Comparison of clinical history and dermatologic findings

in 29 dogs with severe eosinophilic dermatitis:

a retrospective analysis
Blackwell Publishing Ltd

Elizabeth A. Mauldin*, Brian S. Palmeiro*,

Michael H. Goldschmidt and
Daniel O. Morris*
*Department of Clinical Studies and Pathobiology, University of
Pennsylvania, School of Veterinary Medicine, Philadelphia,
Pennsylvania, USA
This study was supported by the Laboratory of Pathology and
Toxicology, University of Pennsylvania, School of Veterinary Medicine.
Correspondence: Elizabeth A. Mauldin, Assistant Professor,
University of Pennsylvania School of Veterinary Medicine, Laboratory
of Pathobiology and Toxicology, 3800 Spruce Street, Philadelphia, PA
19104-6051, USA. E-mail: emauldin@vet.upenn.edu

What is known about the topic of this paper

Acute severe eosinophilic dermatitis in dogs develops
secondary to a number of triggering factors (insect
bites, drugs, atopy, etc.).
Prior study showed similarities to Wells syndrome in
people.
What this paper adds to the field of veterinary
dermatology
Many cases followed treatment of severe
gastrointestinal illness.
Clinical lesions may resemble vasculitis or erythema
multiforme.
May have causal drug associations

Abstract
The medical records and histopathological sections
of 29 dogs diagnosed with a unique eosinophilic dermatitis resembling Wells syndrome were reviewed
in an attempt to elucidate the pathogenesis of this
syndrome. The medical records were reviewed for
information on dermatological lesion appearance, systemic signs in other organ systems, clinical analyte
abnormalities, and drug therapy. Histological sections
of dogs with moderate to severe eosinophilic dermatitis
without folliculitis and furunculosis were reviewed
and evaluated for the presence of collagen flame figures.
Three categories of patients were found. Category 1
consisted of 17 dogs treated for vomiting and/or diarrhoea (often haematochezia or haematemesis) prior
(mean: 4.6 days) to the onset of skin lesions. Fourteen
category 1 dogs had erythematous lesions (macules,
papules or plaques) that were most pronounced on
the abdomen. Sixteen of the 17 dogs received multiple
classes of drugs, and 59% were hypoalbuminemic.
Category 2 consisted of five dogs that had skin lesions
338

and gastrointestinal signs at presentation and four

of these dogs were hypoalbuminemic. Category 3
included seven dogs without enteric illness. A positive
drug score was found in six category 1 dogs and
one each from categories 2 and 3. Eighteen cases
had eosinophilic dermatitis without flame figures,
seven cases had early flame figures and four had
well-developed flame figures. These changes did
not correlate with the categories of clinical presentation. More than 50% of the dogs developed eosinophilic dermatitis following treatment for severe
gastrointestinal disease. The authors propose that this
represents a unique syndrome that may have causal
drug association.
Accepted 30 July 2006

Introduction
A unique eosinophilic dermatitis was originally described
in nine dogs at the Matthew J. Ryan Veterinary Teaching
Hospital of the University of Pennsylvania. This disorder
was characterized by an erythematous maculopapular
eruption with marked eosinophilic inflammation, collagen
flame figures and dermal oedema. This disorder was
thought to be similar to eosinophilic cellulitis in humans,
also known as Wells syndrome. All dogs had bright red
papules or plaques on the ventral abdomen, thorax, and
pinnae that often developed from erythematous macules
with central clearing. The stimuli that preceded the
development of lesions included arthropod bites, food
hypersensitivity, drug (metronidazole) and allergic skin
disease.1
In humans, Wells syndrome is a clinical entity diagnosed
by the clinical presentation, disease course, and histological
features. Affected people may develop prodromal pruritus
or burning followed by annular areas of erythema, blisters
or urticaria. The lesions progress to form large, painful, and
often recurrent, erythematous plaques that may have a
rosy to violaceous hue and central clearing. The swellings
resemble infectious cellulitis but do not respond to antimicrobial therapy.2 4 The histological features depend on the
stage of the lesion at the time of biopsy, and it is important
to note that presence of flame figures is not specific for
Wells syndrome. In the acute stage, the dermis is oedematous and infiltrated by eosinophils. At a later stage, the
eosinophils degranulate and the granules and their contents
surround dermal collagen fibres to form flame figures. As
the lesion progresses, macrophages may palisade around

