- Email: [email protected]
Oral Contraceptive Use and the Cardiovascular Health of Canadian Women
"""
'GYNAECOLOGY""'"
ORAL CONTRACEPTIVE USE AND THE CARDIOVASCULAR HEALTH OF CANADIAN WOMEN John A. Collins, MD, FRCSC, 1,2 Joanne Gunby, MSc, 1 1Department
of Obstetrics and Gynaecology, 2Department of Clinical Epidemiology and Biostatistics, McMaster University
ABSTRACT
Objectives: to review the published Uterature on the association between oral contraceptive (OC) use and cardiovascular disease, in particular venous thromboemboUsm (VTE) and acute myocardial infarction (AMI). To determine if there is an increased risk for VTE due to the use of the new progestins. To use the results of the analysis to construct a Canadian model. Methods: meta-regression was used to analyze adjusted relative risks from 18 studies for VTE and from 15 stwiies for AMI. The resulting risks were appUed to the Canadian population of OC users, to calculate event and mortaUty rates for both second and third generation progestin use. Results: the relative risk ofVTE with the use ofOCs consisting of low dose ethinyl estradiol (EE) with new progestins was 7.7 and with other progestins was 3.5, compared with non-use. The relative risk of AMI with new progestins was 1.2 and with other progestins was 2.9. In the Canadian model, there were 25 VTE and two AMI events annually/lOO,OOO women for users ofOCs with new progestins, and 11 VTE and four AMI events for users of OCs with other progestins, compared to three VTE and two AMI events for similar nonusers. The exclusive use of OCs with new progestins potentially would decrease by seven the annual number of OC-attributable cardiovascular deaths in Canada. Conclusions: the published results appear to exaggerate both the VTE risk and the AMI benefit associated with the new progestins, because of bias inclwiing differences in the duration of use. Oral contraceptives should be avoided by women with other risk factors for cardiovascular disease; however, the typical OC user should not avoid or choose an OC based on the extremely small shifts in the cardiovascular risks between VTE and AMI. RESUME
Objectifs : Examiner la Utterature publiee sur le Uen entre I' utilisation de contraceptifs oraux (CO) et les maladies cardiovasculaires, en particulier la thromboembolie veineuse (TEV) et l'infarctus aigu du myocarde (lAM). Determiner si l'utilisation de nouveaux progestatifs cause un risque accru de TEV. Utiliser les resultats de l'analyse pour creer un modele canadien. Methodes: Une metaregression a permis d'analyser les risques relatifs d' apres 18 etudes sur Ia TEV et 15 etudes sur I' lAM. Les risques ainsi determines ont ere appUques it Ia population canadienne d' utilisatrices de CO pour calculer les taux d' accident cardiovasculaire et de mortalire decoulant de l'utilisation de progestatifs de deuxierne et de troisieme generation.
JOURNAL SOGe
125
FEBRUARY 1997
, , , Resultats : Le risque relatif de TEV attribuable Ii I'utilisation de CO composes d' ethinyiestradiol (EE) Ii faible dose en association avec de nouveaux progestatifs etait de 7,7 et, avec d'autres progestatifs, il etait de 3,5 par comparaison avec Ia non-utilisation. Le risque relatif d'lAM etait de 1,2 avec les nouveaux progestatifs et de 2,9 avec d'autres progestatifs. Dans Ie modele canadien, 25 TEV et deux lAM survenaient annueUement pour 100 000 femmes chez les utilisatrices de co en association avec de nouveaux progestatifs, ainsi que 11 TEV et quatre lAM chez les utilisatrices de CO en association avec d'autres progestatifs, par comparaison avec trois TEV et deux lAM chez des non-utilisatrices semblables. Au Canada, I'utilisation exclusive de CO en association avec de nouveaux progestatifs pourrait eventuellement reduire de sept fois Ie nombre annuel de deces d' arigine cardiovasculaire attribuables aux CO. Conclusions : Les resultats publies semblent exagerer Ii Ia fois Ie risque de TEV et Ia diminution du risque d'lAM qui sont assoaes Ii Ia prise de nouveaux progestatifs, en raison d' une distorsion et notamment de differences sur Ie plan de Ia duree d' utilisation. Les femmes qui presentent d'autres facteurs de risque de maIadie cardiovasculaire devraient eviter de prendre des contraceptifs oraux; I' utilisatrice type de CO ne devrait pas eviter au choisir un CO en fonction de tres /egeres variations des risques d' ordre cardiovasculaire entre Ia TEV et I'lAM.
J SOGe
1997;19:125-37
KEY WORDS
Contraceptives, oral, thromboembolism, myocardial infarction, meta-analysis, models, statistical. Received on June llnd, 1996. Revised and accepted on August 30th, 1996.
agreed that VTE rates appeared to be higher with these new progestins, compared with standard formulations. In one of these five studies, there was a non-significant trend to lower AMI risks. 5 The present paper describes a review of all the available published studies which evaluated VTE or AMI risk among OC users compared with non-users. Special attention was paid to any change in risks observed with use of the new progestins.
INTRODUCTION
Thirty-six years after steroidal contraception became available, an understanding is just emerging about the rare but serious effects, including cancer and cardiovascular disease. With respect to cancer, the gains appear to outweigh the losses. Although the risks of cervical and breast cancers are slightly increased among oral contraceptive (OC) users, the risks of endometrial and ovarian cancers are substantially reduced.! With respect to cardiovascular health, we remain uncertain about the effects of OC use on stroke, but patterns are emerging for venous thromboembolism (VTE) and acute myocardial infarction (AMI). In Canada today, there are approximately 6.8 million women of reproductive age (15 to 44 years). According to published estimates, about 1.2 million of them currently use OCs, the majority in the 15 to 35 age range. For non-OC users in this age group, the incidence of cardiovascular events is low. Although the baseline risks of cardiovascular events are low and the increases in risk due to OC use are moderate, the fear arising from these risks has the potential to affect a large proportion of the young female population. For this reason, an evaluation of the effects of OC use on cardiovascular health in Canada should be of interest to all physicians who prescribe OCs for their patients and for those concerned with public health issues. Eight recent papers from five studies reported cardiovascular risks among women using oral contraceptives containing a new type of progestin. 2•9 All the papers
JOURNAL SOGe
BACKGROUND
The early formulations of oral contraceptives, referred to as first generation OCs, had ethinyl estradiol (EE) doses of 50 Jlg or more combined with high progestin content, and they were associated with increased risks of cardiovascular events. IO- 16 In the second generation OC, introduced in the 1970s, dosages ofEE and progestin were decreased, with an accompanying reduction in cardiovascular risks. Third generation OCs, available for the past 15 years, have EE doses in the 20 to 35 Jlg range and one of three new progestins: gestodene, desogestrel or norgestimate. Only the latter two are available in Canada. With lower amounts of androgenicity and anti-estrogenic potential, the new progestins were designed to minimize unpleasant side effects and to reduce the carbohydrate and lipid metabolism changes seen with previous formulations.!7.1B The world-wide use of these third generation OCs was increasing before the most recent negative publicity. The negative stories reflected concerns that arose in Europe in the early 1990s that the use of third generation OCs was resulting in an increase in the incidence
126
FEBRUARY 1997
prFOSAMAX® increases bone mass in postmenopausal women
8.8% increase in BMD at the spine l, .. ·t.i
7.8%
3.1%
increase in BMD' at the hip'·t.I
increase in BMD at the wrist (ultradistal forearmj '·t.i
20% of bone mass has been lost by the average postmenopausal wo man FOSAMAX- is a bone metabolism regulator. FOSAMAXe is indicated for the treatment of osteoporosis in postmenopausal women .. Bone Mineral Density ... In clinical studies, over 96% of patients studied for up to three years had a measured increase in spine BMD_I t FOSAMAX* 10 mg daily produced statistically significant and clinically important increases in BMD
at the hip, spine. and wrist (ultradistal forearm) relative to placebo at three years (pSO.OOl )1.2
+Combined data from two large. identically designed. double-blind. placebo-controlled. three-year
multicenter studies in 994 women with osteoporosis. defined as low bone mass. 397 of whom received placebo and 196 of whom received FOSAMAX- 10 mg/day_ To ensure an adequate calcium intake. all patients were supplemented with 500 mg of calcium per day. I I. Liberman UA et at. Effect of oral alendronate on bone mineral density and the incidence of fractures in postmenopausal osteoporosis. N Engll Med 1995:333(221. 1437-43 2 Data on file. Merck Frosst canada Inc.: lWo double-blind. randomized. placebo-controlled. parallelgroup. multicenter studies to evaluate the safety and effect on bone density of daily oral MK-217 for two years in osteopenic postmenopausal women. with a one-year open treatment extension IProtocol No. 035 (US) and 037 (lnternational)l-Three Year Data
Builds bone to build independence --rTademark Merck & Co.. Inc .• Merck Frosst Canada Inc .. licensed user
alendronate sodium
, , , of VTE. Several studies were adapted 2 or initiated4,6,8 to investigate the issue. At first glance, the published results confinned the suspicions; the relative risks for VTE were, on average, 70 percent higher for users of OCs with new progestins than for users of OCs with older progestins and similar low EE doses. Although it appeared that physicians and their patients had good reason to be wary of third generation oral contraceptives, further infonnation emerging from these and other studies provides a more reassuring outlook.
There were 18 studies contributing 41 risk estimates for VTE to the analysis: 19 estimates of OP combined with EE dose ~50 Ilg, 14 for OP with EE dose <50 Ilg, and eight for NP with EE dose <50 Ilg.2-4.6,B,IO,14,20,2) The risk estimates from these studies for each category of OC are shown in Figure 1. All but one of the risk estimates were greater than unity with wide, overlapping confidence intervals. On average, the risk ofVTE for use ofOPs with EE dose 250 Ilg was 5.0-fold higher (95% CI 4.1 to 6.1) than for non-users. The contribution from the use ofNPs increased the risk 2.2-fold (95% CI 1.6 to 3.2). For both types of progestin, the risk was reduced by 30 percent (relative risk 0.7 [95% CI 0.5 to 0.9]) when combined with EE dose <50 Ilg . Thus, the calculated risk of VTE for users of OPs with <50 Ilg EE is 5.0 x 0.7 = 3.5-fold higher than the risk for non-users, and for NP users the risk is 5.0 x 2.2 x 0.7 = 7.7-fold higher (Table 1).
METHODS
To put the problem in perspective, the recent studies were added to past studies, obtained from a thorough search of the literature, which assessed the risk of VTE or AMI in OC users compared to non-users. Each study provided one or more relative risk estimates for OC users, usually adjusted for such co-variates as age, duration of use and smoking. Statistical techniques were used to combine the published risks, weighted according to their sample size and the precision of their risk estimates. Where adjusted estimates were unavailable, the adjusted relative risk was calculated from the raw odds ratio and a study-specific adjustment factor. Meta-regression analysis,19 using the logarithm of the weighted adjusted relative risks as the dependant variable, assessed the effects on the risk of VTE or AMI of estrogen dose <50 Ilg versus ~50 Ilg, the new third generation progestins versus first or second generation progestins (referred to as "other progestins" or "OPs"), and for AMI, the effect of current OC use versus past use.
