- Email: [email protected]
Prosthetic Graft Infection: Diagnostic and Therapeutic Role of Interventional Radiology
Prosthetic Graft Infection: Diagnostic and Therapeutic Role of Interventional ~ a d i o l o g ~ ' Plinio Rossi, MD Flaminia M. Arata, MD Filippo M. Salvatori, MD Mario Bezzi, MD Francesco Speziale, MD Domenico Lauri, MD Enrico Sbarigia, MD
Index terms: Aorta, grafts and prostheses. Grafts, infection. Grafts, interventional procedure
JVIR 1997; 8:271-277
PURPOSE: Aortic graft infection is a rare complication of aortic surgery associated with high mortality and morbidity rates. The aim of the study was to evaluate the authors' experience with diagnostic and interventional procedures in the management of aortic graft infection. METHODS: Six patients with clinical signs suggestive of graft infection were studied. Abdominal computed tomography (CT) was performed in all patients. Diagnostic work-up was performed with perigraft puncture for fluid aspiration and contrast medium injection to confirm clinical signs or CT findings of graft infection, and to detect graft-enteric fistulas. Percutaneous drainage of the periprosthetic abscess was performed as attempted therapy in four patients. RESULTS: In four patients, CT examination findings were considered positive for graft infection. Culture and cytologic examination of aspirates revealed numerous white blood cells but did not show any growth. In three patients, direct contrast medium injection into the perigraft space allowed the detection of a graft-enteric fistula. In three patients, percutaneous drainage was performed as preoperative therapy, allowing improvement of the surgical outcome; in one patient the catheter was left in place as an attempt at definitive therapy, allowing a complete recovery. CONCLUSIONS: The percutaneous approach to prosthetic graft infection permits both the drainage and aspiration of the fluid in the perigraft area for laboratory studies. Percutaneous drainage can offer the possibility of definitive cure in patients for whom surgical management is considered too risky or, at least, represents a temporizing maneuver to improve the patient's general condition before surgery.
'
From the Departments of Radiology (P.R., F.M.A., F.M.s., M.B.) and Vascular Surgery (F.S., E.S.), 111 Cattedra, University of Rome "La Sapienza," Policlinico Umberto I, V.le Regina Elena, 324, 00161 Rome, Italy. Received August 3, 1996; revision requested September 3; revision received and accepted November 5. Address correspondence to P.R.
o SCVIR, 1997
INFECTION of an aortic prosthesis is a rare but life-threatening complication of aortic reconstructive surgery. The reported incidence from 1.5%to 6%, with a high mortality rate ranging from 20% to 75% (1).Graft infections are usually limited to the prosthesis Or with a graftenteric fistula, with different clinical implications. Treatment is usually by means of surgical management with complete excision of
the infected graft and extra-anatomic bypass or in situ graft replacement without any previous drainage. However, mortality and complication rates remain high (2). The clinical signs of prosthetic graft infection may be subtle and progress slowly, rendering early diagnosis difficult. Therefore, clinical suspicion and a thorough diagnostic work-up are necessary to make a prompt diagnosis and improve the prognosis.
271
272
Journal of Vascular and Interventional Radiology
March-April 1997
Radiology offers many imaging modalities with which to achieve a diagnosis, such as computed tomography (CT), ultrasonography (US), magnetic resonance (MR), and nuclear medicine. CT and MR are the most suitable techniques for detecting perigraft fluid collections. Moreover, CT-guided periaortic fluid as~irationand contrast medium injection in the periprosthetic space, as an attempt to visualize a graft-enteric fistula, can provide a more accurate diagnosis. Scintigraphy may reveal a graft infection as an abnormal accumulation of radionuclides. Angiography usually is not needed, but it can add information about the vascular tree preoperatively. The aim of our study was to evaluate our experience in the diagnosis and management of prosthetic graft infection by using radiologic procedures.
1 MATERIALS AND MIITHODS From 1987 to 1996 we studied six patients who had previously undergone aortic grafting for aneurysmal disease and had clinical signs suggestive of prosthetic graft infection, with a time to presentation after the primary vascular procedure ranging from 1 to 9 years (mean, 3.6 years). All patients were men, ranging in age from 39 to 86 years. The clinical presentation was characterized in all patients by aspecific symptoms, such as fever, leukocytosis, increase of erythrocyte sedimentation rate, and abdominal or back pain. More specific signs of prosthetic graft infection, such as peripheral septic emboli and gastrointestinal bleeding, were present in one patient only. All patients underwent CT examination performed with various equipment (DRH Siemens, Erlangen, Germany, or GE Pace; GE Medical Systems, Milwaukee, WI). Contiguous 5-mm-thick sections of the abdomen, proximal and distal to the aortic graft, were obtained during intravenous injection of 150 mL of contrast material, continu-
ously or in multiple boluses a t a rate of 3.0 mL/sec. When a perigraft puncture was being considered, no contrast agent was administered orally before CT performed with the patient in the prone position. Thickened graft wall and perigraft fluid collection were considered indirect signs, whereas gas content in the periprosthetic space was considered evidence of aortic graft infection at CT. Technetium-99m scanning was performed in two patients only. Leukocytes were labeled with 200 MBq Tc-99m HMPAO, and images were obtained with a gamma camera with a large field-of-view at 2-3 hours after administration of the labeled leukocytes. Tc-99m scanning was considered to be positive for infection when increased uptake was seen at the aortic graft. All interventional procedures for detection and management of prosthetic graft infections were performed with use of a 18-gauge needle with a Teflon sheath, a 0.035-inch guide wire (Cook, Bloomington, IN), and a 10-12-F catheter (Medi-techJBoston Scientific, Watertown, MA) when abscess drainage was necessary. All patients underwent perigraft puncture: with CT guidance in five patients, a translumbar approach in four patients, and an anterior approach in one patient. In the sixth patient, the perigraft puncture was performed under US guidance by means of a transperitoneal approach. Perigraft fluid was aspirated to reach a clear diagnosis of infection based on culturing and cytologic examination of the aspirate by identifying either the infectious organisms or the presence of white blood cells. Perigraft puncture and contrast medium injection in the graft bed were performed to evaluate the presence of graft-enteric fistula. The spread of iodinated contrast material from the periaortic lumen into the bowel was considered clear evidence of graft-enteric fistula. Endoscopy was performed in three patients to compare radiologic findings with eventual mucosal abnormalities.
