Use of Linezolid to Treat MRSP Bacteremia and Discospondylitis in a Dog

Introduction

Methicillin-resistant Staphylococcus infections are challenging diseases to treat. Often, the only suitable antibiotics are not commonly used, and there is frequently a lack of evidence for their efficacy in companion animals. The following case demonstrates the successful use of linezolid to treat methicillin-resistant Staphylococcus pseudintermedius (MRSP) bacteremia, bacteriuria, and discospondylitis in a dog. The dose regimen was extrapolated from pharmacokinetic (PK) studies, and designed to reach adequate plasma concentrations to inhibit further growth of the organism isolated in this dog. Plasma drug linezolid levels were also measured to assess the efficacy of the prescribed linezolid dosage in reaching targeted drug concentrations.

This case also exemplifies the ongoing debate regarding proper antibiotic use in animals, particularly regarding “tertiary use” drugs such as linezolid. Despite the controversy, there are no uniform prescribing guidelines for tertiary antibiotic use in veterinary medicine. The purpose of this case report is to provide some clinical evidence regarding drug dosage, expected plasma concentrations, and possible side effects of linezolid.

Case Report

A 1.5 yr old male intact German shepherd dog weighing 33 kg was evaluated at the University of Wisconsin Veterinary Medical Teaching Hospital for progressive fever and lethargy over the previous 3 mo. The patient had been diagnosed with steroid-responsive meningitis arteritis 4 mo earlier and was currently receiving 0.45 mg/kg/day of prednisone per os (PO), which had been tapered from an initial dose of 2 mg/kg/day. In the 3 mo prior to evaluation, the patient had developed recurrent episodes of fever and leukocytosis, for which he was prescribed multiple courses of empirical antibiotic therapy (5 mg/kg enrofloxacin PO q 24 hr for 10 days, 15 mg/kg ciprofloxacin PO q 12 hr for 10 days, 8.6 mg/kg metronidazole PO q 12 hr for 7 days, ampicillin Na/sulbactam Na^a 17/8.6 mg/kg IV once, amoxicillin trihydrate/clavulanate potassium^b 10.3/2.6 mg/kg PO q 12 hr for 10 days, and 9.4 mg/kg clindamycin PO q 12 hr for 14 days); however, no diagnostics were performed to investigate the source of the fever. Two mo prior to presentation, azathioprine was added at 1.7 mg/kg PO q 48 hr due to the concern for relapsing steroid-responsive meningitis arteritis. Despite the numerous antibiotics administered, the patient continued to have recurrent episodes of fever and lethargy that typically improved within 2–3 days.

A complete blood count and biochemical analysis were performed by the primary care veterinarian 1 wk prior to presentation. The total white blood cell count was elevated (26 × 10⁹/L; reference range, 6–17 × 10⁹/L) due to a mature neutrophilia (22.5 × 10⁹/L; reference range, 3–12 × 10⁹/L ). The platelet count was normal. The biochemical analysis demonstrated increased alkaline phosphatase (8.4 μkat/L; reference range, 0.34–2.55 μkat/L) and alanine aminotransferase (ALT) activities (6.99 μkat/L; reference range, 0.17–2.01 μkat /L), and increased blood urea nitrogen (12.9 mmol/L; reference range, 2.5–8.9 mmol/L). Urinalysis revealed a pH of 7.5, specific gravity was 1.025, there were 1–5 white blood cells/high-power field, and 20–50 red blood cells/high-power field, with no bacteria observed.

At the time of referral, the dog was febrile at 39.5°C, and mentally dull but responsive. Postural reactions were normal in the thoracic limbs, but absent in the pelvic limbs. A mild paraparesis and ataxia were also present. Spinal reflexes were normal. Marked pain was elicited upon midthoracic spinal palpation. Cranial nerve examination was normal. The remainder of the physical examination was within normal limits.

