Results of carotid endarterectomy in patients with contralateral internal carotid artery occlusion from the Mid-America Vascular Study Group and the Society for Vascular Surgery Vascular Quality Initiative

From the Midwestern Vascular Surgical Society

Results of carotid endarterectomy in patients with contralateral internal carotid artery occlusion from the Mid-America Vascular Study Group and the Society for Vascular Surgery Vascular Quality Initiative Joseph R. Schneider, MD, PhD,a,b Julia B. Wilkinson, MD,a,b Thea J. Rogers, BA, MPH,a Michael J. Verta, MD,a,b Cheryl R. Jackson, DNP, MS, RN,a and Andrew W. Hoel, MD,b,c Winfield and Chicago, Ill

ABSTRACT Objective: Carotid endarterectomy (CEA) is among the most commonly performed vascular procedures. Some have suggested worse outcomes with contralateral internal carotid artery (ICA) occlusion. We compared patients with and patients without contralateral ICA occlusion using the Society for Vascular Surgery Vascular Quality Initiative database. Methods: Deidentified data were obtained from the Vascular Quality Initiative. Patients with prior ipsilateral or contralateral CEA, carotid stenting, combined CEA and coronary artery bypass graft, or <1-year follow-up were excluded, yielding 1737 patients with and 45,179 patients without contralateral ICA occlusion. Groups were compared with univariate tests, and differences identified in univariate testing were entered into multivariate models to identify independent predictors of outcomes and in particular whether contralateral ICA occlusion is an independent predictor of outcomes. Results: Patients with contralateral ICA occlusion were younger and more likely to be smokers; they were more likely to have chronic obstructive pulmonary disease, preoperative neurologic symptoms (56% vs 47%), nonelective CEA (16% vs 13%), and shunt placement (75% vs 53%; all P < .001). The 30-day ipsilateral stroke risk was 1.3% with vs 0.7% without contralateral ICA occlusion (P ¼ .004). The 30-day and 1-year survival estimates were 99.0% 6 0.5% and 94.1% 6 1.1% with vs 99.6% 6 0.1% and 96.0% 6 0.2% without contralateral ICA occlusion (log-rank, P < .001). Logistic regression analysis identified prior neurologic event (P ¼ .046), nonelective surgery (P ¼ .047), absence of coronary artery disease (P ¼ .035), and preoperative angiotensin-converting enzyme inhibitor use (P ¼ .029) to be associated with 30-day ipsilateral stroke risk, but contralateral ICA occlusion remained an independent predictor in that model (odds ratio, 2.29; P ¼ .026). However, after adjustment for other factors (Cox proportional hazards), risk of ipsilateral stroke (including perioperative) during follow-up was not significantly greater with contralateral ICA occlusion (hazard ratio, 1.21; P ¼ .32). Results comparing propensity score-matched cohorts mirrored those from the larger data set. Conclusions: This study demonstrates likely clinically insignificant differences in early stroke or death in comparing CEA patients with and those without contralateral ICA occlusion. After adjustment for other factors, contralateral ICA occlusion was not associated with a greater risk of ipsilateral stroke (including perioperative) in longer follow-up. Mortality was greater with contralateral ICA occlusion, and this difference was more pronounced at 1 year despite younger age of the contralateral ICA occlusion group. CEA risk remains low even in the presence of contralateral ICA occlusion and appears to be explained at least in part by other factors. CEA should still be considered appropriate in the face of contralateral ICA occlusion. (J Vasc Surg 2019;-:1-10.) Keywords: Carotid stenosis; Carotid endarterectomy; Contralateral internal carotid occlusion; Stroke

Carotid endarterectomy (CEA) is among the most commonly performed noncoronary artery interventions in North America.1 The operation has now been performed for >60 years, but debate continues about the appropriate role of CEA. One focus of the debate is the impact of contralateral internal carotid artery (ICA)

occlusion. Patients with contralateral ICA occlusion may have a higher risk of major adverse events after carotid surgery compared with those with patent contralateral ICA, but they may also have a different characteristic profile compared with patients without contralateral ICA occlusion. Patients with contralateral ICA occlusion were

From the Northwestern Medicine West Region, Winfielda; and the North-

The editors and reviewers of this article have no relevant financial relationships to

western University Feinberg School of Medicine,b and Northwestern Memo-

disclose per the JVS policy that requires reviewers to decline review of any

rial Hospital,c Chicago. Author conflict of interest: none. Presented at the Forty-second Annual Meeting of the Midwestern Vascular Surgical Society, St. Louis, Mo, September 13-15, 2018.

manuscript for which they may have a conflict of interest. 0741-5214 Copyright Ó 2019 by the Society for Vascular Surgery. Published by Elsevier Inc. https://doi.org/10.1016/j.jvs.2019.05.040

Correspondence: Joseph R. Schneider, MD, PhD, Vascular Surgery and Interventional Radiology Partners, Central DuPage Hospital, 202 OSB, 25 North Winfield Rd, Winfield, IL 60190 (e-mail: [email protected]).

1

2

Journal of Vascular Surgery

Schneider et al

---

excluded from some of the legacy trials of CEA, potentially limiting the generalizability of these results, and minimal information is available from the randomized prospective studies. The Society for Vascular Surgery Vascular Quality Initiative (VQI) database now contains data from about 100,000 CEA procedures. The VQI provides a potentially important data source to better detect small differences in outcomes that could have significant impact on decision-making and resource utilization. We therefore sought to use these data to understand any difference in outcome that might result from CEA in the setting of contralateral ICA occlusion.