Percutaneous drainage was performed in four patients in whom the diagnosis of prosthetic graft infection was certain: as preoperative therapy in three patients and as definitive therapy in the fourth patient. In the other two patients, drainage was not feasible because, in one patient, the rim of fluid around the graft was too small and, in the other patient, surgery was scheduled after the diagnostic puncture. Each individual patient's history, management, and follow-up can be easily described, as follows: Case 1.-A 72-year-old man developed abdominal and back pain radiating to the left leg, with fever, leukocytosis, and a rise in erythrocyte sedimentation rate. He had undergone aorto-iliac Dacron grafting 2 years earlier. CT findings (Fig la) were considered indicative of prosthetic graft infection because of the presence of an air-fluid level surrounding the graft. A CT-guided perigraft puncture with direct contrast medium injection was performed (Fig lb), which allowed detection of a fistula between the infected graft and the jejunum (Fig lc, Id). The catheter was left in place for approximately 6 months as an attempt a t definitive therapy because surgical management was considered too risky due to the patient's poor cardiovascular condition (Fig le, If). The patient recovered completely, with disappearance of the periprosthetic abscess, and did well for approximately 2 years, when he underwent CT examination because of fever and leukocytosis. CT showed that the prosthetic graft was not involved: both the graft and the retroperitoneal space were normal (Fig lg). The patient died 2 months later of sepsis of unknown origin. Case 2.-A 56-year-old man, with a history of aortic valve replacement, began to feel ill and developed fever, back pain, and leukocytosis 2 years after aortic Dacron grafting. CT of the abdomen showed a thickened graft wall with an air-fluid level and gas content. Under CT guidance, a perigraft puncture with direct contrast
Rossi et a1
273
Volume 8 Number 2
b.
a.
Figure 1. (a)CT scan of the abdomen in a 72-year-old man. Two years after aortic grafting for aneurysm, he developed pain in the left leg and fever. An abscess was seen along the left limb of the aortic graft with gas bubbles. (b) Translumbar approach to the lesion few days later, with the patient in the right lateral decubitus position. ( c ) Contrast material injection into the lesion shows the abscess cavity and a fistulous communication with the small bowel (Fig 1 continues).
C.
medium injection was performed and showed the spread of contrast agent from the perigraft lumen into the third portion of the duodenum, revealing the presence of a graftenteric fistula not visualized a t a previously performed endoscopy (Fig Za, 2b). Percutaneous drainage was not performed because of the small rim of fluid around the graft. Specific antibiotic therapy was given before surgery. However, the patient died of cerebral hemorrhage, probably related to the anticoagulation therapy. Case 3.-A 64-year-old man was admitted as a n emergency case 5 years after aortic Dacron grafting. The clinical presentation was characterized by gastrointestinal bleeding, peripheral septic emboli, leukocytosis, and fever. CT findings were
considered unreliable with respect to graft infection because only a small rim of soft-tissue attenuation surrounded the graft, with no evidence of frank fluid or gas (Fig 3). At endoscopy, no graft-enteric fistula was visualized. The patient underwent CT-guided puncture for diagnostic aspiration and direct contrast medium injection. Culture of the aspirate showed the presence of numerous white blood cells but revealed no growth, probably related to 2 months of antibiotic therapy. No fistula was seen during contrast medium injection. The patient refused surgery because his clinical condition was improving. However. he returned 5 months later with a severe infection and underwent surgery a s a n emergency procedure when complete excision of
the infected graft and in situ replacement was performed. No fistula was found a t surgery, only inflammatory adhesion with the duodenum. The patient is still alive and well a t 1.5-year follow-up. Case 4.-A 61-year-old man was admitted for evaluation because of clinical suspicion of a prosthetic graft infection 3 years after aortoiliac Dacron grafting. CT showed a perigraft fluid collection with a n air-fluid level. Direct contrast medium injection showed no graftenteric fistula. Percutaneous drainage was performed under CT guidance. The catheter was left in place for 7 days, allowing for stabilization of the patient who subsequently underwent successful surgical management with complete excision of the infected graft and construction
274
Journal of Vascular and Interventional Radiology
March-April 1997
Figure 1 (continued). (d) Simultaneous opacification of the abscess cavity and the small bowel after contrast material injection. (e) Four months later, there is still a residual cavity no longer communicating with the small bowel. The patient was afebrile and doing well (Fig 1 continues).
