We appreciate the interest and discussion from Drs Chandra and Williamson regarding our recent publication. Overall, their comments complement the findings of our study.

The only point of debate, really, is that our study found anticoagulant use to be a risk factor for intraocular (not exclusively vitreous, as they mention) hemorrhage on the first postoperative day. They nicely outline the literature with the findings that publications have appeared on both sides of the argument. The problem, of course, is that all of the studies have different methodologies. None (including our own) have an ideal study design, and published studies defy definitive comparison because of many of the following: the frequency of visually consequential events is low, the proportion of patients with various diagnoses varies, the degree of impeded clotting is not standardized, and the criteria of hemorrhage and the observation time point varies. They describe what all clinicians intuitively believe and the studies with subgroup analyses have corroborated: that conditions characterized by more vascular manipulation (proliferative diabetic retinopathy and, to a lesser extent, rhegmatogenous retinal detachment) have higher event rates, especially with anticoagulants. Strengths of our study include the relatively larger cohort, sampling over different intervals, and multivariate analysis. The ideal prospective, randomized study would seem impossible to undertake for this condition.

All investigators have shared the same conclusion that the risk of substantial ocular morbidity from bleeding after vitreoretinal surgery in patients who are also undergoing antiplatelet or anticoagulation treatment is low. As our study found, the frequency of antiplatelet use is increasing, and most vitreoretinal surgeons have come to realize that their use may not always rest on definitive, evidence-based indications, implying that their discontinuance is less risky. Nevertheless, the ophthalmologist should work in concert with the patient’s medical doctor to determine when the risk of withholding such medications presents an undue risk of much higher systemic morbidity. We believe that proceeding to surgery in such cases can be pursued safely in the vast majority of patients.