- Email: [email protected]
Enterococcal Bacteremia After Transjugular Intrahepatic Portosystemic Shunts (TIPS)
THE AMERICAN JOURNAL OF GASTROENTEROLOGY Copyright © 1998 by Am. Coll. of Gastroenterology Published by Elsevier Science Inc.
Vol. 93, No. 4, 1998 ISSN 0002-9270/98/$19.00 PII S0002-9270(98)00057-4
Case reports Enterococcal Bacteremia After Transjugular Intrahepatic Portosystemic Shunts (TIPS) Robert S. Brown, Jr., M.D., M.P.H., Lynne Brumage, M.D., Hal F. Yee, Jr., M.D., PhD., John R. Lake, M.D., John P. Roberts, M.D., and Kenneth A. Somberg, M.D. Department of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina; and Departments of Medicine, and Surgery, University of California, San Francisco, California
The objective of this study was to analyze a series of patients with Enterococcus faecium infection following transjugular intrahepatic portosystemic shunts (TIPS) in order to define the risk factors, outcome, and role of treatment including hepatic transplantation. This study is a case series from a tertiary referral center for liver transplantation. The medical records of four patients referred to one teaching hospital in San Francisco between 1990 and 1995 for evaluation or management of Enterococcal infection following TIPS were reviewed. A review of the microbiology records of all 314 patients who underwent TIPS at that institution and a MEDLINE search were performed to assess whether any other cases existed. The effect of therapy on survival was assessed, in particular, the repeated use of TIPS and prolonged courses of antibiotics. All four patients had thrombosis of their TIPS at the time of diagnosis of enterococcal bacteremia. All were treated with prolonged courses of intravenous antibiotics. One patient had echocardiographic evidence of subacute bacterial endocarditis with chronic aortic insufficiency. In all cases, liver transplantation was contraindicated in the acute setting because of uncontrolled endovascular infection. Two of four patients survived; these were the only two patients who had had a successful repeat TIPS. Enterococcal bacteremia is a rare complication following TIPS but carries a high mortality. It usually occurs in the setting of technically difficult TIPS with shunt thrombosis. Management should be focused on long term antibiotics and attempts at reestablishment of portal decompression with another TIPS. Liver transplantation should not be considered until the infection is cleared. Prophylaxis for Enterococcus species should be considered in technically difficult or unsuccessful TIPS.
(Am J Gastroenterol 1998;93:636 – 639. © 1998 by Am. Coll. of Gastroenterology)
INTRODUCTION Transjugular intrahepatic portosystemic shunts (TIPS) represent an increasingly used therapy for portal hypertension in patients with cirrhosis. An alternative to the surgical portosystemic shunt, this technique involves the placement of an intrahepatic expandable metal stent between the portal and systemic circulation (1, 2). Early results with transjugular intrahepatic portosystemic shunts (TIPS) suggest that it is effective at lowering portal pressure and acutely controlling variceal bleeding, and can also be effective therapy for ascites that is refractory to diuretic therapy (3, 4). TIPS is also associated with a low rate of procedural complications (1). Complications of TIPS include encephalopathy in ;25% of cases (5, 6), bleeding in ;5% (1, 7), and deterioration of liver function in 0 –7% (1, 8). Infectious complications of TIPS are uncommon. Fever and/or bacteremia have been reported in 1–10% of patients in several large series (7). Most fevers have been self-limited, and the few cases of positive blood cultures have typically involved Staphylococcus aureus or Gram-negative organisms. Enterococcus faecalis is an uncommon organism in the biliary tree and upper digestive tract. It is responsible for 5–10% of episodes of cholangitis (9, 10) and 0 –10% of episodes of spontaneous bacterial peritonitis (11–14). E. Faecalis could be expected to be a potential cause of infection following TIPS, inasmuch as these infections are presumed to be secondary to seeding of the blood from the biliary tree and third generation cephalosporins, which are often used as prophylaxis, do not cover Enterococcus species. However, enterococcal infections following TIPS have never been reported. We report four cases of Enterococcus infection following TIPS, half of which were fatal, and
Received Mar. 21, 1997; accepted Jan. 6, 1998. 636
AJG – April 1998
ENTEROCOCCAL BACTEREMIA
637
discuss the implications for procedural prophylaxis and therapy.
tibiotics and comfort care, and died 2 days following transfer. Autopsy was not performed.
