- Email: [email protected]
Obesity is an Independent Risk Factor for Death and Cardiac Complications after Carotid Endarterectomy
Obesity is an Independent Risk Factor for Death and Cardiac Complications after Carotid Endarterectomy Rubie Sue Jackson, MD, MPH, Anton N Sidawy, MD, MPH, FACS, Richard L Amdur, PhD, Robyn A Macsata, MD, FACS The role of obesity as a risk factor after carotid endarterectomy is not well-described. We undertook a study of the association of obesity with 30-day outcomes after carotid endarterectomy. STUDY DESIGN: After obtaining Institutional Review Board approval, we retrospectively analyzed prospectively collected data from carotid endarterectomies in the 2005⫺2006 Veterans Affairs Surgical Quality Improvement Program database. The association of body mass index (BMI; calculated as kg/m2) on 30-day outcomes was assessed using multivariable logistic regression. RESULTS: From 3,706 carotid endarterectomies, we excluded 22 for missing BMI and 39 for emergency status; 3,645 carotid endarterectomies were analyzed. BMI was underweight (⬍18.5) in 1.6%, normal (18.5 to 24.9) in 31.0%, overweight (25.0 to 29.9) in 40.8%, class I obese (30.0 to 34.9) in 19.3%, class II obese (35.0 to 39.9) in 5.8%, and class III obese (ⱖ40) in 1.6%. On multivariable analysis, class II to III (odds ratio ⫽ 6.95; 95% CI, 1.89⫺25.58; p ⫽ 0.004) obesity was associated with death, and class II to III obesity was associated with cardiac complications (odds ratio ⫽ 3.68; 95% CI, 1.27⫺10.66; p ⫽ 0.02) compared with normal weight. CONCLUSIONS: Obesity is an independent risk factor for death and cardiac complications after carotid endarterectomy. Surgeons should consider this when evaluating the risks and benefits of carotid endarterectomy in obese patients. Carotid artery stenting was not assessed, and future studies are needed to examine its role in management of obese patients. (J Am Coll Surg 2012;214: 148–155. © 2012 by the American College of Surgeons) BACKGROUND:
countries3 and as much as 10% to 13% of deaths throughout Europe.4 Obesity has been identified as a risk factor for multiple vascular pathologies, including coronary artery disease,5,6 carotid atherosclerosis and stroke,7,8 peripheral arterial disease,9,10 and abdominal aortic aneurysm (AAA).11 In recent years, several studies have examined the association of obesity with outcomes after vascular surgery. We previously reported an association between obesity and increased mortality after open and endovascular AAA repair.12 Giles and colleagues also reported a relationship between morbid obesity and increased mortality after open and endovascular AAA repair13 and lower-extremity bypass.14 However, the influence of obesity on perioperative outcomes after carotid endarterectomy is not well-described. Understanding the relationship of obesity to outcomes after carotid endarterectomy is critical because appropriate patient selection depends on accurate assessment of the risks and benefits of carotid endarterectomy in an individual patient. In the 1990s, the North American Symptomatic Carotid Endarterectomy Trial (NASCET) and Asymp-
During the past 50 years, the age-adjusted prevalence of obesity in the US adult population has increased dramatically, from 13.4% in 1960 to 33.9% in 2008.1,2 Beyond the United States, obesity poses a worldwide public health challenge, accounting for 7% of all diseases in developed
Disclosure Information: Authors have nothing to disclose. Timothy J Eberlein, Editor-in-Chief, has nothing to disclose. This article represents the personal viewpoint of the authors and cannot be construed as a statement of official Department of Veterans Affairs or US Government policy. Abstract presented at the American College of Surgeons 94th Annual Clinical Congress, Surgical Forum, San Francisco, CA, October 2008. Received May 19, 2011; Revised October 16, 2011; Accepted October 18, 2011. From the Department of Surgery, Veterans Affairs Medical Center (Jackson, Sidawy, Amdur, Macsata), Department of Surgery, Georgetown University Hospital (Jackson, Amdur, Macsata), and Department of Surgery, George Washington University Medical Center (Sidawy, Amdur), Washington, DC. Correspondence address: Robyn A Macsata, MD, FACS, Department of Surgery, Veterans Affairs Medical Center, 50 Irving St NW, Washington, DC 20010. email: [email protected]
© 2012 by the American College of Surgeons Published by Elsevier Inc.
148
ISSN 1072-7515/12/$36.00 doi:10.1016/j.jamcollsurg.2011.10.017
Vol. 214, No. 2, February 2012
Jackson et al
Obesity and Carotid Endarterectomy
149
Table 1. Inclusion Criteria Abbreviations and Acronyms
AAA BMI OR SSI VASQIP VIF VTE
⫽ ⫽ ⫽ ⫽ ⫽
abdominal aortic aneurysm body mass index odds ratio surgical site infection Veterans Affairs Surgical Quality Improvement Project ⫽ variance inflation factor ⫽ venous thromboembolism
Code
CPT code 35301
ICD-9 diagnosis codes 433.1 433.10
tomatic Carotid Atherosclerosis Study (ACAS) established carotid endarterectomy over best medical management, as defined at the time of the study as the treatment of choice for ⱖ50% symptomatic carotid stenosis and ⱖ60% asymptomatic stenosis, respectively,15-17 For 50% to 69% symptomatic stenosis and ⱖ60% asymptomatic stenosis, carotid endarterectomy was found to afford a relative benefit only if perioperative rates of disabling stroke and death are maintained ⱕ2% to 3%.16,17 In a 2004 US multistate audit, however, the procedural risk of stroke and death for asymptomatic carotid stenosis was found to be 3.8%.18 This might partly reflect deficiencies in operative and postoperative care, but some of the excess complications doubtless result from inappropriate selection of high-risk patients for prophylactic carotid endarterectomy. For asymptomatic patients with advanced age or substantial comorbidities, the risks of carotid endarterectomy probably outweigh the benefits.19 However, which comorbidities increase the risk of death and complications after carotid endarterectomy is not well-described, and definitions of high-risk patients are neither evidence-based nor universally accepted.20 Understanding the influence of patient characteristics, such as obesity, on perioperative risk is crucial for appropriate patient selection and optimization for carotid endarterectomy. We undertook a study to examine the association of obesity with 30-day postoperative morbidity and mortality after carotid endarterectomy.
METHODS Database
We undertook a retrospective analysis of prospectively collected data from the Veterans Affairs Surgical Quality Improvement Program (VASQIP). VASQIP is an ongoing quality-management initiative designed to measure and enhance the quality of surgical care at VA hospitals. All 123 VA Medical Centers that perform major surgery participate in the VASQIP. Trained nurse reviewers prospectively collect preoperative, operative, and postoperative data from concurrent chart review, physician interview, and patient
433.11
Description
Thromboendarterectomy, with or without patch graft; carotid, vertebral, subclavian, by neck incision Occlusion and stenosis of precerebral arteries, carotid artery Occlusion and stenosis of precerebral arteries, carotid artery; without mention of cerebral infarction Occlusion and stenosis of precerebral arteries, carotid artery; with cerebral infarction
CPT, current procedural terminology.
