Introduction

   Since its introduction in 1991, endovascular coil embolization has been accepted as an effective method for the treatment of intracranial aneurysms,^5)6)17) and rapid developments in coil technology and aneurysm remodeling techniques have broadened the application of endovascular therapy to an ever increasing range of intracranial aneurysms.³⁾¹²⁾ Several randomized, multicenter trials comparing endovascular and surgical management of ruptured aneurysm have shown significantly improved 1 year outcomes in patients treated with endovascular coil occlusion, although long term durability remains to be clarified.^5)6)17) 
   There is, however, a general consensus that aneurysm recurrences are more frequent after endovascular treatment than in surgical clipping.^6)10)17) Aneurysm recurrence may occur as a result of incomplete occlusion, compaction of the coil-thrombus complex, or de novo regrowth, which raises concerns about the efficacy of coil embolization to protect aneurysm regrowth and subsequent possible rupture in the long-term.^{1)3)8)9)11)13)} The clinical significance of this important drawback of endovascular treatment are still recruited, and a growing knowledge to identify a target population with lesions at higher risk of recurrence or a subpopulation of patients in whom recurrences are unlikely should be more clarified. 
   The authors attempted to identify the risk factors associated with angiographic recurrences after endovascular coil embolization in retrospective manner in patients with at least 2 years of angiographic follow up. 

Material and Methods 

Patients' population 
   From January 2000 to June 2005, 128 aneurysms in 114 patients were treated with coil embolization. Among them, 54 aneurysms in 51 patients could be followed by repeated intra-arterial angiography at 6 months interval as 6, 12, 18, and 24 months post-embolization, and additional treatment was when the recurrence was morphologically significant. After the follow-up period of 2 years from initial embolization, only plain skull x-ray films were taken to evaluate coil mass deformity.¹⁾ The study population consisted of 15 men and 36 women. The mean age of the patients at the time of presentation was 54.72±11.40 years (range, 30~80years). The endovascular coiling of aneurysm was conducted until complete obliteration was achieved, the last coil could not be safely introduced into the sac, or when any possibility of occlusion of a normal branch with further coil insertion.¹³⁾ 
   Recurrence of an aneurysm was defined when either coil compaction or aneurysm regrowth was identified on follow up angiography.²⁾ Aneurysms were divided into 2 groups; the recurred group and the stable group when recurrence was not seen. There were 12 aneurysms (29%) in the recurred group, and 42 aneurysms (71%) in the stable group. 
   Along with radiographical studies, the clinical parameters such as age, sex, initial Glasgow coma scale (GCS), and Glasgow outcome score (GOS) at the time of the last angiographic follow up were reviewed in order to evaluate the relationship to recurrence. 

Anatomical factors 
   The anatomical location was reviewed in all patients. Geometrical measurements of aneurysmal fundus and neck were done at the maximal respective dimension.⁴⁾ The dome/neck ratio (aspect ratio) was calculated by dividing the aneurysm depth by aneurysm neck width.¹⁹⁾²⁰⁾ The shape of aneurysm was labeled as simple lesion defined by unilocular smooth and spherical aneurysm, and complex lesion by multilocular, irregular aneurysm.¹¹⁾ The inflow pattern into the aneurysm was categorized into terminus, bifurcation and side-wall aneurysms, according to the relationship between the direction of flow in the parent vessel and the neck of the aneurysm. The terminus type was defined as having a 0° angle between the direction of flow in the parent vessel and the neck of the aneurysm, the bifurcation type as having a angle smaller or larger than 90°, and the side-wall type as having a 90° angle.⁸⁾ 

