Septicemia and Infection due to ESBL-producing K. pneumoniae Following Feline Renal Allograft Transplantation

Introduction

The rapid emergence of resistance of Klebsiella spp. to antimicrobials and the spread of extended–spectrum-β-lactamase (ESBL) activity amongst micro-organisms is an ever-present concern in human medicine and an emerging concern in veterinary medicine. Additionally, as feline patients undergoing renal transplantation are growing in number, increased vigilance for the presence of multidrug-resistant bacterial infections in immunocompromised patients is required. The most common multidrug-resistant gram-negative bacterial phenotype isolated amongst human patients is one with cephalosporin resistance and carbapenem susceptibility. The ESBL-producing organism cultured from the patient described in this report was resistant to all evaluated cephalosporins, aminoglycosides, fluoroquinolones, and trimethoprim sulfa, but was susceptible to carbapenems. This case report highlights the possible risk factors in this cat and other veterinary patients for the development of ESBL-producing bacterial infections and outlines diagnostic and treatment strategies for recognizing and combating such infections.

Case Report

A 12 yr old castrated male domestic longhair weighing 4.52 kg was presented for chronic renal failure and evaluation as a candidate for renal allograft transplantation. Chronic renal failure had been suspected approximately 4 yr prior by a primary care veterinarian on the basis of mild azotemia (blood urea nitrogen [BUN], 13.6 mmol/L; reference range, 5.7–12.9 mmol/L; creatinine, 247.5 umol/L mg/dL; reference range, 70.7–212.2 umol/L) without hyposthenuria. No treatment was instituted, and the cat's azotemia reportedly remained stable.

Thirty-nine days prior to presentation to the authors' hospital, the patient presented to a private practice tertiary care facility and was examined by an internal medicine specialist. Historical findings included lethargy and inappetence of 1 wk duration and possible weight loss. Abnormal findings on physical examination included a grade II/VI systolic heart murmur; palpation of small, irregularly marginated kidneys; pale mucous membranes; moderate dehydration; and a poor body condition score (three out of nine). Clinical laboratory abnormalities included elevated BUN (39.6 mmol/L) and creatinine (1122.7 umol/L). The packed cell volume (PCV) was 33% (30–45%, ) and total solids (TS) measured 72 g/L (52–88 g/L). Urinalysis (UA) revealed isosthenuria (urine specific gravity, 1.011; reference range, 1.012–1.040) and trace protein was noted. Culture and sensitivity of the urine yielded no bacterial growth. Thoracic radiography revealed no significant findings, and abdominal ultrasound revealed small, irregular kidneys with poor corticomedullary distinction. Mild pyelectasia (3.4 mm) of the left kidney was noted.

Treatment at that time included hospitalization, IV fluid therapy, ampicillin (22 mg/kg IV q 8 hr), and famotidine (0.25 mg/kg IV q 24 hr). The patient was discharged 8 days later with a BUN of 16.4 mmol/L and a creatinine of 459.7 umol/L. The cat was anemic with a PCV of 25%. In-home therapy included subcutaneous (SC) fluid administration (150 mL q 24 hr), cyproheptadine (2 mg per os [PO] q 12 hr), famotidine (2.5 mg PO q 24 hr), metoclopramide (0.9 mL PO q 8 hr), and feeding a balanced renal-specific diet.

At the time of presentation to the authors' facility, standard transplantation candidate diagnostic testing was performed including a complete blood count (CBC), serum biochemical analysis, blood typing with major and minor cross matching to feline renal donors and packed red blood cell (PRBC) units, UA, urine culture, fecal flotation, serology for Toxoplasma gondii (immunoglobulin [Ig] M and G), thoracic radiography, abdominal ultrasonography, and echocardiography. Results of initial diagnostics revealed a mildly elevated BUN (11.8 mmol/L; reference range, 5.4–11.4 mmol/L) and elevated creatinine (344.8 umol/L; reference range, 88.4–176.8 umol/L), a progressive normocytic normochromic nonregenerative anemia (PCV, 21%), and total weight loss of 1.05 kg over the previous 6 mo. Abdominal ultrasonography confirmed chronic changes to both kidneys as previously noted, but revealed mineralization within the left kidney, diffuse thickening of the muscularis layer of the ileum, and multiple hypoechoic mesenteric lymph nodes. The patient also tested seropositive for Toxoplasma gondii (IgG titer, 1:128). Echocardiography revealed a normal cardiac examination with the exception of very mild tricuspid regurgitation and a physiologic flow murmur. Due to those findings and concern for pathology of the gastrointestinal tract that might preclude transplantation, the patient was prescribed recombinant erythropoietin^a (100 IU/kg SC q 72 hr) and was discharged with recommendations to have full-thickness intestinal biopsies and lymph node biopsies performed. Pending results of those diagnostic tests, further consideration of renal transplantation would be given.

