Vancomycin for multi-drug resistant Enterococcus faecium cholangiohepatitis in a cat

Abstract

A 12-year-old, neutered male domestic shorthair cat was evaluated with a life-long history of intermittent, predominantly small bowel diarrhea and a 3 day history of hematochezia. At presentation, the cat had increased liver enzyme activities and an inflammatory leukogram. Histopathology demonstrated inflammatory bowel disease (IBD), cholangiohepatitis and pancreatitis. The cholangiohepatitis was associated with a multi-drug resistant Enterococcus faecium. Gallbladder agenesis was also documented. Treatment with vancomycin was safely instituted for 10 days. Clinical signs resolved, however, cure of the bacterial cholangiohepatitis was not achieved. The risk of vancomycin resistant enterococci (VRE) in human and veterinary medicine is discussed.

A 12-year-old neutered male domestic shorthair cat (3.6 kg) was evaluated at the Iowa State University Veterinary Teaching Hospital (ISUVTH) for a life-long history of intermittent, predominantly small bowel diarrhea and a 3 day history of hematochezia. Physical examination was unremarkable. Hematological abnormalities included a leukocytosis (36,100 cells/μl; reference range 5500–19,500 cells) with a neutrophilia (30,324 cells/μl; reference range 2500–12,500) and left shift (1083 cells/μl; reference range 0–300), a lymphopenia (1083 cells/μl; reference range 1500–7000), monocytosis (1083 cells/μl; reference range 0–850) and eosinophilia (2527 cells/μl; reference range 0–750). Serum biochemical abnormalities included increased alanine aminotransferase (ALT) (311 IU/l; reference range 20–125) and alkaline phosphatase (ALP) (134 IU/l; reference range 0–60) activities; hyperproteinemia (95 g/l; reference range 61–80), hyperglobulinemia (66 g/l; reference range 26–51) and mildly increased pre- (19.7 μmol/l) and post-prandial (27.1 μmol/l) bile acids (reference range 1–25, pre: 1–13; post: 1 to μmol/l). Urinalysis, serum thyroxine concentration and prothrombin and partial thromboplastin times were within normal limits. Fecal floatation was negative for parasitic ova.

The main differential diagnoses were inflammatory bowel disease with concurrent cholangiohepatitis/cholangitis or inflammatory bowel disease that had progressed to intestinal lymphoma with hepatic involvement. Abdominal radiographs were unremarkable. Ultrasonography revealed a mildly hyperechoic liver with decreased differentiation of the portal veins. No gallbladder was evident and in its usual location, the liver contained multiple duct-like structures containing anechoic fluid. These structures were thought to be dilated common bile ducts. Transabdominal biopsy of the liver was not considered safe. Fine needle aspirates of the liver parenchyma, obtained under general anesthesia, were submitted for anaerobic and aerobic microbiological culture and susceptibility testing (NCCLS 2002). Gastroduodenoscopy abnormalities were increased granularity and friability of the duodenal mucosa. Pinch biopsy samples were collected from the stomach and duodenum.

Duodenal histopathology revealed a moderate infiltrate of plasma cells and lymphocytes in the lamina propria with immature crypt epithelium consistent with mild lymphoplasmacytic enteritis. The aerobic culture from the liver grew a heavy pure growth of an α-haemolytic Streptococcus species susceptible to almost all antibiotics evaluated. A diagnosis of lymphoplasmacytic IBD and presumptive bacterial cholangiohepatitis was made. The cat was treated with amoxicillin–clavulanic acid (Clavamox; Pfizer) (17.4 mg/kg, PO, q 12 h), metronidazole (Metronidazole; Watson Labs) (8.3 mg/kg, PO, q 12 h), S-adenosylmethionine (Denosyl; Nutramax Labs) (90 mg, q 24 h) for its antioxidant effects and ursodeoxycholic acid (Ursodiol; Amide Pharmaceutical) (10 mg/kg, PO, q 24 h).

