• RCC;
  • tumour thrombus;
  • bypass;
  • nephrectomy;
  • thrombectomy


  1. Top of page
  2. Abstract

Study Type – Therapy (case series)

Level of Evidence 4

What’s known on the subject? and What does the study add?

Removal of a renal cell carcinoma with a level IV tumour thrombus is a challenging surgery and generally is performed in a tertiary care centre. Performing these cases generally requires a multi-disciplinary approach consisting of urological and vascular/cardiovascular surgeons.

This study sheds light on the high surgical morbidity and mortality of these cases even at experienced centres. For patients requiring cardiopulmonary bypass, approximately 20% may not survive. In these challenging surgeries, deep hypothermic circulatory arrest may limit mortality and further studies should investigate the protective effect of this modality.


• To review experience with nephrectomy/thrombectomy for a renal cell carcimoma (RCC) with a level IV tumour thrombus and to evaluate the benefit of deep hypothermic circulatory arrest (DHCA) with cardiopulmonary bypass (CPBP).


• A multi-institutional retrospective database was created to assess the outcomes of surgery for RCC and associated level IV tumour thrombus from 1983 to 2007. Patients were identified based on radiographic records/operative findings.

• Only cases using CPBP were analysed. Clinicopathological and operative characteristics including use of DHCA were recorded.

• Overall survival (OS) for all patients and by use of DHCA was assessed. Comparisons of clinical and operative characteristics by use of DHCA were performed.

• A Cox regression model determined predictors of perioperative/in-hospital mortality.


• In all, 63 patients underwent resection with CPBP; overall perioperative mortality was 22.2%.

• There were no significant differences in clinicopathological characteristics, operative duration, estimated blood loss, transfusions, and hospital stay by use of DHCA.

• Perioperative mortality rate was lower in patients undergoing DHCA (8.3% vs 37.5%, P= 0.006).

• The median OS was longer for the patients undergoing DHCA (15.8 vs 7.7 months); however, this failed to reach statistical significance (P= 0.357).

• On multivariate analysis, age of >60 years (hazard ratio [HR] 6.7, 95% confidence interval [CI] 1.5–31.1, P= 0.015) and the use of DHCA (HR 0.13, 95% CI 0.036–0.51, P= 0.003) were independent predictors of perioperative mortality.


• Radical nephrectomy and level IV tumour thrombectomy is associated with significant mortality.

• The use of DHCA does not appear to adversely affect operative characteristics and may limit perioperative mortality.

• Further prospective studies should be performed to confirm the benefit of DHCA.


deep hypothermic circulatory arrest


cardiopulmonary bypass


Eastern Cooperative Oncology Group Performance Status


estimated blood loss


overall survival


hazard ratio


  1. Top of page
  2. Abstract

RCC commonly invades the renal vein and can extend into the vena cava in up to 10% of cases. While the prognostic influence of tumour thrombus extension is debated, the more cephalad extent alters surgical approach. Surgical resection of more extensive tumour thrombi is associated with longer operative duration, greater blood loss, and higher perioperative mortality [1,2].

Several classification schemes have been developed to communicate the level of extension [2–5]. The most widely used system considers supra-diaphragmatic and atrial extension as level IV thrombus [2]. These cases often required a multi-disciplinary surgical approach. The need for proximal control of the supra-diaphragmatic vena cava or right atrium often necessitates cardiopulmonary bypass (CPBP) as most patients cannot tolerate loss of venous preload. While other methods such as venovenous bypass are possible for level IV thrombus, use is limited to tumours with minimal atrial involvement of a free-floating thrombus. The use of deep hypothermic circulatory arrest (DHCA) during CPBP was employed in the early series of level IV tumour thrombus resection. After cooling the patient’s core body temperature to 18 °C, the bypass circuit is closed for the critical components of the operation. This allows for inspection of the vena cava and atrium in a bloodless field and extends safe ischaemic time up to 60 min [6]. Once the resection has been completed, the patient is warmed and the circuit is re-started. While this approach may prevent tumour embolization or hepatic/renal ischemia, there are significant risks associated with DHCA, including: coagulopathy, renal failure, neurologic injury, and retroperitoneal haemorrhage [7,8].

