Introduction

Retinal artery occlusion (RAO) results in sudden painless visual loss often leaving a visual acuity of counting fingers in case of central RAO. RAO is often associated with critical cerebrovascular and cardiovascular disease that may require systemic treatment.[1, 2]

RAO shares vascular risk factors with stroke and is caused by various etiologies.[3] Ophthalmic artery (OA) stenosis or occlusion can cause retinal ischemic symptoms in patients, and this can be a treatable target vessel with intra-arterial treatment.[4] However, no studies showing etiologies of RAO have investigated the details of OA status through intra-arterial evaluations such as transfemoral cerebral angiography (TFCA). In addition, risk of recurrent vascular events according to their etiologies has not been investigated yet.

Therefore, the aim of this study was to investigate the etiologies of RAO and the occurrence of subsequent vascular events according to their etiologies in the year following a RAO.

Materials and methods

Results

Over the 10 years, 200 patients with RAO were evaluated. Among them, 49 patients refused admission and detailed evaluations for risk factors and etiology of RAO were not available. Finally 151 consecutive patients with RAO were enrolled in this study (S1 Fig). RAO patients included 119 (78.8%) central RAOs and 32 (21.2%) branch RAOs with a mean age of 60.8 y (range, 14–92 y). TFCA or IAT was performed in 80 RAO patients. In terms of vascular risk factors, 87 of the 151 patients (57.6%) had hypertension, 35 (23.2%) had diabetes, and 35 (23.2%) had hyperlipidemia. Sixteen patients (10.6%) already had stroke and transient ischemic attack (TIA) before RAO occurrence. Atrial fibrillation or valvular heart disease was present in only 8 patients (6%) (Table 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Demographics of retinal artery occlusion.

https://doi.org/10.1371/journal.pone.0177663.t001

With regard to etiology, 62 patients (41.1%) had LAA, the most common etiological factor. About 40% of our patients had a RAO of undetermined etiology. There was a difference in the proportions of TOAST classification and the type of RAO between the TFCA group and the MRA group. In the TFCA group, the primary etiology was LAA in 48.8% (39/80) of patients, a steno-occlusive lesion of the OA was present in 20% (16/80) of patients, and occlusive disease limited to the OA was present in 8.8% (7/80) of patients. On the other hand, undetermined etiology was reported as a main etiology in the MRA group (Table 2 and S1 Table).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Etiological subtypes of retinal artery occlusion.

https://doi.org/10.1371/journal.pone.0177663.t002

After occurrence of RAO, 18 of 151 (11.9%) patients were lost to follow-up within one year. One-year rate of occurrence for subsequent vascular events was 9.9% (14 events). Ischemic stroke occurred in 13 patients (1-year rate, 8.6%) and no patients experienced hemorrhagic stroke. Non-vascular death occurred in one patient (0.7%), which was censored for the analysis. Over three fourths of ischemic strokes occurred ipsilateral to the RAO. Details of the vascular events occurring during the one year follow-up are presented in Table 3. The rate of vascular events was nearly four times higher in RAO patients with LAA (hazard ratio 3.94, 95% confidence interval 1.21–12.81) than in RAO patients without LAA (Fig 1). In the Cox proportional hazard model for predicting vascular events, LAA was the only independent risk factor for vascular event (Table 4). Ten vascular events occurred in RAO patients attributed to LAA and 4 occurred in undetermined etiology. No events occurred in the cardioembolism, other determined and incomplete etiologies. With regard to vascular event according to period after RAO occurrence, 8 (57.1%) events occurred within 30 d of the RAO, three (21.4%) occurred between 30 and 90 d, and three (21.4%) occurred between 91 and 365 d.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Vascular events in patients with retinal artery occlusion.

Kaplan-Meier curves of vascular event according to period in overall subjects; days 1–30, days 31–90, and days 91–365 (A) and according to etiologies; RAO with LAA vs RAO without LAA (B) within 1 year after retinal artery occlusion occurrence. RAO, retinal artery occlusion; LAA, large artery atherosclerosis.

https://doi.org/10.1371/journal.pone.0177663.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. One-year clinical event rate.

https://doi.org/10.1371/journal.pone.0177663.t003

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Independent variables associated with vascular events by the Cox proportional hazard model.

https://doi.org/10.1371/journal.pone.0177663.t004

Discussion

This study demonstrated that LAA was more frequently associated with RAO than other etiologies and RAO patients with LAA had a nearly four times higher risk of vascular events compared to those without LAA.

