- Email: [email protected]
Regional variation in use and outcomes of combined carotid endarterectomy and coronary artery bypass
Regional variation in use and outcomes of combined carotid endarterectomy and coronary artery bypass Linda J. Wang, MD, MBA,a Emel A. Ergul, MA,a Jahan Mohebali, MD, MPH,a Philip P. Goodney, MD, MS,b Virendra I. Patel, MD, MPH,c Mark F. Conrad, MD, MMSc,a Matthew J. Eagleton, MD,a and W. Darrin Clouse, MD,a Boston, Mass; Lebanon, NH; and New York, NY
ABSTRACT Objective: In treating concomitant carotid and coronary disease, some recommend staged carotid endarterectomy (CEA) and coronary artery bypass grafting, whereas others favor the combined approach (CCAB). Pressure to reduce surgical variation and to improve quality is real, yet little is known about how geographic practice differences affect outcomes. Using the Vascular Quality Initiative (VQI), this study evaluated regional variation in use and outcomes of CCAB. Methods: All CCAB procedures in the VQI from 2003 to 2017 were reviewed and stratified into four regions, as defined by the United States Census Bureau. Primary outcomes included perioperative stroke, death, myocardial infarction (MI), and these as composite (SDM). A c2 analysis was performed. Results: There were 1495 CCAB procedures identified, representing 1.8% of the VQI CEAs. Regions included the following: Midwest (MW), 32%; Northeast (NE), 39%; South (S), 25%; and West (W), 4%. Most were male (70%) and white (92%). There was significant regional variation in proportional volume of CCABs to all CEAs (0.7% [W] to 2.5% [MW]; P < .001). Regional variation in patch use (78% [W] to 93% [MW]; P < .001), shunting (29% [W] to 71% [MW]; P < .001), and electroencephalography monitoring (13% [W] to 52% [NE]; P < .001) was also significant. Overall perioperative stroke was 3.6%; death, 3.0%; and SDM, 6.8%. No regional difference was seen in outcomes of mortality (1.5% [MW] to 4.2% [NE]; P ¼ .05), stroke (2.8% [NE] to 4.4% [MW]; P ¼ .52), and MI (0.6% [MW] to 1.8% [W]; P ¼ .62). When the Bonferroni correction was used, there remained no difference in stroke, MI, or SDM across regions, but mortality became significant. Using the Society for Vascular Surgery guidelines for consideration of CCAB, the minority of patients fell within the symptomatic carotid stenosis (SYMP, 15%; n ¼ 218) or severe ($70%) asymptomatic bilateral carotid disease (BIL, 18%; n ¼ 267) categories. The most common indication was asymptomatic unilateral severe carotid stenosis (UNI, 37%; n ¼ 552). There were no differences in regional outcomes stratified by indication (SYMP, BIL, UNI). Overall, when SYMP and BIL were compared with UNI, UNI had lower rates of stroke (2.4% vs 4.9%; P ¼ .03) but similar MI (0.7% vs 1.2%; P ¼ .40) and mortality (2.2% vs 2.5%; P ¼ .75). Conclusions: Significant variation exists across VQI centers in the use of CCAB. Despite differences in volume and practices, regional perioperative outcomes are similar. UNI is the most commonly used indication and has lower stroke rates relative to SYMP and BIL. CCAB is performed well across the United States, but most patients fall outside of Society for Vascular Surgery guidelines. (J Vasc Surg 2019;-:1-7.) Keywords: Carotid stenosis/surgery; Endarterectomy; carotid; Coronary disease/surgery; Geography/statistics and numerical data; Stroke/etiology
In an effort to reduce variation in practice and to improve patient care, the Society for Vascular Surgery (SVS) has released consensus statements aimed at guiding physicians by clearly defining best practices where strong evidence exists. However, in both the updated SVS and European Society for Vascular Surgery
From the Division of Vascular and Endovascular Surgery, Massachusetts General Hospital, Bostona; the Section of Vascular Surgery, Dartmouth-Hitchcock Medical Center, Lebanonb; and the Division of Vascular Surgery and Endovascular Interventions, Columbia University Medical Center, New York.c Author conflict of interest: none. Correspondence: W. Darrin Clouse, MD, University of Virginia, 1215 Lee St, PO Box 800679, Charlottesville, VA 22908 (e-mail: [email protected]). The editors and reviewers of this article have no relevant financial relationships to disclose per the JVS policy that requires reviewers to decline review of any 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.2019.02.003
guidelines, only level B evidence is cited for the management of combined carotid and coronary disease, and equivalence is given to both staged and combined approaches.1,2 The use of combined carotid endarterectomy (CEA) and coronary bypass thus remains controversial. Some centers recommend staged CEA and coronary artery bypass grafting (CABG), whereas others favor the combined approach (CCAB). In patients with atherosclerosis in both distributions, proponents advocate for simultaneous revascularization to minimize perioperative cardiovascular events3,4; arguments for lower costs, shorter lengths of stay, and lower long-term stroke rates have also been made.5 Not all are convinced, however, citing studies that have demonstrated higher rates of stroke with CCAB compared with each individual operation alone.6,7 The SVS created the Vascular Quality Initiative (VQI) with the goal of improving patient care; however, little is known with respect to how well benchmarks are being 1
2
Journal of Vascular Surgery
Wang et al
---
met across the country. Of the few VQI studies addressing regional variation and outcomes, all comment on well-established treatment paradigms where strong consensus and evidence exists.8-13 However, it is equally important, if not more so, to determine and disseminate information on procedures where clinical uncertainty lies. To that end, using the VQI, this study evaluated regional variation in use and outcomes of combined CEA with CABG.
ARTICLE HIGHLIGHTS d
d
METHODS Patients and data collection. The VQI has been well described and reviewed.14 Deidentified demographic, clinical, and procedural variables are collected on each patient at each participating institution and centralized. Participation in this quality registry as a Patient Safety Organization and quality assurance tool for analysis was approved, and direct informed consent waived, at each center’s Institutional Review Board. More information about the VQI can be found at www.vqi.org. Cohorts, definitions, and end points. All combined CEA and CABG procedures in the VQI database were identified from participating centers in the United States from January 2003 to December 2017. The VQI database affords evaluation only of procedures performed concurrently with the indexed CEA, and thus staged procedures cannot be evaluated in this data set. Patients who underwent additional concurrent procedures were excluded to provide an isolated CCAB cohort. Patients who did not have a regional assignment were also excluded (n ¼ 3). Using the standard United States Census Bureau classification system (Fig 1),15 patients were stratified into four regions: Midwest (MW), Northeast (NE), South (S), and West (W). Demographics, medical comorbidities, and operative details were reviewed. Sixty-eight variables were collected prospectively for each procedure and recorded in the VQI database. The definitions of medical comorbidities and procedure details within the VQI, as modeled after the Vascular Study Group of New England, have previously been described.16 Symptomatic carotid disease was defined as a history of prior ipsilateral neurologic event; a patient was considered entirely asymptomatic if no prior transient ischemic attack (TIA) or stroke was documented. Although documentation of timing of prior neurologic event began in 2012, all symptomatic patients from 2003 onward were included in analyses. The decision to not exclude 2003 to 2011 patients from the indications subanalysis was made as every patient in the cohort with a documented prior stroke/TIA timestamp had symptoms within 2 months of the indexed procedure. The degree of ipsilateral stenosis was defined categorically as moderate (50%-69%) and severe ($70%) stenosis. Postoperative stroke was defined as any new neurologic deficit, regardless of
2019
d
Type of Research: Retrospective review of prospectively collected Vascular Quality Initiative data Key Findings: Patients undergoing combined carotid endarterectomy with coronary artery bypass (CCAB) were compared across the four United States Census regional divisions. In 1495 patients, there was significant variation in practice patterns, but outcomes were similar across regions. The most common indication was asymptomatic unilateral severe carotid disease (37%). Take Home Message: Despite differences in volume and practices in the use of CCAB, regional perioperative outcomes are similar. Most indications for CCAB fall outside of Society for Vascular Surgery practice guidelines.