The patient re-presented 14 days later for surgical biopsies of the ileum and mesenteric lymph nodes. Abdominal exploration was unremarkable with the exception of a mildly thickened ileum and prominent mesenteric lymph nodes. Biopsies of those tissues were taken routinely and without complication and submitted for histopathology. Perioperative antibiotics were not administered. The patient recovered uneventfully and was discharged that evening.

Histopathological evaluation of the ileum and mesenteric lymph node showed mild lymphocytic enteritis and mild lymphoid hyperplasia. Due to the lack of significance or clinical implication of those findings, a decision was made to pursue renal transplantation.

Sixteen days following the biopsy procedure, the patient was admitted for renal transplantation. The BUN and creatinine values at that time were 10.7 mmol/L and 318.2 umol/L, respectively. A PCV of 24% and TS of 68 g/L indicated a mild improvement in anemia that was not attributable to hemoconcentration. Immunosuppressive therapy was instituted via administration of microemulsified cyclosporine solution^b(2.5 mg/kg PO q 12 hr). Antibiotic therapy with clindamycin^c (5.5 mg/kg PO q 12 hr) was initiated when immunosuppression was initiated to avoid reactivation of latent Toxoplasma gondii infection as an unintended consequence of immunosuppression. The patient did not exhibit any noticeable side effects of either cyclosporine or clindamycin administration. Adequate immunosuppression was confirmed by assessment of 12 hr whole blood trough concentrations of cyclosporine measured by high-performance liquid chromatography, and dose adjustments were made to achieve a therapeutic trough concentration of cyclosporine of 249.6–416 nmol/L.¹

On the day of renal transplantation (day 0), the patient was administered prednisolone^d (0.6 mg/kg PO q 12 hr) as an additional immunosuppressive agent. Anesthesia was induced with etomidate^e (1 mg/kg IV) and lidocaine^f (1 mg/kg IV) after premedication with midazolam^g (0.2 mg/kg intramuscularly [IM]), hydromorphone^h (0.1 mg/kg IM), and ketamineⁱ (1.25 mg/kg IM) and was maintained with inhaled isoflurane^j .A 13 cm, 5.5-French triple lumen central venous catheter was introduced into the left jugular vein and secured after sterile preparation of the area. The patient was clipped and aseptically prepared for a ventral midline celiotomy. Perioperative antibiotics were administered (cefazolin^k, 22 mg/kg IV) at anesthetic induction and q 2 hr throughout the procedure. One unit of cross-matched compatible feline PRBCs was administered IV during the procedure.

A ventral midline celiotomy was performed in both the renal donor and recipient. Exploration of the recipient abdomen revealed no abnormalities associated with the previous procedure, and the intestinal and lymph node biopsy sites appeared grossly healed. The left-sided aorta and right-sided vena cava of the recipient were isolated and freed of surrounding connective tissue and adventitia by a combination of blunt and sharp dissection, and the abdomen was prepared for receipt of the donor kidney. Nephrectomy of the donor's left kidney and renal transplantation were performed as described elsewhere.² Intravesicular mucosal appositional ureteroneocystostomy was performed as previously described followed by routine cystotomy closure.³ Allograft nephropexy to the left body wall was performed with simple interrupted sutures of 5-0 polypropylene^l. Following lavage with sterile saline and suctioning, abdominal closure was routine. No intra- or perioperative complications occurred.

The patient recovered uneventfully in the intensive care unit (ICU). The patient was discharged on day 5 with instructions to administer clindamycin (5.5 mg/kg PO q 12 hr), prednisolone (0.6 mg/kg PO q 12 hr), and cyclosporine (3.5 mg/kg PO q 12 hr). The BUN and creatinine at the time of discharge were 14.3 mmol/L and 168 umol/L, respectively, and a USG of 1.015 was measured. The PCV and TS values were similar to the values measured at the time of presentation (PCV, 23%; TS, 61 g/L) without further blood transfusion. The cyclosporine level at the time of discharge was moderately increased at 668.1 nmol/L/L. Oral dosing of cyclosporine was decreased to more accurately target the therapeutic range.

Cyclosporine level, BUN, creatinine, PCV, and TS were assessed q 7 days following discharge for 4 wk. During that time, the patient became more anemic (PCV, 20%; TS, 77 g/L), and erythropoietin was reinstituted (100 U/kg SC twice weekly for 2 wk). Minor changes in BUN and creatinine occurred with the subsequent highest values being 12.9 mmol/L and 176.8 umol/L, respectively. A reticulocyte count was not performed. Cyclosporine levels fluctuated between 137.3 and 426 nmol/L, and the oral dose was adjusted accordingly.