follow-up. Patient outcomes are assessed 30 days postoperatively and deaths are verified against the VA Beneficiary Identification and Records Locator System death records.21 The Washington, DC VA Medical Center Institutional Review Board approved this study. Sample selection
We included all carotid endarterectomy procedures, defined by CPT code 35301 plus ICD-9 postoperative diagnosis code 433.1, 433.10, or 433.11 (Table 1), from January 1, 2005 through December 31, 2006. Because CPT code 35301 includes thromboendarterectomy of the vertebral and subclavian arteries, ICD-9 codes 433.1–11 were used to restrict the analysis to carotid procedures. We excluded procedures for which patient height or weight data were not available. Emergency procedures were excluded because it was believed that a separate analysis would be most appropriate for these cases, and there were too few cases to perform such an analysis. Carotid endarterectomies performed with simultaneous coronary artery bypass graft (CPT 33510, 33511, 33512, 33513, 33514, 33515, 33516, 33517, 33518, 33519, 33521, 33522, 33523, 33530, 33533, 33534, 33535, 33536, 33542, or 33545) were excluded. All patient identifiers were stripped from the database used for analysis by the VA SQIP database steward. Obesity definition
BMI (calculated as weight in kg divided by height in m2) was determined from each patient’s preoperative weight and height variables. BMI was categorized using the National Institutes of Health definition22 as underweight (BMI ⬍18.5), normal weight (BMI 18.5 to 24.9), overweight (BMI 25.0 to 29.9), class I obese (BMI 30.0 to
150
Jackson et al
Obesity and Carotid Endarterectomy
J Am Coll Surg
Table 2. Preoperative Demographics, Behaviors, and Comorbidities, by Weight Category (n ⫽ 3,645)
Demographics Age, y, mean (SD) Male sex, n (%) Behaviors, n (%) Current smoking Current ethanol use Comorbidity, n (%) CVA Without persistent deficit With persistent deficit Transient ischemic attack Diabetes Treated with oral medication Treated with insulin COPD Dyspnea With minimal exertion At rest Dependent Lower-extremity rest pain/ gangrene Dialysis-dependent Weight loss Congestive heart failure
Underweight (n ⴝ 59; 1.6%)
Normal weight (n ⴝ 1,131; 31.0%)
Overweight (n ⴝ 1,483; 40.8%)
Class I obese (n ⴝ 703; 19.3%)
Class II obese (n ⴝ 212; 5.8%)
Class III obese (n ⴝ 57; 1.6%)
p Value
69 (9.5) 57 (96.6)
69 (9.2) 1,118 (98.9)
69 (8.7) 1,463 (98.7)
67 (8.6) 692 (98.4)
65 (7.7) 205 (96.7)
65 (8.5) 57 (100)
⬍0.001 0.14
38 (64.4) 6 (10.2)
596 (52.7) 107 (9.5)
566 (37.4) 119 (8.1)
240 (34.1) 54 (7.7)
62 (29.3) 15 (7.1)
13 (22.8) 1 (1.8)
⬍0.001 0.27
11 (18.6) 12 (20.3) 16 (27.1)
107 (9.5) 181 (16.0) 331 (29.3)
129 (8.7) 260 (17.5) 462 (31.2)
63 (9.0) 88 (12.5) 206 (29.3)
28 (13.2) 29 (13.7) 67 (31.6)
4 (7.0) 7 (12.3) 15 (26.3)
0.05 0.05 0.81
5 (8.5) 4 (6.8) 25 (42.4)
131 (11.6) 61 (5.4) 232 (20.5)
300 (20.2) 149 (10.0) 225 (15.2)
198 (27.9) 117 (16.6) 132 (18.9)
58 (27.4) 60 (28.3) 40 (18.9)
19 (33.3) 18 (31.6) 11 (19.3)
⬍0.001 ⬍0.001
9 (15.6) 2 (3.5) 3 (5.1)
176 (16.0) 12 (1.1) 52 (4.6)
255 (17.5) 9 (0.6) 47 (3.2)
136 (19.8) 11 (1.6) 22 (3.1)
52 (24.8) 5 (2.4) 12 (5.7)
15 (26.8) 1 (1.8) 2 (3.5)
0.003 0.003 0.24
3 (5.1) 1 (1.7) 8 (13.6) 0 (0)
36 (3.2) 1 (0.1) 19 (1.7) 10 (0.9)
41 (2.8) 4 (0.3) 7 (0.5) 13 (0.9)
10 (1.4) 3 (0.4) 4 (0.6) 11 (1.6)
2 (0.9) 0 (0) 0 (0) 1 (0.5)
1 (1.8) 0 (0) 1 (1.8) 1 (1.8)
34.9), class II obese (BMI 35.0 to 39.9), or class III obese (BMI ⱖ40). Obesity overall was defined as BMI ⱖ 30.0. Postoperative outcomes
Complications are recorded in the VASQIP as binary variables or detailed using ICD-9-CM codes. The primary study outcomes were mortality (all-cause), cardiac complications, and stroke. Secondary outcomes were surgical site infection (SSI), respiratory complication, venous thromboembolism (VTE), and operative reintervention. Cardiac complications, SSIs, respiratory complications, and VTE were defined as composites of prespecified VASQIP outcomes. Cardiac complications included MI and cardiac arrest. Surgical site infections included both deep and superficial infections. Respiratory complications included failure to wean from mechanical ventilation for ⬎48 hours postoperatively or unplanned reintubation. VTE included deep venous thrombosis and pulmonary embolus.
0.09 0.17 ⬍0.001 0.52
Adjustment variables
The VASQIP contains presurgical and surgical risk adjustment variables. Variables missing ⱖ10% of data were not analyzed. Preincision operative characteristics (eg, American Society of Anesthesiologists class, anesthesia type) were analyzed, but other intraoperative characteristics (eg, operative time) were not analyzed because these variables could mediate rather than confound the relationship between obesity and postoperative outcomes. The remaining characteristics were compared among BMI categories and used as adjustment variables (Tables 2, 3). In the VASQIP, current smoking is defined as smoking within the year before admission. Current ethanol use is defined as ⱖ2 drinks daily during the 2 weeks before surgery. Diabetes is defined as diabetes not controlled by diet, and is further categorized according to treatment with oral medication vs insulin. Dependent status is defined as requiring assistance from another person for some or all activities of daily living. Lower-extremity rest pain/gangrene is defined as “severe, unrelenting pain aggravated by elevation and often preventing sleep” or “marked skin discolor-
Vol. 214, No. 2, February 2012
Jackson et al
Obesity and Carotid Endarterectomy
151
Table 3. Operative Characteristics and Preoperative Laboratory Values, by Weight Category (n ⫽ 3,645) Underweight (n ⴝ 59; 1.6%)
Procedure, n (%) ASA class 1⫺2 3 4 General anesthesia, n (%) Laboratory, mean (SD) Creatinine, mg/dL Hematocrit, %
1 (1.7) 46 (78.0 12 (20.3) 49 (83.1) 1.2 (0.6) 39.3 (4.8)
Normal weight (n ⴝ 1,131; 31.0%)
Overweight (n ⴝ 1,483; 40.8%)
Class I obese (n ⴝ 703; 19.3%)
Class II obese (n ⴝ 212; 5.8%)
53 (5.4) 899 (79.5) 179 (15.8) 1,010 (89.3)
66 (4.5) 1,185 (79.9) 232 (15.6) 1,328 (89.6)
14 (2.0) 579 (82.4) 110 (15.7) 632 (89.9)
3 (1.4) 170 (80.2) 39 (18.4) 190 (89.7)
1.2 (0.7) 40.5 (5.0)
1.2 (0.4) 41.2 (4.5)
1.3 (0.6) 41.1 (4.6)
1.2 (0.5) 40.5 (4.5)
Class III obese (n ⴝ 57; 1.6%)
3 (5.3) 39 (68.4) 15 (26.3) 53 (93.0) 1.3 (0.3) 41.0 (4.3)
p Value
0.09 0.62 0.12 ⬍0.001
ASA, American Society of Anesthesiologists.
ation and disruption indicative of death and decay of tissues in the extremities due to severe and prolonged ischemia,” respectively. Statistical analysis
Baseline characteristics were compared across BMI categories, using 2-tailed ANOVA for continuous variables and 2 statistic for binary and categorical variables. The same methods were used to compare baseline characteristics between patients who were excluded because of missing BMI and patients who were included in the analysis. Unadjusted outcomes, representing the 30-day incidence of each complication, were calculated within each BMI category and compared using the 2 statistic. Multivariable logistic regression was used to assess the independent association of BMI category with outcomes, after adjusting for patient characteristics. Laboratory values (eg, preoperative hematocrit and creatinine) were dichotomized based on exploratory data analysis because their univariate relationships to study outcomes were nonlinear across the range of values. The relationship of age to study outcomes was essentially linear across the range of values, and it was treated as a continuous variable. Each of the preoperative and operative characteristics were considered as adjustment variables, and backward stepwise regression with an exit criterion of p ⬍ 0.1 was used to select variables for final model inclusion. BMI categories were forced into the model. For all outcomes, the association of obesity (BMI ⱖ30) as well as the association of each of the BMI categories (defined here) were assessed, using normal weight as the reference. In the multivariable models, classes II and III obesity were considered together due to the small sample size of class III obese patients. For outcomes showing a significant association of obesity (BMI ⱖ30), a 2 test for trend was performed to assess the association of increasing BMI category on outcomes. For any composite outcomes (eg, cardiac complications, SSIs, respiratory compli-
cations, and VTE) showing an association of BMI category, multivariable logistic regression as described here was used to determine the association of BMI category on each of the component outcomes. To examine collinearity among variables, weighted regression using all baseline characteristics was used to calculate variance inflation factors (VIFs) and tolerances, and factor reduction was attempted for any variables displaying multicollinearity. VIF and its inverse, tolerance, are indices of the effect of collinearity among variables in a regression model. They measure the increase in the standard error of the parameter estimates that occurs due to collinearity.23 Multicollinearity was defined as VIF ⬎2 or tolerance ⬍0.50. A p value ⱕ0.05 was considered significant. All analyses were completed using Stata 11.0 (Stata Corp).
