Splendore-Hoeppli phenomenon in a cat with osteomyelitis caused by Streptococcus species

Abstract

A 9-month-old male neutered mixed-breed cat had a history of chronic lameness of the right hind limb, which was non-responsive to antibiotic treatment. Hematologic analysis revealed marked neutrophilia and mild monocytosis. Radiography revealed extensive loss of cortical bone, and replacement with irregular and disorganized bone. There was loss of the normal cortico-medullary distinction, and the medullary cavity had an irregular radiodensity suggestive of osteomyelitis. Surgical curettage and antibiotics did not improve the clinical condition, and amputation was performed. Grossly, the skin over the right tibia was ulcerated with a viscous and granular exudate. At histopathology, there was marked diffuse pyogranulomatous dermatitis, myositis, periostitis and osteomyelitis associated with Splendore-Hoeppli phenomenon. In addition, there was marrow osteoproliferation and multifocal cortical loss, reabsorption, fibroplasia and endosteal bone formation. Gram staining revealed myriad slightly elongated Gram-positive bacteria, arranged in pairs or single chains, confirmed by polymerase chain reaction as Streptococcus species.

Case Report

A 9-month-old male neutered mixed-breed cat had a history of lameness of the right hind limb. The owner reported that the lameness had started 5 months previously, several days after the cat had been castrated. However, signs of local infection related to the surgery were not found by the veterinarian. Radiographs (mediolateral and craniocaudal) of the right hind limb were taken, which showed periosteal osteolysis and proliferation of the proximal metaphysis of the tibia. Prednisone and enrofloxacin were prescribed, but improvement was not observed. After 50 days, two draining fistulae were observed on the medial tibia, and topical rifamycin was prescribed. Fifteen days later, more draining fistulae were detected, and treatment with prednisone and enrofloxacin was reinitiated. Owing to the poor response to therapy and progressive weight loss, the cat was referred to the veterinary hospital at the Universidade Federal de Minas Gerais.

On clinical examination, the cat had a temperature of 39^°C, had good body condition and the right hind limb was diffusely thickened. Further radiographs were taken of the right hind limb, and blood was collected for hematological and biochemical analysis. Neutrophilia and mild monocytosis were present (Table 1). The cat was anaesthetized routinely, and surgical biopsy from skin was performed for histopathologic analysis. The radiographs showed diffuse and marked changes in the tibial metaphysis and diaphysis, characterized by extensive loss of cortical bone and replacement with irregular and disorganized bone. There was loss of the normal cortico-medullary distinction, and the medullary cavity had an irregular radiodensity suggestive of osteoproliferation. In addition, marked periosteal reaction and thickening of the soft tissues lateral to the tibia were observed (Figure 1a). Histopathology revealed marked multifocal-to-coalescing pyogranulomatous and necrotic dermatitis associated with numerous granules. Given the presence of dermatitis and myositis, together with the radiographic changes, osteomyelitis was suspected and different therapy was initiated. The cat was treated with the antimicrobials cefazolin, sodium cefovecin and doxycycline to provide broad-spectrum cover against Gram-positive and Gram-negative bacteria. Meloxicam and tramadol were prescribed for their analgesic and anti-inflammatory properties. In addition, the cat was put on fluid therapy because it was inappetent. Four days later, the hematological analysis was repeated (Table 1) and a sample of the draining exudate was collected for microbiological analysis. Pseudomonas aeroginosa was isolated. The multiple fistulae were discharging a hemorrhagic and purulent exudate (Figure 1b), and were curettaged after induction of general anesthesia. The cat responded poorly to treatment, and the results of further blood tests (Table 1) did not show any improvement in the values of the blood parameters analyzed. Owing to the poor prognosis, amputation of the cat’s right hind limb was performed and the amputated limb was submitted for histopathologic analysis. One week after amputation, the clinical condition improved notably and no recurrence of disease was detected.

Table 1.

