Preservation of renal function in juxtarenal and suprarenal abdominal aortic aneurysm repair

Preservation of renal function in juxtarenal and suprarenal abdominal aortic aneurysm repair Brent T. Allen, MD, Charles B. Anderson, MD, Brian G. Rubin, MD, M. Wayne Flye, MD, Dirk S. Baumann, MD, and Gregorio A. Sicard, MD, St. Louis, Mo. Purpose: Deterioration in renal fimction is a common cause of morbidity in patients treated surgically for juxtarenal and suprarenal abdominal aortic aneurysms. We reviewed our experience over the last 8 years with 65 consecutive patients undergoing juxtarenal (n = 31) or suprarenal (n = 34) abdominal aortic aneurysm repair. Methods: The aneurysms were repaired with a transabdominal (n = 8), thoracoabdominal (n = 4), retroperitoneal (n = 22), or thoracoretroperitoneal (n = 31) approach. Proximal aortic damps were placed at the suprarenal, supra-superior mesenteric artery, or supraceliac level. Renal hypothermia with cold heparinized saline solution renal artery perfusion was used to protect renal function in 38 patients with either preoperative renal insufficiency or with anticipated prolonged renal ischemia (> 30 minutes). Concomitant renal artery reconstruction was required in 30 patients. Results: Significant operative morbidity developed in 23 (35.3%) patients. There was one (1.53%) perioperative death (0 to 90 days). Temporary dialysis was necessary in two patients. Preoperative renal insufficiency was a significant risk factor on multivariate analysis for a decline in renal function during the first postoperative week. However, serum creatinine concentration had returned to baseline or improved in all patients but two (3.1%) at the time of discharge. In spite of significantly longer renal ischemia, discharge creatinine levels were, on tmivariate analysis, statistically less than baseline creatinine levels in patients with suprarenal aneurysms, patients requiring renal reconstruction, and patients treated with renal hypothermia. The location of the proximal aortic clamp was not a factor in postoperative morbidity. There was no significant difference between juxtarenal and suprarenal aneurysms with respect to operating room time, transfusion requirements, days intubated, resumption of oral diet, or the length of hospitalization. Conclusions: Careful consideration of the route of exposure, location of the proximal aortic clamp, and the preservation of renal function with renal hypothermia and with the repair of significant renal artery lesions will result in minimal morbidity and mortality in patients requiring surgery for juxtarenal or suprarenal abdominal aortic aneurysms. (J VASC SURG 1993;17:948-59.) Abdominal aortic surgery is a common component of the modern vascular surgery practice. In 1991 approximately 29,656 patients underwent elective abdominal aortic surgery (aneurysmal and occlusive disease) in the United States) Most of these proce- From the Section of Vascular Surgery, Department of Surgery, Washington University School of Medicine, St. Louis. Presented at the Sixteenth Annual Meeting of the Midwestern Vascular Surgical Society, Cleveland, Ohio, Sept. 11-12, 1992. Reprint requests: Brent T. Allen, MD, One Barnes Plaza, Suite 5103 Queeny Tower, St. Louis, MO 63110. Copyright 1993 by The Society for Vascular Surgery and International Society for Cardiovascular Surgery, North American Chapter. 0741-5214/93/$1.00 +.10 24/6/46197 948 dures (92%) were confined to the infrarenal aorta. Once a challenging operation, infrarenal aortic surgery is now commonly performed with mortality rates of less than 5%. 2 In approximately 6% (1719), of patients undergoing elective surgery for abdominal aortic disease in 1991,1 the aortic disease extended proximally to involve the renal or visceral vessel segments of the aorta. Such upper abdominal aortic pathology includes juxtarenal aneurysms that approximate but do not involve the renal vessels and suprarenal aneurysms that involve the origins of at least one or more of the renal or visceral vessel ostia but do not extend above the diaphragm. Operations involving aneurysms of this segment of the aorta are

/OURNAL OF VASCULAR SURGERY Volume 17, Number 5 Allen et al. 949 more formidable for both the patient and surgeon than operations involving infrarenal aortic aneurysms. The higher risk associated with juxtarenal or suprarenal abdominal aortic aneurysm repair is related to (1) increased cardiac and pulmonary demands, (2) more extensive exposure sometimes involving both the abdominal and thoracic cavities, and (3) the necessity for suprarenal aortic clamping leading to possible ischemic injury to the kidneys, visceral organs, and the spinal cord. Postoperative renal insufficiency has been a prom- '~ent complication in most series involving surgery of the upper abdominal aorta and is usually secondary to ischemia-induced acute tubular necrosis, as Several methods have been suggested to minimize this complication, including the intraoperative administration of diuretics, short renal ischemia times, renal hypothermia, and renal artery perfusion with oxygenated blood. 9.n None of these methods has gained uniform acceptance and postoperative renal insufficiency continues to be a major obstacle in the successful treatment of patients with juxtarenal or suprarenal aortic aneurysms. Our policy has been to repair significant renovascular lesions and to use renal hypothermia in patients with preoperative renal insufficiency or when prolonged renal ischemia is anticipated in an attempt to reduce the renal complications associated with the ~epair of these upper abdominal aortic aneurysms. fhis report concerns our recent experience with the treatment of patients with aneurysms of the juxtarenal and suprarenal abdominal aorta. PATIENTS AND METHODS The vascular registry in the Department of Surgery at Washington University School of Medicine retrospectively identified 65 consecutive patients who underwent operation for juxtarenal or suprarenal abdominal aortic aneurysms at Barnes Hospital between September 1984 and January 1992. Ruptured aneurysms were excluded. All patients received more than 6 