RESULTS A total of 3,706 carotid endarterectomy procedures were identified. Of these, we excluded 22 procedures for missing height and/or weight information and 39 emergency procedures. When baseline characteristics (listed in Tables 2, 3) were compared between patients with missing BMI vs nonmissing BMI, the only significant difference was in mean age (68.3 years vs 64.3 years; p ⫽ 0.02). No procedures with simultaneous coronary artery bypass graft were identified. The remaining 3,645 procedures were analyzed. Of these, the majority were overweight (40.8%), followed by normal weight (31.0%), class I obese (19.3%), class II obese (5.8%), class III obese (1.6%), and underweight (1.6%). Preoperative and operative characteristics by weight classification are displayed in Tables 2 and 3. Increasing BMI was associated with decreasing prevalence of smoking, ethanol use, and history of CVA and increasing prevalence of diabetes and dyspnea. The prevalence of COPD and preoperative weight loss were higher among underweight patients compared with other categories. Although there were
152
Jackson et al
Obesity and Carotid Endarterectomy
J Am Coll Surg
Table 4. Unadjusted Outcomes by Body Mass Index Category Overall event Underweight Normal weight Overweight rates (n ⴝ 59; (n ⴝ 1,131; (n ⴝ 1,483; (n ⴝ 3,645) 1.6%) 31.0%) 40.8%)
Death Cardiac complications Stroke Surgical site infection Respiratory complication Venous thromboembolism Return to the operating suite
0.8 (31) 1.4 (51) 2.1 (74) 0.9 (33) 1.8 (65) 0.2 (7) 4.8 (175)
1.7 (1) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 5.1 (3)
0.6 (7) 1.1 (12) 2.2 (25) 1.4 (16) 1.4 (16) 0.2 (2) 5.4 (61)
Class I obese Class II obese (n ⴝ 703; (n ⴝ 212; Class III obese 19.3%) 5.8%) (n ⴝ 57; 1.6%)
0.7 (10) 1.1 (16) 2.1 (31) 0.7 (10) 2.0 (29) 0.1 (1) 4.3 (64)
1.1 (8) 2.4 (17) 1.9 (13) 0.7 (5) 2.1 (15) 0 (0) 4.7 (33)
2.4 (5) 1.9 (4) 2.4 (5) 0 (0) 1.9 (4) 1.4 (3) 3.8 (8)
0 (0) 3.5 (2) 0 (0) 3.5 (2) 1.8 (1) 1.8 (1) 10.5 (6)
Values shown are incidence proportion (number of events). p Values are not shown because the number of events was too small to make statistically meaningful comparisons.
statistically significant differences in preoperative hematocrit, the mean values for each BMI category were within the accepted normal range. Overall incidences of study outcomes are displayed in Table 4. For the unadjusted analysis, statistical significance is not reported because the number of events was too small to make meaningful comparisons. The incidence of death was highest in underweight and class II to III obese patients. The observed incidence of cardiac complications increased with increasing BMI category, and the incidence of VTE was higher among class II and III obese patients compared with other weight categories. The incidence of SSI was highest in class III obese patients. On multivariable analysis, obesity (BMI ⱖ30, class I to III) was significantly associated with a 3-fold higher risk of death (odds ratio [OR] ⫽ 3.42; 95% CI, 1.45⫺8.10; p ⫽ 0.005) and a 2-fold higher risk of cardiac complications (OR ⫽ 2.38; 95% CI, 1.23⫺4.62; p ⫽ 0.01) compared with normal weight. When BMI categories were analyzed separately (Table 5), only class II to III (OR ⫽ 6.95; 95% CI, 1.89⫺25.58; p ⫽ 0.004) obesity was associated with increased risk of death. Only class II to III obesity was significantly associated with increased risk of cardiac complications (OR ⫽ 3.68; 95% CI, 1.27⫺10.66; p ⫽ 0.02). A test for trend showed that increasing BMI category was associated with increasing risk of both death (p ⫽ 0.03) and cardiac complications (p ⬍ 0.001). Increasing age and lower-extremity rest pain/gangrene were risk factors for both death and cardiac complications. Preoperative creatinine ⬎2 mg/dL and current smoking were also associated with increased risk of death, and previous stroke with or without residual deficit was associated with increased risk of cardiac complications. Thirty-day mortality was substantially higher among patients who experienced cardiac complications vs those without cardiac complications (17.0% vs 0.6%; p ⬍ 0.001). However, even after excluding patients with cardiac complications, obesity remained independently associated with an elevated risk of 30-day
mortality (OR ⫽ 7.30; 95% CI, 2.22⫺24.00; p ⫽ 0.001). When the component outcomes of cardiac complications were analyzed, BMI was not associated with the risk of MI, but there was significantly increased adjusted risk of cardiac arrest among obese patients (BMI ⱖ30) compared with nonobese patients (OR ⫽ 3.42; 95% CI, 1.45⫺8.10; p ⫽ Table 5. Adjusted Risk of Death and Cardiac Complications, Showing All Variables Selected (Using Stepwise Regression) for Final Model Inclusion BMI complications
Death Underweight Normal weight Overweight Class I obesity Class II–III obesity* Age, 5-year increase Current smoking Lower-extremity rest pain/gangrene Preoperative creatinine ⬎2 mg/dL Cardiac complications Underweight Normal weight Overweight Class I obesity Class II–III* obesity Age, 5-year increase Lower-extremity rest pain/gangrene Previous stroke, no residual deficit Previous stroke with residual deficit
Odds ratio (95% CI)
p Value
2.98 (0.32⫺27.44) Reference 0.89 (0.28⫺2.86) 2.34 (0.72⫺7.66) 6.95 (1.89⫺25.58) 1.32 (1.01⫺1.72) 2.50 (1.00⫺6.26)
0.33 Reference 0.85 0.16 0.004 0.04 0.03
5.99 (1.64⫺21.88)
0.007
3.61 (1.16⫺11.23)
0.05
Undefined† Reference 0.98 (0.42⫺2.27) 1.95 (0.79⫺4.80) 3.68 (1.27⫺10.66) 1.29 (1.07⫺1.57)
Undefined† Reference 0.97 0.15 0.02 0.009
3.85 (1.12⫺13.23)
0.03
2.39 (1.00⫺5.69)
0.05
2.25 (1.05⫺4.82)
0.04
*Pooled analysis of class II and III obese patients. (Analysis was pooled for both outcomes due to the small number of class III obese patients.) † No events. BMI, body mass index.
Vol. 214, No. 2, February 2012
0.005). The Hosmer-Lemeshow test statistic p value for both death (0.44) and cardiac complications (0.77) was not significant, consistent with good model fit. BMI category was not significantly associated with the risk of stroke, SSI, respiratory complications, VTE, or return to the operating suite. Analysis of collinearity among variables included in the final stepwise regression models demonstrated a maximum VIF of 1.79, indicating low multicollinearity.
DISCUSSION This analysis found that obesity was independently associated with increased risk of death and cardiac complications after carotid endarterectomy. We did not find evidence of an association between BMI and risk of stroke, SSI, respiratory complication, VTE, or return to the operating suite. Very few underweight patients were included, and the association of underweight with study outcomes could not be assessed with statistical precision. Several previous studies have reported an increase in the risk of perioperative mortality among obese patients undergoing vascular surgery. We previously demonstrated a relationship between obesity and increased 30-day mortality after open and endovascular AAA repair among patients represented in the VASQIP.12 A similar analysis of 5,455 patients in the American College of Surgeons National Surgical Quality Improvement Program showed that class III obesity was a risk factor for 30-day mortality after open AAA repair but not after endovascular AAA repair.13 In contrast, a few studies have found evidence supporting an obesity paradox. This concept suggests that although obesity is a risk factor for development of cardiovascular and peripheral vascular disease, mild obesity can confer a survival benefit after certain cardiovascular and peripheral vascular surgical procedures.24 An analysis by Davenport and colleagues of data collected from 7,543 patients who underwent a variety of vascular surgery procedures found a reverse-J shape relationship between BMI and 30-day mortality, with the highest risk in underweight and normalweight patients and the lowest risk in class I obese patients.25 Similarly, an analysis by Giles and colleagues of 7,595 lower-extremity bypasses in the American College of Surgeons National Surgical Quality Improvement Program found a U-shaped relationship between BMI and 30-day mortality, with the lowest mortality in class I and II obese patients and the highest mortality in class III obese patients.14 In both studies, the protective association of mild obesity persisted on multivariable analysis. These reports of a protective association of mild obesity on perioperative mortality contrast with our finding of increasing risk of 30-day mortality with increasing BMI
Jackson et al
Obesity and Carotid Endarterectomy
153
category. Differences in study design might account for these discrepant findings. Most noticeably, we examined only carotid endarterectomies, and cerebrovascular procedures constituted only 17.4% of the procedures studied by Davenport and colleagues25 and none of the procedures examined by Giles and colleagues,14 making direct comparison of our results with theirs rather tenuous. Additionally, the proportion of women in the reports by Davenport and colleagues (40.2%)25 and Giles and colleagues (20.8%)14 were substantially higher than in our study (1.4%), based on the patient population at VA medical centers. If the relationship between BMI and surgical outcomes differs between men and women, this might contribute to the differences between our findings and those of the 2 studies discussed here. Finally, although we were not able to detect a statistically significant association between underweight and adverse postoperative outcomes, this might have been due to the small number of underweight patients, reducing the power to detect such an association. The increased risk of perioperative mortality in obese patients is partially attributable to increased postoperative cardiac complications in this population. Obesity has long been recognized as an independent predictor of cardiovascular morbidity and mortality.5-7 Research suggests that visceral adipose tissue secretes cytokines and chemokines that attract macrophages, the products of which lead to a systemic proinflammatory, proatherosclerotic state associated with development of coronary artery disease.26 Additionally, obesity is associated with subclinical abnormalities in left atrial and ventricular structure and function, independent of its association with age, blood pressure, or heart rate.27 Underlying coronary disease and myocardial dysfunction likely contribute to the association between obesity and cardiac complications after carotid endarterectomy. Our data suggest that mortality risk after carotid endarterectomy is elevated even among patients without documented cardiac complications. Unrecognized postoperative cardiac complications might partly account for this. Additionally, worse access to preventive health care services among obese patients, compared with normal-weight patients, might be contributory. Several studies suggest that obese patients are less likely than nonobese patients to undergo health screening examinations.28,29 As a result, obese patients might suffer from a higher prevalence of undiagnosed and untreated comorbidities, which likely contribute to the association between obesity and mortality after carotid endarterectomy observed in this study. Among vascular surgery patients, one of the most commonly noted morbidities associated with obesity is SSI,12,14,25 and it is therefore relevant that our analysis did
154
Jackson et al
Obesity and Carotid Endarterectomy
not uncover such a relationship after carotid endarterectomy. Although we noted an overall 0.9% infection rate, infection rates in previous studies that identified obesity as a risk factor for SSI were 3.6% (AAA repair),12 6.7% (miscellaneous vascular operations),25 and 10.1% (lowerextremity bypass).14 Lack of association between obesity and SSI in our study is probably due to a lower incidence of wound infection overall among patients undergoing carotid endarterectomy, reducing the statistical power to detect such a relationship. Our analysis of baseline characteristics showed a strong association between increasing weight category and decreasing prevalence of smoking. Previous studies have similarly shown a negative correlation between obesity and current smoking,30 as well as an association between smoking cessation and weight gain.31 The association between underweight and COPD is most likely a result of weight loss due to this comorbidity. The VASQIP was designed as a quality-improvement tool to compare individual institutions’ postoperative outcomes with national benchmarks. As such, there are limitations to its use as a research database. For the years studied, the VASQIP did not contain information on hypertension or dyslipidemia, and our data set did not contain information on coronary intervention or congestive heart failure. Analyses were therefore not adjusted for these comorbidities. We were able, however, to adjust for many other important comorbidities, such as diabetes. Additionally, our study contained mostly males, and additional study will be needed to confirm the relationship between obesity and outcomes after carotid endarterectomy in women. Finally, although some studies have suggested that waist-to-hip ratio or waistto-thigh ratio can correlate more closely with cardiovascular risk and mortality than does BMI,32,33 we were limited to BMI as the measure of obesity because other anthropometric measurements were not included in the database.
