Branch renal artery repair with cold perfusion protection

Patient population 

Patients who underwent renal artery reconstruction were identified from an Institutional Review Board-approved vascular registry. From July 1987 through November 2006, 874 patients had open operative renal artery repair to 1312 kidneys at Wake Forest University Baptist Medical Center. Those patients who had reconstruction using cold perfusion protection were selected.

Branch renal artery reconstruction was done in 77 consecutive patients (62 women, 15 men; mean age, 44 ± 17 years) using either ex vivo or in situ cold perfusion protection for 78 kidneys. Patient demographics and preoperative clinical characteristics are depicted in Table I. The hospital records, imaging studies, and clinic records for all patients were reviewed. Demographic data, surgical technique, perioperative morbidity, and mortality were examined. Postoperative patency of repair was determined by angiography or renal duplex ultrasound (RDUS) imaging, or both.

Table I. Patient characteristics

	Patients (n = 77)	N or mean ± SD
	Age (year)	44±17
	Gender
	Male	15
	Female	62
	Race
	White	63
	African American	11
	Other	3
	Hypertension
	Pre-op SBP (mm Hg)	184±35
	Pre-op DBP (mm Hg)	107±19
	Pre-op BP meds (n)	1.9±1.1
	Renal function
	Pre-op serum creatinine	1.03±0.43
	Pre-op EGFR (ml/[min · 1.73 m2])	79±30

SD, Standard deviation; SBP, systolic blood pressure; DBP, diastolic blood pressure; EGFR, estimated glomerular filtration rate.

Measured clinical outcomes included blood pressure, number of antihypertensive medications, and serum creatinine (SCr). Analysis was performed using data from the preoperative assessment, discharge assessment, and the latest postoperative follow-up. Blood pressure response was estimated as previously described.24 Preoperative and postoperative SCr concentrations were measured, and estimated glomerular filtration rate (EGFR; mL/[min · 1.73 m2]) was calculated by a validated prediction equation.24 Early renal function was considered improved, unchanged, or worsened using a ≥20% change in SCr level at discharge. Acute tubular necrosis (ATN) was defined as a ≥20% increase in the SCr level during hospitalization >1.3 in women and >1.5 in men. Late renal function response was considered improved, unchanged or worsened using a ≥20% change in EGFR at last follow-up.

Surgical technique 

Isolated branch renal artery repair is performed through an extended subcostal incision. The patient is positioned with a roll beneath the ipsilateral flank and the ipsilateral arm is padded and tucked. A kidney rest is elevated and the patient is flexed to increase the distance between the costal margin and pelvic crest.

A right or left visceral mobilization is performed. On the right, adhesions between the anterior and posterior liver are divided first, retracting the liver superiorly. A continuous incision in the posterior peritoneum is made to mobilize the duodenum and the hepatic flexure of the colon, entering a retroperitoneal plane anterior to Gerota’s fascia to expose the right renal vein and vena cava. The duodenum and colon are swept medially and inferiorly to the left of the aorta. On the left, the descending colon and splenic flexure are mobilized at the lateral peritoneal reflection, developing a plane anterior to Gerota’s fascia and the left renal vein. In most cases, the spleen and pancreas can be left in position, mobilizing the inferior border of the pancreas and retracting this organ superiorly.

In branch renal artery exposure for either kidney, early identification and mobilization of the renal vein are key. The renal vein is mobilized from the renal hilum to caval origin. On the left, the adrenal, gonadal, and lumbar vein branches are ligated. When the renal artery pathology involves the distal renal artery and hilar vessels, the proximal renal artery is exposed first. When reconstruction is performed for failed percutaneous angioplasty or failed surgical repair associated with significant periarterial fibrosis, dissection is begun distally at the renal hilum through undisturbed tissue planes. During renal artery dissection small (6.25 to 12.5 g) doses of intravenous mannitol are administered intermittently and then repeated during periods of initial ischemia and reperfusion. Mannitol is given up to a maximum cumulative dose of 1 g/kg of patient body weight.

