Predictors of Renal Dysfunction following Endovascular and Open Repair of Abdominal Aortic Aneurysms

Abstract

Objectives

Renal complications following repair of abdominal aortic aneurysms (AAA) have been associated with increased morbidity and mortality. However, limited data have assessed risk factors for renal complications in the endovascular era. This study aims to identify predictors of renal complications following endovascular (EVAR) and open repair.

Methods

Patients who underwent EVAR or open repair of a non-ruptured infrarenal AAA between 2011 and 2013 were identified in the Targeted Vascular module of the National Surgical Quality Improvement Project. Patients on hemodialysis preoperatively were excluded. Renal complications were defined as new postoperative dialysis or creatinine increase greater than 2mg/dL. Patient demographics, comorbidities, glomerular filtration rate (GFR), operative details, and outcomes were compared using univariate analysis between those with and without renal complications. Multivariable logistic regression was utilized to identify independent predictors of renal complications.

Results

We identified 4503 patients who underwent elective repair of infrarenal AAA (EVAR: 3869, Open: 634). Renal complication occurred in 1% of patients following EVAR and 5% of patients following open repair. There were no differences in comorbidities between patients with and without renal complications. A preoperative GFR < 60 occurred more frequently among patients with renal complications (EVAR: 81% vs. 37%, P < .01; Open: 60% vs. 34%, P < .01). 30-day mortality was also significantly increased (EVAR: 55% vs. 1% P < .01; Open: 30% vs. 4% P < .01). After adjustment, renal complications were strongly associated with 30-day mortality (Odds Ratio (OR): 38.3 95% Confidence Interval (CI): 20.4–71.9). Independent predictors of renal complications included: GFR < 60 (OR: 4.6, 95% CI: 2.4–8.7), open repair (OR: 2.6, 95% CI: 1.3–5.3), transfusion (OR: 6.1, 95% CI: 3.0–12.6), and prolonged operative time (OR: 3.0, 95% CI: 1.6–5.6).

Conclusion

Predictors of renal complications include elevated baseline GFR, open approach, transfusion, and prolonged operative time. Given the dramatic increase in mortality associated with renal complications, care should be taken to employ renal protective strategies, achieve meticulous hemostasis to limit transfusions, and to utilize an endovascular approach when technically feasible.

Introduction

Renal complications following surgery are associated with increased mortality, prolonged hospital length of stay, and higher healthcare costs.^1–3 Following open repair of abdominal aortic aneurysms (AAA), both 30-day and long-term mortality have been strongly associated with postoperative renal complications.⁴ Prior work has shown predictors of renal complications after AAA repair include pre-operative kidney dysfunction and chronic obstructive pulmonary disease.⁴ Additional operative factors such as urgency of presentation, supra-renal clamping, and operative time have also been associated with renal complications.^{2, 4} Despite these findings, previous studies have included predominantly open aneurysm repairs with varying proximal extent of aneurysms and operative urgency. The effects of renal dysfunction on mortality in the endovascular era, and the predictors of such complications following endovascular repair (EVAR) remain unclear.

Therefore, this study aims to identify the rate of post-operative renal complications and subsequent mortality associated with this adverse event among patients undergoing EVAR and open repair of intact infrarenal aneurysms in the endovascular era. Additionally, we intend to identify predictors of renal dysfunction among patients.

Methods

Patients

The Targeted Vascular Module of the American College of Surgeons National Surgical Quality Improvement Program (NSQIP) was utilized to identify all patients undergoing elective repairs for intact infrarenal AAAs from 2011–2014. Patients with juxtarenal, pararenal, and suprarenal AAAs were excluded to minimize the effect of clamp time. Additionally, those on dialysis pre-operatively were excluded from this analysis (n=88). The targeted NSQIP is a national clinical registry developed in 2011, which collects patient demographics, operative details, and 30-day outcomes from patients undergoing surgical procedures at more than 65 self-selected hospitals. Further information is available at www.facs.org/quality-programs/acs-nsqip.

