Anesthetic Technique and Acute Kidney Injury in Endovascular Abdominal Aortic Aneurysm Repair Minjae Kim, MD,* Joanne E. Brady, SM,*† and Guohua Li, MD, DrPH*† Objective: Prior studies suggest that general anesthesia (GA) is associated with worse cardiopulmonary outcomes after endovascular abdominal aortic aneurysm repair (EVAR). Patients undergoing EVAR are at a high risk of developing perioperative acute kidney injury (AKI), and the relationship between anesthetic technique and AKI in these patients is not wellcharacterized. The authors sought to determine if anesthetic technique affected the risk of AKI in patients undergoing EVAR. Design: Retrospective, observational cohort study analyzed using a multivariate logistic regression model to assess the effects of anesthetic technique on renal outcome. Setting: Multiple institutions, mainly in North America. Participants: Patients in the American College of Surgeons National Surgical Quality Improvement Program from 2005-2010 undergoing EVAR.

T

he introduction of endovascular techniques for abdominal aortic aneurysm (AAA) repair (EVAR)1 has allowed for minimally invasive methods in the surgical management of this condition, potentially reducing postoperative complications2 and the risk of perioperative AKI3 compared to patients undergoing open repair. The minimally invasive nature of EVAR has led to the evaluation of alternative techniques to general anesthesia (GA), including local anesthesia/monitored anesthesia care (MAC) techniques,4–6 and there are suggestions that GA may increase cardiopulmonary morbidity after EVAR compared to alternatives.7 Acute kidney injury (AKI) is also a serious complication of AAA repair, leading to increased morbidity and mortality.8 Despite the reduction in the risk of AKI after EVAR compared to open repair, there is still a high risk of perioperative AKI after EVAR from renal ischemia-reperfusion injury (IRI) and surgical trauma as well as contrast nephropathy.8 The effects of anesthetic technique on renal morbidity in patients undergoing EVAR are not well studied, and here, the authors sought to evaluate the effect of anesthetic technique on renal outcome in patients undergoing this procedure. For the authors’ analysis, they obtained data from the 20052010 American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP), a large, multicenter database of surgical outcomes from hospitals throughout North America. The ACS NSQIP provides a large, high-quality9 dataset of patient characteristics as well as 30-day morbidity and mortality, providing the opportunity to characterize anesthetic management in patients undergoing EVAR as well as to investigate the hypothesis that anesthetic technique is associated with the development of perioperative AKI in patients undergoing this procedure.

*

The American College of Surgeons National Surgical Quality Improvement Program and the hospitals participating in the ACS NSQIP are the source of the data used herein; they have not verified and are not responsible for the statistical validity of the data analysis or the conclusions derived by the authors.

Interventions: The authors investigated the association between anesthetic techniques, comparing GA to alternative (non-GA) techniques, and AKI. Measurements and Main Results: AKI was defined as an increase in the creatinine level of 42 mg/dL and/or dialysis. Of 13,026 patients, 84.4% underwent GA and 15.6% underwent non-GA techniques. AKI developed in 2.0% of the GA group and 1.4% of the non-GA group (unadjusted odds ratio [OR] 1.43, p ¼ 0.075; adjusted OR [aOR] 1.00, p ¼ 0.99). Risk factors for AKI include ASA class, ruptured aneurysm, preoperative renal dysfunction, symptomatic cardiovascular disease, and perioperative blood transfusion. Conclusions: Anesthetic technique is not independently associated with the risk of AKI in patients undergoing EVAR. & 2013 Elsevier Inc. All rights reserved. KEY WORDS: acute kidney injury, anesthesia, abdominal aortic aneurysm, endovascular procedures, postoperative complications METHODS The Columbia University Medical Center Institutional Review Board determined that this study was not subject to review as the dataset did not contain individually identifiable health information. The ACS NSQIP* is a validated, prospectively collected national dataset aimed at improving surgical quality and outcomes.10 The data collected include demographic characteristics, presurgical comorbidities, intraoperative variables, and 30-day postoperative morbidity and mortality data. All data are reviewed carefully by each site’s surgical clinical reviewer, and centers not meeting specific criteria for quality are removed from the dataset. The systematic sampling process and criteria for maintaining the high quality of the dataset have been described previously.11 Of note, the ACS NSQIP collects data on major cases, defined as those performed under general, spinal or epidural anesthesia. However, certain procedures, including EVAR, are included regardless of the anesthetic technique utilized. The authors obtained the ACS NSQIP participant use data files for the years 2005-2010. There were 13,286 patients undergoing EVAR as their principal procedure as identified using Current Procedural Terminology (CPT; American Medical Association, Chicago, IL) codes for endovascular repair of infrarenal AAA or dissection (CPT codes 34800 [using aorto-aortic tube prosthesis]; 34802 [using modular bifurcated prosthesis– 1 docking limb]; 34803 [using modular bifurcated prosthesis–2 docking limbs]; 34804 [using unibody bifurcated prosthesis] and 34805 [using aorto-uniliac or aorto-unifemoral prosthesis]), as previously described.7 They excluded 241 patients who required preoperative dialysis or mechanical ventilation, as mechanically ventilated patients are likely to

