Initially, all patients underwent neck dissection in order to expose the cervical ICA. The orbito-zygomatic approach (OZA) was used in most patients. The anterior clinoid process (ACP) was removed extradurally and/or intradurally, and the sphenoid ridge was flattened until the lateral edge of the superior orbital fissure (SOF) was reached. The roof of the optic canal and most of the optic strut was removed for adequate exposure of the proximal neck of the aneurysm and for intracranial proximal control of the ICA. All paraclinoid ICA aneurysms were completely obliterated by direct clipping without bypass. Microsurgical clipping was considered if preoperative angiogram showed that the ophthalmic artery was incorporated from the aneurysm neck, aneurysm had a broad-neck, or if an endovascular approach was otherwise technically not feasible.

All procedures were performed under general anesthesia. Heparinization was only started in unruptured aneurysms with a loading dose of 3000 IU and was maintained with a dose of 1000 IU on an hourly basis after femoral puncture and femoral sheath placement. Using a guiding catheter, microcatheters with proper shapes were introduced over micro-guidewires and navigated into the aneurysms. Aneurysms were packed as densely as possible with coils. In wide-neck aneurysms, multiple catheters, balloons, or stents were used for safe coil deployment. Immediate postoperative and follow up angiographic results were classified as total occlusion, neck remnant, and body filling.31) Endovascular treatment was recommended for patients with significant medical comorbidities or age > 75 years. In this situation, if the aneurysm was amendable to both surgical clipping and endovascular treatment, endovascular treatment was considered under agreement from the patients or their relatives.

The Wilcoxon rank sum test and Fisher's exact test were used for analysis of differences between the two groups. A p value < 0.05 was considered significant. Analyses were performed using statistical analysis system version 9.2 (SAS Institute Inc., Cary, NC, United States).

In the clipping group, the mean age was 49.77 ± 10.84 years, and the mean aneurysm diameter was 6.17 ± 5.09 mm in the greatest dimension (range 3-20 mm), except one patient with a giant aneurysm of 60 mm in the greatest dimension. Twenty four patients were female, and six patients were male. SAH with a ruptured paraclinoid ICA aneurysm was observed in 11 patients. Except for two patients, the initial neurologic examination showed a favorable clinical state (mean GCS was 14.43 ± 1.48) in 28 patients. After surgical clipping in 30 cases, six patients had hydrocephalus, and five had vasospasm. Among six hydrocephalic patients, two cases of hydrocephalus occurred in unruptured cases. One showed obstructive hydrocephalus due to a giant aneurysm, and the other had communicating hydrocephalus due to previous SAH associated with anterior communicating artery aneurysm rupture. Transient hemiparesis and ptosis occurred in four patients, but showed spontaneous recovery within one month. One patient had permanent motor weakness, and two patients had permanent visual complications.

In the coiling group, the mean age was 58.16 ± 10.61 years, and the mean aneurysm diameter was 5.45 ± 2.50 mm (ranged 3-13 mm). Twenty eight patients were female, and their mean GCS was 14.90 ± 0.30. Only one patient had SAH due to a ruptured aneurysm. After endovascular coil embolization, none of the patients had hydrocephalus or vasospasm (Table 2). Cerebellar hemorrhage was found in one patient, which resolved without neurological sequelae. Two patients had transient hemiparesis.

No statistical significance regarding aneurysm size, sex, GCS score, H-H grade, and mRS was observed between the surgical clipping and endovascular coil groups (Table 2). However, this statistical comparison is not valuable, because only one patient with SAH underwent endovascular coiling. Therefore, because the clipping group did not reach a sufficient number for statistical analysis, we performed subgroup frequency analysis in unruptured aneurysm cases (Table 3).

Among 61 patients, two patients died after either surgical clipping or endovascular coiling. One patient was a 58-year-old woman who had an unruptured paraclinoid ICA aneurysm and a superior cerebellar artery (SCA) aneurysm. Her GCS score was 15, and the paraclinoid ICA aneurysm measured 13 mm in its greatest dimension. Stent-assisted coil embolization was performed, and the patient recovered without neurological problems. The next day, the patient was transferred to the general ward, however, sudden mental deterioration was observed seven days after embolization. Follow-up brain CT was performed, and SAH with a massive ICH was detected in the left cerebellar hemisphere. Because of antiplatelet agent usage, the amount of hematoma was much greater than expected. Her family members refused further intervention, and she died nine days after embolization. Another patient, a 51-year-old female, suffered SAH due to rupture of a left ophthalmic artery aneurysm, with a GCS score of 9 and Fisher and H-H grades of IV. Emergent surgical clipping of the aneurysm and decompressive craniectomy were performed in order to relieve severe brain edema. She was treated in the intensive care unit with intracranial pressure (ICP) control, but brain edema and severe vasospasm progressed. She died eight days after surgery.