Despite the inherent risks, surgical resection remains the mainstay of treatment for patients with level IV tumour thrombus. Due to the rarity of RCC associated with level IV tumour thrombus research aimed at determining the outcomes and the optimal approach of resection is frequently hindered. Therefore, we aimed to generate a multi-institutional database to assess the surgical outcome and define the role of DHCA in conjunction with CPBP.


  1. Top of page
  2. Abstract

A multi-institutional retrospective database was created to evaluate the surgical outcome for cases of RCC with associated level IV tumour thrombus over a 25-year period from 1983 to 2007. Each centre obtained Institutional approval before participation in the study. Patients were identified based on radiographic records and/or operative findings and classified according to the classification proposed by Blute et al. [2]. Due to the infrequent use of non-bypass techniques or venovenous bypass, only cases using CPBP were analysed. Patient and pathological characteristics including age, Eastern Cooperative Oncology Group Performance Status (ECOG PS; 0, 1, >1), T classification (T3c, T4), Fuhrman grade, tumour size, and metastatic status (M0, M1) were recorded. The use of DHCA, surgical duration, estimated blood loss (EBL), intraoperative blood transfusions, hospital stay, and perioperative/in-hospital mortality were recorded. Each institution was responsible for providing clinical and pathological data. Perioperative morality was defined as inpatient mortality before initial discharge after nephrectomy. Overall survival (OS) was assessed from the date of surgical resection to the date of death by any cause or date of last follow-up. For assessment of disease-specific survival, events were considered to be death related to advanced RCC and not those attributed to surgical complications.

Comparisons of clinical and operative characteristics were performed between bypass techniques. Categorical patient characteristics were compared by chi-square tests while continuous variables were analysed with Students’t-test. A Cox regression model was applied to determine predictors of perioperative mortality and included the variables of grade (2, 3, and 4), T classification, performance status (0, ≥1), metastatic status, tumour size, use of DHCA, and age. Disease-specific survival and OS were assessed according to bypass technique by the Kaplan–Meier method and groups were compared with the log-rank test.


  1. Top of page
  2. Abstract

In all, 63 patients undergoing surgery were identified. The clinicopathological characteristics of all patients are summarized in Table 1. The mean (median; sd) age was 60.6 (62; 12.1) years and men accounted for 61.9% (39 men) of the cohort. Patients undergoing resection frequently had good performance status, with 91.4% having ECOG PS of 0/1. In all, 21 (33.3%) of the patients had evidence of metastasis at the time of surgery.

Table 1.  The patients’ clinical, pathological and operative characteristics
  • *

    for surviving patients; M, male; F, female.

Number of patients63
Mean (median; sd) age, years60.6 (62; 12.1)
Sex M/F, n (%)39 (61.9)/24 (38.1)
T stage, n (%) 
 T3c54 (91.5)
 T4 5 (8.5)
Mean (median; sd) tumour size, cm 9.8 (10; 3.7)
Metastases, n (%) 
 No42 (66.7)
 Yes21 (33.3)
Fuhrman grade, n (%) 
 2 8 (12.7)
 341 (65.1)
 414 (22.2)
ECOG PS, n (%) 
 032 (55.2)
 121 (36.2)
 >1 5 (8.6)
Mean (median; sd): 
Operative duration, min (n= 57) 379.6 (370; 108.8)
EBL, mL (n= 29)3885 (3000; 4279)
Transfusions, n (n= 58)  16.3 (12.5; 14)
Hospital stay, days (n= 49*)  13.8 (10; 12.8)
N (%): 
Perioperative mortality (n= 63)  14 (22.2)

Renal tumours were a mean (median; sd) size of 9.8 (10.0; 3.7) cm. T classification was available in 59 cases and was T3c and T4 in 54 (91.5%) and five (8.5%), respectively. Tumour grade was most commonly high grade with 41 (65.1%) and 14 (22.2%) cases Fuhrman grade 3 and 4, respectively.

The mean (median; sd) operative duration was 379.6 (370; 108.8) min. and EBL was 3885 (3000; 4279) mL. The mean (median; sd) number of blood transfusion units was 16.3 (12.5; 14) and the hospital stay of surviving patients was 13.8 (10; 12.8) days. The overall perioperative mortality was 22.2%, as 14 of 63 patients did not survive to discharge. The operative characteristics are listed in Table 1.