2.
There were 15 studies contributing 25 risk estimates for AMI to the analysis: four estimates for past use of OPs combined with EE dose ~50 Ilg, seven for past use of OPs with EE dose <50 Ilg, six for current use of OPs with EE dose ~50 Ilg, seven for current use of OPs with EE dose <50 Ilg, and one for current use ofNPs with EE dose <50 Ilg. 5.10,15.l6.21,2B,}7 The risk estimates from these studies for each category of OC are shown in Figure 2. For past use, the risk estimates fall on both sides of unity and the confidence intervals overlap, indicating little or no increased risk on the basis of the limited data available. For current use, all risk estimates except one are greater than unity with wide overlapping confidence intervals.
RESULTS
1.
RISK ESTIMATES FOR ACUTE MYOCARDIAL
INFARCTION
RISK ESTIMATES FOR VENOUS THROMBO-
EMBOLISM
In four studies which estimated relative risk for the three new progestins, on average the risk of VTE was estimated to be 5.7 for gestodene (95% confidence interval [CI] 3.0 to 10.7),5.9 for desogestrel (95% CI 3.4 to 10.2), and 5.6 for norgestimate (95% CI 1.8 to 17.3) users, respectively, compared with non_users. 2,4,6,8 Because the estimates for the three new progestins were similar, they have been combined into the category "new progestins" or "NPs" for the remainder of the present analysis.
JOURNAL SOGe
TABLE 1 SUMMARY OF VTE AND AMI RISKS ASSOCIATED WITH ORAL CONTRACEPTIVE USE, ESTIMATED FROM META-REGRESSIONS Currency
Ethinyl Estradiol Dose
Progestin Type
Past Use
2: 50 1l9 <50 119 2: 50 1l9 <50 119 <50 119
Current Use
128
FEBRUARY 1997
VTE Relative Risk Estimate (compared to non-users)
AMI Relative Risk Estimate (compared to non-users)
OP
-
OP
-
1.2 1.3
OP
5.0 3.5 7.7
2.6 2.9 1.2
OP NP
FOSAMAX®
reduces the risk of fractures There was a trend towards a reduction in the proportion of patients treated with FOSAMAX® experiencing one or more nonvertebral fractures relative to those receiving placebo in pooled analysis (5-20 mg) (p=NS) '·tt
48%
reduction in the proportion of patients treated with FOSAMAX® experiencing one or more vertebral fractures relative to those treated with placebo in pooled analysis (5-20 mg) (p=0.03),·1
Low bone mass is a major pred ictor o f increased risk of osteoporotic fractu res' Vertebral fractures occurred in 62% (221355) of patients who received placebo and 3.2% (171526) of patients who received FOSAMAX. (5 or 10 mg for J years or 20 mg for 2 years followed by 5 mg for I yeaT),1 tt NOl1venebral fractures occurred in 9.6% (381397) of patients who received placebo and 7,5% (45/5971 of patients who received FOSAMAX- (5 or 10 mg for J years or 20 mg lor 2 years followed by 5 mg for I year) A trend towards a reduced number of lIol1vertebral fractures was observed in the patients treated with FOSAMAX·.I 1 Consensus Development Conference Diagnosis. prophylaxis. and treatment of osteoporosis.Am I Med 1993 ~94646·9
FOSAMAX., like other bisphosphonates. may cause local irritation of the upper gastrointestinal mucosa Esophageal adverse experiences. such as esophagitis. esophageal ulcers and esophageal erosions have been reported in patients receiving treatment with FOSAMAXe. In some cases these have been severe and required hospitalization. Healthcare professionals should therefore be alert to any signs or symptoms signaling a possible esophageal reaction and patients should be instructed to discontinue FOSAMAX. immediately and seek medical attention if they develop dysphagia, odynophagia or retrosternal pain FOSAMAXe is contraindicated in patients who have abnormalities of the esophagus which delay esophageal emptying such as stricture or achalasia, who are unable to stand or sit upright for at least 30 minutes. who are hypersensitive to any component of this product. who are hypocalcemic or who suffer from renal insufficiency ,creatinine clearance < 35 mUmin)
alendronate sodium
Builds bone to build independence . , MERCK FROSST
BEFORE PRESCRIBING, PLEASE CONSULT THE PRESCRIBING INFORMATION.
MERCK SHARP & DDHME CANADA DIV. OF MERCK FROSST CANADA INC. KIRKLAND. OUEBEC
FSM-96-CDN-9720b-JA
, , , FIGURE 1 14 12 10 OJ 8 ...., <0 6
··
''::;
i
(fJ
W .::.t
4
~ OJ
>
2
''::;
~ OJ
a::
1
·· •
·!
·· ·
(fJ
··•
· ·
·•
E
risk of AMI (95% CI 0.7 to 1.6). Thus, compared with non-users, the risk of AMI for current users of OPs with <50 ~g EE would be 1.2 x 2.2 x 1.1 = 2.9-fold higher and for current users ofNPs would be 1.2 x 2.2 x 0.4 xLI = l.2-fold higher (Table 1).
I
EFFECT OF NEW PROGESTINS ON RISK OF VTE AND AMI
·
OP
OP
NP
>=50 1i9 EE
<50 1i9 EE
<50 1i9 EE
This analysis has summarized the reported cardiovascular risks associated with the use of OCs. Taken at face value, the new progestins are associated with a two-fold increase in the risk for VTE and a potential 60 percent reduction in the risk for AMI, compared with use of first and second generation OCs. One additional study, which was not included in our analysis, compared second and third generation OC use directly, and also found a doubling in the risk for VTE with third generation progestins. 9 The question to be asked at this point is: are these effects real or simply the result of bias? Before discussing methodological bias, it would be helpful to assess whether there is a biological basis for the observations. As the analysis confirmed, VTE risk has long been associated with higher estrogen dosages. The new progestins, although combined with lower estrogen dosages, have been shown to be more estrogenic than second generation progestins in the totality of their effects on metabolic markers. Five studies indicated that sex hormone binding globulin (SHBG) levels (an estrogen-mediated effect) were higher for third generation versus second generation combinations.J8 -42 Five randomized clinical trials demonstrated that high density lipoprotein (HDL) cholesterol concentrations (a further estrogen-mediated effect} were higher after six months exposure to third generation compared with second generation OC formulations. J8.40•4l-45 Thus, there is biological plausibility for increased estrogenicity and an increase in the risk of VTE among third generation OC users. The observed reduction in OC-associated AMI risk would also be consistent with increased estrogen activity. Estrogen exposure is believed to reduce AMI rates among post-menopausal women, through mechanisms which may involve lipid optimization, platelet effects, and small vessel dilatation. 46.47 The welldocumented estrogenic effect on metabolic markers is, therefore, consistent with an estrogen effect on both clinical events, VTE and AMI.
Relative risk estimates from 18 publ ished studies of venous thromboembolism for users of oral contraceptives (OC) compared with non-users. The estimates are grouped according to the dose of ethinyl estradiol (EEl and type of progestin (other progestins, OP; new progestins, NPl.
On average, the risk of AMI for past use of OPs with EE dose ;:::50 ~g was 1.2-fold higher than for non-users (95% CI 0.8 to 1. 7). The risk associated with current use of such products was 2.2-fold higher (95% CI 1.5 to 3.3) than with past use. The AMI risk with use of NPs was reduced by 60 percent (relative risk 0.4 [95% CI 0.1 to 1.4]). The lower dose range of estrogen was associated with a non-significant 1. I-fold increase in the overall
FIGURE 2
· ·
10 8 OJ ....,
6
E
4
<0
''::;
·
(fJ
w
.::.t (fJ
~
·
2
· ··
OJ
> ''::; <0
Q)
a::
1
0.5
:
..
I
· ·
· I
·
op
op
op
op
>-so"g EE
>-so"g EE
+-- po.t U..
. NP
........---. Current Use - - - . . . .
Relative risk estimates from 15 published studies of acute myocardial infarction for users of oral contraceptives (OCl compared with non-users. The estimates are grouped according to the dose of ethinyl estradiol (EEl and type of progestin (other progestins, OP; new progestins, NPl and past or current use.