ARTICLE HIGHLIGHTS d

d

METHODS d

A deidentified data set was obtained from the VQI national database including all CEA records from 2003 through mid-2017. We confined our analysis to firsttime CEA with no previous ipsilateral or contralateral carotid intervention and excluded a second contralateral CEA in the same patient, combined CEA and coronary artery bypass graft, “redo” CEA, and CEA after carotid stenting. Patients without 1-year follow-up data (defined in the VQI as any time between 9 and 21 months after CEA), including those whose procedures had been performed <1 year before the release of the data set, were also excluded. After these exclusions and inspection and “cleaning” of the data set, the data were imported into the standard statistics platform R.2 The size of the data set allowed exploration of the impact of multiple characteristics of the patients as possible predictors of outcome. We made univariate comparisons of patients’ characteristics and comorbidities as well as of procedural details (including age, sex, race, side of surgery, severity of stenosis, severity of contralateral stenosis, urgency of surgery, symptomatic status, use of a shunt, type of anesthesia, azotemia, dialysis dependence, diabetes mellitus, hypertension, dyslipidemia, smoking status, coronary artery disease [multiple variables], congestive heart failure, and beta-blocker use) and outcomes for patients with and without contralateral ICA occlusion. Symptomatic status was taken from the VQI variable prior NEUROEVENT, defined as history of cortical or ocular transient ischemic attack or stroke, vertebrobasilar transient ischemic attack or stroke, or “other” neurologic symptoms judged to be potentially cerebrovascular in etiology. Although this is not specific with respect to time or neuroanatomic location, the same is true for both study groups. Primary outcomes examined included periprocedural neurologic events, periprocedural death, and later neurologic events or death, all variables for which the VQI data set is specific. The initial analyses (except for length of stay and age) were examinations of simple 2  2 contingency tables, and rates between groups were compared and tested for difference using Fisher exact test or c2 test. Multivariate comparisons were performed to explore whether

2019

Type of Research: Retrospective review of prospectively collected multicenter data from the Vascular Quality Initiative Key Findings: The 1737 patients undergoing carotid endarterectomy had a greater stroke risk with contralateral internal carotid artery (ICA) occlusion than 45,179 patients without contralateral ICA occlusion (1.3% vs 0.7%; P ¼ .004). However, after adjustment for other patient factors including symptomatic status, this difference was not significant (hazard ratio, 1.21; P ¼ .32). Take Home Message: Any difference in neurologic risk in comparing patients with and patients without contralateral ICA occlusion is small and likely to be irrelevant with respect to decision-making for carotid endarterectomy.

patients’ characteristics other than contralateral ICA occlusion have independent predictive value with respect to the risk of adverse events. Kaplan-Meier methods were used to compare freedom from selected outcome variables including neurologic events and mortality. These comparisons were tested with the log-rank test. The inclusion of information from the Social Security Death Index in the VQI database allows examination of survival out to 10 years and more in this data set and also allowed Cox proportional hazards modeling to explore the possible impact of various patient factors on survival. Multivariate modeling used a strategy of backward variable selection with P ¼ .1 for entry and <.05 to retain the variable in the model. The variables included in the multivariate models were checked for multicollinearity. Variance inflation factors for the variables ranged from 1.00 to 1.135, indicating that multicollinearity did not factor in the resulting model. The covariance matrix also indicated similar results. These analyses were also verified with eigenvector analysis. We also tested our models using propensity score-matched subsets of patients with and without contralateral ICA occlusion. This study was approved by the Institutional Review Board of Northwestern Central DuPage Hospital on September 20, 2017 (research protocol 17-097). Because the study involved analysis of pre-existing deidentified data, the requirement for informed consent of the patient was waived.

RESULTS Our initial data set included 86,854 CEA procedures. After exclusions as described before, a total of 46,916 procedures remained for analysis. Although there were some differences between the included and excluded patients based on characteristics examined in Tables I

Journal of Vascular Surgery Volume

-,

Number

Schneider et al

3

-

Table I. Baseline characteristics of patients undergoing carotid endarterectomy (CEA)

Age, years

Contralateral ICA occlusion (n ¼ 1737)

Contralateral ICA patent (n ¼ 45,179)

68.2 6 9.2

70.2 6 9.5

<.0001 <.0001

P value

Sex Male

1265 (72.8)

27,061 (59.9)

Female

472 (27.2)

18,118 (40.1)

White race

1614 (92.9)

41,760 (92.4)

Current

650 (37.4)

12,256 (27.1)

Previous

832 (47.9)

21,323 (47.2)

.50

Smoking status

Never

<.0001

253 (14.6)

11,508 (25.5)

1543 (88.8)

39,770 (88.0)

.38

Diabetes mellitus

591 (34.0)

15,493 (34.3)

.78

Coronary artery disease

477 (27.5)

12,039 (26.7)

.44

History of congestive heart failure

169 (9.7)

4269 (9.5)

Chronic obstructive pulmonary disease

431 (24.8)

9437 (20.9)

Hypertension

On maintenance dialysis

.69 <.0001

17 (1.0)

448 (1.0)

.96

Preoperative living at home

1686 (97.0)

44,484 (98.5)

<.0001

Preoperative statin

1437 (82.7)

35,641 (78.9)

.0001

Preoperative beta blocker

952 (54.8)

24,542 (54.3)

.71

Preoperative aspirin

1434 (82.6)

37,426 (82.8)

.92

Preoperative ACE inhibitor

879 (50.6)

16,983 (37.6)

<.0001

977 (56.3)

21,216 (47.0)

<.0001

1576 (90.7)

40,230 (89.1)

.0269

5863 (13.0)

.0004

Any prior neurologic event Ipsilateral disease severity $70% Urgent or emergent procedure

276 (15.9)

ACE, Angiotensin-converting enzyme; ICA, internal carotid artery. Categorical variables are presented as number (%). Continuous variables are presented as mean 6 standard deviation. Missing values are excluded from percentage calculations. P calculated for continuous variables using Wilcoxon rank sum test. P calculated for categorical variables using Fisher exact test unless the contingency table was larger than 2  2 or the sample size was too large to allow calculation of Fisher exact test, in which case P was calculated using c2 test.