Despite clinical improvement and weight gain (0.23 kg) over 2 weeks, the small bowel diarrhea worsened. As ursodeoxycholic acid and S-adenosylmethionine can cause diarrhea, they were discontinued. However, diarrhea and increased hepatic enzyme activity persisted for 2 months in the face of weight gain. An exploratory celiotomy was performed 6 days after discontinuing antibiotics. At the time of surgery, all clinical signs of disease had resolved. The liver was firm, swollen, pale, and mottled while the pancreas was nodular and firm. The absence of a gallbladder was confirmed and no obvious common bile duct dilation was noted. Liver and pancreatic biopsies were collected for histopathology and the liver parenchyma was submitted for culture.

Hepatic histopathology revealed portal areas diffusely infiltrated by large numbers of degenerate neutrophils, as well as large numbers of lymphocytes and plasma cells. There was moderate biliary proliferation, and many bile ductules were plugged with bile or degenerate neutrophils. There was marked portal and bridging fibrosis. Tissue Gram stain did not demonstrate bacteria within the lesions. The histopathological diagnosis was suppurative and lymphoplasmacytic cholangiohepatitis with biliary hyperplasia, marked portal fibrosis, and cholestasis. Histopathology of the pancreas revealed severe diffuse chronic suppurative and lymphocytic pancreatitis with marked fibrosis. Culture of the hepatic parenchyma resulted in a heavy, pure growth of a multi-drug resistant Enterococcus faecium, resistant to all penicillins, cephalosporins, fluroquinolones, tetracyclines, macrolides, aminoglycosides, and sulfonamides routinely tested at ISUVTH. It was susceptible to vancomycin and chloramphenicol.

Neither vancomycin nor chloramphenicol are ideal drugs to be given to cats, especially for long-term therapy. While not routinely used in veterinary medicine, vancomycin, a bactericidal glycopeptide antibiotic, has been used safely and effectively in a cat with cholangiohepatitis (Jackson et al 1994). Its use is limited by poor oral absorption necessitating intravenous infusion. Vancomycin can be associated with anaphylaxic reactions, hypotension and bradycardia, all of which are related to the rate of infusion. Nephrotoxicity and ototoxicity is also infrequently reported (Jackson et al 1994, Greene and Watson 1998). Chloramphenicol is well absorbed orally and concentrated in the liver and bile. Dosage reduction is recommended in patients with impaired hepatic function. Cats are considered more sensitive to the adverse effects of chloramphenicol possibly because of their limited glucuronidation capabilities necessary for its metabolism in other species (Watson 1978). Adverse effects reported at various doses include anorexia, lethargy, dehydration, vomiting and diarrhea (Watson and Middleton 1978, Watson 1980, 1991). Reversible bone marrow suppression occurs in veterinary patients and routine hematological monitoring is recommended especially when long-term therapy is instituted (Watson 1980, Page 1991). In addition, although rare, chloramphenicol exposure through aerosolization and oral intake can cause fatal aplastic anemia in humans (Page 1991). These concerns were discussed with the owner, a research veterinarian, who declined treatment with chloramphenicol.

The cat was masked down with sevoflurane (Sevofluroane; Abbott Labs) and a 5 French jugular catheter (Multi-lumen jugular catheter, Arrow International) was placed for 10 days of treatment with vancomycin (Vancomycin; American Pharmaceutical Partners) (19.4 mg/kg diluted in 10 ml 0.9% saline (Saline: Abbott Labs), IV over 1 h, q 12 h) (Tzannes et al 2000). The owner was instructed to aseptically administer the drug slowly to avoid anaphylaxis. Renal analytes and serum electrolytes were monitored twice weekly and remained normal. Culture of a hepatic aspirate, obtained 5 days after cessation of therapy, grew a heavy, pure growth of E. faecium with the same susceptibility spectrum. Further treatment with vancomycin or alternative treatment with chloramphenicol was declined by the owner. The cat remains clinically asymptomatic 15 months later.

This case report describes a cat diagnosed with IBD, pancreatitis and cholangiohepatitis, the latter subsequently associated with an extremely resistant strain of enterococcus. A previous report documented a compelling association between cholangiohepatitis, pancreatitis and inflammatory bowel disease (IBD) in 7/18 (39%) cats with cholangiohepatitis. Concurrent inflammation of these organs is speculated to occur in cats because the common bile duct and pancreatic duct join prior to entering the duodenum, allowing bacteria to ascend from the gastrointestinal tract (Weiss et al 1996).

