CORA-study notes on HRT and CHD risk

Overview

• This study (CORA-study) investigates the association between hormone replacement therapy (HRT) use and risk of incident coronary heart disease (CHD) in postmenopausal women, considering risk factors and lifestyle characteristics.
• Design: Case–control study embedded in CORA with 200 consecutive women with incident CHD and 255 age-matched population-based controls (most postmenopausal).
• Key finding: Current HRT use was more common among controls than cases, and current HRT users had substantially lower odds of CHD compared with never users in age-adjusted analyses. After adjustment for conventional and dietary risk factors, current HRT was no longer an independent risk factor.
• Conclusion: Long-term HRT use is not associated with increased CHD risk in this cohort; potential benefits include favorable effects on weight, central adiposity, insulin sensitivity, and blood pressure. Heavy smoking and non-ideal diet can offset these benefits. The study cautions against generalizing a “healthy user effect” for HRT.

Study population and design

• Population: 200 consecutive women aged 30–80 years admitted with first manifestation of CHD (ICD-10: 121, 122, 124, 125) confirmed by angiography, recruited Nov 1997–Mar 2001.
• Exclusions: cancer, severe chronic disease, prior CHD, or dietary advice regarding CHD.
• Controls: For each case, two age-matched population-based controls were invited via the population registry; 255 controls included (67% of eligible).
• Ethical approval: Protocol approved by the Ethical Committee of Hamburg.
• Study aim: Compare current HRT users, ever users, and never users in relation to incident CHD; assess risk factors and lifestyle differences.

Key definitions and measurements

• Postmenopausal status: No regular menses for >1 year or on HRT.
• HRT categories:
  • Current HRT: using HRT at time of survey or prior to event (within last 4 weeks).
  • Ever HRT: used HRT at any time but not within last 4 weeks.
  • Never HRT: no history of HRT.
• Menopause-related data: Age at menopause, ovarectomy (bilateral oophorectomy).
• Anthropometrics: Weight (kg), BMI (kg/m^2), waist/hip measurements; central adiposity defined as waist-to-hip ratio (WHR) ≥ 0.85.
• Blood pressure and metabolic markers: Systolic/diastolic BP, hypertension status, fasting glucose, insulin, C‑peptide, lipid profile (Total cholesterol, HDL, LDL calculated via Friedewald formula), triglycerides, lipoprotein(a) [Lp(a)].
• Insulin resistance: Homeostasis Model Assessment (HOMA-IR) score; defined as HOMA-IR ≥ 3.8 (upper quartile of European population) for insulin resistance given no history of diabetes.
• Diabetes/obesity risk factors: Diabetes/insulin resistance defined by medication history or biomarker criteria; hypertension defined by antihypertensive use or BP thresholds (SBP ≥ 140 mmHg or DBP ≥ 90 mmHg).
• Diet: Semi-quantitative food frequency questionnaire for 146 items; focus on fruit/vegetables vs meat/sausage intake as markers of dietary risk pattern.
• Lifestyle: Smoking status (current/former/never) and smoking intensity (cigarettes per day).
• LDL calculation:
  ext{LDL} = ext{Total cholesterol} - ext{HDL} - rac{ ext{Triglycerides}}{5} ag{Friedewald formula}
• HRT duration: Current users had a median duration of use ext{9.5 years} ext{ (IQR 10.5)}; ever users had shorter durations (median around 2.5 years). Time since starting HRT also recorded.

Statistical analysis

• Baseline characteristics: compared using chi-square or Wilcoxon tests as appropriate.
• Relative risks: estimated by logistic regression; 95% confidence intervals (CIs) reported.
• Power considerations: designed to detect relative risks (RR) of approximately 1.5–1.7 for a marker with prevalence 30 ext{\%} or 10\% in controls, with 80% power and α < 0.05, using a 1:2 case-control ratio.
• Model building: Factors significant at P < 0.05 in univariate analyses entered into multivariate analysis; conditional logistic regression used to account for matching among postmenopausal women.
• Software: SAS Version 9.1.3.

Results: demographic and exposure characteristics

• Postmenopausal status: About 87.9\% of the CORA population were postmenopausal.
• Cases vs controls: Median ages around late 60s; cases slightly younger than controls among current HRT users (Current HRT: cases 61.9 ± 7.3 vs controls 64.1 ± 7.2, not statistically significant).
• HRT use:
  • Current HRT: Cases 20.2\% vs Controls 32.9\% (p = 0.0188).
  • Ever HRT: Cases 29.8\% vs Controls 24.7\%.
  • Never HRT: Cases 50.0\% vs Controls 42.5\%.
• Duration of current HRT (years): Median 9.5 (IQR not given) for both groups; most current users had >3 years of use.
• Among ovarectomized women: Similar patterns but small numbers limit subgroup analyses.