From the *Department of Anesthesiology, Columbia University Medical Center; and †Department of Epidemiology, Columbia University Mailman School of Public Health, New York, NY. Address correspondence to Minjae Kim, MD, Columbia University Medical Center, Department of Anesthesiology, 622 West 168th Street, PH 5, Suite 505C, New York, NY 10032. E-mail: minjae.kim@ columbia.edu © 2013 Elsevier Inc. All rights reserved. 1053-0770/2601-0001$36.00/0 http://dx.doi.org/10.1053/j.jvca.2013.06.001

Journal of Cardiothoracic and Vascular Anesthesia, Vol ], No ] (Month), 2013: pp ]]]–]]]

1

2

receive GA as their anesthetic without consideration of an alternative technique. The followup period for patients participating in the ACS NSQIP is 30 days after their operation. For measuring the outcome of AKI, the authors are limited to the 2 prespecified endpoints available in the dataset: (1) Progressive renal insufficiency defined as a rise in the creatinine level 42 mg/dL above the preoperative value, and/or (2) the need for dialysis in a patient who did not require dialysis prior to the operation. Other measures of postoperative renal function are not available, including laboratory and urine output data. Anesthetic technique was identified through the ACS NSQIP dataset for patients undergoing EVAR procedures. Anesthesia type was identified from the patient's anesthesia record and coded as GA, monitored anesthesia care (MAC), spinal, epidural, regional, local, and none. The category GA will include all cases identified as using GA even if there were other techniques involved (eg, combined epidural and GA). The authors excluded cases coded as “Other” or with missing anesthetic data as they could not clearly determine the nature of the anesthetic management (n ¼ 19). For this study, they were interested mainly in comparing GA with alternative anesthetic techniques. Therefore, they combined patients receiving an anesthetic technique other than GA into a non-GA group. Patient baseline demographic and operative variables were collected directly from the ACS NSQIP dataset. Race/ethnicity was categorized as white versus non-white. Age was categorized as less than 60 years of age, between 60 and 75 years of age, or older than 75 years of age. American Society of Anesthesiologists (ASA) class was categorized into 3 groups: ASA 1-2, ASA 3, or ASA 4-5. Body mass index (BMI) was calculated and categorized into 3 groups: o25, 25 to 30, or 430. The authors determined if the aneurysm was ruptured based on the postoperative International Classification of Diseases, Ninth Revision (ICD9, Centers for Disease Control and Prevention, Hyattsville, MD) diagnosis code (ICD-9 codes 441.1, 441.3, 441.5, 441.6). The estimated glomerular filtration rate (eGFR, mL/min/1.73m2) was calculated based on the Modification of Diet in Renal Disease formula incorporating creatinine, sex, age, and race12 and categorized into 4 groups: o30, 30 to 60, 460, or missing. Missing categories for eGFR (based on creatinine) and hematocrit were included to account for the fact that missing preoperative laboratory work may itself be a prognostic indicator in predicting the risk of developing AKI. Intraoperative and postoperative variables also were collected from the dataset. Total operative time was examined as both a continuous and categoric variable. Based on the authors’ analysis of deciles of operative time, 3 categories were used: low (deciles 0-2; operative time ≤114 min), medium (deciles 3-5; operative time 115-155 min), and high (deciles 6-9; operative time ≥156 min). ACS NSQIP changed its reporting for transfusion variables in 2010. Prior to 2010, the dataset reported separately the number of red blood cell (RBC) units transfused intraoperatively and whether the patient received 44 units of RBCs in the first 72 hours of the postoperative period. Beginning in 2010, transfusion of 1 or more units in the intraoperative period or the first 72 hours of the postoperative period is reported. Patients were identified as having had an intra/postoperative transfusion if any intraoperative or postoperative transfusion was reported in the dataset. ACS NSQIP does not identify individual hospitals or characteristics of the hospitals included; as such, resident involvement was used as a proxy for teaching hospitals. The differences in preoperative patient characteristics and comorbidities between patients receiving GA and non-GA techniques were compared with the χ2 test. The authors used stepwise multivariate logistic regression modeling to identify significant covariates to be included in the analysis of anesthetic technique on AKI. A significance level of 0.12 was required to enter the model and a level of 0.25 was required to remain in the model. In a post-hoc analysis, the authors

KIM ET AL

explored the possibility that operative time modified the effect of anesthetic technique on AKI by including an interaction term in the model. Odds ratios (ORs) and 95% confidence intervals (CIs) of AKI were calculated. Statistical analysis was performed using SAS software version 9.2 (SAS Institute, Cary, NC). In all analyses, statistical significance was determined with a p value o0.05. RESULTS