Cold perfusion protection is used when >30 to 40 minutes of warm renal ischemia is anticipated to complete the renal artery repair. Cold perfusion preservation is initiated after complete mobilization of the kidney from Gerota’s space and dissection of the hilar structures. Mobilization begins at the superior pole of the kidney after a cruciate incision is made in Gerota’s fascia. At the inferior pole of the kidney, the ureter is identified and mobilized to the level of the pelvic brim with abundant periureteric fibroareolar tissue. The ureter is left intact, but periureteric collaterals are controlled throughout the period of ischemia with doubly passed elastic loops. In the final portion of exposure, the posterior renal hilum is cleared of all surrounding tissues, thereby eliminating contamination of the systemic circulation with cold perfusate.

After exposure is completed, the patient is systemically heparinized. The renal artery is ligated and divided. The renal vein is controlled at its juncture with the vena cava. When exposure is adequate, the in situ technique leaves the renal vein intact and a venotomy is made before cold perfusion of the kidney (Fig 1, A). When exposure and reconstruction require elevation of the kidney from Gerota’s space, the ex vivo technique divides the renal vein with a cuff of vena cava (Fig 1, B). In either case, the kidney is placed on a plastic barrier drape, immersed in ice slush, and flushed with a commercially available renal preservation solution with intracellular composition that has been chilled to 4°C. The kidney should appear asanguineous, with clear effluent from the renal vein, which typically requires 300 to 500 mL of perfusate. The remainder of the hilar dissection is then completed sharply. Before final division of the renal artery branches, the kidney is once again infused with chilled perfusate and branch renal artery repair performed.

View Large Image Download to PowerPoint

Fig 1. Branch renal artery repair with cold perfusion preservation. A, In situ repair using cold perfusion. Renal mobilization and reconstruction is performed without transecting the renal vein. B, Ex vivo repair, where an ellipse of vena cava containing the renal vein origin is excised to aid in exposure. C, Syndactilization, or combining, of segmental branches and creation of the distal anastomosis. A proximal branch is sewn to the graft separately. D, Completed reconstruction.

In most instances, branch repair involves combining (or syndactylizing) renal artery branches, especially when multiple segmental renal arteries are reconstructed (Fig 1, C). The segmental vessels are tailored and sutured together to create one or two common orifices bearing the circulation to the kidney. The distal anastomosis between the syndactylized patch and the bypass conduit is then performed.

Once the distal arterial reconstruction is completed, the renal vein is repaired. For ex vivo repair, the kidney is placed in an orthotopic position and an anastomosis from the renal vein to the vena cava is created. For the in situ technique, the venotomy is simply closed.

In the final stage of reconstruction, the proximal bypass graft is anastomosed to the aorta. On the right, the renal artery conduit is routed posterior to the vena cava except in those cases when branch renal reconstruction is required for failed angioplasty or failed surgical repair. In these cases, there is often a retrocaval cicatrix that makes dissection in this area hazardous. Otherwise, the retrocaval space is mobilized, and one or two pairs of lumbar veins are ligated and divided. The infrarenal aorta is mobilized, preserving all lumbar arteries, and then controlled with clamps. An anterolateral ellipse of aortic wall is removed using several applications of an aortic punch. The renal conduit is then spatulated and sewn in an end-to-side fashion to the aorta (Fig 1, D).

For the entirety of the ischemic period, the kidney is maintained in ice slush. Before reperfusion, the kidney is removed from slush, the renal vessels are released simultaneously, and the ureter loop is released. Ureteric peristalsis resumes within 15 minutes in most cases.

Regardless of the method of reconstruction, each repair is evaluated with intraoperative RDUS imaging. B-scan images are obtained from all areas of operative dissection and vascular reconstruction. A B-scan defect is imaged in both longitudinal and transverse projection, and Doppler samples are obtained proximal and distal to an imaged defect to determine their contribution to flow disturbance. Color-flow is taken from the renal parenchyma in the upper, middle, and inferior kidney, and Doppler parenchymal signals are obtained. The criteria for major B-scan defects requiring immediate revision are (1) a focal peak systolic velocity ≥1.8 m/s with a distal turbulent waveform and (2) no Doppler-shifted signal from the renal artery (consistent with occlusion).

Statistical methods 

Data were examined for all patients who underwent branch renal artery reconstruction using cold perfusion protection. Descriptive statistics were computed, including means ± standard deviations of continuous variables and frequencies and percents of categoric variables. Changes in blood pressure and renal function (SCr or EGFR) were assessed using paired t tests. Censored outcomes (primary renal artery patency and patient survival) were estimated by life-table methods. Graphic depiction of patency was constructed using product-limit estimates (Fig 2).

