Activities of Daily Living Is a Critical Factor in Predicting Outcome After Carotid Endarterectomy in Asymptomatic Patients
Background and Purpose—Ability to perform basic daily activity represented by functional status (FNS) before surgery can be assessed in the clinic for determining health status of the patient. We sought to study the effect of FNS on postoperative outcomes after carotid endarterectomy (CEA) in a national data set.
Methods—National Surgical Quality Improvement Project is a national data set, which includes data from >300 hospitals. Patients who underwent CEA were identified by Current Procedural Terminology code and divided into 3 categories based on FNS: independent, partially dependent, and dependent. Thirty-day postoperative stroke, death, and other postoperative complications were identified as the study end point. We used multivariate logistic regression to estimate odds ratio for outcomes while controlling for sex, race, diabetes mellitus, cardiovascular disease, smoking, and other confounders.
Results—Of 19 748 CEAs, 19 348 (97.97%) were functionally independent, 377 (1.99%) were functionally partially dependent, and 23 (0.12%) were functionallydependent. In functionally independent group, there were 196 (1.01%) strokes, 84 (0.43%) deaths, and 1416 (7.17%) other complications, whereas in the functionally partially dependent group, there were 14 (3.71%) strokes, 10 (2.65%) deaths, and 80 (21.22%) other complications. In multivariable risk-adjusted model, using functionally independent as reference, functionally partially dependent was associated with death (odds ratio, 3.3; 95% confidence interval, 1.6–6.8; P<0.001), stroke (odds ratio, 3; 95% confidence interval, 1.7–5.4; P<0.001), and other complications (odds ratio, 2.5; 95% confidence interval, 1.9–3.2; P<0.001).
Conclusions—In this national data set, patient’s inability to perform basic activities of independent living is associated with adverse postoperative outcomes after CEA. Hence, FNS should be vigilantly assessed in clinic for risk stratification along with other objective factors for gauging risk of adverse outcomes after CEA.
A landmark study from 1990s, asymptomatic Carotid Atherosclerosis Study (ACAS)1 established the benefit of carotid endarterectomy (CEA) in primary stroke prevention among patients with asymptomatic internal carotid artery stenosis. However, in the ACAS study, elderly patients aged ≥80 years and patients with chronic cardiovascular or other debilitating conditions, including congestive heart failure, unstable angina, uncontrolled atrial fibrillation, severe valvular heart disease, uncontrolled diabetes mellitus, renal insufficiency, respiratory insufficiency, hepatic disease, uncontrolled hypertension, and previous major surgery within 30 days, were excluded. These factors have been shown to be independently associated with adverse postoperative outcomes in recent studies.2,3 Also, substantial improvement in medical management of patients with antiplatelet therapy and statin has raised question about the appropriateness of CEA in asymptomatic patients.4,5
Conventional models of stroke prevention after CEA in asymptomatic patients are not calibrated to predict the outcomes in the patients accurately with inability to perform daily activities of living.2 Functional status (FNS) is a validated measure to assess patient’s ability to perform routine daily activities and is recorded in National Surgical Quality Improvement Project (NSQIP), a national data set. Our study aims to assess the effect of preoperative FNS on postoperative outcomes, including 30-day mortality after CEA in asymptomatic patients, performed for stroke prevention with the primary hypothesis that preoperative functional debility is associated with worsened postoperative outcomes and in turn survival of the patients. We performed this analysis in NSQIP, a national data set.
American College of Surgeon-NSQIP (ACS-NSQIP) is a risk-adjusted data collection mechanism that collects and analyzes clinical outcomes data.6 Participating hospitals use their collected data to develop quality initiatives that can improve surgical care and to identify elements in provided health care that can be improved when compared with other institutions. ACS-NSQIP collects data on a variety of clinical variables, including preoperative risk factors and intraoperative variables, and tracks patients for 30 days postoperative to record mortality and morbidity outcomes for patients undergoing major surgical procedures in both the inpatient and the outpatient setting.6 ACS-NISQIP chapter 4 provides standard definitions of these variables.7 Every site’s surgical clinical nurse reviewer captures outcomes data using a variety of methods, including medical chart abstraction. ACS-NSQIP monitors accrual rates and data sampling methodologies and conducts inter-rater reliability audits on a random basis. Hospitals that are flagged by internal diagnostics are also audited. The ACS-NSQIP training, data collection, and auditing process have been shown to be highly reliable.
Using the ACS-NSQIP database, we identified asymptomatic patients undergoing CEA for stroke prevention from 2005 to 2010. The American Medical Association’s Current Procedural Terminology (CPT, Chicago, IL) code 35301 that describes CEA was used to identify the patients who had received this operation. To include only asymptomatic patients in the analysis selectively, we excluded patients who had history of stroke, paraplegia, hemiplegia, quadriplegia, transient ischemic attacks, and coma or patients who underwent emergency surgery.
