See the corresponding editorial in this issue, pp 1022–1023.
J Neurosurg 121:1024–1038, 2014 ©AANS, 2014
Age-related differences in unruptured intracranial aneurysms: 1-year outcomes Clinical article Kelly B. Mahaney, M.D., M.S.,1 Robert D. Brown Jr., M.D., M.P.H., 2 Irene Meissner, M.D., 2 David G. Piepgras, M.D., 3 John Huston III, M.D., 4 Jie Zhang, M.S., 5 and James C. Torner, Ph.D., 5 for the ISUIA Investigators Department of Neurosurgery, Carver College of Medicine, University of Iowa, Iowa City, Iowa; Departments of 2Neurology, 3Neurologic Surgery, and 4Radiology, Mayo Clinic, Rochester, Minnesota; and 5Department of Epidemiology, College of Public Health, University of Iowa, Iowa City, Iowa
1
Object. The aim of this study was to determine age-related differences in short-term (1-year) outcomes in patients with unruptured intracranial aneurysms (UIAs). Methods. Four thousand fifty-nine patients prospectively enrolled in the International Study of Unruptured Intracranial Aneurysms were categorized into 3 groups by age at enrollment: < 50, 50–65, and > 65 years old. Outcomes assessed at 1 year included aneurysm rupture rates, combined morbidity and mortality from aneurysm procedure or hemorrhage, and all-cause mortality. Periprocedural morbidity, in-hospital morbidity, and poor neurological outcome on discharge (Rankin scale score of 3 or greater) were assessed in surgically and endovascularly treated groups. Univariate and multivariate associations of each outcome with age were tested. Results. The risk of aneurysmal hemorrhage did not increase significantly with age. Procedural and in-hospital morbidity and mortality increased with age in patients treated with surgery, but remained relatively constant with increasing age with endovascular treatment. Poor neurological outcome from aneurysm- or procedure-related morbidity and mortality did not differ between management groups for patients 65 years old and younger, but was significantly higher in the surgical group for patients older than 65 years: 19.0% (95% confidence interval [CI] 13.9%–24.4%), compared with 8.0% (95% CI 2.3%–13.6%) in the endovascular group and 4.2% (95% CI 2.3%–6.2%) in the observation group. All-cause mortality increased steadily with increasing age, but differed between treatment groups only in patients < 50 years of age, with the surgical group showing a survival advantage at 1 year. Conclusions. Surgical treatment of UIAs appears to be safe, prevents 1-year hemorrhage, and may confer a survival benefit in patients < 50 years of age. However, surgery poses a significant risk of morbidity and death in patients > 65 years of age. Risk of endovascular treatment does not appear to increase with age. Risks and benefits of treatment in older patients should be carefully considered, and if treatment is deemed necessary for patients older than 65 years, endovascular treatment may be the best option. (http://thejns.org/doi/abs/10.3171/2014.6.JNS121179)
A
Key Words • unruptured aneurysm • age • outcome • hemorrhage • International Study of Unruptured Intracranial Aneurysms • vascular disorders
has been shown to be an important predictor of outcome in patients with aneurysmal subarachnoid hemorrhage (SAH) from a ruptured intracranial aneurysm. Several studies have demonstrated that older patients fare worse than their younger counterparts, regardless of treatment strategy.2,10,12,13 However, the impact of age on outcomes of patients with unruptured inge
Abbreviations used in this paper: ACA = anterior cerebral artery; ACoA = anterior communicating artery; BA = basilar artery; HR = hazard ratio; ICA = internal carotid artery; ISUIA = International Study of Unruptured Intracranial Aneurysms; MCA = middle cerebral artery; MMSE = Mini-Mental State Examination; PCA = posterior cerebral artery; PCoA = posterior communicating artery; SAH = subarachnoid hemorrhage; TICS = Telephone Interview for Cognitive Status; UIA = unruptured intracranial aneurysm.
