We read with great interest the article by Oh and associates, which adds further debate to the question of the management of antiplatelet and anticoagulant medication in patients undergoing vitreoretinal surgery.

Traditional beliefs suggest that platelets and factors affecting primary hemostasis are more crucial than suppression of clotting factors through anticoagulants in minute surgical trauma such as intraocular surgery. The converse is thought to be true in moderate or severe surgical trauma (such as ocular muscle and lid surgery). However, Oh and associates suggest that anticoagulation is a risk factor for vitreous hemorrhage occurring after vitrectomy, whereas antiplatelet medication poses no increased risk. Most recently, El-Sanhouri and associates showed a similar effect on vitreous hemorrhage associated with posterior vitreous detachment. Although we did not show an increase in hemorrhagic complications in vitrectomy associated with warfarin (in particular, suprachoroidal hemorrhage), we did reveal an increased rate of vitreous hemorrhage in patients with rhegmatogenous retinal detachment associated who underwent warfarin treatment. Conversely, Brown and Mahmoud recently showed no increase in hemorrhage occurring in the perioperative or postoperative periods in diabetics receiving antiplatelet or anticoagulant therapy.

The differences in these findings, we suggest, lies in the causes of vitreous hemorrhage and the possible effects of antithrombotic agents on these causes. When the causes primarily are a consequence of disruption to retinal vessels secondary to posterior vitreous detachment or rhegmatogenous retinal detachment, anticoagulants seem to pose a slightly greater risk than antiplatelets. This would be more akin to the moderate surgical trauma in Pandolfi’s categorization. However, in hemorrhages occurring after vitrectomy in diabetics, the source is unclear, with multiple suggestions. These include fibrovascular proliferation from sclerotomy sites or vitreous base and incomplete removal of fibrovascular tissue. Such vessels may be less vulnerable to anticoagulants, being minor surgical trauma vessels. This may explain the findings by Brown and associates, who excluded patients with rhegmatogenous retinal detachment from their cohort. Moreover, only 15.5% of the cohort presented by Oh and associates underwent vitrectomy for diabetic indications. Most of the indications were either retinal detachment or other. In these cases, it is possible to envisage a disruption at the vitreoretinal interface to retinal vessels. This may explain their findings with anticoagulants. The findings by others thus also may be explained. Larger, more targeted studies are needed to confirm such a concept.

Finally, it is important that Oh and associates emphasized that the increased postoperative hemorrhagic risk of patients taking anticoagulants was very low. Additionally, no patients taking anticoagulants or antiplatelets demonstrated any ocular morbidity. They suggest caution if surgeons consider withholding these medications for vitreoretinal surgery. We suggest that the risks of withholding such medications are far greater than any perceived benefits of reduced complications and add our support to such a caution.