WHI: Relative risk or benefit

1
Presented byJudith Hsia, M.D.at the December
2, 2004 meeting of theAdvisory Committee for
Reproductive Health Drugs
2
Approaches to assessing risks benefits
Lessons from postmenopausal hormone therapy
studies

• Biomarkers
• Observational studies
• Randomized trials
• Intermediate outcomes

3
Change in Lipids After Menopause
Total Cholesterol
HDL-C
of level at -6 months before menopause
-24
6
-18
-12
-6
0
-18
-24
6
-12
-6
0
LDL-C
Triglycerides
of level at -6 months before menopause
-24
6
-18
-12
-6
0
-24
6
-18
-12
-6
0
Months
Months
Jensen J et al. Maturitas 199012321-31.
4
Estrogen Progestin and Intermediate Outcomes
( change, EP minus Placebo)
NEJM 2003349523-34
5
Observational Studies with Estrogen Progestin
NEJM 20032487
6
WHI Hormone Program Baseline Hypotheses
Risk
Benefit
Stroke?
Coronary Artery Disease
Breast Cancer

• Additional Benefits
• Bone (Hip) Fractures
• Overall Mortality

• Additional Risks
• VTE (PE, DVT)

Plan to follow to 2005 (average 8.5 years)
7
Womens Health Initiative Hormone Trials
Initiated screening (N 373,092)
Women who had no uterus at start of study
Women who had a uterus at start of study
N 10,739
N 16,608
CEEdaily MPA Placebo
CEE Placebo
8
WHI Clinical outcomes in the Estrogen Plus
Progestin Trial
Various WHI papers
9
WHI EP Absolute risk or benefit
Events per 10,000 woman-years
10
WHI EP Trial Findings, July 2002 (avg 5.2 y)
Risks
Benefits
105 Increase Dementia
Fracture Reduction (Hip 23)
24 Increase CHD
39 Reduction Colorectal Cancer

31 Increase Stroke
111 Increase Pulmonary Emboli
24 Increase Breast Cancer
Threshold Level
STOPPED Early, Clear Harm
Stopped 3.3 yrs early
Also DVTs
JAMA. 2002288321-333
11
Observational Study vs Randomized Trial Results
NEJM 20032487
Various WHI papers
12
Possible explanations

• Confounding due to healthy user effect
• Compliance bias women adherent to hormones may
  also adhere to other healthful behaviors
• Outcomes identification bias
• Incomplete capture of early clinical events

NEJM 20033487
13
CEE vs CEEMPA
14
WHI Relative risk or benefit
JAMA 20042911701-12 JAMA 20042912947-58
Various WHI papers
15
WHI Absolute risk or benefit
E Alone
Events per 10,000 woman-years
EP
16
WHI E Alone Trial Findings, 2/04 (avg 6.8 y)
Neutral for CHD Neutral for breast cancer
Risks
49 Increase Dementia
Benefits
39 Increase Stroke
Fracture Reduction (Hip 39)
34 Increase Pulmonary Emboli
Threshold Level
STOPPED Early, suggestion of harm
Stopped 1.7 yrs early
Also DVTs
JAMA 20042912947-58
17
Impact of added androgen may be difficult to
predict
18
Intermediate Outcomes
19
Estrogen Trials with Intermediate Outcomes

• Coronary angiography 3 randomized trials
  demonstrated no benefit (or harm) with PHT
• Carotid ultrasound 1 randomized trial
  demonstrated benefit with estradiol
• Coronary calcification no trial data

NEJM 2000343522 JAMA 20022882432 NEJM
2003349535 Ann Intern Med 2001135939
20
Approaches to evaluating risk

• Biomarkers mixed picture may not be predictive
• Observational studies subject to bias
  confounding suitable cohorts may not be
  available
• Randomized trials with intermediate outcomes
  potentially useful
• Randomized trials with clinical outcome long
  expensive