ELITE Trial: Methods and Baseline Data

Design and Objectives

• ELITE stands for Early versus Late Intervention Trial with Estradiol. It is the only randomized controlled trial specifically designed to test the HT timing hypothesis in relation to subclinical atherosclerosis and cognitive changes.
• Primary question: whether postmenopausal hormone therapy (HT) reduces progression of subclinical atherosclerosis and CAD when started soon after menopause (<6 years) compared with initiation distant from menopause (≥10 years).
• Design: randomized, double-blind, placebo-controlled with a 2 × 2 factorial design. Trial registration: NCT00114517.
• Population: 643 healthy, postmenopausal women without cardiovascular disease, aged roughly mid-50s to mid-60s, stratified by time since menopause and hysterectomy status, to receive active HT or placebo.
• Endpoints include: subclinical atherosclerosis measured by carotid intima-media thickness (CIMT) and coronary imaging (CAC and CCTA); cognitive endpoints reported separately.

Treatment and Intervention Details

• For women with an intact uterus: randomized to oral micronized 17β-estradiol (1 mg/day) plus vaginal micronized progesterone gel (45 mg/day for 10 days each month) or to placebo matching both components.
• For women without an intact uterus (no hysterectomy): randomized to oral micronized 17β-estradiol (1 mg/day) alone or to placebo.
• Route/dose specifics: HT = estradiol 1 mg/day; progesterone gel used cyclically in those with uterus intact.
• Blinding: all participants, investigators, imaging staff, and data monitors were masked to treatment assignment.

Eligibility Criteria

• Inclusion: postmenopausal women without clinical evidence of CVD; serum estradiol < 25 pg/mL; cessation of regular menses for ≥ 6 months; time since menopause < 6 years or ≥ 10 years at randomization.
• Exclusion: inability to determine time since menopause; TG > 500 mg/dL; diabetes with fasting glucose > 140 mg/dL; creatinine > 2.0 mg/dL; uncontrolled HTN (> 160/110 mm Hg); untreated thyroid disease; life-threatening disease prognostic < 5 years; history of DVT/PE or breast cancer; current HT use within 1 month of screening.

Randomization and Stratification

• Stratification factors: time since menopause (<6 y or ≥10 y); hysterectomy status (yes/no); baseline CIMT (<0.75 mm or ≥0.75 mm).
• Randomization ratio: 1:1 to HT or placebo within each stratum.
• Method: stratified blocked randomization; opaque, blinded study products prepared according to the stratification lists. Assignment concealed from clinic staff and participants.

Follow-Up and Assessments

• Visit schedule: monthly for the first 6 months, then every other month until trial completion.
• Baseline visits occurred before randomization and included two CIMT assessments to establish a baseline and randomization stratum (2 CIMTs 2–4 weeks apart).
• Laboratory measurements (every 6 months): plasma estradiol (radioimmunoassay with solvent extraction and Celite chromatography), SHBG (Immunlite), lipids (total cholesterol, TG, LDL-C, VLDL-C, HDL-C, HDL-TG) by ultracentrifugation and enzymatic assays standardized to CDC Lipid Research Clinic protocol, HbA1c (HPLC-IE), fasting glucose; yearly CBC and chemistry panels (Quest Diagnostics).
• Anthropometrics and vitals: BP, pulse rate, weight baseline and at each visit; waist/hip circumference every 6 months; height at baseline.
• Other assessments: yearly 12-lead ECG; medical history, angina, claudication, physical activity, smoking, alcohol per 6 months; 3-day dietary records; reproductive history; depression and WHQ at baseline and every 6 months; hot flash, vaginal bleeding, cramping, and breast pain diaries; mammography and gynecologic exams baseline, annually, and as indicated; endometrial biopsy if indicated.

