Skip Navigation LinksHome > February 2013 - Volume 116 - Issue 2 > Incidence, Predictors, and Outcomes of Perioperative Stroke...
Anesthesia & Analgesia:
doi: 10.1213/ANE.0b013e31826a1a32
Neuroscience in Anesthesiology and Perioperative Medicine

Incidence, Predictors, and Outcomes of Perioperative Stroke in Noncarotid Major Vascular Surgery

Sharifpour, Milad MD, MS*; Moore, Laurel E. MD*; Shanks, Amy M. MS*; Didier, Thomas J. MD*; Kheterpal, Sachin MD, MBA*; Mashour, George A. MD, PhD*

Free Access
Article Outline
Collapse Box

Author Information

From the *Department of Anesthesiology, University of Michigan Medical School, Ann Arbor, MI.

Accepted for publication July 3, 2012.

This study was supported by departmental funds.

Dr. Milad Sharifpour and Dr. Laurel E. Moore contributed equally to the manuscript.

Milad Sharifpour, MD, MS, is currently with the Department of Anesthesia, Critical Care and Pain Medicine, Massachusetts General Hospital, Boston, MA.

The authors declare no conflicts of interest.

Reprints will not be available from the authors.

This report was previously presented, in part, at the 2010 annual ASA meeting, 2010 Annual SNACC Meeting.

Address correspondence to Laurel E. Moore, MD, Division of Neuroanesthesiology, Department of Anesthesiology, University of Michigan Medical School, 1H247 UH/Box 5048, 1500 East Medical Center Drive, Ann Arbor, MI 48109–5048. Address e-mail to laurelmo@med.umich.edu.

Collapse Box

Abstract

BACKGROUND: Perioperative stroke is a potentially catastrophic complication of surgery. Patients undergoing vascular surgery suffer from systemic atherosclerosis and are expected to be at increased risk for this complication. We studied the incidence, predictors, and outcomes of perioperative stroke after noncarotid major vascular surgery using the American College of Surgeons National Quality Improvement Program database.

METHODS: Forty-seven thousand seven hundred fifty patients undergoing noncarotid vascular surgery from 2005 to 2009 at nonVeterans Administration hospitals were identified from the American College of Surgeons National Quality Improvement Program database. An analysis of patients undergoing elective lower extremity amputation, lower extremity revascularization, or open aortic procedures was performed to determine the incidence, independent predictors, and 30-day mortality of perioperative stroke.

RESULTS: The overall incidence of perioperative stroke within 30 days of surgery (n = 37,927) was 0.6%. Multivariate analysis revealed that each 1-year increase in age [odds ratio 1.02, 95% confidence interval (CI) (1.01 to 1.04)], cardiac history [1.42, (1.07 to 1.87)], female sex [1.47, (1.12 to 1.93)], history of cerebrovascular disease [1.72, (1.29 to 2.29)], and acute renal failure or dialysis dependence [2.03, (1.39 to 2.97)] were independent predictors of stroke. Only 15% (95% CI, 11%–20%) of strokes occurred on postoperative day 0 or 1. Perioperative stroke was associated with a 3-fold increase in 30-day all-cause mortality [3.36, (1.77 to 6.36)] and an increased median surgical length of stay from 6 (95% CI, 2 to 28) to 13 (95% CI, 3 to 43) days (P < 0.001, WMWodds 2.5, 95% CI, 2.0 to 3.2) in a matched-cohort assessment.

CONCLUSION: Perioperative stroke is an important source of morbidity and mortality, as reflected by significant increases in median surgical length of stay and all-cause 30-day mortality. The independent predictors of stroke that we have identified in this population are not readily modifiable and the majority of strokes occurred after postoperative day 1. Additional studies are required to identify potentially modifiable intraoperative or postoperative risk factors of perioperative stroke.

Perioperative stroke is a potentially catastrophic complication of vascular surgery.1,2 Most perioperative strokes are ischemic and related to systemic atherosclerosis.3–6 Patients undergoing vascular surgery suffer from systemic atherosclerosis and many of the other known risk factors for stroke including hypertension, coronary artery disease, and diabetes mellitus. These patients are therefore expected to be at increased risk for the complication.1,2

In a recent study of perioperative stroke in adult patients undergoing noncardiac, nonvascular, and nonneurologic surgery, Mashour et al. found an incidence of 0.1% using the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) database.7 Patients undergoing noncarotid vascular surgery were excluded from that study because they are a higher risk population with an incidence of perioperative stroke between 0.5% and 3%.8,9 Of note, this is comparable to the incidence of perioperative myocardial infarction (MI) in the same population.2 Thus, perioperative stroke is an important adverse outcome in vascular patients and has been associated with an increased length of hospital stay (LOS), increased intensive care unit stay, and up to a 6-fold increased postoperative mortality.2,4,10 Although the risk factors for stroke after coronary artery bypass graft11–15 and carotid endarterectomy16 have been extensively studied and risk stratification models developed,13,14 there is a paucity of data regarding risk factors and predictors of perioperative stroke in patients undergoing noncarotid major vascular surgery.