months of follow-up. Current follow-up was available on 59 (95.2%) patients. There were 24 women and 41 men with an average age of 67.6 _+ 1.4 years (5C + SEM) (range 48 to 86 years). Juxtarenal aortic aneurysms (n = 31) were defined as aneurysms that approximated but did not involve the renal arteries such that they could not be repaired with an infrarenal aortic clamp. Suprarenal aortic aneurysms (n = 34) involved one or more of the renal or proximal visceral vessels (superior mesenteric artery [SMA], celiac axis) but did not extend above the diaphragm. Thirty-nine patients (60%) presented with an asymptomatic aneurysm. Abdominal or back pain led to the diagnosis of the aneurysm in 12 patients (18.4%). An evaluation for dandication or renovascular disease led to the diagnosis in 15.4% (five patients each) of patients. The aneurysm was noted to be increasing in size on serial examination in two patients (3.1%), and the aneurysm was identified during a workup for atheroemboli in two patients (3.1%). Three patients in this series had previous aortic grafting and underwent operation for a pseudoaneurysm at the juxtarenal or suprarenal anastomosis. One patient had an aborted attempt at aneurysm repair before referral to our institution. Preoperative evaluation. Preoperative general medical evaluations were performed by the surgical staff. Formal medical consultations were obtained selectively. Thirteen patients (20%) had preoperative evidence of renal insufficiency, defined as a serum creatinine level >1.8 mg/dl. A history of heart disease was identified in 37 patients (56.9%), including 22 (33.8%) patients who had a previous myocardial infarction, 17 (26.2%) who had a history of congestive heart failure, and 10 patients (15.4%) who had undergone a previous coronary artery bypass. An objective cardiac evaluation (radionucleotide ventriculogram, thallium stress test, echocardiogram, or cardiac angiogram) was obtained before operation in 37 patients (56.9%). The general medical characteristics of the patients are contained in Table I. The extent of aneurysmal disease and assessment of the renal and visceral vessels were determined by one or more of the following methods: angiography (62; 95.3%), computed tomography (51; 78.5%), and ultrasonography (15; 23.1%). Magnetic resonance imaging was used selectively. Associated severe renal artery disease was identified in 10 patients (15.3%) (nine with renal artery stenoses, one with renal artery aneurysm). Ulcerative atherosclerotic aortic lesions were noted to be the source of atheroemboli in two patients. Perioperative management. All patients were routinely monitored during operation with electrocardiography, pulmonary artery catheters, radial artery catheters, oxygen saturation probes, and expired gas mass spectrometry. Recent patients were selectively assessed with intraoperative transesophageal echocardiography, and some patients received epidural catheters to assist in anesthetic administration and postoperative pain control. Cerebrospinal fluid drainage was not routinely used. Shed blood collected during surgery was filtered, washed, and reinfused. Arterial blood gasses, electrolytes, blood

JOURNAL OF VASCULAR SURGERI I 950 Allen et al May 1993 Table I. Patient population Age (X -+ SEM) 67.6 + 1.4 yr Sex Male 41 (63%) Female 24 (37%) Abdominal aortic aneurysm size 6.68 + 0.23 cm (X + SEM) Cardiac disease 37 (56.9%) Previous myocardial infarction 22 (33.8%) Previous congestive heart failure 17 (26.2%) Previous coronary artery bypass 10 (15.4%) grafting Hypertension 49 (75.4%) Diabetes 8 (12.3%) Previous stroke 9 (13.8%) Pulmonary disease 19 (29.2%) History of smoking 46 (70.8%) Renal insufficiency (creatinine 13 (20%) level > 1.8 mg/dl) counts, and coagulation profiles were serially monitored and corrected as needed. Heparin (4000 to 7000 units) was administered intravenously before the aorta was cross-clamped in 25 (80.6%) patients with juxtarenal aneurysms and 18 (52.9%) patients with suprarenal aneurysms, according to the surgeon's preference. All patients were monitored in the surgical intensive care unit for at least 24 hours after surgery. Exposure. One of four operative approaches was used: transabdominal, retropefitoneal, thoracoabdominal, or thoracoretroperitoneal. The choice of incision was made by the surgeon based on the location of the aneurysm, associated vascular disease, habitus, and previous operative procedures. Transabdominal aneurysm repairs were done with a midline xiphoid to pubis incision, and the aneurysm was repaired through the peritoneal cavity. Retroperitoneal incisions were performed with the patient's left shoulder and trunk rotated anteriorly 45 and extended from the tip of the tenth, eleventh, or twelfth rib to the lateral border of the left rectus muscle midway between the umbilicus and pubis. The thoracic cavity was not entered. The left kidney was either left in the renal fossa, mobilized circumferentially, or mobilized anteriorly with the peritoneum, depending on the location of the aneurysm, the level of aortic clamping, and associated renal artery disease. The parietal peritoneum was peeled off the hemidiaphragm, and the crus of the diaphragm at the aortic hiatus was incised as needed to expose the visceral vessel segment of the aorta for crossclamping. If additional exposure was required in the pelvis, the incision was extended to the midline across the left rectus muscle, and if necessary, a separate fight lower quadrant incision was made lateral to the edge of the fight rectus muscle to allow retroperitoneal access to the distal right external lilac artery. Thoracoretroperitoneal incisions were performed in a manner similar to retroperitoneal incisions but extended on to the chest wall over the ninth or tenth ribs. The thoracic cavity was entered through the bed of the partially resected rib or through the intercostal interspace. The hemidiaphragm was completely divided only when necessary to expose the supraceliac aorta at the aortic hiatus or above the diaphragm. The chest was drained after operation with an intercost~ ~-~ tube. Thoracoabdominal incisions were made with