- Email: [email protected]
Incidence and Management of Seroma after Arteriovenous Graft Placement
Incidence and Management of Seroma after Arteriovenous Graft Placement Danielle M Dauria, MD, Pawel Dyk, BA, Paul Garvin, MD, FACS Perigraft seromas are rare complications of insertion of PTFE hemodialysis grafts. They are often difficult to treat and recurrence is common. This study evaluates the incidence, potential etiologic variables, and management strategies for seromas after prosthetic arteriovenous graft (AVG) placement. STUDY DESIGN: A retrospective analysis of all patients undergoing AVG placement between August 2002 and December 2005 was performed to identify all patients diagnosed with seroma requiring surgical intervention. Multiple variables were analyzed to determine potential risk factors for seroma formation and outcomes of various forms of surgical management. RESULTS: In this interval, 535 AVG were inserted in 427 patients. Ten patients presented with a seroma and underwent surgical treatment. Overall incidence of seroma formation was 1.7%. There was no significant difference in seroma formation based on gender, age, diabetes, lower extremity versus upper extremity placement, or loop forearm versus straight forearm grafts. A statistically significant difference was found between upper arm (p ⫽ 0.007) and lower arm grafts (p ⫽ 0.04), with upper arm grafts more prone to seroma formation. Patients undergoing bypass of the seromatous segment of graft have not had a recurrence, compared with those who were simply evacuated and have had a mean patency of 402 days. CONCLUSIONS: Seroma complications after AVG insertion are higher in patients with upper arm grafts. To minimize this complication, meticulous operative technique is required. If a seroma develops, the graft might still be salvageable with aggressive management, including bypass of the involved segment. (J Am Coll Surg 2006;203:506–511. © 2006 by the American College of Surgeons) BACKGROUND:
There are approximately 300,000 people in the United States currently receiving longterm hemodialysis for end-stage renal disease. This number is expected to double by the year 2010. Despite the documented advantages of primary arteriovenous fistulas, approximately 60% of these patients still depend on prosthetic grafts for permanent access. In patients with end-stage renal disease, ⬎25% of all hospitalizations are vascular-access–related complications, with an eightfold higher mortality in patients with prosthetic grafts versus fistulas. One of these complications is seroma formation. A seroma is a collection of sterile, clear, ultrafiltered serum, surrounded by a nonsecretory fibrous soft tis-
sue pseudocapsule. Seromas, if untreated, can lead to wound infection, dehiscence, skin necrosis, graft thrombosis, or loss of the available graft puncture area. The incidence of seroma after arteriovenous graft placement for hemodialysis has been poorly defined in the current literature and has ranged from approximately 2% to 4%. It is poorly defined because some seromas resolve without treatment and are not reported, and others are grouped into hematomas and low-grade infection problems. There have been many different treatments proposed, but recurrence has remained a difficult problem. The goal of this study was to determine the incidence and outcomes of surgical management of seroma formation after arteriovenous graft insertion and to identify potential etiologic variables that can contribute to this complication.
Competing Interests Declared: None.
METHODS This study is a retrospective review of all patients who received an arteriovenous graft in a single large urban academic center between August 2002 and December
Received March 30, 2006; Revised May 27, 2006; Accepted June 7, 2006. From the Department of Surgery, St Louis University Hospital, St Louis, MO. Correspondence address: Danielle M Dauria, MD, Department of Surgery, St Louis University Hospital, 3635 Vista Ave, DesLoge Towers, 3rd Fl, St Louis, MO 63110. email: [email protected]
© 2006 by the American College of Surgeons Published by Elsevier Inc.
506
ISSN 1072-7515/06/$32.00 doi:10.1016/j.jamcollsurg.2006.06.002
Vol. 203, No. 4, October 2006
Figure 1. Patient with two seromas in a single right upper extremity loop forearm arteriovenous graft. One seroma is at the venous anastomosis and one seroma is at the arterial anastomosis.