CASE 1
CASE 2
Patient 1 was a 45-yr-old man with alcoholic cirrhosis and a history of alcoholic hepatitis who presented to an another institution following two episodes of upper GI bleeding from portal hypertensive gastropathy despite b-blocker therapy. Admission laboratory values included a hematocrit of 26%, ALT 21 mg/dl, alkaline phosphatase 321 mg/dl, total bilirubin 9.2 mg/dl, prothrombin time (PT) 5 16.6 s, and albumin 2.5 g/dl. Physical examination was significant for ascites, peripheral edema, jaundice, a II/VI systolic murmur and no evidence of asterixis. Upper endoscopy revealed 11 esophageal varices and marked portal hypertensive gastropathy without acute bleeding noted. Ultrasound revealed flow in the left and right portal veins but no definite flow was seen in the main portal vein. The patient underwent a TIPS. At angiography, the portal vein was patent; however, several small filling defects were seen near the portal vein end of the shunt, which were thought to represent thrombi. A large esophageal varix was embolized with 5-mm coils. Because of the question of portal vein thrombus, the patient underwent repeat portal angiography the following day, which revealed no evidence of thrombus. The TIPS was patent with a portosystemic gradient of 10 mm Hg. The patient was discharged on neomycin, lactulose, Aldactone, and Lasix. The patient was initially well after TIPS, with no further bleeding and improved ascites but with mildly worsened encephalopathy. Seven months following TIPS he was admitted with melena and confusion. He underwent upper endoscopy, which showed blood oozing from the duodenal bulb and proximal stomach, both of which were cauterized. Blood cultures were positive for Enterococcus in four of four bottles. The patient was transferred to UCSF for transplant evaluation. There was no history of intravenous drug use and it was unclear whether he had been having fevers at home. On exam he had anasarca, jaundice, and a III/VI systolic with a II/VI diastolic murmur at the left sternal border radiating throughout the precordium. Laboratory tests revealed white count 17.1 K, hematocrit 22.7%, platelets 46 K, BUN 84 mg/dl, creatinine 2.0 mg/dl, PT 18.7 s, and albumin 1.2 g/dl. Repeat upper endoscopy showed no varices nor evidence of bleeding. Ultrasound with Doppler revealed a patent TIPS with a velocity of 1.8 m/s. Transesophageal echo showed a vegetation on the aortic valve and wide open aortic insufficiency. The echocardiographic findings and the degree of compensation of his aortic insufficiency supported a diagnosis of subacute bacterial endocarditis (SBE). Treatment was initiated with ampicillin and gentamycin. Valve replacement was contraindicated based on his severe liver disease. Liver transplantation was not possible because of his endovascular infection and cardiac status. He had recurrent melena and refused endoscopy. He was transferred back to his primary care physician for an-
Patient 2 was a 57-yr-old man with end stage liver disease secondary to alcohol. He had a TIPS placed at another institution for recurrent episodes of variceal bleeding despite sclerotherapy, and for refractory ascites. Angiographically, the stent clotted immediately. Urokinase was infused without success. The thrombosis included the portal vein, superior mesenteric vein, and the splenic vein. Postprocedure he developed fevers, and blood cultures were positive for Enterococcus faecium and Citrobacter amalonaticus diversus. He was treated with intravenous vancomycin and gentamycin for 2 wk and discharged without fevers. One week later he was readmitted with fevers, lethargy, and progressive renal insufficiency (creatinine 3.9 mg/dl versus 1.2 mg/dl before TIPS). Blood cultures were once again positive for Enterococcus and Citrobacter. The former was sensitive to vancomycin and high dose streptomycin; the latter was sensitive to most antibiotics. The patient was treated with vancomycin, high dose streptomycin, and ceftriaxone. A Hickman catheter was placed with a plan for 5 wk of antibiotic therapy. His renal insufficiency resolved off gentamycin and he was discharged 1 wk later. Nine days postdischarge he developed melena without hemodynamic instability. His hematocrit was 29%. Ultrasound of his liver with Doppler analysis revealed no flow through the TIPS. Upper endoscopy revealed two oozing bulbar duodenal ulcers (one with a visible vessel), which were treated with heater probe cautery and epinephrine injection with good hemostasis. He had nonbleeding grade I–II esophageal varices and gastric varices. He was sent to UCSF for evaluation for liver transplantation to relieve his portal hypertension and to remove his infected TIPS. Transplantation was not possible because of his concurrent endovascular infection, and he was sent home for prolonged intravenous antibiotics with a plan to return for transplant evaluation following 4 – 6 wk of intravenous antibiotics. He died at home 2 wk after discharge. Autopsy was not performed. CASE 3 Patient 3 is a 66-yr-old man with alcoholic cirrhosis who had a TIPS placed at UCSF in June 1994 for recurrent massive bleeding from gastric varices. Two months later, flow through the TIPS was noted to be decreased (0.1 meters/s) on routine ultrasound. The patient was admitted and a venogram revealed nonocclusive thrombus within the TIPS shunt. Stent revision and parallel Wallstent placement were unsuccessful on two occasions. The patient was discharged after unsuccessful TIPS revisions. Two days after discharge, the patient was admitted for a massive variceal bleed. Physical exam was significant for jaundice, massive ascites, and melena on rectal exam. He was lethargic, but
638
BROWN et al.