CONCLUSIONS This report identifies obesity as an independent risk factor for death and cardiac complications after carotid endarterectomy. When considering carotid endarterectomy in obese patients, surgeons should assess the risks and benefits of the procedure, considering the increased risk of death and cardiac complications in obese patients. If the decision is made to proceed with carotid endarterectomy, surgeons should undertake additional risk assessment and modification strategies, including more rigorous preoperative cardiac evaluation and consideration for postoperative telemetry or ICU monitor-
J Am Coll Surg
ing for early recognition and treatment of cardiac or other complications. Author Contributions Study conception and design: Sidawy, Amdur, Macsata Acquisition of data: Sidawy, Amdur Analysis and interpretation of data: Jackson, Amdur Drafting of manuscript: Jackson Critical revision: Sidawy, Amdur, Macsata Acknowledgment: The authors would like to acknowledge the VA Surgical Quality Data Use Group (SQDUG) for its role as scientific advisors and for the critical review of data use and analysis presented in this article.
REFERENCES 1. Flegal KM, Carroll MD, Ogden CL, Johnson CL. Prevalence and trends in obesity among US adults, 1999–2000. JAMA 2002;288:1723–1727. 2. Flegal KM, Carroll MD, Ogden CL, Curtin LR. Prevalence and trends in obesity among US adults, 1999–2008. JAMA 2010; 303:235–241. 3. World Global Infobase Team. The SuRF Report 2. Surveillance and chronic disease risk factors: country-level data and comparative estimates. Geneva: WHO; 2005. 4. WHO. WHO European Ministerial Conference on Counteracting Obesity, 15⫺17 November 2006, Istanbul, Turkey. Conference report. Geneva: WHO; 2007. 5. Wilson PW, D’Agostino RB, Sullivan L, et al. Overweight and obesity as determinants of cardiovascular risk: the Framingham experience. Arch Intern Med 2002;162:1867–1872. 6. Krauss RM, Winston M. Obesity: impact on cardiovascular disease. Circulation 1998;98:1472. 7. Hubert HB, Feinleib M, McNamara PM, Castelli WP. Obesity as an independent risk factor for cardiovascular disease: a 26year follow-up of participants in the Framingham Heart Study. Circulation 1983;67:968–977. 8. Yatsuya H, Yamagishi K, North KE, et al for the ARIC Investigators. Associations of obesity measures with subtypes of ischemic stroke in the ARIC study. J Epidemiol 2010;20:347–354. 9. Planas A, Clara A, Pou JM, et al. Relationship of obesity distribution and peripheral arterial occlusive disease in elderly men. Int J Obes Relat Metab Disorder 2001;25:1068–1070. 10. Vogt MT, Cauley JA, Kuller LH, Hulley SB. Prevalence and correlates of lower extremity arterial disease in elderly women. Am J Epidemiol 1998;137:559–568. 11. Kent KC, Zwolak RM, Egorova NN, et al. Analysis of risk factors for abdominal aortic aneurysm in a cohort of more than 3 million individuals. J Vasc Surg 2010;52:539–548. 12. Johnson ON, Sidawy AN, Scanlon JM, et al. Impact of obesity on outcomes after open surgical and endovascular abdominal aortic aneurysm repair. J Am Coll Surg 2010;210:166–177. 13. Giles KA, Wyers MC, Pomposelli FB, et al. The impact of body mass index on perioperative outcomes of open and endovascular abdominal aortic aneurysm repair from the National Surgical Quality Improvement Program, 2005–2007. J Vasc Surg 2010;52: 1471–1477.
Vol. 214, No. 2, February 2012
14. Giles KA, Hamdan AD, Pomposelli FB, et al. Body mass index: surgical site infections and mortality after lower extremity bypass from the National Surgical Quality Improvement Program 2005⫺2007. Ann Vasc Surg 2010;24:48–56. 15. The North American Symptomatic Carotid Endarterectomy Trial (NASCET) Investigators. Clinical alert: benefit of carotid endarterectomy for patients with high-grade stenosis of the internal carotid artery. Stroke 1991;22:816–817. 16. Barnett HJM, Taylor W, Eliasqiw M, et al. Benefit of carotid endarterectomy in patients with symptomatic moderate or severe stenosis. N Engl J Med 1998;339:1415–1425. 17. Executive Committee for the Asymptomatic Carotid Atherosclerosis Study. Endarterectomy for asymptomatic carotid artery stenosis. JAMA 1995;273:1421–1428. 18. Kresowik TF, Bratzler DW, Kresowik RA. Multistate improvement in process and outcomes of carotid endarterectomy. J Vasc Surg 2004;39:372–380. 19. Nagaki T, Sato K, Yoshida T, Yoshimoto Y. Benefits of carotid endarterectomy for symptomatic and asymptomatic severe carotid artery stenosis: a Markov model based on data from randomized controlled trials. J Neurosurg 2009;111:970–977. 20. Jim J, Sicard GA. Who is unfit for carotid endarterectomy? Perspect Vasc Surg Endovasc Ther 2010;22:40–45. 21. Khuri SF, Daley J, Henderson W, et al. The Department of Veterans Affairs’ NSQIP: the first national, validated, outcomebased, risk-adjusted, and peer-controlled program for the measurement and enhancement of the quality of surgical care. Ann Surg 1998;228:491–507. 22. Clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults: executive summary. Expert Panel on the Identification, Evaluation, and Treatment of Overweight in Adults. Am J Clin Nutr 1998;68:899–917.
