From the Eastern Vascular Society
Postoperative complications after common femoral endarterectomy Bao-Ngoc Nguyen, MD, Richard L. Amdur, PhD, Mustafa Abugideiri, BS, Rodeen Rahbar, MD, Richard F. Neville, MD, and Anton N. Sidawy, MD, MPH, Washington, D.C. Background: Common femoral endarterectomy (CFE) for limited arterial occlusive disease is considered a fairly low-risk operation of short duration. This study investigated the timing of 30-day outcomes as they related to hospital discharge and predicted the risk of operative mortality of this procedure. Methods: All patients in the National Surgical Quality Improvement Program database who underwent isolated CFE between 2005 and 2010 were selected for the test sample. We identified postoperative mortality and morbidities occurring before and after hospital discharge. A risk calculator for 30-day mortality, developed in the test sample using logistic regression, was validated in a new sample of cases from 2011 to 2012. Results: A total of 1843 CFEs reported from 2005 to 2010 met the inclusion and exclusion criteria. The average operative time was 146 6 69.5 minutes (median, 133; interquartile range, 98-179 minutes), and 10% of patients needed to return to the operating room. The average length of stay was 4 6 7.5 days (median, 3; interquartile range, 2-5 days); 91% of patients were discharged #1 week of surgery. Occurrences of cardiovascular events, renal dysfunction, and pulmonary complication were relatively low. There was 3.4% mortality and 8% wound-related complications, 30% and 86% of which occurred after hospital discharge, respectively. Overall, there was a 15% risk of combined mortality/morbidity, and >60% of these events occurred after discharge. The independent predictors of 30-day mortality were age, nonindependent functional status, preoperative dialysis, sepsis, emergency status, and American Society of Anesthesiologists Physical Status Classification 4 or 5, and the association between risk strata and death in the validation sample was strong (4 [ 0.29) and significant (P < .001). Conclusions: CFE is not as “benign” a procedure as previously believed. The risks of death and wound complications are not insignificant, and a high percentage of these complications occurred after patients were discharged from the hospital. Patients should be carefully selected, especially in the elderly population, and close postoperative follow-up should be considered. (J Vasc Surg 2015;61:1489-94.)
Isolated common femoral arterial lesions are rare in lower extremity arterial occlusive disease, and the sole treatment of common femoral artery disease is quite effective in relieving the symptoms of claudication and rest pain, with high rates of limb salvage.1-5 On one hand, because the common femoral artery is located at a high mobility area, stenting could lead to stent fracture and also preclude a potential access site for subsequent endovascular interventions. On the other hand, because the common femoral artery is easily accessible surgically, even under local anesthesia, most surgeons treat the disease with open endarterectomy because the procedure is considered to be technically straightforward and of short duration, with minimal postoperative complications. From the Department of Surgery, George Washington University. Author conflict of interest: none. Presented at the Twenty-seventh Annual Meeting of the Eastern Vascular Society, White Sulfur Spring, WVa, September 19-22, 2013. Additional material for this article may be found online at www.jvascsurg.org. Reprint requests: Bao-Ngoc Nguyen, MD, 2150 Pennsylvania Ave #6B411, Washington, D.C. 20037 (e-mail: [email protected]). The editors and reviewers of this article have no relevant financial relationships to disclose per the JVS policy that requires reviewers to decline review of any manuscript for which they may have a conflict of interest. 0741-5214 Copyright Ó 2015 by the Society for Vascular Surgery. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jvs.2015.01.024
Most single-center studies on isolated common femoral endarterectomy (CFE) have reported a very limited number of patients. In this study, we used the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) database to analyze a large volume of procedures, allowing in-depth study of the postoperative outcomes of this procedure and focusing specifically on the timing of the complications and the predictors of 30-day mortality. METHODS We performed a retrospective cohort study using data from the ACS-NSQIP database. The George Washington University Institutional Review Board approved protocol for the NSQIP study. Informed consents were waived. All CFEs in the NSQIP database between 2005 and 2010 were selected by the Current Procedural Terminology (American Medical Association, Chicago, Ill) code 35,371. Profunda and superficial femoral endarterectomies were accepted as concurrent codes, but we excluded other major vascular procedures such as aneurysm repair, infrainguinal bypass, major amputation, and thrombectomy. Also excluded were reoperative or hybrid procedures that involved additional endovascular treatment of the inflow or outflow arteries. Preoperative patient characteristics examined included age, weight, sex, race, American Society of Anesthesiologists (ASA) Physical Status Classification, cardiac risks 1489
