| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From the Section of Surgical Oncology, Department of Surgery, Robert C. Byrd Health Science Center, and Mary Babb Randolph Cancer Center of West Virginia University, Morgantown, West Virginia.
Correspondence: Address correspondence and reprint requests to: Stephen P. Povoski, MD, FACS, Division of Surgical Oncology, Department of Surgery, Arthur G. James Cancer Hospital and Richard J. Solove Research Institute, The Ohio State University, N-924 Doan Hall, 410 West 10th Ave., Columbus, OH 43210; Fax: 614-293-3465; E-mail: povoski-1@ medctr.osu.edu.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Methods: This study was a retrospective analysis of selective use of preoperative venous duplex ultrasound and intraoperative venography in 248 consecutive cancer patients undergoing central venous access placement.
Results: Ninety patients had a history of central venous access placement. Eleven had a history of deep venous thrombosis of an upper extremity or central vein. One hundred three underwent preoperative ultrasound. Previous central venous access placement was not associated with an abnormal preoperative ultrasound; however, previous central venous access with deep venous thrombosis was (P = .014). Thirty patients underwent intraoperative venography, of which 18 also had preoperative ultrasound. Fifty percent of patients with an abnormal intraoperative venogram had no abnormal findings on preoperative ultrasound.
Conclusions: Routine preoperative ultrasound is unnecessary. We advocate the selective use of preoperative ultrasound in those with a history of central venous access associated with deep venous thrombosis. We advocate the use of intraoperative venography when there is difficulty advancing the guidewire or catheter or when preoperative ultrasound is negative despite a history of central venous access with deep venous thrombosis.
Key Words: Central venous access • Ultrasound • Venography • Deep venous thrombosis
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
PATIENTS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Patients with one or more such preexisting factors underwent bilateral preoperative venous duplex US of the upper-extremity deep and superficial veins. Preoperative venous duplex US consisted of combined real-time b-mode US imaging and pulsed Doppler spectral analysis (duplex scanning) with color flow, using a Hewlett-Packard Sonos 5500® imaging system (Hewlett-Packard Corp., Andover, MA) with a Hewlett-Packard linear-array transducer probe (4.5 to 7.0 MHz). Veins imaged included the cephalic, basilic, axillary, subclavian, and internal jugular veins. Abnormal vascular findings on preoperative venous duplex US were classified as occlusion/thrombosis, significant focal stricture/nonocclusive thrombus, or unilateral small-caliber vein. A unilateral small caliber vein was defined as a vein that had a 
30% difference in caliber uniformly along its entire length, as compared with the contralateral vein.
IOV was used to map venous anatomy in all patients in whom there were intraoperative difficulties with CVAD placement. IOV was performed by using either a venous cutdown approach (cephalic vein or external jugular vein) or a percutaneous subclavian vein approach. Portable C-arm fluoroscopic imaging equipment available in the operating room at West Virginia University was used. The contrast medium used was either Hypaque®-76 (Nycomed, Inc., Princeton, NJ) or Omnipaque®-350 (Nycomed). Four types of CVADs were placed: BardPort® titanium-implanted single-lumen (9.6F or 6.6F) ports (Bard Access Systems, Salt Lake City, UT), Leonard® (10F) or Hickman® (12F) dual-lumen tunneled central venous catheters (Bard Access Systems), Arrow® 7F triple-lumen nontunneled catheters (Arrow International, Inc., Reading, PA), or Arrow® 12F triple-lumen nontunneled catheters (Arrow International, Inc.).
All patients were prospectively entered into a CVAD computerized database. The software program SPSS© for Windows (version 8.0; SPSS, Inc., Chicago, IL) was used for all statistical analyses. Pearson 
2 analysis was used for univariate comparison of all categorical variables. When the expected frequency of one or more cells of a given 2 x 2 table for univariate comparison was 
5, Fisher’s exact test was used. A P value of .05 was considered statistically significant.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Of those 248 consecutive cancer patients taken to the operating room for attempted CVAD placement, 90 (36%) had a history of CVAD placement, of whom 47 (52%) had a single CVAD placed in the past and 43 (48%) had multiple CVADs placed in the past. Of those 90 patients with a history of CVAD placement, 10 (11%) had a documented history of a CVAD-related infection. A history of DVT of an upper extremity or central vein was reported in 11 (12%) of 90 with a history of CVAD placement as compared with 0 (0%) of 158 with no history of CVAD placement (P < .001).
A CVAD was successfully placed in the same trip to the operating room in 247 (99.6%) of 248 patients. This included 123 implanted ports, 62 tunneled catheters, and 62 nontunneled catheters. The CVAD was placed by way of a venous cutdown approach in 151 (61%) patients and by way of a percutaneous approach in 96 (39%) patients.
