![]() |
Abstract |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Key words: hormone replacement therapy/oral contraceptives/screening
![]() |
Screening methodology |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
The basic requirements of a screening programme have been well documented (Wilson and Dungree, 1968) (Table I
). There must be evidence that: early detection will affect the natural history of the disease; the performance characteristics of the test (e.g. sensitivity, specificity and predictive value) must be known; the test should be cost-effective, acceptable to users and providers, and should influence clinical decisions; and `treatment' should exist for abnormal results and detected disease. Finally, if OC or HRT users are deemed to require differential screening there must be clear evidence that they are at different risk of disease than other women. Simply collecting information about risk factors without any influence on management is unacceptable (Hannaford and Webb, 1996a
).
|
Taking all these considerations into account, very few screening procedures are prerequisites for the safe provision of either OC or HRT. A recent international workshop reviewed the current evidence relating to OC (Hannaford and Webb, 1996b) and concluded that there are only two prerequisites for the safe provision of OC (Hannaford and Webb, 1996a
). These are: a careful personal and family history with particular attention to cardiovascular risk factors, and an accurate blood pressure (BP) measurement. The BP should also be checked at regular intervals during OC use. The recommendations for BP are based on good evidence that high BP increases the risk of cardiovascular disease in the general population (MacMahon et al., 1990
) and that a history of hypertension (and sometimes non-checking of BP) increases the risk of cardiovascular disease in OC users [World Health Organization (WHO), 1996a,b, 1997]. Furthermore, in developed countries at least, the measurement of BP is cheap and readily available.
Recommendations for screening of HRT users might be different, because this group of women is of an age when the incidence of disease is generally much higher. This will alter the performance characteristics and cost-effectiveness of screening procedures. Compared with OC, however, there is often less information about the relationship between HRT and disease risk markers. The extrapolation of information obtained from the observation of OC users to HRT users may not be valid. For example, current evidence suggests that HRT users may have a lower risk of cardiovascular disease than non-users of HRT, whereas the effect, if any, tends to be in the opposite direction in OC users.
In developed countries, it is reasonable to assess all peri-menopausal and menopausal women for their cardiovascular risk status. This should be based initially on a careful personal and family history with particular attention to cardiovascular risk factors, and a BP check. If relevant history is disclosed or the BP is elevated, then further tests will be indicated. Such procedures should be available to all women, irrespective of their desire to use HRT. Likewise, current evidence suggests that policies for screening for breast and genital tract cancers should be the same as for similarly aged non-users of HRT. These procedures are potentially embarrassing and may make some women hesitate to seek advice about HRT if they are thought to be prerequisites for the provision of HRT. Care must be taken, therefore, not to make women feel that they will not receive HRT if these procedures are not undertaken. The poor performance characteristics of the bimanual vaginal examination and the clinical breast examination argue against incorporating these procedures into screening protocols for HRT users.
Outstanding research questions include: whether the use of HRT affects the sensitivity and specificity of mammography, the psychological impact of various screening programmes, and how women and their advisors perceive risk when making decisions about whether to undergo a sceening procedure.
![]() |
Testing hepatic function |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
The studies suggesting a link between OC use and liver adenoma or hepatocellular carcinoma have tended to find an increasing risk with prolonged duration of use (perhaps 5 years or more). Although the relative risks (RR) may be increased, the rarity of both of these conditions means that the absolute risk is extremely low. The available limited evidence suggests that current patterns of OC use do not alter the risk of liver cancer in areas where hepatitis B is probably the main risk factor.
Most studies of liver disease were conducted at a time when the hormonal content of the OC was relatively high. The limited evidence that exists relating to currently available preparations suggests that the lower dose (less than ethinyl oestradiol 50 µg) formulations have a lower risk potential than higher dose products.
Testing asymptomatic women for hepatic function before or during OC use will not identify the few women likely to develop liver disease (assuming that there is a true causal relationship between OC use and liver disease exists). Screening for other markers of liver disease risk (such as hepatitis B or C) is also inappropriate, as there is no evidence that OC use in such women is less safe than in users without these markers. Although there is little information about the progress of women known to have liver disease who concurrently use OC, current recommendations are that such women avoid these preparations while the liver disease is active. This advice is based on the wish to avoid diagnostic problems that might arise if complications occur.