Twenty-nine days following transplantation, the patient's creatinine level increased to 203.3 umol/L while the BUN remained stable at 12.9 mmol/L. The cyclosporine level of 77.4 nmol/L was subtherapeutic. Ultrasonography of the allograft showed mild renal enlargement (5.5 cm in length) with mild pyelectasia (2.3 mm) and slight perirenal effusion. Allograft rejection was suspected, and the patient was treated with methylprednisolone Na succinate^m (10 mg/kg IV once) and enrofloxacinⁿ (5 mg/kg IV once). The oral prednisolone dosage was increased from 0.6 mg/kg q 12 hr to 0.8 mg/kg PO q 12 hr, and the oral cyclosporine dosage was increased from 2.5 mg/kg q 12 hr to 4.85 mg/kg q 12 hr. The patient recovered well from that event without hospitalization.

Re-examination on day 36 revealed an increased BUN (17.1 mmol/L) and a stable creatinine (203.3 umol/L). The cyclosporine level was in the therapeutic range at 391 nmol/L. Serum blood glucose (BG) measured 12.4 mmol/L (reference range, 3.7–9.3 mmol/L). The patient was doing well clinically, and no additional changes in treatment were instituted.

Renal values, cyclosporine levels, PCV, TS, and BG were documented twice weekly for three more visits then q 1 mo thereafter for 8 mo. Renal values remained stable and cyclosporine levels during that time fluctuated between 167.2 and 881.1 nmol/L. Oral doses of cylcosporine were adjusted accordingly. Serum BG values fluctuated from 5.7 to 9.0 mmol/L. Overall, the patient was clinically well, his anemia had resolved (PCV was 35%), and his body weight had increased by 1.8 kg over 8 mo.

On day 376 following transplantation, the patient presented with a chief complaint of decreased appetite and weight loss of 0.5 kg. Physical examination revealed flea infestation and cestode parasitism not noted previously. Abdominal palpation was not performed according to the service protocol following renal allograft transplantation. Renal values at that time were stable (BUN, 10.4 mmol/L; creatinine, 159.1 umol/L). A CBC revealed no significant findings, and the cyclosporine level was significantly elevated at 1399.4 nmol/L. The cat was treated with nitenpyram^o (11.4 mg PO once) and fipronil^p (98 mg) with methoprene (118 mg) topically q 1 mo for the flea infestation. The patient was also administered praziquantal (27.3 mg) and pyrantel pamoate^q (108.9 mg) PO once for cestode parasitism. Instructions were given to skip the evening dose of cyclosporine and reinstitute cyclosporine therapy the following morning at a dose of 1.4 mg/kg.

The patient re-presented on day 381 for anorexia and lethargy. The patient was slightly febrile with a temperature of 39.2°C (reference range, 37.8–39.2°C). Initial diagnostic evaluation included a CBC and serum biochemical analysis. Biochemical abnormalities included an increased BUN (22.5 mmol/L), hypocalcemia (2.1 mmol/L; reference range, 2.3–2.8 mmol/L), hyponatremia (133 mmol/L; reference range, 146–157 mmol/L), and hypoalbuminemia (21 g/L; reference range, 24–38 g/L). Serum creatinine measured within the normal range at 176.8 mmol/L but was slightly elevated compared to the previous visit (159.1 mmol/L). The cyclosporine level at that time was severely elevated at 1151.5 nmol/L. A CBC revealed a normal total leukocyte and neutrophil count (12.5 × 10⁹ cells/L; reference range, 2.3–14 × 10⁹ cells/L), but a left shift (3.4 × 10⁹ band neutrophils/L, 33%) was present and moderate toxic changes to the neutrophils were noted. A septic process was suspected, and coagulation testing, UA, urine culture, thoracic radiography, and abdominal ultrasonography were performed. Results of the coagulation profile showed a mild prolongation of the partial thromboplastin time (PTT) at 20.7 sec (reference range, 12.6–15.7 sec). A serum titer for Toxoplasma IgG antibody was performed and remained positive at 1:64. Abdominal ultrasonography revealed a 4 cm hypoechoic mass associated with the small intestine, prominent mesenteric lymph nodes, and marked echogenic peritoneal effusion. The patient was sedated, fine-needle aspiration of the mass and lymph nodes was performed, and the effusion was sampled. Samples of all aspirates were submitted for cytological evaluation. All three samples were demonstrative of septic, suppurative inflammation with a nucleated cell count of >111 × 10⁹ cells/L noted in the effusion, and frequent intracellular cocci and bacilli were noted within neutrophils and macrophages. No atypical cells characteristic of neoplasia were seen. Based on a confirmed diagnosis of septic peritonitis, the patient was prepared for general anesthesia and exploratory laparotomy.

Anesthesia was induced with midazolam (0.3 mg/kg IV), methadone (0.5 mg/kg IV), lidocaine (1 mg/kg IV), and etomidate (0.36 mg/kg IV). The patient was intubated, and anesthesia was maintained with isoflurane in 100% O₂ and a constant rate infusion of remi-fentanyl^r at 0.3–0.5 ug/kg/min. Perioperative ticarcillin diNa/clavulanate potassium^s (40 mg/kg) were administered IV at induction and q 90 min throughout the procedure. A 13 cm 5.5-French triple lumen central venous catheter was placed into the right jugular vein after sterile preparation of the area. During the procedure, the patient became hypotensive (systolic arterial pressure, 62–92 mm Hg) and multiple 5 mL/kg boluses of hydroxyethylstarch^t were administered IV to expand the intravascular volume and provide oncotic support. The hypotension persisted, and the patient was started on vasopressor therapy (dopamine^u, 7–12 ug/kg/min). Persistent hypotension resulted in the addition of norepinephrine^v (0.1 mg/kg/min). Additionally, the patient was transfused with one unit of feline fresh frozen plasma (FFP, 20 mL/kg IV).

