I read with interest the article by Okoroh et al, regarding the association between venous thromboembolism (VTE) and polycystic ovary syndrome (PCOS). The authors concluded that oral contraceptive use might be protective against VTE. In my opinion, this is a controversial study. Recent meritorious studies report 2.0-fold increased risk of VTE among women with PCOS who were taking oral contraceptive pills (OCPs), and a 1.5-fold increased risk among those not taking OCPs. Physicians should be aware of the increased risk of VTE when prescribing OCPs to women with PCOS.

OCP use was identified as ≥1 filled OCP prescription during the study period. This approach may lead to a significant bias as there is no assurance that the prescription was taken regularly, and young women often experience OCP side effects and discontinue use. Reports and some important studies show that the OCP discontinuation rate may be as high as 62% among young women at a 6 months interval. The current study does not include a time interval between OCP use and the VTE event. Thus, we cannot definitively claim that there is a protective effect of OCP against VTE.

Data derived from recent studies related to VTE among users of OCPs in PCOS are of borderline statistical significance because important factors such as dose, duration of OCP use, coexistence of other VTE risk factors, and other prescription bias are not considered.

Okoroh et al, have not mentioned factors likely to contribute to the decreased risk of VTE among OCP users in PCOS. The risk of VTE resulting in hospitalization is not as rare as estimated. It has been confirmed by a well-designed study that the incidence of VTE among young non-OCP users was very low compared to OCP users and that a significant part of OCP-related VTE might have been avoided. From this point of view one can understand the importance of a take-home message about OCP use in PCOS being protective or harmful. Currently available data on the correlation between OCP and VTE risk are not conclusive, and other randomized, controlled, well-designed, and prospective studies in a large population are needed to answer this debating point.