and II, these differences were in every case small and likely to be of no clinical significance. In particular, any in-hospital stroke was observed in 1.67% of the included patients and 2.08% of the excluded patients. Although this difference was significant (P < .001), this and the other differences may be explained by the fact that patients such as those undergoing combined CEA and coronary bypass, redo CEA, and CEA after previous carotid stenting, who might be expected to be at greater risk of stroke, were among those excluded. In any case, these small differences seem unlikely to be related to any sort of systematic reporting bias. Selected characteristics of the patients are listed in Table I. Patients with contralateral ICA occlusion tended to be younger. They were more likely to be male, to be tobacco smokers, to have suffered symptoms thought to be related to their carotid disease, and to have undergone elective CEA. Procedural and outcome details are listed in Table II. Not surprisingly, patients with contralateral ICA occlusion were more likely to have had a temporary shunt placed during CEA. In addition, they were more likely to have

had surgery using general anesthesia and to have received protamine at the end of the procedure. With the exception of use of a shunt, these differences were numerically small. Based on available data, operations were otherwise conducted in a similar way in the two groups. The risk of perioperative ipsilateral stroke was greater among patients with contralateral ICA occlusion (Table II). Because strokes related to emboli arising from the area of CEA or technical defects, dissection, thrombosis, or other complications at the site of CEA would conceivably put both sides of the brain at risk, we also compared the risk of any perioperative (30 days) stroke (either side); univariate comparisons yielded a greater risk of any perioperative stroke with contralateral ICA occlusion. Kaplan-Meier comparison of freedom from stroke and any ipsilateral neurologic event showed that despite the slightly greater perioperative risk of stroke with contralateral ICA occlusion, there was no continuing divergence of freedom from stroke, that is, there was not a continuing greater risk during follow-up (Fig 1, A

4

Journal of Vascular Surgery

Schneider et al

---

2019

Table II. Operative details and outcomes of patients undergoing carotid endarterectomy (CEA) Contralateral ICA occlusion (n ¼ 1737) Local or regional anesthesia Use of a shunt

Contralateral ICA patent (n ¼ 45,179)

129 (7.4)

4104 (9.1)

1300 (74.8)

23,876 (52.9)

P value .018 <.001

Heparin use

1721 (99.1)

44,597 (98.7)

.34

Protamine use

1196 (68.9)

28,789 (63.7)

<.001

Dextran use

207 (11.9)

Use of a drain

694 (40.0)

15,292 (33.9)

439 (25.3)

13,445 (29.8)

4933 (10.9)

.20 <.001

Intraoperative monitoring EEG Stump pressure

112 (6.5)

3901 (8.6)

1186 (68.3)

27,883 (61.6)

Doppler ultrasound

990 (57.0)

26,016 (57.6)

Duplex ultrasound

283 (16.3)

7852 (17.4)

39 (2.3)

1028 (2.3)

Unknown

<.001

Intraoperative completion study

Angiography Unknown

.33

425 (24.5)

10,283 (22.8)

Procedure time, minutes

122 6 51

116 6 45

Any postoperative myocardial infarction

12 (0.7)

305 (0.7)

Any cranial nerve injury

41 (2.4)

1549 (3.4)

.016

Any new ipsilateral stroke (in hospital)

27 (1.3)

397 (0.7)

.004

Any stroke (in hospital)

43 (2.5)

742 (1.6)

.009

<.001 .94

Perioperative death (30 days)

17 (1.0)

197 (0.4)

.003

Perioperative return to operating room for neurologic event

10 (0.6)

140 (0.3)

.054

Perioperative return to operating room for bleeding

17 (1.0)

396 (0.9)

.65

Point estimate (lower and upper 95% CI limits) 30-day survival

99.0% (98.6-99.5)

99.6% (99.5-99.6)

.001

1-year survival

94.1% (93.0-95.2)

96.0% (95.8-96.2)

<.001

1-year freedom from reoperation

99.6% (99.3-99.9)

99.8% (99.7-99.8)

.17

CI, Confidence interval; EEG, electroencephalography; ICA, internal carotid artery. Categorical variables are presented as number (%). Continuous variables are presented as mean 6 standard deviation. P calculated for continuous variables using Wilcoxon rank sum test. P calculated for categorical variables using Fisher exact. P calculated for survival and freedom from adverse events at 30 days and 1 year using log-rank test.

and B). However, Kaplan-Meier comparison of estimated survival showed a continued divergence of survival over time, that is, patients with contralateral ICA occlusion suffered greater risk of death throughout follow-up despite that group’s younger average age (Fig 1, C). Multivariate analysis (logistic regression) determined that prior neurologic event, nonelective surgery, and preoperative use of angiotensin-converting enzyme inhibitor were associated with 30-day ipsilateral stroke risk; but after adjustment for these confounders, contralateral ICA occlusion remained an independent predictor with an odds ratio of 2.29 (Table III). Curiously, this analysis suggested that the presence of coronary artery disease was associated with a lower risk of perioperative ipsilateral stroke, that is, that coronary artery disease was in some way “protective.” Multivariate analysis over time (Cox proportional hazards model) found that after adjustment for confounders

including prior neurologic event, smoking, diabetes mellitus, preoperative aspirin use, preoperative angiotensinconverting enzyme inhibitor use, ipsilateral stenosis $70%, urgent or emergent CEA, and perioperative myocardial infarction, there was not a significant difference in ipsilateral stroke in comparing patients with and those without contralateral ICA occlusion (hazard ratio, 1.21; P ¼ .32; Table IV). This analysis also suggested a lower long-term risk of ipsilateral stroke among patients with ipsilateral stenosis $70% (hazard ratio estimate, 0.73), that is, that a very high grade of stenosis would be protective in the longer term. Because the sizes of the two groups were significantly different and there were significant differences in the overall characteristics of the two groups, we also used propensity score matching to select similar subsets of patients in the two groups. Variables included in the propensity score-matched model were prior neurologic

