Predictors and Outcomes of Intraprocedural Rupture in Patients Treated for Ruptured Intracranial Aneurysms
The CARAT Study
Background and Purpose— Intraprocedural rupture (IPR) is a well known complication of intracranial aneurysm treatment. Risks and predictors of IPR and its impact on outcome have not been clearly established.
Methods— Potential predictors of IPR were evaluated in patients treated in the Cerebral Aneurysm Rerupture After Treatment (CARAT) study using multivariate logistic regression with stepwise elimination stratified by treatment modality. Periprocedural death or disability was defined as death or a change of ≥2 points on the Modified Rankin Scale at discharge compared to before treatment.
Results— IPR occurred in 14.6% of 1010 patients (299 coiled, 711 clipped): 19% with clipping and 5% with coiling (P<0.001). Among those clipped, 31% with IPR had periprocedural death or disability compared to 18% without IPR (P=0.001); among those coiled, 63% with IPR had periprocedural death or disability compared to 15% without IPR (P<0.001). Overall, coronary artery disease and initial lower Hunt and Hess Grade were independent predictors of IPR. For those undergoing coiling, independent predictors of IPR were Asian race, black race, COPD, and lower initial Hunt and Hess Grade. Among those undergoing clipping, hyperlipidemia and lower initial Hunt and Hess Grade were both independent predictors of IPR.
Conclusions— IPR was common in patients undergoing treatment of ruptured aneurysms, particularly with surgical clipping. The frequency of IPR with new disability was similar in the surgical and endovascular treatment groups. Coronary artery disease, hyperlipidemia, race, COPD, and lower Hunt and Hess Grade were associated with greater risk of IPR, which may reflect differences in vessel fragility but requires further confirmation.
Intraprocedural rupture (IPR) has been reported as a frequent complication of intracranial aneurysm treatment. However, its incidence, risk factors, and consequences are not well described. Large-scale prospective studies of endovascular and open surgical treatments have not reported data on IPR.1,2 Several retrospective case series from single institutions found varying rates of IPR, ranging from 2% to 4.5% for coiling and from 7.6 to 34.9% for clipping.3–10 Among patients with IPR, mortality rates have also varied considerably, ranging 0% to 40% with coiling and 0% to 33% with clipping.
Prior studies of IPR are discordant on the importance of aneurysm location, size, and technical/anatomic factors that may influence the risk of IPR.5,7,11,12 Furthermore, these studies have been relatively small single center studies with limited follow-up and demographic data, nonstandardized assessment of procedure related disability, thus the ability to identify risk factors and consequences of IPR has been limited. One report found that those with previously ruptured aneurysms were more likely to experience IPR than those with unruptured aneurysms, whether patients were treated with clipping or coiling.11 In addition, single center experience may not reflect wider practice.
CARAT is an ambidirectional cohort study of 1010 unselected patients presenting with ruptured intracranial aneurysms treated by coil embolization or surgical clipping at 9 high-volume centers in the United States from 1996 to 1998 followed for more than 5 years.13 The goal of the present study was to quantify the risk and clinical impact of IPR, and to identify predictors of IPR, for coil embolization and surgical clipping.
The patient selection criteria and procedures for data collection have been described in detail in the original CARAT manuscript.13 Briefly, 9 US hospitals affiliated with 8 medical centers with expertise in both surgical clipping and endovascular coiling (>30 patients treated per year) were invited and agreed to participated in CARAT. All patients discharged from January 1, 1996 to December 31, 1998 with a diagnosis of subarachnoid hemorrhage were identified from hospital administrative databases to assure complete case selection. These patient charts were reviewed to determine eligibility, which required treatment of a ruptured saccular aneurysm by surgical clipping, wrapping, or endovascular coil embolization. Patients were not randomized to treatment type; each center determined the treatment modality according to local practice norms so that we could capture usual care.
IPR was identified at the time of first treatment by the operating neurosurgeon or neurointerventionalist, with adjudication by a team of 3 independent physicians (a neurologist, a neurointerventionalist, and a neurosurgeon) reviewing detailed medical records. Agreement of at least 2 of 3 reviewers was required to confirm the event as an IPR. Rankin scores at presentation, postoperatively, and on discharge were derived from medical records by trained analysts using standard criteria.14
Characteristics of the patient, presentation, and aneurysm were evaluated as predictors of IPR. Univariate comparisons were tested with Fisher’s exact test for dichotomous variables and the Wilcoxon Rank Sum Test for continuous and ordinal variables. Multivariate logistic regression was performed stratified by treatment modality with stepwise elimination of variables with P>0.10, beginning with a model that included all potential predictors. Variables were chosen for the model based on previously established importance in regards to outcome for subarachnoid hemorrhage (eg, Hunt and Hess Grade), posing a technical risk for IPR (eg, size), and epidemiological factors known to influence vascular biology that may influence aneurysm rupture (eg, smoking, sex). Aneurysm size was not included in the main model because it was not measured in some patients treated surgically; in a sensitivity analysis, it was added back to the final multivariable models. New periprocedural death or disability was defined as postprocedural in-hospital death or a change of ≥2 points on the Modified Rankin Scale (mRS) at discharge compared to just before treatment. Wilcoxon Rank Sum Test was performed to determine whether operator experience using the number of cases treated by a practitioner during the study period, as a measure of experience and practice, influenced IPR. All statistical analysis was performed with STATA (Version 8.0).
