Data collection and analysis reported in this study were approved by the Institutional Review Board of our hospital. All patients diagnosed with OphA aneurysms using digital subtraction angiography (DSA) at our hospital between January 2010 and August 2018 were retrospectively reviewed. Aneurysms clearly associated with the OphA that could be clearly distinguished from other paraclinoid aneurysms were selected for this study. Only those patients who underwent treatment were included in the analysis and were divided either into the clipping group or the coiling group based on the treatment method. The treatment decision was made by each patient after discussion with neurointerventional physicians and neurosurgeons, based on their own characteristics. The data of patient characteristics, aneurysm features, procedural factors, aneurysm occlusion outcomes, post-treatment visual deficits, and clinical outcomes for the 2 groups were collected for comparison and analysis.

Before the treatment, all aneurysms were evaluated by DSA in both groups, and a wide-neck aneurysm was defined as an aneurysm with a neck width >4 mm or dome-to-neck ratio <1.5. The direction of the aneurysm was classified as superior, medial, or inferior based on the angular relationship between the aneurysm and the ICA as evaluated from the anterior projection on 3D rotational angiography. The relationship between the OphA and aneurysm was classified as “neck” or “sac” based on whether the OphA originated from the aneurysmal neck or was incorporated into the aneurysmal sac [8].

Immediate post-procedural aneurysmal occlusion status and preservation of antegrade flow of the OphA were evaluated on the basis of the final angiogram recorded during the procedure in the coiling group and on the basis of the follow-up DSA, usually 1–2 days post-surgery, in the clipping group. Imaging follow-up by either DSA or computed tomographic angiography in the clipping group and by either DSA or magnetic resonance angiography in the coiling group was scheduled at 6–12 months post-surgery. The aneurysm occlusion status in the coiling group was evaluated by the Raymond–Roy classification. Raymond Class III in the coiling group and residual sac in the clipping group were regarded as incomplete occlusion of the aneurysm. Complete occlusion of OphA was defined as disappearance of previously antegrade blood flow, as seen on ipsilateral ICA angiography, while near OphA occlusion was defined as significantly slow yet antegrade blood flow.

Visual outcomes were carefully assessed for all patients before and after treatment. Visual deficit included any worsening of visual acuity, visual field defects, and blindness. In cases with pre-treatment visual deficit, the change of pre-treatment visual deficit was evaluated and classified into (1) improved and (2) unchanged. Post-treatment visual complications were defined as any newly developed visual deficit after the treatment.

Clinical outcomes were evaluated 1 month after the patient’s discharge using the modified Rankin Scale (mRS) and were then classified as good (mRS score 0–2) or poor (mRS score 3–6).

For convenient analysis, each patient with OphA aneurysms on both sides of the ICA was considered as 2 patients of the same age and sex but with OphA aneurysms on different sides. SPSS 24.0 (IBM Corp., Armonk, NY, USA) was used for all statistical analyses, and the significance level for the analyses was fixed at 0.05. Quantitative variables were expressed as mean±standard deviation and were compared using an independent-sample t-test (normally distributed) or a ranksum test (abnormally distributed). Chi-square test or Fisher’s exact test was performed to evaluate covariates for the binary categorical dependent variables. Logistic regression analysis using the forward stepwise method was used to evaluate the association between newly appeared visual deficits and suspected factors, and only factors with a P-value <0.2 in the univariate analysis were included in the multivariate analysis.

In total, 267 OphA aneurysms in 264 patients were confirmed in our hospital between January 2010 and August 2018 using DSA. Among them, 56 aneurysms in 55 patients treated by microsurgical clipping and 82 aneurysms in 80 patients treated by coiling were included in the analysis. No aneurysm was ruptured. The baseline characteristics of the 135 patients are shown in Table 1.

The features of the patients and OphA aneurysms in the 2 groups are summarized in Table 2. There were no differences between the 2 groups in terms of baseline and aneurysm characteristics. Among the patients treated by coiling, a stent was used in 52 patients.

Immediate incomplete aneurysm occlusion was confirmed in 12 (21.4%) patients who underwent microsurgical clipping and in 18 (22.0%) patients who underwent coiling. No difference was found between the 2 groups in terms of immediate incomplete occlusion (P=0.942). Imaging follow-up data were available for 45 (80.4%) aneurysms treated by clipping and 55 (67.1%) aneurysms treated by coiling. The median follow-up time was 13.1 months in the clipping group and 19.0 months in the coiling group. At the latest follow-up, incomplete occlusion was found in 24.4% (11/45) of aneurysms treated by clipping and in 23.6% (13/55) of aneurysms treated by coiling (P=0.925). There was no incomplete occlusion among patients with total or near OphA occlusion in the coiling group.

