Article Text

Original research
Safety and cost of stent-assisted coiling of unruptured intracranial aneurysms compared with coiling or clipping
1. Jennifer A Frontera1,
2. Joseph Moatti2,
3. Kenneth M de los Reyes2,
4. Stephen McCullough2,
5. Henry Moyle2,
6. Joshua B Bederson2,
7. Aman Patel2
1. 1Cerebrovascular Center, Cleveland Clinic, Cleveland, Ohio, USA
2. 2Department of Neurosurgery, Mount Sinai School of Medicine, New York, New York, USA
1. Correspondence to Dr Jennifer A Frontera, Cerebrovascular Center, Cleveland Clinic, 9500 Euclid Ave. S80, Cleveland, OH 44195, USA; frontej{at}ccf.org

## Abstract

Objective Stent-assisted coiling (SAC) of unruptured intracranial aneurysms is a treatment alternative to clipping or coiling, although high complication and procedure-related mortality rates have been reported.

Methods A retrospective study was conducted of patients undergoing SAC, coiling or clipping of unruptured intracranial aneurysms between 2003 and 2010. Rates of residual aneurysm, recanalization, complications, cost (adjusted to 2010), length of stay (LOS) and outcome were compared between groups.

Results Of 116 subjects, 47 underwent SAC, 33 coiling and 36 clipping. The groups were similar in age, gender and aneurysm location, although the SAC group had significantly larger aneurysms with wider necks (p=0.001). Patients who underwent SAC had more residual aneurysm after initial treatment than those treated with coiling or clipping (75%, 52% and 19%, respectively, p<0.0001), but this difference was smaller at follow-up angiography (50%, 50% and 17% residual, respectively) and was not significant after adjusting for baseline aneurysm and neck size. SAC was not associated with increased recanalization, requirement for additional treatment, mortality or complications after adjusting for aneurysm and neck size. Patients who underwent SAC and those who underwent coiling were more likely to have a good discharge disposition than patients treated with clipping (100% vs 91%, p=0.042). LOS was significantly shorter for patients who underwent SAC or coiling compared with those treated with clipping (p<0.0001). The overall direct cost was higher for patients who underwent SAC than for those treated with coiling or clipping (median $22 544 vs$12 933 vs $14 656, p=0.001), even after adjusting for aneurysm and neck size, LOS and retreatment. Conclusions SAC is a safe alternative to coiling or clipping of unruptured aneurysms but it is currently more expensive. • Aneurysm • Angiography • Coil • Stent View Full Text ## Statistics from Altmetric.com ## Introduction Stent-assisted coiling (SAC) has become an attractive treatment option for large wide-necked or side wall aneurysms or aneurysms located in the supraclinoid internal cerebral artery or basilar tip. A recent retrospective study of SAC in 216 patients found that, although SAC was associated with significantly lower angiographic recurrence rates than coiling alone, complication rates were substantially higher. Permanent neurological complications occurred in 7.4% and procedure-related mortality was 4.6% among patients who underwent SAC, which was significantly higher than for those treated with coiling alone.1 This study prompted an editorial citing the ‘alarmingly high serious complication rate of SAC’ in this cohort of largely unruptured aneurysms with a benign natural history.2 Others have found SAC to be associated with morbidity rates ranging from 0% to 4.2%.3–5 However, all these series are retrospective and some contain mixed populations of ruptured and unruptured aneurysms. Additionally, none compared the rates of obliteration, recanalization, complications or cost with surgical clipping. In this study we examined patients with unruptured aneurysms who underwent SAC, coiling alone or clipping and compared the safety, residual and recanalization rates and cost of each treatment modality. We hypothesized that, after adjusting for baseline aneurysm characteristics, SAC would have similar rates of residual aneurysm, recanalization and morbidity to coiling alone or clipping. ## Materials and methods ### Study population A single-center retrospective study was conducted of all consecutive patients with unruptured intracranial aneurysms treated at our medical center between December 2003 and August 2010. Inclusion criteria for analysis were defined as: unruptured cerebral aneurysm, attempted aneurysm repair using SAC, coiling alone or surgical clipping, presence of at least one digital subtraction angiogram following aneurysm repair and age ≥18 years. ### Repair procedure All patients underwent baseline digital subtraction cerebral angiography to characterize the aneurysm and assess potential treatment. The choice of repair procedure was based on the aneurysm location and morphology as well as the patient's clinical status and wishes. SAC was considered particularly in patients with unfavorable dome to neck ratios or in those with necks that were relatively wide compared with the parent artery. Aneurysm obliteration was the goal for all aneurysm repairs, although staged procedures were planned on an individual basis depending on aneurysm morphology, location and mass effect on eloquent structures. All procedures were performed under general anesthesia. Endovascular procedures were performed in an angiography