Liver disease also appears to be rare in HRT users. Screening asymptomatic women for hepatic function, or markers of liver disease, will not identify women who will develop liver disease, and so should not be done.
![]() |
Blood pressure screening in OC users |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
The WHO study of Cardiovascular Disease and Steroid Hormone Contraception conducted in 12 developing and five European countries (WHO, 1995a) revealed higher overall risk of ischaemic stroke among OC users in developing countries than in those in European countries (WHO, 1996a). This difference in risk appeared to be related to the types of OC used as well as to the frequency with which users reported that their BP had been checked prior to or during OC use. A similar observation was made in the Transnational study conducted mainly in the UK and Germany (Heinemann et al., 1997). In contrast, two recent studies conducted in Washington State (Schwartz et al., 1997
) and California (Petitti et al., 1996
) in which women are more carefully screened before and during OC use, showed almost no increase in risk. Similar observations of a gradient in risk according to the health care setting were made for MI (Sidney et al., 1996
; Lewis et al., 1997
; WHO, 1997). Since the incidence rates for stroke and MI are low in women of reproductive age, the higher risk among OC users who did not have their BP checked represents only a small increase in risk associated with use of OC (attributable risk). Nevertheless, such observations raise questions about the safety of providing OC without regular BP screening or in providing OC in community-based distribution programmes in which only rudimentary screening of potential and current OC users is possible. However, the wider access to a reliable, reversible contraceptive method may have substantial benefits in populations with poor health care services, in particular by avoiding the risks associated with unwanted pregnancy. Although precise information from which to assess the balance of risks is lacking, the OC-attributable risk of cardiovascular disease in women with different underlying disease incidence rates can be estimated. These attributable risks must be compared with the risk which women experience daily, and with the risks attributable to other voluntary exposures such as smoking.
Cardiovascular risks attributable to OC use and smoking
Using information about cardiovascular disease incidence and mortality rates, and RR associated with current use of OC, Farley and colleagues estimated the increased cardiovascular incidence and mortality attributable to OC use in women of reproductive age in different regions of the world (Farley et al., 1998). These OC-attributable risks are presented according to age and whether or not users had their BPs checked and are shown in the context of the smoking attributable risks.
Mortality rates
Nationally reported deaths due to maternal and external causes in women at ages 1524, 2534 and 3544 years were extracted from the 19921996 World Health Statistics Annual (WHO, 1993WHO, 1994WHO, 1995b, 1996b, 1998b), for countries that reported only the numbers of deaths, not mortality rates, since the size of the underlying population is unknown. The ratio of maternal to external cause of mortality was computed. For areas in which the population included at least 10 million women in the age range 1544 years the most recently reported data were used; otherwise incidence rates were averaged over the most recent years to provide more stable estimates for each country. Countries were grouped according to development status and region (United Nations, 1996).
OC-attributable mortality is small (<20 deaths per million OC-users annually) for women under 35 years of age, even if they smoke. The excess mortality at all ages is less than the mortality attributable to smoking. While BP checking reduces the OC-attributable mortality, the impact is small (in the range 15 deaths per million users annually in non-smokers and 310 in smokers), Only above age 35 years does the OC-attributable mortality exceed 20 deaths per million users per year, and the impact of screening women for BP is more marked, particularly among smokers (60100 deaths annually per million OC-users who smoke).
Mortality for women due to external causes is approximately constant throughout the reproductive age and ranges from 100 to 400 deaths per million women annually depending on the country. Maternal mortality rates in developed countries are in the range 26 per million women annually, with maternal mortality ratios in the range 210 per 100 000 live births. In contrast, maternal mortality rates and ratios are up to 10-fold greater in those developing countries that report such rates. In countries that do not report cause of death information to WHO, estimated maternal mortality ratios are another 10-fold higher. In all regions of the world, about one sixth of maternal deaths are a result of unsafe abortion.