Hematological and clinical chemistry data of a cat with osteomyelitis

Hematological and clinical chemistry (units)	First blood analysis	Second blood analysis	Four days after surgery	Mean reference intervals
Red blood cells (106 µl)	6.0	4.0	6.0	5–10
Hemoglobin (g/dl)	8.7	4.6	6.7	8–15
PCV (%)	26.5	17	24	24–45
MCV (fl)	44.2	45.9	40	39–55
MCH (pg)	32.8	25	27.9	30–36
MCHC (g/dl)	14.5	11.3	11.1	12.5–17.5
Total leukocytes (103/µl)	30.1	45.1	20.7	5.5–19.5
Immature neutrophils (103/µl)	1.8	2.7	NR	0–0.3
Mature neutrophils (103/µl)	20.7	39.6	18.6	2.5–12.5
Lymphocytes (103/µl)	4.5	3.2	1.5	1.5–7.0
Monocytes (103/µl)	0.9	1.2	0.2	0–0.8
Eosinophils (103/µl)	2.1	0.6	NR	0–1.5
Platelets (103/µl)	0.2	0.15	0.58	0.25–0.8
AST (IU/l)	137.7	42.6	NR	23–43
Alkaline phosphatase (IU/l)	12.4	19.5	NR	25–93
BUN (mg/dl)	32	27	NR	40–60
Creatinine (mg/dl)	0.22	0.5	NR	0.8–1.8

PCV = packed cell volume; MCV = mean cell volume; MCH = mean cell hemoglobin; MCHC = MCH concentration; AST = aspartate aminotransferase; BUN = blood urea nitrogen; NR = not realized

Figure 1.

(a) Mediolateral radiography of the right tibia, cat. The normal corticomedullary definition has been lost and there is diffuse and widespread replacement with trabecular bone of irregular radiodensity. Marked diffuse periosteal reaction is present. (b) The right hind limb of the cat. There are multifocal coalescing fistulae exuding granular and viscous reddish–yellow exudate; (inset) the appearance of the leg after surgical curettage. (c) The right tibia of the cat. Formalin-fixed longitudinal section showing replacement of the medullary cavity with firm, white, woven bone. There is no differentiation between the medullary cavity and cortical bone. (d) Histopathology of the subcutaneous tissues and muscles of right hind limb. Extensive areas composed of colonies of bacteria, characteristic of the Splendore-Hoeppli reaction. Adjacent muscle is necrosed. hematoxylin and eosin (HE), 20×. (e) Histopathology of the right tibia. The compact cortical bone has been extensively infiltrated by inflammation and there is widespread bone reabsorption. Multiple areas of Splendore-Hoeppli reaction can be seen. HE, 10×. (f) Histopathology of the bone marrow of right tibia. Splendore-Hoeppli reaction surrounded by neutrophils and remnants of necrotic trabeculae. HE, 40×. (g) Histopathology of the bone marrow of right tibia. Splendore-Hoeppli reaction surrounded by neutrophils, multinucleated giant cells and foamy macrophages. HE, 40×. (h) Histopathology of the bone marrow of right tibia. Gram-positive bacteria grouped in the central area or associated with the filamentous, radially-arranged material. Many coccoid bacteria grouped in pairs or single chains can be observed. Goodpasture special stain, 60×.

