Hormone replacement therapy and ischaemic heart disease: continuing questions but still no answers

Two years ago it was widely accepted that in both primary and secondary prevention of ischaemic heart disease, hormone replacement therapy (HRT) would provide a useful adjunct to existing treatments. This belief was backed up by overwhelming observational data, with more than 30 studies demonstrating benefit. Unassailable evidence in the form of a randomised control trial was not, however, available. In 1997, the Heart Estrogen Replacement Study (HERS) reported no effect on the incidence of major cardiac events in women with established ischaemic heart disease.1 This was a secondary prevention study in a group of 2763 women with established coronary disease, treated over 4.5 years with conjugated oestrogen (Premarin) 0.625 mg and medroxyprogesterone 2.5 mg in continuous combination. Since this publication there has been a reappraisal of the data regarding HRT, with a focus on the inherent bias in the observational data as an explanation for the cardioprotective effect of female sex hormones. Before HERS, most secondary prevention studies had been with unopposed oestrogen. Sullivan et al examined 1178 women (1108 non-users) with coronary angiograms demonstrating severe coronary artery disease.2 The study showed an 84% adjusted reduction in subsequent mortality among the oestrogen users. In primary prevention most studies of varying design in which unopposed oestrogen was the principal HRT underwent a meta-analysis by Stampfer and Colditz that demonstrated an overall risk reduction for oestradiol of 44%.3 This benefit does not appear to be attenuated by the addition of progesterone.4 Not all the data have been conclusively positive and it is instructive that the New England Journal of Medicine in 1985 carried articles from two well respected cohorts, the Framingham5 and Nurses health study populations,6 which yielded conflicting results. The former demonstrated an excess of risk of 50% in cardiovascular morbidity and the latter a mortality reduction of 50%. A just criticism of the observational data is the question of bias, the two most important sources being healthy user bias, i.e. sick women give up their HRT,7 and selection bias, where women who choose to take HRT are healthier and have a lower long-term cardiac risk.8, 9 These factors highlight the need for well-conducted randomised control-led trials. HERS is the first of these controlled trials to report and its overall negative result requires further exploration. The study was relatively underpowered to detect anything less than major differences in mortality and assumed that 50% of any HRT benefit was mediated through non-lipid effects; furthermore there was a large proportion of dropouts and crossovers (25%) in the HRT limb of the study. In the ‘as treated’ as opposed to ‘intention to treat’ analysis, there was a 13% non-significant risk reduction in favour of HRT. The most interesting point to come out of HERS was the difference in risk associated with different years of treatment. In the first year there was a 52% increased risk of reaching a cardiovascular endpoint. In subsequent years this increased risk was attenuated, translating into a risk reduction in the last two years of the treatment of 33%. The critical question that remains unresolved is whether the pattern of early risk with later benefit is simply due to high-risk patients reaching end-points early in the study, in which case the later benefit would be likely to slacken with longer duration of treatment. Alternatively, there may be more than one mechanism at play, in which case protracted treatment might in time translate into mortality benefit. The mechanism for early excess risk remains obscure but there are a number of proposals. HERS conclusively demonstrated an increase in risk of venous thromboembolism. It may be that spontaneous intra-arterial thrombosis was also more likely, resulting in early increased risk of myocardial infarction. Another hypothesis is that the documented effect of oestradiol on cellular adhesion molecules, previously felt to inhibit atherogenesis, may in fact result in a reduced structural integrity of the atheromatous plaque resulting in an increased risk of plaque rupture and subsequent myocardial infarction.10 Steiner et al, in this issue (pages 475–477), highlight the potential importance of this early increased risk. In their case the initiation of HRT demonstrated a good temporal relationship with the development of an acute myocardial infarction.11 The dose of Premarin in this patient was higher than in HERS and the possibility of a dose response must be considered. The presence of the subsequent normal coronary arteriogram suggests the possibility of spontaneous intra-arterial thrombosis to explain this myocardial infarction, rather than plaque rupture. Vascular spasm is also less likely as oestrogens possess predominantly vasorelaxative properties. Many questions remain to be answered. It may be that with prolonged use of HRT the early excess of cardiac risk will be outweighed by subsequent benefit. This is a more important question for women who have not yet developed heart disease, who have a better prognosis than women with coronary disease. Indeed, HRT may prove to be a paradoxical form of treatment which works better as primary than as secondary prevention. The forthcoming Women's Health Initiative and Medical Research Council studies will, it is hoped, resolve these issues. The effect of initiating HRT early after an acute event such as a myocardial infarction is also not addressed by HERS. Finally, HRT does not represent a homogenous form of treatment and the effect of alternative progestins, alternative routes of delivery and other agents such as tibolone and selective oestrogen receptor modulators remains to be elucidated. But how, at present, should women be advised regarding HRT and the heart? In primary prevention it is still reasonable to advise that long-term use may be associated with reduced long-term cardiac risk, particularly in those without major cardiac risk factors in whom the early excess in risk is likely to be negligible. Women with ischaemic heart disease who are already taking HRT need not stop therapy, because presumably they have survived early excess risk and will experience only benefit. Women with heart disease who have an important reason for taking HRT may take it, as over five years the effect on risk will be neutral. At present, however, outside the context of a clinical trial, HRT should not be given to women with ischaemic heart disease specifically with the objective of reducing long-term cardiac risk.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.