An Increased Risk of DVT with Vaginal Ring Contraception — True or Flawed Studies?
Abstract & Commentary
By Jeffrey T. Jensen, MD, MPH , Leon Speroff, Professor and Vice Chair for Research, Department of Obstetrics and Gynecology, Oregon Health and Science University, Portland, is Editor for OB/GYN Clinical Alert.
Synopsis:Use of the contraceptive ring or patch was associated with a two-fold increase in venous thrombosis when compared to a standard oral contraceptive pill containing levonorgesterel
Source: Lidegaard O, et al. Venous thrombosis in users of non-oral hormonal contraception: Follow-up study, Denmark 2001-10. BMJ 2012;344:e2990. PMID:22577198.
The authors assessed the risk of venous thrombosis (vt) in current users of non-oral hormonal contraception using four national registries in Denmark. Non-pregnant women aged 15-49 years (n = 1,626,158) with no prior diagnosis of thrombotic disease or cancer were followed from 2001 to 2010. The main outcome measures were the incidence rate of VT in users of transdermal, vaginal, intrauterine, or subcutaneous hormonal contraception. The incidence rates were used to calculate the relative risk of VT in users of non-oral hormonal contraception compared with non-users of hormonal contraception and with a reference low dose (30-40 mcg) ethinyl estradiol levonorgestrel (LNG) oral contraceptive. Diagnoses were confirmed by records indicating at least 4 weeks of anticoagulation therapy after the diagnosis. In the study population, a total of 5,287 VT diagnoses were recorded in 9,429,128 woman-years of observation, and 3,434 were confirmed. The baseline rate of confirmed VT events in non-users of hormonal contraception was 2.1 per 10,000 woman years. The incidence of VT was elevated in users of the transdermal patch (9.7/10,000) and intravaginal ring (7.8/10,000), yielding adjusted relative risks (RR) of confirmed VT of 7.9 (95% confidence interval [CI] 3.5-17.7) for the patch and 6.5 (CI 4.7-8.9) for the ring. The RR also was increased in women who used subcutaneous implants (RR 1.4; CI 0.6-3.4; not statistically significant) but not in those who used the LNG intrauterine system (RR 0.6; CI 0.4-0.8). Compared with users of the reference low-dose LNG combined oral contraceptive, the adjusted RR of VT was doubled in users of the patch (RR 2.3; CI 1.0-5.2) and ring (RR 1.9; CI 1.3-2.7).
The Danish National Database studies provide a gift that keeps on giving. Unfortunately, it is a gift that no one wants to receive and results in a distorted message that confuses providers and patients. The result may be the discontinuation of hormonal contraceptive methods, and an increase in unintended pregnancies, abortions, and pregnancy-related venous thrombosis!
Although it is well-established that combined hormonal contraception (CHC) is associated with an increase in the risk of VT, the absolute risk of an event remains small in most otherwise healthy young users.1 The risk is higher in obese women and in those with inherited thrombophilias. The importance of the interaction with obesity must be emphasized, since the proportion of obese women is growing in our population.
We also know that the risk is related to estrogen dose, and that the risk with all CHC methods is lower than that seen in pregnancy (where estrogen levels are much higher). But should anyone believe that the contraceptive vaginal ring or patch are actually associated with a higher risk of VT than a low dose LNG pill?
The most recent Lidegaard paper uses the same meticulous methodology of early reports from the same group2 that has linked drospirenone to an increased risk of VT (reviewed in previous issues of OB/GYN Clinical Alert). The database study design allows linkage of prescription data with subsequent events, such that a "retrospective" prospective design is possible. These studies are sometimes called TROHOC (spells COHORT backwards). Unfortunately, these are not true prospective studies. Although the investigator can define the cohorts (in this case by using prescription and health records), there is little control over important variables that are risk modifiers. For VT risk and hormonal contraception, the most important issues are the healthy user effect and prescription bias. Since the risk of VT is highest in the first 6 months after starting a combined hormonal method, these are not insignificant handicaps.2 Also, providers typically see a new product as potentially safer, and this can adversely affect the incidence of events as providers switch less healthy individuals to these newer methods. Only a true prospective study can overcome these biases. Results of large, well-designed, prospective studies in the United States and in Europe (EURAS) have shown no increased risk of VT with drospirenone pills compared to LNG pills or other pills.2 The review by Heinemann and Heinemann provides an excellent discussion of the epidemiology of VT and the biases that emerge in studying this rare event in users of hormonal contraception.1
Since Lidegaard applies the same methodology in each study, it is not surprising that results from this study lead us in exactly the same direction as his previous work. Imagine an explorer with a flawed compass that reliably reads 20 degrees east of true north. Following this compass will lead the explorer on an accurate course in entirely the wrong direction. If you repeated the adventure, starting at the same point and using the same compass, you would find yourself heading in the same wrong direction. Consistency of the results does not imply they are correct. This is the difference between internal and external validity. Methodology matters. Using a better instrument will lead you on the correct course. The results of database studies can be useful when no better data are available, but should be weighted below those of well-designed prospective studies. Repeating a lie (or a misconception) does not make it a truth.
Fortunately, the new Lidegaard paper has received little attention in the American media and among U.S. clinicians, but you can bet the lawyers are interested. My opinion is that there is no clinically important increase in the risk of VT among patch, ring, and implant users compared to users of low-dose LNG pills. We can look forward to publication of another great prospective study by Jürgen Dinger and the ZEG Institute that will help clarify this. This paper was presented at the 2012 ACOG meeting.3 The "Transatlantic Active Surveillance on Cardiovascular Safety of NuvaRing" (TASC) study prospectively enrolled 33,704 new users (starters, switchers, or restarters) of the vaginal ring or combined oral contraceptives in the United States and five European countries and followed them for 2-4 years. Self-reported clinical outcomes were systematically validated. The preliminary results showed no difference in the crude (0.9; CI 0.5-1.8) or adjusted (age, body mass index, duration of use, family history) (0.8; CI 0.4-1.6) hazard ratio for the vaginal ring vs combined oral contraceptives.
The last Lidegaard paper led to FDA hearings and a labeling change for drospirenone pills. The new package insert is actually helpful, as it presents data from the database and prospective studies, but few clinicians and patients will take the time to evaluate this information. Let's hope that the FDA checks its compass, and shows more prudence in the response to this current study.
- Heinemann K, Heinemann LA. Comparative risks of venous thromboembolism among users of oral contraceptives containing drospirenone and levonorgestrel. J Fam Plan Reprod Health Care 2011;37:132-135.
- Dinger JC, et al. The safety of a drospirenone-containing oral contraceptive: Final results from the European Active Surveillance Study on oral contraceptives based on 142,475 women-years of observation. Contraception 2007;75:344-354.
- Dinger JC, Pineda AA. Risk of VTE in users of an etonogestrel-containing vaginal ring and combined oral contraceptives. Abstract presented at the 61st Annual ACOG meeting; 2012.