JOURNAL OF VASCULAR SURGERY June 2015
1490 Nguyen et al
(smoking, diabetes, congestive heart failure, myocardial infarction, previous percutaneous coronary intervention, or coronary artery bypasses), chronic obstructive pulmonary disease, acute kidney injury, dialysis, the presence of rest pain, and surgical emergency status. Operative details recorded included procedure duration and the amount of intraoperative blood transfusion. Thirty-day outcomes included death, cardiac, pulmonary (prolonged intubation, pulmonary embolism, reintubation, or pneumonia), renal complications (acute kidney injury or dialysis occurred in patients without previous dialysis), sepsis, and graft failure. Wound-related complications in the NSQIP database are defined as superficial infection, deep wound infection, organ space infection, or wound dehiscence, which are combined into one main category of “wound complication” during data analysis. In addition to reporting the incidence of postoperative outcomes, this study pursued a more focused analysis of the timing of major postoperative events before and after hospital discharge during a 30-day period. We took this approach for two main reasons. Clinically, it is important to examine whether there is a high risk of postdischarge complications that would warrant a closer postoperative follow-up. From a research perspective, it is important to know whether databases providing only inpatient outcomes can provide valid conclusions. The NSQIP variable indicating days from surgery to discharge was used, together with the outcome event timing variables, to determine whether mortality and morbidity of the respective events occurred before or after hospital discharge. Another objective of our study was to find a prediction model for death and then validate it in a new patient sample. We used a stepwise logistic regression model. Initial variables that were considered for model entry included a wide variety of presurgery patient characteristics (sex, race, weight, albumin, emergency status, smoking, ASA class 4 or 5, rest pain, diabetes mellitus, surgery time, ventilator status, chronic obstructive pulmonary disease, pneumonia, ascites, esophageal varices, congestive heart failure, myocardial infarction, previous percutaneous coronary intervention or cardiac surgery, angina, hypertension, peripheral vascular disease, acute kidney injury, dialysis, impaired sensorium, coma, hemiplegia, cerebrovascular accident [CVA], transient ischemic attack, central nervous system tumor, wound infection, steroid use, weight loss >10%, paraplegia or quadriplegia, chemotherapy or radiotherapy, bleeding disorder, transfusion, sepsis, pregnancy, previous surgery #30 days). Stepwise regression was used to select relevant predictors, using P < .05 as a cutoff for model entry and exit. The regression parameters were then used to compute a simplified risk score for each patient by dividing each regression parameter by the smallest parameter and rounding off to the nearest integer. Risk strata with increasing likelihood of having the outcome were created, and their association with the outcome was examined using c2. To cross-validate the prediction model in a new sample, we used ACS-NSQIP data from 2011 and 2012.
Table I. Patient demographics and comorbidities Variables Age,a years Nonindependent statusa Weight, kg Albumin < 3 mg/dL Female sex Race Black White Other Emergency surgerya ASA class 4 or 5a Pre-op sepsisa Diabetes Hypertension Peripheral vascular disease Rest pain Congestive heart failure History of Angina Myocardial infarction PCI Cardiac surgery Transient ischemic attack Stroke Smoker COPD Pneumonia Steroid use Acute kidney injury Dialysisa
Mean 6 SD or No. (%) (N ¼ 1843) 69.2 6 11.6 239 (13) 78 6 18.6 113 (6.1) 742 (40.3) 122 1138 583 236 378 91 614 1558 772 527 56
(6.6) (61.8) (31.6) (12.8) (20.5) (5.0) (33.3) (84.5) (41.9) (28.6) (3.0)
81 88 428 467 154 116 636 252 9 93 20 66
(4.4) (4.8) (23.2) (25.3) (8.4) (6.3) (34.5) (13.7) (0.5) (5.1) (1.1) (3.6)
ASA class, American Society of Anesthesiologists Physical Status Classification; COPD, chronic obstructive pulmonary disease; PCI, percutaneous coronary intervention; SD, standard deviation. a Independent predictors of 30-day mortality.