Results: anthropometric, clinical and dietary factors by HRT use (Table 2)

• Weight and central adiposity: Current HRT users had significantly lower weight, BMI, and WHR than never users;
  • BMI: current vs never users, 24.8 \,\pm\, 4.2 vs 26.5 \,\pm\, 4.6\, ext{kg/m}^2; p=0.0007}; BMI < 25 kg/m^2: 63.9\% vs 40.7\%.
  • WHR: current 0.84 \,\pm\, 0.10 vs never 0.88 \,\pm\, 0.10; WHR ≥ 0.85: 34.6\% vs 51.9\%.
• Blood pressure and metabolic risk:
  • Systolic BP: 133.2 ± 14.7 mmHg (current) vs 139.3 ± 16.2 mmHg (never) in the postmenopausal subgroup; hypertension: 62.0% (current) vs 78.6% (never).
  • Insulin resistance (HOMA-IR) and diabetes: insulin resistance lower in current users; diabetes/insulin resistance: 23.2% (current) vs 40.1% (never).
• Lipids and lipoproteins:
  • HDL-cholesterol higher in current users among controls; overall trend toward more favorable lipids in current users, but smoking mitigates HDL effects.
  • LDL-cholesterol and triglycerides: current users had lower LDL in controls; triglycerides similar overall but some subgroups differed.
• Diet and smoking:
  • Fruit/vegetables: higher intake in current users for controls and cases where present; per 100 g increase associated with lower CHD risk in multivariate model.
  • Meat/sausage: current HRT users ate less meat/sausage than never users; cases ate more meat/sausage overall, with never HRT cases showing the highest meat intake.
  • Smoking: higher prevalence among current HRT users who developed CHD; cigarettes per day higher in current HRT cases than never HRT cases.
• Overall pattern: Current HRT users showed a favorable cardiovascular risk profile on several risk factors, but some subgroups (especially smoking) counteracted potential benefits.

Results: multivariate analysis (Table 3)

• Postmenopausal women in a conditional logistic regression model adjusting for established risk factors:
  • WHR > 0.85: OR = 3.17, 95% CI [1.81, 5.61], p < 0.0001.
  • Hypertension: OR = 3.17, 95% CI [1.56, 6.72], p = 0.002.
  • Diabetes/insulin resistance: OR = 2.36, 95% CI [1.31, 4.30], p = 0.005.
  • Fruit/vegetables per 100 g: OR = 0.66, 95% CI [0.48, 0.88], p = 0.007 (protective effect).
  • Lipoprotein(a) > 25 mg/dl: OR = 2.08, 95% CI [1.16, 3.78], p = 0.01.
  • Meat/sausage per 100 g: OR = 2.33, 95% CI [1.14, 4.78], p = 0.02.
  • HDL-cholesterol > 50 mg/dl: OR = 0.49, 95% CI [0.26, 0.92], p = 0.03 (protective).
  • Smoking: OR = 1.76, 95% CI [0.97, 3.22], p = 0.07 (not statistically significant).
  • Current HRT: OR = 0.70, 95% CI [0.37, 1.33], p = 0.28 (not statistically significant after adjustment).
• Overall interpretation: After accounting for conventional and dietary factors, current HRT was no longer an independent predictor of CHD risk; however, there was a trend toward protection (OR < 1) that did not reach statistical significance in the fully adjusted model.
• Notably, the combination of HRT with smoking did not emerge as an independent predictor of risk, suggesting that any potential interaction may be limited or require larger samples to detect.

Discussion and interpretation

• Primary takeaway: The CORA-study does not support the view that HRT increases CHD risk; instead, it shows a non-significant trend toward protective effects after adjustment and a clear set of lifestyle/disease factors that influence risk.
• Consistency with other evidence:
  • The finding aligns with two of the three major intervention trials or arms that do not show increased CHD risk with HRT in similar age ranges.
  • The WHI estrogen-only arm suggests potential benefit when started early, while combined estrogen/progestin regimens in WHI/HERS show more nuanced results; timing appears critical.
• Timing and age at initiation:
  • Early initiation of HRT (near menopause) may confer cardiovascular benefits; benefits appear attenuated when started later in menopause, consistent with WHI subgroup analyses and other meta-analyses.
  • The CORA cohort largely consisted of women who started HRT around the perimenopausal period or shortly after menopause, resembling groups where cardiovascular benefits have been observed in other studies.
• Mechanistic interpretation:
  • HRT was associated with favorable profiles in weight, central adiposity (lower WHR), insulin sensitivity (lower insulin resistance), and blood pressure, which collectively could reduce CHD risk.
  • HDL-C was not consistently higher in HRT users in this study, possibly due to high smoking rates among HRT users which depress HDL-C levels.
  • Lipoprotein(a) and other lipids showed some favorable trends but not uniformly across groups.
• Diet and lifestyle implications:
  • HRT users tended to eat more fruit/vegetables and less meat/sausage, patterns linked to lower CHD risk in several analyses and consistent with the CORA findings.
  • Smoking remained a key adverse habit; current HRT users who smoked did not show a robust independent protective effect, highlighting lifestyle as a major modifier.
• Limitations:
  • Moderate sample size (200 cases, 255 controls) and single-center design limit the generalizability and power, particularly for subgroup analyses.
  • Heterogeneity of HRT regimens (mostly combined therapy) makes it difficult to separate effects of estrogen vs progestin components.
  • Possible recall bias in dietary data and self-reported HRT use; timing since stopping/starting HRT could be misclassified.
• Implications for practice and future research:
  • Findings support cautious interpretation of observational data suggesting no increased CHD risk with HRT, especially when started near menopause.
  • Emphasize comprehensive risk-factor management (weight, central adiposity, insulin resistance, BP, lipids) and smoking cessation alongside any HRT considerations.
  • Future work could stratify by specific HRT formulations and timing relative to menopause, ideally via larger prospective cohorts or randomized trials with longer follow-up.