Among 1010 patients treated (299 coiled, 711 clipped), IPR occurred in 148 (14.6%). IPR was more frequent in those treated with clipping (19%) compared to coiling (5%, P<0.001). Race (P=0.03) and lower initial Hunt and Hess Grade (P=0.02) were also associated with IPR in univariate analysis (Table 1).
Several predictors of IPR were identified in multivariate analysis (Table 2). Lower initial Hunt and Hess Grade on presentation predicted IPR across all groups: for all patients considered together (P=0.001) and in models limited to those treated surgically (P=0.012) and to those treated endovascularly (P=0.015). In all patients regardless of treatment type, coronary artery disease (OR 1.93, 1.06 to 3.51, P=0.031) and the group of unknown/other race (2.18, 1.30 to 3.65, P=0.003) both predicted IPR. In the endovascular group, chronic obstructive pulmonary disease (13.9, 3.19 to 60.6, P<0.0001), black race (11.3, 2.16 to 58.8, P=0.004), and Asian race (25.9, 4.49 to 150, P<0.0001) were predictors of IPR. Among surgical patients, hyperlipidemia (2.54, 1.19 to 5.4, P=0.016) and other/unknown race (1.95, 1.12 to 3.39, P=0.018) predicted IPR. Aneurysm size (per centimeter) was not predictive of IPR in the whole cohort (0.66, 0.44 to 1.00, P=0.07); nor was it predictive when patients were considered by treatment modality (endovascular, 0.44, 0.06 to 1.11 P=0.36; surgical, 0.66, 0.44 to 1.10, P=0.10). Aneurysm size by category (<3 mm, 3 to 5 mm, >5 to 10 mm, >10 mm) was also not predictive of IPR in the whole cohort or by treatment modality (0.84, 0.66 to 1.09, P=0.19, endovascular 0.91, 0.38 to 2.20, P=0.84, and surgical 0.81, 0.63 to 1.05, P=0.11). Operator volume did not influence the rate of IPR in endovascular or surgical treated patients, P=0.25 and P=0.22, respectively.
Overall, IPR was associated with an increased risk of periprocedural death/disability (34% among those with IPR compared to 17% without, P<0.0001). The association was present in patients treated either with surgery or endovascular therapy (Figure 1). Among surgically clipped patients, periprocedural death/disability occurred in 31% with IPR versus 18% without IPR (P=0.001). In the endovascular group, periprocedural death/disability occurred in 63% of those with IPR versus 15% without (P<0.001). Furthermore, the risk of new periprocedural death/disability in patients with IPR was greater among those treated with endovascular coiling compared to those with IPR from surgical clipping (Figure 2). When all patients are considered, the risk of periprocedural death/disability with IPR was similar in the 2 treatment groups (endovascular, 3.3% versus 5.8% surgical, P=0.12).
We found that IPR was common in patients undergoing treatment of ruptured aneurysms, particularly among those treated with surgical clipping. Our report is consistent with previous literature on IPR from single centers that characterizes it as a relatively frequent adverse event.11,15 The rates of IPR in this study for surgical clipping (19%) and endovascular coiling (5%) are within the wide range reported in previously published single-center series.11 Data from International Subarachnoid Hemorrhage Trial corresponds with our findings, reporting IPR rates of 5.4% and 19% for coiling and clipping respectively (Personal Correspondence, Dr Andrew Molyneux, 2007). Thus, IPR remains an important complication at high-volume centers throughout Europe and North America that have extensive expertise in aneurysm treatment. This finding is reinforced by the fact that surgical/endovascular volume in CARAT did not influence the rate of IPR. Thus, the high IPR rates in ISAT do not appear to be indicative of poor surgical skills in the centers that participated relative to high volume centers in the US.
