← Back

ISAT

International Subarachnoid Aneurysm Trial (ISAT) of neurosurgical clipping versus endovascular coiling in 2143 patients with ruptured intracranial aneurysms: a randomised comparison of effects on survival, dependency, seizures, rebleeding, subgroups, and aneurysm occlusion

Year of Publication: 2005

Authors: Andrew J Molyneux, Richard S C Kerr, Ly-Mee Yu, ..., for the International Subarachnoid Aneurysm Trial (ISAT) Collaborative Group

Journal: Lancet

Citation: Lancet 2005; 366: 809-17

PDF: https://www.researchgate.net/profile/And...saWNhdGlvbiJ9fQ

Clinical Question

In patients with ruptured intracranial aneurysms suitable for both treatments, does a policy of endovascular detachable-coil treatment compared with a policy of neurosurgical clipping reduce the proportion of patients who are dead or dependent at 1 year?

Bottom Line

In patients with ruptured intracranial aneurysms suitable for both treatments, endovascular coiling is more likely to result in independent survival at 1 year than neurosurgical clipping, with an absolute risk reduction of 7.4% (95% CI 3.6-11.2, p=0.0001) and relative risk reduction of 23.9% (12.4-33.9). The survival benefit continues for at least 7 years (log rank p=0.03). The risk of epilepsy is substantially lower with endovascular treatment (RR 0.52, 95% CI 0.37-0.74). The risk of late rebleeding is low (0.2% per patient-year after first year) but slightly more common after endovascular coiling than after neurosurgical clipping. At first follow-up angiography, 66% of coiled aneurysms showed complete occlusion compared to 82% of clipped aneurysms (in selected patients who had follow-up angiography).

        Major Points
        Multicenter randomized controlled trial at 42 neurosurgical centres, mainly in UK and Europe
2143 patients with ruptured intracranial aneurysms randomized between recruitment closure
Patients eligible only if responsible neurosurgeon and neuroradiologist were uncertain about best treatment (equipoise)
88% of patients in good clinical grade (WFNS 1-2) at enrollment
95% of aneurysms in anterior circulation, 90% smaller than 10 mm
Primary outcome: death or dependency (mRS 3-6) at 1 year
At 1 year: 23.5% dead or dependent in endovascular group vs 30.9% in neurosurgery group (RR 0.76, 95% CI 0.66-0.87, p=0.0001)
Absolute risk reduction 7.4% (3.6-11.2), relative risk reduction 23.9% (12.4-33.9)
Case fatality at 1 year: 8.0% endovascular vs 9.9% neurosurgery
Survival advantage maintained for up to 7 years (log rank p=0.03)
Mean follow-up 4 years, with 6542 patient-years of follow-up after 1 year
Rebleeding from target aneurysm after 1 year: 7 patients endovascular vs 2 neurosurgery (0.2% per patient-year)
Seizures significantly reduced with endovascular treatment (RR 0.52, 95% CI 0.37-0.74)
Follow-up angiography in 881 of 988 surviving endovascular patients (89%): 66% complete occlusion, 26% neck remnant/subtotal, 8% incomplete
Follow-up angiography in 450 of 965 surviving neurosurgery patients (47%): 82% complete occlusion, 12% neck remnant, 6% incomplete
Treatment effect heterogeneous by age, WFNS grade, and aneurysm location, but no robust evidence that neurosurgery has advantages for any subgroup
Trial closed early after interim analysis showed benefit of endovascular treatment

      

Design

Study Type: Prospective randomized controlled trial, partially blinded

Randomization: 1

Blinding: Outcome assessors (certified examiners and neuropsychologist) were unaware of treatment assignments. Blinded assessment of rebleeding not possible because clips, coils, and craniotomy evidence visible on CT scans. Two independent investigators (neurosurgeon and neuroradiologist) judged all rebleeding and intracranial hemorrhage events.

Enrollment Period: Not explicitly stated in excerpts, but recruitment closed after interim analysis; trial enrolled 2143 patients

Follow-up Duration: 1 year primary outcome, with continued follow-up to 7 years (mean 4 years)

Centers: 42

Countries: United Kingdom, Europe, Other international sites

Sample Size: 2143

Analysis: Intention-to-treat analysis. Kaplan-Meier method used to analyze time to death, with log-rank test to compare mortality. Statistical test of interaction to assess treatment effect consistency across prespecified subgroups. Relative risks describe direction and magnitude of treatment effect. Pre-specified subgroups: WFNS grade, age by decade, Fisher grade, aneurysm location, aneurysm lumen size.

