1 Emeritus Professor of Epidemiology and 2 Department of Epidemiology and Biostatistics, McGill University, Montreal, Canada and 3 Professor of Medical Statistics, University of Surrey, Guildford, Surrey, UK
![]() |
Abstract |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Key words: casecontrol study/epidemiology/myocardial infarction/oral contraceptives
![]() |
Introduction |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Initial research findings about associations of cardiovascular and cerebrovascular adverse events with the use of OC dealt with a generation of the earliest marketed products. The included doses of ethinyl estradiol (EE) >50 µm, often 75 or 100 µg, frequently designated today as the first generation (1gen) OC. The subsequent generations are defined by type of progestogen and not by dose of estrogen, i.e. both second (2gen) and third (3gen) have <50 µg. The generations according to progestins are defined as follows. 1gen OC: norethisterone, norethisteroneacetate, ethynodioldiacetate, lynesterol; 2gen OC: norgestrel, levonorgestrel (LNG); 3gen OC: desogestrel, gestodene (norgestimate under debate because of the close relation with levonorgestrel).
The generations are also discernible by their progestogenic and androgenic activity: 2gen increases progestogenic binding and 3gen reduces androgenic receptor binding compared with their predecessors. Third generation products all appeared on the market most recently; chronology has been important in taxonomy.
Recent controversies about the safety of oral contraceptives have dwelt upon venous cardiovascular side-effects (venous thromboembolism, VTE), which has the lowest morbidity and mortality (World Health Organization, 1995a,b
; Edwards and Cohen, 1999
). The initial interpretations were that the newest (3gen) OC on the market had higher risk of VTE than for users of the older (2gen), more established, OC. Later studies and analyses requested by the Committee on Proprietary Medicines and Products (CPMP) of the European Medicines' Evaluation Agency (EMEA) in March of 1996, conducted or reported between 1996 and 1999 (Farmer et al., 1997
, 1998
, 1999
; Suissa et al., 1997
; Lewis et al., 1999
; Lidegaard et al., 1998
; Todd et al., 1999
; Suissa et al., 2000
), have offered objective opinions about the controversial data (Farmer et al., 1997
; Benagiano and Primiero, 1999
; Cohen and Edwards, 1999
; International Federation of Fertility Societies, 1999
). Less attention has been paid to the safety record of newer OC products in respect to arterial events. The low incidence of stroke among users and the undetectable differences between users of 2gen and 3gen OC are important (Petitti et al., 1996
; World Health Organization, 1996a
,b
; Schwartz et al., 1997
; Heinemann et al., 1998
; Lidegaard and Kreiner, 1998
).
The aim of this paper is to aggregate and present the findings of recent research on the risk of 2gen and 3gen OC for acute MI. Because the absolute risk of this disorder among young women aged <45 years, and especially <25 years, is very low, the category of event appears much less frequently in the literature. Comparative recent studies, except the Transnational Case Control Study of Oral Contraceptives and Health (Lewis et al., 1996b, 1997a
) all lack statistical power if analysed as stand-alone projects. Initially, we considered all investigations reported from 1996 to 1999 and meta-analysed the findings. As an historical reference and baseline we derived medians of pertinent studies from 1965 to 1966. With delays in usual rigorous peer review, we have now been able to extend the coverage of the aggregating process to two major papers published during 2001. The main findings of this data aggregation are the pooled odds ratios (OR, relative risks) of the newer 3gen OC as determinants of MI. The referents are older 2gen OC or `no use' of OC.
![]() |
Materials and methods |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Before 1996, we found one single randomized controlled trial (Fuertes de la Haba et al., 1973). All other studies included were observational and controlled. Given the appreciable differences in the methods, definitions and approach to analysis of pre-1996 reports over three decades, we have only derived medians within category, when feasible, as well as the range of the values estimated. Aggregation prior to 1996 was not a meta-analysis in this paper. Formal comparisons of findings before and after 1996 were not part of the objective. For the casecontrol studies between January 1996 and December 2001, we used the meta-analytic method that Petitti has adapted from others (Greenland, 1987
; Greenland and Longnecker, 1992
; Petitti, 2000
). Her quantitative methods, now in a second edition, have become mainstream when synthesizing observational studies (Petitti, 2000
). The aggregation of the data as a meta-analysis is justified because we only include casecontrol studies. The main pooled OR is presented four times, including and excluding one study (MICA, myocardial infarction and oral contraceptives) and then including and excluding Dunn 2001 and finally including and excluding the Danish study (Lidegaard and Kreiner, 1998
). We reiterate that the reader can choose the combination of studies he/she deems most appropriate.