Variables

Patient demographics, age, and comorbid conditions were compared between those with and without renal complications. Smoking was defined as current tobacco use. Glomerular filtration rate (GFR - mL/minute per 1.73m²) was calculated in accordance with the Modification of Diet in Renal Disease (MDRD) equation, and chronic kidney disease was identified according the Kidney Disease: Improving Global Outcomes (KDIGO) and Acute Kidney Injury Network (AKIN) Clinical Practice Guidelines.^5–7

The operative variables compared are listed in Table II and include: aneurysm diameter, transfusion, renal revascularization, and lower extremity revascularization, as defined by NSQIP. Transfusion was defined as any transfusion within 72 hours of the initial operation. Prolonged operative time was defined as greater than 2 standard deviations from the mean (greater than 180 minutes for EVAR and greater than 360 minutes for open repair).

Table II.

Operative Characteristics

Outcome	EVAR			Open
	No Renal Complication N=3836	Renal Complication N=33	P-Value	No Renal Complication N=604	Renal Complication N=30	P-Value
Operative Time: min, median (IQR)	132 (103–171)	183 (130–275)	< .01	221 (168–285)	374 (220–457)	<.01
Diameter: cm, median (IQR)	5.5 (5.1–6.0)	5.8 (5.1–7.7)	0.15	5.8 (5.2–6.7)	6.6 (5.8–8.6)	<.01
Transfusion	373 (10%)	24 (72%)	<.01	417 (69%)	23 (77%)	0.38
Renal Revascularization	170 (4%)	4 (12%)	0.06	25 (4%)	2 (7%)	0.50
LE Revascularization	138 (4%)	4 (13%)	0.04	39 (7%)	5 (17%)	0.04

IQR: Interquartile Range, cm: centimeters, LE: Lower Extremity

All outcomes measured occurred within 30-days of operation. A renal complication was defined as a creatinine increase greater than 2 mg/dL from baseline or new dialysis in the 30-day post-operative period, as defined by NSQIP. A pulmonary complication was defined as pneumonia, failure to wean from mechanical ventilation within 48 hours, re-intubation, or pulmonary embolism. Prolonged length of stay was defined as great than 2 days following EVAR and greater than 7 days following open repair.

Statistics

All statistical analyses were performed using the SPSS statistical package (version 21.0). Univariate analysis was stratified by EVAR or open repair, and patients with and without renal complications were compared using chi-square and Fisher’s exact tests for categorical variables, as appropriate. The Student’s t-test and Mann-Whitney U-test were utilized to assess continuous variables, as appropriate. All pre-operative variables and outcomes compared had less than 2% missing data, with the exception of lower extremity revascularization (10.9% missing data). Independent predictors of renal complications, mortality, and prolonged length of stay were established using multivariable logistic regression. Purposeful selection was utilized to select variables for inclusion.⁸ This included all variables with P < 0.1 on univariate analysis as well as those variables shown to be predictive of the outcome of interest in previous studies. The Hosmer-Lemeshow goodness of fit test was used to evaluate each model. A P-value < 0.05 was considered significant. The institutional review board of Beth Israel Deaconess Medical Center approved this study and consent was waived due to the de-identified nature of the NSQIP database.

Results

We identified 4503 patients, 3869 of whom underwent EVAR and 634 had open repair. Renal complications, as defined by NSQIP, occurred in 33 patients (1%) following EVAR and 30 patients (5%) after open repair. Dialysis was initiated in 22 (0.6%) patients following EVAR and 26 (4%) of patients following open repair.

Baseline Characteristics

Among those treated with EVAR, patients with renal complications were older (80 years vs. 75 years, P = 0.01), less commonly male (67% vs. 81%, P = 0.03), and more commonly had a GFR < 60 (81% vs. 37%, P < .01). Among patients undergoing open repair, only GFR < 60 differed between patients with and without renal complications (60% vs. 34%, P < .01, respectively) (Table I).

Table I.