1024
tracranial aneurysms (UIAs) is less clear. Much of what is known is derived from hospital discharge records—namely, the Nationwide Inpatient Sample—and is restricted to comparisons of surgical clipping and endovascular coiling.1,3,5 Increasing age has been associated with worse outcomes following treatment of UIAs, and specifically, worse outcomes with surgical compared with endovascular treatment have been observed with increasing age.1,3,5 Recently, elderly patients (> 65 years of age) were shown to be at higher risk of treatment-related complications following endovascular treatment of UIAs compared with This article contains some figures that are displayed in color online but in black-and-white in the print edition.
J Neurosurg / Volume 121 / November 2014
Age-related differences in 1-year outcomes of UIAs nonelderly patients.9 However, the factors associated with poor outcomes in patients of differing ages are not well described, and outcomes beyond initial hospitalization for treatment are lacking. Furthermore, little is known regarding the influence of age on outcomes of patients who are managed conservatively. In this study we sought to assess the relationship between age and 1-year outcomes of patients with UIAs. Specifically, we sought to define associations of age with hemorrhage risk, procedural and in-hospital morbidity and mortality, poor neurological outcome (related to aneurysmal hemorrhage or aneurysm treatment), and all-cause mortality at 1 year. We also sought to identify risk factors for procedural and in-hospital morbidity and mortality, poor neurological outcome, and all-cause mortality in persons of different ages with UIAs. We have previously described age-related differences in presentation of UIAs. In this study we address 1-year outcomes of surgically treated, endovascularly treated, and observed patients enrolled in the International Study of Unruptured Intracranial Aneurysms (ISUIA).
Methods Study Design
The ISUIA is an international multicenter study of treated and untreated patients with UIAs. The ISUIA study involved a retrospective arm and a prospective arm; only the prospective cohort is considered in this analysis. In the prospective cohort, a total of 4059 patients were enrolled at 61 medical centers in the US, Canada, and Europe between 1991 and 1998 (Phases I and II). The last follow-up of the prospective cohort was conducted between 2004 and 2007 (Phase III). Among the treated patients, 1917 patients were treated surgically and 451 patients were treated with an endovascular procedure. An additional 1691 patients were managed conservatively. For analysis of the association of age with outcomes in the first year of follow-up, patients were categorized according to age at the time of diagnosis of a UIA: < 50 years, 50–65 years, and > 65 years of age. The age groups were specified prior to analysis and were chosen because of their relevant clinical significance.
Patient Inclusion Criteria
Patients enrolled in the study had at least 1 saccular UIA of at least 2 mm in maximum diameter confirmed by cerebral arteriography. Some patients had other aneurysms presenting with SAH in the past; these aneurysms were required to be definitively treated prior to enrollment in the study. Patients were classified by group with regard to SAH history. Those who were identified to have had prior SAH from another intracranial aneurysm were classified as Group II and those with no history of prior SAH as Group I. Patients with fusiform, traumatic, or mycotic aneurysms were not included in the study. Patients were required to be functionally independent (Rankin11 scale score of 1 or 2) at the time of enrollment; thus, patients with a neurologically devastating prior hemorrhage were excluded from the study. Patients in whom the sole UIA was previously manipulated by wrapping,
J Neurosurg / Volume 121 / November 2014
packing, coil placement, proximal arterial ligation, bypass, balloon occlusion, or clip placement before entry into the study were not eligible. Patients with a history of intracranial hemorrhage from an unrepaired underlying structural lesion, primary intracerebral hemorrhage (without an underlying structural lesion), or SAH from an undetermined origin were excluded from the study. Patients with a malignant brain tumor were also excluded from the study. Patients were classified as “operated” if a surgical or endovascular procedure to treat at least 1 intracranial aneurysm was scheduled within 30 days of enrollment. Otherwise, patients were classified as “unoperated.” Follow-Up Evaluation
Baseline characteristics were documented for all enrolled patients via in-person interview with the patient and a comprehensive review of medical records. Unoperated patients were followed-up by routine examination or by questionnaire at yearly intervals. Operated patients were assessed at 7 days after the procedure, at hospital discharge, at 30 days after the procedure, and at yearly intervals thereafter. Neurological status as measured by the Rankin scale11 was collected at each time point, and cognitive status was measured by the Mini-Mental State Examination (MMSE)6 or the Telephone Interview for Cognitive Status (TICS)4 at the same time intervals. Complications related to the aneurysm procedure were assessed at the time of the procedure and at hospital discharge.