JOURNAL SOGe
130
FEBRUARY 1997
, , , On the other hand, a methodological bias may be partly responsible for the OC generation effect on VTE and AMI. An inherited mutation, the Factor V Leiden defect, present in about six percent of the European population,48 is known to increase the risk of VTE about eight-fold. s Oral contraceptive use may increase further the susceptibility to VTE in women with Factor V Leiden defect. 49 Other hereditary factors, which also may be pro-coagulant, are known to occur among young women, including protein C or S deficiency. If a woman is genetically susceptible to suffering a VTE event, it will probably occur in the first few months of OC use, after which she will discontinue use. Continuing users, therefore, have a lower risk ofVTE than new users. New drugs are more likely to be prescribed to new users; thus, the third generation OC group will include proportionally more new users than the second generation OC group. As a result, the duration of use will be longer, on average, in second generation users compared with third generation users. The presence of this bias required the authors of
two large studies to adjust some of their estimates for duration of use.3,4 It appears, therefore, that duration of use bias may account in part for the observed increase in VTE risk for third generation users. This same duration of use bias also may be affecting the results of the AMI analysis. Longer duration of OC use is associated with a higher risk of AMI,29 thus putting second generation OC users, with longer duration, at a disadvantage. With respect to AMI risk, a further caution is needed: only one published study estimated the effects of third generation OC use on risks of AMI, and the number of cases that had used third generation OCs was only six. 5 Ideally, larger numbers are preferable, but if third generation OC use truly reduces AMI risk, such cases will continue to be rare. Our estimate of AMI risk reduction with NPs is not Significant; however, it is consistent with the findings in a published study that did not meet our entry criteria.9 Although it should be noted that the recently published risks of VTE and AMI for third generation OC
On ce you pre crib e Tripha ii , yo u' ll rind th a t mo t p ati e nts
and re du ces eITors.'·' A
Passport for FIRST TIME
well , only Triphasil ha th e Pill I ass port
for J?irst Tim e U es, r a ERS
comprehen ive patie nt edu cation
program . Send he r hom e with a kit that in c lud e
rill e d compute r di s k, bookl e t
that te ll he r how to tak e Tl'ipha il and what
to ex pect durin g he r fir t ph ys ica l exam, and informati on about th Start He lp Lin e . Odd l! '
anada '
I'\'Jo
h e ca n lea rn all a bout Tl'ipha il in th e co mfort
are when yo u pre c rib e T'l'ipha ii , th ey le arn about a go od rh ythm. Presc ribe Tripha ii , first a nd la st. Pr e
c l'ib e d
Oral
Con lr acep Li
Please rcf('r to Ih(' product llIonograph for· delai ll'd
JOURNAL SOGC
131
a fa ct-
~ilr(;'t~
ilifonn
FEBRUARY 1997
e'
, , , TABLE 2 ANNUAL CARDIOVASCULAR EVENT RATES PER 100,000 WOMEN IN CANADA BY ORAL CONTRACEPTIVE USE AND AGE GROUP Age Range (years)
Venous Thromboembolism Rates Non-OC Users
OP Users
NP Users
Acute Myocardial Infarction Rates Non-OC Users
OP Users
NP Users
lS-19
2.3
7 .6
17.0
0.7
1 .7
0.7
20-24
2.7
8.9
20.0
0.8
2.1
0.9
25-29
3.2
10.8
24.2
1 .1
2.8
1.1
30-34
4.1
13.8
30.8
1 .5
4.1
1.7
35-39
5.7
18.9
42.4
2.8
7.4
3.0
40-44
9.0
30.2
67.7
15.0
40.0
16.3
users may be influenced in part by bias and should not be taken at face value, these studies provide the only information that is currently available on VTE and AMI risk with the new progestins. THE CANADIAN MODEL: VTE AND AMI RISKS APPLIED TO CANADIAN OC USERS
Although there is good reason to believe that both the hazards and the benefits may be exaggerated in the study results, it is these published results that form the basis of the calculated model to follow. To fit the current published results into a Canadian context, we estimated the number of OC users in Canada by age group, and determined the baseline VTE and AMI rates for young Canadian women who do not use 0Cs. From this information, plus the estimates of relative risk derived from the regression model, the number of events among OC users were calculated. We also estimated the mortality that might be associated with these events. Table 2 shows the baseline cardiovascular rates for six age ranges of women who do not use OCs. 50 To obtain the rates of VTE and AMI for users of OCs containing OP or NP, each age-specific baseline rate was multiplied by the appropriate relative risk calculated from the meta-regression results (Table 1). All relative risks used in these calculations apply to OCs with EE dose <50 f..lg as the lower estrogen doses are used most commonly today. Of course, the data in Table 2 relate to 100,000 OC users at each age range, but OC use changes with age as do the baseline risks for VTE and AMI. It follows that the next step is to estimate the effect on cardiovascular events within a population sample in which the number of women and the proportion of OC use in each age
JOURNAL SOGC
range are taken into account. Figure 3 shows these figures, estimated for 1995 based on Statistics Canada population data and reported OC use in Canada (1984 and 1993) and the UK (1993).6.51.52 Figure 4 shows the projected numbers of annual cardiovascular events for a representative sample of 100,000 Canadian OC users, for both OPs and NPs. To obtain these estimates, the proportion of total users in each age range (derived from the data shown in Figure 3) was multiplied by the appropriate event rate from Table 2. The total number of events per 100,000 OC users were obtained by summing over all age ranges. This was compared to the predicted number of events in a similarly age-distributed group of non-users (3.4 VTE events and 1.6 AMI events, open bars). Cardiovascular events in OC users in excess of the non-user rates are referred to as "attributable" events, because they are associated with the use of 0Cs. Table 3 shows that the attributable risk of OP use was eight VTE events and two AMI events per 100,000 users. The attributable risk of NP use was 22 VTE events and zero AMI events per 100,000 users. If 100,000 second generation OC users switched to third generation products, there would be 14 more attributable VTE events and two fewer attributable AMI events in the next year. To obtain the figures for the Canadian population, these numbers are multiplied by 12, because there are an estimated 1.2 million OC users in Canada.
FIGURE 3 14
6' 0 0
12
0
10
c5
~ c
II>
E
0
~
8 6
'+-
0
Q;
4
-D
E 2 ~
Z
-
-
-
15-19
20-24
25-29
'--------
30-34
~
35-39
40-44
Age g roup (years) Distribution of the Canadian female population of reproductive age by age group (hatched bars) and the number of OC users (solid bars) in each group, estimated for 1995.
132
FEBRUARY 1997
, , , TABLE 3
annual numbers of deaths in the 1.2 million Canadian OC user population are less than one due to VTE and five due to AMI. Annually in the Canadian population, the predicted additional number of deaths attributable to OC use would be one from VTE and nine from AMI if all women used second generation products, or three from VTE and zero from AMI if all women used third generation products (Table 3). There are zero AMI deaths attributable to third generation OC use because the AMI event rate is the same for users and non-users (Figure 4). If all Canadian women were using second generation OCs and switched to third generation products, there would be two more VTE deaths and nine fewer AMI deaths in Canada annually, for a net reduction of seven in the annual number of deaths among OC users.
PREDICTED ANNUAL ATTRIBUTABLE CARDIOVASCULAR EVENTS AND DEATHS IN ORAL CONTRACEPTIVE USERS IN CANADA OP Users
NP Users
VTE
AMI
Total
VTE
AMI
Total
Attributable events/100,000*
8
2
10
22
0
22
Attributable deaths among 1.2 million OC users
1
9
10
3
0
3
* a representative sample across age groups of OC users
It is also important to take into account the case mortality ratio associated with cardiovascular events, which is much higherfor AMI. To examine the cardiovascular mortality associated with oral contraceptive use, we adopted published estimates that approximately 1.25 percent of VTE events S3 •S4 and 30 percent of AMI events SS will result in death. The resulting mortality rates are shown graphically in Figure 4 as a solid bar at the base of each corresponding event bar. The baseline expected
IMPLICATIONS FOR CANADIAN OC USERS AND PRESCRIBERS
In addition to the obvious benefit of contraceptive efficacy, there are many other direct and indirect benefits
SOMETIMES, PULLING OUT
DOESN'T WORK.
If you ' re thinkin g o f pulling out of th e tri ed and true oral contra ceptive think again. With over 34 million cycle
of u e in Canada,' Tripha iI offers exce ll ent cycle control
with a low in cid e nce of ide e ffect .... A
Pass'port for FIR T T IME
for Fir t Time ER
in favour of a n ew one,
well , onl y Tripha iI ha
the Pill Pa port
e r , acomprehen ive pati ent education
program . end her hom e with a kit that include
a fact-filled
compute r di k, bookle t that tell he r how to take Tripha il and what to expect durin g her fir t physical exam, and information about the 1-8 88 Ju Line. h e can learn all about Tripha il in th Cycle tte packa ging help
comfort of her hom .
mak e compli a nce easy and redu ce
new OC offer all of that? Prescribe Tripha ii, first and last. Ca n a d a ' Mos t
Pr es
rib e d
Or a l
Co nt race pti ve'
Pl ease re fer 10 th e product rnonop-a ph for detailed sa rety 10 fonn8110n.
JOURNAL SOGe
133
FEBRUARY 1997
error .
, , , FIGURE 4
ARE THE VTE RISKS SIMILAR FOR ALL THIRD GENERATION PRODUCTS?
25
(/)
All three of the third generation progestins (desogestrel, gestodene, and norgestimate) recently introduced in Europe were associated with a two-fold increase in VTE risk, compared to second generation progestins, usually levonorgestreL
20
+"
C
Q)
>
UJ
'+-
15
0
....Q)
.D
E
10
::J
Z
SHOULD WOMEN SWITCH OR DISCONTINUE
5
0
ORAL CONTRACEPTIVES?
VTE
Women at high risk for either VTE or AMI should use non-hormonal methods of contraception. For the typical oral contraceptive user, however, cardiovascular morbidity and mortality rates are so low that they should not discourage OC use or influence the choice of one product over another.
AMI
Predicted annual number of cardiovascular events (venous thromboembolism [VTE] and acute myocardial infarction [AMI]) per 100,000 women in Canada, of an age-distribution similar to that of OC users. Number of events are shown for non-users (open bars), users of OCs with other progestins (diagonally hatched bars), and users of OCs with new progestins (crosshatched bars). The predicted annual numbers of deaths from these causes in each group are shown by the solid bars.
Is THE POPULATION EFFECT A CANADIAN PUBLIC HEALTH CONCERN?
Cardiovascular events in young women are rare, even considering the additional events due to OC use. If the published increased risks due to OC use were accepted at face value, the difference between second and third generation products would do little more than shift the health care burden between VTE and AMI. None of the studies published during the last year was conducted in North America, where there may be as many as 15 million ex::: users, making it necessary to generalize from the results of studies in other countries. It is premature, therefore, to regard as proven any excess risk ofYTE with third generation products or any excess risk of AMI with second generation products.
from oral contraceptive use, including the reduction in the incidence of ovarian and endometrial cancers. l Having noted this, how important for ex::: users and prescribers are the cardiovascular risks described in this paper? DOES ORAL CONTRACEPTIVE USE CARRY AN INCREASED RISK OF CARDIOVASCULAR DISEASE?
This analysis has shown that the use of oral contraceptives is associated with 10 to 22 additional cardiovascular events per 100,000 users annually, depending on whether OCs with second or third generation progest ins are used_ The risk of VTE is increased four to eight-fold and the risk of AMI is increased up to threefold during current OC use. Past OC use does not have a significant effect on cardiovascular health.
REFERENCES
1. Is THE VTE RISK DOUBLED FOR THIRD GENERATION OC USERS?
2.
There is a biological rationale to explain partially the increased VTE risk with third generation progestins (increased estrogenicity), yet there are also sound epidemiological reasons to be sceptical about the magnitude of this risk. Because the currently available published risks may be susceptible to bias and should not be taken at face value, we need to await the results of future welldesigned studies to judge more accurately whether there is truly an increased risk.
JOURNAL saGC
3.
134
Schlesselman JJ. Net effect of oral contraceptive use on the risk of cancer in women in the United States. Obstet GynecoI1995;85:793-801. Meirik O. World Health Organization Collaborative Study of Cardiovascular Disease and Steroid Hormone Contraception. Effect of different progestogens in low oestrogen oral contraceptives on venous thromboembolic disease. Lancet 1995;346:1582-8. Poulter NR. World Health Organization Collaborative Study of Cardiovascular Disease and Steroid Hormone Contraception. Venous thromboembolic disease and combined oral contraceptives: results of international multicentre case-control study. Lancet 1995;346: 1575-82.
FEBRUARY 1997
, , , 10. Inman W, Vessey MP. Investigation of deaths from pulmonary, coronary, and cerebral thrombosis and embolism in women of child-bearing age. Br Med J 1968;2:193-9. 11. Vessey MP, Doll R. Investigation of relation between use of oral contraceptives and thromboembolic disease. A further report. Br Med J 1969;2:651-7. 12. Sartwell PE, Masi AT, Arthes FG, Greene GR, Smith HE. Thromboembolism and oral contraceptives: an epidemiologic case-control study. Am J EpidemioI1969;90:365-80. 13. Boston Collaborative Drug Surveillance Programme. Oral contraceptives and venous thromboembolic disease, surgically confirmed gallbladder disease, and breast tumours. Lancet 1973;1 :1399-404. 14. Royal College of General Practitioners' Oral Contraception Study. Oral contraceptives, venous thrombosis, and varicose veins. Journal of the Royal College of General Practitioners 1978;28:393-9. 15. Mann J, Inman W. Oral contraceptives and death from myocardial infarction. Br Med J 1975;2:245-8. 16. Mann J, Vessey M, Thorogood M, Doll R. Myocardial infarction in young women with special reference to oral contraceptive practice. Br Med J 1975;2:241-5.