Axelrod et al. reported the incidence and risk factors of perioperative stroke in 20,037 male patients who underwent noncarotid vascular surgery at Veterans Affairs (VA) hospitals from 1997 through 2000, using the VA patient treatment file and the VA NSQIP database. The study found a 0.4% to 0.6% incidence of stroke in patients undergoing lower extremity revascularization, open abdominal aortic surgeries, and lower extremity amputations. Independent predictors of perioperative stroke in that population were previous stroke or transient ischemic attack (TIA), preoperative mechanical ventilation, postoperative MI, and need to return to the operating room.2 However, this study was limited by the fact that it was focused only on males in the unique VA population. Because risk factors for stroke may be influenced by sex,17–19 it is important to assess the incidence and predictors using a model that incorporates sex as a covariate. Furthermore, it is unclear if the findings of Axelrod et al. can be generalized to the non-VA population. Therefore, the primary objective of this study was to identify the incidence, predictors, and outcomes of perioperative stroke (defined as occurring within 30 days after surgery) in a diverse patient population undergoing elective noncarotid major vascular surgery, using the prospectively gathered clinical dataset derived from the ACS-NSQIP. Because the Axelrod et al. study was limited to open procedures, and clinical practice is moving increasingly toward an endovascular approach to vascular disease, a secondary and exploratory analysis was performed to assess the influence of surgical approach [open versus endovascular abdominal aortic aneurysm (AAA) repair] on the incidence and predictors of perioperative stroke.

Back to Top | Article Outline

METHODS

Patients and Data Collection

IRB approval (University of Michigan Medical School, Ann Arbor, Michigan) was obtained for analysis of these prospectively collected, deidentified data that are publicly available. The requirement for written informed consent was waived by the IRB. The ACS-NSQIP methodology has been described in detail20–22 and is summarized here. Operations that require general, epidural, or spinal anesthesia are prospectively divided into 8-day cycles and the first 40 consecutive general and vascular surgery operations within each cycle are included. Each cycle begins on a different day of the week, which is done to ensure that the selected sample appropriately represents each hospital’s case variety. Additional operations on the same patient within 30 days of the initial procedure are not included.

For each operation, trained nurse data reviewers prospectively collect data on 136 variables in a standardized fashion. These variables include patient demographics, preoperative comorbidities, operative information, selected intraoperative elements, and 30-day postoperative morbidities and mortality. The nurses are trained in clinical medicine and quality assurance and are required to complete standardized training on data collection and the definitions of study variables to qualify for participation in ACS-NSQIP. Data uniformity and reliability are ensured by conducting periodic conference calls, annual meetings, and interrater reliability audits. If an interrater reliability audit score demonstrates ≥5% disagreement, the participating site is excluded from the ACS-NSQIP participant use data file (PUF). The PUF is a Health Insurance Portability and Accountability Act compliant data file containing cases submitted to the ACS-NSQIP. The 2005 to 2009 PUF is a compilation of operations from 211 participating non-VA United States medical centers (representing 41 states) submitted over a 5-year period.a To maintain institutional, provider, and patient anonymity, no site- or region-specific data elements are included in the PUF.

Back to Top | Article Outline
ACS-NSQIP Study Population and Variables Analyzed
Inclusion and Exclusion Criteria

We included adult patients (18 years or older) in the database who underwent elective lower extremity major amputation, lower extremity revascularization, open abdominal aortic procedures, and endovascular AAA repair (EVAR), as identified by primary procedure Current Procedural Terminology codes (CPT; Appendix 1). CPT codes were hand-reviewed to ensure the appropriateness of procedures included for analysis and were chosen to make our study comparable to the only other large study of the incidence and predictors of perioperative stroke in patients undergoing noncarotid vascular surgery.2 The cases analyzed in the current study were excluded from our prior study of perioperative stroke, which focused on low-risk populations;7 patient samples were therefore nonoverlapping. For our primary cohort we combined patients who underwent AAA repair, aortofemoral bypass procedures, and aortovisceral bypass procedures into a single category of open abdominal aortic procedures. Likewise, we combined all lower extremity revascularization procedures to treat occlusive disease and aneurysmal disease into the lower extremity revascularization procedure subset. Finally, we grouped all amputations above the ankle (both above-knee and below-knee) into lower extremity amputation procedures. Although trauma patients are excluded from the NSQIP database, it is possible that some patients with lower extremity amputations related to tumor or other etiologies unrelated to peripheral vascular disease may have been included in the analysis. The following case types were excluded from the primary cohort analysis: (1) procedures performed by the cardiac surgery, neurosurgery, urology, ophthalmology, and obstetrics services; (2) carotid endarterectomies; (3) procedures involving the ascending aorta and vertebral arteries; (4) amputations below the ankles; and (5) endovascular graft procedures. Patients <18 years of age, on preoperative ventilator support, or undergoing emergent operations were excluded. Variables of interest such as atrial fibrillation and valvular disease are not included in the ACS-NSQIP database and could not be analyzed.

Appendix 1
Appendix 1
Image Tools

Because of changing clinical practice since the Axelrod et al. study, a secondary and exploratory analysis was performed on a cohort of patients undergoing elective open AAA versus EVAR. Using ACS-NSQIP data from 2005 to 2008, Raval and Eskandari reported no difference in the incidence of stroke in patients older than 80 years of age between patients undergoing open AAA repair and EVAR.23 We broadened these entry criteria to include all patients 18 years of age and older and specifically compared the incidence and independent predictors of postoperative stroke in elective open AAA repair and EVAR. CPT codes are listed in Appendix 1 under secondary analysis. A summary of the surgical procedures, the variables analyzed, and the statistics performed for each cohort is included in Table 1.