Patients were selected in the same way as described above, using the same exclusions. We computed risk scores for each patient using the prediction model and cut-points for risk strata that were developed in the 2005 to 2010 data set and examined the association of predicted risk with actual outcomes. In addition to 30-day mortality, predictors for other complications with incidence greater than 5% were also examined. All analyses were done using SAS 9.3 software (SAS Institute Inc, Cary, NC) with P < .05 considered significant. RESULTS A total of 1843 CFEs were reported during 2005 to 2010. Patient demographics and preoperative comorbidities are summarized in Table I. The average age was 69 6 12 years, and 40% were women. Twenty-three percent were positive for rest pain, and 13% were emergencies with critical limb ischemia. Eighty percent of the patients had hypertension, 33% were diabetic, and 34% were smokers. Although 23% to 25% of patients had undergone cardiac surgery or coronary stenting, 60% of these occurred after hospital discharge. The prediction model for 30-day mortality found six independent predictors for death: age, nonindependent functional status, preoperative dialysis, preoperative sepsis, emergency surgery, and ASA class >3 (Table IV). On the basis of the regression model, the risk score for 30-day mortality was computed for each individual based on the equation: Mortality risk ¼ years of age þ 7 ðnon-independent functional statusÞ þ 5 ðpreoperative dialysisÞ þ 5 ðpreoperative sepsisÞ þ 3 ðemergency surgeryÞ þ 2 ðASA 4 or 5Þ In this equation, years of age is a continuous variable, whereas the other five variables are binary (1 if present and 0 if not present). This model had excellent discrimination (C ¼ 0.92). We stratified risk scores into increments of 10s with rising risk of death (Fig 3). In the lower end of the spectrum, where the risk score was 100 (Fig 3). The validation sample, drawn from ACS-NSQIP patients in 2011 to 2012, with the same inclusion and exclusion criteria used in the test sample, had 1009 patients with all relevant variables, of which 17 died (1.7%). Risk strata
Nguyen et al 1491
Table II. Intraoperative and postoperative outcomes Outcomes Intraoperative Operative time, hours Transfusion >4 units Postoperative Mortality Return to the operating room Wound complications Superficial infection Deep wound infection Wound dehiscence Pneumonia Prolonged intubation Genitourinary tract infection Sepsis Septic shock Graft failure Cardiac arrest Myocardial infarction Acute kidney injury Dialysis Deep vein thrombosis Pulmonary embolism
Mean 6 SD or No. (%) 2.4 6 1.16 51 (2.8) 62 188 147 109 37 15 29 26 29 30 19 21 17 11 5 11 15 2
(3.4) (10.2) (8) (5.9) (2.0) (0.8) (1.6) (1.4) (1.6) (1.6) (1.0) (1.1) (0.9) (0.6) (0.3) (0.6) (0.8) (0.1)
SD, Standard deviation.
Fig 1. Thirty-day post-operative complications after common femoral endarterectomy (CFE). Major organ dysfunction (cardiac, pulmonary, renal) was rare, but wound-related complications and operative reintervention rates were high. The mortality rate was relatively high at 3.4%. OR, Operating room.
were computed using the risk score model and risk cutpoints developed in the test sample. The association between risk strata and death was strong (4 ¼ .29) and significant (P < .0001). Because the incidence of wound-related complications and return to the operating room were both >5%, we examined predictors for these complications. Significant predictors positively associated with wound complications included operation time (P ¼ .0002), weight (P < .0001), female sex (P ¼ .0009), diabetes mellitus (P ¼ .03), dialysis (P ¼ .0016), and chronic steroid use (P ¼ .0074). Independent predictors positively associated with return to the operating room included weight (P ¼ .0006), operation time (P ¼ .0003), emergency procedure
JOURNAL OF VASCULAR SURGERY June 2015
1492 Nguyen et al
Table III. Distribution of occurrence of postoperative complications during 30-day period Weeks after CFE Outcomes
First, No. (%)
Mortality Cardiovascular events Pulmonary complications Renal Wound complications Composite
31 19 38 10 14 77
Second, No. (%)
(53.5) (70.4) (63.3) (62.5) (9.2) (27.8)
12 4 11 4 60 85
(20.7) (14.8) (18.3) (25.0) (39.2) (30.7)
Third, No. (%) 4 1 8 2 45 63
(6.9) (3.7) (13.3) (12.5) (29.4) (22.7)
Fourth, No. (%) 11 3 3 0 34 52
(19.0) (11.1) (5.0) (0) (22.2) (18.8)
CFE, Common femoral endarterectomy.
Fig 2. Distribution of major complications before and after hospital discharge. This graph provides another perspective to the distribution of postoperative complications reported in Table III. A significant percentage (30%) of deaths and cardiac and pulmonary complications occurred after hospital discharge. Most woundrelated complications occurred after hospital discharge.
Fig 3. Prediction model for 30-day mortality. Age and nonindependent functional status carried the highest weights as predictive factors for death. Risk scores were calculated for the patients in the test sample and tabulated to the observed death rates to estimate the expected death rate for a given patient. The range bars show the standard deviation. ASA, American Society of Anesthesiologists.