Conclusion
Although there is little systematic information from the Least Developed Countries, in which the wider use of reliable and safe contraceptive methods has the greatest potential to reduce the impact of unwanted pregnancy, the patterns and balance of risks in developed and developing countries for which information is available point to clear conclusions. The additional risk of cardiovascular death attributable to lack of BP screening in OC users is negligible compared with other daily risks and the risks associated with unwanted pregnancy. Although this is true for non-smoking women, even in developed countries, this should not be taken to imply that BP screening should not be performed. Where it is possible, such a simple screening tool should be included as part of general preventive health care. The key question concerns health care situations where the requirement for BP screening might prevent access to the method, and hence expose potential users to the health consequences of unwanted pregnancy. In such health care settings, it is clear that the small cardiovascular risk attributable to OC use is negligible compared with the risks of unwanted pregnancy.
![]() |
Thrombosis and thrombophilia |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Several parameters of the haemostatic system of patients who had had a first episode of VTE differ significantly from those of healthy controls [Leiden Thrombophilia Study (LETS)]. However, RR for most of these markers are very low (<3) (van der Meer et al., 1997). The highest RR (78) were found for deficiencies of the coagulation inhibitors (antithrombin III, protein C and protein S) and the recently detected functional abnormality of the protein C anticoagulant pathway associated with the factor V Leiden mutation. The mutant factor V is protected against the proteolytic action of activated protein C (APC resistance). Whereas all carriers of the mutation were found to be APC resistant, up to 20% of APC-resistant individuals may carry other genetic abnormalities (Koster et al., 1993
), whose thrombogenic potential has not entirely been elucidated. APC resistant women carrying the mutation are at a particular risk of thrombotic disease when taking the pill. Their risk was found to be more than 30-fold higher than that of users of non-hormonal contraceptives who do not carry the factor V Leiden mutation (Vandenbroucke et al., 1994
). OC induce effects on protein S, factor VIII and X activity which interfere with the APC assessment, inducing an `acquired APC resistance' in some women (De Ronde and Bertina, 1994
). The thrombogenicity of this condition is unknown. Thus, if APC resistance is found in women on the pill, analysis of the factor V Leiden mutation may be required. For the time being, the interpretation of results in women off the pill is standing on much more solid ground (Osterud et al., 1994
).
The value of screening for thrombophilia may be in improving public health by reducing the number of women per year suffering from pill-associated thromboembolism. However, screening should also be an aid to the individual pill user helping her to make her contraceptive choices easy and straightforward. A measure of the former is the sensitivity and specificity of the screening test, i.e. the ability to identify women at risk and the ability not to identify falsely women who do not carry any risk. A measure of the latter is the predictive value, i.e. the probability with which a positive result of the assay indicates that this woman would suffer from thromboembolic complications if she were to use OC (Table II)
|
It is obvious that even if all women with evidence of inhibitor deficiencies or APC resistance had refrained from using OC, more than 50% of cases would not have been prevented. Moreover, among the 6 milion, more than 300 000 women would screened false positives. Thus, inspite of a quite impressive specificity in the range of 9599%, the predictive value of these laboratory tests is in the range of 103 to 104. If women with APC resistance or any of the inhibitor deficiencies choose to use OC, their risk of VTE is less than 1/1000 in the former and less than 1/500 in the latter. Considering the numerous non-contraceptive benefits of OC use and the contraceptive and non-contraceptive risks of alternative contraceptive methods, it is reasonable to assume that the value of a general screening both for the individual user and for the pill population in general is rather limited.
To reduce the number of false positive results there is only one option: to perform the laboratory screening only after a preselection step, i.e. only for women with a positive family history. In principle, the family history should be the ideal tool to preselect for congenital thrombophilia. Given a cumulative thrombosis incidence of 70% in inhibitor deficient individuals and about 30% in APC resistant individuals, family history is likely to yield a rather limited sensitivity. Especially when only first degree relatives are taken into consideration, a high rate of false negative results must be expected. Moreover, given the fact that there are numerous other, acquired conditions that may cause thrombosis, a considerable number of false positive results is to be expected as well (Briët et al., 1994).