Table 1
Table 1
Image Tools
Back to Top | Article Outline
Outcomes

Postoperative stroke and 30-day all-cause mortality were analyzed as the primary and secondary outcomes of interest, respectively. Postoperative stroke is defined by ACS-NSQIP as a new embolic, thrombotic, or hemorrhagic cerebrovascular event with motor, sensory, or cognitive dysfunction that persists for ≥24 hours within 30 days of the operation. Thirty-day mortality can be from any mechanism of death and the cause of death is not reported.

Back to Top | Article Outline
Patient Variables

Basic demographic data including age, sex, body mass index (BMI), and the American Society of Anesthesiologists (ASA) physical classification status were analyzed. Risk factors for perioperative stroke were identified using previously reported literature4,6,7,10,11,14,24,25 and included current tobacco use, diabetes mellitus (requiring oral or insulin therapy), congestive heart failure within 30 days before surgery, history of MI within 6 months before surgery, previous cardiac interventions, acute renal failure, dialysis dependence, history of bleeding disorders, history of hemiplegia, history of TIA, and history of stroke with or without neurologic deficit. Standardized definitions of ACS-NSQIP preoperative patient comorbidities that were analyzed in this study are available in Appendix 2.

Appendix 2
Appendix 2
Image Tools
Back to Top | Article Outline
Statistical Analysis

Statistical analysis was performed using PASW version 18 (SPSS, Chicago, IL). Descriptive statistics were performed on all categorical and continuous data elements using Pearson’s χ2, Fisher’s exact test, Student’s t-test, or Mann-Whitney U-test, where appropriate, to determine univariate associations with postoperative stroke. To build the logistic regression model, colinearity diagnostics and Pearson correlations were evaluated for all preoperative variables in Table 2, with the exception of ASA physical status. This variable was not included because the comorbidities included in the model constitute the individual criteria for the ASA physical status. Because of the possibility of artifactual BMI data, we classified valid BMIs as 10 to 60; any BMI outside this range was treated as a missing data element. Diabetes oral therapy and diabetes insulin therapy were collapsed into a single diabetes positive category and compared against no diabetes. Similarly, history of stroke with neurologic deficit, history of stroke without neurologic deficit, history of TIA, and preoperative hemiplegia were combined into a single category of history of neurologic event. Finally, we collapsed history of MI within 6 months before surgery, history of cardiac revascularization, and congestive heart failure within 30 days of surgery into a single category of cardiac history. To investigate the timing of the strokes, we collapsed the data into categories of stroke events occurring on postoperative day (POD) 0 or 1, POD 2 to 8, and POD 9 to 30. The 95% confidence interval (CI) was calculated using Newcombe’s method without a correction for continuity.

Table 2
Table 2
Image Tools

For the primary cohort analysis, all variables were entered into a nonparsimonious logistic regression model to identify independent predictors of perioperative stroke in patients undergoing open abdominal aortic, lower extremity revascularization, and lower extremity amputations. Variables with P < 0.05 were considered independent predictors of perioperative stroke. Effect size for each independent predictor was assessed using adjusted odds ratios and 95% CI. The model’s goodness of fit was tested using the Hosmer and Lemeshow test. This test groups the data into deciles based on risk levels of developing the outcome. The incidence of the observed outcome (stroke) is compared to the incidence of the predicted outcome (stroke) using the proposed model within each decile risk level. A P value that is not significant indicates that the observed strokes are no different than the predicted strokes based on the covariates entered into the model and therefore the fit of the model is deemed good. The model’s predictive value was evaluated using a receiver operating characteristic area under the curve (ROC AUC).26

For the primary cohort analysis, we developed a risk-matched cohort to assess whether the occurrence of perioperative stroke was associated with an increase in 30-day all-cause mortality. We used the perioperative stroke propensity score from the nonparsimonious logistic regression model to minimize the effect of confounding preoperative comorbidities on the 30-day all-cause mortality. The propensity score for each patient in this cohort was the predicted probability (0 to 1) of experiencing the primary outcome, a perioperative stroke. A 3-digit propensity score was used to match patients with perioperative stroke with patients who did not experience this outcome. Coding was created to randomly rank each patient in the nonstroke population with the same propensity score. The patients in the nonstroke population were then matched to the patients in the stroke population based on the same propensity score; we performed univariate analysis to assess the quality of the matching (Table 3). After the matching, we assessed the association between perioperative stroke and 30-day all-cause mortality using a Pearson χ2. Effect size is reported as odds ratio and 95% CI. We also investigated if perioperative stroke patients had longer surgical LOS using the Mann-Whitney U-test in the matched cohort.7,21 WMWodds were calculated to demonstrate the median increase in LOS with odds ratio and 95% CI using R studio.

Table 3
Table 3
Image Tools

For the secondary cohort analysis, the variables summarized in Table 1 were entered into a nonparsimonious logistic regression model to identify independent predictors of perioperative stroke in patients undergoing elective open AAA and EVAR procedures. Variables with P < 0.05 were considered independent predictors of perioperative stroke. Effect size for each independent predictor was assessed using adjusted odds ratios and 95% CI. The model’s goodness of fit was tested using the Hosmer and Lemeshow test. The model’s predictive value was evaluated using a ROC AUC.26 Given the nonrandom assignment of patients to these surgical subsets, univariate associations were performed to compare comorbidities between open AAA and EVAR patients (Table 4). All-cause 30-day mortality and surgical LOS were investigated for associations with perioperative stroke. Because of the smaller sample size, it was not feasible to perform a risk-adjusted cohort for these comparisons. A P value of <0.05 was considered statistially significant.