The authors identified 13,026 patients in the 2005-2010 ACS NSQIP undergoing EVAR as their primary procedure who met the inclusion criteria. Of these, 10,989 (84.4%) received GA while 2,037 (15.6%) received a non-GA technique (Table 1). Among those not undergoing GA, 1,337 (65.6%) patients received a neuraxial technique (spinal/epidural/ regional) while 700 (34.4%) received MAC (including local anesthesia and none). Over time, the proportion of EVAR procedures performed under GA steadily increased from 67.1% in 2005 to 86.8% in 2010 (Table 1). There were differences in the baseline comorbidities and demographic characteristics between patients undergoing EVAR with GA compared to those receiving non-GA techniques (Table 2). The GA cohort had a greater proportion of patients who were o60 or between 60 and 75 years of age and who were ASA 4-5. In addition, there were higher rates of emergency procedures, ruptured aneurysms, current smoking, bleeding disorders, chronic anticoagulation, and significant blood transfusions in the 72 hours prior to surgery in the GA group. The non-GA cohort had a greater proportion of those 475 years of age and those who were ASA 1-2 as well as higher rates of pulmonary comorbidities (dyspnea or chronic obstructive pulmonary disease [COPD]) and resident involvement. Finally, patients with an aorto-aortic tube prosthesis (CPT 34800) had higher rates of GA (10.0% v 5.6%) while those with modular bifurcated prosthesis with 1 docking limb (CPT 34802) had higher rates of non-GA anesthesia (50.3% v 45.7%). Among patients undergoing EVAR, AKI developed in 2.0% of the GA group and in 1.4% of the non-GA group (p ¼ 0.075) (Table 3) for an unadjusted OR of 1.43 [0.96, 2.13]. However, after adjusting for patient factors, the OR of AKI associated with GA was reduced to 1.00 [0.66, 1.52] (Table 4). Independent predictors of AKI were ASA class, ruptured aneurysm, histories of angina and rest pain/gangrene, functional dependence, and intra/postoperative transfusion (Table 4). Compared to patients with an eGFR 460, those with lower eGFR (o30 or 30-60) had significantly higher odds of developing AKI. Surgical factors played a role as those procedures performed with a modular bifurcated prosthesis with 2 docking limbs (CPT 34803) had reduced odds of developing AKI compared to those performed with an aortoaortic tube prosthesis (CPT 34800) (Table 4). In addition, operative time was an important predictor of AKI as, compared to medium length operative times, low operative times were associated with significantly reduced odds of developing AKI while high operative times were associated with significantly greater odds of developing AKI. In analyzing the data, it appeared as if the effects of anesthetic technique on AKI rates might vary with operative

3

ACUTE KIDNEY INJURY IN AORTIC ANEURYSM REPAIR

Table 1. Anesthetic Technique for Endovascular Abdominal Aortic Aneurysm Repair Procedures, American College of Surgeons National Surgical Quality Improvement Program, 2005-2010 GA

2005 2006 2007 2008 2009 2010 Total†

243 809 1,765 2,407 2,800 2,965 10,989

(67.1%) (77.1%) (81.2%) (85.2%) (87.4%) (86.8%) (84.4%)

Non-GA*

119 240 408 417 404 449 2,037

(32.9%) (22.9%) (18.8%) (14.8%) (12.6%) (13.2%) (15.6%)

Abbreviations: GA, general anesthesia. *Non-GA includes spinal, epidural, regional, monitored anesthesia care, local anesthesia, and none. †p o 0.0001 Mantel-Haenszel χ2 test.

time. Indeed, stratification analysis demonstrated that in patients in the low operative time category (n ¼ 3,876), the incidence of AKI was 0.7% for those receiving GA (n ¼ 3,254) and 0.3% for those receiving non-GA techniques (n ¼ 622; p ¼ 0.30). In patients in the medium operative time category (n ¼ 3,946), the incidence of AKI was 1.0% for the GA group (n ¼ 3,295) and 2.0% for the non-GA group (n ¼ 651; p ¼ 0.03). In patients in the high operative time category (n ¼ 5,204), the incidence of AKI was 3.6% for the GA group (n ¼ 4,440) and 1.7% for the non-GA group (n ¼ 764; p o 0.01). To allow for the effects of GA to vary with operative time in the authors’ model, they included an interaction term between these variables, and this term was statistically significant in both simple (p ¼ 0.003) and adjusted models (p ¼ 0.02). In the simple model including anesthetic technique, operative time and the interaction term, the OR for GA in the low operative time category was 2.11 [0.50, 8.98], the OR for GA in the medium operative time category was 0.50 [0.26, 0.95], and the OR for GA in the high operative time category was 2.16 [1.22, 3.82]. The ORs in the adjusted model were 1.50 [0.35, 6.49], 0.43 [0.22, 0.85] and 1.48 [0.82, 2.69], respectively. Progressive renal insufficiency developed in 0.8% of patients in both the GA and non-GA groups (Table 3), for an unadjusted OR of 1.06 [0.62, 1.80] and an adjusted OR of 0.79 [0.46, 1.37]. The interaction between anesthetic technique and operative time was not significant for this outcome. Postoperative dialysis was required in 1.3% of the GA group and 0.8% in the non-GA group (Table 3), for an unadjusted OR of 1.68 [0.998, 2.82] (p ¼ 0.051) and an adjusted OR of 1.10 [0.64, 1.90]. After incorporating the interaction between GA and operative time (p ¼ 0.01 for interaction term), the adjusted ORs for GA were 0.33 [0.15, 0.74] in the medium operative time group and 2.03 [0.86, 4.78] in the high operative time group. The OR for the low operative time group could not be calculated because there were no cases of postoperative dialysis in the non-GA group. DISCUSSION

