Venous thromboembolism is egregiously underestimatedBMJ 2011; 342 doi: https://doi.org/10.1136/bmj.d3344 (Published 31 May 2011) Cite this as: BMJ 2011;342:d3344
- Samuel Shapiro, epidemiologist and emeritus professor1
In Jick and Hernandez’s report the incidence rates of idiopathic plus non-idiopathic VTE in drosperinone and levonorgestrel users were 5.0 per 10 000 women years (idiopathic cases comprised 61% of total cases) and 2.1 per 10 000 women years, respectively.1 In Parkin and colleagues’ report the corresponding rates of idiopathic VTE were 2.3 and 0.9 per 10 000 women years.2 In this second study the combined idiopathic plus non-idiopathic incidence rates cannot be precisely calculated because the non-idiopathic exclusions are only described in the methods, with no numbers given. However, for drosperinone and levonorgestrel users, the combined rates cannot be much more than about 61% higher—about 3.7 and 1.5 per 10 000 women years, respectively.
It is well established that for oral contraceptive users the overall incidence of VTE is 9-10 per 10 000 women years.3 For the reference drug, levonorgestrel, the incidence rates reported by Jick and Hernandez and by Parkin and colleagues were about four and six times too low. Even among the drosperinone users the rates were two to three times too low.
In the US and the UK, the allegation that drosperinone is more thrombogenic than levonorgestrel has been given considerable publicity. The major under-ascertainment of the incidence of VTE in oral contraceptive users, particularly users of levonorgestrel, in these studies makes bias not only possible, but likely.
Cite this as: BMJ 2011;342:d3344
Competing interests: SS currently consults, and in the past has consulted, with manufacturers of oral contraceptives, including Bayer Schering, the manufacturer of drosperinone.