suite located in the radiology department. No coiling or SAC procedures were performed in a biplane operating room suite as this was not available during the time period studied (2003–2010). #### Stent-assisted coiling From 2003 to 2006, Neuroform (Stryker) self-expanding stents were preferentially used and from 2007 to 2010 Enterprise (Cordis) self-expanding stents were used. All patients received clopidogrel 75 mg orally four times a day and aspirin 325 mg orally four times a day beginning a minimum of 5 days prior to the procedure. During the procedure 4000 U intravenous heparin was administered and redosed throughout the case at 1000 U every hour. All patients received stent placement followed by coiling during a single procedure. Stents were deployed under roadmap guidance. After confirming the stent position with a follow-up angiogram, coil embolization of the aneurysm was performed using either Gugliemi detachable coils (Boston Scientific, Massachusetts, USA) or Orbit coils (Codman, Massachusetts, USA). Patients were instructed to continue a combination of clopidogrel 75 mg four times a day and aspirin 325 mg four times a day for a minimum of 6 weeks followed by aspirin 81 mg four times daily alone indefinitely. #### Coiling alone Intravenous heparin was administered during the procedure as above and either Gugliemi detachable coils or Orbit coils were deployed through a standard microcatheter approach to pack the aneurysm. In some cases where a branch artery at the aneurysm base needed protection, balloon-assisted coil embolization using a hyperglide balloon (eV3, Plymouth, Minnesota, USA) was used during embolization. #### Clipping Open microsurgical repair with clipping was performed with the guidance of microvascular Doppler to assess blood flow and intraoperative somatosensory evoked potentials and EEG to identify any neurophysiological aberrations during temporary clipping. #### Post-repair clinical management Patients who underwent SAC or coiling alone stayed overnight in the postoperative recovery room and were discharged the following morning unless significant comorbidities or procedural complications warranted care in the Neuro ICU. All clipped patients were admitted overnight to the Neuro ICU. ### Follow-up angiograms In accordance with the protocol, follow-up angiograms were performed immediately after surgery for patients treated with clipping (within 24 h, even if intraoperative angiography was performed) and at 6 months for patients treated with SAC or coiling. A second angiogram was performed at 1 year for all patients (SAC, coiling and clipping). ### Data collection and outcome measures Patient demographics including age, gender and ethnicity as well as aneurysm size (maximum diameter determined by three-dimensional rotational angiography), neck size and location (anterior or posterior location) were documented. Residual aneurysm after the first treatment was diagnosed by digital subtraction angiography using the last run after SAC or coiling or the immediate postoperative angiogram in patients treated with clipping. Residual aneurysm on the follow-up angiogram was coded if present on the 6-month or 1-year follow-up study. Residual aneurysm was categorized as neck, dome or interstices (between the coil mass) filling (figure 1) by two endovascular neurosurgeons performing the angiograms. Recanalization was coded if there was angiographic evidence of recurrent aneurysm filling at the 6-month or 1-year angiogram after previous angiographic documentation of obliteration or thrombosis. The requirement for additional treatment (categorized as planned/staged or unplanned) was recorded. Treatment complications were categorized as asymptomatic angiographic complications, symptomatic complications and periprocedural aneurysm rupture. Discharge disposition was characterized as dead, good (discharged home, home with health aid or to an acute rehabilitation center) or poor (discharged to a nursing home or a subacute rehabilitation center). Hospitalization costs were categorized as total direct costs (sum of all direct hospital costs including nursing, diagnostic, imaging, ICU, surgical, angiographic, laboratory, emergency department, blood bank, pharmacy, rehabilitation and respiratory therapy costs), overheads (defined as costs related to the overall operation of the hospital, not directly related to patient care including human resources, payroll, information technology, finance, legal administration and facilities costs), ICU direct costs, surgical direct costs and angiography direct costs. Hospitalization costs do not include professional fees. The cost of materials to perform a patient care-related procedure is not included in the overheads. Surgical instruments and clips are bundled with surgical costs, while stents and coils are bundled with angiographic costs. Overheads are not included in the total direct costs. A separate assessment of coil costs was performed using the median cost of Gugliemi detachable coils ($1140.44 per coil in 2010) and the standard cost of Orbit coils ($1253.02 per coil in 2010). If both types of coils were used, the average cost of each brand was applied ($1196.73 per coil in 2010). Material costs were provided by the manufacturers and were specific to Mount Sinai Hospital. All cost data were adjusted to 2010.

