Concurrent spinal epidural empyema and endocarditis in a dog

Abstract

A 9-year-old neutered male Rhodesian ridgeback cross dog was evaluated for progressive non-ambulatory paraparesis, fever, and leukocytosis. The dog was diagnosed with spinal epidural empyema (SEE) and infectious endocarditis (IE) of the mitral valve based on the findings of contrast-enhanced computed tomography (CT), CT myelography, echocardiography, and bacterial culture. The report herein describes the clinical presentation, CT findings, clinical and surgical management of this case, together with the electrocardiography, and echocardiography findings. To the authors’ knowledge, this is the first reported case of spinal epidural empyema likely to be caused by infectious endocarditis of the mitral valve in a dog.

Spinal epidural empyema (SEE) is a rare condition defined as an accumulation of purulent material in the epidural space. The accumulation can cause mechanical compression of the spinal cord and often causes fever, spinal pain, and a rapidly progressive myelopathy (1–4). Advanced imaging including magnetic resonance imaging (MRI), contrast enhanced computed tomography (CT) or CT myelography is essential for the diagnosis of SEE and to help guide treatment decisions (2,3). Spinal epidural empyema has been managed successfully with antibiotics and supportive care, and with or without surgery to decompress the spinal cord (2,4). Spinal epidural empyema is often caused by extension of a local infectious process such as discospondylitis, although it has occasionally been reported in conjunction with distant infections (1–11). Infectious endocarditis (IE) is also a rare condition in dogs and is caused by bacterial colonization of the heart valves or endocardium. The aortic and mitral valves are most commonly affected, and IE can lead to serious multi-organ disease due to septic thromboembolism, including disease of the central and peripheral nervous system (12–15). Early diagnosis and treatment of IE may help limit the extent of systemic complications. Furthermore, identification of multi-organ disease or localized infection potentially caused by bacteremia, should prompt investigation for IE.

Case description

A 9-year-old 52.5 kg neutered male Rhodesian ridgeback cross dog was referred to the Adelaide Animal Emergency and Referral Centre (AAERC) following a 13-day history of pyrexia, lethargy, and progressive spinal pain with associated neurological deficits. He was initially examined at the referring veterinary clinic 13 d before referral for polyuria, polydipsia, lethargy, and inappetence. The dog’s rectal temperature was 41.8°C and a complete blood (cell) count (CBC) revealed a neutrophilia. He was treated with dexamethasone (Dexadresson 2 mg/mL; Intervet, Victoria, Australia), 0.11 mg/kg body weight (BW), SC, enrofloxacin (Baytril 50; Bayer Animal Health, NSW, Australia), 5.2 mg/kg BW, SC, and 5 d of enrofloxacin tablets at 5.7 mg/kg BW, PO, q24h. The signs partially improved. Five days before referral, however, back pain was newly noted. Rectal temperature was not reported at that time and carprofen (Carprieve Injection; Norbrook, Victoria, Australia), 3.9 mg/kg BW, SC, was administered followed by carprofen tablets (dose unknown) PO for 5 d. Two days before referral, pelvic limb dysfunction was seen along with a rectal temperature of 40.4°C. The day before referral, non-ambulatory paraparesis was reported and an indwelling urinary catheter was placed.

