PATIENTS AND METHODS
- Top of page
- PATIENTS AND METHODS
- CONFLICT OF INTEREST
After institutional review board approval, patients were identified who had undergone radical nephrectomy with IVC tumour thrombectomy at our institution between May 1997 and February 2011. Initial diagnosis was made by a contrast-enhanced four-phase CT scan, with subsequent MRI for tumour thrombus characterization. The cranial extent of the tumour was initially defined as per the Neves & Zincke  system, and level III thrombi were subclassified using our modified definition .
We have previously described the details of our surgical technique for the management of RCC with tumour thrombus [5,6]. Tumour or bland IVC thrombus without total venous occlusion was initially managed via a bloodless cavatomy with thrombectomy. However, the presence of iliac vein extension of a bland thrombus, or residual thrombus after an attempted bland thrombectomy, was managed with intraoperative placement of an IVC filter through the cavatomy. Surgical interruption of the IVC was performed in the presence of a bland thrombus totally occluding the IVC and not amenable to thrombectomy or safe deployment of a filter. Intraoperative trans-oesophageal echocardiography (TEE) was used selectively in patients with thrombus level ≥ III to accurately determine the cranial extent of the thrombus.
Circumferential interruption of the IVC was preformed by either stapling, suture ligation or oversewing. This was extended to a segmental resection of the IVC or en bloc IVC resection with the involved kidney when we could not identify an interface between the tumour and bland thrombus in the face of total IVC occlusion. In the event of IVC interruption, collateral venous flow was maintained by minimizing the ligation of venous tributaries. In the early phase of this series, patients with segmental resection of the IVC were reconstructed with a vascular PTFE graft.
Preoperative anticoagulation was prescribed in cases of a bland thrombus or a thromboembolic event. All patients received lower extremity sequential compressive devices during surgery and until after the patient was fully ambulating. All patients were started on 325 mg of aspirin 48 h after the surgery unless contraindicated.
Statistical analysis was performed using PASW version 18.0 (IBM Corporation, Somners, NY, USA). Groups were compared using Mann–Whitney or Pearson chi-squared test. Univariate and multivariate logistic regression analyses were performed to identify significant predictors of bland thrombus coexistence, as well as the need for surgical interruption of the IVC. P < 0.05 was considered to indicate statistical significance. All P values were two-sided.
- Top of page
- PATIENTS AND METHODS
- CONFLICT OF INTEREST
Our cohort included 129 patients. Patient demographics are outlined in Table 1. Three patients (2%) with level IV thrombus underwent cardiopulmonary bypass (CPB) for removal of tumour thrombus from the right atrium. The perioperative mortality rate was 2.3%. Although no intraoperative deaths were noted, one of the deaths occurred on the first postoperative day in the level III group as a result of refractory cardiac arrhythmia. The second death occurred at 1 week as a result of sepsis and respiratory failure in a patient with chronic obstructive pulmonary disease, and the third death occurred at 2 weeks as a result of hepatic failure in a patient with level IV thrombus and Budd–Chiari syndrome after CPB.
