Effect of Perigraft Seroma Fluid on Fibroblast Proliferation In Vitro Victor E. Pricolo, MD, Fabio Potenti, MD, Clarence H. Soderberg, MD, Providence, Rhode Island
Different theories have been postulated in the last 10 years in an attempt to explain the pathogenetic mechanisms involved in perigraft seroma formation. This paper reports our experience with in vitro analysis of the cellular composition and the effect of perigraft seroma fluid on fibroblast proliferative activity. Our results indicate that the cellular component of the fluid was almost exclusively represented by polymorphonuclear leukocytes. Perigraft seroma fluid was stimulatory on fibroblast growth, while plasma samples from the patient and normal controls were inhibitory. These findings suggest that a perigraft seroma may represent a complex phenomenon which cannot be explained exclusively in terms of fibroblast stimulation or inhibition. This pattern displays a close similarity to experimental observations in the early phases of wound healing. Perigraft seroma formation could be mediated by an intense and persistent chemotactic stimulus on leukocytes, which does not allow normal progression of wound healing through fibroblast growth and collagen deposition. (Ann Vasc Surg 1991 ;5:462-466). KEY WORDS:
Perigraft seroma; fibroblast proliferation; grafts.
Perigraft seroma represents an uncommon but potentially disastrous complication of arterial bypass surgery, which may occur in several different locations [1-4]. By definition, the fluid collection surrounding the vascular graft must be sterile, persistent, or enlarging for o v e r a month, and usually contained within a nonsecretory fibrous membrane [1]. Presence of prosthetic material undoubtedly plays a significant causal role, since nearly all seromas are associated with polytetrafluoroethylene (PTFE) or Dacron grafts [5]. However, the precise pathogenetic mechanisms involved in the formation of perigraft seroma are far from being elucidated. Excessive graft porosity with transudation o f serum has been postulated as a possible cause [6]. An immunologic or allergic phenomenon
in response to the prosthesis has been subsequently proposed [7]. Other authors attributed failure to incorporate the graft to excessive graft " w e e p i n g " caused by inadvertent soilage with fat, blood, alcohol or povidone-iodine [4,8]. More recently, inhibition of wound fibroblasts by factors present in the patient's own serum have been implicated as a novel hypothesis in perigraft seroma formation [5,9]. However, to our knowledge, the effects of the perigraft seroma fluid itself on fibroblast function in vitro have not been investigated. In this paper, our clinical and laboratory experience with a case of perigraft seroma is presented and discussed.
From the Department of Surgery, Rhode Island Hospital and Brown University, Providence, Rhode Island. Reprint requests: Victor E. Pricolo, MD, Department of Surgery, Rhode Island Hospital, APC Room 108, 593 Eddy Street, Providence, Rhode Island 02903.
A 60-year-old man had undergone abdominal aortic aneurysmectomy and reconstruction with a Dacron bifurcated graft in February of 1978. His postoperative course was complicated by an infection with Staphylococcus epidermidis of the left limb of the graft, which was
CASE REPORT
462
VOLUME5
E F F E C T OF P E R I G R A F T S E R O M A F L U I D O N F I B R O B L A S T S
No 5 - 1991
Fig. 1. Intraoperative photograph of left axillofemoral Dacron prosthesis with perigraft seroma in its midportion. Medially, planned placement of new polytetrafluoroethylene graft is outline on patient's skin.
excised and reconstructed with a saphenous vein obturatot bypass in July of the same year, In October 1978, this graft occluded and the patient required a PTFE femorofemoral and left femoropopliteal reversed saphenous vein bypass, with symptomatic relief of his disabling left leg claudication. He was readmitted to the hospital one year later with paresthesia and pain in his left leg. An arteriogram demonstrated occlusion of the femorofemoral and femoropopliteal grafts, profunda collaterals, and single calf vessel runoff. A lumbar sympathectomy was done with some symptomatic improvement. He continued to work for about six years; however, because of acceleration of symptoms and development of rest pain in his left foot, an arteriogram was repeated. The left profunda femoris was perfusing his left leg through collaterals in the thigh, in February 1986, a left axilloprofunda femoris bypass graft with Dacron was successfully performed. However, in the ensuing months, the patient noticed an area of swelling in the midportion of the axillofemoral graft. Computed tomographic (CT) scan showed a localized fluid collection around the graft and a diagnosis of perigraft seroma was made. The fluid was aspirated on five different occasions but reaccumulated rapidly and was causing discomfort. No aspirated samples showed microbial growth. In October 1989, the entire left axillofemoral Dacron graft was removed and replaced with a PTFE graft placed in a more medial virgin subcutaneous tunnel (Figs. 1,2). At one-year follow-up, the patient is doing well and the new graft remains well incorporated (Fig. 3).
