Evidence Table 3

Outcomes in Women's Health Initiative Hormone Therapy Studies

StudyInclusion criteriaPopulationInterventionSample size Follow-up period (years)Primary endpoint: efficacy or safetyOther endpointsPopulation subgroupsAttrition AdherenceConclusions
Cardiovascular outcomes
CEE + MPA
Rossouw 2002
(prior review)
Women 50–70 years; intact uterus; could have taken HT previously (3-m wash-out)

Exclusion criteria: Menses ≤6m previously (≤ 12m if 50–54y); life expectancy <3y; history of breast cancer, melanoma, other cancers in last 10y; low hematocrit; alcoholism; dementia
Age: Mean 63.6 (SD 7.1)
Race: non-Hispanic white: 75%
Education: 24% college education
History of MI, angina, stroke, PE: each <3.0%
CEE 0.625 + MPA 2.5 mg qdTotal N: 16,608
CEE: 8506
Placebo: 8102

Average F/U: 5.2 (stopped early due to concerns regarding increased breast cancer and some increase in CHD, stroke, and PE)
CHD events: HR 1.29 (95% CI, 1.02–1.63)
CHD deaths: HR 1.18 (95% CI, 0.70 – 1.97)
Global index (earliest occurrence of CHD, invasive breast cancer, stroke, PE, endometrial cancer, colorectal cancer, hip fracture, death due to other causes): HR 1.15 (95% CI,1.03 – 1.28)

Safety:
Invasive breast cancer: HR 1.26 (95% CI, 1.00 –1.59)
Total deaths: HR 0.98 (95% CI, 0.82 – 1.18)
Strokes: HR 1.41 (95% CI, 1.07 – 1.85)
Venous thromboembolic disease: HR 2.11 (95% CI, 1.58 – 2.82)

Colorectal cancer: HR 0.63 (95% CI, 0.43 – 0.92)

Total fractures: HR 0.76 (95% CI, 0.69 – 0.89)
Hip fractures: HR 0.66 (95% CI, 0.45–0.98)
CHD at baseline (N=400): HR for subsequent CHD in CEE vs placebo: 1.28 (95% CI, 0.64–2.56)

Prior HT use for 5–10y: HR for breast cancer in CEE vs placebo: 4.61 (95% CI, 1.01 – 21.02)
Data available on 96.5%

At end of study 42% of CEE and38% of placebo had stopped taking study medication
CEE + MPA after mean follow-up of 5.2 y:
Increased: CHD events, invasive breast cancer, stroke, PE
Decreased: colorectal cancer, hip and vertebral fractures
Total mortality and endometrial cancer did not differ significantly between groups
CEE
Anderson 2004Women 50–70 years; hysterectomy >3m prior; could have taken HRT previously (3-m wash-out)

Exclusion: As above
Age: 63.6 (SD 7.3)
Race: Non-Hispanic white: 75.3%
Prior MI: 4.1%
CEE 0.625 mg qdTotal: 10,739
CEE: 5,310
Placebo: 5,429

Average F/U: 6.8 (range5.7 to 10.7y)
Incidence per 10,000 person-years
CHD events: CEE 49, placebo 54 (p>0.05); HR 0.91 (95% CI, 0.75–1.12)
Total CVD events: CEE 225, placebo 201; HR1.12 (95% CI, 1.01 – 1.24)

Global index of health risks and benefits: HR1.01 (95% CI, 0.91–1.12)

Safety:
Invasive breast cancer: CEE 26, placebo 33(p=0.06); HR 0.77 (95% CI, 0.59 – 1.10)
Total mortality: HR 1.04 (95% CI, 0.88 – 1.22)
Incidence per 10,000 person-years
Hip fractures: CEE 11, placebo 17 (p=0.01), HR 0.61 (95% CI, 0.41 – 0.91)
Total osteoporotic fractures: CEE 139, placebo 195 (p<0.001), HR 0.70 (95% CI, 0.63 – 0..79)

Stroke: CEE 44, placebo 32 (p=0.007); HR1.39 (95% CI, 1.10 – 1.77)
VTE (DVT and PE): CEE 28, placebo 21(p>0.05); HR 1.33 (95% CI, 0.99–1.79)

