Unruptured Aneurysms: A Surgical Perspective

CHIRAG D. GANDHI, AMAN B. PATEL, AND JOSHUA B. BEDERSON

Although the last two decades have seen significant advances in the surgical and medical management of aneurysmal subarachnoid hemorrhage (SAH), the overall morbidity and mortality rates are as high as 40 to 50%.^1,2 Poor neurological outcomes after SAH are due to the effects of the initial hemorrhage, early rebleeding, and delayed cerebral ischemia. Prevention of SAH by surgical treatment of unruptured aneurysms is a definitive strategy by which to reduce poor outcomes. However, because all treatments carry at least some risk, there remains significant controversy as to the best treatment for unruptured intracranial aneurysms (UIAs). Autopsy studies have shown that the overall frequency of UIAs in the general population is ~1 to 5%.³ The incidence of SAH is ~10 cases per 100,000 persons per year⁴; hence, it is likely that most intracranial aneurysms do not rupture. Detection of incidental intracranial aneurysms has increased recently because of improved diagnostic technology such as magnetic resonance imaging (MRI) and angiography, computerized tomography (CT) angiography, and digital subtraction angiography (which remains the “gold standard” for diagnosis), making the management of UIAs a more common neurosurgical dilemma.

Various attempts have been made to establish rational treatment protocols for UIAs. Such protocols should take into account the risk of treatment as well as the risk of rupture of an UIA (its natural history). Studies have demonstrated that both the natural history of UIAs and the outcomes of treatment are strongly influenced by three factors: (1) patient characteristics such as previous SAH, age, and other medical comorbidities; (2) aneurysm characteristics such as size, location, and morphology³; and (3) management factors such as experience of the surgical team and treating hospital. These complex and interrelated factors have made evidencebased assessments of bleeding rate and treatment efficacy very difficult. In light of the lack of convincing prospective randomized studies comparing observation with treatment, UIAs are currently managed based on our best understanding of natural history versus risk of treatment.

Aneurysm characteristics such as size, location, and morphology influence the natural history of UIAs. A review of 945 patients found that ruptured aneurysms have a greater diameter than unruptured ones, 10.8 mm versus 7.8 mm.⁵ Additionally, autopsy studies and retrospective reviews have found that aneurysms in certain locations such as the anterior communicating artery and pericallosal artery may be at a higher risk for rupture than other locations.^5,6 Morphological variations such as multilobulation and loculations may also increase the rupture risk.⁷

The natural history of patients with incidental aneurysms without prior SAH can be separated according to those that present with symptoms other than SAH (i.e., third cranial nerve palsy, headaches, orbital pain, seizures) and those that are truly incidental. In the 1969 Cooperative Aneurysm Study, 165 patients with symptomatic aneurysms were followed, and 79% of these aneurysms were treated surgically.⁸ The remaining 34 cases were left untreated, and the natural course of the disease was studied. Of these remaining patients, 25% had a SAH at 3 months to 3 years. All of the aneurysms that hemorrhaged were 7 to 10 mm in diameter, and all were fatal. Aneurysms that were smaller than 7 mm did not bleed. Other studies have suggested that the risk of rupture is higher for symptomatic aneurysms with high mortality rates, although the magnitude of the reported increase in risk varies.^9–11 Very few studies following the natural history of this specific subgroup exist because they are selected for treatment. Aneurysms that present with symptoms of mass effect are likely to be larger. Nevertheless, the data suggest that symptomatic aneurysms have a higher rate of rupture that should be addressed somewhat urgently. Based on the recommendations of the Stroke Council, all intradural symptomatic UIAs should be considered for treatment, especially those that have recently increased in size.¹²

