Fatal Aortic Endocarditis Associated with Community-Acquired Serratia marcescens Infection in a Dog

Introduction

Originally, Serratia spp. were considered nonpathogenic, saprophytic bacteria; however, these bacterial species are now recognized as an important cause of nosocomial outbreaks in both human and veterinary medicine. In contrast, nonhospital-associated infections are rare and appear to be limited to patients with immune deficiency, chronic debilitating disease, or in human medicine, intravenous (IV) drug use.

This case report provides evidence that community-acquired Serratia spp. are capable of causing life-threatening infections in dogs that lack known predisposing immunosuppressive factors or the historical possibility of nosocomial transmission.

Case Report

A 12 yr old, 29 kg, castrated male Dalmatian was referred to the Veterinary Teaching Hospital, College of Veterinary Medicine at North Carolina State University (NCSU-VTH) for evaluation of fever, lethargy, and neurologic signs that had begun acutely approximately 1 wk earlier. According to the owners, the first clinical signs included a head tilt to the right, falling to the left, and nystagmus. After being examined by the primary veterinarian, the dog was presumptively diagnosed with canine idiopathic vestibular disease and sent home the same day with no treatment. The dog was not hospitalized, no catheterization procedures were performed, and no drugs were administered IV. Subsequently, vestibular signs progressed and the dog started vomiting. Five days after the initial examination, when re-evaluated by the primary veterinarian, the dog was febrile (105°F), a heart murmur was ausculted for the first time, and a tick was removed from his right rear foot. These findings resulted in immediate referral for additional diagnostic evaluation. No additional diagnostic tests were performed prior to referral and the dog was not hospitalized.

The owners had moved from Florida to North Carolina 3 wk before the onset of clinical signs. This dog was the only pet in the household and remained mainly indoors. According to the owners, the dog was previously healthy, received year-round heartworm prevention (ivermectin) and vaccination status was current. Except for surgical neutering at 4 mo of age, there had been no other surgical procedures.

On admission to NCSU-VTH, the dog was depressed, the body condition score was estimated at 5 out of 9, mucous membranes were pink, and the capillary refill time was <2 sec. Rectal temperature was 106.5°F, the heart rate was 140 beats/min, and the dog was panting. There were no skin lesions or wounds and lymph nodes were normal in size. There was moderate dental tartar. A grade 5 out of 6 systolic heart murmur was ausculted, which was loudest at the left base. Femoral arterial pulses were hyperdynamic, but no gallop sounds or arrhythmias were ausculted. A resting electrocardiogram revealed sinus tachycardia. The dog was hypertensive (multiple Doppler systolic blood pressures ranged from 180–240 mm Hg). Neurologic abnormalities consistent with paradoxical vestibular disease included ataxia, right head tilt, head bob, a truncal sway, falling to the left, right ventral strabismus, and bilateral proprioceptive deficits. Postural reflexes were normal. Multifocal central nervous system disease was suspected. Differentials included vascular (i.e., hypertensive brain damage, septic emboli, and hemorrhage), infectious (meningitis), inflammatory, or neoplasia.

Hematologic abnormalities included leukocytosis characterized by neutrophilia (22.3×10³/μL; reference range, 2.53–12.88×10³/μL), a left shift (bands 531×10³/μL), and monocytosis (2.39×10³/μL; reference range, 0.21–1.05×10³/μL). Slight toxicity was noted on the blood smear. The dog was severely thrombocytopenic (19×10³/μL; reference range, 190–468×10³/μL) with a high mean platelet volume of 27.2 fL (reference range, 7.9–13.8 fL). The dog was azotemic (creatinine was 2.8 mg/dL; reference range, 0.5–1.6 mg/dL and blood urea nitrogen was 57 mg/dL; reference range, 8–27 mg/dL), hyperphosphatemic (8.6 mg/dL; reference range, 2–6.7 mg/dL), hypoalbuminemic (2.3 g/dL; reference range, 2.8–4 g/dL), hyperbilirubinemic (0.4 mg/dL; reference range, <0.2 mg/dL), and hyponatremic (143 mmol/L; reference range, 147–154 mmol/L). Serum alkaline phosphatase activity was mildly increased (299 IU/L; reference range, 14–120 IU/L). There was hyperfibrinogenemia (400 mg/dL; reference range, 100–300 mg/dL), prolonged activated partial thromboplastin time (18.1 sec; reference range, 7.8–14.4 sec), and an elevated D-dimer concentration (500–1,000 ng/mL; reference range, <250 ng/mL). Urinalysis findings included a urine specific gravity of 1.029, proteinuria (2+ on the dipstick), hematuria (10–15 red blood cells per high power field), rare granular casts, and occasional bilirubin crystals. A urine culture was negative.