Determination of Events
Detailed information was obtained on all end points (definite or questionable subarachnoid or intracerebral hemorrhage and death). Comprehensive adjudication of ISUIA patients was performed centrally for all hemorrhages, ischemic strokes, and deaths on the basis of uniform criteria, with the use of available clinical, radiological, and autopsy findings and other information, and hemorrhages were classified according to the location of the rupture. Subarachnoid or intracerebral hemorrhage was classified as follows: 1) definite (symptoms of subarachnoid or intracerebral hemorrhage and positive findings on CT or MRI at surgery or at autopsy); 2) highly probable (symptoms and positive findings on CSF analysis); or 3) probable (symptoms only). In this analysis, definite or highly probable hemorrhages were included; probable hemorrhages were excluded. Evidence of procedure-related cerebral infarction, hemorrhage, or death was confirmed at the ISUIA coordinating center at the Mayo Clinic in Rochester, Minnesota. Perioperative hemorrhage included hemorrhages occurring 3 days or fewer after aneurysm treatment. Neurological deficits occurring 30 days to 1 year after treatment were evaluated for their relationship to treatment or coexisting neurological disorders. Deficits clearly related to a coexisting disorder were not attributed to aneurysm treatment. Poor treatment outcome was defined as Rankin scale score of 3, 4, or 5 (moderate to severe neurological disability) at the time of discharge. Aneurysmal morbidity was defined as worsening in neurological status attributed to aneurysmal hemorrhage 1025
K. B. Mahaney et al. or aneurysm procedure from the time of diagnosis to the first occurrence of a neurological change. This worsening in neurological status was considered as a Rankin scale score of 3, 4, or 5 (moderate to severe neurological disability), a score less than 24 on the MMSE, or a score less than 27 on the TICS. Mortality was assessed by date and cause of death. Interval to death was defined as time from date of diagnosis (untreated) or date of procedure (treated) to date of death. Cause of death was determined by contacting the local physician or by obtaining medical records, including death certificate and autopsy information when available. Procedure-related 1-year deaths included 1) death occurring within 30 days of the aneurysm treatment procedure; 2) death from complications of the aneurysm procedure any time during the year; 3) death from a repeat procedure to treat the aneurysm; or 4) death from aneurysmal hemorrhage after the procedure. A combined end point of aneurysmal morbidity and mortality attributable to aneurysmal hemorrhage or aneurysm procedure was defined as death prior to 366 days after diagnosis or assessment of a neurological deficit (Rankin scale score of 3, 4, or 5; TICS score < 27; or MMSE score < 24) at the observation closest to 365 days (within a ± 180-day window). Statistical Analysis Univariate Analysis. Univariate associations between age group and baseline characteristics, stratified by management group, were tested using chi-square and Fisher exact tests for categorical variables and ANOVA for continuous variables. Differences in case-fatality rates among management groups within each age group were tested using the chi-square test. In both the surgical group and the endovascular group, associations between age group and occurrence of procedural and in-hospital morbidities were tested using the chi-square and Fisher exact tests. In all prospective patients, within each age category, differences in hemorrhage rates, aneurysm- and procedure-related morbidity and mortality rates, and all-cause mortality rates were compared between management groups utilizing the log-rank test. Subgroup analysis was completed to determine if aneurysm site and aneurysm size influenced differences between management groups in hemorrhage rates and aneurysm- and procedure-related morbidity and mortality rates within each age group.