This study of a large, national surgical outcomes dataset from 2005-2010 demonstrated that the use of GA for patients undergoing

the EVAR procedure has been increasing over time. Patients receiving GA appeared to be a sicker group as evidenced by a greater proportion of patients with high ASA class, higher rates of smoking, prior myocardial infarctions, bleeding disorders, and preoperative transfusion requirements as well as a greater likelihood of undergoing an emergency procedure with a ruptured aneurysm. Those receiving non-GA techniques were older and had higher rates of pulmonary comorbidities (dyspnea and COPD). Overall, GA was associated with a higher (although not statistically significant) rate of perioperative AKI compared to non-GA techniques (crude OR ¼ 1.43, p ¼ 0.076). However, this increased risk can be explained by patient comorbidities and operative factors because after adjustment, there was no association between GA and AKI (adjusted OR ¼ 1.00, p ¼ 0.99). Thus, GA does not appear to be associated with the risk of developing AKI after EVAR. Several risk factors for the development of AKI in EVAR patients were identified. Not surprisingly, preoperative renal function was a significant predictor of AKI as well as ASA class, ruptured aneurysm, functional dependence, and blood transfusions. Interestingly, symptomatic cardiovascular diseases (angina and rest pain/gangrene) were significant predictors of AKI, but prior revascularizations (coronary and peripheral) were not. The authors cannot determine here if the risk of AKI would decrease if these patients delayed their procedure for revascularization therapy, and this is a potential area for further study. However, it is possible that they presented for urgent/emergent procedures that necessitated immediate surgical intervention of the aneurysm. Increasing operative time was also a significant predictor of AKI. It is possible that the length of the procedure itself affects AKI rates directly, but it more likely serves as a useful proxy for factors that increase AKI risk, such as increased surgical complexity. Though the utilization of GA increased over time, there was no significant change in the rate of AKI over time (not shown). In addition, certain risk factors for AKI identified in general surgical patients,13 including ascites, congestive heart failure, hypertension, male sex, diabetes, and age, were not significant predictors in the authors’ model. While preoperative anemia is a risk factor for AKI in patients undergoing cardiac surgery,14 it is not significant in their analysis. It is possible that there are unique factors associated with anemia and cardiopulmonary bypass that are not present in those undergoing EVAR. It might be surmised that non-GA techniques would offer clear, significant benefits over GA. Indeed, GA was associated with higher rates of AKI compared to regional anesthesia in patients undergoing total joint arthroplasty.15 On the other hand, local anesthesia has not been shown to reduce postoperative stroke or death after carotid endarterectomy compared to GA,16 and there are no clinically important differences between regional anesthesia and GA after hip fracture surgery.17 Anesthetic technique also may affect other aspects of care, such as postoperative cognitive function, pain control, quality of life, and costs associated with healthcare.18 Although an early review found no differences in cardiopulmonary morbidity and mortality between GA or local anesthesia after EVAR,19 several reports suggest that alternatives to GA are preferable for EVAR,20–22 including a recent study demonstrating that GA was associated with an increased risk of cardiac complications compared to locoregional anesthesia.23 The impact

4

KIM ET AL

Table 2. Characteristics of Patients Undergoing Endovascular Abdominal Aortic Aneurysm Repair by Anesthetic Technique, American College of Surgeons National Surgical Quality Improvement Program, 2005-2010

Age

Female White ASA class

Emergency Ruptured aneurysm Functionally dependent Body mass index

Estimated glomerular filtration rate (mL/min/1.73m2)

Current smoker Dyspnea Chronic obstructive pulmonary disease Current pneumonia Diabetic Angina Hypertension Coronary revascularization (PCI or CABG) Congestive heart failure Myocardial infarction PVD requiring intervention Rest pain/gangrene Stroke (with or without neuro deficit) Hematocrit (%)

