Human Reproduction Update Advance Access originally published online on June 28, 2006
Human Reproduction Update 2006 12(5):483-497; doi:10.1093/humupd/dml028
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Hormones and cardiovascular health in women
The ESHRE Capri Workshop Group1To whom correspondence should be addressed: P.G.Crosignani, II Department of Obstetrics and Gynecology, University of Milano, Via Commenda 12, 20122 Milano, Italy. E-mail: piergiorgio.crosignani{at}unimi.it
 |
Abstract |
|---|
Abstract
IntroductionCardiovascular diseases (CVDs) may have their origin before birth: the combination of being small at birth and having an overly rich post-natal diet increases the likelihood of obesity and of acquiring a specific metabolic syndrome in adulthood that carries an increased risk of CVD. The incidence of CVD and mortality is very low in women of reproductive age but rises to a significant level in older women. In this article, we discuss CVD in relation to hormonal contraception, pregnancy and polycystic ovarian syndrome (PCOS) in younger women and menopause in older women. Women with PCOS have a higher risk of diabetes and hypertension, but studies to date have not shown an effect on CVD events. Use of combined hormonal contraception has only small effects on CVD because of the low baseline incidence of myocardial infarction (MI), stroke and venous thromboembolism (VTE) among young women. Women with existing risk factors or existing CVD, however, should consider alternative contraception. In pregnancy, CVD is rare, although, in the West, it now accounts for a significant proportion of maternal mortality as the frequency of obstetrical causes of mortality has substantially declined. The frequency of VTE is 15 per 10 000 during pregnancy and the post-partum period. In older women, menopause causes a slightly higher risk of MI after allowing for age, although there is substantial heterogeneity in the results of studies on menopause and age at menopause and MI. A larger effect might have been expected, because estrogen reduces the risk of developing atherosclerosis in premenopausal women, whereas in post-menopausal women who may have established atherosclerotic disease, estrogen increases the risk of myocardial disease through the effects on plaque stability and clot formation. Recent trial results indicate that hormone treatment in menopause does not favourably affect the risk of MI, stroke or other vascular disease. Thus, prevention of CVD should rely on diet and fitness, low-dose aspirin and treatment of hypertension, hyperglycaemia and hyperlipidaemia.
Key words: cardiovascular disease / contraception / menopause / polycystic ovary syndrome / pregnancy
|
Introduction |
|---|
Cardiovascular deaths (CVDs) due to myocardial infarction (MI), stroke and venous thromboembolism (VTE) account for the leading number of deaths among women as well as men. Coronary heart disease (CHD) due to atherosclerosis is the cause of MI and is the leading cause of death in men and women worldwide. Strokes due to venous or arterial thrombosis are more frequent in women of menopausal age, whereas strokes due to cerebral haemorrhage, although less common, are also seen in younger women because of their frequent origin in cerebrovascular anomalies. Venous thrombosis includes superficial thromboses that are self-limited and deep thromboses, most commonly in the popliteal and femoral veins. In about 10% of cases, part or all of a thrombus may detach and form a pulmonary embolus.
CHD arises in the coronary arteries, where atherosclerotic lesions evolve from an initial accumulation of foam cells in the arterial endothelium leading to a fatty streak. This is followed by the accumulation of fatty deposits including cholesterol, creating a true atheroma. Once an atheroma has formed, collagen in the fibrous cap stabilizes the plaque and prevents the plaque from rupturing. Matrix metalloproteinases (MMPs), which are produced by the inflammatory cells in the atheroma, degrade collagen; if the capsule then ruptures, the resulting thrombus may occlude the coronary artery (Pepine, 1998
; Heistad, 2003) (Figures 1 and 2).
|
|
Estradiol (E2) reduces the development of early lesions of atherosclerosis, in part through the effects on lipid metabolism which reduce lipid deposits in the endothelium. Once the atheroma is established, however, estrogens increase MMP expression, which may promote disruption of the fibrous cap and subsequent rupture of the plaque. If the capsule does rupture causing turbulent blood flow, E2 is thrombogenic and clot formation may occlude the arterial lumen (Phillips and Langer, 2005). Through different mechanisms, therefore, estrogens inhibit early development of atherosclerosis but increase the risk of damage once atherosclerosis has been established. Atherosclerotic lesions in the carotid and cerebral vessels may be affected by similar mechanisms, so that in comparison with men, women are relatively protected from thrombotic stroke before menopause. With respect to VTE, any impact from hormones is through changes in coagulation, anticoagulation and thrombolytic factors.
The clinical issues regarding cardiovascular disease (CVD) in women concern the effects during pregnancy, the cardiovascular prognosis with polycystic ovarian syndrome (PCOS), questions about cardiovascular risks with hormonal contraception and the quandary over whether hormonal treatment of menopausal symptoms increases or decreases cardiovascular risk.
|
Incidence of CVD |
|---|
The World Health Organization has collected mortality data for MI, stroke and VTE from most countries around the world (http://www3.who.int/whosis/mort/). These data show that rates vary in different parts of the world because both incidence and mortality of CVDs are influenced by many factors.