![]() |
Results |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Cohort studies, relative risks
The cohort project of The Royal College of General Practitioners (UK) published several relative risks (current users versus non-users) for MI and for non-rheumatic diseases of the heart (Beral, 1977; Royal College of General Practitioners, 1977
, 1981
, 1983
; Kay, 1982
). They represent a useful benchmark. We cite four statistically significant risk ratios: 5.2 (in 1974) (Royal College of General Practitioners, 1974
); 3.2 (in 1977) (Royal College of General Practitioners, 1977
); 6.4 (1982) (Kay, 1982
); and 2.0 (1983) (Royal College of General Practitioners, 1983
). Vessey, comparing ever users to non-users in the Oxford Family Planning Study (19681980), reported 4.7 (NS) for fatal MI (Vessey et al., 1977
). In the USA, Ramcharan's reported risk ratio for non-fatal MI was 1.1 (NS) from the Walnut Creek Study (19681977) (Ramcharan et al., 1981
). The median point estimate for relative risks derived from early cohort projects was 3.2 (Figure 1
).
|
Cohort studies, absolute risks
Absolute risks reported in the cohort studies were as follows for the Oxford, RCGP and Walnut Creek cohort studies: among never users, considering both fatal and non-fatal events, the range was 1.4 to 23 MI per 100 000 women years (WY) (Beral, 1977; Vessey et al., 1977
, 1981
; Ramcharan et al., 1981
; Royal College of General Practitioners, 1981
, 1983
). The median was 1.5 MI per 100 000 WY. In users, the values ranged from 7.5 to 30 per 100 000 WY and the median was 13.0 per 100 000 WY. The range for absolute attributable risks can also be derived from Realini and Goldzieher's large review. It was from 4.0 to 15 per 100 000 WY; the median was 5.7 (Realini and Goldzieher, 1985
).
Summarizing cohort and casecontrol studies, for any OC on the market, when the early studies were combined, the overall relative risk hovered around 2.0. It was 3.0 when comparing current users or ever users against non-users or never users. There appears to be no material difference in rates between fatal and non-fatal MI. None of the early studies compared categories of combined OC or individual drugs classified by progestin content head-to-head against one another. There was an appreciable difference in the absolute rates of MI comparing users of any OC and non-users (Figure 2).
|
|
The Danish casecontrol ctudy
In an ongoing Danish casecontrol study (Lidegaard and Kreiner, 1998) with three cardiovascular outcomes, an analysis with data useable in a meta-analysis was published as an abstract in 1996 (Lidegaard and Edstrom, 1996
). For 88 cases of MI (19941995) and 1045 controls compared with non-use, the adjusted OR were 4.1 (1.5510.9) for first generation OC and 1.9 (0.74.9) for 2gen OC, and for 3gen OC users the OR was 1.9 (0.49.2) in the first year, 0.5 (0.13.9) in years 15 and 0.8 (0.22.7) for use >5 years. Based on these point estimates and CI we calculated an overall OR of 0.51 (0.151.72) for 3gen OC users compared with 2gen OC use (Table I
). There was another follow-on abstract published after March 2001 and we found no subsequent full paper. Thus we also show aggregate estimates of the recent studies excluding the Danish work for that reason (Table I
, Figure 1
).
The World Health Organization study
Overall, 21 centres in this study recruited 368 cases of MI and 941 controls matched by age within 5 year bands (World Health Organization, 1997). We refer only to the European findings here. There had been 198 cases accrued in Europe and 480 controls. In a scientific meeting at Geneva, three cases and five controls exposed to 3gen OC and 24 cases and 22 controls to levonorgestrel (LNG) were reported. The resulting unpublished crude OR for MI in users of 3gen OC versus 2gen OC is 0.55 (NS). Eventually, having eliminated the Hungarian centre, the data for the UK and Germany (World Health Organization, 1997
) reduced the cases on LNG to 13 cases and 17 controls with an OR of 0.59 (0.093.75) for 3gen combined OC versus 2gen LNG.
The Transnational study on oral contraceptives and the health of young women
The Transnational study had 182 cases of MI and 635 matched controls within 5 year age bands (Spitzer et al., 1993; Lewis et al., 1996a
,b
, 1997a
). The OR for MI in 1gen, 2gen and 3gen OC compared with non-use were 4.7 (1.5214.3), 2.99 (1.515.91), and 0.85 (0.302.39). The OR (3gen versus 2gen) was <1.0 and statistically significant: 0.24 (0.070.78).