Neuroanatomic localization was supportive of a third thoracic vertebra to third lumbar vertebra myelopathy. Spinal radiographs were obtained and demonstrated collapse of the seventh the eighth thoracic vertebral disc space, with endplate sclerosis and vertebral body lysis of both the seventh and eighth thoracic vertebrae, suggestive of discospondylitis (Figure 1). To identify the causative organism of the discospondylitis, a urine sample was collected via cystocentesis and submitted for aerobic culture and susceptibility, and blood was collected for culture using aseptic technique from the left jugular and right medial saphenous veins (20 mL/site), and inoculated into aerobic and anaerobic blood culture bottles^c. Because the patient’s clinical signs had not changed since evaluation by the referring veterinarian 1 wk earlier, a complete blood cell count and biochemical analysis were not repeated. The patient was discharged from the University of Wisconsin Veterinary Medical Teaching Hospital with a prescription for cephalexin (22 mg/kg PO q 12 hr) pending final culture results. The dose of prednisone was decreased to 0.3 mg/kg PO q 24 hr, with the intention of tapering and discontinuing that medication over the subsequent few weeks. The azathioprine was promptly discontinued.

View Figure

• Download as Powerpoint
• Download Figure

The patient’s urine and aerobic blood cultures both grew > 100,000 colony-forming units/mL of a β-hemolytic Staphylococcus spp., belonging to the MRSP group. The isolates shared the same susceptibility panel. Based on Clinical and Laboratory Standards Institute (CLSI) standard antibiotic minimum inhibitory concentration (MIC) breakpoints, both isolates were resistant to all reported antibiotics (i.e., chloramphenicol, ciprofloxacin, clindamycin, erythromycin, gentamicin, oxacillin, penicillin, rifampin, tetracycline, trimethoprim/sulphamethoxazole), except nitrofurantoin. Methicillin resistance was determined due to both isolates’ resistance to oxacillin. Because approximately 50% of the total dose of nitrofurantoin is excreted in its active form into the urine, nitrofurantoin would have likely been efficacious to treat the MRSP bacteriuria, but likely ineffective to achieve adequate concentrations within blood and bone to treat discospondylitis.¹ That is why nitrofurantoin is not indicated for the treatment of any vancomycin-resistant Enterococcus spp. infections other than lower urinary tract infections.² Therefore, MICs were requested for additional antimicrobials.

MIC breakpoints were determined via serial broth dilution for tigecycline, quinupristin/alfopristin, vancomycin, daptomycin, and linezolid, with CLSI susceptibility standards extrapolated from human performance standards.³ An MIC was reported for tigecycline; however, that drug had not been included in the CLSI standards since 2009. Therefore, its efficacy was uncertain. Both MRSP isolates were reported to be sensitive to quinupristin/dalfopristin, vancomycin, and linezolid based on the published human breakpoints. Susceptibility for quinupristin/dalfopristin applies only to methicillin-susceptible Staphylococcus aureus, not to MRSP, and vancomycin has published standards against Staphylococcus aureus and coagulase-negative Staphylococcus spp., but not to Staphylococcus pseudintermedius.³ Because those antibiotics did not have MIC data for Staphylococcus pseudintermedius or even a genus-wide Staphylococcus susceptibility, their efficacy was also uncertain. In addition, quinupristin/dalfopristin and vancomycin both require IV administration. Prolonged IV catheterization posed a risk for creating a nidus of infection due to the MRSP bacteremia, and the long-term duration of therapy required to treat discospondylitis made IV antimicrobial therapy impractical. An MIC was also reported for daptomycin, which was relevant for genus-wide Staphylococcus susceptibility; however, only IV pharmacokinetic data for that drug were available for dogs.^3,4

As for linezolid, which is in the oxazolidinone class of antibiotics, the MIC data were applicable to genus-wide Staphylococcus susceptibility, which was relevant to the dog’s infection.³ Importantly, linezolid can be administered orally, and pharmacokinetic data for oral administration were available in dogs.⁵ Therefore, that drug was chosen for treatment. The recommended dose of linezolid in humans is 600 mg PO q 12 hr (approximately 8.5 mg/kg q 12 hr for a 70 kg person), with reported mean peak steady state concentrations of 21 μg/mL.⁶ Multiple dose administration of linezolid in dogs (20 mg/kg of linezolid PO q 12 hr) resulted in plasma peak and trough concentrations of 19.7 μg/mL and 2.14 μg/mL, respectively, with steady state reached by at least 15 days of dosing.⁵ This dose in dogs was predicted to provide plasma concentrations greater than twice the reported MIC for the MRSP isolated in this dog (≤ 1 μg/mL), even at trough concentrations. Based on those calculations, the dog was started on linezolid^d at 20 mg/kg PO q 12 hr and the cephalexin was discontinued.