MI mortality in women rises exponentially with age (Figure 3—upper left panel). Because there is a fairly constant 50% mortality to case ratio in developed countries, the mortality trends for these countries, at least, reflect the incidence of MI (Farley et al., 1998). MI mortality is more than 2-fold higher in the Americas than in the Western Pacific. MI mortality rates are low in the reproductive years, ranging from one to seven per 100 000 women per year at ages 35–44 years. Women of reproductive age have 3- to 5-fold lower rates than men, and although MI mortality rates are more similar beyond the age of 65 years, they remain lower in women at all years in all regions (Figure 3—upper right panel).
|
Stroke mortality also rises exponentially with age, and the rates are lowest in developed countries and 2- to 3-fold higher in the so-called countries with ‘Economies in Transition’ (the countries in former socialist block of Eastern Europe) than in developed countries (Figure 3—left lower panel). Stroke mortality rates are low in the reproductive years, ranging from four to 12 per 100 000 women per year at ages 35–44 years. Women have lower rates than men by 30–50%, except in the highest age group (Figure 3—lower right panel).
As with other CVDs, VTE mortality rates also rise exponentially with age. VTE mortality rates are much lower than for MI or stroke, in part because the VTE case mortality ratio is <2% (Farley et al., 1998). The highest regional rates of VTE mortality are in countries with ‘Economies in Transition’ which are 2- to 3-fold higher than in developed countries. VTE mortality is less than one per 100 000 women per year at ages 35–44 years. The low mortality rates mean that any difference that might exist between age-specific mortality rates in women and men is obscured by chance variation in the national statistics.
To put mortality rates per 100 000 women into perspective, note that until the age of 35 years, mortality rates for CVDs are less than those which occur due to accidents of all kinds (10 per 100 000 deaths) (http://www3.who.int/whosis/mort/).
|
The metabolic syndrome: developmental origins |
|---|
Recently, a new syndrome called ‘the metabolic syndrome’ has been identified (Grundy et al., 2004
). It involves abdominal obesity, raised serum cholesterol, hypertension, insulin resistance with or without impaired glucose tolerance, a pro-inflammatory state with raised C-reactive protein and a pro-thrombotic state (raised plasma fibrinogen and clotting factors) (Grundy et al., 2004). Because patients with metabolic syndrome are at high risk of developing CHD, understanding its aetiology, e.g. why smaller babies have a higher adult risk of metabolic syndrome, may help to clarify many risk factors for heart disease.
More than 90% of human existence and evolution as a species occurred before the development of agriculture and its associated cultural and dietary changes (Figure 4) (Cordain et al., 2002). Thus, understanding human biology and the capacity to respond to the environment must take evolutionary history into account. Knowledge about the recent past is also important because such life-history approaches, particularly the environmental conditions experienced by parents and grandparents and in early life, provide additional insights into biological risks (Figure 5). These elements combine into the concept of ‘developmental origins of adult disease’ (Gluckman and Hanson, 2004, 2005b).
|
|
One genotype can generate a range of phenotypes through the processes of developmental plasticity. In mammals, developmental plasticity involves epigenetic processes, including DNA methylation (Lillycrop et al., 2005), histone acetylation and perhaps small RNAs. These mechanisms induce important developmental influences on the regulatory processes that control blood pressure, metabolism and hormonal responses.
A key concept in evolutionary biology is a phenotypic adaptive immediate advantage, where responses are induced which make that particular organism more likely to survive and reproduce successfully. However, there is a second class of response in which the adaptive advantage is deferred, in the expectation of the future environment, introducing the concept of predictive adaptive responses (Gluckman and Hanson, 2004; Gluckman et al., 2005).
For the fetus, prediction of the future environment depends on transplacental signals from the mother: these are mainly nutritional or hormonal in origin. If nutrient delivery from the mother is poor, the fetus will predict a deficient post-natal environment and develop its metabolic regulation accordingly.
Epidemiological studies have shown that the risk of developing metabolic syndrome is inversely related to birthweight (Table I). This correlation within the normal weight range emphasizes that the phenomenon is associated with developmental plasticity rather than developmental disruption. Birth size itself is not part of the causal pathway but is an indirect measure of the quality of the intrauterine environment. Signals other than those affecting fetal growth can nonetheless induce predictive adaptive responses. After fetal life, there are childhood mechanisms, as the risk is greater in those who show rapid weight gain or earlier adiposity rebound in childhood (Barker, 1998).
|
Fetal growth is not regulated solely by fetal genetic factors. ‘Maternal constraint’ matches fetal growth to a size appropriate for vaginal delivery (Gluckman and Hanson, 2005a). Fetal nutrition is the result of maternal nutrition, maternal metabolism, uteroplacental blood flow, placental metabolism and transfer to the fetus. The fidelity with which information about the nutritional environment of the mother is transferred to the fetus can be low, particularly in maternal or placental disease where fetal nutrition is reduced. Hence, the fetus will match its development to the expectation of living in a nutritionally deprived environment. This would have conferred an adaptive advantage in the evolving hominid because it would bias development towards the assumption of poor nutrition in post-natal life. Maternal constraint however is now disadvantageous, especially as nutrition becomes more abundant and lifestyle more sedentary.