The MICA study
The MICA study (myocardial infarction and oral contraceptives, a retrospective casecontrol study in England and Scotland) (Dunn et al., 1997, 1999
) was conducted mostly after October 1995 and was vulnerable to the publicity bias generated by the British pill scare. The eligible MI events had occurred mostly before October 1995; the interviews on cases and controls were all done after that date. It is noteworthy that the study showed large differences between the findings of interview data and those based on the data of general practitioners who had written records on the study subjects. The difference almost certainly reflects the publicity bias. Another indication of systematic error is the considerable excess of healthy controls that were observed to be users of levonorgestrel-containing OC when compared with readily available data from UK prescription databases. It is a concern that four separate sequential analyses with pronounced fluctuations of OR had been performed prior to submission of the research report for publication, raising doubts about the validity and appropriateness of the final findings based on acceptable criteria for a suitable analysis plan. The MICA findings proved to be the outliers.
The Dunn mortality study in Great Britain
The study (Dunn et al., 2001) evaluated issues similar to those of MICA and followed it on with a second casecontrol study. However, cases were limited to fatalities and the controls were survivors. An important bias seen in the earlier morbidity study was corrected in that clinical notes from the doctors' surgeries were used to determine exposure to the risk factor. There was incompatibility of definitions in the choice of cases and controls when comparing with the other six studies. Despite its improved design, its care in minimizing bias and its adequate size (432 subjects), the study was both included and excluded in the aggregated analyses we did (Figure 1
). Exclusion was primarily justified by the unconventional choice of controls (`survivors' as distinct from `unaffected'). The reader can judge which of the pooled values is best.
The Tanis oral contraceptive and risk of MI study
A large casecontrol study was conducted in The Netherlands coordinated from Leiden (Tanis et al., 2001). There were 248 cases with a first MI and 925 controls in an unmatched design. The adjusted OR for 2gen OC versus no use was 2.5 (1.54.1); for 3gen OC versus no use it was 1.3 (0.72.5). From data in the paper we estimated an OR of 0.52 (0.231.2) for 3gen OC versus 2gen OC. We did not find any important biases of methods or field work in the published paper. The OR for `any OC' versus no use was 2.0 (1.52.8). Despite the large size of the study with 1173 subjects the authors report it as inconclusive.
Aggregating the seven recent studies
We conducted a meta-analysis of seven casecontrol studies published since January 1, 1996 (Jick et al., 1996; Lewis et al., 1997b
; World Health Organization, 1997
; Lidegaard et al., 1998
; Dunn et al., 1999
; Tanis et al., 2001
). A summary of pertinent findings from those seven casecontrol studies is in Table I
. The outcome is MI. The key finding is the OR (relative risk) contrasting 3gen and 2gen OC with levonorgestrel only (LNG) as well as 3gen OC versus no use. We included MICA in two meta-analyses of the data and excluded it from two others. We included and excluded the Dunn 2001 study. The Danish study was also excluded from two aggregations. The reader can choose which resulting estimate he/she deems best (Figure 1
). The seven studies published from 1996 to 2001 included 6464 subjects (Table I
). The four studies we aggregated that seem to us to be least affected by bias, design or field problems are WHO, GPRD, Transnational study and Leiden (Tanis et al., 2001
). They accrued 2723 patients. We propose that the best estimates of an OR of 3gen versus 2gen OC are 0.44 (0.240.89) with four studies and 0.62 (0.380.99) with all seven studies (Figure 1
).
Absolute rates of MI in recent publications were only reported in the WHO study (World Health Organization, 1997), the GPRD study (Jick et al., 1996
), and the Danish study (Lidegaard and Edstrom, 1996
). Both the WHO and the Jick studies report 1.8 events of MI per 100 000 WY. The Danish study reports 4.8 MI per 100 000 WY among users of 2gen OC products and 0.6 per 100 000 WY among users of 3gen OC products. A formal aggregation is not appropriate but a cautious and conservative estimate is 1.3 events per 100 000 WY (Figure 2
) combining users and non-users.
![]() |
Discussion |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
The aetiology of OC-associated MI has been the subject of lengthy debate. Endothelial damage does not seem to be a prerequisite for formation of a venous thrombus (Thomas et al., 1985). In contrast, haemodynamic considerations alone make it unlikely that an arterial thrombus will accumulate on an entirely intact endothelium. An imbalance in the normal cycle of arterial endothelial damage and repair, or the endothelial and intimal lesions seen in OC users, could result in platelet adhesion and activation and thrombus accumulation in an apparently normal artery. Either way this process will be opposed by circulating factors which promote endothelial integrity and will be furthered by those factors which disrupt endothelial function. There are three major candidates for involvement in these processes: serum triglycerides, insulin and high density lipoprotein (HDL) (Godsland et al., 1995
).