of an extra-anatomic bypass. He was subsequently lost to follow-up. Case 5.-A 39-year-old man developed fever and general malaise 1year after aortic polytetrafluoroethyene grafting for aneurysmal disease, with a blood pressure of 80150 mm Hg. Abdominal CT showed a thickened graft wall and a perigraft fluid collection without gas content. The patient underwent Tc-99m imaging, which showed an increased uptake of labeled leukocytes a t the aortic graft. Periaortic fluid was aspirated and the culture
showed numerous white blood cells. No fistula was seen during contrast material injection. US-guided percutaneous drainage was carried out before surgical treatment. Total removal of the infected graft with in situ replacement was then performed, and the patient had no major complications at a 5-year follow-up. Case 6.-An 86-year-old man was admitted with a 6-month history of recurrent septic fever 9 years after aortic Dacron grafting. CT confirmed the clinical suspicion of prosthetic graft infection by revealing the presence of a large perigraft fluid collection with an enhancing irregular wall consistent with " manulation tissue. He underwent Tc-99m scanning, which showed an increased uptake of labeled leukocytes at the aortic graft. Endoscopy did not allow the visualization of a graft-enteric fistula, but revealed the presence of pus in the duodenum. A CT-guided perigraft puncture with direct contrast medium injection was also performed and a secondary fistula between the infected graft and the duodenum was detected. Because of the severity of the infection and the patient's age and poor general condition, a percutaneous drainage was left in place as preoperative therapy. Two weeks later, the patient underwent emergency surgery for active, massive gastrointestinal hemorrhage. At surgery, active bleeding into the duodenum was found at the site of the proximal anastomosis, and partial removal of the infected graft was performed. Because of clinical signs suggestive of pulmonary emboli, a vena cava filter was subsequently placed. The patient is still alive and well at 6-month follow-up.
1 RESULTS Of the six patients clinically suspected of having prosthetic graft infection who underwent CT examination, five had positive CT findings consisting of a perigraft fluid collection with or without gas content. In two of these five pa-
tients, a Tc-99m radionuclide scan showed an increased uptake of labeled leukocytes at the aortic graft, consistent with the presence of infection. In the sixth patient, in whom only a small rim of softtissue attenuation surrounded the graft, with no evidence of fluid or gas content, the CT findings were considered suggestive of graft infection but not completely reliable. Perigraft fluid, after lavage, was aspirated and collected to allow culture and cytologic examinations that revealed the presence of numerous white blood cells but did not show any growth. The patient refused both nuclear medicine scanning and surgical treatment. The presence of a graft-enteric fistula, suggested on CT scans by adhesions between the graft and the enteric loops, was easily proved in three patients with radiologic procedures. Perigraft puncture and direct contrast medium injection under CT guidance were performed, allowing graft-enteric fistula to be correctly diagnosed by the spread of contrast agent from the perigraft lumen into the bowel. Endoscopy was performed only in two patients; in one patient the finding was negative, and in the other, intraduodenal pus was found. In one patient, there was no extravasation of iodinated contrast material into the bowel but the patient had a graft-enteric fistula found at surgery; endoscopy was negative as well. The last two patients did not undergo endoscopic examination. Percutaneous drainage was performed as a temporizing treatment before surgery in three patients to improve the their general condition and was then followed by surgical management. In one patient the drain was left in place for 6 months as attempted definitive therapy because surgery was not indicated due to the high cardiovascular risk.
Despite improved surgical procedures, aortic graft infection remains
Rossi et a1
275
Volume 8 Number 2
f. Figure 1 (continued).
g.
(f) The last control image obtained before the catheter was removed, approximately 6 months after the beginning of the drainage, shows a small tract t h a t closed spontaneously. ( g ) CT scan of the abdomen 2 years after removal of the catheter when t h e patient returned for sepsis.
a. b. Figure 2. (a, b) CT scan of the abdomen showed fluid collection i n the periaortic graft area. A translumbar puncture of t h e periaortic area was made; after aspiration of t h e fluid, contrast material was injected, which visualized the periaortic area and the third portion of t h e duodenum.
a rare but serious complication associated with high mortality and complication rates. Surgical management consists of total removal of the infected graft and performance of a n extra-anatomic bypass or in situ graft replacement (2-5). The correct management of patients with minimal clinical signs and low-grade infection is less clear; in
these cases more conservative surgery such as partial graft excision, local debridement, and antibiotic irrigation can be performed. These procedures may be equally effective and do not carry the high complication rates associated with total excision (6,7). Many imaging modalities are available for the detection of infection
of an aortic prosthesis (ie, US, CT, MR, scintigraphy, and, less frequently, angiography) (8-10). Various scintigraphy techniques have been used to detect prosthetic graft infections; Tc-99m scanning has been recently proposed as a noninvasive technique for the diagnosis of lowgrade graft infection in patients with minimal clinical signs (11).
276
Journal of Vascular and Interventional Radiology -
-
-
March-April 1997
Figure 3. CT scan of the abdomen showed a small rim of soft-tissue attenution with no evidence of fluid or gas content.