463
Fig. 2. Intraoperative view of perigraft seroma cavity with Dacron graft and surrounding fibrous membrane.
incorporated (postoperative plasma). Perigraft seroma fluid was obtained by needle aspiration preoperatively under aseptic conditions. Heparinized blood samples were also collected from two healthy volunteers, one male (plasma control 1), and one female (plasma control 2). Cell-free supernatants were obtained by centrifugation from all samples and stored at -80~ until assayed.
Seroma fluid cytology The cell pellet obtained from centrifugation of the seroma fluid was resuspended and washed in culture medium (RPMI) and 0.15 M NHaCI-10 mM KHCO3 to lyse the erythrocytes (5 to 10 minutes incubation with constant shaking). After two washings with RPMI, 100/~1 aliquots of the cell suspension were centrifuged onto a microscopic slide, as
O
LABORATORY METHODS
Collection of samples Heparinized blood samples were drawn from the patient before the perigrafl seroma was operated upon (preoperative plasma) and two months postoperatively, when the new graft appeared to be well
Fig. 3, Postoperative result at one-year follow-up shows well healed scar (lateral to nipple) and well incorporated new graft (medial),
EFFECT OF PERIGRAFT SEROMA FLUID O N FIBROBLASTS
464
ANNALS OF VASCULAR SURGERY
FIBROBLAST PROLIFERATION ASSAY seroma fluid
20000 [3H]-THYM1DINE
UPTAKE(CPM) baseline
1000O
preop, plasma
control 1 control 2
postop, plasma
Fig. 4. Fibroblast proliferation in vitro results. Plasma samples from healthy controls (1 and 2). Preoperative and postoperative plasma samples demonstrate a marked inhibitory effect, as compared with baseline fibroblast proliferation rate. Perigraft seroma fluid is strongly stimulatory.
the medium was absorbed by a blotter (Cytospin 2*). The cells in monolayer were stained (WrightGiemsa), identified, and counted as a percentage composition of polymormonucleates, lymphocytes and macrophages.
Very few mononucleates (mostly lymphocytes) were seen. Over 95% of the granulocytes were neutrophils, with occasional eosinophils and rare basophils observed. All samples of perigraft seroma fluid sent to the microbiology laboratory showed no . bacterial or fungal growth.
Fibroblast proliferation assay
Minced wound fibroblasts were isolated, plated in culture medium, passaged for one week, then frozen in liquid nitrogen until needed [10]. Filtered (0.45 ixm) aliquots of plasma control 1, plasma control 2, preoperative plasma, postoperative plasma, and seroma fluid were used in triplicates in 10% and 20% concentrations of the total volume of each microtiter well, in cell culture medium complete with fetal bovine serum. Cells were cultured for 48 hours, pulsed for 18 hours with 0.5 txCi [3H]-thymidine, then harvested. The fibroblast proliferation rate is expressed as [3H]-thymidine uptake, measured as counts per minute (CPM) in a liquid scintillation counter t.
RESULTS Seroma fluid cytology
The perigraft seroma fluid collected on three separate occasions showed a consistent cellular composition pattern. The cell sediment, after centrifugation, accounted for approximately 25% of the total fluid volume. The Cytospin analysis revealed a leukocytic cellular infiltrate almost exclusively composed of polymorphonuclear cells (>98%). *Shandon, Pittsburgh, Pennsylvania. *Pharmacia LKB Nuclear Inc., Gaithersburg, Maryland.
Fibroblast proliferation assay
The influence of 20% concentrated perigraft seroma fluid, patient's preoperative and postoperative plasma, as well as plasma samples from two healthy volunteers on fibroblast proliferation in vitro is reported in Figure 4. At 10% concentration of the total well volume, remarkably similar findings were observed. Baseline counts resulted from addition of complete medium, instead of sample, to the culture, and provide a reference point to demonstrate an inhibitory or a stimulatory effect. Results are expressed as means of count values obtained in triplicate; scanning electron microscopy (SEM) did not exceed 10% of the mean. Plasma samples from two healthy volunteers (control 1 and control 2), and the patient's plasma, both preoperative and postoperative, were all inhibitory on fibroblast proliferation. On the other hand, the perigraft seroma fluid exerted a marked stimulatory effect. Although a statistical analysis is precluded, the consistency of results in repeated experiments at different concentrations, with a greater than 90% difference, is remarkable.