View Large Image Download to PowerPoint

Fig 2. Product-limit estimates of primary patency (blue) and primary-assisted patency (red) in 68 arteries. The standard error of the estimate never exceeds 10%.

Operative management 

Cold perfusion was used to repair 78 renal arteries by ex vivo (49 kidneys) or in situ (29 kidneys) techniques (Table II). Of the 29 patients with bilateral disease or disease to a solitary kidney, bilateral renal artery repair was performed in eight patients, with 13 repairs to solitary kidneys. Bilateral renal artery repairs were performed in a staged fashion (mean interval, 8.6 months). One patient had bilateral cold perfusion renal artery reconstruction and seven had unilateral reconstruction using cold perfusion protection and contralateral warm in situ reconstruction. Renal artery disease included renal artery aneurysm (RAA) in 50 patients, fibromuscular dysplasia (FMD) in 37, atherosclerosis in 5, and arteritis in 2; 16 patients had both FMD and RAA. Renal artery reconstruction was performed in six patients who had failed PTRA (4 ex vivo, 2 in situ). All six patients had FMD and significant hypertension.

Table II. Characteristics of renal artery repairs

	Characteristic	RAs
	Unilateral repair	57
	Right	40
	Left	17
	Bilateral repair	8
	Bilateral cold perfusion	1
	Unilateral cold perfusion	7
	Solitary kidneys	13
	Associated conditions
	Fibromuscular dysplasia	37
	Atherosclerosis	5
	RAA	50
	Arteritis	2
	Ex vivo	49
	In situ	29
	Branch number⁎	2.8 ± 1.6
	Cold perfusion after:
	Failed PTRA	6
	Failed open RA repair	1

RA, Renal artery; RAA, renal artery aneurysm; PTRA, percutaneous transluminal renal artery angioplasty.

⁎ Branch number expressed as mean ± standard deviation.

Of 78 renal artery repairs, saphenous vein was used in 69, hypogastric artery in 3, composite vein/polytetrafluoroethylene (PTFE) in 2, PTFE in 2, cephalic vein in 1, and aneurysmorrhaphy in 1. One planned nephrectomy was performed at the time of contralateral ex vivo renal artery reconstruction for unreconstructable renal artery disease. No unplanned nephrectomies were required for intended reconstruction. Each renal artery reconstruction required branch dissection and reconstruction (mean 2.8 ± 1.6 branches were repaired). Mean operative time was 7.75 ± 1.2 hours (median, 8 hours; range, 5 to 11 hours). Mean cold ischemia time was 125 ± 40 minutes. Each kidney was reconstructed in an orthotopic fashion.

Intraoperative duplex imaging (48 kidneys) and early postoperative angiography/RDUS imaging (18 patients) defined significant technical defects in 12 patients (18.2%). Seven renal arterial reconstructions with major B-scan defects noted on intraoperative RDUS imaging were revised immediately, with completion duplex imaging to insure a satisfactory result. Postoperative angiography or RDUS imaging, or both, demonstrated five occlusions.

Morbidity and mortality 

Perioperative complications that extended the hospital stay occurred in 5.2% of patients. These included ileus in 1 patient, wound infection in 2, sacral decubitus ulcer in 1, and respiratory failure, pneumonia, and need for temporary dialysis in 1 patient. The latter patient underwent renal artery bypass to a solitary kidney, which thrombosed in the immediate perioperative period requiring emergent surgical revision. Renal failure developed despite a patent graft, and he required 1 week of hemodialysis. After 1 month of hospitalization, the patient was discharged home off dialysis, with a stable EGFR. No additional hemodialysis was required during 19 months of follow-up.

Five early repair failures required three nephrectomies (Table III). One perioperative graft occlusion to a solitary kidney was reoperated on with successful kidney salvage (as noted previously). The remaining patients were asymptomatic, and their graft occlusions were identified on routine postoperative angiography or RDUS imaging, or both.