2005. There were a total of 535 graft insertions in 427 patients. The technique for initial graft insertion was standard throughout this study interval. All grafts were 7.0-mm PTFE grafts and all procedures were performed under the guidance of the same attending surgeon. Arterial inflow was determined to be adequate if there was strong pulsatile blood flow from the artery. Venous outflow was determined to be adequate if, after venotomy, a 4F Fogarty catheter passed easily proximally into the vein, the vein dilated to 3.5 mm with a coronary dilator, and the vein flushed easily with heparinized saline. After this, a subcutaneous (SC) tunnel was formed with a tunneler in a loop, C-shape, or straight path. The PTFE graft with a preattached plastic sheath and tunneler tip was drawn through the tunnel and then the plastic sheath was removed. The grafts were anastomosed to both vessels in an end-to-side manner using 6-0 Gore-Tex (WL Gore & Assoc) suture in a continuous running fashion. The incisions were closed with 3-0 Vicryl (Ethicon, Inc) SC sutures and 4-0 nylon skin sutures. In
Figure 3. New graft-to-graft anastomosis in the forearm and graftto-vein anastomoses in the upper arm are illustrated.
Dauria et al
Seroma after Arteriovenous Graft Placement
507
Figure 2. Both segments of graft involving the seromas were bypassed. Brachial artery and accompanying vein in the upper arm were isolated. End-to-side anastomoses between the new graft segments and the artery and vein in the upper arm were performed.
four patients with obvious graft weeping at the time of initial insertion, which did not cease after multiple attempts with various hemostatic agents, the weeping segment was removed and replaced with a new segment of graft. Clinical variables including age, gender, diagnosis of diabetes, length of time on hemodialysis, graft location, and anatomic graft type were analyzed. Graft status in all patients was determined at 3, 6, and 12 months. For this study, inclusion criteria for diagnosis of seroma were a progressive increase in size requiring surgical intervention and surgical findings of clear fluid surrounded by a pseudocapsule consistent with seroma. All treated seromas were ⬎4 cm ⫻ 5 cm and expanding (Fig. 1). Treatment consisted of incision and evacuation of the seroma, complete excision of the entire graft, or primary bypass of the involved graft segment only (Figs. 2, 3, 4, 5). All
Figure 4. Graft-to-graft anastomoses of the arterial limb in the forearm are shown. Also, an incision was made over the seromas in the middle of the patient’s forearm to drain the seromas.
508
Dauria et al
Seroma after Arteriovenous Graft Placement
Figure 5. The surrounding capsules of the seroma were removed after the area was bypassed.
patients who were surgically treated for seromas had drains placed and received perioperative IV antibiotics (Fig. 6). Multiple variables were analyzed as potential risk factors for seromas using the Kaplan-Meier test for statistical analysis. RESULTS In the study period, a total of 535 arteriovenous grafts were inserted into 427 patients. Grafts were in a forearm (68%), upper arm (25%), or thigh (7%) position. Ten patients (2.3%) presented with a seroma and all underwent surgical intervention. This represents an overall incidence of 1.7% of the total number of graft insertions. Table 1 summarizes the characteristics of the patients presenting with seromas. Nearly all patients with seromas were on dialysis at the time of insertion. There was no significant difference in seroma formation based on gender, the presence of diabetes, or age (51.2 years ⫾ 15.3 years with seroma; 54.7 years ⫾ 14.6 years without seroma; p ⬎ 0.05). There were nine other patients who presented with small, perigraft fluid collections that were thought to be hematomas; these resolved spontaneously. Table 2 summarizes the surgical interventions and case results. All seromas occurred near an anastomosis, except in one patient in whom the seroma formed along the entire length of the graft. Based on the anatomic type of graft (two forearm loops, seven upper-arm C-grafts, and one thigh graft) there was no significant difference between forearm loop grafts or straight forearm grafts (p ⬎ 0.05). There was a difference in seroma formation between an upper arm C-graft and both a forearm loop graft and forearm straight graft with seromas forming more often in upper arm grafts (p ⫽ 0.007, p ⫽ 0.04,
J Am Coll Surg
Figure 6. The incision over the seromas was closed over a small drain. The distal loop of graft in the forearm is able to be accessed for hemodialysis immediately after operation (circled segment).
