Purpose Aneurysm recurrence following coil occlusion is well recognized. However, there is controversy as to how long these patients should be followed up after coiling to detect reopening. We aimed to identify the rate of late reopening and the risk factors for reopening in a large single-center cohort of ruptured aneurysms that appeared adequately occluded at 6 months. We also aimed to assess whether rates of recurrence have altered over time with improving coil and angiographic technology.
Methods Patients treated between 1996 and 2010 were assessed and those with both 6-month initial and subsequent long-term follow-up with either digital subtraction angiography or magnetic resonance angiography were included. Aneurysms were stratified by features such as size, neck width, anatomical location and time of treatment: 1996–2005 (cohort 1) and 2006–2010 (cohort 2). ORs for risk of recurrence were calculated for aneurysm features and rates of recurrence in each cohort were compared using a χ2 test.
Results 437 patients with 458 adequately occluded aneurysms at 6 months had mean long-term follow-up of 31 months; 57 (12.4%) were large (≥10 mm) and 104 (22.7%) were wide-necked (>4 mm). Nine aneurysms (2%) showed significant late anatomical deterioration whereby retreatment was considered or undertaken. The risk was greater for large aneurysms (≥10 mm) (OR 15.61, 95% CI 3.79 to 64.33, p=0.0001) or wide-necked aneurysms (>4 mm) (OR 12.70, 95% CI 2.60 to 62.13, p=0.0017). The frequency of significant late anatomical deterioration and retreatment was also less common in those treated in cohort 2 (p<0.05). No completely occluded aneurysm at 6 months demonstrated significant late recurrence.
Conclusions Most aneurysms adequately occluded at 6 months did not show evidence of late recurrence. Large and wide-neck aneurysms are, however, at greater risk of later recurrence.
- Magnetic Resonance Angiography
Statistics from Altmetric.com
Endovascular coiling is the preferred treatment for ruptured cerebral aneurysms, but systematic review of the literature suggests that approximately 20% of aneurysms will show reopening or recurrence following treatment.1 Intuitively, the principal indication for following aneurysms with imaging after coil occlusion is to identify aneurysm recurrence (either through coil compaction, migration into thrombus or aneurysm regrowth) and to pre-empt rehemorrhage through retreatment. A number of studies have demonstrated an inverse relationship between the degree of occlusion and the risk of rerupture.2 ,3 However, there is little consensus in the literature as to the most appropriate timing of follow-up or length of follow-up, and long-term follow-up is not without detriment; there is evidence that those patients are more prone to anxiety and depression.4 Follow-up also brings with it a financial cost either to health systems or patients.
Although previous series with variable initial follow-up time have suggested that long-term follow-up is mandatory to detect late aneurysm reopening,5––7 recent evidence provided by a study using a fixed initial follow-up date of 6 months post-treatment in adequately occluded aneurysms suggests differently.8 In 400 patients with 440 aneurysms, the rate of aneurysm reopening was very low (2.6%) on long-term follow-up (mean 6 years) and consequently the risk of delayed rebleeding in adequately treated aneurysms is reported to be extremely low.2 ,9 ,10 An additional question is whether some aneurysms (eg, large, wide-necked or those at specific anatomical locations) are more prone to recurrence than others, even when adequately occluded at 6 months, and it may be that longer follow-up should be focused on these subgroups.
At our center we follow aneurysms routinely at 6 months and then at 30 months, but continue for longer until stability is achieved or a remnant is retreated. We aimed to assess the rate of late reopening in adequately occluded aneurysms (defined as those either completely occluded or with a small neck remnant at 6 months follow-up) in patients treated between 1996 and 2010 to assess whether late follow-up was necessary in all of these patients. We also aimed to assess whether reopening or retreatment rates for patients with adequately occluded aneurysms at initial follow-up have changed with evolving coil and angiographic technology.