Bleeding disorders Transfusions 44 Units 72 hours prior to surgery Preop SIRS/sepsis/septic shock

o60 60-75 475 Yes No Yes No 1-2 3 4-5 Yes No Yes No Yes No o25 25-30 430 o30 30-60 460 Missing Yes No Yes No Yes No Yes No Yes No Yes No Yes No Yes No Yes No Yes No Yes No Yes No Yes No 430 ≤30 Missing Yes No Yes No Yes No

GA

Non-GA

n ¼ 10,989

n ¼ 2,037

627 5,349 5,013 1,968 9,021 9,177 1,812 680 7,895 2,408 697 10,292 563 10,426 549 10,436 2,896 4,142 3,435 365 3,900 6,376 348 3,335 7,654 2,539 8,450 2,044 8,945 18 10,971 1,658 9,331 208 10,781 8,715 2,274 4,192 6,797 145 10,844 135 10,854 634 10,355 123 10,866 1,627 9,362 10,203 448 338 1,309 9,680 43 10,946 252 10,602

(5.7) (48.7) (45.6) (17.9) (82.1) (83.5) (16.5) (6.2) (71.9) (21.9) (6.3) (93.7) (5.1) (94.9) (5.0) (95.0) (27.0) (38.7) (32.1) (3.3) (35.5) (58.0) (3.2) (30.4) (69.7) (23.1) (76.9) (18.6) (81.4) (0.2) (99.8) (15.1) (84.9) (1.9) (98.1) (79.3) (20.7) (38.2) (61.9) (1.3) (98.7) (1.2) (98.8) (5.8) (94.2) (1.1) (98.9) (14.8) (85.2) (92.9) (4.1) (3.1) (11.9) (88.1) (0.4) (99.6) (2.3) (97.7)

86 932 1,019 358 1,679 1,669 368 214 1,418 405 77 1,960 70 1,967 90 1,947 546 787 586 54 750 1,137 96 543 1,494 536 1,501 454 1,583 7 2,030 282 1,755 41 1,996 1,594 443 765 1,272 35 2,002 14 2,023 116 1,921 15 2,022 298 1,739 1,856 68 113 142 1,895 2 2,035 40 1,979

(4.2) (45.8) (50.0) (17.6) (82.4) (81.9) (18.1) (10.5) (69.6) (19.9) (3.8) (96.2) (3.4) (96.6) (4.4) (95.6) (27.7) (40.0) (29.8) (2.7) (36.8) (55.8) (4.7) (26.7) (73.3) (26.3) (73.7) (22.3) (77.7) (0.3) (99.7) (13.8) (86.2) (2.0) (98.0) (78.3) (21.8) (37.6) (62.4) (1.7) (98.3) (0.7) (99.3) (5.7) (94.3) (0.7) (99.3) (14.6) (85.4) (91.1) (3.3) (5.6) (7.0) (93.0) (0.1) (99.9) (2.0) (98.0)

P*

o0.001†

0.718 0.080 o0.0001†

o0.0001† 0.001† 0.266 0.131

0.001†

0.001† 0.002† 0.0001† 0.089 0.148 0.717 0.282 0.613 0.157 0.035† 0.894 0.121 0.837 o0.0001†

o0.0001† 0.038† 0.345

5

ACUTE KIDNEY INJURY IN AORTIC ANEURYSM REPAIR

Table 2 (continued )

Resident involvement

Yes No

Current procedural terminology Group Using aortoaortic tube prosthesis (34800) Using modular bifurcated prosthesis [1 docking limb] (34802) Using modular bifurcated prosthesis [2 docking limbs] (34803) Using unibody bifurcated prosthesis (34804) Using aortouniliac or aorto-unifemoral prosthesis (34805)

GA

Non-GA

n ¼ 10,989

n ¼ 2,037

P*

7,113 (64.8) 3,867 (35.2)

1,405 (69.2) 625 (30.8)

o0.001†

1,099 5,021 3,556 787 526

113 1,024 642 137 121

o0.0001†

(10.0) (45.7) (32.4) (7.2) (4.8)

(5.6) (50.3) (31.5) (6.7) (5.9)

NOTE. Categoric variables expressed as number (%). Abbreviations: ASA, American Society of Anesthesiologists; CABG, coronary artery bypass graft; GA, General Anesthesia; PCI, percutaneous coronary intervention; PVD, peripheral vascular disease; SIRS, systemic inflammatory response syndrome. *Differences among categoric variables analyzed with χ2 test. †p o 0.05.