Clinical guidelines
Several highly specific haemostaseological assays are now available to identify women at risk of OC-associated thromboembolic complications. However, more than 50% of cases cannot be predicted by these assays. Due to the very low absolute risk of thrombosis in OC users the predictive value of these assays is low, thus limiting both the public and the individual benefit of a general screening protocol. Women with a positive family history may selectively be screened for antithrombin III, protein C and S deficiency as well as for APC resistance. A positive family history in the presence of a thrombophilic abnormality of coagulation is a strong argument for a non-hormonal contraceptive method.
![]() |
Screening for lipids |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
HRT should be considered for the prevention and treatment of CHD in postmenopausal women at increased risk for CHD. Epidemiological studies have consistently shown that HRT reduces the incidence of CHD by 4050%, both in healthy postmenopausal women and in those with established CHD. The results of a randomized clinical trial of HRT for the secondary prevention of CHD were recently published (Hulley et al., 1998). This study showed no overall difference in cardiac events between the HRT and placebo groups. However, it should be appreciated that by years 4 and 5 there was a reduction of events in the HRT group of 35% compared with placebo, but this was counterbalanced by an increase in events during the first few months of the study. This study has been criticized (Stevenson, 1998) because of its lack of power, and because of the dose of oestrogen and choice of progestogen which may have been responsible for the transient increase in cardiac events observed early in the study. Further studies using different formulations and doses are clearly needed. There are many mechanisms through which oestrogens and progestogens influence the cardiovascular system, including effects on lipids and lipoproteins. Lipid and lipoprotein abnormalities are important in the pathogenesis of atheromatous disease, and it is therefore appropriate to measure lipid and lipoprotein profiles in women who may be at increased risk for CHD prior to starting any treatment with sex hormones. A major problem is that the relative importance of each of these various mechanisms has not yet been established, nor indeed what is the relative importance of the various lipids and lipoproteins with respect to cardiovascular disease risk.
Lipids, lipoproteins and CHD risk
There is considerable evidence that lipids and lipoproteins are involved in the development of CHD in women (Stevenson, 1996). It is well established that high circulating levels of cholesterol are associated with increased incidence of CHD. Low density lipoproteins (LDL) account for the major portion of total plasma cholesterol in most individuals and increased concentrations of LDL cholesterol lead to increased risk of CHD. However, while the lowering of LDL concentrations reduces the risk of CHD in men, it is not clear that the same effect can be achieved in women. The clearance of LDL from the circulation is effected by hepatic receptors. This process is slow and operates at near saturation of the receptors. Thus, if LDL concentrations are high, the LDL will have a relatively long half-life in the circulation. This makes them more susceptible to modification or damage, and damaged LDL are more likely to be retained intramurally in the arteries. Postprandial lipoprotein remnants are also atherogenic and thus the efficiency of the clearance of such remnants is also important.
LDL comprises various subclasses of different size, chemical composition and density. Lipoprotein (a) is an LDL which contains two distinct apolipoproteins (B and a). The latter is structurally a giant mutant of plasminogen. It has been suggested that lipoprotein (a) is an independent lipoprotein risk marker for CHD, with high levels being associated with increased risk for CHD. However, this is not an invariable finding and it may be that high lipoprotein (a) concentrations signify increased CHD risk only when LDL levels are also raised. Lipoprotein (a) is atherogenic largely because of its propensity to be retained in the arterial wall. It binds avidly to arterial proteoglycans, and also enhances arterial LDL retention. Lipoprotein (a) is also potentially thrombogenic because of its structural homology with plasminogen. It may thus compete with plasminogen for binding sites and inhibit fibrinolysis, and it also binds fibrin.