On gross pathological evaluation, the skin over the right tibia was affected by extensive, multifocal to coalescing ulceration. A reddish-yellow viscous exudate was observed in the subcutis and adjacent muscles. The right tibia was enlarged with moderate new bone formation along the metaphysis and diaphysis. A longitudinal and transverse section showed loss of the medullary cavity and replacement with firm, white woven bone interspersed with yellowish -white and granular material. There was no evidence of differentiation between the medullar cavity and cortical bone (Figure 1c). Histologically, the epidermis was extensively ulcerated and replaced with many neutrophils and cellular debris. In the subjacent dermis, hypodermis and muscles, there were irregular aggregates of variable sizes characterized by basophilic centers and eosinophilic filamentous material radiating into the periphery (Splendore-Hoeppli phenomenon) surrounded by numerous neutrophils, macrophages and giant multinucleated cells. Adjacent to the pyogranulomatous reaction there was fibrous connective tissue and neovascularization. In areas where the infection had infiltrated deeply, there was marked necrosis and loss of muscle fibers (Figure 1d). The tibial periosteum was thickened owing to marked fibrous tissue proliferation and inflammation. The compact cortical bone of the diaphysis was multifocally replaced by fibrous proliferation interspersed with inflammatory reaction and trabecular-like bone (Figure 1e). The Harvesian canals were enlarged and there were areas of bone necrosis, characterized by loss of osteoblasts and absence of a mineralized osteoid layer. The bone was also replaced with multiple Splendore-Hoeppli reactions surrounded by neutrophils, foamy macrophages and giant cells (several containing 20–60 nuclei) (Figure 1f, g). Multifocal areas of new networks of endosteal bone were found among trabecular remnants. The popliteal lymph node showed a marked cortical lymphoid hyperplasia. Within the lymphatic sinus there were numerous macrophages and giant cells containing foamy cytoplasm. Selected tissues from tibia, lymph node and adjacent muscles were subjected to Goodpasture, Giemsa, periodic acid–Schiff (PAS), Ziehl Neelsen and Grocott’s methenamine silver (GMS) special stains. Goodpasture stained Gram-positive bacteria within colonies in the bone marrow, periosteum, muscles, fascia, skin and lymph nodes. This special stain showed many coccoid bacteria grouped in the central area or associated with the filamentous, radially-arranged material. In the periphery of reactive zone, cocci were grouped in pairs or single chains (Figure 1h). Formalin-fixed, paraffin-embedded tissue samples, including skin and bone, were selected for DNA extraction using Qiagen DNA Extraction kits according to the manufacturer’s instructions. Polymerase chain reactions (PCR) reactions were made according to previous publications for Enterococcus species, ¹ Staphylococcus aureus, ² and Streptococcus species. ³ PCR was negative for Enterococcus species and S aureus, and positive for Streptococcus species.

Bacterial pseudomycetoma (botryomycosis) is a pyogranulomatous skin disease characterized by an unusual, presumed immunologic, reaction to non-filamentous bacteria.^4,5 Some specific bacteria elicit the Splendore-Hoeppli reaction (composed of antigen–antibody complexes), which are characterized by the presence of radiating, club-shaped eosinophilic material around infectious and non-infectious agents. In addition, there is inflammatory cell debris, glycoproteins, lipid and calcium derived from host leukocytes. However, the exact mechanism of the formation of the Splendore-Hoeppli reaction is still unknown. ⁵ Grossly, purulent material discharged from fistulae frequently contains white-to-yellow sand-like granules. ⁶ The characteristic formation of the peribacterial or perifungal Splendore-Hoeppli reaction probably prevents phagocytosis and intracellular killing of the insulting agent, leading to chronicity of infection. ⁵ The antigen–antibody complex, a morphologically unique reaction, was first described in sporotrichosis by Splendore and in schistosomiasis by Hoeppli. ⁵ The Splendore-Hoeppli reaction can be caused by filamentous bacteria, including Actinomyces species and Nocardia species.^4,7 Also, non-filamentous bacteria, including Pseudomonas, Proteus, Escherichia coli, ⁵ Staphylococcus species and Streptococcus species, ⁴ are described and should be considered in the differential diagnosis. Gram staining for bacteria can be used for differentiating between Gram-positive (Actinomyces, Nocardia, Staphylococcus, Streptococcus) and Gram-negative (Pseudomonas, Proteus and E coli) organisms. ⁵ Furthermore, Nocardia species can be identified by Ziehl Neelsen, ⁷ PAS and GMS stains. ⁶ In cats, infections with Actinomyces viscosus ⁸ and Nocardia species^7,9 have been reported. Osteomyelitis associated with S aureus and β-hemolytic streptococci has been reported in dogs and cats. ¹⁰ However, reports of non-filamentous bacteria causing Splendore-Hoeppli reaction in the skin, muscles and bone in cats were not found.