Table 4
Table 4
Image Tools
Back to Top | Article Outline

RESULTS

Incidence and Temporal Distribution of Perioperative Stroke: Primary Cohort Analysis

A total of 47,750 high-risk patients undergoing elective noncarotid vascular surgery from 2005 to 2009 at non-VA hospitals were identified from the ACS-NSQIP PUF. These included 8077 lower extremity amputations, 21,962 lower extremity revascularizations, 7888 open abdominal aortic procedures and 9823 EVAR procedures. For the primary cohort analysis, lower extremity amputations, lower extremity revasularizations, and open abdominal aortic procedures were included. Table 2 demonstrates demographics and univariate associations with perioperative stroke in these patients (n = 37,927). The overall incidence of perioperative stroke within 30 days after operation in the primary cohort was 0.6% (n = 228). The incidence of perioperative stroke across different surgical categories was as follows: 0.5% in lower extremity revascularizations (n = 109), 0.7% in amputations (n = 55), and 0.8% in open abdominal aortic procedures (n = 64). Fifteen percent (95% CI, 11%–20%) of strokes were identified during POD 0 and 1 (0.09% overall incidence of perioperative stroke), whereas 48.7% (95% CI, 42%–55%) of strokes occurred between POD 2 and 8 (Fig. 1).

Figure 1
Figure 1
Image Tools
Back to Top | Article Outline
Independent Predictors of Perioperative Stroke: Primary Cohort Analysis

Colinearity diagnostics did not demonstrate a condition index above 30 and therefore no bivariate correlation matrix was needed. All variables in Table 2 except ASA classification were therefore entered into the nonparsimonious logistic regression model. The model included 36,900 (97.3%) of the patients. Advancing age (each additional year of age, keeping all other variables constant), cardiac history, female sex, history of neurologic event, and acute renal failure or dialysis dependence (Fig. 2) were identified as independent predictors of perioperative stroke (P < 0.05). The Hosmer and Lemeshow test demonstrated a P = 0.601, suggesting a better than random fit. The ROC AUC was 0.66 ± 0.02.

Figure 2
Figure 2
Image Tools
Back to Top | Article Outline
30-Day Mortality Analysis Using a Matched Cohort: Primary Cohort Analysis

Two hundred twenty-five patients with perioperative stroke were matched in a 1:1 ratio to patients without perioperative stroke based on the previously derived propensity score. The matched cohort revealed no significant associations in the univariate analysis of the preoperative predictors of stroke (Table 3) and thus was matched for comorbid risk factors. Univariate analysis demonstrated that for all patients perioperative stroke was associated with a 3-fold increase in perioperative mortality within 30 days (95% CI, 1.8 to 6.4). The lower extremity revascularizations demonstrated an almost 7-fold increase in perioperative mortality within 30 days (95% CI, 2.4 to 18.1); amputations and open abdominal aortic procedures did not demonstrate statistical significance.

The matched cohort analysis also revealed that perioperative stroke is associated with increased median surgical LOS. For the entire cohort, perioperative stroke increased the median surgical LOS from 6 (95% CI, 2 to 28) to 13 (95% CI, 3 to 43) days (P < 0.001, WMWodds 2.8, 95% CI, 2.4 to 3.3). Median surgical LOS was increased from 9 days (95% CI, 3 to 35) to 13 (95% CI, 3 to 48) days (P = 0.03, WMWodds 1.8, 95% CI, 1.3 to 2.5) in patients undergoing major amputations complicated by perioperative stroke, and from 5 (95% CI, 1 to 24) days to 12 (95% CI, 2 to 38) days (P < 0.001, WMWodds 3.1, 95% CI, 2.4 to 4.1) in patients undergoing lower extremity revascularization procedures experiencing perioperative stroke. Similarly, perioperative stroke increased median surgical LOS from 7 days (95% CI, 3 to 49) to 15 (95% CI, 5 to 45) days (P < 0.001,WMWodds 3.1, 95% CI, 2.3 to 4.6) in open abdominal aortic procedures.

Back to Top | Article Outline
Incidence and Independent Predictors of Perioperative Stroke: Secondary Cohort Analysis