of anesthetic technique on renal outcomes in EVAR patients has not been defined clearly. In addition, there are limitations to nonGA techniques in lengthier procedures as patients may not tolerate local/MAC or neuraxial techniques for extended periods of time. Indeed, of the 67 cases with operative times 4500 min, only 1 was performed with a non-GA technique. However, a sensitivity analysis demonstrated that these extremes of operative time did not affect the main result (not shown). A unique aspect of GA involves the use of volatile anesthetics (VAs) and they have been shown to reduce the risk of AKI in animal24 and in vitro25 models. The clinical data focus on cardiac surgery where sevoflurane preconditioning reduced levels of the renal biomarker cystatin c26 but did not affect creatinine levels,27 postoperative oliguria, or dialysis rates.28 In other organ systems, VAs have been shown to be protective in the heart29,30 and liver.31,32 However, some recent studies also demonstrated a lack of VA-mediated protection, including a lack of cardioprotection in noncardiac surgery33 and renal protection in cardiac surgery.28 Here, the authors are unable to determine the proportion of patients receiving VAs for the maintenance of GA, and this is an issue that will require further study. As the anesthetic decision was not randomized in the retrospective cohort study, the authors are concerned about possible selection bias that could influence the outcome. In order to account for this potential confounding, they conducted separate analyses (not shown) using propensity score methods34 to determine the likelihood of receiving GA, and the findings were consistent with their main analysis. However, even the use of propensity score methods cannot account for variables influencing the anesthetic decision that are not captured in the dataset, such as factors related to surgeons, anesthesiologists, and institutions as well as vascular anatomy, and this remains a potential source of bias. Despite the lack of an overall association between GA and AKI in the authors’ study, an interesting post-hoc analysis demonstrated that the effect of anesthetic technique on the risk of AKI varied by the length of the procedure as indicated by operative time. Using a model allowing for the effect of GA on AKI to vary with operative time, the authors found a significant association between GA and reduced AKI rates in those with a moderate length of procedure (115-155 min), indicating that

there may be a subset of patients in whom anesthetic technique affects the risk of AKI. Further study is necessary to explore and validate these findings. In the present study, those receiving a technique other than GA were combined into 1 group, as the authors mainly were interested in evaluating the effects of GA on AKI. When separating the non-GA group into a neuraxial anesthesia group and an MAC group, they again found no significant associations between anesthetic technique and AKI, although there was a lack of statistical power to detect a difference among the 3 groups. Although the authors recognize that there are differences between neuraxial and MAC anesthesia, based on their analysis, they felt that combining them into one non-GA group was appropriate for the purposes of this study. In addition to limitations regarding retrospective analyses of large medical datasets,35 there are some additional important limitations to the ACS NSQIP relevant to this study. The dataset defines progressive renal insufficiency as an increase in the postoperative creatinine level of 42 mg/dL although creatinine changes as small as 0.3 mg/dL may constitute clinically significant AKI.36,37 As such, there is likely an underreporting of the total rate of AKI by accounting only for severe creatinine changes following surgery. Many important variables affecting the risk of AKI are not collected, such as intraoperative hemodynamics,38 nephrotoxic radiocontrast dye, and fenestrated grafts.39 In addition, conversions from one technique to another (eg, MAC to GA) are not captured. Despite these limitations, the

Table 3. Renal Outcomes by Anesthetic Technique, American College of Surgeons National Surgical Quality Improvement Program, 2005-2010 GA 10,989

Acute Kidney Injury Renal Insufficiency Dialysis

215 (2.0) 91 (0.8) 144 (1.3)

Non-GA 2,037

28 (1.4) 16 (0.8) 16 (0.8)

Total 13,026

243 (1.9) 107 (0.8) 160 (1.2)

NOTE. Categoric variables expressed as number (%). Abbreviations: GA, General Anesthesia *Differences analyzed with χ2 test. †p o 0.05.

P*

0.075 0.845 0.048†

6

KIM ET AL

Table 4. Logistic Regression of Acute Kidney Injury by Anesthetic Technique, American College of Surgeons National Surgical Quality Improvement Program, 2005-2010 Adjusted Model* OR

General anesthesia ASA class (ref ¼ ASA 3) ASA 1-2 ASA 4-5 Ruptured aneurysm eGFR (mL/min/1.73m2) (ref ¼ 460) o30 30-60 Missing Angina Rest pain/gangrene Functionally dependent Chronic obstructive pulmonary disease Hematocrit (%) (ref ¼ ≤30) 430 Missing Resident involvement Intra/postoperative transfusion Operative time (ref ¼ Medium [115-154 min]) Low [≤114 min] High [≥155 min] CPT (Ref ¼ Using aorto-aortic tube prosthesis [34800]) Using modular bifurcated prosthesis [1 docking limb] (34802) Using modular bifurcated prosthesis [2 docking limbs] (34803) Using unibody bifurcated prosthesis (34804) Using aorto-uniliac or aorto-unifemoral prosthesis (34805) c-statistic Hosmer-Lemeshow p-value

95% CI

1.00

[0.66, 1.52]

0.74 1.50† 2.04†

[0.30, 1.84] [1.10, 2.04] [1.39, 3.00]