Table 1. Patient demographics and analysis of bland thrombus and IVC interruption. Data are n, n(%) or mean (range), unless noted otherwise
|Study variable||All patients||No bland thrombus||Bland thrombus present|| P value||No interruption of IVC||Interruption of IVC|| P value|
|Patients||129||114 (88)||15 (12)||–||110 (85)||19 (15)||–|
|Age*, years||63 (25–84)||61.7 ± 11.4||56.22 ± 9.7||0.175||62.2 ± 11.2||54.87 ± 10.2||0.03|
|Tumour size*, mL|| ||10.36 ± 4.4||12.5 ± 4.2||0.465||10.3 ± 4.3||12.02 ± 5||0.171|
|EBL† (cc)||500 (200–16 000)||1328 ||2305 ||0.007||1176 ||2896 ||0.03|
|Metastasis at presentation||9 (7)||9 (100)||0 (0)|| ||9 (100)||0 (0)|| |
|Sex|| || || ||0.785|| || ||0.083|
| Male||70 (54)||61 (54)||9 (60)|| ||56 (51)||14 (74)|| |
| Female||59 (46)||53 (46)||6 (40)|| ||54 (49)||5 (26)|| |
|Tumour location|| || || ||0.781|| || ||0.068|
| Right||82 (64)||73 (64)||9 (60)|| ||66 (60)||16 (84)|| |
| Left||47 (36)||41 (36)||6 (40)|| ||44 (40)||3 (16)|| |
|Level of thrombus|| || || ||0.068|| || ||0.003|
| I||38 (29)||38 (33)||0 (0)|| ||38 (35)||0 (0)|| |
| II||17 (13)||14 (12)||3 (20)|| ||16 (14)||1 (5)|| |
| III||62 (48)||52 (46)||10 (66)|| ||48 (44)||14 (74)|| |
| III a (retrohepatic)||25 (19)||21 (41)||4 (40)|| ||20 (42)||5 (36)|| |
| III b (hepatic veins)||21 (16)||16 (31)||5 (50)|| ||14 (28)||7 (50)|| |
| III c (suprahepatic)||9 (7)||8 (15)||1 (10)|| ||7 (15)||2 (14)|| |
| III d (infradiaphragmatic)||7 (5)||7 (13)||0 (0)|| ||7 (15)||0 (0)|| |
| IV||12 (9)||10 (9)||2 (14)|| ||8 (7)||4 (21)|| |
|Histological type|| || || ||0.041|| || ||0.957|
| Clear cell||98 (76)||89||9|| ||82||16|| |
| Clear cell granular||8 (6)||7||1|| ||7||1|| |
| Papillary||9 (7)||6||3|| ||8||1|| |
| Chromophobe||1 (1)||0||1|| ||1||0|| |
| Sarcomatoid||5 (4)||5||0|| ||5||0|| |
| Poorly differentiated||7 (5)||6||1|| ||6||1|| |
| Collecting duct||1 (1)||1||0|| ||1||0|| |
|Nuclear grade|| || || ||0.826|| || ||0.114|
| I||3 (2)||3 (3)||0 (0)|| ||3 (3)||0 (0)|| |
| II||25 (19)||23 (20)||2 (13)|| ||24 (22)||1 (5)|| |
| III||60 (47)||52 (45)||8 (53)|| ||52 (47)||8 (42)|| |
| IV||41 (32)||36 (32)||5 (34)|| ||31 (28)||10 (53)|| |
|Bland thrombus|| || || || || || ||<0.001|
| No||114 (88)||–||–||–||103 (94)||11 (58)|| |
| Yes||15 (12)|| || || ||7 (6)||8 (42)|| |
| Intraoperative||16 (12)||12||4||0.093||9||7||0.003|
| Postoperative||13 (10)||11||2||0.648||8||5||0.024|
Pulmonary embolus at the time of referral was identified in four patients (3%). Preoperative anticoagulation was necessary in two of these patients, and the remaining two patients underwent CPB for removal of major tumour emboli from the pulmonary artery. After surgery, two patients developed a pulmonary embolus (PE). Clinical characteristics of patients with pre- and postoperative PE are outlined in Table 2. One patient with a level IIIa thrombus had a prophylactic IVC filter placed before surgery at another institution and was referred to us for surgical intervention. He underwent segmental resection of the IVC with removal of the tumour thrombus and the IVC filter. None of the remaining 128 patients had an IVC filter placed preoperatively.