A total of 14,222 patients underwent open AAA repair or EVAR with an overall stroke incidence of 0.6%. Nine thousand eight hundred twenty-three (69.1%) of patients underwent EVAR and 4399 (30.9%) underwent an open repair of their AAA. The proportion of patients undergoing EVAR relative to open AAA increased each year between 2005 and 2009 (Fig. 3). The incidence of stroke in open procedures (0.8%) was more than in endovascular procedures (0.5%, P = 0.007) with an odds ratio of 1.8 (1.2 to 2.8). Table 4 shows the univariate analysis of preoperative risk factors between the two surgical groups. There were no differences in the incidence of ASA physical classification of 3, 4, or 5 or history of TIA or stroke between the 2 surgical groups (Table 4). The endovascular patients were older (74 ± 9 year vs 71 ± 9, P < 0.001) and had a higher incidence of renal failure or dialysis (P = 0.018), cardiac history (combined congestive heart failure within 30 days of procedure, history of MI within 6 months of procedure and history of previous cardiac intervention, P = 0.004), and diabetes (requiring oral or insulin therapy, P < 0.001). The open AAA group had a larger proportion of women (26% vs 17%, P < 0.001) and a larger proportion of patients with current tobacco use (P < 0.001) and treated hypertension (P < 0.001). The open AAA group also had a longer median operative time (216 vs 142 min, P < 0.001) and a higher incidence of patients receiving 4 or more units of packed red blood cells (P < 0.001). Independent predictors of perioperative stroke in this cohort (n = 14,222) included hypertension, history of renal failure or dialysis, prior neurologic event, female sex, open AAA (versus EVAR), diabetes, and age (P < 0.05; Fig. 4). The Hosmer and Lemeshow test demonstrated a P = 0.725, suggesting a good model fit. The ROC AUC was 0.73 ± 0.03. Perioperative stroke increased 30-day all cause mortality from 1.3% to 13.0% (P < 0.001) in the EVAR group (n = 9823) and from 3.7% to 10.8% (P = 0.025) in the open AAA group (n = 4399). Surgical LOS (both surgical groups, n = 14,222) increased from 3 (95% CI, 1 to 16) to 10 days (95% CI, 1 to 38) in those patients suffering stroke (P < 0.001, WMWodds 4.2, 95% CI, 3.1 to 6.1).

Figure 3
Figure 3
Image Tools
Figure 4
Figure 4
Image Tools
Back to Top | Article Outline

DISCUSSION

Using a national clinical database, we found a perioperative stroke incidence of 0.5% in lower extremity revascularizations, 0.7% in major lower extremity amputations, and 0.8% in open abdominal aortic reconstructions. We identified advancing age (each additional year of age, keeping all other variables constant), cardiac history, female sex, history of neurologic event, and acute renal failure or dialysis dependence as independent risk factors for perioperative stroke in this patient population. Of note, perioperative strokes tend to occur between POD 2 and 8 in this cohort (Fig. 1). Perioperative stroke is associated with a significantly worsened outcome for patients including a significantly longer median surgical LOS and a 3-fold increase in 30-day all-cause mortality.

Our results are consistent with the only other large study of the incidence and predictors of perioperative stroke in noncarotid vascular surgery. Also using the ACS-NSQIP database, Axelrod et al. found a perioperative stroke incidence of 0.4% to 0.6%, with an associated 6-fold increase in 30-day mortality and significantly longer LOS.2 Consistent with their results, we found the incidence of perioperative stroke was highest among patients undergoing abdominal aortic reconstruction and lowest in lower extremity revascularizations. The apparent discrepancy in the mortality rates (6-fold in Axelrod et al. study versus 3-fold in the primary cohort of the current study) is likely due to the fact that our comparison was performed in a comorbidity-matched cohort, which would be expected to yield a smaller difference in mortality. Furthermore, the design of the current study enabled us to identify female sex as a risk factor for stroke in this vascular surgery population.

Advanced age is associated with reduced cerebrovascular reserve as well as increased incidence and severity of stroke-associated comorbidities such as hypertension, cardiac arrhythmias, and atherosclerosis.4,6,24 Congestive heart failure puts patients at increased risk of embolic strokes.16 Furthermore, decreased left ventricular systolic function increases the risk of cerebral hypoperfusion and could increase the risk of perioperative stroke.15 Female sex has been shown to be a risk factor for perioperative stroke.4,10,14,27,28 However, the reasons for this increase in risk remain unclear. Possible explanations include higher rate of atrial fibrillation,18,19 higher embolic potential,29,30 incomplete inhibition of platelet aggregation with aspirin,31 and increased early restenosis after carotid endarterectomy, suggestive of differential vascular remodeling.32,33 However, these data should be interpreted with caution because none of the reported studies was designed to investigate the mechanisms explaining the observed outcome difference between females and males.

Preoperative hemiplegia reflects the severity of previous stroke, which has been identified as a strong risk factor for perioperative stroke in multiple studies and case reviews.4,6,7,10 This could reflect tenuous brain vascular reserve and impaired cerebral autoregulation in patients with previous stroke. Renal disease is a prominent risk factor for perioperative stroke in general and cardiovascular surgeries.10,14,15 This may, in part, be explained by the accelerated atherosclerosis that results from renal disease.34 Dialysis dependence also places patients at risk for repeated episodes of hypotension and cerebral hypoperfusion. Consistent with the study of Axelrod et al.,2 hypertension, diabetes mellitus, and smoking were not associated with perioperative stroke in our study. One likely explanation is that patients undergoing noncarotid vascular procedures have so many of these risk factors in common35,36 that making distinctions between them is difficult, even when using a large database such as ACS-NSQIP.

The majority of perioperative strokes occurred after POD 2, but within 8 days of the operation; only 15% of the strokes occurred between POD 0 and 1. This temporal distribution is consistent with findings from other studies of perioperative stroke2,24,37 and suggests that the postoperative rather than intraoperative period may be critical for stroke development in this population. None of the independent predictors of stroke identified in this study or the study by Axelrod et al. is potentially modifiable. This highlights the critical importance of conducting further prospective investigations to identify intraoperative and, given the timing of perioperative strokes, postoperative risk factors of perioperative stroke in the vascular surgery population.

