Venous thromboembolism: small increased risk for women taking oral contraceptives.

Clinical bottom line (level 3b)

  1. The risk of venous thromboembolism during pregnancy was higher than the risk on any pill.
  2. The risk of venous thromboembolism in healthy women < 40 was very small (0.003%).
  3. Oral contraceptives increased the risk slightly (NNH = 5700 at 4 years) .
  4. Third generation contraceptives were slightly more risky than second generation (NNH = 7000 at 4 years) . This was due to the progestogen component.
Jick et al: Lancet 1995; 346: 1589-1593
Expires September 2003

The study

Case-control study with objective outcomes, adjusted for confounding factors, not validated in an independent set of patients.

Setting: 365 GP practices, UK

238130 patients (aged ?, 100% female) healthy women who received oral contraceptives containing < 35 mg ethynylestradiol with either levonorgestrel, desogestrel, or gestodene from 1991-94.

Excluded if
  • history of PE or DVT
  • acute myocardial infarction
  • stroke
  • cancer
  • epilepsy
  • diabetes
  • hypertension
  • hyperlipidaemia
  • cystic fibrosis
  • >40 years


    • Cases: 107180 patients (% female, mean age ): deep vein thrombosis or pulmonary embolism
    Controls: 130590 patients (% female, mean age ): matched for age and current use of a study oral contraceptive


    Odds ratios were adjusted for age and calender year.

    Outcomes studied:
  • DVT or PE PE or DVT; diagnosed by computer records and discharge letters. Events defined as probable if no definitive study (v/q scan, or angiogram for PE, venogram, doppler or ultrasound for DVT) reported, or definite if definitive study reported.
  • probable PE/DVT
  • definite PE/DVT

    • Women were classified into users and past users of each type of OC. Smoking and body mass index controlled in analysis.
    • 100% were followed for four years.

    The evidence

    Patient expected event rate for DVT or PE: 0.0038%
    Patient expected event rate for probable PE/DVT: 0.033%
    risk factor for
    probable PE/DVT     		adjusted OR
    (95% CI)    		 NNH
    (95% CI)
    levonorgestrel 1.0
    ( to )
    desogestrel 2.2
    (1.1 to 4.4)     		5200
    (6200 to 12000)
    gestodene 2.1
    (1.0 to 4.4)     		5700
    (1000 to infinity)

    Patient expected event rate for definite PE/DVT: 0.021%
    risk factor for
    definite PE/DVT     		adjusted OR
    (95% CI)    		 NNH
    (95% CI)
    levonorgestrel 1.0
    ( to )
    desogestrel or gestodene 2.2
    (1.0 to 4.7)     		7000
    (2300 to infinity)

    Comments

    1. There were fewer smokers in the control group (risk factor for VTE - i.e. slight overestimation of oral contraceptive harm).
    2. Risk of PE/ DVT for pregnancy and the post-partum period is approximately 60/100 000 woman-years (NNF = 1700 per year) i.e. still higher than third generation contraceptives).

    Citation

    1. Jick H, Jick SS, Gurewich V, et al: Risk of idiopathic cardiovascular death and nonfatal venous thromboembolism in women using oral contraceptives with differing progestogen components. Lancet 1995; 346: 1589-1593
    Search Terms:
    Contributor: Chris Ball and Musab Hayatli, Unknown Month 2000
    Reviewer: Daniel Sontheimer

    Clinical Question.
    Patient healthy women
    Intervention or Exposure pregnant or on contraceptive pill
    Outcome venous thromboembolism