13.1† 2.91† 1.86 2.36† 2.84† 1.97† 1.33

[8.46, [2.10, [0.57, [1.21, [1.36, [1.35, [0.97,

20.2] 4.04] 6.09] 4.63] 5.93] 2.87] 1.82]

1.50 0.80 0.72† 5.41†

[0.97, [0.21, [0.53, [3.97,

2.33] 3.03] 0.96] 7.39]

0.59† 1.88†

[0.35, 0.99] [1.32, 2.69]

0.68 0.57* 0.69 0.86 0.876 0.681

[0.45, [0.37, [0.37, [0.49,

1.03] 0.89] 1.31] 1.49]

Abbreviations: ASA, American Society of Anesthesiologists; CI, confidence interval; CPT, Current Procedural Terminology; eGFR, estimated glomerular filtration rate; OR, odds ratio. *Multivariable model excludes 26 records with missing values in 1 or more covariates in the model. †p o 0.05.

dataset still provides valuable insight of the effects of anesthetic technique on perioperative AKI. The authors have demonstrated using a large-scale national dataset that, in patients undergoing EVAR, GA is not associated with the risk of AKI compared to alternative anesthetic techniques and that the utilization of GA for this procedure has been increasing over time. Understanding the clinical implications of anesthetic decisions on perioperative

outcomes is critical as anesthesiologists continue to care for the increasing number of patients presenting for endovascular management of their AAAs.

ACKNOWLEDGMENT The authors would like to thank Charles W. Emala, MD for valuable assistance with manuscript preparation.

REFERENCES 1. Parodi JC, Palmaz JC, Barone HD: Transfemoral intraluminal graft implantation for abdominal aortic aneurysms. Ann Vasc Surg 5: 491-499, 1991 2. Prinssen M, Verhoeven ELG, Buth J, et al: A randomized trial comparing conventional and endovascular repair of abdominal aortic aneurysms. N Engl J Med 351:1607-1618, 2004 3. Wald R, Waikar SS, Liangos O, et al: Acute renal failure after endovascular vs open repair of abdominal aortic aneurysm. J Vasc Surg 43:460-466, 2006 4. Verhoeven ELG, Cina CS, Tielliu IFJ, et al: Local anesthesia for endovascular abdominal aortic aneurysm repair. J Vasc Surg 42: 402-409, 2005

5. Verhoeven ELG, Prins TR, van den Dungen JJAM, et al: Endovascular repair of acute AAAs under local anesthesia with bifurcated endografts: A feasibility study. J Endovasc Ther 9:729-735, 2002 6. Henretta JP, Hodgson KJ, Mattos MA, et al: Feasibility of endovascular repair of abdominal aortic aneurysms with local anesthesia with intravenous sedation. J Vasc Surg 29:793-798, 1999 7. Edwards MS, Andrews JS, Edwards AF, et al: Results of endovascular aortic aneurysm repair with general, regional, and local/ monitored anesthesia care in the American College of Surgeons National Surgical Quality Improvement Program database. J Vasc Surg 54:1273-1282, 2011

ACUTE KIDNEY INJURY IN AORTIC ANEURYSM REPAIR

8. Walsh SR, Tang TY, Boyle JR: Renal consequences of endovascular abdominal aortic aneurysm repair. J Endovasc Ther 15: 73-82, 2008 9. Shiloach M, Frencher SK Jr, Steeger JE, et al: Toward robust information: Data quality and inter-rater reliability in the American College of Surgeons National Surgical Quality Improvement Program. J Am Coll Surg 210:6-16, 2010 10. Fink AS, Campbell DA Jr, Mentzer RM Jr, et al: The National Surgical Quality Improvement Program in non-veterans administration hospitals: Initial demonstration of feasibility. Ann Surg 236:344-353, 2002 11. Hua M, Brady J, Li G: The epidemiology of upper airway injury in patients undergoing major surgical procedures. Anesth Analg 114: 148-151, 2012 12. Levey AS, Coresh J, Greene T, et al: Using standardized serum creatinine values in the modification of diet in renal disease study equation for estimating glomerular filtration rate. Ann Intern Med 145: 247-254, 2006 13. Kheterpal S, Tremper KK, Heung M, et al: Development and validation of an acute kidney injury risk index for patients undergoing general surgery: Results from a national data set. Anesthesiology 110: 505-515, 2009 14. Karkouti K, Wijeysundera DN, Yau TM, et al: Acute kidney injury after cardiac surgery: focus on modifiable risk factors. Circulation 119:495-502, 2009 15. Weingarten TN, Gurrieri C, Jarett PD, et al: Acute kidney injury following total joint arthroplasty: retrospective analysis. Can J Anaesth 59:1111-1118, 2012 16. Rerkasem K, Rothwell PM: Local versus general anaesthesia for carotid endarterectomy. Cochrane Database Syst Rev CD000126, 2008 17. Parker MJ, Handoll HHG, Griffiths R: Anaesthesia for hip fracture surgery in adults. Cochrane Database Syst Rev CD000521, 2004 18. Bodenham AR, Howell SJ: General anaesthesia vs local anaesthesia: An ongoing story. Br J Anaesth 103:785-789, 2009 19. de Virgilio C, Romero L, Donayre C, et al: Endovascular abdominal aortic aneurysm repair with general versus local anesthesia: A comparison of cardiopulmonary morbidity and mortality rates. J Vasc Surg 36:988-991, 2002 20. Bettex DA, Lachat M, Pfammatter T, et al: To compare general, epidural and local anaesthesia for endovascular aneurysm repair (EVAR). Eur J Vasc Endovasc Surg 21:179-184, 2001 21. Geisbusch P, Katzen BT, Machado R, et al: Local anaesthesia for endovascular repair of infrarenal aortic aneurysms. Eur J Vasc Endovasc Surg 42:467-473, 2011 22. Ruppert V, Leurs LJ, Steckmeier B, et al: Influence of anesthesia type on outcome after endovascular aortic aneurysm repair: An analysis based on EUROSTAR data. J Vasc Surg 44:16-21, 2006 23. Bakker EJ, van de Luijtgaarden KM, van Lier F, et al: General anaesthesia is associated with adverse cardiac outcome after endovascular aneurysm repair. Eur J Vasc Endovasc Surg 44:121-125, 2012