Publications from 1982 onwards show consistently that OC containing 3gen progestogens improve the HDL-cholesterol to total cholesterol ratio (Bergink et al., 1982, 1983
, 1984
; Pentilla et al., 1983
). The progestogen component of OC had been identified as an important risk factor for arterial disease. Bergink wrote, `in view of the available epidemiological evidence a low ratio of HDL-cholesterol to total cholesterol should be avoided and an important consideration in the choice of an oral contraceptive is the anticipated directional change in serum lipoprotein composition' (Bergink et al., 1982
). Bergink's concern that low HDL-cholesterol ratios be avoided has been corroborated in a recent report (Brown et al., 2001
). They linked regression of coronary artery disease to selected drugs and exercise. The empirical data from Brown's major paper justify theoretical postulates about a favourable effect of OC on HDL to a considerable extent. It had already been noted (Solyom, 1972
) that androgens are known to lower HDL-cholesterol. Following the discovery of the difference of intrinsic androgenicity between desogestrel and levonorgestrel as typical representatives of 3gen and 2gen progestagens (Bergink et al., 1981
) it was the intention to develop OC with a favourable lipid profile which would be positive risk factors for arterial disease. What has been achieved for arterial disease, as demonstrated by recent epidemiological research, is not only true with respect to MI but also stroke, as indicated in the introduction of this paper (World Health Organization, 1996a
). Attainment of equivalency between the risks of exposure to 3gen OC and non-use of any OC product in respect to MI allowed goals of laboratory investigators involved in the development of newer and safer preparations for many years to be realized (Petitti et al., 1996
; Schwartz et al., 1997
).
The later MICA Study (Dunn et al., 1997, 1999
), which followed our Transnational protocol closely and which accrued 2176 subjects, failed to attain statistical significance. It is encouraging that other investigative groups have tested the same question. None of six studies (Jick et al., 1996
; Lidegaard and Edstrom, 1996
; Lewis et al., 1997b
; World Health Organization, 1997
; Lidegaard et al., 1998
; Dunn et al., 1999
, 2001
; Tanis et al., 2001
) attained significant estimates of risks for MI in comparisons of risks other than the Transnational study (Table I
).
We do not conduct meta-analyses or syntheses to test hypotheses. We only do such overviews to corroborate findings or to generate hypotheses, especially when aggregating observational epidemiological studies. Petitti's adaptation of Greenland and Longnecker's method was found to be particularly appropriate for this analytic synthesis (Petitti, 2000). The pooled OR from all studies since 1995 is 0.62 (0.380.99). For the four most rigorous studies it is 0.44 (0.240.80), The MICA study alone corroborates the Transnational Study. MICA was conducted exclusively in the UK and the Transnational findings for the UK were very similar. Clearly, at worst, the findings show no difference between 3gen and 2gen OC and the absolute risks are all very low. It has now been repeatedly demonstrated that 3gen OC do not add risk for MI compared with non-use. Moreover, interpreting moderate benefit is consistent with the findings of six of the seven studies. This was a goal of the basic scientists who developed the product.
The findings with respect to checking blood levels when counselling about OC and the impact on MI are also clinically important to the doctor and the woman counselled.
The greater danger of smoking as a risk factor for arterial diseases becomes clear in the studies where the lifestyle was assessed. As Tanis has emphasized, it cannot be ignored (Tanis et al., 2001).
What are the implications for clinicians? (i) Since OC have become the mainstay of prevention of pregnancy the secular trends of the pill as a risk factor for MI, beginning in the 1960s, have demonstrated increasing safety. This applies to all OC on the market (except 1gen OC) and applies to 3gen OC preparations in particular. Nevertheless, none of the data presented here, nor our conclusions in this discussion, should be interpreted as recommendations against 2gen OC or as strong recommendations favouring 3gen OC in any clinical context as policy. The choice of an approved OC should always be that of the counselling physician, based primarily on clinical judgement of one patient at a time. The 3gen OC preparations are clearly safe in respect to MI compared with non-use. The Transnational results and the corroborating meta-analyses of epidemiological studies now demonstrate benefit in comparison with 2gen OC. Only the British MICA study shows no difference. The subset(s) of older women, of women with a family history of MI or of an unfavourable profile of cardiovascular risk factors should be candidates for 3gen OC as first line reversible prevention of pregnancy. (ii) Smoking, adjusted or not for OC use, is a very strong risk factor for MI. If a woman cannot or will not cease smoking and insists on oral contraception, the data favour 3gen OC (Lewis et al., 1996b). (iii) Women with elevated blood pressure should be prescribed OC only with great caution and after it has been brought down. That is true for both 2gen and 3gen OC products. The evidence shows that measuring the blood pressure when administering OC favours the woman counselled (Lewis et al., 1996b
).
With respect to arterial adverse events in relation to 3gen OC, the recent literature offers grounds for reassurance, especially in regard to acute MI. That is true whether the interpretation is either benefit or no difference. The aggregated findings presented here suggest that it would be imprudent to curtail the armamentarium of OC for first-line use.