There were 16 patients (9 in the clipping group and 7 in the coiling group) with pre-treatment visual deficits. The 9 patients in the clipping group presented with decreased visual acuity in 6 (66.7%) and visual field defect in 3 (33.3%) before treatment. After clipping, the deficits improved in 5 (55.6%) patients and were unchanged in 4 (44.4%) patients. The 7 patients in the coiling group presented with decreased visual acuity in 5 (71.4%) and visual field defect in 2 (28.6%) before treatment. After coiling, the visual deficits improved in 3 (42.9%) and were unchanged in 4 (57.1%) patients. No patient showed aggravation of the pre-existing deficits in either group. There was no difference in the frequency of visual deficit improvement between the 2 groups (P>0.05).

OphA patency was observed in all patients treated by microsurgical clipping. However, post-treatment visual deficits occurred in 9 patients (16.1%) in this group, and the deficits included decreased visual acuity in 2, visual field defect in 4, both in 2, and monocular blindness in 1. All post-treatment visual deficits occurred in patients who had undergone anterior clinoidectomy (n=50) during clipping.

In the coiling group, total occlusion of the ipsilateral OphA, either intentional or unintentional, was observed in 6 patients, and near occlusion was observed in 3 patients. Intact collateral circulation was confirmed by BOT in 8 of these patients before treatment, and BOT was not performed in the other patient. However, ECA angiography after coiling demonstrated intact collateral circulation in this patient, and no visual deficit was observed. Post-treatment visual deficits occurred in 2 patients (2.4%) in the coiling group. A patient with both visual field defect and reduction in visual acuity was found to show near OphA occlusion on final angiography; however, BOT before coiling showed intact collateral circulation in this patient. The other patient with only a reduction in visual acuity was found to have a well-preserved OphA during the procedure; however, this patient complained of foggy vision after waking up from the anesthesia, and small infarctions were observed on the retina by fundus examination. This patient did not undergo the BOT before the procedure.

The test results of 51 patients who underwent BOT in the 2 groups and their treatment outcomes are shown in Table 3. There were no complications in all 51 patients who underwent BOT and intact collateral circulation was demonstrated in 43 (84.3%) patients. Total occlusion of ipsilateral OphA was found in 5 patients with adequate collaterals during coiling, and they showed no post-treatment visual deficits. A patient with near occlusion of ipsilateral OphA experienced visual deficits after coiling, although intact collateral circulation was confirmed in this patient by the BOT. No delayed visual complications occurred in either group during follow-up. A significant difference in the frequency of post-treatment visual deficits was found between the 2 groups: 16.1% (9/56) in the clipping group and 2.4% (2/82) in the coiling group (P=0.010). Univariate analysis for the incidence of post-treatment visual deficits in all included patients is shown in Table 4. Only microsurgical clipping exhibited a significant difference in the incidence of post-treatment visual deficits (odds ratio 7.660, 95% confidence interval 1.587–36.962, P=0.011) by logistic regression analysis (Table 5). Patients with newly developed visual deficits and their follow-up outcomes are presented in Table 6. During follow-up, the deficits disappeared completely in 2 patients, improved in 7 patients, and remained unchanged in 2 patients.

In the clipping group, premature rupture of a previously unruptured aneurysm occurred in 1 patient, but without neurological deficits after clipping. In the coiling group, 1 patient developed a retroperitoneal hematoma requiring surgery. A delayed ischemic complication involving monoparesis of the upper arm 10 days after discharge occurred in this group. On the clinical follow-up at 1 month post-hospital discharge, all patients, except 1 in the clipping group with pre-existing thalamic hemorrhage, showed excellent outcomes.

The reported incidence of post-treatment visual complications for OphA aneurysms treated by clipping is between 10.6% and 28.5% [2,11], while it is between 0 and 9.3% for patients treated by coiling [3,8,9,12]. Our results are consistent with those of previous studies in which the incidence of newly appeared visual deficits was significantly higher in patients who underwent clipping than in those who underwent coiling (16.1% vs. 2.4%, P=0.010). Multivariate analysis also revealed that only microsurgical clipping resulted in a significant increase in the incidence of post-treatment visual deficits.