laterality or symptomatic duration. Postoperative myocardial infarction (MI) was diagnosed by either troponin levels or electrocardiographic or clinical evaluation. With respect to troponin levels, at least two levels must have been obtained, with associated rise or fall, as well as at least one value above the 99th percentile upper reference limit. Electrocardiographic or clinically evaluated MI was reported when there was evidence of myocardial necrosis in a clinical setting consistent with acute myocardial ischemia. All perioperative outcomes reported occurred either within 30 days of the indexed procedure (stroke, death) or before discharge (MI). Primary end points were perioperative stroke, death, MI, and these as composite (SDM). Statistical analysis. Statistical analysis was performed with Stata/SE 15 software (StataCorp LLC, College Station, Tex). Dichotomous variables are described as a percentage of the cohort. Continuous variables are expressed as mean 6 standard deviation. Univariate analysis was performed to compare clinical and demographic features and postoperative complications between regions. The Bonferroni correction was used as a multiple comparison correction method in analyzing outcomes. For each comparison, the c2 test was used. A P value of <.05 was considered significant. Forest plots were used to illustrate variables of interest.
RESULTS Demographics. There were 1495 CCABs performed during the study period that met inclusion criteria. These procedures represented 1.8% of all CEAs performed in the VQI, and the proportion of CCABs to CEAs performed from 2003 (1.4%) through 2017 (1.5%) remained relatively constant. Regional breakdown included the following: 482 (32%) MW, 578 (39%) NE, 379 (25%) S, and 56 (4%) W. Proportional volume of CCABs to total CEAs varied
Journal of Vascular Surgery Volume
-,
Number
Wang et al
3
-
Fig 1. Census regions of the United States, broken down by state.
between regions (0.7% [W] to 2.5% [MW]; P < .001). Across all VQI centers, 155 (47%) performed CCAB during the study period: MW, 42; NE, 38; S, 56; and W, 19 (Fig 1). Of these, 42 centers performed 10 or more CCABs and accounted for 1183 (79%) of all patients in this study; 63 centers performed 5 or more CCABs. Overall, 70% of the study population were male, 92% were white, 91% had hypertension, 77% were either former or current smokers, and 25% had chronic obstructive pulmonary disease. Average age was 69 6 9 years (range, 40-90 years). Fig 2 shows the baseline characteristics of patients who underwent CCAB. Patients’ demographics were similar across regions with the exception of proportion of white patients (87% [S] to 95% [MW]; P < .001) and patients with congestive heart failure (17% [MW] to 27% [W]; P ¼ .02). There was significant regional variation in the type of carotid imaging obtained as well as in preoperative medications (Fig 3). Notably, preoperative computed tomography angiography use ranged from 27% (W) to 57% (MW; P < .001). Overall duplex ultrasound use was 92% with a significant difference seen between regions (88% [S] to 94% [NE]; P ¼ .004). There was also variation in the use of preoperative aspirin (77% [MW] to 89% [NE]; P < .001), P2Y12 inhibitors (13% [NE] to 22% [MW; P ¼ .001), and beta blockade (73% [W] to 85% [NE]; P ¼ .001).
Operative characteristics. There was significant regional variation in the operative techniques used and neuromonitoring practices employed (Fig 4). The use of the eversion technique (7% [W] to 20% [NE]; P < .001) and shunting (29% [NE] to 71% [W]; P < .001) varied significantly. Electroencephalography monitoring (13% [W] to 52% [NE]; P < .001) and measurement of stump pressures (5% [NE] to 13% [MW]; P < .001) also differed between regions. There was also significant regional variation in the percentage of urgent/emergent procedures (23% [MW] to 37% [NE]; P < .001). Perioperative outcomes. Overall perioperative stroke was 3.6% (n ¼ 54) and when stratified by laterality, the ipsilateral perioperative stroke rate was 2.0% (n ¼ 30). No difference was seen between regions with respect to perioperative stroke (2.8% [NE] to 4.4% [MW]; P ¼ .52). Overall 30-day mortality was 3.0% (n ¼ 45), with no difference seen between regions (1.5% [MW] to 4.2% [NE]; P ¼ .05). Perioperative MI did not vary (0.6% [MW] to 1.8% [W]; P ¼ .62), nor did total perioperative stroke/death/MI rates (overall, 6.8%; 5.4% [W] to 7.3% [NE]; P ¼ .92). When the Bonferroni correction was used, there remained no difference in stroke, MI, or SDM across regions, but mortality became significant (P ¼ .01). Additional regional variation was found with respect to
4
Journal of Vascular Surgery
Wang et al
---
Fig 2. Baseline characteristics of patients undergoing carotid endarterectomy with coronary artery bypass (CCAB). Each symbol on the line represents a region. CHF, Congestive heart failure; COPD, chronic obstructive pulmonary disease; HTN, hypertension. *Significant variation (to P < .001).
2019
Fig 4. Operative characteristics of patients undergoing carotid endarterectomy with coronary artery bypass (CCAB). EEG, Electroencephalography. *Significant variation (to P < .001).
Fig 5. Perioperative complications of patients undergoing carotid endarterectomy with coronary artery bypass (CCAB). Before Bonferroni correction, no statistical difference was seen in regional outcomes. After Bonferroni correction, only perioperative death was statistically different between regions (P ¼ .01). MI, Myocardial infarction. Fig 3. Preoperative carotid artery imaging and medications of patients undergoing carotid endarterectomy with coronary artery bypass (CCAB). CTA, Computed tomography angiography; MRA, magnetic resonance angiography. *Significant variation (to P < .001).
perioperative rhythm disturbances requiring treatment (11% [W] to 22% [MW]; P ¼ .04). Fig 5 illustrates variation in perioperative outcomes between regions. Indications. Using the SVS guidelines for patients in whom staged procedures or CCAB should be considered,1 patients who were symptomatic (SYM) or who had severe asymptomatic bilateral carotid artery stenosis (BIL) were identified; 218 patients (15%) represented SYM and 267 patients (18%) fell within the BIL category. When SYM and BIL were combined (SYM-BIL) to comprise patients who fell within SVS guidelines, there was significant regional variation in guideline adherence (27% [NE] to 41% [W]; P ¼ .001). The most common indication for CCAB fell outside of SVS guidelines and represented patients with asymptomatic unilateral severe disease (UNI, 37%; n ¼ 552). When patients were stratified by indication (SYM, BIL, UNI), there remained no differences in outcomes (stroke, death, MI) between regions. SYM
and BIL were then compared with UNI to ascertain differences in perioperative outcomes. UNI had lower overall perioperative stroke rates (2.4% vs 4.9%; P ¼ .03) but similar 30-day mortality (2.2% vs 2.5%; P ¼ .75) and perioperative MI (0.7% vs 1.2%; P ¼ .40). The SYM-BIL cohort had higher rates of the combined stroke/death/ MI end point (7.8% vs 4.5%; P ¼ .04; Table). The remaining patients who did not fall into SYM, BIL, or UNI were missing contralateral carotid stenosis data and thus were unable to be analyzed within this subanalysis.
