Development of post-vaccination sarcomas, including fibrosarcomas, is well documented in many species, particularly cats,1–9 in which the estimated incidence is approximately 1 in 10,000 cats vaccinated. Vaccination-site fibrosarcomas have also been reported in dogs,10 ferrets11 and a lion.12 Sarcomas arising at vaccination sites have a fibroblastic, myofibroblastic, myoblastic, osteoblastic, chondroblastic or histiocytic origin.4,7,13,14 To the authors' knowledge, there are no reports of a post-vaccination sarcoma developing in a horse.
We describe a fibrosarcoma in a 12-year-old Quarterhorse × Arabian gelding as a sequela to equine influenza vaccination. Shortly after the second vaccination, swelling at the site was noticed by the owner and it continued to increase in size over the following 6 months. Biopsy of the mass indicated a fibrosarcoma had developed at the vaccination site. It was approximately 20 cm in diameter and elevated well above the level of the skin. There was no clinical evidence of metastases to the lungs or local lymph nodes. Surgical resection of the mass was performed and the wound healed by first and second intention. Histopathological examination and immunohistochemical staining confirmed a myofibroblastic fibrosarcoma with multifocal osseous metaplasia. To the authors' knowledge, this is the first equine case of a vaccine-associated fibrosarcoma.
A 12-year-old Quarterhorse × Arab gelding was referred to the Redlands Veterinary Clinic with a large swelling on the left side of the neck (Figure 1). The horse had received two vaccinations for equine influenza (Proteq Flu; Merial), 4 weeks apart, 6 months prior to presentation. The owner reported a swelling at the vaccination site approximately 2 weeks after the second injection. At that time there was no associated discomfort, heat or discharge and the horse's temperature was within normal limits. The swelling progressively increased over the following 4 months and the horse resented palpation of the swelling and movement of the neck. A biopsy of the mass was performed by the referring veterinarian and the histopathological diagnosis was a high-grade fibrosarcoma.
Biopsy samples were fixed in 10% formalin, processed and stained with haematoxylin–eosin for routine light microscopic examination. The biopsied tissue was almost exclusively composed of large, plump ovoid to spindle-shaped neoplastic cells in densely cellular, broad interlacing bundles (Figure 2). Interwoven fascicles and occasional perivascular whorls were also present. The cells had very large, plump ovoid nuclei with a fine to vesicular chromatin pattern and indistinct cell borders within a minimal collagenous stroma. There were prominent nucleoli, marked anisokaryosis and a high mitotic rate (21 per 10 400 x magnification fields) and occasional bi- and multinucleated neoplastic cells. Areas of multifocal necrosis and moderate neutrophilic inflammation were scattered throughout the neoplasm.
There were also multiple, often large, foci of woven bone in interconnecting trabeculae throughout the samples. The bone was lined by a single prominent layer of plump osteoblasts (Figure 3). Individual osteocytes were present within each lacuna and the trabeculae were separated by loose, non-neoplastic collagenous stroma. Where peripheral/capsular regions of the neoplasm were present for examination, the neoplastic infiltrate had a slightly multinodular appearance and was well demarcated from adjacent normal soft tissue. A diagnosis of fibrosarcoma with multifocal osseous metaplasia was made, with the mitotic rate being consistent with a rapidly growing, higher grade neoplasm.
Immunohistochemistry using immunoperoxidase staining (DakoCytomation, Campbellfield, Victoria, Australia) was performed to further categorise the fibrosarcoma. There was strong positive staining of neoplastic cells for vimentin, with the vast majority of cells (>85%) also staining positive for α-smooth-muscle actin (Figure 4), including the multinucleated cells. Neoplastic cells were negative for cytokeratin (wide-spectrum screening), desmin, von Willebrand factor and S-100 protein. Species-specific control tissues stained appropriately in all instances. These findings confirmed the neoplasm was a myofibroblastic fibrosarcoma.
On presentation, the swelling on the left side of the neck was approximately 25 cm in diameter and raised approximately 10 cm above the level of the surrounding skin (Figure 1). Palpation was strongly resented. There was a purulent discharge from the centre of the lesion where the biopsy had been performed. Ultrasonographic examination of the mass revealed a markedly hyperechoic periphery and the discharging sinus communicated directly with the centre of a 15-cm diameter soft tissue mass. Radiographic examination of the cervical vertebrae and soft tissues revealed a radio-opaque mass approximately 15 cm in diameter. There were spicules of increased radio-opacity throughout the mass (Figure 5). There was no clinical evidence of metastases to the lungs or local lymph nodes.
