Peritoneal EMH in a Dog with Immune-Mediated Hemolytic Anemia

Introduction

Extramedullary hematopoiesis (EMH) is the process by which normal blood cells are produced outside the bone marrow. In humans, EMH effusions are rare and are characterized by the presence of megakaryocytes, immature erythrocytes, immature leukocytes, or combinations of those cells. To the authors’ knowledge, this is the first report to describe a case of peritoneal EMH effusion in a dog.

Case Report

A 5 yr old castrated male shorthaired dachshund presented with a 2 day history of pigmenturia and inappetence. Abnormalities detected on physical examination included moderate periodontal disease, icterus, and red-brown urine. The patient’s mentation was depressed; however, vital parameters were within normal limits. The complete blood count demonstrated marked agglutination that did not disperse with a 1:9 saline dilution and a regenerative anemia (hematocrit, 13%; reference range, 37–55%) that was characterized by the presence of scattered polychromatophilic erythrocytes and frequent rubricytes extending back to the basophilic rubricyte. Marked spherocytosis confirmed the presence of immune-mediate hemolytic anemia. Additionally, an acute inflammatory leukogram was characterized by a neutrophilia (20.3 × K/µL; reference range, 3.0–11.5 × K/µL), a regenerative left shift (band neutrophil concentration, 2.9 × K/µL; reference range, 0.0–0.3 × K/µL), and monocytosis (3.5 × K/µL; reference range, 0.1–0.8 × K/µL). Moderate thrombocytopenia (based on subjective assessment of the blood smear) was also noted. Serum biochemical abnormalities included hypoglycemia (2.27 mmol/L; reference range, 4.05–6.27 mmol/L), azotemia (blood urea nitrogen, 29.27 mmol/L; reference range, 3.21–11.78 mmol/L), hypoalbuminemia (28 g/L; reference range, 34–42 g/L), total hypocalcemia (2.35 mmol/L; reference range, 2.43–3.03 mmol/L), hypokalemia (3 mmol/L; reference range, 3.6–5.3 mmol/L), decreased bicarbonate (14 mmol/L; reference range, 18–29 mmol/L), increased alanine transaminase (3,168 U/L; reference range, 28–171 U/L), hyperphosphatasemia (817 U/L; reference range, 1–142 U/L), increased creatine kinase (504 U/L; reference range, 128–328 U/L), and hyperbilirubinemia (372.78 µmol/L; reference range, 1.71–5.13 µmol/L). Urinalysis, collected via catheterization, revealed brown, cloudy urine, a specific gravity of 1.029 (measured via refractometry), bilirubinuria (3+ dipstick, confirmed via ictotest) and proteinuria (3+ dipstick, confirmed by sulfasalicylic acid test). Sediment analysis of urine showed moderate bilirubin crystalluria, scant struvite crystalluria, and erythrocytes (10–50/high-power field). Aerobically cultured urine was negative for growth of microorganisms.

A Snap 4DX test^a on peripheral blood was negative for the Dirofilaria immitus antigen and Borrelia burgdoferi, Anaplasma phagocytophilum, and Ehrlichia canis antibodies. Babesia spp. polymerase chain reaction testing^b performed on serum was negative. Thoracic radiographs were unremarkable, but abdominal ultrasonography revealed the presence of a moderate amount of highly echogenic peritoneal fluid. Ultrasound-guided fine-needle aspiration yielded approximately 5 mL of serosanguineous fluid.

Peritoneal fluid analysis demonstrated that the total nucleated cell count was 9.4 × K /µL, the protein concentration by refractometry was 31 g/L and the hematocrit of the fluid was < 3%. Direct, line, and cytocentrifuged preparations of the peritoneal fluid were examined (Figure 1). The cytocentrifuged preparation had high nucleated cellularity and many mature and polychromatophilic erythrocytes. A count of 500 of the nucleated cells revealed a granulocytic/erythroid ratio of 1.4:1. The cells in the granulocytic series were almost exclusively of the neutrophil cell line and included ∼ 60% nondegenerate segmented neutrophils, ∼ 30% band forms, ∼ 8% metamyelocytes, and ∼ 2% myelocytes. The erythroid series exhibited complete maturation with ∼ 15% of the erythroid precursor cells in the early (nonhemoglobinized) pool (including ∼ 2% rubriblasts, ∼ 5% prorubricytes, and ∼ 8% basophilic rubricytes) and ∼ 85% of the cells in the late (hemoglobinized) pool, with approximately equal numbers of polychromatophilic rubricytes and metarubricytes. Small mature lymphocytes and macrophages comprised ∼ 3% and 1% of the population, respectively. Neither organisms nor cells with features of malignancy were seen.

View Figure

• Download as Powerpoint
• Download Figure

A diagnosis of primary immune-mediated hemolytic anemia with peritoneal EMH effusion was presumptively established, and the patient was treated accordingly. A whole blood transfusion (17 mL/kg) was administered, as well as the following medications: prednisone^c (1.4 mg/kg per os [PO] q 12 hr), doxycycline^d (7.2 mg/kg PO q 12 hr), unfractionated heparin^e (102 IU/kg subcutaneously q 8hr), acetylsalicylic acid^f (1.5 mg/kg PO q 24 hr), biologically modified cyclosporine^g (5 mg/kg PO q 12 hr), and maropitant^h (1 mg/kg subcutaneously q 24 hr). IV fluid composed of 0.9% saline with 5% dextrose was also administered.