Experimental and empirical evidence
In experimental animals, post-natal changes in metabolic and cardiovascular control, in appetite, exercise propensity and body composition, can be readily induced by manipulating maternal nutrition or administering glucocorticoids to the mother. These manipulations need not reduce birth size. An interaction between the prenatal and post-natal environments can be demonstrated: for example, the nature of the prenatal environment determines how the animal responds post-natally to high-fat diet. In some cases, these models mimic human metabolic syndrome. In the forced experiment of the Dutch winter famine of 1944–45, the fetuses exposed to famine in late gestation had a higher risk of later developing insulin resistance and obesity. Empirically, prematurity is another potential response to a deficient intrauterine environment, and it is not surprising that it may also lead to long-term consequences. In addition to the effects of nutritional deprivation in the first trimester of pregnancy, during development there may be the effects of environmental influences persisting from a previous generation (Drake and Walker, 2004). Intriguingly, these effects may be transmitted through either male or female intermediate generations, suggesting that not all epigenetic changes are erased at meiosis. Such non-genomic inheritance mechanisms make evolutionary sense in assisting a species to survive a transient environmental change lasting longer than one reproductive cycle.
In affluent populations, the focus should remain on promoting exercise and healthy diets in children, adolescents and adults. But in developing societies, significant health gains may result from strategies that improve the health of young girls and women of reproductive age, before and during pregnancy.
|
PCOS as a metabolic disease |
|---|
PCOS should also be viewed as a metabolic disease with an increased risk of type 2 diabetes and CVD (Guzick, 2004
). According to a recent international consensus on diagnostic criteria (The Rotterdam ESHRE/ASRM-Sponsored PCOS Consensus Workshop Group, 2004), PCOS is characterized by oligomenorrhoea/amenorrhoea, hyperandrogenaemia and polycystic ovaries. In addition, PCOS is frequently associated with obesity and insulin resistance. Although women with PCOS seek medical care for irregular menstrual cycles, hirsutism or infertility, the appreciation of their increased CVD risk has changed the clinical emphasis to include prevention of long-term disease conditions such as diabetes and heart disease (Sharpless, 2003).
Recent clinical studies in PCOS focusing on surrogate outcomes for CVD have established the existence of metabolic risk factors in these women. The surrogate outcomes included endothelial dysfunction, platelet dysfunction, increased leucocyte counts or C-reactive protein levels, increased coronary artery calcification and carotid intima-media thickness. A study in 161 PCOS patients also concluded that components of the metabolic syndrome are common in PCOS (Apridonidze et al., 2005). With respect to intermediate outcomes, 33 women of 40–59 years of age who had histologically proven PCOS at wedge resection and were followed for 22–31 years had more central obesity and an increased prevalence of diabetes and hypertension compared with age-matched controls (Dahlgren et al., 1992).
Regarding cardiovascular event outcomes, although a correlation with the metabolic syndrome has been clearly established, the limited studies conducted in women with PCOS yielded conflicting results (Table II). It must be stressed that most studies are retrospective, with small numbers of patients, short periods of follow-up and many questions about the control groups (Wild, 2002; Legro, 2003). Large prospective cohort studies such as the Framingham or the Nurses Health Study, which focused on hard outcomes such as heart disease or cancer, have failed to identify hyperandrogenaemia and anovulation as a separate risk phenotype. A retrospective study from the UK involving 786 PCOS women diagnosed between 1930 and 1979 failed to establish increased cardiovascular mortality (Pierpoint et al., 1998) or morbidity (Wild et al., 2000).
|
Hence, despite the existence of classical risk factors for CVD, such as diabetes, hypertension and dyslipidaemia, in PCOS and the impact of PCOS on surrogate outcomes for CVD, studies to date have failed to establish an effect of PCOS on cardiovascular markers (Table III). Well-designed, sufficiently powered and closely matched long-term follow-up studies are warranted to evaluate potential long-term health consequences in PCOS. However, this should not prevent the physician from focusing on the importance of life style and weight reduction in these high-risk women. Further clinical proof of the usefulness of preventive medication (such as insulin-sensitizing agents) is required before this approach should be recommended to women in clinical practice.
|
|
Hormonal contraception and CVD |
|---|
CVD is uncommon among women of reproductive age. The first report of CVD as a side effect of contraception occurred in 1961 soon after the combined oral contraceptives (COCs) came onto the market (Jordan, 1961
). Since the 1960s, the doses of both estrogen and progestogen in the contraceptive pill have been dramatically reduced and new, theoretically safer, progestogens have been developed. Substantial data exist on the effects of the COC on CVD, but there is less evidence for the effects of progestogen only contraceptives (POCs).
A recently published meta-analysis of 23 studies of COC use and MI estimated an overall odds ratio (OR) of 2.5 [95% confidence interval (CI): 1.9–3.2] for current users compared with never users (Khader et al., 2003). The risk was directly related to the dose of estrogen but, nevertheless, still increased for women using low-dose pills. The Nurses’ Health Study reported an increased risk of developing MI (95% CI: 1.0–9.2) among women with no known risk factors for CVD (Stampfer et al., 1990). Smoking and hypertension both substantially increase the risk of MI among COC users, and some data suggest an increased risk among women with diabetes, hypercholesterolaemia or a history of pregnancy-induced hypertension or pre-eclampsia. The MI risk does not appear to be modified by age, nor is it associated with past use. There is still debate over a differential risk of MI in association with the different types of progestogen used in hormonal contraception.
