Previous studies have provided conflicting results regarding the effect of drospirenone-containing oral contraceptive pills (OCPs) on the risk of venous and arterial thrombosis.
Previous studies have provided conflicting results regarding the effect of drospirenone-containing oral contraceptive pills (OCPs) on the risk of venous and arterial thrombosis.
To conduct a systematic review to assess the risk of venous thromboembolism (VTE), myocardial infarction (MI), and stroke in individuals taking drospirenone-containing OCPs.
We systematically searched CINAHL, the Cochrane Library, Dissertation & Abstracts, EMBASE, HealthStar, Medline, and the Science Citation Index from inception to November 2012.
We included all case reports, observational studies, and experimental studies assessing the risk of venous and arterial thrombosis of drospirenone-containing OCPs.
Data were collected independently by two reviewers.
A total of 22 studies [six case reports, three case series (including 26 cases), and 13 comparative studies] were included in our systematic review. The 32 identified cases suggest a possible link between drospirenone-containing OCPs and venous and arterial thrombosis. Incidence rates of VTE among drospirenone-containing OCP users ranged from 23.0 to 136.7 per 100 000 woman-years, whereas those among levonorgestrel-containing OCP users ranged from 6.64 to 92.1 per 100 000 woman-years. The rate ratio for VTE among drospirenone-containing OCP users ranged from 4.0 to 6.3 compared with non-users of OCPs, and from 1.0 to 3.3 compared with levonorgestrel-containing OCP users. The arterial effects of drospirenone-containing OCPs were inconclusive.
Our systematic review suggests that drospirenone-containing OCP use is associated with a higher risk for VTE than both no OCP use and levonorgestrel-containing OCP use.
Oral contraceptive pills (OCPs) are associated with an increased risk of thrombotic events.[1-3] Fourth-generation OCPs were introduced to the North American market in 2000. This new generation is characterised by the addition of the progestin drospirenone, which was believed to be associated with a lower risk of thrombosis. Drospirenone-containing OCPs are currently the only available oral contraceptive with three indications: contraception; the treatment of premenstrual dysphoric disorder; and the treatment of moderate acne. However, recent observational studies have provided conflicting results regarding the effects of drospirenone-containing OCPs on the risk of venous thrombosis.[7-10] In addition, the effect of drospirenone-containing OCPs on the risk of arterial thrombosis remains controversial.[10, 11] We therefore conducted a systematic review to synthesise the available data regarding drospirenone-containing OCPs and the risk of venous and arterial thrombotic events, including deep vein thrombosis (DVT), pulmonary embolism (PE), myocardial infarction (MI), and stroke.
We systematically searched the CINAHL (from 1981 to November 2012), Cochrane Library (from 1898 to November 2012), Dissertation & Abstracts (from 1861 to November 2012), EMBASE (from 1947 to November 2012), HealthStar (from 1966 to November 2012), Medline (from 1946 to November 2012), and the Science Citation Index (from 1900 to November 2012) databases to identify all reports of thrombotic events in women taking OCPs (Appendix S1). In this systematic review, OCPs pertain to hormonal oral contraceptive pills containing a combination of estrogen and progestin. Keywords used were levonorgestrel, desogestrel, gestodene, norgestimate, and drospirenone. In addition, we searched www.clinicaltrialresults.org for potentially relevant randomised controlled trials (RCTs). We limited our search to studies conducted in the female adult population, and reported in English or French. The references of included studies were hand-searched to identify any additional potentially relevant publications.
Studies were included if they: (1) were case reports, case series, or comparative studies of women taking drospirenone-containing OCPs; (2) reported at least one of the venous and arterial thrombotic outcomes of interest [DVT, PE, MI, and cerebrovascular events, such as stroke or transient ischemic attack (TIA)]; and (3) were published in English or French. All studies failing to meet these criteria were excluded.