Colorectal cancer: HR 1.08 (95% CI, 0.75–1.55)
Prior MI (n=441): HR MI: 1.04 (95% CI, 0.63–1.71)
Those without prior MI: HR 0.91 (95% CI, 0.73 – 1.14)
At study termination: 53% had already stopped taking study medication; NSD between groupsCEE in women with a hysterectomy increases the risk of stroke (12 cases per 10,000 P-Y) , reduces the risk of hip (6 cases per 10,000 P-Y) and other fractures, and does not significantly affect CHD event rates or overall mortality. There was a nonsignificant reduction in breast cancer (7 cases per 10,000 P-Y)
Symptoms
Barnabei 2005Women 50–70 years; intact uterus; could have taken HRT previously (3-m wash- out)Age: mean 63 (range 50–79)
Race: Non-hispanic white: 84%; non- Hispanic black: 6.8%
Education: 35% college educated
Post-menopausal: mean 13.4y
CEE 0.625 + MPA 2.5 mg qdTotal: 16,608 F/U 5.6Relief/improvement of symptoms at 1y:
Symptomatic at baseline:
Hot flashes, night sweats, breast tenderness, vaginal/genital dryness, joint pain/stiffness: improved (p<0.05)
Vaginal/genital discharge, irritation/itching, headaches, mood swings, extremity swelling: NSD

Asymptomatic at baseline:
Hot flashes, night sweats, vaginal dryness, joint pain/stiffness, general aches or pains: improved (p<0.05)

Safety:
Vaginal bleeding: most frequently reported treatment effect in CEE+MPA (42.5% and51.0% at 6w and 6m); placebo < 5% throughout study
Weight 1y: higher proportion lost weight with CEE+MPA than placebo (no statistics)

Breast tenderness, vaginal irritation and discharge, headaches: increased (p<0.05)
Mood swings, extremity swelling: NSD
8.6% subsample at year 3:
Moderate-to-severe symptoms at baseline: NSD hot flashes, various genital and musculoskeletal symptoms

Asymptomatic at baseline: decreased joint pain or stiffness(p=0.04); NSD other symptoms
NRCEE+MPA decreased vasomotor symptoms, especially in younger women. Vaginal or genital dryness and joint aches and pains were also decreased. Vaginal bleeding in the CEE+MPA group was common, especially in the first 6 months.
In a subsample examined at 3y, women who were asymptomatic at baseline were less likely to report vaginal or genital dryness; there were no significant differences between groups in other symptoms either for women asyptomatic or symptomatic at basline.
Bone
CEE+MPA
Cauley 2003
(prior review)
Women 50–70 years; intact uterus; could have taken HT previously (3-m wash-out)

Exclusion: As above If femoral neck BMD >3 SD below the age-specific mean
Mean age: 63
Race: Non-Hispanic white: 84%
CEE 0.625 + MPA 2.5 mg qdTotal N: 16,608
Patients with BMD measurements: 1024
F/U 5.6 (average)
F/U average 5.6y
Total fractures: CEE+MPA 8.6%, placebo 11.1%; HR.76 (95% CI, 0.69–0.83)
Hip fracture: HR 0.67 (95% CI, 0.47–0.96)
BMD at 3y:
Total hip: increased 3.7% in CEE+MPA vs 0.14% increase in placebo (p<0.001)
Data available on 96.5% of participants
At end of study 42% of CEE+MPA and 38% of placebo had stopped taking study medication
CEE+MPA increases BMD and reduces the risk of fractures in health postmenopausal women, regardless of fracture risk
CEE
Jackson 2006
Update of Anderson 2004
Women 50–70 years;
hysterectomy >3m prior;
could have taken HRT previously (3-m wash-out)
Age: Mean 63.6
Race: non-Hispanic white: 75%
Education: 24% college education
CEE 0.625 mg qdTotal N: 10,739
I: 5310
C: 5429

F/U: mean 7.1y
Hip fracture: HR 0.65 (95% CI, 0.45–0.94)
Total fracture: HR 0.71 (95% CI, 0.64–0.80)
BMD lumbar spine (n=938): CEE increase 7.1%, placebo increase 1.9% (p<0.0001)
Interaction between fracture risk and global index NS (p=0.42)