4. Spitzer WO, Lewis MA, Heinemann lAJ, Thorogood M, MacRae KD. Third generation oral contraceptives and risk of venous thromboembolic disorders: an international case-control study. Br Med J 1996;312:83-8. 5. Lewis MA, Spitzer WO, Heinemann lAJ, MacRae KD, Thorogood M. Third generation oral contraceptives and risk of myocardial infarction: an international case-control study. Br Med J 1996;312:88-90. 6. Farmer RDT, Preston TD. The risk of venous thromboembolism associated with low oestrogen oral contraceptives. J Obstet GynaecoI1995;15:195-200. 7. Farmer RDT. Oral contraceptives and thromboembolism. Lancet 1996;347:259. 8. Bloemenkamp KWM, Rosendaal FR, Helmerhorst FM, Buller HR, Vandenbroucke JP. Enhancement by factor V Leiden mutation of risk of deep-vein thrombosis associated with oral contraceptives containing a third-generation progestagen. Lancet 1995;346:1593-6. 9. Jick H, Jick SS, Gurewich V, Myers MW, Vasilakis C. Risk of idiopathic cardiovascular death and non-fatal venous thromboembolism in young women using oral contraceptives with differing progestagen components. Lancet 1995;346: 1589-93.
WE DON'T LET
COITUS
INTERRUPT US.
Your pa ti e nl wanl a co nlra ce ptive lhal won 'l inle rrupl lh e ir lives. \' ilh l'ripha s il , yo u ca n feel
onfidenl
lhal you're prescribin g a lri ed and lrue oral co nlrace plive. Th e mo l presc rib e d in Canada. ' Tt'iphasil ha s a low in c id e nce of id e e ffecl and our uniqu e
I
Pass'port
ror FIRST TIME
yc le lle pac ka gin g help m a ke com pli ance
ea y and r edu ces e rrors .'" As \ ell , only Triphasil has lh e ER
Pill Passporl fo r Fi r l Tim e
e r , aco mpre he ns ive pali e nl
e du calion prog ram . Send he r hom e wilh a killhal includ e
a fa cl-fill e d co mpule r
di s k, bookl e ls lhalle ll h e r how lo lake Tripha il and whallo ex pecl d phy ica l xam , and information about th e 1-888 Jump
larl lI e lp Line.
all aboul Tripha il in th e co mforl of h e r hom e. This ma y impro ve pati wilh fe we r inle rrupti on s. Presc rib e Tt'iph as il , fir sl and la l. a n ad a'
Mo
P,' e
c rib ed Pl ca ~ c
Oral
Co nl racept i ve '
re fe r (0 th e product m onograph for del ailcd safcl~ infonnalion.
JOURNAL SOGe
135
FEBRUARY 1997
, , , 17. Darney PD. The androgenicity of progestins. Am J Med 1995;98: 1045-11 OS. 18. Jones EE. Androgenic effects of oral contraceptives: implications for patient compliance. Am J Med 1995;98:1165-1195. 19. Greenland S. Quantitative methods in the review of epidemiologic literature. Epidemiol Rev 1987;9:1-30. 20. Ludwig H. Anovulatory agents and venous disease. Ergebnisse der Angologie und Phlebologie 1970;4: 81-102. 21. Fuertes-de la Haba A, Curet JO, Pelegrina I, Bangdiwala I. Thrombophlebitis among oral and normal contraceptive users. Obstet Gynecol 1971 ;38:259-63. 22. Maguire MG, Tonascia J, Sartwell PE, Stolley PD, Tockman MS. Increased risk of thrombosis due to oral contraceptives: a further report. Am J Epidemiol 1979; 110: 188-95. 23. Petitti DB, Wingerd J, Pellegrin F, Ramcharan S. Risk of vascular disease in women. Smoking, oral contraceptives, noncontraceptive estrogens, and other factors. ·JAMA 1979;242: 1150-4. 24. Porter JB, Hunter JR, Danielson DA, Jick H, Stergachis A. Oral contraceptives and nonfatal vascular diseaserecent experience. Obstet Gynecol 1982;59:299-302. 25. Porter J, Hunter J, Jick H, Stergachis A. Oral contraceptives and nonfatal vascular disease. Obstet Gynecol 1985;66: 1-4. 26. Helmrich SP, Rosenberg L, Kaufman DW, Strom B, Shapiro S. Venous thromboembolism in relation to oral contraceptive use. Obstet GynecoI1987;69:91-5. 27. Vessey MP. Epidemiologic studies of oral contraception. Int J Fertil 1989;34:64-70. 28. Jick H, Dinan BJ, Rothman KJ. Oral contraceptives and nonfatal myocardial infarction. JAMA 1978;239: 1403-6. 29. Rosenberg L, Hennekens CH, Rosner B, Belanger C, Rothman KJ, Speizer FE. Oral contraceptive use in relation to non-fatal myocardial infarction. Am J Epidemiol 1980; 111 :59-66. 30. Krueger DE, Ellenberg 5, Bloom 5, Calkins BM, Maliza C, Nolan DC, et a/. Fatal myocardial infarction and the role of oral contraceptives. Am J Epidemiol 1980;111 :655-74. 31. Adam SA, Thorogood M, Mann JI. Oral contraception and myocardial infarction revisited: the effects of new preparations and prescribing patterns. Br J Obstet Gynaecol 1981 ;88:838-45. 32. Slone D, Shapiro 5, Kaufman D, Rosenberg L, Miettinen 0, Stolley P. Risk of myocardial infarction in relation to current and discontinued use of oral contraceptives. N Engl J Med 1981;305:420-4. 33. La Vecchia C, Franceschi 5, Decarli A, Pampa Ilona 5, Tognoni G. Risk factors for myocardial infarction in young women. Am J EpidemioI1987;125:832-43. 34. Ananijevic-Pandey J, Vlajinac V. Myocardial infarction in young women with reference to oral contraceptive use. Int J Epidemiol 1989; 18:585-8.
JOURNAL SOGe
35. Croft P, Hannaford Pc. Risk factors for acute myocardial infarction in women: evidence from the Royal College of General Practitioners' Oral Contraception Study. Br Med J 1989;298:165-8. 36. Rosenberg L, Palmer JR, Lesko SM, Shapiro S. Oral contraceptive use and the risk of myocardial infarction. Am J EpidemioI1990;131:1009-16. 37. Thorogood M, Mann J, Murphy M, Vessey M. Is oral contraceptive use still associated with an increased risk of fatal myocardial infarction? Report of a case-control study. Br J Obstet Gynaecol 1991 ;98: 1245-53. 38. Ball MJ, Ashwell E, Jackson M, Cillmer MDG. Comparison of two triphasic contraceptives with different progestogens: effects on metabolism and coagulation proteins. Contraception 1990;41 :363-76. 39. Becker H. Supportive European data on a new oral contraceptive containing norgestimate. Acta Obstet Gynecol Scand 1990 (Suppl);152:33-9. 40. Chapdelaine A, Desmarais J-L, Derman RJ. Clinical evidence of the minimal androgenic activity of norgestimate.lntJ FertiI1989;34:347-52. 41. Kloosterboer HJ, van Wayjen RGA, van den Ende A. Effects of three low-dose oral contraceptive combinations on sex hormone binding globulin, corticosteroid binding globulin and antithrombin III activity in healthy women: two monophasic desogestrel combinations (containing 0.020 or 0.030 mg ethynlestradiol) and one triphasic levonorgestrel combination. Acta Obstet Gynecol Scand 1987 (Suppl);144:41-4. 42. Refn H, Kjaer A, Lebech A-M, Borggaard B, Schierup L, Bremmelgaard A. Metabolic changes during treatment with two different progestogens. Am J Obstet Gynecol 1990;163:374-7. 43. Harvengt C, Desager JP, Gaspard U, Lepot M. Changes in lipoprotein composition in women receiving two low-dose oral contraceptives containing ethinyl estradiol and gonane progestins. Contraception 1988;37:565-75. 44. Ho PC, Liu WT, Kwan MSW. Serum lipids in Chinese patients using oral contraceptive pills. Contraception 1990;41 :55-61. 45. Kjaer A, Lebech A-M, Borggaard B, Refn H, Pedersen LR, Schierup L, et al. Lipid metabolism and coagulation of two contraceptives: correlation to serum concentrations of levonorgestrel and gestodene. Contraception 1989;40:665-73. 46. The Writing Group for the PEP I Trial. Effects of estrogen or estrogen/progestin regimens on heart disease risk factors in postmenopausal women: the postmenopausal estrogen/progestin interventions (PEPI) trial. JAMA 1995;273:199-208. 47. Chester AH, Jiang C, Borland JA, Yacoub MH, Collins P. Oestrogen relaxes human epicardial coronary arteries through non-endothelium-dependent mechanisms. Coron Artery Dis 1995;6:417-22.
136
FEBRUARY 1997
, , , 48. Dahlback B. Physiological anticoagulation: resistance to activated Protein C and venous thromboembolism. J Clin Invest 1994;94:923-7. 49. Henkens CMA, Born VJJ, Seinen AJ, van der Meer J. Sensitivity to activated Protein C: influence of oral contraceptives and sex. Thromb Haemost 1995;73:402-4. 50. Lis Y, Spitzer WO, Mann RD, Cockburn I, Chukwujindu J, Thorogood M, et al. A concurrent cohort study of oral contraceptive users from the VAMP research bank. Pharmacoepidemiol Drug Safety 1993;2:51-63. 51. Balakrishnan TR, Krotki K, Lapierre-Adamcyk E. Contraceptive use in Canada, 1984. Fam Plann Perspect 1985;17:209-15. 52. Boroditsky R, Fisher W, Sand M. The Canadian Contraception Study. J Soc Obstet Gynaecol Can 1995; 17(Suppl): 1-28. 53. Creager MA, Dzau VJ. Vascular diseases of the extremities. In: Isselbacher KJ, Braunwald E, Wilson JD, Martin JB, Fauci AS, Kasper DL (Eds). Harrison's Principles of Internal Medicine. 13th ed. New York: McGraw-HiII,lnc. 1994:1135-41. 54. Moser KM. Pulmonary thromboembolism. In: Isselbacher KJ, Braunwald E, Wilson JD, Martin JB, Fauci AS, Kasper DL (Eds). Harrison's Principles of Internal Medicine. 13th ed. New York: McGraw-HiII,lnc. 1994:1214-25. 55. Pasternak RC, Braunwald E. Acute myocardial infarction. In: Isselbacher KJ, Braunwald E, Wilson JD, Martin JB, Fauci AS, Kasper DL (Eds). Harrison's Principles of Internal Medicine. 13th ed. New York: McGraw-HiII,lnc. 1994:1066-97.