![]() |
Notes |
---|
4 To whom correspondence should be addressed at: 4920 Bath Rd., Bath by Kingston, Ontario, K0H 1G0, Canada.
![]() |
References |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Arthes, F.G. and Masi, A.T. (1976) Myocardial infarction in younger women. Associated clinical features and relationship to use of oral contraceptive drugs. Chest, 70, 574583.[Abstract]
Benagiano, G. and Primiero, F.M. (1999) Safety of modern oral contraception: the options for women: lessons to be learned. Hum. Reprod. Update, 5, 633638.
Beral, V. (1977) Mortality among oral-contraceptive users. Royal College of General Practitioners' Oral Contraceptives Study. Lancet, 2, 727731.[Medline]
Bergink, E.W., Hamburger, A.D., de Jager, E. and van der Vies, J. (1981) Binding of a contraceptive progestogen ORG 2969 and its metabolites to receptor proteins and human sex hormone binding globulin. J. Steroid Biochem., 14, 175183.[ISI][Medline]
Bergink, E.W., Borglin, N.E., Klottrup, P. and Liukko, P. (1982) Effects of desogestrel and levonorgestrel in low-dose oestrogen oral contraceptives on serum lipoproteins. Contraception, 25, 477485.[ISI][Medline]
Bergink, E.W., Bos, E.S., Kloosterboer, H.J., Lund, L., Nummi, S. and van der Vies, J. (1983) Oral Contraceptives and Lipoproteins. IHF, Geneva, pp. 5264.
Bergink, E.W., Kloosterboer, H.J., Lund, L. and Nummi, S. (1984) Effects of levonorgestrel and desogestrel in low-dose oral contraceptive combinations on serum lipids, apolipoproteins A-I and B and glycosylated proteins. Contraception, 30, 6172.[ISI][Medline]
Boyce, J., Fawcett, J.W. and Noall, E.W.P. (1963) Coronary thrombosis and Conovid. Lancet, i, 111.
Brown, B.G., Zhao, X.Q., Chait, A., Fisher, L.D., Cheung, M.C., Morse, J.S., Dowdy, A.A., Marino, E.K., Bolson, E.L., Alaupovic, P. et al. (2001) Simvastatin and niacin, antioxidant vitamins, or the combination for the prevention of coronary disease. N. Engl. J. Med., 345, 15831592.
Cohen, J. and Edwards, R.G. (1999) Conclusions: the relative safety of modern oral contraceptives. Hum. Reprod. Update, 5, 756771.
Dunn, N.R., Thorogood, M., de Caestecker, L. and Mann, R.D. (1997) Myocardial infarction and oral contraceptives, a retrospective casecontrol study in England and Scotland (MICA Study). Pharmacoepidemiol. Drug Safety, 6, 283289.[ISI]
Dunn, N.R., Thorogood, M., Faragher, B., de Caestecker, L., MacDonald, T.M., McCollum, C., Thomas, S. and Mann, R.D. (1999) Oral contraceptives and myocardial infarction: results of the MICA casecontrol study. Br. Med. J., 318, 15791584.
Dunn, N.R., Arscott, A. and Thorogood, M. (2001) The relationship between use of oral contraceptives and myocardial infarction in young women with fatal outcome, compared with those who survive: results from the MICA casecontrol study. Contraception, 63, 6569.[ISI][Medline]
Edwards, R.G. and Cohen, J. (1999) The recent saga of cardiovascular disease and safety of oral contraceptives. Hum. Reprod. Update, 5, 565620.
Farmer, R.D., Lawrenson, R.A., Thompson, C.R., Kennedy, J.G. and Hambleton, I.R. (1997) Population-based study of risk of venous thromboembolism associated with various oral contraceptives [see comments]. Lancet, 349, 8388.[ISI][Medline]
Farmer, R.D., Todd, J.C., Lewis, M.A., MacRae, K.D. and Williams, T.J. (1998) The risks of venous thromboembolic disease among German women using oral contraceptives: a database study. Contraception, 57, 6770.[ISI][Medline]
Farmer, R.D., Lawrenson, R.A., Todd, J.C., Williams, T.J. and MacRae, K. (1999) Oral contraceptives and venous thromboembolic disease. Analyses of the UK General Practice Research Database and the UK Mediplus database. Hum. Reprod. Update, 5, 688706.