Key numerical references and formulas (summary)

• Postmenopausal status in CORA population: ext{Postmenopause} \% \approx 87.9\%.
• Current HRT use:
  • Cases: 20.2\%\,
  • Controls: 32.9\%\,
• Ever HRT use: Cases 29.8\%; Controls 24.7\%; Never HRT: Cases 50.0\%; Controls 42.5\%.
• Median duration of current HRT: 9.5\text{ years} in both groups; most >3 years.
• Time since starting HRT (current users): Cases 12.5\pm10.0\text{ years}; Controls 14.0\pm10.5\text{ years}.
• Multivariate odds ratio for current HRT (age-adjusted):
  ext{OR} = 0.428,
  ewline 95 ext{\% CI} = [0.206, 0.860],
  ewline p = 0.0196.
• After adjusting for conventional and dietary risk factors:
  ext{OR}_{ ext{HRT}} = 0.70,
  ewline 95 ext{\% CI} = [0.37, 1.33] (not statistically significant).
• Table 3 (key adjusted associations):
  • WHR > 0.85: ext{p} < 0.0001, ext{ OR} = 3.17, ext{ 95 ext{\% CI}}=[1.81,5.61].
  • Hypertension: p = 0.002, ext{ OR} = 3.17, ext{ 95 ext{\% CI}}=[1.56,6.72].
  • Diabetes/insulin resistance: p = 0.005, ext{ OR} = 2.36, ext{ 95 ext{\% CI}}=[1.31,4.30].
  • Fruit/vegetables per 100 g: p = 0.007, ext{ OR} = 0.66, ext{ 95 ext{\% CI}}=[0.48,0.88].
  • Lipoprotein(a) > 25 mg/dl: p = 0.01, ext{ OR} = 2.08, ext{ 95 ext{\% CI}}=[1.16,3.78].
  • Meat/sausage per 100 g: p = 0.02, ext{ OR} = 2.33, ext{ 95 ext{\% CI}}=[1.14,4.78].
  • HDL-cholesterol > 50 mg/dl: p = 0.03, ext{ OR} = 0.49, ext{ 95 ext{\% CI}}=[0.26,0.92].
  • Smoking: p = 0.07, ext{ OR} = 1.76, ext{ 95 ext{\% CI}}=[0.97,3.22].
  • Current HRT: p = 0.28, ext{ OR} = 0.70, ext{ 95 ext{\% CI}}=[0.37,1.33].
• Overall interpretation: Several traditional risk factors retained significance; current HRT not independently predictive after full adjustment.

Connections to broader literature

• WHI and WHI-like trials show mixed results depending on estrogen-alone vs estrogen-plus-progestin regimens, and timing relative to menopause; CORA's findings of potential benefit with early initiation align with some subgroup analyses suggesting reduced CHD risk with earlier start.
• The notion of a “healthy user effect” is not supported by the CORA data; however, favorable risk-factor profiles among HRT users persist alongside potential offsetting adverse lifestyle habits (notably smoking).

Practical takeaways

• For postmenopausal women, HRT is not clearly associated with increased CHD risk in this observational cohort, especially when started near menopause and used long-term, but this does not imply universal cardiovascular protection.
• Lifestyle modification remains crucial: avoid heavy smoking, maintain a healthy diet (higher fruit/vegetables, lower meat/sausage intake), manage BP and insulin resistance, and monitor lipids.
• Clinical decisions about HRT should weigh benefits for menopausal symptoms and bone health against cardiovascular risks, individual risk factor profiles, and patient preferences; avoid attributing cardiovascular risk solely to HRT without considering lifestyle factors.