respectively). There was no statistical significance between the thigh graft and any of the upper extremity grafts. The time to first surgical intervention ranged from 28 to 144 days (mean 83 ⫾ 40 days). Of the four patients who underwent evacuation, two eventually had the graft completely excised because of infection. One patient underwent complete primary excision initially because of seroma formation along the entire length of the graft. Five patients underwent primary bypass of only the involved graft segment. Primary bypass of the involved graft segment has resulted in longterm patency in four of five grafts. These remained functional at the end of the study period. The mean number of days functioning since last surgical intervention of all currently patent grafts is 403 days (⫾366 days SD). On review of the current literature, a comparative analysis of seroma formation can be found for other types of vascular grafts, which shows results that are similar to our dialysis grafts. In this previously published analysis, removal of seroma and replacement of the graft led to a 92% cure rate, and evacuation alone had a 72% cure rate, but it also had the highest rate of infection (12%). Seromas undergoing multiple aspirations led to a 70% cure rate. Those that were simply observed had a 65% cure rate.1,2 Synthetic arterial grafts generally become encapsulated in fibrous tissue 6 weeks postoperatively.3 The primary suggested cause of seroma is failure of the surrounding connective tissue to incorporate the graft. This has been documented histologically by a fibrous pseudomembrane lining the seroma wall and immature fibroblasts lining the graft. In contrast, a
Vol. 203, No. 4, October 2006
Dauria et al
Seroma after Arteriovenous Graft Placement
509
Table 1. Patient Summaries Patient No.
1 2 3 4 5 6 7 8 9 10 Total (n ⫽ 10)
Gender
Age at time of placement (y)
Diabetes status at time of placement
Dialysis duration up to placement (mo)
Prior vascular access
F F F M F F M M F F (p ⬎ 0.05)
56 64 22 44 68 50 55 72 35 46 (p ⬎ 0.05)
Y Y N Y Y N N N N N (p ⬎ 0.05)
20.00 96.00 48.00 12.00 NOD 36.00 120.00 18.00 120.00 108.00 (p ⬎ 0.05)
Y Y Y Y N Y Y Y Y Y (p ⬎ 0.05)
F, female; M, male; N, no; NOD, not on dialysis; Y, yes.
normal graft will display a firmly adherent fibrous tissue covering and a healthy wall matrix lining. Several reasons for this failure to incorporate have been suggested. A transient fibroblast inhibitor, found in the seroma fluid of patients in whom this complication develops, disappears when the seroma resolves. An allergic reaction to the foreign material of the graft has also been postulated, as after removal of the graft the fibroblast inhibitory properties in the seroma fluid are no longer detectable. This suggests that the graft can induce production of some type of fibroblast inhibitor.3 Graft “wetting,” ie, when the hydrophobic surface of the PTFE graft becomes hydrophilic when in contact with blood or body fluid, might also be a contributing cause. A graft is supposed to seal by deposition of proteins and fibrous tissue attachment externally and internally through neointima formation. If “wetting” occurs, then this process can be hindered. Hemodynamic factors, such as flow rate (velocity) through the graft of ⬎1 L/minute might play a role more so than blood pressure. In previous studies, the rate of graft weeping becomes more pressuredependent with decreasing hematocrit. There has been a threefold increase in graft weeping with a decrease in hematocrit by one-half. Decreased oncotic pressure in malnourished patients and use of substances like povidone-iodine or alcohol around the graft can also lead to weeping. Extensive manipulation of the graft with instruments in the operating room causes graft injury, which can lead to weeping and seroma formation. A 41% increase in leakage through “bent” segments of graft has been seen, but
there is no relation between anastomotic angle and leakage. No consistent relationship between heparin administration and seroma formation has been found.4 It is likely that biochemical, mechanical, and structural issues act synergistically in the formation of seromas. Our study yields a statistically significant difference in the rate of seroma formation between an upper arm graft (p ⫽ 0.007) and a forearm graft (p ⫽ 0.04). This is likely a result of an increased flow rate through upper arm grafts, which causes increased pressure on the graft wall. Another study has also suggested that fat liquification can increase the risk of graft wetting and weeping.2 Grafts placed in an upper arm position are surrounded by more SC fat. These fat cells can be injured during vessel isolation and graft tunneling, resulting in fat liquification. A higher concentration of liquefied fat in the upper arm can contribute to higher seroma incidence. Every effort should be made to place a graft in the forearm position when possible. To prevent seromas, avoid graft wetting with chemicals, serum, liquefied fat, or blood. Materials most commonly involved with this complication have been shown to be knitted Dacron (Invista) and PTFE, with a higher percentage involving Dacron grafts.2 We use a PTFE graft that is preattached to a tunneler tip and covered with a protective sheath to assist with atraumatic tunneling and prevent stretching of the graft. This slider graft, by reducing graft stretching and wetting, has been shown to have statistically significant (p ⫽ 0.0005) lower rates of graft sweating and seroma development than grafts without a protective sheath.5
510
Dauria et al
Seroma after Arteriovenous Graft Placement
J Am Coll Surg
Table 2. Patient Results Patient No.