Table 2. Clinical characteristics of patients with pre and post operative pulmonary embolus
|Patient no||Age, years||Sex||Primary||Level of thrombus||EBL, mL||Tumour size, cm||Grade||Pathology||Metastasis at presentation||PE||Bland thrombus||Anticoagulation||Surgical intervention||CPB||Resection of IVC|
|1||67||Male||Right||IIIB||2100||9||4||clear cell||No||Preoperative PE||No||Preoperative coumadin||RRN, cavatomy and thrombectomy, CPB and removal of tumour thrombus from pulmonary arteries||Yes||No|
|2||59||Male||Right||IIIA||250||9||4||clear cell||No||Preoperative PE||No||Preoperative coumadin||RRN, cavatomy and thrombectomy, CPB and removal of tumour thrombus from pulmonary arteries||Yes||No|
|3||73||Male||Left||I||2000||5||3||clear cell||No||Preoperative PE||No||Preoperative coumadin||LRN, cavatomy and thrombectomy||No||No|
|4||50||Male||Right||IIIA||4000||9||3||clear cell||No||Preoperative PE||Yes||Preoperative heparin||RRN, resection of IVC with stapling||No||Yes|
|5||38||Male||Left||IIIA||1000||14||4||Clear cell||No||Preoperative IVC filter||Yes||None||LRN, segmental resection of IVC with thrombus incorporated IVC filter and oversewn||No||Yes|
|6||64||Female||Right||II||6000||11||4||Clear cell||No||Postoperative PE||No||Postoperative coumadin||RRN, cavatomy and thrombectomy||No||No|
|7||60||Male||Right||II||200||10||3||Papillary||No||Postoperative PE||No||Postoperative coumadin||RRN, cavatomy and thrombectomy||No||No|
Of the 129 patients in the cohort, 19 (15%) underwent surgical IVC interruption. Table 3 shows the type of IVC interruptions according to level of thrombus in the entire cohort and in patients with associated bland thrombus. A total of 114 patients (88%) in our cohort had tumour thrombus without bland thrombus; 103 of them (90%) underwent complete removal of the tumour thrombus and primary repair of the IVC. Eleven (9.6%) patients had tumour thrombus infiltrating the wall of the IVC and underwent segmental resection of the IVC. None of these patients had a perioperative IVC filter placed.
Table 3. Perioperative complications
| || N |
| Liver injury and repair||1|
| Splenic injury and repair||5|
| Diaphragm injury and repair||2|
| Hepatic vein injury and repair||2|
| Opposite renal vein injury and repair||2|
| Thrombus incorporated into IVC filter||1|
| Prolonged intubation (I)||3|
| Tracheostomy (II)||1|
| Refractory cardiac arrhythmia (V)||1|
| Acute tubular necrosis (I)||1|
| Abdominal wall haematoma (IIIa)||1|
| Post-operative pulmonary Embolism (IIIa)||2|
| Abdominal collection (IIIb)||1|
| Small bowel obstruction (I)||1|
| Liver failure (V)||1|
| Sepsis (V)||1|
In all, 15 patients (12%) had bland thrombus associated with the tumour thrombus. A bland thrombus was identified before surgery in two (13%) patients with a level IIIb tumour thrombus. On MRI, the thrombus extended distally to the bifurcation of the IVC on both patients, and into the contralateral renal vein in one patient. The level of thrombus involvement and the type of intervention are given in Table 4. Three had a small bland thrombus distal to the tumour thrombus, which was removed during surgery. Postoperatively heparin anticoagulation was used to maintain their partial thromboplastin time between 50 and 60 s. Residual bland thrombus was present distally after removal of tumour thrombus in four patients. These patients received an intraoperative IVC filter below the opposite renal vein through the cavotomy. Eight patients had bland and tumour thrombus organized and adherent to the wall of the IVC where complete bland thrombectomy was not feasible. Of these, five patients had the IVC interrupted by ligation below the tumour thrombus. The remaining three patients underwent segmental resection of the IVC using vascular stapling due to extensive tumour thrombus infiltration. None of the 15 patients had a clinical bland thrombus PE during or after surgery.