Baseline Demographics and Comorbidities

Outcome	EVAR			Open
	No Renal Complication N=3836	Renal Complication N=33	P-Value	No Renal Complication N=604	Renal Complication N=30	P-Value
Age: median (SD)	75 (8.6)	80 (9.3)	0.01	70 (9.3)	69 (10.3)	0.46
Male Gender	3121 (81%)	22 (67%)	0.03	451 (75%)	24 (80%)	0.51
White Race	3318 (87%)	27 (82%)	0.43	484 (80.1)	25 (83%)	0.67
GFR < 60	1435 (37%)	27 (81%)	<.01	203 (34%)	18 (60%)	<.01
Diabetes	616 (16%)	6 (18%)	0.74	693 (16%)	8 (13%)	0.53
COPD	106 (18%)	6 (20%)	0.81	796 (18%)	15 (24%)	0.23
CHF	9 (2%)	1 (3%)	0.39	67 (2%)	1 (2%)	0.96
Hypertension	493 (82%)	25 (83%)	0.81	3590 (81%)	55 (87%)	0.20
Smoking	259 (43%)	14 (47%)	0.71	1410 (32%)	26 (41%)	0.11

SD: Standard Deviation

Operative Characteristics

Following EVAR, patients with renal complications had longer operative times (183 minutes vs. 132 minutes, P < .01), more lower extremity revascularizations (13% vs. 4% P < .01) and transfusions (70% vs. 10%, P < .01). There were no significant differences in AAA diameter or proportion of patients undergoing renal revascularization.

Following open repair, patients with renal complications had longer operative times (374 minutes vs. 221 minutes, P <. 01), larger AAA diameters (6.6 cm vs. 5.8 cm, P < .01), and more lower extremity revascularizations (15% vs. 5%, P = 0.03). There were no differences in transfusions or proportion with concurrent renal revascularization (Table II).

Outcomes

Both morbidity and mortality increased among patients with renal complications. Among EVAR patients, 30-day mortality was 55% in patients with renal complications compared to 1% without renal complications (P < .01). Major complications, including myocardial infarction (21% vs. 1%, P < .01), pulmonary complications (49% vs. 2%, P < .01), ischemic colitis (15% vs. 0.3%, P < .01), and lower extremity ischemia (15% vs. 1%, P < .01) were also more common among patients with renal complications. Median hospital stay was 8 days among patients with renal complications and 2 days among those without (P < .01).

Following open repair, 30-day mortality was 30% among those with renal complications and 4% among those without (P < .01). Similar to EVAR, pulmonary complications (80% vs. 13%, P < .01), ischemic colitis (23% vs. 2%, P < .01), and lower extremity ischemia (17% vs. 2%, P < .01) were increased among patients with renal complications. The median hospital stay was 19 days among patients with renal complications compared to 7 days among patients without (P < .01) (Table III).

Table III.

Univariate Outcomes

Outcome Number (%)	EVAR			Open
	No Renal Complication N=3836	Renal Complication N=33	P-Value	No Renal Complication N=604	Renal Complication N=30	P-Value
30-day Mortality	38 (1%)	18 (55%)	<.01	23 (4%)	9 (30%)	<.01
Pulmonary Complication	73 (2%)	16 (49%)	<.01	81 (13%)	24 (80%)	<.01
Ischemic Colitis	13 (0.3%)	5 (15%)	<.01	10 (2%)	7 (23%)	<.01
Lower Extremity Ischemia	45 (1%)	5 (15%)	<.01	11 (2%)	5 (17%)	<.01
Myocardial Infarction	46 (1%)	7 (21%)	<.01	12 (2%)	1 (3%)	0.47
Re-operation	22 (1%)	1 (3%)	0.18	12 (2%)	1 (3%)	0.47
Hospital Stay: median (IQR)	2 days (1–3)	8 days (4–20)	<.01	7 days (5–9)	19 days (12–29)	<.01

IQR: Interquartile Range

In multivariable analysis, adjusting for patient demographics, comorbidities, and operative approach, renal complications were predictive of both 30-day mortality (Odds Ratio (OR): 38.3 95% Confidence Interval (CI): 20.4–71.9) and prolonged length of stay (OR: 8.3, CI: 4.2–16.4).