On presentation to the AAERC, the dog showed non-ambulatory paraparesis, reduced tibial and withdrawal reflexes, and absent patellar reflexes. Cranial nerve and thoracic limb function were normal, and no spinal pain was elicited. Rectal temperature was 40.4°C and clinical examination results were otherwise unremarkable. Myelopathy localizing to the L4-S3 spinal cord segments was suspected with differential diagnoses to include spinal epidural empyema, discospondylitis, osteomyelitis, paraspinal abscess, neoplasia, and meningomyelitis. Methadone (Methone; Ceva Animal Health, NSW, Australia), 0.3 mg/kg BW, IV, was administered and general anesthesia was induced using propofol (Propofol Sandoz; Sandoz, NSW, Australia) titrated to effect and maintained with 1% to 2% isoflurane (Isoflurane USP; Piramal Critical Care, Bethlehem, Pennsylvania, USA) in 100% oxygen via an endotracheal tube. Computed tomography of the vertebral column from T1 to the sacrum was performed before and after administration of iohexol 240 mg/mL (Omnipaque 240; GE Healthcare, NSW, Australia), 1.9 mL/kg BW, IV. Post-IV contrast CT of the trunk was then obtained and a CT myelogram (T1-sacrum) was subsequently performed via intrathecal injection of a total of 0.42 mL/kg BW iohexol 240 mg/mL (Omnipaque 240; GE Healthcare) at an L4–L5 inter-arcuate space. Cerebrospinal fluid (CSF) collection was attempted at the time of myelography; however, there was insufficient flow of CSF to obtain a diagnostic sample. Throughout the lumbar spine, the epidural fat had increased attenuation and reduced definition and/or was effaced by soft tissue-attenuating material (Figure 1). The spinal cord and cauda equina similarly had reduced definition and in several places were compressed by thickened epidural tissues; the most severe compression was at L5–L6 on the left. There was mild contrast enhancement of the meninges within the lumbar spinal canal (Figure 1). Several sites of mild chronic degenerative intervertebral disc disease (IVDD) were present through the spine. However, no definitive evidence of discospondylitis or extradural spinal cord compression secondary to disc herniation was seen. Ventral to the sacrum and tail base, there was moderate asymmetry of the midline muscle bellies, with the right containing a few small hypodense regions. There was mild bilateral medial iliac and internal iliac (sub-lumbar) lymphadenopathy. Both kidneys had multiple variably sized wedge-shaped non-contrast-enhancing cortical lesions (Figure 2). The thorax was unremarkable.

Figure 1.

Dorsal (top left), transverse (top right), and sagittal (bottom) plane computed tomography (CT) images of the lumbar spine, post-IV contrast (WL, 40; WW, 320). The transverse image is at the level of L5–6. Thickened epidural tissues with abnormal attenuation (black arrows), meningeal contrast enhancement (black arrowheads), and site of most severe extradural neural compression at L5–6 on the left side (white asterisk) are shown.

Figure 2.

Dorsal (left) and transverse (right) computed tomography (CT) images of the abdomen, post-IV contrast (WL, 40; WW, 320). Red arrows depict examples of wedge-shaped, hypo-enhancing, cortical lesions consistent with infarcts in the left kidney.

The CT findings supported the clinical neuroanatomical localization of an L4-S3 myelopathy with the most significant extradural cord compression seen at L5–6 on the left. The presence of thickened epidural tissues with increased attenuation, meningeal enhancement, and regional lymphadenopathy and myopathy, together with the history of a fever and inflammatory leukogram, raised spinal epidural empyema (SEE) as the major differential diagnosis for the neurological signs. The presence of renal infarcts raised concern for bacteremia or a pro-thromboembolic state.

The following morning, prior to surgery, blood was collected for an in-house CBC, biochemistry profile, and electrolytes (ProCyte Dx, Catalyst Dx, and VetStat; IDEXX, NSW, Australia). The following abnormalities were detected: white blood cell count [WCC, 29.36 × 10⁹/L; reference range (RR): 5.05 to 16.76 × 10⁹/L], neutrophils (23.10 × 10⁹/L; RR: 2.95 to 11.64 × 10⁹/L), monocytes (2.75 × 10⁹/L; RR: 0.16 to 1.12 × 10⁹/L), platelets (87 × 10³/μL; RR: 148 to 484 × 10³/μL), sodium (161 mmol/L; RR: 144 to 160 mmol/L), potassium (3.4 mmol/L; RR: 3.5 to 5.8 mmol/L), gamma-glutamyl transferase (GGT, 16 IU/L; RR: 0 to 11 IU/L). A blood smear was not assessed to confirm the low platelet count.

The dog was pre-medicated with methadone (Methone; Ceva Animal Health), 0.3 mg/kg BW, IV, and anesthesia was induced with propofol (Propofol Sandoz; Sandoz), 150 mg IV. Anesthesia was maintained with 1% to 2% isoflurane (Isoflurane USP; Piramal Critical Care) in 100% oxygen via an endotracheal tube. Intravenous fluids were provided throughout the procedure. Blood was collected aseptically from the left jugular vein and urine was collected from the indwelling catheter for microbial culture.