Patient details were retrieved from two institutional databases containing all ruptured aneurysms coiled at our institution between 1996 and 2005 (cohort 1) and all those coiled between 2006 and 2010 (cohort 2). Only patients with both short- and long-term imaging follow-up were included in the primary study population. These patients were stratified into those with adequately occluded aneurysms (complete occlusion or only a small neck remnant) at 6-month angiographic follow-up and those with an aneurysmal remnant; this was determined by occlusion status as recorded on the neuroradiological reports. Subsequent long-term digital subtraction angiography (DSA) or magnetic resonance angiography (MRA) follow-up was then assessed for evidence of anatomical deterioration. Occlusion status was re-adjudicated by the study personnel through consensus between two neuroradiologists (AMM and SAR). Further retreatment was also recorded.
The endovascular procedure
All procedures were performed by consultant interventional neuroradiologists (SAR and MDB). The aim was to place coils sequentially into the aneurysmal sac to the point of angiographic occlusion. The vast majority of coils deployed were bare platinum (Boston Scientific (now Stryker), Kalamazoo, Michigan, USA; Micrus (now Codman Neurovascular), Raynham, Massachusetts, USA; ev3 (now Covidien), Dublin, Ireland). A minority of aneurysms in the study cohort were coiled with balloon remodeling and none with stent assistance. In only a small number of aneurysms were bioactive coils used. The coiling technique did not alter over the study period. The aneurysm responsible for hemorrhage was identified by blood distribution on CT, aneurysm appearance, and vasospasm distribution. If it was not possible to clearly identify the ruptured aneurysm, all possible candidates were treated and aneurysms were classified as ruptured. The degree of initial aneurysm occlusion was recorded on the basis of post-coiling angiography.
Imaging follow-up protocol
Prior to 2005 we followed aneurysms with DSA. After 2005 our institutional protocol changed to following with 3D time-of-flight (TOF) MRA, often favored as a non-invasive tool for detection of reopening.11 ,12 ,13 MRI examinations were performed on 1.5 T or 3 T systems (Philips Healthcare). The MRI protocol included axial T2-weighted fast spin echo and multiple overlapping thin slab acquisition 3D TOF MRA sequences. Axial images and triplanar maximum intensity projections were assessed. If the initial MRA was performed using 1.5 T, the subsequent scan would have been either 1.5 T or 3 T. If the initial scan was 3 T, subsequent scans were also at 3 T. A 3 T scan never preceded a 1.5 T scan.
Anatomical deterioration was defined as increased flow on DSA or MRA at the base or within the aneurysm, either caused by compaction of the coil mesh or by aneurysm growth, when compared with the initial MRA. The aneurysms were dichotomized between adequately occluded aneurysms (completely occluded or small neck remnant) and incompletely occluded aneurysms. Mild deterioration (but continued adequate occlusion) was classified as deterioration to a small neck remnant or mild enlargement of a neck remnant (see figure 1). Significant deterioration was classified as deterioration to aneurysm residuum (see figure 2).
The results were compared statistically (Medcalc V.12.4.0, Ostend, Belgium). ORs were used to compare patients with and without aneurysm reopening in terms of factors such as aneurysm size and neck width, early recanalization and location. Analysis of variance was used for comparison of age between the two cohorts.
χ2 test was used for statistical comparison of proportions for patient sex, aneurysm size ≥10 mm, neck size >4 mm, aneurysm location and rates of recurrence in each of the two cohorts selected on the basis of year of procedure.
A summary of patient selection is shown in figure 3. Between 1996 and 2010, 1106 patients with subarachnoid hemorrhage underwent endovascular coiling. The primary study population included 437 patients with 458 adequately occluded aneurysms at 6 months with mean long-term follow-up of 31 months. This population was subdivided into those aneurysms treated between 1996 and 2005 (cohort 1; 197 patients with 198 aneurysms) and those treated between 2006 and 2010 (cohort 2; 241 patients with 260 aneurysms).