Inclusion Criteria

Subarachnoid hemorrhage due to intracranial aneurysm
Aneurysm suitable for both endovascular and neurosurgical treatment
Uncertainty by responsible neurosurgeon and neuroradiologist about best treatment (equipoise)
Consensus that either technique would be suitable treatment option
Consensus that it was uncertain whether ruptured aneurysm should be treated by neurosurgical or endovascular means
Written informed consent from able patients or assent from relatives for patients who could not give written consent (where ethics committees allowed)

Exclusion Criteria

Insufficient uncertainty about best treatment (neurosurgeon and neuroradiologist must both be uncertain)
Aneurysm not suitable for either treatment modality
Aneurysm clearly more suitable for one specific treatment approach

Baseline Characteristics

Characteristic	Control	Active
Total patients	1070	1073
WFNS Grade 1-2 (good grade)	88%	88%
Anterior circulation aneurysm	95%	95%
Aneurysm size <10 mm	90%	90%
Age distribution	Distributed across decades from <40 to ≥70 years	Distributed across decades from <40 to ≥70 years

Arms

Field	Control	Endovascular Coiling
Intervention	Craniotomy and surgical clipping of ruptured intracranial aneurysm. Of 1070 patients allocated to neurosurgery, 19 died before first procedure, 39 had coiling as first procedure (crossover), 7 treated conservatively, and 2 missing information. Of first procedures performed: 977 clipped (96.5%), 13 wrapped (1.3%), 14 not completed (partial clipping or wrapping, 1.4%), 8 not attempted (0.8%). Total 1012 first procedures. Performed by experienced neurosurgeons at 42 centers.	Endovascular coil embolization using detachable platinum coils (Guglielmi detachable coil system). Of 1073 patients allocated to endovascular treatment, 7 died before procedure, 9 underwent clipping as first procedure (crossover). Of first procedures performed: 1014 completed (92.6%), 29 failed to catheterize target aneurysm (2.6%), 37 aneurysm catheterized but anatomy unsuitable (3.4%), 15 not attempted (1.4%). Total 1095 first procedures. Performed by experienced interventional neuroradiologists at 42 centers. Follow-up angiography requested in all patients at approximately 6 months.
Duration	Acute treatment with long-term follow-up to 7 years	Acute treatment with long-term follow-up to 7 years