Within 2 wk of initiation of treatment with linezolid, clinical signs of pain and lameness were completely resolved. Further clinical and biochemical monitoring in the dog was based on reported side effects of linezolid in humans, which included myelosuppression (i.e., anemia, leukocytosis, thrombocytopenia), lactic acidosis, serotonin syndrome, convulsions, and gastrointestinal upset.⁶ A complete blood count was rechecked 1 mo into therapy and revealed a moderate normocytic, normochromic anemia (hematocrit was 31.5%; reference range, 37–55%) and thrombocytopenia (slide estimated to be between 50 and 120 × 10⁹/L; reference range 164–510 × 10⁹/L). The previously observed increases in alkaline phosphatase and blood urea nitrogen had normalized, and the ALT had improved (2.36 μkat/L; reference range, 0.085 to 1.82 μkat/L).

Three wk later (7 wk into therapy), another complete blood count showed a persistent but mild thrombocytopenia (114 × 10⁹/mL; reference range, 175–500 × 10⁹/mL) with an increased mean platelet volume (15.4 fL; reference range, 7.9–14.4 fL). The total white blood cell count was normal, with a continued mild normocytic normochromic anemia (hematocrit was 34%; reference range, 40–59%). Because linezolid had been associated with lactic acidosis in human patients, a venous blood gas was also performed.⁶ That test demonstrated a normal blood pH (7.41; reference range, 7.38–7.45), with a compensated lactic acidosis. Lactate was 4.9 mmol/L (reference range, 0–2.9 mmol/L), the partial pressure of CO₂ was 22.7 mm Hg (reference range, 26–40.1 mm Hg), bicarbonate was 14.7 mmol/L (reference range, 20–24 mmol/L), and the base excess was −10.1 mmol/L (reference range, −4–2 mmol/L). The patient had not been involved in any strenuous exercise for at least 24 hr prior to testing and was clinically hydrated; therefore, the compensated lactic acidosis was considered to be a possible linezolid side effect.

At the same 7 wk recheck appointment, peak and trough serum linezolid levels were submitted to the National Jewish Health Laboratory^e to assess the linezolid dose regimen. The trough level, obtained 12 hr after the previous dose (6.3 μg/mL), was within the expected steady-state range established for humans (2–9 μg/mL)^f. The measured peak serum concentration, obtained 2 hr after chronic oral dose administration, was 9.6 μg/mL, which was slightly below targeted peak concentrations in humans (12–26 μg/mL). However, that concentration was still predicted to be effective in treating the MRSP isolated from the patient based on the isolates’ MIC of ≤ 1 μg/mL.⁵ The serum concentrations observed in the patient were above the MIC at both trough and peak concentration times, correlating to a time > MIC index of 100%.

Blood work was repeated 14 wk after initiation of linezolid therapy. The complete blood count showed resolution of the anemia and thrombocytopenia, despite no changes in linezolid dosing regimen. A serum biochemical analysis showed only a static, mildly increased ALT (2.29 μkat/L; reference range, 0.085–1.82 μkat/L), and the venous blood gas showed resolution of the compensated lactic acidosis (lactate was 1.8 mmol/L; reference range, 0–2.9 mmol/L). At each of those follow-up appointments, the patient was afebrile and strongly ambulatory.

Spinal radiographs were performed 8 wk and 14 wk into treatment with linezolid and showed static vertebral lysis with progressive bone remodeling consistent with healing discospondylitis.⁷ Linezolid treatment was continued for a total of 23 wk. The decision to discontinue antibiotics was based on complete recovery of clinical signs of discospondylitis and static radiographic changes observed at the site of vertebral infection at 23 wk (Figure 2). Blood and urine cultures were found to be negative at 2 wk, 12 wk, and 32 wk after discontinuation of linezolid therapy. The patient remained asymptomatic 36 mo after drug discontinuation.