1 2 3 4 5 6 7 8 9 10
Graft type
First surgical intervention, d
Additional complications
Primary treatment
LUA C-graft RUA C-graft LUA C-graft Left forearm loop Left forearm loop RUA C-graft LUA C-graft Thigh SaphV Right forearm C-graft RUA C-graft
64 134 63 104 44 144 44 96 113 28
N N Y† N N N N N N Y†
Bypassed Bypassed Bypassed Bypassed Bypassed Evacuated Evacuated Evacuated Evacuated Excised
Recurrence
Second surgical intervention, d
Secondary treatment
Current status of grafts
Last treatment, d*
N N N N N N Y† N Y† NA
— — — — — — 42 — 264 —
— — — — — — Evacuated — Excised —
Functioning Functioning Excised† Functioning Functioning Functioning Excised† Functioning Excised Excised
729 626 — 199 8 31 — 824 — —
LUA, left upper arm; N, no; NA, not applicable; RUA, right upper arm; SaphV, saphenous vein; Y, yes. *Days functioning since last treatment of current patient grafts. † With infection.
Prosthetic angulation or bending should be avoided and smaller, tighter tunnels should be used. Effort to avoid excessive handling of the graft with forceps or other instruments should be taken to avoid graft damage. We also, when removing the clamps after the anastomoses, remove the venous clamps first to avoid the rapid filling causing weeping that is noted when releasing the arterial clamps first against closed venous clamps.6 In several patients, weeping could not be controlled with various hemostatic agents. In these patients, in an effort to avoid future seromas, the weeping segment was removed and replaced by a new segment during the initial operation. Seromas do not develop in these patients. Operation is generally indicated for expanding seromas before pressure necrosis and erosion through the skin occurs, which can result in infection and loss of graft. In 25% of patients, clinical evidence of perigraft seroma is seen within the first month postoperatively, but it can occur up to 1 year later.1 To treat a seroma, we suggest excising the seroma capsule and bypassing only that portion of the involved graft segment. Po-Jen and colleagues7 have also published two cases that were treated successfully in this manner. This method is simple to do technically, can be done under local anesthesia, and the part of the graft left behind can be accessed immediately, avoiding the morbidities associated with placing temporary lines for dialysis. The literature also suggests changing the type of graft material used from that which was used in the initial operation and rerouting through a new anatomic route. Although we did reroute the grafts, we did not change to a different type of graft material
and so far have not found it necessary to do so in our patients. In refractory cases of patients presenting late with a seroma, removal of a possible fibroblastinhibiting factor with plasmapheresis has been suggested. Sladen and colleagues3 use a protocol that calls for 3 exchanges a week for 2 weeks, twice a week for the next 2 weeks, and then once a week for 2 weeks. This is a total of 12 plasma exchanges in a 6-week period. There have been sporadic reports of using a collagen hemostat, replacement of the graft with umbilical/saphenous vein, or wrapping the graft in collagen fleece soaked in fibrin glue.7 After simple bypass of the involved graft segment we have not had recurrence. In conclusion, our incidence of seroma formation was found to be 1.7%. Although the relative rate of seroma formation is small in our series, it contributes to graft loss and increased morbidity in the end-stage renal disease patient. To minimize this complication, meticulous operative technique is required. Even if this complication occurs, the graft can still be salvageable. Aggressive management, including bypass of the involved segment, can lead to prolonged graft survival. Author Contributions
Study conception and design: Garvin Acquisition of data: Dauria, Dyk, Garvin Analysis and interpretation of data: Dauria, Dyk, Garvin Drafting of manuscript: Dauria Critical revision: Dauria, Garvin
Vol. 203, No. 4, October 2006
REFERENCES 1. Allaria PM, Lucatello A, Gandini E, et al. Relapsing seroma in a uremic patient bearing a PTFE graft as vascular access. J Vasc Access 2001;2:28–31. 2. Blumenberg RM, Gelfand ML, Dale A. Perigraft seromas complicating arterial grafts. Surgery 1985;97:194–203. 3. Sladen JG, Mandl MAJ, Grossman L, Denegri JF. Fibroblast inhibition: a new and treatable cause of prosthetic graft failure. Am J Surg 1985;149:587–590.