Degree of occlusion 
   The degree of occlusion of aneurysm sac after coil embolization was classified into 3 categories through the immediate post-treatment angiography; complete occlusion, 100% occlusion; near complete occlusion with a small neck remnant, greater than 95% occlusion; partial occlusion with evidence of residual filling, less than 95% occlusion according to the classification by Kole, et al (Fig. 1).⁸⁾ 
   To assess the coil packing density, the size of the aneurysm in three planes (height, length, and width) was measured by digital subtraction angiography, and the volume of the aneurysm sac before embolization was calculated. Percentage occlusion at the end of the procedure was evaluated by coil packing density. Coil packing density was calculated from using the following algebraic equation: Coil packing density = (volume of the embolized coil)/(volume of the aneurysm). The volume of the coil is calculated approximately on the supposition that the coil is a cylinder. The algebraic equation to calculate the volume of the coil is as follows: volume of coil = πX (diameter of coil/2)² X length of coil. The volume of the aneurysm is also approximately calculated based on the supposition that the aneurysm is ellipsoid: volume of the aneurysm = 4π/3 X (width/2) X (length/2) X (height/2).⁷⁾¹⁶⁾ 

Statistical analysis 
   All values are expressed as mean±standard deviation. Student's t test, Chi square test and Fisher's exact test were used for statistical analysis and a p value of less than 0.05 was considered significant. 

Results 

Patient demographics 
   The majority of recurrences were found after 6 to 12 months post-embolization (Table 1). Mean age was 55.17± 10.60 years (range 30~80 years) in the recurred group, and 54.33±12.60 years (range 33~68 years) in the stable group (p=0.836). The initial GCS scores and GOS scores did not show a difference between the two groups (12.95±2.34 and 4.57±0.88, respectively, in the stable group, 12.67±2.74 and 4.58±0.79, respectively, in the recurred group). The male to female ratio in the stable group was 11 to 30, and in the recurred group, 4 to 6. 

Anatomical factors 
   In our study, aneurysms were mainly located in the anterior circulation (Table 2). The diameter of the largest dimension of the fundus was larger in the recurred group than the stable group (7.33±2.26mm vs. 5.87±1.93mm, p=0.048) and a significantly higher rate of recurrence was seen in fundus size of same or more than 10mm (41.7% vs 14.3%, p=0.038). There was a trend that the mean maximum size of the neck was larger in the recurred group than stable group, and a higher rate of recurrence was seen in neck size of same or more than 4mm, but did not reach significance (36.0% vs 22.9%, p=0.084) (Table 3, 4). Complex shaped aneurysm showed more recurrence then simple ones (41.7% vs. 16.7%), and there was 25% recurrence in terminus inflow pattern and 20% in the bifurcation pattern (Table 5). 

Degree of occlusion 
   Among the 54 aneurysms, we were able to perform complete occlusion by coil embolization in 26 aneurysms (48.1%), near complete 17 (31.5%), and partial occlusion 11 (20.4%). Follow up angiographic evaluation showed as high as twice the recurrence rate in partial occlusion group (36.4%) than complete (19.2%) or near complete group (17.6%) (Table 6). The coil packing density was lower in the recurred group in all occlusion statuses compared to the stable group, showing a significant correlation between coil packing density and recurrence (Table 7). A significantly higher rate of recurrence was seen in patients with a coil packing density below 25% (47.0% vs 12.9%, p=0.017) (Table 8). 

