Combined oral contraceptive pills (COCPs) containing desogestrel, gestodene or drospirenone appear to confer double the risk of venous thromboembolism (VTE) compared with those containing levonorgestrel, a large population-based study in the BMJ has found.

Data on all women in Denmark aged 15-49 years were identified from four national registries during the period 2001-2009. After exclusions for confounders such as malignant disease, pregnancy and coagulopathy almost 1.3 million women with more than 8 million women-years of data were analysed. Around a third had never used the COCP.

A total of 4,307 thromboembolic events were recorded in this cohort of which 4,246 were included and 2,847 (67%) were deemed certain cases as anticoagulant use was also confirmed. The analysis controlled for confounders such as BMI, smoking and social class.

Among women with a record of both VTE and anticoagulant treatment (the most definitive group), around one in 5,000 never-users of the COCP had a VTE event. Using this figure for reference the relative risk (RR) of confirmed VTE in users of COCPs containing 30-40 µg ethinylestradiol and levonorgestrel was 2.9.

Respective RRs for norgestimate, desogestrel, gestodene, drospirenone and cyproterone containing pills were 3.52, 6.61, 6.24, 6.37 and 6.35. Drospirenone maintained its higher VTE risk with the 20µg ethinylestradiol formulation pills as well; RR=6.95.

RR values were generally 2-3-fold higher in the confirmed than the unconfirmed group in whom the use of anticoagulants was not known and therefore the veracity of the VTE diagnosis may be less robust.

The authors conclude that in order to prevent one VTE event in one year around 2,000 women would have to change from using COCPs with desogestrel, gestodene or drospirenone to formulations with levonorgestrel.

It is encumbent on the prescriber to counsel carefully about the slight individual increase in VTE risk with third-generation progestogens and to avoid prescribing them at all where there are historical factors or comorbidities that suggest the background risk of venous thromboembolic disease is higher.

Dr Chris Barclay

• Lidegaard Ø, Nielsen LH, Skovlund CW et al. Risk of venous thromboembolism from use of oral contraceptives containing different progestogens and oestrogen doses: Danish cohort study, 2001-9.BMJ 2011;343:d6423