asterixis was not present. Laboratory exams included a white count of 10 K, a bilirubin of 3.8 mg/dl, and a normal BUN and creatinine. Three further attempts at TIPS revision were made; the third resulted in successful placement of a parallel stent. The original TIPS was noted to be totally occluded at that time. Blood cultures at admission were positive for Enterococcus, and the patient was started on vancomycin and then changed to high dose ampicillin after several days. This isolate had relatively high resistance to aminoglycosides. Echocardiogram revealed no evidence of vegetations. The patient received a total course of 14 days of intravenous ampicillin. The patient has had no fevers or other evidence of bacteremia, nor has he had any further variceal bleeding with 1 yr of follow-up. CASE 4 Patient 4 is a 43-yr-old man with cirrhosis secondary to alcohol abuse and hepatitis C, who presented to an outside hospital with an acute upper GI bleed. At that time, endoscopy revealed two duodenal ulcers, distal esophageal ulceration, and portal hypertensive gastropathy. Two months later, the patient had recurrence of bleeding and underwent multiple sessions of sclerotherapy, which were not successful in controlling his recurrent portal hypertensive bleeding. He had a TIPS placed at that time, which was complicated by Streptococcus mitis and Staphylococcus epidermidis septicemia which was treated with vancomycin. He was readmitted 10 days later with hematemesis. Endoscopy showed large esophageal varices and a blood clot at the gastroesophageal junction. A TIPS venogram revealed thrombosis of the portal vein with extension of thrombus into the main portal trunk down to the confluence of the inferior mesenteric vein (IMV), superior mesenteric vein (SMV), and splenic veins. Multiple attempts by interventional radiology to open the shunt were unsuccessful. The patient was then transferred to UCSF for transplant evaluation. There was no history of injection drug use. On exam he was awake and alert and did not show any evidence of encephalopathy. His abdomen was nontender, obese, without obvious ascites. He had a hematocrit of 24.0, white count 8.0, normal prothrombin time and transaminases, and a bilirubin of 3.3. Repeat endoscopy showed marked portal hypertension, gastropathy, small gastric varices, and 21 esophageal varices with two ulcers at the GE junction. These ulcers were clean-based and consistent with previous sclerotherapy treatments. A repeat venogram was done, which was consistent with the findings at the outside hospital. The patient then underwent a TIPS revision and stenting through a portal clot. Blood cultures from the day of admission grew enterococcus from two of four bottles and the patient was started on vancomycin and gentamycin. The patient underwent an ultrasound after TIPS revision, which showed patent TIPS with maximal velocity of 1.2 meters/s, hepatofugal flow in the right and left portal veins, and a patent hepatic vein. The patient was determined not to be a
AJG – Vol. 93, No. 4, 1998 transplant candidate secondary to his continued alcohol consumption. He was then transferred back to the referring hospital for continued care. He completed a 4-wk course of antibiotics for his bacteremia. Subsequently, he has done well since his hospitalization. He is currently working and enjoying a good quality of life. DISCUSSION We present four cases of endovascular infection with E. faecalis following TIPS. Each case was associated with some degree of portal vein or shunt thrombosis. This is a potentially important point, as bile duct injury, which is more common with technically difficult shunt procedures and thrombosis, has been associated with shunt stenosis (15, 16) and would also be a possible source of enterococcal seeding of the TIPS. Although it is impossible to unequivocally implicate the TIPS as the cause of bacteremia (particularly in patient 1, for whom the time interval was long between the placement of the TIPS and the positive cultures), there were no other risk factors for enterococcal bacteremia. Specifically, none of the patients had a history of intravenous drug use, recent dental procedures, or other parenteral risk factors. In addition, none of the patients were found to have cholecystitis, cholangitis, spontaneous bacterial peritonitis (SBP), or positive ascites cultures for enterococcus. Thus, the clinical opinion of the hepatologists, surgeons, and interventional radiologists caring for the patient was that the bacteremia was likely related to the TIPS procedure. Following TIPS, three of the four patients presented with bleeding and were found to have positive blood cultures and clotted TIPS. The fourth patient had fevers following thrombosis of his TIPS. Chronic infective endocarditis was present in one patient, and the deaths of two of four patients were related to persistent infection. It is interesting to note that both patients who were successfully treated were the only two to have had a successful repeat TIPS procedure following thrombosis and occlusion of the original shunt. Whether the lack of portal decompression played a role in the poor outcome of the first two patients, or whether the inability to successfully place another TIPS was a reflection of a more extensive septic thrombosis, is uncertain. Whether these cases could have been prevented with a prophylactic regimen with activity against Enterococcus (i.e., vancomycin or ampicillin), in addition to a third generation cephalosporin, is also unknown. Given the rarity of this infection, which occurred in one of 314 TIPS performed at our institution, it will be difficult (if not impossible) to study. However, prophylaxis with ampicillin or vancomycin might be warranted in technically complicated TIPS that involve a large numbers of passes, shunt or portal vein thrombosis, or biliary opacification. In this scenario, the severity of the complication must be weighed against the possibility of the emergence and prevalence of vancomycin-resistant Enterococcus species. Vancomycin prophylaxis in patients under-
AJG – April 1998
ENTEROCOCCAL BACTEREMIA
going cardiac and vascular surgery has been shown to reduce postoperative infections with enterococcal species (17), a rare complication following other vascular procedures. Enterococcal vascular graft infections are usually treated by surgical removal and antibiotics with good results (18, 19); however a case report of E. faecium infection following aortobifemoral bypass that was not removed resulted in the death of the patient (20). In the case of TIPS, surgical removal is not possible without transplantation, and transplantation is usually considered to be contraindicated in the presence of an uncontrolled endovascular infection. In the case of a documented infection with enterococcus, echocardiography should be considered to rule out valvular vegetations that would require prolonged courses of intravenous antibiotics. The patient should not be viewed as a candidate for liver transplantation until negative blood cultures have been documented off antibiotics. Finally, intravascular infection should be considered in the differential diagnosis when a patient presents with recurrent bleeding and shunt thrombosis following TIPS, particularly in the case of technical difficulty or thrombosis associated with the initial shunt procedure, or in the presence of fevers or new cardiac murmurs. ACKNOWLEDGMENT The authors acknowledge the assistance of Dr. G. Sandhu in reviewing and revising the manuscript. Dr. Brown is supported by a grant from the Glaxo Institute for Digestive Health. Reprint requests and correspondence: Robert S. Brown, Jr., M.D., M.P.H., Division of Digestive Diseases, 708 Burnett Womack Building, Box 7080, University of North Carolina, Chapel Hill, NC 27599-7080.