Fuertes de la Haba, A., Cuervas, J.O., Pelegrina, I. and Bangdiwala, I. (1970) Deaths among users or oral and nonoral contraceptives. Obstet. Gynecol., 36, 597602.[ISI][Medline]
Fuertes de la Haba, A., Bangdiwala, I.S. and Pelegrina, I. (1973) Success of randomization in a controlled contraceptive experiment. J. Reprod. Med., 11, 142148.[ISI][Medline]
Godsland, I.F., Crook, D., Devenport, M. and Wynn, V. (1995) Relationships between blood pressure, oral contraceptive use and metabolic risk markers for cardiovascular disease. Contraception, 52, 143149.[ISI][Medline]
Greenland, S. (1987) Quantitative methods in the review of epidemiologic literature. Epidemiol. Rev., 9, 130.[ISI][Medline]
Greenland, S. and Longnecker, M.P. (1992) Methods for trend estimation from summarized dose-response data, with applications to meta-analysis. Am. J. Epidemiol., 135, 13011309.[Abstract]
Guillebaud, J. (1997) The Pill: and Other Forms of Hormonal Contraception. Oxford University Press Oxford, vol. 5.
Heinemann, L.A.J., Lewis, M.A., Spitzer, W.O., Thorogood, M., Guggenmoos-Holzmann, I. and Bruppacher, R. (1998) Thromboembolic stroke in young women. The Transnational Research Group of Oral Contraceptives and the Health of Young Women. A European casecontrol study on oral contraceptives. Contraception, 57, 2937.[ISI][Medline]
Inman, W.H.W. and Vessey, M.P. (1968) Investigation of deaths from pulmonary, coronary and cerebral thrombosis and embolism in women of child-bearing age. Br. Med. J., 2, 193199.[Medline]
International Federation of Fertility Societies (1999) Consensus conference on combination oral contraceptives and cardiovascular disease. Fertil. Steril., 71, 1S6S.[ISI]
Jick, H., Dinan, B. and Rothman, K.J. (1978a) Oral contraceptives and nonfatal myocardial infarction. J. Am. Med. Assoc., 239, 14031406.[Abstract]
Jick, H., Dinan, B., Herman, R. and Rothman, K.J. (1978b) Myocardial infarction and other vascular diseases in young women: role of estrogens and other factors. J. Am. Med. Assoc., 240, 25482552.[Abstract]
Jick, H., Jick, S.S., Gurewich, V., Myers, M.W. and Vasilakis, C. (1995) Risk of idiopathic cardiovascular death and nonfatal venous thromboembolism in women using oral contraceptives with differing progestagen components. Lancet, 346, 8388.
Jick, H., Jick, S.S., Myers, M.W. and Vasilakis, C. (1996) Risk of acute myocardial infarction and low-dose combined oral contraceptives (Letter). Lancet, 347, 627628.[ISI][Medline]
Jordan, W.M. (1961) Pulmonary embolism. Lancet, 2, 11461147.
Kay, C.R. (1982) Progestogens and arterial disease (evidence from the Royal College of General Practitioners' Study). Am. J. Obstet. Gynecol., 142, 762765.[ISI][Medline]
Krueger, D.E., Ellenberg, S.S., Bloom, S., Calkins, B.M., Maliza, C., Nolan, D.C., Phillips, R., Rios, J.C., Rosin, I., Shekelle, R.B. et al. (1980) Fatal myocardial infarction and the role of oral contraceptives. Am. J. Epidemiol., 111, 655674.[Abstract]
Lewis, M.A., Assman, A., Heinemann, L.A.J. and Spitzer, W.O. (1996a) Trinational casecontrol study or oral contraceptives and health. Approved revisions through September 1995. Pharmacoepidemiol. Drug Safety, 5, 4351.
Lewis, M.A., Spitzer, W.O., Heinemann, L.A.J., MacRae, K., Bruppacher, R. and Thorogood, M. (1996b) Third generation oral contraceptives and risk of myocardial infarction: an international casecontrol study. Transnational Research Group on Oral Contraceptives and the Health of Young Women. Br. Med. J., 312, 8890.
Lewis, M.A., Heinemann, L.A.J., Spitzer, W.O., MacRae, K., Bruppacher, R. and Transnational Research Group on Oral Contraceptives and the Health of Young Women (1997a) 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]
Lewis, M.A., Spitzer, W.O., Heinemann, L.A.J., MacRae, K. and Bruppacher, R. (1997b) Lowered risk of dying of heart attack with third generation pill may offset risk of dying of thromboembolism. Br. Med. J., 315, 679680.
Lewis, M.A., MacRae, K.D., Kuhl-Habichl, D., Bruppacher, R., Heinemann, L.A. and Spitzer, W.O. (1999) The differential risk of oral contraceptives: the impact of full exposure history. Hum. Reprod., 14, 14931499.
Lidegaard, O. and Edstrom, B. (1996) Oral contraceptives and acute myocardial infarction: a casecontrol study (Abstract). Eur. J. Contracept. Reprod. Health Care, 1, 74.