On the natural history of asymptomatic unruptured aneurysms, a recent study in Japan reported on 62 patients observed for more than 6 months to follow their saccular, nonthrombotic aneurysms. The aneurysms had been diagnosed on angiograms obtained for causes other than SAH.¹³ Patients were followed for 6 months to 7 years; in seven patients SAH was documented on CT scans at a mean interval of 4.89 years. Six of the seven patients died from SAH, and the remaining patient suffered severe neurological deficit. Based on this, the cumulative risk for all aneurysms at 5 and 10 years were 7.5% and 22.5%, respectively. For aneurysms smaller than 10 mm, the 5- and 10-year rates were 4.5% and 13.9%, respectively; in aneurysms greater than 10 mm, the 5- and 10-year rates were 33.5% and 55.9%, respectively. These rates of aneurysm rupture are higher than other reports from North America. In a subset of the North American Carotid Endarterectomy trial, 66 patients (mean age of 66 years) were found to have unruptured, noncavernous sinus, intracranial aneurysms, and only one of the follow-up group of 58 patients suffered a SAH.¹⁴ Although both studies have similar patient demographics, the rate of aneurysm rupture was significantly less in the North American study.

In another review by Weibers et al, 130 patients were followed for 8.3 years, during which time 15 patients suffered a SAH.¹¹ All 15 ruptured aneurysms were greater than 10 mm in diameter. The 102 aneurysms measuring less than 10 mm remained unruptured. The subsequent analysis determined that size was the only predictor of aneurysm rupture. However, 10 aneurysms were 8 to 9 mm, and 36 aneurysms were 6 to 9 mm; 10 mm could not definitively be established as the critical size for rupture. This finding promoted the further investigation of the influence of size on bleeding risk.

The International Study of Unruptured Intracranial Aneurysms (ISUIA) is the largest and most comprehensive natural history study of unruptured intracranial aneurysms conducted with a large number of patients. The study also allows for statistical significance and the analysis of secondary subgroups.¹⁵ The preliminary report from this continuing study has generated much controversy and raises questions about conclusions drawn from prior studies in regards to the natural history and the indications for surgery in patients with UIAs. The retrospective component of the ISUIA evaluated natural history and included 1449 patients divided into two groups—patients with and without a prior history of SAH from a different treated aneurysm. Group I consisted of 727 patients with no history of SAH who were followed for an average of 7.5 years. Rates of rupture for aneurysms less than 10 mm they were 0.05% per year, for those greater than 10 mm they were 1% per year, and for those greater than 25 mm the rate of rupture was 6% in the first year. Predictors of rupture were increasing size and location, specifically at the basilar tip, posterior cerebral or vertebrobasilar distribution, or the origin of the posterior communicating artery.

Patients with a history of SAH from a different aneurysm were designated as Group 2 in the ISUIA. In Group 2, aneurysm size alone was not a predictor of rupture rate. Basilar tip location was the only predictor among these patients. In patients with aneurysms less than 10 mm, the risk of rupture was 11 times higher in patients with prior SAH than in patients without SAH (0.5% vs. 0.05% per year). In aneurysms greater than 10 mm, the rate was found to be equal between the two groups. The presence of multiple intracranial aneurysms was not a predictive factor for future rupture even though unruptured aneurysms in Group 2 were more likely to bleed.

The rate of rupture found in the ISUIA is significantly lower than prior reports of 1 to 2% per year.^13,16 Critics of the ISUIA argue that the retrospective portion of the study is invalid because it excludes those patients who are more likely to undergo treatment rather than conservative management, specifically younger patients. Additionally, the inclusion of intracavernous and proximal carotid artery aneurysms, which are much less likely to bleed, may have contributed to the lower rupture rate in this study. Critics also note that although the reported mortality rate of SAH is ~45%, the rate in the ISUIA study was 83%. This fact could be a result of wide confidence intervals or may reflect selection bias by having a larger number of older, medically frail patients who were not eligible for treatment. Additionally, the retrospective branch of the study had a variety of inclusion and exclusion criteria that some argue introduce further bias. Although the results of this study did not clearly support treatment of Group 1 patients with UIAs less than 10 mm in diameter, these conclusions have been challenged because of the criteria for inclusion in the study as discussed above.