LDL particle sizes vary and can be classified into subgroups accordingly. It has been shown that LDL particle size is clinically important. Patients with CHD, and especially female patients, have a greater proportion of small dense LDL than healthy controls, a pattern described as subtype B. These smaller dense LDL particles are more atherogenic, perhaps because they are more readily cleared by scavenger mechanisms than by the apoB100 receptors, and also because they may be more susceptible to oxidative damage. Small dense LDL are found in individuals with high triglycerides and low HDL (high density lipoproteins) and are a feature of the metabolic syndrome.
High density lipoproteins may participate in reverse cholesterol transport, in which they remove cholesterol from tissues such as arterial walls and return it to the liver for excretion or resecretion. Thus, HDL levels are inversely associated with CHD risk, with high levels considered to be protective. The HDL2 subfraction is considered to be the subfraction which confers the major benefit, and HDL2 can be catabolized by hepatic lipase. Conditions that increase hepatic lipase activity, such as increased adiposity or increased androgen activity, are associated with low HDL2 levels. Triglyceride levels are inversely related to HDL and HDL2 levels. Increased hepatic triglyceride production leads to increased levels of triglyceride-rich very low density lipoproteins (VLDL). Reduced VLDL clearance results in increased plasma residence of VLDL remnants, decreased HDL2 and increased IDL levels, an atherogenic lipoprotein profile. Thus, low levels of HDL and HDL2 may reflect a reduction in reverse cholesterol transport or a reduction in VLDL remnant clearance, and it is not clear as to which is more important. The role of the HDL3 cholesterol subfraction is less clear.
Lipids, lipoproteins and HRT or OC
Changes seen in lipids and lipoproteins with combined OC vary according to progestogen type (Godsland, 1990). OC containing low dose norethisterone or desogestrel lower LDL and increase HDL as do those containing gestodene. Those containing levonorgestrel and high dose norethisterone have no effect on LDL but lower HDL, particularly HDL2. Most OC increase triglycerides but this effect is opposed by higher doses of levonorgestrel. The OC-induced increase in triglycerides is due to increased synthesis rather than decreased elimination. Postprandial triglyceride elimination time appears to be increased in OC users. Progestogen-only OC have minimal effects on lipids and lipoproteins, although there is a tendency to decrease HDL.
HRT produced changes in lipids and lipoproteins that depended both on the types of steroids used and their route of administration (Stevenson, 1997a). Oestrogen causes a reduction in total cholesterol due to a lowering of LDL. It also causes an increase in HDL, and particularly in the HDL2 subfraction. This latter effect appears to be less pronounced with transdermal oestradiol than with oral oestrogens. The addition of an androgenic progestogen does not impede the lowering of LDL by oestrogen but may blunt or reverse the beneficial increase in HDL. The use of less androgenic C-21 progestogens does not usually impede the oestrogen-induced HDL rise. However, C-21 progestogens do not lower triglycerides while androgenic progestogens do. This can be an important effect when using conjugated equine oestrogens which increase triglycerides, as increased triglyceride concentrations may be of particular importance in the development of CHD in women. Oral oestradiol may also elevate triglycerides but transdermal oestradiol actually lowers their levels. Oral oestradiol also increases postprandial lipoprotein clearance, and causes a decrease in lipoprotein (a). There is also evidence that oestrogens may protect LDL against oxidation.
Practical guidelines
Specific changes in lipids and lipoproteins can therefore be achieved by choosing the type of HRT or OC regimen used. For example, in women who may have high triglycerides, such as diabetics or the obese, the use of transdermal oestradiol and an androgenic progestogen in HRT would be appropriate. In women with low HDL, oral oestradiol and a non-androgenic progestogen would be more suitable. A pretreatment screening for lipid abnormalities would therefore be useful for women who have any evidence of increased CHD risk, such as those with a strong family history of CHD or hypercholesterolaemia, those with diabetes, obesity or hypertension, and those who smoke.
![]() |
Osteoporosis |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
HRT is as effective as any other currently available agent for both the prevention and the treatment of postmenopausal osteoporosis. It reduces bone turnover, and increases bone density and reduces fracture incidence in all the important skeletal sites. The usual bone-conserving doses of different oestrogens have been determined for women in the early postmenopause, although lower doses may be effective for older women (Stevenson, 1997b). OC use may be associated with small increases in bone density but the magnitude of such increases is small. They may occasionally be used in place of HRT in young women who have had premature ovarian failure and who do not wish to take HRT.