Table IV. Predictors of 30-day mortality Predictors
ORa
95% CI
P
Age Nonindependent status Dialysis Sepsis Emergency ASA class 4/5
1.04 6.72 4.95 4.61 3.27 2.31
1.01-1.07 3.64-12.44 2.06-11.92 2.30-9.35 1.72-6.21 1.26-4.25
.0046 80 years who are not independent. Close postoperative follow-up should be exercised to monitor complications after discharge. AUTHOR CONTRIBUTIONS Conception and design: BN Analysis and interpretation: BN, RA, RR Data collection: BN, RA, MA Writing the article: BN, RA Critical revision of the article: AS, RA, RN Final approval of the article: AS, RA Statistical analysis: RA Obtained funding: Not applicable Overall responsibility: BN REFERENCES 1. Archie JP Jr, Feldtman RW. Common femoral artery atherosclerotic occlusion. Difficult to diagnose but easy to treat. Am Surg 1982;48: 339-40. 2. Mukherjee D, Inahara T. Endarterectomy as the procedure of choice for atherosclerotic occlusive lesions of the common femoral artery. Am J Surg 1989;157:498-500. 3. Springhorn ME, Kinney M, Littooy FN, Saletta C, Greisler HP. Inflow atherosclerotic disease localized to the common femoral artery: treatment and outcome. Ann Vasc Surg 1991;5:234-40.
1494 Nguyen et al
4. Cardon A, Aillet S, Jarno P, Bensalah K, Le Du J, Idrissi A, et al. [Endarteriectomy of the femoral tripod: long-term results and analysis of failure factors]. Ann Chir 2001;126:777-82. 5. Desai M, Tsui J, Davis M, Myint F, Wilson A, Baker DM, et al. Isolated endarterectomy of femoral bifurcation in critical limb ischemia: is restoration of inline flow essential? Angiology 2011;62:119-25. 6. Kechagias A, Ylonen K, Biancari F. Long-term outcome after isolated endarterectomy of the femoral bifurcation. World J Surg 2008;32: 51-4. 7. Ballotta E, Gruppo M, Mazzalai F, Da Giau G. Common femoral artery endarterectomy for occlusive disease: an 8-year single-center prospective study. Surgery 2010;147:268-74. 8. Derksen WJ, Verhoeven BA, van de Mortel RH, Moll FL, de Vries JP. Risk factors for surgical-site infection following common femoral artery endarterectomy. Vasc Endovascular Surg 2009;43:69-75.
JOURNAL OF VASCULAR SURGERY June 2015
9. Greenhalgh RM, Brown LC, Kwong GP, Powell JT, Thompson SG. Comparison of endovascular aneurysm repair with open repair in patients with abdominal aortic aneurysm (EVAR trial 1), 30-day operative mortality results: randomised controlled trial. Lancet 2004;364:843-8. 10. Kang JL, Patel VI, Conrad MF, Lamuraglia GM, Chung TK, Cambria RP. Common femoral artery occlusive disease: contemporary results following surgical endarterectomy. J Vasc Surg 2008;48: 872-7.
Submitted Sep 10, 2014; accepted Jan 13, 2015.
Additional material for this article may be found online at www.jvascsurg.org.
REQUEST FOR SUBMISSION OF SURGICAL ETHICS CHALLENGES ARTICLES The Editors invite submission of original articles for the Surgical Ethics Challenges section, following the general format established by Dr. James Jones in 2001. Readers have benefitted greatly from Dr. Jones’ monthly ethics contributions for more than 6 years. In order to encourage contributions, Dr. Jones will assist in editing them and will submit his own articles every other month, to provide opportunity for others. Please submit articles under the heading of “Ethics” using Editorial Manager, and follow the format established in previous issues.
JOURNAL OF VASCULAR SURGERY Volume 61, Number 6
Nguyen et al 1494.e1
Supplementary Table (online only). Risk calculator for 30-day mortalitya Predictors Age, years Nonindependent Pre-op dialysis Pre-op sepsis Emergency surgery ASA class 4 or 5 Risk ¼ Likelihood of 30-day mortality, mean 6 SD %
Predictor weight
Healthy 80 yo
Healthy 90 yo
1 7 5 5 3 2
80 0 0 0 0 0 80 3.2 6 1.6
90 0 0 0 0 0 90 7.8 6 3.6
Nonindependent 90 yo 90 1 0 0 0 0 97 32.1 6 12.2
Nonindependent 75 yo emergency 75 1 0 0 1 0 85 w5.5
ASA class, American Society of Anesthesiologists Physical Status Classification; SD, Standard deviation; yo, years old. a Instructions: To compute the risk score for death, enter the patient’s age, and 1 for each factor present. Multiple each number by the predictor weight. Take the sum and round to the nearest 10. Use this number together with Fig 3 to estimate the likelihood of 30-day mortality.