REFERENCES 1. LaBerge JM, Ring EJ, Gordon RL, et al. Creation of transjugular intrahepatic portosystemic shunts with the wallstent endoprosthesis: Results in 100 patients. Radiology 1993;187:413–20. 2. Rossle M, Haag K, Ochs A, et al. The transjugular intrahepatic portosystemic stent-shunt procedure for variceal bleeding [see comments]. N Engl J Med 1994;330:165–71.
639
3. Ochs A, Rossle M, Haag K, et al. The transjugular intrahepatic portosystemic stent-shunt procedure for refractory ascites. N Engl J Med 1995;332:1192–7. 4. Somberg KA, Lake JR, Tomlanovich SJ, et al. Transjugular intrahepatic portosystemic shunts for refractory ascites: Assessment of clinical and hormonal response and renal function. Hepatology 1995;21: 709 –16. 5. Sanyal AJ, Freedman AM, Shiffman ML, et al. Portosystemic encephalopathy after transjugular intrahepatic portosystemic shunt: Results of a prospective controlled study. Hepatology 1994;20:46 –55. 6. Somberg KA, Riegler JL, LaBerge JM, et al. Hepatic encephalopathy after transjugular intrahepatic portosystemic shunts: incidence and risk factors [see comments]. Am J Gastroenterol 1995;90:549 –55. 7. Freedman AM, Sanyal AJ, Tisnado J, et al. Complications of transjugular intrahepatic portosystemic shunt: A comprehensive review. Radiographics 1993;13:1185–210. 8. Noeldge G, Richter GM, Roessle M, et al. Morphologic and clinical results of the transjugular intrahepatic portosystemic stent-shunt (TIPSS). Cardiovasc Intervent Radiol 1992;15:342– 8. 9. Boey JH, Way LW. Acute cholangitis. Ann Surg 1980;191:264 –70. 10. Shimada H, Nakagawara G, Kobayashi M, et al. Pathogenesis and clinical features of acute cholangitis accompanied by shock. Jpn J Surg 1984;14:269 –77. 11. Akriviadis EA, Runyon BA. Utility of an algorithm in differentiating spontaneous from secondary bacterial peritonitis [see comments]. Gastroenterology 1990;98:127–33. 12. Runyon BA. Monomicrobial nonneutrocytic bacterascites: A variant of spontaneous bacterial peritonitis. Hepatology 1990;12:710 –5. 13. Runyon BA, Canawati HN, Akriviadis EA. Optimization of ascitic fluid culture technique. Gastroenterology 1988;95:1351–5. 14. Runyon BA, Umland ET, Merlin T. Inoculation of blood culture bottles with ascitic fluid. Improved detection of spontaneous bacterial peritonitis. Arch Intern Med 1987;147:73–5. 15. Benner K, Saxon R, Flora K, et al. The achilles heel of TIPS, parenchymal tract stenosis and occlusion, is caused by biliary-TIPS fistulae. Hepatology 1995;22:297A. 16. LaBerge JM, Ferrell LD, Ring EJ, et al. Histopathologic study of stenotic and occluded transjugular intrahepatic portosystemic shunts. J Vasc Interv Radiol 1993;4:779 – 86. 17. Maki DG, Bohn MJ, Stolz SM, et al. Comparative study of cefazolin, cefamandole, and vancomycin for surgical prophylaxis in cardiac and vascular operations. A double-blind randomized trial. J Thorac Cardiovasc Surg 1992;104:1423–34. 18. Soyer R, Bessou JP, Bouchart F, et al. Surgical treatment of infected composite graft after replacement of ascending aorta. Ann Thorac Surg 1994;58:425– 8. 19. Baxter BT, Mesh CL, McGee GS, et al. Limb-threatening ischemia complicated by perigenicular infection. J Surg Res 1993;54:163–7. 20. Chirurgi VA, Oster SE, Goldberg AA, et al. Endovascular infections caused by enterococci highly resistant to ampicillin. South Med J 1991;84:1435–7.