JOURNAL SOGC
137
FEBRUARY 1997
'GYNAECOLOGY""'"
ORAL CONTRACEPTIVE USE AND THE CARDIOVASCULAR HEALTH OF CANADIAN WOMEN John A. Collins, MD, FRCSC, 1,2 Joanne Gunby, MSc, 1 1Department
of Obstetrics and Gynaecology, 2Department of Clinical Epidemiology and Biostatistics, McMaster University
ABSTRACT
Objectives: to review the published Uterature on the association between oral contraceptive (OC) use and cardiovascular disease, in particular venous thromboemboUsm (VTE) and acute myocardial infarction (AMI). To determine if there is an increased risk for VTE due to the use of the new progestins. To use the results of the analysis to construct a Canadian model. Methods: meta-regression was used to analyze adjusted relative risks from 18 studies for VTE and from 15 stwiies for AMI. The resulting risks were appUed to the Canadian population of OC users, to calculate event and mortaUty rates for both second and third generation progestin use. Results: the relative risk ofVTE with the use ofOCs consisting of low dose ethinyl estradiol (EE) with new progestins was 7.7 and with other progestins was 3.5, compared with non-use. The relative risk of AMI with new progestins was 1.2 and with other progestins was 2.9. In the Canadian model, there were 25 VTE and two AMI events annually/lOO,OOO women for users ofOCs with new progestins, and 11 VTE and four AMI events for users of OCs with other progestins, compared to three VTE and two AMI events for similar nonusers. The exclusive use of OCs with new progestins potentially would decrease by seven the annual number of OC-attributable cardiovascular deaths in Canada. Conclusions: the published results appear to exaggerate both the VTE risk and the AMI benefit associated with the new progestins, because of bias inclwiing differences in the duration of use. Oral contraceptives should be avoided by women with other risk factors for cardiovascular disease; however, the typical OC user should not avoid or choose an OC based on the extremely small shifts in the cardiovascular risks between VTE and AMI. RESUME
Objectifs : Examiner la Utterature publiee sur le Uen entre I' utilisation de contraceptifs oraux (CO) et les maladies cardiovasculaires, en particulier la thromboembolie veineuse (TEV) et l'infarctus aigu du myocarde (lAM). Determiner si l'utilisation de nouveaux progestatifs cause un risque accru de TEV. Utiliser les resultats de l'analyse pour creer un modele canadien. Methodes: Une metaregression a permis d'analyser les risques relatifs d' apres 18 etudes sur Ia TEV et 15 etudes sur I' lAM. Les risques ainsi determines ont ere appUques it Ia population canadienne d' utilisatrices de CO pour calculer les taux d' accident cardiovasculaire et de mortalire decoulant de l'utilisation de progestatifs de deuxierne et de troisieme generation.
JOURNAL SOGe
125
FEBRUARY 1997
, , , Resultats : Le risque relatif de TEV attribuable Ii I'utilisation de CO composes d' ethinyiestradiol (EE) Ii faible dose en association avec de nouveaux progestatifs etait de 7,7 et, avec d'autres progestatifs, il etait de 3,5 par comparaison avec Ia non-utilisation. Le risque relatif d'lAM etait de 1,2 avec les nouveaux progestatifs et de 2,9 avec d'autres progestatifs. Dans Ie modele canadien, 25 TEV et deux lAM survenaient annueUement pour 100 000 femmes chez les utilisatrices de co en association avec de nouveaux progestatifs, ainsi que 11 TEV et quatre lAM chez les utilisatrices de CO en association avec d'autres progestatifs, par comparaison avec trois TEV et deux lAM chez des non-utilisatrices semblables. Au Canada, I'utilisation exclusive de CO en association avec de nouveaux progestatifs pourrait eventuellement reduire de sept fois Ie nombre annuel de deces d' arigine cardiovasculaire attribuables aux CO. Conclusions : Les resultats publies semblent exagerer Ii Ia fois Ie risque de TEV et Ia diminution du risque d'lAM qui sont assoaes Ii Ia prise de nouveaux progestatifs, en raison d' une distorsion et notamment de differences sur Ie plan de Ia duree d' utilisation. Les femmes qui presentent d'autres facteurs de risque de maIadie cardiovasculaire devraient eviter de prendre des contraceptifs oraux; I' utilisatrice type de CO ne devrait pas eviter au choisir un CO en fonction de tres /egeres variations des risques d' ordre cardiovasculaire entre Ia TEV et I'lAM.
J SOGe
1997;19:125-37
KEY WORDS
Contraceptives, oral, thromboembolism, myocardial infarction, meta-analysis, models, statistical. Received on June llnd, 1996. Revised and accepted on August 30th, 1996.
agreed that VTE rates appeared to be higher with these new progestins, compared with standard formulations. In one of these five studies, there was a non-significant trend to lower AMI risks. 5 The present paper describes a review of all the available published studies which evaluated VTE or AMI risk among OC users compared with non-users. Special attention was paid to any change in risks observed with use of the new progestins.
INTRODUCTION
Thirty-six years after steroidal contraception became available, an understanding is just emerging about the rare but serious effects, including cancer and cardiovascular disease. With respect to cancer, the gains appear to outweigh the losses. Although the risks of cervical and breast cancers are slightly increased among oral contraceptive (OC) users, the risks of endometrial and ovarian cancers are substantially reduced.! With respect to cardiovascular health, we remain uncertain about the effects of OC use on stroke, but patterns are emerging for venous thromboembolism (VTE) and acute myocardial infarction (AMI). In Canada today, there are approximately 6.8 million women of reproductive age (15 to 44 years). According to published estimates, about 1.2 million of them currently use OCs, the majority in the 15 to 35 age range. For non-OC users in this age group, the incidence of cardiovascular events is low. Although the baseline risks of cardiovascular events are low and the increases in risk due to OC use are moderate, the fear arising from these risks has the potential to affect a large proportion of the young female population. For this reason, an evaluation of the effects of OC use on cardiovascular health in Canada should be of interest to all physicians who prescribe OCs for their patients and for those concerned with public health issues. Eight recent papers from five studies reported cardiovascular risks among women using oral contraceptives containing a new type of progestin. 2•9 All the papers
JOURNAL SOGe
BACKGROUND
The early formulations of oral contraceptives, referred to as first generation OCs, had ethinyl estradiol (EE) doses of 50 Jlg or more combined with high progestin content, and they were associated with increased risks of cardiovascular events. IO- 16 In the second generation OC, introduced in the 1970s, dosages ofEE and progestin were decreased, with an accompanying reduction in cardiovascular risks. Third generation OCs, available for the past 15 years, have EE doses in the 20 to 35 Jlg range and one of three new progestins: gestodene, desogestrel or norgestimate. Only the latter two are available in Canada. With lower amounts of androgenicity and anti-estrogenic potential, the new progestins were designed to minimize unpleasant side effects and to reduce the carbohydrate and lipid metabolism changes seen with previous formulations.!7.1B The world-wide use of these third generation OCs was increasing before the most recent negative publicity. The negative stories reflected concerns that arose in Europe in the early 1990s that the use of third generation OCs was resulting in an increase in the incidence
126
FEBRUARY 1997
prFOSAMAX® increases bone mass in postmenopausal women
8.8% increase in BMD at the spine l, .. ·t.i
7.8%
3.1%
increase in BMD' at the hip'·t.I
increase in BMD at the wrist (ultradistal forearmj '·t.i
20% of bone mass has been lost by the average postmenopausal wo man FOSAMAX- is a bone metabolism regulator. FOSAMAXe is indicated for the treatment of osteoporosis in postmenopausal women .. Bone Mineral Density ... In clinical studies, over 96% of patients studied for up to three years had a measured increase in spine BMD_I t FOSAMAX* 10 mg daily produced statistically significant and clinically important increases in BMD
at the hip, spine. and wrist (ultradistal forearm) relative to placebo at three years (pSO.OOl )1.2
+Combined data from two large. identically designed. double-blind. placebo-controlled. three-year
multicenter studies in 994 women with osteoporosis. defined as low bone mass. 397 of whom received placebo and 196 of whom received FOSAMAX- 10 mg/day_ To ensure an adequate calcium intake. all patients were supplemented with 500 mg of calcium per day. I I. Liberman UA et at. Effect of oral alendronate on bone mineral density and the incidence of fractures in postmenopausal osteoporosis. N Engll Med 1995:333(221. 1437-43 2 Data on file. Merck Frosst canada Inc.: lWo double-blind. randomized. placebo-controlled. parallelgroup. multicenter studies to evaluate the safety and effect on bone density of daily oral MK-217 for two years in osteopenic postmenopausal women. with a one-year open treatment extension IProtocol No. 035 (US) and 037 (lnternational)l-Three Year Data
Builds bone to build independence --rTademark Merck & Co.. Inc .• Merck Frosst Canada Inc .. licensed user
alendronate sodium
, , , of VTE. Several studies were adapted 2 or initiated4,6,8 to investigate the issue. At first glance, the published results confinned the suspicions; the relative risks for VTE were, on average, 70 percent higher for users of OCs with new progestins than for users of OCs with older progestins and similar low EE doses. Although it appeared that physicians and their patients had good reason to be wary of third generation oral contraceptives, further infonnation emerging from these and other studies provides a more reassuring outlook.
There were 18 studies contributing 41 risk estimates for VTE to the analysis: 19 estimates of OP combined with EE dose ~50 Ilg, 14 for OP with EE dose <50 Ilg, and eight for NP with EE dose <50 Ilg.2-4.6,B,IO,14,20,2) The risk estimates from these studies for each category of OC are shown in Figure 1. All but one of the risk estimates were greater than unity with wide, overlapping confidence intervals. On average, the risk ofVTE for use ofOPs with EE dose 250 Ilg was 5.0-fold higher (95% CI 4.1 to 6.1) than for non-users. The contribution from the use ofNPs increased the risk 2.2-fold (95% CI 1.6 to 3.2). For both types of progestin, the risk was reduced by 30 percent (relative risk 0.7 [95% CI 0.5 to 0.9]) when combined with EE dose <50 Ilg . Thus, the calculated risk of VTE for users of OPs with <50 Ilg EE is 5.0 x 0.7 = 3.5-fold higher than the risk for non-users, and for NP users the risk is 5.0 x 2.2 x 0.7 = 7.7-fold higher (Table 1).
METHODS
To put the problem in perspective, the recent studies were added to past studies, obtained from a thorough search of the literature, which assessed the risk of VTE or AMI in OC users compared to non-users. Each study provided one or more relative risk estimates for OC users, usually adjusted for such co-variates as age, duration of use and smoking. Statistical techniques were used to combine the published risks, weighted according to their sample size and the precision of their risk estimates. Where adjusted estimates were unavailable, the adjusted relative risk was calculated from the raw odds ratio and a study-specific adjustment factor. Meta-regression analysis,19 using the logarithm of the weighted adjusted relative risks as the dependant variable, assessed the effects on the risk of VTE or AMI of estrogen dose <50 Ilg versus ~50 Ilg, the new third generation progestins versus first or second generation progestins (referred to as "other progestins" or "OPs"), and for AMI, the effect of current OC use versus past use.
