Contemporary Results of Carotid Endarterectomy for Asymptomatic Carotid Stenosis

Methods

The National Surgical Quality Improvement Program (NSQIP)* is a nationally validated, risk-adjusted, outcomes-based program to measure the quality of surgical care. NSQIP collects data on 136 variables, including preoperative risk factors, intraoperative variables, and 30-day postoperative mortality and morbidity outcomes. The data are collected, validated, and submitted by a trained Surgical Clinical Reviewer at each site. CEA is included on the list of Current Procedural Terminology codes that the NSQIP tracks.¹³

NSQIP is the only national surgical database that contains data with this level of detail and is entered by a Surgical Clinical Reviewer who is trained and employed for this purpose. NSQIP also has the advantage of identifying morbidity by independent chart review after patient discharge versus the VascularRegistry (Society of Vascular Surgery, Chicago, Ill). The VascularRegistry tracks outcomes of carotid stenting and CEA but depends on self-reporting.

NSQIP has been validated through 2 studies, which specifically examined sites with higher than expected and lower than expected risk-adjusted outcomes.^14,15 All data are continuously submitted and monitored through data audits and the www.acsnsqip.org web site. In 2007, there were 183 participating sites. For 2005 and 2006, data are available from 121 hospitals. Approximately half of the participating sites are community hospitals.

The 2005, 2006, and 2007 NSQIP databases were searched for the Current Procedural Terminology code for CEA (35301) in combination with a postoperative International Classification of Diseases, 9th Revision, Clinical Modification diagnosis code of occlusion and stenosis of the carotid artery without mention of cerebral infarction (433.10). To eliminate any potentially symptomatic patients, those with impaired sensorium, coma, hemiplegia/hemiparesis, a history of transient ischemic attack, cerebrovascular accident with or without residual neurological deficit, tumor involving the central nervous system, paraplegia/paraparesis, or quadriplegia/quadriparesis were excluded. Subsequently, to eliminate any confounding factors that may affect outcomes, the following exclusion criteria were also applied: disseminated cancer, open wound with or without infection, emergency case classification, chemotherapy within 30 days, radiotherapy within 90 days, systemic sepsis within 48 hours, contaminated/dirty wound class, cases with a simultaneous procedure, and previous operation within 30 days.

Patient characteristics recorded included gender, age, dyspnea (at rest or with moderate exertion), history of chronic obstructive pulmonary disease (COPD), history of congestive heart failure (CHF) within 30 days, diabetes, history of myocardial infarction (MI) within 6 months, history of angina within 30 days, and renal failure with chronic dialysis dependence. Preoperative functional status was recorded and is defined by NSQIP as the level of self-care demonstrated in the 30 days before surgery and is categorized as independent, partially dependent, and totally dependent. Postoperative complications analyzed included peripheral nerve injury, wound infection, pneumonia, stroke, MI, and death.

Univariate analysis was performed using the χ² test for categorical variables and logistic regression for continuous variables. Multivariate analysis was performed using multivariable logistic regression. Statistical analysis was performed using the software package Statistical Analysis System (SAS) Version 7.0 (Cary, NC) and the Systat Software Version 11.0 (Chicago, Ill).

Results

Review of the database identified 10 423 CEAs. After applying the exclusion criteria, 5009 CEAs for asymptomatic carotid stenosis remained. The mean age was 71 years. Table 1 describes the clinical characteristics of the patients. (Data were not available on the function status of 16 patients.) General anesthesia was used in 4121 patients (82%), regional anesthesia in 677 (14%), local anesthesia with monitored anesthesia care in 168 (3%), and other in 43 (1%).

In the 30-day postoperative period, there were 48 strokes, 28 deaths, and 11 MIs. The stroke, death, and MI rates were 0.96%, 0.56%, and 0.22%, respectively. The combined stroke and death rate was 1.4% and the combined stroke, death, and MI rate was 1.6%. The incidence of peripheral nerve injury, wound infection, and pneumonia was 0.32%, 0.68%, and 0.66%, respectively.

Univariate analysis showed that COPD was associated with an increased risk of stroke (Table 2). Dyspnea, COPD, history of CHF, and history of MI were associated with an increased risk of death and the combined outcome of stroke and death (Table 2). Increased risk of combined stroke, death, and MI was associated with presence of dyspnea, COPD, history of CHF, history of MI, and chronic dialysis dependence (Table 2). Analysis was not performed on the complication of MI due to the small number of events.

