We appreciate the interest of Lazzeri and associates in our recent article entitled “Iatrogenic Retinal Artery Occlusion Caused by Cosmetic Facial Filler Injections,” published in the American Journal of Ophthalmology . We apologize if our article was not clear enough to readers; through this correspondence, we would like to clarify various points.

First, Lazzeri and associates had a concern over whether the short follow-up (3 days) of Patient 6 was adequate to allow conclusions about this patient’s outcome. For standard cases of central retinal artery occlusion (CRAO), we agree with their opinion. Herein, we provide outcomes for Patients 1, 4, and 6, for whom longer follow-up durations of 302, 605, and 405 days after the event, respectively, apply. In all of these patients, their final vision was still classified as no light perception. All 7 patients with ophthalmic artery occlusion caused by facial filler injections had a final visual acuity of no light perception after 6 months, which is enough time for defining final vision. Thus, the initial presentation of blindness caused by filler-associated ophthalmic artery occlusion should be regarded as being nearly definite, and ophthalmic artery occlusion should be differentiated from CRAO, which usually is associated with a remnant temporal visual field. The term relative afferent pupillary defect that we used for Patient 6 (Table 1) is a neuro-ophthalmologic term for the asymmetric pupillary response to light, and is not a term for the degree of pupillary defect. In fact, Patient 6 showed no pupillary response to light at presentation. Because Patients 1 through 4 were in a state of pharmacologic pupillary mydriasis at the initial visit, we could not evaluate the relative afferent pupillary defect in these patients.

Second, regarding the intra-arterial thrombolysis mentioned in our study, we can clarify the interval between the onset of symptoms and intra-arterial thrombolysis therapy in 4 patients: 2.5 hours in Patient 1, 2.5 hours in Patient 2, 3 hours in Patient 3, and 4 hours in Patient 4. A treatment interval within 24 hours is included in the inclusion criteria for intra-arterial thrombolysis for CRAO in our hospital, as mentioned in the Methods section. As Lazzeri and associates pointed out, intervention within 90 minutes after symptom onset is difficult to conduct, because most patients are transferred from clinics or other hospitals. More important than the timing of treatment was that compared with the usual successful reperfusion achieved in noniatrogenic CRAO patients (from our unpublished data), intra-arterial thrombolysis could not restore retinal perfusion in patients with filler-associated ophthalmic artery occlusion. Thus, intra-arterial thrombolysis should not be considered as a therapeutic option for these patients currently.

Third, Lazzeri and associates pointed out that intra-arterial thrombolysis has significant side effects including cerebral bleeding. In the EAGLE (European Assessment Group for Lysis in the Eye) trial, the incidence of major complications, such as cerebral infarct or hemorrhage, was not different between the thrombolysis and conservative treatment groups. To obtain the exact risk of major complications of selective intra-arterial thrombolysis for CRAO, data from a larger sample size are required. From our experience, we believe that selective intra-arterial thrombolysis has a risk of cerebrovascular complications comparable with that of standard transfemoral cerebral angiography.

Fourth, vision improvement in patients with retinal artery occlusion caused by facial fillers should be differentiated according to the obstruction site and the extent of reperfusion. It is natural that most CRAO patients experience visual improvement as retinal reperfusion occurs in the early period and functional improvement occurs in the reversibly damaged retina. In patients with ophthalmic artery occlusion resulting from cosmetic fillers, there was no retinal reperfusion in the early period (usually within 1 month), which we thought was the major reason for the absence of visual improvement. However, in cases of CRAO and branch retinal artery occlusion caused by facial filler injections, there was partial visual improvement (as in noniatrogenic retinal artery occlusion). This indicates that obstruction by filler emboli was less severe than in ophthalmic artery occlusion. However, we cannot understand the difference between the 6-step therapy in the EAGLE trial mentioned by Lazzeri and associates and that used in our study. Intra-arterial thrombolysis is always performed as a single-step emergent therapy to restore retinal perfusion, and early intervention is the most important factor.

Lazzeri and associates conducted a good systematic review of cases with iatrogenic retinal artery occlusion caused by facial filler injections and suggested some precautionary measures to be taken before and during injection procedures (Figure 2 in their article). We agree that their suggestions could be helpful to reduce the risk of blindness resulting from facial filler injections; however, to date, there is no evidence to support their suggestions. Most important of all may be that patients, as well as doctors, should understand the possibility of this serious complication before the procedure.