The following day, the patient showed evidence of intracranial disease; mental obtundation was observed, as well as systemic hypertension and bilateral pupil miosis with hippus. The client declined further intervention, and the patient was euthanized. Necropsy examination was not permitted.

Discussion

EMH describes the production of normal blood cells outside the bone marrow environment. In humans, EMH effusions are uncommon and are characterized by the presence of megakaryocytes, immature erythrocytes or leukocytes, and (in rare cases) evidence of cells from two or all three lineages.¹ EMH is most often detected in reticuloendothelial organs, such as the liver and spleen; however, myriad other sites of EMH have been reported in humans. Nonhepatosplenic EMH has been documented in the mediastinum, lymph nodes, pancreas, pleura, lungs, gastrointestinal tract, kidney, prostate, endometrium, as well as within the choroid plexus and mammary neoplasms. Interestingly, the latter two have been described in the dog.^1–7

Confirmatory diagnostic tests for EMH effusions include radionuclide bone marrow scanning, contrast-enhanced computed tomography, MRI, and cytogenetic analysis of cells within the fluid. Although the gold standard test for the diagnosis of an EMH effusion is considered to be a biopsy of the EMH lesion, many authors clearly state that when a patient with a predisposing hematologic condition has an effusion with characteristic cytologic features, no further diagnostics are necessary.^4–9 In humans, EMH effusions diagnosed by cytologic examination of the fluid have been reported in the pleural cavity, the peritoneal cavity, and both concurrently.^1,4,5,9–12 The case described herein meets the criteria outlined in the human literature for diagnosis of EMH effusions, and to the authors’ knowledge, this is the first report describing EMH peritoneal effusion in a dog.

The human literature reports that nonhepatosplenic EMH effusions most often occur in association with myelofibrosis and myeloid metaplasia.^4,7,11 EMH effusions have been diagnosed in people with neoplasia, most often myeloproliferative disorders and β-thalassemia, hepatitis, and tuberculosis.^1,4,5,10,11 Interestingly, there is one report of a human patient with malignant EMH ascites and agnogenic myeloid metaplasia (i.e., chronic idiopathic myelofibrosis), as well as a history of Coombs positive hemolytic anemia.¹³

To the authors’ knowledge, only one relevant abstract is available in the veterinary literature. A small number of erythroid and myeloid precursors were observed in the peritoneal fluid of a young, female cat that was diagnosed with myelofibrosis and EMH.¹⁴

The incidence of EMH effusions in humans is low. One retrospective report found 8 EMH effusions from 5 patients out of 20,793 patients with cavitary effusions over a 21 yr period.¹ Nonetheless, diagnosis of EMH effusions is clinically important. For example, normal megakaryocytes within peritoneal fluid have been misinterpreted as atypical cells associated with carcinoma.⁴

At this time, there is no satisfactory explanation of the pathogenesis of nonhepatosplenic EMH. Contemporary hypotheses that may be pertinent to the case described in this report include the following: EMH may be a compensatory response to the replacement of bone marrow elements, or EMH may manifest in the presence of a circulating factor(s) that induces adult stem cell populations to differentiate into cells of the hematopoietic lineages.⁷

Inadvertent splenic aspiration could explain the unique findings in this case, but the authors of this report believe this to be highly unlikely. Abdominocentesis was completed with ultrasound guidance by a board-certified radiologist, and the large volume of fluid removed during aspiration was inconsistent with the small volume samples that are usually obtained from the spleen. The possibility that the patient’s effusion was hemorrhagic is also considered to be unlikely because the hematocrit of the patient’s peripheral blood (13%) was substantially greater than that of the peritoneal fluid sample (< 3%) and there was no evidence of erythrophagocytosis, hemosiderin, or hematoiden in the effusion that would support previous hemorrhage. Finally, the effusion was noted to be highly echogenic prior to sampling, which is consistent with the elevated total nucleated cell count demonstrated by cytologic examination.

Another unusual finding in this case was the presence of hypoglycemia. There are several plausible explanations for this abnormality, including increased utilization by poorly perfused tissues employing anaerobic glycolysis and/or reduced hepatic gluconeogenesis secondary to acidosis-induced decreased delivery of precursors. In addition, it has been reported that cells in septic patients consume more glucose and in the absence of either similar or contradictory cases in the literature, the authors of this report theorize that the cells within the peritoneal EMH effusion may have behaved similarly.¹⁵

Why this patient acutely developed signs of the intracranial disease is a matter of speculation. The primary differential diagnoses considered for this patient’s neurologic deterioration included a thromboembolic effect, an adverse drug reaction, or intracranial EMH. Although choroid plexus EMH has been described in five dogs, the majority of those cases demonstrated seizure activity at the time of initial presentation.³

Some authors have suggested that the diagnosis of an EMH effusion in humans carries a poor prognosis when it is not merely a compensatory response.¹ In such cases, it is believed that EMH effusions herald the presence of an aggressive neoplastic process within the bone marrow.

Conclusion

This case highlights a unique cytologic finding in a dog with immune-mediated hemolytic anemia. Veterinarians should be aware that EMH can occur outside of the liver and spleen under some conditions. The diagnosis of an EMH effusion is made when characteristic cytologic findings are detected in cavitary fluid from a patient with a hematologic condition likely to result in EMH.