Journal of Vascular Surgery Volume

-,

Number

A

Schneider et al

Freedom from Stroke with 95% Hall-Wellner Bands 1.00

Censored

0.98

0.96

0.94

0.92

0.90 73

0

292

146 219 Time to Stroke (Days) Total Occlusion

No

365

Yes

Number of subjects entering the next interval Occluded Not Occluded

B

30 days

180 days

365 days

1694 44460

1685 44300

1677 44083

Freedom from Ipsilateral Event with 95% Hall-Wellner Bands 1.00

0.98

0.96

0.94

0.92

0.90

Censored

73

0

292

146 219 Days to Ipislateral Event Total Occlusion

No

365

Yes

Number of subjects entering the next interval Occluded Not Occluded

C

30 days

180 days

365 days

1716 44956

1683 44315

1627 1683

Survival with 95% Hall-Wellner Bands 1.00

Censored

0.98

0.96

0.94

0.92

0.90 73

0

292

146 219 Time to Death (Days) Total Occlusion

No

365

Yes

Number of subjects entering the next interval Occluded Not Occluded

5

-

30 days

180 days

365 days

1720 44989

1688 4371

1634 43369

Fig 1. Kaplan-Meier estimates of (A) freedom from stroke, (B) any ipsilateral event, and (C) survival from the entire data set of 1737 patients with and 45,179 patients without contralateral internal carotid artery (ICA) occlusion. Standard error of the mean remains <10% to the limit of the graph at 365 days of follow-up.

event, race, sex, smoking status, hypertension, diabetes, coronary artery disease, congestive health failure, chronic obstructive pulmonary disease, active dialysis, location before admission, preoperative statin use, preoperative beta-blocker use, preoperative aspirin use, ipsilateral occlusion >70%, emergency procedure, use of general anesthesia, shunt use, heparin use, protamine use, dextran use, postoperative myocardial infarction, cranial nerve injury, return to the operating room during the perioperative stay, and discharge to long-term care or skilled nursing facility. The area under the curve for the propensity score-matched model was 0.697 (95% confidence interval, 0.684-0.709). There were 1702 casecontrol pairs matched on propensity score. There were 35 contralateral ICA occlusion patients who were unable to be matched by propensity score. Univariate comparisons of the resultant groups (1702 patients with and 1702 patients without contralateral ICA occlusion) using all of the same variables listed in the comparisons in Table I showed that with the exception of age (68.2 6 9.2 years with vs 69.2 6 9.4 years without contralateral ICA occlusion; P < .001), there were no significant differences between the two cohorts, that is, the preoperative differences seen in the unmatched sample were no longer present after propensity score matching. Comparison of the same variables listed in Table II showed that with the exception of heparin use (99.1% vs 98.7%; P ¼ .049) and procedure time (122 minutes vs 116 minutes; P ¼ .017), the two propensity-matched groups were similar with respect to perioperative and early postoperative outcomes, and these differences are clearly numerically small in any case. With respect to any perioperative stroke, the trend was similar to Table II (2.5% with vs 1.6% without contralateral occlusion), but this difference was not significant (P ¼ .09) in the comparison of the propensity scorematched groups. Kaplan-Meier comparison of any stroke and of any ipsilateral event again showed no significant divergence of risk during the first year (Fig 2, A and B). Multivariate modeling was also performed with patients nested within propensity score matches. Cox proportional hazards analysis suggested that white race and congestive heart failure were predictive of a greater risk of stroke and that male sex was actually associated with a lower risk of stroke (Table V). However, contralateral ICA occlusion was again not an independent predictor of stroke among the propensity score-matched patients (Table V; hazard ratio, 1.28; P ¼ .19). The consistently greater mortality risk estimate with contralateral ICA occlusion persisted even after propensity score matching (Fig 2, C). Nesting by surgery center did not find any evidence for center effect on these outcomes. Nesting by a center’s contralateral ICA occlusion volume quartiles did not produce any significant change in the results of hypothesis testing within the propensity score-matched groups.

6

Journal of Vascular Surgery

Schneider et al

---

2019

Table III. Odds ratio estimates for any ipsilateral perioperative stroke from multivariate logistic regression model 95% Confidence interval limits Estimate

Lower

Upper

P value

Contralateral ICA occlusion

2.29

1.10

4.76

.026

Prior neurologic event

1.60

1.01

2.53

.046

Coronary artery disease

0.55

0.31

0.96

.035

Effect

Preoperative use of ACE inhibitor

1.60

1.05

2.44

.029

Urgent or emergent CEA

1.70

1.01

2.86

.047

ACE, Angiotensin-converting enzyme; CEA, carotid endarterectomy; ICA, internal carotid artery.

Table IV. Hazard ratio estimates for ipsilateral stroke recorded at any time from surgery to last follow-up from multivariate Cox proportional hazards model 95% Confidence interval limits Effect

Lower

Contralateral ICA occlusion

1.21

0.83

1.74

.32

Prior neurologic event

1.72

1.45

2.05

<.0001

Preoperative smoking

1.13

1.02

1.26

.024

Diabetes mellitus

1.18

1.00

1.39

.045

0.81

0.67

0.99

.038 .008

Preoperative aspirin Preoperative ACE inhibitor

Upper

P value

Estimate

1.24

1.06

1.46

0.73

0.59

0.91

.005

Urgent or emergent CEA

1.67

1.38

2.03

<.0001

Postoperative MI

3.64

2.21

5.98

<.0001

Ipsilateral carotid stenosis $70%

ACE, Angiotensin-converting enzyme; CEA, carotid endarterectomy; ICA, internal carotid artery; MI, myocardial infarction.