Multivariate Analysis. Multivariate models were completed for several endpoints.
One-Year Hemorrhage. Cox proportional hazard regression was used for multivariate models of aneurysmal hemorrhage risk. All baseline variables associated with risk of aneurysmal hemorrhage (at the p < 0.10 level) were entered into the multivariate model. Backward model selection was used to retain variables significant at the p < 0.05 level in the final model. A model was completed for each age group as well as two separate models inclusive of all prospective untreated patients (all ages). In the models inclusive of all prospective untreated patients, the variable age was forced; 1 model was completed with age as a cat1026
egorical variable and 1 model with age as a continuous variable.
Poor Treatment Outcome (Discharge Rankin Scale Score). For the end point of poor treatment outcome—de-
fined as Rankin scale score of 3, 4, or 5 at discharge following surgical or endovascular treatment—multivariate logistic regression models were completed for each age group within the surgical group and each age group within the endovascular group. Baseline characteristics, intraoperative events, and in-hospital events that were associated with poor treatment outcome in univariate analysis (at the p < 0.10 level) were entered into multivariate logistic regression models of poor treatment outcome. Backward model selection was used, keeping variables significant at the p < 0.05 level in the final models.
One-Year Aneurysm- and Procedure-Related Morbidity and Mortality. Multivariate logistic regression analysis
was used to model 1-year aneurysm- and procedure-related morbidity and mortality within each age group. All baseline variables with a univariate association (at the p < 0.10 level) with age or with occurrence of aneurysm- and procedure-related morbidity and mortality were entered into multivariate models of aneurysm- and procedurerelated morbidity and mortality, with 1 model for each age group. Backward model selection was used, retaining variables significant at the p < 0.05 level in the final models. The variables “treatment group” and “baseline Rankin scale score” were forced in the final models. One-Year All-Cause Mortality. Cox proportional hazard regression was used to model risk of 1-year mortality (from any cause) within each age group. All baseline variables with a univariate association (at the p < 0.10 level) with age or with 1-year all-cause mortality were entered into the models, with 1 model for each age group. Backward model selection was used, retaining variables significant at the p < 0.05 level in the final models.
Baseline Characteristics
Results
Age-related variability in baseline patient and aneurysm characteristics in patients with UIAs are reported (Tables 1 and 2). Differences in baseline characteristics between management groups in patients of the same age group are presented in Table 1. In all 3 age groups, the observation group had a significantly higher proportion of patients with prior SAH (UIA Group II), compared with surgical and endovascular groups (the endovascular group had the lowest fraction). The observation group also had a significantly higher proportion of patients with baseline Rankin scale scores of 2, compared with surgical and endovascular patients, which was observed in all 3 age groups. In patients younger than 50 years old, there was no significant difference in history of ischemic cerebrovascular disease between management groups. However, in patients 50–65 and > 65 years of age, history of ischemic cerebrovascular disease was significantly more common in the observation group than in the surgical or endovascular groups. History of hypertension varied significantly between management groups only for patients > 65 years J Neurosurg / Volume 121 / November 2014
Age-related differences in 1-year outcomes of UIAs TABLE 1: Patient baseline characteristics by age category and management group Characteristic no. of patients UIA Group II (SAH) 65 yrs old baseline Rankin score = 2 65 yrs old history of ischemic cerebrovascular disease 65 yrs old history of hypertension 65 yrs old smoking history 65 yrs old never current former recent aspirin use§ 65 yrs old yes no