Group of 103 Patients Undergoing Preoperative Venous Duplex US
Of the 248 patients taken to the operating room for attempted CVAD placement, 103 (42%) underwent preoperative venous duplex US. The major indication for performing preoperative venous duplex US was a history of CVAD placement in 90 (87%), a history of surgery of the thoracic cavity or neck region in 5 (5%), significant intrathoracic adenopathy in 3 (3%), significant intrathoracic adenopathy and a history of surgery of the thoracic cavity or neck region in 2 (2%), a history of clavicle fracture in 2 (2%), and dwarf stature with significant deformity of the upper extremities and torso in 1 (1%).
Of the 103 preoperative venous duplex US studies performed, 37 (36%) were reported to demonstrate 1 or more abnormal venous findings. Table 1 lists the major abnormal venous findings demonstrated on preoperative venous duplex US for all veins examined, as well as individually for deep veins (axillary, subclavian, and internal jugular vein) and for superficial veins (cephalic and basilic vein). In all three categories of veins, the major abnormal venous finding of preoperative venous duplex US was occlusion or thrombosis.
|
|
Group of 30 Patients Undergoing IOV
Of those 248 patients taken to the operating room for attempted CVAD placement, 30 (12%) underwent IOV. The main indication for IOV was inability to advance the catheter by way of the cephalic vein cutdown approach in 20 (67%), inability to advance the catheter by way of the external jugular vein cutdown approach in 3 (10%), inability to advance the guidewire by way of a percutaneous subclavian vein approach (infraclavicular) in 5 (17%), inability to advance the catheter by way of a percutaneous subclavian vein approach (infraclavicular) once the guidewire had been passed in 1 (3%), and verification of the supraclavicular percutaneous subclavian vein approach after inability to cannulate the subclavian vein by way of the infraclavicular approach in 1 (3%). The venous route used for performing IOV was the cephalic vein cutdown approach in 20 (67%), the percutaneous subclavian vein approach in 7 (33%), and the external jugular vein cutdown approach in 3 (10%). Of the 30 IOVs performed, an abnormal venous finding was demonstrated in 29 (97%). The major abnormal venous finding seen on IOV was the presence of a difficult angle to transverse between the junction of the cephalic vein and subclavian vein in 11 (38%), a significant focal stricture of the subclavian vein or innominate vein in 9 (31%), a difficult angle to transverse between the junction of the subclavian vein and innominate vein in 3 (10%), placement of the catheter or guidewire into a prominent azygos vein in 3 (10%), complete occlusion of the subclavian-innominate vein junction in 1 (3%), compression of the catheter at the junction of the first rib and clavicle in 1 (3%), and a small-caliber subclavian vein in 1 (3%). It is interesting to note that the patient with complete occlusion of the subclavian-innominate vein junction had metastatic esophageal cancer with no history of CVAD placement or significant intrathoracic adenopathy.
A CVAD was successfully placed in the same trip to the operating room in all 30 patients (100%) who underwent IOV. The final site for CVAD placement was altered in 29 (100%) of 29 cases when IOV revealed one or more abnormal venous findings as compared with 0 (0%) of 1 case when IOV findings were normal (P = .033).
Group of 18 Patients Undergoing Both Preoperative Venous Duplex US and IOV
Of those 30 patients undergoing IOV, 18 (60%) had also undergone preoperative venous duplex US. The major abnormal venous finding seen on IOV in those 18 cases included a significant focal stricture of the subclavian vein or innominate vein in 7 (39%), the presence of a difficult angle to transverse between the junction of the cephalic vein and subclavian vein in 6 (33%), placement of the catheter or guidewire in a prominent azygos vein in 2 (11%), a difficult angle to transverse between the junction of the subclavian vein and innominate vein in 1 (6%), compression of the catheter at the junction of the first rib and clavicle in 1 (6%), and a small-caliber subclavian vein in 1 (6%). Of those 18 patients, 9 (50%) with an abnormal venous finding on IOV had no abnormal venous findings on preoperative venous duplex US. The major abnormal venous finding seen on IOV in those 9 cases in which no abnormal venous finding was demonstrated on preoperative venous duplex US included significant focal stricture of the subclavian vein or innominate vein in 3 (33%), the presence of a difficult angle to transverse between the junction of the cephalic vein and subclavian vein in 2 (22%), placement of the catheter or guidewire in a prominent azygos vein in 2 (22%), a difficult angle to transverse between the junction of the subclavian vein and innominate vein in 1 (11%), and compression of the catheter at the junction of the first rib and clavicle in 1 (11%).
