Presumptive Subdural Empyema in a Dog

Introduction

Subdural empyema is a focal collection of purulent material between the dura mater and the arachnoid mater. While spinal empyema is usually epidural, intracranial empyema is most commonly subdural due to the limited epidural space in the cranial cavity.^1–9 Subdural empyema has been infrequently reported in small animals (six reports in cats and one report in a dog), with an extremely high mortality rate.^7–9 Of the veterinary reports regarding subdural empyema, only one cat was successfully treated and was doing well at the time of follow-up 4 yr later.⁹ In that case, the diagnosis was made via MRI and decompressive craniectomy followed by cytology and culture.⁹ In all other veterinary reports, the diagnosis had been confirmed via postmortem examination.^7–9

Although uncommon in people, subdural empyema has been extensively described in the human medical literature. Subdural empyema is a progressive, fatal disease in humans if left untreated; however, the mortality rate has significantly decreased with the introduction of antibiotics.¹⁰ MRI is the preferred modality for the diagnosis of subdural empyema in people.^10,11 Decompressive surgery and culture are considered the gold standard in management of either intracranial or spinal subdural empyema in human cases.¹⁰

The purpose of this report is to describe a case of subdural empyema that was successfully treated with antibiotics after a presumptive diagnosis was made based on the clinical presentation and characteristic imaging features.

Case Report

A 13 mo old castrated male golden retriever/standard poodle mixed-breed dog was referred for progressive encephalopathy. The dog had been treated with a 7 day course of amoxicillin trihydrate/clavulanate potassium (16.7 mg/kg per os [PO] q 12 hr) and meloxicam (0.1 mg/kg PO q 24 hr) for a possible tooth root abscess approximately 6 wk earlier. Clinical signs of oral pain improved but recurred 3 wk after finishing antibiotics. Thus, a longer course (2 wk) of antibiotics was prescribed. Despite antibiotics, the clinical signs continued to progress. The dog developed a fever (40.6 C), left-sided facial swelling, exophthalmos, and mucopurulent to serosanguinous discharge from the left eye. The patient was given IV fluids and subcutaneous buprenorphine and enrofloxacin at another hospital before being referred for additional diagnostics. The dog was previously healthy with no significant medical history and was up-to-date on vaccines and flea and heartworm preventative medications.

At the time of referral, body temperature was slightly elevated (39.1 4C); the dog was painful on oral examination, but no abnormalities were noted; both globes retropulsed normally; and the dog was alert but mentally inappropriate and intermittently head pressed. Gait was considered normal; however, there was a compulsive tendency to circle to the left. Postural reactions were significantly delayed in the right thoracic and pelvic limbs with equivocal deficits on the left side. Neuroanatomic localization was the left prosencephalon. The differential diagnoses for neurologic signs included meningoencephalitis, neoplasia (primary brain tumor or round cell tumor), and either intracranial or brain malformation. Retrobulbar abscess (secondary to direct penetrating injury, tooth root abscess, or hematogenous spread), orbital cellulitis, and orbital tumor were considered for the etiology of ocular signs.

Diagnostics

Advanced Imaging

MRI of the brain was performed using a 1.0 T magnet^a. The study was comprised of transverse T1-weighted (T1W), transverse and sagittal T2-weighted (T2W), transverse T2*-weighted (T2*W), and transverse fluid-attenuated inversion recovery (FLAIR) images of the head. T1W images were obtained in the transverse, sagittal, and dorsal plane following IV contrast administration (gadopentetate dimeglumine)^b. There was a large collection of T2 hyperintense signal (maximally 3 mm) along the surface of the entire left cerebrum (Figure 1A), causing compression of the ipsilateral ventricle, falcine shift to the right, and transtentorial brain herniation; the majority of the signal did not suppress on FLAIR images (Figure 1B). The fluid accumulation was slightly hyperintense to ventricular cerebrospinal fluid (CSF) and hypointense compared to white or gray matter on T1W images (Figure 1C). On T2W and FLAIR images, there was an increase in conspicuity of sulci along the left cerebrum potentially due to sulcal widening or meningeal hyperintensity. No abnormal signal void was noted on T2*W images. Following contrast medium administration, there was marked enhancement of the meninges adjacent to the left cerebrum and between the hemispheres. Within the meninges over the left cerebrum, there was an isointense region on T1W images that did not enhance with contrast, suggesting an intrameningeal collection of cell/protein-rich fluid (Figure 1D). The fluid was considered subdural in location based on its appearance as a crescent conforming to the surface of the entire left cerebrum. There was a similar accumulation of fluid between the olfactory bulbs with equally intense enhancement of the meninges (Figure 2A). An ill-defined mass that was hyperintense on T2W images and hypointense on T1W images was noted in the retrobulbar space adjacent to the calvaria. The periphery of the mass enhanced following contrast administration (ring enhancement) with confluent enhancement of the orbital bone and meninges, suggesting a retrobulbar abscess with intracranial extension (Figure 2A). A small amount of gravity-dependent fluid was present within the left frontal sinus. Postcontrast, there was a slight enhancement of the lining of the sinus indicating sinusitis. The MRI findings indicated a marked left-sided subdural fluid accumulation along the cerebrum and between the olfactory bulbs, pronounced meningitis around the subdural fluid accumulation, sinusitis, and retrobulbar fluid accumulation with secondary exophthalmos. The presumptive diagnosis was subdural empyema secondary to a retrobulbar abscess.