Figure 1

Digital subtraction angiography demonstrating residual aneurysm neck (thin white arrow), residual aneurysm dome (thick white arrow) and residual interstices filling between the coil mass (black arrow).

### Statistical analysis

Demographics, aneurysm characteristics, the presence and type of residual aneurysm (after initial treatment and on follow-up angiogram), aneurysm recanalization, the requirement for additional aneurysm treatment, complications and discharge disposition were compared between patients treated with SAC, coiling or clipping using the Pearson χ2 two-sided test. The cost and length of stay (LOS) were compared between groups using the Kruskal–Wallis and Mann–Whitney U non-parametric test for continuous non-normally distributed variables. Multivariate logistic regression analysis was performed to assess the association between the repair procedure and the presence of residual aneurysm, recanalization, the requirement for additional aneurysm treatment and cost (dichotomized at the median) after adjusting for aneurysm and neck size. All analyses were performed using SPSS V.19 (Chicago, Illinois, USA). Significance was set at p<0.05.

## Results

A total of 116 patients with unruptured cerebral aneurysms were included in the analysis. Of these, 47 (41%) underwent SAC, 36 (31%) had coiling alone and 33 (28%) were treated by clipping. Among the patients who underwent SAC, 34 (72%) received an Enterprise stent (Cordis) and 13 (28%) received a Neuroform stent. All clipped patients had a postoperative follow-up angiogram performed a median of 1 day (range 1–492) following surgery. At 6 months a follow-up angiogram was performed in 98% of patients treated with SAC and 97% of patients treated with coiling. The median time to the 6-month follow-up angiogram was 183 days for patients who underwent SAC and 225 days for patients treated with coiling alone. One-year follow-up angiograms were performed in 42% (N=15) of patients treated by clipping at a median of 378 days, 42% (N=15) of patients treated by coiling at a median of 584 days and 40% (N=19) of patients who underwent SAC at a median of 363 days.

There were no significant differences in age, gender, ethnicity or aneurysm location between the repair groups, but those who underwent SAC had significantly larger aneurysm and neck sizes than those who were coiled and a trend towards larger aneurysm size than in those treated with clipping (table 1).

Table 1

Demographics and aneurysm characteristics

After initial aneurysm repair, significantly more patients who underwent SAC had some residual aneurysm compared with those who were coiled or clipped. However, there was no significant difference in neck or dome residual aneurysm, only in interstices residual aneurysm (table 2). A third of SAC patients with residual aneurysm after initial treatment had complete obliteration on follow-up angiogram at 6 months or 1 year, and overall there was no significant difference in the prevalence of residual aneurysm between patients who underwent SAC and those treated with coiling at this time point. Patients in the SAC group were significantly more likely to have residual interstices filling at 6 months or 1 year than those treated with clipping. After adjusting for aneurysm and neck size, SAC was significantly associated with residual aneurysm after initial treatment (adjusted OR (aOR) 4.7, 95% CI 1.8 to 12.6, p=0.002), but SAC did not significantly predict residual aneurysm on the 6-month or 1-year follow-up angiogram (aOR 1.2, 95% CI 0.4 to 3.4, p=0.604) compared with coiling or clipping. Although coiling alone was not associated with residual aneurysm following initial treatment after adjustment for baseline aneurysm characteristics, it was significantly associated with residual aneurysm on the 6-month or 1-year follow-up angiogram (aOR 9.0, 95% CI 2.1 to 37.7, p=0.003) after adjusting for aneurysm and neck size. Conversely, clipping was protective against residual aneurysm after initial treatment (aOR 0.1, 95% OR 0.01 to 0.3, p<0.0001) but not at follow-up angiography after adjusting for aneurysm and neck size and time to follow-up angiography.

