Renal artery embolization: clinical indications and experience from over 100 cases


  • Michael J. Schwartz,

    1. Department of Urology and Division of Cardiovascular Interventional Radiology, The New York-Presbyterian Hospital, Weill Medical College of Cornell University, NY, USA
    Search for more papers by this author
  • Eric B. Smith,

    1. Department of Urology and Division of Cardiovascular Interventional Radiology, The New York-Presbyterian Hospital, Weill Medical College of Cornell University, NY, USA
    Search for more papers by this author
  • David W. Trost,

    1. Department of Urology and Division of Cardiovascular Interventional Radiology, The New York-Presbyterian Hospital, Weill Medical College of Cornell University, NY, USA
    Search for more papers by this author
  • E. Darracott Vaughan Jr

    1. Department of Urology and Division of Cardiovascular Interventional Radiology, The New York-Presbyterian Hospital, Weill Medical College of Cornell University, NY, USA
    Search for more papers by this author

Michael J. Schwartz, The New York-Presbyterian Hospital, Weill Cornell Medical Center, Department of Urology, 525 East 68th Street, Starr 900, New York, NY 10021, USA. e-mail:


Renal artery embolization is used for various reasons, and authors from the USA review their experience of using this technique before nephrectomy in patients with large renal masses. They found it to be safe and effective, with minimal morbidity. They note the lack of published prospective randomized trials, and suspect that this contributes to it being underused.


To review current indications and techniques for renal artery embolization (RAE) and more specifically to review cases of RAE before nephrectomy for treating patients with a large renal mass.


All RAEs done at our institution between May 1993 and December 2005 were reviewed. Patients were identified using a database assembled by the Division of Cardiovascular Interventional Radiology. Indications, techniques and RAE-related complications were then obtained from a retrospective review of medical records. Additional data for patients undergoing preoperative infarction were acquired, including estimated blood loss (EBL), transfusion requirement, pathological size, subtype, grade, stage, and level of tumour thrombus if present.


In all, there were 121 RAEs, 69 in males and 52 in females (mean age 57.6 years, range 11–89). Metallic microcoils were the most often used embolization agent, followed by acrylic microspheres (embospheres), polyvinyl alcohol particles, absolute ethanol, and Gelfoam (Pharmacia & Upjohn, USA). The most common indication for RAE was infarction before nephrectomy (54.5%). Other indications included symptomatic angiomyolipomas, palliation of unresectable renal cancer, haemorrhage, perinephric bleeding in end-stage renal disease, vascular lesions, malignant hypertension, and sequelae of end-stage renal disease. RAE-associated complications including coil migration, incomplete embolization, and groin haematoma (in 5.0%). Symptoms of post-infarction syndrome were common, with 74.4% of patients having flank pain, nausea, or vomiting; the vast majority of these symptoms were mild and self-limited. In patients having nephrectomy after RAE the median (range) interval from RAE was 2 (0–78) days. The mean tumour size was 11.2 (3.5–25) cm and 46% of patients had tumour thrombus present in either the renal vein or inferior vena cava (IVC). The mean (median) overall EBL in patients having nephrectomy after RAE was 1048 (725) mL. The mean transfusion requirement over the course of hospitalization was 3.9 units of packed red blood cells.


RAE is a safe and effective therapeutic tool for many urological, renal and vascular conditions. Its use has increased at our institution due to improved techniques, embolization materials, and our increasing use of RAE as an adjuvant procedure for patients requiring nephrectomy with or without IVC thrombectomy. There are many potential operative advantages for patients having RAE before surgery, with minimal morbidity. It is likely that the lack of prospective randomized trials is the primary reason why it is underutilized in the preoperative setting.


renal artery embolization




estimated blood loss


inferior vena cava.


The initial indications developed in the 1970s for renal artery embolization (RAE) were limited to symptomatic haematuria and palliation for metastatic renal cancer [1,2]. With technical advances and growing experience the indications have broadened to include conditions such as vascular malformations, medical renal disease, angiomyolipomas (AMLs), and preoperative infarction. The introduction of smaller delivery catheters and more precise embolic agents has drastically improved the morbidity associated with this technique [3]. RAE has continued to gain popularity as a minimally invasive approach for various urological conditions.