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

Discussion

Previous reports of subdural empyema in the veterinary literature are very limited, with six reports in cats and one report in a dog.^7–9 MRI findings have been described in three of those cats but not in a dog.^8,9 In one cat, the diagnosis of subdural empyema based on MRI characteristics was confirmed by decompressive craniectomy followed by cytology and culture.⁹ All other cases were euthanized due to progressive neurologic dysfunction, and diagnosis was established by postmortem examination.^7–9 Foramen magnum brain herniation was identified on MRI in one cat and necropsy in three other cases.^7–9

On MRI, subdural fluid collections are usually found spread widely over the convexities and interhemispherically.^9,10,12 In general, epidural fluid accumulations tend to be either biconvex or lentiform, whereas subdural fluid tends to be crescent-shaped.^13,14 Unlike epidural fluid accumulation, subdural fluid can cross suture lines due to lack of dural tethering by the cranial sutures.^12–14 The intracranial subdural space is lined by a single layer of endothelial cells where it is bounded by the dura mater externally and the arachnoid internally, with no septations other than where the arachnoid granules are imbedded in the dura.¹⁵ Because the subdural potential space is not restricted by the cranial sutures, subdural fluid accumulations may extend over an entire cerebral hemisphere. They are, however, limited by the falx cerebri and tentorium cerebelli and do not cross the midline because of the meningeal reflections.^12–15 Empyema appears slightly hyperintense to CSF and hypointense relative to brain parenchyma on T1W images and either iso- or hyperintense relative to CSF on T2W images.^9–12,16 In contrast, sterile subdural effusion is isointense to CSF on both T1W and T2W images, and chronic subdural hematoma is isointense to brain parenchyma in T2W images.¹⁷ The MRI findings in the current case were consistent with what is seen in humans; however, some areas of fluid accumulation on T2W images were hypointense relative to CSF rather than iso- or hyperintense. That may be due to heterogeneity of the purulent material with variability in the protein and cellular constituents of the fluid. The fluid in subdural empyema is typically surrounded by a contrast-enhancing rim, which is due to formation of a membrane of granulomatous tissue on the leptomeninges and inflammation in the adjacent cerebral cortex.^9,10,12 A chronic subdural hematoma in a dog with ceroid lipofuscinosis was also found to have marked enhancement of the meninges; however, subdural hematoma is readily distinguished from empyema by the presence of signal void on T2*W images.¹⁵ Diffusion-weighted imaging (DWI) may be valuable in evaluating a subdural fluid collection. Similar to brain abscess, empyema has high signal on DWI, most likely because the high viscosity of the purulent fluid restricts proton mobility.^12,16,17 The apparent diffusion coefficient map usually shows decreased diffusion, seen as low signal intensity, within empyemata.^11,16 Subdural hematoma may also appear as an area of high signal on DWI, although signal void on T2*W images is seen in hematomas but not empyemata.^12,16,17 Sterile subdural effusion tends to have low signal on DWI similar to that of CSF.^12,16,17

In the current case, presumptive diagnosis of subdural empyema was based on characteristic MRI features and supported by positive clinical response to antibiotic therapy and resolution on subsequent imaging. A very mild neutrophilic pleocytosis was consistent with localization of infection primarily outside of the subarachnoid space. Decompressive craniectomy for meningeal culture and lavage was considered because that is the treatment of choice in humans; however, the owner of the dog had reservations about the cost and risks associated with cranial surgery. In veterinary cases, surgical treatment may not be pursued due to client concern regarding the invasiveness of the procedure and associated risks, unavailability of appropriate equipment, and cost. In this case, decompressive craniectomy was recommended because surgery tends to provide a more predictable and favorable outcome in humans with subdural empyema. Pending the client’s decision regarding surgery, the dog subsequently made a profound clinical response to antibiotics and craniectomy was, therefore, deemed unnecessary.

