A Case of Canine Streptococcal Meningoencephalitis Diagnosed Using Universal Bacterial Polymerase Chain Reaction Assay

Introduction

Bacterial infections of the central nervous system (CNS) occur in several different forms (e.g., encephalitis, meningitis, meningoencephalitis).¹ The mortality rate (87%) associated with these forms is very high in dogs, and many dogs that recover are left with permanent neurological sequelae (e.g., circling, ataxia, blindness, head tilt, proprioceptive deficits, cranial nerve palsies, seizures).^2–8

Much of the mortality associated with these diseases arises from the difficulty in diagnosing bacterial CNS infections. Antemortem diagnosis has been based on culture or cytology that demonstrates bacteria within the cerebrospinal fluid (CSF).^9–11 Bacterial cultures appear to be fairly sensitive in people, with an estimated 71% to 95% positive cultures in cases of bacterial meningoencephalitis.^12,13 In contrast, CSF cultures are positive in <20% of dogs with histopathologically confirmed intracranial bacterial infections.⁴ Direct visualization of bacteria on cytological examination of CSF is rare.¹²

Several factors may be responsible for the low rate of positive cultures from dogs. The relatively small size of many dogs limits both the volume of CSF that can be collected and the ability to detect etiological organisms that generally are found in low numbers in the CSF with bacterial CNS infections.¹ Some organisms also grow very slowly or are difficult to culture. Furthermore, dogs often receive antibiotics before the sample is collected, which decreases the chance of a positive culture.¹⁴

Cerebrospinal fluid samples from dogs with bacterial CNS infections typically have elevated white blood cell counts (WBC), neutrophils, and protein concentrations.¹ However, these characteristic changes are not seen in many cases of CNS bacterial infection.¹

Polymerase chain reaction (PCR) assays using bacterial primers universal to all eubacteria may be useful in diagnosing bacterial CNS infections, because such assays are less influenced by the above factors.⁹ The gene for the 16S ribosomal subunit contains conserved segments of deoxyribonucleic acid (DNA) common to all bacteria, as well as divergent sequences unique to each genus or species of bacteria.¹⁵ A small piece of this sequence that is complementary to a conserved region can act as a “universal primer” to nonselectively amplify any bacterial DNA in a PCR assay.¹⁶ Once the DNA has been amplified, the PCR product is then stained with ethidium bromide and visualized by electrophoresis on agarose gel.¹⁷ Amplified DNA can be directly sequenced for identification of any known bacteria (including those found in CNS infections) that are available for comparison with the sequenced DNA.¹¹

In human bacterial meningitis, universal bacterial PCR has a reported sensitivity of 86% to 100%, a specificity of 97% to 100%, a positive predictive value of 80% to 94%, and a negative predictive value of 98% to 100%.^9,11 Several human studies have evaluated this technique for identifying various bacterial species. These studies have shown that universal primers are able to identify most genera of eubacteria, including those commonly found in canine bacterial CNS infections.¹⁶

The authors’ case report describes the use of universal bacterial PCR for identifying CNS infection in a dog with consistent clinical signs and laboratory data but negative CSF cultures.

Case Report

A 3-year-old, spayed female, mixed-breed dog was presented to The Ohio State Veterinary Emergency Service with a 5-day history of dysphagia and lethargy. The dog was treated with injectable penicillin and methylprednisone 1 day prior to presentation by the referring veterinarian. On examination, the dog was drooling and had complete loss of tone and motor function of the jaw. Body temperature (38.6°C) was normal. Radiographs of the skull were unremarkable, and a diagnosis of bilateral trigeminal nerve paralysis was made. An esophagostomy tube was inserted, and the dog was discharged.

The dog was presented to the emergency service 3 days later for severe lethargy and tremors. On physical examination, rectal temperature was 39.5°C, and there were bilateral trigeminal and facial nerve pareses. A complete blood count (CBC) was normal. Serum biochemical abnormalities included mildly increased alanine amino-transferase ([ALT] 67 IU/L, reference range 10 to 55 IU/L), aspartate aminotransferase ([AST] 104 IU/L, reference range 12 to 40 IU/L), and creatine kinase (3012 IU/L, reference range 50 to 400 IU/L). No abnormalities were detected on thoracic radiography or abdominal ultrasonography.

The dog became obtunded over the course of the day, presumably from increased intracranial pressure. Mentation improved after treatment with dexamethasone sodium phosphate^a (0.2 mg/kg intravenously [IV]), clindamycin^b (10 mg/kg IV), and mannitol^c (500 mg/kg IV over 15 minutes, then 250 mg/kg IV over 15 minutes).

Cerebrospinal fluid was obtained from the cisterna magna under general anesthesia. The CSF was colorless and hazy, with an increased protein concentration (174 mg/dL, reference range 0 to 25 mg/dL). The CSF also showed neutrophilic pleocytosis (387 WBC/μL, reference range 0 to 5 WBC/μL), with 71% nondegenerate neutrophils, 14% large monocytes, and 15% lymphocytes. The cytological findings suggested infectious encephalitis, although immune-mediated encephalitis could not be excluded from the differential diagnosis.