a lower midline transabdominal incision that curved in the epigastrium to the costal margin. The incision extended over and through the ninth or tenth intercostal space. The abdominal viscera and left kidney were all reflected to the right. The hemidiaphragm was completely divided in a radial fashion from the chest wall to the aortic hiatus. Ischemic renal protection. Renal hypothermia obtained by cold heparinized saline solution renal artery perfusion was used selectively in patients because of preoperative renal insufficiency or when prolonged renal ischemia ( > 30 minutes) was anticipated. Perfusion was performed with a 6F or 9F balloon-tipped Pruitt irrigation occlusion catheter (Ideas for Medicine, Inc., Tampa, Fla.) inserted ig each renal artery orifice (Fig. 1). Cold (4 C) heparinized (1 unit/ml) saline solution was flushed through the catheter throughout the ischemic period to promote renal hypothermia. When possible in retroperitoneal exposures the left kidney was packed in ice slush to further ensure renal hypothermia. Mannitol (12.5 to 25 gm) and lasix (20 to 40 mg) were administered before suprarenal aortic occlusion. Additional doses of diuretics were given after renal reperfusion to maintain a brisk urine output. Indigo carmine was routinely administered intravenously after renal reperfusion to confirm renal filtration. Analysis. Patient charts were reviewed, and the following variables were recorded: route of exposure, ' location of proximal aortic clamp, duration of renal ischemia, use of renal perfusion, postoperative core body temperature, use of renal reconstruction, serial serum creatinine concentration, site of distal anastomosis, early and late complications, and death. The duration of renal ischemia was calculated from the time of proximal aortic occlusion to the time of reperfusion of both kidneys. Unlvariate statistical comparisons ofserttm creatinine concentrations, core body temperatures, and ischemia times were per-

'JOURNAL OF VASCULAR SURGERY Volume 17, Number 5 Allen et al. 951 i / Fig. 1. Renal hypothermia is performed by cannulating renal artery ostia with balloon-tipped catheter and perfusing kidneys with syringe containing cold heparinized saline solution. Backbleeding from visceral vessels is controlled with balloon-tipped occlusion catheter. formed with Student's t test. Regression analysis was used to determine independent multivariate predictors of a significant change in serum creatinine concentration between preoperative and 7 day postoperative or discharge values. The variables analyzed were age, sex, cardiac history, chronic obstructive pulmonary disease (COPD), history of smoking, stroke, aneurysm location, renal ischemia less than 30 minutes, renal hypothermia, renal artery reconstruction, operative exposure, and preoperative renal insufficiency. For multivariate analysis an increase in creatinine concentration was defined as a normal preoperative creatinine level ( < 1.8 mg/dl) increasing to abnormal levels after operation ( > 1.8 mg/dl) or an abnormal creatinine level increasing more than 20% after operation. Comparisons between the methods of exposure, operating room time, transfusion requirements, days intubated, days with nothing by mouth, or duration of hospitalization were made with an analysis of variance. RESULTS The route of exposure and location of the proximal aortic clamp are displayed in Table II. A transabdominal incision was used in 8 patients (12.3%), a retroperitoneal incision was used in 22 patients (33.8%), a thoracoretroperitoneal incision was used in 31 patients (47.7%), and a thoracoabdominal incision was used in 4 patients (6.2%). Juxtarenal aneurysms were usually repaired with a retroperitoneal incision (16 of 31; 51.6%), and suprarenal aneurysms were usually repaired with a thoracoretroperitoneal incision (24 of 34; 70.6%). A thoracoabdominal incision was not used for juxtarenal aneurysms and suprarenal aneurysms were not repaired with a transabdominal incision. The proximal aortic clamp was placed above the renal arteries in 26 patients (40%), above the SMA in 12 patients (18.5%), and above the celiac axis in 27 patients (41.5%). Juxtarenal aneurysms were usually repaired with a suprarenal aortic clamp (19 of 31;

JOURNAL OF VASCULAR SURGERY: 952 Allen et al. May 1993 Table II. Route of exposure and level of proximal aortic clamp Level of clamp Transabdominal Retroperitoneal Thoracoretroperitoneal Thoracoabdominal Total Juxtarenal aneurysms (n = 31) Above renals 5 13 1 0 19 (61.3%) Above SMA 0 3 2 0 5 (16.1%) Above celiac 3 --0-4 0 7 (22.6%) Total 8 (25.8%) 16 (51.6%) 7 (22.6%) 0 (0%) 31 (100%) Suprarenal aneurysms (n = 34) Above renals 0 4 3 0 7 (20.6%) Above SMA 0 2 5 0 7 (20.6%) Above celiac 0 "-0 1--6 "-4 20 (58.8%) Total 0 (0%) 6 (17.6%) 24 (70.6%) 4 (11.8%) 34 (100%) Table lii. Surgical procedure No. % Site of proximal anastomosis Juxtarenal 31 47.7 Above renals 15 23.1 Above SMA 9 13.8 Above celiac 10 15.4 Renal and visceral vessel reconstruction Renal artery reconstruction 30 46.1 Reimplantation (n = 24) Endarterectomy (n = 6) Suprarenal endarterectomy (atheroemboli) 2 3.1 Site of distal anastomosis Aorta 40 61.5 Iliac arteries 17 26.2 Femoral arteries 8 12.3 61.3%). Suprarenal aneurysms were usually repaired with a supraceliac aortic clamp ( 20 of 34; 58.8 %). An aortic clamp between the SMA and celiac axis was not used in any patient undergoing transabdominal aortic repair. Supraceliac aortic occlusion was not used with a retroperitoneal approach. The aortic clamp was placed at the supraceliac site in all patients undergoing thoracoabdominal exposure. All three sites for aortic damping were used in patients undergoing thoracoretroperitoneal exposures but the supraceliac location was selected most frequently in both groups of aneurysms repaired with this approach. Vascular reconstruction. All juxtarenal aneurysms (n = 31) were repaired with the proximal anastomosis just below the renal arteries (Table III). Additional renal artery reconstruction was required in seven patients with juxtarenal aneurysms. Four patients had associated severe bilateral renal artery ostial stenosis