Two reviewers independently extracted the data from the studies included. Disagreements were resolved by consensus or, when necessary, by a third reviewer. Study characteristics such as study design, study period, population, and country of origin were extracted. For each outcome of interest, we extracted incidence rates (IRs) by exposure status and comparative effect measures, including hazards ratios (HRs), odds ratios (ORs), and rate ratios (RRs). Outcome data were extracted with corresponding 95% confidence intervals (95% CIs).
We performed this systematic review according to the MOOSE (Meta-analysis of Observational Studies in Epidemiology) statement, as all included studies were observational. The results of our systematic search are detailed in a flow chart that follows the guidelines outlined by the PRISMA statement (Figure 1).
Our search identified 9148 potentially relevant articles (Figure 1). Of these, 9123 were excluded because they were irrelevant to the subject of study (n = 9013), were editorials or commentaries (n = 62), or were other review articles (n = 48). A total of 25 full-text articles were retrieved for further review. Three additional studies were excluded: one presented the rationale and design for a prospective study, and the two others were subgroup analyses of a study already included. A total of 22 studies [six case reports, three case series (including 26 cases), and 13 comparative studies] were included in our systematic review. No interventional studies met our inclusion criteria.
The six case reports and three case series contained a total of 32 cases of thrombotic events that occurred in drospirenone-containing OCPs users (Table 1). All reports occurred in women residing in Europe, and were published between 2003 and 2012. A total of 31 women were taking a combination of 30 μg of ethinyl estradiol and 30 mg of drospirenone; one woman was taking a combination of 20 μg of ethinyl estradiol and 30 mg of drospirenone. The median age of women was 33.5 years (range: 17–50 years), and the median duration of drospirenone-containing OCP use before the thrombotic event was 150.5 days (range: 15–2557 days). Twenty of the 32 women described in the case reports and case series included had at least one known risk factor for thrombotic disease, including an age of >35 years, diabetes mellitus, family history of thrombotic disease, hyperlipidemia, hypertension, immobilisation, obesity, pregnancy/delivery, smoking, and surgical intervention. Six women also reported a genetic predisposition for thrombotic disease: factorV Leiden mutation; prothrombin G20210A mutation; or positive IgG anticardiolipin antibodies. A total of 27 women experienced VTEs, including two reports of venous thrombosis,[14, 15] nine DVTs,[14-18] two pulmonary thromboses,[14, 15] 12 PEs (one fatal),[15, 17] and two women with both DVT and PE. Risk factors were unspecified in 12 of the 27 women with confirmed venous thrombosis. Arterial thrombotic events were reported in four women, three of which had an MI,[5, 19, 20] and one of which had a TIA. All four women had at least one of three risk factors: smoking, family history of MI, and recent surgery.
|Studya||Year||Country||Age (years)||Past OCP||Duration (days)b||Risk factors for thrombotic events||Event|
|van Grootheest||2003||Netherlands||17||NR||183||None specified||Fatal PE|
|35||NR||17||Patient had given birth 4 months earlier||PT|
|Vaya||2003||Spain||21||NR||15||Smoking 15 cigarettes/day||TIA|
|Pearce||2005||UK||21||None||155||Smoking, homozygous for factor V Leiden||DVT|
|29||CPA/EE||42||Smoking, DVT in a sibling, heterozygous for factor V Leiden||DVT|
|30||DSG/EE||361||Heterozygous for prothrombin gene mutation and moderate positive IgG anticardiolipin antibodies||PE|
|34||NR||71||Smoking, varicose veins||PE|
|41||NR||263||Obesity (BMI = 30 kg/m)||PE|
|Cabou||2006||France||33||LNG/EE||15||Smoking, 15 days before MI, she had undergone a left ovarian cyst excision and left salpingectomy for an ectopic pregnancy||MI|
|Orti||2007||Spain||39||NR||731||Smoking, father died at 40 years old of MI||MI|
|Lopez||2009||Spain||43||NR||45||Obesity (BMI = 46 kg/m)||PE|
|37||NR||365||Smoking, hypertension, hyperlipidemia, diabetes, obesity (BMI = 38.22 kg/m)||PT|
|28||NR||21||Was immobilised for several weeks, obesity (BMI = 35.15 kg/m2)||PE|
|35||CPA/EE||122||Heterozygous carrier of prothrombin G20210A mutation||DVT, PE|
|21||NR||152||Heterozygous carrier of prothrombin G20210A mutation||DVT, PE|
|Zehir||2011||Turkey||36||NR||2557||Smoking 10 cigarettes/day||MI|
|Marti Gil||2012||Spain||36|| |