Patients not included in the primary study population
A total of 669 patients were not included in the primary study population for the following reasons: (1) no angiographic follow-up; (2) follow-up at 6 months only; (3) aneurysms incompletely occluded at 6 months.
Two hundred and eighty-five patients had no follow-up; 163 patients died in the acute period and 122 had no follow-up for the following reasons: advanced age (37), severe clinical outcome (24), out of region follow-up (14), and other reasons (19). Twenty-eight patients refused follow-up.
Three hundred and three patients with 306 aneurysms had a single follow-up DSA or MRA available for assessment. The occlusion grades at initial follow-up in these aneurysms were complete occlusion 80.4% (246/306) and small neck remnant 19.3% (59/306). All but one patient were therefore adequately occluded. There was no subsequent follow-up available for reasons such as advanced age, out of region follow-up, severe outcome, and patient choice.
Eighty-one patients with 82 aneurysms were deemed incompletely occluded at initial 6-month follow-up. Of these, 70 had shown anatomical deterioration and 12 were stable but incompletely occluded. Wide neck and large size represented significant risk factors for incomplete occlusion at 6 months: 31 (37.8%) were large compared with 95 of the 764 (12.4%) adequately occluded aneurysms at initial follow-up (OR 4.28, 95% CI 2.61 to 7.03, p<0.0001) and 38 (46%) were wide-necked compared with 190 of the 764 (24.9%) adequately occluded aneurysms at initial follow-up (OR 2.61, 95% CI 1.64 to 4.15, p=0.0001). Fifty-five of these aneurysms (67% of incompletely occluded aneurysms as adjudged at 6 months) were retreated. Those that were not retreated were managed conservatively for reasons such as patient age, anatomical configuration, and estimated procedural risk.
Analysis of the primary study population: overall outcomes for adequately occluded aneurysms at 6 months
A total of 437 patients with 458 adequately occluded aneurysms at 6 months had mean long-term follow-up of 31 months. Two hundred and ninety-nine (68.4%) were women. Mean age was 50.3 years. Table 1 summarizes the aneurysm characteristics: 57 (12.4%) were large (≥10 mm) and 104 (22.7%) were wide-necked (>4 mm).
Immediate post-coiling angiography demonstrated adequate occlusion in 97.8%. In cohort 1, adequate occlusion was 97.4% (193/198 aneurysms) and, in cohort 2, adequate coiling was 98.1% (255/260 aneurysms). Therefore, although immediate angiography suggested some residual aneurysm in a minority, by definition, all went on to adequate occlusion by 6 months.
Overall, 28 aneurysms (6%) showed any anatomical deterioration between 6 months and 30 months of follow-up, but most remained adequately occluded at late follow-up with only nine (2.0%) showing significant deterioration to aneurysmal remnant whereby retreatment was either considered or performed. Six of the nine were retreated, one which was stable between 30 and 42 months opted for neurosurgical clipping, and the other two continue to be followed.
In the group that deteriorated to aneurysmal remnant, seven of nine (78%) were wide-necked (OR 12.70, 95% CI 2.60 to 62.13, p=0.0017) and six (60%) were large (OR 15.61, 95% CI 3.79 to 64.33, p=0.0001). The anatomical locations were as follows: two in the anterior communicating artery (22.2%), two in the posterior communicating artery (22.2%), two in the basilar tip (22.2%, OR 3.28, 95% CI 0.66 to 16.37, p=0.1479), and one each in the ophthalmic artery, posterior inferior cerebellar artery, and distal anterior cerebral artery.
Sixty-one of the 430 aneurysms (14.2%) that had long-term stability showed minor deterioration between the initial post-coiling angiogram and 6-month follow-up compared with five of the 28 aneurysms (17.9%) that did show late deterioration (OR 1.32, 95% CI 0.48 to 3.59, p=0.5930).
