Oral Contraceptives and CVD Epidemiologic Effects

1
Oral Contraceptives and CVDEpidemiologic Effects

• TMM Farley
• Department of Reproductive Health and Research
• World Health Organization
• Geneva

2
Overview

• Rationale for WHO Collaborative Study of
  Cardiovascular Disease and Steroid Hormone
  Contraception
• Venous thromboembolism, stroke and myocardial
  infarction
• Supplement with data from other recent studies
• Overall cardiovascular risk

3
WHO Study of Cardiovascular Disease and Steroid
Hormone Contraception - 1986

• OC composition and patterns of use have changed
  since late 1970s
• Most information refers to older products
• No information available on risks in women from
  developing countries
• What are the cardiovascular risks associated
  with modern OCs used in modern ways?

4
WHO Study of Cardiovascular Disease and Hormonal
Contraception
5
WHO CVD Study - Design

• Hospital-based case-control study
• 17 countries (12 developing, 5 in Europe)
• Conducted February 1989 - January 1993
• First time cases of stroke, AMI or idiopathic VTE
  in women aged 20-44 years
• 3 controls per case, matched on age, hospital and
  time period
• 2,242 stroke, 368 AMI, 1,143 VTE, 10,025 controls

6
Venous Thromboembolism
Adjusted odds ratio (95 CI)
WHO, Lancet 1995 3461575
7
Risk Factors for Idiopathic VTE

• Increased relative risk with
• OC use
• elevated body mass index
• hypertension in pregnancy
• No effect on relative risk
• smoking, age, hypertension
• duration of OC use
• previous OC use

WHO, Lancet 1995 3461575
8
VTE and Low Estrogen OCs
Adjusted odds ratios (95 CI)
WHO, Lancet 1995 3461575
9
VTE and Low Estrogen OCs

• Excess risk with desogestrel gestodene compared
  with levonorgestrel
• About 2.5 ? higher risk
• Similar excess risk for the two products
• Bias or confounding unlikely explanation
• Unexpected, Unexplained, Unwelcome, Uncomfortable
• Must be confirmed by independent research

10
3rd Gen vs. Levonorgestrel
Overall risk ratio (95 CI)
1.9 (1.5, 2.2)
Adjusted risks relative to non-users (crude risk)
Lancet 1995 3461582, 1589, 1593 BMJ 1996 312
83 Contraception 1998 57 291
11
Ischaemic Stroke

• Major risk factors
• smoking
• hypertension
• rheumatic heart disease
• diabetes
• Overall risk with OC use 2.9 (2.2 - 3.9)
• No effect of past OC use or duration of use

12
Ischaemic Stroke - smoking
Pooled adjusted odds ratio (95 CI) controls
WHO, Lancet 1996 348 498
13
Ischaemic Stroke - Hypertension
Pooled adjusted odds ratio (95 CI) controls
WHO, Lancet 1996 348 498
14
Ischaemic Stroke and Low Estrogen OCs

• Kaiser, CA 1.2 (0.5, 2.6)
• Washington State 1.4 (0.5, 3.8)
• Denmark 1.6 (1.1, 2.4)
• WHO Europe 1.4 (0.6, 3.1) Developing
  countries 3.4 (2.2, 3.1)
• TransNational 2.8 (2.0, 3.8)

RR (95 CI) compared with non-users
15
Ischaemic Stroke and Low Estrogen OCs

• With Blood Pressure check WHO Europe 1.3 (0.5,
  3.5) WHO Developing 2.1 (1.1, 3.8)
  TransNational 2.1 (1.4, 3.1)
• Without Blood Pressure check WHO Europe 1.5
  (0.5, 4.6) WHO Developing 5.2 (2.9, 9.1)
  TransNational 4.5 (2.6, 8.0)

RR (95 CI) compared with non-users
16
Ischaemic stroke - Conclusion

• Some excess risk associated with low estrogen
  dose OCs
• Smoking and hypertension potentiate OC-associated
  risk
• Lower risk when screened for hypertension
• No evidence of difference in risk according to OC
  type (2nd vs. 3rd generation)

17
Haemorrhagic Stroke

• No difference according to BP checking
• No impact on risk in women lt 35 years
• About 2 ? risk in women over 35 years
• Higher (relative) risk among older women,
  smokers, women with hx of hypertension
• Smoking has greater impact on risk of
  haemorrhagic than ischaemic stroke
• Consistent with data from Kaiser, CA (Petitti,
  1996)

18
Myocardial Infarction

• Major risk factors
• smoking, hypertension, rheumatic heart disease,
  diabetes, hyperlipidaemia
• Overall risk with OC use 4.9 (3.1 - 7.8)
• No effect of past OC use or duration of use
• Lower risks with low compared with high dose OCs

19
AMI - Hypertension OC use
Pooled adjusted odds ratio (95 CI) controls
WHO, Lancet 1997 349 1202
20
AMI - smoking OC use
Pooled adjusted odds ratio (95 CI) controls
WHO, Lancet 1997 349 1202
21
AMI - Conclusion

• Majority of cases (78) occur in smokers
• Lower risk with low dose OCs, in women without CV
  risk factors and who reported BP check (similar
  observations in TransNational study)
• Among women with no cardiovascular risk factors
  who do not smoke, RR 1.1 in women with BP check

22
AMI and Type of OC
Lancet 1997 349 1202 Contraception 1997 56
129 BMJ 1999 318 1579 NEJM 2001 345 1787
23
Overall Cardiovascular Risk

• Different risk factors for VTE and stroke or MI
• Age distribution of VTE, stroke and MI cases very
  different over 15 - 44 age range
• Any reduction in MI risk for third generation OC
  users more important for older women and smokers

24
Observed CVD Incidence Oxford
VTE
Incidence (per 100 000 wyrs)
Haemorrhagic stroke
AMI
Ischaemic stroke
Age group (years)
J Epidemiol Comm Health 1998 52 775
25
CVD Incidence - Non-smoker
Non-OC user
OC user
Events per 106 wyrs
26
CVD Mortality - Non-smoker
Non-OC user
OC user
Deaths per 106 wyrs
27
OCs and CVD

• OCs most widely studied pharmacologic agent
• In young women without cardiovascular risk
  factors, OCs are safe
• Excess risk seen in older women, smokers and
  those with pre-existing risk factors
• Risk-benefit of 2nd vs. 3rd generation OCs
• VTE risk more important in younger women
• MI risk more important in older women and smokers