2.
There were 15 studies contributing 25 risk estimates for AMI to the analysis: four estimates for past use of OPs combined with EE dose ~50 Ilg, seven for past use of OPs with EE dose <50 Ilg, six for current use of OPs with EE dose ~50 Ilg, seven for current use of OPs with EE dose <50 Ilg, and one for current use ofNPs with EE dose <50 Ilg. 5.10,15.l6.21,2B,}7 The risk estimates from these studies for each category of OC are shown in Figure 2. For past use, the risk estimates fall on both sides of unity and the confidence intervals overlap, indicating little or no increased risk on the basis of the limited data available. For current use, all risk estimates except one are greater than unity with wide overlapping confidence intervals.
RESULTS
1.
RISK ESTIMATES FOR ACUTE MYOCARDIAL
INFARCTION
RISK ESTIMATES FOR VENOUS THROMBO-
EMBOLISM
In four studies which estimated relative risk for the three new progestins, on average the risk of VTE was estimated to be 5.7 for gestodene (95% confidence interval [CI] 3.0 to 10.7),5.9 for desogestrel (95% CI 3.4 to 10.2), and 5.6 for norgestimate (95% CI 1.8 to 17.3) users, respectively, compared with non_users. 2,4,6,8 Because the estimates for the three new progestins were similar, they have been combined into the category "new progestins" or "NPs" for the remainder of the present analysis.
JOURNAL SOGe
TABLE 1 SUMMARY OF VTE AND AMI RISKS ASSOCIATED WITH ORAL CONTRACEPTIVE USE, ESTIMATED FROM META-REGRESSIONS Currency
Ethinyl Estradiol Dose
Progestin Type
Past Use
2: 50 1l9 <50 119 2: 50 1l9 <50 119 <50 119
Current Use
128
FEBRUARY 1997
VTE Relative Risk Estimate (compared to non-users)
AMI Relative Risk Estimate (compared to non-users)
OP
-
OP
-
1.2 1.3
OP
5.0 3.5 7.7
2.6 2.9 1.2
OP NP
FOSAMAX®
reduces the risk of fractures There was a trend towards a reduction in the proportion of patients treated with FOSAMAX® experiencing one or more nonvertebral fractures relative to those receiving placebo in pooled analysis (5-20 mg) (p=NS) '·tt
48%
reduction in the proportion of patients treated with FOSAMAX® experiencing one or more vertebral fractures relative to those treated with placebo in pooled analysis (5-20 mg) (p=0.03),·1
Low bone mass is a major pred ictor o f increased risk of osteoporotic fractu res' Vertebral fractures occurred in 62% (221355) of patients who received placebo and 3.2% (171526) of patients who received FOSAMAX. (5 or 10 mg for J years or 20 mg for 2 years followed by 5 mg for I yeaT),1 tt NOl1venebral fractures occurred in 9.6% (381397) of patients who received placebo and 7,5% (45/5971 of patients who received FOSAMAX- (5 or 10 mg for J years or 20 mg lor 2 years followed by 5 mg for I year) A trend towards a reduced number of lIol1vertebral fractures was observed in the patients treated with FOSAMAX·.I 1 Consensus Development Conference Diagnosis. prophylaxis. and treatment of osteoporosis.Am I Med 1993 ~94646·9
FOSAMAX., like other bisphosphonates. may cause local irritation of the upper gastrointestinal mucosa Esophageal adverse experiences. such as esophagitis. esophageal ulcers and esophageal erosions have been reported in patients receiving treatment with FOSAMAXe. In some cases these have been severe and required hospitalization. Healthcare professionals should therefore be alert to any signs or symptoms signaling a possible esophageal reaction and patients should be instructed to discontinue FOSAMAX. immediately and seek medical attention if they develop dysphagia, odynophagia or retrosternal pain FOSAMAXe is contraindicated in patients who have abnormalities of the esophagus which delay esophageal emptying such as stricture or achalasia, who are unable to stand or sit upright for at least 30 minutes. who are hypersensitive to any component of this product. who are hypocalcemic or who suffer from renal insufficiency ,creatinine clearance < 35 mUmin)
alendronate sodium
Builds bone to build independence . , MERCK FROSST
BEFORE PRESCRIBING, PLEASE CONSULT THE PRESCRIBING INFORMATION.
MERCK SHARP & DDHME CANADA DIV. OF MERCK FROSST CANADA INC. KIRKLAND. OUEBEC
FSM-96-CDN-9720b-JA
, , , FIGURE 1 14 12 10 OJ 8 ...., <0 6
··
''::;
i
(fJ
W .::.t
4
~ OJ
>
2
''::;
~ OJ
a::
1
·· •
·!
·· ·
(fJ
··•
· ·
·•
E
risk of AMI (95% CI 0.7 to 1.6). Thus, compared with non-users, the risk of AMI for current users of OPs with <50 ~g EE would be 1.2 x 2.2 x 1.1 = 2.9-fold higher and for current users ofNPs would be 1.2 x 2.2 x 0.4 xLI = l.2-fold higher (Table 1).
I
EFFECT OF NEW PROGESTINS ON RISK OF VTE AND AMI
·
OP
OP
NP
>=50 1i9 EE
<50 1i9 EE
<50 1i9 EE
This analysis has summarized the reported cardiovascular risks associated with the use of OCs. Taken at face value, the new progestins are associated with a two-fold increase in the risk for VTE and a potential 60 percent reduction in the risk for AMI, compared with use of first and second generation OCs. One additional study, which was not included in our analysis, compared second and third generation OC use directly, and also found a doubling in the risk for VTE with third generation progestins. 9 The question to be asked at this point is: are these effects real or simply the result of bias? Before discussing methodological bias, it would be helpful to assess whether there is a biological basis for the observations. As the analysis confirmed, VTE risk has long been associated with higher estrogen dosages. The new progestins, although combined with lower estrogen dosages, have been shown to be more estrogenic than second generation progestins in the totality of their effects on metabolic markers. Five studies indicated that sex hormone binding globulin (SHBG) levels (an estrogen-mediated effect) were higher for third generation versus second generation combinations.J8 -42 Five randomized clinical trials demonstrated that high density lipoprotein (HDL) cholesterol concentrations (a further estrogen-mediated effect} were higher after six months exposure to third generation compared with second generation OC formulations. J8.40•4l-45 Thus, there is biological plausibility for increased estrogenicity and an increase in the risk of VTE among third generation OC users. The observed reduction in OC-associated AMI risk would also be consistent with increased estrogen activity. Estrogen exposure is believed to reduce AMI rates among post-menopausal women, through mechanisms which may involve lipid optimization, platelet effects, and small vessel dilatation. 46.47 The welldocumented estrogenic effect on metabolic markers is, therefore, consistent with an estrogen effect on both clinical events, VTE and AMI.
Relative risk estimates from 18 publ ished studies of venous thromboembolism for users of oral contraceptives (OC) compared with non-users. The estimates are grouped according to the dose of ethinyl estradiol (EEl and type of progestin (other progestins, OP; new progestins, NPl.
On average, the risk of AMI for past use of OPs with EE dose ;:::50 ~g was 1.2-fold higher than for non-users (95% CI 0.8 to 1. 7). The risk associated with current use of such products was 2.2-fold higher (95% CI 1.5 to 3.3) than with past use. The AMI risk with use of NPs was reduced by 60 percent (relative risk 0.4 [95% CI 0.1 to 1.4]). The lower dose range of estrogen was associated with a non-significant 1. I-fold increase in the overall
FIGURE 2
· ·
10 8 OJ ....,
6
E
4
<0
''::;
·
(fJ
w
.::.t (fJ
~
·
2
· ··
OJ
> ''::; <0
Q)
a::
1
0.5
:
..
I
· ·
· I
·
op
op
op
op
>-so"g EE
>-so"g EE
+-- po.t U..
. NP
........---. Current Use - - - . . . .
Relative risk estimates from 15 published studies of acute myocardial infarction for users of oral contraceptives (OCl compared with non-users. The estimates are grouped according to the dose of ethinyl estradiol (EEl and type of progestin (other progestins, OP; new progestins, NPl and past or current use.