Assessment of Risk Factors and End Points
We divided selected patients into 3 categories based on preoperative FNS. This variable focuses on the patient’s abilities to perform activities of daily living (ADLs) in 30 days before surgery. ADLs are defined as the activities usually performed in the course of a normal day in a patient’s life. ADLs include bathing, feeding, dressing, toileting, and mobility. Categories included functionally independent (FNS-I), functionally partially dependent (FNS-PD), and functionally completely dependent (FNS-D; Table 1). Preoperatively assessed demographic factors (age, sex, and race), smoking status, body mass index, medical history (history of hypertension, diabetes mellitus, symptoms of angina within 1 month, cardiac surgery, or revascularization and history of chronic obstructive pulmonary disease [COPD]), and perioperative factors (American Society of Anesthesiology [ASA] classification, operation time, and resident participation) were considered to be potential covariates/confounders. Thirty-day postoperative stroke, mortality, and other postoperative complications were identified as study end point. We created composite end points to facilitate a better understanding of adverse outcomes. These composite end points are renal complications, pulmonary complications, cardiovascular complications, and wound complications (Table 2). Online-only Data Supplement outlines the definition of postoperative factors assessed in this study in detail.
We described categorical data as absolute numbers and percentage prevalence (%) in the study cohort and continuous variables as means±SD. To assess the association of potential covariates with exposure variable (FNS status), we performed χ2 or Fisher exact test for categorical variables and 2-tailed Student t test or the Wilcoxon Mann–Whitney test for continuous variables, depending on the fulfillment of normality assumption by each variable. The normality assumption was evaluated using criteria of skewness (absolute value, <1), kurtosis (absolute value, <1), and Kolmogorov–Smirnov test (P value of the test, >0.05).
The distributions of individual postoperative complications among different FNS groups were assessed using the χ2 or Fisher exact test. Multivariable logistic regression modeling was used to assess the association between FNS status and postoperative complications while controlling for possible confounders. Multivariable models were created by adding candidate covariates previously identified as risk factors for adverse events after CEA based on the literature review (age, female sex, race, coronary artery disease [defined as history of angina], smoking, COPD, ASA classification, and history of revascularization/amputation2,3,8,9 to the model containing FNS as main exposure of interest). Candidate covariates were evaluated for inclusion in adjusted models using backward selection method with exclusion of the potential confounders from the final model if P value is >0.05, and exclusion does not change the odds ratio (OR) of main predictor variable by >10%.
All statistical analyses were conducted using SAS version 9.3 for Windows (SAS, Cary, NC). A value of P<0.05 was considered statistically significant.
Patients Sample and Resident Participation
From 2005 to 2010, a total of 35 698 patients underwent CEAs; of these, 19 748 CEAs met the study inclusion criteria; the mean age was 71.02±9.03 years; 41.8% of patients were women and 3.5% were black. The distribution of FNS before surgery was 97.97% independent (19 348), 1.99% partially dependent (377), and 0.12% dependent (23). When compared with patients undergoing CEA with FNS-I, those with FNS-PD had a greater percentage of female subjects (41.7% versus 47.5%), higher mean age (73.94±9.32 versus 70.96±9.01), higher number of black (8.2% versus 3.4%), a higher rate of history of peripheral revascularization (16.2% versus 10.2%), a higher frequency of ASA class IV (29.4% versus 10.1%), greater prevalence of COPD (19.4% versus 9.9%), and higher reported history of angina 1 month before surgery (7.4% versus 2.4%). All other baseline characteristics were similar in these 2 groups (Table 3).
Thirty-Day Perioperative Outcomes
Incidence of mortality after elective CEA in asymptomatic patients within 30 days was 106 (0.5%), 211 (1.1%) subjects had stroke, and 1507 (7.6%) subjects had other surgery-related complication. Rates of perioperative outcome categorized by FNS before surgery are shown in Table 4. Mortality rates were 0.5% among patients undergoing CEA with FNS-I versus 2.7% of patients with FNS-PD; however, there were no deaths recorded in FNS-D cohort within 30 days, which might be because of a small sample size; postoperative stroke rate among FNS-I was 1% versus 3.7% and 4.3% among FNS-PD and FNS-D individuals; postoperative cardiovascular complication rates were 1% in FNS-I versus 2.9% in FNS-PD and 4.3% in FNS-D; and all the other adverse events had higher occurrence in FNS-PD and FNS-D when compared with FNS-I group (21.2% and 47.8% versus 7.3%).