Dauria et al
Seroma after Arteriovenous Graft Placement
511
4. Bolton W, Cannon JA. Seroma formation associated with PTFE vascular grafts used as arteriovenous fistulae. Dial Transplant 1981;10:60–66. 5. Schild AF, Baltodano NM, Alfieri K, et al. New graft for low friction tunneling in vascular access surgery. J Vasc Access 2004;5:19–24. 6. Johnson JM. Serous fluid leakage through PTFE grafts: a possible explanation. J Thoracic Cardiovasc Surg 1985;89:469–471. 7. Po-Jen K, Yun-Hen L, Chu JJ, Lin PJ. Effective treatment for recurrent perigraft seromas of upper arm polytetrafluoroethylene grafts: report of two cases. Chang Gung Med J 2003;26:440–442.
There are approximately 300,000 people in the United States currently receiving longterm hemodialysis for end-stage renal disease. This number is expected to double by the year 2010. Despite the documented advantages of primary arteriovenous fistulas, approximately 60% of these patients still depend on prosthetic grafts for permanent access. In patients with end-stage renal disease, ⬎25% of all hospitalizations are vascular-access–related complications, with an eightfold higher mortality in patients with prosthetic grafts versus fistulas. One of these complications is seroma formation. A seroma is a collection of sterile, clear, ultrafiltered serum, surrounded by a nonsecretory fibrous soft tis-
sue pseudocapsule. Seromas, if untreated, can lead to wound infection, dehiscence, skin necrosis, graft thrombosis, or loss of the available graft puncture area. The incidence of seroma after arteriovenous graft placement for hemodialysis has been poorly defined in the current literature and has ranged from approximately 2% to 4%. It is poorly defined because some seromas resolve without treatment and are not reported, and others are grouped into hematomas and low-grade infection problems. There have been many different treatments proposed, but recurrence has remained a difficult problem. The goal of this study was to determine the incidence and outcomes of surgical management of seroma formation after arteriovenous graft insertion and to identify potential etiologic variables that can contribute to this complication.
Competing Interests Declared: None.
METHODS This study is a retrospective review of all patients who received an arteriovenous graft in a single large urban academic center between August 2002 and December
Received March 30, 2006; Revised May 27, 2006; Accepted June 7, 2006. From the Department of Surgery, St Louis University Hospital, St Louis, MO. Correspondence address: Danielle M Dauria, MD, Department of Surgery, St Louis University Hospital, 3635 Vista Ave, DesLoge Towers, 3rd Fl, St Louis, MO 63110. email: [email protected]
© 2006 by the American College of Surgeons Published by Elsevier Inc.
506
ISSN 1072-7515/06/$32.00 doi:10.1016/j.jamcollsurg.2006.06.002
Vol. 203, No. 4, October 2006
Figure 1. Patient with two seromas in a single right upper extremity loop forearm arteriovenous graft. One seroma is at the venous anastomosis and one seroma is at the arterial anastomosis.
2005. There were a total of 535 graft insertions in 427 patients. The technique for initial graft insertion was standard throughout this study interval. All grafts were 7.0-mm PTFE grafts and all procedures were performed under the guidance of the same attending surgeon. Arterial inflow was determined to be adequate if there was strong pulsatile blood flow from the artery. Venous outflow was determined to be adequate if, after venotomy, a 4F Fogarty catheter passed easily proximally into the vein, the vein dilated to 3.5 mm with a coronary dilator, and the vein flushed easily with heparinized saline. After this, a subcutaneous (SC) tunnel was formed with a tunneler in a loop, C-shape, or straight path. The PTFE graft with a preattached plastic sheath and tunneler tip was drawn through the tunnel and then the plastic sheath was removed. The grafts were anastomosed to both vessels in an end-to-side manner using 6-0 Gore-Tex (WL Gore & Assoc) suture in a continuous running fashion. The incisions were closed with 3-0 Vicryl (Ethicon, Inc) SC sutures and 4-0 nylon skin sutures. In
Figure 3. New graft-to-graft anastomosis in the forearm and graftto-vein anastomoses in the upper arm are illustrated.