Endovascular Treatment (%)
Surgical Treatment (%)
Untreated (%)
451*
1917†
1691‡
20 (11.6) 21 (11.2) 1 (1.1)
182 (21.5) 125 (15.0) 19 (8.1)
323 (54.1) 225 (33.7) 67 (15.7)
J Neurosurg 121:1024–1038, 2014 ©AANS, 2014
Age-related differences in unruptured intracranial aneurysms: 1-year outcomes Clinical article Kelly B. Mahaney, M.D., M.S.,1 Robert D. Brown Jr., M.D., M.P.H., 2 Irene Meissner, M.D., 2 David G. Piepgras, M.D., 3 John Huston III, M.D., 4 Jie Zhang, M.S., 5 and James C. Torner, Ph.D., 5 for the ISUIA Investigators Department of Neurosurgery, Carver College of Medicine, University of Iowa, Iowa City, Iowa; Departments of 2Neurology, 3Neurologic Surgery, and 4Radiology, Mayo Clinic, Rochester, Minnesota; and 5Department of Epidemiology, College of Public Health, University of Iowa, Iowa City, Iowa
1
Object. The aim of this study was to determine age-related differences in short-term (1-year) outcomes in patients with unruptured intracranial aneurysms (UIAs). Methods. Four thousand fifty-nine patients prospectively enrolled in the International Study of Unruptured Intracranial Aneurysms were categorized into 3 groups by age at enrollment: < 50, 50–65, and > 65 years old. Outcomes assessed at 1 year included aneurysm rupture rates, combined morbidity and mortality from aneurysm procedure or hemorrhage, and all-cause mortality. Periprocedural morbidity, in-hospital morbidity, and poor neurological outcome on discharge (Rankin scale score of 3 or greater) were assessed in surgically and endovascularly treated groups. Univariate and multivariate associations of each outcome with age were tested. Results. The risk of aneurysmal hemorrhage did not increase significantly with age. Procedural and in-hospital morbidity and mortality increased with age in patients treated with surgery, but remained relatively constant with increasing age with endovascular treatment. Poor neurological outcome from aneurysm- or procedure-related morbidity and mortality did not differ between management groups for patients 65 years old and younger, but was significantly higher in the surgical group for patients older than 65 years: 19.0% (95% confidence interval [CI] 13.9%–24.4%), compared with 8.0% (95% CI 2.3%–13.6%) in the endovascular group and 4.2% (95% CI 2.3%–6.2%) in the observation group. All-cause mortality increased steadily with increasing age, but differed between treatment groups only in patients < 50 years of age, with the surgical group showing a survival advantage at 1 year. Conclusions. Surgical treatment of UIAs appears to be safe, prevents 1-year hemorrhage, and may confer a survival benefit in patients < 50 years of age. However, surgery poses a significant risk of morbidity and death in patients > 65 years of age. Risk of endovascular treatment does not appear to increase with age. Risks and benefits of treatment in older patients should be carefully considered, and if treatment is deemed necessary for patients older than 65 years, endovascular treatment may be the best option. (http://thejns.org/doi/abs/10.3171/2014.6.JNS121179)
A
Key Words • unruptured aneurysm • age • outcome • hemorrhage • International Study of Unruptured Intracranial Aneurysms • vascular disorders
has been shown to be an important predictor of outcome in patients with aneurysmal subarachnoid hemorrhage (SAH) from a ruptured intracranial aneurysm. Several studies have demonstrated that older patients fare worse than their younger counterparts, regardless of treatment strategy.2,10,12,13 However, the impact of age on outcomes of patients with unruptured inge
Abbreviations used in this paper: ACA = anterior cerebral artery; ACoA = anterior communicating artery; BA = basilar artery; HR = hazard ratio; ICA = internal carotid artery; ISUIA = International Study of Unruptured Intracranial Aneurysms; MCA = middle cerebral artery; MMSE = Mini-Mental State Examination; PCA = posterior cerebral artery; PCoA = posterior communicating artery; SAH = subarachnoid hemorrhage; TICS = Telephone Interview for Cognitive Status; UIA = unruptured intracranial aneurysm.