Other studies have found that the rate of rupture in patients with prior SAH is higher than stated in the ISUIA study. Juvela et al retrospectively examined the natural history and the predictive risk factors for aneurysm rupture in 142 patients, 131 of whom had a prior SAH.¹⁰ Median follow-up was 19.7 years, and the overall annual incidence of bleeding was 1.3% with a rate of 2.6% in symptomatic aneurysms and 1% in incidental aneurysms. Increasing size, age at diagnosis (inversely), and cigarette smoking were significant predictors of outcome. They assert that surgery should be considered in young and middle-age patients; smoking cessation may be an alternative in the elderly. However, the small number of cases weakened the statistical power of the study

Various studies have concluded that most spontaneous SAHs are as a result of aneurysms that are 7 to 10 mm in diameter.^8,11 However, the ISUIA suggested that 10 mm is the critical size for rupture. This disagreement suggests that either there are a significantly higher number of 7 to 10 mm aneurysms, there is a decrease in aneurysm size at rupture, or that aneurysms that are more likely to rupture will do so at a smaller critical diameter. No clear evidence exists for any single theory.

Current evidence in natural history studies suggests that the risk of aneurysmal rupture increases with aneurysm size and the clinical presentation of the patient (previous SAH from different aneurysm, symptomatic vs. asymptomatic). Additionally, the risk is also likely influenced by location.¹⁷ Existing studies reflect at least some bias in patient selection, and many lack statistical power because of limited patient cohorts. Many others have suggested that a prospective, randomized study comparing conservative management and intervention would be helpful. Although such a study would be very appealing, it remains impractical. It would require years of follow-up, and because of low rupture rates, a vast number of patients would need to be enrolled to permit adequate statistical analysis and stratification for specific risk factors.

Recent reviews suggest that aneurysms are more prevalent and much less likely to rupture in older individuals, ^5,12 and that females are more likely to bleed.⁹ The natural history of UIAs is also modified by environmental factors such as cigarette smoking. Weir et al found that ~3500 patients from five prospective studies of aneurysm rupture had smoking rates 2.5 times higher than the expected rates based on European and U.S. national surveys.¹⁸ In another long-term study in Finland, active smoking status at the time of diagnosis was a significant risk factor for subsequent aneurysm rupture with a risk ratio of 1.46.¹⁶ Active smoking as a time-dependent covariant was the most important risk factor for rupture, with a risk ratio of 3.04. Based on these results, the authors conclude that the cessation of smoking may be a good alternative to surgery in some older patients.

The discussion above suggests that treatment over observation should be considered in several scenarios. These are outlined in a recent review published by the American Heart Association’s Stroke Council.¹² The AHA recommendations include a few keys points. The higher risk of treatment and shorter life expectancy of older individuals favors observation in the oldest patients with the smallest asymptomatic aneurysms. In other words, patient age is perhaps the most important factor in recommending treatment versus conservative management. For example, we would recommend treatment of even small aneurysms (i.e., 5 to 9 mm) in young patients, while even large aneurysms (i.e., >9 mm) might be treated conservatively in older individuals. Symptomatic intradural aneurysms of all sizes should be considered for treatment, especially those with new symptoms that should be treated with relative urgency. Coexisting aneurysms of any size in patients with prior SAH from another aneurysm should be considered for treatment. Asymptomatic aneurysms greater than 10 mm in size merit strong consideration for treatment. Special consideration for aneurysm treatment should be given for any young patients, patients with aneurysms with daughter sac formation or other unique hemodynamic factors, and patients with a family history of aneurysms or SAH.

Several studies have estimated the surgical morbidity and mortality rate of elective treatment of UIAs to be 0 to 4% and 0 to 7.2%, respectively.^19–22 The morbidity rates for giant aneurysms have been shown to be as high as 20%, with size as the key predictor of outcome. King et al in a meta-analysis combined 28 previous studies and in 733 patients found the overall mortality and morbidity rate to be 1% and 4.1%, respectively.²³