Exploratory laparotomy revealed a diffusely tan liver with multifocal, red, punctuate areas and a 5 cm × 3 cm diffusely tan/pink, firm mass adjacent to the aboral jejunum and ileum showing extensive hyperemia and focal areas of white, purulent material over its exterior. On gross inspection, the mass was intimately associated with the ileocecal segment of the intestine, and confirmation of either invasion or involvement of the intestinal lumen could not be made (Figure 1). Additionally, the mesenteric lymph nodes were prominent. A jejunocolic resection and anastomosis, partial omentectomy, and liver biopsies were performed routinely. All excised tissues were submitted for histopathology, and aerobic and anaerobic bacterial culture and sensitivity testing was performed on swabs of the abdomen and mass (Figure 2). The abdomen was lavaged aggressively with sterile saline, and two sterile closed suction drains were placed in the abdomen, exiting the abdominal flank. The abdomen was closed routinely and an abdominal bandage was placed postoperatively.

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

The patient recovered in the ICU. Shortly after extubation, the patient suffered respiratory arrest; however, resuscitative efforts were successful. He was then maintained in the ICU on supportive care and antibiotic therapy including enrofloxacin (5 mg/kg IV q 24 hr), ticarcillin diNa/clavulate potassium (40 mg/kg IV q 6 hr), and clindamycin (10 mg/kg IV q 12 hr).

Over the course of the next day, the patient remained vasopressor-dependent and was treated with dopamine (5–8 ug/kg/min IV). Worsening anemia (PCV, 15%) and coagulopathy (prothrombin time, 13.3 sec; reference range, 9.6–13.2 sec; PTT, 47.8 sec) were treated with units of cross-matched PRBCs and plasma transfusions as needed. The total leukocyte count was elevated at 32.300 × 10⁹ cells/L (reference range, 4.0–18.7 × 10⁹ cells/L) with 21.3 × 10⁹ segmented neutrophils/L and 8.7 × 10⁹ band neutrophils/L (41%) with marked toxic change. Total parenteral nutrition was initiated at 50% of the resting energy requirement. Considering the patient was unable to tolerate oral administration of immunosuppressive agents, cyclosporine^w was dosed empirically at 1 mg/kg IV q 12 hr and dexamethasone^x was administered at 0.05 mg/kg IV q 24 hr.

A CBC obtained 24 hr later revealed a worsening peripheral neutrophilia with 41.3 × 10⁹ segmented neutrophils/L and 9.5 × 10⁹ band neutrophils/L (30%) present. Grossly, a yellow-colored turbid discharge was collected in the abdominal drains (2.1 mL/kg/hr over the previous 24 hr). Further, BG and lactate levels of the abdominal fluid (5.8 mmol/L and 12.9 mmol/L, respectively) compared to those of the peripheral blood (12.3 mmol/L and 4.7 mmol/L, respectively) in conjunction with cytology results from fluid collected from the abdominal drain were suspicious of persistent septic peritonitis.⁴

The patient was anesthetized for a second exploratory laparotomy. At surgery, small areas of purulent material were noted at the previous sites of liver biopsy and omentectomy. The jejunocolic anastomosis site was intact and not leaking. The devitalized portions of liver and mesentery were debrided, and a serosal patch was placed at the intestinal anastomosis site. The two previous abdominal drains were removed. The abdomen was lavaged aggressively with sterile saline, and one new closed-suction drain was placed within the abdomen exiting the abdominal flank. The subcutaneous tissues around the incision were also debrided and lavaged before routine closure. The patient recovered in the ICU but remained dependent on dopamine (10–14 ug/kg/min) therapy. Additionally, multiple transfusions of FFP and PRBCs were required to maintain the systolic arterial blood pressure between 60 and 95 mm Hg. The patient received a total of 8 PRBC and 12 FFP transfusions over the course of hospitalization. Due to persistent septic peritonitis and worsening neutrophilia, antimicrobial therapy was empirically changed from enrofloxacin and ticarcillin diNa/clavulanate potassium to meropenem^y (10 mg/kg IV q 8 hr). The clindamycin was continued at 10 mg/kg IV q 12 hr.