Endpoints

• Primary Endpoint: rate of change in the right distal common carotid artery far-wall intima-media thickness (CIMT) measured by high-resolution B-mode ultrasound from two baseline CIMTs and serial follow-ups every 6 months. Measurements use standardized acquisition and a dedicated SEMI-automated analysis platform.
  • Primary measure: $ext{rate of CIMT change} = \frac{ ext{CIMT}<em>{t} - ext{CIMT}</em>{0}}{t} ext{ (mm/y)}$ for the right distal common carotid artery far wall.
• Secondary Endpoints:
  • Cognition assessments (detailed elsewhere).
  • Subclinical coronary artery atherosclerosis: CAC from cardiac CT and CCTA for stenosis, with endpoints including total CAC score, total plaque score, total stenosis score, and plaque composition across coronary segments.
  • Endpoints assessed in a single CT session at trial end using a GE 64-slice MDCT scanner; CAC and CCTA endpoints adjusted for visit timing.
• Safety Endpoints:
  • Major clinical events: cardiovascular events (MI fatal/nonfatal, silent MI, sudden death), hospitalization for unstable angina, revascularization (CABG, PCI).
  • Stroke; venous thromboembolism (DVT/PE); cancer (breast, uterine, ovarian, GI, lung);
  • Fractures; all-cause and noncoronary mortality.
• Statistical Analysis Plan:
  • Baseline comparisons: demographics, risk factors, HT history, estradiol and progesterone, CIMT; methods include two-sample t-tests or nonparametric analogs for continuous variables and chi-square tests for categorical variables.
  • On-trial laboratory variables vs treatment: analyzed with marginal models using generalized estimating equations; factors include treatment, time since menopause, stratification variables; baseline levels as covariates.
  • Primary CIMT analyses: linear mixed-effects models with participant-specific intercept and slope; factors include treatment group, time-since-menopause group; interactions for treatment-by-time, menopause-by-time, and three-way interaction.
  • Endpoints CAC and CCTA: generalized linear models with covariates for stratification factors and visit indicators; analyses include all randomized participants and adherent participants (≥80% estrogen therapy compliance).
  • Safety analyses: exact methods comparing events across four groups: early HT, early placebo, late HT, late placebo.

Sample Size and Power

• Projected mean CIMT progression differences used for power calculations:
  • Early (<6 y) group: expected difference of $0.0144$ mm/y with SD $0.02$.
  • Late (≥10 y) group: expected difference of $0.0021$ mm/y with SD $0.02$.
• Required: $83$ participants per cell to achieve 80% power at $\beta=0.20$ and two-sided \a = 0.05 for the interaction between treatment and time since menopause.
• Planned sample size was adjusted for dropout: $126$ participants per cell (anticipating 25% dropout).
• Actual randomized: $n = 643$ participants.

Recruitment and Screening Outcomes

• Recruitment period: May 2005 to Sep 30, 2008; first participant randomized July 13, 2005; last randomized Sept 30, 2008.
• Pre-randomization screening: telephone screen (n = 2166) followed by clinic screen (n = 895).
• Ineligible or declines: 1,523 (70%). Primary reasons for ineligibility/decline included current HT use, time since menopause (6–9 years), signs/symptoms of CVD, diabetes (glucose ≥ 140 mg/dL), history of breast cancer, and not postmenopausal by criteria.
• Randomized: 643 participants randomized across strata.

Baseline Characteristics (Table 1 summary)

• Age (mean, SD):
  • <6 years since menopause: 55.4 ext{ y} ext{ (SD }4.1)
  • ≥10 years since menopause: 65.4 ext{ y} ext{ (SD }6.0)
• Time since menopause (median, IQR):
  • <6 y: $3.5 ext{ y} ext{ (1.9–5.0)}$ [n ≈ 269]
  • ≥10 y: $14.3 ext{ y} ext{ (11.5–18.7)}$ [n ≈ 344]
• Race/ethnicity: White, non-Hispanic: 64.2 o 71.5 ext{%}; Black, non-Hispanic: $8.9 o 9.7 ext$; Hispanic: $15.1 o 13.2 ext$; Asian: $11.8 o 5.7 ext$; overall, minority representation ~32%; education: College graduate higher in early group (72.0%) vs late group (62.6%); high school/some college ~27.7% vs 36.8%; less than high school ~0.4% vs 0.5%.
• Smoking history: Current smokers $4.1 ext$ vs $3.0 ext$; Former smokers $32.8 ext$ vs $39.5 ext$; Never smoked $63.1 ext$ vs $57.5 ext$.
• Physical activity: Mean hours/week: $7.0$ (SD $8.4$) vs $5.5$ (SD $6.2$); p = $0.01$.
• Alcohol use (weekly): distribution across categories (None, >0 to 1 unit, >1 to 2 units, >2 units).
• Menopause type: Natural $96.7 ext$ vs $83.9 ext$; Surgical $3.3 ext$ vs $16.1 ext$.
• Hot flashes (any in prior month): more common in the early postmenopause group; data show differences by stratum (exact counts reported in Table 1).
• Past HT use, current hypertension meds, and current lipid-lowering meds higher in late postmenopause group.