Discussion 

   In the treatment of aneurysmal subarachnoid hemorrhage, rebleeding and vasospasm have been the major complication with negative impact on morbidity and mortality. Rates of aneurysmal remnants or rests after surgical clipping ranged from 4 to 6% on post-operative follow-up angiography.¹⁾¹⁵⁾ Rebleeding in patients with aneurysmal remnants or rests after surgical clipping showed an annual risk of 0.38 to 0.79% compared to overall annual rebleeding rates after surgical clipping of 0 to 0.03% in surgical clipping cases.^5)6)17) On the other hand, in cases with coil embolization, although considerable achievement devoted to obtaining the highest degree of occlusion without leaving residual filling or a neck remnant have been made than before, recent studies have revealed that some larger part of the patients, i.e. 38~64% of the aneurysms treated with endovascular coil insertion are incomplete at the time of immediate posttreatment angiography, though the degree of occlusion was not always specified identically, as is our result (54%).^2)8)9)11)13) The incidence of overall aneurysm rebleeding after coil embolization was between 0.8% and 3.3% per year, but in case of incomplete coil occlusion, rebleeding incidence increased up to between 1.5% and 7% per year.^{1)5)6)8)9)17)} Overall rebleeding rates were slightly higher in patients treated with coil embolization, and as might be expected, many authors have suggested that incomplete occlusion of aneurysmal sac revealed on immediate post-embolization may be directly related to aneurysmal recurrence and possible subsequent bleeding.^1)2)9)11)13) While angiographical recurrences of aneurysms after surgical clipping have been reported to be 1.4% to 2.9%, recurrence after coil embolization ranged from 22% to 34%, which was similar to our results (29%).^1)10)13)18) 
   Many anatomic geometric variables, such as aneurysm location, fundus diameter, neck size, aspect ratio, inflow pattern, occlusion status and coil packing density have been identified as risk factors of recurrence. Among theses, some authors have described aneurysm location of the aneurysm as a possible risk factor of recurrence.^2)4)17) The main advantage of detachable coils is its ability to adapt ideally to the shape of the aneurysm, but conversely, this flexibility prevents the coil from resisting the blood flow at the neck level, resulting in the flow pushing the coils into the sac and eventual recanalization. Therefore, the amount of flow in the parent vessel, which depends upon the location, can influence the recurrence rate. For example, aneurysms of the basilar or internal carotid artery bifurcations are known to be more prone to the arterial flow stresses than the middle cerebral or anterior communicating artery aneurysms, which in turn may lead to a higher recurrence rates.⁴⁾ However, others have not found a significant impact of location on recurrence, although this may be due to insufficient number of cases in the posterior circulation.⁸⁾¹³⁾ 
   It is well known that coil embolization of large and giant aneurysms frequently fail to obtain a total and durable occlusion of the aneurysm sac, which may lead to less favorable initial results and a higher recurrence rate. Regarding the size of aneurysm, the larger the size, more increased rate of recurrence (20~50%) were noted.^1)2)8)11)13) Raymond, et al., reported that larger sized aneurysms showed a strong correlation with recurrence, and specified that lesion of same or more than 10mm in fundus diameter had a recurrence rate of 50.6%, compared to 21.3% in smaller aneurysms, although initial occlusion rates were not described according to initial fundus size.¹³⁾ In our study, there was significantly higher rate of recurrence in fundus size of same or more than 10 (41.7% vs 14.3%, p=0.038). 
   Aneurysms with a neck larger than 4mm, or wide necked aneurysms, are more likely to have a immediate posttreatment neck remnant or residual filling, and a higher rate of recurrence in angiographical follow-up.^4)8)9)13) Murayama, et al., reported that aneurysms with wide necks had a recurrence rate of 20% compared to 5.1%, in small necked aneurysms.⁹⁾ In our study, there was a trend that the neck size was larger in the recurred group than in the stable group (Table 3), and the recurrence rate in neck size of same or more than 4mm was 36%, which was higher than the 29% overall recurrence rate. 
   It is generally accepted that increased aspect ratio is closely related with high incidence of recurrence after coil embolization. Theoretically, a more sluggish blood flow and longer period of blood stasis inside the aneurysm can be anticipated in aneurysms with a larger depth compared to the neck diameter, suggesting that the aspect ratio may be a good parameter by which to express the intra-aneurysmal flow condition.¹⁹⁾²⁰⁾ Cognard, et al., showed that an increase in incomplete occlusion after initial coil embolization and a higher rate of recurrence was seen in patients with larger aspect ratios, and others have yielded similar results.^2)9)20) Thus, a smaller aspect ratio may correlate to a faster blood flow into the aneurysm, which in turn may lead to a greater chance of recurrence. An aspect ratio of less than 1:2 is usually acknowledged as the critical point of less favorable outcome after coil embolization.