Biopsy specimens from the first exploratory laparotomy revealed that the mass associated with the ileum did not penetrate the intestinal lumen but did cause degeneration of the muscularis. The mass was diagnosed as an abscess within the mesenteric adipose tissue composed of degenerating neutrophils and large numbers of macrophages containing bacteria. The liver contained diffuse fatty infiltration of the hepatocytes. Bacterial culture results of the patient's urine as well as the mesenteric abscess produced moderate growth of an organism identified as an ESBL-producing Klebsiella pneumoniae (K. pneumonia) isolate. Both isolates exhibited a similar sensitivity pattern. Specifically, the organism was resistant to all first-, second-, and third-generation cephalosporin antibiotics evaluated including cefazolin, cefoxitin, cefotaxime, ceftazidime, and ceftriaxone. The isolates also showed resistance to aminoglycosides, fluoroquinolones (including ciprofloxacin, marbofloxacin, and enrofloxacin), and ticarcillin diNa/clavulanate potassium, amoxicillin trihydrate/clavulanate potassium, and trimethoprim sulfa. The organism was intermediately susceptible to piperacillin/tazobactam, and susceptible to imipenem at a minimal inhibitory concentration of 93 umol/L..

The following day, a CBC revealed bacteremia with rod-shaped bacteria noted within 2% of circulating neutrophils. Blood cultures were not performed. The segmented and band neutrophil counts had decreased to 34.2 × 10⁹ and 3.9 × 10⁹ cells/L (11%), respectively. The abdominal drain produced 1.3 mL/kg/hr of a yellow to pink, clear to slightly turbid material over 24 hr. Supportive care and treatment with IV meropenem was continued; however, fresh total parenteral nutrition was reformulated to eliminate the previous infusion as a potential source of bacteremia. An echocardiogram was performed to rule out endocarditis and cardiac causes of hypotension. Mild mitral and tricuspid regurgitation were noted, but no evidence of endocarditis was seen.

Forty-eight hours later, the cat appeared subjectively brighter. The segmented neutrophil count decreased to 27.4 × 10⁹ cells/L with 4.2 × 10⁹ band neutrophils/L (15%). The BUN and creatinine were within normal limits (6.4 mmol/L and 88.4 umol/L, respectively). Coagulation parameters had also improved, with the prothrombin time measuring 10.9 sec; however, the PTT remained mildly prolonged at 21.1 sec. The cyclosporine level was still elevated at >1165 nmol/L. Cyclosporine administration was discontinued for 2 days.

One week following admission, the patient exhibited signs of volume overload, including increased respiratory effort, auscultation of a Grade II/VI systolic heart murmur and gallop rhythm, and an elevated central venous pressure (8–11 cm H₂O). Fluid therapy was adjusted, and diuretic therapy was initiated with furosemide, with a total of 10.5 mg/kg furosemide^z given IV over 24 hr. Cyclosporine administration was reinstituted at a lower dose of 0.25 mg/kg IV q 12 hr.

The following day, the patient was depressed and uncomfortable. Despite multiple PRBC transfusions, anemia persisted (PCV, 12%; TS, 50 g/L). No signs of either bleeding or hemolysis were present. Colloid osmotic pressure measured 17.4 mm Hg (reference range, 21–34 mm Hg). The patient was prescribed oxyglobin^aa (0.5 mL/kg/hr) to increase O₂ carrying capacity and provide continued oncotic support. The patient remained hypotensive (systolic arterial blood pressure, 80 mm Hg) and became tachycardic (240–280 beats/min) despite judicious fluid therapy and support with both dopamine (12–16 ug/kg/min) and norepinephrine (0.1–0.4 mg/kg/min). Renal values at that time were within normal limits (BUN, 7.5 mmol/L; creatinine, 97.2 umol/L). The patient developed focal facial seizures, and midazolam (0.2 mg/kg IV) was administered to control seizure activity. The patient continued to decline and suffered cardiopulmonary arrest. Resuscitative efforts were unsuccessful.

A full necropsy was declined by the client; however, a cosmetic necropsy was performed and the renal allograft was submitted for histopathological evaluation.^bb Findings were not considered consistent with rejection but could represent inflammation induced by sepsis, return of native disease, and early nephropathy associated with cyclosporine toxicity.

Discussion

Renal allograft transplantation is a treatment option for end-stage acute kidney injury and chronic renal failure in cats.^1,2,5 Renal transplantation offers 6 and 36 mo survival rates of 59% and 42%, respectively and has been shown to improve survival and quality of life compared to medical management of renal disease.^1,5,6 Complications associated with immunosuppressive therapy include the development of infection, diabetes mellitus, and neoplasia.^1,6–8 Further, cats that become diabetic have a significantly increased risk of developing clinically relevant infections following renal transplantation.^8,9 In two separate studies evaluating the prevalence of infection in feline renal transplant recipients, infection was second only to rejection as a cause of death, occurring most often in the first 3–6 mo following transplantation. Six of 66 cats in one study and 43 of 169 cats in another study developed infections following renal transplantation.^6,9 Bacterial infections appear to be most common, comprising 53% of infections in the latter study, followed by viral, fungal, and protozoal infections.⁹ Bacterial infections most commonly affect the urinary tract, although infections at esophagostomy tube sites, and intra-abdominal, respiratory, bone, and disseminated infections have been reported.^1,9 With regard to such infections, the infecting species of bacteria and other characteristics of the bacterial inoculum have not been widely reported. Moreover, to the authors' knowledge, septic peritonitis and bacteremia caused by ESBL-producing K. pneumoniae has not been previously described in the feline renal transplant recipient or other small animal patients.