View Figure

• Download as Powerpoint
• Download Figure

Discussion

This case represents the first report of discospondylitis attributed to MRSP in a dog, and highlights the risks of empirical use of multiple antibiotics, without a definitive diagnosis, in potentially selecting for antibiotic-resistant strains of bacteria. In a previous report, two dogs developed iatrogenic methicillin-resistant Staphylococcus aureus (MRSA) discospondylitis following hemilaminectomy for thoracolumbar intervertebral disc protrusion.⁸ However, both MRSA isolates were sensitive to other commonly used first line veterinary antibiotics. The original source of infection in the dog in this report was unknown and was complicated by the 3 mo delay between the onset of clinical signs and diagnosis of bacteremia. Prior to the initiation of antimicrobial therapy in febrile animals, an appropriate work-up is necessary to localize and further characterize the presence of an actual bacterial infection, particularly in patients that are receiving immunosuppressive therapy. The administration of six different antibiotics to this patient within the 2 mo preceding the diagnosis of bacteremia likely contributed to the selection of antibiotic resistant bacteria. Carriage of MRSP in the nasopharynx and skin of dogs is associated with previous hospitalization and antibiotic administration (and to a lesser extent, previous glucocorticoid therapy); however, not all dogs carrying the organism will develop clinical infections.⁹ The patient in this report had all three risk factors.

In humans, the comparable disease, infectious spondylodiscitis, is typically diagnosed by a combination of imaging and blood culture results. The radiographic findings and positive blood and urine culture results in the dog described in this report led to the clinical diagnosis of discospondylitis due to MRSP, similar to accepted guidelines in humans.¹⁰ MRI is considered the best imaging modality in humans, with a sensitivity and specificity of 96% and 93%, respectively.¹¹ However, radiographs and clinical response were consistent with the diagnosis in the patient described herein, so MRI was not pursued. Computed tomographic-guided vertebral biopsy for tissue culture can be performed in humans if blood cultures are negative.¹² Fluoroscopic-guided percutaneous fine-needle aspiration of the intervertebral disc space has been reported in dogs, with positive cultures obtained in 75% of samples; however, that approach was not considered necessary given the dog’s positive blood and urine cultures.¹³

Standard therapy for MRSA spondylodiscitis in humans involves initial treatment with IV antibiotics (typically vancomycin) for 2–8 wk, followed by oral therapy (i.e., linezolid, ciprofloxacin, rifampin) for another 2–12 wk; however, specific guidelines for antibiotic duration have not been established.¹⁰ Reported criteria for discontinuation of antibiotic treatment in humans include resolution of clinical signs, normalization of acute phase proteins (such as C-reactive protein), or a reduction in erythrocyte sedimentation rate.¹⁴ A retrospective study of dogs with discospondylitis described a mean antibiotic duration of 54 wk, wherein those dogs documented to have Staphylococcus infections had a mean duration of 74 wk.¹⁵ However, those dogs were treated until no further radiographic changes were observed, a treatment endpoint that is not used in humans. Reported radiographic signs of infection resolution in canine discospondylitis include disappearance of vertebral lysis and either replacement with bridging or fusion of the involved vertebrae.¹⁶ In adult dogs, radiographic signs can lag behind clinical signs by 3–9 wk, and due to the variability in radiographic changes seen in successfully treated cases of discospondylitis, treatment efficacy is often based on clinical response rather than on radiographic findings.⁷ At this time, there are no established standard criteria for duration of antibiotic therapy in canine discospondylitis.

Linezolid belongs to the oxazolidinone class of antibiotics, which are bacteriostatic through inhibition of protein synthesis via binding to the 50S subunit.¹⁷ Linezolid has adequate penetration into lung and other soft issues at concentrations above the MICs for most susceptible pathogens, including MRSA.¹⁸ Based on the demonstrated efficacy of linezolid for MRSA spondylodiscitis in humans and previously reported pharmacokinetic data for this drug in dogs, the study authors found linezolid to be an attractive antibiotic choice for the patient described herein, given that the MRSP isolated was reported to be resistant to all other lower category antibiotics with anticipated efficacy.¹⁹ It is important to note that the MIC breakpoints for many antibiotics in veterinary use are actually adapted from human studies rather than based on species-specific pharmacokinetic PK data. Therefore, it is possible that some of the antibiotics reported as resistant based on human-adapted MIC standards might actually have been effective in this patient if significant PK differences exist for each antibiotic between dogs and humans. However, in the absence of canine PK data for most antibiotics tested and given the patient’s bacteremia, the authors could not risk the choice of an ineffective antibiotic. Of all the other antibiotics for which this patient’s isolates were tested against, only daptomycin and linezolid were suggestive of being efficacious based on human MIC breakpoints that included Staphylococcus pseudintermedius. Linezolid was chosen because previously reported PK data after oral administration were available to assist in dose calculation and the patient’s blood and urine isolates were found to be susceptible with a low MIC based on data extrapolated from human performance standards.