Many cats with suppurative cholangiohepatitis have underlying predisposing structural or functional hepatobiliary defects (Weiss et al 1997, Center 2000). All the cats in a report by Hirsch and Doige (1983) had an abnormality in their extrahepatic biliary system or pancreas including anatomic abnormalities of the gallbladder. Gallbladder agenesis may have predisposed our patient to an ascending infection from the intestinal tract. In order to diagnose agenesis, a careful search of the abdomen during surgery must be made to rule out ectopic locations. Sites of an ectopic gallbladder in humans include the abdominal wall, retroperitoneal space, falciform ligament and left side of the liver (Richards et al 1993). Extensive exploratory laparotomy did not identify an ectopic gallbladder in our patient. Additionally, surgery did not confirm dilation of the common bile ducts suggesting partial resolution of the initial bacterial cholangiohepatitis accounting for the lack of clinical signs at the time of surgery.

Most bacteria carry genes enabling them to be resistant to a given antimicrobial prior to any drug exposure. The continued presence of an antibiotic creates an environment whereby bacterial mutants may be selected (Hoffman 2001). The organisms cultured in our patient were typical of those previously implicated in feline cholangiohepatitis (Day 1995, Center 2000). However, the α-haemolytic Streptococcus species grown initially was supplanted by a resistant enterococcus possibly secondary to selection pressure exerted by prolonged amoxicillin–clavulanic acid treatment.

Enterococcus species are facultative, anaerobic Gram-positive cocci that demonstrate intrinsic and acquired resistance to multiple antibiotics including macrolides, sulfonamides and aminoglycosides (DeLisle and Perl 2003). Routine vancomycin use in human patients over the past 15 years has created serious antimicrobial resistance problems. In 1993, the National Nosocomial Infections Surveillance system documented a 20-fold increase in the percentage of patients found to have an infection with vancomycin resistant enterococci (VRE) between 1989 and 1993 (CDC 1993). DeLisle and Perl (2003) recently reported that as of December 2000, 26.3% of enterococci isolated from human intensive care unit patients were resistant to vancomycin. This is a 31% increase over the data averaged for the years 1995 to 1999. The most important factor leading to transmission of VRE from patient to patient via hospital personnel, contaminated hospital equipment and contaminated surfaces is the unknown carrier state followed by the use of antibiotic therapy which increases fecal bacteria concentration (Martone 1998). Monitoring for VRE includes routine fecal cultures in high risk patients who are immunocompromised or receiving antimicrobials (DeLisle and Perl 2003). A fecal sample was collected in our patient after vancomycin therapy in an attempt to culture the enterococcus species found in the liver. The culture yielded normal flora; however, this could reflect intermittent shedding of the organism.

The treatment goal for bacterial cholangiohepatitis is elimination of the offending microorganism based on culture and antimicrobial susceptibility testing and control of the secondary immune response. Bile is preferentially cultured in suspected cases of cholangiohepatitis in human medicine and is commonly recommended in veterinary patients. In our case, bile could not be collected, as neither a gallbladder nor enlarged bile ducts could be visualized for sampling. Adverse side effects associated with cholecystocentesis are bile peritonitis, hemorrhage and rarely, vagal hypotension (Center 1990, 2000, Savary-Bataille et al 2003). Savary-Bataille et al (2003) found percutaneous ultrasound-guided cholecystocentesis to be safe in healthy cats but there have been no studies in cats with cholangiohepatitis. No veterinary or human medical studies have evaluated the utility of bile culture compared with hepatic parenchymal culture. In our case, hepatic parenchyma was cultured both for initial diagnosis and monitoring.

