Drospirenone-containing oral contraceptives and risk of venous thromboembolism

Combined oral contraceptives (COCs) are widely used by women to prevent unwanted pregnancies. The fourth generation of COCs, containing the progestin drospirenone, came on the market in 2000. In the last decade, there have been concerns regarding the potential venous thromboembolytic (VTE) risk of drospirenone-containing COCs. The many studies that assessed this relationship reported variable results which may be due to differences in their methodological quality. I aimed to synthesize the literature regarding drospirenone-containing COCs and the risk of VTE with a specific focus on its methodological quality. Subsequently, I aimed to describe and illustrate the methodological challenge of studying the risk of VTE among first-time users of drospirenone-containing COCs in a healthcare database.In the first manuscript, I systematically searched seven databases from inception to November 2015 to identify all observational studies assessing the risk of VTE among users of drospirenone-containing COCs. I identified 17 studies that met the inclusion criteria (11 cohort and 6 case-control). Based on the ACROBAT-NRSI, one study had low risk of bias, nine had a moderate risk, three had a serious risk, and four had a critical risk. Ten studies included prevalent user bias, five studies included inappropriate comparator groups, seven studies had VTE misclassification, and five failed to adjust for two or more important confounders. In the second manuscript, I conducted an original research study that examined the VTE risk of drospirenone-containing COCs. I used data from the Clinical Practice Research Datalink (CPRD) to create two cohorts. The first-time user cohort included all women aged 16 to 45 years who received a first ever prescription of drospirenone- or levonorgestrel-containing COCs between May 1st, 2002 and March 31st, 2015. The restarter cohort included all women aged 16 to 45 years who restarted a COC after a period of non-use of at least 6 months. For both cohorts, women were followed until VTE (defined by hospitalization for VTE or an outpatient diagnosis of VTE with subsequent prescription of warfarin, INR testing, or death within 90 days), discontinuation of COC use, switching to any other form of hormonal contraception (including the other study COC), pregnancy, arterial thromboembolism, death, departure from the CPRD, or the end of the study period, whichever occurred first. High-dimensional propensity scores (HDPS) were estimated to contrast users of drospirenone- and levonorgestrel-containing COCs. Cox proportional hazards models were used to estimate hazard ratios (HRs) and corresponding 95% confidence intervals (CIs). The final cohorts comprised 55,139 first-time ever users (3,582 drospirenone and 51,557 levonorgestrel) and 162,959 restarter users (23,191 drospirenone and 139,768 levonorgestrel). Among first-time ever users, drospirenone-containing COCs were associated with a substantially higher risk of VTE than levonorgestrel-containing COCs (adjusted HR: 3.19, 95% CI: 1.12-9.08). Among restarters, the adjusted HR was 1.96 (95% CI: 1.12-3.41). Both cohorts produced non-proportional hazards and all sensitivity analyses were consistent with the main findings. The most important limitation of previous studies examining the association between drospirenone-containing COCs and the risk of VTE was the inclusion of prevalent users. To overcome these limitations, I examined two cohorts: a first-time ever user cohort and a restarter cohort, the latter of which represents a novel approach to studying the VTE risk of COCs. Although left truncation of healthcare databases is a concern for the identification of first-time users, the use of a well-defined restarter cohort suggests a doubling of VTE risk with drospirenone-containing COCs relative to levonorgestrel-containing COCs. Physicians and patients should be aware of this increased risk when considering the most appropriate choice of contraception.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.