DISCUSSION This study determined that the difference in neurologic event rates between patients undergoing CEA with and without contralateral ICA occlusion is small and is in part explained by other confounders. The difference is likely to be of no practical significance, and outcomes for patients with contralateral ICA occlusion have lower rates than the thresholds recommended in current guidelines.3 Our analysis suggests that surgeons should make their recommendations for CEA independent of the status of the contralateral ICA and that event rates will be low in any case and within the recommended guidelines. This is a large “real-world” experience and not likely to be vulnerable to reporting bias. We believe that the only similarly sized carotid registry data set that has been examined was amalgamated from nearly 400 mainly European centers by Menyhei et al.4 Our findings are also generally similar to those in a study performed 5 years earlier using data from the Vascular Study Group of New England.5 The precedent literature regarding comparative outcomes among patients with contralateral ICA occlusions is too large to cite every article, but in general, singlecenter studies have reported no significant differences in stroke outcome between patients with and patients without contralateral ICA occlusions.6-36 In fact, some single-center studies have reported no strokes among

patients with contralateral ICA occlusion.9,12,14,15,25,27,36,37 However, a small number of single-institution studies have found inferior outcomes in patients with contralateral ICA occlusion.38-41 Furthermore, some randomized prospective trials, including the North American Symptomatic Carotid Endarterectomy Trial (NASCET) and the Asymptomatic Carotid Atherosclerosis Study (ACAS), noted an increased risk of neurologic events after CEA in patients with contralateral ICA occlusion.42,43 However, data from registries have been less uniform, but at least some have concluded that perioperative neurologic risk is greater in the case of contralateral ICA occlusion. For examples, analysis of data from the National Surgical Quality Improvement Program identified a trend toward worse neurologic outcomes in patients with contralateral ICA occlusion,44 and data from the National Cardiovascular Data Registry’s Carotid Artery Revascularization and Endarterectomy (CARE) registry indicated worse outcomes with contralateral ICA occlusion.45 On the other hand, analyses of data from other registries have concluded that there is no difference in neurologic outcomes in these groups,46,47 although Ricotta et al,47 using data from the Society for Vascular Surgery registry for carotid intervention, interestingly found significantly worse neurologic risk with contralateral ICA occlusion among symptomatic but not among asymptomatic patients, seeming to

Journal of Vascular Surgery -,

Volume

A

Number

Schneider et al

Freedom from Stroke with 95% Hall-Wellner Bands 1.00

Censored

0.98 0.96 0.94 0.92 0.90 73

0

292

146 219 Time to Stroke (Days)

Total Occlusion

No

365

Yes

Number of subjects entering the next interval 30 days

1659 1674

Occluded Not Occluded

B

180 days

365 days

1650 1668

1642 1663

Freedom from Ipsilateral Event with 95% Hall-Wellner Bands

1.00 0.98 0.96 0.94 0.92

0.90

Censored

0

73

292

146 219 Days to Ipislateral Event

Total Occlusion

365

Yes

No

Number of subjects entering the next interval Occluded Not Occluded

C

30 days

1688 1696

180 days

365 days

1687 1696

1685 1695

Survival with 95% Hall-Wellner Bands

1.00

Censored

0.98 0.96 0.94 0.92 0.90 73

0

292

146 219 Time to Death (Days)

Total Occlusion

No

365

Yes

Number of subjects entering the next interval 30 days

Occluded Not Occluded

7

-

1685 1694

180 days

1653 1665

365 days

1600 1618

Fig 2. Kaplan-Meier estimates of (A) freedom from stroke, (B) any ipsilateral event, and (C) survival comparing propensity score-matched subsets of 1702 patients with and 1702 patients without contralateral internal carotid artery (ICA) occlusion. Standard error of the mean remains <10% to the limit of the graph at 365 days of follow-up.

conflict with the observation of Baker et al43 from the ACAS. Population-based administrative data sets and registries from other countries have most often concluded that outcomes are worse in those with contralateral ICA occlusion,4,48,49 although there have been exceptions.50 Studies examining data from the Vascular Study Group of New England have concluded that outcomes are worse with contralateral ICA occlusion.51 Rockman,52 in an invited review in 2004, concluded that the literature in general supported the view that there is no neurologic outcome difference. However, more formal pooled analyses (focused reviews and meta-analyses) have generally concluded that there is greater risk of neurologic events in patients with contralateral ICA occlusions.40,41,53 The apparent discrepancy between conclusions from the single-center articles and the pooled analyses may be as simple as publication bias (inferior results tend not to be published)54 or, alternatively, statistically insignificant trends toward worse outcomes with ICA occlusion in many of the single-center studies vs statistically significant trends from the pooled analyses because pooling increases the sample size adequately to more securely answer the research question. In general, large registries are regarded as less likely to reflect publication bias. With the exception of the remarkable findings from NASCET and ACAS,42,43 the actual differences in outcomes after CEA with and without contralateral ICA occlusion, if any, have been numerically small and probably of marginal clinical significance. The outcomes and riskbenefit comparisons may be different among symptomatic vs asymptomatic patients considered for CEA in the face of contralateral ICA occlusion. The role of shunts in CEA remains controversial as well, and the possible differential role of shunts in patients with contralateral ICA occlusion remains controversial. Some argue that no patient needs to be shunted (regardless of the status of the contralateral ICA)18,19,34,37,55 or alternatively that all patients need to be shunted, that every patient with contralateral ICA occlusion should be shunted,56 or that shunting decisions can be made with standard selection algorithms using electroencephalography monitoring25 or stump pressures or by performing CEA in an awake patient. The status of the contralateral ICA may be used in the decision-making algorithm by those who practice selective shunting. However, Goodney et al57 concluded that a practice of selective shunting (as opposed to routine shunting) may actually increase the risk of morbid events in the face of CEA with contralateral ICA occlusion and explored the possible implication that familiarity with shunting gained by a practice of routine shunting may make shunting safer “when you need it.” Suffice it to say that there is a broad range of practice with respect to shunting. Finally, with the exception of ACAS,43 most studies have concluded that the natural history of patients with an

8

Journal of Vascular Surgery

Schneider et al

---

2019

Table V. Hazard ratio estimates for ipsilateral stroke recorded at any time from surgery to last follow-up from multivariate Cox proportional hazards model, propensity score-matched subsets 95% Confidence interval limits Estimate

Lower

Upper

P value

1.14

0.90

1.44

.29

Male sex

0.76

0.59

0.98

.04

White race

2.24

1.21

4.13

.01

Congestive heart failure

2.03

1.48

2.79

<.001

Effect Contralateral ICA occlusion

ICA, Internal carotid artery.