Baseline Clinical and Laboratory Characteristics (Table 2)

• Carotid CIMT (mm):
  • <6 y: $0.748 ext{ (SD }0.095)$
  • ≥10 y: $0.787 ext{ (SD }0.109)$
  • P < 0.001
• BMI (kg/m²): $27.2 ext{ vs } 27.4 ext{; }P = 0.7$
• Pulse rate (beats/min): $65.5 vs 66.1; P = 0.1$
• Systolic BP (mm Hg): $116.8 vs 118.7; P = 0.06$
• Diastolic BP (mm Hg): $76.1 vs 74.3; P = 0.001$
• Lipids (lipoproteins): LDL-C (mg/dL): $139.4 vs 134.3; P = 0.04$; HDL-C: $64.4 vs 67.1; P = 0.06$; Total cholesterol: $224.8 vs 223.0; P = 0.5$; TG: $105.1 vs 108.0; P = 0.5$; TC-to-HDL-C ratio: $3.7 vs 3.5; P = 0.07$
• Glucose (mg/dL): $95.5 vs 94.5; P = 0.2$
• HbA1c (%): $5.5 ext{ (0.5) [n=269]} ext{ vs } 5.7 ext{ (0.4) [n=369]}$; P = $0.002$
• Serum concentrations (median, IQR):
  • Total estradiol (pg/mL): $8.0 (5.8-11.4)$ vs $7.7 (5.8-10.5)$; P = $0.2$
  • Free estradiol (pg/mL): $0.20 (0.14-0.31)$ vs $0.21 (0.14-0.31)$; P = $0.9$
  • Estrone (pg/mL): $29.1 (22.7-38.1)$ vs $27.7 (20.2-37.1)$; P = $0.4$
  • Progesterone (ng/mL): 0.2 (<0.2-0.3) vs 0.2 (<0.2-0.3); P = $0.1$
  • Total testosterone (ng/dL): $20.5 (15.5-27.9)$ vs $7.7 (5.8-10.5)$; P = $0.18$
  • Free testosterone (pg/dL): $3.8 (2.9-5.5)$ vs ???; P = $0.04$
  • SHBG (nmol/L): $46.1 (32.5-61.1)$ vs $51.1 (37.7-66.9)$; P = $0.01$
• All values are means (SD) unless noted otherwise; medians for some variables; comparisons by strata used t-tests or Wilcoxon tests as appropriate.

Baseline CIMT Associations by Postmenopause Strata (Table 3; baseline longitudinal associations)

• Overall pattern: CIMT increases with age and with traditional CV risk factors across both strata, but certain relationships differ by time since menopause (interaction effects discussed).
• Age: mean CIMT increases by $0.026 ext{ mm}$ per every $5 ext{ years}$ of age; SE = $0.003$; P < 0.001.
• SBP: CIMT increases by $0.002 ext{ mm}$ per each 1 mm Hg increase in SBP in both strata (early and late). SEs: early $0.0004$; late $0.0005$; P values not explicitly provided here but significant for SBP effect.
• BMI: CIMT increases by $0.002 ext{ mm}$ per 1 kg/m² increase in BMI (early); by $0.003 ext{ mm}$ per 1 kg/m² (late).
• LDL-C: CIMT increases by $0.0004 ext{ mm}$ per 1 mg/dL increase in LDL-C in the early stratum; effect not evident in the late stratum.
• TG (log-transformed): CIMT increases by $0.066 ext{ mm}$ per unit increase in log(TG) in the early stratum.
• TC-to-HDL-C ratio: CIMT increases by $0.018 ext{ mm}$ per unit increase; SE $0.005$.
• HDL-C (inverse relation): CIMT decreases by $0.001 ext{ mm}$ per 1 mg/dL increase in HDL-C in the early stratum; SE $0.0003$; in late stratum, effect is smaller with SE $0.0003$.
• Total testosterone (inverse in early stratum): CIMT decreases by $0.054 ext{ mm}$ per unit increase in log(Total testosterone) (ng/dL); SE $0.027$; P ≈ $0.05$.
• SHBG (inverse in early stratum): CIMT decreases by $0.082 ext{ mm}$ per unit increase in log(SHBG) (nmol/L); SE $0.027$; P = 0.002 in early stratum; interaction with stratum not significant overall (P ≈ 0.1).
• Estradiol: Serum total estradiol shows a negative association with CIMT in the early stratum (mean CIMT decrease of $0.032$ per unit increase in log(total estradiol); SE $0.019$; P ≈ 0.09), but not in the late stratum (P ≈ 0.2); interaction P ≈ 0.02.
• Estrone and Progesterone: weaker or non-significant associations with CIMT; interactions generally non-significant.
• Pulse rate: Positive association with CIMT in the early stratum (increase of $0.003$ per additional beat/min; SE $0.001$; P = 0.002); not in the late stratum.
• Notes: The table reports separate regression coefficients (β) and SEs for each stratum with P for interaction to indicate whether the relationships differ by time since menopause.