K. pneumoniae is a gram-negative, rod-shaped organism belonging to the Enterobacteriaceae family. It is ubiquitous in the environment, found in soil and water. Although a commensal organism found on the mucosal surfaces of mammals and in the gastrointestinal, respiratory, and genitourinary tracts of humans, K. pneumoniae has been implicated in causing infection of the respiratory and urinary tracts, abdomen, surgical sites, bloodstream, and brain.^10–16 ESBL-producing K. pneumoniae was identified in the urinary tract and abdomen of the patient described herein. Although rod-shaped bacteria were noted within circulating WBCs, blood cultures were not performed. Therefore, the cause of septicemia in this patient could not be confirmed as an ESBL-producing K. pneumoniae, although such a cause is likely.

Historically, effective treatments for K. pneumonia included penicillin and third-generation cephalosporin antibiotics; however, the rapid emergence of resistance of Klebsiella spp. to those and other antimicrobials is an ever-present concern.^10,17–20 Resistance is conferred by plasmid-mediated enzymes termed β-lactamases, and recently, ESBLs capable of hydrolyzing the oxyamino β-lactam antibiotics, such as third-generation cephalosporins and monobactams have been identified. Plasmids encoding ESBL genes frequently carry other resistance genes, and concurrent resistance to antibiotics, including the aminoglycosides, fluoroquinolones, and sulfonamides, is common.^10,21,22 Additionally, plasmid transmission between members of the Enterobacteriaceae family has been documented and may be responsible for several outbreaks of multidrug resistant K. pneumoniae and Escherichia coli (E. coli) described in the human medical literature.^{12,15,20,23,24} The most common multidrug resistant gram-negative bacterial phenotype isolated amongst human patients is one of cephalosporin resistance and carbapenem susceptibility.^12,21,25,26 The ESBL-producing organism cultured from the patient presented in this report was resistant to all evaluated cephalosporins, aminoglycosides, fluoroquinolones, and trimethoprim sulfa, but was susceptible to carbapenems.

ESBL-producing K. pneumoniae has become a growing concern for practitioners of human and veterinary medicine since first recognized in the US in 1989.^10,27–31 Outbreaks of infection with ESBL-producing K. pneumoniae have been most commonly reported in human neonatal and pediatric ICU settings.^15,32 Reports describing localized infection and bacteremia with ESBL-producing K. pneumoniae have also emerged from studies of adult ICUs, of patients afflicted with hematologic malignancies, and of recipients of solid-organ transplants, including liver and kidney transplants.^{14,22,33–37} Those reports have identified numerous risk factors associated with ESBL-producing K. pneumoniae and E. coli infection in people, including the previous use of antibiotics, especially β-lactam/β-lactamase inhibitor combinations, third-generation cephalosporins, carbapenems, and fluoroquinolones. The presence of either chronic/immunosuppressive diseases (e.g., diabetes, renal insufficiency, cancer) or immunosuppressive treatment (especially corticosteroid treatment, history of organ transplantation [pancreas and double kidney transplantation], prolonged hospitalization >5–7 days, recurrent hospitalization, admission to ICUs, previous surgery, urinary or central venous catheterization, and mechanical ventilation) were also identified as risk factors.^{10–12,15,17,26,34–38}

In the cat described in this report, it was unclear where the ESBL-producing K. pneumoniae infection originated and why the intestinal- and mesenteric-associated abscesses developed. In cases of either primary or secondary endogenous infection in humans, colonization of the intestinal tract with K. pneumoniae is considered a prerequisite for infection.²⁶ However, exogenous acquisition of infection is also possible, and the infection could have been acquired from other pets, from contact with the owners, or from contact with hospital personnel. Although the digestive tract of the colonized patient is thought to be the most significant reservoir of the organism, transmission via the hands of nursing staff has been documented.²⁶ Amongst humans, fecal carriage of ESBL-producing Enterobacteriaceae was documented in 47.3% of inpatients and 15.2% of outpatients in one study.²⁶ Another study reported at least 9.5% of patients' stool samples cultured positive for ESBL-producing Enterobacteriaceae after 5 days of hospitalization.³⁹

Although little information exists documenting either colonization or infection of veterinary patients by ESBL-producing K. pneumoniae or E. coli, one study found isolates carrying ESBL genes in the feces of two cows, one dog, and one pig.²⁷ Another study reported a high prevalence of ESBL-producing Enterobacteriaceae in 86 fecal samples obtained from 43 puppies in two kennels in Japan.²⁹ Another study evaluating the fecal flora of dogs treated with cephalexin for various skin disorders found that 62% of Enterobacteriaceae isolates from treated dogs contained ESBL genes.²⁸ ESBL-producing Enterobacteriaceae have also been isolated from the feces of healthy dogs in England.³⁰ In one study evaluating 24 ESBL-producing K. pneumoniae isolates from pets treated at a single veterinary hospital in France, 17 of the isolates were found in patients that underwent urinary tract procedures, such as cystotomy and perineal urethrostomy.³¹ Because the particular isolate was also found in humans in the same study, questions remain whether pets may acquire K. pneumoniae bacteria from their owners or caregivers.