However, opinions on the role of preoperative RAE in the management of patients with RCC are controversial. Although a significant number of studies on RAE are reported in these patients, there is no consensus on the benefits and morbidity associated with the procedure [4–6]. Moreover, many large studies on the use of RAE were conducted in the 1980s, before the development of improved techniques and imaging.

If there is a real benefit to be gained, most proponents of preoperative RAE cite the facilitation of nephrectomy through decreased operative blood loss, ease of dissection secondary to the development of oedema in tissue planes, and decreased operative time [7,8]. For those patients with significant tumour thrombus there might be a beneficial effect of decreasing the size or extent of tumour thrombus before surgery [9]. Interestingly, there might also be an advantage in the form of immunomodulation, whereby RAE-induced tumour necrosis stimulates a tumour-specific response [10–12]. It is likely that RAE is underutilized, perhaps because of a lack of prospective randomized studies demonstrating these potential benefits.

The purpose of the present study was to review and revisit current indications and techniques for RAE, with particular emphasis on preoperative RAE of renal tumours. We present our experience with RAE for various indications in 121 patients over the last 13 years.


We retrospectively assessed all RAEs done between January 1993 and December 2005 at our institution. Patients were identified using a database assembled by the Division of Cardiovascular Interventional Radiology. The indications, techniques and complications were then obtained from patients’ medical records. For patients undergoing subsequent surgery for renal tumours, additional peri-operative and pathological data were analysed. These data included estimated blood loss (EBL), transfusion requirement, presence of tumour thrombus, and pathological tumour size, subtype, grade and stage.


In all, 121 RAEs were done at our institution in the study period; 69 patients were male and 52 were female (mean age 57.6 years, range 11–89). Table 1 summarizes the different indications for RAE; the most common (73.6%) was for renal tumours (Fig. 1), including 66 (54.5%) before nephrectomy, 15 (12.4%) for symptomatic AMLs (Fig. 2), and eight (6.6%) for palliation of unresectable renal cancer. Seventeen percent were performed for persistent haemorrhage (nine iatrogenic, eight haematuria, four perinephric bleeding in end-stage renal disease, two retroperitoneal bleeding from renal tumours), and 12% were for vascular lesions (Fig. 3; six arteriovenous fistulae, four arteriovenous malformations, four pseudoaneurysms, three before placing an endograft for abdominal aortic aneurysm repair). Renal disease accounted for 4% (three malignant hypertension, one severe nephrotic syndrome, one severe hydronephrosis); 16 patients had multiple indications.

Table 1. 
The indications for RAE
Indications*N (%)
  • *

    16 patients had more than one indication. AAA, abdominal aortic aneurysm.

Before nephrectomy for renal mass66 (54.5)
Persistent bleeding23 (19.0)
AML15 (12.4)
Palliation 8 (6.6)
Arteriovenous fistula 6 (5.0)
Arteriovenous malformation 4 (3.3)
End-stage renal disease 4 (3.3)
Before endograft placed for AAA repair 3 (2.5)
Pseudoaneurysm 3 (2.5)
Malignant hypertension 3 (2.5)
Severe hydronephrosis 1 (0.8)
Severe nephrotic syndrome 1 (0.8)
Figure 1.

(A) Aortogram showing contrast medium in bilateral renal vessels and downward displacement of the left kidney by a large upper-pole tumour. (B) Collateral blood supply to the same upper-pole renal tumour. (C) An aortogram after RAE showing no residual flow to left kidney via the main renal artery or collateral vessels.

Figure 2.

(A) A renal angiogram showing the entire arterial tree and lower-pole AML. (B) A selective angiogram showing the lower-pole AML with the vascular supply in greater detail. (C) An angiogram after RAE showing no blood flow to the tumour and preservation of blood supply to the normal portion of the kidney.

Figure 3.

(A) A renal angiogram showing an arteriovenous fistula. The white arrowhead marks venous filling extending into the IVC during the arterial phase as the parenchyma are just beginning to enhance. (B) Selective embolization of the same arteriovenous fistula with a coil, seen in the arterial branch (white arrowhead). A small parenchymal wedge-shaped defect is seen, resulting from the embolization procedure.