Secondary Cognitive and CAC/CCTA Endpoints (Context)

• Cognitive baseline data reported elsewhere (not in this article).
• CAC and CCTA endpoints are planned as end-of-study measures to assess coronary atherosclerosis across all coronary segments, including total plaque score and stenosis measures, with end-of-study CAC and CCTA outcomes compared between treatment groups and strata.

Recruited Sample and Baseline Separation (Key Observations)

• Intent: Well-separated strata by time since menopause to maximize the ability to detect differential HT effects by initiation timing.
• Early postmenopause group ( <6 y ) characteristics roughly: younger age, shorter time since menopause, higher hot flash prevalence, lower baseline CIMT compared with late postmenopause group.
• Late postmenopause group ( ≥10 y ) characteristics roughly: older age, higher CIMT, greater baseline use of antihypertensive and lipid-lowering medications, higher prevalence of prior HT use and natural vs surgical menopause differences.
• The mean CIMT at baseline was higher in the late postmenopause group (0.787 mm) than in the early postmenopause group (0.748 mm).

Context and Implications (Discussion Highlights)

• Baseline data suggest that HT timing may modulate responses of atherosclerosis progression to HT, potentially through differential associations with lipids (LDL-C, HDL-C), SHBG, and estradiol levels between strata.
• The early postmenopause group showed associations consistent with HT-mediated lipid improvements contributing to reduced CIMT progression; SHBG increase with HT might link to reduced CIMT progression.
• The inverse relationship between total estradiol and CIMT in the early stratum supports the timing hypothesis: raising estradiol levels near menopause could slow atherosclerosis progression; such a relationship was not observed in the late stratum.
• These baseline observations position ELITE to test whether HT initiation shortly after menopause reduces CIMT progression and CAD risk compared with initiation later.
• Prior trials EPAT and WELL-HART showed divergent results by stage of atherosclerosis, supporting the timing hypothesis and motivating ELITE’s design to evaluate the window of opportunity.
• Broader context includes evidence from meta-analyses suggesting magnitude of CHD and mortality benefits with younger initiation (<60 y and/or <10 y since menopause) and potential harm or null effects with older initiation; ELITE aims to contribute mechanistic and subclinical-atherosclerosis data to these clinical outcomes.

Conclusions (Integrated View)

• ELITE baseline data indicate well-separated strata by time since menopause, with expected differences in age, CIMT, baseline risk factors, and medication use.
• Baseline associations suggest that LDL-C, HDL-C, SHBG, and estradiol levels may modulate CIMT progression differently by timing of HT initiation, aligning with the timing hypothesis and providing a mechanistic basis for potential differential HT effects on atherosclerosis progression.
• The study design and baseline data position ELITE to test whether early HT initiation (<6 years since menopause) reduces CIMT progression and CAD burden relative to late initiation (≥10 years since menopause).

Supplemental and References (as noted in the article)

• Supplemental Digital Content 1 contains detailed ultrasound CIMT procedures and CT imaging methodologies.
• Key references cited include groundwork on the timing hypothesis, EPAT and WELL-HART results, and meta-analytic syntheses of HT effects by age and years since menopause.
• Acknowledgments list the ELITE Research Group and supporting cores (Imaging, Data Coordinating Center, Lipid Lab, Cardiac CT Core, Neurocognition, etc.).

Quick glossary of key terms used

• CIMT: Carotid artery intima-media thickness, a measure of subclinical atherosclerosis.
• CAC: Coronary artery calcium score, a CT-based measure of calcified plaque.
• CCTA: Coronary CT angiography, imaging for coronary artery stenosis and plaque composition.
• SHBG: Sex hormone-binding globulin, a transport protein for sex hormones.
• LDL-C, HDL-C: Low-density and high-density lipoprotein cholesterol.
• E2: Estradiol (total estradiol); NETA: norethynisterone (in trial nomenclature for progestin combos in some contexts).
• DVT/PE: Deep vein thrombosis / pulmonary embolism.
• WHQ: Women's Health Questionnaire.
• RCT: Randomized controlled trial.

LaTeX-formatted numerical references (examples)

• Primary CIMT rate definition: $\frac{ ext{CIMT}<em>t - ext{CIMT}</em>0}{t} ext{ (mm/y)}$
• Age effect on CIMT: eta = 0.026 ext{ mm per } 5 ext{ y}, ext{ SE }=0.003, ext{ P}<0.001
• Baseline CIMT by strata: $ext{CIMT}<em>{<6y} = 0.748 ext{ mm} ext{ (SD }0.095)$, $$ ext{CIMT}{\