Lidegaard, O. and Kreiner, S. (1998) Cerebral thrombosis and oral contraceptives. A casecontrol study. Contraception, 57, 303314.[ISI][Medline]
Lidegaard, O., Edstrom, B. and Kreiner, S. (1998) Oral contraceptives and venous thromboembolism. A casecontrol study. Contraception, 57, 291301.[ISI][Medline]
Maguire, M.G., Tonascia, J., Sartwell, P.E., Stolley, P.D. and Tockman, M.S. (1979) Increased risk of thrombosis due to oral contraceptives: a further report. Am. J. Epidemiol., 110, 188195.[Abstract]
Mann, J.I. and Inman, W.H.W. (1975) Oral contraceptives and death from myocardial infarction. Br. Med. J., 2, 245248.[Medline]
Mann, J.I., Vessey, M.P., Thorogood, M. and Doll, S.R. (1975a) Myocardial infarction in young women with special reference to oral contraceptive practice. Br. Med. J., 2, 241245.[Medline]
Mann, J.I., Thorogood, M., Waters, W.E. and Powell, C. (1975b) Oral contraceptives and myocardial infarction in young women: a further report. Br. Med. J., 3, 631632.[Medline]
Mann, J.I., Inman, W.H.W. and Thorogood, M. (1976) Oral contraceptive use in older women and fatal myocardial infarction. Br. Med. J., 2, 445447.[Medline]
Mant, D., Villard-Mackintosh, L., Vessey, M.P. and Yeates, D. (1987) Myocardial infarction and angina pectoris in young women. J. Epidemiol. Community Health, 41, 215219.[Abstract]
Pentilla, I.M., Bergink, E.W., Holma, P., Hulkko, S., Makkonen, M., Pyorala, T. and Castren, O. (1983) Serum lipids and proteins during treatment with a new oral contraceptive combination containing desogestrel. Eur. J. Obstet. Gynecol. Reprod. Biol., 16, 275281.[ISI][Medline]
Petitti, D.B. (2000) Meta-analysis, decision analysis and cost-effectiveness analysis: methods for quantitative synthesis in medicine. Monographs in Epidemiology and Biostatistics. Oxford University Press, New York, vol. 2.
Petitti, D.B., Sidney, S., Bernstein, A., Wolf, S., Quesenberry, C. and Ziel, H.K. (1996) Stroke in users of low-dose oral contraceptives. N. Engl. J. Med., 335, 815.
Pincus, G. (1965) The Control of Fertility. Academic Press, New York.
Ramcharan, S., Pellegrin, F.A., Ray, R. and Hsu, J.P. (1981) The Walnut Creek Contraceptive Drug Study. Center for Population Research, Bethesda, MD.
Realini, J.P. and Goldzieher, J.W. (1985) Oral contraceptives and cardiovascular disease: a critique of the epidemiologic studies. Am. J. Obstet. Gynecol., 152, 729798.[ISI][Medline]
Rosenberg, L., Armstrong, B. and Jick, H. (1976) Myocardial infarction and estrogen therapy in premenopausal women (Letter). N. Engl. J. Med., 294, 12901291.[ISI][Medline]
Rosenberg, L., Hennekens, C.H. and Rosner, B. (1980) Oral contraceptive use in relation to nonfatal myocardial infarction. Am. J. Epidemiol., 111, 5966.[Abstract]
Royal College of General Practitioners (1974) Oral Contraception Study. Oral Contraceptives and Health: an Interim Report from the Royal College of General Practitioners. Pitman Medical, London.
Royal College of General Practitioners (1977) Effect on hypertension and benign breast disease of progestagen component in combined oral contraceptives. Royal College of General Practitioners' Oral Contraceptive Study. Lancet, i, 624.
Royal College of General Practitioners (1981) Further analysis of mortality in oral contraceptive users. Royal College of General Practitioners' Oral Contraceptive Study. Lancet, 1, 541546.[Medline]
Royal College of General Practitioners (1983) Incidence of arterial disease among oral contraceptive users. Royal College of General Practitioners' Oral Contraceptive Study. J. R. Coll. Gen. Pract., 33, 7582.[ISI][Medline]
Schwartz, S.M., Siscovick, D.A., Longstreth, W.T., Psary, B.M., Beverly, R.K., Raghunatan, T.E., Lin, D. and Koepsell, T.D. (1997) Use of low-dose oral contraceptives and stroke in young women. Ann. Intern. Med., 127, 596603.
Shapiro, S., Slone, D., Rosenberg, L., Kaufman, D.W., Stolley, P.D. and Miettinen, O.S. (1979) Oral contraceptive use in relation to myocardial infarction. Lancet, 1, 743747.[ISI][Medline]
Sidney, S., Petitti, D.B., Quesenberry, C.P. Jr, Klatsky, A.L., Ziel, H.K. and Wolf, S. (1996) Myocardial infarction in users of low-dose oral contraceptives. Obstet. Gynecol., 88, 939944.