The number of embolization procedures at our institution increased over time (Fig. 4). Various embolic agents were used, including metallic coils, polyvinyl alcohol particles, ethanol, Gelfoam (Pharmacia & Upjohn, USA), and more recently, acrylic microspheres (embospheres). Our embolization agents of choice in the 1990s were absolute ethanol or microcoils, or a combination of the two. Ethanol now has a very limited role in our options for embolization, as embospheres have emerged as the new agent of choice. However, metallic coils maintain a dominant role and 87% of the agents used in 2005 were either embospheres or microcoils, while ethanol was not used at all (Fig. 5).

Figure 4.

The number of RAEs with time.

Figure 5.

The use of embolization agents for RAE with time.

Of the 66 patients treated by RAE before nephrectomy, all went on to surgery; the median (range) interval between RAE and surgery was 2 (0–78) days. Pathology was available for 63 of these 66 patients; the mean (range) tumour size was 11.2 (3.5–25) cm. The final pathological analysis of these tumours is summarized in Table 2. Information about presence or absence of tumour thrombus was available for 63 patients, 29 (46%) of whom had tumour thrombus involving the renal vein (seven) or inferior vena cava (IVC, 22). The staging classification according to the TNM staging system for RCC is summarized in Table 3 for those tumours with conventional clear cell, papillary and chromophobe histology. The EBL was available in 63 of 65 patients (one was excluded from EBL statistics, who had a planned simultaneous abdominal aortic aneurysm repair). We included both the mean and median EBL to account for several significant outliers. The overall mean (range) EBL was 1048 (100–5000) mL and the median 725 mL. In cases with no tumour thrombus, the mean EBL was 647 mL with a median EBL of 425 mL (Table 4). The overall average transfusion requirement over the patient’s hospital course was 3.9 units of packed red blood cells.

Table 2.  The pathology of nephrectomy specimens after RAE
PathologyN (%)
  • *

    spindle cell carcinoma, liposarcoma, adrenal carcinoma, unspecified RCC, adult polycystic kidney disease.

Clear cell carcinoma44 (69.8)
Papillary carcinoma 6 (9.5)
Oncocytoma 2 (3.2)
Chromophobe 2 (3.2)
AML 4 (6.3)
Other* 5 (7.9)
Total63 (100)
Table 3.  TNM staging classification for cases of RCC
Table 4. 
The EBL by level of tumour thrombus
Tumour thrombus involvementNumber of patientsEBL, mL mean (median)
No involvement34 647 (425)
Renal vein only 7 700 (400)
Subhepatic IVC 71250 (1150)
Retrohepatic IVC 91974 (1200)
Atrial 62350 (2500)
Overall mean631048 (725)

There were six RAE-related complications in 121 patients, i.e. incomplete embolization in two patients, who required a repeat procedure to achieve complete infarction of bleeding AMLs, a coil migration in two patients (during each case the coil was retrieved using an endovascular grasper with no complication) and two patients developed small groin haematomas, neither of which required transfusion. Symptoms of post-infarction syndrome, including nausea and flank pain, were reported by ≈ 75% of patients, most being mild and self-limited (Table 5). Excluding those patients with post-infarction syndrome, the overall complication rate was 5%. There were no cases of adjacent organ injury or RAE-related death.

Table 5.  The morbidity and complications of RAE
Morbidity/complicationN (%)
Post-infarction syndrome:90 (74.4)
 Flank pain74 (61.1)
 Fever45 (37.2)
 Nausea29 (24.0)
 Vomiting 11 (9.1)
 Incomplete embolization 2 (1.7)
 Coil migration 2 (1.7)
 Groin haematoma 2 (1.7)
 Adjacent organ necrosis 0
 Total 6 (5.0)


Unlike in similar reviews, most of the present patients (54.5%) had RAE before surgical resection of a large renal mass. Despite a large mean tumour size of 11.2 cm and that 46% of surgical cases involved tumour thrombi within the renal vein or IVC, the overall mean and median EBL were only 1048 and 725 mL, respectively. While it can be extremely difficult to accurately assess and compare blood loss between embolized and non-embolized groups, the consensus at our institution is that there is a trend towards less blood loss in embolized patients, especially in those with tumours of >10 cm. Other vascular benefits include the potential for early ligation of the renal vein before dividing the renal artery. This option might be particularly advantageous in patients with significant peri-hilar disease or adenopathy, where early control of the renal artery is difficult or would considerably lengthen the surgery.

