- Where did the story come from?
- What are contraceptive pills and what is venous thromboembolism?
- What were the results of the case-control study?
- What were the results of the cohort study?
- How did the clinical review put these studies into context?
- What interpretations did the editorial draw from these results?
- What does the NHS Knowledge Service make of this study?
"Women are not routinely using the safest brand of contraceptive pill,” reported The Guardian . The newspaper said that all types of combined contraceptive pill carry a risk of blood clots, but some have a higher risk than others. It said that researchers have found that the safest pills had combined low-dose oestrogen and levonorgestrel.
As the researchers say, all types of combined oral contraceptives carry a small increased risk of clots. This risk is very low and fewer than about one in 1,000 users will be affected. For every 100,000 women who take the pill over a year, there is an absolute risk that 15-25 of them having a clot, compared to five in every 100,000 women who are not on the pill.
This risk can be mitigated by the type of pill that women take and some pills are safer than others. However, there could be good reasons why some women have been put on ‘riskier’ pills, and so advice from a healthcare practitioner should be sought before changing.
Where did the story come from?
All four papers are published in the peer-reviewed British Medical Journal (BMJ):
- The MEGA case control study was carried out by Dr A van Hylckama Vlieg and colleagues from the Leiden University Medical Center in the Netherlands.
- The national cohort study in Denmark was conducted by Professor Øjvind Lidegaard and colleagues from the Gynaecological Clinic, Rigshospitalet, Copenhagen University.
- The clinical review was written by Dr Jean-Jacques Amy from Belgium as editor in chief of the European Journal of Contraception and Reproductive Health Care, and Vrijesh Tripathi, a lecturer at the University of West Indies.
- The editorial was written by Dr Nick Dunn, a senior lecturer in medical education at the University of Southampton Medical School.
What are contraceptive pills and what is venous thromboembolism?
There are several types, brands and generations of oral contraceptive pill and 26 varieties are listed in the British National Formulary. They differ in which hormones they contain and the exact forms of the hormones used. Some contain low-strength oestrogen (20 micrograms) combined with synthetic progestogen such as norethisterone, desogestrel, drospirenone or gestodene. Others contain more oestrogen (30 or 35 micrograms) combined with the above or levonorgestrel or norgestimate ((two other types of synthetic progesterone).
Since 1961, several large studies have shown a two- to sixfold increased risk of deep venous thrombosis associated with oral contraceptive use. This increased risk was assumed to be related to the oestrogen content of the pills. As a result, the oestrogen dose in combined oral contraceptives has been reduced. However, there was still uncertainty as to which of the different types of hormonal contraceptives is safest with regard to the risk of venous thrombosis. The current studies provide evidence about this question.
Venous thromboembolism is one of the most serious side effects of taking the oral contraceptive pill and occurs when a blood clot forms in a vein, usually one of the deep veins in the leg. Although rare, it is possible, unless treated with anticoagulation, for the clot to travel through the veins, lodge in the lungs and cause more serious complications (pulmonary embolism).
What were the results of the case-control study?
This study focused on the dose of oestrogen and the type of progestogen in oral contraceptives available in the Netherlands. The researchers used data from the MEGA study (multiple environmental and genetic assessment of risk factors for venous thrombosis-study). This was a large, population-based, case-control study on risk factors for venous thrombosis that ran between March 1999 and September 2004. The researchers identified 1,524 women from six participating anticoagulation clinics in the Netherlands who had had a venous thromboembolism in the leg. These women had not yet had the menopause and were all less than 50 years old. They were also not pregnant or within four weeks of having had a baby and were not using a hormone-excreting intrauterine device (IUD) or long-acting injection form of contraceptive. These women were matched to 1,760 controls who were similar but had not had a clot.
The researchers then calculated the risk of venous thrombosis for each type of pill compared to women not on the pill, women on hormone-excreting IUDs and those on the long-acting injection form of contraceptive.
They found that overall, taking oral contraceptive pills produced a fivefold increase in risk compared with non-use (odds ratio 5.0, 95% confidence interval 4.2 to 5.8), with the exact level of risk varying by type of progestogen and dose of oestrogen.