Greater reduction in hip fracture in women >20y after menopause
Nonadherence to study medications over 7.1y of trial: I 57.5%, C 57.7%CEE in hysterectomized women after menopause significantly reduces incident fractures at the hip, spine and wrist, as well as total fractures. BMD increased significantly persisting for 6y of F/U. Positive effects were consistent largely irrespective of individual risk factors for osteoporosis or fracture. global index was not related to fracture risk, suggesting that even in women at highest risk for fracture, the global index was balanced with no evidence of overall benefit or risk.
Health-related quality of life
CEE + MPA
Hays 2003
(prior review)
Women 50–70 years; intact uterus; could have taken HT previously (3-m wash-out)Age: mean 63.2 (range50-79)
Race: non-hispanic white: 84; non-Hispanic black: 6.8
Education: 35% college educated
CEE 0.625 + MPA 2.5 mg qdCEE+MPA: 8,506
Placebo: 8,102
F/U: 1y
HRQL (SF-36, 8 subscales)
1 y: CEE+MPA > placebo for physical function, bodily pain, sleep disturbance (all p<0.001)
3 y: NSD between CEE+MPA and placebo (9% subsample)
Subgroup analyses: no significant interactions between baseline age, race, BMI, symptoms and outcomes
Age 50–59: same findings as main study
Moderate-to-severe vasomotor symptoms at baseline: same findings as main study
Post hoc analyses:
Age 50–54 and moderate-to- severe symptoms at baseline: improved sleep; NSD other outcomes
Loss to follow-up: 0.1%
Stopped study medications:
CEE+MPA 9.7%, placebo 6.6%
Adherence (taking ≥ 80% of study medications): CEE+MPA74%, placebo 81%
In postmenopausal women CEE+MPA did not have a clinically significant effect on HRQL
CEE
Brunner 2005Women 50–70 years; hysterectomy >3m prior; could have taken HT previously (3-m wash-out)Age: Mean 63.6
Race: non-Hispanic white: 75%
Education: 24% college education
CEE 0.625 mg qdI: 5310
C: 5429
F/U: 1 and 3 y
1y:
Sleep disturbance: positive effect CEE vs placebo (absolute effect 2%) (p<0.001)
SF-36: negative effect of CEE on social functioning (p=0.003); NSD other measures

3y: NSD any HRQL measure (8.6% subsample)
Global QOL rating: NSD in distributions of scores between CEE and placeboModerate-to-severe symptoms at baseline:
% not reporting symptoms at 1-y
F/U: I 72%, C 56% (p<0.001)

Post hoc analyses:
Moderate-to-severe symptoms at baseline and age 50–54y:
SF-36 subscales: NSD
1y
Loss to F/U: 0%
Drug discontinuation: CEE: 8.4%, placebo: 8.0%
Adherence (taking 80% of study medications): CEE: 78%, placebo: 82%

3y
Adherence: I and C: 59%
Estrogen therapy alone in women with a hysterectomy did not improve HRQL to a clinically significant degree compared to placebo at up to 3- y follow-up
Cognition and dementia
CEE + MPA
Rapp 2003

WHIMS
Subset of WHI CEE+MPA study: women ≥ 65y; intact uterus; could have taken HRT previously (3-m wash- out)
English-speaking

Exclusion: As above Probable dementia
Age: mean NR; 46% 65-59y
Race: Non-Hispanic white: 90%
Completed college: 34%
CEE 0.625 + MPA 2.5 mg qdTotal N: 4,532

F/U mean 4.2 (range, 0.9 – 6.4)
Safety:
Rates of change in 3MSE (global cognitive function):
Both groups increased over the first 4y, then decreased; Y3 and Y4 scores for placebo > CEE (p<0.05); NSD Y5 and Y6
Strokes: NSD between groups (p=0.62)
Probable dementia: CEE 40, placebo 21 (p=0.01)
Rates of change in 3MSE for subgroups (age, race, BMI, diabetes, education): NSD between groups
Decrease in score of >2SD: CEE 6.7%, placebo 4.8% (p=0.008)
Attrition: 3.2% (no F/U data available)
Adherance: higher in placebo than CEE (p<0.01)
Mean rate of change in global cognition scores was slightly less favorable in the CEE group than in placebo over average F/U of 4.2y, a result that is not clinically significant. CEE offers no benefit for global cognitive function or no negative effect. There may be a subgroup of women who suffer a detrimental effect.
Schumaker 2003

WHIMS
Subset of WHI CEE+MPA study: women ≥ 65y; intact uterus; could have taken HRT previously (3-m wash- out)
English-speaking

Exclusion: As above Probable dementia
As above for Rapp 2003CEE 0.625 + MPA 2.5 mg qdTotal N: 4,532

F/U mean: 4.05y (SD1.19)
Safety:
incidence probable dementia:
CEE + MPA vs placebo: HR 2.05 (95% CI, 1.21–3.48) (p=0.01)
Mild cognitive impairment: CEE + MPA vs placebo:
HR 1.07 (95% CI, 0.74–1.55)
Probable dementia or mild cognitive impairment: CEE vs placebo:
HR 1.37 (95% CI, 0.99 – 1.89)
Exclude those with high risk dementia at baseline (3MSE scores below screening cut point):
HR CEE+MPA vs placebo: 2.64 (95% CI, 1.26–5.53)
Risk dementia aged 70–74y vs 65–69y: HR 3.54 (95% CI, 1.57–8.00)
Risk dementia in those with high risk dementia at baseline: HR 24.84 (95% CI, 13.19 – 47.75)
NR
Resnick 2006, 2004