A total of 13 comparative studies evaluating the risk of thrombotic events related to the use of drospirenone-containing OCPs were identified (Table 2). Nine of the 13 identified studies were cohort studies, and the remaining four were case–control studies. No RCTs were identified. The total patient populations in the individual studies ranged from 867 to 1 626 158 women. Studies were reported [either published or included in Food and Drug Administration (FDA) briefing material] between 2007 and 2012, and included data from databases of developed countries, notably the National Registry of Medicinal Products Statistics, National Registry of Patients, Statistics of Denmark, the European Active Surveillance Study (EURAS), German outpatient offices, Ingenix Research Data Mart, the Multiple Environmental and Genetic Assessment study (MEGA), the UK General Practice Research Database (GPRD), the US PharMetrics database, Kaiser Permanente Northern California, Kaiser Permanente Southern California, US State Medicaid databases, and the Israeli Clalit Clinical database. The duration of follow-up ranged from 12 to 180 months, and occurred from 1995 to 2011. There was heterogeneity in inclusion criteria and user definitions, with six studies including prevalent users and seven involving new users or initiators (Appendix S2).
|Studya||Study design||n||Data origin||Study period||Study population|
|Dinger 2007||Prospective cohort||42 875||EURAS study||2000–2005||Initiators of OCP treatment (first-ever users or switchers to a new product)|
|Seeger 2007||Prospective, claims-based, cohort||67 287||Ingenix Research Data Mart||2001–2004||Aged 10–59 years starting DRSP or other OCs|
|Lidegaard 2009||National cohort||NS||Four Danish registries||1995–2005||Aged 15–49 years with no history of cardiovascular or malignant disease|
|Vlieg 2009||Population-based, case–control||1556||MEGA study||1999–2004||Aged <50 years who were not pregnant, not within 4 weeks postpartum, not using a hormone-excreting IUD or depot contraceptive|
|Dinger 2010||Community-based, case–control||366||German outpatient offices||2002–2008||Aged 15–49 years, with a clinical diagnosis of VTE|
|Parkin 2011||Nested case–control||276||UK General Practice Research Database||2002–2009||Aged 15–44 years without major risk factors for VTE who started a new episode of use of an OCP|
|Jick 2011||Nested case–control and cohort||867||US PharMetrics database||2002–2008||Aged 15–44 years and current users of OCPs|
|Lidegaard 2011||National cohort||1 436 130||Four Danish registries||2001–2009||Aged 15–49 years with no previous venous, arterial thrombotic events, or cancers|
|FDA 2011||Population-based cohort||835 826||KPNC, KPSC, and two State Medicaids||2001–2007||Aged 10–55 years and current users of OCPs|
|Gronich 2011||Population-based cohort||329 995||Israeli Clalit clinical database||2002–2009||Aged 12–50 years with no previous diagnoses of thrombotic events|
|LASS 2011||Prospective cohort||47 799b||EURAS study + 5–year extended follow-up||2000–2011||Initiators of OCP treatment (first-ever users or switchers to a new product)|
|Leppee 2012||Cohort||1 050 000||HALMED||2008–2010||Aged 15–49 years|
|Lidegaard 2012||National historical cohort||1 626 158||Four Danish registries||1995–2009||Aged 15–49 years, not pregnant, with no history of cardiovascular disease, cancer, venous or arterial thrombotic event, coagulopathy, bilateral oophorectomy, unilateral oophorectomy two times, hysterectomy, or sterilisation procedure|
The primary endpoint was VTE for 12 of the comparative studies included (Table 3), and arterial thrombosis for one of the studies included (Table 4). Eight studies compared the risk of VTE between drospirenone-containing and levonorgestrel-containing OCP users. The incidence rates for VTE ranged from 23.0 to 136.7 per 100 000 women-years for drospirenone-containing OCP users, and from 6.64 to 92.1 per 100 000 woman-years for levonorgestrel-containing OCP users. Drospirenone-containing OCP users had an increased risk of VTE compared with users of levonorgestrel-containing OCPs, with relative risks ranging from 1.0 to 3.3. In the eight studies comparing the risk of VTE between levonorgestrel- and drospirenone-containing OCPs, five reported a greater risk for VTE among users of drospirenone-containing OCPs,[7, 8, 11, 22, 23] whereas the three other studies were inconclusive.[9, 24, 25] Two studies examined these associations in both ‘all users’ and a subgroup of ‘new users’ of drospirenone-containing OCPs.[11, 23] In both studies, the ‘new user’ analysis produced results that were consistent with those of the ‘all user’ analysis with respect to VTE (Table 3).