Group of 12 Patients Undergoing IOV Without Preoperative Venous Duplex US
Of those 30 patients undergoing IOV, 12 (40%) did not undergo preoperative venous duplex US. Of those 12 patients, a major abnormal venous finding was seen on IOV in 11 (92%). The major abnormal venous finding seen on IOV included the presence of a difficult angle to transverse between the junction of the cephalic vein and subclavian vein in 5 (46%), significant focal stricture of the subclavian vein or innominate vein in 2 (18%), a difficult angle to transverse between the junction of the subclavian vein and innominate vein in 2 (18%), placement of the catheter or guidewire in a prominent azygos vein in 1 (9%), and complete occlusion of the subclavian-innominate vein junction in 1 (9%).
Group of 11 Patients With Documented History of DVT of an Upper Extremity or Central Vein
Of those 11 patients with a documented history of a DVT of an upper extremity or central vein, 8 (73%) had 1 or more abnormal venous findings on preoperative venous duplex US. This included occlusion or thrombosis of a deep vein in 5, unilateral small-caliber deep vein in 1, significant focal stricture or nonocclusive thrombus of a deep vein in 1, and occlusion or thrombosis of a superficial vein in 1. Of those 11 patients with a documented history of a DVT of an upper extremity or central vein, 3 (27%) had no abnormal venous finding reported on preoperative venous duplex US.
Of those 11 patients with a documented history of a DVT of an upper extremity or central vein, only 2 (18%) required IOV for successful CVAD placement. These two patients had also undergone preoperative venous duplex US, with demonstration of an abnormal venous finding on both preoperative venous duplex US and IOV. In these two cases, the abnormal venous finding demonstrated on IOV was located on the contralateral side of the abnormal venous finding demonstrated on preoperative venous duplex US. Those three patients with a documented history of a DVT of an upper extremity or central vein who had no abnormal venous finding reported on preoperative venous duplex US did not require an IOV for successful CVAD placement on the side contralateral to that of the location of the previous DVT of an upper extremity or central vein.
The final site for CVAD placement was altered in 6 (55%) of 11 cases when there was a documented history of a DVT of an upper extremity or central vein, as compared with 40 (17%) of 237 cases when there was no documented history of a DVT of an upper extremity or central vein (P = .007).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
In this study and as one would expect, a history of CVAD placement was clearly associated with a documented history of a DVT of an upper extremity or central vein. Likewise, the final site for CVAD placement was altered significantly by a documented history of a DVT of an upper extremity or central vein. With regard to preoperative venous duplex US, our analysis has demonstrated that a history of CVAD placement itself was not associated with the finding of an abnormal venous finding on preoperative venous duplex US; however, subgroup analysis revealed that a history of CVAD placement with a documented history of a DVT was associated with the finding of an abnormal venous finding on preoperative venous duplex US. From these results, preoperative venous duplex US does not seem to be routinely indicated in cancer patients with a history of CVAD placement alone. It is our opinion that these findings support more selective use of preoperative venous duplex US in only those cancer patients with a history of CVAD placement in which there is a documented history of a DVT. With regard to IOV, this study shows that the selective use of IOV commonly identifies abnormalities in central venous anatomy, which could otherwise preclude successful CVAD placement. Likewise, 50% of the patients in this study who had an abnormal venous finding on IOV had no abnormal venous findings on preoperative venous duplex US. It is well documented that preoperative venous duplex US is a highly accurate imaging modality for the assessment of the internal jugular, brachial, axillary, and lateral segment of the subclavian vein. However, it is not ideal for the assessment of the medial segment of the subclavian vein, innominate vein, and superior vena cava7–12 or for evaluation of the difficult angle between the cephalic vein and the subclavian vein. It is our opinion that the high discordance between preoperative venous duplex US and IOV was a result of the inability of preoperative venous duplex US to detect venous abnormalities located more centrally along the medial segment of the subclavian vein and innominate vein and within the superior vena cava. Such more centrally located venous abnormalities are well known to be accurately detected by IOV,6,7,9 as is indicated in this study.
Two studies in the literature support the routine use of preoperative venous duplex US in cancer patients with a history of a CVAD placement to evaluate venous anatomy before placement of a new CVAD.2,3 Both studies warrant discussion.