that were treated with bilateral transaortic endarterectomy. In these patients a longitudinal aortotomy was made from the neck of the aneurysm to above the renal arteries to fully expose the renal artery ostia. The aortotomy was closed after the endarterectomy, and the graft was implanted in the infrarenal location. The left renal artery was divided and implanted directly into the aortic graft (with a prosthetic or saphenous vein interposition graft when necessary) in three patients-two with renal artery stenosis and one with a renal artery aneurysm. The inferior mesenteric artery was reimplanted in one patient. The proximal end of the graft was cut at an angle in patients with suprarenal aneurysms to allow inclusion of the renal and visceral vessel orifices in the proximal anastomosis. The proximal beveled anastomosis extended to just above the renal arteries in 15 patients (44%) and just above the origin of eithey_ visceral vessel (9 SMA, 10 celiac) in 19 patients (56%). The left renal artery was frequently disconnected from the aorta in patients with suprarenal aneurysms to simplify and reduce the size of the proximal anastomosis. The left renal artery was implanted in the aortic graft after completion of the proximal aortic anastomosis (Fig. 2). Thus 20 (58.8%) patients with suprarenal aneurysms underwent reimplantation of the left renal artery. Both renal arteries were reimplanted in one patient. A right renal endarterectomy and bilateral renal endarterectomy were required in one patient each. Two patients underwent suprarenal aortic endarterectomy about the renal and visceral vessels as treatment for atheroemboli. Distal vascular reconstruction used a tube graft in 40 (61.5%) patients. Aneurysmal or occlusive iliac disease necessitated bypass to the iliac vessels in 17 (26.2%) patients and the femoral vessels in eight (12.3%) patients. Renal ischemia and preservation of renal function. The relationship of renal ischemia and serum creatinine levels are shown in Table IV. The average duration of suprarenal aortic occlusion in all patients

)OURNAL OF VASCULAR SURGERY Volume 17, Number 5 Allen et al. 953 Fig. 2. Division of left renal artery simplifies and expedites proximal aortic anastomosis in suprarenal aneurysms. Left kidney is protected from ischemic injury with hypothermic perfusion until proximal aortic anastomosis is completed and left renal artery is implanted in aortic graft. was 35.40 2.90 minutes. Patients undergoing suprarenal aneurysm repair had significantly longer renal ischemia (44.13 + 4.25 minutes) than patients undergoing juxtarenal aneurysms (26.04 _+ 3.21 minutes; p < 0.001). Twenty-one (61.8%) patients undergoing suprarenal aneurysm repair required greater than 30 minutes, whereas only four (12.9%) patients in the juxtarenal aneurysm group required more than 30 minutes of renal ischemia. Renal hypothermia with cold heparinized saline solution renal artery perfusion was used in 38 (58.5%) patients. Eight patients had juxtarenal aneurysms and 30 underwent operation for suprarenal aneurysms. Renal hypothermia was used in 10 of the 13 patients with preoperative renal insufficiency. The average duration of renal ischemia in patients receiving renal hypothermia was significantly longer than in patients not being treated with renal hypothermia (42.91 _+ 4.08 minutes vs 23.96 +_ 2.60 minutes; p < 0.001). Thirty-one (47.7%) of the patients treated with hypothermia required renal artery reconstruction; renal artery bypass (n = 24), renal endarterectomy (n = 5), or suprarenal aortic endarterectomy (n = 2). Renal reconstruction required an average of 48.0 -+ 5.14 minutes of ischemia. Preoperative renal insufficiency was noted in six patients requiring renal artery reconstruction. Patients not requiring renal reconstruction had an average renal ischemia time of 25.8 -+ 2.26 minutes (p < 0.001). There was an overall trend for a decline in the creatinine concentration from baseline to discharge. However, the comparisons between baseline creatinine (1.47 _+ 0.11), 7 day creatinine (1.45 _+ 0.13) and discharge creatinine levels (1.33 + 0.09) were not statistically significant. One patient regained sufficient renal function after renal revascularization to discontinue his preoperative requirement for dialysis. Univariate analysis. When baseline and discharge creatinines were compared according to aneurysm location, there was a significant drop in the serum

JOURNAL OF VASCULAR SURGER'~ 954 Allen et al. May 1993 Table IV. Preservation of renal function (univariate analysis) Creatinine (mg/dl) ~ No. Renal hypothermia Yes 38 No 27 Renal arte~ reconstructaon Yes 30 No 35 Baseline creatinine > 1.8 mg/dl 13 < 1.8 mg/dl 52 Renal ischemia > 30 rain 25 < 30 min 40 Renal ischemia (min) 42.91 ± 4.08 23.96 + 2.60 (p = 0.001) 48.0 ± 5.14 25.8 + 2.26 (p = 0.001) 27.63 ± 11.75 36.17 -± 3.41 (NS) 54.0 ± 4.40 21.3 + 1.16 (p = 0.0001) Baseline Discharge p Value 1.62 _+ 0.18 1.38 _+ 0.13 p < 0.02 1.28 ± 0.07 1.26 ± 0.11 NS 1.73-0.19 1.37 _ 0.13 p < 0.009 1.28 ± 0.11 1.30 ± 0.12 NS 2.88 ± 0.33 2.26 ± 0.31 NS 1.18 ± 0.05 1.13 _ 0.05 NS 1.42 ± 0.19 1.29 ± 0.13 NS 1.47 _+ 0.14 1.34 ± 0.13 NS NS, Not significant. ~Baseline to discharge creatinine comparison. Values expressed as mean + SEM. creatinine value in patients with suprarenal aneurysms (1.54 _+ 0.17vs 1.28 _+ 0.11;p < 0.04), but not for juxtarenal aneurysms (1.41 _+ 0.12 vs 1.38 -+ 0.13;p < 0.74). There was a significant fall from baseline to discharge creatinine concentration in patients treated with renal hypothermia (1.62 _ 0.18 vs 1.38 _+ 0.13, respectively; p < 0.02), and in patients undergoing renal artery reconstruction (1.73 _ 0.19 vs 1.37 _+ 0.13 respectively; p < 0.009). There was no significant change in serum creatinine concentration from baseline to discharge in patients with preoperative renal insufficiency or in patients requiring more than 30 minutes of renal ischemia during operation. Multivariate analysis. Multivariate analysis identified preoperative renal insufficiency (serum creatinine 2.88 -+ 0.33 mg/dl) as an independent predictor of a decline in renal function when measured seven days after