The newer contraceptive pills and venous thromboembolism risk

The 19th century German physician Rudolf Virchow described the mechanism of venous thromboembolism (VTE) in terms of interactions between haemodynamics, vein wall disruption and hypercoagulability. An understanding of this interplay is particularly important in a women’s health context since changes in the hormonal milieu, such as those seen in pregnancy and the use of exogenous oestrogen therapy, may act in concert to further increase VTE risk.

In 1961, less than a year after the introduction of the combined oral contraceptive pill (COCP), a 40-year-old British nurse was the first to be diagnosed with a VTE believed to be secondary to its use. From that point on, discussion of this increased risk became an integral part of any contraceptive pill counselling. However, that risk can be difficult to accurately quantify since it varies with ethnic background and from country to country. There are higher rates of thromboembolism for instance in Scandinavia and in Holland, because the background incidence of factor V Leiden carriage there is around 15%,1 compared with only 2%–4% in Australia.2 The most widely accepted estimate, however, is that the background risk of VTE in a woman of reproductive age (15–44 years), who is nether pregnant nor using combined contraception, is around 4 per 10 000 per year,3 and consensus holds that this risk is two to three times higher in those using combined contraception. It is also important to appreciate that the risks of VTE are highest in the first 12 months of COCP use. Such an increase has generally been regarded as clinically acceptable, however, given that the absolute number of events in this group is very low and certainly much lower than that seen during pregnancy.

However, three epidemiological studies published in the Lancet in 1995 suggested that COCPs containing the newer progestogens clustered at the higher end of the VTE risk range when compared with older combinations containing norethisterone or levonorgestrel. This resulted in the newer products being temporarily withdrawn from the British market, despite an acknowledgement that various biases and confounders may have affected the published results. One issue is that the “marketing glow” associated with the introduction of a new preparation may persuade prescribers to use it preferentially in new patients or in those with acknowledged risk factors. A temporary increase in VTE rates, for instance, was noted in women commenced on gestodene-containing COCPs when these were first introduced into Austria in the mid 1990s, but this difference disappeared within 10 years.4 Antiandrogenic preparations are particularly subject to prescriber bias due to their more frequent use in women with polycystic ovarian syndrome — a condition with a higher background risk of VTE.5

The 1995 media coverage of the perceived increased risks associated with the newer COCPs led many women in the United Kingdom to discontinue hormonal contraception. Despite a steady reduction over the preceding 5 years, 1995–1996 saw an 8% increase in abortion rates in Britain.6 In the same year, deliveries increased by around 25%7 — particularly ironic since pregnancy and the postpartum period hold significantly higher risks of VTE than any COCP. Marketing restrictions on the newer preparations were eventually lifted in 1996, with the one proviso that the older preparations be regarded as “first choice”.

Neither drospirenone- nor cyproterone acetate-containing preparations figured in the 1995 studies, but by 2002 there were suggestions that they too had a higher VTE risk. Between 2007 and 2014, eight studies had reached conflicting conclusions about this risk, sparking an intense debate over the significance and validity of their various findings. Four of the six retrospective studies8-11 showed a two- to three-times increase in VTE risk in women taking these newer progestogens (compared with levonorgestrel-containing pills), while two showed no difference.12,13 Neither of the large well designed prospective studies14,15 (which had a higher evidentiary level but were both pharma-sponsored) showed significant differences in VTE risk between any of the preparations in routine clinical use. In a case of duelling epidemiologists, one side stands accused of potential commercial bias and the other of poor control selection, non-significance, missing data and neglecting the effect of bias and confounders.

Media commentary on recent scientific research on the pill and VTE has been characterised by limited understanding of the scientific process and a strong focus on individual experiences and outcomes. There is also no doubt that an alarmist approach to such research creates far more impact than a considered one. In January 2013, for instance, there was extensive reporting of the French medicines agency (National Agency for the Safety of Medicine and Health Products) deregistering cyproterone acetate-containing pills after reports of four deaths attributed their use. Yet, with their reintroduction to the French market in January 2014 after a comprehensive scientific review by the European Medicines Agency, there was barely a media whisper.

When considering the public health impact of emerging data on contraceptive safety, it appears we have learnt little from the “pill scares” of the past, and women across the world continue to be concerned and confused by media misinformation and misrepresentation. Undue pressure is often placed on various regulatory authorities to act before proper evaluation of the evidence, though thankfully regulators in Australia have chosen to take a very measured approach to date. Despite the contradictory evidence, the case against the newer progestogens is already regarded as a fait accompli by the legal profession, and a number of lawsuits are presently in train. Settlement of such cases, even when for commercial rather than scientific reasons, reinforces the perception of unacceptable risk. This raises the distinct possibility that contraceptive choices in the future will narrow as manufacturers are driven by medicolegal rather than scientific considerations. Unlike epidemiologists, clinicians are only too aware that “one pill does not fit all”, and that a wide range of contraceptive options is required to best serve the needs of the community.

Every woman deserves an individualised discussion and risk analysis if she is contemplating the use of a COCP, and this discussion should take account of both her medical history and her preferences. As well as covering the VTE risks associated with COCP use, the patient should be made aware of the fact that long-acting methods such implants, injectables and intrauterine devices are not only more effective but do not increase the risk of VTE. We also need to develop better ways of communicating risk to patients. Comparison tables, even those that adopt the “worst-case scenario” as shown in the Box, are often useful tools for putting VTE risk into perspective.

Though a “doubling” of the VTE risk in newer pill users sounds alarming, it actually translates into an additional four to six attributable cases per 10 000 users per year.17 The risk of death from VTE is about one-hundredth of that and could be negated, albeit only statistically, by choosing to drive for 2 hours less each year.18

Professional bodies such as the Faculty of Sexual and Reproductive Healthcare in the UK provide evidence-based guidance on contraceptive use and VTE risk.19 Previous VTE remains an absolute contraindication to all COCP use. The signs and symptoms of VTE should be discussed with all prospective COCP users, and a downloadable fact sheet intended for consumer use is available through Family Planning New South Wales.20 If a woman chooses a COCP as her preferred option, it is preferable that she be commenced on a preparation containing norethisterone or levonorgestrel. In Australia, this also makes good economic sense since such preparations are subsidised. However, if the initial choice does not suit, or non-contraceptive benefits drive the decision, the clinical threshold for suggesting a trial of one of the newer preparations should not be unduly high.

Scientific research is an iterative process with each study adding gradually to the pool of knowledge until a more objective whole is reached. It will require further dispassionate analysis of existing research and a commitment to filling in the gaps with more well designed prospective studies in order to gauge the true impact of the newer progestogens on VTE risk. As we continue along this path, however, we should never lose sight of the simple fact that for most women, the benefits of combined contraceptives, regardless of their composition, far outweigh their risks.