johnson, Martin, group analyses in this trial, which are of less than 6 years and our results can- Margolis, Curb, Howard, Lewis, Wactawski-Wende.
beyond the scope of the current not be extrapolated to longer or shorter Drafting of the manuscript: LaCroix, Manson, Aragaki.
Critical revision of the manuscript for important in- study, will explore this issue.
tellectual content: LaCroix, Chlebowski, Manson, It is unlikely that diagnostic delay ex- Our results emphasize the need to Johnson, Martin, Margolis, Stefanick, Brzyski, Curb,Howard, Lewis, Wactawski-Wende.
plains our breast cancer results be- counsel women about hormone Statistical analysis: LaCroix, Aragaki.
cause CEE only modestly influenced therapy differently depending on their Obtained funding: LaCroix, Chlebowski, Manson, breast density21 and mammogram di- age and hysterectomy status. A post- Johnson, Stefanick, Curb, Howard, Lewis, Wactawski-Wende.
agnostic performance.22 In terms of bio- menopausal woman who has had a Administrative, technical, or material support: logical plausibility, preclinical23,24 and hysterectomy and is considering ini- Chlebowski, Manson, Johnson, Brzyski, Curb, Lewis,Wactawski-Wende.
clinical25 studies suggest that the adap- tiation of CEE should be counseled Study supervision: LaCroix, Johnson, Margolis, Curb, tive changes to gene expression pro- about the increased risks of venous files that occur during estrogen expo- thromboembolism and stroke during Conflict of Interest Disclosures: All authors have
completed and submitted the ICMJE Form for Dis-
sure and after estrogen deprivation26 treatment, which diminish with treat- closure of Potential Conflicts of Interest. Dr LaCroix may render mammary tumors suscep- ment cessation. Among younger reported serving on scientific advisory committeesfor research studies funded by Warner Chilcott and tible to inhibition by estrogen. In con- women, no new safety concerns sanofi-aventis, Amgen, and Pfizer. Dr Chlebowski trast to these results from the Estrogen- emerged and some risk reductions reported being a consultant for AstraZeneca, Novar-tis, Amgen, and Pfizer; receiving funding support Alone Trial, the WHI combined became apparent during the postinter- from Amgen; and serving on speaker's bureaus Estrogen Plus Progestin Trial showed vention period. Among older women, for AstraZeneca and Novartis. Dr Howard reportedreceiving payment for lectures, serving on a that treatment impeded mammo- risks of colorectal cancer, death, and speaker's bureau, and being a consultant for graphic accuracy, and was associated the global index of chronic diseases Merck/Schering-Plough Wyeth Research; and with significant increase in rates of both were elevated over the cumulative receiving research support through the donation ofdrugs from Merck/Schering-Plough. None of the breast cancer incidence and breast can- follow-up period. The risks and ben- other authors reported any financial disclosures.
cer mortality.27-29 efits of CEE use for periods of longer Funding/Support: Wyeth Ayerst donated the study
drugs. The Women's Health Initiative program is
With extended follow-up, hip frac- than 5 to 6 years cannot be inferred funded by the National Heart, Lung, and Blood Insti- ture cumulative incidence was the same from these data for any age group.
tute, the National Institutes of Health, and the US De-partment of Health and Human Services through con- in the CEE and placebo groups. Rates Mechanisms underlying the reduced tracts N01WH22110, 24152, 32100-2, 32105-6, of hip fracture were somewhat higher risks of breast cancer in all women, 32108-9, 32111-13, 32115, 32118-32119, 32122, among women in the CEE group com- and coronary events in younger but 42107-26, 42129-32, and 44221.
Role of the Sponsor: Committees of the Women's
pared with those in the placebo group not older women, warrant further Health Initiative investigators that included represen- after stopping the intervention. These tatives from the National Heart, Lung, and Blood In-stitute (the study's sponsor) had a role in the design results are consistent with studies show- Author Affiliations: Public Health Sciences, Fred
and conduct of the study; collection, management, ing accelerated bone loss30 and a short- Hutchinson Cancer Research Center, Seattle, Wash- analysis, and interpretation of the data; and prepara-tion and review of the manuscript. The Women's term increased risk of hip fracture ington (Dr LaCroix and Mr Aragaki); Los Angeles Bio-medical Research Institute at Harbor-UCLA Medical Health Initiative Project Office at the National Heart, among women who discontinue hor- Center, Torrance, California (Dr Chlebowski); Divi- Lung, and Blood Institute approved the manuscript but mone therapy,31 and no fracture risk re- sion of Preventive Medicine, Brigham and Women's had no other role.