The delay to surgery after RAE in the present series was several hours to 78 days, with a median of 2 days. The optimum delay would: (i) ideally maximize the benefits of tissue oedema after RAE, for ease of dissection; (ii) allow the surgeon to proceed before collateral vessels formed; and (iii) minimize the period of post-infarction syndrome for the patient. Most likely this optimum delay is 24–48 h, but in patients with tumour thrombi extending into the right atrium or retrohepatic IVC, the optimum delay to surgery might be longer. Preoperative embolization might allow the size and level of the tumour thrombus to decrease, potentially sparing a patient the otherwise obligatory cardiopulmonary bypass or liver mobilization, and significant operative morbidity. In patients with tumour thrombi, there has been discussion about the theoretical risk of pulmonary embolus secondary to RAE. Of 29 patients in the present series with tumour thrombi and treated by RAE before surgery, there were no cases of pulmonary embolus after infarction, including those with a significant interval between RAE and nephrectomy.

Several investigators have cited immunological benefits for patients undergoing RAE before surgery, including augmentation of natural killer cell activity and lymphoproliferative responses [10,11]. In addition, a retrospective study of RAE followed by nephrectomy vs a case-matched control group of nephrectomy alone showed a significant survival benefit in the former group [12]. There was a survival benefit in patients with pT2 and pT3 disease, and in patients with positive lymph nodes at the time of surgery. However, several studies in the 1980s showed no survival benefit in patients with metastatic disease undergoing RAE with or with no subsequent nephrectomy [13,14]. Further investigation through prospective randomized trials is certainly warranted to elucidate the potential survival benefits in patients with and without metastatic disease.

RAE has now become the initial treatment of choice for most symptomatic AMLs [15–18]. As evidence of this trend, 12.4% of the procedures in our review were for this indication, most within the last 5 years. Other conditions have also been effectively treated with RAE at our institution. Renal diseases, vascular malformations and iatrogenic injuries were a significant proportion (18%) of the indications, as in other reports [19,20]. As progress is made in this field the indications will probably continue to increase.

RAE has changed significantly over the past 15 years; imaging capabilities have dramatically improved, allowing better location of target arteries. Likewise, the embolic material and delivery catheters have improved, allowing for more precise embolization [21]. Together, these advances have resulted in more effective treatments with fewer complications. In a review of our 13-year experience, the most common complaints after RAE were symptoms of the well-described post-infarction syndrome, including nausea, vomiting, fevers and flank pain [22,23]. These symptoms were self-limited and easily controlled with narcotic and anti-emetic medications, although most patients required overnight hospitalization. We found no significant difference in post-infarction syndrome based on the type of embolization agent used. In the present series there were six complications in 121 procedures. Complications due to incomplete embolization or coil migration could both be expected to decrease with improving technology and techniques. An example of such new technology is the increasing use of acrylic microspheres at our institution over the past 6 years. This new embolic material allows for more effective embolization with no risk of coil migration, and because of the extremely precise delivery, there is less risk of injury to adjacent normal tissues.

In conclusion, RAE is an effective therapeutic and adjuvant tool for many urological conditions, and in patients with vascular pathology or end-stage renal disease. With current techniques it is well tolerated with few complications. In addition, it might facilitate the dissection of large renal tumours and tumours with extensive involvement around the renal hilum. RAE has the potential to decrease operative blood loss and operative time, leading to lower overall morbidity and transfusion requirement. While these trends have been recorded at our institution, additional prospective studies are needed to draw any definitive conclusions one the role of preoperative RAE for either intraoperative or immunological benefit. It is likely that the lack of these prospective studies is the primary reason that RAE is not used more often before surgery. RAE will probably continue to become safer and more effective in the future, providing urologists with an alternative treatment for various genitourinary conditions.


None declared.