The horse was premedicated with acepromazine (0.01 mg/kg IV), xylazine (1.1 mg/kg IV), procaine penicillin (20 mg/kg IM), gentamicin (6.6 mg/kg IV), phenylbutazone (4.4 mg/kg IV) and tetanus prophylaxis. General anaesthesia was induced with ketamine (2.5 mg/kg IV) and diazepam (0.1 mg/kg IV), and maintained with halothane and oxygen. Dobutamine hydrochloride (2 µg/kg/min given to effect) was used as required to ensure that the intra-operative mean arterial blood pressure remained above 70 mmHg.
A 40-cm elliptical incision was made over the mass, which was adhered to all surrounding tissue, making removal extremely difficult. Sharp dissection using both a scalpel and Metzenbaum scissors was required to separate the mass from the surrounding tissue, including the nuchal ligament and cervical musculature, which resulted in severe haemorrhage during the procedure. However, the blood pressure remained above 70 mmHg for the duration of surgery and blood transfusion was not required. The volume depletion was corrected by administration of 5 L colloids, 2 L hypertonic saline and 20 L Hartmann's solution over a 40-min period. Ventrally, the mass was closely adhered to the third and fourth cervical vertebrae and only minimal margins were taken. The resected tumour was elliptical, approximately 20 cm long and 15 cm wide, with very well defined margins (Figure 6). After removal of the mass, the necrotic skin over the site was resected. In order to control haemorrhage, the dead space was tightly packed with 15 rolls of sterile gauze bandage (conforming gauze bandage; Zebravet). The wound was closed primarily using vertical mattress sutures of USP 1 monofilament nylon (Ethilon), with some areas left unsutured to allow for drainage and removal of the packing. As there was considerable pressure on the suture line in the dorsal and middle sections of the incision, a stent bandage was applied over the wound. The surgical area was then compressed with a sterile combined dressing and an adherent elastic bandage to limit further haemorrhage during the anaesthetic recovery period.
Postoperatively, the horse was treated with procaine penicillin (20 mg/kg IM twice daily), gentamicin (6.6 mg/kg IV once daily), and phenylbutazone (2.2 mg/kg PO twice daily) for 7 days and it made an excellent recovery. The sterile packing was removed at 72 h and the wound was lavaged with dilute chlorhexidine. The packed cell volume of the horse was 0.14 L/L (normal range 0.35–0.50 L/L) 24 h postoperatively and gradually returned to normal. The ventral and caudal aspect of the wound healed by first intention and the cranial aspect by second intention (Figure 7a,b). Four months postoperatively the wound had healed well and there was no evidence of tumour recurrence.
The myoblastic fibrosarcoma with osseus metaplasia reported in this horse demonstrated behavioural, histological and immunohistochemical similarities to both canine and feline vaccination-site fibrosarcomas. The neoplasm in this case was immunoreactively positive to vimentin,3,9,10,13–15 and α-smooth-muscle actin9,10,13-15 and negative for desmin, cytokeratin, S-100 protein and factor VIII,3,9,10,14 which suggests that equine post-vaccination fibrosarcomas can occur and to the authors' knowledge, this is the first report of a myofibroblastic fibrosarcoma arising after intramuscular vaccination in a horse.
Myofibroblastic fibrosarcomas appears to be the most common subtype of vaccine-associated sarcoma, as neoplastic cells with myofibroblastic features were identified in 16 of 20,9 28 of 44,13 and 20 of 2010 vaccination-site fibrosarcomas in cats and in 10 of 15 cases in dogs.10
Myofibroblasts are fibroblasts that proliferate during granulation tissue formation.16 The role of myofibroblasts in vaccination-associated sarcomas has not been fully elucidated, but they may reflect a continuum of the initial inflammatory response,9 and differentiation into the myofibroblastic phenotype is regulated by cytokines.13 Osseous metaplasia of multipotential mesenchymal cells of non-osseous origin can occur, probably related to production of osteoinductive cytokines by the neoplastic cells.17
Vaccine-associated sarcomas appear to result from an inflammatory response, which causes myofibroblast proliferation and the uncontrolled growth of these transformed mesenchymal cells results in a neoplasm1,4,6,7,18 Feline vaccine-associated sarcomas often, but not always, contain evidence of concurrent inflammation.4,9,10 Furthermore, histological lesions that appear in transition from chronic granulomatous inflammation to neoplasia have been reported in cats.6 Inflammation was evident within the neoplasm described here and may have been induced by necrosis as the proliferating neoplastic cells outgrew their blood supply. However, given the intense inflammatory reaction seen in this horse after the second vaccination, it is possible that an inflammatory reaction to the vaccine initiated the neoplastic transformation.