Outcomes for completely occluded aneurysms
In all, 290 aneurysms were completely occluded at 6 months (see table 2), 50 (17.2%) of which were wide-necked and 27 (9.3%) were large; 16 (5.5%) of these aneurysms showed deterioration to small neck remnant on late follow-up, but none of these aneurysms required retreatment. All remain adequately occluded. Three of these aneurysms were large (OR 2.40, 95% CI 0.64 to 9.03, p=0.1940), two were wide-necked, and six of the 16 were posterior communicating artery aneurysms (OR 2.50, 95% CI 0.87 to 7.19, p=0.0886). There have been no cases of delayed rebleeding from aneurysms completely occluded at 6 months initially treated from January 1996 to date.
Comparison of anatomical outcomes over time
A summary of the populations treated between 1996 and 2005 (cohort 1) and 2006 and 2010 (cohort 2) is shown in table 3. There was no significant difference in either patient or aneurysm characteristics. In cohort 1, 16 aneurysms showed late anatomical deterioration (8%). Seven showed significant deterioration to the aneurysm residuum and six (3%) were retreated. In cohort 2, 12 aneurysms showed late anatomical deterioration (4.6%). Only two of these cases (0.8%) showed significant deterioration, both stable on later follow-up, but one awaits neurosurgical clipping. The difference in the rate of significant late recurrence reached significance (p=0.0345), with a smaller number of significant deteriorations in cohort 2. The rate of retreatment was also significantly lower in cohort 2 (p=0.0221). Comparison of aneurysms that exhibited reopening versus those that did not within each cohort is displayed in table 4. In cohort 1, aneurysm size ≥10 mm significantly predisposed to any recurrence (OR 4.72, 95% CI 1.46 to 15.27, p=0.0097) and in cohort 2 aneurysm size ≥10 mm also increased the risk of recurrence significantly (OR 3.38, 95% CI 1.02 to 12.51, p=0.0474). In cohort 2, posterior communicating artery aneurysms tended to show more recurrence and anterior cerebral artery aneurysms showed less recurrence, but this was not reproduced in cohort 1 so the validity of this is questionable.
The results of this study suggest that the vast majority of patients with adequately coil-occluded cerebral aneurysms at initial follow-up of 6 months remain well occluded at long-term follow-up. Overall the rate of any reopening was 6%, but the majority of these remained well occluded and only 2% showed significant deterioration to aneurysmal remnant. Completely occluded aneurysms showed no cases of significant deterioration (5.5% showed mild deterioration but they remained well occluded), and the rate of significant deterioration in all adequately occluded aneurysms was lower in those aneurysms treated after 2006 with only one case in this cohort awaiting retreatment. Our results are broadly in agreement with those of a recent investigation of 400 patients with 440 aneurysms followed for a mean period of 6 years.8 This investigation found that the rate of aneurysm reopening after adequate initial occlusion identified at 6 months with 3 T TOF MRA was 2.6% (11 patients) and only three patients were retreated. In a series of 126 patients with fixed follow-up intervals of 6 and 18 months after coiling, all reopened aneurysms were found at 6-month angiography.14 In another study, 111 aneurysms adequately occluded at 6 months underwent 3 T MRA 5–11 years after coiling. Three showed minor reopening and only one aneurysm, which initially contained intraluminal thrombus, showed major reopening.15 A difference between these studies and our own is that they included patients with both ruptured and unruptured aneurysms whereas we confined our analysis to ruptured aneurysms that, in theory, may be more prone to recurrence due to a thinner more friable wall and associated thrombus. Our results suggest that this is not necessarily the case. Likewise, a study by Gallas et al10 reported that 96% of ruptured aneurysms completely occluded at 1 year remained stable, with a mean follow-up period of 36 months. They also reported that no patient with a completely occluded aneurysm at 2 years demonstrated recanalization at 3 years.