DISCUSSION As discussions surrounding health care have shifted toward efficient resource use and outcome optimization, variation in surgical practice has become a focus of debate.12 Studies have shown that the likelihood of undergoing an operation depends as much on one’s geographic location as on one’s clinical condition.17,18 Despite increasing evidence of the impact of regional variation, geographic differences in rates of surgical procedures do not appear to be decreasing.10 Determinants of surgical regional variation are likely to be multifactorial. Birkmeyer et al have argued that illness burden and diagnostic practices
Journal of Vascular Surgery Volume
-,
Number
Wang et al
5
-
Table. Perioperative outcomes of patients undergoing carotid endarterectomy with coronary artery bypass (CCAB) stratified by indication SYM-BIL (n ¼ 485 [32%]), No. (% total cohort)
UNI (n ¼ 552 [37%]), No. (% total cohort)
P value
All stroke
24 (4.9)
13 (2.4)
.03
Death
12 (2.5)
12 (2.2)
.75
Stroke/death
33 (6.8)
23 (4.2)
.06
6 (1.2)
4 (0.7)
.40
38 (7.8)
25 (4.5)
.03
Perioperative outcomes
MI SDM
MI, Myocardial infarction; SDM, stroke/death/myocardial infarction; SYM-BIL, symptomatic patients (SYM) and patients with asymptomatic bilateral carotid artery stenosis (BIL); UNI, patients with asymptomatic unilateral severe carotid disease.
explain only a small degree of the variation, and the majority can be better correlated to the physician’s beliefs and the patient’s preferences.17 Differences in how surgeons apply evidence likely relates to clinical uncertainty and gaps in current scientific knowledge.18 Indeed, debate on best practice for management of combined coronary and carotid disease is ongoing. Critics of CCAB, who advocate for a staged approach for the treatment of concomitant carotid and coronary disease, cite higher reported rates of perioperative stroke and death with the combined procedure.7,19 Using the Society of Thoracic Surgeons Database, Prasad et al6 found CCAB to carry higher rates of operative mortality and neurologic complications compared with isolated CABG in patients with a history of carotid disease. Yet, many have argued for the safety of combined CCAB.3,4 In a recent metaanalysis, Sharma et al20 analyzed 25,021 patients across 12 studies undergoing either staged CEA and CABG or CCAB and found no difference in perioperative stroke or death between operative strategies. Using the National Inpatient Sample, Gopaldas et al21 similarly found no differences in mortality or neurologic complications in 6153 patients undergoing staged vs combined (CCAB) procedures but did demonstrate a higher risk of overall complications in the staged cohort. Despite these data, there remains a lack of large randomized trials comparing treatment paradigms and differing institutional strategies. To that end, across societal guidelines, equivalence is given to staged and combined strategies.1,2,22 This has become more complex with the rise of percutaneous coronary interventions and carotid stenting alternatives.23 We found that across the United States, despite significant variation in selection of patients, volume, and preoperative and intraoperative management, CCAB outcomes remain similar across regions. We found an overall perioperative SDM rate of 6.8% (range, 5.4%-7.3%; P ¼ .92), which is lower than previously reported rates of 9% to 12% in older retrospective reviews.24 This is likely to be a testament to advances in medical and surgical management but may also highlight the trend of care consolidation as more advanced procedures are being performed exclusively at select centers of excellence. Whereas proportional CCAB to CEA regional
volume did not correlate with outcomes, it is important to highlight that only 47% (n ¼ 155) of all VQI centers performed this procedure during the study period; 42 centers accounted for 79% of the patients in this analysis. National trends in coronary revascularization as well as in CCAB institutional experience can provide some insight into these findings. In a study of Medicare beneficiaries, Culler et al25 found a significant decrease in the number of CABGs performed from 2008 through 2012, and this was accompanied by a 20% increase in the number of facilities performing percutaneous coronary intervention compared with a 2% increase in the number of sites performing CABG. These dates are of particular interest as 2012 saw nearly a doubling of the number of VQI sites, and correspondingly, 110 (71%) of the sites in our study joined in 2012 or later. We would also argue that the fact that less than half of all VQI sites perform the combined procedure is a testament to (1) the importance of institutional experience as not all facilities have the resources to perform this procedure, (2) the impact of clinical uncertainty on practice patterns, and (3) the role of the physician’s beliefs in selection of treatment strategy. For the management of combined carotid and coronary disease, the SVS guidelines highlight two groups of patients who should be considered for CEA before or concomitantly with CABG: patients with symptomatic carotid stenosis and patients with severe bilateral asymptomatic disease.1 The more recently published European Society for Vascular Surgery guidelines add a temporal component and stenosis criteria for the first group, recommending staged or synchronous revascularization in patients with a history of stroke or TIA in the preceding 6 months and a $50% stenosis.2 In this analysis, 15% of CCABs carried a symptomatic diagnosis, whereas 18% fell into the second category. In combining these groups and comparing across regions, there was significant variability in the proportion of patients who met either criterion; 27% of CCABs in the Northeast had an indication of SYM or BIL compared with 41% of CCABs performed in the West. Indeed, the large range of reported percentages of symptomatic patients undergoing CCAB (4%-44%) may in part be explained by differences in regional practices.23,26,27
6
Journal of Vascular Surgery
Wang et al
---
Regional variation in the treatment of carotid artery disease is known to exist. Using the VQI, Shean et al9 analyzed 57,555 patients undergoing carotid artery revascularization and demonstrated widespread variation in selection of patients and preoperative management. Using Medicare data, Goodney et al28 demonstrated a sevenfold difference in regional variation in the use of CEA. These differences and the factors leading to them may also influence CCAB practices. As most CCAB patients have asymptomatic carotid stenosis, concerns about this practice have been voiced.29 The effect of symptomatic status on combined procedural outcomes has been debated. Some have found unacceptably high rates of perioperative stroke (9%-12%) in CCAB populations with high proportions of symptomatic patients.27,30 Others found no difference in neurologic CCAB outcomes when stratifying by symptomatic status.26 A randomized controlled trial performed in 185 UNI patients demonstrated a perioperative stroke benefit to patients who underwent CEA before or concomitantly with CABG compared with those who underwent CEA after CABG (1% vs 7.7%; P ¼ .01).31 Our analysis demonstrated that UNI had lower rates of stroke (2.4% vs 4.9%; P ¼ .03) and SDM (4.5% vs 7.8%; P ¼ .03) compared with SYM-BIL. These data suggest that carotid symptomatic status does affect CCAB outcomes and should factor into preoperative risk profile assessment. Despite differences in selection of patients and guideline adherence, no differences were seen in outcomes across regions. After Bonferroni correction, a difference in mortality was seen across regions (P ¼ .01) but did not correlate with absolute or proportional volume. Furthermore, when subgroup analysis was performed within each indication group (SYM, BIL, UNI), despite regional variance in selection of patients, no differences in outcomes between regions was noted. Whereas the data presented in this analysis suggest that CCAB is performed well across the United States, the large proportion of CCABs performed for asymptomatic unilateral disease suggests there is a continued role for quality projects aimed at either improved guideline adherence or data to better clarify contemporary guideline appropriateness. However, whereas the goal should be approaching standardization of care and outcomes, areas of clinical equipoise will continue to existdand with that, regional variation in surgical experiences and practice as well as individual patient management. This study has several limitations. One limitation lies in its retrospective review of a prospectively maintained database, prone to entry errors and missing data; the VQI performs regular audits to minimize these issues. Statistical analysis was designed to assess whether variation across all regions existed and did not identify the regions that led to the differences seen. In addition, true regional volume may not be captured in regions with
2019
fewer participating VQI centers. The VQI represents a minority of CEAs performed in the United States during this time, and thus many CCABs and their effect may be lost. As an extension of this, the VQI includes only a subset of hospital centers and does not include the Veterans Affairs population; thus, these data may not be fully representative of national practice patterns and outcomes. Furthermore, our data do not account for geographic variation in disease burden. Last, the true reasons for variation in practice patterns can only be speculated on; further study of system, provider, and patient preferences would be needed to address this question, and further investigation is warranted.
CONCLUSIONS Significant regional variation exists across VQI centers in use of combined CEA and CABG. Despite differences in volume and practices, perioperative outcomes were similar. Asymptomatic unilateral severe carotid stenosis represents the most common carotid indication, and significantly lower 30-day stroke rates were seen in UNI compared with the SYMP-BIL cohorts. These data suggest that CCAB is performed well across the United States, but most patients fall outside of societal guidelines. As clinical uncertainty about best practices in management of combined carotid and coronary disease continues to exist, so too will regional variation in surgical application. However, these differences do not necessarily correlate with worse CCAB outcomes.