1024
tracranial aneurysms (UIAs) is less clear. Much of what is known is derived from hospital discharge records—namely, the Nationwide Inpatient Sample—and is restricted to comparisons of surgical clipping and endovascular coiling.1,3,5 Increasing age has been associated with worse outcomes following treatment of UIAs, and specifically, worse outcomes with surgical compared with endovascular treatment have been observed with increasing age.1,3,5 Recently, elderly patients (> 65 years of age) were shown to be at higher risk of treatment-related complications following endovascular treatment of UIAs compared with This article contains some figures that are displayed in color online but in black-and-white in the print edition.
J Neurosurg / Volume 121 / November 2014
Age-related differences in 1-year outcomes of UIAs nonelderly patients.9 However, the factors associated with poor outcomes in patients of differing ages are not well described, and outcomes beyond initial hospitalization for treatment are lacking. Furthermore, little is known regarding the influence of age on outcomes of patients who are managed conservatively. In this study we sought to assess the relationship between age and 1-year outcomes of patients with UIAs. Specifically, we sought to define associations of age with hemorrhage risk, procedural and in-hospital morbidity and mortality, poor neurological outcome (related to aneurysmal hemorrhage or aneurysm treatment), and all-cause mortality at 1 year. We also sought to identify risk factors for procedural and in-hospital morbidity and mortality, poor neurological outcome, and all-cause mortality in persons of different ages with UIAs. We have previously described age-related differences in presentation of UIAs. In this study we address 1-year outcomes of surgically treated, endovascularly treated, and observed patients enrolled in the International Study of Unruptured Intracranial Aneurysms (ISUIA).
Methods Study Design
The ISUIA is an international multicenter study of treated and untreated patients with UIAs. The ISUIA study involved a retrospective arm and a prospective arm; only the prospective cohort is considered in this analysis. In the prospective cohort, a total of 4059 patients were enrolled at 61 medical centers in the US, Canada, and Europe between 1991 and 1998 (Phases I and II). The last follow-up of the prospective cohort was conducted between 2004 and 2007 (Phase III). Among the treated patients, 1917 patients were treated surgically and 451 patients were treated with an endovascular procedure. An additional 1691 patients were managed conservatively. For analysis of the association of age with outcomes in the first year of follow-up, patients were categorized according to age at the time of diagnosis of a UIA: < 50 years, 50–65 years, and > 65 years of age. The age groups were specified prior to analysis and were chosen because of their relevant clinical significance.
Patient Inclusion Criteria
Patients enrolled in the study had at least 1 saccular UIA of at least 2 mm in maximum diameter confirmed by cerebral arteriography. Some patients had other aneurysms presenting with SAH in the past; these aneurysms were required to be definitively treated prior to enrollment in the study. Patients were classified by group with regard to SAH history. Those who were identified to have had prior SAH from another intracranial aneurysm were classified as Group II and those with no history of prior SAH as Group I. Patients with fusiform, traumatic, or mycotic aneurysms were not included in the study. Patients were required to be functionally independent (Rankin11 scale score of 1 or 2) at the time of enrollment; thus, patients with a neurologically devastating prior hemorrhage were excluded from the study. Patients in whom the sole UIA was previously manipulated by wrapping,
J Neurosurg / Volume 121 / November 2014
packing, coil placement, proximal arterial ligation, bypass, balloon occlusion, or clip placement before entry into the study were not eligible. Patients with a history of intracranial hemorrhage from an unrepaired underlying structural lesion, primary intracerebral hemorrhage (without an underlying structural lesion), or SAH from an undetermined origin were excluded from the study. Patients with a malignant brain tumor were also excluded from the study. Patients were classified as “operated” if a surgical or endovascular procedure to treat at least 1 intracranial aneurysm was scheduled within 30 days of enrollment. Otherwise, patients were classified as “unoperated.” Follow-Up Evaluation
Baseline characteristics were documented for all enrolled patients via in-person interview with the patient and a comprehensive review of medical records. Unoperated patients were followed-up by routine examination or by questionnaire at yearly intervals. Operated patients were assessed at 7 days after the procedure, at hospital discharge, at 30 days after the procedure, and at yearly intervals thereafter. Neurological status as measured by the Rankin scale11 was collected at each time point, and cognitive status was measured by the Mini-Mental State Examination (MMSE)6 or the Telephone Interview for Cognitive Status (TICS)4 at the same time intervals. Complications related to the aneurysm procedure were assessed at the time of the procedure and at hospital discharge.