DISCUSSION A clearer understanding of the pathogenetic steps involved in the occurrence of perigraft seroma may have significant therapeutic implications. In fact,
VOLUME5
N o 5 - 1991
EFFECT OF PERIGRAFT SEROMA FLUID O N FIBROBLASTS
the detection of an inhibitor of fibroblast growth in their patients' serum has led Sladen and associates to successfully treat two cases of perigraft seroma with plasmapheresis [9]. Ahn and colleagues also reported their observation of a fibroblast inhibitor present in three patients' sera, which may have prevented adequate incorporation of the graft. In addition, the same authors observed disappearance of the inhibition after surgical removal of the graft (one case), or spontaneous resolution of the seroma (another case). Their conclusion was that effective therapy of perigraft seroma may include fibroblast modulation, graft removal, or both [5]. The latter of these two publications stimulated our study of the effect of the seroma fluid on fibroblast proliferative activity in vitro, under the assumption that the inhibitor may be produced at the site of injury. However, our results point out several noticeable differences: all tested plasma samples were inhibitory, and the seroma fluid was actually stimulatory. This apparent paradox may be explained by additional considerations. First, it is very difficult to compare studies with such a small number of cases and draw any general conclusions. Moreover, our methodology allows an objective quantification of fibroblast proliferative activity. In the above mentioned studies [5,9], tissue culture techniques were very similar to ours (RPMI and fetal bovine serum), but fibroblast proliferation was assessed by photomicrography and individual interpretation of cell culture density. Finally, the presence of an ~qnhibitor," the nature of which has not been clarified, in serum which normally should not surround the graft, does not necessarily establish a causal relationship. Previous work from our institution has addressed the question of regulation of fibroblast function in experimental wound models [10,w Our results, in longitudinal studies of changes in the wound environment over time, indicate that an early stimulatory activity (Days 1-3) is later replaced by an inhibitory activity on fibroblast growth (Days 1015). The inhibitory activity was found in the
Different theories have been postulated in the last 10 years in an attempt to explain the pathogenetic mechanisms involved in perigraft seroma formation. This paper reports our experience with in vitro analysis of the cellular composition and the effect of perigraft seroma fluid on fibroblast proliferative activity. Our results indicate that the cellular component of the fluid was almost exclusively represented by polymorphonuclear leukocytes. Perigraft seroma fluid was stimulatory on fibroblast growth, while plasma samples from the patient and normal controls were inhibitory. These findings suggest that a perigraft seroma may represent a complex phenomenon which cannot be explained exclusively in terms of fibroblast stimulation or inhibition. This pattern displays a close similarity to experimental observations in the early phases of wound healing. Perigraft seroma formation could be mediated by an intense and persistent chemotactic stimulus on leukocytes, which does not allow normal progression of wound healing through fibroblast growth and collagen deposition. (Ann Vasc Surg 1991 ;5:462-466). KEY WORDS:
Perigraft seroma; fibroblast proliferation; grafts.
Perigraft seroma represents an uncommon but potentially disastrous complication of arterial bypass surgery, which may occur in several different locations [1-4]. By definition, the fluid collection surrounding the vascular graft must be sterile, persistent, or enlarging for o v e r a month, and usually contained within a nonsecretory fibrous membrane [1]. Presence of prosthetic material undoubtedly plays a significant causal role, since nearly all seromas are associated with polytetrafluoroethylene (PTFE) or Dacron grafts [5]. However, the precise pathogenetic mechanisms involved in the formation of perigraft seroma are far from being elucidated. Excessive graft porosity with transudation o f serum has been postulated as a possible cause [6]. An immunologic or allergic phenomenon
in response to the prosthesis has been subsequently proposed [7]. Other authors attributed failure to incorporate the graft to excessive graft " w e e p i n g " caused by inadvertent soilage with fat, blood, alcohol or povidone-iodine [4,8]. More recently, inhibition of wound fibroblasts by factors present in the patient's own serum have been implicated as a novel hypothesis in perigraft seroma formation [5,9]. However, to our knowledge, the effects of the perigraft seroma fluid itself on fibroblast function in vitro have not been investigated. In this paper, our clinical and laboratory experience with a case of perigraft seroma is presented and discussed.