In adjusted multivariable modeling controlling for age, sex, race, smoking status, history of COPD, angina symptoms, history of revascularization, and ASA class, FNS before CEA was significantly associated with 30-day risk of adverse events. Individuals in FNS-PD and FNS-D cohort had, respectively, 3.11 (95% confidence interval [CI], 1.73–5.60; P<0.001) and 3.15 (95% CI, 0.41–24.33; P=0.27) higher odds of mortality when compared with those in FNS-I cohort. FNS-PD had 3.18 (95% CI, 1.55–6.52; P=0.002) and FNS-D had 4.02 (95% CI, 0.51–31.93; P=0.19) times higher odds of stroke when compared with individuals in FNS-I cohort. Similar association was noticed with other postoperative complications, with FNS-PD having 2.54 (95% CI, 1.95–3.32; P<0.001) times higher and FNS-D having 6.50 (95% CI, 2.64–15.97; P<0.001) times higher odds when compared with those with FNS-I (Table 5).
In our adjusted model, other variables independently associated with mortality were age (OR, 1.05; 95% CI, 1.02–1.07; P=0.000), black race (OR, 2.34; 95% CI, 1.06–5.19; P=0.03), and history of COPD (OR, 3.63; 95% CI, 2.28–5.76; P=0.000). Other variables independently associated with stroke were history of angina within 30 days (OR, 2.58; 95% CI, 1.49–4.48; P=0.000), history of COPD (OR, 1.50; 95% CI, 1.02–2.22; P=0.03), and history of revascularization/amputation (OR, 1.52; 95% CI, 1.04–2.24; P=0.02). Other variables independently associated with other complication variables (defined in Table 2) were black race (OR, 1.55; 95% CI, 1.21–1.99; P=0.000), active smoker (OR, 1.21; 95% CI, 1.07–1.38; P=0.003), history of COPD (OR, 1.53; 95% CI, 1.31–1.78; P=0.000), history of revascularization (OR, 1.31; 95% CI, 1.12–1.53; P=0.000), ASA class III (OR, 1.26; 95% CI, 1.02–1.55; P=0.02), ASA class IV (OR, 2.62; 95% CI, 2.08–3.32; P=0.000), and angina within 30 days (OR, 3.14; 95% CI, 2.91–3.93; P=0.000). Female sex was associated with favorable postoperative other complication variables (OR, 0.86; 95% CI, 0.77–0.96; P=0.008).
After CEA for asymptomatic carotid stenosis in our study, we analyzed the effect of patients preoperative ability to perform ADLs (FNS) on postoperative complications using national registry in ≥19 000 patients. We found that those patients who are partially or completely dependent for ADLs have ≥3× higher risk of 30-day mortality when compared with independent patients. We also found that stroke rates were 3× and other complication (as defined in Table 2) rates were 2.5× higher in partially dependent patients when compared with independent patients. Whereas 30-day mortality and stroke and myocardial infarction rates in FNS-I cohort were 0.5%, 1%, and 0.6%, respectively, which modestly correlate to data present in the literature.2,10
To our knowledge, this is the first study to report association of FNS before surgery and postoperative complications after asymptomatic CEA. FNS/ADL is an important parameter for preoperative risk stratification and a critical outcome predictor that can be used for patient selection undergoing CEA in asymptomatic population. Preoperative inability to function independently is associated with worse postoperative outcome, which clearly falls outside guideline stipulated by the Society for Vascular Surgery and the American Heart Association for asymptomatic CEA.11,12 The Society for Vascular Surgery and the American Heart Association guidelines for asymptomatic state that patients with at least a 60% carotid artery stenosis should be considered for CEA only if the patients have a predicted risk of perioperative stroke/death of ≤3% and a minimum life expectancy of 3 to 5 years.