Compared to non-use, taking oral contraceptives increased risk of venous thrombosis by:
- 3.6-fold for levonorgestrel containing pills,
- 5.6-fold for gestodene containing pills,
- 7.3-fold for desogestrel containing pills,
- 6.8-fold for cyproterone acetate containing pills, and
- 6.3-fold for drospirenone containing pills.
The risk of venous thrombosis increased with increasing oestrogen dose. The risk of a venous thrombosis was highest in the first months of oral contraceptive use irrespective of the type of oral contraceptive.
What were the results of the cohort study?
Danish women aged 15-49 with no history of cardiovascular or malignant disease were recruited to this study. It linked the data of all women in a National Registry of Patients collected from all Danish hospitals since 1977 with data on their oral contraceptive use from a national registry of prescriptions. A total of 10.4million ‘woman years’ of data were recorded. A ‘woman year’ is a statistical concept of one woman’s data collected for a year. In this concept, five women followed for one year contribute the same amount of data to this study as one woman followed for five years.
The analysis included 3.4million woman years of current use of oral contraceptives, 2.3million woman years of former use, 4.8million woman years of never use, giving a total of about 10.4million woman years of observation.
A total of 4,213 first-time venous thrombotic events were recorded during the study period. Of these, 2,045 were current users of hormonal contraception. The venous thrombotic events included deep vein leg thrombosis (61.8%), pulmonary embolism (26.2%), femoral vein thrombosis (4.7%), portal thrombosis (1.2%), caval or renal thrombosis (0.8%) and unspecified deep vein thrombosis (5.4%).
Following analysis, the authors conclude, “the risk of venous thrombosis in current users of combined oral contraceptives decreases with duration of use and decreasing oestrogen dose”.
They also found that in women who had been taking oral contraceptives for the same length of time, and which contained the same dose of oestrogen, those whose pills contained desogestrel, gestodene, or drospirenone had a significantly higher risk of venous thrombosis than those whose pills contained levonorgestrel.
How did the clinical review put these studies into context?
The review goes through the clinical process of discussing contraception in a consultation and describes how each of the possible contraceptive methods work. The reviewers recommend oral contraceptives containing levonorgestrel or norethisterone, with as low a dose of oestrogen as possible. They say that all of the more recent progestogens, with the possible exception of norgestimate, seem to be at a disadvantage with regard to venous thromboembolism.
The reviewers also clarify that the absolute risk of having venous thromboembolism is low. For women not on the pill, the risk of having a clot is about five per 100,000 women, over a year. This is compared to about 15-25 per 100,000 women who take the pill, over a year.
What interpretations did the editorial draw from these results?
The author of the editorial discusses the strengths and weaknesses of these studies. He explains that all oral contraceptives are effective in preventing pregnancy if they are taken correctly, so the choice of which one to use rests on the profile of side effects. He says the chance of developing a venous thromboembolism is low enough to consider a range of contraceptive pills when dealing with individual patients.
The author suggests that, for some individuals, a pill containing a newer progestogen type or one with a higher dose of oestrogen may still be appropriate, but that patients with a personal or family history of venous thromboembolism should not take combined oral contraceptives at all.
What does the NHS Knowledge Service make of this study?
The primary research studies provide reliable estimates of the risk of developing venous thromboembolism in women taking a variety of contraceptive pills, and have been interpreted by the clinical reviewers with care. The authors do mention some limitations to relying on observational studies such as these:
- The Danish cohort study, for example, was carried out by merging national databases. As such, the authors could not control for a family history of venous thromboembolic disease, or the presence of inherited clotting disorders. The case control study was able to do this.
- Both of these studies were observational and were therefore prone to confounding and bias associated with this type of study. For example, body weight or BMI can influence the risk of thromboembolism and was not controlled or adjusted for in the Danish study.
There may be good reason why some women have been prescribed pills with higher risks of venous thromboembolism. It is important that women considering changing their contraceptive should consult their doctors to fully discuss these issues.
Overall, this edition of the BMJ will be useful for prescribers who are used to considering a range of possible side effects alongside the individual profiles and preferences of women in prescribing decisions.