2006 The Authors. Journal compilation 2006 European Society of Veterinary Dermatology. 17; 338347

Canine eosinophilic dermatitis

the material and sometimes form giant cells.5 8 A variety

of triggering factors have been documented such as,
insect bites, drugs, vaccination, intestinal parasitism and
myeloproliferative disorders.214
After the initial report of Wells syndrome-like disease in
dogs, the authors noticed similar clinical and histological
features in several consecutive cases medically treated
for severe vomiting and diarrhoea 2 to 7 days prior to the
onset of skin lesions. The aim of this retrospective study
was to further investigate and characterize inciting causes
of acute eosinophilic dermatitis in dogs.

Materials and methods

This study was designed in a retrospective manner to evaluate dogs
that had a histopathological diagnosis of moderate to severe
eosinophilic dermatitis. A computer search for cases of moderate
to severe eosinophilic dermatitis in dogs was performed from the
University of Pennsylvania, Laboratory of Pathology and Toxicology
biopsy database from 1 January 2000 to 31 July 2005. Histological
sections were re-evaluated in all cases.

Inclusion criteria
Inclusion criteria were intentionally broad and consisted of the
following:
1 Histological findings of moderate to severe eosinophilic dermatitis with or without collagen flame figures and an absence of
eosinophilic folliculitis and furunculosis.
2 Acute onset of severe erythema or erythroderma with or without
wheals.
3 Persistence of lesions for greater than 48 h.
4 Follow up on each case via medical records and phone calls to
local veterinarians or owners.

Drug score
A drug score was assigned for each case using a modified method
previously described.15,16 This system assigns a numerical score (+3
to 3) based on the following criteria:
1 A score of +1 (suggestive) was given if the lesions appeared
more than 7 days after first administration of the drug or less
than 1 day after re-administration.
2 A score of +1 was given if the lesions resolved after removal of
the drug with no other therapeutics. A score of 0 (inconclusive)
was assigned if other drugs were given simultaneously. A
score of 1 (incompatible) was given if the lesions persisted
despite drug withdrawal or if the patient improved without drug
withdrawal.
3 A score of 1+ was assigned if the drug was re-administered and
the patient relapsed; a score 0 if no rechallenge occurred and a
score 1 if no recurrence was seen after challenge.
4 A positive drug score was deemed suggestive of a casual drug
association. A zero score was inconclusive. A negative score
indicated that a drug association was unlikely.

Results
Twenty-nine cases met the inclusion criteria. The cases
were further subdivided into three categories. Category 1
consisted of dogs presenting for the treatment of severe
vomiting and/or diarrhoea that preceded the development of skin lesions. Category 2 consisted of dogs with
vomiting and/or diarrhoea concurrent with the onset of
skin lesions. Category 3 consisted of dogs evaluated by
a veterinarian for skin disease without gastrointestinal
signs.

Category 1 Gastrointestinal signs prior to skin

lesions
Signalment
Seventeen dogs met the category 1 criterion (Table 1).
Twelve were spayed females, and five were castrated males.
Their median age was 4 years (range: 0.5812 years). Ten
different breeds were represented with Labrador retrievers
(5) and Shih Tzus (2) being represented more than once.
There were two mixed breed dogs.
Clinical presentation
Fourteen dogs were hospitalized for severe vomiting and/
or diarrhoea with or without haematochezia (8), haematemesis (2) or both (2). One dog was treated symptomatically
on an outpatient basis for vomiting and diarrhoea. Two dogs
were hospitalized for lethargy and skin lesions after having
diarrhoea that improved with medical treatment.
Initial therapy
All dogs were treated for the gastrointestinal illness 1
to10 days (mean SD: 4.6 2.5 days) prior to the onset of
skin lesions (Table 1). Only one dog received single-drug
therapy (metronidazole). Eleven dogs were treated with
metronidazole in addition to multiple other drugs (one
received metronidazole and loperamide only). Ten cases
received beta-lactam antibiotics in combination with other
therapeutics. Dog 15 was treated for vomiting and diarrhoea 4 days following routine vaccination. At the same
visit, the dog received a maintenance allergen immunotherapy injection (given monthly for 2 years) from a new
allergen vial. Skin lesions appeared approximately 1224 h
after treatment of vomiting and diarrhoea. Dog 17 was
prescribed doxycycline for lameness attributed to probable
Lyme disease. This medication was prescribed for a similar
problem 6 months earlier. The dog subsequently developed
vomiting and retching and was prescribed cough tablets.
This was followed by severe diarrhoea and swollen joints.
The skin rash was noted upon repeat examination by the
referring veterinarian a few days following worsening of
gastrointestinal signs.
Skin lesions
In 15 cases, the dermatological lesions consisted of severe
erythema with erythematous to haemorrhagic macules or
plaques which were most pronounced on the ventral
abdomen and limbs (Figs 1, 2). Four dogs had erythematous macules with central clearing (targetoid macules). In
addition, three dogs had wheals, one dog was reported to
have facial swelling, and two had generalized oedema.
Four dogs were reported to have pruritus with the onset
of skin lesions.
Laboratory data
The most consistent change in the laboratory data was low
serum protein (59% cases). Seven dogs had hypoalbuminaemia prior to fluid therapy; while three additional dogs
developed hypoalbuminaemia during fluid therapy. One
dog had hypocholesterolaemia. None of the dogs had
alterations in blood eosinophils. One dog had thrombocytopenia and another had a history of thrombocytopenia but
platelets were normal at presentation (current therapy