Hospital, Harvard University, Boston, Massachusetts Women's Health Initiative Investigators: A full list-
duction or elevation in past hormone (Dr Manson); Health Science Center, Department of ing of Women's Health Initiative investigators can therapy users.32,33 Preventive Medicine, University of Tennessee, Mem- be found at http://whiscience.org/publications phis (Dr Johnson); Department of Cardiology, George Our results suggest that women ran- Washington University, Washington, DC (Dr Mar- Online-Only Material: eTable 1 and eTable 2 are avail-
domized to CEE while in their 50s had tin); Berman Center for Clinical Research, University able at http://www.jama.com.
fewer CHD events than those random- of Minnesota, Minneapolis (Dr Margolis); Stanford Pre- Additional Contributions: We thank the Women's
vention Research Center, Stanford University School Health Initiative investigators and staff for their out- ized to placebo, findings that are sup- of Medicine, Stanford, California (Dr Stefanick); Health standing dedication and commitment.
2011 American Medical Association. All rights reserved.
(Reprinted) JAMA, April 6, 2011—Vol 305, No. 13 1313
Downloaded From: http://jama.jamanetwork.com/ on 07/19/2013
HEALTH OUTCOMES AFTER STOPPING ESTROGEN THERAPY mone therapy and risk of breast cancer. JAMA. 2003; positive breast cancer in postmenopausal women.
J Clin Oncol. 2010;28(7):1161-1167.
1. Anderson GL, Limacher M, Assaf AR, et al. Effects
14. Kerlikowske K, Miglioretti DL, Ballard-Barbash R,
27. Chlebowski RT, Hendrix SL, Langer RD, et al. In-
of conjugated equine estrogen in postmenopausal et al. Prognostic characteristics of breast cancer among fluence of estrogen plus progestin on breast cancer women with hysterectomy: the Women's Health Ini- postmenopausal hormone users in a screened and mammography in healthy postmenopausal tiative randomized controlled trial. JAMA. 2004; population. J Clin Oncol. 2003;21(23):4314-4321.
women. JAMA. 2003;289(24):3243-3253.
15. Schairer C, Lubin J, Troisi R, et al. Menopausal es-
28. Chlebowski RT, Kuller LH, Prentice RL, et al. Breast
2. The Women's Health Initiative Study Group. De-
trogen and estrogen-progestin replacement therapy cancer after use of estrogen plus progestin in post- sign of the Women's Health Initiative clinical trial and and breast cancer risk. JAMA. 2000;283(4):485- menopausal women. N Engl J Med. 2009;360 observational study. Control Clin Trials. 1998; 16. Calle EE, Feigelson HS, Hildebrand JS, et al. Post-
29. Chlebowski RT, Anderson GL, Gass M, et al. Es-
3. Curb JD, McTiernan A, Heckbert SR, et al. Out-
menopausal hormone use and breast cancer associa- trogen plus progestin and breast cancer incidence and comes ascertainment and adjudication methods in the tions differ by hormone regimen and histologic subtype.
mortality in postmenopausal women. JAMA. 2010; Women's Health Initiative. Ann Epidemiol. 2003; 17. Chen WY, Manson JE, Hankinson SE, et al. Un-
30. Greendale GA, Espeland M, Slone S, et al. Bone
4. Heiss G, Wallace R, Anderson GL, et al. Health risks
opposed estrogen therapy and the risk of invasive mass response to discontinuation of long-term hor- and benefits 3 years after stopping randomized treat- breast cancer. Arch Intern Med. 2006;166(9):1027- mone replacement therapy. Arch Intern Med. 2002; ment with estrogen and progestin. JAMA. 2008; 18. Beral V, Reeves G, Bull D, et al. Breast cancer risk
31. Yates J, Barrett-Connor E, Barlas S, et al. Rapid
5. Cox DR. Regression analysis and life tables. J R Stat
in relation to the interval between menopause and loss of hip fracture protection after estrogen cessa- Soc [Ser A]. 1972;34:187-220.
starting hormone therapy. J Natl Cancer Inst. 2011; tion: evidence from the National Osteoporosis Risk 6. Prentice RL, Caan B, Chlebowski RT, et al. Low-
Assessment. Obstet Gynecol. 2004;103(3):440- fat dietary pattern and risk of invasive breast cancer: 19. Fournier A, Mesrine S, Boutron-Ruault MC,
the Women's Health Initiative Randomized Con- Clavel-Chapelon F. Estrogen-progestin menopausal 32. Cauley JA, Seeley DG, Ensrud K, et al. Estrogen
trolled Dietary Modification Trial. JAMA. 2006; hormone therapy and breast cancer: does delay from replacement therapy and fractures in older women.
menopause onset to treatment initiation influence risks? Ann Intern Med. 1995;122(1):9-16.