Thoracic and abdominal radiographs were unremarkable. Two dimensional, M-mode, color flow, and spectral Doppler echocardiography demonstrated vegetative lesions on the aortic valve, severe aortic insufficiency, and mitral regurgitation. Vegetative lesions of the aortic valve were best visualized in the right parasternal long-axis (Figure 1) and short-axis (Figure 2) views. Based on these findings, aortic vegetative endocarditis and mitral valve insufficiency were diagnosed.

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

Three blood samples (10 mL each) collected aseptically at 1 hr intervals from different anatomic sites on the dog were submitted for aerobic and anaerobic blood culture. The results of the Snap 4Dx^a for detection of Anaplasma spp., Borrelia burgdorferi, Ehrlichia canis antibodies, and Dirofilaria immitis antigen were negative. In addition, antibody reactivity was not found by indirect fluorescent antibody testing using Babesia canis, E. canis, Rickettsia rickettsii, Bartonella henselae, and Bartonella vinsonii subsp. berkhoffii antigens. After the blood culture samples were obtained, antibiotic therapy was initiated. The dog was prescribed ampicillin-sulbactam (22 mg/kg IV q 8 hr) and doxycycline (10 mg/kg per os q 12 hr). Famotidine (0.5 mg/kg IV q 12 hr) and metoclopramide (0.5 mg/kg subcutaneously q 8 hr) were administered for vomiting in conjunction with conservative fluid therapy. During the next 5 hr, the dog remained febrile, became progressively hypotensive, and developed multiple areas of petechiation on the abdomen. Fresh frozen plasma was transfused to treat suspected DIC. Over the next 3 hr, the dog deteriorated rapidly and developed labored breathing and crackles. Despite diuretic administration for congestive heart failure, the dog continued to deteriorate, developed opisthotonus, and died.

Gross necropsy examination confirmed severe aortic valve endocarditis characterized by multiple, friable, white, verrucous, vegetative lesions. There was endocardiosis of the mitral valve. Petechiae and ecchymoses were found scattered throughout the lungs, intestines, meninges, and brain. A coronary artery was partially occluded with a large thrombus and multiple myocardial arterioles were occluded by aggregates of fibrin. There was an acute, focal splenic infarct accompanied by fibrin exudation and necrosis. Microscopically, the aortic valve was expanded by multifocal infiltrates of degenerative neutrophils and fibrin. There was also neutrophilic myocarditis and myocardial necrosis; multifocal lymphoplasmacytic infiltrates in the kidneys and pancreas; acute necrosis and hemorrhage in the brain; and multifocal neutrophilic infiltrates throughout the meninges, cerebrum, and cerebellum. No remarkable lesions were found in the stomach and there was no evidence of either adrenal hyperplasia or neoplasia. The clinical and postmortem findings were consistent with heart failure, septicemia, and DIC due to severe bacterial endocarditis. In less than 12 hr, all three blood cultures yielded a heavy growth and pure culture of S. marcescens. Based on the sensitivity results, the organism was resistant to ampicillin-sulbactam and cephazolin, but was sensitive to quinolones, third generation cephalosporins, imipenem, and aminoglycosides. Culture of a portion of the aortic valve in Bartonella alpha-proteobacteria growth medium resulted in a heavy growth of S. marcescens, which was confirmed by 16S rDNA sequencing. Bartonella spp. DNA was not amplified from the either the aortic valve or BAPGM culture medium.