7

24. Lee HT, Ota-Setlik A, Fu Y, et al: Differential protective effects of volatile anesthetics against renal ischemia-reperfusion injury in vivo. Anesthesiology 101:1313-1324, 2004 25. Kim M, Kim M, Park SW, et al: Isoflurane protects human kidney proximal tubule cells against necrosis via sphingosine kinase and sphingosine-1-phosphate generation. Am J Nephrol 31:353-362, 2010 26. Julier K, da Silva R, Garcia C, et al: Preconditioning by sevoflurane decreases biochemical markers for myocardial and renal dysfunction in coronary artery bypass graft surgery: A double-blinded, placebo-controlled, multicenter study. Anesthesiology 98:1315-1327, 2003 27. Lorsomradee S, Cromheecke S, De Hert SG: Effects of sevoflurane on biomechanical markers of hepatic and renal dysfunction after coronary artery surgery. J Cardiothorac Vasc Anesth 20:684-690, 2006 28. Sindhvananda W, Phisaiphun K, Prapongsena P: No renal protection from volatile-anesthetic preconditioning in open heart surgery. J Anesth 27:48-55, 2013 29. Belhomme D, Peynet J, Louzy M, et al: Evidence for preconditioning by isoflurane in coronary artery bypass graft surgery. Circulation 100:II-340-II-344, 1999 30. Landoni G, Biondi-Zoccai GGL, Zangrillo A, et al: Desflurane and sevoflurane in cardiac surgery: A meta-analysis of randomized clinical trials. J Cardiothorac Vasc Anesth 21:502-511, 2007 31. Beck-Schimmer B, Breitenstein S, Urech S, et al: A randomized controlled trial on pharmacological preconditioning in liver surgery using a volatile anesthetic. Ann Surg 248:909-918, 2008 32. Ko JS, Gwak MS, Choi SJ, et al: The effects of desflurane and propofol-remifentanil on postoperative hepatic and renal functions after right hepatectomy in liver donors. Liver Transpl 14:11501158, 2008 33. Lurati Buse GAL, Schumacher P, Seeberger E, et al: Randomized comparison of sevoflurane versus propofol to reduce perioperative myocardial ischemia in patients undergoing noncardiac surgery. Circulation 126:2696-2704, 2012 34. Williamson E, Morley R, Lucas A, et al: Propensity scores: From naive enthusiasm to intuitive understanding. Stat Methods Med Res 21: 273-293, 2012 35. Ward RA, Brier ME: Retrospective analyses of large medical databases: What do they tell us? J Am Soc Nephrol 10:429-432, 1999 36. Mehta RL, Kellum JA, Shah SV, et al: Acute Kidney Injury Network: Report of an initiative to improve outcomes in acute kidney injury. Crit Care 11:R31, 2007 37. Bihorac A, Yavas S, Subbiah S, et al: Long-term risk of mortality and acute kidney injury during hospitalization after major surgery. Ann Surg 249:851-858, 2009 38. Tallgren M, Niemi T, Poyhia R, et al: Acute renal injury and dysfunction following elective abdominal aortic surgery. Eur J Vasc Endovasc Surg 33:550-555, 2007 39. Brooks CE, Middleton A, Dhillon R, et al: Predictors of creatinine rise post-endovascular abdominal aortic aneurysm repair. ANZ J Surg 81:827-830, 2011