|Study||DRSP na||Comparator n||DRSP users||Comparator||Effect measure||Point estimate||95% CI|
|IRb||95% CI||IRb||95% CI|
|Drospirenone- versus levonorgestrel-containing OCPs|
|Dinger 2007||16 534||26 341||91||59–33||80||52–117||HR||3.3||0.9–10|
|Parkin 2011||NRd||NR||23.0||13.4–36.9||9.1||6.6, 12.2||OR||3.3||1.4–7.6|
|Jick 2011||NRe||NR||30.8||25.6–36.8||9.6||9.6, 15.9||OR||2.4||1.7–3.4|
|FDA 2011 (all users)||142 166||198 839||102.2g||NR||6.64g||NR||RR||1.45||1.15–1.83|
|FDA 2011 (new users)||NR||NR||136.7g||NR||92.1g||NR||RR||1.57||1.13–2.18|
|Gronich 2011 (all users)||73 629||21 546h||86i||NR||69i||NR||RR||1.65||1.02–2.65|
|Gronich 2011 (new users)||NR||NR||NR||NR||NR||NR||RR||1.67||0.98–2.86|
|LASS 2011||NR||NR||107||81–139||92||69, 120||HR||1.1||0.8–1.7|
|Drospirenone-containing OCPs versus other OCP users|
|Seeger 2007||22 429||44 858||130||80–200||NR||NR||RR||0.9||0.5–1.6|
|Leppee 2012||NR||NR||NRj||NR||NRk||NR||Incidence RR||6.4||NR|
|Drospirenone-containing OCPs versus non-users of OCPs|
|Study||DRSP n a||Comparator n||DRSP users||Comparator||Effect measure||Point estimate||95% CI|
|IRb||95% CI||IRb||95% CI|
|Drospirenone versus levonorgestrel-containing OCP users|
|FDA 2011 (all users)||142 166||198 839||10.8||NR||16.4||NR||HR||0.81||0.45–1.44|
|FDA 2011 (new users)||109 070||137 311||25.5||NR||22.8||NR||HR||1.64||0.79–3.40|
|Gronich 2011||73 629||21 546c||58d||NR||123d||NR||RR||0.87||0.56–1.33|
|LASS 2011||NR||NR||13||5, 28||38||24, 58||HR||0.4||0.2–0.9|
|Drospirenone-containing OCP versus other OCP users|
|FDA 2011 (all users)||142 166||586 278||10.8||NR||14.4||NR||HR||0.99||0.58–1.69|
|FDA 2011 (new users)||109 070||383 151||25.5||NR||17.6||NR||HR||2.01||1.06–3.81|
|LASS 2011||NR||NR||13||5, 28||32||22, 45||HR||0.4||0.2–0.8|
|Drospirenone-containing OCP versus non-users of OCPs|
|Lidegaard 2012 (stroke)e||NR||NR||18.1||NR||24.2||NR||Relative Risk||1.64||1.24–2.18|
|Lidegaard 2012 (MI)f||NR||NR||6.3||NR||13.2||NR||Relative Risk||1.65||1.03–2.63|
Two studies investigated the risk of VTE in drospirenone-containing OCP users compared with that in users of other oral contraceptives.[26, 27] One study involved 18 cases of VTE among drospirenone users and 39 among users of other oral contraceptives. The comparison between these different formulations of oral contraceptives was inconclusive because of the sparse data (RR 0.9, 95% CI 0.5–1.6). The other study involved 17 cases of VTE among drospirenone users and four among norgestimate and desogestrel users. The authors reported an incidence rate ratio of 6.4.
