Introduction

Immune-mediated neutropenia (IMN) is an uncommon disease in veterinary medicine, and the early stages of the disease can be easily mistaken for other conditions such as sepsis or bone marrow suppressing disorders.^1,2 Obtaining a definitive diagnosis can be challenging and has historically relied on clinical response to immuno-suppressive therapy after excluding all other causes of disease.^1,3 Several attempts have been made at developing anti-neutrophil antibody flow cytometry to match the gold standard of diagnosis in human medicine; however, these have not proven to be reliable diagnostic techniques on the veterinary side.^4,5 Bone marrow evaluation can potentially provide confirmation of a diagnosis of IMN but has traditionally been costly and often impractical in the private practice setting.²

This case report describes the clinical presentation of IMN, the challenges associated with diagnosis, and a cheap and reliable method of obtaining bone marrow aspirates to help provide a definitive diagnosis.

Case Reports

A 10 yr old spayed female toy poodle was referred to the Colorado State University Veterinary Teaching Hospital after a 10 day history of waxing and waning lethargy, vomiting, diarrhea, and anorexia. Current treatments from her referring veterinarian included maropitant, ampicillin/sulbactam, and IV fluids. Previously, she has had similar episodes occurring every 1–2 mo over the last 7 mo prior to presentation that were resolved with fluid therapy and IV antibiotics.

At presentation to the hospital, she was evaluated by the emergency service. On physical examination, she was lethargic and febrile (103.4°F, 39.7°C) with a body weight of 3.78 kg (8.3 lb). No additional pertinent physical exam findings were noted.

Serum biochemical analysis revealed a marked hypoglycemia (42 mg/dL; reference interval 70–115 mg/dL), mildly low creatinine (0.5 mg/dL; reference interval 0.6–1.6 mg/dL) and magnesium (1.5 mg/dL; reference interval 1.8–2.4 mg/dL), and a mildly increased alkaline phosphatase (288 IU/L; reference interval 15–140 IU/L).

A complete blood count (CBC) revealed a mild leukopenia (4.2 × 10³/μL; reference interval 4.5–15.0 × 10³/μL) characterized by a marked neutropenia (1.2 × 10³/μL; reference interval 2.6–11.0 × 10³/μL) with band cells (0.2 × 10³/μL; reference interval 0–0.2 × 10³/μL), a mild monocytosis (1.6 × 10³/μL; reference interval 0.2–1.0 × 10³/μL), and an eosinopenia (0.0 × 10³/μL; reference interval 0.1–1.2 × 10³/μL). Moderate toxic changes to the neutrophils were noted on pathologist readout. A moderate thrombocytopenia (108 × 10³/μL; reference interval 200–500 × 10³/μL) with rare clumping was also noted.

A urinalysis was performed on a voided sample and significant findings included a urine specific gravity of 1.010, a pH of 8, trace proteinuria, and an inactive sediment.

Thoracic radiographs were performed and were unremarkable. Abdominal ultrasound revealed bilateral decreased renal corticomedullary distinction, a hyperechoic nodule in the liver, and an accessory spleen. No other abdominal abnormalities were noted. The renal changes were consistent with chronic kidney disease, whereas differentials for the hepatic changes included nodular hyperplasia, chronic granuloma, hematoma, and vacuolar hepatopathy, with neoplasia considered less likely.

Additional diagnostics performed by the emergency department included an Adrenocorticotrophic hormone stimulation test (pre = 6.64 μg/dL, reference interval 1.0–6.2 ug/dL; post = 21.60 μg/dL, reference interval 10.5–20.0 μg/dL), an insulin concentration (<2.0 μIU/mL; reference interval 5–20 μIU/mL), and bile acids (preprandial = 18 μmol/L, reference interval 0–9 μmol/L; postprandial: 34 μmol/L, reference interval 0–19 μmol/L). These diagnostics excluded hypoadrenocorticism, insulinoma, and liver failure as causes of her clinical signs. Prothrombin time/partial thromboplastin time (within the reference intervals) and a tickborne disease panel (negative) were also performed^a.

After sample collection, a dextrose bolus was given (0.5 mL/kg) IV, which increased her glucose level to 135 mg/dL^b. She presented on a Friday afternoon and was hospitalized in the intensive care unit on IV fluids (Norm-R + 20 mEq KCl/L + 5% dextrose)^c, maropitant (1 mg/kg IV q 24 hr)^d, pantoprazole (1 mg/kg IV q 12 hr)^e, and ampicillin/sulbactam (50 mg/kg IV q 8 hr)^f over the weekend with supportive care and monitoring of her blood glucose. A nasogastric tube was placed to provide enteral nutrition.

The following Monday, she was transferred to the Internal Medicine Service. A CBC and serum chemistry were repeated. Over the weekend, her dextrose supplementation had been weaned from 5 to 2.5%, and she was continuing to maintain her glucose at appropriate levels. She was receiving 75% of her daily caloric needs via the nasogastric tube. Dextrose supplementation was discontinued at 10 a.m., and she continued to maintain an adequate blood glucose throughout the day.