AUTHOR CONTRIBUTIONS Conception and design: LW, WC Analysis and interpretation: LW, EE, JM, PG, VP, MC, ME, WC Data collection: LW, EE, WC Writing the article: LW, WC Critical revision of the article: LW, EE, JM, PG, VP, MC, ME, WC Final approval of the article: LW, EE, JM, PG, VP, MC, ME, WC Statistical analysis: LW, EE, WC Obtained funding: Not applicable Overall responsibility: WC
REFERENCES 1. 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. J Vasc Surg 2011;54:e1-31. 2. Naylor AR, Ricco JB, de Borst GJ, Debus S, de Haro J, Halliday A, et al. Editor’s choicedmanagement of atherosclerotic carotid and vertebral artery disease: 2017 Clinical Practice Guidelines of the European Society for Vascular Surgery (ESVS). Eur J Vasc Endovasc Surg 2018;55:3-81. 3. Kougias P, Kappa JR, Sewell DH, Feit RA, Michalik RE, Imam M, et al. Simultaneous carotid endarterectomy and coronary artery bypass grafting: results in specific patient groups. Ann Vasc Surg 2007;21:408-14. 4. Akins CW, Hilgenberg AD, Vlahakes GJ, Madsen JC, MacGillivray TE, LaMuraglia GM, et al. Late results of combined carotid and coronary surgery using actual versus actuarial methodology. Ann Thorac Surg 2005;80:2091-7.
Journal of Vascular Surgery Volume
-,
Number
Wang et al
7
-
5. Darling RC, Dylewski M, Chang BB, Paty PS, Kreienberg PB, Lloyd WE, et al. Combined carotid endarterectomy and coronary artery bypass grafting does not increase the risk of perioperative stroke. Cardiovasc Surg 1998;6:448-52. 6. Prasad SM, Li S, Rankin JS, O’Brien SM, Gammie JS, Puskas JD, et al. Current outcomes of simultaneous carotid endarterectomy and coronary artery bypass graft surgery in North America. World J Surg 2010;34:2292-8. 7. Borger MA, Fremes SE, Weisel RD, Cohen G, Rao V, Lindsay TF, et al. Coronary bypass and carotid endarterectomy: does a combined approach increase risk? A metaanalysis. Ann Thorac Surg 1999;68:14-20; discussion: 21. 8. Shean KE, O’Donnell TF, Deery SE, Pothof AB, Schneider JR, Rockman CB, et al. Regional variation in patient outcomes in carotid artery disease treatment in the Vascular Quality Initiative. J Vasc Surg 2018;68:749-59. 9. Shean KE, McCallum JC, Soden PA, Deery SE, Schneider JR, Nolan BW, et al. Regional variation in patient selection and treatment for carotid artery disease in the Vascular Quality Initiative. J Vasc Surg 2017;66:112-21. 10. Zettervall SL, Buck DB, Soden PA, Cronenwett JL, Goodney PP, Eslami MH, et al. Regional variation exists in patient selection and treatment of abdominal aortic aneurysms. J Vasc Surg 2016;64:921-7. 11. Zettervall SL, Soden PA, Buck DB, Cronenwett JL, Goodney PP, Eslami MH, et al. Significant regional variation exists in morbidity and mortality after repair of abdominal aortic aneurysm. J Vasc Surg 2017;65:1305-12. 12. Soden PA, Zettervall SL, Shean KE, Vouyouka AG, Goodney PP, Mills JL, et al. Regional variation in outcomes for lower extremity vascular disease in the Vascular Quality Initiative. J Vasc Surg 2017;66:810-8. 13. Soden PA, Zettervall SL, Curran T, Vouyouka AG, Goodney PP, Mills JL, et al. Regional variation in patient selection and treatment for lower extremity vascular disease in the Vascular Quality Initiative. J Vasc Surg 2017;65:108-18. 14. Cronenwett JL, Kraiss LW, Cambria RP. The Society for Vascular Surgery Vascular Quality Initiative. J Vasc Surg 2012;55:1529-37. 15. United States Census Bureau. Available at: https://www. census.gov/geo/reference/webatlas/regions.html. Accessed April 3, 2018. 16. Cronenwett JL, Likosky DS, Russell MT, Eldrup-Jorgensen J, Stanley AC, Nolan BW. A regional registry for quality assurance and improvement: the Vascular Study Group of Northern New England (VSGNNE). J Vasc Surg 2007;46:1093-102.e1. 17. Birkmeyer JD, Reames BN, McCulloch P, Carr AJ, Campbell WB, Wennberg JE. Understanding of regional variation in the use of surgery. Lancet 2013;382:1121-9. 18. Reames BN, Shubeck SP, Birkmeyer JD. Strategies for reducing regional variation in the use of surgery. Ann Surg 2014;259:616-27.
19. Dubinsky RM, Lai SM. Mortality from combined carotid endarterectomy and coronary artery bypass surgery in the US. Neurology 2007;68:195-7. 20. Sharma V, Deo SV, Park SJ, Joyce LD. Meta-analysis of staged versus combined carotid endarterectomy and coronary artery bypass grafting. Ann Thorac Surg 2014;97:102-9. 21. Gopaldas RR, Chu D, Dao TK, Huh J, LeMaire SA, Lin P, et al. Staged versus synchronous carotid endarterectomy and coronary artery bypass grafting: analysis of 10-year nationwide outcomes. Ann Thorac Surg 2011;91:1323-9. 22. Windecker S, Kolh P, Alfonso F, Collet JP, Cremer J, Falk V, et al. 2014 ESC/EACTS guidelines on myocardial revascularization. Rev Esp Cardiol (Engl Ed) 2015;68:144. 23. Timaran CH, Rosero EB, Smith ST, Valentine RJ, Modrall JG, Clagett GP. Trends and outcomes of concurrent carotid revascularization and coronary bypass. J Vasc Surg 2008;48: 355-61.e1. 24. Naylor AR, Cuffe RL, Rothwell PM, Bell PR. A systematic review of outcomes following staged and synchronous carotid endarterectomy and coronary artery bypass. Eur J Vasc Endovasc Surg 2003;25:380-9. 25. Culler SD, Kugelmass AD, Brown PP, Reynolds MR, Simon AW. Trends in coronary revascularization procedures among Medicare beneficiaries between 2008 and 2012. Circulation 2015;131:362-70; discussion: 370. 26. Jones DW, Stone DH, Conrad MF, Baribeau YR, Westbrook BM, Likosky DS, et al. Regional use of combined carotid endarterectomy/coronary artery bypass graft and the effect of patient risk. J Vasc Surg 2012;56:668-76. 27. Brown KR, Kresowik TF, Chin MH, Kresowik RA, Grund SL, Hendel ME. Multistate population-based outcomes of combined carotid endarterectomy and coronary artery bypass. J Vasc Surg 2003;37:32-9. 28. Goodney PP, Travis LL, Malenka D, Bronner KK, Lucas FL, Cronenwett JL, et al. Regional variation in carotid artery stenting and endarterectomy in the Medicare population. Circ Cardiovasc Qual Outcomes 2010;3:15-24. 29. Venkatachalam S, Gray BH, Mukherjee D, Shishehbor MH. Contemporary management of concomitant carotid and coronary artery disease. Heart 2010;97:175-80. 30. Mackey WC, Khabbaz K, Bojar R, O’Donnell TF. Simultaneous carotid endarterectomy and coronary bypass: perioperative risk and long-term survival. J Vasc Surg 1996;24:58-64. 31. Illuminati G, Ricco JB, Caliò F, Pacilè MA, Miraldi F, Frati G, et al. Short-term results of a randomized trial examining timing of carotid endarterectomy in patients with severe asymptomatic unilateral carotid stenosis undergoing coronary artery bypass grafting. J Vasc Surg 2011;54:993-9; discussion: 998-9.