Vol. 93, No. 4, 1998 ISSN 0002-9270/98/$19.00 PII S0002-9270(98)00057-4
Case reports Enterococcal Bacteremia After Transjugular Intrahepatic Portosystemic Shunts (TIPS) Robert S. Brown, Jr., M.D., M.P.H., Lynne Brumage, M.D., Hal F. Yee, Jr., M.D., PhD., John R. Lake, M.D., John P. Roberts, M.D., and Kenneth A. Somberg, M.D. Department of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina; and Departments of Medicine, and Surgery, University of California, San Francisco, California
The objective of this study was to analyze a series of patients with Enterococcus faecium infection following transjugular intrahepatic portosystemic shunts (TIPS) in order to define the risk factors, outcome, and role of treatment including hepatic transplantation. This study is a case series from a tertiary referral center for liver transplantation. The medical records of four patients referred to one teaching hospital in San Francisco between 1990 and 1995 for evaluation or management of Enterococcal infection following TIPS were reviewed. A review of the microbiology records of all 314 patients who underwent TIPS at that institution and a MEDLINE search were performed to assess whether any other cases existed. The effect of therapy on survival was assessed, in particular, the repeated use of TIPS and prolonged courses of antibiotics. All four patients had thrombosis of their TIPS at the time of diagnosis of enterococcal bacteremia. All were treated with prolonged courses of intravenous antibiotics. One patient had echocardiographic evidence of subacute bacterial endocarditis with chronic aortic insufficiency. In all cases, liver transplantation was contraindicated in the acute setting because of uncontrolled endovascular infection. Two of four patients survived; these were the only two patients who had had a successful repeat TIPS. Enterococcal bacteremia is a rare complication following TIPS but carries a high mortality. It usually occurs in the setting of technically difficult TIPS with shunt thrombosis. Management should be focused on long term antibiotics and attempts at reestablishment of portal decompression with another TIPS. Liver transplantation should not be considered until the infection is cleared. Prophylaxis for Enterococcus species should be considered in technically difficult or unsuccessful TIPS.
(Am J Gastroenterol 1998;93:636 – 639. © 1998 by Am. Coll. of Gastroenterology)
INTRODUCTION Transjugular intrahepatic portosystemic shunts (TIPS) represent an increasingly used therapy for portal hypertension in patients with cirrhosis. An alternative to the surgical portosystemic shunt, this technique involves the placement of an intrahepatic expandable metal stent between the portal and systemic circulation (1, 2). Early results with transjugular intrahepatic portosystemic shunts (TIPS) suggest that it is effective at lowering portal pressure and acutely controlling variceal bleeding, and can also be effective therapy for ascites that is refractory to diuretic therapy (3, 4). TIPS is also associated with a low rate of procedural complications (1). Complications of TIPS include encephalopathy in ;25% of cases (5, 6), bleeding in ;5% (1, 7), and deterioration of liver function in 0 –7% (1, 8). Infectious complications of TIPS are uncommon. Fever and/or bacteremia have been reported in 1–10% of patients in several large series (7). Most fevers have been self-limited, and the few cases of positive blood cultures have typically involved Staphylococcus aureus or Gram-negative organisms. Enterococcus faecalis is an uncommon organism in the biliary tree and upper digestive tract. It is responsible for 5–10% of episodes of cholangitis (9, 10) and 0 –10% of episodes of spontaneous bacterial peritonitis (11–14). E. Faecalis could be expected to be a potential cause of infection following TIPS, inasmuch as these infections are presumed to be secondary to seeding of the blood from the biliary tree and third generation cephalosporins, which are often used as prophylaxis, do not cover Enterococcus species. However, enterococcal infections following TIPS have never been reported. We report four cases of Enterococcus infection following TIPS, half of which were fatal, and
Received Mar. 21, 1997; accepted Jan. 6, 1998. 636
AJG – April 1998
ENTEROCOCCAL BACTEREMIA
637
discuss the implications for procedural prophylaxis and therapy.
tibiotics and comfort care, and died 2 days following transfer. Autopsy was not performed.