Table 4. Type of surgical intervention according to the level of thrombus
|Level of tumour thrombusSurgical approach||Level I||Level II||Level III||Level IV||Total|
|All cases||Bland thrombus||All cases||Bland thrombus||All cases||Bland thrombus||All cases||Bland thrombus||All cases||Bland thrombus|
|Classic cavatomy and thrombectomy||38||0||14||0||43||0||8||0||103||0|
|Thrombectomy and anticoagulation||0||0||1||1||2||2||0||0||3||3|
|Tumour thrombectomy and IVC filter||0||0||1||1||3||3||0||0||4||4|
|Ligation of IVC||0||0||0||0||4||4||1||1||5||5|
|Resection of IVC||0||0||1||1||10||1||3||1||14||3|
Surgical interruption of the IVC was performed in eight (53%) patients with bland thrombus and 11 (10%) patients without bland thrombus. Surgical interruption of the IVC was more frequent in patients with an associated bland thrombus (P < 0.001). We further analysed the data to identify the risk factors for bland thrombus and IVC interruption. Patient age (P= 0.202) and sex (P= 0.636), as well as tumour size (P= 0.377), side (P= 0.760) and nuclear grade (P= 0.515) were not predictors of coexisting bland thrombus. Advanced level of tumour thrombus was the only significant factor predicting association of bland thrombus (OR = 2.09, 95% CI: 1.082–4.037, P= 0.028). Univariate analysis identified younger age (OR = 1.06, 95% CI: 1.019–1.115, P= 0.005), level of tumour thrombus (OR −3.54, 95% CI: 1.66–7.56, P= 0.001) and associated bland thrombus (OR = 10.7, 95% CI: 3.25–35.16, P < 0.001) as significant predictors of IVC interruption. On multivariate analysis, level of thrombus (OR = 3.1, 95% CI: 1.30–7.74, P= 0.011) and association of bland thrombus (OR = 9.07, 95% CI: 2.42–34.01, P= 0.001) were significant for IVC interruption.
There were 29 (22%) perioperative complications recorded. Of these, 16 (12%) were intraoperative and 13 (10%) were postoperative complications (Table 3). Coexistence of bland thrombus did not increase the incidence of complications (P= 0.102). Patients who underwent surgical interruption of the IVC had a significantly higher incidence of intraoperative (P= 0.003) as well as post operative (P= 0.024) complications (Table 1).
- Top of page
- PATIENTS AND METHODS
- CONFLICT OF INTEREST
It is not uncommon for bland thrombus to be associated with tumour thrombus in cases of RCC (Fig. 1) . The surgical significance of coexisting bland thrombus is not well reported. In our experience, small distal bland thrombus is usually amenable to removal without difficulty (Fig. 2). However, it is not routinely recommended because of the risk of venous disruption and incomplete clearance . Three patients in the early phase of our series underwent successful bland as well as tumour thrombectomy and were started on postoperative anticoagulation. Nowadays we prefer to place an intraoperative IVC filter in patients with a distal bland thrombus irrespective of successful bland thrombectomy. This can be done through the cavatomy or through a separate needle puncture after closure of the cavatomy. The added advantage of this approach is the ability to avoid long-term anticoagulation.
Figure 1. Right upper pole kidney tumour with associated level II tumour thrombus partially occluding the IVC. The inset shows tumour thrombus in place and associated bland thrombus in proximal and distal positions.
Download figure to PowerPoint
Several case reports have shown that the preoperative placement of an IVC filter at the proximal end of the tumour thrombus successfully prevents embolization [8,9]. The most important factor concerning the surgical team is the possibility of the filter being incorporated into the thrombus [7,10]. Preoperative IVC filter placement could play a role in patients with level I or II IVC tumour thrombus associated with distal bland thrombus and in cases when anticoagulation is not feasible. In our experience with cases of level III thrombus, the placement of a filter should be avoided due to the risk of thrombus incorporation into the filter.
Extensive infrarenal bland thrombus partly or completely occluding the IVC is usually not amenable to thrombectomy. We have found that the bland thrombus often extends distally below the iliac bifurcation. In these cases we aim for a negative margin by identifying the distal margin of the tumour thrombus and then proceeding to ligate and divide the IVC. Moreover, we perform a segmental resection of the IVC when the tumour thrombus is adherent to the vessel wall of if there is no identifiable interface between tumour and bland thrombus (Fig. 3).
Figure 3. (A) Mobilization of the liver and dissection of the IVC; (B) en bloc resection specimen which includes a segment of the IVC; (C) removed segment of IVC showing tumour and bland thrombi.