Table 2

Comparison of residual aneurysm and recanalization after SAC, coiling alone or clipping

Recanalization was significantly more common after SAC or coiling than after clipping (table 2). However, after adjusting for aneurysm and neck size, neither SAC (aOR 1.2, 95% CI 0.2 to 6.4, p=0.803) nor coiling (aOR 5.0, 95% CI 0.8 to 31.8, p=0.090) significantly predicted recanalization. Clipping was also not associated with a reduced risk of recanalization after adjustment in a multivariate model.

Patients who underwent SAC required more additional aneurysm treatments than those treated with coiling or clipping, although most of these additional treatments were planned or staged (table 2). Among SAC patients, seven (16%) required additional treatment compared with two (6%) of those treated with coiling alone and none of those treated with clipping (p=0.030). The maximum number of treatments in the SAC group was two. After adjusting for aneurysm and neck size, SAC was not significantly associated with the need for additional aneurysm treatment (aOR 3.3, 95% CI 0.3 to 36.9, p=0.330). Clipping and coiling alone were also not significantly associated with the need for additional treatment.

Complications including periprocedural aneurysm rupture, asymptomatic and symptomatic complications did not differ significantly between treatment groups (p=0.387 for any complication). Periprocedural aneurysm rupture occurred in one patient treated with SAC (2%) and one treated with clipping (3%), (p=0.816). Asymptomatic angiographic complications occurred in two patients who underwent SAC (4%) including thrombus formation on a coil mass which cleared with intravenous abciximab infusion and carotid dissection related to deployment of a Neuroform stent (p=0.235 vs other treatment groups). Symptomatic complications occurred in one patient treated with SAC (2%) with a periprocedural aneurysm rupture, in two patients treated with clipping (6%) (one with a periprocedural rupture and one with a procedure-related stroke) and no patients treated with coiling alone had a symptomatic complication (p=0.349 across all groups). Procedure-related symptomatic complications were predicted by aneurysm size (OR 1.2, 95% CI 1.0 to 1.5, p=0.049) and neck size (OR 1.5, 95% CI 1.1 to 2.0, p=0.022).

Mortality rates did not differ significantly between treatment groups; no deaths occurred in the SAC or coiling alone groups and one patient treated with clipping died. All patients treated with SAC or coiling had good discharge disposition compared with 91% of patients in the clipping group (p=0.042, table 3). Both ICU and hospital LOS were significantly longer in patients treated with clipping than in those who underwent SAC or coiling (table 4).

Table 3

Discharge disposition following SAC, coiling alone or clipping

Table 4

ICU and hospital length of stay following SAC, coiling alone or clipping

Hospital costs including total direct costs, overhead and angiographic costs were significantly higher for patients undergoing SAC than for those treated by coiling or clipping (table 5). ICU and surgical costs were significantly higher for clipped patients. Among patients with SAC, the median number of coils was 7 (range 1–24) at a median cost of $7983 (range$1253–30 072) per aneurysm treatment while, for patients with coiling alone, the median number of coils was 4.5 (range 2–14) at a median coil cost per treatment of $5385 (range$2393–16 754). Patients who underwent SAC had significantly more coils (p=0.046) and incurred significantly more expense for these coils (p=0.035) compared with the coiling alone group. This may contribute, in part, to the higher overall direct cost and angiographic cost of SAC compared with coiling alone. A comparison of large (>10 mm) and small (≤10 mm) aneurysms revealed a significant difference in coil cost but no differences in direct costs, overheads, angiographic, ICU or surgical costs (table 6). Comparing costs among retreated patients and those requiring only one treatment, any retreatment was associated with higher angiographic and coil costs although the total direct cost, overhead, ICU and surgical costs were not significantly different (table 7). After adjusting for aneurysm and neck size, ICU and hospital LOS and the need for retreatment, SAC was independently predictive of a higher overall direct cost (aOR 17.1, 95% CI 3.4 to 86.8, p=0.001), overhead (aOR 38.6, 95% CI 5.9 to 254.2, p<0.001) and angiographic cost (aOR 52.9, 95% CI 6.3 to 446.0, p<0.001), while clipping was inversely related to a higher direct cost (aOR 0.01, 95% CI 0.001 to 0.4, p=0.012) and coiling was not related to the overall direct cost after adjustment.