Submitted Nov 7, 2018; accepted Feb 1, 2019.
ABSTRACT Objective: In treating concomitant carotid and coronary disease, some recommend staged carotid endarterectomy (CEA) and coronary artery bypass grafting, whereas others favor the combined approach (CCAB). Pressure to reduce surgical variation and to improve quality is real, yet little is known about how geographic practice differences affect outcomes. Using the Vascular Quality Initiative (VQI), this study evaluated regional variation in use and outcomes of CCAB. Methods: All CCAB procedures in the VQI from 2003 to 2017 were reviewed and stratified into four regions, as defined by the United States Census Bureau. Primary outcomes included perioperative stroke, death, myocardial infarction (MI), and these as composite (SDM). A c2 analysis was performed. Results: There were 1495 CCAB procedures identified, representing 1.8% of the VQI CEAs. Regions included the following: Midwest (MW), 32%; Northeast (NE), 39%; South (S), 25%; and West (W), 4%. Most were male (70%) and white (92%). There was significant regional variation in proportional volume of CCABs to all CEAs (0.7% [W] to 2.5% [MW]; P < .001). Regional variation in patch use (78% [W] to 93% [MW]; P < .001), shunting (29% [W] to 71% [MW]; P < .001), and electroencephalography monitoring (13% [W] to 52% [NE]; P < .001) was also significant. Overall perioperative stroke was 3.6%; death, 3.0%; and SDM, 6.8%. No regional difference was seen in outcomes of mortality (1.5% [MW] to 4.2% [NE]; P ¼ .05), stroke (2.8% [NE] to 4.4% [MW]; P ¼ .52), and MI (0.6% [MW] to 1.8% [W]; P ¼ .62). When the Bonferroni correction was used, there remained no difference in stroke, MI, or SDM across regions, but mortality became significant. Using the Society for Vascular Surgery guidelines for consideration of CCAB, the minority of patients fell within the symptomatic carotid stenosis (SYMP, 15%; n ¼ 218) or severe ($70%) asymptomatic bilateral carotid disease (BIL, 18%; n ¼ 267) categories. The most common indication was asymptomatic unilateral severe carotid stenosis (UNI, 37%; n ¼ 552). There were no differences in regional outcomes stratified by indication (SYMP, BIL, UNI). Overall, when SYMP and BIL were compared with UNI, UNI had lower rates of stroke (2.4% vs 4.9%; P ¼ .03) but similar MI (0.7% vs 1.2%; P ¼ .40) and mortality (2.2% vs 2.5%; P ¼ .75). Conclusions: Significant variation exists across VQI centers in the use of CCAB. Despite differences in volume and practices, regional perioperative outcomes are similar. UNI is the most commonly used indication and has lower stroke rates relative to SYMP and BIL. CCAB is performed well across the United States, but most patients fall outside of Society for Vascular Surgery guidelines. (J Vasc Surg 2019;-:1-7.) Keywords: Carotid stenosis/surgery; Endarterectomy; carotid; Coronary disease/surgery; Geography/statistics and numerical data; Stroke/etiology
In an effort to reduce variation in practice and to improve patient care, the Society for Vascular Surgery (SVS) has released consensus statements aimed at guiding physicians by clearly defining best practices where strong evidence exists. However, in both the updated SVS and European Society for Vascular Surgery
From the Division of Vascular and Endovascular Surgery, Massachusetts General Hospital, Bostona; the Section of Vascular Surgery, Dartmouth-Hitchcock Medical Center, Lebanonb; and the Division of Vascular Surgery and Endovascular Interventions, Columbia University Medical Center, New York.c Author conflict of interest: none. Correspondence: W. Darrin Clouse, MD, University of Virginia, 1215 Lee St, PO Box 800679, Charlottesville, VA 22908 (e-mail: [email protected]). The editors and reviewers of this article have no relevant financial relationships to disclose per the JVS policy that requires reviewers to decline review of any 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.2019.02.003
guidelines, only level B evidence is cited for the management of combined carotid and coronary disease, and equivalence is given to both staged and combined approaches.1,2 The use of combined carotid endarterectomy (CEA) and coronary bypass thus remains controversial. Some centers recommend staged CEA and coronary artery bypass grafting (CABG), whereas others favor the combined approach (CCAB). In patients with atherosclerosis in both distributions, proponents advocate for simultaneous revascularization to minimize perioperative cardiovascular events3,4; arguments for lower costs, shorter lengths of stay, and lower long-term stroke rates have also been made.5 Not all are convinced, however, citing studies that have demonstrated higher rates of stroke with CCAB compared with each individual operation alone.6,7 The SVS created the Vascular Quality Initiative (VQI) with the goal of improving patient care; however, little is known with respect to how well benchmarks are being 1
2
Journal of Vascular Surgery
Wang et al
---
met across the country. Of the few VQI studies addressing regional variation and outcomes, all comment on well-established treatment paradigms where strong consensus and evidence exists.8-13 However, it is equally important, if not more so, to determine and disseminate information on procedures where clinical uncertainty lies. To that end, using the VQI, this study evaluated regional variation in use and outcomes of combined CEA with CABG.
ARTICLE HIGHLIGHTS d
d
METHODS Patients and data collection. The VQI has been well described and reviewed.14 Deidentified demographic, clinical, and procedural variables are collected on each patient at each participating institution and centralized. Participation in this quality registry as a Patient Safety Organization and quality assurance tool for analysis was approved, and direct informed consent waived, at each center’s Institutional Review Board. More information about the VQI can be found at www.vqi.org. Cohorts, definitions, and end points. All combined CEA and CABG procedures in the VQI database were identified from participating centers in the United States from January 2003 to December 2017. The VQI database affords evaluation only of procedures performed concurrently with the indexed CEA, and thus staged procedures cannot be evaluated in this data set. Patients who underwent additional concurrent procedures were excluded to provide an isolated CCAB cohort. Patients who did not have a regional assignment were also excluded (n ¼ 3). Using the standard United States Census Bureau classification system (Fig 1),15 patients were stratified into four regions: Midwest (MW), Northeast (NE), South (S), and West (W). Demographics, medical comorbidities, and operative details were reviewed. Sixty-eight variables were collected prospectively for each procedure and recorded in the VQI database. The definitions of medical comorbidities and procedure details within the VQI, as modeled after the Vascular Study Group of New England, have previously been described.16 Symptomatic carotid disease was defined as a history of prior ipsilateral neurologic event; a patient was considered entirely asymptomatic if no prior transient ischemic attack (TIA) or stroke was documented. Although documentation of timing of prior neurologic event began in 2012, all symptomatic patients from 2003 onward were included in analyses. The decision to not exclude 2003 to 2011 patients from the indications subanalysis was made as every patient in the cohort with a documented prior stroke/TIA timestamp had symptoms within 2 months of the indexed procedure. The degree of ipsilateral stenosis was defined categorically as moderate (50%-69%) and severe ($70%) stenosis. Postoperative stroke was defined as any new neurologic deficit, regardless of
2019
d
Type of Research: Retrospective review of prospectively collected Vascular Quality Initiative data Key Findings: Patients undergoing combined carotid endarterectomy with coronary artery bypass (CCAB) were compared across the four United States Census regional divisions. In 1495 patients, there was significant variation in practice patterns, but outcomes were similar across regions. The most common indication was asymptomatic unilateral severe carotid disease (37%). Take Home Message: Despite differences in volume and practices in the use of CCAB, regional perioperative outcomes are similar. Most indications for CCAB fall outside of Society for Vascular Surgery practice guidelines.
laterality or symptomatic duration. Postoperative myocardial infarction (MI) was diagnosed by either troponin levels or electrocardiographic or clinical evaluation. With respect to troponin levels, at least two levels must have been obtained, with associated rise or fall, as well as at least one value above the 99th percentile upper reference limit. Electrocardiographic or clinically evaluated MI was reported when there was evidence of myocardial necrosis in a clinical setting consistent with acute myocardial ischemia. All perioperative outcomes reported occurred either within 30 days of the indexed procedure (stroke, death) or before discharge (MI). Primary end points were perioperative stroke, death, MI, and these as composite (SDM). Statistical analysis. Statistical analysis was performed with Stata/SE 15 software (StataCorp LLC, College Station, Tex). Dichotomous variables are described as a percentage of the cohort. Continuous variables are expressed as mean 6 standard deviation. Univariate analysis was performed to compare clinical and demographic features and postoperative complications between regions. The Bonferroni correction was used as a multiple comparison correction method in analyzing outcomes. For each comparison, the c2 test was used. A P value of <.05 was considered significant. Forest plots were used to illustrate variables of interest.