We consider CT to be the diagnostic method of choice. So far, CT is the best and safest way of performing interventional procedures such as periaortic fluid aspiration, contrast medium injection, and drainage. In selected cases with large collections extending toward the anterior abdominal wall, a safe route for puncture also can be found with US, and the drainage be performed under such guidance. In all six patients, the prosthetic graft infection was first suspected clinically; all of the patients were studied with CT to evaluate the possibility of graft involvement. The presence of a perigraft fluid collection and gas content was suggestive of infection, while the presence of a minimally thickened graft wall, in default of other findings, necessitated further examinations. In patients in whom CT findings were not definitive for infection, we performed percutaneous needle aspiration. The perigraft puncture was preferably performed with use of the same technique as is used for translumbar angiography, with the patient in the prone position to avoid both the bowel and the native aortic wall. Anterior transperitoneal puncture can also be performed. Examination and culture of the aspirates can be useful in assessing the presence of white blood cells and in determining the causative organism. Unfortunately, as was the case in our series, the culture may result sterile, most likely because of antibiotic therapy. In
~ a t i e n t sin whom there was a fluid collection, we proceeded to drainage, using the same approach as was described earlier for the diagnostic puncture. The drainage material was sent for culture analysis as well. The presence of a graft-enteric fistula is usually suspected on a clinical basis. Although endoscopy is important for excluding other causes of gastrointestinal bleeding, it often fails to detect a graftenteric fistula (12). In our experience, CT-guided perigraft puncture with injection of contrast medium into the graft bed allowed, in three patients, a correct diagnosis of a graft-enteric fistula by showing the spread of contrast material from the perigraft lumen into the bowel. We recommend avoiding oral administration of contrast agent for CT whenever this technique is employed, to facilitate detection of bowel opacification. The choice among different surgical procedures in the management of prosthetic graft infection may depend on the severity of the infection (13). Total graft removal and extra-anatomic bypass are considered mandatory in patients with severe prosthetic infection (2-5), even though the mortality and morbidity are high. Aortic stump disruption remains a catastrophic event after complete excision of the infected prosthesis (2,14). In situ graft replacement has been proposed for the treatment of selected patients with a low-grade infection in whom a less aggressive approach may improve the surgical outcome (13,15). In our patients, interventional procedures have proved valuable in the diagnosis and management of prosthetic graft infections. Percutaneous needle aspiration under CT guidance has provided a safe way of diagnosing infection in patients with unclear clinical signs and ambiguous CT findings. Moreover, we consider CT-guided perigraft puncture with direct contrast agent injection to be a suitable radiologic procedure for detecting graft-enteric fistulas.
Percutaneous treatment of infected prostheses, recently proposed as an alternative conservative aDproach in selected patients (161, is not usually considered by surgeons to be a safe procedure. Nevertheless, we suggest that percutaneous drainage of perigraft fluid collections be employed as a preoperative measure to "cool off' the infection, or in any patient for whom surgical management is considered too risky because of associated heart, lung, or kidney disease, or because of advanced age. Moreover, we speculate that preoperative percutaneous drainage, by ameliorating the patient's general condition, may help to improve the surgical outcome in patients with severe infection. References 1. Buckels JAC, Wilson SE. Prevention and management of prosthetic graft infection. In: Veith FJ, Hobson RW 11, Williams RA, Wilson SE, eds. Vascular surgery. 2nd ed. New York: McGraw-Hill, 1994; 10811089. 2. Reilly LM, Altman H, Lusby RJ, Kersh RA, Erhenfeld WK, Stoney RJ. Late results following surgical management of vascular graft infection. J Vasc Surg 1984; 1:36-44. 3. Trout H 111, Kozloff L, Giordano JM. Priority of revascularization in patients with graft enteric fistulas, infected arteries, or infected arterial prosthesis. Ann Surg 1984; 199:669-683. 4. Edwards MJ, Richardson JD, Klamer TW. Management of aortic prosthetic infections. Am J Surg 1988; 155:327-330. 5. Calligaro KD, Veith FJ. Diagnosis and management of infected prosthetic aortic grafts. Surgery 1991; 110:805-813. 6. Sharp WJ, Hoballah J J , Mohan CR, et al. The management of the infected aortic prosthesis: a current decade of experience. J Vasc Surg 1994; 19:844-850. 7. Gordon A, Conlon C, Collin J , et al. An eight year experience of conservative management for aortic graft sepsis. Eur J Vasc Surg 1994; 8:611-616. 8. Haaga JR, Baldwin NG, Reich NE, et al. CT detection of infected synthetic grafts: preliminary report of a new sign. Am J Roentgen01 1978; 131:317-320. 9. Hilton S, Megibow AJ, Naidich DP,
Rossi et a1
277
Volume 8 Number 2
Bosniak MA. Computed tomography of the postoperative abdominal aorta. Radiology 1982; 145:403-407. 10. Low RN, Wall SD, Brooke JR, Sollitto RA, Reilly LM, Tierney LM. Aortoenteric fistula and perigraft infection: evaluation with CT. Radiology 1990; 175:157162. 11. Fiorani P, Speziale F, Rizzo L, et al. Detection of aortic graft infection with leukocytes labeled with technetium 99m-hexam
etazime. J Vasc Surg 1993; 17:8796. 12. Thomas WEG, Baird RN. Secondary aorto-enteric fistulae: towards a more conservative approach. Br J Surg 1986; 73:875-878. 13. Jacobs MJHM, Reul G, Gregoric I, Cooley DA. In-situ replacement and extra-anatomic bypass for the treatment of infected abdominal aortic grafts. Eur J Vasc Surg 1991; 5233-86. 14. O'Brien T, Collin J. Prosthetic
vascular graft infection. Br J Surg 1992; 79:1262-1267. 15. Bandyk DF, Bergamini TM, Kinney EV, Seabrook GR, Towne JB. In situ replacement of vascular prosthesis infected by bacterial biofilms. J Vasc Surg 1991; 13:575-583. 16. Pistolese GR, Ippoliti A, Tuccimei I, Lorido A. Conservative treatment of aortic graft infection. International Symposium on management of aortofemoral graft infection. Rome. December 1995.