Download figure to PowerPoint
The findings in the present series clearly indicate that the presence of bland thrombus in association with the tumour thrombus should alert the surgical team to a possibly complex and challenging surgical encounter. We failed to observe an instance of bland thrombus in any patient with a level I tumour thrombus. This reiterates that the tumour thrombus in the IVC must cause significant flow changes that will evolve the bland thrombus. In all, 15–20% of patients with a level II–IV tumour thrombus had a coexisting bland thrombus. One of two (50%) patients with associated bland thrombus needed surgical interruption of the IVC; 27% of patients with bland thrombus received an intraoperative IVC filter; 80% of patients with bland thrombus needed at least one additional surgical manoeuvre.
Blute et al.  from the Mayo Clinic report the type of surgical intervention utilized in 160 cases with level II–IV tumour thrombus. IVC interruption was performed in 25% of patients. These included placement of a Greenfield IVC filter in 2.5% of cases, IVC ligation in 14.4% and segmental resection in 8%. These authors devised a preoperative classification based on the imaging that could predict the need for IVC interruption. The two most important factors that decided the need for IVC interruption intraoperatively were presence of bland thrombus and the degree of venous occlusion. The present series differs slightly from this work in the following aspects. In the Mayo Clinic series, all the patients who had IVC ligation or resection had bland thrombus in association with tumour thrombus . In the present series, eight of 19 patients (42%) who had ligation or resection of IVC had associated bland thrombus. Eleven patients had tumour thrombus adherent to the wall of the IVC necessitating IVC interruption. Younger age, coexisting bland thrombus and level III and IV tumour thrombus were the predictors of IVC interruption. The reason for more younger patients undergoing IVC resection is not clear.
The degree of venous obstruction is an important factor that determines the feasibility of surgical interruption of the IVC. Clinically, venous obstruction can be evident by lower extremity oedema and the development of dilated abdominal wall venous collaterals. Radiographic evaluation is more precise and could provide evidence of peripheral network venous collaterals such as dilated azygos-hemiazygos and lumbar veins . Patients with chronic obstruction of the IVC have less haemodynamic instability than those with acute obstruction. The major concerns of interrupting the IVC are morbidity associated with venous congestion and lymphatic extravasation causing third space retention. Generally, with chronic IVC occlusion these complications are manageable . Blute et al.  recommended avoiding ligation of competitive collaterals that have developed with time, particularly on the contralateral side of the great vessels. Minimizing the segment of resection, preserving the contralateral renal vein and preserving the collateral channels are the key factors to minimize morbidity.
From our experience, patients with symptoms of a preoperative pulmonary embolism should have a radiographic evaluation of the pulmonary arteries followed by evaluation by a cardiothoracic team for the possibility of CPB for removal of tumour fragments from the pulmonary arteries. In the present series, two patients with level III thrombus with a preoperative diagnosis of PE underwent CPB for removal of the tumour thrombus from the pulmonary arteries.
In expert hands, intraoperative TEE provides vital information about the extent of the tumour thrombus. In many instances, authors have witnessed a change in the level of tumour thrombus during intraoperative TEE after ligating the renal artery. This finding allows the surgeon to avoid a CPB in cases where the change in levels allows the surgical team to gain control above the proximal end of the tumour thrombus. Hence, we recommend that TEE should be available in all cases of vena caval tumour thrombus and should be performed in the presence of any haemodynamic changes, or cases of complicated level II or higher level thrombi, preferably immediately before and during the surgical intervention .
The present study is retrospective in nature and not without limitations. First, the number of patients in this series is small. Secondly, since the present series spans over two decades, the surgical techniques and the management protocols were constantly changing, which prevents any robust statistical analysis pertaining to a single surgical technique. Large-sample, multicentre pooled analysis is necessary to validate the predictors of bland thrombus and IVC resection.
In conclusion, surgical interruption of IVC is a feasible option in selected patients with chronic obstruction of the IVC secondary to tumour thrombus. One in five patients with level II–IV tumour thrombus might require IVC interruption. Association of bland thrombus with the tumour thrombus should alert the surgical team to a potentially challenging surgery. About half of the patients with bland thrombus required IVC ligation or segmental resection. Precise preoperative imaging to assess the degree of venous obstruction and to differentiate bland from tumour thrombus is a key step in achieving the surgical goal with minimal morbidity.