(This study was commissioned in conjunction with both the FDA and Johnson & Johnson, makers of the contraceptive patch, but researchers had full control over the data and results and were not required to consult with Johnson & Johnson when reporting findings.)
There were 49,048 woman-years of exposure to the patch and 202,344 woman-years of exposure to the pill, with an incidence of VTE of 40.8 and 18.3 for every 100,000 woman-years in patch and pill users, respectively. The incidence of AMI was 6.1 and 3.5 for every 100,000 woman-years in patch and pill users, respectively. No ischemic strokes were noted in patch users.
The adjusted incidence ratio for VTE in patch users compared with pill users was 2.2, and for AMI it was 1.8. Following publication of this study, the FDA issued a statement in January of this year that women using the patch face an increased risk of VTE, compared with women using the pill. Package labeling was changed to reflect this heightened risk.
Reasons for different findings
The studies by Jick and colleagues and Cole and associates present very different findings. The studies by Jick and colleagues give the impression that there is no increased risk of VTE in patch users compared with pill users, but the studies have significant flaws. First, Jick and colleagues do not confirm the diagnosis of VTE in the medical record. This is particularly problematic because the reported number of pulmonary emboli (PE) is very high, compared with the number of deep vein thromboses. The 2006 study found 42 cases of PE and only 26 cases of deep vein thrombosis. Because the latter is more common than PE, this could indicate that deep vein thrombosis was underdiagnosed.
Another shortcoming is that Jick and colleagues included only nonfatal thromboembolic events, which may mean that they missed many cases of fatal VTE because they were not looking for this information. The inclusion of new initiators only also may have skewed the data. This would mean that former users of an OC may have been included in the patch group but were ineligible for inclusion in the pill group. This may bias the data toward experienced hormonal contraceptive users in the patch group, thereby falsely lowering the VTE rate.
The study by Cole and associates also has limitations. It included long-term users of hormonal contraceptives in both the patch and the OC groups, which may bias the data toward lower rates of VTE, AMI, and stroke for the same reasons cited above. One would assume that this bias was corrected, because prior use was allowed in both groups, making the bias equally distributed, but there is no way to confirm this with any degree of certainty.
All three studies have some flaws in common
All three studies used prescription information to determine exposure, but there is no guarantee that the women who filled the prescriptions actually used the agents. Patients given drug samples by their clinicians were overlooked because these samples are not tracked through pharmacy data.
Because the data were collected from insurance claims information of privately insured patients, it is impossible to generalize these findings to the general population. We cannot use the findings to determine whether the same results would be seen in uninsured women or women insured through nonprivate programs such as Medicaid or the Veterans Administration.
So what’s the bottom line?
Health-care providers should be cautious about citing these studies as “evidence” when advising patients about the risk of VTE while using the patch. The twofold increased risk of VTE observed in patch users and the almost twofold increased risk of AMI observed by Cole and associates cannot be completely ignored, however, particularly because this study was better designed than those by Jick and colleagues.
It is more important to remember that the incidence of VTE in patch users is extremely low. If a patient has been using the patch, is happy with the method, and has had no adverse effects, there is no reason, based on these findings, to discontinue it. When counseling new initiators, the best that can be done is to explain the potential risks and side effects associated with the method and allow the patient to make an informed choice using the information that is available.
If the increased risk of VTE is accurate, it would still be equal to or lower than the risk during pregnancy. A recent review found the overall incidence of VTE in pregnancy or the postpartum period to be 200 for every 100,000 woman-years.7