Practical guidelines
Widespread unselected screening for osteoporosis is not cost-effective, but bone density measurements are very useful for women selected because of their likelihood of increased osteoporotic risk. These include women who have undergone a premature menopause, those with a strong family history of osteoporosis, previous secondary amenorrhoea such as with anorexia, and previous oral corticosteroid use. Bone density measurements may also be useful to help decision-making about whether or not to advise HRT for a woman who would only take it if she needed osteoprosis prevention. However, for a woman who has already decided to go onto HRT, there is no reason to screen for osteoporosis since HRT would be the appropriate therapy for the disease if it were present. Similarly, there is no indication for any skeletal assessment before starting OC.
![]() |
Is there a real need for a screening before either OC or HRT? |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
![]() |
Notes |
---|
* A meeting was organized by ESHRE (Capri, August 2728, 1998) with financial support from Schering S.p.A. to discuss the above subjects. The speakers included J.Collins (Hamilton), E.Diczfalusy (Rönninge), T.Farley (WHO), P.Hannaford (Aberdeen), S.Skouby (Copenhagen), J.C.Stevenson (London) and U.Winkler (Essen). D.R.Mishell (Los Angeles) was not able to participate but contributed to the manuscript. The discussants included P.G.Crosignani (Milano), L.A.J.Heinemann (Berlin), C.La Vecchia (Milano), M.Meschia (Milano), T.Norpoth (Berlin), G.Rosano (Roma) and M.Whitehead (London). This report was prepared by P.G.Crosignani and B.L.Rubin.
![]() |
References |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Campos, H., Sacks, F.M., Walsh, B.W. et al. (1993) Differential effects of estrogen on low-density lipoprotein subclasses in healthy postmenopausal women. Metabolism, 42, 11531158.[ISI][Medline]
Crook, D., Cust, M.P., Gangar, K.F. et al. (1992) Comparison of transdermal and oral estrogen/progestin hormone replacement therapy: effects on serum lipids and lipoproteins. Am. J. Obstet. Gynecol., 166, 950955.[ISI][Medline]
Crook, D., Godsland, I.F., Worthington, M. et al. (1993) A comparative metabolic study of two low-estrogen-dose oral contraceptives containing desogestrel or gestodene progestins. Am. J. Obstet. Gynecol., 169, 11831189.[ISI][Medline]
Crook, D., Godsland, I.F., Hull, J. and Stevenson, J.C. (1997) Hormone replacement therapy with dydrogesterone and oestradiol-17ß: effects on serum lipoproteins and glucose tolerance. Br. J. Obstet. Gynaecol., 104, 298304.[ISI][Medline]
Delmas, P.D. and Garnero, P. (1998) Biological markers of bone turnover in osteoporosis. In Stevenson, J.C. and Lindsay, R. (eds), Osteoporosis. Chapman and Hall, London, pp. 117136.
de Ronde, H. and Bertina, R.M. (1994) Laboratory diagnosis of APC-resistance. A critical evaluation of the test and the development of diagnostic criteria. Thrombos. Haemost., 72, 880886.[ISI][Medline]
Farley, T.M.M., Collins, J. and Schlesselman, J.J. (1998) Hormonal contraception and risk of cardiovascular disease: an international perspective. Contraception, 57, 211230.[ISI][Medline]
Farmer, R.D.T. and Preston, T.D. (1995) The risk of venous thromboembolism associated with low oestrogen oral contraceptives. J. Obstet. Gynaecol., 15, 195200.