Postoperative complications after common femoral endarterectomy Bao-Ngoc Nguyen, MD, Richard L. Amdur, PhD, Mustafa Abugideiri, BS, Rodeen Rahbar, MD, Richard F. Neville, MD, and Anton N. Sidawy, MD, MPH, Washington, D.C. Background: Common femoral endarterectomy (CFE) for limited arterial occlusive disease is considered a fairly low-risk operation of short duration. This study investigated the timing of 30-day outcomes as they related to hospital discharge and predicted the risk of operative mortality of this procedure. Methods: All patients in the National Surgical Quality Improvement Program database who underwent isolated CFE between 2005 and 2010 were selected for the test sample. We identified postoperative mortality and morbidities occurring before and after hospital discharge. A risk calculator for 30-day mortality, developed in the test sample using logistic regression, was validated in a new sample of cases from 2011 to 2012. Results: A total of 1843 CFEs reported from 2005 to 2010 met the inclusion and exclusion criteria. The average operative time was 146 6 69.5 minutes (median, 133; interquartile range, 98-179 minutes), and 10% of patients needed to return to the operating room. The average length of stay was 4 6 7.5 days (median, 3; interquartile range, 2-5 days); 91% of patients were discharged #1 week of surgery. Occurrences of cardiovascular events, renal dysfunction, and pulmonary complication were relatively low. There was 3.4% mortality and 8% wound-related complications, 30% and 86% of which occurred after hospital discharge, respectively. Overall, there was a 15% risk of combined mortality/morbidity, and >60% of these events occurred after discharge. The independent predictors of 30-day mortality were age, nonindependent functional status, preoperative dialysis, sepsis, emergency status, and American Society of Anesthesiologists Physical Status Classification 4 or 5, and the association between risk strata and death in the validation sample was strong (4 [ 0.29) and significant (P < .001). Conclusions: CFE is not as “benign” a procedure as previously believed. The risks of death and wound complications are not insignificant, and a high percentage of these complications occurred after patients were discharged from the hospital. Patients should be carefully selected, especially in the elderly population, and close postoperative follow-up should be considered. (J Vasc Surg 2015;61:1489-94.)
Isolated common femoral arterial lesions are rare in lower extremity arterial occlusive disease, and the sole treatment of common femoral artery disease is quite effective in relieving the symptoms of claudication and rest pain, with high rates of limb salvage.1-5 On one hand, because the common femoral artery is located at a high mobility area, stenting could lead to stent fracture and also preclude a potential access site for subsequent endovascular interventions. On the other hand, because the common femoral artery is easily accessible surgically, even under local anesthesia, most surgeons treat the disease with open endarterectomy because the procedure is considered to be technically straightforward and of short duration, with minimal postoperative complications. From the Department of Surgery, George Washington University. Author conflict of interest: none. Presented at the Twenty-seventh Annual Meeting of the Eastern Vascular Society, White Sulfur Spring, WVa, September 19-22, 2013. Additional material for this article may be found online at www.jvascsurg.org. Reprint requests: Bao-Ngoc Nguyen, MD, 2150 Pennsylvania Ave #6B411, Washington, D.C. 20037 (e-mail: [email protected]). The editors and reviewers of this article have no relevant financial relationships to disclose per the JVS policy that requires reviewers to decline review of any manuscript for which they may have a conflict of interest. 0741-5214 Copyright Ó 2015 by the Society for Vascular Surgery. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jvs.2015.01.024
Most single-center studies on isolated common femoral endarterectomy (CFE) have reported a very limited number of patients. In this study, we used the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) database to analyze a large volume of procedures, allowing in-depth study of the postoperative outcomes of this procedure and focusing specifically on the timing of the complications and the predictors of 30-day mortality. METHODS We performed a retrospective cohort study using data from the ACS-NSQIP database. The George Washington University Institutional Review Board approved protocol for the NSQIP study. Informed consents were waived. All CFEs in the NSQIP database between 2005 and 2010 were selected by the Current Procedural Terminology (American Medical Association, Chicago, Ill) code 35,371. Profunda and superficial femoral endarterectomies were accepted as concurrent codes, but we excluded other major vascular procedures such as aneurysm repair, infrainguinal bypass, major amputation, and thrombectomy. Also excluded were reoperative or hybrid procedures that involved additional endovascular treatment of the inflow or outflow arteries. Preoperative patient characteristics examined included age, weight, sex, race, American Society of Anesthesiologists (ASA) Physical Status Classification, cardiac risks 1489