Several studies5 ,6 ,7 ,12 have reached conclusions that contradict the findings of our study and that of Ferns et al.8 In these studies, prolonged imaging follow-up was recommended to detect more first-time recurrences, but the findings of these studies are hindered by wide variations in the time intervals of follow-up angiography; it is possible that initial reopening may have been identified sooner with earlier follow-up. When initial follow-up is fixed at 6 months, very few recurrences are first recognized beyond the 6-month follow-up.8 ,14
Aneurysm rupture,16––18 larger aneurysm size,16––18 wide aneurysm neck,1 ,16––20 posterior circulation location,1 intraluminal thrombus,15 low packing density,21 ,22 and initial incomplete occlusion6 have previously been implicated in conferring a greater risk for aneurysm recurrence. We found that large aneurysm size (≥10 mm) (p<0.001) and wide aneurysm neck (>4 mm) (p=0.0001) were significant risk factors for early recurrence. We have also demonstrated that aneurysms of ≥10 mm size (p=0.0001) and neck width >4 mm (p=0.0017) are significant risk factors for significant late recurrence (reopening to residual aneurysm). Any influence of aneurysm location on the predisposition to recurrence is more controversial.5 ,19 We noted that the posterior communicating artery location tended to confer an increased risk, but this is of debatable significance.
Our results suggest that, for adequately occluded aneurysms, the rate of late recurrence was lower in those coiled between 2006 and 2010 than in those coiled between 1995 and 2005 (p<0.05). We speculate that this coincided with the introduction of a biplane unit and 3D rotational angiography to our department that we feel has improved assessment of the aneurysm neck and local anatomy and also the introduction of next generation coil technology including 360 (Boston, now Stryker) and Axium 3D (eV3/Covidien) coils. Although we did not formally assess packing density, complex-shaped platinum coils have been shown to increase packing density.23 Nevertheless, we accept that there are multiple flaws in comparing the data of these two cohorts, not least that the imaging modalities used were different, that the missing data could have altered the results, and that attitudes to retreatment may have changed over time. This therefore limits the conclusions that can be drawn.
The results do indicate that, for adequately occluded aneurysms at 6 months, long-term follow-up may not be necessary for all aneurysms. Patients with small narrow-necked aneurysms that are completely occluded at 6 months may not necessarily benefit from long-term follow-up. Indeed, through evaluation of the effects on mood and level of anxiety from long-term follow-up MRA in comparison to general population norms, aneurysm follow-up has been implicated in prolonging patient anxiety,4 although there are no data about follow-up imaging and depression levels in a ruptured aneurysm population between those who underwent imaging and those who did not.
The purpose of detecting recurrence is to identify those patients who may benefit from retreatment in order to reduce the risk of hemorrhage. Although this is a well-accepted concept, only a small number of studies have related the degree of aneurysm coil occlusion to the subsequent risk of rehemorrhage. The CARAT investigators2 demonstrated that the risk of rehemorrhage in completely occluded aneurysms was 1.1%, but was as high as 17.6% for those with <70% occlusion. Similarly, Sherif and colleagues3 demonstrated that the risk of rebleeding ranged from 0% for completely occluded aneurysms to 16.7% for those with <70% occlusion. Furthermore, the risk of rehemorrhage is greatest in the first year following treatment.2 Current evidence suggests that the risk of rehemorrhage beyond the first year is low and lies between 0.11%2 and 0.21%9 per year. The very low incidence of aneurysm rebleeding therefore questions the validity and cost-effectiveness of performing routine long-term follow-up imaging of intracranial aneurysms completely obliterated after 6–12 months. Indeed, in the International Subarachnoid Aneurysm Trial (ISAT), two of 988 patients had rehemorrhage from an aneurysm that was completely occluded at 6-month follow-up angiography.24 What confounds this argument is the effects of retreatment (7.7% in the first year and 4.5% in the second year2 in CARAT and 17.4% overall in ISAT7) that requires a robust follow-up system to guide the need for re-intervention. In this study we specifically investigated those aneurysms that were adequately occluded at 6 months. In ISAT, 5.8% of 584 aneurysms that were well occluded at 6-month angiographic follow-up were retreated. If retreatment is used as a surrogate measure of recurrence, this somewhat contradicts the findings of our study, particularly as the mean time to retreatment was 17 months, but what is difficult to quantify is the effect of retreatment in this group on the rate of rehemorrhage. It is noticeable that there may now be a trend to manage small recurrences or neck remnants more conservatively. This trend may be reflected in the proportion of retreated patients in the more recent Cerecyte25 and HELPS (HydroCoil Endovascular Aneurysm Occlusion and Packing Study)26 trials compared with ISAT (5.5% and 3%, respectively, vs 17.4%).