The dog was placed in sternal recumbence and a left-sided L5–6 hemilaminectomy was performed. A large amount of red granulation tissue was found within the spinal canal. Samples were taken for microbial culture and histopathology with the remaining abnormal tissue suctioned copiously. No specific complications were noted during the surgery. Treatment with ampicillin (Austrapen; Alphapharm, QLD, Australia), 30 mg/kg BW, IV, q8h, metronidazole (Baxter Healthcare, NSW, Australia), 15 mg/kg BW, slow IV, q12h and enrofloxacin (Baytril 50; Bayer Animal Health), 10 mg/kg BW, IV, q24h, was commenced following acquisition of the tissue samples. Post-operative pain was managed with a fentanyl infusion (DBL Fentanyl Injection; Hospira Australia, Victoria, Australia) at 2 to 4 μg/kg BW per hour, IV, and tramadol (Tramadol Sandoz; Sandoz), 4 mg/kg BW, SQ, q8h.

During the evening following surgery, a cardiac dysrhythmia was noted. Electrocardiography (ECG) revealed a ventricular tachycardia with a ventricular rate of 140 to 160 beats/min (Figure 3). A normal sinus rhythm was present during inspiration and ventricular tachycardia was noted during expiration. No specific anti-arrhythmic therapy was provided, and the tachyarrhythmia resolved over the ensuing 4 d. Two days after surgery echocardiography was conducted to further investigate the arrhythmia and assess for endocarditis. Vegetative lesions were identified on both mitral valve leaflets, with an oscillating lesion on the septal leaflet (Figures 4a, 4b, 4c) that was up to 1.4 × 1.2 cm in the right parasternal 4-chamber view. There was a moderate volume regurgitant jet through the mitral valve. The left atrial to aortic ratio was 1.33 (RR: < 1.6) in the right parasternal short axis view (16). The echocardiogram findings were consistent with a diagnosis of mitral valve endocarditis. Despite the moderate regurgitant volume, a heart murmur was not auscultated and there was no evidence of left-sided congestive heart failure. Following the echocardiogram, an esophageal feeding tube was surgically placed as the dog remained inappetent.

Figure 3.

Lead II ECG: Fusion beat followed by 4 normal QRS complexes and a segment of ventricular tachycardia.

Figure 4.

a and b — Modified right parasternal long axis view of the septal mitral valve leaflet with vegetative lesion (*). c — Right parasternal short axis view of the mitral valve showing vegetative lesion (*).

Both the blood culture and the swab obtained from the spinal canal at surgery returned a heavy growth of Staphylococcus aureus with identical antimicrobial susceptibilities for the 13 antibiotics tested on both samples. The urine culture returned a heavy growth of Pseudomonas aeruginosa presumed to be iatrogenic due to the placement of a urinary catheter in a recumbent dog. All bacteria were susceptible to enrofloxacin, so this was continued as single antibiotic therapy. The histopathology of the epidural adipose tissue (red granulation tissue) revealed subacute fibrinosuppurative inflammation with congestion, hemorrhage, fat necrosis, and mild fibroplasia. No bacteria were visualized within the histopathology samples after Gram and periodic acid-Schiff (PAS) stains.

The dog remained in hospital for 12 d following surgery. He was managed with enrofloxacin (Enrofloxacin 150; Apex Laboratories, NSQ, Australia), 300 mg, q24h, gabapentin (Neurontin; Pfizer, NSW, Australia), 400 mg, q12h, and tramadol (Tramal SR; Seqirus, Victoria, Australia), 200 mg, q8h, all administered orally or via the feeding tube. His appetite returned by 7 d after surgery and the initial signs of systemic infection were all well-controlled. Post-surgery CT was discussed to monitor the evolution of the SEE; however, this was declined by the owner due to fiscal constraints. Unfortunately, motor function to the pelvic limbs had not significantly improved by day 12 and the owners elected to have the dog euthanized. Necropsy was declined by the owners.