Vaccine-associated sarcomas in small animals are mostly associated with rabies virus and feline leukaemia virus vaccines,8,18 although other vaccines, including parvo-, herpes- and calicivirus, have been reported.2 The interval between vaccination and tumour appearance ranges from 3 months to 3 years.18 Aluminium-based vaccine adjuvants may cause a persistent inflammatory reaction and predispose to derangements in the fibrous tissue repair response, thus playing a role in the aetiology of vaccine-associated sarcomas.3,18 Aluminium has been demonstrated by electron probe microanalysis in the area of central necrosis, in macrophages and giant cells of feline3,9,18 and canine10 post-vaccination fibrosarcomas. However, Kass et al. found no significant difference in the incidence of sarcomas between aluminium-adjuvant and non-adjuvant products.8
There has been a report of inflammatory reaction in horses following vaccination with a strangles vaccine containing an aluminium adjuvant19 and Zimmel et al. also reported an anaphylactic reaction to a multidose adjuvant-containing vaccine in an individual horse.20 Aluminium was not responsible for the neoplasm in the present horse because there is no aluminium contained in the Proteq equine influenza vaccine (Merial, personal communication). The adjuvant used in the Proteq vaccine is carbopol and the vaccine uses a canarypox vector and there have been no reports of adverse reactions to either the adjuvant or vector in any species to the authors' knowledge.
A possible explanation for the lack of reports of vaccination-site fibrosarcomas in horses and other large animals may be intramuscular administration of vaccines as opposed to subcutaneous administration in small animals. The irritant material is deposited into muscle rather than subcutaneous connective tissue where mesenchymal cells can be stimulated to proliferate and potentially transform into neoplastic cells.
Sarcomas are also reported to develop at the site of microchip implantation in many species, including cats,21 dogs,15 rodents22 and mahogany gliders (Petaurus gracilis) (IPC Wilkie, Veterinary Pathology, School of Veterinary Science, The University of Queensland, St Lucia, Australia, 2008, personal communication). Microchips are sealed in a bioglass capsule composed of silicone15, 21 and the horse in question had a microchip implanted at the time of the first vaccination, but it was not associated with the tumour and remained intact on the left side of the neck after the tumour was resected.
Fibrosarcomas are not common tumours in horses,23 comprising 2% of cutaneous and musculocutaneous neoplasms.24 They have been described in the premaxilla,25 rostral mandible,26 paranasal sinuses,27,28 tarsal groove,23 humerus,29 epaxial musculature,30 omentum,31 mesentery,32 abdominal cavity and wall,33 kidney,34 thoracic vertebrae and wall,35 mammary gland36 and prepuce.37 A fibrosarcoma has also been reported following a burn injury in a horse.38 However, there is some confusion in the older literature concerning the classification of equine fibroblastic sarcoids and fibrosarcomas;23 equine fibrosarcomas are usually locally invasive, with only rare reports of widespread metastases.29
Using grading systems adapted for feline13 and canine39 fibrosarcomas, this particular neoplasm would be classified as a high-grade (grade III/III) malignancy, but use of such grading systems for prognostication in equine fibrosarcoma is untested. High-grade feline and canine vaccination-associated sarcomas commonly are locally invasive and aggressive and have a high incidence of local recurrence, especially if surgical excision is incomplete.4,13,14,10 Metastasis to distant sites is also reported1,5,14 and a high histological grade increases the chance of distant metastasis.40 Given the histological similarities and rapid growth, similar behaviour to that seen in canine and feline high-grade vaccination-associated sarcomas might be expected in horses.
A positive correlation has also been reported between the presence of neoplastic multinucleated cells, as seen within the neoplasm described here, and tumour grade in feline vaccination-associated fibrosarcomas.13 However, although the mass was rapidly growing it was well encapsulated within a bony shell and not locally invasive, and there has been no clinical evidence of metastases to date.
It must be emphasised again that the previously reported histological grading systems13 may have limited prognostic value in feline vaccine/injection-site sarcomas, because clinical progression is highly variable and a low histological grade does not exclude the development of either local recurrence or distant metastasis.40 The prognostic merit of histological grading of feline vaccine-associated sarcomas is therefore still unclear and it is obviously impossible to offer any comment regarding the value of grading such tumours in horses based on this single report.
Approximately 140,000 horses received 300,000 doses of the Proteq vaccine during the 2007 equine influenza outbreak in Australia. The only other serious adverse effect documented was a case of possible polyarthritis reaction. Other cases of adverse reaction were consistent with injection and vaccination technique rather than being product related (Merial, personal communication). Therefore, despite this documentation of an adverse reaction, the Proteq flu vaccine should still be considered very safe and effective (NJ Kannegieter, unpublished data) and can be used with confidence in horses for controlling and eradicating particular strains of equine influenza. In rare cases where post-injection swelling persists and can not be explained by local infection, a diagnosis of fibrosarcoma should be considered and early surgical intervention will have a successful outcome.
The authors thank Dr Peter Agnew for referring the case and the staff at Redlands Veterinary Clinic who assisted in the care of the horse.