Three of the included studies compared the risk of VTE in drospirenone-containing OCP users with non-users of OCPs.[22, 25, 28] The incidence rate for VTE ranged from 78.3 to 93 per 100 000 woman-years among drospirenone-containing OCP users, and from 37 to 54.7 per 100 000 woman-years among non-users of OCPs. After adjusting for potential confounding factors (Appendix S3), drospirenone-containing OCP users had a substantially higher risk of VTE (relative risk ranging from 4.0 to 6.3) compared with non-users.
Our literature search identified four studies that compared the risk of arterial thrombosis between drospirenone-containing and other OCP users (Table 4). Incidence rates for arterial thrombosis ranged from 6.3 to 58 per 100 000 woman-years among drospirenone-containing OCP users, and from 13.2 to 123 per 100 000 woman-years among levonorgestrel-containing OCP users. In the Long-term Active Surveillance Study (LASS), drospirenone-containing OCP users had a substantial reduction in the risk of arterial thrombosis compared with levonorgestrel-containing OCP users (HR 0.4, 95% CI 0.2–0.9), whereas the results of Gronich, and the FDA analysis of all users, were inconclusive, with relative risks ranging from 0.81 to 0.87, and the limits of their 95% CIs including both clinically important harms and benefits. In contrast, when the FDA analysis was restricted to new users, the HR increased to 1.64 (95% CI 0.79–3.40), although the broad 95% CIs arising from sparse data prevent strong conclusions from being drawn from this analysis.
The comparison of the arterial thrombotic effects of drospirenone-containing OCPs with those of other OCPs also produced heterogeneous results (Table 4). The LASS found that drospirenone-containing OCP users had a substantial reduction in arterial thrombosis (HR 0.4, 95% CI 0.2–0.8), whereas the FDA's analysis of all users resulted in an HR of 0.99 (95% CI 0.58–1.69). Restriction to new users in the FDA study resulted in an increased risk of arterial thrombosis among drospirenone-containing OCPs users, compared with users of other OCPs (HR 2.01, 95% CI 1.06–3.81).
The 2012 Lidegaard study compared arterial thrombotic risk in drospirenone-containing OCP users with that in non-users. In this study, drospirenone-containing OCPs were associated with an increased risk of stroke (RR 1.64, 95% CI 1.24–2.18) and MI (RR 1.65, 95% CI 1.03–2.63).
Our study was designed to summarise the available evidence regarding the venous and arterial thrombotic risk of drospirenone-containing OCPs. The evidence to date suggests that drospirenone-containing OCPs may increase the risk of VTE compared with levonorgestrel-containing OCPs and non-use of OCPs. The effects of drospirenone-containing OCPs on the risk of arterial thrombosis remain unclear, with the studies included in this review providing conflicting results: some suggested a protective effect; and others suggested a doubling of risk.
Twenty out of the 32 cases identified in case reports and case series had at least one concomitant risk factor for thrombotic events, highlighting the need to screen for thrombotic risk factors before initiating OCPs. Furthermore, although the duration of OCP use varied among cases, the majority of thrombotic events occurred during the first year of OCP use (28 out of 32 cases). Comparative studies involving women starting OCP therapy also had greater incidence rates for VTE,[24, 26] compared with studies involving prevalent OCP users.[7, 8] This trend is consistent with conclusions drawn from previous studies investigating thrombotic risk with the use of second- and third-generation OCPs.