Discussion

Spinal epidural empyema is a rare diagnosis in dogs, with only a small number of cases reported in the veterinary literature (1–11). The first case in which mitral valve endocarditis was the likely primary source of infection is reported herein. A detailed description and figures demonstrating the CT findings of SEE, examples of which are also limited in the literature, are provided.

Only 3 previous studies have reported more than 2 cases of SEE in dogs. One study reported case management including decompressive surgery in 7 dogs, 1 reported successful medical management of 5 dogs, with the other study focusing on the MRI findings in 5 dogs (2–4). In most of the reported cases in these 3 studies, and in all of the single case reports, dogs with SEE had a concurrent adjacent infectious process such as discospondylitis, osteomyelitis, or paraspinal abscess (2–4,6–11). There have only been 4 reported cases with distant infections likely contributing to the SEE, including 1 dog each with: pyothorax; pleuritis and pneumonia; hemorrhagic enteritis; and 1 dog with prostatitis, discospondylitis, and SEE all due to E. coli infection (1,2,4). One of the dogs treated successfully with medical management only had a documented urinary tract infection, a problem that is common in dogs that are non-ambulatory due to spinal disease, casting some doubt over the direct link. A primary cause for empyema was not identified in 5 of the 26 reported cases (1–4). In no previous case report or series was echocardiography discussed or reported.

Infectious endocarditis (IE) is also a rare disease in dogs, though it is more widely reported than SEE (12–15). Sykes et al (12) described the largest series of cases, with 71 dogs reported (incidence of 0.05% compared with the hospital population). Endocarditis is often difficult to identify and diagnose before the development of severe clinical implications such as congestive heart failure, immune-mediated polyarthritis or glomerulonephritis, and the sequelae of septic embolization. Due to both septic embolization and formation of immune complexes, multi-organ disease is very common with IE. Of relevance to this report, both brain and spinal neurological deficits have been reported in association with bacterial endocarditis in up to 24% of dogs (12). In the same study, 4 dogs with endocarditis had abnormal cerebrospinal fluid including 1 with a positive culture for Streptococcus canis. An additional 5 dogs were confirmed to have inflammatory central nervous system lesions at necropsy due to both infarction and immune-mediated processes. These findings help support the link between bacterial endocarditis and SEE as seen in the current case. The diagnosis of IE in this case was based on the modified Duke’s criteria (17,18). There was a positive echocardiogram with an oscillating vegetative lesion on the mitral valve together with valvular insufficiency, considered a major criterion. While it is likely that the mitral regurgitation was new, this could not be confirmed as there was no audible murmur present, and echocardiography demonstrating a normal mitral valve had not been performed previously. The absence of a heart murmur was unexpected based on the echocardiographic findings. However, in this case, the lack of murmur may be attributed to inadequate auscultation because of the dog’s large size and the fact he was recumbent during the hospital stay. Alternatively, there may have been no murmur due to the central direction of the regurgitant jet. The dog also met 4 minor criteria including pyrexia; large breed; evidence of thromboembolic disease; and a single positive blood culture with Staphylococcus aureus, a known causative organism. Staphylococcus aureus was also cultured from the epidural space, suggesting endocarditis was causative for the SEE. A study by Curry Jr et al (19) found 100% concordance between the results of blood and epidural abscess cultures when both were positive (n = 24) in human patients, with Staphylococcus aureus the most common causative organism. Additionally, the same study revealed that 10 of 48 patients with epidural abscess had distant infections, 2 of which were endocarditis. Thrombocytopenia, as was seen in this patient, is reported in about 50% of dogs with endocarditis due to immune-mediated destruction or consumption (12). The presence of ventricular tachycardia also adds support to the diagnosis of IE, though arrhythmias are not considered within the modified Duke’s criteria.

Based on the previous reports of SEE in dogs, the clinical signs of disease typically progress rapidly, with spinal pain, fever, and progressive neurological deficits common. Most reported cases had clinical signs for < 7 d before presentation at a referral center, and most dogs had neurological deficits ranging from mild ataxia to severe, non-ambulatory tetraparesis. The case reported here had signs of systemic infection for 2 weeks before presentation, with more than 1 week of illness before the development of signs of spinal disease, suggesting the SEE was likely secondary to bacteremia and septic embolization.