Jackson et al
Obesity and Carotid Endarterectomy
155
23. Allison P. Logistic regression using the SAS system: Theory and application. Cary, NC: SAS Institute, Inc.; 1999. 24. Amundson DE, Djurkovic S, Matmiyoff GN. The obesity paradox. Crit Care Clin 2010;26:538–596. 25. Davenport DL, Xenos ES, Hosokawa P, et al. The influence of body mass index obesity status on vascular surgery 30-day morbidity and mortality. J Vasc Surg 2009;49:140–147. 26. Ohman MK, Wright AP, Wickenheiser KJ, et al. Visceral adipose tissue and atherosclerosis. Curr Vasc Pharmacol 2009;7: 169–179. 27. Tumukulu MM, Etikan I, Kisacik B, Kayikcioglu M. Effects of obesity on left ventricular structure and myocardial systolic function: assessment by tissue Doppler imaging and strain/ strain rate imaging. Echocardiography 2007;8:802–809. 28. Maruthur NM, Bolen S, Francati FL, Clark JM. Obesity and mammography: a systematic review and meta-analysis. J Gen Intern Med 2009;24:665–677. 29. Amy NK, Aalborg A, Lyons P, Keranen L. Barriers to routine gynecological cancer screening for white and African-American obese women. Int J Obes 2006;30:147–155. 30. Healton CG, Vallone D, McCausland KL, et al. Smoking, obesity, and their co-occurrence in the United States: cross sectional analysis. BMJ 2006;333:25–26. 31. Flegal KM. The influence of smoking cessation on the prevalence of overweight in the United States. N Engl J Med 1995; 333:1165–1170. 32. Reis JP, Araneta MR, Wingard DL, et al. Overall obesity and abdominal adiposity as predictors of mortality in US white and black adults. Ann Epidemiol 2009;19:134–142. 33. Bodenant M, Kuulasmaa K, Wagner A, et al. Measures of abdominal adiposity and the risk of stroke: The MOnica Risk, Genetics, Archiving and Monograph (MORGAM) Study. Stroke 2011;42:2872–2877.
countries3 and as much as 10% to 13% of deaths throughout Europe.4 Obesity has been identified as a risk factor for multiple vascular pathologies, including coronary artery disease,5,6 carotid atherosclerosis and stroke,7,8 peripheral arterial disease,9,10 and abdominal aortic aneurysm (AAA).11 In recent years, several studies have examined the association of obesity with outcomes after vascular surgery. We previously reported an association between obesity and increased mortality after open and endovascular AAA repair.12 Giles and colleagues also reported a relationship between morbid obesity and increased mortality after open and endovascular AAA repair13 and lower-extremity bypass.14 However, the influence of obesity on perioperative outcomes after carotid endarterectomy is not well-described. Understanding the relationship of obesity to outcomes after carotid endarterectomy is critical because appropriate patient selection depends on accurate assessment of the risks and benefits of carotid endarterectomy in an individual patient. In the 1990s, the North American Symptomatic Carotid Endarterectomy Trial (NASCET) and Asymp-
During the past 50 years, the age-adjusted prevalence of obesity in the US adult population has increased dramatically, from 13.4% in 1960 to 33.9% in 2008.1,2 Beyond the United States, obesity poses a worldwide public health challenge, accounting for 7% of all diseases in developed
Disclosure Information: Authors have nothing to disclose. Timothy J Eberlein, Editor-in-Chief, has nothing to disclose. This article represents the personal viewpoint of the authors and cannot be construed as a statement of official Department of Veterans Affairs or US Government policy. Abstract presented at the American College of Surgeons 94th Annual Clinical Congress, Surgical Forum, San Francisco, CA, October 2008. Received May 19, 2011; Revised October 16, 2011; Accepted October 18, 2011. From the Department of Surgery, Veterans Affairs Medical Center (Jackson, Sidawy, Amdur, Macsata), Department of Surgery, Georgetown University Hospital (Jackson, Amdur, Macsata), and Department of Surgery, George Washington University Medical Center (Sidawy, Amdur), Washington, DC. Correspondence address: Robyn A Macsata, MD, FACS, Department of Surgery, Veterans Affairs Medical Center, 50 Irving St NW, Washington, DC 20010. email: [email protected]
© 2012 by the American College of Surgeons Published by Elsevier Inc.
148
ISSN 1072-7515/12/$36.00 doi:10.1016/j.jamcollsurg.2011.10.017
Vol. 214, No. 2, February 2012
Jackson et al
Obesity and Carotid Endarterectomy
149
Table 1. Inclusion Criteria Abbreviations and Acronyms
AAA BMI OR SSI VASQIP VIF VTE
⫽ ⫽ ⫽ ⫽ ⫽
abdominal aortic aneurysm body mass index odds ratio surgical site infection Veterans Affairs Surgical Quality Improvement Project ⫽ variance inflation factor ⫽ venous thromboembolism
Code
CPT code 35301
ICD-9 diagnosis codes 433.1 433.10
tomatic Carotid Atherosclerosis Study (ACAS) established carotid endarterectomy over best medical management, as defined at the time of the study as the treatment of choice for ⱖ50% symptomatic carotid stenosis and ⱖ60% asymptomatic stenosis, respectively,15-17 For 50% to 69% symptomatic stenosis and ⱖ60% asymptomatic stenosis, carotid endarterectomy was found to afford a relative benefit only if perioperative rates of disabling stroke and death are maintained ⱕ2% to 3%.16,17 In a 2004 US multistate audit, however, the procedural risk of stroke and death for asymptomatic carotid stenosis was found to be 3.8%.18 This might partly reflect deficiencies in operative and postoperative care, but some of the excess complications doubtless result from inappropriate selection of high-risk patients for prophylactic carotid endarterectomy. For asymptomatic patients with advanced age or substantial comorbidities, the risks of carotid endarterectomy probably outweigh the benefits.19 However, which comorbidities increase the risk of death and complications after carotid endarterectomy is not well-described, and definitions of high-risk patients are neither evidence-based nor universally accepted.20 Understanding the influence of patient characteristics, such as obesity, on perioperative risk is crucial for appropriate patient selection and optimization for carotid endarterectomy. We undertook a study to examine the association of obesity with 30-day postoperative morbidity and mortality after carotid endarterectomy.
METHODS Database
We undertook a retrospective analysis of prospectively collected data from the Veterans Affairs Surgical Quality Improvement Program (VASQIP). VASQIP is an ongoing quality-management initiative designed to measure and enhance the quality of surgical care at VA hospitals. All 123 VA Medical Centers that perform major surgery participate in the VASQIP. Trained nurse reviewers prospectively collect preoperative, operative, and postoperative data from concurrent chart review, physician interview, and patient
433.11
Description
Thromboendarterectomy, with or without patch graft; carotid, vertebral, subclavian, by neck incision Occlusion and stenosis of precerebral arteries, carotid artery Occlusion and stenosis of precerebral arteries, carotid artery; without mention of cerebral infarction Occlusion and stenosis of precerebral arteries, carotid artery; with cerebral infarction
CPT, current procedural terminology.
follow-up. Patient outcomes are assessed 30 days postoperatively and deaths are verified against the VA Beneficiary Identification and Records Locator System death records.21 The Washington, DC VA Medical Center Institutional Review Board approved this study. Sample selection
We included all carotid endarterectomy procedures, defined by CPT code 35301 plus ICD-9 postoperative diagnosis code 433.1, 433.10, or 433.11 (Table 1), from January 1, 2005 through December 31, 2006. Because CPT code 35301 includes thromboendarterectomy of the vertebral and subclavian arteries, ICD-9 codes 433.1–11 were used to restrict the analysis to carotid procedures. We excluded procedures for which patient height or weight data were not available. Emergency procedures were excluded because it was believed that a separate analysis would be most appropriate for these cases, and there were too few cases to perform such an analysis. Carotid endarterectomies performed with simultaneous coronary artery bypass graft (CPT 33510, 33511, 33512, 33513, 33514, 33515, 33516, 33517, 33518, 33519, 33521, 33522, 33523, 33530, 33533, 33534, 33535, 33536, 33542, or 33545) were excluded. All patient identifiers were stripped from the database used for analysis by the VA SQIP database steward. Obesity definition
BMI (calculated as weight in kg divided by height in m2) was determined from each patient’s preoperative weight and height variables. BMI was categorized using the National Institutes of Health definition22 as underweight (BMI ⬍18.5), normal weight (BMI 18.5 to 24.9), overweight (BMI 25.0 to 29.9), class I obese (BMI 30.0 to
150
Jackson et al
Obesity and Carotid Endarterectomy
J Am Coll Surg
Table 2. Preoperative Demographics, Behaviors, and Comorbidities, by Weight Category (n ⫽ 3,645)
Demographics Age, y, mean (SD) Male sex, n (%) Behaviors, n (%) Current smoking Current ethanol use Comorbidity, n (%) CVA Without persistent deficit With persistent deficit Transient ischemic attack Diabetes Treated with oral medication Treated with insulin COPD Dyspnea With minimal exertion At rest Dependent Lower-extremity rest pain/ gangrene Dialysis-dependent Weight loss Congestive heart failure
Underweight (n ⴝ 59; 1.6%)
Normal weight (n ⴝ 1,131; 31.0%)
Overweight (n ⴝ 1,483; 40.8%)
Class I obese (n ⴝ 703; 19.3%)
Class II obese (n ⴝ 212; 5.8%)
Class III obese (n ⴝ 57; 1.6%)
p Value
69 (9.5) 57 (96.6)
69 (9.2) 1,118 (98.9)
69 (8.7) 1,463 (98.7)
67 (8.6) 692 (98.4)
65 (7.7) 205 (96.7)
65 (8.5) 57 (100)
⬍0.001 0.14
38 (64.4) 6 (10.2)
596 (52.7) 107 (9.5)
566 (37.4) 119 (8.1)
240 (34.1) 54 (7.7)
62 (29.3) 15 (7.1)
13 (22.8) 1 (1.8)
⬍0.001 0.27
11 (18.6) 12 (20.3) 16 (27.1)
107 (9.5) 181 (16.0) 331 (29.3)
129 (8.7) 260 (17.5) 462 (31.2)
63 (9.0) 88 (12.5) 206 (29.3)
28 (13.2) 29 (13.7) 67 (31.6)
4 (7.0) 7 (12.3) 15 (26.3)
0.05 0.05 0.81
5 (8.5) 4 (6.8) 25 (42.4)
131 (11.6) 61 (5.4) 232 (20.5)
300 (20.2) 149 (10.0) 225 (15.2)
198 (27.9) 117 (16.6) 132 (18.9)
58 (27.4) 60 (28.3) 40 (18.9)
19 (33.3) 18 (31.6) 11 (19.3)
⬍0.001 ⬍0.001
9 (15.6) 2 (3.5) 3 (5.1)
176 (16.0) 12 (1.1) 52 (4.6)
255 (17.5) 9 (0.6) 47 (3.2)
136 (19.8) 11 (1.6) 22 (3.1)
52 (24.8) 5 (2.4) 12 (5.7)
15 (26.8) 1 (1.8) 2 (3.5)
0.003 0.003 0.24
3 (5.1) 1 (1.7) 8 (13.6) 0 (0)
36 (3.2) 1 (0.1) 19 (1.7) 10 (0.9)
41 (2.8) 4 (0.3) 7 (0.5) 13 (0.9)
10 (1.4) 3 (0.4) 4 (0.6) 11 (1.6)
2 (0.9) 0 (0) 0 (0) 1 (0.5)
1 (1.8) 0 (0) 1 (1.8) 1 (1.8)
34.9), class II obese (BMI 35.0 to 39.9), or class III obese (BMI ⱖ40). Obesity overall was defined as BMI ⱖ 30.0. Postoperative outcomes
Complications are recorded in the VASQIP as binary variables or detailed using ICD-9-CM codes. The primary study outcomes were mortality (all-cause), cardiac complications, and stroke. Secondary outcomes were surgical site infection (SSI), respiratory complication, venous thromboembolism (VTE), and operative reintervention. Cardiac complications, SSIs, respiratory complications, and VTE were defined as composites of prespecified VASQIP outcomes. Cardiac complications included MI and cardiac arrest. Surgical site infections included both deep and superficial infections. Respiratory complications included failure to wean from mechanical ventilation for ⬎48 hours postoperatively or unplanned reintubation. VTE included deep venous thrombosis and pulmonary embolus.