Although data suggest that both the mortality and morbidity rates may be higher than previously reported, the inclusion criteria of these studies remain controversial. Raaymakers et al in a meta-analysis of 2460 patients from 61 studies reported a mortality rate of 2.6% and a morbidity rate of 10.9%.²⁴ The higher rates maybe a result of inclusion of more symptomatic and larger aneurysms. Similar rates for morbidity and mortality were also supported by the prospective branch of the ISUIA study in which 1172 patients with newly diagnosed unruptured aneurysms were followed.¹⁵ Similar to the retrospective arm, patients were divided into two groups consisting of those without a prior history of SAH (Group 1 with 961 patients) and those with a prior history of SAH from a different aneurysm that had been successfully repaired (Group 2 with 211 patients). Within Group 1, 83% of patients underwent surgery, and the remaining patients were treated by endovascular coiling, with a resulting 1-month and 1-year mortality of 2.3% and 3.8%, respectively. Similarly, in Group 2, 94% of patients were treated surgically with a 1-month and 1-year mortality rate of 0% and 1%, respectively. Both groups were found to have a morbidity of ~12%. This is significantly higher than previously reported morbidity rates, which the ISUIA investigators attributed to more validated measures of neurological assessment, but their critics attributed to inadequate study design.

The surgical risks are modified by a variety of risk factors that include patient and aneurysm characteristics. Increasing age was found to be the only predictor of poor outcome in the prospective arm of ISUIA study, with combined morbidity and mortality rates doubling among the older age groups; 6.5% for patients younger than 45, 14.4% for patients 45 to 64, and 32% for patients older than 65 years of age.¹⁵ These results should be considered in older patients who have a shorter risk period but a higher surgical morbidity. Any other presurgical comorbidities may also affect the outcome. In the ISUIA, it remains unclear whether a cohort with similar comorbid conditions managed conservatively might not also have a similar decrease in the Rankin score.

There is evidence to suggest that aneurysm size influences surgical risk, with larger aneurysms having greater surgical risks than smaller ones.^19,20,25 Additionally, both aneurysm morphology and location have been found to affect outcomes secondary to the need for greater microsurgical expertise. Those with atherosclerotic or calcified walls are considered to pose a significant risk for ischemic complications, as shown in one study in which among six patients, there was a 50% rate of ischemic complication following clipping.²⁶ Partially thrombosed vessels may be a source of emboli if not treated with great care. Additionally, large, poorly defined aneurysms, those located in the posterior fossa, those involving major vessel bifurcations, those partially located in the cavernous sinus, or ones originating from the basilar artery have been associated with worse outcomes.^3,24,27 The results, however, are inconsistent; other studies have found that location is only important in the case of giant aneurysms²⁰ and that surgical risk may be no higher in posterior circulation UIAs.²⁸

Unruptured aneurysms presenting with symptoms of mass effect or cerebral ischemia also carry a greater surgical morbidity as compared with asymptomatic aneurysms.^20,25 Symptoms can be effectively treated in many cases to help relieve the mass effect and are considered an indication for intervention in appropriately healthy patients.²⁷

Surgical clipping is associated with low rates of postoperative aneurysm rupture and has traditionally been considered highly efficacious. However, the results are not absolute, as shown by studies that have considered both the incidence of residual aneurysm necks and subsequent rates of SAH following clipping. David et al evaluated 102 patients with 160 surgically treated aneurysms with late angiographic follow-up at a mean of 4.4 years postsurgery.²⁹ Of 135 clipped aneurysms that initially demonstrated no residual necks, 1.5% later exhibited recurrence. Of the 12 aneurysms with known residual necks, 25% enlarged during the follow-up period. Among the 102 patients, one SAH developed, resulting in a risk of hemorrhage of 1.9% per year. Eight new aneurysms were also found in six patients, resulting in a de novo aneurysm-formation rate of 1.8%. This data confirm the long-term efficacy of clipping with an aneurysm obliteration rate of 98.5%. However, even with surgery, residual aneurysm necks do occur and pose a risk for SAH.

In another study, 115 patients with unruptured aneurysms were surgically treated and followed for a median of 8.8 years.²¹ Four patients suffered SAH, two from de novo aneurysms, one from regrowth of a clipped aneurysm, and one from rupture of a wrapped aneurysm. The cumulative risk for SAH after treatment was 1.4% at 10 years, which is low but requires consideration.

