Primary Prevention of Cardiovascular Disease With HRT

Kate Maclaran; John C Stevenson


Women's Health. 2012;8(1):63-74. 

In This Article

How Can the Differences Between the Observational Data & WHI be Explained?

The surprising findings of the WHI RCTs led to much critique of the studies to try and address why they had generated findings so out of keeping with previous preclinical and epidemiological data.

One suggestion has been that the observational studies would have been subject to several forms of bias and, therefore, the results must have been inaccurate. Of particular relevance in HRT studies is the effect of 'healthy user' bias. Women who take estrogen are generally from a higher social class, better educated and have fewer risk factors for CVD than nonusers.[33] In addition, users of HRT are more likely to be subject to regular health checks and take further precautions to maintain their general well-being. However, in the NHS,[27] bias did not appear to have a significant role as adjustment for confounding variables between users and nonusers did not affect the results, so this does not adequately explain the differing results with WHI. Another concern raised was that initial analyses of NHS data may have missed early cardiovascular events occurring in the 2 years following hormone initiation; however, further sensitivity analysis found that incomplete event capture did not account for the results of reduced CVD in HRT users.[34]

Crucial differences in the study populations are likely to help explain many of the discordant findings. The observational studies generally involved women who started HRT around the time of the menopause for symptomatic relief. Subjects tended to continue treatment consistently and were followed-up for a long duration, often 10–15 years. By contrast, women in the WHI studies were started on HRT at an advanced age (average 63 years), often with a significant delay following menopause. Furthermore, subjects had elevated BMI, were not using HRT for symptom relief (only 12–17% had moderate-to-severe vasomotor symptoms[9]) and generally had much shorter duration of treatment and follow-up. The importance of age at initiation was highlighted in an 'intention-to-treat' comparison between WHI and NHS, which showed that HRs are in much greater agreement when only comparing women who commence HRT within 10 years of menopause.[35]

The presence or absence of vasomotor symptoms in study populations is important as hot flushes are increasingly being recognized as a determinant of vascular health.[36] Hot flushes have been associated with risk factors for CVD, including increased cholesterol, elevated BMI and elevated blood pressure compared with nonflushers.[37] In addition, women with hot flushes have been shown to have adverse vascular changes such as impaired endothelial function,[38,39] increased aortic calcification[38] and higher carotid artery IMT.[40,41] Conflicting data exist[42] and further evidence is needed to help fully understand the mechanisms by which vasomotor symptoms may influence cardiovascular risk.

Differing durations of follow-up between observational studies and RCTs may impact on findings. The progression of fatty streaks to clinically significant plaques can take up to 5 years and, therefore, if HRT prevents formation of new plaques, it may take several years for the benefit of HRT to become evident. This is supported by animal studies which observed that estrogen inhibits initiation but not progression of established lesions in mice.[43]

Observational studies, which tend to have a much longer duration of follow-up, may therefore be better placed to detect benefit than RCTs with a shorter follow-up. This theory is supported by further analysis of the WHI estrogen-only arm, which demonstrated that lower cardiovascular event rates in women receiving estrogen compared with placebo only appeared to emerge from 7 years onwards (HR years 1–6: 1.08, 95% CI: 0.86–1.36; years 7–8: 0.46, 95% CI: 0.28–0.78).[44] Similarly, data from the WHI estrogen plus progestogen arm showed that CVD benefit only appears in younger women after at least 6 years; although this trend did not reach statistical significance.[45] These findings suggesting that cardiovascular benefit from HRT requires prolonged exposure are more consistent with preclinical and observational studies with longer follow-up. Furthermore, this has more widespread implications for HRT use, given that many guidelines recommend that HRT should be used for the shortest possible duration, often interpreted as less than 5 years.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.