Re: Venous thrombosis in users of non-oral hormonal contraception: follow-up study, Denmark 2001-10
We read with interest the article by Lidegaard et al, published on 10 May 2012, titled ‘Venous thrombosis in users of non-oral hormonal contraception: follow-up study, Denmark 2011-10 (1). The findings from this analysis of a retrospective cohort of 1,626,158 Danish women aged 15-49 years suggest that women using either the transdermal combined contraceptive patch or the combined hormonal vaginal ring are at significantly higher risk of venous thrombosis than women who do not use hormonal contraception. Risk of venous thromboembolism (VTE) was nearly 8-fold greater among women using the patch, and ring users were reported to have a 6.5 fold greater risk, compared with non-hormonal users. The analysis showed women using a subcutaneous implant had a greater risk (relative risk = 1.4), but the estimate was not statistically significant, and the LNG-releasing intrauterine device had a protective effect compared with non-hormonal users (RR = 0.6).
We are concerned that these new data, as they are presented in the article, will have effect on prescribing practices and upon women themselves, which may lead to a new (unfounded) scare regarding several methods of contraception, as witnessed in 1995 and attributed with the subsequent increased rates of abortion observed in countries such as UK, Norway and Sweden (2-6). Moreover, problems with contraceptive compliance and uncertainty among users about the safety of contraceptive methods may emerge. This new epidemiological study concentrates on risk of venous thrombosis and compares different combined hormonal contraceptive methods, which in turn contain different progestogens. The study claims differences in risk among the methods of hormonal contraception and goes so far as to make recommendations for clinical decision-making.
Regarding the evidence on VTE risk presented by Lidegaard et al., we want to remind the professional community that it is essential to take into account the following issues when interpreting findings reported in publications on this topic: types of interventions investigated; characteristics of the users; the outcomes measured; the study design; and biological plausibility. Comparing different combined hormonal contraceptive methods needs to take into account not only the methods being compared, but also patterns of use (initiation, duration of use, discontinuation, and re-initiation), and the pharmacological properties of the hormones in the methods.
With regards to VTE, there are a large number of known risk factors that predispose women to an elevated risk of VTE, which may in turn complicate (or bias) comparison between different cohorts of users, especially when they are based upon retrospective cohorts where data on risk factors may not be available (7).
Thromboembolic events are difficult to diagnose. Currently, the prevalence and incidence rates in healthy population are contradictory and remain under debate. For example, VTE incidence among non-pregnant, non-hormonal contraceptive users reported in the literature ranges from 2 to 13 events per 10,000 women-years (8). It has been suggested that there is a high likelihood of false positive diagnoses being reported (9). Observational studies analyzing records from databases established for administrative, rather than scientific, purposes reasons are highly vulnerable to bias. This is particularly true when comparing associations between VTE and exposure to contraceptives, where the possibility for misclassification of VTE diagnosis and under-reporting of contraceptive use are likely (9). Further, inferring a causal association based upon epidemiological findings warrants an explanation of the mechanism linking cause and effect. For example, we are unclear as to how the lower pharmacokinetic exposure of ethanol estradiol released by vaginal ring, compared with combined hormonal pills, could result in a greater risk of VTE.
Taking into account these methodological issues, we believe that Lidegaard et al. overstate the implications of their results and readers should interpret the assertions they make with caution.
Venous thrombosis is a serious condition. While Lidegaard et al.’s analysis identified an elevated risk of VTE among users of certain hormonal contraceptive methods compared with women who did not use a hormonal method, in each group, it is important to note that the absolute risk of VTE was low for all methods (incidence rates were below 10/10,000 women-years exposure regardless of the method). More troubling, there was no consideration of weighing the beneficial effects of contraception on avoiding intended or unwanted pregnancy, or the consequences of pregnancy, in that pregnancy carries a much high risk of VTE (7). Further, it is difficult to ascertain how a practitioner would find the computations of the number of women who should switch methods helpful as they counsel their patients on what method of contraception is right for them. It is imperative that clinicians providing contraception have the evidence they need to counsel their patients on medical eligibility, so that women are able to make informed decisions regarding the contraceptive method that meets their personal needs.
1. Lidegaard O, Nielsen LH, Skovlund CW, Lokkegaard E. Venous thrombosis in users of non-oral
hormonal contraception: follow-up study, Denmark 2001-10. BMJ 2012; 344:e2990.
2. Dillner L. Pill scare linked to risk in abortions. BMJ 1996; 312:996.
3. Wood R, Botting B, Dunnell K. Trends in conceptions before and after the 1995 pill scare. Popul Trends 1997;89:5-12.
4. Skjeldestad FE. Increased number of induced abortions in Norway after media coverage of adverse vascular events from the use of third-generation oral contraceptives. Contraception 1997;55:11-4.
5. Milsom I, Odlind V, Hedberg C, Lidegaard O. [Alarming increase in the number of abortions. Does fear of thrombosis limit the use of oral contraceptives?] Lakartidningen 1997;94:4735-6.
6. Nilsen ST, Iversen OE. [Negative reports on oral contraceptives – increased number of interrupted pregnancies] Tidsskr Nor Laegeforen 1996;116:3486-8.
7. Dinger J, Shapiro S. Combined oral contraceptives, venous thromboembolism, and the problem of interpreting large but incomplete datasets. J Fam Plann Reprod Health Care 2011;doi:10.1136.
8. Dinger JC, Heinemann LAJ, Kul-Habisch D. The safety of a drospirenone-containing oral contraceptive: final results from the European Active Surveilance study on Oral Contraceptives based on 142,475 women-years of observation. Contraception 2007;72:344-54.
9. Severinsen MT, Kristensen SR, Overad K, et al. Venous thromboembolism discharge diagnoses in the Danish National Patient Registry should be used with caution. J Clin Epidemiol 2010; 63:223-228.
Competing interests: No competing interests