Table III. Outcomes

	Outcome	N (%)
	Hypertension response⁎
	Cured	11(15)
	Improved	47(65)
	Failed	14(20)
	Renal function
	Early
	Improved	21(35)
	Unchanged	29(48)
	Worsened	10(17)
	Late
	Improved	13(37)
	Unchanged	16(46)
	Worsened	6(17)
	Patency†
	Primary	85±5
	Primary-assisted	93±4
	Occlusions	5±7.4
	Anatomic follow-up‡	34±7
	Mortality
	In-hospital	1(1.3±1.3)
	Long-term mortality	3(3.9±2.2)

⁎ Hypertension response, renal function response, occlusions, and mortality expressed as number of patients (percentages). † Patency at 12 months, expressed as percentages with standard error. ‡ Expressed as months ± standard deviation.

One perioperative death occurred (1.3%) in a patient who had advanced cirrhosis and a large, symptomatic RAA. Postoperative hemorrhage led to multisystem organ dysfunction, acute respiratory distress syndrome, and death.

No late graft occlusions occurred, and no patients progressed to dialysis-dependence. Long-term follow-up for vital status was a mean 7.3 years (range, 0 to 17 years; median, 7.7 years). Two late deaths occurred, one at 8.4 years and one at 14.2 years. The product-limit estimate of survival was 96% at 10 years.

Blood pressure response 

Preoperative hypertension was present in 93.5% of patients (mean blood pressure, 184 ± 35/107 ± 19 mm Hg; mean number medications, 1.9 ± 1.1). Among the 72 hypertensive patients, 15% were cured, 65% were improved, and 20% were considered failed after renal artery repair (Table III).24 Considered collectively, hypertensive patients had decreased systolic and diastolic blood pressures (141 ± 21/83 ± 12 mm Hg; P < .0001) as well as decreased medication requirements for adequate blood pressure control (medications, 1.5 ± 1.5; P < .0001). Of the six patients who underwent repair after failed PTRA, hypertension was cured in four and improved in two. Bilateral disease or disease to a solitary kidney was noted in 29 patients, of which eight had bilateral repair, including one with bilateral cold perfusion. Eight patients with bilateral disease had only unilateral repair when blood pressure was cured or improved after the first operation.

Early renal function 

Early renal function was improved in 35% of patients, preserved in 48%, and worse in 17% after renal artery repair (mean preoperative SCr, 1.0 ± 0.36; mean discharge SCr, 0.91 ± 0.45; Table III). Four patients (7.1%) had ATN in the postoperative period. Mean preoperative SCr in these patients was 1.83 ± 0.72 compared with a postoperative SCr of 2.23 ± 0.67. One of the four patients with ATN required temporary hemodialysis (as described previously). Two of the patients with ATN had early graft occlusions, one of which was noted on routine postoperative angiography.

Renal function response 

Late renal function response was available for 35 of 77 patients (Table III). The average time to follow-up EGFR was 12 months (median, 3.4 months; range, 0.70 to 80 months). The EGFR was improved in 37%, unchanged in 46%, and worse in 17% (mean preoperative EGFR, 76.2 ± 31.7; mean follow-up EGFR, 83.0 ± 41.9; mean change EGFR, 6.8 ± 30.5). In the eight patients with preoperative renal insufficiency and follow-up, renal function was improved in four, unchanged in three, and worse in one (mean preoperative EGFR, 41.3 ± 14.7; mean follow-up EGFR, 55.1 ± 28.2). Three of the four patients with perioperative ATN had follow-up SCr for analysis. Of those, two had unchanged EGFR on last follow-up and one had a slow decline in EGFR over 3 years despite a patent graft (preoperative EGFR, 53.8; follow-up EGFR, 26.0). No patients have required permanent dialysis in a mean follow-up of 34 months.

Estimated patency of repair 

Of the 77 patients, 68 had anatomic follow-up (mean, 34 ± 47 months; range, 1 week to 198 months). Estimated primary patency of renal artery repair using angiography or RDUS imaging, or both, was 85% ± 5% at 12 months (Table III). Graft stenoses were demonstrated in five patients, three on surveillance duplex imaging and two on follow-up angiography performed for recurrent hypertension. Of these, 1 refused reintervention, 2 underwent successful operative revision at the time of a contralateral renal artery repair, 1 had successful angioplasty and endoluminal stenting, and 1 patient had operative revision to a solitary kidney. The product-limit estimate of primary assisted patency was 93% ± 4% at 12 months. (Fig 2, Appendix, online only.)