Determination of Events
Detailed information was obtained on all end points (definite or questionable subarachnoid or intracerebral hemorrhage and death). Comprehensive adjudication of ISUIA patients was performed centrally for all hemorrhages, ischemic strokes, and deaths on the basis of uniform criteria, with the use of available clinical, radiological, and autopsy findings and other information, and hemorrhages were classified according to the location of the rupture. Subarachnoid or intracerebral hemorrhage was classified as follows: 1) definite (symptoms of subarachnoid or intracerebral hemorrhage and positive findings on CT or MRI at surgery or at autopsy); 2) highly probable (symptoms and positive findings on CSF analysis); or 3) probable (symptoms only). In this analysis, definite or highly probable hemorrhages were included; probable hemorrhages were excluded. Evidence of procedure-related cerebral infarction, hemorrhage, or death was confirmed at the ISUIA coordinating center at the Mayo Clinic in Rochester, Minnesota. Perioperative hemorrhage included hemorrhages occurring 3 days or fewer after aneurysm treatment. Neurological deficits occurring 30 days to 1 year after treatment were evaluated for their relationship to treatment or coexisting neurological disorders. Deficits clearly related to a coexisting disorder were not attributed to aneurysm treatment. Poor treatment outcome was defined as Rankin scale score of 3, 4, or 5 (moderate to severe neurological disability) at the time of discharge. Aneurysmal morbidity was defined as worsening in neurological status attributed to aneurysmal hemorrhage 1025
K. B. Mahaney et al. or aneurysm procedure from the time of diagnosis to the first occurrence of a neurological change. This worsening in neurological status was considered as a Rankin scale score of 3, 4, or 5 (moderate to severe neurological disability), a score less than 24 on the MMSE, or a score less than 27 on the TICS. Mortality was assessed by date and cause of death. Interval to death was defined as time from date of diagnosis (untreated) or date of procedure (treated) to date of death. Cause of death was determined by contacting the local physician or by obtaining medical records, including death certificate and autopsy information when available. Procedure-related 1-year deaths included 1) death occurring within 30 days of the aneurysm treatment procedure; 2) death from complications of the aneurysm procedure any time during the year; 3) death from a repeat procedure to treat the aneurysm; or 4) death from aneurysmal hemorrhage after the procedure. A combined end point of aneurysmal morbidity and mortality attributable to aneurysmal hemorrhage or aneurysm procedure was defined as death prior to 366 days after diagnosis or assessment of a neurological deficit (Rankin scale score of 3, 4, or 5; TICS score < 27; or MMSE score < 24) at the observation closest to 365 days (within a ± 180-day window). Statistical Analysis Univariate Analysis. Univariate associations between age group and baseline characteristics, stratified by management group, were tested using chi-square and Fisher exact tests for categorical variables and ANOVA for continuous variables. Differences in case-fatality rates among management groups within each age group were tested using the chi-square test. In both the surgical group and the endovascular group, associations between age group and occurrence of procedural and in-hospital morbidities were tested using the chi-square and Fisher exact tests. In all prospective patients, within each age category, differences in hemorrhage rates, aneurysm- and procedure-related morbidity and mortality rates, and all-cause mortality rates were compared between management groups utilizing the log-rank test. Subgroup analysis was completed to determine if aneurysm site and aneurysm size influenced differences between management groups in hemorrhage rates and aneurysm- and procedure-related morbidity and mortality rates within each age group.