Because clinical practice is changing and the use of EVAR is increasing over time,23,38 we performed a secondary cohort analysis to investigate the incidence of stroke and determine independent predictors in these patients. Although there have been numerous studies comparing morbidity and mortality in EVAR patients compared to open AAA repair, there is only one study that reports the incidence of stroke in such a large population. Raval and Eskandari23 compared the outcomes of patients ≥80 years undergoing EVAR or open AAA repair to patients younger than 80 years of age using the 2005 to 2008 ACS-NSQIP PUF. Not surprisingly, their overall incidence of stroke for all patients (all ages, both EVAR and open AAA repair) was identical to that found in our study (0.6%). Unlike our study, their comparison of the incidence of postoperative stroke in EVAR (0.8%) versus open AAA repair (1.8%) involved only patients ≥80 years (n = 2034) and there was no significant difference between groups. Our study has the advantage of a more diverse patient population (all patients 18 years or older), a larger number of patients and the inclusion of more recent data from ACS-NSQIP. In the current study there were a number of differences related to comorbid conditions in the EVAR versus open AAA cohort, which precluded propensity score matching. As such, no firm conclusions can be drawn regarding the relative risk of perioperative stroke in this patient cohort. Our results should therefore be regarded as hypothesis-generating for future studies in which open AAA and EVAR patients are appropriately matched.

The advantages of analyzing large nationwide clinical databases such as ACS-NSQIP include prospective standardized data collection, large number of cases included (which is ideal for detecting infrequent complications such as stroke), and diversity of the patient population. However, there are several limitations associated with such analyses. First, detailed clinical data and preoperative variables that have been shown to influence the risk of stroke in other patient populations such as atrial fibrillation and other arrhythmias,4,10,39 cardiac valvular diseases,10 perioperative anticoagulation,4 the use of perioperative β-adrenergic blockers,40,41 perioperative hypotension,42 and preoperative statin use43,44 are not elements in the ACS-NSQIP database. This limits the ability to address the contribution of these variables to perioperative stroke in our study. Similarly, because of the limited nature of intraoperative variables included in ACS-NSQIP database, factors such as intraoperative hypotension, management of anticoagulation, glycemic control, and intraoperative arrhythmias could not be addressed in our study. However, the low incidence of strokes on POD 0 and 1 in the current study suggests a limited contribution of these intraoperative variables to the overall risk of perioperative stroke. Another limitation of the ACS-NSQIP database is that no brain imaging studies are available; ischemic strokes could therefore not be distinguished from hemorrhagic strokes. The ACS-NSQIP database does not contain data on the prevalence and severity of preexisting carotid artery disease and therefore it was not possible to determine whether previous stroke was caused by an anatomical defect that had since been repaired.2 Basic laboratory data were not analyzed because of the large proportion of missing data elements. Finally, because of the limited cohort size for patients undergoing open AAA and EVAR, we were unable to perform propensity score matching as we were with the primary analysis. Conclusions based on this cohort are limited because patients were not randomly assigned to the 2 surgical subsets.

In conclusion, the incidence of perioperative stroke in the noncarotid vascular surgery population is low at 0.6% to 0.8%. Despite this, perioperative stroke is an important source of morbidity and mortality, as reflected by the significantly increased median surgical LOS and 30-day all-cause postoperative mortality. Furthermore, despite the low incidence of perioperative stroke in patients undergoing noncarotid vascular surgery, these patients have an approximately 6-fold increased risk for this complication compared to patients undergoing noncardiac, nonvascular, and nonneurologic surgery.2 The current investigation demonstrates the need for prospective studies to develop risk models for stroke after noncarotid vascular surgery. Given the limited value of preoperative predictors and the fact that in this high-risk population nearly half of strokes occur between POD 2 and 8, further studies are needed to identify modifiable postoperative risk factors to reduce the incidence and sequelae of stroke.

Back to Top | Article Outline

RECUSE NOTE

Dr. George Mashour is the Cover Editor for Anesthesia & Analgesia. This manuscript was handled by Dr. Gregory J. Crosby, Section Editor for Neuroscience in Anesthesiology and Perioperative Medicine, and Dr. Mashour was not involved in any way with the editorial process or decision.

Back to Top | Article Outline

DISCLOSURES

Name: Milad Sharifpour, MD, MS.

Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.

Attestation: Milad Sharifpour has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

Name: Laurel E. Moore, MD.

Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.

Attestation: Laurel E. Moore has seen the original study data, reviewed the analysis of the data, approved the final manuscript, and is the author responsible for archiving the study files.

Name: Amy M. Shanks, MS.

Contribution: This author helped conduct the study, analyze the data, and write the manuscript.

Attestation: Amy M. Shanks has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

Name: Thomas J. Didier, MD.

Contribution: This author helped conduct the study and analyze the data.

Attestation: Thomas J. Didier has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

Name: Sachin Kheterpal, MD, MBA.

Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.

Attestation: Sachin Kheterpal has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

Name: George A. Mashour, MD, PhD.

Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.

Attestation: George A. Mashour has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

This manuscript was handled by: Gregory J. Crosby.