The risk of ischaemic stroke is increased among current users of the combined pill [OR 2.7 (95% CI: 2.2–3.3)] (Gillum et al., 2000; Chan et al., 2004). The authors of one review commented that a true association between low-dose COC use and stroke was doubtful because of the low magnitude of the ORs, the methodological limitations of the studies and the ORs of <1.0 in the cohort studies (Chan et al., 2004).
There is some evidence for an increased risk among past users but no apparent relationship with duration of use (World Health Organization Collaborative Study of Cardiovascular Disease and Steroid Hormone Contraception, 1996a). Data on the effect of estrogen dose and progestogen type are conflicting. There is general agreement that smoking and hypertension increase the risk of ischaemic stroke among pill-users and the risk is also increased among women who suffer from migraine (Etminan et al., 2005). The combined pill is relatively contraindicated for women with diabetes or obesity, and therefore data are scarce; however, one study reports an increased risk of ischaemic stroke among women with obesity or hypercholesterolaemia (Kemmeren et al., 2002). Most studies have found no statistically significant increase in the risk of haemorrhagic stroke among COC users without other risk factors. There is some suggestion that the risk may be increased among women aged over 35 years, smokers and women with hypertension (World Health Organization Collaborative Study of Cardiovascular Disease and Steroid Hormone Contraception, 1996b).
Combined hormonal contraception delivered by a different route (transdermal, vaginal and particularly injectables) may have a different effect on the risk of MI and stroke, but data are scarce.
Data on heart attack and stroke with the use of progestogen-only contraception are limited but to date, demonstrate no increased risk. This is reassuring because women with other risk factors for CVD are often prescribed POC preferentially.
VTE is much more common among women of reproductive age than stroke or heart attack but much less likely to be fatal. Combined pill users with no other risk factors for CVD are at increased risk of VTE. The risk is 3- to 6-fold higher than that of non-users and is highest in the first year of use, probably as a result of the unmasking of unrecognized thrombogenic mutations. The magnitude of VTE risk is greatest for carriers of factor V Leiden mutations (up to 20 times the risk of women with the mutation who do not take the pill). However, because the prevalence of mutations is low and the number of women using the pill enormous, routine screening is not cost-effective.
The risk of VTE among COC users is probably increased by obesity (Nightingale et al., 2000) and may be increased by immobilization (Sydney et al., 2004) and air travel (Martinelli et al., 2003), but not with smoking or hypertension. The debate over a differential risk with different types of progestogen still rumbles on, but if there is a difference, it is small (one to two extra cases/10 000 women years) (Kemmeren et al., 2001).
Studies of progestogen-only contraception and VTE risk have all demonstrated either no effect or a non-significant effect (World Health Organization Collaborative Study of Cardiovascular Disease and Steroid Hormone Contraception, 1998).
Combined hormonal contraception delivered transdermally (and perhaps by other non-oral routes) may be associated with a lower risk of VTE as has been suggested for hormone replacement therapy (Scarabin et al., 2003), but as yet, there is no evidence to support this.
The absolute risk of any cardiovascular event is small and must be balanced against the risk of pregnancy itself (associated with an increased risk of CVD). Where contraceptives are simultaneously used for the relief of other conditions (such as menorrhagia, dysmenorrhoea or hyperandrogenism), the risk benefit ratio changes.
|
CVD during pregnancy |
|---|
MI in pregnancy
The incidence of MI during pregnancy is increasing in Western populations. The risk is considered to be around 1 in 10 000 women, but this may be an underestimate (Wen et al., 2005). Most MIs occur in multigravid women, and case maternal mortality rates in pregnancy range from 7 to 21%. Although infarction may occur at any point during gestation, the majority of deaths occur at, or within 2 weeks of, the infarction and often during labour or delivery. In subjects in whom a post-mortem was carried out, coronary atherosclerosis was evident in 43% of subjects and probable coronary thrombosis, without evidence of atherosclerotic disease, in 21%. Atherosclerosis was more commonly found in subjects with an ante-partum MI than in those with a post-partum event. Dissection of the coronary artery was found in 16% of subjects and was more often associated with infarction in the puerperium (Clark and Greer, 2005). No detectable evidence of coronary occlusion was found in around 30% of ante-partum and 75% of post-partum events. The absence of detectable occlusive disease may indicate a significant role of transient coronary artery spasm in these subjects. It is likely that the physiological prothrombotic changes in coagulation increase the risk of arterial thrombosis. In cases of coronary dissection, the aetiology remains unknown but may relate to changes in arterial histology, such as alteration in elastic and reticular fibres and muco-polysaccharide composition, which is a feature of normal gestation. This may be compounded by pre-existing risk factors for CHD, increasing blood volume/heart rate, myocardial oxygen requirements and anaemia (Clark and Greer, 2005; Kaaja and Greer, 2005).