RESULTS Demographics. There were 1495 CCABs performed during the study period that met inclusion criteria. These procedures represented 1.8% of all CEAs performed in the VQI, and the proportion of CCABs to CEAs performed from 2003 (1.4%) through 2017 (1.5%) remained relatively constant. Regional breakdown included the following: 482 (32%) MW, 578 (39%) NE, 379 (25%) S, and 56 (4%) W. Proportional volume of CCABs to total CEAs varied
Journal of Vascular Surgery Volume
-,
Number
Wang et al
3
-
Fig 1. Census regions of the United States, broken down by state.
between regions (0.7% [W] to 2.5% [MW]; P < .001). Across all VQI centers, 155 (47%) performed CCAB during the study period: MW, 42; NE, 38; S, 56; and W, 19 (Fig 1). Of these, 42 centers performed 10 or more CCABs and accounted for 1183 (79%) of all patients in this study; 63 centers performed 5 or more CCABs. Overall, 70% of the study population were male, 92% were white, 91% had hypertension, 77% were either former or current smokers, and 25% had chronic obstructive pulmonary disease. Average age was 69 6 9 years (range, 40-90 years). Fig 2 shows the baseline characteristics of patients who underwent CCAB. Patients’ demographics were similar across regions with the exception of proportion of white patients (87% [S] to 95% [MW]; P < .001) and patients with congestive heart failure (17% [MW] to 27% [W]; P ¼ .02). There was significant regional variation in the type of carotid imaging obtained as well as in preoperative medications (Fig 3). Notably, preoperative computed tomography angiography use ranged from 27% (W) to 57% (MW; P < .001). Overall duplex ultrasound use was 92% with a significant difference seen between regions (88% [S] to 94% [NE]; P ¼ .004). There was also variation in the use of preoperative aspirin (77% [MW] to 89% [NE]; P < .001), P2Y12 inhibitors (13% [NE] to 22% [MW; P ¼ .001), and beta blockade (73% [W] to 85% [NE]; P ¼ .001).
Operative characteristics. There was significant regional variation in the operative techniques used and neuromonitoring practices employed (Fig 4). The use of the eversion technique (7% [W] to 20% [NE]; P < .001) and shunting (29% [NE] to 71% [W]; P < .001) varied significantly. Electroencephalography monitoring (13% [W] to 52% [NE]; P < .001) and measurement of stump pressures (5% [NE] to 13% [MW]; P < .001) also differed between regions. There was also significant regional variation in the percentage of urgent/emergent procedures (23% [MW] to 37% [NE]; P < .001). Perioperative outcomes. Overall perioperative stroke was 3.6% (n ¼ 54) and when stratified by laterality, the ipsilateral perioperative stroke rate was 2.0% (n ¼ 30). No difference was seen between regions with respect to perioperative stroke (2.8% [NE] to 4.4% [MW]; P ¼ .52). Overall 30-day mortality was 3.0% (n ¼ 45), with no difference seen between regions (1.5% [MW] to 4.2% [NE]; P ¼ .05). Perioperative MI did not vary (0.6% [MW] to 1.8% [W]; P ¼ .62), nor did total perioperative stroke/death/MI rates (overall, 6.8%; 5.4% [W] to 7.3% [NE]; P ¼ .92). When the Bonferroni correction was used, there remained no difference in stroke, MI, or SDM across regions, but mortality became significant (P ¼ .01). Additional regional variation was found with respect to
4
Journal of Vascular Surgery
Wang et al
---
Fig 2. Baseline characteristics of patients undergoing carotid endarterectomy with coronary artery bypass (CCAB). Each symbol on the line represents a region. CHF, Congestive heart failure; COPD, chronic obstructive pulmonary disease; HTN, hypertension. *Significant variation (to P < .001).
2019
Fig 4. Operative characteristics of patients undergoing carotid endarterectomy with coronary artery bypass (CCAB). EEG, Electroencephalography. *Significant variation (to P < .001).
Fig 5. Perioperative complications of patients undergoing carotid endarterectomy with coronary artery bypass (CCAB). Before Bonferroni correction, no statistical difference was seen in regional outcomes. After Bonferroni correction, only perioperative death was statistically different between regions (P ¼ .01). MI, Myocardial infarction. Fig 3. Preoperative carotid artery imaging and medications of patients undergoing carotid endarterectomy with coronary artery bypass (CCAB). CTA, Computed tomography angiography; MRA, magnetic resonance angiography. *Significant variation (to P < .001).
perioperative rhythm disturbances requiring treatment (11% [W] to 22% [MW]; P ¼ .04). Fig 5 illustrates variation in perioperative outcomes between regions. Indications. Using the SVS guidelines for patients in whom staged procedures or CCAB should be considered,1 patients who were symptomatic (SYM) or who had severe asymptomatic bilateral carotid artery stenosis (BIL) were identified; 218 patients (15%) represented SYM and 267 patients (18%) fell within the BIL category. When SYM and BIL were combined (SYM-BIL) to comprise patients who fell within SVS guidelines, there was significant regional variation in guideline adherence (27% [NE] to 41% [W]; P ¼ .001). The most common indication for CCAB fell outside of SVS guidelines and represented patients with asymptomatic unilateral severe disease (UNI, 37%; n ¼ 552). When patients were stratified by indication (SYM, BIL, UNI), there remained no differences in outcomes (stroke, death, MI) between regions. SYM
and BIL were then compared with UNI to ascertain differences in perioperative outcomes. UNI had lower overall perioperative stroke rates (2.4% vs 4.9%; P ¼ .03) but similar 30-day mortality (2.2% vs 2.5%; P ¼ .75) and perioperative MI (0.7% vs 1.2%; P ¼ .40). The SYM-BIL cohort had higher rates of the combined stroke/death/ MI end point (7.8% vs 4.5%; P ¼ .04; Table). The remaining patients who did not fall into SYM, BIL, or UNI were missing contralateral carotid stenosis data and thus were unable to be analyzed within this subanalysis.