Back to Top | Article Outline

RECUSE NOTE

Dr. George Mashour is the Cover Editor for Anesthesia & Analgesia. This manuscript was handled by Dr. Gregory J. Crosby, Section Editor for Neuroscience in Anesthesiology and Perioperative Medicine, and Dr. Mashour was not involved in any way with the editorial process or decision.

a Available at: http://www.acsnsqip.org/main/aboutacs/about_overview.jsp. Accessed August 8, 2011. Cited Here...

Back to Top | Article Outline

REFERENCES

1. Harris EJ Jr, Moneta GL, Yeager RA, Taylor LM Jr, Porter JM. Neurologic deficit following noncarotid vascular surgery. Am J Surg. 1992;163:537–40

2. Axelrod DA, Stanley JC, Upchurch GR Jr, Khuri S, Daley J, Henderson W, Demonner S, Henke PK. Risk for stroke after elective noncarotid vascular surgery. J Vasc Surg. 2004;39:67–72

3. Smith WS, English JD, Johnston SCFauci A, Braunwald E, Kasper DL, Hauser SL, Longo DL, Jameson JL, Loscalzo J. Cerebrovascular diseases. In: Harrison’s Principles of Internal Med. 200817th ed New York McGraw-Hill:2513–35

4. Selim M. Perioperative stroke. N Engl J Med. 2007;356:706–13

5. Kelley RE, Kovacs AG. Mechanism of in-hospital cerebral ischemia. Stroke. 1986;17:430–3

6. Kam PC, Calcroft RM. Perioperative stroke in general surgical patients. Anaesthesia. 1997;52:879–83

7. Mashour GA, Shanks AM, Kheterpal S. Perioperative stroke and associated mortality after noncardiac, nonneurologic surgery. Anesthesiology. 2011;114:1289–96

8. Johnston KW, Scobie TK. Multicenter prospective study of nonruptured abdominal aortic aneurysms: I. Population and operative management. J Vasc Surg. 1988;7:69–81

9. Cruz CP, Drouilhet JC, Southern FN, Eidt JF, Barnes RW, Moursi MM. Abdominal aortic aneurysm repair. J Vasc Surg. 2001;35:335–44

10. Bateman BT, Schumacher HC, Wang S, Shaefi S, Berman MF. Perioperative acute ischemic stroke in noncardiac and nonvascular surgery: incidence, risk factors, and outcomes. Anesthesiology. 2009;110:231–8

11. Bucerius J, Gummert JF, Borger MA, Walther T, Doll N, Onnasch JF, Metz S, Falk V, Mohr FW. Stroke after cardiac surgery: a risk factor analysis of 16,184 consecutive adult patients. Ann Thorac Surg. 2003;75:472–8

12. Ascione R, Reeves BC, Chamberlain MH, Ghosh AK, Lim KH, Angelini GD. Predictors of stroke in the modern era of coronary artery bypass grafting: a case control study. Ann Thorac Surg. 2002;74:474–80

13. McKhann GM, Grega MA, Borowicz LM Jr, Baumgartner WA, Selnes OA. Stroke and encephalopathy after cardiac surgery: an update. Stroke. 2006;37:562–71

14. Charlesworth DC, Likosky DS, Marrin CA, Maloney CT, Quinton HB, Morton JR, Leavitt BJ, Clough RA, O’Connor GT. Development and validation of a prediction model for strokes after coronary artery bypass grafting. Ann Thorac Surg. 2003;76:436–43

15. D’Ancona G, Saez de Ibarra JI, Baillot R, Mathieu P, Doyle D, Metras J, Desaulniers D, Dagenais F. Determinants of stroke after coronary artery bypass grafting. Eur J Cardiothorac Surg. 2003;24:552–6

16. Tu JV, Wang H, Bowyer S, Green L, Fang J, Kucey D. Risk factors for death or stroke after carotid endarterectomy: observations from the Ontario Carotid Endarterectomy Registry. Stroke. 2003;34:2568–73

17. Forster A, Gass A, Kern R, Wolf ME, Ottomeyer C, Zohsel K, Hennerici M, Szabo K. Gender differences in acute ischemic stroke: etiology, stroke patterns and response to thrombolysis. Stroke. 2009;40:2428–32

18. Roquer J, Campello AR, Gomis M. Sex differences in first-ever acute stroke. Stroke. 2003;34:1581–5

19. Holroyd-Leduc JM, Kapral MK, Austin PC, Tu JV. Sex differences and similarities in the management and outcome of stroke patients. Stroke. 2000;31:1833–7

20. Khuri SF, Henderson WG, Daley J, Jonasson O, Jones RS, Campbell DA Jr, Fink AS, Mentzer RM Jr, Steeger JE. The patient safety in surgery study: background, study design, and patient populations. J Am Coll Surg. 2007;204:1089–102

21. Kheterpal S, Tremper KK, Heung M, Rosenberg AL, Englelsbe M, Shanks AM, Campbell DA Jr. Development and validation of an acute kidney injury risk index for patients undergoing general surgery: results from a national data set. Anesthesiology. 2009;110:505–15

22. Shiloach M, Frencher SK Jr, Steeger JE, Rowell KS, Bartzokis K, Tomeh MG, Richards KE, Ko CY, Hall BL. Toward robust information: data quality and inter-rater reliability in the Am College of Surgeons National Surgical Quality Improvement Program. J Am Coll Surg. 2010;210:6–16