From the Department of Surgery, Rhode Island Hospital and Brown University, Providence, Rhode Island. Reprint requests: Victor E. Pricolo, MD, Department of Surgery, Rhode Island Hospital, APC Room 108, 593 Eddy Street, Providence, Rhode Island 02903.
A 60-year-old man had undergone abdominal aortic aneurysmectomy and reconstruction with a Dacron bifurcated graft in February of 1978. His postoperative course was complicated by an infection with Staphylococcus epidermidis of the left limb of the graft, which was
CASE REPORT
462
VOLUME5
E F F E C T OF P E R I G R A F T S E R O M A F L U I D O N F I B R O B L A S T S
No 5 - 1991
Fig. 1. Intraoperative photograph of left axillofemoral Dacron prosthesis with perigraft seroma in its midportion. Medially, planned placement of new polytetrafluoroethylene graft is outline on patient's skin.
excised and reconstructed with a saphenous vein obturatot bypass in July of the same year, In October 1978, this graft occluded and the patient required a PTFE femorofemoral and left femoropopliteal reversed saphenous vein bypass, with symptomatic relief of his disabling left leg claudication. He was readmitted to the hospital one year later with paresthesia and pain in his left leg. An arteriogram demonstrated occlusion of the femorofemoral and femoropopliteal grafts, profunda collaterals, and single calf vessel runoff. A lumbar sympathectomy was done with some symptomatic improvement. He continued to work for about six years; however, because of acceleration of symptoms and development of rest pain in his left foot, an arteriogram was repeated. The left profunda femoris was perfusing his left leg through collaterals in the thigh, in February 1986, a left axilloprofunda femoris bypass graft with Dacron was successfully performed. However, in the ensuing months, the patient noticed an area of swelling in the midportion of the axillofemoral graft. Computed tomographic (CT) scan showed a localized fluid collection around the graft and a diagnosis of perigraft seroma was made. The fluid was aspirated on five different occasions but reaccumulated rapidly and was causing discomfort. No aspirated samples showed microbial growth. In October 1989, the entire left axillofemoral Dacron graft was removed and replaced with a PTFE graft placed in a more medial virgin subcutaneous tunnel (Figs. 1,2). At one-year follow-up, the patient is doing well and the new graft remains well incorporated (Fig. 3).
463
Fig. 2. Intraoperative view of perigraft seroma cavity with Dacron graft and surrounding fibrous membrane.
incorporated (postoperative plasma). Perigraft seroma fluid was obtained by needle aspiration preoperatively under aseptic conditions. Heparinized blood samples were also collected from two healthy volunteers, one male (plasma control 1), and one female (plasma control 2). Cell-free supernatants were obtained by centrifugation from all samples and stored at -80~ until assayed.
Seroma fluid cytology The cell pellet obtained from centrifugation of the seroma fluid was resuspended and washed in culture medium (RPMI) and 0.15 M NHaCI-10 mM KHCO3 to lyse the erythrocytes (5 to 10 minutes incubation with constant shaking). After two washings with RPMI, 100/~1 aliquots of the cell suspension were centrifuged onto a microscopic slide, as
O
LABORATORY METHODS
Collection of samples Heparinized blood samples were drawn from the patient before the perigrafl seroma was operated upon (preoperative plasma) and two months postoperatively, when the new graft appeared to be well
Fig. 3, Postoperative result at one-year follow-up shows well healed scar (lateral to nipple) and well incorporated new graft (medial),
EFFECT OF PERIGRAFT SEROMA FLUID O N FIBROBLASTS
464
ANNALS OF VASCULAR SURGERY
FIBROBLAST PROLIFERATION ASSAY seroma fluid
20000 [3H]-THYM1DINE
UPTAKE(CPM) baseline
1000O
preop, plasma
control 1 control 2
postop, plasma
Fig. 4. Fibroblast proliferation in vitro results. Plasma samples from healthy controls (1 and 2). Preoperative and postoperative plasma samples demonstrate a marked inhibitory effect, as compared with baseline fibroblast proliferation rate. Perigraft seroma fluid is strongly stimulatory.
the medium was absorbed by a blotter (Cytospin 2*). The cells in monolayer were stained (WrightGiemsa), identified, and counted as a percentage composition of polymormonucleates, lymphocytes and macrophages.