JOURNAL SOGe
130
FEBRUARY 1997
, , , On the other hand, a methodological bias may be partly responsible for the OC generation effect on VTE and AMI. An inherited mutation, the Factor V Leiden defect, present in about six percent of the European population,48 is known to increase the risk of VTE about eight-fold. s Oral contraceptive use may increase further the susceptibility to VTE in women with Factor V Leiden defect. 49 Other hereditary factors, which also may be pro-coagulant, are known to occur among young women, including protein C or S deficiency. If a woman is genetically susceptible to suffering a VTE event, it will probably occur in the first few months of OC use, after which she will discontinue use. Continuing users, therefore, have a lower risk ofVTE than new users. New drugs are more likely to be prescribed to new users; thus, the third generation OC group will include proportionally more new users than the second generation OC group. As a result, the duration of use will be longer, on average, in second generation users compared with third generation users. The presence of this bias required the authors of
two large studies to adjust some of their estimates for duration of use.3,4 It appears, therefore, that duration of use bias may account in part for the observed increase in VTE risk for third generation users. This same duration of use bias also may be affecting the results of the AMI analysis. Longer duration of OC use is associated with a higher risk of AMI,29 thus putting second generation OC users, with longer duration, at a disadvantage. With respect to AMI risk, a further caution is needed: only one published study estimated the effects of third generation OC use on risks of AMI, and the number of cases that had used third generation OCs was only six. 5 Ideally, larger numbers are preferable, but if third generation OC use truly reduces AMI risk, such cases will continue to be rare. Our estimate of AMI risk reduction with NPs is not Significant; however, it is consistent with the findings in a published study that did not meet our entry criteria.9 Although it should be noted that the recently published risks of VTE and AMI for third generation OC
On ce you pre crib e Tripha ii , yo u' ll rind th a t mo t p ati e nts
and re du ces eITors.'·' A
Passport for FIRST TIME
well , only Triphasil ha th e Pill I ass port
for J?irst Tim e U es, r a ERS
comprehen ive patie nt edu cation
program . Send he r hom e with a kit that in c lud e
rill e d compute r di s k, bookl e t
that te ll he r how to tak e Tl'ipha il and what
to ex pect durin g he r fir t ph ys ica l exam, and informati on about th Start He lp Lin e . Odd l! '
anada '
I'\'Jo
h e ca n lea rn all a bout Tl'ipha il in th e co mfort
are when yo u pre c rib e T'l'ipha ii , th ey le arn about a go od rh ythm. Presc ribe Tripha ii , first a nd la st. Pr e
c l'ib e d
Oral
Con lr acep Li
Please rcf('r to Ih(' product llIonograph for· delai ll'd
JOURNAL SOGC
131
a fa ct-
~ilr(;'t~
ilifonn
FEBRUARY 1997
e'
, , , TABLE 2 ANNUAL CARDIOVASCULAR EVENT RATES PER 100,000 WOMEN IN CANADA BY ORAL CONTRACEPTIVE USE AND AGE GROUP Age Range (years)
Venous Thromboembolism Rates Non-OC Users
OP Users
NP Users
Acute Myocardial Infarction Rates Non-OC Users
OP Users
NP Users
lS-19
2.3
7 .6
17.0
0.7
1 .7
0.7
20-24
2.7
8.9
20.0
0.8
2.1
0.9
25-29
3.2
10.8
24.2
1 .1
2.8
1.1
30-34
4.1
13.8
30.8
1 .5
4.1
1.7
35-39
5.7
18.9
42.4
2.8
7.4
3.0
40-44
9.0
30.2
67.7
15.0
40.0
16.3
users may be influenced in part by bias and should not be taken at face value, these studies provide the only information that is currently available on VTE and AMI risk with the new progestins. THE CANADIAN MODEL: VTE AND AMI RISKS APPLIED TO CANADIAN OC USERS
Although there is good reason to believe that both the hazards and the benefits may be exaggerated in the study results, it is these published results that form the basis of the calculated model to follow. To fit the current published results into a Canadian context, we estimated the number of OC users in Canada by age group, and determined the baseline VTE and AMI rates for young Canadian women who do not use 0Cs. From this information, plus the estimates of relative risk derived from the regression model, the number of events among OC users were calculated. We also estimated the mortality that might be associated with these events. Table 2 shows the baseline cardiovascular rates for six age ranges of women who do not use OCs. 50 To obtain the rates of VTE and AMI for users of OCs containing OP or NP, each age-specific baseline rate was multiplied by the appropriate relative risk calculated from the meta-regression results (Table 1). All relative risks used in these calculations apply to OCs with EE dose <50 f..lg as the lower estrogen doses are used most commonly today. Of course, the data in Table 2 relate to 100,000 OC users at each age range, but OC use changes with age as do the baseline risks for VTE and AMI. It follows that the next step is to estimate the effect on cardiovascular events within a population sample in which the number of women and the proportion of OC use in each age
JOURNAL SOGC
range are taken into account. Figure 3 shows these figures, estimated for 1995 based on Statistics Canada population data and reported OC use in Canada (1984 and 1993) and the UK (1993).6.51.52 Figure 4 shows the projected numbers of annual cardiovascular events for a representative sample of 100,000 Canadian OC users, for both OPs and NPs. To obtain these estimates, the proportion of total users in each age range (derived from the data shown in Figure 3) was multiplied by the appropriate event rate from Table 2. The total number of events per 100,000 OC users were obtained by summing over all age ranges. This was compared to the predicted number of events in a similarly age-distributed group of non-users (3.4 VTE events and 1.6 AMI events, open bars). Cardiovascular events in OC users in excess of the non-user rates are referred to as "attributable" events, because they are associated with the use of 0Cs. Table 3 shows that the attributable risk of OP use was eight VTE events and two AMI events per 100,000 users. The attributable risk of NP use was 22 VTE events and zero AMI events per 100,000 users. If 100,000 second generation OC users switched to third generation products, there would be 14 more attributable VTE events and two fewer attributable AMI events in the next year. To obtain the figures for the Canadian population, these numbers are multiplied by 12, because there are an estimated 1.2 million OC users in Canada.
FIGURE 3 14
6' 0 0
12
0
10
c5
~ c
II>
E
0
~
8 6
'+-
0
Q;
4
-D
E 2 ~
Z
-
-
-
15-19
20-24
25-29
'--------
30-34
~
35-39
40-44
Age g roup (years) Distribution of the Canadian female population of reproductive age by age group (hatched bars) and the number of OC users (solid bars) in each group, estimated for 1995.
132
FEBRUARY 1997
, , , TABLE 3
annual numbers of deaths in the 1.2 million Canadian OC user population are less than one due to VTE and five due to AMI. Annually in the Canadian population, the predicted additional number of deaths attributable to OC use would be one from VTE and nine from AMI if all women used second generation products, or three from VTE and zero from AMI if all women used third generation products (Table 3). There are zero AMI deaths attributable to third generation OC use because the AMI event rate is the same for users and non-users (Figure 4). If all Canadian women were using second generation OCs and switched to third generation products, there would be two more VTE deaths and nine fewer AMI deaths in Canada annually, for a net reduction of seven in the annual number of deaths among OC users.
PREDICTED ANNUAL ATTRIBUTABLE CARDIOVASCULAR EVENTS AND DEATHS IN ORAL CONTRACEPTIVE USERS IN CANADA OP Users
NP Users
VTE
AMI
Total
VTE
AMI
Total
Attributable events/100,000*
8
2
10
22
0
22
Attributable deaths among 1.2 million OC users
1
9
10
3
0
3
* a representative sample across age groups of OC users
It is also important to take into account the case mortality ratio associated with cardiovascular events, which is much higherfor AMI. To examine the cardiovascular mortality associated with oral contraceptive use, we adopted published estimates that approximately 1.25 percent of VTE events S3 •S4 and 30 percent of AMI events SS will result in death. The resulting mortality rates are shown graphically in Figure 4 as a solid bar at the base of each corresponding event bar. The baseline expected
IMPLICATIONS FOR CANADIAN OC USERS AND PRESCRIBERS
In addition to the obvious benefit of contraceptive efficacy, there are many other direct and indirect benefits
SOMETIMES, PULLING OUT
DOESN'T WORK.
If you ' re thinkin g o f pulling out of th e tri ed and true oral contra ceptive think again. With over 34 million cycle
of u e in Canada,' Tripha iI offers exce ll ent cycle control
with a low in cid e nce of ide e ffect .... A
Pass'port for FIR T T IME
for Fir t Time ER
in favour of a n ew one,
well , onl y Tripha iI ha
the Pill Pa port
e r , acomprehen ive pati ent education
program . end her hom e with a kit that include
a fact-filled
compute r di k, bookle t that tell he r how to take Tripha il and what to expect durin g her fir t physical exam, and information about the 1-8 88 Ju Line. h e can learn all about Tripha il in th Cycle tte packa ging help
comfort of her hom .
mak e compli a nce easy and redu ce
new OC offer all of that? Prescribe Tripha ii, first and last. Ca n a d a ' Mos t
Pr es
rib e d
Or a l
Co nt race pti ve'
Pl ease re fer 10 th e product rnonop-a ph for detailed sa rety 10 fonn8110n.
JOURNAL SOGe
133
FEBRUARY 1997
error .
, , , FIGURE 4
ARE THE VTE RISKS SIMILAR FOR ALL THIRD GENERATION PRODUCTS?
25
(/)
All three of the third generation progestins (desogestrel, gestodene, and norgestimate) recently introduced in Europe were associated with a two-fold increase in VTE risk, compared to second generation progestins, usually levonorgestreL
20
+"
C
Q)
>
UJ
'+-
15
0
....Q)
.D
E
10
::J
Z
SHOULD WOMEN SWITCH OR DISCONTINUE
5
0
ORAL CONTRACEPTIVES?
VTE
Women at high risk for either VTE or AMI should use non-hormonal methods of contraception. For the typical oral contraceptive user, however, cardiovascular morbidity and mortality rates are so low that they should not discourage OC use or influence the choice of one product over another.
AMI
Predicted annual number of cardiovascular events (venous thromboembolism [VTE] and acute myocardial infarction [AMI]) per 100,000 women in Canada, of an age-distribution similar to that of OC users. Number of events are shown for non-users (open bars), users of OCs with other progestins (diagonally hatched bars), and users of OCs with new progestins (crosshatched bars). The predicted annual numbers of deaths from these causes in each group are shown by the solid bars.
Is THE POPULATION EFFECT A CANADIAN PUBLIC HEALTH CONCERN?
Cardiovascular events in young women are rare, even considering the additional events due to OC use. If the published increased risks due to OC use were accepted at face value, the difference between second and third generation products would do little more than shift the health care burden between VTE and AMI. None of the studies published during the last year was conducted in North America, where there may be as many as 15 million ex::: users, making it necessary to generalize from the results of studies in other countries. It is premature, therefore, to regard as proven any excess risk ofYTE with third generation products or any excess risk of AMI with second generation products.
from oral contraceptive use, including the reduction in the incidence of ovarian and endometrial cancers. l Having noted this, how important for ex::: users and prescribers are the cardiovascular risks described in this paper? DOES ORAL CONTRACEPTIVE USE CARRY AN INCREASED RISK OF CARDIOVASCULAR DISEASE?
This analysis has shown that the use of oral contraceptives is associated with 10 to 22 additional cardiovascular events per 100,000 users annually, depending on whether OCs with second or third generation progest ins are used_ The risk of VTE is increased four to eight-fold and the risk of AMI is increased up to threefold during current OC use. Past OC use does not have a significant effect on cardiovascular health.
REFERENCES
1. Is THE VTE RISK DOUBLED FOR THIRD GENERATION OC USERS?
2.
There is a biological rationale to explain partially the increased VTE risk with third generation progestins (increased estrogenicity), yet there are also sound epidemiological reasons to be sceptical about the magnitude of this risk. Because the currently available published risks may be susceptible to bias and should not be taken at face value, we need to await the results of future welldesigned studies to judge more accurately whether there is truly an increased risk.
JOURNAL saGC
3.
134
Schlesselman JJ. Net effect of oral contraceptive use on the risk of cancer in women in the United States. Obstet GynecoI1995;85:793-801. Meirik O. World Health Organization Collaborative Study of Cardiovascular Disease and Steroid Hormone Contraception. Effect of different progestogens in low oestrogen oral contraceptives on venous thromboembolic disease. Lancet 1995;346:1582-8. Poulter NR. World Health Organization Collaborative Study of Cardiovascular Disease and Steroid Hormone Contraception. Venous thromboembolic disease and combined oral contraceptives: results of international multicentre case-control study. Lancet 1995;346: 1575-82.
FEBRUARY 1997
, , , 10. Inman W, Vessey MP. Investigation of deaths from pulmonary, coronary, and cerebral thrombosis and embolism in women of child-bearing age. Br Med J 1968;2:193-9. 11. Vessey MP, Doll R. Investigation of relation between use of oral contraceptives and thromboembolic disease. A further report. Br Med J 1969;2:651-7. 12. Sartwell PE, Masi AT, Arthes FG, Greene GR, Smith HE. Thromboembolism and oral contraceptives: an epidemiologic case-control study. Am J EpidemioI1969;90:365-80. 13. Boston Collaborative Drug Surveillance Programme. Oral contraceptives and venous thromboembolic disease, surgically confirmed gallbladder disease, and breast tumours. Lancet 1973;1 :1399-404. 14. Royal College of General Practitioners' Oral Contraception Study. Oral contraceptives, venous thrombosis, and varicose veins. Journal of the Royal College of General Practitioners 1978;28:393-9. 15. Mann J, Inman W. Oral contraceptives and death from myocardial infarction. Br Med J 1975;2:245-8. 16. Mann J, Vessey M, Thorogood M, Doll R. Myocardial infarction in young women with special reference to oral contraceptive practice. Br Med J 1975;2:241-5.