Surgical experience of both the surgeon and the hospital may affect postoperative outcomes. At the Cleveland Clinic, 449 aneurysms in 366 patients were clipped by 10 different surgeons.³⁰ At the 6-month follow-up there was found to be a statistically significant relationship between the surgeon and patient outcome. Similarly, there is thought to be a direct relationship between the volume of craniotomies per year for aneurysms at an institution and better outcomes. Hospitals that performed 10 or more per year had an overall mortality rate of 5.3%, and those with fewer than 10 craniotomies per year had a mortality rate of 11.2%.³¹ When assessing patients for treatment of unruptured aneurysms, it is of paramount importance for the individual surgeon to realize their technical capabilities and treat patients accordingly.

Over the past decade, endovascular coiling has been increasingly utilized in both the initial and adjunctive management of UIAs. With improvements in technology and technique, the complication rates have been steadily declining and the treatment results steadily improving. In a systematic review of 48 studies with 1383 patients from 1990 to 1997, the permanent complication rate was 3.7%,³² and mortality rates were from 0.05 to 1%.^33,34 However, despite the low morbidity rate the incomplete obliteration rate has been found to be as high as 46%.³² Based on such high recanalization rates and the low rates of UIA rupture, critics suggest that endovascular coiling may not alter the natural history of UIAs.

There are multiple factors that make the utility of embolization treatment difficult to assess. Current studies that compare coiling versus surgery are not randomized; furthermore, they compare unequal patient groups. For example, the initial patient population undergoing endovascular treatment was traditionally older and more medically frail. Alternatively, a patient with a posterior fossa giant aneurysm randomized to a surgical group would certainly face higher risks for morbidity. The rapid advances in endovascular therapies make it difficult to standardize techniques within a given study.

The safety of endovascular treatment was reviewed in recent retrospective studies by Johnston et al, who demonstrated that coiling of UIAs resulted in lower rates of morbidity and mortality when compared with surgery.^34,35 Although the safety of coiling has been demonstrated, its efficacy still requires further investigation. In a single institutional review of 42 unruptured aneurysms, Wanke et al reported a complete or near complete occlusion in 34 of 38 patients.³⁶ Similarly, a larger review of 115 patients with a 120 aneurysms located in a variety of locations and with varying sizes and morphology reported a 91% rate of complete or near complete aneurysm obliteration.³⁷ However, angiographic followup in this study revealed that 32% of the patients with a small neck remnant demonstrated recanalization. This suggests that an aneurysm that undergoes complete obliteration has a much better long-term outcome than one that undergoes only near complete obliteration. Some critics even argue that near complete obliteration may not change the natural history. In the follow-up study by Hayakawa et al the incompletely embolized aneurysms were further scrutinized for rates of progressive thrombosis and rates of recanalization.³⁸ The overall progressive thrombosis rate was 25%, with higher rates in small aneurysms with small necks. The overall recanalization rate was 49%; with large and giant aneurysms, the rate approached 90%. This suggests that although small aneurysms with small necks and, in some cases, wide necks can be successfully treated with endovascular coiling, larger aneurysms may not be adequately secured with this treatment and may need to be considered for surgery.

The University of California, Los Angeles (UCLA)-based interventional group discussed above published a study of their 11-year experience with aneurysm coiling in which they reviewed 818 patients harboring 916 aneurysms.³⁹ The patients were divided into Group A, which was the initial 5 years’ experience, with 230 patients harboring 251 aneurysms, and Group B, which was the later 6 years’ experience, with 588 patients harboring 665 aneurysms. The rate of complete embolization was higher in Group B than in Group A (56.8% vs. 50.2%), and the rates of recanalization were lower in Group B than in Group A (17.2% vs. 26.1%). As in previous studies, the rate of recanalization was related to the size of the aneurysm dome and neck. Additionally, the overall rate of aneurysm rupture was 1.6%, but within the past 5 years, that rate had fallen to 0.5%. This study demonstrates that both the efficacy and safety of aneurysm embolization have improved significantly over the past decade but that recanalization remains a major limitation of this technique. Recent advances in embolization technology including neurostents and bioactive coils may improve on the problems of aneurysm recanalization seen with traditional platinum coils, but these studies are still pending. Based upon the studies detailed above, the appropriate treatment option requires an adequate assessment of various characteristics, the most important of which is the morphology of the aneurysm. Consideration should also include the location and size of the aneurysm as well as the age and medical condition of the patient.