Index terms: Aorta, grafts and prostheses. Grafts, infection. Grafts, interventional procedure
JVIR 1997; 8:271-277
PURPOSE: Aortic graft infection is a rare complication of aortic surgery associated with high mortality and morbidity rates. The aim of the study was to evaluate the authors' experience with diagnostic and interventional procedures in the management of aortic graft infection. METHODS: Six patients with clinical signs suggestive of graft infection were studied. Abdominal computed tomography (CT) was performed in all patients. Diagnostic work-up was performed with perigraft puncture for fluid aspiration and contrast medium injection to confirm clinical signs or CT findings of graft infection, and to detect graft-enteric fistulas. Percutaneous drainage of the periprosthetic abscess was performed as attempted therapy in four patients. RESULTS: In four patients, CT examination findings were considered positive for graft infection. Culture and cytologic examination of aspirates revealed numerous white blood cells but did not show any growth. In three patients, direct contrast medium injection into the perigraft space allowed the detection of a graft-enteric fistula. In three patients, percutaneous drainage was performed as preoperative therapy, allowing improvement of the surgical outcome; in one patient the catheter was left in place as an attempt at definitive therapy, allowing a complete recovery. CONCLUSIONS: The percutaneous approach to prosthetic graft infection permits both the drainage and aspiration of the fluid in the perigraft area for laboratory studies. Percutaneous drainage can offer the possibility of definitive cure in patients for whom surgical management is considered too risky or, at least, represents a temporizing maneuver to improve the patient's general condition before surgery.
'
From the Departments of Radiology (P.R., F.M.A., F.M.s., M.B.) and Vascular Surgery (F.S., E.S.), 111 Cattedra, University of Rome "La Sapienza," Policlinico Umberto I, V.le Regina Elena, 324, 00161 Rome, Italy. Received August 3, 1996; revision requested September 3; revision received and accepted November 5. Address correspondence to P.R.
o SCVIR, 1997
INFECTION of an aortic prosthesis is a rare but life-threatening complication of aortic reconstructive surgery. The reported incidence from 1.5%to 6%, with a high mortality rate ranging from 20% to 75% (1).Graft infections are usually limited to the prosthesis Or with a graftenteric fistula, with different clinical implications. Treatment is usually by means of surgical management with complete excision of
the infected graft and extra-anatomic bypass or in situ graft replacement without any previous drainage. However, mortality and complication rates remain high (2). The clinical signs of prosthetic graft infection may be subtle and progress slowly, rendering early diagnosis difficult. Therefore, clinical suspicion and a thorough diagnostic work-up are necessary to make a prompt diagnosis and improve the prognosis.
271
272
Journal of Vascular and Interventional Radiology
March-April 1997
Radiology offers many imaging modalities with which to achieve a diagnosis, such as computed tomography (CT), ultrasonography (US), magnetic resonance (MR), and nuclear medicine. CT and MR are the most suitable techniques for detecting perigraft fluid collections. Moreover, CT-guided periaortic fluid as~irationand contrast medium injection in the periprosthetic space, as an attempt to visualize a graft-enteric fistula, can provide a more accurate diagnosis. Scintigraphy may reveal a graft infection as an abnormal accumulation of radionuclides. Angiography usually is not needed, but it can add information about the vascular tree preoperatively. The aim of our study was to evaluate our experience in the diagnosis and management of prosthetic graft infection by using radiologic procedures.
1 MATERIALS AND MIITHODS From 1987 to 1996 we studied six patients who had previously undergone aortic grafting for aneurysmal disease and had clinical signs suggestive of prosthetic graft infection, with a time to presentation after the primary vascular procedure ranging from 1 to 9 years (mean, 3.6 years). All patients were men, ranging in age from 39 to 86 years. The clinical presentation was characterized in all patients by aspecific symptoms, such as fever, leukocytosis, increase of erythrocyte sedimentation rate, and abdominal or back pain. More specific signs of prosthetic graft infection, such as peripheral septic emboli and gastrointestinal bleeding, were present in one patient only. All patients underwent CT examination performed with various equipment (DRH Siemens, Erlangen, Germany, or GE Pace; GE Medical Systems, Milwaukee, WI). Contiguous 5-mm-thick sections of the abdomen, proximal and distal to the aortic graft, were obtained during intravenous injection of 150 mL of contrast material, continu-
ously or in multiple boluses a t a rate of 3.0 mL/sec. When a perigraft puncture was being considered, no contrast agent was administered orally before CT performed with the patient in the prone position. Thickened graft wall and perigraft fluid collection were considered indirect signs, whereas gas content in the periprosthetic space was considered evidence of aortic graft infection at CT. Technetium-99m scanning was performed in two patients only. Leukocytes were labeled with 200 MBq Tc-99m HMPAO, and images were obtained with a gamma camera with a large field-of-view at 2-3 hours after administration of the labeled leukocytes. Tc-99m scanning was considered to be positive for infection when increased uptake was seen at the aortic graft. All interventional procedures for detection and management of prosthetic graft infections were performed with use of a 18-gauge needle with a Teflon sheath, a 0.035-inch guide wire (Cook, Bloomington, IN), and a 10-12-F catheter (Medi-techJBoston Scientific, Watertown, MA) when abscess drainage was necessary. All patients underwent perigraft puncture: with CT guidance in five patients, a translumbar approach in four patients, and an anterior approach in one patient. In the sixth patient, the perigraft puncture was performed under US guidance by means of a transperitoneal approach. Perigraft fluid was aspirated to reach a clear diagnosis of infection based on culturing and cytologic examination of the aspirate by identifying either the infectious organisms or the presence of white blood cells. Perigraft puncture and contrast medium injection in the graft bed were performed to evaluate the presence of graft-enteric fistula. The spread of iodinated contrast material from the periaortic lumen into the bowel was considered clear evidence of graft-enteric fistula. Endoscopy was performed in three patients to compare radiologic findings with eventual mucosal abnormalities.