The CBC revealed a persistent mild leukopenia (3.9 × 10³/μL; reference interval 4.5–15.0 × 10³/μL) characterized by a neutropenia (0.9 × 10³/μL; reference interval 2.6–11.0 × 10³/μL) and no band cells at this time as well as a monocytosis (1.5 × 10³/μL; reference interval 0.2–1.0 × 10³/μL). A mild to moderate, normocytic, normochromic, nonregenerative anemia, with a noted absence of a reticulocyte response, was noted (31%; reference interval 40–55%). A mild thrombocytopenia (134 × 10³/μL; reference interval 200–500 × 10³/μL) with moderate clumping was still noted.

A serum biochemical analysis revealed her glucose level to be within the reference interval (94 mg/dL; reference interval 70–115 mg/dL); mildly decreased creatinine (0.4 mg/dL; reference interval 0.6–1.6 mg/dL), calcium (8.8 mg/dL; reference interval 9.0–11.5 mg/dL), and magnesium (1.5 mg/dL; reference interval 1.8–2.4 mg/dL); moderately decreased albumin (2.1 g/dL; reference interval 3.0–4.3 g/dL); and a mildly increased alkaline phosphatase (174 IU/L; reference interval 15–140 IU/L).

Because of the history of chronic gastrointestinal signs, upper gastrointestinal endoscopy was performed. She was anesthetized with methadone (0.3 mg/kg IV)^g and alfaxalone (1.6 mg/kg IV)^h. Upon visualization, her stomach appeared grossly normal. Her duodenal mucosa appeared erythematous and friable with multifocal pale regions. Biopsies were obtained from the stomach and duodenum and helicobacter cultures were collected. She recovered well from anesthesia and was maintained on her current medications overnight.

The following day, preliminary histopathology indicated a mild lymphoplasmacytic gastroenteritis. To further evaluate the pancytopenia noted on CBC, bone marrow aspirates from the 11th and 12th costochondral junction on the right side were obtained. The patient was sedated with butorphanol (0.2 mg/kg IV)ⁱ, the right lateral thorax was clipped and cleaned, and a local lidocaine block was administered^j. The area was then sterilely prepared, and once aseptically gloved, the sampler’s left hand was used to stabilize the 11th rib just above the costochondral junction. A 22-gauge hypodermic needle attached to a 3.0 mL syringe was inserted, beveled up at the costochondral junction of the 11th rib, and advanced dorsally into the medullary cavity. Gentle aspiration was used to attain a hub sample of marrow. Anticoagulant (ethylenediaminetetraacetate) was not used as smears for cytologic evaluation were immediately prepared.

Six slides of moderate to high cellularity were obtained for evaluation. Cytologic evaluation of the bone marrow showed marked myeloid hyperplasia with orderly maturation until the band stage, which is suggestive of immune-mediated destruction of late-stage, mature neutrophils. The erythroid and megakaryocytic cell maturation was orderly and appropriate with adequate numbers of erythroid cells and possibly increased megakaryocyte numbers.

A blood sample was sent to the University of Tennessee Veterinary Diagnostic Lab for neutrophil surface antibody flow cytometry. The sample was negative for IgG, IgM, and C3 binding to neutrophils.

Considering the bone marrow aspirate findings and the most likely diagnosis of IMN, IV dexamethasone (2 mg/kg prednisone equivalent)^k therapy was initiated, and she was maintained on her other treatments overnight. The following morning, her attitude and appetite had improved. A repeat CBC showed a more pronounced left shift with metamyelocytes (0.3 × 10³/μL; reference interval 0–0.001 × 10³/μL), band cells (0.8 × 10³/μL; reference interval 0–0.2 × 10³/μL), a persistent neutropenia (0.8 × 10³/μL; reference interval 2.6–11 × 10³/μL), lymphopenia (0.6 × 10³/μL; reference interval 1–4.8 × 10³/μL), monocytosis (2.2 × 10³/μL; reference interval 0.2–1 × 10³/μL), and an eosinopenia (0.0 × 10³/μL; reference interval 1–1.2 × 10³/μL). She still exhibited a mild anemia (31%; reference interval 40–55%) and moderate thrombocytopenia (92 × 10³/μL; reference interval 200–500 × 10³/μL) with clumping and giant platelets noted on blood smear evaluation. At this time, mycophenolate therapy was initiated (10 mg/kg IV q 12 hr)^l as a second-line immunosuppressive to allow for more rapid clinical response and to allow for tapering of the prednisone sooner.

The following day, she exhibited marked improvement in both attitude and appetite, and that afternoon, she was discharged to her owners for continued care at home. She was discharged on the following medications: mycophenolate capsules (8 mg/kg per os [PO] q 12 hr), prednisolone (0.7 mg/kg PO q 12 hr)^m, amoxicillin/clavulanic acid oral suspension (13.75 mg/kg PO q 12 hr)ⁿ, mirtazapine (1 mg/kg PO q 24 hr pro re nata)^o, and ondansetron (1 mg/kg PO q 12 hr pro re nata)^m,n,o,p.