Predictors of Renal Complications

Following multivariable adjustment for only those characteristics available to surgeons pre-operatively, GFR < 60 (OR: 5.7, 95% CI: 3.0–10.6), AAA diameter (OR: 1.1, 95% CI: 1.02–1.2), and open repair (OR: 6.0, 95% CI: 3.5–10.3) were predictive of renal complications. When intraoperative characteristics were added to the same model, GFR < 60 (OR: 4.6, 95% CI: 2.4–8.7), open repair (OR: 2.6, 95% CI: 1.3–5.3), transfusion (OR: 6.1, 95% CI: 3.0–12.6), and prolonged operative time (OR: 3.0, 95% CI: 1.6–5.6) were independently predictive of renal complications (Table IV).

Table IV.

Multivariable Predictors of Renal Complications

	Odds Ratio	95% Confidence Interval	P-Value
Open Repair	2.6	1.3–5.3	<.01
GFR < 60	4.6	2.4–8.7	<.01
Transfusion	6.1	3.0–12.6	<.01
Prolonged Operative Time	3.0	1.6–5.6	<.01
Age (decade)	0.9	0.7–1.3	0.63
Female gender	1.0	0.5–1.9	0.93
CHF	0.7	0.1–5.3	0.71
COPD	1.4	0.7–2.7	0.33
Diabetes	0.9	0.4–2.0	0.73
Diameter	1.0	0.9–1.1	0.23
Renal Revascularization	1.4	0.5–3.9	0.48
Lower Extremity Revascularization	1.2	0.5–3.0	0.73

Discussion

This study found that post-operative renal complications, defined as an increase in creatinine of 2.0mg/dL from baseline or new dialysis, occur in 1% of elective infra-renal EVARs and 5% of open repairs and are associated with a significant increase in mortality, morbidity, and prolonged length of stay compared to those patients without renal complications. Moreover, a baseline GFR < 60, open operative approach, transfusion, and prolonged operative time are independently predictive of renal complications.

The reported rates of renal complications vary in current literature. Following open repair, reported rates have ranged from 5–11%.^{2, 4, 9} Lower rates have been reported following EVAR, occurring in 2–7% of patients. ^9–11 Our study found a similar rate of renal complications following open repair to that reported by Grant et al., in a study of 2347 consecutive repairs, (6%) but was lower than other prior studies.² This variation was likely due to differences in study population and the definition of renal dysfunction. Our study evaluated infrarenal aneurysms only, with a renal complication defined by NSQIP as an increase in creatinine > 2.0mg/dL from baseline or new onset dialysis. This differs from previous work by both Patel and Ellenberg, who had a less stringent definition of renal complications, defined as all those with a creatinine increase > 0.5 mg/dL, and included all elective open repairs including those utilizing a suprarenal clamp, which is known to be independently associated with increased renal complications.^{4, 12} Fewer studies have directly addressed renal function following EVAR, however Mehta reported rates of 3–7% following EVAR among patients treated in the first years of EVAR utilization (1996–2000) and also included physician-made grafts for patients with complex anatomy, both of which likely explain the increased renal complication rate compared to our study.¹⁰ In a more recent study, Saratzis found a rate of renal complications of 19%. This rate was likely higher than our work, and previous studies, due to their use of the highly sensitive KDIGO definition of renal dysfunction which included: increase in creatinine > 0.3mg/dL or well as low urine output.¹³

Given the infrequency of renal complications large databases are necessary to adequately power studies on acute kidney injury. However, a common definition of renal complications has not been widely utilized by any major databases including Vascular Quality Initiative, NSQIP, Medicare, or NIS leading to variable reports of such complications. The 2012 KDIGO and Acute Kidney Injury Network (AKIN) guidelines define acute kidney injury as an increase in creatinine > 0.3mg/dL, 50% increase in creatinine from baseline, or reduction in urine output to less than 0.5mL/kg per hour for more than 6 hours, are the most widely utilized guidelines for acute kidney injury.^{7, 14} However, like many alternative definitions, the utility of this definition is challenged by the difficulty and reliability or urine collection at many institutions and the potential for fluid shifts among surgical patients. Nonetheless, NSQIP and other large databases would be improved by reporting of post-operative creatinine and GFR levels to more uniformly evaluate post-operative renal dysfunction.