23. Raval MV, Eskandari MK. Outcomes of elective abdominal aortic aneurysm repair among the elderly: endovascular versus open repair. Surgery. 2012;151:245–60

24. Kim J, Gelb AW. Predicting perioperative stroke. J Neurosurg Anesthesiol. 1995;7:211–5

25. Limburg M, Wijdicks EF, Li H. Ischemic stroke after surgical procedures: clinical features, neuroimaging, and risk factors. Neurology. 1998;50:895–901

26. Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology. 1982;143:29–36

27. Goldstein LB, Samsa GP, Matchar DB, Oddone EZ. Multicenter review of preoperative risk factors for endarterectomy for asymptomatic carotid artery stenosis. Stroke. 1998;29:750–3

28. Sarac TP, Hertzer NR, Mascha EJ, O’Hara PJ, Krajewski LP, Clair DG, Karafa MT, Ouriel K. Gender as a primary predictor of outcome after carotid endarterectomy. J Vasc Surg. 2002;35:748–53

29. Hayes PD, Payne DA, Evans NJ, Thompson MM, London NJ, Bell PR, Naylor AR. The excess of strokes in female patients after CEA is due to their increased thromboembolic potential—analysis of 775 cases. Eur J Vasc Endovasc Surg. 2003;26:665–9

30. Laman DM, Wieneke GH, van Duijn H, van Huffelen AC. High embolic rate early after carotid endarterectomy is associated with early cerebrovascular complications, especially in women. J Vasc Surg. 2002;36:278–84

31. Cavallari LH, Helgason CM, Brace LD, Viana MA, Nutescu EA. Sex difference in the antiplatelet effect of aspirin in patients with stroke. Ann Pharmacother. 2006;40:812–7

32. Johnson CA, Tollefson DF, Olsen SB, Andersen CA, McKee-Johnson J. The natural history of early recurrent carotid artery stenosis. Am J Surg. 1999;117:433–6

33. den Hartog AG, Algra A, Moll FL, de Borst GJ. Mechanisms of gender-related outcome differences after carotid endarterectomy. J Vasc Surg. 2010;52:1062–71

34. Schiffrin EL, Lipman ML, Mann JF. Chronic kidney disease: effects on the cardiovascular system. Circulation. 2007;116:85–97

35. Cruz CP, Eidt JF, Capps C, Kirtley L, Moursi MM. Major lower extremity amputations at a Veterans Affairs hospital. Am J Surg. 2003;186:449–54

36. Nowygrod R, Egorova N, Greco G, Anderson P, Gelijns A, Moskowitz A, McKinsey J, Morrissey N, Kent KC. Trends, complications, and mortality in peripheral vascular surgery. J Vasc Surg. 2006;43:205–16

37. Ng JLW, Chan MTV, Gelb AW. Perioperative stroke in noncardiac, nonneurosurgical surgery. Anesthesiology. 2011;115:879–90

38. Mani K, Lees T, Beiles B, Jensen LP, Venermo M, Simo G, Palombo D, Halbakken E, Troeng T, Wigger P, Bjorck M. Treatment of abdominal aortic aneurysm in nine countries 2005–2009: a Vascunet report. Eur J Vasc Endovasc Surg. 2011 Jul 8 [Epub ahead of print]

39. Harthun NL, Stukenborg GJ. Atrial fibrillation is associated with increased risk of perioperative stroke and death from carotid endarterectomy. J Vasc Surg. 2010;51:330–6

40. Devereaux PJ, Yang H, Yusuf S, Guyatt G, Leslie K, Villar JC, Xavier D, Chrolavicius S, Greenspan L, Pogue J, Pais P, Liu L, Xu S, Malaga G, Avezum A, Chan M, Montori VM, Jacka M, Choi P. Effects of extended-release metoprolol succinate in patients undergoing non-cardiac surgery (POISE trial): a randomized controlled trial. Lancet. 2008;371:1839–47

41. van Lier F, Schouten O, van Domburg RT, van der Geest PJ, Boersma E, Fleisher LA, Poldermans D. Effect of chronic beta-blocker use on stroke after noncardiac surgery. Am J Cardiol. 2009;104:429–33

42. Gerraty RP, Gilford EJ, Gates PC. Watershed cerebral infarction associated with perioperative hypotension. Clin Exp Neurol. 1993;30:82–9

43. Bushnell CD, Griffin J, Newby LK, Goldstein LB, Mahaffey KW, Graffagnino CA, Harrington RA, White HD, Simes RJ, Califf RM, Topol EJ, Easton JD. Statin use and sex-specific stroke outcomes in patients with vascular disease. Stroke. 2006;37:1427–31

44. McGirt MJ, Perler BA, Brooke BS, Woodworth GF, Coon A, Jain S, Buck D, Roseborough GS, Tamargo RJ, Heller J, Freischlag JA, Williams GM. 3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitors reduce the risk of perioperative stroke and mortality after carotid endarterectomy. J Vasc Surg. 2005;42:829–36

Cited By:

This article has been cited 1 time(s).

British Journal of Anaesthesia
Perioperative database research: possibilities and pitfalls
Fleischut, PM; Mazumdar, M; Memtsoudis, SG
British Journal of Anaesthesia, 111(4): 532-534.
10.1093/bja/aet164
CrossRef
Back to Top | Article Outline

© 2013 International Anesthesia Research Society

Login

Become a Society Member