| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Editorials |
1 From the Department of Radiology, R-109, West Wing 2, Rm 279, University of Miami School of Medicine, 1600 NW 10th Ave, Miami, FL 33136. Received December 21, 1998; accepted January 4, 1999. Address reprint requests to the author.
Index terms: Aneurysm, intracranial, 17.73 • Aneurysm, rupture, 17.73 • Aneurysm, surgery, 17.1267 • Aneurysm, therapy, 17.1264, 17.1267 • Editorials • Interventional procedures, utilization, 17.1264
The article by Vanninen et al (1) in the current issue of Radiology is important for a variety of reasons. The first and foremost reason is the study design. To my knowledge, the study by Vanninen et al is the first prospective, randomized comparison between embolization with the Guglielmi detachable coil (GDC) and surgery for the treatment of acutely ruptured intracranial aneurysm. A second reason is related to the selection criteria, the essential criterion being that a given aneurysm could be successfully treated with either technique. A third reason is the comparability of the clinical outcomes at 3 months. The results of this study indicate that endovascular treatment should be considered an effective alternative to surgery rather than a secondary treatment for difficult aneurysms or in sick patients. The importance of this study, for both the current treatment of aneurysms and the evolution of endovascular therapy, demands a critical analysis relative to studies reported previously in the surgery and radiology literature.
Previous multicenter studies of the use of GDCs were neither prospective nor randomized. These studies included patients with aneurysms considered to be too difficult for a neurosurgeon to clip without the potential for substantial morbidity. Hence, 57% of the 403 patients treated with the GDC at eight centers in the United States, reported by Viñuela et al (2), had aneurysms of the posterior circulation. The expected number is 8%, according to Kassell et al (3,4), who reported the results of the International Cooperative Study on the Timing of Aneurysm Surgery, which represent the most widely accepted database of the epidemiology of aneurysms.
These nonrandomized studies also included a large number of patients judged to be very sick on the basis of their level of consciousness and neurologic status. In the study by Vanninen et al, the patients in the endovascular treatment and surgical treatment groups were equal in terms of aneurysm location and size, amount (according to the Fisher scale) of subarachnoid hemorrhage (SAH), and neurologic grade (according to the Hunt and Hess classification). These variables also were similar to those detailed in the international cooperative study (3,4). Hence, the treatment groups in the report by Vanninen et al represent "reality."
The most important selection criterion in the study by Vanninen et al was that a given aneurysm must have been treatable by using either a surgical or an endovascular approach, as determined on the basis of the combined expertise of a neurosurgeon and an interventional neuroradiologist. Of 242 consecutive patients with proved aneurysmal SAH, 111 (46%) qualified as suitable candidates for random assignment to endovascular or surgical treatment. However, only 33 (25%) of the remaining 131 patients were excluded because of an aneurysmal neck that was considered to be too wide for GDC therapy. The rest were excluded because of age, time since bleeding, presence of a hematoma that necessitated surgery, and other less common situations. In fact, Vanninen et al were extremely liberal with regard to the criterion of neck size: It need only be equal to or smaller than the width of the aneurysmal fundus. Other investigators (5–7) have suggested that successful endovascular obliteration of an aneurysm is greatly facilitated when the fundus-to-neck size ratio is no less than 2:1 or even 3:1.
Why have there been no previous prospective, randomized studies of the effectiveness of GDC embolization relative to that of surgery? Why has the GDC been reserved for the very difficult aneurysm or the very sick patient? Neurosurgeons pride themselves on their success rate in terms of complete aneurysmal obliteration, which has been reported (8,9) to be as high as 96%. The persistence of a small remnant after incomplete surgical obliteration may result in a substantial repeat hemorrhage rate (9,10). Feuerberg et al (11) estimated the risk of bleeding due to a remnant to be 0.8% per year. Even in experienced hands, the complete obliteration rate is 71% in those aneurysms that are most amenable to obliteration with GDCs (2)—namely, small aneurysms (<10-mm diameter) with small necks (<4-mm diameter, without regard to a fundus-to-neck ratio). Remnants are even more common in large (10–25-mm, 57% obliteration rate) and giant (>25-mm, 50% obliteration rate) aneurysms (2). Therefore, given the lower percentage of complete obliteration with the GDC and the high hemorrhage rate due to a surgical remnant, many neurosurgeons have been reluctant to offer GDC embolization as an equal alternative to surgery.
Is the natural history of a GDC embolization remnant the same as that of a surgical remnant? Malisch et al (12) suggested that the bleeding rate from a remnant after GDC embolization may differ from that after surgery. Patients in whom there was a small neck remnant after GDC embolization were followed up for 2 to 6 years (average, 3.5 years). The rate of bleeding was 0% in patients with a small aneurysm, 4% in those with a large aneurysm, and 33% in those with a giant aneurysm.