The management of MI in pregnancy is comparable with that of non-pregnant subjects (Ray et al., 2004; Verhaert and Van Acker, 2004). Although thrombolytic therapy is usually contraindicated in pregnancy, there is no evidence that streptokinase crosses the placenta. The majority of reports using streptokinase or tissue plasminogen activator have shown a favourable fetal outcome, although maternal haemorrhage and an increased risk of pre-term delivery and fetal loss have been reported. Angioplasty and stents can be used. The risk benefit ratio in MI is clearly important to consider. Low-dose aspirin, low molecular weight heparin (LMWH) and clopidogrel can be used. Statins are generally avoided, especially in the first trimester. If possible, delivery should be postponed until 2–3 weeks after the acute event. It is usually preferable to plan delivery by Caesarean section to avoid the extra cardiac demands of labour. The risk associated with future pregnancy is dependent on the aetiology of the event, the presence of persisting myocardial ischaemia and the degree of residual impairment of left ventricular function.
There is a great variation in the estimate of risk of ischaemic stroke during pregnancy. This may be due to an absence of computerized tomography (CT) scanning in early studies, a lack of distinction between venous and arterial events, small study size and referral bias. This results in a quoted risk of ischaemic stroke in pregnancy varying from 1 in 5000 to 1 in 20 000 pregnancies, with the latter figure suggesting that pregnancy may be associated with only a marginal risk when compared with non-pregnant subjects (at approximately five per 100 000 per women-years). Most strokes occur in the third trimester or puerperium, and in over 50%, there may be evidence of major arterial occlusion. In a substantial number of cases, no evidence of major occlusion is found at angiogram or post-mortem (Sibai and Coppage, 2004). A number of common risk factors relate to stroke in pregnant subjects as they do in non-pregnant subjects. These include hypertension, smoking, premature atherosclerosis and valvular heart disease. Pre-eclampsia and eclampsia can also be responsible for stroke. Stroke in pregnancy is also more common in older women, in subjects with sepsis and in those who have undergone Caesarean delivery. In addition, rarer causes, such as choriocarcinoma, paradoxical embolism, sickle cell disease, lupus inhibitor-related thrombosis and homocysteinuria, have also been reported to underlie pregnancy-related stroke.
A fatal maternal outcome occurs in 0–26% of cases. Risk of recurrent stroke is around 2-fold higher in the puerperium when compared with the ante-partum period. The overall outcome of subsequent pregnancies was similar, however, to that expected in the general population (Clark and Greer, 2003).
Low-dose aspirin therapy is employed during pregnancy, and the delivery should be managed to minimize the risk of venous and arterial thrombosis. In particular, a prolonged second stage of labour should be avoided and, if there is reduced mobility due to hemiplegia, there may be a requirement for LMWH thromboprophylaxis (Greer, 2004).
Epidemiology and risk factors for VTE in pregnancy
The incidence of VTE is around 15 per 10 000 pregnancies, approximately one-fourth are pulmonary embolism and half occur post-partum (Macklon and Greer, 1996; Heit et al., 2005). Pulmonary embolism is the major cause of maternal death in the UK (Confidential Enquiry into Maternal and Child Health, 2004). In addition to mortality and acute morbidity, pulmonary embolism also increases the risk of future deep venous thrombosis (DVT) and venous insufficiency (McColl et al., 2000). The incidence of non-fatal events in pregnancy has been difficult to gauge due to problems with screening studies. The risk of DVT after Caesarean section has been estimated at 1–2%. Symptomatic DVT in pregnancy is usually left-sided (85 versus 55% in the non-pregnant) and ilio-femoral (72 versus 9% in the non-pregnant) (Greer, 1999). Previous thrombotic event(s), age, obesity, thrombophilia and operative delivery are the most important risk factors (Table IV). In a 10-year review of 650 000 pregnancies in Scotland, the antenatal incidence of DVT was greater than post-natal incidence: the rate doubled with age over 35 years and with operative delivery (Macklon and Greer, 1996; Greer, 2000).
|
The pathophysiology of venous thrombosis in pregnancy is associated with the procoagulant changes of pregnancy, venous stasis and trauma to the pelvic veins at delivery (Clark et al., 1998; Greer, 1999). Congenital thrombophilia underlies up to 50% of pregnancy-related thromboses. These include antithrombin (AT), protein C or S deficiencies, activated protein C resistance with underlying factor V Leiden and the prothrombin gene variant (McColl et al., 1997). Interestingly, hyperhomocysteinaemia due to the MTHFRC677T mutation is not a risk factor for VTE in pregnancy, possibly because homocysteine concentration is lower in pregnancy and folic acid, which corrects hyperhomocysteinaemia, is usually supplemented in pregnancy. In a study involving more than 90 000 pregnancies, the risks of VTE in the presence of thrombophilia were 1/437 for factor V Leiden, 1/2.8 for AT deficiency type 1, 1/42 for AT deficiency type 2 and 1/113 for protein C deficiency, consistent with other studies (McColl et al., 2000; Martinelli et al., 2002). Acquired thrombophilic traits such as anticardiolipin antibody syndrome also place the patient at risk. Acquired resistance to activated protein C occurs as a feature of normal pregnancy, and this may be one of the key changes in the coagulation system in pregnancy, changes that predispose the mother to thrombosis.
Despite the association between thrombophilia and VTE and pregnancy complications, there is no evidence at present to support routine universal screening of all pregnant women for thrombophilia. The natural history of many of these conditions, particularly in asymptomatic women, is not established, and the appropriate intervention remains unclear. Universal screening for factor V Leiden mutation in pregnancy or a range of other thrombophilias is not cost-effective. However, selective screening of women with VTE in pregnancy or who have a personal or family history of VTE may be useful, as around 50% of such women will have a thrombophilia (Clark et al., 2002; Wu et al., 2005).