The dog was treated with clindamycin (10 mg/kg IV q 12 hours), doxycycline^d (10 mg/kg IV q 12 hours), ceftriaxone^e (22 mg/kg IV q 12 hours), dexamethasone sodium phosphate (0.1 mg/kg IV q 24 hours), and IV-administered fluids. ^f The dog’s mentation and activity level improved substantially overnight. An ophthalmological examination the next day was normal.

Intermittent vomiting developed, prompting a second CBC and serum biochemical profile on the 4th day of hospitalization. Thrombocytopenia (67,200 cells/μL, reference range 106,000 to 424,000/μL) and increased levels of ALT (742 IU/L), AST (309 IU/L), alkaline phosphatase ([ALP] 2222 IU/L, reference range 15 to 120 IU/L), and total bilirubin (3.99 mg/dL, reference range 0.1 to 0.4 mg/dL) were detected. Bile acid concentrations were markedly increased before (449 μmol/L, reference range <15.5 μmol/L) and after eating (319 μmol/L, reference range 5 to 20 μmol/L). A coagulation profile was normal.

Ceftriaxone was discontinued, because the drug has been reported to cause cholestatic disease in dogs and thrombocytopenia in people.¹⁸ Cefpodoxime^g (5 mg/kg per os [PO] q 12 hours), vitamin K^1h (0.5 mg/kg PO q 24 hours), metoclopramideⁱ (0.04 mg/kg per hour IV), and S-adenosylmethionine ([SAM-e^j] 225 mg PO q 24 hours) were added to the treatment regimen.

On day 9, the vomiting resolved and the dog ate well. Bacterial cultures of the blood, urine, and CSF were negative. Serum total bilirubin concentration had decreased to 1.90 mg/dL. The dog was discharged from the hospital on clindamycin (10 mg/kg PO q 12 hours), doxycycline (5 mg/kg PO q 12 hours), cefpodoxime (5 mg/kg PO q 12 hours), SAM-e (225 mg PO q 24 hours), dexamethasone (0.1 mg/kg PO q 24 hours), vitamin K¹ (0.5 mg/kg PO q 24 hours), and metoclopramide (0.2 mg/kg PO q 12 hours).

Three days after discharge, the dog was again presented to the emergency service with tetraparesis, truncal ataxia, head pressing, a left head tilt, ptosis in the left eye, and a fever (39.2°C). A platelet count done at that time was 38,000/μL. The following day, the ataxia and head pressing were more pronounced, and the dog began circling to the right and had two generalized, tonic-clonic seizures. The dog was treated with phenobarbital^k (10 mg/kg IV followed by 1.4 mg/kg PO q 12 hours), ampicillin^l (22 mg/kg IV q 8 hours), and enrofloxacin^m (10 mg/kg IV q 24 hours). Dexamethasone, SAM-e, and vitamin K¹ were discontinued.

The dog slowly improved over several days but then developed more seizures, along with hyphema and panuveitis in both eyes. A CBC showed anemia (hematocrit 29%, reference range 36% to 54%), with a reticulocyte count of 74,100/μL (reference range <60,000/μL) and persistent thrombocytopenia (61,000/μL). A Coombs’ test was positive at a 1:32 dilution. Serum liver enzymes had improved (ALT 84 IU/L, AST 34 IU/L, ALP 544 IU/L), and total bilirubin concentration (0.34 mg/dL) was normal. Levetiracetamⁿ (10 mg/kg PO q 12 hours), prednisone^o (one drop in each eye q 6 hours), and atropine^p (one drop in each eye q 24 hours) were added to the therapy. Serum titers for Neospora caninum, Toxoplasma gondii, Ehrlichia canis, Histoplasma capsulata, Blastomyces dermatitidis, and Cryptococcus neoformans were negative.

Because the dog’s clinical findings and CSF cytology results were most consistent with infectious encephalitis, but immune-mediated encephalitis could not be reliably eliminated from the differential diagnosis, a PCR using universal bacterial primers was performed on pretreatment CSF that had been stored under sterile conditions. Deoxyribonucleic acid was extracted from the sample using routine proteinase-K procedures.¹⁷ Polymerase chain reactions were prepared using universal bacterial primers (27f 5′-AGAGTTTGATCMTGGCTCAG and 1525r 5′-AAGGAGGTGWTCCARCC), positive and negative controls, and an internal control for DNA extraction. Reactions were amplified in a thermal cycler.^q (The amplification process consisted of initial denaturation at 94°C for 5 minutes, followed by 30 cycles of denaturation at 94°C for 30 seconds, annealing at 58°C for 30 seconds, extension at 68°C for 30 seconds, and a final extension at 68°C for 7 minutes.) Reaction products were visualized on 1% agarose gel stained with ethidium bromide.

A 1500 base-pair segment of DNA was identified from the dog’s CSF sample. The PCR product was gel-purified and submitted for direct genetic sequencing^r. Results were compared with bacterial sequences in the Genbank sequence database, and the DNA recovered from the dog’s sample was identified as a species of Streptococcus [see Figure].