It can be argued that an average follow-up of 3.5 years is too short to determine the potential for repeat hemorrhage. In the surgical studies (9–11), patients were followed up for many years to determine the high repeat hemorrhage rate. However, the GDC mass may change the hemodynamics of aneurysmal inflow and outflow and, secondarily, the sheer stresses on the walls of the remnant. It can also be theorized that there is more elasticity in a GDC, so that the hemodynamic forces in a remnant may lead to compression of the coil rather than growth and rupture of the aneurysm. Coil compression can be detected at follow-up angiography, and repeat aneurysmal packing can be performed with essentially no morbidity or mortality (12), whereas repeat surgery to complete aneurysmal obliteration has a morbidity rate of 7% (9) or higher.
It can be argued, therefore, that endovascular therapy should be viewed as a feasible alternative to surgery, given the knowledge that the rate of complete obliteration is lower with GDC embolization but that the remnant has a relatively low potential for catastrophe and can be reembolized with no morbidity. For the acutely ruptured aneurysm, GDC embolization will produce aneurysm stabilization to prevent rebleeding and allow treatment of vasospasm, with a complication rate that is no higher than that which would occur if the aneurysm was embolized later (2). If there is a remnant, repeat embolization can be performed as soon as the patient is clinically more stable.
The potential for complete obliteration of an aneurysm, plus the morbidity associated with a remnant, must be balanced against the morbidity and mortality due to the initial therapeutic procedure used to achieve that goal—that is, safety versus effectiveness. Surgical morbidity and mortality in cases of an acutely ruptured intracranial aneurysm are dependent on a number of factors, including aneurysm location, neurologic condition, amount of SAH, patient age, concomitant systemic disease, and the timing of surgery after hemorrhage.
The morbidity and mortality due to surgery must be separated from that due to the disease itself, including the neurologic effects of the initial hemorrhage, the morbidity and mortality due to rebleeding while awaiting surgery, and the ischemia due to vasospasm. Numerous authors (3,4,13,14) have shown the benefits of early surgery (within 3 days of hemorrhage) to prevent rebleeding and allow treatment of vasospasm, thereby markedly increasing the number of patients with a good outcome at 6 months. Surgical morbidity and mortality vary according to the proportion of patients with poorer neurologic grades. Ohman and Heiskanen (14) reported a 6% mortality rate and an additional 3% morbidity rate for all neurologic grades in the early surgery group in their randomized study.
These results (3,4,13,14), however, do not take into account the neuropsychologic and cognitive impairments that may be produced by the surgical procedure. The results of a recently published study (15) of the natural history of and the surgical morbidity and mortality associated with unruptured aneurysms show that surgery itself produces important cognitive changes. At 1 month after surgery in patients without any previous aneurysmal hemorrhage or surgery, there were cognitive impairments in 5.5% of the patients, neurologic abnormalities in 3.6%, and both in 6.1%, for an overall morbidity rate of 15.3%. The neurologic impairments improved somewhat during the next year, but the degree of the cognitive impairments did not. Cognitive impairments were even more marked in a separate group of patients who had had another aneurysm that had ruptured and been clipped. The 1-year mortality rate in the group without any previous hemorrhage or surgery was 3.8% (15). Thus, even in the best of situations, surgical treatment for an aneurysm produces substantial cognitive impairment, neurologic morbidity, and mortality.
The complication rate associated with GDC embolization was 9% in the series of Viñuela et al (2), in which embolization was performed at variable times (up to 15 days after hemorrhage), but not all complications resulted in permanent deficits. The mortality rate directly related to these complications was 1.75%. The overall morbidity rate in these patients was 9%, but it was difficult to separate the morbidity related to the procedure from that due to the disease. Therefore, GDC embolization has a lower mortality rate and, most likely, a lower morbidity rate than does surgery. This difference is all the more dramatic given that GDC embolization usually is reserved for more difficult aneurysms in sicker patients. What might the outcomes be if the populations of surgical and endovascular treatment candidates were similar?
It is obvious that a comparative randomized trial has been needed. Such a trial would be justified because the higher rate of aneurysmal remnants after GDC embolization is balanced by the lower morbidity and mortality associated with endovascular therapy and the ability to reembolize those remnants at follow-up. Enter the study by Vanninen et al. How effective was GDC embolization relative to surgery and to historical results?
The initial rate of complete obliteration with GDC embolization was 61% in small aneurysms with small necks (1), as compared with the rate of 71% in the study by Viñuela et al (2), but spontaneous thrombosis of four small remnants and repeat embolization of three others increased the rate in the study by Vanninen et al to 66% at 3 months. Vanninen et al allowed the inclusion of aneurysms with a neck width equal to the fundal width, and such a liberal criterion for neck width might have led to the lower rate of complete obliteration. Viñuela et al used 4 mm as the division between a large and a small neck and found a significant difference between the small- and large-necked aneurysm groups in terms of the success of obliteration. In the study of Viñuela et al, 54% of the patients had a small-necked aneurysm, but in the study by Vanninen et al, even more of the patients (76%) had a small-necked aneurysm. Hence, the liberal criterion for neck size in the Vanninen et al study probably was not a factor in the lower percentage of complete aneurysmal obliteration after GDC embolization.