Anticoagulants for VTE in pregnancy
Treatment and prophylaxis of VTE requires the use of anticoagulants, for which there are special considerations in pregnancy. Warfarin is associated with teratogenesis and a significant risk of bleeding in utero, particularly at delivery. In the longer term, there is an association with neurodevelopmental problems (Greer, 1999; Wesseling et al., 2001). Unfractionated heparin also causes particular problems in pregnancy, such as allergic reactions, thrombocytopaenia and heparin-induced osteoporosis. More than 2% women on long-term heparin will have symptomatic vertebral fractures (Greer, 1999; Bates et al., 2004). LMWH is now used extensively for prophylaxis and treatment in pregnancy because there are substantially fewer side effects such as heparin-induced thrombocytopaenia and osteoporosis (Pettila et al., 2002; Greer and Nelson-Piercy, 2005). LMWH may be combined with graduated elastic compression stockings (Greer, 2004). Dextran has been associated with severe anaphylactoid reactions at Caesarean section, which can precipitate perinatal death or severe neurological handicap and so should be avoided (Greer, 1999). Treatment regimens depend on the patient’s history and the risk factors in each case.
Women undergoing Caesarean section and vaginal delivery should have a risk assessment for VTE. For Caesarean section, LMWH thromboprophylaxis and graduate elastic compression stockings are indicated with any additional risk factors such as emergency section in labour, age over 35 years or high BMI. With vaginal delivery, LMWH prophylaxis is indicated if there are two or more additional risk factors or a major risk factor such as morbid obesity. Prolonged prophylaxis should be considered in women with significant ongoing risk factors as many VTEs occur following discharge from hospital (Greer, 2004).
If VTE is suspected, objective diagnosis is essential, and duplex ultrasound venography and ventilation-perfusion lung scans are the first-line investigations. However, investigations such as spiral CT are increasingly being used for pulmonary thromboembolism.
|
Menopause, age at menopause and CHD |
|---|
The hypothesis that menopause and consequent biological modifications are related to the risk of CHD derived from the observation that incidence and mortality rates for CHD in women are substantially lower than in men before menopause but tend to rise approaching those of men at older ages (Heller and Jacobs, 1978
) (Figure 3—right upper panel). However, it is difficult to disentangle the effect of age from that of menopause on CHD, because the two variables are strongly related and an apparent higher risk from menopause may simply be due to the rise of CHD incidence and mortality with increasing age.
The overall epidemiological evidence on the relationship between menopause per se rather than age and CHD is still controversial. Most information derives from eight cohort studies and three case–control studies. As for the cohort studies, the 20-year follow-up of the Framingham Study showed a 2-fold increase in relative risk (RR) in post-menopausal versus premenopausal women (Kannel et al., 1976); in the 24-year follow-up of the same cohort, based on 43 cases of fatal and non-fatal MI, the MI incidence rate was 1.4 in premenopausal and 3.9 in post-menopausal women (Gordon et al., 1978). In a cohort of Swedish women (Lapidus et al., 1985), based on 25 cases of MI, the RRs were 2.0 (95% CI: 0.2–19.1) for women aged 
40 years at menopause, 2.2 (95% CI 0.7–7.4) and 1.4 (95% CI, 0.5–3.8) for women aged 45 and 50 years, compared with premenopausal women. In the 6-year follow-up of the American Nurses’ Health Study (Colditz et al., 1987), after strict allowance for age, compared with premenopausal women, never HRT users with natural menopause had an RR of 1.1, and those with surgical menopause had an RR of 1.7. In a Dutch study of 12 195 women including 824 deaths from CVD (van der Schouw et al., 1996), the overall RR was 0.982 per year delay of menopause, and the inverse relation was greater at younger age. In a study from Norway, including 2767 cases of CHD (Jacobsen et al., 1997), the RR was 0.84 (95% CI: 0.65–1.08) for women aged 
53 years at menopause compared with those aged <40 years. In the US National Health and Examination Survey (NHANES) I Study, based on 84 cases of fatal acute MI, a moderate and not significant association was observed with age at menopause, the RR being 1.50 (95% CI: 0.67–3.36) for women with menopause at age <40 years compared with 50 years (Cooper and Sandler, 1998). In the Menstruating and Reproductive History Study, Cooper et al. (1999) found RRs of 3.2 in women with natural menopause and of 2.7 in those with surgical menopause at age 45 years, compared with women with natural menopause when aged 51 years. At the 13-year follow-up of the California Seventh-Day Adventist Study, including 308 cases of fatal CHD, Jacobsen et al. (1999) found increased risks of CHD in women with menopause either at young (35–40 years) or later age (56–60 years), the association being stronger in non-HRT users.
At least three case–control studies investigated the relation between menopause and MI. Two of these were from North America. Rosenberg et al. (1983) in a study involving 255 non-fatal MI cases found an OR of 0.6 (95% CI: 0.4–1.0) in post-menopausal women compared with premenopausal ones after allowance for age, and Palmer et al. (1992) in a study involving 858 non-fatal MI cases found an OR of 0.5 (95%CI: 0.2–0.9) in pre- and peri-menopausal women compared with post-menopausal ones, with higher risks in women with menopause at age <45 years than in women with menopause at age 50 years.