Kamide et al. [2] analyzed 208 OphA aneurysms treated by microsurgical clipping and found that anterior clinoidectomy during surgery was associated with the incidence of new postoperative deficits. In our series, all 9 patients with newly acquired visual deficits had undergone anterior clinoidectomy during microsurgical clipping, and 89.3% of patients in the clipping group underwent anterior clinoidectomy, which was much higher than that in their study (75.3%). This could be a possible explanation for the higher incidence of post-treatment visual deficits in our clipping group (16.1% vs. 10.6%). However, during the clipping of OphA aneurysms, anterior clinoidectomy is sometimes unavoidable because complete removal of this structure is required to expose the proximal neck of the aneurysm. Injury to the optic nerve during drilling of anterior clinoid process, ischemia of the optic pathway from perforator occlusion, or direct compression of the optic nerve by an applied clip may occur during this process, and all of these can cause visual deficits [2,13,14].

In our series, 9 patients in the clipping group and 7 in the coiling group had pre-treatment visual deficits. Previous studies have indicated that microsurgical clipping is superior to endovascular treatment in the context of improvement or resolution of pre-treatment visual deficits, due to its ability to simultaneously eliminate compression and pulsation caused by the aneurysm [13,15,16]. However, in our series, the frequency of improvement of pre-treatment visual deficits did not differ between the 2 groups.

A previous study has suggested that coiling of OphA aneurysms is associated with a high rate of incomplete aneurysm occlusion, which was 60.5% on angiography immediately after coiling [2,17]. This could be due to attempts made by surgeons to save the OphA during the procedure, such as intentional partial coiling and avoidance of dense coil packing in the aneurysm. However, the incomplete aneurysm occlusion rate observed immediately or during follow-up in our series was comparable between the clipping and coiling groups. One possible reason for this is that the incomplete aneurysm occlusion rate in our coiling group was low as the BOT was performed to help treatment. This is substantiated by evidence indicating the lack of incomplete occlusion among patients with total or near OphA occlusion.

Previous studies have indicated that the BOT is helpful in predicting vision outcomes after coiling. Kim et al. [8] reported that intact collateral circulation was demonstrated in 26 of 28 (92.9%) patients who underwent the BOT in their series, and that no specific visual symptoms were observed in 3 patients with total occlusion of the OphA after coiling. Ahn et al. [9] reported that an adequate collateral supply was identified in 12 of 14 (85.7%) patients, and that there were no visual deficits in 4 patients with total OphA occlusion. Consistent with the results of these studies, we showed that intact collateral circulation was observed in 84.3% of patients, and there were no visual deficits in all 6 patients with total OphA occlusion. Therefore, we suggest that total occlusion of the OphA with an adequate collateral supply from the ECA, confirmed by the BOT, is safe during coiling for OphA aneurysms.

However, unpredicted thromboembolism events may occur during coiling for OphA aneurysms and can also cause post-treatment visual complications [4,5], which occurred in 2 patients in our coiling group, ie, 1 with good anterograde flow in the OphA and the other with significantly slow anterograde flow in the OphA. Retinal and/or optic ischemia could occur not only due to decreased perfusion of the completely occluded OphA, but also due to the variable degree of thromboembolic phenomenon from the orifice of the OphA during surgical or endovascular manipulation or the presence of any foreign material in the presence of any residual flow. As intentional total occlusion of the OphA is technically not easy with only coil packing, further studies are needed to determine whether partial occlusion of the OphA with an adequate collateral supply could increase the incidence of visual deficits.

The methods used to perform the BOT between studies are different [3,8,18]. Undoubtedly, it is more precise to evaluate the collateral flow from ECA during the BOT if the OphA aneurysmal neck is occluded by a balloon catheter. However, this procedure also carries the risk of rupturing the aneurysm, and passing the balloon over the cavernous segment is frequently difficult. That is why we just put the balloon catheter at the petrous or cavernous segment of the ipsilateral ICA during BOT. Due to the presence of anterior and posterior communicating artery, the rate of real collateral circulation from ECA to OphA might be higher than 83.4% in the present study.

There are several limitations to this study. First, this is a retrospective study, and there could be a selection bias in the choice of treatment modality. Second, the size of the aneurysms included in this study is relatively small, and no giant aneurysm was included, which could mask some risk factors associated with post-treatment visual complications. Third, we did not address the usefulness of flow diverters for the treatment of OphA aneurysms. In recent years, flow-diverter stents have emerged as a promising option in the management of intracranial aneurysms. Di Maria et al. [19] conducted a comparative study between flow diverters and coiling for the treatment of carotid-ophthalmic aneurysms and found that aneurysms treated by flow diverters showed better long-term occlusion outcomes, though procedural complications and visual deficits seem higher in the patients treated by flow diverters [12,19]. However, not all patients with OphA aneurysms could be treated with flow diverters because of the cost or regional policies, and therefore, our result that coiling is safer than clipping for the management of OphA aneurysms is still valuable.