CASE 1
CASE 2
Patient 1 was a 45-yr-old man with alcoholic cirrhosis and a history of alcoholic hepatitis who presented to an another institution following two episodes of upper GI bleeding from portal hypertensive gastropathy despite b-blocker therapy. Admission laboratory values included a hematocrit of 26%, ALT 21 mg/dl, alkaline phosphatase 321 mg/dl, total bilirubin 9.2 mg/dl, prothrombin time (PT) 5 16.6 s, and albumin 2.5 g/dl. Physical examination was significant for ascites, peripheral edema, jaundice, a II/VI systolic murmur and no evidence of asterixis. Upper endoscopy revealed 11 esophageal varices and marked portal hypertensive gastropathy without acute bleeding noted. Ultrasound revealed flow in the left and right portal veins but no definite flow was seen in the main portal vein. The patient underwent a TIPS. At angiography, the portal vein was patent; however, several small filling defects were seen near the portal vein end of the shunt, which were thought to represent thrombi. A large esophageal varix was embolized with 5-mm coils. Because of the question of portal vein thrombus, the patient underwent repeat portal angiography the following day, which revealed no evidence of thrombus. The TIPS was patent with a portosystemic gradient of 10 mm Hg. The patient was discharged on neomycin, lactulose, Aldactone, and Lasix. The patient was initially well after TIPS, with no further bleeding and improved ascites but with mildly worsened encephalopathy. Seven months following TIPS he was admitted with melena and confusion. He underwent upper endoscopy, which showed blood oozing from the duodenal bulb and proximal stomach, both of which were cauterized. Blood cultures were positive for Enterococcus in four of four bottles. The patient was transferred to UCSF for transplant evaluation. There was no history of intravenous drug use and it was unclear whether he had been having fevers at home. On exam he had anasarca, jaundice, and a III/VI systolic with a II/VI diastolic murmur at the left sternal border radiating throughout the precordium. Laboratory tests revealed white count 17.1 K, hematocrit 22.7%, platelets 46 K, BUN 84 mg/dl, creatinine 2.0 mg/dl, PT 18.7 s, and albumin 1.2 g/dl. Repeat upper endoscopy showed no varices nor evidence of bleeding. Ultrasound with Doppler revealed a patent TIPS with a velocity of 1.8 m/s. Transesophageal echo showed a vegetation on the aortic valve and wide open aortic insufficiency. The echocardiographic findings and the degree of compensation of his aortic insufficiency supported a diagnosis of subacute bacterial endocarditis (SBE). Treatment was initiated with ampicillin and gentamycin. Valve replacement was contraindicated based on his severe liver disease. Liver transplantation was not possible because of his endovascular infection and cardiac status. He had recurrent melena and refused endoscopy. He was transferred back to his primary care physician for an-
Patient 2 was a 57-yr-old man with end stage liver disease secondary to alcohol. He had a TIPS placed at another institution for recurrent episodes of variceal bleeding despite sclerotherapy, and for refractory ascites. Angiographically, the stent clotted immediately. Urokinase was infused without success. The thrombosis included the portal vein, superior mesenteric vein, and the splenic vein. Postprocedure he developed fevers, and blood cultures were positive for Enterococcus faecium and Citrobacter amalonaticus diversus. He was treated with intravenous vancomycin and gentamycin for 2 wk and discharged without fevers. One week later he was readmitted with fevers, lethargy, and progressive renal insufficiency (creatinine 3.9 mg/dl versus 1.2 mg/dl before TIPS). Blood cultures were once again positive for Enterococcus and Citrobacter. The former was sensitive to vancomycin and high dose streptomycin; the latter was sensitive to most antibiotics. The patient was treated with vancomycin, high dose streptomycin, and ceftriaxone. A Hickman catheter was placed with a plan for 5 wk of antibiotic therapy. His renal insufficiency resolved off gentamycin and he was discharged 1 wk later. Nine days postdischarge he developed melena without hemodynamic instability. His hematocrit was 29%. Ultrasound of his liver with Doppler analysis revealed no flow through the TIPS. Upper endoscopy revealed two oozing bulbar duodenal ulcers (one with a visible vessel), which were treated with heater probe cautery and epinephrine injection with good hemostasis. He had nonbleeding grade I–II esophageal varices and gastric varices. He was sent to UCSF for evaluation for liver transplantation to relieve his portal hypertension and to remove his infected TIPS. Transplantation was not possible because of his concurrent endovascular infection, and he was sent home for prolonged intravenous antibiotics with a plan to return for transplant evaluation following 4 – 6 wk of intravenous antibiotics. He died at home 2 wk after discharge. Autopsy was not performed. CASE 3 Patient 3 is a 66-yr-old man with alcoholic cirrhosis who had a TIPS placed at UCSF in June 1994 for recurrent massive bleeding from gastric varices. Two months later, flow through the TIPS was noted to be decreased (0.1 meters/s) on routine ultrasound. The patient was admitted and a venogram revealed nonocclusive thrombus within the TIPS shunt. Stent revision and parallel Wallstent placement were unsuccessful on two occasions. The patient was discharged after unsuccessful TIPS revisions. Two days after discharge, the patient was admitted for a massive variceal bleed. Physical exam was significant for jaundice, massive ascites, and melena on rectal exam. He was lethargic, but
638
BROWN et al.