The bacteria isolated in this case (P aeruginosa) could not be considered when special stain sections were visualized. To assess the primary causes and avoid secondary microorganisms or contaminants (which tend to grow in moist, draining sinus tracts), needle aspiration, instead of swabs from draining sinuses, should be sent for microbiologic examination.

In general, skin and subcutis infections probably develop as a result of wound contamination or trauma, such as bites, lacerations or puncture wounds with foreign bodies. Thus, infections localized in the skin and subcutis may extend to the deeper tissues to involve bone and muscle. ⁶ Bacterial infections of bones usually starts in vascular areas, such as the periosteum (resulting in periostitis) and the medullary cavity (resulting in osteomyelitis). During bacteremia or septicemia, bacteria can become localized in many organs. In bones, there is a strong predilection for sites of active endochondral ossification within the metaphyses and epiphyses of long bones and vertebral bodies. The medullary sinusoidal capillaries are fenestrated, permitting ready escape of bacteria into the bone marrow. ¹¹ Apparently, lack of blood leukocytes and incompetence of the tissue-based phagocytes of the cartilage matrix predispose to the development of hematogenous osteomyelitis in young animals or kittens. ¹² Thus, a combination of physeal, metaphyseal, or epiphyseal injury and concurrent bacteremia may be involved in the pathogenesis of hematogenous osteomyelitis. ¹³ Probably, extension of infection to periosteum and bone marrow from adjacent tissue (skin and muscles) occurred in this cat. Evidence of other foci of infection was not detected. However, according to the owner, the lameness started several days after orchiectomy. In dogs, manifestation of slowly-developing osteomyelitis 7–15 months after orthopedic surgery has been described. ¹⁰ Bacteria can be localized in the bone marrow, and the infection can disseminate to adjacent tissues. Clinical manifestations of osteomyelitis may not develop until several months later, when the bone lesion becomes extensive enough to cause pain or disfigurement of the bone or, perhaps, result in pathological fracture.^10,13

The osteomyelitis treatment should include debridement, sequestrectomy, lavage, open wound drainage and administration of appropriate antimicrobial drugs. Bacteriological investigations and antimicrobial susceptibility testing should be undertaken. ¹⁴ Empiric therapy is used while culture and antimicrobial susceptibility is pending or when cultures fail. ¹¹ In the present case, open wound drainage, lavage and identification of the cause of the infection should have been established earlier, and this might have resulted in a different outcome other than amputation. The clinical and radiological appearance of the lesion was consistent with osteomyelitis at the first examination. But, by the time the cat was referred, it was too late to cure the patient without amputation. Surgical curettage and systemic antibiotics were successful in eliminating the infection in dogs and cats with osteomyelitis. ¹⁰ Nevertheless, some animals require surgery for complete resolution of disease. ¹² In the present case, the chronicity of the condition and delay in instituting curettage likely contributed to the unsuccessful outcome and ineffectiveness of the antibiotic therapy. Also, bacterial isolation from the immune system (Splendore-Hoeppli reaction) and fibrous tissue can cause persistent infection and impaired antibiotic tissue penetration and concentration. ¹¹

Hematologic data were characterized by marked leukocytosis and mild monocytosis, in particular, which were persistent even after antibiotic therapy and surgical curettage. A decrease in the leukocyte count was only seen after amputation. The changes seen on hematologic analysis were useful in following the progression of the disease in the case reported herein.

Conclusions

The recognition of inflammatory conditions associated with the Splendore-Hoeppli phenomenon and the subsequent differentiation from other causative agents (bacteria or fungi) are important for determining therapeutic measures. ¹⁰ Also, aseptic collection of exudate using needle aspiration for obtaining a definitive culture and associated antibiogram is important for success in the treatment. Presumptive clinical and radiologic diagnosis of osteomyelitis should drive the early treatment, which might result in resolution of infection.