The elevated VTE risk that occurs following the initiation of OCP use has important implications for the design and analysis of observational studies of this association. With the greatest risk occurring following the start of therapy among first-time users, the failure to properly account for history of OCP use may result in spurious findings.[30, 31] In addition, the inclusion of prevalent or current users may result in an important underestimation of treatment effects, as those who experienced events early after the start of therapy (but before the study period) are excluded for having a history of thrombosis. User definitions used in the studies included varied (Appendix S3), which may explain some of the observed heterogeneity of results. For example, the restriction to new users of drospirenone-containing OCPs in the FDA-funded study resulted in substantially higher risks of arterial thrombosis. Moreover, the estrogen dose, although known to be associated with a higher risk of both venous and arterial thrombosis,[25, 29] was unspecified in several of the studies included. These potential methodological limitations of the studies included need to be considered when weighing the strength of the evidence supporting the association between drospirenone-containing OCPs and thrombotic risk.
Importantly, although drospirenone-containing OCPs appear to increase the risk of VTE, and have unclear effects on the risk of arterial thrombosis, the absolute risk of thrombosis when using these agents remains low. Among drospirenone-containing OCP users, the incidence rate ranged from 23.0 to 136.7 per 100 000 woman-years for VTE, and from 6.3 to 58 per 100 000 woman-years for arterial thrombosis. Hence, there is probably insufficient evidence to recommend discontinuing the use of drospirenone-containing OCPs, particularly among long-term users. However, women with VTE are also at risk for developing arterial thrombotic events, and women should be provided with our current understanding of the risks and benefits associated with the use of these agents to allow for informed decision-making.
In 2011, the Medicines and Healthcare products Regulatory Agency (MHRA) in the UK, the US FDA, and Health Canada conducted reviews concluding that drospirenone-containing OCPs may be associated with a 1.5–3 times higher risk of VTE, and warning labels have been revised to adequately reflect this risk.[11, 34-36] These results are supported by the findings of our systematic review. It should be noted that the statements released by these regulatory agencies dealt only with venous effects, and that the arterial effects of drospirenone-containing OCPs remain under-investigated.
Our systematic review was the first to evaluate the safety of drospirenone-containing OCPs with respect to both venous and arterial thrombotic outcomes. The inclusion of detailed case reports allows for a clinically relevant examination of thrombotic risk factors among exposed cases, and the inclusion of comparative studies allows for rigorous statistical adjustment for potential confounding factors and uses a comparison group to account for the underlying thrombotic risk in this population. The effect of OCPs, including that of drospirenone-containing OCPs, on the risk of venous thrombosis was recently examined in two systematic reviews and meta-analyses.[37, 38] However, the literature searches for these two previous reviews were conducted in April–May 2010, and nine studies have since been completed. Furthermore, given the heterogeneity across studies, the meta-analysis of these data is questionable.
Our study has several potential limitations. First, because of the heterogeneity of comparators, user definitions, and effect measures reported, we were unable to pool data across studies to derive a single overall summary estimate. Secondly, our systematic search did not identify any interventional studies examining this issue. Given the observational nature of the included studies, there is the possibility of confounding by indication. In addition, based on the anti-mineralocorticoid and anti-androgenic properties of drospirenone, OCPs containing this progestin may have been preferentially prescribed to women with conditions associated with a higher risk of VTE and arterial thrombosis. Furthermore, despite the use of rigorous statistical adjustment (Appendix S2), the possibility of residual confounding remains. All of the studies included contain various degrees of switching between OCPs, and the inadequate adjustment for prior use is likely to result in an overestimation of the risk of thrombosis. In addition, the present systematic review was limited to studies published in English or French, and may thus be affected by language bias. There is widespread awareness of the association between VTE, which is often asymptomatic, and OCP use. Thus, the studies included may be affected by detection bias.
