Hormone replacement therapy and acute coronary outcomes: methodological issues between randomized and observational...

Area of Science:

• Cardiovascular Science
• Clinical Trials
• Epidemiology

Background:

• Observational studies and basic research suggested hormone replacement therapy (HRT) protects against acute coronary outcomes.
• The Women's Health Initiative (WHI) randomized trial reported a 29% increased risk of acute coronary outcomes with estrogen-progestin therapy.
• Discrepancies between observational data and the WHI study necessitate investigation into potential biases.

Purpose of the Study:

• To explore methodological reasons for the conflicting results regarding HRT and acute coronary outcomes.
• To assess the impact of potential biases on the findings of both observational studies and the WHI trial.

Main Methods:

• Analysis of potential methodological flaws in observational studies, including healthy user bias and reference group selection.
• Evaluation of potential differential unblinding in the WHI study and its effect on outcome detection rates.
• Statistical modeling to adjust for identified biases and re-evaluate risk ratios.

Main Results:

• Alterations in diagnostic patterns due to unblinding in the WHI study could reduce the observed risk ratio from 1.28 to 1.02.
• Adjusting for healthy user bias and inappropriate reference groups in observational studies increased the rate ratio from 0.67 to 0.82.
• Both observational and randomized trial data may be influenced by significant biases.

Conclusions:

• Conflicting findings on HRT and acute coronary events may stem from biases inherent in both observational studies and the WHI trial.
• Differential unblinding and altered diagnostic patterns are potential sources of bias in the WHI study.
• Healthy user bias and suboptimal reference group selection may inflate protective effects in observational studies.