JOURNAL OF VASCULAR SURGERY June 2015
1490 Nguyen et al
(smoking, diabetes, congestive heart failure, myocardial infarction, previous percutaneous coronary intervention, or coronary artery bypasses), chronic obstructive pulmonary disease, acute kidney injury, dialysis, the presence of rest pain, and surgical emergency status. Operative details recorded included procedure duration and the amount of intraoperative blood transfusion. Thirty-day outcomes included death, cardiac, pulmonary (prolonged intubation, pulmonary embolism, reintubation, or pneumonia), renal complications (acute kidney injury or dialysis occurred in patients without previous dialysis), sepsis, and graft failure. Wound-related complications in the NSQIP database are defined as superficial infection, deep wound infection, organ space infection, or wound dehiscence, which are combined into one main category of “wound complication” during data analysis. In addition to reporting the incidence of postoperative outcomes, this study pursued a more focused analysis of the timing of major postoperative events before and after hospital discharge during a 30-day period. We took this approach for two main reasons. Clinically, it is important to examine whether there is a high risk of postdischarge complications that would warrant a closer postoperative follow-up. From a research perspective, it is important to know whether databases providing only inpatient outcomes can provide valid conclusions. The NSQIP variable indicating days from surgery to discharge was used, together with the outcome event timing variables, to determine whether mortality and morbidity of the respective events occurred before or after hospital discharge. Another objective of our study was to find a prediction model for death and then validate it in a new patient sample. We used a stepwise logistic regression model. Initial variables that were considered for model entry included a wide variety of presurgery patient characteristics (sex, race, weight, albumin, emergency status, smoking, ASA class 4 or 5, rest pain, diabetes mellitus, surgery time, ventilator status, chronic obstructive pulmonary disease, pneumonia, ascites, esophageal varices, congestive heart failure, myocardial infarction, previous percutaneous coronary intervention or cardiac surgery, angina, hypertension, peripheral vascular disease, acute kidney injury, dialysis, impaired sensorium, coma, hemiplegia, cerebrovascular accident [CVA], transient ischemic attack, central nervous system tumor, wound infection, steroid use, weight loss >10%, paraplegia or quadriplegia, chemotherapy or radiotherapy, bleeding disorder, transfusion, sepsis, pregnancy, previous surgery #30 days). Stepwise regression was used to select relevant predictors, using P < .05 as a cutoff for model entry and exit. The regression parameters were then used to compute a simplified risk score for each patient by dividing each regression parameter by the smallest parameter and rounding off to the nearest integer. Risk strata with increasing likelihood of having the outcome were created, and their association with the outcome was examined using c2. To cross-validate the prediction model in a new sample, we used ACS-NSQIP data from 2011 and 2012.
Table I. Patient demographics and comorbidities Variables Age,a years Nonindependent statusa Weight, kg Albumin < 3 mg/dL Female sex Race Black White Other Emergency surgerya ASA class 4 or 5a Pre-op sepsisa Diabetes Hypertension Peripheral vascular disease Rest pain Congestive heart failure History of Angina Myocardial infarction PCI Cardiac surgery Transient ischemic attack Stroke Smoker COPD Pneumonia Steroid use Acute kidney injury Dialysisa
Mean 6 SD or No. (%) (N ¼ 1843) 69.2 6 11.6 239 (13) 78 6 18.6 113 (6.1) 742 (40.3) 122 1138 583 236 378 91 614 1558 772 527 56
(6.6) (61.8) (31.6) (12.8) (20.5) (5.0) (33.3) (84.5) (41.9) (28.6) (3.0)
81 88 428 467 154 116 636 252 9 93 20 66
(4.4) (4.8) (23.2) (25.3) (8.4) (6.3) (34.5) (13.7) (0.5) (5.1) (1.1) (3.6)
ASA class, American Society of Anesthesiologists Physical Status Classification; COPD, chronic obstructive pulmonary disease; PCI, percutaneous coronary intervention; SD, standard deviation. a Independent predictors of 30-day mortality.