Discussion

This case report describes the clinical, hematologic, echocardiographic, and pathologic abnormalities in a dog diagnosed with community-acquired S. marcescens bacteremia and aortic valve endocarditis. Historically, S. marcescens, a gram-negative bacillus belonging to the family Enterobacteriaceae, was considered an innocuous, nonpathogenic, saprophytic, water organism. Because of the easily recognized red pigmentation of certain strains, it was used extensively as a biologic marker; however, since the first cases of nosocomial infections were described in people in 1951, Serratia spp. have been identified as increasing in medical importance. In the human medical literature, S. marcescens infection is almost uniformly caused by nosocomial transmission of these bacteria within the hospital environment.^1–4 Serratia spp. have been implicated in hospital-acquired conjunctivitis, endocarditis, meningitis, urinary tract, and wound infections. Based on investigations of nosocomial outbreaks, S. marcescens infections appear to be transmitted through hand-to-hand contact by hospital personnel, via solutions used for medical procedures, and by catheterization or needle puncture.^2–5

Three studies published in the veterinary literature also support nosocomial transmission of S. marcescens to severely ill dogs located in the intensive care units of large veterinary medical centers.^6–8 The first documentation of nosocomial Serratia spp. septicemia in dogs was reported in 1973 in a case series involving six dogs.⁶ Although the source of infection was not determined, all dogs had two risk factors in common: fever developed while hospitalized under intensive medical care with an indwelling jugular catheter; and various debilitating diseases were diagnosed prior to S. marcescens septicemia. Also, a recent surgery had been performed in three of the six dogs. In a subsequent study from the United States, S. marcescens was isolated from 50% of all contaminated IV catheters removed from cats and dogs at a veterinary teaching hospital during a 1 yr period.⁷ The high prevalence of S. marcescens in that hospital population was due to contaminated benzalkonium chloride sponge pots that were located throughout the hospital. Serotyping by the Centers for Disease Control and Prevention in Atlanta confirmed the same 010:H11 S. marcescens serotype in isolates obtained from the sponge pots and catheters. In a study involving dogs with endocarditis, myocarditis, and arrhythmias associated with B. vinsonii subsp. berkhoffii infection, S. marcescens was isolated from an aortic mass from one dog.⁸ In that case, alpha-proteobacteria DNA was also amplified from blood and valve tissue, and S. marcescens was suspected to be a postmortem contaminant; however, hospital-associated transmission or a secondary infection due to chronic bartonellosis could not be ruled out. In a subsequent study from South Africa, multiple bacteria were isolated from 22 IV catheters from 100 dogs with parvoviral enteritis. Of the bacteria isolated, 3 of 22 catheter cultures yielded S. marcescens.⁹ Recently, a dog developed fatal necrotizing fascitis in association with S. marcescens after a tooth extraction.¹⁰ Oral surgery or skin contamination at antibiotic injection sites were considered the most likely transmission routes. In that dog, similar to the case described herein, multi-organ failure and DIC were evident at necropsy.

In humans, nonhospital-associated infections due to S. marcescens are limited to patients with chronic, debilitating diseases and IV drug users.^2–5 Nonhospital-acquired transmission of these bacteria in dogs has only been described in association with canine granulocytopathy syndrome, an autosomal recessive condition in Irish setters that is characterized by a defective bactericidal oxidative burst, which predisposes these dogs to recurrent opportunistic bacterial infections.¹¹ A case report of S. marcescens endocarditis in a horse was described in 1992. Although information regarding previous hospitalizations or injectable medications was lacking, nosocomial transmission was thought to be unlikely and no underlying disease was diagnosed in that horse.¹² A comprehensive review of the literature by the authors of the present study found only one previous report of aortic endocarditis due to a nosocomial S. marcescens infection in a dog, which was diagnosed at necropsy.⁶ In the patient described in this report, an antemortem echocardiographic diagnosis of S. marcescens endocarditis was possible and the organism was independently isolated from both blood and the aortic valve.