7. Rossouw JE, Prentice RL, Manson JE, et al. Post-
J Clin Oncol. 2009;27(31):5138-5143.
33. Banks E, Beral V, Reeves G, et al. Fracture inci-
menopausal hormone therapy and risk of cardiovas- 20. Chlebowski RT, Anderson GL. The influence of
dence in relation to the pattern of use of hormone cular disease by age and years since menopause. JAMA.
time from menopause and mammography on hor- therapy in postmenopausal women. JAMA. 2004; mone therapy-related breast cancer risk assessment.
8. Stefanick ML, Anderson GL, Margolis KL, et al. Ef-
J Natl Cancer Inst. 2011;103(4):284-285.
34. Mendelsohn ME, Karas RH. Molecular and cel-
fects of conjugated equine estrogens on breast can- 21. McTiernan A, Chlebowski RT, Martin C, et al. Con-
lular basis of cardiovascular gender differences. Science.
cer and mammography screening in postmeno- jugated equine estrogen influence on mammo- pausal women with hysterectomy. JAMA. 2006; graphic density in postmenopausal women in a sub- 35. Manson JE, Allison MA, Rossouw JE, et al. Estro-
study of the Women's Health Initiative randomized gen therapy and coronary-artery calcification. N Engl 9. Prentice RL, Chlebowski RT, Stefanick ML, et al.
trial. J Clin Oncol. 2009;27(36):6135-6143.
J Med. 2007;356(25):2591-2602.
Conjugated equine estrogens and breast cancer risk 22. Chlebowski RT, Anderson G, Manson JE, et al. Es-
36. Mikkola TS, Clarkson TB. Estrogen replacement
in the Women's Health Initiative clinical trial and ob- trogen alone in postmenopausal women and breast therapy, atherosclerosis, and vascular function. Car- servational study. Am J Epidemiol. 2008;167(12): cancer detection by means of mammography and diovasc Res. 2002;53(3):605-619.
breast biopsy. J Clin Oncol. 2010;28(16):2690- 37. Grodstein F, Clarkson TB, Manson JE. Under-
10. Collaborative Group on Hormonal Factors in Breast
standing the divergent data on postmenopausal hor- Cancer. Breast cancer and hormone replacement 23. Santen RJ, Song RX, Zhang Z, et al. Adaptive hy-
mone therapy. N Engl J Med. 2003;348(7):645- therapy: collaborative reanalysis of data from 51 epi- persensitivity to estrogen. J Steroid Biochem Mol Biol.
demiological studies of 52,705 women with breast can- 38. Manson JE, Bassuk SS. Invited commentary: hor-
cer and 108,411 women without breast cancer. Lancet.
24. Jeng MH, Shupnik MA, Bender TP, et al. Estro-
mone therapy and risk of coronary heart disease why gen receptor expression and function in long-term es- renew the focus on the early years of menopause? Am 11. Colditz GA, Hankinson SE, Hunter DJ, et al. The
trogen-deprived human breast cancer cells.
J Epidemiol. 2007;166(5):511-517.
use of estrogens and progestins and the risk of breast 39. Grodstein F, Manson JE, Stampfer MJ. Hormone
cancer in postmenopausal women. N Engl J Med. 1995; 25. Ellis MJ, Gao F, Dehdashti F, et al. Lower-dose
therapy and coronary heart disease: the role of time vs high-dose oral estradiol therapy of hormone re- since menopause and age at hormone initiation. J Wo- 12. Beral V; Million Women Study Collaborators. Breast
ceptor-positive, aromatase inhibitor-resistant ad- mens Health (Larchmt). 2006;15(1):35-44.
cancer and hormone-replacement therapy in the Mil- vanced breast cancer. JAMA. 2009;302(7):774- 40. Herna´n MA, Alonso A, Logan R, et al. Observa-
lion Women Study. Lancet. 2003;362(9382):419- tional studies analyzed like randomized experiments: 26. Dunbier AK, Anderson H, Ghazoui Z, et al. Re-
an application to postmenopausal hormone therapy 13. Li CI, Malone KE, Porter PL, et al. Relationship be-
lationship between plasma estradiol levels and estro- and coronary heart disease. Epidemiology. 2008; tween long durations and different regimens of hor- gen-responsive gene expression in estrogen receptor- 1314 JAMA, April 6, 2011—Vol 305, No. 13 (Reprinted)
2011 American Medical Association. All rights reserved.
Downloaded From: http://jama.jamanetwork.com/ on 07/19/2013

Source: http://www.fengclinic.com/files/2014/10/AMA-713-Estrogen.pdf