2006 The Authors. Journal compilation 2006 European Society of Veterinary Dermatology.

339

 2006 The Authors. Journal compilation 2006 European Society of Veterinary Dermatology.

Skin lesion
onset (days)*

Lesion type

Location

Treatment

Amoxicillin, aminopentamine
hydrogen sulphate,
metronidazole, penicillin,
reglan

Erythema, wheals

Ventrum, pinnae,
generalized

Hydroxyzine,
mesalamine, limited
antigen diet

Vomiting, diarrhoea,
haematochezia

Cholodin, metaclopromide,
metronidazole, ursodiol

Erythema, erythroderma,
wheals (dorsum), oedema
(pinnae and limbs)

Ventrum, pinnae,
generalized

Predisone
pentoxifylline

Vomiting, diarrhoea,
haematochezia,
haematemesis

Ampicillin, baytril,
cimetidine, hetastarch

Erythema, wheals,
oedema, facial swelling

Ventrum, legs, face

Prednisone
diphenhydramine,
sucralfate,
supportive care

Mixed breed

Diarrhea, haematochezia

Amoxicillin, metronidazole,
pepcid, sucralfate

Erythematous macules,
papules

Ventrum, legs, dorsum

Prednisone

Shiz tzu

Vomiting, diarrhoea

Doxycycline, metaclopromide,
metronidazole, prednisone

Erythema with targetoid

red macules

Generalized

Prednisone

Boxer

Vomiting, haematemesis,
diarrhoea

Ampicillin, chlorpromazine,
famotidine, metaclopromide,
metronidazole

Erythema with targetoid

red macules

Ventrum to generalized

Cephalexin, prednisone,
betamethasone/
gentocin spray

Italian
greyhound

Vomiting, diarrhoea,
haematochezia

Ampicillin, enrofloxacin,
famotidine, metronidazole,
hetastarch, fresh frozen plasma

Erythematous macules

Ventrum, neck, thorax

Diphenhydramine,
prednisone

Labrador
retriever

Vomiting, diarrhoea

Morphine, metaclopromide,
sucralfate

10

Erythematous macules

Ventrum

Diphenhydramine

Labrador
retriever

Vomiting, diarrhoea,
haematochezia

Metronidazole, metaclopromide,
hetastarch

Haemorrhagic to
erythematous plaques

Ventrum

Cefpodoxime,
metoclopromide,
prednisone

10

Basset hound

12

Vomiting, diarrhoea,
haematochezia

Ampicillin, enrofloxacin,
famotidine, metronidazole,
prednisone, sucralfate

Erythroderma, papules,
targetoid macules

Ventrum, limbs

Prednisone,
diphenhydramine,
enrofloxacin, novel
protein diet

11

Labrador
retriever

0.58

Vomiting, haematemesis,
diarrhoea

Amoxicillin, famotidine;
metronidozole
(after skin lesions)

Erythematous papules

Axillae, groin, ventrum

0.015% triamcinolone
acetonide, Purina HA

Case

Breed

Jack Russell
terrier

Age
(year)

Chief complaint

Drugs administered

1.5

Vomiting, diarrhoea,
haematochezia

Doberman

Labrador
retriever

Mauldin et al.