DISCUSSION As discussions surrounding health care have shifted toward efficient resource use and outcome optimization, variation in surgical practice has become a focus of debate.12 Studies have shown that the likelihood of undergoing an operation depends as much on one’s geographic location as on one’s clinical condition.17,18 Despite increasing evidence of the impact of regional variation, geographic differences in rates of surgical procedures do not appear to be decreasing.10 Determinants of surgical regional variation are likely to be multifactorial. Birkmeyer et al have argued that illness burden and diagnostic practices
Journal of Vascular Surgery Volume
-,
Number
Wang et al
5
-
Table. Perioperative outcomes of patients undergoing carotid endarterectomy with coronary artery bypass (CCAB) stratified by indication SYM-BIL (n ¼ 485 [32%]), No. (% total cohort)
UNI (n ¼ 552 [37%]), No. (% total cohort)
P value
All stroke
24 (4.9)
13 (2.4)
.03
Death
12 (2.5)
12 (2.2)
.75
Stroke/death
33 (6.8)
23 (4.2)
.06
6 (1.2)
4 (0.7)
.40
38 (7.8)
25 (4.5)
.03
Perioperative outcomes
MI SDM
MI, Myocardial infarction; SDM, stroke/death/myocardial infarction; SYM-BIL, symptomatic patients (SYM) and patients with asymptomatic bilateral carotid artery stenosis (BIL); UNI, patients with asymptomatic unilateral severe carotid disease.
explain only a small degree of the variation, and the majority can be better correlated to the physician’s beliefs and the patient’s preferences.17 Differences in how surgeons apply evidence likely relates to clinical uncertainty and gaps in current scientific knowledge.18 Indeed, debate on best practice for management of combined coronary and carotid disease is ongoing. Critics of CCAB, who advocate for a staged approach for the treatment of concomitant carotid and coronary disease, cite higher reported rates of perioperative stroke and death with the combined procedure.7,19 Using the Society of Thoracic Surgeons Database, Prasad et al6 found CCAB to carry higher rates of operative mortality and neurologic complications compared with isolated CABG in patients with a history of carotid disease. Yet, many have argued for the safety of combined CCAB.3,4 In a recent metaanalysis, Sharma et al20 analyzed 25,021 patients across 12 studies undergoing either staged CEA and CABG or CCAB and found no difference in perioperative stroke or death between operative strategies. Using the National Inpatient Sample, Gopaldas et al21 similarly found no differences in mortality or neurologic complications in 6153 patients undergoing staged vs combined (CCAB) procedures but did demonstrate a higher risk of overall complications in the staged cohort. Despite these data, there remains a lack of large randomized trials comparing treatment paradigms and differing institutional strategies. To that end, across societal guidelines, equivalence is given to staged and combined strategies.1,2,22 This has become more complex with the rise of percutaneous coronary interventions and carotid stenting alternatives.23 We found that across the United States, despite significant variation in selection of patients, volume, and preoperative and intraoperative management, CCAB outcomes remain similar across regions. We found an overall perioperative SDM rate of 6.8% (range, 5.4%-7.3%; P ¼ .92), which is lower than previously reported rates of 9% to 12% in older retrospective reviews.24 This is likely to be a testament to advances in medical and surgical management but may also highlight the trend of care consolidation as more advanced procedures are being performed exclusively at select centers of excellence. Whereas proportional CCAB to CEA regional
volume did not correlate with outcomes, it is important to highlight that only 47% (n ¼ 155) of all VQI centers performed this procedure during the study period; 42 centers accounted for 79% of the patients in this analysis. National trends in coronary revascularization as well as in CCAB institutional experience can provide some insight into these findings. In a study of Medicare beneficiaries, Culler et al25 found a significant decrease in the number of CABGs performed from 2008 through 2012, and this was accompanied by a 20% increase in the number of facilities performing percutaneous coronary intervention compared with a 2% increase in the number of sites performing CABG. These dates are of particular interest as 2012 saw nearly a doubling of the number of VQI sites, and correspondingly, 110 (71%) of the sites in our study joined in 2012 or later. We would also argue that the fact that less than half of all VQI sites perform the combined procedure is a testament to (1) the importance of institutional experience as not all facilities have the resources to perform this procedure, (2) the impact of clinical uncertainty on practice patterns, and (3) the role of the physician’s beliefs in selection of treatment strategy. For the management of combined carotid and coronary disease, the SVS guidelines highlight two groups of patients who should be considered for CEA before or concomitantly with CABG: patients with symptomatic carotid stenosis and patients with severe bilateral asymptomatic disease.1 The more recently published European Society for Vascular Surgery guidelines add a temporal component and stenosis criteria for the first group, recommending staged or synchronous revascularization in patients with a history of stroke or TIA in the preceding 6 months and a $50% stenosis.2 In this analysis, 15% of CCABs carried a symptomatic diagnosis, whereas 18% fell into the second category. In combining these groups and comparing across regions, there was significant variability in the proportion of patients who met either criterion; 27% of CCABs in the Northeast had an indication of SYM or BIL compared with 41% of CCABs performed in the West. Indeed, the large range of reported percentages of symptomatic patients undergoing CCAB (4%-44%) may in part be explained by differences in regional practices.23,26,27
6
Journal of Vascular Surgery
Wang et al
---
Regional variation in the treatment of carotid artery disease is known to exist. Using the VQI, Shean et al9 analyzed 57,555 patients undergoing carotid artery revascularization and demonstrated widespread variation in selection of patients and preoperative management. Using Medicare data, Goodney et al28 demonstrated a sevenfold difference in regional variation in the use of CEA. These differences and the factors leading to them may also influence CCAB practices. As most CCAB patients have asymptomatic carotid stenosis, concerns about this practice have been voiced.29 The effect of symptomatic status on combined procedural outcomes has been debated. Some have found unacceptably high rates of perioperative stroke (9%-12%) in CCAB populations with high proportions of symptomatic patients.27,30 Others found no difference in neurologic CCAB outcomes when stratifying by symptomatic status.26 A randomized controlled trial performed in 185 UNI patients demonstrated a perioperative stroke benefit to patients who underwent CEA before or concomitantly with CABG compared with those who underwent CEA after CABG (1% vs 7.7%; P ¼ .01).31 Our analysis demonstrated that UNI had lower rates of stroke (2.4% vs 4.9%; P ¼ .03) and SDM (4.5% vs 7.8%; P ¼ .03) compared with SYM-BIL. These data suggest that carotid symptomatic status does affect CCAB outcomes and should factor into preoperative risk profile assessment. Despite differences in selection of patients and guideline adherence, no differences were seen in outcomes across regions. After Bonferroni correction, a difference in mortality was seen across regions (P ¼ .01) but did not correlate with absolute or proportional volume. Furthermore, when subgroup analysis was performed within each indication group (SYM, BIL, UNI), despite regional variance in selection of patients, no differences in outcomes between regions was noted. Whereas the data presented in this analysis suggest that CCAB is performed well across the United States, the large proportion of CCABs performed for asymptomatic unilateral disease suggests there is a continued role for quality projects aimed at either improved guideline adherence or data to better clarify contemporary guideline appropriateness. However, whereas the goal should be approaching standardization of care and outcomes, areas of clinical equipoise will continue to existdand with that, regional variation in surgical experiences and practice as well as individual patient management. This study has several limitations. One limitation lies in its retrospective review of a prospectively maintained database, prone to entry errors and missing data; the VQI performs regular audits to minimize these issues. Statistical analysis was designed to assess whether variation across all regions existed and did not identify the regions that led to the differences seen. In addition, true regional volume may not be captured in regions with
2019
fewer participating VQI centers. The VQI represents a minority of CEAs performed in the United States during this time, and thus many CCABs and their effect may be lost. As an extension of this, the VQI includes only a subset of hospital centers and does not include the Veterans Affairs population; thus, these data may not be fully representative of national practice patterns and outcomes. Furthermore, our data do not account for geographic variation in disease burden. Last, the true reasons for variation in practice patterns can only be speculated on; further study of system, provider, and patient preferences would be needed to address this question, and further investigation is warranted.
CONCLUSIONS Significant regional variation exists across VQI centers in use of combined CEA and CABG. Despite differences in volume and practices, perioperative outcomes were similar. Asymptomatic unilateral severe carotid stenosis represents the most common carotid indication, and significantly lower 30-day stroke rates were seen in UNI compared with the SYMP-BIL cohorts. These data suggest that CCAB is performed well across the United States, but most patients fall outside of societal guidelines. As clinical uncertainty about best practices in management of combined carotid and coronary disease continues to exist, so too will regional variation in surgical application. However, these differences do not necessarily correlate with worse CCAB outcomes.