Percutaneous drainage was performed in four patients in whom the diagnosis of prosthetic graft infection was certain: as preoperative therapy in three patients and as definitive therapy in the fourth patient. In the other two patients, drainage was not feasible because, in one patient, the rim of fluid around the graft was too small and, in the other patient, surgery was scheduled after the diagnostic puncture. Each individual patient's history, management, and follow-up can be easily described, as follows: Case 1.-A 72-year-old man developed abdominal and back pain radiating to the left leg, with fever, leukocytosis, and a rise in erythrocyte sedimentation rate. He had undergone aorto-iliac Dacron grafting 2 years earlier. CT findings (Fig la) were considered indicative of prosthetic graft infection because of the presence of an air-fluid level surrounding the graft. A CT-guided perigraft puncture with direct contrast medium injection was performed (Fig lb), which allowed detection of a fistula between the infected graft and the jejunum (Fig lc, Id). The catheter was left in place for approximately 6 months as an attempt a t definitive therapy because surgical management was considered too risky due to the patient's poor cardiovascular condition (Fig le, If). The patient recovered completely, with disappearance of the periprosthetic abscess, and did well for approximately 2 years, when he underwent CT examination because of fever and leukocytosis. CT showed that the prosthetic graft was not involved: both the graft and the retroperitoneal space were normal (Fig lg). The patient died 2 months later of sepsis of unknown origin. Case 2.-A 56-year-old man, with a history of aortic valve replacement, began to feel ill and developed fever, back pain, and leukocytosis 2 years after aortic Dacron grafting. CT of the abdomen showed a thickened graft wall with an air-fluid level and gas content. Under CT guidance, a perigraft puncture with direct contrast
Rossi et a1
273
Volume 8 Number 2
b.
a.
Figure 1. (a)CT scan of the abdomen in a 72-year-old man. Two years after aortic grafting for aneurysm, he developed pain in the left leg and fever. An abscess was seen along the left limb of the aortic graft with gas bubbles. (b) Translumbar approach to the lesion few days later, with the patient in the right lateral decubitus position. ( c ) Contrast material injection into the lesion shows the abscess cavity and a fistulous communication with the small bowel (Fig 1 continues).
C.
medium injection was performed and showed the spread of contrast agent from the perigraft lumen into the third portion of the duodenum, revealing the presence of a graftenteric fistula not visualized a t a previously performed endoscopy (Fig Za, 2b). Percutaneous drainage was not performed because of the small rim of fluid around the graft. Specific antibiotic therapy was given before surgery. However, the patient died of cerebral hemorrhage, probably related to the anticoagulation therapy. Case 3.-A 64-year-old man was admitted as a n emergency case 5 years after aortic Dacron grafting. The clinical presentation was characterized by gastrointestinal bleeding, peripheral septic emboli, leukocytosis, and fever. CT findings were
considered unreliable with respect to graft infection because only a small rim of soft-tissue attenuation surrounded the graft, with no evidence of frank fluid or gas (Fig 3). At endoscopy, no graft-enteric fistula was visualized. The patient underwent CT-guided puncture for diagnostic aspiration and direct contrast medium injection. Culture of the aspirate showed the presence of numerous white blood cells but revealed no growth, probably related to 2 months of antibiotic therapy. No fistula was seen during contrast medium injection. The patient refused surgery because his clinical condition was improving. However. he returned 5 months later with a severe infection and underwent surgery a s a n emergency procedure when complete excision of
the infected graft and in situ replacement was performed. No fistula was found a t surgery, only inflammatory adhesion with the duodenum. The patient is still alive and well a t 1.5-year follow-up. Case 4.-A 61-year-old man was admitted for evaluation because of clinical suspicion of a prosthetic graft infection 3 years after aortoiliac Dacron grafting. CT showed a perigraft fluid collection with a n air-fluid level. Direct contrast medium injection showed no graftenteric fistula. Percutaneous drainage was performed under CT guidance. The catheter was left in place for 7 days, allowing for stabilization of the patient who subsequently underwent successful surgical management with complete excision of the infected graft and construction
274
Journal of Vascular and Interventional Radiology
March-April 1997
Figure 1 (continued). (d) Simultaneous opacification of the abscess cavity and the small bowel after contrast material injection. (e) Four months later, there is still a residual cavity no longer communicating with the small bowel. The patient was afebrile and doing well (Fig 1 continues).