0.09 0.17 ⬍0.001 0.52
Adjustment variables
The VASQIP contains presurgical and surgical risk adjustment variables. Variables missing ⱖ10% of data were not analyzed. Preincision operative characteristics (eg, American Society of Anesthesiologists class, anesthesia type) were analyzed, but other intraoperative characteristics (eg, operative time) were not analyzed because these variables could mediate rather than confound the relationship between obesity and postoperative outcomes. The remaining characteristics were compared among BMI categories and used as adjustment variables (Tables 2, 3). In the VASQIP, current smoking is defined as smoking within the year before admission. Current ethanol use is defined as ⱖ2 drinks daily during the 2 weeks before surgery. Diabetes is defined as diabetes not controlled by diet, and is further categorized according to treatment with oral medication vs insulin. Dependent status is defined as requiring assistance from another person for some or all activities of daily living. Lower-extremity rest pain/gangrene is defined as “severe, unrelenting pain aggravated by elevation and often preventing sleep” or “marked skin discolor-
Vol. 214, No. 2, February 2012
Jackson et al
Obesity and Carotid Endarterectomy
151
Table 3. Operative Characteristics and Preoperative Laboratory Values, by Weight Category (n ⫽ 3,645) Underweight (n ⴝ 59; 1.6%)
Procedure, n (%) ASA class 1⫺2 3 4 General anesthesia, n (%) Laboratory, mean (SD) Creatinine, mg/dL Hematocrit, %
1 (1.7) 46 (78.0 12 (20.3) 49 (83.1) 1.2 (0.6) 39.3 (4.8)
Normal weight (n ⴝ 1,131; 31.0%)
Overweight (n ⴝ 1,483; 40.8%)
Class I obese (n ⴝ 703; 19.3%)
Class II obese (n ⴝ 212; 5.8%)
53 (5.4) 899 (79.5) 179 (15.8) 1,010 (89.3)
66 (4.5) 1,185 (79.9) 232 (15.6) 1,328 (89.6)
14 (2.0) 579 (82.4) 110 (15.7) 632 (89.9)
3 (1.4) 170 (80.2) 39 (18.4) 190 (89.7)
1.2 (0.7) 40.5 (5.0)
1.2 (0.4) 41.2 (4.5)
1.3 (0.6) 41.1 (4.6)
1.2 (0.5) 40.5 (4.5)
Class III obese (n ⴝ 57; 1.6%)
3 (5.3) 39 (68.4) 15 (26.3) 53 (93.0) 1.3 (0.3) 41.0 (4.3)
p Value
0.09 0.62 0.12 ⬍0.001
ASA, American Society of Anesthesiologists.
ation and disruption indicative of death and decay of tissues in the extremities due to severe and prolonged ischemia,” respectively. Statistical analysis
Baseline characteristics were compared across BMI categories, using 2-tailed ANOVA for continuous variables and 2 statistic for binary and categorical variables. The same methods were used to compare baseline characteristics between patients who were excluded because of missing BMI and patients who were included in the analysis. Unadjusted outcomes, representing the 30-day incidence of each complication, were calculated within each BMI category and compared using the 2 statistic. Multivariable logistic regression was used to assess the independent association of BMI category with outcomes, after adjusting for patient characteristics. Laboratory values (eg, preoperative hematocrit and creatinine) were dichotomized based on exploratory data analysis because their univariate relationships to study outcomes were nonlinear across the range of values. The relationship of age to study outcomes was essentially linear across the range of values, and it was treated as a continuous variable. Each of the preoperative and operative characteristics were considered as adjustment variables, and backward stepwise regression with an exit criterion of p ⬍ 0.1 was used to select variables for final model inclusion. BMI categories were forced into the model. For all outcomes, the association of obesity (BMI ⱖ30) as well as the association of each of the BMI categories (defined here) were assessed, using normal weight as the reference. In the multivariable models, classes II and III obesity were considered together due to the small sample size of class III obese patients. For outcomes showing a significant association of obesity (BMI ⱖ30), a 2 test for trend was performed to assess the association of increasing BMI category on outcomes. For any composite outcomes (eg, cardiac complications, SSIs, respiratory compli-
cations, and VTE) showing an association of BMI category, multivariable logistic regression as described here was used to determine the association of BMI category on each of the component outcomes. To examine collinearity among variables, weighted regression using all baseline characteristics was used to calculate variance inflation factors (VIFs) and tolerances, and factor reduction was attempted for any variables displaying multicollinearity. VIF and its inverse, tolerance, are indices of the effect of collinearity among variables in a regression model. They measure the increase in the standard error of the parameter estimates that occurs due to collinearity.23 Multicollinearity was defined as VIF ⬎2 or tolerance ⬍0.50. A p value ⱕ0.05 was considered significant. All analyses were completed using Stata 11.0 (Stata Corp).
RESULTS A total of 3,706 carotid endarterectomy procedures were identified. Of these, we excluded 22 procedures for missing height and/or weight information and 39 emergency procedures. When baseline characteristics (listed in Tables 2, 3) were compared between patients with missing BMI vs nonmissing BMI, the only significant difference was in mean age (68.3 years vs 64.3 years; p ⫽ 0.02). No procedures with simultaneous coronary artery bypass graft were identified. The remaining 3,645 procedures were analyzed. Of these, the majority were overweight (40.8%), followed by normal weight (31.0%), class I obese (19.3%), class II obese (5.8%), class III obese (1.6%), and underweight (1.6%). Preoperative and operative characteristics by weight classification are displayed in Tables 2 and 3. Increasing BMI was associated with decreasing prevalence of smoking, ethanol use, and history of CVA and increasing prevalence of diabetes and dyspnea. The prevalence of COPD and preoperative weight loss were higher among underweight patients compared with other categories. Although there were
152
Jackson et al
Obesity and Carotid Endarterectomy
J Am Coll Surg
Table 4. Unadjusted Outcomes by Body Mass Index Category Overall event Underweight Normal weight Overweight rates (n ⴝ 59; (n ⴝ 1,131; (n ⴝ 1,483; (n ⴝ 3,645) 1.6%) 31.0%) 40.8%)
Death Cardiac complications Stroke Surgical site infection Respiratory complication Venous thromboembolism Return to the operating suite
0.8 (31) 1.4 (51) 2.1 (74) 0.9 (33) 1.8 (65) 0.2 (7) 4.8 (175)
1.7 (1) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 5.1 (3)
0.6 (7) 1.1 (12) 2.2 (25) 1.4 (16) 1.4 (16) 0.2 (2) 5.4 (61)
Class I obese Class II obese (n ⴝ 703; (n ⴝ 212; Class III obese 19.3%) 5.8%) (n ⴝ 57; 1.6%)
0.7 (10) 1.1 (16) 2.1 (31) 0.7 (10) 2.0 (29) 0.1 (1) 4.3 (64)
1.1 (8) 2.4 (17) 1.9 (13) 0.7 (5) 2.1 (15) 0 (0) 4.7 (33)
2.4 (5) 1.9 (4) 2.4 (5) 0 (0) 1.9 (4) 1.4 (3) 3.8 (8)
0 (0) 3.5 (2) 0 (0) 3.5 (2) 1.8 (1) 1.8 (1) 10.5 (6)
Values shown are incidence proportion (number of events). p Values are not shown because the number of events was too small to make statistically meaningful comparisons.