The median (95% CI) OS for the cohort was 10.8 (6.6–21.7) months (Fig. 1a). In all, 42 (66.7%) patients died during the study period. The median (range) follow-up of the surviving patients was 24.6 (1–92) months. The median (95% CI) disease-specific survival was 21.7 (10.8 to ‘not yet reached’) months (Fig. 1b). The OS for those without documented metastatic disease was improved (13.9 vs 8.9 months); however, this failed to achieve statistical significance (P= 0.12).


Figure 1. The OS (a) and disease-specific survival (b).

Download figure to PowerPoint

Information on usage of DHCA was available for 60 patients; 36 (60%) underwent DHCA. Comparisons of patient characteristics based on bypass technique are given in Table 2. There were no clinical or pathological differences between patients undergoing either bypass technique. The mean operative duration, EBL, transfusions, and hospital stay were also similar between patients by bypass technique (Table 2). Perioperative mortality was significantly decreased for patients undergoing DHCA (8.3% vs 37.5%, P= 0.006). There was no difference in OS by use of DHCA (P= 0.357; not shown).

Table 2.  The overall and operative characteristics of patients compared by use of DHCA
  1. NS, not statistically significant; M, male; F, female.

Number of patients2436 
Mean (sd) age, years59.5 (12.4)60.7 (11.6)NS (0.712)
Sex M/F, n (%)14 (58.3)/10 (41.7)23 (63.9)/13 (36.1)NS (0.665)
T stage, n (%)  NS (0.067)
 T3c20 (83.3)33 (97.1) 
 T4 4 (16.7) 1 (2.9) 
Mean (sd) tumour size, mL 9.7 (3.3)10.1 (3.9)NS (0.682)
Metastases, n (%)  NS (0.576)
 No15 (62.5)25 (69.4) 
 Yes 9 (37.5) 11 (30.6) 
Fuhrman grade, n (%)  NS (0.606)
 2 4 (16.7) 3 (8.3) 
 315 (62.5)24 (66.7) 
 4 5 (20.8) 9 (25.0) 
ECOG PS, n (%)  NS (0.144)
 010 (41.7)21 (67.8) 
 112 (50) 8 (25.8) 
 >1 2 (8.3) 2 (6.5) 
Operative duration, min  NS (0.489)
 n20  34 
 Mean (median) 364.7 (367) 386.3 (371) 
EBL, mL  NS (0.398)
 n  13  14 
 Mean (median)2500 (2777)3240 (3398) 
Transfusions, n  NS (0.169)
 n20  35 
 Mean (median)12.5 (11.5)  16.7 (13) 
Hospital stay, days  NS (0.121)
 n15  33 
 Mean (median)10.7 (9)  15.4 (11) 
Perioperative mortality, n  0.006
 n24  36 
 N (%) deaths 9 (37.5)   3 (8.3) 

Risk factors potentially contributing to perioperative mortality were assessed. Higher Fuhrman grade, T stage, worse ECOG PS, larger tumour size, advanced age, and CPBP without DHCA were all associated with increased perioperative mortality. A Cox multivariate regression model showed that age and DHCA status were independent predictors of perioperative mortality. Patients aged >60 years were at an increased risk of perioperative mortality (hazard ratio [HR] 6.71, 95% CI 1.45–31.1; P= 0.015) while use of DHCA was protective (HR 0.13, 95% CI 0.036–0.510; P= 0.003).


  1. Top of page
  2. Abstract

Few series assess the operative characteristics, perioperative mortality, and long-term survival for patients with RCC with level IV tumour thrombus. Those series that report outcomes are limited by small patient numbers and limited follow-up. Additionally, several series combine patients with different bypass techniques and include both level III and IV thrombi. The present series is unique as it evaluates a large series of patients with level IV tumour thrombus from RCC from several tertiary referral centres and evaluates the outcomes with the standard surgical approach, CPBP. Interestingly, the current series suggested a decreased rate of perioperative mortality with the use of DHCA.

Surgical resection of renal tumours with level IV tumour thrombi is clearly a challenging endeavour. To minimize mortality, these cases should probably only be attempted in experienced centres in conjunction with a cardiothoracic surgery team. However, even in centres of excellence, in the present series mortality was relatively high at 22% in patients undergoing CPBP. Despite this high mortality rate, immediate surgical intervention remains the first-line of therapy in patients with level IV tumour thrombus given the palliation and potential cure offered with surgical resection. While there is some evidence suggesting that neoadjuvant usage of targeted therapy is safe [9–12], only a few case reports document a possible advantage (thrombus downsizing) in patients with tumour thrombus [13,14].