Dauria et al
Seroma after Arteriovenous Graft Placement
507
Figure 2. Both segments of graft involving the seromas were bypassed. Brachial artery and accompanying vein in the upper arm were isolated. End-to-side anastomoses between the new graft segments and the artery and vein in the upper arm were performed.
four patients with obvious graft weeping at the time of initial insertion, which did not cease after multiple attempts with various hemostatic agents, the weeping segment was removed and replaced with a new segment of graft. Clinical variables including age, gender, diagnosis of diabetes, length of time on hemodialysis, graft location, and anatomic graft type were analyzed. Graft status in all patients was determined at 3, 6, and 12 months. For this study, inclusion criteria for diagnosis of seroma were a progressive increase in size requiring surgical intervention and surgical findings of clear fluid surrounded by a pseudocapsule consistent with seroma. All treated seromas were ⬎4 cm ⫻ 5 cm and expanding (Fig. 1). Treatment consisted of incision and evacuation of the seroma, complete excision of the entire graft, or primary bypass of the involved graft segment only (Figs. 2, 3, 4, 5). All
Figure 4. Graft-to-graft anastomoses of the arterial limb in the forearm are shown. Also, an incision was made over the seromas in the middle of the patient’s forearm to drain the seromas.
508
Dauria et al
Seroma after Arteriovenous Graft Placement
Figure 5. The surrounding capsules of the seroma were removed after the area was bypassed.
patients who were surgically treated for seromas had drains placed and received perioperative IV antibiotics (Fig. 6). Multiple variables were analyzed as potential risk factors for seromas using the Kaplan-Meier test for statistical analysis. RESULTS In the study period, a total of 535 arteriovenous grafts were inserted into 427 patients. Grafts were in a forearm (68%), upper arm (25%), or thigh (7%) position. Ten patients (2.3%) presented with a seroma and all underwent surgical intervention. This represents an overall incidence of 1.7% of the total number of graft insertions. Table 1 summarizes the characteristics of the patients presenting with seromas. Nearly all patients with seromas were on dialysis at the time of insertion. There was no significant difference in seroma formation based on gender, the presence of diabetes, or age (51.2 years ⫾ 15.3 years with seroma; 54.7 years ⫾ 14.6 years without seroma; p ⬎ 0.05). There were nine other patients who presented with small, perigraft fluid collections that were thought to be hematomas; these resolved spontaneously. Table 2 summarizes the surgical interventions and case results. All seromas occurred near an anastomosis, except in one patient in whom the seroma formed along the entire length of the graft. Based on the anatomic type of graft (two forearm loops, seven upper-arm C-grafts, and one thigh graft) there was no significant difference between forearm loop grafts or straight forearm grafts (p ⬎ 0.05). There was a difference in seroma formation between an upper arm C-graft and both a forearm loop graft and forearm straight graft with seromas forming more often in upper arm grafts (p ⫽ 0.007, p ⫽ 0.04,
J Am Coll Surg
Figure 6. The incision over the seromas was closed over a small drain. The distal loop of graft in the forearm is able to be accessed for hemodialysis immediately after operation (circled segment).
respectively). There was no statistical significance between the thigh graft and any of the upper extremity grafts. The time to first surgical intervention ranged from 28 to 144 days (mean 83 ⫾ 40 days). Of the four patients who underwent evacuation, two eventually had the graft completely excised because of infection. One patient underwent complete primary excision initially because of seroma formation along the entire length of the graft. Five patients underwent primary bypass of only the involved graft segment. Primary bypass of the involved graft segment has resulted in longterm patency in four of five grafts. These remained functional at the end of the study period. The mean number of days functioning since last surgical intervention of all currently patent grafts is 403 days (⫾366 days SD). On review of the current literature, a comparative analysis of seroma formation can be found for other types of vascular grafts, which shows results that are similar to our dialysis grafts. In this previously published analysis, removal of seroma and replacement of the graft led to a 92% cure rate, and evacuation alone had a 72% cure rate, but it also had the highest rate of infection (12%). Seromas undergoing multiple aspirations led to a 70% cure rate. Those that were simply observed had a 65% cure rate.1,2 Synthetic arterial grafts generally become encapsulated in fibrous tissue 6 weeks postoperatively.3 The primary suggested cause of seroma is failure of the surrounding connective tissue to incorporate the graft. This has been documented histologically by a fibrous pseudomembrane lining the seroma wall and immature fibroblasts lining the graft. In contrast, a
Vol. 203, No. 4, October 2006
Dauria et al
Seroma after Arteriovenous Graft Placement
509
Table 1. Patient Summaries Patient No.