operation (3.71 + 0.63; p = 0.03) but not at discharge (2.26 + 0.31). None of the other variables (age, sex, cardiac history, COPD, history of smoking, stroke, aneurysm location, renal ischemia greater than 30 minutes, renal hypothermia, renal artery reconstruction, or operative exposure) were independent multivariate predictors of postoperative renal function at the 7 day or discharge time point. There was an increase in creatinine concentration after operation in eight (12.3%) patients as defined by a normal preoperative creatinine level (< 1.8 mg/dl) increasing to abnormal levels after operation (> 1.8 mg/di) or an abnormal creatinine level increasing more than 20% after operation. Four of the patients had normal renal function before operation and in four patients the baseline creatinine level ranged from 2.0 to 4.3 mg/dl. Four of these patients received renal hypothermia, and three required renal construction. Renal function returned to baseline or lower values by discharge in all patients but two. One patient with a baseline creatinine level of 1.6 mg/dl had congestive heart failure after a juxtarenal aneurysm repair and was discharged with a creatinine level of 1.9 mg/dl. The other patient had preoperative renal insufficiency and was one of two patients (neither treated with hypothermia) who required temporary dialysis after operation. At discharge he was no longer receiving dialysis, and he had a creatinine level of 3.7 mg/dl and continues to do well without receiving dialysis on 2 year follow-up. The other patient requiring postoperative dialysis had a single kidney and underwent suprarenal aneurysm repair with a supraceliac clamp. He had multisystem failure after operation and died of a myocardial infarction as his renal function was improving. Operating room requirements and hospital course. Core body temperatures were available in 44 (67.7%) of the patients. Patients treated with hypothermic saline solution renal perfusion had a significantly lower core body temperature immediately after operation than patients not receiving renal'' hypothermia (34.9 _ 0.22 vs 35.8-0.19, respectively; p = 0.005). There were no significant differences in operative time, transfusion requirements, intraoperative crystalloid administration, days intubated, days with nothing by mouth, or length of postoperative hospitalization among patients with juxtarenal or suprarenal aneurysms (Table V). Valid comparisons of these variables with respect to the surgical approach was limited by the small number of patients undergoing a transabdominal (n = 8) or thoracoabdominal (n = 4) approach. However,

,JOURNAL OF VASCULAR SURGERY Volume 17, Number 5 Allen et al 955 Table V. Operating room requirements and hospital course Juxtarenal Suprarenal Anesthetic time (min) Packed red blood cells (units) Crystalloid (ml) Days intubated Days with nothing by mouth Postoperative hospitalization (days) Values expressed as mean +_ SEM. 344.35-19.22 4.35 + 1.06 6423.91 + 612.52 2.76 + 0.79 (range 0-18) 5.18-1.10 17.5 + 2.32 (range6-61) 374.0 _+ 14.74 4.71 + 0.45 6136.54 + 482.12 4.48 + 2.07 (range 1-52) 4.21 + 0.49 19.2 + 3.18 (range 7-90) Table VI. Complications Juxtarenal aneurysm (n = 31) Suprarenal aneurysm (n = 34) Total (n = 65) Morbidity Reoperation for bleeding 2 0 2 Ventilator > 48 hrs. 6 5 11 Paraparesis 0 1 1 Wound 1 1 2 Temporary dialysis 1 1 2 Small bowel obstruction 2 1 3 Pulmonary emboli 0 1 1 Myocardial infarction 0 1" 1 8 (25.8%) 8 (23.5%) Mortality Early (0 to 90 days) 1" Late (90 days to 6.5 years) 3 7 *Patient died of myocardial infarction. 23 (35.3%) 1 (1.53%) 10 (15.4%) none of the comparisons of approach versus these variables were significant. Complications. Twenty-three (35.3%) patients experienced complications (Table VI). The most frequent complication was prolonged ( > 48 hours) mechanical ventilation accounting for 11 of the 23 (47.8%) complications. Six of these patients had juxtarenal aneurysm repair, and five had suprarenal aneurysms. All but one of these patients had a history of smoking and a preoperative diagnosis of COPD was noted in six. Two patients required a tracheostomy for ventilator weaning, one of which required a tracheostomy during a previous aortic operation. All patients were successfully weaned from the ventilator, except one who eventually died (see below). Three patients had a small bowel obstruction in the perioperative period. The juxtarenal aneurysm in two of these patients was repaired transabdominally. The source for the small bowel obstruction was an adhesion in both cases. Segmental small bowel infarction 14 days after operation was the cause of the obstruction in the third patient who underwent suprarenal aneurysm repair via a thoracoretroperitoneal incision with a proximal clamp placed between the celiac axis and SMA. One patient who presented with atheroemboli from a 5 cm suprarenai aneurysm had paraparesis, which slowly improved after operation. Her aneurysm repair included a suprarenal endarterectomy, with a supraceliac occlusion time of 30 minutes. Two patients with juxtarenal aneurysms required repeat exploration for bleeding. One patient initially underwent operation retroperitoneally, and the other underwent aneurysm repair transabdominauy because of a left lower quadrant colostomy. Two patients had wound complications. One had a superficial infection in a thoracoretroperitoneal incision, and the other patient had a groin incision lymph fistula. Both wounds healed with treatment. One patient (1.53%) had multiorgan system failure (cardiac, pulmonary, renal) after a suprarenal aneurysm repair that required prolonged hospitalization that ultimately ended in death as a result of myocardial infarction. Continued contact with these patients through the vascular registry for up to 6.5 years after surgery has identified late death in 10 patients. Six patients have died of cardiac causes, and four patients died of unknown causes. DISCUSSION Table VII displays several recent series of patients undergoing abdominal aortic surgery requiring suprarenal cross-clamping. The incidence of postoperative renal insufficiency varies from 12% to 31%.