Some suggest that an additional reason for following patients with treated cerebral aneurysms is also to detect de novo aneurysms and to depict growth of untreated aneurysms. This is an unresolved issue. We have not detected any de novo aneurysms in our cohort of patients imaged using MRA (unpublished observations), nor have we observed the growth of additional small untreated aneurysms. The results of previous large follow-up studies indicate that, in the first 5 years after coiling (and probably also in the first 10 years), both the risk of de novo aneurysm formation and the risk of growth of existing untreated aneurysms is very low. Gallas et al27 found five de novo aneurysms in 731 patients over a 5-year period of follow-up and Sprengers et al28 describe a 1.5% incidence after 5 years. The risk of subarachnoid hemorrhage from such aneurysms is extremely low.27––29 There is no evidence that treatment of small unruptured aneurysms is beneficial. Where appropriate at our institution, additional aneurysms are treated either at the time of the initial procedure or 3–6 months after the subarachnoid hemorrhage, following recovery.
There are multiple problems with the methodology employed in this study and other retrospective observational studies that have reached similar conclusions.8 ,10 ,14 ,15 Each study has been relatively small and suffered from missing data. This has previously been thoroughly critically appraised by Raymond.30 Grading of aneurysm occlusion was initially based on operator reports. It is well-accepted that there is wide variation in the reporting of aneurysm occlusion between operators and independent assessors.31––33 We used different imaging modalities to follow patients (DSA prior to 2005 and 1.5 T or 3 T TOF MRA for follow-up after 2005). This could well have impacted on our comparison of the two cohorts. It is generally accepted that DSA is the most sensitive method of detection for recurrence and therefore any differences in the rate of reported recurrence between the two cohorts should take into account the inherent differences in the sensitivity of DSA and MRA. The evidence is conflicting with regard to the differences between 1.5 and 3 T MRA and is probably not significant.34 ,35 The study could also have been skewed by missing data. A total of 122 patients were not followed up (11% of the overall population treated) and 303 patients with 306 aneurysms (27% of the overall population treated) had only 6-month follow-up available for review and were therefore not included in the primary study cohort, so limiting our ability to draw firm conclusions on the basis of the available data: we may have failed to capture all recurrences or alternatively, the rate of recurrence may have been overestimated since many of the aneurysms that underwent only 6-month follow-up were completely occluded at that stage. An additional limitation is that long-term follow-up in most cases was limited to only 30 months. It is possible that later reopening could have occurred but would not have been identified by the timing of our follow-up. Significantly, however, no cases of rebleeding have occurred from the two cohorts of adequately coil-occluded aneurysms.
This study concurs with others and proposes that most aneurysms adequately occluded at 6 months do not require further follow-up. However, current data suggest that this policy should perhaps be reserved for small, narrow-necked and completely occluded aneurysms while large and wide-necked aneurysms are at greater risk of recurrence.
Contributors All authors contributed to the manuscript and study.
Competing interests None.
Ethics approval North Bristol NHS Trust Audit and Quality Improvement Department.
Provenance and peer review Not commissioned; externally peer reviewed.
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.