We found that IPR was associated with a greater risk of periprocedural death/disability, whether patients were treated with surgical clipping or endovascular coil embolization. For surgery patients, IPR nearly doubled the risk of new major disability or death. One prior study of 59 surgically clipped patients also found that risk of poor outcome was greater in those with IPR, with mortality doubling from 15% in those without IPR to 31% in those with it; however, the differences were not significant in this small study.7 Other studies have not shown an impact of IPR on surgical outcomes.9,16 The CARAT surgical group included more patients with IPR (n=132) than prior studies and thus had more power to assess the impact of IPR on outcome. The present results suggest that IPR has an important detrimental effect on surgical outcome, and that efforts to reduce its risk and manage its consequences are well justified.16
IPR during coil embolization had an even greater impact on risk of periprocedural death/disability, with a 4-fold increased risk among those with IPR (63%) compared to those without (15%). This finding is consistent with the majority of previous reports that identified a high morbidity and mortality with IPR during coil embolization, with 39% morbidity and 33% mortality in 1 prior review.11 The greater impact of IPR during endovascular treatment compared to open surgery may be related to the ability to control bleeding with clips during surgery or to reduce the impact of new subarachnoid hemorrhage by suctioning and preventing elevated intracranial pressure.
Chronic obstructive pulmonary disease (COPD), coronary artery disease (CAD), and hyperlipidemia were identified as predictors of IPR. These findings suggest that processes that damage the arterial endothelium may predispose patients to IPR. COPD and aneurysm rupture share several important epidemiological and pathophysiologic mechanisms. Smoking is a recognized risk factor for aneurysm growth and rupture that is positively correlated with the intensity of smoking,17,18 and COPD may be associated with risk of IPR through this association with smoking. Genetic and acquired α-1-antitrypsin deficiency, as well as increased levels of matrix metalloproteinases, have also been implicated in the pathogenesis of ruptured intracranial aneurysms.18–21 These findings suggest that patients with COPD may have more fragile vessels. In our study, we found that COPD increased the risk of IPR in patients undergoing coil embolization but not during surgery. The risk of IPR with COPD may be smaller with clipping because vessels are not approached intraluminally or simply because risk is predicted more strongly by other unidentified factors that obscure an association with COPD. COPD patients in CARAT where more commonly treated with coiling, so it is also possible that we did not have sufficient statistical power to describe the relationship between these two variables in the surgical group.
We found that atherosclerosis and its risk factors were also associated with IPR. Specifically, CAD was associated with IPR overall and hyperlipidemia was associated with IPR in surgical patients. A recent pathological study reported increased expression of matrix metalloproteinase-9 in the walls of ruptured atherosclerotic aneurysms, which could increase fragility.22 Systemic hyperlipidemia may be a surrogate marker of aneurysms that are often technically difficulty to treat surgically because of their large size and irregular morphology with calcified necks and domes.23
The relationship between lower Hunt and Hess grade and a greater risk of IPR is perplexing. More rapid treatment of aneurysms with lower Hunt and Hess grade was not the explanation because IPR was not related to timing of treatment (P>0.34). We cannot rule out the possibility that IPR is simply less frequently reported in patients with greater pretreatment morbidity.
In the coiling group, IPR was significantly more frequent in patients of either Asian or black race. This finding is not attributable to selection bias as both patient groups were distributed equally among the 2 treatment groups. The risk of IPR was also higher in the unknown/other race group for all patients and clipped patients. These findings require validation with prospective data as a biological mechanism that would explain the risk of IPR in these diverse groups is not apparent.
CARAT has several important limitations. Data were not collected prospectively and patients were not randomized to treatment type. Additionally, IPR may not have been reported completely, particularly for small hemorrhages; thus the incidence of IPR may have been underreported and its consequences exaggerated by selective reporting of severe hemorrhages. Other than size and location, morphological features of aneurysms were not recorded and these may be important in determining the risk of IPR. Future study should consider the presence of multi-lobed or daughter aneurysms and the orientation of the aneurysm dome relative to the surrounding brain. These morphological factors may be important in regards to treatment modality selection and technical aspects of each procedure as they relate to the risk of IPR. Information regarding the circumstances surrounding IPR and the practioners’ response, for example reversal of anticoagulation with protamine or proximal vascular control with temporary clips, was not systematically recorded and would be useful in better defining strategies to mitigate the consequences of this feared complication. We evaluated 16 variables as potential predictors of IPR, and it is possible that some associations were simply attributable to chance. Finally, these results reflect the experience of tertiary referral centers; thus, the outcomes may not be generalizable to community practice.24 Despite these limitations, CARAT represents the largest cohort to date of both clipped and coiled patients in which data on IPR and its potential risk factors have been systematically collected, and illustrates the importance of studying this complication to reduce its frequency and impact.