Although studies examining the thrombotic effects of drospirenone-containing OCPs have methodological limitations, our systematic review suggests that users of these oral contraceptives may be at greater risk for VTE than either non-users of OCPs or users of levonorgestrel-containing OCPs. Despite the observed increased VTE risk, the absolute risk of thrombosis remains low. Doctors should therefore consider the indication for use and the risk–benefit profile of the individual woman prior to prescribing these OCPs. With the available studies providing conflicting results, the effect of drospirenone-containing OCPs on arterial thrombosis remains unclear. Further studies on the arterial thrombotic effects of these OCPs are warranted.
The authors declare that they have no competing interests to disclose.
CQW conducted the literature search, extracted data, and drafted the article. All authors contributed to the study design, interpretation of data, and critically reviewed the article for important intellectual content.
This study involved published data, and so did not require ethics approval.
This study was funded by the Canadian Institutes of Health Research (CIHR; grant number MOP-119362). CQW is supported by the Mach-Gaensslen Foundation of Canada Student Grant, funded through the McGill Research Bursary Program. KBF is a CIHR New Investigator.
We would like to thank Renee Atallah, Tara Dourian, Caroline Franck, and Elana Folbe for their help with data abstraction.
Combined oral contraceptive pills (OCP) containing drospirenone are among the most popular pills worldwide. Although cardiovascular events such as deep vein thrombosis (DVT), myocardial infarction, and stroke are rare among OCP users, they are potentially life-threatening, and so robust data on the risk associated with use of OCPs containing different progestogens is important. However, when interpreting the data in systematic reviews such as that of Wu et al. (BJOG 2013;120: DOI: 10.1111/1471-0528.12210), we need to remain cognisant of key epidemiological concepts. Firstly, outcomes such as DVT are rare and so huge numbers of subjects must be studied. There are therefore no RCTs, and so suboptimal methodologies have to be used instead, such as case–control studies or the examination of large databases or registries. Newer OCPs containing progestogens such as drospirenone are often marketed as being ‘safer’, and so may be selectively prescribed to women with risk factors. Many of the retrospective studies included in this review have previously been the subject of extensive critique because they lack data on important confounding variables such as body mass index, smoking, or family history of DVT (Jensen and Trussell Contraception 2012;86:327–329). It has also been argued that the prospective comparative studies should rank higher in terms of ‘validity’ than retrospective ones, and these have not shown a higher risk of DVT with drospirenone containing OCPs. Wu et al. also stress that the risk of DVT is greatest when a woman starts taking an OCP for the first time, or restarts it, and so bias can be introduced in studies comparing ‘new users’ of OCPs with drospirenone and established users of OCPs containing older progestogens (such as levonorgestrel). So, given these potential biases, flaws, and epidemiological criticisms, what can clinicians advise women about the use of OCPs containing drospirenone and cardiovascular risks such as DVT? Well, we can advise women that the risk of DVT is low, and even if the use of an OCP doubles or trebles this risk, it is still a rare event. We should also advise them that the risk of a DVT with the use of any OCP is much less than during pregnancy or postpartum, and so they should not discontinue an OCP because of their fear of DVT. We can put risk into context by using the absolute risk and advising them that the incidence of DVT is estimated to be 9–10 per 10 000 women-years for an OCP user (compared with 5–10 per 10 000 for non-users), but approximately 29 per 10 000 during pregnancy and 300–400 per 10 000 postpartum (combined hormonal contraception, guideline 2012, Faculty of Sexual and Reproductive Healthcare, www.fsrh.org.uk). As the effectiveness of the OCP depends on compliance, which depends on user satisfaction, it is important that women remain on an OCP with which they are satisfied (even if that OCP contains drospirenone). However, the extent (if any) to which the type of progestogen may affect the risk of DVT (or other rarer cardiovascular outcomes) in OCP users remains, unfortunately, difficult to assess.
The author has received funding from several pharmaceutical companies (Exelgyn, HRA Pharma, and Pfizer), but not from the manufacturer of OCPs containing drospirenone. The author does not routinely prescribe OCPs containing drospirenone for contraception, as they are not approved by the Scottish Medicines Consortium on a cost-effectiveness basis.
Consultant, Chalmers Sexual and Reproductive Health Service, Edinburgh, UK