Outcomes

Outcome	Type	Control	Intervention	HR / OR / RR	P-value
Death or dependency defined by modified Rankin Scale (mRS) score of 3-6 at 1 year	Primary	326 of 1055 patients (30.9%)	250 of 1063 patients (23.5%)	7.40%	0.0001
mRS 0 (no symptoms) at 1 year	Secondary	187 of 1055 (17.7%)	260 of 1063 (24.5%)
mRS 1 (minor symptoms) at 1 year	Secondary	292 of 1055 (27.7%)	301 of 1063 (28.3%)
mRS 2 (some restriction in lifestyle) at 1 year	Secondary	250 of 1055 (23.7%)	252 of 1063 (23.7%)
mRS 0-2 (independent) at 1 year	Secondary	729 of 1055 (69.1%)	813 of 1063 (76.5%)
mRS 3 (significant restriction) at 1 year	Secondary	141 of 1055 (13.4%)	107 of 1063 (10.1%)
mRS 4 (partly dependent) at 1 year	Secondary	42 of 1055 (4.0%)	30 of 1063 (2.8%)
mRS 5 (fully dependent) at 1 year	Secondary	38 of 1055 (3.6%)	28 of 1063 (2.6%)
mRS 6 (death) at 1 year	Secondary	105 of 1055 (9.9%)	85 of 1063 (8.0%)
Cumulative mortality to 7 years	Secondary	Higher mortality (log rank p=0.03)	Lower mortality		0.03
Death or dependency at 2 months (mRS 3-6)	Secondary	392 of 1063 (36.9%)	278 of 1065 (26.1%)	RR 0.71, 95% CI 0.62-0.80	<0.0001
Rebleeding before first procedure	Secondary	28 patients (19 deaths)	17 patients (7 deaths)	RR 0.60, 95% CI 0.33-1.10
Rebleeding after first procedure and before 30 days	Secondary	8 patients (4 deaths)	20 patients (9 deaths)	RR 2.46, 95% CI 1.09-5.57
Rebleeding 30 days to 1 year	Secondary	3 patients (1 death)	8 patients (6 deaths)	RR 2.64, 95% CI 0.70-9.93
Total rebleeding during first year	Secondary	39 patients	45 patients	RR 1.15, 95% CI 0.75-1.75
Rebleeding after 1 year from target aneurysm	Secondary	2 patients (2 deaths)	7 patients (2 deaths)	0.2% per patient-year
Cumulative rebleeding to 7 years	Secondary	Lower rate	Higher rate	Log rank p=0.22 (not significant)	0.22
Seizures before first treatment	Secondary	11 patients (6 deaths)	3 patients (3 deaths)
Seizures after procedure to before discharge	Secondary	33 patients (2 deaths)	16 patients (2 deaths)
Seizures discharge to 1 year	Secondary	44 patients	27 patients (1 death)
Seizures after 1 year	Secondary	24 patients	14 patients (1 death)
Overall seizure risk after first procedure	Secondary	Higher risk	Lower risk	RR 0.52, 95% CI 0.37-0.74	<0.001
Complete aneurysm occlusion on first follow-up angiography	Secondary	370 of 450 with angiography (82%)	584 of 881 with angiography (66%)
Neck remnant or subtotal occlusion on follow-up angiography	Secondary	55 of 450 (12%)	228 of 881 (26%)
Incomplete occlusion on follow-up angiography	Secondary	25 of 450 (6%)	69 of 881 (8%)
Death 2-12 months (complication of severe dependent survival)	Adverse	15 patients	7 patients
Death after 1 year (complication of severe dependent survival)	Adverse	5 patients	1 patient
Death 2-12 months from treated aneurysm rebleeding	Adverse	1 patient	1 patient
Death after 1 year from treated aneurysm rebleeding	Adverse	2 patients	2 patients
Death from another aneurysm hemorrhage after 1 year	Adverse	0 patients	3 patients
Death from cardiac causes after 1 year	Adverse	10 patients	5 patients
Death from cancer after 1 year	Adverse	12 patients	9 patients
Death from suicide after 1 year	Adverse	1 patient	2 patients

Subgroup Analysis

Treatment effect was heterogeneous by age (p=0.04), WFNS grade (p=0.01), and aneurysm location (p=0.01). However, no robust evidence that neurosurgery has advantages over endovascular treatment for any subgroup. Endovascular treatment showed benefit across all aneurysm locations. For good grade patients (WFNS 1-3), RR 0.71 (95% CI 0.61-0.83); for poor grade (WFNS 4-6), RR 1.11 (95% CI 0.84-1.46). Age showed heterogeneity but no consistent trend. Small numbers in extreme age groups limit interpretation. Fisher grade (blood amount on CT) showed no significant interaction (p=0.3). Aneurysm lumen size showed no significant interaction (p=0.4).

Criticisms

Only patients for whom there was uncertainty about best treatment were enrolled - approximately 80% of screened aneurysms excluded
Highly variable enrollment rates between centers
Selection bias concerns - patients clearly more suitable for one treatment were not randomized
Questions about whether surgical expertise represented the best that neurosurgery had to offer
Applicability concerns - results may not generalize to all patients with subarachnoid hemorrhage
More patients in surgical group had rebleeding before treatment (23 vs 14)
Follow-up angiography not mandatory after neurosurgical clipping (only 47% had angiography), creating potential selection bias in angiographic outcome comparison
Blinded assessment of rebleeding not possible due to visibility of clips and coils on imaging
Long-term durability of coil occlusion uncertain beyond 7 years
Angiographic occlusion rates better with clipping (82% complete) than coiling (66% complete)
Small numbers in some subgroups (elderly, poor grade) limit interpretation
Trial predominantly enrolled good-grade patients (88% WFNS 1-2) and small anterior circulation aneurysms (90% <10mm)
Limited data on giant or complex aneurysms
No data on cost-effectiveness presented in this publication

Funding

Pilot phase: Oxford Regional Health Authority Research and Development (1994-97). Main trial: Medical Research Council UK; Programme Hospitalier de Recherche Clinique 1998 of French Ministry of Health (AOM 98150) sponsored by Assistance Publique-Hôpitaux de Paris (AP-HP); Canadian Institutes of Health Research; Stroke Association of UK (for neuropsychological assessments)

Based on: ISAT (Lancet, 2005)