Introduction

Cutaneous adverse drug reactions (ADRs) in domestic animals are considered uncommon and account for 1–2% of cases presented to specialty dermatology clinics.^12–3 They are difficult to diagnose because the clinical signs can mimic many skin diseases and a true cause and effect relationship is often challenging to prove.⁴ Idiosyncratic ADRs in cats have been reportedly caused by many systemic and topical drugs, but there have been no previous reports of cutaneous ADRs to non-steroidal anti-inflammatory drugs (NSAIDs).^1,3,4 The aim of this case report is to describe a probable cutaneous ADR to piroxicam in a cat.

Case description

A 9-year-old male neutered Devon Rex cat presented to the University of Minnesota, College of Veterinary Medicine (UMNCVM) dermatology service, with symmetrical, ulcerative lesions localized to the axillae, inguinal areas and ventral abdomen. The lesions developed 10 days previously and had since progressed rapidly. The cat had a history of seasonal pruritus and miliary dermatitis that occurred from spring through autumn, which was responsive to oral prednisolone and injections of cefovecin.

Ten months prior to presentation, an oral mass developed on the right side of the buccal surface of the lower lip. Six weeks later the mass was surgically excised by the referring veterinarian (rDVM) with narrow margins. The histopathologic diagnosis was salivary gland adenocarcinoma. A second surgery was performed by the rDVM 11 weeks later to remove a 5 mm new mass that appeared at the surgical site.

Three weeks after the second surgery, the cat was seen by the oncology service at the UMNCVM for further work-up. Piroxicam (compounded 10 mg/ml, University of Minnesota) 0.33mg/kg PO q48h was prescribed (day 0) for its anti-tumor activity. A CT scan was performed 2 weeks after starting piroxicam (day 14) and revealed mildly enlarged right retropharyngeal and right deep cervical lymph nodes.

Two weeks later (day 28) the cat was evaluated by the surgery department at UMNCVM for removal of the enlarged lymph nodes and crusting was noted on the ventral abdomen. Surgery was scheduled 4 weeks later; however, this was postponed owing to the development of ulcerative skin lesions 2 months after starting piroxicam (day 60).

Skin lesions were treated by the rDVM with oral amoxicillin/clavulanate potassium (Clavamox; Zoetis) at 14.9 mg/kg PO q12h for 7 days. No response to antibiotic therapy was noted and an aerobic skin bacterial culture and sensitivity test revealed no growth. Piroxicam was then discontinued (day 67). Ten days after discontinuing piroxicam, the ulcerative skin lesions improved, and cefovecin (Convenia; Zoetis) at 7.6 mg/kg SC and prednisolone at a reducing anti-inflammatory oral dose were prescribed by the rDVM (day 77) for a presumed flare of allergic dermatitis.

Three weeks after stopping the piroxicam (day 88), surgery was performed at the UMNCVM to remove the right mandibular salivary gland along with draining regional lymph nodes and a thoracic wall lymph node for staging purposes. No evidence of neoplasia was found. Moreover, there was no report of skin lesions at that time. The cat had an additional surgery 5 weeks later (day 123) and 8 weeks after discontinuation of piroxicam, to obtain wider margins at the initial surgery site. Histopathology of a lymph node revealed a focus of metastatic adenocarcinoma.

The cat was evaluated by the oncology service 2 weeks later (day 137) and was once more prescribed piroxicam at a dosage of 0.39 mg/kg PO q48h for palliative treatment of the adenocarcinoma.

The cat presented to the UMNCVM dermatology service 12 days later (day 149). After a thorough history, it became evident that the lesions developed 2 days after re-starting piroxicam and 5 months since the first treatment with piroxicam (Figure 1). Two days prior to presentation at the UMNCVM dermatology service, an aerobic skin bacterial culture was taken by the rDVM, which revealed no growth.

Figure 1

Timeline of drug administration and onset of clinical signs

The general physical examination was unremarkable. The dermatological examination revealed large areas of ulceration symmetrically distributed to the axilla, ventral abdomen, inguinal areas and medial aspect of both front- and hindlimbs (Figure 2). There was a tan-colored mucopurulent exudate covering the lesions and a thick black crusting at the periphery. The cat was noted to obsessively lick the affected areas.

Figure 2

Ulcerated lesion at presentation. There is a large ulceration in the axilla with central exudation and thick peripheral crusting

Cytology of the exudate revealed numerous degenerate neutrophils, but no microorganisms. Complete blood cell count revealed mild segmented neutrophilia (13.89 × 10³ µ/l; reference interval [RI] 1.2–13.2 ×10³ µ/l). Serum biochemical abnormality showed mild hypophosphatemia (2.9 mg/dl; RI 3.3–7.8 mg/dl). Differential diagnoses included a severe flare of allergic dermatitis, eosinophilic granuloma complex and cutaneous ADR.

Piroxicam was discontinued (day 149) and prednisolone (compounded 10 mg/ml; University of Minnesota) 1.8 mg/kg PO q24h was prescribed after a 3 day washout period (day 152) for presumptive severe allergic dermatitis. No improvement was noted after 6 days of treatment. Prednisolone was then discontinued and a biopsy was performed 2 days later (day 160). Skin biopsies were taken under general anesthesia with a 6 mm punch and samples were fixed in 10% buffered formalin. Sections were paraffin-embedded and stained with hematoxylin and eosin following standard protocols. Histopathology showed epidermal ulceration covered with a thick fibrinonecrotic exudate (Figure 3a). A focal area of intact epidermis revealed marked hyperplasia with compact orthokeratotic hyperkeratosis, hypergranulosis, acanthosis and spongiosis. There was focal dyskeratotic keratinocytes within the granular layer and remnants of these cells were also found in the stratum corneum (Figure 3b). The underlying dermis was moderately edematous. There was mild dermal infiltrate of eosinophils and mast cells admixed with variable numbers of lymphocytes, plasma cells and fibroblasts. The histopathologic findings indicated a non-specific ulcerative dermatitis.