asterixis was not present. Laboratory exams included a white count of 10 K, a bilirubin of 3.8 mg/dl, and a normal BUN and creatinine. Three further attempts at TIPS revision were made; the third resulted in successful placement of a parallel stent. The original TIPS was noted to be totally occluded at that time. Blood cultures at admission were positive for Enterococcus, and the patient was started on vancomycin and then changed to high dose ampicillin after several days. This isolate had relatively high resistance to aminoglycosides. Echocardiogram revealed no evidence of vegetations. The patient received a total course of 14 days of intravenous ampicillin. The patient has had no fevers or other evidence of bacteremia, nor has he had any further variceal bleeding with 1 yr of follow-up. CASE 4 Patient 4 is a 43-yr-old man with cirrhosis secondary to alcohol abuse and hepatitis C, who presented to an outside hospital with an acute upper GI bleed. At that time, endoscopy revealed two duodenal ulcers, distal esophageal ulceration, and portal hypertensive gastropathy. Two months later, the patient had recurrence of bleeding and underwent multiple sessions of sclerotherapy, which were not successful in controlling his recurrent portal hypertensive bleeding. He had a TIPS placed at that time, which was complicated by Streptococcus mitis and Staphylococcus epidermidis septicemia which was treated with vancomycin. He was readmitted 10 days later with hematemesis. Endoscopy showed large esophageal varices and a blood clot at the gastroesophageal junction. A TIPS venogram revealed thrombosis of the portal vein with extension of thrombus into the main portal trunk down to the confluence of the inferior mesenteric vein (IMV), superior mesenteric vein (SMV), and splenic veins. Multiple attempts by interventional radiology to open the shunt were unsuccessful. The patient was then transferred to UCSF for transplant evaluation. There was no history of injection drug use. On exam he was awake and alert and did not show any evidence of encephalopathy. His abdomen was nontender, obese, without obvious ascites. He had a hematocrit of 24.0, white count 8.0, normal prothrombin time and transaminases, and a bilirubin of 3.3. Repeat endoscopy showed marked portal hypertension, gastropathy, small gastric varices, and 21 esophageal varices with two ulcers at the GE junction. These ulcers were clean-based and consistent with previous sclerotherapy treatments. A repeat venogram was done, which was consistent with the findings at the outside hospital. The patient then underwent a TIPS revision and stenting through a portal clot. Blood cultures from the day of admission grew enterococcus from two of four bottles and the patient was started on vancomycin and gentamycin. The patient underwent an ultrasound after TIPS revision, which showed patent TIPS with maximal velocity of 1.2 meters/s, hepatofugal flow in the right and left portal veins, and a patent hepatic vein. The patient was determined not to be a
AJG – Vol. 93, No. 4, 1998 transplant candidate secondary to his continued alcohol consumption. He was then transferred back to the referring hospital for continued care. He completed a 4-wk course of antibiotics for his bacteremia. Subsequently, he has done well since his hospitalization. He is currently working and enjoying a good quality of life. DISCUSSION We present four cases of endovascular infection with E. faecalis following TIPS. Each case was associated with some degree of portal vein or shunt thrombosis. This is a potentially important point, as bile duct injury, which is more common with technically difficult shunt procedures and thrombosis, has been associated with shunt stenosis (15, 16) and would also be a possible source of enterococcal seeding of the TIPS. Although it is impossible to unequivocally implicate the TIPS as the cause of bacteremia (particularly in patient 1, for whom the time interval was long between the placement of the TIPS and the positive cultures), there were no other risk factors for enterococcal bacteremia. Specifically, none of the patients had a history of intravenous drug use, recent dental procedures, or other parenteral risk factors. In addition, none of the patients were found to have cholecystitis, cholangitis, spontaneous bacterial peritonitis (SBP), or positive ascites cultures for enterococcus. Thus, the clinical opinion of the hepatologists, surgeons, and interventional radiologists caring for the patient was that the bacteremia was likely related to the TIPS procedure. Following TIPS, three of the four patients presented with bleeding and were found to have positive blood cultures and clotted TIPS. The fourth patient had fevers following thrombosis of his TIPS. Chronic infective endocarditis was present in one patient, and the deaths of two of four patients were related to persistent infection. It is interesting to note that both patients who were successfully treated were the only two to have had a successful repeat TIPS procedure following thrombosis and occlusion of the original shunt. Whether the lack of portal decompression played a role in the poor outcome of the first two patients, or whether the inability to successfully place another TIPS was a reflection of a more extensive septic thrombosis, is uncertain. Whether these cases could have been prevented with a prophylactic regimen with activity against Enterococcus (i.e., vancomycin or ampicillin), in addition to a third generation cephalosporin, is also unknown. Given the rarity of this infection, which occurred in one of 314 TIPS performed at our institution, it will be difficult (if not impossible) to study. However, prophylaxis with ampicillin or vancomycin might be warranted in technically complicated TIPS that involve a large numbers of passes, shunt or portal vein thrombosis, or biliary opacification. In this scenario, the severity of the complication must be weighed against the possibility of the emergence and prevalence of vancomycin-resistant Enterococcus species. Vancomycin prophylaxis in patients under-
AJG – April 1998
ENTEROCOCCAL BACTEREMIA
going cardiac and vascular surgery has been shown to reduce postoperative infections with enterococcal species (17), a rare complication following other vascular procedures. Enterococcal vascular graft infections are usually treated by surgical removal and antibiotics with good results (18, 19); however a case report of E. faecium infection following aortobifemoral bypass that was not removed resulted in the death of the patient (20). In the case of TIPS, surgical removal is not possible without transplantation, and transplantation is usually considered to be contraindicated in the presence of an uncontrolled endovascular infection. In the case of a documented infection with enterococcus, echocardiography should be considered to rule out valvular vegetations that would require prolonged courses of intravenous antibiotics. The patient should not be viewed as a candidate for liver transplantation until negative blood cultures have been documented off antibiotics. Finally, intravascular infection should be considered in the differential diagnosis when a patient presents with recurrent bleeding and shunt thrombosis following TIPS, particularly in the case of technical difficulty or thrombosis associated with the initial shunt procedure, or in the presence of fevers or new cardiac murmurs. ACKNOWLEDGMENT The authors acknowledge the assistance of Dr. G. Sandhu in reviewing and revising the manuscript. Dr. Brown is supported by a grant from the Glaxo Institute for Digestive Health. Reprint requests and correspondence: Robert S. Brown, Jr., M.D., M.P.H., Division of Digestive Diseases, 708 Burnett Womack Building, Box 7080, University of North Carolina, Chapel Hill, NC 27599-7080.
