Computed tomographic appearance of epidural empyema in a dog

Abstract

The computed tomographic (CT) appearance of epidural empyema in a 5.5-month-old, intact male boxer dog is described. Epidural empyema was diagnosed by means of CT, surgery, and bacterial culture. The imaging and pathogenesis of epidural abscesses and the pitfalls of differentiating caudal fossa lesions from severe cranial cervical lesions are discussed.

A 5.5-month-old, intact male, boxer dog was admitted to the Cornell University Hospital for Animals for evaluation of mild depression and stiffness of 1-day duration that had rapidly progressed to severe neck pain and recumbency prior to referral. Four days prior to admission, the dog sustained a bite wound to the right hip region. Two days later the referring veterinarian detected no substantial abnormality during physical examination and the dog was vaccinated at an undescribed site. At the time of admission, the dog had severe neck pain and fever (40.6°C), and was in lateral recumbency. Neurological examination revealed general proprioceptive deficits in all 4 limbs with the left being side more severely affected than the right, and left-sided hypertonicity. The neuroanatomic localization was to the caudal fossa or cervical spinal cord segments (C1–5). A complete blood cell (CBC) count revealed a mild neutrophilia (15.5 × 10⁹ cells/L; reference range, 3.4 to 9.5 × 10⁹ cells/L) with a left shift and toxic changes, monocytosis (3.4 × 10⁹ cells/L; reference range, 0.1 to 1 × 10⁹ cells/L), and thrombocytopenia (151 × 10⁹ cells/L; reference range, 179 to 483 × 10⁹ cells/L). The results of serum chemical analysis and coagulation panel were within reference range, aside from moderately elevated fibrinogen (4.61 g/L; reference range 1.09 to 260 g/L). Based on the neurological examination and the CBC count, the differential diagnoses included meningitis, severe extradural spinal cord compression (due to intervertebral disc disease, abscess, granuloma, or neoplasia), and a caudal fossa lesion (abscess, granuloma, neoplasia). Cerebrospinal fluid (CSF) collection was planned; however, prior to induction of anesthesia, respiratory arrest occurred and bilateral miosis was detected. Mechanical ventilation was instituted and an emergency CT examination of the head and neck was performed.

Computed tomographic examination of the brain pre- and postcontrast administration, revealed no substantial abnormal finding. Computed tomographic examination of the neck revealed a questionable soft tissue density in the left dorsal aspect of the vertebral canal at C3–4. Computed tomographic myelography confirmed the presence of an extradural compression at this site (Figure 1). No bone or other soft tissue abnormality was detected. Analysis of CSF obtained from the lumbar cistern revealed mildly elevated protein (0.27 g/L; reference range, < 0.25 g/L) and the results of the cytologic examination were normal. The clinical and neurological signs were attributed to this extradural lesion, and the differential diagnoses included empyema, hematoma, cyst, granuloma, and neoplasm.

Figure 1. Transverse CT myelography at the level of the C3–4 intervertebral disk. In the left dorsal aspect of the epidural space, there is a soft-tissue mass (a) that results in loss of the normal low attenuating epidural fat. This lesion also causes attenuation of the ring of contrast medium and extradural compression of the spinal cord (b).

Emergency surgical exploration was performed via a dorsal approach to the vertebral column centered at C3–4. No cutaneous abnormalities were noted in the region; however, a 5-cm oblong cystic structure filled with a clear fluid was found within the SC tissue dorsal to the dorsal spinous processes of C3–4. A second fluctuant structure was located deep to the cystic structure and within the epaxial musculature. Upon dissection of this deeper structure, the wall ruptured and purulent material was expressed. A band of pale, discolored, friable tissue was found extending from the left lateral epaxial musculature toward the left lateral aspect of the vertebral column. A standard dorsal laminectomy was performed and large amounts of friable tissue were removed from the extradural space at C3, dorsal and lateral to the spinal cord. No penetration of the dura was evident. Culture of the fluid from the pockets within the SC tissue yielded pure beta-hemolytic Streptococcus, group G. The patient remained on the ventilator post operatively, but cardiac arrest occurred approximately 1 h after surgery and resuscitation attempts were unsuccessful. On necropsy, widespread petechial and ecchymotic hemorrhages and a paravertebral abscess were observed adjacent to C3–4. Histologic examination of the cervical spinal cord revealed a moderate, multifocal, subacute, suppurative, hemorrhagic meningitis.