ESBL-producing K. pneumoniae in the cat described above was cultured from the urinary tract, mesenteric lymph nodes, and mesenteric mass. The most common site of infection in human patients is within the urinary tract due to proximity to the anus and gastrointestinal tract, followed by the respiratory tract, wounds, and bloodstream.^10,15 Additionally, in one report, 21.5% of bloodstream infections originated in the urinary tract.²⁵ Intra-abdominal infections and infections at surgical sites have also been reported.^10–12 Studies of human patients also reveal that recipients of solid organ transplants have an increased risk of developing infections caused by ESBL-producing organisms, presumably due to the large number of confirmed risk factors in those patients.^22,23,33,34 Moreover, of solid organ transplant recipients in one study, recipients of kidney transplants developed ESBL-related infections more than recipients of any other solid organ, possibly owing to the high incidence of urinary tract infections in those patients.^36,38 Indeed, one study demonstrated a 20% incidence of infection with ESBL-producing Enterobacteriaceae with a urinary tract source in renal transplant recipients.⁴⁰ Human renal transplant recipients that also underwent pancreas transplantation were thought to be at increased risk for development of ESBL infections due to the prolonged contact with enteric bacteria that occurs during pancreas transplantation.³⁸ Contact with intestinal bacteria occurred in the patient described herein during surgery to obtain full-thickness intestinal biopsies 2 wk prior to initiation of immunosuppression. Additionally, the mesenteric abscess was intimately associated with the intestinal wall in the region of the ileum from which the biopsies had been obtained. Finally, the histopathology report of the resected intestine associated with the mass noted areas of “fibrosis, probably [reflective] of prior surgical intervention.” It is possible that contact with enteric bacteria during the intestinal biopsy played a role in the development of the mesenteric infection with ESBL-producing K. pneumoniae that occurred in this patient.

Additional surgical factors that may have allowed for intra-abdominal infection with K. pneumoniae include failure to administer perioperative antibiotics for the clean-contaminated surgery to collect the biopsy samples. Although aseptic technique was used and precautions were taken to protect the abdomen from either spillage or contamination with gastrointestinal bacteria, the surgeons may have seeded the abdomen with K. pneumoniae. Although the biopsy sites appeared grossly intact at the time of renal transplantation, polydioxanone sutures were used in the intestine, which are known to persist for weeks to months. The sutures may have acted as a conduit for intestinal pathogens to migrate transmurally and colonize the mesenteric area.

In human renal transplant recipients, one or more infections develops in over two-thirds of patients, often within 3–6 mo of transplantation, a period during which immunosuppression is often maintained at higher levels to prevent allograft rejection.⁴¹ Infections developing after 6 mo are often thought to be associated with either excessive immunosuppression/concurrent disease or to be community acquired from the general population.^9,41 It has been reported that 70% of feline renal transplant recipients with infection had whole blood cyclosporine levels in excess of the target range within 2 wk prior to diagnosis of infection.⁹

Concurrent conditions significantly associated with infections in feline renal transplant patients include rejection, neoplasia, and diabetes mellitus.⁹ This patient underwent treatment for acute rejection 11 mo prior to a diagnosis of septic peritonitis. Renal values normalized and clinical signs of rejection resolved shortly after treatment. It is unlikely that one episode of significant immunosuppression was directly related to the development of sepsis; however, the cat's cyclosporine level at the time of diagnosis of septic peritonitis was 1151.5 nmol/L, approximately 5–6 times; the recommended level 1 yr following transplantation (recommended level, 208 nmol/L).¹ Additionally, the whole blood cyclosporine level of this patient was documented to be in excess of the target range for the time period following transplantation 6 times; despite regular and moderately frequent monitoring. It is possible that such chronic excessive immunosuppression may have played a role in the development of infection.⁴¹

Although neoplasia has been associated with the development of infection in feline renal transplant recipients, neoplasia was not noted on any biopsy in this patient.⁹ Finally, although the cat did not develop overt diabetes mellitus, he did present with significant hyperglycemia at several times during follow-up examinations. Alterations in neutrophil chemotaxis and the presence of glucosuria may contribute to the development of infection in patients with significant hyperglycemia, and high tissue glucose concentrations may allow for more rapid bacterial growth.^8,9