Multivariate Analysis. Multivariate models were completed for several endpoints.
One-Year Hemorrhage. Cox proportional hazard regression was used for multivariate models of aneurysmal hemorrhage risk. All baseline variables associated with risk of aneurysmal hemorrhage (at the p < 0.10 level) were entered into the multivariate model. Backward model selection was used to retain variables significant at the p < 0.05 level in the final model. A model was completed for each age group as well as two separate models inclusive of all prospective untreated patients (all ages). In the models inclusive of all prospective untreated patients, the variable age was forced; 1 model was completed with age as a cat1026
egorical variable and 1 model with age as a continuous variable.
Poor Treatment Outcome (Discharge Rankin Scale Score). For the end point of poor treatment outcome—de-
fined as Rankin scale score of 3, 4, or 5 at discharge following surgical or endovascular treatment—multivariate logistic regression models were completed for each age group within the surgical group and each age group within the endovascular group. Baseline characteristics, intraoperative events, and in-hospital events that were associated with poor treatment outcome in univariate analysis (at the p < 0.10 level) were entered into multivariate logistic regression models of poor treatment outcome. Backward model selection was used, keeping variables significant at the p < 0.05 level in the final models.
One-Year Aneurysm- and Procedure-Related Morbidity and Mortality. Multivariate logistic regression analysis
was used to model 1-year aneurysm- and procedure-related morbidity and mortality within each age group. All baseline variables with a univariate association (at the p < 0.10 level) with age or with occurrence of aneurysm- and procedure-related morbidity and mortality were entered into multivariate models of aneurysm- and procedurerelated morbidity and mortality, with 1 model for each age group. Backward model selection was used, retaining variables significant at the p < 0.05 level in the final models. The variables “treatment group” and “baseline Rankin scale score” were forced in the final models. One-Year All-Cause Mortality. Cox proportional hazard regression was used to model risk of 1-year mortality (from any cause) within each age group. All baseline variables with a univariate association (at the p < 0.10 level) with age or with 1-year all-cause mortality were entered into the models, with 1 model for each age group. Backward model selection was used, retaining variables significant at the p < 0.05 level in the final models.
Baseline Characteristics
Results
Age-related variability in baseline patient and aneurysm characteristics in patients with UIAs are reported (Tables 1 and 2). Differences in baseline characteristics between management groups in patients of the same age group are presented in Table 1. In all 3 age groups, the observation group had a significantly higher proportion of patients with prior SAH (UIA Group II), compared with surgical and endovascular groups (the endovascular group had the lowest fraction). The observation group also had a significantly higher proportion of patients with baseline Rankin scale scores of 2, compared with surgical and endovascular patients, which was observed in all 3 age groups. In patients younger than 50 years old, there was no significant difference in history of ischemic cerebrovascular disease between management groups. However, in patients 50–65 and > 65 years of age, history of ischemic cerebrovascular disease was significantly more common in the observation group than in the surgical or endovascular groups. History of hypertension varied significantly between management groups only for patients > 65 years J Neurosurg / Volume 121 / November 2014
Age-related differences in 1-year outcomes of UIAs TABLE 1: Patient baseline characteristics by age category and management group Characteristic no. of patients UIA Group II (SAH) 65 yrs old baseline Rankin score = 2 65 yrs old history of ischemic cerebrovascular disease 65 yrs old history of hypertension 65 yrs old smoking history 65 yrs old never current former recent aspirin use§ 65 yrs old yes no
Endovascular Treatment (%)
Surgical Treatment (%)
Untreated (%)
451*
1917†
1691‡
20 (11.6) 21 (11.2) 1 (1.1)
182 (21.5) 125 (15.0) 19 (8.1)
323 (54.1) 225 (33.7) 67 (15.7)