PK and pharmacodynamic data should be considered when choosing and dosing any antibiotic, particularly nontraditional antibiotic choices. Indices such as area under the concentration-time curve (AUC) over 24 hr at steady state divided by the MIC (AUC/MIC), and the cumulative percentage of a 24 hr period that the drug concentration exceeds the MIC under steady-state PK conditions (percent time of dosing interval > MIC) are useful in predicting the efficacy of an antibiotic. For linezolid, clinical cure and bacterial eradication in human patients were associated with a linezolid AUC/MIC of > 100 and percent time of dosing interval > MIC of > 85.²⁰ In the absence of canine data, those criteria were used to support the dosage prescribed to this dog.

Linezolid treatment was successful in resolving the bacteremia, bacteriuria, and clinical signs of discospondylitis in this dog. Anemia, thrombocytopenia, and lactic acidosis, all of which are reported side effects of linezolid in humans, were observed transiently in the first several weeks of linezolid therapy; however, all of those abnormalities resolved without any change in linezolid dose. The use in dogs of extralabel antibiotics that are typically reserved for resistant human bacterial infections (i.e., “tertiary use” antimicrobials) is controversial. Some physicians, public health specialists, and even veterinarians might advocate for euthanasia of a veterinary patient with a resistant infection rather than administer one of those antimicrobials. However, the American College of Veterinary Internal Medicine consensus statement on antibiotic use advises that “veterinarians are ethically obligated to use antimicrobial drugs, when indicated, to aid in the promotion of the health and well-being of animals.”²¹ Efforts should be made to choose “primary use” antibiotics when possible. Those are older antibiotics with a narrower spectrum of activity. “Secondary use” antibiotics are typically newer and are used for more serious infections in people. Those drugs often have a broader spectrum of activity, and their use should be based on culture and susceptibility results. Linezolid would qualify as a “tertiary use” antibiotic (i.e., used for very resistant human infections). The same consensus statement concludes “tertiary use” drugs “should not be employed in patients that are likely to recover without treatment, in patients that are as likely to be helped through treatment with lower category drugs, or in patients that are unlikely to survive regardless of the therapeutic regimen.”²¹ Because the dog described herein showed progressive clinical signs for 3 mo despite multiple antibiotic regimens, had a documented resistant MRSP bacteremia, and had a disease with a good prognosis for recovery with appropriate treatment, the study authors elected to use a “tertiary use” antibiotic, but only after careful consideration.^15,22 Despite the apparent efficacy of linezolid in successfully treating the dog’s MRSP discospondylitis, the study authors do not advocate the empirical use of the drug under any circumstances. Ideally, the decision to use linezolid or other “tertiary use” antibiotics should be guided by hospital-wide, evidence-based infectious disease control programs that are developed for each institution.

Conclusion

This report documents the successful use of linezolid for clinical resolution of MRSP bacteremia, bacteriuria, and discospondylitis in a dog. To the authors’ knowledge, this is the first report of the use of linezolid to treat naturally occurring disease in a dog. The linezolid dose of 20 mg/kg PO q 12 hr was effective to treat the MRSP isolate with a linezolid MIC of < 1 μg/mL. For Staphylococcus pseudintermedius infections that are resistant to all “primary use” and “secondary use” antibiotics, linezolid should be considered as a treatment option when its use if justified, as outlined above in the American College of Veterinary Internal Medicine consensus statement.

This case underscores the importance of performing a thorough evaluation for infectious etiologies in febrile immunosuppressed patients, as well as considering the risks that may accompany empirical use of antibiotics. It is possible that the patient’s broad antibiotic resistance would not have been selected for if the patient had not received six different antibiotics without appropriate diagnostics. Tests such as a complete blood count, biochemical analysis, urinalysis, blood and urine cultures, as well as imaging (thoracic radiographs, spinal radiographs, abdominal ultrasound, echocardiography, MRI, etc.) should be considered part of the diagnostic regimen for investigation of the cause of fever. Antibiotic administration should be based on culture susceptibility results, where use of “secondary use” and “tertiary use” antibiotics should only be considered if no other effective options are available.