Increased mortality among patients with renal complications following open AAA repair was also demonstrated in prior work.^{2, 4, 13, 15, 16} Our study found 30-day mortality rates of 30% following open repair and 55% following EVAR amongst patients with renal complications. These rates are similar to those reported by Grant et al. who found a 30-day mortality rate of 35% in their study of 2378 open repairs. However, mortality rates among patients with renal complications vary tremendously in the literature and range from 9–58% following open repair. Much of this variation is likely due to differing definitions of renal dysfunction, with lower mortality rates seen in those studies that used the lower cutoff of 0.5 mg/dL increase from baseline as their definition of renal complication. Following EVAR, few studies have evaluated the mortality rates among patients with renal complications, and additional research is warranted to confirm our findings. Saratzis et al. found a mortality rate of 32% following EVAR; however this study utilized more sensitive definition of acute kidney injury including a significantly lower increase of serum creatinine.⁵

Despite differing rates of renal complications, we found similar predictors of this adverse outcome compared to prior work in patients undergoing open repair.^{2, 4, 12, 16, 17} Only one previous study, from Wald et al., identified predictors following EVAR and open repair using the Nationwide Inpatient Sample (NIS) and found open repair, chronic kidney disease, and congestive heart failure to be associated with post-operative renal complications. However, due to limitations of the NIS database, the authors were unable to account for operative and anatomic characteristics including transfusion, operative time, and aneurysm extent.¹⁶ Additionally, the NIS is considered a suboptimal dataset for evaluation of post-operative morbidity, as it is an administrative dataset reliant on coding, rather than chart review, and cannot identify events occurring after discharge. Finally, previous work has shown administrative databases to be inferior to NSQIP and chart review in identifying perioperative complications.^{18, 19} Following open repair, other authors have identified baseline kidney dysfunction, transfusion, urgency, clamp location, and renal ischemia as other predictors of renal complications; however such studies included suprarenal and ruptured aneurysms which have significantly different risks as compared to elective infrarenal aneurysms, and as such we elected to exclude them for this study.^{2, 4, 12, 16, 17}

There are important clinical implications to the results in this study. Both chronic kidney disease and operative approach are characteristics known to the surgeon in the pre-operative period and should be utilized for patient education and risk assessment pre-operatively. Furthermore, given the risk of open repair in those with chronic kidney disease, surgeons should utilize an EVAR-first approach for patients with suitable anatomy. Additionally, in all patients, but particularly those with a GFR < 60, surgeons should take care to limit the volume of contrast used to avoid further renal deterioration and contrast nephropathy. Transfusion and operative time are characteristics reflective of challenging cases and may not be avoidable; however, given their strong association with renal complications, particular care to minimize blood loss and to ensure complete hemostasis at the closure of the case should be taken.

This study has multiple limitations, which must be noted. First, it is subject to generic limitations of a clinical registry including errors in coding, missing data, and limited variable definitions. Therefore, it is possible that other confounders including blood loss, clamp time, neck length, angulation, and thrombus may impact this study and cannot be accounted for. In the current era, open repairs are often more technically challenging due to poor anatomy for EVAR; however, in this analysis we excluded those patients with short necks (suprarenal, pararenal, and juxtarenal clamps). As a result the rates of renal dysfunction following open repair may not be reflective of all open AAA repairs. Additionally, this study was unable to assess the long-term effects of renal complications. This study was also unable to account for the volume of contrast used; however, contrast volume is often not known in the pre-operative period and as such does not assist with pre-operative risk stratification. Finally, in this study renal dysfunction is restricted to the VSGNE definition of renal dysfunction and characterized by a large increase in creatinine of > 2mg/dL, which neglects to include those patients with less severe renal dysfunction; therefore, the effects of mild kidney injury and exact cause of dysfunction are unable to be evaluated.

Conclusions

Predictors of renal complications include elevated baseline GFR, open approach, transfusion, and prolonged operative time. Given the dramatic increase in mortality associated with renal complications, care should be taken to employ renal protective strategies, achieve meticulous hemostasis to limit transfusions, and to utilize an endovascular approach when technically feasible.