REFERENCES 1. LaBerge JM, Ring EJ, Gordon RL, et al. Creation of transjugular intrahepatic portosystemic shunts with the wallstent endoprosthesis: Results in 100 patients. Radiology 1993;187:413–20. 2. Rossle M, Haag K, Ochs A, et al. The transjugular intrahepatic portosystemic stent-shunt procedure for variceal bleeding [see comments]. N Engl J Med 1994;330:165–71.
639
3. Ochs A, Rossle M, Haag K, et al. The transjugular intrahepatic portosystemic stent-shunt procedure for refractory ascites. N Engl J Med 1995;332:1192–7. 4. Somberg KA, Lake JR, Tomlanovich SJ, et al. Transjugular intrahepatic portosystemic shunts for refractory ascites: Assessment of clinical and hormonal response and renal function. Hepatology 1995;21: 709 –16. 5. Sanyal AJ, Freedman AM, Shiffman ML, et al. Portosystemic encephalopathy after transjugular intrahepatic portosystemic shunt: Results of a prospective controlled study. Hepatology 1994;20:46 –55. 6. Somberg KA, Riegler JL, LaBerge JM, et al. Hepatic encephalopathy after transjugular intrahepatic portosystemic shunts: incidence and risk factors [see comments]. Am J Gastroenterol 1995;90:549 –55. 7. Freedman AM, Sanyal AJ, Tisnado J, et al. Complications of transjugular intrahepatic portosystemic shunt: A comprehensive review. Radiographics 1993;13:1185–210. 8. Noeldge G, Richter GM, Roessle M, et al. Morphologic and clinical results of the transjugular intrahepatic portosystemic stent-shunt (TIPSS). Cardiovasc Intervent Radiol 1992;15:342– 8. 9. Boey JH, Way LW. Acute cholangitis. Ann Surg 1980;191:264 –70. 10. Shimada H, Nakagawara G, Kobayashi M, et al. Pathogenesis and clinical features of acute cholangitis accompanied by shock. Jpn J Surg 1984;14:269 –77. 11. Akriviadis EA, Runyon BA. Utility of an algorithm in differentiating spontaneous from secondary bacterial peritonitis [see comments]. Gastroenterology 1990;98:127–33. 12. Runyon BA. Monomicrobial nonneutrocytic bacterascites: A variant of spontaneous bacterial peritonitis. Hepatology 1990;12:710 –5. 13. Runyon BA, Canawati HN, Akriviadis EA. Optimization of ascitic fluid culture technique. Gastroenterology 1988;95:1351–5. 14. Runyon BA, Umland ET, Merlin T. Inoculation of blood culture bottles with ascitic fluid. Improved detection of spontaneous bacterial peritonitis. Arch Intern Med 1987;147:73–5. 15. Benner K, Saxon R, Flora K, et al. The achilles heel of TIPS, parenchymal tract stenosis and occlusion, is caused by biliary-TIPS fistulae. Hepatology 1995;22:297A. 16. LaBerge JM, Ferrell LD, Ring EJ, et al. Histopathologic study of stenotic and occluded transjugular intrahepatic portosystemic shunts. J Vasc Interv Radiol 1993;4:779 – 86. 17. Maki DG, Bohn MJ, Stolz SM, et al. Comparative study of cefazolin, cefamandole, and vancomycin for surgical prophylaxis in cardiac and vascular operations. A double-blind randomized trial. J Thorac Cardiovasc Surg 1992;104:1423–34. 18. Soyer R, Bessou JP, Bouchart F, et al. Surgical treatment of infected composite graft after replacement of ascending aorta. Ann Thorac Surg 1994;58:425– 8. 19. Baxter BT, Mesh CL, McGee GS, et al. Limb-threatening ischemia complicated by perigenicular infection. J Surg Res 1993;54:163–7. 20. Chirurgi VA, Oster SE, Goldberg AA, et al. Endovascular infections caused by enterococci highly resistant to ampicillin. South Med J 1991;84:1435–7.