JOURNAL OF VASCULAR SURGER]7" 956 Allen et al. May 1993 Table VII. Suprarenal aortic clamping for abdominal aortic disease Renal Patients ischemia Author Year (no.) Aortic disease Renal protection (mean) Crawford et al.4 1985 101 Qvarfordt et al.7 1986 77 Green et al. s 1989 52 Shepard et al.s 1991 43 Breckwoldt et al.3 1992 39 Poulias et al. 6 1992 38 Present series 1992 65 Total AAA, Abdominal aortic aneurysm; NS, not stated. Abnormal creatinine > 2.0 mg/dl. t Abnormal creatinine level > 1.2 mg/dl. :~ Abnormal creatinine level > 50% baseline. Median renal ischemia time. II Abnormal creatinine level > 168 ~mol/l. Dialysis not permanent. # Abnormal creatinine level > 1.5 mg/di. luxtarenal AAA NS 19 rnin Pararenal AAA Renal hypothermia 25 rain AAA with difficult proximal aortic cuff Renal hypothermia NS Complex abdominal aortic reconstruc- Renal hypothermia 31 min tions (n = 11) Juxtarenal AAA, suprarenal AAA, diffi- None 48.5 min cult AAA Juxtarenal AAA None 13 to 48 rain Juxtarenal AAA, suprarenal AAA Renal hypothermia 35 rain (n = 38) Mean 31.7 min Dialysis, either temporary or permanent, was required in 23 (5.5%) of the 415 patients in the combined series. In most cases the decline in renal function was temporary but may have delayed overall recovery and prolonged hospitalization. The cause of postoperative renal insufficiency in previous series is largely acute tubular necrosis related to prolonged renal ischemia. Prolonged renal ischemia was not an important factor in postoperative renal function in this series. The average duration of renal ischemia in patients with juxtarenal aneurysms was significantly shorter than in patients with suprarenal aneurysms (26.04 vs 44.13, respectively;p < 0.001). In spite of the longer renal ischemia, the baseline serum creatinine level in patients with suprarenal aneurysm had decreased significantly at discharge (1.54 ~ 1.28 mg/dl; p < 0.04) and was not significantly changed in juxtarenal aneurysms. Renal hypothermia was used in most (88%) of the suprarenal aneurysms and in eight (26%) of the juxtarenal aneurysms when extended renal ischemia was anticipated or when preoperative renal insufficiency (creatinine level > 1.8) was noted. The use of hypothermic renal protection during ischemia was developed primarily in the field of renal transplantation. Hypothermic renal preservation was reviewed recently by Marberger and Dreikorn.12 Experimental data suggests that renal oxygen consumption is reduced to 40% with cooling of the renal parenchyma to 30 C, to 15% at 20 C, and to less than 5% at 10 C. 13-is A parenchymal temperature of 15 C to 25 C will allow approximately 2 hours of reversible ischemia. Dass, et al. 16 have shown that in situ flushing of canine kidneys with 250 ml of lactated Ringer's at 4 C lowered renal cortical temperatures to 13 C. 16 Svensson et al.n have reported their clinical experience with renal perfusion to preserve renal function in a series of 1233 patients undergoing repair of thoracic or thoracoabdominal aortic aneurysms. Univariate analysis indicated that perfusion O f the left renal artery with cold Ringer's lactate in, subgroup of 103 patients did not seem to reduce postoperative renal dysfunction (dialysis or creatinine level > 2.0 mg/dl). However, multivariate analysis indicated renal perfusion with blood or Ringer's lactate was associated with a lower incidence of postoperative serum creatinine level >2.0 mg/di. The left renal core temperature was monitored in 56 of their patients treated with cold Ringer's lactate perfusion and noted to average 17.2 C (range 12 C to 25 C). More recently Svensson et al. 17 reported that bilateral perfusion of the renal arteries with cold Ringer's lactate in 24 of 271 patients undergoing operation for thoracoabdominal aortic aneurysms with associated visceral or renal artery occlusive disease resulted on univariate analysis in a significantly lower incidence of renal complications (dialysis or creatinine level > 3.0 mg/dl). We found in this series that none of the patients treated with hypothermia required dialysis and only one (2.6%) had a significant increase in preoperative (1.6 mg/dl) to postoperative (1.9 mg/dl) creatinine levels. The value of renal hypothermia and, when appropriate, renal artery reconstruction in preserving renal function is

fournal OF VASCULAR SURGERY Volume 117, Number 5 Allen et al. 957 Preoperative Postoperative Temporary renal renal and permanent insufficiency insufficiency dialysis Death 19" 16 (15.8%)* 8 (7.9%) 8 (7.9%) 421 18 (23%)t 2 (2.5%) 1 (1.3%) NS NS 6 (11.5%) 8 (15.4%) 5 6 (i4%)~: 0 1 (2.3%) 9 12 (31%)11 1 (2.6%) 1 (2.6%) 6# 9 (24%)# 4 (11%) 2 (5.2%) 13 8 (12%) 2 (3%) 1 (1.5%) 94 (26%) 69 (19%) 23 (5.5%) 22 (5.3%) emphasized by the significant decline on univariate analysis in preoperative to discharge sermn creatinine concentration in patients receiving either treatment. Our general method is to give 150 ml heparinized saline solution over 3 to 5 minutes simultaneously to each kidney. Subsequent infusions are slowed to appro:dmately 5 ml/min. When possible the perinephric fat is elevated offthe left kidney, and ice slush is applied directly to the renal capsule. Renal hypothermia resulted in a slight (0.9 C) but statistically significant lower core body temperature when compared with patients not receiving hypothermia. This small difference in temperature was not associated "-,ith complications and seemed to be well tolerated. One of the major determinants of postoperative renal dysfimction in other studies has been preexisting renal insufficiency (Table VII). In this series renal hypothermia was used in 77% of patients with preoperative renal insufficiency and renal artery reconstruction was performed in six (46%). Multivariate analysis of our data indicated that preoperative renal insufficiency was the only independent predictor of a postoperative decline in renal function. Importantly, we found that the decline in renal fixnction was transient, occurring during the first 7 days after surgery. There was no significant difference in preoperative and discharge creatinine levels in patients with preexisting renal insufficiency. Green et al.s identified atheroemboli as another important cause of postoperative renal insufficiency in patients treated with proximal aortic clamping just above the renal arteries or between the SMA and celiac axis. The renal and visceral vessel segment of the aorta frequently has significant atherosclerotic disease that may embolize when manipulated or clamped. In addition placement of a clamp