To limit the risks associated with surgery, it is imperative to periodically assess outcomes and see if further refinement in technique can be performed. For the initial series involving level IV tumour thrombus, CPBP frequently used DHCA. DHCA was recommended, as it was believed to prevent tumour thrombus embolization and thought to limit hepatic and renal ischaemia. However, as thrombi were noted to usually be adherent only at the renal vein, extraction could often be performed without a bloodless field. Thus, some groups opted to avoid DHCA and the inherent surgical risks.

Direct comparisons of the operative characteristics of RCC with level IV tumour thrombus have been limited in part due to the fortunate rarity of these cases. This series is the first to evaluate operative characteristics and outcomes with and without DHCA for level IV tumour thrombus. While generally believed to increase operative durations by cooling and warming, in the present study there were no significant differences between techniques. Additionally while concern for coagulopathy exists with hypothermia, in the present study there were no differences in EBL or rate of transfusions for cases involving DHCA. For surviving patients, hospital stay was also similar between groups. The most significant finding of the present study was the significant difference in associated perioperative mortality based on usage of DHCA. Perioperative mortality decreased from 37.5% to 8.3% for patients in which DHCA was utilized. Even when controlling for other risk factors associated with perioperative mortality including age, ECOG PS, tumour size, or stage, those that did not have DHCA performed had a seven-fold increased risk of perioperative mortality.

Use of DHCA has been extensively studied in aortic surgery, as there is concern for neurological sequelae with limited cerebral perfusion. In the setting of nephrectomy and thrombectomy, concern for neurological perfusion is not a major concern. Therefore, a major benefit of this technique may not be beneficial for this type of surgery. However, there was a significant decrease in perioperative mortality in the present study with this technique. A clear explanation why performing circulatory arrest would limit the perioperative mortality does not exist but one theory involves the use of systemic hypothermia. The deep hypothermia in conjunction with circulatory arrest is thought to limit renal and hepatic ischaemia in aortic aneurysm repair [15]. The need for dialysis from acute renal failure in cardiac surgery with CPBP is only 2% but it is associated with nearly 50% mortality [16]. Perhaps the risk of renal failure during cases involving RCC with level IV tumour thrombus is significantly higher with the removal of one kidney and use of CPBP. If hypothermia could protect the contralateral kidney and limit the need for postoperative dialysis, this could potentially reduce perioperative mortality.

If systemic hypothermia is the protective mechanism, further prospective studies can test its benefit by applying its usage to routine CPBP. Moderate systemic hypothermia with CPBP without circulatory arrest can be performed and has been shown to be safe in cases of level IV tumour thrombus [17]. This would potentially avoid known risks of circulatory arrest and allow the benefits of hypothermia. Perhaps this would even further decrease the associated perioperative mortality.

Several limitations must be addressed including the retrospective and multi-institutional nature of the series. Differences in surgeon volume, surgical approach, and patient selection probably influenced surgical morbidity and mortality. The detailed patient factors influencing the use of DHCA cannot be interpreted in a retrospective fashion and may influence the present conclusions. While ECOG PS was accounted for, we did not have any control for patient comorbdities, which probably had an influence on surgical mortality. A retrospective assessment of surgical complications was not performed due to non-uniform definition of complications between centres and the lack of detailed patient information for some centres. Additionally, information on the use of perioperative embolization was not available in the database. Most centres reported rarely using embolization; however, it could have influenced the operative characteristics as recent evidence suggests its usage may increase perioperative mortality [18]. A future prospective analysis is recommended to assess if DHCA or the use of hypothermia could limit complications and perioperative mortality.

In conclusion, surgery for RCC and level IV tumour thrombus is associated high perioperative mortality. Despite use of CPBP and surgery being performed in major tertiary referral centres, 22% of patients did not survive the hospital stay. The use of DHCA does not appear to adversely affect operative characteristics. Advanced age and lack of use of DHCA were independent predictors of worse mortality in the present cohort. Further prospective studies should confirm the benefit of DHCA to further reduce the associated risk with this complex surgery.


  1. Top of page
  2. Abstract