4. Spitzer WO, Lewis MA, Heinemann lAJ, Thorogood M, MacRae KD. Third generation oral contraceptives and risk of venous thromboembolic disorders: an international case-control study. Br Med J 1996;312:83-8. 5. Lewis MA, Spitzer WO, Heinemann lAJ, MacRae KD, Thorogood M. Third generation oral contraceptives and risk of myocardial infarction: an international case-control study. Br Med J 1996;312:88-90. 6. Farmer RDT, Preston TD. The risk of venous thromboembolism associated with low oestrogen oral contraceptives. J Obstet GynaecoI1995;15:195-200. 7. Farmer RDT. Oral contraceptives and thromboembolism. Lancet 1996;347:259. 8. Bloemenkamp KWM, Rosendaal FR, Helmerhorst FM, Buller HR, Vandenbroucke JP. Enhancement by factor V Leiden mutation of risk of deep-vein thrombosis associated with oral contraceptives containing a third-generation progestagen. Lancet 1995;346:1593-6. 9. Jick H, Jick SS, Gurewich V, Myers MW, Vasilakis C. Risk of idiopathic cardiovascular death and non-fatal venous thromboembolism in young women using oral contraceptives with differing progestagen components. Lancet 1995;346: 1589-93.
WE DON'T LET
COITUS
INTERRUPT US.
Your pa ti e nl wanl a co nlra ce ptive lhal won 'l inle rrupl lh e ir lives. \' ilh l'ripha s il , yo u ca n feel
onfidenl
lhal you're prescribin g a lri ed and lrue oral co nlrace plive. Th e mo l presc rib e d in Canada. ' Tt'iphasil ha s a low in c id e nce of id e e ffecl and our uniqu e
I
Pass'port
ror FIRST TIME
yc le lle pac ka gin g help m a ke com pli ance
ea y and r edu ces e rrors .'" As \ ell , only Triphasil has lh e ER
Pill Passporl fo r Fi r l Tim e
e r , aco mpre he ns ive pali e nl
e du calion prog ram . Send he r hom e wilh a killhal includ e
a fa cl-fill e d co mpule r
di s k, bookl e ls lhalle ll h e r how lo lake Tripha il and whallo ex pecl d phy ica l xam , and information about th e 1-888 Jump
larl lI e lp Line.
all aboul Tripha il in th e co mforl of h e r hom e. This ma y impro ve pati wilh fe we r inle rrupti on s. Presc rib e Tt'iph as il , fir sl and la l. a n ad a'
Mo
P,' e
c rib ed Pl ca ~ c
Oral
Co nl racept i ve '
re fe r (0 th e product m onograph for del ailcd safcl~ infonnalion.
JOURNAL SOGe
135
FEBRUARY 1997
, , , 17. Darney PD. The androgenicity of progestins. Am J Med 1995;98: 1045-11 OS. 18. Jones EE. Androgenic effects of oral contraceptives: implications for patient compliance. Am J Med 1995;98:1165-1195. 19. Greenland S. Quantitative methods in the review of epidemiologic literature. Epidemiol Rev 1987;9:1-30. 20. Ludwig H. Anovulatory agents and venous disease. Ergebnisse der Angologie und Phlebologie 1970;4: 81-102. 21. Fuertes-de la Haba A, Curet JO, Pelegrina I, Bangdiwala I. Thrombophlebitis among oral and normal contraceptive users. Obstet Gynecol 1971 ;38:259-63. 22. Maguire MG, Tonascia J, Sartwell PE, Stolley PD, Tockman MS. Increased risk of thrombosis due to oral contraceptives: a further report. Am J Epidemiol 1979; 110: 188-95. 23. Petitti DB, Wingerd J, Pellegrin F, Ramcharan S. Risk of vascular disease in women. Smoking, oral contraceptives, noncontraceptive estrogens, and other factors. ·JAMA 1979;242: 1150-4. 24. Porter JB, Hunter JR, Danielson DA, Jick H, Stergachis A. Oral contraceptives and nonfatal vascular diseaserecent experience. Obstet Gynecol 1982;59:299-302. 25. Porter J, Hunter J, Jick H, Stergachis A. Oral contraceptives and nonfatal vascular disease. Obstet Gynecol 1985;66: 1-4. 26. Helmrich SP, Rosenberg L, Kaufman DW, Strom B, Shapiro S. Venous thromboembolism in relation to oral contraceptive use. Obstet GynecoI1987;69:91-5. 27. Vessey MP. Epidemiologic studies of oral contraception. Int J Fertil 1989;34:64-70. 28. Jick H, Dinan BJ, Rothman KJ. Oral contraceptives and nonfatal myocardial infarction. JAMA 1978;239: 1403-6. 29. Rosenberg L, Hennekens CH, Rosner B, Belanger C, Rothman KJ, Speizer FE. Oral contraceptive use in relation to non-fatal myocardial infarction. Am J Epidemiol 1980; 111 :59-66. 30. Krueger DE, Ellenberg 5, Bloom 5, Calkins BM, Maliza C, Nolan DC, et a/. Fatal myocardial infarction and the role of oral contraceptives. Am J Epidemiol 1980;111 :655-74. 31. Adam SA, Thorogood M, Mann JI. Oral contraception and myocardial infarction revisited: the effects of new preparations and prescribing patterns. Br J Obstet Gynaecol 1981 ;88:838-45. 32. Slone D, Shapiro 5, Kaufman D, Rosenberg L, Miettinen 0, Stolley P. Risk of myocardial infarction in relation to current and discontinued use of oral contraceptives. N Engl J Med 1981;305:420-4. 33. La Vecchia C, Franceschi 5, Decarli A, Pampa Ilona 5, Tognoni G. Risk factors for myocardial infarction in young women. Am J EpidemioI1987;125:832-43. 34. Ananijevic-Pandey J, Vlajinac V. Myocardial infarction in young women with reference to oral contraceptive use. Int J Epidemiol 1989; 18:585-8.
JOURNAL SOGe
35. Croft P, Hannaford Pc. Risk factors for acute myocardial infarction in women: evidence from the Royal College of General Practitioners' Oral Contraception Study. Br Med J 1989;298:165-8. 36. Rosenberg L, Palmer JR, Lesko SM, Shapiro S. Oral contraceptive use and the risk of myocardial infarction. Am J EpidemioI1990;131:1009-16. 37. Thorogood M, Mann J, Murphy M, Vessey M. Is oral contraceptive use still associated with an increased risk of fatal myocardial infarction? Report of a case-control study. Br J Obstet Gynaecol 1991 ;98: 1245-53. 38. Ball MJ, Ashwell E, Jackson M, Cillmer MDG. Comparison of two triphasic contraceptives with different progestogens: effects on metabolism and coagulation proteins. Contraception 1990;41 :363-76. 39. Becker H. Supportive European data on a new oral contraceptive containing norgestimate. Acta Obstet Gynecol Scand 1990 (Suppl);152:33-9. 40. Chapdelaine A, Desmarais J-L, Derman RJ. Clinical evidence of the minimal androgenic activity of norgestimate.lntJ FertiI1989;34:347-52. 41. Kloosterboer HJ, van Wayjen RGA, van den Ende A. Effects of three low-dose oral contraceptive combinations on sex hormone binding globulin, corticosteroid binding globulin and antithrombin III activity in healthy women: two monophasic desogestrel combinations (containing 0.020 or 0.030 mg ethynlestradiol) and one triphasic levonorgestrel combination. Acta Obstet Gynecol Scand 1987 (Suppl);144:41-4. 42. Refn H, Kjaer A, Lebech A-M, Borggaard B, Schierup L, Bremmelgaard A. Metabolic changes during treatment with two different progestogens. Am J Obstet Gynecol 1990;163:374-7. 43. Harvengt C, Desager JP, Gaspard U, Lepot M. Changes in lipoprotein composition in women receiving two low-dose oral contraceptives containing ethinyl estradiol and gonane progestins. Contraception 1988;37:565-75. 44. Ho PC, Liu WT, Kwan MSW. Serum lipids in Chinese patients using oral contraceptive pills. Contraception 1990;41 :55-61. 45. Kjaer A, Lebech A-M, Borggaard B, Refn H, Pedersen LR, Schierup L, et al. Lipid metabolism and coagulation of two contraceptives: correlation to serum concentrations of levonorgestrel and gestodene. Contraception 1989;40:665-73. 46. The Writing Group for the PEP I Trial. Effects of estrogen or estrogen/progestin regimens on heart disease risk factors in postmenopausal women: the postmenopausal estrogen/progestin interventions (PEPI) trial. JAMA 1995;273:199-208. 47. Chester AH, Jiang C, Borland JA, Yacoub MH, Collins P. Oestrogen relaxes human epicardial coronary arteries through non-endothelium-dependent mechanisms. Coron Artery Dis 1995;6:417-22.
136
FEBRUARY 1997
, , , 48. Dahlback B. Physiological anticoagulation: resistance to activated Protein C and venous thromboembolism. J Clin Invest 1994;94:923-7. 49. Henkens CMA, Born VJJ, Seinen AJ, van der Meer J. Sensitivity to activated Protein C: influence of oral contraceptives and sex. Thromb Haemost 1995;73:402-4. 50. Lis Y, Spitzer WO, Mann RD, Cockburn I, Chukwujindu J, Thorogood M, et al. A concurrent cohort study of oral contraceptive users from the VAMP research bank. Pharmacoepidemiol Drug Safety 1993;2:51-63. 51. Balakrishnan TR, Krotki K, Lapierre-Adamcyk E. Contraceptive use in Canada, 1984. Fam Plann Perspect 1985;17:209-15. 52. Boroditsky R, Fisher W, Sand M. The Canadian Contraception Study. J Soc Obstet Gynaecol Can 1995; 17(Suppl): 1-28. 53. Creager MA, Dzau VJ. Vascular diseases of the extremities. In: Isselbacher KJ, Braunwald E, Wilson JD, Martin JB, Fauci AS, Kasper DL (Eds). Harrison's Principles of Internal Medicine. 13th ed. New York: McGraw-HiII,lnc. 1994:1135-41. 54. Moser KM. Pulmonary thromboembolism. In: Isselbacher KJ, Braunwald E, Wilson JD, Martin JB, Fauci AS, Kasper DL (Eds). Harrison's Principles of Internal Medicine. 13th ed. New York: McGraw-HiII,lnc. 1994:1214-25. 55. Pasternak RC, Braunwald E. Acute myocardial infarction. In: Isselbacher KJ, Braunwald E, Wilson JD, Martin JB, Fauci AS, Kasper DL (Eds). Harrison's Principles of Internal Medicine. 13th ed. New York: McGraw-HiII,lnc. 1994:1066-97.
JOURNAL SOGC
137
FEBRUARY 1997