In a previous report, resolution of clinical signs associated with cholangiohepatitis caused by E. faecium was achieved using vancomycin. However, elimination of the organism was not proven (Jackson et al 1994). In our case, vancomycin was used based on the presumed resolution of the infection in that cat and our owner's reluctance to use chloramphenicol. Unfortunately, despite continued lack of clinical signs in our patient, a microbiological cure was not attained. This could have been partially due to the need for extended treatment. Clinical resolution is expected early in the therapy for cholangiohepatitis, however, successful resolution of the bacterial component often requires 3–6 months of antimicrobial therapy. Confirmation of cure is by repeated hepatic parenchymal or bile cultures. Longer treatment with vancomycin is problematic because of the need for intravenous administration and the potential for increasing renal toxicity with cumulative dosages. Therefore, vancomycin may not be an appropriate antimicrobial for the long-term treatment of cholangiohepatitis in the cat.

The apparent spontaneous resolution of clinical signs despite persistent E. faecium infection remains unexplained. Bacterial infection may not have caused the hematochezia and increased hepatic enzyme activities. The initial presenting signs may have been related to IBD and extended treatment with amoxicillin–clavulanic acid may have caused persistent diarrhea. Subsequent discontinuation of this drug and response of the IBD to metronidazole could explain the resolution of all clinical signs before surgery. Treatment of an asymptomatic, culture positive patient may not be necessary and long-term treatment with antimicrobials could increase selection pressure and preferential growth of resistant organisms. Although recommended, repeat hepatic culture after extensive antimicrobial therapy is often not pursued. Instead, clinicians rely on resolution of clinical, hematological and biochemical changes. This report suggests that some patients may remain culture-positive despite clinical resolution. Further study is warranted to determine if it is necessary to continue therapy to the point of obtaining a negative culture.

The potential development of antimicrobial resistance in humans and other species, including dogs and cats, is of significant concern. Although reports of VRE infections in companion animals are uncommon, the evolution of multi-drug resistance in hospital infections in dogs has been documented as a direct result of the pattern of antimicrobial administration (Warren et al 2001, Prescott et al 2002). The use of vancomycin in small animal practice could provide sufficient selection pressure to increase the prevalence of VRE isolates. This case demonstrates a hepatic infection with a multi-drug resistant E. faecium in a cat. While the infection in this case did not involve a VRE, and VRE infections have not been documented in cats, it certainly demonstrates the gravity of the decisions made in selecting antimicrobial therapy.

The consequence of vancomycin selection pressure in veterinary hospitals would presumably be determined by the prevalence of VRE in veterinary environments. In a study of clinically healthy dogs presented to an urban veterinary practice in The Netherlands, 26% (6/23) were VRE carriers (van Belkun et al 1996). In the United States, a VRE infection associated with the urinary tract of a dog has been reported (Simjee et al 2002). Current research at the Colorado State University Veterinary Teaching Hospital has found VRE to be prevalent in routine monitoring. Cultures revealed VRE carriage in 15% of dogs and approximately 75% of environmental samples (Morley et al 2004, Morley, personal communication). These data suggest that VRE carriage poses a risk to other hospitalized patients as well as the potential for human colonization with VRE through transmission from companion animals.

This case demonstrates the potential consequence of antimicrobial use and illustrates the need to critically evaluate the use of critical antimicrobials such as vancomycin, imipenem, linezolid, fluroquinolones and third and fourth generation cephalosporins in companion animal medicine. At the time of this writing, the American College of Veterinary Internal Medicine is developing a consensus statement concerning this issue (Morley, personal communication). The authors suggest that practicing veterinarians utilize this position statement to facilitate discussion regarding antimicrobials used in their practices.

In summary, the cat of this case report tolerated treatment with vancomycin without adverse effects. However, treatment was ineffective and the cat likely remains persistently colonized. Inflammatory bowel disease combined with aberrant bile flow associated with gallbladder agenesis may have predisposed this cat to cholangiohepatitis and pancreatitis. Vancomycin use in this patient stimulated a number of important questions about the risk involved with treating veterinary patients with multi-drug resistant bacterial infections. In retrospect, the use of vancomycin in this patient was probably inappropriate. We would caution future use of this drug in cats with susceptible cholangiohepatitis. In addition, the lack of clinical signs for over a year, suggests that infection with E. faecium may not equate with disease and this should also be considered when deciding on treatment protocols for cats with bacterial cholangiohepatitis.