Paraclinoid ICA aneurysms have imposed a significant technical burden on neurovascular surgeons. If possible, direct obliteration of the aneurysm is the most preferable treatment available. Before the development of microsurgical cranial base approaches, the ACP hindered the surgeon's view and thus restricted the angle of attack used in treatment of the aneurysm. Removal of the ACP allows for at least 6 mm of additional exposure of the distal ICA without entering the true CS.26) This technique has improved the approach to paraclinoid ICA aneurysms for direct surgical clipping.

Paraclinoid ICA aneurysms have different anatomical relationships with the dural folds and bony part of the cranial base and their surgical treatment requires specific surgical techniques such as exposure of the ICA in the neck, extra- and intradural ACP resection, ICA exposure of the clinoidal segment, opening of the distal dural ring, and hemostasis on the CS.5)

Raco et al., who studied the long-term efficacy of surgical clipping in treatment of ophthalmic segment aneurysms, 28) reported that clipping is a complete treatment for ophthalmic segment aneurysms with an acceptable morbidity rate (5.8%). In other literature, favorable outcomes (grades 4 and 5 of the GOS) ranging from 58.7% to 96.4% have been reported for surgical treatment of paraclinoid ICA aneurysms.12)3)6)8)9)12)15)16)18)22)24) The outcome in the current series was comparable to that reported in the published data. The results did not show a statistically significant difference when compared with the coil embolization group (Table 2).

Because of the complex anatomy around the aneurysms and the characteristic risks associated with microsurgery, endovascular techniques have been popularized for treatment of paraclinoid ICA aneurysms.10)14)17)30)33) Endovascular embolization for paraclinoid ICA aneurysms has emerged as a major therapeutic option in selected patients, with comparable morbidity and mortality rates to surgery, as well as markedly reduced hospital stays.26)36)

Endovascular treatment is specifically suited for patients with the following conditions: a high medical risk for surgical clipping, worsening clinical grade, an aneurysm with no mass effect to be relieved or extension deep into the cavernous sinus, vasospasm before diagnostic angiography, and those who choose endovascular treatment.5)34) In this study, a statistically significant difference in age was observed between the clipping and coiling groups, which, as mentioned earlier, is expected due to the increased surgical risk associated with old age. In addition, it is also the reason that most elderly people are reluctant to undergo open surgery.

In a previous study, transient morbidity ranged from 0% to 4.8%, and permanent morbidity ranged from 0% to 3.5%.11)21)23)35) In the current study, most patients had good outcomes (97% with grades 0 to 2 of the mRS). However, for large or giant paraclinoid ICA aneurysms, dense coiling is difficult to achieve and can result in recanalization of the aneurysm. Aneurysm size larger than 10 mm is a significant predictor of recurrence, and, therefore, long-term angiographic follow-up is mandatory.29)

Overall transient or permanent postoperative complications have been reported at rates between 21.3% and 41.7% for surgical treatment of paraclinoid ICA aneurysms and from 27.3% to 50.0% for patients with giant aneurysms.1)6)8)9)15)18)22)24) The most frequent postoperative deficits or worsening status reported in the literature are attributed to compromise of the optic and oculomotor nerves. These complications are rarely seen in coil embolization; therefore, these low complications might be one of the advantages of coil embolization. Colli et al. reported vasospasm in 20.1% of patients with ruptured aneurysms and hydrocephalus in 4.8%.5) In this literature, routine opening of the lamina terminalis in aneurysm surgery might be attributed in part to the relatively low incidence of hydrocephalus.5)

In the current study, vasospasm (17%) and hydrocephalus (20%) were observed only in the clipping group. A statistically significant difference was observed between the clipping and coiling groups, however, it was believed to result from the discrepancy in the number of patients with aneurysm rupture in each group. In the clipping group, 11 of the 30 patients had ruptured aneurysms, but only one of 31 patients in the coiling group had a ruptured aneurysm. In this study, the clinical outcomes of microsurgical clipping for large or giant paraclinoid ICA aneurysms were favorable. Therefore, microsurgical clipping can be considered as the first-line treatment option in the case of large and/or broad-neck paraclinoid ICA aneurysms.

First, the percentage of patients with ruptured aneurysms differed significantly between the groups, which might indicate a selection bias. Second, if we further divide patients based on location of the aneurysm, information about complications would provide more comprehensive data. Finally, future studies should include long-term, prospective allocation of all the available data in order to overcome the shortcomings encountered in the current study.