In an Italian case–control study of 429 incident cases of MI and 863 controls, post-menopausal women were not at higher risk of MI than pre-/peri-menopausal ones, after adjustment for age and other selected covariates (multivariate OR, 0.99) (Table V). With reference to age at menopause, compared with women reporting menopause when <45 years, the multivariate ORs were 1.54 for those aged 45–49 years at menopause, 1.36 for those aged 50–52 years and 0.97 for those aged 53 years, in the absence of any trend in risk. No meaningful relation emerged with time since menopause (OR 0.85 for <10 years since menopause). The results were similar in women aged <60 and 60 years at MI (Fioretti et al., 2000).
|
Thus, there is some suggestion that, after allowing for age, post-menopausal women are at higher risk of CHD, although there is substantial heterogeneity in the results across various studies on menopause, and age at menopause and CHD. This is not easily explained by the different type of studies (cohort or case–control), the inclusion of fatal or non-fatal diseases, the inclusion of hormone therapy users, the cut-points selected for age at menopause and other identified factors. Part of these discrepancies may depend on difficulties in the collection and analysis of epidemiological data on menopause. Besides uncertainties in the definition of the peri-menopausal period, age at menopause is difficult to establish in women after hysterectomy and in those using HRT. Moreover, and similar to any other time factor, it is important to make an extremely strict age-adjustment to obtain an unbiased quantification of risk (Pike, 1987). The studies are sufficiently robust, however, to exclude a strong effect of menopause and age at menopause on CHD (as there is for breast cancer), and the effect is probably limited to a short period after menopause.
|
Effects of menopausal hormonal treatment on the cardiovascular system |
|---|
Although the epidemiological studies do not show a large immediate effect of menopause on CHD events, ovarian hormone deprivation after menopause is associated with an increase in CVD risk factors (Gordon et al., 1978
; Colditz et al., 1987). Estrogen given soon after the menopause improves these cardiovascular risk factors and cardiovascular function (Greendale et al., 1996; Herrington et al., 1999; Vitale et al., 2001). A large body of evidence suggests that hormone replacement therapy may be protective for atherosclerosis and other cardiovascular risk factors (Greendale et al., 1996) but may also be thrombogenic if given late after the menopause, as suggested by the excess risk after starting use in the Heart and Estrogen/progestin Replacement Study (HERS) (Daly et al., 1996; Jick et al., 1996; Perez Gutthann et al., 1997) and by the data of the Women’s Health Initiative (WHI) (Rossouw et al., 2002). Other potential effects include modifications of plasma lipid profile (Hulley et al., 1998), fibrinolysis (Petitti, 1998) and blood pressure (Hazzard, 1989). However, the potential role of such modifications on the risk of CHD remains undefined and warrants an exploration of the role of estrogen and progestogen on the cardiovascular system.
Some cardiovascular effects of estrogen may be counteracted by progestogens. Progesterone receptors are present in the arterial wall, and there is evidence that the arterial effects of progestogens are mediated through progesterone receptors as well as through down-regulation of E2 receptors. Progestogen therapy can stabilize arteries in a state of vasomotor instability but may also induce vasoconstriction of estrogenized vessels and precipitate arrhythmia. According to their chemical structure, progestogens have different metabolic and vascular effects that may enhance or abolish the effects induced by estrogen therapy on cardiovascular risk factors and on vascular functions (Adams et al., 1997).
Lipid, glucose and insulin metabolism are improved by estrogen replacement therapy (Hulley et al., 1998), but this effect may be reversed by the combination of estrogen with androgenic progestogens, whereas combination with non-androgenic progestogens has a more favourable metabolic profile.
Estrogens improve endothelial function and vascular reactivity and reduce the progression of coronary atherosclerosis both in animals and in early post-menopausal women (Williams et al., 1994; Darling et al., 1997; Sbarouni et al., 1998; Vitale et al., 2001). When administered in combination with estrogens, progestogens may, in some instances, interfere with the endothelial effect of estrogens. In post-menopausal women, data on the anti-atherogenetic effect of progestogens are scanty and mainly limited to medroxyprogesterone acetate and gestodene. Combining more androgenic progestogens with estrogens also negatively affects peripheral vascular resistance and vascular reactivity.
It is well known that endogenous and exogenous estrogens stimulate the hepatic synthesis of angiotensin that in turn increases plasma levels of aldosterone through an activation of the renin angiotensin system. The main effect of aldosterone is sodium reabsorption in the kidney. Therefore, in predisposed women estrogens may cause sodium and water retention. Progestogens have different effects on sodium metabolism that may range from extreme sodium retention to sodium excretion. Synthetic progestogens cause an increase of hepatic angiotensin and plasmatic angiotensin, thereby enhancing sodium retention. Progesterone competes with aldosterone at kidney level causing a dose-dependent natriuretic effect. Similar effect on kidney sodium excretion is shared by dydrogesterone, whereas a newer progestogen, drospirenone, has a more complex effect on sodium balance having direct anti-aldosterone activity. Drospirenone has a powerful antimineralcorticoid effect that is effective in counterbalancing the increase of aldosterone that may be induced by estrogen administration especially in predisposed women and in those predisposed to develop arterial hypertension. Furthermore, in hypertensive women drospirenone is effective in reducing blood pressure either alone or in combination with other anti-hypertensive agents (Archer et al., 2005; Oelkers, 2005; Sitruk-Ware, 2005; White et al., 2005; Ylikorkala, 2005).