Garnero, P., Hausherr, E., Chapuy, M.C. et al. (1996) Markers of bone resorption predict hip fracture in elderly women. J. Bone Min. Res., 11, 15311538.[ISI][Medline]
Gerstman, B.B., Piper, J.M., Tomita, D.K. et al. (1991) Oral contraceptive estrogen dose and the risk of deep venous thromboembolic disease. Am. J. Epidemiol., 133, 3237.[Abstract]
Godsland, I.F., Crook, D., Simpson, R. et al. (1990) The effects of different formulations or oral contraceptive agents on lipid and carbohydrate metabolism. N. Engl. J. Med., 232, 13751381.
Hannaford, P.C. and Webb, A.M.C. on behalf of participants of an international workshop (1996a) Evidence-guided prescribing of combined oral contraceptives: consensus statement. Contraception, 54, 125129.[ISI][Medline]
Hannaford P.C. and Webb, A.M.C. (eds) (1996b) Evidence-guided Prescribing of the Pill. Parthenon, Carnforth, UK.
Hannaford, P.C., Kay, C.R., Vessey, M.P. et al. (1997) Combined oral contraceptives and liver disease. Contraception, 55, 145151.[ISI][Medline]
Heinemann, L.A.J., Lewis, M.A., Thorogood, M. et al. (1997) Oral contraceptives and risk of thromboembolic stroke: results from international study on oral contraceptives and health of young women. Br. Med. J., 315, 15021504.
Hulley, S., Grady, D., Bush, T. et al. (1998) Randomized trial of estrogen plus progestin for secondary prevention of coronary heart disease in postmenopausal women. J. Am. Med. Assoc., 280, 605613.
Koster, T., Rosendaal, F.R., de Ronde, H. et al. (1993) Venous thrombosis due to poor anticoagulant response to activated protein C: Leiden thrombophilia study. Lancet, 342, 15031506.[ISI][Medline]
Krauss, R.M. (1991) The tangled web of coronary risk factors. Am. J. Med., 90 (Suppl. 2A), 3641.
Lees, B., Banks, L.M. and Stevenson, J.C. (1998) Bone mass measurements. In Stevenson, J.C. and Lindsay, R. (eds), Osteoporosis. Chapman and Hall, London, pp. 137160.
Lewis, M.A., Heinemann, L.A.J., Spitzer, W.O. et al. for the Transnational Research Group on Oral Contraceptives and the Health of Young Women (1997) The use of oral contraceptives and the occurrence of acute myocardial infarction in young women. Results from the Transnational Study on Oral Contraceptives and the Health of Young Women. Contraception, 56, 129140.[ISI][Medline]
MacMahon, S., Peto, R., Cutler, J. et al. (1990) Blood pressure, stroke and coronary heart disease: Part 1, prolonged differences in blood pressure: prospective observational studies corrected for the regression dilution bias. Lancet, 335, 765774.[ISI][Medline]
Mant, D. and Fowler, G. (1990) Mass screening: theory and ethics. Br. Med. J., 300, 916918.[ISI][Medline]
Miller, N.E. (1987) Associations of high-density lipoprotein subclasses and apolipoproteins with ischaemic heart disease and coronary atherosclerosis. Am. Heart J., 113, 589597.[ISI][Medline]
Osterud, B., Robertsen, A., Asvang, G.B. and Thyssen, F. (1994) Resistance to activated protein C is reduced in women using oral contraceptives. Blood Coag. Fibrin., 5, 853854.[ISI][Medline]
Petitti, D.B., Sidney, S., Bernstein, A. et al. (1996) Stroke in users of low-dose oral contraceptives. N. Engl. J. Med., 335, 815.
Schwartz, S.M., Siscovick, D.S., Longstreth, W.T., et al. (1997) Use of low-dose oral contraceptives and health of young women. Ann. Intern. Med., 127, 596603.
Shelton, J.D., Angle, M.A. and Jacobstein, R.A. (1982) Medical barriers to access to family planning. Lancet, 340, 13341335.
Sidney, S., Petitti, D.B., Quesenberry, C.P. et al. (1996) Myocardial infarction in users of low-dose oral contraceptives. Obstet. Gynecol., 88, 939944.