Table 5

Hospital costs for SAC, coiling alone or clipping adjusted to 2010

Table 6

Cost comparison of large and small aneurysm treatment

Table 7

Cost comparison of patients requiring aneurysm retreatment compared with those requiring only one treatment

## Discussion

In this study of unruptured intracranial aneurysms we found SAC to be efficacious and safe compared with clipping or coiling alone, although significantly more expensive. This study is unique because it compares SAC with both coiling and clipping, which has not been reported before to our knowledge. Since aneurysm morphology that is amenable to SAC may not be treatable by coiling alone, it is important to compare SAC with surgical clipping. Another unique aspect of this study is the comparison of the cost for each treatment option.

Patients who underwent SAC did not have increased complication rates compared with those treated with clipping or coiling. The periprocedure aneurysm rupture rate was low across all groups and similar to the rates of 2–3% quoted in the literature.1 ,6 Our symptomatic complication rate among patients who underwent SAC (2%) was lower than that reported in other studies, which has been as high as 7.4–9.5%.1 ,5 Mortality rates across the treatment groups were similar and, reassuringly, 0% among patients treated with SAC or coiling compared with procedure-related mortality rates up to 4.6% among SAC patients in some studies,1 which is relatively high given the natural history of unruptured aneurysms. The SAC and coiling groups also had better discharge dispositions and shorter LOS than patients treated with clipping. Shorter LOS has been described among patients with unruptured aneurysms treated with coiling compared with those treated with clipping7–10 but, to our knowledge, no literature exists comparing LOS among patients treated with SAC, clipping or coiling alone. Our data attest to the safety of SAC compared with coiling and clipping.

Overall hospitalization cost (direct cost) was significantly higher for patients who underwent SAC than for those treated with coiling or clipping, even after adjusting for aneurysm and neck size, LOS and retreatment. Although the surgical and ICU costs were higher among patients treated with clipping, angiographic costs comprised the largest share of the differences between the procedures, contributing to an overall higher direct cost for SAC. Part of the increased cost of SAC is related to the cost of the stent itself, ranging from approximately $5200 (Enterprise stent) to$5315 (Neuroform stent). Additionally, patients treated with SAC required significantly more coils and had a higher procedural coil cost than patients who underwent coiling alone. Overhead costs were also higher for patients who underwent SAC, which may reflect longer procedure time, longer anesthesia time, higher staffing needs and higher facilities costs. Since SAC was associated with significantly more repeated treatments, this may have contributed somewhat to angiographic and overhead costs although, after adjusting for retreatment, aneurysm and neck size and LOS, SAC was still an independent predictor of the overall direct cost as well as overhead and angiographic costs.

The location where procedures are performed (ie, operating room vs angiography suite) may also contribute to the differences in overheads. The costs of performing an endovascular procedure in a biplane operating room versus an angiography suite may vary from institution to institution based on the types of staff that are required for each location (ie, scrub nurses, nurse anesthetists, anesthesiologists), the time required to perform a procedure in each location and unmeasured costs or savings such as the impact on the volume of other non-vascular procedures that may be performed if the operating room or radiology suite was not being occupied for an aneurysm repair procedure. Additionally, centers that routinely perform combined surgery and angiography procedures (ie, aneurysm clipping followed immediately by post-clipping angiography) may see particular cost savings for a biplane operating room. Since all of our endovascular procedures were performed in an angiography suite, we are unable to estimate if there would be any difference in cost based on the location in which the procedure is performed.