Multivariate analysis demonstrated that COPD, history of CHF, and history of MI increased the risk of death as well as the combined outcome of stroke or death (Table 3). COPD and history of MI increased the risk of the combined outcome stroke, death, or MI (Table 3).

Discussion

The medical management of asymptomatic carotid stenosis has undergone a major shift over the last decade, in particular with respect to the use of statins. Since ACAS and ACST, there have been no randomized direct comparisons between intervention for asymptomatic carotid stenosis and best medical management.

As a result, a great deal of controversy exists regarding the role of CEA for asymptomatic carotid stenosis.^16–19 Some have even gone so far as to suggest that asymptomatic carotid stenosis is being intentionally overtreated for financial reasons.¹⁸ To address this issue, we analyzed the outcomes of CEA for asymptomatic carotid stenosis from a large contemporary series and examined the literature in an effort to establish stroke rates with current best medical management of asymptomatic carotid stenosis.

The stroke and death rate in this contemporary series of CEA for asymptomatic carotid stenosis from the NSQIP database is comparable to that of ACAS if risk of angiography is not included.¹ The excellent perioperative stroke and death rate in ACAS has been criticized as not reflective of communitywide practice due to the very stringent selection of participating surgeons, which required a rate of perioperative stroke or death of no more than 3% among asymptomatic patients.²⁰ In ACST, participation criteria for surgeons were less stringent, requiring a <6% postoperative stroke or death rate and the 30-day stroke or death rate in that study was 3.1%. These results called into question whether it is feasible for the results seen in ACAS to be replicated in the general community. Based on the stroke and death rate of this contemporary series representing results from both community- and university-based hospitals, it would appear that current outcomes approximate the results demonstrated in the ACAS trial.

COPD, CHF, and history of MI within 6 months preoperatively were all associated with adverse outcomes. CHF and history of recent MI are widely accepted as major predictors of severe to fatal postoperative cardiac complications.^21–24 CHF has also been shown to be associated with an increased risk of 30-day stroke or death after CEA specifically.^25,26 This may indicate that for patients with these conditions, the risks of CEA for asymptomatic carotid stenosis outweighs the benefits.

Our results demonstrated no difference in the incidence of perioperative stroke or death between males and females with approximately 50% of the patients in this NSQIP analysis being female. Other authors have also found no gender influence on clinical outcomes after CEA.²⁷ Previously, however, women have been shown to have a higher operative stroke/death risk resulting in a significantly smaller relative risk reduction. The Cochrane Review of CEA for asymptomatic carotid stenosis, which pooled the results of the Veterans’ Administration (VA) trial, ACAS and ACST found that the relative risk reduction from CEA was 51% for men but only 4% for women.²⁸ Women also have a lower incidence of spontaneous strokes than men.²⁸ However, there were significantly fewer women than men in these studies and the lack of efficacy may be due to Type II error.

The controversy surrounding CEA for asymptomatic carotid stenosis is largely based on the increasing volume of data, which suggests that the stroke-reducing effects of statins extends beyond simply improving the lipid profile. As early as 1998, data were emerging that statin treatment reduces levels of C-reactive protein, a marker of inflammation, which has been shown to predict risk of peripheral arterial disease.^28,29 Subsequently, statins were shown to have a plaque-stabilizing effect in human carotid plaques.⁶

Consequently, data emerged from the Stroke Prevention by Aggressive Reduction in Cholesterol Levels trial, which demonstrated that high-dose atorvastatin after stroke or transient ischemic attack was associated with a 16% relative risk reduction for nonfatal or fatal stroke.¹⁰ Even more compelling are the recent results from the Justification for the Use of Statins in Prevention: an Intervention Trial Evaluating Rosuvastatin (JUPITER). The JUPITER trial enrolled apparently healthy patients and randomized them to treatment with rosuvastatin or placebo. There was a 48% risk reduction in the risk of stroke in the rosuvastatin group.⁹