near the origins of these vessels may damage their already diseased ostia, resulting in stenosis or thrombosis. We have found the preoperative angiogram extremely helpful in providing important information regarding atherosclerotic disease in this segment of the aorta. Additionally it demonstrates accessory renal arteries, the status of the inferior mesenteric artery, and disease in the iliofemoral segment. Preoperative angiography was performed in all but three (95.3%) of our patients and was very useful in selecting the best site for aortic clamping. Thus placement of the proximal aortic clamp at the suprarenal or supra-sma levels in 38 (58.5%) did not seem to cause significant problems with peripheral embolization or renal function. Indeed, of the eight patients who had significant deterioration in renal function measured at: 1 week after operation, only three had the proximal aortic clamp placed on the visceral vessel segment of the aorta. Renal function at discharge in these three patients had returned to baseline. Most patients in this series underwent exploration through a retroperitoneal or thoracoretroperitoneal approach. Our group has previously reported a favorable experience with the retroperitoneal approach for infrarenal abdominal aortic aneurysm surgery when compared retrospectively with the transperitoneal approach28 Shepard et al. 8 have pointed out the advantages of the retroperitoneal approach for difficult aortic reconstructions including juxtarenal and suprarenal aneurysms and for patients requiring renal or visceral endarterectomy. We have found the left-sided retroperitoneal approach very useful in the treatment of proximal abdominal aortic aneurysms as well. Sixteen (67%) of the juxtarenal aneurysms and six (42.9%) of the suprarenal aneurysms that were repaired with a clamp above the renal arteries or SMA were approached retroperitoneally. If a retroperitoneal incision does not provide adequate exposure for suprarenal or supra-sma aortic occlusion or if a supraceliac damp is required, then the incision can be extended over and through the ninth or tenth ribs with a thoracoretroperitoneal approach. This incision provides excellent visualization of the renal and visceral vessel segment of the aorta in patients where exposure is limited because the costal margin approximates the iliac crest and in patients that require exposure of the supraceliac aorta for cross.-clamping either above or below the hemidiaphragm. Thoracoretroperitoneal incisions were used in 11 (29%) of our patients requiring suprarenal or supra-sma clamping and in 20 (74%) of the patients requiring supraceliac aortic exposure. There was no difference in the operative morbidity rate in retroperitoneal versus thoracoret-

958 Allen et al. JOURNAL OF VASCULAR SURGER~ I May 1993 roperitoneal incisions. We currently reserve transabdominal incisions for juxtarenal aneurysms in thin patients without previous abdominal surgery and in those that require right renal artery reconstruction. Thoracoabdominal incisions are infrequently used but are occasionally necessary for suprarenal aneurysms that have had previous left nephrectomies or other left retroperitoneal surgery. We have found division of the left renal artery facilitates in the repair of many suprarenal aneurysms. By dividing the left renal artery and mobilizing the kidney anteriorly, the exposure of the pararenal aorta is markedly improved. The management ofjuxtarenal atherosclerotic lesions is simplified, and the aortic anastomosis is expedited and made technically easier. The left kidney is protected with hypothermia until completion of the proximal aortic anastomosis, and then the renal artery is attached to the aortic graft either directly or with an interposition graft. Hypothermic protection of the left kidney in this setting has been satisfactory, and we have not noted persistent abnormalities on postoperative renal scans. In summary careful consideration of the route of exposure, the site of proximal aortic clamping, the need for renal artery reconstruction, and the need for renal protection in patients requiring prolonged renal ischemia and in those with preexisting renal insufficiency will allow the repair of juxtarenal and suprarenal abdominal aortic aneurysms with acceptable operative morbidity rates, preservation of renal function, and minimal mortality rates. REFERENCES 1. Department of Health & Human Services, Health Care Financing Administration, 6325 Security Blvd., Baltimore, Maryland 21207. 2. I-Iollier LH, Taylor LM, Ochsner J. Recommended indications for operative treatment of abdominal aortic aneurysms. J VASC SUttG 1992;15:1046-56. 3. Breckwoldt WL, Mackey WC, Belkin M, O'Donnell TF Jr. The effect of suprarenal cross-clamping on abdominal aortic aneurysm repair. Arch Surg 1992;127:520-4. 4. Crawford ES, Beckett WC, Greer MS. Juxtarenal infrarenal abdominal aortic aneurysm: special diagnostic and therapeutic considerations. Ann Surg 1986;203:661-70. 5. Green RM, Ricotta JJ, Ouriel K, DeWeese JA. Results of supraceliac aortic clamping in the difficult elective resection of infrarenal abdominal aortic aneurysm, l VASC SURG 1989;9: 125-34. 6. Poulias GE, Doundoulakis N, Skoutas B, et al. Juxtarenal abdominal aneurysmectomy. J Cardiovasc Surg 1992;33:324-30. 7. Qvarfordt PG, Stoney RJ, Reilly LM, Skioldebrand CG, Goldstone J, Ehrenfeld WK. Management of pararenal aneurysms of the abdominal aorta. J VASC SUING 1986;3:84-93. 8. Shepard AD, Tollefson DFJ, Reddy DJ, et al. Left fla~ -~ retroperitoneal exposure: a technical aid to complex aortic reconstruction, l VASC SURG 1991;14:283-91. 9. Hanley MJ, Davidson K. Prior mannitol and furosemide infusion in a model of ischemic acute renal failure. Am l Physiol 1981;241:F556-64. 10. Ochsner JL, Mills NL, Gardner PA. A technique for renal preservation during suprarenal abdominal aortic operations. Surg Gynecol Obstet 1984;159:388-90. 11. Svensson LG, Coselli JS, Sail HI, Hess KR, Crawford ES. Appraisal of adjuncts to prevent acute renal failure after surgery on the thoracic or thoracoabdominal aorta. J VASC SURG 1989;10:230-9. 12. Marberger M, Dreickorn K. Renal preservation. In: Marberger M, Dreickorn K, eds. International perspectives in urology, vol 8. Baltimore: Williams and Wilkins, 1983. 13. Levy M. Oxygen consumption and blood flow in the hypothermic perfused kidney. Am J Physio11959;197:1111-4. 14. Harvey RB. Effects of temperature on fimction of isolated dog kidney. Am J Physiol 1959;197:181-6. 