The ACAS study in the 1990s has shown benefit of CEA but had stringent exclusion criteria, which does not reflect real-world population, which has nearly twice comorbidities when compared with ACAS study population, leading to significantly higher perioperative complications.13 And, with better-defined risk factors and advent of medical management with new platelet inhibitor, statins, and lifestyle changes, some researchers have questioned the appropriateness of CEA in asymptomatic patients.4,5 A meta-analysis reported that average annual rates of ipsilateral stroke among patients receiving vascular disease medical intervention alone fell below patients who received CEA in ACAS.4 Similar evidence published by Spence et al5 found that decline in microemboli coincided with intensive medical therapy, which included the following: motivating patients to adhere to smoking cessation, exercise, medication, and a Mediterranean diet; increasing the dose of statins to the maximum tolerated dose, regardless of low-density lipoprotein levels (eg, 80 mg of atorvastatin or 40 mg of rosuvastatin); in patients already at their maximum tolerated dose of statins, adding ezetimibe; in those already taking maximum doses of statins and ezetimibe, adding niacin for patients who did not have diabetes mellitus or adding fibrates for patients with diabetes mellitus or those unable to take niacin or slow-release niacin because of flushing; ensuring that the patients were taking an angiotensin-converting enzyme inhibitor, or for those not able to take them because of cough or angioedema, ensuring that they were taking an angiotensin-receptor blocker; optimizing blood pressure control by individualizing therapy according to the renin/aldosterone profile; and in some patients with insulin resistance (defined by a high fasting insulin level with normal serum glucose level), metformin or pioglitazone was added before the onset of diabetes mellitus. In the above study, 17.6% had stroke, death, myocardial infarction, or CEA for symptoms before this intensive regimen versus 5.6% patients in intensive therapy.5 Although debate over surgical management versus medical management persists, we think that aggressive medical management is the acceptable option for high-risk patients at risk of postoperative stroke or mortality.
This study is valuable because it includes a robust sample of patients from ≥250 hospitals across the nation. The data were collected prospectively with rigorous attention to details and with standardized definitions for preoperative variables and complications. Because of the group of hospitals that contribute data to the ACS-NSQIP, this study can be thought of as representing more real-world outcomes of elective asymptomatic CEA. However, the data shown here are not without limitations. The variables that were analyzed were limited to those that could be captured by the ACS-NSQIP data set. Notably, the reason for functional dependence in patients undergoing CEA was not included, and the degree of stenosis and flow velocities in internal carotid artery/common carotid artery was not included in the study because of nonavailability of this information in the data set. Other adjuncts to the procedure, such as intraoperative shunt to internal carotid during anastomosis, were also not included. This applies as well to reasons for return to the operating room and the incidence of postoperative cranial nerve injury, a significant potential complication of CEA. Data beyond 30 days postoperatively were not available, so long-term follow-up was not possible with this cohort. The ACS-NSQIP is a quality improvement program designed to improve patient care quality by providing data for comparison among vascular and general surgery population. Thus, limitations because of the database design are expected when a specific surgical entity or issue is addressed. And also, small sample size of FNS-D group was the suspected reason behind nonsignificance of results in multivariate analysis, despite higher percentage of postoperative complications in this group with 3.15, 4.02, and 6.50 ORs for death, stroke, and other complications, respectively. However, this does not take away the potential contributions of this study, and inability to perform ADLs must be considered along with advancing age, black, history of COPD, angina within 30 days, and history of revascularization in decision to operate versus aggressive medical management of disease. In addition to providing robust national data, our study can be a valuable tool to help the patients and their families in making informed decision.
Our study provides real-world postoperative outcomes data of CEA in asymptomatic patients and demonstrates that preoperative inability to perform ADL is clearly associated with significantly higher mortality, stroke, myocardial infarction, and other complications after CEA in asymptomatic patients, which falls outside stipulated guideline provided by the Society for Vascular Surgery and the American Heart Association. This information will potentially help the concerned surgeons, patients, and family members make true informed decision about management of asymptomatic carotid artery disease. We think that ADL is a critical factor in predicting outcomes after CEA in asymptomatic patients, and aggressive medical management is the acceptable option for high-risk patients, such as partially dependent patients for prevention of future strokes.
We acknowledge the National Surgical Quality Improvement Project (NSQIP) surgical clinical nurse reviewers and administrators for their dedication to ensure the integrity of NSQIP data.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.113.003956/-/DC1.
- Received October 25, 2013.
- Revision received March 28, 2014.
- Accepted April 1, 2014.
- © 2014 American Heart Association, Inc.
- Bekelis K,
- Bakhoum SF,
- Desai A,
- Mackenzie TA,
- Goodney P,
- Labropoulos N
- Abbott AL
- 6.↵ACS NSQIP introduction: The American College of Surgeons National Surgical Quality Improvement Program. http://site.acsnsqip.org/. Accessed February 4, 2014.
- 7.↵ACS NSQIP variables: Variable Definition the American College of Surgeons National Surgical Quality Improvement Program. http://nsqip.healthsoftonline.com/lib/Documents/Ch_4_Variables_Definitions_062810.pdf. Accessed February 4, 2014.
- Calvillo-King L,
- Xuan L,
- Zhang S,
- Tuhrim S,
- Halm EA
- Rajamani K,
- Kennedy KF,
- Ruggiero NJ,
- Rosenfield K,
- Spertus J,
- Chaturvedi S
- Biller J,
- Feinberg WM,
- Castaldo JE,
- Whittemore AD,
- Harbaugh RE,
- Dempsey RJ,
- et al