Debate exists regarding the proper treatment of such cases; however, antibiotic therapy should be based on bacterial culture and sensitivity. Identification of ESBL-producing bacteria can be performed in a laboratory setting by conducting either disk diffusion or double disk synergy testing.^10,25 Data summarized from worldwide reports name the carbapenems, including imipenem and meropenem, as the most effective class of antibiotics against ESBL-producing strains.^15,17,18,20 Unfortunately, carbapenemases have also recently been documented, and 3.9–10.9% of ESBL-producing isolates have been reported to be insensitive to carbapenems, such resistance being conferred by plasmid-mediated AmpC β-lactamase production and porin loss.^14,19,42

To combat the threat of antibiotic resistance in humans, some authors have suggested novel treatment and prevention modalities. One of those novel treatments includes the potential use of zinc oxide nanoparticles, which have been shown to decrease growth of K. pneumoniae by 98.6% and of E. coli by 99.3% in vitro.⁴³ Also, as digestive tract colonization with ESBL-producing organisms is a prerequisite for either primary or secondary endogenous infection to occur, selective digestive decontamination of at-risk patients has been recommended by some institutions to decrease carriage of resistant aerobic gram-negative bacilli, such as ESBL- producing strains, and improve survival.^26,44 In one report, selective digestive decontamination was conducted in children presenting to the pediatric ICU and known to be harboring ESBL- producing Enterobacteriaceae based on fecal sample positivity. Decontamination consisted of parenteral cefotaxime to combat primary endogenous infection by respiratory tract pathogens, such as Streptococcus pneumoniae, Staphylococcus aureus, and Haemophilus influenzae. Enteral polymyxin E and tobramycin were also administered. The authors reported that the addition of the enteral antibiotics reduced carriage and infection rates due to ESBL-producing K. pneumoniae from 19.6 and 9% to 1 and 0%, respectively. Infection control strategies, including hand hygiene and the use of topical antibiotics, were also advocated.⁴⁴

Recognition and surgical treatment of septic peritonitis occurred promptly after presentation in the patient described herein; however, empirical antimicrobial selection was proven inappropriate following speciation and specific characterization of the infecting organism. Because of unanticipated resistance, many human patients with infection caused by ESBL-producing bacteria experience some delay in the prescription of appropriate antibiotic therapy.^10,25,35 Although that may result in a worsened prognosis, it is likely that outcome is also dependent on the primary site of the infection and the severity of the underlying disease. Therapy with meropenem was initiated within 48 hr of presentation in the patient described above. Signs of resolution of the infection were noted, including decreasing band cell counts on hematologic evaluation and improving nucleated cell count and character on cytological evaluation of abdominal effusion. However, the cat continued to decline clinically and eventually succumbed to cardiopulmonary arrest, presumably due to consequences of bacteremia and sepsis, including systemic inflammatory response syndrome, ongoing hypotension, and multiorgan dysfunction.

A systematic review of the literature regarding the relationship between ESBL production and mortality in patients with bacteremia showed an almost two-fold increase in mortality associated with ESBL production in patients with bacteremia caused by Enterobacteriaceae.³⁵ Moreover, inadequate empirical therapy for the serious infections with ESBL-producing organisms was independently associated with increased mortality in one cohort of ICU patients with bacteremia.¹⁴ In patients with hematologic malignancies treated with carbapenems for ESBL-related infections, the 30 day mortality rate was 44.8% (compared with 17.6% in patients without malignancy), suggesting either concurrent disease or treatment with immunosuppressive agents may worsen prognosis even in the face of appropriate antimicrobial therapy.³⁵ Finally, although not significant for this patient, mortality was increased by 2.5 in in one subgroup of patients with bacteremia that had infections resistant to carbapenem.¹⁴ Overall, human mortality rates associated with ESBL-producing K. pneumoniae infections ranged from 22–25% of infected patients.^10,14

Conclusion

Immunocompromised human solid organ transplant recipients have an increased risk of developing infections caused by ESBL-producing organisms. Veterinary transplant recipients may be at increased risk also. Risk factors predisposing to the development of infection in the cat described in this report may include prior receipt of antibiotics, lengthy hospital stay, ICU care, central venous catheter use, immunosuppression, and previous surgery of the gastrointestinal tract. Treatment of sepsis and bloodstream infections should be directed toward elimination of predisposing factors, combined with pathogen-directed antibiotic therapy dictated by antimicrobial susceptibility testing. Indwelling vascular and urinary catheters should be removed, intra-abdominal abscesses drained or excised, and other potential sources of infection surgically corrected, if possible. Immunosuppressive therapy should prevent allograft rejection yet avoid excessive immune impairment that may have adverse infectious and other complications. Antimicrobial stewardship should be employed to minimize selective pressure on microbes, and hand hygiene and other environmental preventive measures should be practiced routinely, especially in facilities caring for immunocompromised patients and those with other risk factors. Finally, selective digestive decontamination may be considered for feline renal transplantation candidates with one or more risk factors associated with the development of ESBL-related infections.