ICA occlusion treated with best medical therapy is one of a greater risk of strokes.42,58,59 Thus, even if the morbidity is greater for CEA in the face of contralateral ICA occlusion, the net benefit compared with best medical therapy for these patients may be substantial and may still justify CEA in these patients. Because the VQI does not contain any information about the outcomes in medically treated patients, we were unable to explore the question of possible different relative benefit (or harm) of CEA among patients treated with CEA vs best medical therapy. Two observations that surprised us were the apparent long-term protective effect of coronary artery disease and ipsilateral stenosis $70%. With respect to coronary artery disease, we can only surmise that this may reflect more aggressive management of atherosclerosis among patients with identified coronary artery disease and that this may be protective around the time of CEA. With respect to ipsilateral stenosis $70%, we are unable to suggest a plausible explanation. Our observation of a greater mortality risk over time among the patients with contralateral occlusion despite the younger average age is remarkable. This difference persisted after multivariate adjustment and even after comparing the propensity score-matched cohorts. No information is available about causes of death in the VQI database. However, it seems likely that these patients suffer from some greater systemic disease, possibly greater atherosclerotic disease burden. Limitations. Whereas important findings arise from our analysis, this study does have limitations. As a nonrandomized, retrospective analysis of prospectively collected data, it is possible that selection of patients for intervention varies significantly across VQI participating centers, resulting in an increased risk that the two study samples are different. Propensity scoreweighted analysis provides some increased confidence that any difference or the lack thereof is not due to heterogeneity across participating centers. The VQI data set is presently limited to 1-year follow-up, and achieving complete 1-year follow-up has been a challenge. We have addressed the follow-up problem by excluding patients without 1-year follow-up.

Furthermore, the stroke/mortality/major adverse event rate is so low in most CEA studies that it may be difficult to identify a difference in end points between patients with and patients without contralateral ICA occlusion. However, the size of the VQI data set provides a high level of confidence that a true difference of any magnitude would likely be detected. Finally, we have not compared CEA with carotid artery stenting, but Nejim et al60 made such a comparison using VQI data and concluded that neurologic risk is consistently greater for carotid artery stenting than for CEA in the face of contralateral ICA occlusion. Sample size modeling has not been performed because the precedent literature is likely to be subject to publication bias, and there is no legitimate way to predict any possible outcome differences.

CONCLUSIONS This large study of CEA in the setting of contralateral occlusion demonstrates nominal and likely clinically insignificant differences in stroke or death compared with CEA in the absence of a contralateral occlusion. Outcome differences noted in the absence of propensity score matching suggest that patient factors other than carotid occlusion contribute to adverse outcomes. Furthermore, given an observation of consistently increased medium-term mortality despite younger average age, our data suggest that worse outcomes observed in the setting of contralateral ICA occlusion may be a marker of patients with more advanced comorbidities. Therefore, we conclude that CEA may still be considered appropriate in the setting of contralateral ICA occlusion.

AUTHOR CONTRIBUTIONS Conception and design: JS Analysis and interpretation: JS, TR Data collection: JS, JW, MV, CJ, AH Writing the article: JS Critical revision of the article: JS, JW, TR, MV, CJ, AH Final approval of the article: JS, JW, TR, MV, CJ, AH Statistical analysis: JS, TR Obtained funding: Not applicable Overall responsibility: JS

Journal of Vascular Surgery Volume

-,

Number

Schneider et al

9

-

REFERENCES 1. National Institute of Neurological Disorders and Stroke, National Instutues of Health. Questions and answers about carotid endarterectomy. Available at: http://www.ninds.nih. gov/disorders/stroke/carotid_endarterectomy_backgrounder. htm. Accessed May 1, 2013. 2. The R Project for Statistical Computing. Available at: https:// www.r-project.org/. Accessed July 27, 2017. 3. Ricotta JJ, Aburahma A, Ascher E, Eskandari M, Faries P, Lal BK, et al. Updated Society for Vascular Surgery guidelines for management of extracranial carotid disease: executive summary. J Vasc Surg 2011;54:832-6. 4. Menyhei G, Bjorck M, Beiles B, Halbakken E, Jensen LP, Lees T, et al. Outcome following carotid endarterectomy: lessons learned from a large international vascular registry. Eur J Vasc Endovasc Surg 2011;41:735-40. 5. Pothof AB, Soden PA, Fokkema M, Zettervall SL, Deery SE, Bodewes TC, et al. The impact of contralateral carotid artery stenosis on outcomes after carotid endarterectomy. J Vasc Surg 2017;66:1727-34.e2. 6. Mackey WC, O’Donnell TF Jr, Callow AD. Carotid endarterectomy contralateral to an occluded carotid artery: perioperative risk and late results. J Vasc Surg 1990;11:778-83; discussion: 784-5. 7. Mattos MA, Barkmeier LD, Hodgson KJ, Ramsey DE, Sumner DS. Internal carotid artery occlusion: operative risks and long-term stroke rates after contralateral carotid endarterectomy. Surgery 1992;112:670-9; discussion: 679-80. 8. Perler BA, Burdick JF, Williams GM. Does contralateral internal carotid artery occlusion increase the risk of carotid endarterectomy? J Vasc Surg 1992;16:347-52; discussion: 352-3. 9. Redekop G, Ferguson G. Correlation of contralateral stenosis and intraoperative electroencephalogram change with risk of stroke during carotid endarterectomy. Neurosurgery 1992;30:191-4. 10. McCarthy WJ, Wang R, Pearce WH, Flinn WR, Yao JS. Carotid endarterectomy with an occluded contralateral carotid artery. Am J Surg 1993;166:168-71; discussion: 171-2. 11. Meyer FB, Fode NC, Marsh WR, Piepgras DG. Carotid endarterectomy in patients with contralateral carotid occlusion. Mayo Clin Proc 1993;68:337-42. 12. Deriu GP, Franceschi L, Milite D, Calabro A, Saia A, Grego F, et al. Carotid artery endarterectomy in patients with contralateral carotid artery occlusion: perioperative hazards and late results. Ann Vasc Surg 1994;8:337-42. 13. Adelman MA, Jacobowitz GR, Riles TS, Imparato AM, Lamparello PJ, Baumann FG, et al. Carotid endarterectomy in the presence of a contralateral occlusion: a review of 315 cases over a 27-year experience. Cardiovasc Surg 1995;3: 307-12. 14. Cao P, Giordano G, De Rango P, Ricci S, Zannetti S, Moggi L. Carotid endarterectomy contralateral to an occluded carotid artery: a retrospective case-control study. Eur J Vasc Endovasc Surg 1995;10:16-22. 15. Coyle KA, Smith RB 3rd, Salam AA, Dodson TF, Chaikof EL, Lumsden AB. Carotid endarterectomy in patients with contralateral carotid occlusion: review of a 10-year experience. Cardiovasc Surg 1996;4:71-5. 16. Aungst M, Gahtan V, Berkowitz H, Roberts AB, Kerstein MD. Carotid endarterectomy outcome is not affected in patients with a contralateral carotid artery occlusion. Am J Surg 1998;176:30-3. 17. Julia P, Chemla E, Mercier F, Renaudin JM, Fabiani JN. Influence of the status of the contralateral carotid artery on the outcome of carotid surgery. Ann Vasc Surg 1998;12: 566-71.