The positive association between retinal stroke or emboli and a subsequent vascular event has been presented in prior studies.[7–10] In our study, the rate of vascular events and strokes in the first year following a RAO was 9.9% and 8.6%, respectively. Our results are similar to those of previous studies reporting a range of 7.8–26%, although these prior results were based on heterogeneous study populations.[7, 9–12] This result was also similar to those of studies using an ischemic stroke cohort.[13]

RAO patients attributed to LAA had about a four times higher risk for the occurrence of vascular events during the one year follow-up after RAO than those with other etiologies. There has been no study focused on recurrent risk of subsequent vascular event according to RAO etiologies. However, Helenius et al. showed that approximately one out of every four patients with monocular visual loss caused by ischemia in the retinal circulation had a concurrent acute brain infarct with diffusion-weighted imaging and their probability was higher in patients with an identified etiology than those with an undetermined one. Major subtype of identified etiology was LAA. [14] Lovett et al. reported risk of recurrent stroke by etiologic subtype after ischemic stroke occurrence and clearly demonstrated that the rate of recurrent stroke was also highest in ischemic stroke patients attributed to LAA.[15]

Over half of subsequent vascular events occurred within one month of the RAO event. Moreover, in most patients the ischemic stroke occurred ipsilateral to the RAO. Previous studies have reported that the incidence of experiencing a subsequent stroke was highest in the first month after retinal ischemia.[7, 11, 12, 16] Considering these findings, early treatment and appropriate interventions, such as carotid endarterectomy or carotid stenting, may be carefully considered for the prevention of a second vascular event.

Our study provided data regarding the incidence of steno-occlusion of the OA, in addition to carotid artery disease, using a protocol based on TFCA. The main etiology presented a contrast between two groups according to receiving TFCA or not: LAA in the TFCA group vs. UD-negative in the MRA group. Sixteen patients (20%) of TFCA group had steno-occlusion of the OA and seven patients (8.8%) had pure steno-occlusion of OA, without evidence of carotid artery atherosclerosis. Babikian et al. reported that the most common identifiable etiology (22.1%) of RAO in American subjects was stenosis of the extracranial carotid artery and OA stenosis was reported in only 2.6%.[11] However, Asians have a higher incidence of intracranial atherosclerosis than Caucasians. Anatomically, the retinal artery arises from the OA, the first intracranial branch of the internal carotid artery. Our study implies that steno-occlusive lesions of the OA are an important, but less often recognized etiology, especially in Asian RAO patients. Several studies have reported pure steno-occlusion at the origin of the OA, with no evidence of carotid stenosis, in patients with transient mono-ocular blindness.[4, 17, 18] The OA may be a potential therapeutic target vessel in patients with an abrupt decline of vision from reduced retinal perfusion due to steno-occlusion of the OA.[4] However, adverse reactions related with endovascular treatment of intracranial arteries should be carefully considered in clinical decision making.

Systemic vascular risk factors for RAO overlapped with those for ischemic stroke. Hypertension was the most common vascular risk factor in RAO patients followed by diabetes mellitus and hyperlipidemia. In previous studies, the vascular risk factors that predispose patients to atherosclerotic disease were present in a larger proportion of patients with RAO than that seen in the general population.[8, 11, 19] Cardioembolism as an etiology, such as atrial fibrillation and valvular heart disease, was present in 6% of RAO patients. Leisser et al. reported about 4.3% of patients with vegetations on the cardiac valves and 10.3% of patients with other reasons for possible cardioembolism diagnosed with TEE. Atrial fibrillation was 9.4%.[20] These finding was similar to our results.

Major vascular risk factors for RAO are modifiable and can predispose patients to atherosclerotic disease. Most of subsequent vascular events occurred within 1 month after the RAO occurrence. However, only one third of ophthalmologists will immediately evaluate patients with an acute RAO; most ophthalmologists send patients to the emergency room for inclusion in clinical trials.[21] Although we utilized a clinical pathway for immediate evaluation in patients with RAO, about one in four RAO patients (49/200) refused any evaluation for risk factors and etiology, and 11.9% were lost to follow-up at one year. An effort to raise awareness of RAO among patients, as well as physicians, is necessary.

This study had several limitations. First, only 151 patients were included, which is not a sufficient number for a statistically significant evaluation. However, many patients underwent a more detailed evaluation of the OA by TFCA, allowing a comparison of the two groups regarding etiology. Second, TOAST classification based on stroke etiology has limitation about causal relationship. Leisser et al. reported about a probability of 12.6% or lower for clots to be washed into the OA, which means that at least 87.4% of clots in the carotid artery are washed into the brain.[22] Based on this finding, it is unclear whether RAO is caused by LAA or other reasons in even RAO patients with LAA. Nevertheless, it is definite that LAA is a common comorbidity in RAO patients. Third, our patient population was Asian; as such, our results should be interpreted with caution. These finding need to be validated among other ethnic cohorts before our findings are generalizable. Finally, many patients with RAO were lost to follow-up at one year, possibly because of a lack of insight into the disease by both patients and physicians.

Because RAO patents have a high risk of a subsequent vascular event, an evaluation for risk factors and the etiology of the RAO should be performed. LAA is a major etiology for the occurrence of subsequent vascular events and prompt aggressive management should be instituted to decrease the risk of a subsequent vascular event. Further investigations of the etiologies and vascular events associated with RAO in larger ethnically diverse populations are necessary.

Supporting information