Patients were selected in the same way as described above, using the same exclusions. We computed risk scores for each patient using the prediction model and cut-points for risk strata that were developed in the 2005 to 2010 data set and examined the association of predicted risk with actual outcomes. In addition to 30-day mortality, predictors for other complications with incidence greater than 5% were also examined. All analyses were done using SAS 9.3 software (SAS Institute Inc, Cary, NC) with P < .05 considered significant. RESULTS A total of 1843 CFEs were reported during 2005 to 2010. Patient demographics and preoperative comorbidities are summarized in Table I. The average age was 69 6 12 years, and 40% were women. Twenty-three percent were positive for rest pain, and 13% were emergencies with critical limb ischemia. Eighty percent of the patients had hypertension, 33% were diabetic, and 34% were smokers. Although 23% to 25% of patients had undergone cardiac surgery or coronary stenting, 60% of these occurred after hospital discharge. The prediction model for 30-day mortality found six independent predictors for death: age, nonindependent functional status, preoperative dialysis, preoperative sepsis, emergency surgery, and ASA class >3 (Table IV). On the basis of the regression model, the risk score for 30-day mortality was computed for each individual based on the equation: Mortality risk ¼ years of age þ 7 ðnon-independent functional statusÞ þ 5 ðpreoperative dialysisÞ þ 5 ðpreoperative sepsisÞ þ 3 ðemergency surgeryÞ þ 2 ðASA 4 or 5Þ In this equation, years of age is a continuous variable, whereas the other five variables are binary (1 if present and 0 if not present). This model had excellent discrimination (C ¼ 0.92). We stratified risk scores into increments of 10s with rising risk of death (Fig 3). In the lower end of the spectrum, where the risk score was 100 (Fig 3). The validation sample, drawn from ACS-NSQIP patients in 2011 to 2012, with the same inclusion and exclusion criteria used in the test sample, had 1009 patients with all relevant variables, of which 17 died (1.7%). Risk strata
Nguyen et al 1491
Table II. Intraoperative and postoperative outcomes Outcomes Intraoperative Operative time, hours Transfusion >4 units Postoperative Mortality Return to the operating room Wound complications Superficial infection Deep wound infection Wound dehiscence Pneumonia Prolonged intubation Genitourinary tract infection Sepsis Septic shock Graft failure Cardiac arrest Myocardial infarction Acute kidney injury Dialysis Deep vein thrombosis Pulmonary embolism
Mean 6 SD or No. (%) 2.4 6 1.16 51 (2.8) 62 188 147 109 37 15 29 26 29 30 19 21 17 11 5 11 15 2
(3.4) (10.2) (8) (5.9) (2.0) (0.8) (1.6) (1.4) (1.6) (1.6) (1.0) (1.1) (0.9) (0.6) (0.3) (0.6) (0.8) (0.1)
SD, Standard deviation.
Fig 1. Thirty-day post-operative complications after common femoral endarterectomy (CFE). Major organ dysfunction (cardiac, pulmonary, renal) was rare, but wound-related complications and operative reintervention rates were high. The mortality rate was relatively high at 3.4%. OR, Operating room.
were computed using the risk score model and risk cutpoints developed in the test sample. The association between risk strata and death was strong (4 ¼ .29) and significant (P < .0001). Because the incidence of wound-related complications and return to the operating room were both >5%, we examined predictors for these complications. Significant predictors positively associated with wound complications included operation time (P ¼ .0002), weight (P < .0001), female sex (P ¼ .0009), diabetes mellitus (P ¼ .03), dialysis (P ¼ .0016), and chronic steroid use (P ¼ .0074). Independent predictors positively associated with return to the operating room included weight (P ¼ .0006), operation time (P ¼ .0003), emergency procedure
JOURNAL OF VASCULAR SURGERY June 2015
1492 Nguyen et al
Table III. Distribution of occurrence of postoperative complications during 30-day period Weeks after CFE Outcomes
First, No. (%)
Mortality Cardiovascular events Pulmonary complications Renal Wound complications Composite
31 19 38 10 14 77
Second, No. (%)
(53.5) (70.4) (63.3) (62.5) (9.2) (27.8)
12 4 11 4 60 85
(20.7) (14.8) (18.3) (25.0) (39.2) (30.7)
Third, No. (%) 4 1 8 2 45 63
(6.9) (3.7) (13.3) (12.5) (29.4) (22.7)
Fourth, No. (%) 11 3 3 0 34 52
(19.0) (11.1) (5.0) (0) (22.2) (18.8)
CFE, Common femoral endarterectomy.
Fig 2. Distribution of major complications before and after hospital discharge. This graph provides another perspective to the distribution of postoperative complications reported in Table III. A significant percentage (30%) of deaths and cardiac and pulmonary complications occurred after hospital discharge. Most woundrelated complications occurred after hospital discharge.
Fig 3. Prediction model for 30-day mortality. Age and nonindependent functional status carried the highest weights as predictive factors for death. Risk scores were calculated for the patients in the test sample and tabulated to the observed death rates to estimate the expected death rate for a given patient. The range bars show the standard deviation. ASA, American Society of Anesthesiologists.