of an extra-anatomic bypass. He was subsequently lost to follow-up. Case 5.-A 39-year-old man developed fever and general malaise 1year after aortic polytetrafluoroethyene grafting for aneurysmal disease, with a blood pressure of 80150 mm Hg. Abdominal CT showed a thickened graft wall and a perigraft fluid collection without gas content. The patient underwent Tc-99m imaging, which showed an increased uptake of labeled leukocytes a t the aortic graft. Periaortic fluid was aspirated and the culture
showed numerous white blood cells. No fistula was seen during contrast material injection. US-guided percutaneous drainage was carried out before surgical treatment. Total removal of the infected graft with in situ replacement was then performed, and the patient had no major complications at a 5-year follow-up. Case 6.-An 86-year-old man was admitted with a 6-month history of recurrent septic fever 9 years after aortic Dacron grafting. CT confirmed the clinical suspicion of prosthetic graft infection by revealing the presence of a large perigraft fluid collection with an enhancing irregular wall consistent with " manulation tissue. He underwent Tc-99m scanning, which showed an increased uptake of labeled leukocytes at the aortic graft. Endoscopy did not allow the visualization of a graft-enteric fistula, but revealed the presence of pus in the duodenum. A CT-guided perigraft puncture with direct contrast medium injection was also performed and a secondary fistula between the infected graft and the duodenum was detected. Because of the severity of the infection and the patient's age and poor general condition, a percutaneous drainage was left in place as preoperative therapy. Two weeks later, the patient underwent emergency surgery for active, massive gastrointestinal hemorrhage. At surgery, active bleeding into the duodenum was found at the site of the proximal anastomosis, and partial removal of the infected graft was performed. Because of clinical signs suggestive of pulmonary emboli, a vena cava filter was subsequently placed. The patient is still alive and well at 6-month follow-up.
1 RESULTS Of the six patients clinically suspected of having prosthetic graft infection who underwent CT examination, five had positive CT findings consisting of a perigraft fluid collection with or without gas content. In two of these five pa-
tients, a Tc-99m radionuclide scan showed an increased uptake of labeled leukocytes at the aortic graft, consistent with the presence of infection. In the sixth patient, in whom only a small rim of softtissue attenuation surrounded the graft, with no evidence of fluid or gas content, the CT findings were considered suggestive of graft infection but not completely reliable. Perigraft fluid, after lavage, was aspirated and collected to allow culture and cytologic examinations that revealed the presence of numerous white blood cells but did not show any growth. The patient refused both nuclear medicine scanning and surgical treatment. The presence of a graft-enteric fistula, suggested on CT scans by adhesions between the graft and the enteric loops, was easily proved in three patients with radiologic procedures. Perigraft puncture and direct contrast medium injection under CT guidance were performed, allowing graft-enteric fistula to be correctly diagnosed by the spread of contrast agent from the perigraft lumen into the bowel. Endoscopy was performed only in two patients; in one patient the finding was negative, and in the other, intraduodenal pus was found. In one patient, there was no extravasation of iodinated contrast material into the bowel but the patient had a graft-enteric fistula found at surgery; endoscopy was negative as well. The last two patients did not undergo endoscopic examination. Percutaneous drainage was performed as a temporizing treatment before surgery in three patients to improve the their general condition and was then followed by surgical management. In one patient the drain was left in place for 6 months as attempted definitive therapy because surgery was not indicated due to the high cardiovascular risk.
Despite improved surgical procedures, aortic graft infection remains
Rossi et a1
275
Volume 8 Number 2
f. Figure 1 (continued).
g.
(f) The last control image obtained before the catheter was removed, approximately 6 months after the beginning of the drainage, shows a small tract t h a t closed spontaneously. ( g ) CT scan of the abdomen 2 years after removal of the catheter when t h e patient returned for sepsis.
a. b. Figure 2. (a, b) CT scan of the abdomen showed fluid collection i n the periaortic graft area. A translumbar puncture of t h e periaortic area was made; after aspiration of t h e fluid, contrast material was injected, which visualized the periaortic area and the third portion of t h e duodenum.
a rare but serious complication associated with high mortality and complication rates. Surgical management consists of total removal of the infected graft and performance of a n extra-anatomic bypass or in situ graft replacement (2-5). The correct management of patients with minimal clinical signs and low-grade infection is less clear; in
these cases more conservative surgery such as partial graft excision, local debridement, and antibiotic irrigation can be performed. These procedures may be equally effective and do not carry the high complication rates associated with total excision (6,7). Many imaging modalities are available for the detection of infection
of an aortic prosthesis (ie, US, CT, MR, scintigraphy, and, less frequently, angiography) (8-10). Various scintigraphy techniques have been used to detect prosthetic graft infections; Tc-99m scanning has been recently proposed as a noninvasive technique for the diagnosis of lowgrade graft infection in patients with minimal clinical signs (11).
276
Journal of Vascular and Interventional Radiology -
-
-
March-April 1997
Figure 3. CT scan of the abdomen showed a small rim of soft-tissue attenution with no evidence of fluid or gas content.