Haire et al.2 preoperatively evaluated 22 patients with a history of subclavian vein CVAD by using preoperative venous duplex US in a blinded fashion with subsequent CVAD placement, without knowledge of the preoperative venous duplex US results. A normal preoperative venous duplex US was seen in 19 patients. Of these, 18 patients underwent successful CVAD placement. The one patient with unsuccessful CVAD placement underwent IOV, which showed significant focal stricture of the subclavian vein. Of the 22 patients with a history of subclavian vein CVAD placement who underwent a preoperative venous duplex US, 3 (14%) had a preoperative venous duplex US that demonstrated evidence of subclavian vein occlusion. CVAD placement was unsuccessful in all three cases. The authors concluded that a normal preoperative venous duplex US finding in a patient with placement of a subclavian vein CVAD predicts successful CVAD placement in 95% (18 of 19) of cases. In contrast, our study demonstrated that only 86% (57 of 66) of the patients with a normal preoperative venous duplex US underwent uneventful CVAD placement, whereas 14% (9 of 66) of the patients required additional venous imaging with IOV to further evaluate the venous anatomy before successful CVAD placement. This is supported by our previous analysis, which demonstrated that a history of CVAD placement alone was not associated with the finding of an abnormal venous finding on preoperative venous duplex US.
Because 50% of the patients in our study with an abnormal venous finding on IOV had no abnormal venous findings on preoperative venous duplex US, the utility of preoperative venous duplex US remains unclear. Kraybill and Allen3 preoperatively evaluated 32 patients with preoperative venous duplex US, including 27 patients with a history of CVAD placement. Of those 27 patients, 12 (44%) preoperative venous duplex US studies demonstrated thrombosis of the internal jugular, subclavian, axillary, or brachial veins. The authors concluded that preoperative venous duplex US (PVDU) was effective in directing the surgeon to nonthrombosed central veins for successful venous access and, therefore, that patients with a history of CVAD should have PVDU to identify central venous thrombosis and direct the surgeon to the most appropriate site for venous access. In contrast, our study demonstrated that only 22% (20 of 90) of patients with a history of CVAD placement were found to have significant focal venous stricture/nonocclusive thrombus or venous occlusion/thrombosis by PVDU. However, if we performed the same subset comparison for only those patients with a history of CVAD placement in whom there was a documented history of a DVT of an upper extremity or central vein, then 64% (7 of 11) were found to have significant focal venous stricture/nonocclusive thrombus or venous occlusion/thrombosis by preoperative venous duplex US (P = .005). This specific subset analysis of our data clearly recognizes the contribution of a documented history of a DVT and supports more selective use of preoperative venous duplex US in only those patients with a history of CVAD placement in whom there is a documented history of a DVT, rather than routine use in all patients with a history of CVAD placement, as suggested by Kraybill and Allen.3
One could postulate that solely a history of CVAD placement in which there is a documented history of a DVT of an upper extremity or central vein would be adequate information to potentially eliminate the need for performing preoperative venous duplex US. In such situations, use of the contralateral side for CVAD placement may be sufficient without the need for further venous imaging. This is supported by the fact that three patients in our study with a documented history of a DVT of an upper extremity or central vein who had no abnormal venous finding reported on preoperative venous duplex US required no further venous imaging for successful CVAD placement on the contralateral side. Likewise, such a history might also warrant consideration of preoperative venography to rule out unexpected central venous abnormalities that may not be seen on preoperative venous duplex US. However, we do not necessarily advocate this approach, nor have we attempted to evaluate it.
In this study, IOV was performed by way of a cephalic or external jugular vein cutdown approach in 23 (77%) of 30 patients, as opposed to a percutaneous subclavian vein approach in 7 (23%) of 30. In these difficult cases in which abnormalities of central venous anatomy are suspected, the process of simply attempting to percutaneously cannulate a deep vein by the venipuncture needle for IOV can be as difficult as is the process of attempting to advance the guidewire or catheter for eventual CVAD placement. Therefore, we generally favor the venous cutdown approach to CVAD placement for all implanted or tunneled devices13 because the venous cutdown approach offers a simplified access route for possible IOV.
It is well recognized that subclavian vein stenosis is a significant problem in patients with end-stage renal failure who have a history of placement of a temporary subclavian vein hemodialysis catheter.12,14–16 The significant risk of subclavian vein stenosis in these patients may be related to the physical properties of the temporary subclavian vein hemodialysis catheters. Therefore, it is the general recommendation that preoperative venous duplex US should be performed for hemodialysis access planning on all end-stage renal failure patients with a history of placement of a temporary subclavian vein hemodialysis catheter.12,14–16 In this article, we have not attempted to address this particular issue in end-stage renal failure patients and have strictly limited our discussion and recommendation to CVAD placement in cancer patients.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
Acknowledgments |
|---|
Received for publication October 30, 2001. Accepted for publication March 5, 2002.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  M. P. e Albuquerque A Modified Internal Jugular Vein Access for Long-Term Catheter Placement in Cancer Patients Ann. Surg. Oncol., February 1, 2007; 14(2): 937 - 941. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