Higher procedure costs related to SAC seemed to overshadow reductions in LOS compared with clipping. In contrast to other studies of unruptured aneurysm repair which have found coiling to be more expensive than clipping,7 ,8 ,11 we did not observe a difference in cost between these treatment modalities. This may be due to differences in patient comorbidities or hospital billing procedures between cohorts. Of note, the cost for clipping or coiling in our study was similar to the average Medicare reimbursement rate of $12 599, while others have reported much higher median hospitalization costs of$23 574 for clipping and \$25 734 for coiling of unruptured aneurysms.11

When examining the prevalence of residual aneurysm after SAC, we found that, after adjusting for aneurysm and neck size, SAC was associated with a significantly increased risk of residual aneurysm after initial treatment but was not associated with 6-month or 1-year residual aneurysm compared with coiling alone or clipping. Our rates of complete occlusion after initial treatment with SAC were only 25%, which is low, but not dissimilar from previously reported rates of 26.5–46%.1 ,5 ,12 Low rates of obliteration after initial SAC treatment may be due to the fact that coil packing after stent placement is more difficult due to restricted microcatheter maneuverability, and the use of dual antiplatelet therapy and heparin during the procedure may limit thrombosis around the coil mass.1

Our overall obliteration rates for SAC and coiling at 6 months or 1 year were only 50%, which is lower than reported rates of 53–67% in various studies with angiographic follow-up occurring over a median 12–24 months.3 ,5 ,13 Our shorter time frame for angiographic follow-up may in part explain the lower obliteration rate. Since our protocol did not include angiography at 6 months for patients treated by clipping, the data are skewed to favor clipped patients with a longer follow-up period. After adjusting for aneurysm and neck size as well as time to follow-up angiography, clipping was no different from other treatments in producing aneurysm obliteration.

Although some authors have found significantly lower recanalization rates among patients treated with SAC compared with those treated by coiling,1 we found similar rates among both groups of patients (13% in each group). The patients who underwent SAC in our cohort had similar recanalization rates to those reported in other series (12–17%),1 ,4 ,12 ,14 although our coiled group had much lower rates than the literature estimates of 34–36%.1 ,12

Patients who underwent SAC were more likely to undergo staged treatment than those treated with coiling or clipping largely due to the aneurysm size, location and compression of eloquent structures. Our rates of retreatment, either planned (12%) or unplanned (4%), are not dissimilar from other series reporting retreatment rates of 6–14% after SAC.3 ,15 ,16 Again, after adjusting for aneurysm and neck size, patients treated with SAC were no more likely to require additional treatment than those treated by clipping or coiling.

Some limitations of this study should be mentioned. First, this was a single-center retrospective study and treatment approaches may change over time as endovascular technology changes. Second, heterogeneous aneurysm morphology and location can contribute to treatment bias, although we did not find significant differences in aneurysm location between groups and we attempted to adjust for morphological differences (aneurysm and neck size) in multivariate analysis. Third, we had limited angiographic follow-up on patients treated by clipping which may have led to an underestimation of recanalization rates. Fourth, our angiographic follow-up only extended up to 1 year, limiting our ability to compare the long-term durability of each treatment modality. Finally, we focused only on hospitalization costs related to the initial treatment and did not include long-term follow-up costs which may differ between SAC, clipping and coiling. However, other studies following cost out to 12 months in patients with ruptured aneurysms (which included costs related to follow-up procedures, rehabilitation or nursing home stays) found that 85–96% of the cost was incurred during the initial treatment and hospitalization.17 ,18 Thus, focusing on predictors of cost related to the initial hospitalization may be the most productive approach.

## Conclusions

SAC is a safe treatment option for unruptured intracranial aneurysms, although it is more expensive than clipping or coiling even after adjusting for aneurysm and neck size, LOS and the number of treatment procedures. Lower coil and stent costs over time may allow SAC to become a more affordable treatment strategy.

View Abstract

## Footnotes

• Contributors JAF: study design, analysis, interpretation of data, drafting the article, final approval. JM: study conception, acquisition of data, critical revision, final approval. KMdlR: acquisition of data, critical revision, final approval. SMcC: acquisition of data, critical revision, final approval. HM: study conception, acquisition of data, critical revision, final approval. JBB: study design, acquisition of data, critical revision, final approval. AP: study conception and design, acquisition of data, critical revision, final approval.

• Funding This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.

• Ethics approval Ethics approval was obtained from the Institutional Review Board.

• Competing interests None.

• Provenance and peer review Not commissioned; externally peer reviewed.

## Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.