statistically significant differences in preoperative hematocrit, the mean values for each BMI category were within the accepted normal range. Overall incidences of study outcomes are displayed in Table 4. For the unadjusted analysis, statistical significance is not reported because the number of events was too small to make meaningful comparisons. The incidence of death was highest in underweight and class II to III obese patients. The observed incidence of cardiac complications increased with increasing BMI category, and the incidence of VTE was higher among class II and III obese patients compared with other weight categories. The incidence of SSI was highest in class III obese patients. On multivariable analysis, obesity (BMI ⱖ30, class I to III) was significantly associated with a 3-fold higher risk of death (odds ratio [OR] ⫽ 3.42; 95% CI, 1.45⫺8.10; p ⫽ 0.005) and a 2-fold higher risk of cardiac complications (OR ⫽ 2.38; 95% CI, 1.23⫺4.62; p ⫽ 0.01) compared with normal weight. When BMI categories were analyzed separately (Table 5), only class II to III (OR ⫽ 6.95; 95% CI, 1.89⫺25.58; p ⫽ 0.004) obesity was associated with increased risk of death. Only class II to III obesity was significantly associated with increased risk of cardiac complications (OR ⫽ 3.68; 95% CI, 1.27⫺10.66; p ⫽ 0.02). A test for trend showed that increasing BMI category was associated with increasing risk of both death (p ⫽ 0.03) and cardiac complications (p ⬍ 0.001). Increasing age and lower-extremity rest pain/gangrene were risk factors for both death and cardiac complications. Preoperative creatinine ⬎2 mg/dL and current smoking were also associated with increased risk of death, and previous stroke with or without residual deficit was associated with increased risk of cardiac complications. Thirty-day mortality was substantially higher among patients who experienced cardiac complications vs those without cardiac complications (17.0% vs 0.6%; p ⬍ 0.001). However, even after excluding patients with cardiac complications, obesity remained independently associated with an elevated risk of 30-day
mortality (OR ⫽ 7.30; 95% CI, 2.22⫺24.00; p ⫽ 0.001). When the component outcomes of cardiac complications were analyzed, BMI was not associated with the risk of MI, but there was significantly increased adjusted risk of cardiac arrest among obese patients (BMI ⱖ30) compared with nonobese patients (OR ⫽ 3.42; 95% CI, 1.45⫺8.10; p ⫽ Table 5. Adjusted Risk of Death and Cardiac Complications, Showing All Variables Selected (Using Stepwise Regression) for Final Model Inclusion BMI complications
Death Underweight Normal weight Overweight Class I obesity Class II–III obesity* Age, 5-year increase Current smoking Lower-extremity rest pain/gangrene Preoperative creatinine ⬎2 mg/dL Cardiac complications Underweight Normal weight Overweight Class I obesity Class II–III* obesity Age, 5-year increase Lower-extremity rest pain/gangrene Previous stroke, no residual deficit Previous stroke with residual deficit
Odds ratio (95% CI)
p Value
2.98 (0.32⫺27.44) Reference 0.89 (0.28⫺2.86) 2.34 (0.72⫺7.66) 6.95 (1.89⫺25.58) 1.32 (1.01⫺1.72) 2.50 (1.00⫺6.26)
0.33 Reference 0.85 0.16 0.004 0.04 0.03
5.99 (1.64⫺21.88)
0.007
3.61 (1.16⫺11.23)
0.05
Undefined† Reference 0.98 (0.42⫺2.27) 1.95 (0.79⫺4.80) 3.68 (1.27⫺10.66) 1.29 (1.07⫺1.57)
Undefined† Reference 0.97 0.15 0.02 0.009
3.85 (1.12⫺13.23)
0.03
2.39 (1.00⫺5.69)
0.05
2.25 (1.05⫺4.82)
0.04
*Pooled analysis of class II and III obese patients. (Analysis was pooled for both outcomes due to the small number of class III obese patients.) † No events. BMI, body mass index.
Vol. 214, No. 2, February 2012
0.005). The Hosmer-Lemeshow test statistic p value for both death (0.44) and cardiac complications (0.77) was not significant, consistent with good model fit. BMI category was not significantly associated with the risk of stroke, SSI, respiratory complications, VTE, or return to the operating suite. Analysis of collinearity among variables included in the final stepwise regression models demonstrated a maximum VIF of 1.79, indicating low multicollinearity.
DISCUSSION This analysis found that obesity was independently associated with increased risk of death and cardiac complications after carotid endarterectomy. We did not find evidence of an association between BMI and risk of stroke, SSI, respiratory complication, VTE, or return to the operating suite. Very few underweight patients were included, and the association of underweight with study outcomes could not be assessed with statistical precision. Several previous studies have reported an increase in the risk of perioperative mortality among obese patients undergoing vascular surgery. We previously demonstrated a relationship between obesity and increased 30-day mortality after open and endovascular AAA repair among patients represented in the VASQIP.12 A similar analysis of 5,455 patients in the American College of Surgeons National Surgical Quality Improvement Program showed that class III obesity was a risk factor for 30-day mortality after open AAA repair but not after endovascular AAA repair.13 In contrast, a few studies have found evidence supporting an obesity paradox. This concept suggests that although obesity is a risk factor for development of cardiovascular and peripheral vascular disease, mild obesity can confer a survival benefit after certain cardiovascular and peripheral vascular surgical procedures.24 An analysis by Davenport and colleagues of data collected from 7,543 patients who underwent a variety of vascular surgery procedures found a reverse-J shape relationship between BMI and 30-day mortality, with the highest risk in underweight and normalweight patients and the lowest risk in class I obese patients.25 Similarly, an analysis by Giles and colleagues of 7,595 lower-extremity bypasses in the American College of Surgeons National Surgical Quality Improvement Program found a U-shaped relationship between BMI and 30-day mortality, with the lowest mortality in class I and II obese patients and the highest mortality in class III obese patients.14 In both studies, the protective association of mild obesity persisted on multivariable analysis. These reports of a protective association of mild obesity on perioperative mortality contrast with our finding of increasing risk of 30-day mortality with increasing BMI
Jackson et al
Obesity and Carotid Endarterectomy
153
category. Differences in study design might account for these discrepant findings. Most noticeably, we examined only carotid endarterectomies, and cerebrovascular procedures constituted only 17.4% of the procedures studied by Davenport and colleagues25 and none of the procedures examined by Giles and colleagues,14 making direct comparison of our results with theirs rather tenuous. Additionally, the proportion of women in the reports by Davenport and colleagues (40.2%)25 and Giles and colleagues (20.8%)14 were substantially higher than in our study (1.4%), based on the patient population at VA medical centers. If the relationship between BMI and surgical outcomes differs between men and women, this might contribute to the differences between our findings and those of the 2 studies discussed here. Finally, although we were not able to detect a statistically significant association between underweight and adverse postoperative outcomes, this might have been due to the small number of underweight patients, reducing the power to detect such an association. The increased risk of perioperative mortality in obese patients is partially attributable to increased postoperative cardiac complications in this population. Obesity has long been recognized as an independent predictor of cardiovascular morbidity and mortality.5-7 Research suggests that visceral adipose tissue secretes cytokines and chemokines that attract macrophages, the products of which lead to a systemic proinflammatory, proatherosclerotic state associated with development of coronary artery disease.26 Additionally, obesity is associated with subclinical abnormalities in left atrial and ventricular structure and function, independent of its association with age, blood pressure, or heart rate.27 Underlying coronary disease and myocardial dysfunction likely contribute to the association between obesity and cardiac complications after carotid endarterectomy. Our data suggest that mortality risk after carotid endarterectomy is elevated even among patients without documented cardiac complications. Unrecognized postoperative cardiac complications might partly account for this. Additionally, worse access to preventive health care services among obese patients, compared with normal-weight patients, might be contributory. Several studies suggest that obese patients are less likely than nonobese patients to undergo health screening examinations.28,29 As a result, obese patients might suffer from a higher prevalence of undiagnosed and untreated comorbidities, which likely contribute to the association between obesity and mortality after carotid endarterectomy observed in this study. Among vascular surgery patients, one of the most commonly noted morbidities associated with obesity is SSI,12,14,25 and it is therefore relevant that our analysis did
154
Jackson et al
Obesity and Carotid Endarterectomy
not uncover such a relationship after carotid endarterectomy. Although we noted an overall 0.9% infection rate, infection rates in previous studies that identified obesity as a risk factor for SSI were 3.6% (AAA repair),12 6.7% (miscellaneous vascular operations),25 and 10.1% (lowerextremity bypass).14 Lack of association between obesity and SSI in our study is probably due to a lower incidence of wound infection overall among patients undergoing carotid endarterectomy, reducing the statistical power to detect such a relationship. Our analysis of baseline characteristics showed a strong association between increasing weight category and decreasing prevalence of smoking. Previous studies have similarly shown a negative correlation between obesity and current smoking,30 as well as an association between smoking cessation and weight gain.31 The association between underweight and COPD is most likely a result of weight loss due to this comorbidity. The VASQIP was designed as a quality-improvement tool to compare individual institutions’ postoperative outcomes with national benchmarks. As such, there are limitations to its use as a research database. For the years studied, the VASQIP did not contain information on hypertension or dyslipidemia, and our data set did not contain information on coronary intervention or congestive heart failure. Analyses were therefore not adjusted for these comorbidities. We were able, however, to adjust for many other important comorbidities, such as diabetes. Additionally, our study contained mostly males, and additional study will be needed to confirm the relationship between obesity and outcomes after carotid endarterectomy in women. Finally, although some studies have suggested that waist-to-hip ratio or waistto-thigh ratio can correlate more closely with cardiovascular risk and mortality than does BMI,32,33 we were limited to BMI as the measure of obesity because other anthropometric measurements were not included in the database.