Therefore, the overall effect of hormone replacement therapy on blood pressure is related, on one hand, to the individual response to the activation of the renin angiotensin system and, on the other hand, to the dose and type of molecules used. Higher doses of estrogens may induce sodium retention as do synthetic and androgenic progestogens. Micronized progesterone, dydrogesterone and drospirenone have an antimineralcorticoid effect (that is higher for drospirenone) and therefore antagonize the sodium retention effect of estrogens. These progestogens should be preferred in women with borderline hypertension, in those with arterial hypertension well controlled by anti-hypertensive therapy and in women with a tendency to sodium retention. Evidence from many observational studies suggests that estrogen replacement therapy after menopause can provide protection against heart disease (Colditz et al., 1987; Gruchow et al., 1988; Sullivan et al., 1988; McFarland et al., 1989; Stampfer and Colditz, 1991; Grodstein and Stampfer, 1995; Grodstein and Stampfer, 1998). However, the results of randomized studies using estrogen and progestogens in women averaging >60 years of age failed to confirm the results of the observational studies.
|
Prevention of CVD in women |
|---|
CVD is the leading cause of mortality in men and women in most developed countries despite research-based advances in its management and treatment. Mortality from CVD has been steadily decreasing for men, but the same is not true for women (Peltonen et al., 2000
; American Heart Association, 2005). CHD accounts for the majority of CVD deaths in women, and it represents the prime target for prevention, because CVD is for the most part preventable (Mosca et al., 2004). Prevention is especially important because most CVD treatments are only palliative.
CHD may differ in men and women with respect to epidemiology, diagnosis, management and prognosis (Table VI). Several studies have found that women hospitalized with an acute MI had higher unadjusted short-term mortality rates than men (Greenland et al., 1991; Becker et al., 1994; Weaver et al., 1996; Malacrida et al., 1998; de Gevigney et al., 2001; Rosengren et al., 2001; Heer et al., 2002; Nicolau et al., 2004). Older age, more frequent co-morbidities and less aggressive management are hypothesized as the major reasons contributing to the poorer prognosis in women. Using data from Unité de Soins Intensifs Coronaires (USIC) I–II registries (Danchin et al., 1997; Vaur et al., 2003), a nationwide registry of MI which followed 4347 unselected men and women hospitalized within 48 h of the onset of MI in France, there was a significantly higher risk of death after MI in women, particularly in the younger age group (30–67 years), after taking into account co-morbidities, severity of MI and treatment strategies (Simon et al., 2006). A worse prognosis was found in MI but not in congestive heart failure. In the CIBIS II trial, a randomized double-blind study in 2647 patients with class III and class IV heart failure, females had significant reduced survival regardless of baseline clinical profile and treatment (Simon et al., 2001).
|
There is a misperception of the prevalence of coronary artery disease and the impact of CV risk factors in women. CVD is for the most part preventable, because nine identifiable risk factors account for the majority of MI cases. In a study evaluating a population of middle-aged women, those who did not smoke cigarettes, were not overweight, maintained a healthful diet, exercised for half an hour a day and consumed alcohol moderately had an incidence of coronary events that was >80% lower than that in the rest of the population (Stampfer et al., 2000). Of the reduction, 70% was due to stopping smoking. Unfortunately, awareness and knowledge about CVD risk and prevention are suboptimal, not only among women but also among some of the physicians in charge of their management, supporting the need for targeted educational programmes about CVD risk factors in women, as well as optimal access to diagnosis, treatment and care.
Several efficient treatments such as angiotensin-converting enzyme (ACE) inhibitors and angiotensin receptor blocker (ARB), diuretics and ß-blockers are available to control blood pressure in women who are still hypertensive despite changes in lifestyle habits. For women who do not meet target lipid levels despite diet modification, results from randomized clinical trials clearly support the use of statins for those at high risk of CVD. However, for women at lower or intermediate risk, the data available to date are insufficient for a conclusion on the use of statins. The results of the CASHMERE trial, a randomized double-blind study currently evaluating the effects of atorvastatin in the progression of carotid intima-media thickness and arterial stiffness in menopausal women with hypercholesterolaemia, will be interesting in this perspective (Simon et al., 2004). With respect to prevention among women with diabetes, long-term control of glucose is needed to maintain a HbA1C level <7%.
Previous understandings of favourable effect of hormone replacement therapy in the prevention of coronary disease, derived from observational studies, have been challenged by the results of randomized trials. Further trials are necessary to better evaluate the benefit/risk ratio in recently menopausal members of the targeted population. Until then, prevention of events due to atherosclerosis should rely on diet and fitness and low-dose aspirin therapy. Where indicated, of course, treatment of hypertension and hyperlipidaemia will help to prevent CVD.
|
Hormone treatment and MI: methodological issues between randomized and observational studies |
|---|
 TOPAbstractIntroductionIncidence of CVDThe metabolic syndrome:...PCOS as a metabolic...Hormonal contraception and CVD |
|---|