The dog was then treated with amoxicillin^s (50 mg/kg PO q 12 hours) for 4 months. Follow-up CSF evaluation to monitor therapy was declined by the owner. The CBC and serum biochemical profiles were normal by day 22. Neurological and ophthalmological signs completely resolved by day 38. Although the dog’s brain-stem signs resolved, seizures continued, requiring a combination of phenobarbital (6 mg/kg PO q 12 hours) and potassium bromide^t (26 mg/kg PO q 24 hours) for control.

Discussion

A definitive diagnosis of bacterial encephalitis could not be made in this dog, because bacteria were not identified in CSF using cytology or bacterial culture. However, the neutrophilic pleocytosis and response to antibiotic therapy were consistent with bacterial encephalitis. In addition, the streptococcal DNA recovered from the CSF, as well as the resolution of most clinical signs after treatment with an antibiotic effective against Streptococcus spp., were supportive of a bacterial infection. Immune-mediated disease was unlikely, because this would not have resolved without prolonged immunosuppressive therapy.¹⁹ The negative CSF cultures may have resulted from missing growth requirements of the specific organism and/or from the antibiotic therapy used before the CSF sample was collected.²⁰

Ceftriaxone, a third-generation cephalosporin, was initially chosen for its broad spectrum of bactericidal activity and good penetration into the CNS.¹⁸ Side effects associated with ceftriaxone administration in people include thrombocytopenia, biliary sludging (from dosages >100 mg/kg per day), and elevated liver enzymes.¹⁸ The package insert^u also indicates that calcium salts of ceftriaxone can form concretions in the gallbladder of treated dogs.^u The authors attributed all non-neurological signs in this case to ceftriaxone. In addition, when cephalosporins were withdrawn, the liver enzyme elevations, anemia, and thrombocytopenia resolved, supporting this conclusion.

Ampicillin, which has fewer adverse effects and is less expensive than ceftriaxone, would have been given initially if Streptococci spp. had been identified earlier. Penicillins are the drugs of choice for streptococcal infections of the CNS, although some Streptococci spp. are also susceptible to third-generation cephalosporins.¹ This susceptibility pattern may be the reason this dog initially responded to treatment with ceftriaxone and then relapsed when the antibiotic was changed to cefpodoxime. Also, the cefpodoxime might not have penetrated into the CNS, or perhaps the dosage was inadequate. Penetration of cefpodoxime into the CNS is generally poor, although CNS concentrations are thought to be above the minimum inhibitory concentration for most pathogens.²¹

An unusual aspect of this case was the trigeminal nerve involvement. Trigeminal nerve deficits are uncommon in people and dogs with bacterial encephalitis, but Streptococcus pneumoniae has been reported to cause dysphagia and infection of the trigeminal ganglion in chimpanzees and mice.^22–25 Evidence has also shown that encephalitis can be associated with trigeminal neuropathy in dogs. In a study of 29 dogs with trigeminal neuropathy, CSF abnormalities were found in seven of eight dogs in which CSF analysis was performed.²⁶

Polymerase chain reaction techniques using universal bacterial primers may be able to overcome the diagnostic limitations associated with CSF bacterial cultures. Polymerase chain reaction assays do not depend on the culture requirements of the bacteria, can amplify DNA from living or dead bacteria, and may be less influenced by previous drug therapy.⁹ Each bacterium also has several copies of the 16S ribosomal ribonucleic acid (RNA) gene, so the theoretical sensitivity may be as low as one organism for certain bacteria.¹⁵ In addition, PCR requires only a small sample volume (approximately 100 μL), and results are typically available sooner than results of bacterial cultures, especially with slow-growing bacteria.⁹

Limitations to universal bacterial PCR must be considered. Strict laboratory and sample collection procedures are required, because sample contamination can easily lead to false-positive results.¹¹ This technique is appropriate only for single-agent infections and cannot be used in mixed infections or when samples contain normal flora.⁹ Furthermore, PCR assays do not provide information about susceptibility to antibacterial drugs, so specific antibiotic treatment must be chosen empirically.⁹ Universal bacterial PCR may also be unable to differentiate bacteria beyond the genus level,²⁷ especially when bacterial species have closely related 16S ribosomal subunit genes (e.g., Streptococcus spp.).¹⁵ This is why the specific Streptococcus sp. could not be identified in this case.

Even with these limitations, the advantages of universal bacterial PCR likely outweigh its disadvantages. Further studies are required to evaluate the diagnostic performance of universal bacterial PCR for canine CNS infections and to optimize assay protocols for use in dogs.

Conclusion

Universal bacterial PCR was used to diagnose streptococcal infection in the CSF of a young dog with multifocal neurological signs. The dog experienced severe side effects from ceftriaxone therapy but recovered when treatment was changed to prolonged administration of amoxicillin. Further studies are warranted to evaluate the diagnostic performance of this assay in dogs with encephalitis, particularly in those with negative bacterial cultures.

Acknowledgments

The authors thank Dr. Elaine Hughes for her contributions to this animal’s care.