1 2 3 4 5 6 7 8 9 10 Total (n ⫽ 10)
Gender
Age at time of placement (y)
Diabetes status at time of placement
Dialysis duration up to placement (mo)
Prior vascular access
F F F M F F M M F F (p ⬎ 0.05)
56 64 22 44 68 50 55 72 35 46 (p ⬎ 0.05)
Y Y N Y Y N N N N N (p ⬎ 0.05)
20.00 96.00 48.00 12.00 NOD 36.00 120.00 18.00 120.00 108.00 (p ⬎ 0.05)
Y Y Y Y N Y Y Y Y Y (p ⬎ 0.05)
F, female; M, male; N, no; NOD, not on dialysis; Y, yes.
normal graft will display a firmly adherent fibrous tissue covering and a healthy wall matrix lining. Several reasons for this failure to incorporate have been suggested. A transient fibroblast inhibitor, found in the seroma fluid of patients in whom this complication develops, disappears when the seroma resolves. An allergic reaction to the foreign material of the graft has also been postulated, as after removal of the graft the fibroblast inhibitory properties in the seroma fluid are no longer detectable. This suggests that the graft can induce production of some type of fibroblast inhibitor.3 Graft “wetting,” ie, when the hydrophobic surface of the PTFE graft becomes hydrophilic when in contact with blood or body fluid, might also be a contributing cause. A graft is supposed to seal by deposition of proteins and fibrous tissue attachment externally and internally through neointima formation. If “wetting” occurs, then this process can be hindered. Hemodynamic factors, such as flow rate (velocity) through the graft of ⬎1 L/minute might play a role more so than blood pressure. In previous studies, the rate of graft weeping becomes more pressuredependent with decreasing hematocrit. There has been a threefold increase in graft weeping with a decrease in hematocrit by one-half. Decreased oncotic pressure in malnourished patients and use of substances like povidone-iodine or alcohol around the graft can also lead to weeping. Extensive manipulation of the graft with instruments in the operating room causes graft injury, which can lead to weeping and seroma formation. A 41% increase in leakage through “bent” segments of graft has been seen, but
there is no relation between anastomotic angle and leakage. No consistent relationship between heparin administration and seroma formation has been found.4 It is likely that biochemical, mechanical, and structural issues act synergistically in the formation of seromas. Our study yields a statistically significant difference in the rate of seroma formation between an upper arm graft (p ⫽ 0.007) and a forearm graft (p ⫽ 0.04). This is likely a result of an increased flow rate through upper arm grafts, which causes increased pressure on the graft wall. Another study has also suggested that fat liquification can increase the risk of graft wetting and weeping.2 Grafts placed in an upper arm position are surrounded by more SC fat. These fat cells can be injured during vessel isolation and graft tunneling, resulting in fat liquification. A higher concentration of liquefied fat in the upper arm can contribute to higher seroma incidence. Every effort should be made to place a graft in the forearm position when possible. To prevent seromas, avoid graft wetting with chemicals, serum, liquefied fat, or blood. Materials most commonly involved with this complication have been shown to be knitted Dacron (Invista) and PTFE, with a higher percentage involving Dacron grafts.2 We use a PTFE graft that is preattached to a tunneler tip and covered with a protective sheath to assist with atraumatic tunneling and prevent stretching of the graft. This slider graft, by reducing graft stretching and wetting, has been shown to have statistically significant (p ⫽ 0.0005) lower rates of graft sweating and seroma development than grafts without a protective sheath.5
510
Dauria et al
Seroma after Arteriovenous Graft Placement
J Am Coll Surg
Table 2. Patient Results Patient No.
1 2 3 4 5 6 7 8 9 10
Graft type
First surgical intervention, d
Additional complications
Primary treatment
LUA C-graft RUA C-graft LUA C-graft Left forearm loop Left forearm loop RUA C-graft LUA C-graft Thigh SaphV Right forearm C-graft RUA C-graft
64 134 63 104 44 144 44 96 113 28
N N Y† N N N N N N Y†
Bypassed Bypassed Bypassed Bypassed Bypassed Evacuated Evacuated Evacuated Evacuated Excised
Recurrence
Second surgical intervention, d
Secondary treatment
Current status of grafts
Last treatment, d*
N N N N N N Y† N Y† NA
— — — — — — 42 — 264 —
— — — — — — Evacuated — Excised —
Functioning Functioning Excised† Functioning Functioning Functioning Excised† Functioning Excised Excised
729 626 — 199 8 31 — 824 — —
LUA, left upper arm; N, no; NA, not applicable; RUA, right upper arm; SaphV, saphenous vein; Y, yes. *Days functioning since last treatment of current patient grafts. † With infection.