AUTHOR CONTRIBUTIONS Conception and design: LW, WC Analysis and interpretation: LW, EE, JM, PG, VP, MC, ME, WC Data collection: LW, EE, WC Writing the article: LW, WC Critical revision of the article: LW, EE, JM, PG, VP, MC, ME, WC Final approval of the article: LW, EE, JM, PG, VP, MC, ME, WC Statistical analysis: LW, EE, WC Obtained funding: Not applicable Overall responsibility: WC
REFERENCES 1. 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. J Vasc Surg 2011;54:e1-31. 2. Naylor AR, Ricco JB, de Borst GJ, Debus S, de Haro J, Halliday A, et al. Editor’s choicedmanagement of atherosclerotic carotid and vertebral artery disease: 2017 Clinical Practice Guidelines of the European Society for Vascular Surgery (ESVS). Eur J Vasc Endovasc Surg 2018;55:3-81. 3. Kougias P, Kappa JR, Sewell DH, Feit RA, Michalik RE, Imam M, et al. Simultaneous carotid endarterectomy and coronary artery bypass grafting: results in specific patient groups. Ann Vasc Surg 2007;21:408-14. 4. Akins CW, Hilgenberg AD, Vlahakes GJ, Madsen JC, MacGillivray TE, LaMuraglia GM, et al. Late results of combined carotid and coronary surgery using actual versus actuarial methodology. Ann Thorac Surg 2005;80:2091-7.
Journal of Vascular Surgery Volume
-,
Number
Wang et al
7
-
5. Darling RC, Dylewski M, Chang BB, Paty PS, Kreienberg PB, Lloyd WE, et al. Combined carotid endarterectomy and coronary artery bypass grafting does not increase the risk of perioperative stroke. Cardiovasc Surg 1998;6:448-52. 6. Prasad SM, Li S, Rankin JS, O’Brien SM, Gammie JS, Puskas JD, et al. Current outcomes of simultaneous carotid endarterectomy and coronary artery bypass graft surgery in North America. World J Surg 2010;34:2292-8. 7. Borger MA, Fremes SE, Weisel RD, Cohen G, Rao V, Lindsay TF, et al. Coronary bypass and carotid endarterectomy: does a combined approach increase risk? A metaanalysis. Ann Thorac Surg 1999;68:14-20; discussion: 21. 8. Shean KE, O’Donnell TF, Deery SE, Pothof AB, Schneider JR, Rockman CB, et al. Regional variation in patient outcomes in carotid artery disease treatment in the Vascular Quality Initiative. J Vasc Surg 2018;68:749-59. 9. Shean KE, McCallum JC, Soden PA, Deery SE, Schneider JR, Nolan BW, et al. Regional variation in patient selection and treatment for carotid artery disease in the Vascular Quality Initiative. J Vasc Surg 2017;66:112-21. 10. Zettervall SL, Buck DB, Soden PA, Cronenwett JL, Goodney PP, Eslami MH, et al. Regional variation exists in patient selection and treatment of abdominal aortic aneurysms. J Vasc Surg 2016;64:921-7. 11. Zettervall SL, Soden PA, Buck DB, Cronenwett JL, Goodney PP, Eslami MH, et al. Significant regional variation exists in morbidity and mortality after repair of abdominal aortic aneurysm. J Vasc Surg 2017;65:1305-12. 12. Soden PA, Zettervall SL, Shean KE, Vouyouka AG, Goodney PP, Mills JL, et al. Regional variation in outcomes for lower extremity vascular disease in the Vascular Quality Initiative. J Vasc Surg 2017;66:810-8. 13. Soden PA, Zettervall SL, Curran T, Vouyouka AG, Goodney PP, Mills JL, et al. Regional variation in patient selection and treatment for lower extremity vascular disease in the Vascular Quality Initiative. J Vasc Surg 2017;65:108-18. 14. Cronenwett JL, Kraiss LW, Cambria RP. The Society for Vascular Surgery Vascular Quality Initiative. J Vasc Surg 2012;55:1529-37. 15. United States Census Bureau. Available at: https://www. census.gov/geo/reference/webatlas/regions.html. Accessed April 3, 2018. 16. Cronenwett JL, Likosky DS, Russell MT, Eldrup-Jorgensen J, Stanley AC, Nolan BW. A regional registry for quality assurance and improvement: the Vascular Study Group of Northern New England (VSGNNE). J Vasc Surg 2007;46:1093-102.e1. 17. Birkmeyer JD, Reames BN, McCulloch P, Carr AJ, Campbell WB, Wennberg JE. Understanding of regional variation in the use of surgery. Lancet 2013;382:1121-9. 18. Reames BN, Shubeck SP, Birkmeyer JD. Strategies for reducing regional variation in the use of surgery. Ann Surg 2014;259:616-27.
19. Dubinsky RM, Lai SM. Mortality from combined carotid endarterectomy and coronary artery bypass surgery in the US. Neurology 2007;68:195-7. 20. Sharma V, Deo SV, Park SJ, Joyce LD. Meta-analysis of staged versus combined carotid endarterectomy and coronary artery bypass grafting. Ann Thorac Surg 2014;97:102-9. 21. Gopaldas RR, Chu D, Dao TK, Huh J, LeMaire SA, Lin P, et al. Staged versus synchronous carotid endarterectomy and coronary artery bypass grafting: analysis of 10-year nationwide outcomes. Ann Thorac Surg 2011;91:1323-9. 22. Windecker S, Kolh P, Alfonso F, Collet JP, Cremer J, Falk V, et al. 2014 ESC/EACTS guidelines on myocardial revascularization. Rev Esp Cardiol (Engl Ed) 2015;68:144. 23. Timaran CH, Rosero EB, Smith ST, Valentine RJ, Modrall JG, Clagett GP. Trends and outcomes of concurrent carotid revascularization and coronary bypass. J Vasc Surg 2008;48: 355-61.e1. 24. Naylor AR, Cuffe RL, Rothwell PM, Bell PR. A systematic review of outcomes following staged and synchronous carotid endarterectomy and coronary artery bypass. Eur J Vasc Endovasc Surg 2003;25:380-9. 25. Culler SD, Kugelmass AD, Brown PP, Reynolds MR, Simon AW. Trends in coronary revascularization procedures among Medicare beneficiaries between 2008 and 2012. Circulation 2015;131:362-70; discussion: 370. 26. Jones DW, Stone DH, Conrad MF, Baribeau YR, Westbrook BM, Likosky DS, et al. Regional use of combined carotid endarterectomy/coronary artery bypass graft and the effect of patient risk. J Vasc Surg 2012;56:668-76. 27. Brown KR, Kresowik TF, Chin MH, Kresowik RA, Grund SL, Hendel ME. Multistate population-based outcomes of combined carotid endarterectomy and coronary artery bypass. J Vasc Surg 2003;37:32-9. 28. Goodney PP, Travis LL, Malenka D, Bronner KK, Lucas FL, Cronenwett JL, et al. Regional variation in carotid artery stenting and endarterectomy in the Medicare population. Circ Cardiovasc Qual Outcomes 2010;3:15-24. 29. Venkatachalam S, Gray BH, Mukherjee D, Shishehbor MH. Contemporary management of concomitant carotid and coronary artery disease. Heart 2010;97:175-80. 30. Mackey WC, Khabbaz K, Bojar R, O’Donnell TF. Simultaneous carotid endarterectomy and coronary bypass: perioperative risk and long-term survival. J Vasc Surg 1996;24:58-64. 31. Illuminati G, Ricco JB, Caliò F, Pacilè MA, Miraldi F, Frati G, et al. Short-term results of a randomized trial examining timing of carotid endarterectomy in patients with severe asymptomatic unilateral carotid stenosis undergoing coronary artery bypass grafting. J Vasc Surg 2011;54:993-9; discussion: 998-9.
Submitted Nov 7, 2018; accepted Feb 1, 2019.