Very few mononucleates (mostly lymphocytes) were seen. Over 95% of the granulocytes were neutrophils, with occasional eosinophils and rare basophils observed. All samples of perigraft seroma fluid sent to the microbiology laboratory showed no . bacterial or fungal growth.
Fibroblast proliferation assay
Minced wound fibroblasts were isolated, plated in culture medium, passaged for one week, then frozen in liquid nitrogen until needed [10]. Filtered (0.45 ixm) aliquots of plasma control 1, plasma control 2, preoperative plasma, postoperative plasma, and seroma fluid were used in triplicates in 10% and 20% concentrations of the total volume of each microtiter well, in cell culture medium complete with fetal bovine serum. Cells were cultured for 48 hours, pulsed for 18 hours with 0.5 txCi [3H]-thymidine, then harvested. The fibroblast proliferation rate is expressed as [3H]-thymidine uptake, measured as counts per minute (CPM) in a liquid scintillation counter t.
RESULTS Seroma fluid cytology
The perigraft seroma fluid collected on three separate occasions showed a consistent cellular composition pattern. The cell sediment, after centrifugation, accounted for approximately 25% of the total fluid volume. The Cytospin analysis revealed a leukocytic cellular infiltrate almost exclusively composed of polymorphonuclear cells (>98%). *Shandon, Pittsburgh, Pennsylvania. *Pharmacia LKB Nuclear Inc., Gaithersburg, Maryland.
Fibroblast proliferation assay
The influence of 20% concentrated perigraft seroma fluid, patient's preoperative and postoperative plasma, as well as plasma samples from two healthy volunteers on fibroblast proliferation in vitro is reported in Figure 4. At 10% concentration of the total well volume, remarkably similar findings were observed. Baseline counts resulted from addition of complete medium, instead of sample, to the culture, and provide a reference point to demonstrate an inhibitory or a stimulatory effect. Results are expressed as means of count values obtained in triplicate; scanning electron microscopy (SEM) did not exceed 10% of the mean. Plasma samples from two healthy volunteers (control 1 and control 2), and the patient's plasma, both preoperative and postoperative, were all inhibitory on fibroblast proliferation. On the other hand, the perigraft seroma fluid exerted a marked stimulatory effect. Although a statistical analysis is precluded, the consistency of results in repeated experiments at different concentrations, with a greater than 90% difference, is remarkable.
DISCUSSION A clearer understanding of the pathogenetic steps involved in the occurrence of perigraft seroma may have significant therapeutic implications. In fact,
VOLUME5
N o 5 - 1991
EFFECT OF PERIGRAFT SEROMA FLUID O N FIBROBLASTS
the detection of an inhibitor of fibroblast growth in their patients' serum has led Sladen and associates to successfully treat two cases of perigraft seroma with plasmapheresis [9]. Ahn and colleagues also reported their observation of a fibroblast inhibitor present in three patients' sera, which may have prevented adequate incorporation of the graft. In addition, the same authors observed disappearance of the inhibition after surgical removal of the graft (one case), or spontaneous resolution of the seroma (another case). Their conclusion was that effective therapy of perigraft seroma may include fibroblast modulation, graft removal, or both [5]. The latter of these two publications stimulated our study of the effect of the seroma fluid on fibroblast proliferative activity in vitro, under the assumption that the inhibitor may be produced at the site of injury. However, our results point out several noticeable differences: all tested plasma samples were inhibitory, and the seroma fluid was actually stimulatory. This apparent paradox may be explained by additional considerations. First, it is very difficult to compare studies with such a small number of cases and draw any general conclusions. Moreover, our methodology allows an objective quantification of fibroblast proliferative activity. In the above mentioned studies [5,9], tissue culture techniques were very similar to ours (RPMI and fetal bovine serum), but fibroblast proliferation was assessed by photomicrography and individual interpretation of cell culture density. Finally, the presence of an ~qnhibitor," the nature of which has not been clarified, in serum which normally should not surround the graft, does not necessarily establish a causal relationship. Previous work from our institution has addressed the question of regulation of fibroblast function in experimental wound models [10,w Our results, in longitudinal studies of changes in the wound environment over time, indicate that an early stimulatory activity (Days 1-3) is later replaced by an inhibitory activity on fibroblast growth (Days 1015). The inhibitory activity was found in the