Figure 3

(a) Hematoxylin and eosin photomicrograph (× 10) revealing abrupt transition from severe epidermal hyperplasia, orthokeratotic hyperkeratosis and focal parakeratosis to ulceration covered with fibrinonecrotic exudate. Moderate perivascular-to-interstitial dermal inflammation and edema. (b) Hematoxylin and eosin photomicrograph (× 20) showing multiple individualized apoptotic/dyskeratotic keratinocytes present in the stratum granulosum. Mild perivascular to interstitial dermal inflammation. There is no lymphocyte satellitosis

The skin lesions significantly improved approximately 3 and 2 weeks after piroxicam and prednisolone discontinuation, respectively. Dermatological examination showed only few small crusts on the axillae, ventral abdomen and inguinal regions. Most of the crusts were accompanied by a few, mildly erythematous, serpiginous lesions extending from the central areas of crusting (Figure 4). Erythematous papules were noted on the dorsal aspect of the head and caudal neck and the owner reported that the cat was itching at these lesions. Prednisolone was restarted at 1.8 mg/kg PO q24h and tapered to 1 mg/kg PO q48h to control a flare of seasonal allergic dermatitis. There was no recurrence of ulcerative lesions 6 months after discontinuation of piroxicam (Figure 5). The Naranjo score (see the supplementary material) was 6 indicating piroxicam as a probable cause for the cutaneous lesions.⁵

Figure 4

Ulcerated lesion 3 weeks after discontinuing piroxicam. Note the erythematous, serpiginous lesion extending from a central area of crusting in the axilla

Figure 5

Scar in axilla 6 months after discontinuing piroxicam

Discussion

To our knowledge, this is the first report of a probable cutaneous ADR to piroxicam in a cat. The most convincing evidence that supported this diagnosis was the repeatable distinctive ulcerative lesions seen 2 days after piroxicam was re-administered to the patient. Generally, re-administration of a drug suspected to cause an adverse reaction is not recommended as the subsequent reaction can be more severe and even life-threatening; however, piroxicam was unintentionally re-challenged in this case.⁶

NSAIDs have been previously shown to cause cutaneous ADRs in both dogs and people.^{78910111213141516171819–20} Carprofen and firocoxib have been associated with skin eruptions resembling Sweet’s syndrome in four dogs.^89–10 Carprofen has also been associated with toxic epidermal necrolysis in a dog.¹¹ In another dog, meloxicam caused vasculitis, erosions and ulcers.¹² In people, reports of cutaneous reactions to piroxicam include photosensitivity, urticaria, erythema multiforme, toxic epidermal necrolysis, pemphigus vulgaris, fixed drug eruption, drug rash with eosinophilia and systemic symptoms (DRESS syndrome), and acute generalized pustulosis.^{1314151617–18} Most cases resolved rapidly following discontinuation of piroxicam.^{1314151617–18} ADRs from NSAIDs are thought to be caused by a hypersensitivity reaction, which can be classified as either cross-reactive among chemically unrelated NSAIDs or selective to a specific NSAID.^19,20 Cross-reactive reactions occur through non-immunological mechanisms related to arachidonic acid metabolism.^19,20 Selective reactions are induced by immunological mechanisms and are either IgE- or T-cell-mediated reactions.^19,20

Piroxicam has been prescribed for long-term use in cats with a variety of neoplasms for its anti-tumor effects believed to be related to the inhibition of cyclooxygenase and, ultimately, of prostaglandins.^{212223242526–27}

A limitation of this case report is the various drugs administered to the cat before and after the onset of distinctive ulcerative lesions. However, the Naranjo scale used in this case to estimate the probability of an ADR to piroxicam revealed a score of 6, indicating a probable causal relation of piroxicam and the skin lesions. In addition, 3 months prior to presentation, similar ulcerative lesions were noted by the rDVM after the cat had received piroxicam for 2 months. At that time, the skin healed after discontinuing piroxicam and treatment with prednisolone and a single injection of cefovecin. Moreover, the rapid development of similar skin lesions after repeat drug exposure and the lack of response to systemic glucocorticoids after piroxicam discontinuation further supported a probable cutaneous ADR. In fact, the use of glucocorticoids in treating drug reactions is controversial as response can be poor.^4,6 After skin biopsy was performed, no treatment was initiated and the skin lesions rapidly improved over the following 2 weeks, adding to the evidence that piroxicam likely caused the ulcerative cutaneous lesions in this case.

Skin biopsies are often performed when suspecting an ADR to rule out other possible dermatoses and support a presumptive diagnosis. The histopathologic findings in this case were non-specific and it is possible the systemic glucocorticoid administered a few days before skin biopsies were performed altered the inflammatory reaction pattern observed histologically. Nevertheless, not all histologic changes associated with cutaneous ADRs are characteristic of a specific disease.

Conclusions

This is the first known report of a NSAID causing cutaneous ADR in a cat. Discontinuation of piroxicam resulted in complete resolution of skin lesions within 4 weeks. Cutaneous ADRs should be suspected when new skin lesions develop after a medication is initiated. Any suspected drug reactions should be communicated to the owner and all attending veterinarians and should be well documented in the medical record.

Supplemental Material