Prosthetic angulation or bending should be avoided and smaller, tighter tunnels should be used. Effort to avoid excessive handling of the graft with forceps or other instruments should be taken to avoid graft damage. We also, when removing the clamps after the anastomoses, remove the venous clamps first to avoid the rapid filling causing weeping that is noted when releasing the arterial clamps first against closed venous clamps.6 In several patients, weeping could not be controlled with various hemostatic agents. In these patients, in an effort to avoid future seromas, the weeping segment was removed and replaced by a new segment during the initial operation. Seromas do not develop in these patients. Operation is generally indicated for expanding seromas before pressure necrosis and erosion through the skin occurs, which can result in infection and loss of graft. In 25% of patients, clinical evidence of perigraft seroma is seen within the first month postoperatively, but it can occur up to 1 year later.1 To treat a seroma, we suggest excising the seroma capsule and bypassing only that portion of the involved graft segment. Po-Jen and colleagues7 have also published two cases that were treated successfully in this manner. This method is simple to do technically, can be done under local anesthesia, and the part of the graft left behind can be accessed immediately, avoiding the morbidities associated with placing temporary lines for dialysis. The literature also suggests changing the type of graft material used from that which was used in the initial operation and rerouting through a new anatomic route. Although we did reroute the grafts, we did not change to a different type of graft material
and so far have not found it necessary to do so in our patients. In refractory cases of patients presenting late with a seroma, removal of a possible fibroblastinhibiting factor with plasmapheresis has been suggested. Sladen and colleagues3 use a protocol that calls for 3 exchanges a week for 2 weeks, twice a week for the next 2 weeks, and then once a week for 2 weeks. This is a total of 12 plasma exchanges in a 6-week period. There have been sporadic reports of using a collagen hemostat, replacement of the graft with umbilical/saphenous vein, or wrapping the graft in collagen fleece soaked in fibrin glue.7 After simple bypass of the involved graft segment we have not had recurrence. In conclusion, our incidence of seroma formation was found to be 1.7%. Although the relative rate of seroma formation is small in our series, it contributes to graft loss and increased morbidity in the end-stage renal disease patient. To minimize this complication, meticulous operative technique is required. Even if this complication occurs, the graft can still be salvageable. Aggressive management, including bypass of the involved segment, can lead to prolonged graft survival. Author Contributions
Study conception and design: Garvin Acquisition of data: Dauria, Dyk, Garvin Analysis and interpretation of data: Dauria, Dyk, Garvin Drafting of manuscript: Dauria Critical revision: Dauria, Garvin
Vol. 203, No. 4, October 2006
REFERENCES 1. Allaria PM, Lucatello A, Gandini E, et al. Relapsing seroma in a uremic patient bearing a PTFE graft as vascular access. J Vasc Access 2001;2:28–31. 2. Blumenberg RM, Gelfand ML, Dale A. Perigraft seromas complicating arterial grafts. Surgery 1985;97:194–203. 3. Sladen JG, Mandl MAJ, Grossman L, Denegri JF. Fibroblast inhibition: a new and treatable cause of prosthetic graft failure. Am J Surg 1985;149:587–590.
Dauria et al
Seroma after Arteriovenous Graft Placement
511
4. Bolton W, Cannon JA. Seroma formation associated with PTFE vascular grafts used as arteriovenous fistulae. Dial Transplant 1981;10:60–66. 5. Schild AF, Baltodano NM, Alfieri K, et al. New graft for low friction tunneling in vascular access surgery. J Vasc Access 2004;5:19–24. 6. Johnson JM. Serous fluid leakage through PTFE grafts: a possible explanation. J Thoracic Cardiovasc Surg 1985;89:469–471. 7. Po-Jen K, Yun-Hen L, Chu JJ, Lin PJ. Effective treatment for recurrent perigraft seromas of upper arm polytetrafluoroethylene grafts: report of two cases. Chang Gung Med J 2003;26:440–442.