18. Samson RH, Showalter DP, Yunis JP. Routine carotid endarterectomy without a shunt, even in the presence of a contralateral occlusion. Cardiovasc Surg 1998;6:475-84. 19. Dimakakos PB, Antoniou A, Papasava M, Mourikis D, Rizos D. Carotid endarterectomy without protective measures in patients with occluded and non occluded contralateral carotid artery. J Cardiovasc Surg (Torino) 1999;40:849-55. 20. AbuRahma AF, Robinson P, Holt SM, Herzog TA, Mowery NT. Perioperative and late stroke rates of carotid endarterectomy contralateral to carotid artery occlusion: results from a randomized trial. Stroke 2000;31:1566-71. 21. Locati P, Socrate AM, Lanza G, Tori A, Costantini S. Carotid endarterectomy in an awake patient with contralateral carotid occlusion: influence of selective shunting. Ann Vasc Surg 2000;14:457-62. 22. Karmeli R, Lubezky N, Halak M, Loberman Z, Weller B, Fajer S. Carotid endarterectomy in awake patients with contralateral carotid artery occlusion. Cardiovasc Surg 2001;9:334-8. 23. Ballotta E, Da Giau G, Baracchini C. Carotid endarterectomy contralateral to carotid artery occlusion: analysis from a randomized study. Langenbecks Arch Surg 2002;387:216-21. 24. Pulli R, Dorigo W, Barbanti E, Azas L, Russo D, Matticari S, et al. Carotid endarterectomy with contralateral carotid artery occlusion: is this a higher risk subgroup? Eur J Vasc Endovasc Surg 2002;24:63-8. 25. Schneider JR, Droste JS, Schindler N, Golan JF, Bernstein LP, Rosenberg RS. Carotid endarterectomy with routine electroencephalography and selective shunting: influence of contralateral internal carotid artery occlusion and utility in prevention of perioperative strokes [see comment]. J Vasc Surg 2002;35:1114-22. 26. Rockman CB, Su W, Lamparello PJ, Adelman MA, Jacobowitz GR, Gagne PJ, et al. A reassessment of carotid endarterectomy in the face of contralateral carotid occlusion: surgical results in symptomatic and asymptomatic patients. J Vasc Surg 2002;36:668-73. 27. Ballotta E, Renon L, Da Giau G, Barbon B, Terranova O, Baracchini C. Octogenarians with contralateral carotid artery occlusion: a cohort at higher risk for carotid endarterectomy? J Vasc Surg 2004;39:1003-8. 28. Cinar B, Goksel OS, Karatepe C, Kut S, Aydogan H, Filizcan U, et al. Is routine intravascular shunting necessary for carotid endarterectomy in patients with contralateral occlusion? A review of 5-year experience of carotid endarterectomy with local anaesthesia. Eur J Vasc Endovasc Surg 2004;28:494-9. 29. Grego F, Antonello M, Lepidi S, Zaramella M, Galzignan E, Menegolo M, et al. Is contralateral carotid artery occlusion a risk factor for carotid endarterectomy? Ann Vasc Surg 2005;19:882-9. 30. Dalainas I, Nano G, Bianchi P, Casana R, Malacrida G, Tealdi DG. Carotid endarterectomy in patients with contralateral carotid artery occlusion. Ann Vasc Surg 2007;21:16-22. 31. Tan TW, Garcia-Toca M, Marcaccio EJ Jr, Carney WI Jr, Machan JT, Slaiby JM. Predictors of shunt during carotid endarterectomy with routine electroencephalography monitoring. J Vasc Surg 2009;49:1374-8. 32. Chiriano J, Abou-Zamzam AM Jr, Nguyen K, Molkara AM, Zhang WW, Bianchi C, et al. Preoperative carotid duplex findings predict carotid stump pressures during endarterectomy in symptomatic but not asymptomatic patients. Ann Vasc Surg 2010;24:1038-44. 33. Kretz B, Abello N, Astruc K, Terriat B, Favier C, Bouchot O, et al. Influence of the contralateral carotid artery on carotid surgery outcome. Ann Vasc Surg 2012;26:766-74. 34. Samson RH, Cline JL, Showalter DP, Lepore MR, Nair DG. Contralateral carotid artery occlusion is not a