Appendix CARAT Investigators
University of California, San Francisco: S. Claiborne Johnston, MD, PhD, Principal Investigator (PI); Christopher F. Dowd, MD (Co-PI); Michael T. Lawton, MD (Co-PI); Daryl R. Gress, MD; Randall T. Higashida, MD; Van V. Halbach, MD; Shoujun Zhao, MD, PhD; Katherine H. Katsura, BS; Kristin J. Fong, BS; Vanja C. Douglas, MD; Rosalyn Ventura, MD; Jacob S. Elkins, MD; Mai N. Nguyen-Huynh, MD.
Barrow Neurological Institute of St Joseph’s Hospital and Medical Center: Cameron G. McDougall, MD (Site PI); Robert F. Spetzler, MD; Joseph M. Zabramski, MD; Heidi K. Jahnke, RN, BSN. Mayo Clinic: David G. Piepgras, MD (Site PI); Douglas A. Nichols, MD; Denise R. Gravenhof; Debra Herzig, RN. Houston Methodist Hospital: Michel E. Mawad, MD (Site PI); Denise Meyer, RN. Stanford University Medical Center: Gary K. Steinberg, MD, PhD (Site PI); Michael P. Marks, MD; Desiree Luu, RN; Hanna Yi, RN. University of California, Los Angeles: Gary R. Duckwiler, MD (Site PI); Neil A. Martin, MD; Henry Adapon, MD. University of Southern California: Steven L. Giannotta, MD (Site PI); Donald W. Larsen, MD; George P. Teitelbaum, MD; Dawn Fishback, PA-C; Evangeline Thomson, RN. University of Texas, Southwestern: Duke S. Samson, MD (Site PI); Phillip D. Purdy, MD; Robert E. Replogle, MD; Jerri Thomas, BS.
We thank Trinh Pham, MA for providing statistical expertise.
Sources of Funding
The original CARAT study was supported by an unrestricted grant from Boston Scientific. The sponsor had no role in the concept, design, statistical analysis, or preparation of the original or present manuscript.
Dr Johnston has received research support from Boston Scientific and Johnson & Johnson. Dr Higashida has received compensation for consulting/advisor board services for Cordis Neurovascular. Dr Duckwiler has no disclosures; his employer, UCLA, receives royalty payments from endovascular coil manufacturers.
↵*For list of investigators, see Appendix.
- Received September 14, 2007.
- Accepted September 27, 2007.
Molyneux A, Kerr R, Stratton I, Sandercock P, Clarke M, Shrimpton J, Holman R. International subarachnoid aneurysm trial (isat) of neurosurgical clipping versus endovascular coiling in 2143 patients with ruptured intracranial aneurysms: A randomised trial. Lancet. 2002; 360: 1267–1274.
Tummala RP, Chu RM, Madison MT, Myers M, Tubman D, Nussbaum ES. Outcomes after aneurysm rupture during endovascular coil embolization. Neurosurgery. 2001; 49: 1059–1066;discussion 1066–1057.
Doerfler A, Wanke I, Egelhof T, Dietrich U, Asgari S, Stolke D, Forsting M. Aneurysmal rupture during embolization with guglielmi detachable coils: Causes, management, and outcome. AJNR Am J Neuroradiol. 2001; 22: 1825–1832.
Cloft HJ, Kallmes DF. Cerebral aneurysm perforations complicating therapy with guglielmi detachable coils: A meta-analysis. AJNR Am J Neuroradiol. 2002; 23: 1706–1709.
Kopitnik TA, Horowitz MB, Samson DS. Surgical management of intraoperative aneurysm rupture. In: Schmidek HH, Sweet WH, editors. Operative neurosurgical techniques, vol 2. Philadelphia: WB Saunders; 2000. p. 1275–81.
Rates of delayed rebleeding from intracranial aneurysms are low after surgical and endovascular treatment. Stroke. 2006; 37: 1437–1442.
Bonita R, Beaglehole R. Recovery of motor function after stroke. Stroke. 1988; 19: 1497–1500.
Lawton MT, Du R. Effect of the neurosurgeon’s surgical experience on outcomes from intraoperative aneurysmal rupture. Neurosurgery. 2005; 57: 9–15;discussion 19–15.
Giannotta SL, Oppenheimer JH, Levy ML, Zelman V. Management of intraoperative rupture of aneurysm without hypotension. Neurosurgery. 1991; 28: 531–535;discussion 535–536.
Juvela S, Poussa K, Porras M. Factors affecting formation and growth of intracranial aneurysms: A long-term follow-up study. Stroke. 2001; 32: 485–491.
Mohr JP CD, Grotta JC, Weir B, Wolf P. Stroke Pathophysiology, Diagnosis, and Management. Philadelphia: Elsevier; 2004.
Bardach NS, Olson SJ, Elkins JS, Smith WS, Lawton MT, Johnston SC. Regionalization of treatment for subarachnoid hemorrhage: A cost-utility analysis. Circulation. 2004; 109: 2207–2212.