Table IV. Predictors of 30-day mortality Predictors
ORa
95% CI
P
Age Nonindependent status Dialysis Sepsis Emergency ASA class 4/5
1.04 6.72 4.95 4.61 3.27 2.31
1.01-1.07 3.64-12.44 2.06-11.92 2.30-9.35 1.72-6.21 1.26-4.25
.0046 80 years who are not independent. Close postoperative follow-up should be exercised to monitor complications after discharge. AUTHOR CONTRIBUTIONS Conception and design: BN Analysis and interpretation: BN, RA, RR Data collection: BN, RA, MA Writing the article: BN, RA Critical revision of the article: AS, RA, RN Final approval of the article: AS, RA Statistical analysis: RA Obtained funding: Not applicable Overall responsibility: BN REFERENCES 1. Archie JP Jr, Feldtman RW. Common femoral artery atherosclerotic occlusion. Difficult to diagnose but easy to treat. Am Surg 1982;48: 339-40. 2. Mukherjee D, Inahara T. Endarterectomy as the procedure of choice for atherosclerotic occlusive lesions of the common femoral artery. Am J Surg 1989;157:498-500. 3. Springhorn ME, Kinney M, Littooy FN, Saletta C, Greisler HP. Inflow atherosclerotic disease localized to the common femoral artery: treatment and outcome. Ann Vasc Surg 1991;5:234-40.
1494 Nguyen et al
4. Cardon A, Aillet S, Jarno P, Bensalah K, Le Du J, Idrissi A, et al. [Endarteriectomy of the femoral tripod: long-term results and analysis of failure factors]. Ann Chir 2001;126:777-82. 5. Desai M, Tsui J, Davis M, Myint F, Wilson A, Baker DM, et al. Isolated endarterectomy of femoral bifurcation in critical limb ischemia: is restoration of inline flow essential? Angiology 2011;62:119-25. 6. Kechagias A, Ylonen K, Biancari F. Long-term outcome after isolated endarterectomy of the femoral bifurcation. World J Surg 2008;32: 51-4. 7. Ballotta E, Gruppo M, Mazzalai F, Da Giau G. Common femoral artery endarterectomy for occlusive disease: an 8-year single-center prospective study. Surgery 2010;147:268-74. 8. Derksen WJ, Verhoeven BA, van de Mortel RH, Moll FL, de Vries JP. Risk factors for surgical-site infection following common femoral artery endarterectomy. Vasc Endovascular Surg 2009;43:69-75.
JOURNAL OF VASCULAR SURGERY June 2015
9. Greenhalgh RM, Brown LC, Kwong GP, Powell JT, Thompson SG. Comparison of endovascular aneurysm repair with open repair in patients with abdominal aortic aneurysm (EVAR trial 1), 30-day operative mortality results: randomised controlled trial. Lancet 2004;364:843-8. 10. Kang JL, Patel VI, Conrad MF, Lamuraglia GM, Chung TK, Cambria RP. Common femoral artery occlusive disease: contemporary results following surgical endarterectomy. J Vasc Surg 2008;48: 872-7.
Submitted Sep 10, 2014; accepted Jan 13, 2015.
Additional material for this article may be found online at www.jvascsurg.org.
REQUEST FOR SUBMISSION OF SURGICAL ETHICS CHALLENGES ARTICLES The Editors invite submission of original articles for the Surgical Ethics Challenges section, following the general format established by Dr. James Jones in 2001. Readers have benefitted greatly from Dr. Jones’ monthly ethics contributions for more than 6 years. In order to encourage contributions, Dr. Jones will assist in editing them and will submit his own articles every other month, to provide opportunity for others. Please submit articles under the heading of “Ethics” using Editorial Manager, and follow the format established in previous issues.
JOURNAL OF VASCULAR SURGERY Volume 61, Number 6
Nguyen et al 1494.e1
Supplementary Table (online only). Risk calculator for 30-day mortalitya Predictors Age, years Nonindependent Pre-op dialysis Pre-op sepsis Emergency surgery ASA class 4 or 5 Risk ¼ Likelihood of 30-day mortality, mean 6 SD %
Predictor weight
Healthy 80 yo
Healthy 90 yo
1 7 5 5 3 2
80 0 0 0 0 0 80 3.2 6 1.6
90 0 0 0 0 0 90 7.8 6 3.6
Nonindependent 90 yo 90 1 0 0 0 0 97 32.1 6 12.2
Nonindependent 75 yo emergency 75 1 0 0 1 0 85 w5.5
ASA class, American Society of Anesthesiologists Physical Status Classification; SD, Standard deviation; yo, years old. a Instructions: To compute the risk score for death, enter the patient’s age, and 1 for each factor present. Multiple each number by the predictor weight. Take the sum and round to the nearest 10. Use this number together with Fig 3 to estimate the likelihood of 30-day mortality.