Although there is a considerable amount of literature on epidural abscesses in humans, there have been only a few reports in dogs (1,2,3). The prevalence of spinal epidural abscesses in humans is 0.2 to 2 per 10 000 hospital admissions and, in most cases, these are associated with a recent spinal procedure (4,5). Known predisposing factors in humans include diabetes mellitus, trauma, chronic renal failure, neoplasia, immune compromise, corticosteroid administration, skin infections, paraspinal abscesses, IV drug use, and alcoholism (2,4,6,7,8). Infection can occur through direct inoculation, such as epidural injection, epidural catheter placement, or surgical intervention; extension from local infection; lymphatic spread; or hematogenous spread (4,5,6). Cervical spinal epidural abscesses in humans are most often from local extension of osteomyelitis or discitis (6). Clinical signs include fever, spinal pain, weakness, spinal cord dysfunction, and nonspecific malaise (1,3,4,6,7). The progression of clinical signs occurs at variable rates; however, rapid recognition of the disease process is essential for a favorable outcome (6,8,9). Prognosis is proportional to the rapidity of diagnosis and the severity of the clinical signs prior to treatment (4,10). As a result, after neurological deficits develop, it is uncommon for humans to return to a normal neurological status following therapy (5,6,8,9). Additionally, without rapid intervention, clinical signs can progress rapidly to death (6).

Historically, myelography has been used to diagnose spinal epidural abscesses; however, both CT and MRI are now commonly used (5,10). Conventional radiography is frequently unremarkable, because detectable bony changes can take several weeks to develop; therefore, this test is of low sensitivity (6,7). Myelography is excellent for detecting extradural compressive lesions; however, this is not a specific test, because it does not allow the differentiation of epidural abscesses from other extradural lesions (6,7). The advantage of CT is that it has the ability to detect bone changes earlier than does conventional radiography and the cross sectional anatomy is helpful for surgical planning; however, delineation of the spinal cord from the epidural space can be difficult and the contrast resolution of the surrounding soft tissues is poor (6). Computed tomographic myelography increases the diagnostic accuracy and has a sensitivity of 92% in humans (6). The additional use of IV contrast medium can aid in the detection of enhancing lesions both in the epidural space and in the surrounding soft tissues. The reported survey CT findings for spinal epidural abscesses include loss of the low attenuating epidural fat, poorly defined hypodense lesions in the paraspinal region, ring enhancement (post IV contrast medium administration), and gas in the soft tissues (7,11,12). Readministration of IV contrast medium for the cervical CT examination was not performed in this case, because, in our experience, contrast enhancement is detectable for 15 to 20 min and that time frame was not exceeded between the administration of contrast medium for the brain examination and the cervical scan. Magnetic resonance imaging is currently considered the most accurate and least invasive method of diagnosing epidural abscesses and, in most cases, is superior to CT in determining the extent of the lesion due to the excellent contrast resolution (8).

Although the initial clinical signs of neck pain and depression in this dog could have been caused by either a cervical spinal cord or a caudal fossa lesion, the rapidly progressive tetraparesis and respiratory distress in the absence of cranial nerve deficits suggested the former. However, subsequent development of profound depression, progressive miosis, and respiratory arrest raised the index of suspicion for a caudal fossa lesion, brain herniation, or both. Bilateral miosis might have been due to a lesion in the cervical spinal cord (upper motor neuron (UMN) Horner's syndrome) or intracranial disease, because the UMN sympathethic pathway (hypothalamo-tectotegmental spinal system) originates in the diencephalon. Furthermore, brain herniation can lead to miosis because of loss of cortical UMN influences on the oculomotor nucleus. Epidural empyema should be considered as an uncommon differential diagnosis in dogs that are presented with fever and a rapidly progressive myelopathy.

The use of noncontrast CT examination did reveal a loss of the low attenuating epidural fat at C4; however, CT myelography was helpful in confirming that this was due to an extradural lesion. The inability to detect the SC lesions is likely due to the limited contrast resolution of CT. The use of IV contrast medium may have permitted delineation of the SC abscesses detected at surgery. The cause of the epidural empyema was not definitively determined in this patient. Hypotheses include undetected bite wound or previous injection. It is additionally hypothesized that the infection originated in the paraspinal region and tracked along the left C4 nerve root to invade the vertebral canal. A common cause of epidural empyema in veterinary patients is plant awn migration; however, given the age, geographic location, and history of a previous bite wound and injection, this pathogenesis was considered unlikely in this patient. CVJ