The most common bacteria associated with meningitis or meningoencephalitis in dogs and cats are Staphylococcus spp., Pasteurella multocida, Nocardia spp., and Actinomyces spp., and various anaerobic species.^18–20 In contrast, cultures from subdural empyemata in prior veterinary case reports most commonly found polymicrobial infection with at least one anaerobic bacteria (Fusobacterium spp., Bacterioides spp., and Actinomyces spp.).^7–9 Culture of CSF is rarely useful in identifying the causative agent for bacterial meningoencephalomyelitis in dogs and cats.¹⁸ In this case, CSF culture was not submitted because of the marginal changes on CSF analysis and the subdural location of infection.

Cefotaxime, a third-generation IV cephalosporin, was initially chosen because of its broad-spectrum activity against gram-positive and gram-negative bacteria, including resistant bacterial isolates, along with its capacity to cross the blood-brain barrier.²¹ Cefpodoxime is the only third-generation cephalosporin with an oral formulation available for use in small animals. Cefpodoxime has a spectrum of activity that resembles a first-generation cephalosporin along with a more limited capacity for penetration into the central nervous system and was, therefore, not considered adequate for treatment of an intracranial infection.²² A combination of clindamycin and enrofloxacin was chosen for oral therapy because both drugs cross the blood-brain barrier and, when administered together, provide excellent broad-spectrum antimicrobial activity. Enrofloxacin was discontinued after 1 mo because an aerobic gram-negative bacterial infection was considered less likely based on previous veterinary case reports.^7–9 Clindamycin was chosen for long-term antibiotic therapy due to its affordability and activity against gram-positive bacteria and anaerobes. Subdural fluid presumed to be empyema had resolved on MRI after 10 wk of antibiotic therapy; however, there was mild meningeal enhancement and hyperostosis of the calvaria. It was unclear if those changes represented either mild osteomyelitis/meningitis or were reactive to prior infection; therefore, clindamycin was continued for an additional 6 wk. The total length of treatment was 4 mo, which is comparable to treatment times reported in human cases of subdural empyema that were successfully treated with medical therapy alone.²³ The length of therapy was also comparable to the minimum recommended treatment time for other forms of bacterial osteomyelitis (e.g., vertebral osteomyelitis, diskospondylitis).^23,24 MRI was not repeated after discontinuation of clindamycin because of the cost of reimaging.

Subdural empyema in humans typically develops as direct extension of paranasal sinusitis or otitis media/interna, but other possible causes include hematogenous spread from distant sites, cranial surgery, trauma, retrograde septic thrombophlebitis, or secondary infection of subdural effusion or hematoma.^10,12 Common sources of infection in either meningitis or meningoencephalitis are similar to those implicated in subdural empyema and include otitis interna, tooth root abscesses, retrobulbar abscesses, and sinusitis; however, hematogenous spread from extracranial source such as endocarditis, pneumonia, and prostatitis has also been reported.^7–9,19,20 In cases of subdural empyema secondary to sinusitis, the infection spreads intracranially through thrombophlebitis of the emissary veins, which drain the external skull into the dural sinuses.¹² Direct extension of the infection may also occur through the Haversian canals within bones of the skull. In small animals, subdural empyema developed secondary to direct inoculation from a puncture wound in one cat and from local extension from otitis media/interna in another cat.^8,9 The underlying etiology was undetermined in the remaining cases.^7–9

In the current case, a retrobulbar abscess with intracranial extension was considered the most likely source of intracranial infection. Retrobulbar abscess can result from a penetrating injury, tooth root abscess, infections in sinuses or zygomatic glands, or hematogenous spread.²⁵ There is limited literature on orbital infection with intracranial extension in animals. In a recent study of four dogs with evidence of orbital inflammation with intracranial extension on MRI, retrobulbar abscess was found in three of the four dogs, and thickened periorbital tissue with contrast enhancement was found in close anatomic relationship to cavernous sinus in two of the dogs.²⁶ Intracranial extension of retrobulbar blastomycosis has also been reported in a dog in which infection followed the optic canal and orbital fissure into the ventral aspect of diencephalon.²⁷ In humans, intracranial extension of orbital inflammation is most commonly established via the orbital fissure subsequently entering the middle cranial fossa and cavernous sinus.^28,29 Sinusitis, which is the most common cause of subdural empyema in children, was noted in the current case, although the degree of sinusitis was very mild.^10–12 Retrobulbar abscess with intracranial extension was considered more likely based on the confluence of enhancement between the retrobulbar space and meninges.

Conclusion

This report describes the clinical presentation and MRI findings of presumptive subdural empyema in a dog with resolution following antibiotic therapy.