15. Semb G, Krog J, Johansen K. Renal metabolism and blood flow during local hypothermia, studied by means of rer~i -* perfusion in situ. Acta Chir Scand Supp11960;253:196-202. 16. Dass S, Maggio AJ, Sacks A. In sire flushing of donor kidneys; its technique and rational. J Urol 1979;121:262-4. 17. Svensson LG, Crawford ES, Hess KR, Coselli JS, Sail HJ. Thoracoabdominal aortic aneurysms associated with celiac, superior mesenteric, and renal artery occlusive disease: methods and analysis of results in 271 patients. J VASC SUltG 1992;16:378-90. 18. Sicard GA, Allen BT, Munn JS, Anderson CB. Retroperitoneal versus transperitoneal approach for repair of abdominal aortic aneurysms. Surg Clin North Am 1989;69:795-806. Submitted Sept. 14, 1992; accepted Feb. 1, 1993. DISCUSSION Dr. William Smead (Columbus, Ohio). The group from Washington University has done much to popularize the left flank retroperitoneal approach to the abdominal aorta for routine aortic anenrysmectomy and offer several good arguments to support its use for more complex problems involving the visceral segment. Their periopera- five mortality rate of 1.53% and the low incidence of renal complications and statistically significant renal function improvement in many groups present a target for many of us to aim at. The article presents a careful description of the perioperative considerations in this group of patients

JOURNAL OF VASCULAR SURGERY Volume 17, Number 5 Allen et al. 959 detailing the exposure and approaches to the suprarenal aorta, the advantages and liabilities of each of these options. Considering the wide variety of anatomy and disease, one has to be prepared to vary one's approach as appropriate. In contrast to earlier reports neither death nor kidney failure increased in patients with perioperative renal insufficiency or when concomitant renal revascularization was required. What factors were responsible for these improved results? The authors would respond that renal cold perfusion was responsible for preserving renal function in selected patients with preoperative renal insufficiency or expected prolonged clamp times. This report is ~rtainly not a randomized prospective study that would prove the point; however, when compared with most historical controls, these results suggest that something beneficial is going on here. Drs. Crawford and Hollier have reported no protective effect from renal hypothermia. How can we be certain that these excellent results are not related to other factors, such as clamp time? In this study clamp times were short, 44 minutes for suprarenal aneurysm repair and 26 minutes for juxtarenal lesions. No difference was observed when 30 minutes was taken as the discriminating time. I suspect that clamp time might become significant if 45 or 60 minutes were taken as the discriminating time. Short clamp times reflect good preoperative assessment, careful planning, and lots of' experience. Another factor may be their appropriately aggressive stance toward renal revascularization and perhaps also careful technique to prevent atheroembolism, improved anesthetic techniques, fluid management, and so on. Most of us use renal hypothermia in these settings, probably for the same reasons, but I have often wondered how effectively the renal parenchyma is cooled in these situations when surface cooling is awkward or impossible and collateral blood flow cannot be interrupted. This situation is not directly analogous to the renal transplantation situation where the kidney is in a pan. Do you have any data relating to renal core temperatures throughout the clamp interval when using your techniques or techniques like this? How do you deal with right renal artery lesions extending beyond the orifice when using the left flank approach when transaortic endarterectomy may be less than optimal? Do you have any tricks to help us prevent kinking at the reimplanted left renal artery? We have often fotmd that a short graft originating from the distal aortic reconstruction is effective. Do you have any experience with the hepatorenal or splenorenal techniques of renal revascularization that may minimize individual renal ischemic times and atheroembolism when combined with distal aortic reconstruction? Dr. Brent T. Allen. I agree that it is important to minimize when possible the duration of renal ischemia. The actual duration of renal ischemia in the entire series of our patients was 35 minutes. The average renal ischemia time in other large series is closer to 20 or 25 minutes. We did not measure any core renal temperatures in either kidney during the operation on any of these patients. Experimental data suggest that: you can reduce renal oxygen consumption down to about 15% by cooling the renal parenchyma to 15 C. That temperature can be achieved by peffusing the renal artery with between 200 and 500 ml cold heparinized saline solution. Our typical technique is to rapidly perfuse each kidney with about 150 ml cold heparinized saline solution at 4 C, and then throughout the remainder of the ischernic interval we irrigate at a slower rate, 50 ml every 5 minutes continuously. This minimizes venous backbleeding and renal warming. Dr. Smead asked about how to deal with the right renal artery when operating through the left retroperitoneal approach. He has identified one of the disadvantages in the left retroperitoneal approach. I think the best technique to deal with distal renal artery lesions is to divide the aorta right above the renal artery ostia and mobilize the renal artery underneath the cava. The renal artery can then be endarterectomized and reimplanted, usually with a vein graft or prosthetic interposition graft. He also was concerned about the possibility of kinking when reimplanting the left renal artery and mentioned that he uses an interposition graft and implants the renal artery more distally on the aorta. We also like to use interposition grafts when necessary, although in many of these patients it was not necessary. I think k3nking is minimized when the left renal artery is implanted higher on the aorta and more on the left side so that when the kidney rolls over the artery does not kink. We have had some experience with hepatorenal and splenorenal reyascularizations in patients undergoing operations primarily for renal wlscular lesions, not aneurysms or aortic occlusive disease. We generally tend to reimplant the renal arteries directly in the aortic graft with or without the use of a vein or prosthetic interposition segment and not to go into the additional dissection and exposure that is required for hepatorenal or splenorenal bypasses.