Surgery resulted in a total obliteration rate of 74%; 9% of the patients had a large remnant at postoperative angiography (two were completely obliterated with GDCs), and 16% had a small remnant (after one patient had undergone GDC closure of that remnant) (1). This surgical obliteration rate was low compared with rates reported in the literature (8,9), and it was not much higher than the rate achieved with GDC embolization. Surgery was significantly better than GDC embolization for closure of anterior cerebral artery aneurysms (P = .005), but GDC embolization was superior for posterior circulation lesions, and surgical and endovascular treatment were equally successful for middle cerebral artery and internal carotid artery aneurysms (1).
There was a 17% rate of technical complications associated with GDC embolization in the study by Vanninen et al, including a 6% perforation rate, which is high compared with that in the study by Viñuela et al (2). However, only three of the technical complications necessitated surgery or resulted in neurologic problems, which amounted to 6% of the group and was comparable to the overall surgical morbidity and mortality rate of 7%. The procedure-related mortality rate was 4% in the surgery group and 2% in the endovascular treatment group, which compares favorably with results reported in the literature (2,4,15). It is important to note that the 3-month clinical outcome, as measured with the Glasgow Outcome Scale, was the same for both groups. Hence, the effectiveness of the procedure and the morbidity and mortality rates in the two groups were similar.
There are three important considerations that temper these interesting results. First, the rate of complete obliteration after surgery was low compared with that reported in the literature (1,8,9). The balance between the safety of GDC embolization and the effectiveness of surgery is easily tilted toward embolization, if surgical effectiveness is at this level.
Second, no neuropsychologic or cognitive testing was performed in these patients, but such assessments will reportedly be performed at future follow-up. The issue of whether surgery (in comparison with relatively less invasive endovascular therapy) imposes cognitive and neuropsychologic insults in addition to those caused by the hemorrhage is an extremely important one that may weigh heavily in the balance. The authors of a recent study (15) of unruptured aneurysms strongly suggest this to be true. Additional neuropsychologic morbidity due to surgery would make GDC embolization appear to be even more appealing.
Third, 3 months is too short a time to evaluate the potential for rebleeding from aneurysmal remnants. Results from previous noncomparative trials (2,6,12) suggest that the frequency of such remnants is substantially higher with GDC embolization than with surgery. The higher than expected frequency of remnants in the surgery group in the study by Vanninen et al is another factor that changes the balance. The rate of hemorrhage in these remnants will have to be determined with long-term follow-up. The potential for endovascular treatment of the remnants in either group may decrease the importance of such remnants (16).
The techniques for endovascular treatment of aneurysms are evolving. Coil embolization of a wide-necked aneurysm may be aided by using a balloon that is temporarily inflated across the mouth of the aneurysm. A three-dimensional configuration of the GDC is about to be released by the manufacturer (Boston Scientific/Target, Natick, Mass). The potential for filling aneurysms with other materials has always been apparent. A flexible stent may be developed that would allow intracranial passage to bridge a wide aneurysmal mouth. These and other developments surely will lead to more aneurysms being treated primarily by endovascular means.
Even without these expected developments, the comparability of the outcomes and the procedure-related morbidity in the GDC embolization and surgical groups in the study by Vanninen et al may give added support to those who advocate that the initial attempt at aneurysm obliteration should be with endovascular techniques in the majority of patients. If endovascular treatment fails or a hematoma must be evacuated, the surgical option can be exercised. Most important, the comparability of these results supports the ethical appropriateness of performing more prospective randomized studies.
Footnotes
See also the article by Vanninen et al (pp 325–336 ) in this issue.
References
Related Article
This article has been cited by other articles:
|
|
 |
|
  T. Koivisto, R. Vanninen, H. Hurskainen, T. Saari, J. Hernesniemi, and M. Vapalahti Outcomes of Early Endovascular Versus Surgical Treatment of Ruptured Cerebral Aneurysms : A Prospective Randomized Study Stroke, October 1, 2000; 31(10): 2369 - 2377. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Wilms, F. v. Calenbergh, L. Stockx, P. Demaerel, J. van Loon, and J. Goffin Endovascular Treatment of a Ruptured Paraclinoid Aneurysm of the Carotid Syphon Achieved Using Endovascular Stent and Endosaccular Coil Placement AJNR Am. J. Neuroradiol., April 1, 2000; 21(4): 753 - 756. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