We consider CT to be the diagnostic method of choice. So far, CT is the best and safest way of performing interventional procedures such as periaortic fluid aspiration, contrast medium injection, and drainage. In selected cases with large collections extending toward the anterior abdominal wall, a safe route for puncture also can be found with US, and the drainage be performed under such guidance. In all six patients, the prosthetic graft infection was first suspected clinically; all of the patients were studied with CT to evaluate the possibility of graft involvement. The presence of a perigraft fluid collection and gas content was suggestive of infection, while the presence of a minimally thickened graft wall, in default of other findings, necessitated further examinations. In patients in whom CT findings were not definitive for infection, we performed percutaneous needle aspiration. The perigraft puncture was preferably performed with use of the same technique as is used for translumbar angiography, with the patient in the prone position to avoid both the bowel and the native aortic wall. Anterior transperitoneal puncture can also be performed. Examination and culture of the aspirates can be useful in assessing the presence of white blood cells and in determining the causative organism. Unfortunately, as was the case in our series, the culture may result sterile, most likely because of antibiotic therapy. In
~ a t i e n t sin whom there was a fluid collection, we proceeded to drainage, using the same approach as was described earlier for the diagnostic puncture. The drainage material was sent for culture analysis as well. The presence of a graft-enteric fistula is usually suspected on a clinical basis. Although endoscopy is important for excluding other causes of gastrointestinal bleeding, it often fails to detect a graftenteric fistula (12). In our experience, CT-guided perigraft puncture with injection of contrast medium into the graft bed allowed, in three patients, a correct diagnosis of a graft-enteric fistula by showing the spread of contrast material from the perigraft lumen into the bowel. We recommend avoiding oral administration of contrast agent for CT whenever this technique is employed, to facilitate detection of bowel opacification. The choice among different surgical procedures in the management of prosthetic graft infection may depend on the severity of the infection (13). Total graft removal and extra-anatomic bypass are considered mandatory in patients with severe prosthetic infection (2-5), even though the mortality and morbidity are high. Aortic stump disruption remains a catastrophic event after complete excision of the infected prosthesis (2,14). In situ graft replacement has been proposed for the treatment of selected patients with a low-grade infection in whom a less aggressive approach may improve the surgical outcome (13,15). In our patients, interventional procedures have proved valuable in the diagnosis and management of prosthetic graft infections. Percutaneous needle aspiration under CT guidance has provided a safe way of diagnosing infection in patients with unclear clinical signs and ambiguous CT findings. Moreover, we consider CT-guided perigraft puncture with direct contrast agent injection to be a suitable radiologic procedure for detecting graft-enteric fistulas.
Percutaneous treatment of infected prostheses, recently proposed as an alternative conservative aDproach in selected patients (161, is not usually considered by surgeons to be a safe procedure. Nevertheless, we suggest that percutaneous drainage of perigraft fluid collections be employed as a preoperative measure to "cool off' the infection, or in any patient for whom surgical management is considered too risky because of associated heart, lung, or kidney disease, or because of advanced age. Moreover, we speculate that preoperative percutaneous drainage, by ameliorating the patient's general condition, may help to improve the surgical outcome in patients with severe infection. References 1. Buckels JAC, Wilson SE. Prevention and management of prosthetic graft infection. In: Veith FJ, Hobson RW 11, Williams RA, Wilson SE, eds. Vascular surgery. 2nd ed. New York: McGraw-Hill, 1994; 10811089. 2. Reilly LM, Altman H, Lusby RJ, Kersh RA, Erhenfeld WK, Stoney RJ. Late results following surgical management of vascular graft infection. J Vasc Surg 1984; 1:36-44. 3. Trout H 111, Kozloff L, Giordano JM. Priority of revascularization in patients with graft enteric fistulas, infected arteries, or infected arterial prosthesis. Ann Surg 1984; 199:669-683. 4. Edwards MJ, Richardson JD, Klamer TW. Management of aortic prosthetic infections. Am J Surg 1988; 155:327-330. 5. Calligaro KD, Veith FJ. Diagnosis and management of infected prosthetic aortic grafts. Surgery 1991; 110:805-813. 6. Sharp WJ, Hoballah J J , Mohan CR, et al. The management of the infected aortic prosthesis: a current decade of experience. J Vasc Surg 1994; 19:844-850. 7. Gordon A, Conlon C, Collin J , et al. An eight year experience of conservative management for aortic graft sepsis. Eur J Vasc Surg 1994; 8:611-616. 8. Haaga JR, Baldwin NG, Reich NE, et al. CT detection of infected synthetic grafts: preliminary report of a new sign. Am J Roentgen01 1978; 131:317-320. 9. Hilton S, Megibow AJ, Naidich DP,
Rossi et a1
277
Volume 8 Number 2
Bosniak MA. Computed tomography of the postoperative abdominal aorta. Radiology 1982; 145:403-407. 10. Low RN, Wall SD, Brooke JR, Sollitto RA, Reilly LM, Tierney LM. Aortoenteric fistula and perigraft infection: evaluation with CT. Radiology 1990; 175:157162. 11. Fiorani P, Speziale F, Rizzo L, et al. Detection of aortic graft infection with leukocytes labeled with technetium 99m-hexam
etazime. J Vasc Surg 1993; 17:8796. 12. Thomas WEG, Baird RN. Secondary aorto-enteric fistulae: towards a more conservative approach. Br J Surg 1986; 73:875-878. 13. Jacobs MJHM, Reul G, Gregoric I, Cooley DA. In-situ replacement and extra-anatomic bypass for the treatment of infected abdominal aortic grafts. Eur J Vasc Surg 1991; 5233-86. 14. O'Brien T, Collin J. Prosthetic
vascular graft infection. Br J Surg 1992; 79:1262-1267. 15. Bandyk DF, Bergamini TM, Kinney EV, Seabrook GR, Towne JB. In situ replacement of vascular prosthesis infected by bacterial biofilms. J Vasc Surg 1991; 13:575-583. 16. Pistolese GR, Ippoliti A, Tuccimei I, Lorido A. Conservative treatment of aortic graft infection. International Symposium on management of aortofemoral graft infection. Rome. December 1995.