CONCLUSIONS This report identifies obesity as an independent risk factor for death and cardiac complications after carotid endarterectomy. When considering carotid endarterectomy in obese patients, surgeons should assess the risks and benefits of the procedure, considering the increased risk of death and cardiac complications in obese patients. If the decision is made to proceed with carotid endarterectomy, surgeons should undertake additional risk assessment and modification strategies, including more rigorous preoperative cardiac evaluation and consideration for postoperative telemetry or ICU monitor-
J Am Coll Surg
ing for early recognition and treatment of cardiac or other complications. Author Contributions Study conception and design: Sidawy, Amdur, Macsata Acquisition of data: Sidawy, Amdur Analysis and interpretation of data: Jackson, Amdur Drafting of manuscript: Jackson Critical revision: Sidawy, Amdur, Macsata Acknowledgment: The authors would like to acknowledge the VA Surgical Quality Data Use Group (SQDUG) for its role as scientific advisors and for the critical review of data use and analysis presented in this article.
REFERENCES 1. Flegal KM, Carroll MD, Ogden CL, Johnson CL. Prevalence and trends in obesity among US adults, 1999–2000. JAMA 2002;288:1723–1727. 2. Flegal KM, Carroll MD, Ogden CL, Curtin LR. Prevalence and trends in obesity among US adults, 1999–2008. JAMA 2010; 303:235–241. 3. World Global Infobase Team. The SuRF Report 2. Surveillance and chronic disease risk factors: country-level data and comparative estimates. Geneva: WHO; 2005. 4. WHO. WHO European Ministerial Conference on Counteracting Obesity, 15⫺17 November 2006, Istanbul, Turkey. Conference report. Geneva: WHO; 2007. 5. Wilson PW, D’Agostino RB, Sullivan L, et al. Overweight and obesity as determinants of cardiovascular risk: the Framingham experience. Arch Intern Med 2002;162:1867–1872. 6. Krauss RM, Winston M. Obesity: impact on cardiovascular disease. Circulation 1998;98:1472. 7. Hubert HB, Feinleib M, McNamara PM, Castelli WP. Obesity as an independent risk factor for cardiovascular disease: a 26year follow-up of participants in the Framingham Heart Study. Circulation 1983;67:968–977. 8. Yatsuya H, Yamagishi K, North KE, et al for the ARIC Investigators. Associations of obesity measures with subtypes of ischemic stroke in the ARIC study. J Epidemiol 2010;20:347–354. 9. Planas A, Clara A, Pou JM, et al. Relationship of obesity distribution and peripheral arterial occlusive disease in elderly men. Int J Obes Relat Metab Disorder 2001;25:1068–1070. 10. Vogt MT, Cauley JA, Kuller LH, Hulley SB. Prevalence and correlates of lower extremity arterial disease in elderly women. Am J Epidemiol 1998;137:559–568. 11. Kent KC, Zwolak RM, Egorova NN, et al. Analysis of risk factors for abdominal aortic aneurysm in a cohort of more than 3 million individuals. J Vasc Surg 2010;52:539–548. 12. Johnson ON, Sidawy AN, Scanlon JM, et al. Impact of obesity on outcomes after open surgical and endovascular abdominal aortic aneurysm repair. J Am Coll Surg 2010;210:166–177. 13. Giles KA, Wyers MC, Pomposelli FB, et al. The impact of body mass index on perioperative outcomes of open and endovascular abdominal aortic aneurysm repair from the National Surgical Quality Improvement Program, 2005–2007. J Vasc Surg 2010;52: 1471–1477.
Vol. 214, No. 2, February 2012
14. Giles KA, Hamdan AD, Pomposelli FB, et al. Body mass index: surgical site infections and mortality after lower extremity bypass from the National Surgical Quality Improvement Program 2005⫺2007. Ann Vasc Surg 2010;24:48–56. 15. The North American Symptomatic Carotid Endarterectomy Trial (NASCET) Investigators. Clinical alert: benefit of carotid endarterectomy for patients with high-grade stenosis of the internal carotid artery. Stroke 1991;22:816–817. 16. Barnett HJM, Taylor W, Eliasqiw M, et al. Benefit of carotid endarterectomy in patients with symptomatic moderate or severe stenosis. N Engl J Med 1998;339:1415–1425. 17. Executive Committee for the Asymptomatic Carotid Atherosclerosis Study. Endarterectomy for asymptomatic carotid artery stenosis. JAMA 1995;273:1421–1428. 18. Kresowik TF, Bratzler DW, Kresowik RA. Multistate improvement in process and outcomes of carotid endarterectomy. J Vasc Surg 2004;39:372–380. 19. Nagaki T, Sato K, Yoshida T, Yoshimoto Y. Benefits of carotid endarterectomy for symptomatic and asymptomatic severe carotid artery stenosis: a Markov model based on data from randomized controlled trials. J Neurosurg 2009;111:970–977. 20. Jim J, Sicard GA. Who is unfit for carotid endarterectomy? Perspect Vasc Surg Endovasc Ther 2010;22:40–45. 21. Khuri SF, Daley J, Henderson W, et al. The Department of Veterans Affairs’ NSQIP: the first national, validated, outcomebased, risk-adjusted, and peer-controlled program for the measurement and enhancement of the quality of surgical care. Ann Surg 1998;228:491–507. 22. Clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults: executive summary. Expert Panel on the Identification, Evaluation, and Treatment of Overweight in Adults. Am J Clin Nutr 1998;68:899–917.
Jackson et al
Obesity and Carotid Endarterectomy
155
23. Allison P. Logistic regression using the SAS system: Theory and application. Cary, NC: SAS Institute, Inc.; 1999. 24. Amundson DE, Djurkovic S, Matmiyoff GN. The obesity paradox. Crit Care Clin 2010;26:538–596. 25. Davenport DL, Xenos ES, Hosokawa P, et al. The influence of body mass index obesity status on vascular surgery 30-day morbidity and mortality. J Vasc Surg 2009;49:140–147. 26. Ohman MK, Wright AP, Wickenheiser KJ, et al. Visceral adipose tissue and atherosclerosis. Curr Vasc Pharmacol 2009;7: 169–179. 27. Tumukulu MM, Etikan I, Kisacik B, Kayikcioglu M. Effects of obesity on left ventricular structure and myocardial systolic function: assessment by tissue Doppler imaging and strain/ strain rate imaging. Echocardiography 2007;8:802–809. 28. Maruthur NM, Bolen S, Francati FL, Clark JM. Obesity and mammography: a systematic review and meta-analysis. J Gen Intern Med 2009;24:665–677. 29. Amy NK, Aalborg A, Lyons P, Keranen L. Barriers to routine gynecological cancer screening for white and African-American obese women. Int J Obes 2006;30:147–155. 30. Healton CG, Vallone D, McCausland KL, et al. Smoking, obesity, and their co-occurrence in the United States: cross sectional analysis. BMJ 2006;333:25–26. 31. Flegal KM. The influence of smoking cessation on the prevalence of overweight in the United States. N Engl J Med 1995; 333:1165–1170. 32. Reis JP, Araneta MR, Wingard DL, et al. Overall obesity and abdominal adiposity as predictors of mortality in US white and black adults. Ann Epidemiol 2009;19:134–142. 33. Bodenant M, Kuulasmaa K, Wagner A, et al. Measures of abdominal adiposity and the risk of stroke: The MOnica Risk, Genetics, Archiving and Monograph (MORGAM) Study. Stroke 2011;42:2872–2877.