Slone, D., Shapiro, S. and Kaufman, D.W. (1981) Risk of myocardial infarction in relation to current and discontinued use of oral contraceptives. N. Engl. J. Med., 305, 420424.[Abstract]
Solyom, A. (1972) Effect of androgens on serum lipids and lipoproteins. Lipids, 7, 100105.[ISI][Medline]
Spitzer, W.O., Thorogood, M. and Heinemann, L.A.J. (1993) Trinational casecontrol study of oral contraceptives and health. Pharmacoepidemiol. Drug Safety, 2, 2131.
Stampfer, M.J., Willett, W.C., Colditz, G.A., Speizer, F.E. and Hennekens, C.H. (1988) A prospective study of past use of oral contraceptive agents and risk of cardiovascular diseases. N. Engl. J. Med., 319, 13131317.[Abstract]
Stolley, P.D., Tonascia, J.A., Tockman, M.S., Sartwell, P.E., Rutledge, A.H. and Jacobs, M.P. (1975) Thrombosis with low-estrogen oral contraceptives. Am. J. Epidemiol., 102, 197208.[Abstract]
Suissa, S., Blais, L., Spitzer, W.O., Cusson, J., Lewis, M. and Heinemann, L. (1997) First-time use of newer oral contraceptives and the risk of venous thromboembolism [see comments]. Contraception, 56, 141146.[ISI][Medline]
Suissa, S., Spitzer, W.O., Rainville, B., Cusson, J., Lewis, M. and Heinemann, L. (2000) Recurrent use of newer oral contraceptives and the risk of venous thromboembolism. Hum. Reprod., 15, 817821.
Tanis, B.C., van den Bosch, M.A., Kemmeren, J.M., Cats, V.M., Helmerhorst, F.M., Algra, A., van der, Graaf Y. and Rosendaal, F.R. (2001) Oral contraceptives and the risk of myocardial infarction. N. Engl. J. Med., 345, 17871793.
Thomas, D.P., Merton, R.E., Wood, R.D. and Hockley, D.J. (1985) The relationship between vessel wall injury and venous thrombosis: an experimental study. Br. J. Haematol., 59, 449457.[ISI][Medline]
Thorogood, M., Mann, J.I., Murphy, M. and Vessey, M. (1991) Is oral contraceptive use still associated with an increased risk of fatal myocardial infarction? Report of a casecontrol study. Br. J. Obstet. Gynaecol., 98, 12451253.[ISI][Medline]
Todd, J., Lawrenson, R., Farmer, R.D., Williams, T.J. and Leydon, G.M. (1999) Venous thromboembolic disease and combined oral contraceptives: A re-analysis of the MediPlus database. Hum. Reprod., 14, 15001505.
Vessey, M.P. and Doll, R. (1968) Investigation of relation between oral contraceptives and thromboembolic disease. A further report. Br. Med. J., 2, 199205.
Vessey, M.P. and Doll, R. (1969) Investigation between oral contraceptives and thromboembolic disease: a further report. Br. Med. J., 2, 651657.[ISI][Medline]
Vessey, M.P., McPherson, K. and Johnson, B. (1977) Mortality among women participating in the Oxford/Family Planning Association contraceptive study. Lancet, 2, 731733.[Medline]
Vessey, M.P., McPherson, K. and Yeates, D. (1981) Mortality in oral contraceptive users (Letter). Lancet, 1, 549550.[Medline]
World Health Organization (1995a) Venous thromboembolic disease and combined oral contraceptives: results of international multicentre casecontrol study. WHO Collaborative Study of Cardiovascular Disease and Steroid Hormone Contraception. Lancet, 346, 15751582.[ISI][Medline]
World Health Organization (1995b) Effect of different progestagens in low eostrogen oral contraceptives on venous thromboembolic disease. WHO Collaborative Study of Cardiovascular Disease and Steroid Hormone Contraception. Lancet, 346, 15821588.[ISI][Medline]
World Health Organization (1996a) Haemorrhagic stroke, overall stroke risk and combined oral contraceptives: results of an international, multicentre, casecontrol study. WHO Collaborative Study of Cardiovascular Disease and Steroid Hormone Contraception. Lancet, 348, 505510.[ISI][Medline]
World Health Organization (1996b) Ischaemic stroke and combined oral contraceptives: results of an international, multicentre, casecontrol study. WHO Collaborative Study of Cardiovasular Disease and Steroid Hormone Contraception. Lancet, 348, 498505.[ISI][Medline]
World Health Organization (1997) Acute myocardial infarction and combined oral contraceptives: results of an international multicentre casecontrol study. WHO Collaborative Study of Cardiovascular Disease and Steroid Hormone Contraception. Lancet, 349, 12021209.[ISI][Medline]
Submitted on December 12, 2001; accepted on April 2, 2002.