In examining the literature, it appears that from the results of trials, studies, and reports, there has been a gradual reduction in the average annual risk of ipsilateral stroke from approximately 2.5% in the mid-1980s to approximately 1% by 2008.¹⁷ The medical arm of ACAS had an 11% risk of ipsilateral stroke over 5 years. Similarly, the medical group of the VA trial had a 9.4% incidence of ipsilateral stroke over a mean follow-up of 47.9 months. The 2006 result of the Second Manifestations of Arterial Disease Study (SMART) study showed that 6 of 221 patients (3%) with asymptomatic carotid artery stenosis of ≥50% had an ischemic stroke over a mean follow-up period of 3.6 years.³⁰ In addition, only 63% of the patients in this group were taking antiplatelet agents and 45% were taking lipid-lowering agents.³⁰ The recent data from JUPITER demonstrating a significantly lower risk of stroke in patients taking rosuvastatin suggests that if the asymptomatic carotid stenosis group in the SMART study had all been taking a statin, the incidence of stroke would have been even lower.

The most recent large trial of CEA for asymptomatic carotid stenosis is the ACST.¹¹ The ACST demonstrated a 5-year stroke risk after CEA of 3.8%, excluding perioperative stroke and death. If the 0.96% postoperative stroke rate from our contemporary NSQIP analysis is combined with the 3.8% risk from ACST, this would result in a stroke rate of 4.8% over 5 years, which translates into an annual stroke risk of approximately 1%. This is comparable to the 0.8% annual stroke risk seen in the SMART study. However, there are some additional considerations with respect to this comparison. ACST included patients with >70% carotid stenosis, whereas approximately one third of patients in the SMART trial had carotid stenosis of 50% to 70% resulting in a lower stroke risk overall in that trial. If routine use of statins and current medical management had been used in the ACST trial and with greater frequency in the SMART trial, the 5-year stroke risk in both of the trials may have been lower. Nevertheless, the available data from these trials support the concept that the 5-year risk of stroke due to carotid stenosis is remarkably similar in patients managed medically and surgically.

Although the NSQIP database is detailed and has 30-day outcomes, it does have limitations for this study. One major weakness of our study is that we are unable to obtain from the database the degree of preoperative carotid stenosis. The low stroke rate in the 50% carotid stenosis group in the SMART study is to be expected given that the recommendation from ACAS was CEA for asymptomatic stenosis of ≥60%. If the data were available, the most accurate comparison would be to compare patients from the SMART study and the NSQIP database having the same degree of carotid stenosis. We also do not have information with regard to the laterality or cerebral distribution of the postoperative strokes, whether the strokes were ischemic or hemorrhagic, or how disabling the strokes were. NSQIP does not provide information on preoperative CT/MRI imaging. As such, it is possible that patients with a stroke by imaging could have been included in the study. In addition, NSQIP does not provide technical details of the operation such as whether shunts or patches were used and whether antiplatelet agents were administered.

Another weakness of our study is the very low incidence of peripheral nerve injuries compared with the 4% to 8% reported in the literature.^31–34 Peripheral nerves that can be injured during CEA include the hypoglossal nerve, the vagus nerve, the marginal mandibular nerve, and the glossopharyngeal nerve. This difference calls into question the ability of the database to capture subtleties. There are certain diagnoses that are clear and easily identified such as stroke and death. The nurses who enter the data may not recognize a situation such as a mild hypoglossal nerve injury with a slight tongue deviation as a peripheral nerve injury. Finally, there is a very low incidence of MI, which may be related to the fact that NSQIP does not record troponin levels for every patient as the Stenting and Angioplasty with Protection in Patients at High Risk for Endarterectomy trial did.³⁵

Conclusion

Although the comparison of NSQIP outcomes with ACST and SMART outcomes has significant limitations, the results of our study suggest that contemporary outcomes of CEA for asymptomatic carotid stenosis are nearly identical to those of the ACAS study that was published 14 years ago. On the other hand, medical management of asymptomatic carotid stenosis has evolved significantly since ACAS with impressive results. There has been a great deal of enthusiasm in recent years to compare the outcomes of CEA and carotid stenting. However, the results of this study indicate that the more relevant question in management of asymptomatic carotid stenosis is whether the outcomes of any intervention are superior to current best medical management. There are no trials either in progress or planned that include a best medical treatment arm. Plans for the Transatlantic Asymptomatic Carotid Intervention Trial (TACIT) were initiated in 2004 and intended to randomize 2500 patients to carotid stenting, CEA, and best medical therapy. To date, funding for the TACIT has not been secured (Katzen, personal communication). The findings in this study emphasize the importance of going forward with such a trial.