A recheck exam and CBC were performed 9 days later. At presentation, she was bright, alert, and interactive. Her physical exam was unchanged when compared with her exam the day of discharge, save for the changes in attitude and activity. The CBC revealed a leukocytosis (26.9 × 10³/μL; reference interval 4.5–15 × 10³/μL) characterized by a neutrophilia (23.7 × 10³/μL; reference interval 2.6–11 × 10³/μL), monocytosis (1.9 × 10³/μL; reference interval 0.2–1 × 10³/μL), and eosinopenia (0.0 × 10³/μL; reference interval 1–1.2 × 10³/μL). Her hematocrit had improved (36%; reference interval 40–55%) with an increase in reticulocytes (103.4 × 10³/μL; reference interval 0–100.0 × 10³/μL). Her platelet count had normalized (231 × 10³/μL; reference interval 200–500 × 10³/μL) as well. She continued on her current therapy, and additional follow-ups were scheduled to discuss tapering in the future.

Discussion

IMN is a rare autoimmune disease of dogs that has primary and secondary origins, similar to, and often in tandem with, other immune-mediated disease.^1,2 As with other immune-mediated disease, all secondary causes such as neoplasia, drug sensitivity, and infectious causes must be excluded before a diagnosis of primary IMN can be attained. In the case of this patient, a thorough history review revealed no medications known to cause bone marrow suppression. In addition, screening for neoplasia and tick-borne disease was performed and was negative. Her most recent vaccination history was greater than 6 mo prior to presentation. No underlying secondary cause could be determined for this patient.

The gold standard of diagnosis in humans is anti-neutrophil flow cytometry; however, even this test has a relatively low sensitivity in human medicine, thought to be because of neutrophil fragility, low numbers of anti-neutrophil antibodies, and poor binding affinity of the antibodies to the neutrophils.⁶ In veterinary medicine, one study reported promising results in developing a canine-specific anti-neutrophil antibody test; however, a low sample population, problems with cell fragility, and leukocyte aggregation in vitro call in to question the diagnostic utility of this test in the clinical setting.⁴ In addition, a recent retrospective study evaluating IMN found that, of nine dogs who had anti-neutrophil antibody testing performed, four (44%) had a negative test result, despite a presumptive diagnosis of IMN based on response to treatment.⁵ Based on these factors, it is generally accepted that, after exclusion of all secondary causes of neutropenia, a primary immune-mediated cause can be determined based on rapid response to immunosuppressive therapy in veterinary medicine.^1,3 Supportive diagnostic testing such as flow cytometry, bone marrow cytology, and/or histopathology can help the clinician feel more confident in a diagnosis of IMN, but it should not be relied upon as the sole diagnostic technique.

Bone marrow evaluation can be useful diagnostic tool for the clinician when suspicious of marrow-directed disease. Bone marrow aspirates of IMN patients, as was the case with this patient, will typically have a left-shifted myeloid series with an orderly erythroid and megakaryocytic line and an abnormal myeloid:erythroid ratio.² Traditional sample sites include the humerus, iliac crest, femur, and sternum but require general anesthesia and specialized equipment in the case of core biopsies.⁷ The extra cost and associated risk of anesthesia, particularly in a compromised patient, may make obtaining these samples a challenge. As an alternative, performing chostochondral rib aspirates can be done with significant cost and time savings and requires only light sedation and local anesthesia and no specialized equipment. In addition, many more sites are readily available to sample. Trade-offs include limited sample size per site, risks of puncturing underlying structures such as liver or lung, and inability to evaluate marrow architecture. Evaluation of the diagnostic quality of these samples has shown similar results to traditional sites.^8,9

It should be mentioned that patients with compromised neutrophil function are at a higher risk for secondary bacterial infections and sepsis. In the case of this patient, it was suspected that her hypoglycemia and febrile state at presentation were likely the result of a systemic bacterial infection. Often, these cases will need antibiotic therapy and will show improvement in clinical signs with antibiotics alone prior to making a diagnosis of IMN.¹⁰

One final point to consider is that these patients can present with thrombocytopenia and/or anemia.^1,11 Antibodies directed against cell surface proteins may not be specific to neutrophils and can result in destruction of platelets or red blood cells.¹² Although most reported cases show only mild to moderate decreases in platelet and red blood cell numbers, some may be severe enough to require transfusions.^{2,3,5,6,10,11}

Conclusion

This case illustrates not only the challenges associated with diagnosing IMN, but it also serves to illustrate a useful diagnostic technique for obtaining bone marrow samples that can be performed in the private practice setting. In addition, the clinician should be aware of concurrent disease processes that can make diagnosis and treatment more challenging in the initial phase. In the long run, although frustrating to diagnose, once appropriately treated, these cases can be very rewarding.