By definition, a nosocomial infection is acquired in a hospital or health care facility.¹³ For most bacterial infections, an onset of symptoms more than 48 hr after admission is evidence for nosocomial acquisition. In the case presented, no hospital-acquired transmission was identified and several facts rule out the possibility of a hospital-acquired infection. Specifically, the dog was not hospitalized prior to referral, total hospitalization time following referral was less than 12 hr, no urinary or IV catheterization was performed prior to obtaining the blood cultures, and clinical signs had begun 1 wk prior to admission. There was no history of immunosuppressive drug administration, and no underlying pre-existing disease was found at necropsy. Hematologic and histopathologic abnormalities detected were all consistent with septicemia. In addition, no tick-borne disease, including canine bartonellosis, was diagnosed. Whether the dog suffered from an acquired immune deficiency could not be determined; however, this dog did not have any previous history of infections and genetically predisposed immune deficiency syndromes have not been reported in the Dalmatian breed. Therefore, an inherited immune deficiency seemed unlikely.

Bacterial endocarditis should be suspected in cases of fever of unknown origin, particularly when there is a heart murmur of recent onset, a defined source of infection, or embolic phenomena.¹⁴ The diagnosis of endocarditis can be challenging, requiring the synthesis of clinical, microbiologic, and echocardiographic criteria to arrive at a definitive diagnosis. Generally, causative bacteria are identified in only 50–60% of the cases.¹⁴ Staphylococcus spp., Streptococcus spp., Erysipelothrix spp., Corynebacterium spp., and Escherichia coli are the most common bacterial isolates obtained from dogs with endocarditis. Recently, Bartonella spp. have become a more frequently recognized cause of endocarditis in dogs and humans. Serratia spp. have been recognized as causative agents of endocarditis in human patients with histories of IV drug abuse and as nosocomial infection in hospitalized patients.¹⁵ Other predisposing factors include an immune deficiency, IV or urinary catheterizations, corticosteroid therapy, or broad-spectrum antibiotics.¹⁵ Therefore, the ability of S. marcescens to cause endocarditis was thought to be limited to patients with chronic debilitating disorders or hospital interventions. Interestingly, and similar to both the dog described here and the canine S. marcescens endocarditis case of 1973, human cases appear to be associated with a high rate of thromboembolic complications. Antibiotic therapies often fail to clear the infection, resulting in a fatal outcome. In fact, the initial neurologic signs in this patient were most likely a component of multifocal septic embolism, which was confirmed at necropsy.

Generally, a long course of bactericidal antibiotics is recommended for the treatment of bacterial endocarditis. Initial treatment usually consists of in-hospital administration of parenteral antibiotic combinations. As soon as the fever resolves and clinical improvement is evident (within the next 3–5 days), oral antibiotic treatment is continued on an outpatient basis.¹³ Antibiotic selection should always be based on blood culture and sensitivity results, but until they are available, empirical antibiotic therapy is generally started using either a combination of a fluoroquinolone and a penicillin-based antibiotic or a penicillin and an aminoglycoside.¹⁴ Because of the potential of a concurrent tick-borne infection (such as ehrlichiosis or Rocky Mountain spotted fever) doxycycline was used in conjunction with ampicillin-sulbactam to treat this dog while awaiting blood culture results. High resistance to a variety of antibiotics makes the treatment of S. marcescens infections difficult. Serratia spp. have been found to be resistant to beta-lactams, aminoglycosides, and quinolones.^14,15 Currently, third generation cephalosporins or a commercial product containing equal quantities of imipenem and cilastatin are the antibiotics of choice. In this case, Serratia spp. was resistant to ampicillin-sulbactam by in vitro testing, but susceptible to fluoroquinolones. The dog described in this report failed to respond to therapy and died within 6 hr after starting antibiotics, most likely because of the advanced stage of the disease.

Conclusion

Although risk factors for S. marcescens infections include long-term use of IV or urinary catheters; receiving intensive care in a hospital setting; or having a chronic, debilitating disease prior to admission, this case report provides evidence that community-acquired Serratia spp. are capable of causing life-threatening infections, such as endocarditis, in dogs lacking known predisposing immunosuppressive factors or the historical possibility of nosocomial transmission.