Similar to the current case, most reported dogs with empyema were managed with a combination of surgical spinal cord decompression, broad-spectrum antibiotics, pain relief, and supportive therapies. However, the recent study by Monteiro et al (4) reporting successful medical management of 5 dogs raises some doubt as to the optimal management guidelines for dogs with empyema. Neurological improvement was marked in most of those dogs, including 1 dog which improved from being deep pain negative in the thoracic limbs to residual ataxia and ambulatory paraparesis within 7 d. Complete neurological recovery was noted in 3 of the 4 dogs available for follow-up past hospital discharge. The prognosis for complete neurological recovery with medical management of spinal epidural abscess (the comparable disease process in humans) with severe neurological deficits, is poor (19). Based on the guidelines for surgical intervention in human medicine (20), surgery would have been recommended for most of the dogs in the aforementioned study. These findings highlight that while we can learn from human medicine, the response of dogs to a similar disease process may be vastly different. Further studies are warranted to identify the optimal management approach for dogs with SEE.

Gadolinium-enhanced MRI is considered the imaging modality of choice to diagnose SEE in dogs, and epidural abscess in humans (3,20). Magnetic resonance imaging provides superior contrast resolution compared with CT, and better defines the extent of the disease, spinal cord, and regional soft tissue changes, and alternative pathologies. The sensitivities of both CT myelography and MRI, however, are equivalent in the diagnosis of epidural abscess in human medicine. Despite similar sensitivities, myelography is avoided when possible in human medicine due to the theoretical risk of introducing bacteria into the intrathecal space in cases of lumbar epidural abscess. Myelography also has general risks such as neural damage and seizures. Lumbar intrathecal administration of contrast material was performed in the dog herein without clinical evidence of spread of the infectious process. There have been few reports detailing contrast enhanced CT and CT myelography for SEE diagnosis in dogs, with no studies of sufficient power to offer a sensitivity for the tests. In this case, myelography added little to the diagnostic utility of the contrast-enhanced CT images alone, in part due to variable opacification of both the subarachnoid and epidural spaces. Contrast enhanced CT and CT myelogram descriptions of SEE in dogs have included a hypodense area within the vertebral canal (6), a soft tissue mass that results in loss of the normal low attenuating epidural fat, ring enhancement following IV contrast, and extradural compression of the spinal cord (7). In support of the study by Nykamp et al (7), we similarly identified a locally extensive increase in attenuation of epidural fat and/or effacement of the epidural fat by fluid-to-soft tissue attenuating material, contrast-enhancement of the spinal meninges, and extradural spinal cord compression. The contrast-enhanced CT images also revealed changes in the adjacent sublumbar musculature and mild enlargement of the sublumbar lymph nodes attributed to the local infection and inflammatory response. With MR imaging, increased signal in the paraspinal muscles has been seen in T2W images and post-contrast T1W images (3). While the muscle changes on CT in this case were subtle, they supported the suspicion of inflammatory disease adjacent to the vertebral canal. In addition to the local spinal changes seen in this case, renal infarcts were also identified. The presence of renal infarcts, ventricular tachycardia, and a systemic infection were all contributing factors as to why the dog was eventually investigated for endocarditis. If the abdomen had not been included in the CT study allowing identification of renal infarction, the eventual diagnosis of mitral valve endocarditis may have been missed.

We have described the clinical presentation together with CT and surgical findings in a dog with SEE. Contrast-enhanced CT and CT myelography were highly useful for defining the epidural disease process in the lumbar spine, localizing extradural spinal cord compression which guided surgical intervention, and in documenting evidence of regional inflammation and distant thromboembolic disease. Echocardiography was necessary to confirm the diagnosis of infectious endocarditis, and this was considered the likely primary source of infection in this dog. The findings of a ventricular arrhythmia together with renal infarcts prompted investigation of the heart, whereas the initial finding of SEE did not trigger a search for underlying cardiac disease. Spinal epidural empyema does not appear to occur as a solitary infectious process, and the authors recommend that echocardiography be considered for dogs with SEE when a primary source of infection is not otherwise readily identified. CVJ