Speroff, L., DeCherney, A. and the Advisory Board for the New Progestins (1993) Evaluation of a new generation of oral contraceptives. Obstet. Gynecol., 81, 10341047.[Abstract]
Stevenson, J.C. (1996) Metabolic effects of the menopause and oestrogenreplacement. In Barlow, D.H. (ed.), Ballière's Clinical Obstetrics and Gynaecology. The Menopause: Key Issues. Baillière Tindall, London. pp. 449467.
Stevenson, J.C. (1997a) Hormone replacement therapy and lipids. Menopause Rev., 2, 1520.
Stevenson, J.C. (1997b) Gonadal hormones. Osteoporosis Int., 7 (Suppl. 1), 5860.[ISI]
Stewart-Brown, S. and Farmer, A. (1997) Screening could seriously damage your health. Br. Med. J., 314, 533534.
United Nations (1996) World Economic and Social Survey 1996: Trends and Policies in the World Economy. United Nations, New York.
Vandenbroucke, J.P., Koster, T., Briët, E. et al. (1994) Increased risk of venous thrombosis in oral contraceptive users who are carriers of factor Leiden mutation. Lancet, 344, 14531457.[ISI][Medline]
van der Meer, F.J.M., Koster, T., Vandenbroucke, J.P. et al. (1997) The Leiden Thrombophilia Study (LETS). Thrombos. Haemostas., 78, 631635.[ISI][Medline]
van der Vange, N., Kloosterboer, H.J. and Haspels, A.A. (1987) Effect of seven low-dose combined oral contraceptive preparations on carbohydrate metabolism. Am. J. Obstet. Gynecol., 156, 918922.[ISI][Medline]
Westerveld, H.T., Kock, L.A.W., van Rijn, J.M. et al. (1995) 17ß-estradiol improves postprandial lipid metabolism in postmenopausal women. J. Clin. Endocrinol. Metab., 80, 249253.[Abstract]
Wilson, J.M.G. and Dungree, G. (1968) Principles and Practice of Screening for Disease. World Health Organization, Geneva.
Winkler, U.H. (1996) Role of screening for vascular disease in pill users: the hemostatic system. In Hannaford, P.C. and Webb, A.M.C. (eds), Evidence-guided Prescribing of the Pill. Parthenon, London, pp. 109120.
World Health Organization (1993) World Health Statistics Annual 1992. World Health Organization, Geneva.
World Health Organization (1994) World Health Statistics Annual 1993. World Health Organization, Geneva.
World Health Organization Collaborative Study of Cardiovascular Disease and Steroid Hormone Contraception (1995a) A multinational casecontrol study of a cardiovascular disease and steroid hormone contraceptives. Description and validation of methods. J. Clin. Epidemiol., 48, 15131547.[ISI][Medline]
World Health Organization (1995b) World Health Statistics Annual 1994. World Health Organization, Geneva.
World Health Organization Collaborative Study of Cardiovascular Disease and Steroid Hormone Contraception (1996) Haemorrhagic stroke, overall stroke risk, and combined oral contraceptives: results of an international, multicentre, casecontrol study. Lancet, 348, 505510.[ISI][Medline]
World Health Organization Collaborative Study of Cardiovascular Disease and Steroid Hormone Contraception (1996a) Ischaemic stroke and combined oral contraceptives: results of an international multicentre, casecontrol study. Lancet, 348, 498505.[ISI][Medline]
World Health Organization (1996b) World Health Statistics Annual 1995. World Health Organization, Geneva.
World Health Organization Collaborative Study of Cardiovascular Disease and Steroid Hormone Contraception (1997) Acute myocardial infarction and combined oral contraceptives: results of an international multicentre casecontrol study. Lancet, 349, 12021209.[ISI][Medline]
World Health Organization Scientific Group on Cardiovascular Disease and Steroid Hormone Contraception (1998a) Cardiovascular Disease and Steroid Hormone Contraception: Report of a WHO Scientific Group. WHO Technical Report Series. World Health Organization, Geneva, p. 89.
World Health Organization (1998b) World Health Statistics Annual 1996. World Health Organization, Geneva.
Submitted on June 21, 1999; accepted on November 12, 1999.