WHISCA
As for Rapp 2003
Subset of WHIMS participants
Age: 73.9 (SD 3.8)
Race: Non-Hispanic white: 92%
Completed college: 35%
CEE 0.625 + MPA 2.5 mg qdTotal N: 1,416

Mean F/U: 1.35
Study started 3y after WHI randomization
Safety:
Verbal memory: CEE negative impact vs placebo (p<0.01)
Figural memory: CEE positive impact vs placebo (p=0.012)
Other cognitive domains, affect, depressive symptoms: NSD
Attrition
2nd annual assessment F/U: I 93.3%, C 92.7%
3rd annual assessment F/U: I42.2%, C 44.1%
Adherence
2-y F/U: I 47.4%, C 61.2%
CEE + MPA effect on cognitive function varies across cognitive domains in women over 65y.
CEE alone
Espeland 2004

WHIMS
Subset of WHI CEE study: women ≥ 65y; hysterectomy; could have taken HRT previously (3-m wash-out)
English-speaking

Exclusion: As above Probable dementia
Age: mean NR; 45% 65-69
Race: non-Hispanic white: 83%
Completed college: 24%
CEE 0.625 mg qdCEE: 1,387
Placebo: 1,421
Mean F/U: 5.4y
Safety:
Rates of change in 3MSE (global cognitive function):
Both groups increased over the first 4y, then decreased; NSD between groups for each year
Overall mean 3MSE score: placebo slightly higher than CEE (p=0.04)
Largest declines in scores occurred more frequently in CEE than placebo: relative risk of decline of 10 units in 3MSE with CEE vs placebo: 1.47 (95% CI, 1.04 – 2.07)Baseline 3MSE score lowest had greatest decline in cognitive function (p<0.01)Cumulative drop-out rates year7: CEE 59.4%, placebo 54.2%During F/U of mean 5.4y, global cognitive function decreased with CEE compared to placebo(p=0.04); the adverse effect was more pronounced among women with lower cognitive function at baseline.
Schumaker 2004

WHIMS
As above for Espeland 2004As above for Espeland 2004CEE 0.625 + MPA 2.5 mg qd
or
CEE 0.625 mg qd
CEE alone: 2,947
Pooled data (CEE alone and CEE+MPA): 7,479

F/U CEE alone: 5.21y
(SD 1.73)
F/U Pooled data: 4.05y
(SD 1.19)
Safety:
Incidence of probable dementia:
CEE alone: HR 1.49 (95% CI, 0.83 – 2.66)
Pooled data: HR 2.05 (95% CI, 1.21 – 3.48)
NSD between CEE alone CEE+MPA (p=0.11)
Mild cognitive impairment:
CEE alone: HR 1.34 (95% CI, 0.95 – 1.89)
Pooled data: HR 1.25 (95% CI, 0.97 – 1.60)
Adherence rates dropped over time; at Y7: CEE 36.8%, placebo 45.1%CEE does not reduce dementia or mild cognitive impairment incidence; Pooling data for CEE alone and CEE+MPA increased risk for both endpoints. HT is not recommended to prevent dementia or cognitive decline in women ≥ 65y.

Abbreviations: BMI= body mass index (kg/m2), BMD=bone mineral density, CVD=cardiovascular disease, CEE=conjugated equine estrogens, C=control, CI=confidence interval, CHD=coronary heart disease, DVT=deep vein thrombosis, F/U=follow-up, HR=hazard ratio, HRQL=health-related quality of life, HT=hormone therapy, I=intervention, MPA= medroxyprogesterone acetate, MSE=mini-mental state examinations, m=month, MI=myocardial infarction, N=sample size, NSD=no significant difference, NR=not reported, p=patients, PE=pulmonary embolism, qd=daily, QOL=quality of life, SD=standard deviation, VTE=Venous thromboembolism, WHI=the women's health initiative, WHIMS=the women's health initiative memory study, WHISCA= the women's health initiative study of cognitive aging, y=year

From: Evidence Tables

Cover of Drug Class Review: Hormone Therapy for Postmenopausal Women or Women in the Menopausal Transition Stage
Drug Class Review: Hormone Therapy for Postmenopausal Women or Women in the Menopausal Transition Stage: Final Report Update # 3 [Internet].
Nelson HD, Nygren P, Freeman M, et al.
Portland, (OR): Oregon Health & Science University; 2007 Oct.
Copyright © 2007, Oregon Health & Science University, Portland, Oregon.

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.