10

Schneider et al

Journal of Vascular Surgery ---

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

contraindication to carotid endarterectomy even if shunts are not routinely used. J Vasc Surg 2013;58:935-40. Hokari M, Nakayama N, Kazumata K, Osanai T, Nakamura T, Yasuda H, et al. Surgical outcomes for cervical carotid artery stenosis: treatment strategy for bilateral cervical carotid artery stenosis. J Stroke Cerebrovasc Dis 2015;24:1768-74. Taboada CR, Duran Mariño JL, Garcia Colodro JM, Pena Holguin J, Martinez Gallego EL. Clinical outcomes after carotid endarterectomy in patients with contralateral carotid occlusion. Ann Vasc Surg 2016;32:83-7. Frawley JE, Hicks RG, Gray LJ, Niesche JW. Carotid endarterectomy without a shunt for symptomatic lesions associated with contralateral severe stenosis or occlusion. J Vasc Surg 1996;23:421-7. Gasparis AP, Ricotta L, Cuadra SA, Char DJ, Purtill WA, Van Bemmelen PS, et al. High-risk carotid endarterectomy: fact or fiction. J Vasc Surg 2003;37:40-6. Reed AB, Gaccione P, Belkin M, Donaldson MC, Mannick JA, Whittemore AD, et al. Preoperative risk factors for carotid endarterectomy: defining the patient at high risk. J Vasc Surg 2003;37:1191-9. Maatz W, Köhler J, Botsios S, John V, Walterbusch G. Risk of stroke for carotid endarterectomy patients with contralateral carotid occlusion. Ann Vasc Surg 2008;22:45-51. Faggioli G, Pini R, Mauro R, Freyrie A, Gargiulo M, Stella A. Contralateral carotid occlusion in endovascular and surgical carotid revascularization: a single centre experience with literature review and meta-analysis. Eur J Vasc Endovasc Surg 2013;46:10-20. Gasecki AP, Eliasziw M, Ferguson GG, Hachinski V, Barnett HJ. Long-term prognosis and effect of endarterectomy in patients with symptomatic severe carotid stenosis and contralateral carotid stenosis or occlusion: results from NASCET. North American Symptomatic Carotid Endarterectomy Trial (NASCET) Group. J Neurosurg 1995;83:778-82. Baker WH, Howard VJ, Howard G, Toole JF. Effect of contralateral occlusion on long-term efficacy of endarterectomy in the asymptomatic carotid atherosclerosis study (ACAS). ACAS Investigators. Stroke 2000;31:2330-4. Bennett KM, Scarborough JE, Cox MW, Shortell CK. The impact of intraoperative shunting on early neurologic outcomes after carotid endarterectomy. J Vasc Surg 2015;61: 96-102. Wimmer NJ, Spertus JA, Kennedy KF, Anderson HV, Curtis JP, Weintraub WS, et al. Clinical prediction model suitable for assessing hospital quality for patients undergoing carotid endarterectomy. J Am Heart Assoc 2014;3: e000728. Goldstein LB, McCrory DC, Landsman PB, Samsa GP, Ancukiewicz M, Oddone EZ, et al. Multicenter review of preoperative risk factors for carotid endarterectomy in patients with ipsilateral symptoms. Stroke 1994;25:1116-21.

2019

47. Ricotta JJ 2nd, Upchurch GR Jr, Landis GS, Kenwood CT, Siami FS, Tsilimparis N, et al. The influence of contralateral occlusion on results of carotid interventions from the Society for Vascular Surgery Vascular Registry. J Vasc Surg 2014;60: 958-64; discussion: 964-5. 48. Tu JV, Wang H, Bowyer B, Green L, Fang J, Kucey D, et al. Risk factors for death or stroke after carotid endarterectomy: observations from the Ontario Carotid Endarterectomy Registry. Stroke 2003;34:2568-73. 49. Kragsterman B, Logason K, Ahari A, Troëng T, Parsson H, Bergqvist D. Risk factors for complications after carotid endarterectomyda population-based study. Eur J Vasc Endovasc Surg 2004;28:98-103. 50. da Silva AF, McCollum P, Szymanska T, de Cossart L. Prospective study of carotid endarterectomy and contralateral carotid occlusion. Br J Surg 1996;83:1370-2. 51. Goodney PP, Likosky DS, Cronenwett JL; Vascular Study Group of Northern New England. Factors associated with stroke or death after carotid endarterectomy in Northern New England. J Vasc Surg 2008;48:1139-45. 52. Rockman C. Carotid endarterectomy in patients with contralateral carotid occlusion. Semin Vasc Surg 2004;17: 224-9. 53. Antoniou GA, Kuhan G, Sfyroeras GS, Georgiadis GS, Antoniou SA, Murray D, et al. Contralateral occlusion of the internal carotid artery increases the risk of patients undergoing carotid endarterectomy. J Vasc Surg 2013;57: 1134-45. 54. Dickersin K. The existence of publication bias and risk factors for its occurrence. JAMA 1990;263:1385-9. 55. Whitney DG, Kahn EM, Estes JW, Jones CE. Carotid artery surgery without a temporary indwelling shunt. 1,917 consecutive procedures. Arch Surg 1980;115:1393-9. 56. Baker WH, Littooy FN, Hayes AC, Dorner DB, Stubbs D. Carotid endarterectomy without a shunt: the control series. J Vasc Surg 1984;1:50-6. 57. Goodney PP, Wallaert JB, Scali ST, Stone DH, Patel V, Shaw P, et al. Impact of practice patterns in shunt use during carotid endarterectomy with contralateral carotid occlusion. J Vasc Surg 2012;55:61-71.e1. 58. Nicholls SC, Bergelin R, Strandness DE. Neurologic sequelae of unilateral carotid artery occlusion: immediate and late. J Vasc Surg 1989;10:542-7; discussion: 547-8. 59. AbuRahma AF. Treatment of asymptomatic carotid stenosis with contralateral carotid occlusion. Adv Surg 2004;38: 247-65. 60. Nejim B, Dakour Aridi H, Locham S, Arhuidese I, Hicks C, Malas MB. Carotid artery revascularization in patients with contralateral carotid artery occlusion: stent or endarterectomy? J Vasc Surg 2017;66:1735-48.e1.

Submitted Sep 16, 2018; accepted May 4, 2019.