340
Table 1. Clinical features of category 1 dogs

 2006 The Authors. Journal compilation 2006 European Society of Veterinary Dermatology.

Table 1. Continued

Case

Breed

Age
(year)

Chief complaint

Drugs administered

Skin lesion
onset (days)*

Lesion type

Location

Treatment

12

Mixed breed

Diarrhoea

Metronidazole

Erythematous macules

Ventrum, pinnae

Cephalexin,
diphenhydramine,
prednisone

13

Labrador
retriever

1.4

Vomiting, diarrhoea,
haematochezia

Ampicillin, erythromycin,
enrofloxacin, metronidazole

Targetoid erythematous
macules, plaques,
oedema (face)

Ventrum, legs, face

Intensive care;
prednisone;
methylprednisolone;
Purina HA

14

Shih Tzu

Diarrhoea

Loperamide, metronidazole

Erythroderma, crusting
(head)

Ventrum, pinna, head

Diphenhydramine,
hydoxyzine

15

Golden retriever

Vomiting, haematochezia

Enrofloxacin, tagamet,
sucralfate

Erythema, oedema

Generalized

Prednisone,
diphenhydramine,
sucralfate, misoprostol

16

Akita

Vomiting, haematamesis,
haematochezia

Amoxicillin clavulanic acid,

sucralfate

Erythema, oedema,
crusting (pinnae/head)

Pinnae, head,
hindlimbs

Prednisone,
sucralfate, cimetidine,
misoprostol

17

Miniature
schnauzer

12

Vomiting, diarrhoea,
joint swelling

Doxycycline, cough tablets

Erythema, macules,
papules

Ventrum, limbs

Intensive care;
enrofloxacin

*Time period (days) between drug administration for treatment of the gastrointestinal problem and the development of skin lesions.
In all cases, treatment consisted of withdrawal of any medication prescribed prior to the onset of skin lesions.

Canine eosinophilic dermatitis

341

Mauldin et al.

Figure 1. Clinical photograph: Labrador retriever, category 1.

Severe erythema with slightly raised coalescing papules in the
abdominal skin.

Figure 2. Clinical photograph: Italian greyhound, category 1.

Coalescing erythematous macules on the abdomen.

0.25 mg kg1 other day). One dog each had leukocytosis,

leukopenia, neutropenia and mildly increased alkaline
phosphatase (ALP) and alanine aminotransferase (ALT). All
dogs documented as having a faecal exam performed, had
negative results for parasites. Dog 4 had a history of recurrent giardiasis, but faecal flotation and Giardia antigen test
were negative at the time of presentation.
Additional diagnostics
Three dogs had endoscopic biopsy results consistent with
inflammatory bowel disease: mild to moderate lymphoplasmacytic ileitis/typhlitis/colitis (case 1), moderate lymphoplasmacytic and eosinophilic gastroenteritis and colitis
(case 8), and severe eosinophilic gastroenteritis and colitis
(case 10).
Drug score
Six cases had a drug score of +1, whereas 10 cases had
an equivocal score (0). Dog 2 was euthanized within days
of the onset of skin lesions; therefore, a drug score was
not assigned.
342

Treatment
All 17 dogs were uniformly treated by withdrawal of any
drug, or classes of drugs (e.g. beta-lactam antibiotics) that
were prescribed prior to the onset of skin lesions; however, none was treated solely with drug withdrawal. Drugs
(usually multiple) were prescribed for the gastrointestinal
signs, skin disease or both and included gastrointestinal
protectants, antibiotics, antihistamines, and glucocorticoids. Twelve dogs received prednisone (dose range: 0.8
2 mg kg1) or methylprednisolone (1.7 mg kg1). One dog
had topical concurrent betamethasone/gentamicin spray
and one dog received only topical 0.015% triamcinolone
acetonide glucocorticoid (Genesis topical spray, Virbac
AH, Inc., Fort Worth, TX, USA). Eight dogs were prescribed
antihistamines (hydroxyzine or diphenhydramine), five
dogs oral antibiotics (cephalosporins, fluoroquinolones)
and four dogs a limited antigen diet or hydrolysed protein
diet. In dog 17, the gastrointestinal signs and erythema
decreased with withdrawal of doxycycline, but the dog
remained severely lethargic. This aged dog was thought
to have multiple problems including leukocytosis of undetermined cause, positive Lyme titre, decreased albumin
and proteinuria. This dog was subsequently treated with
famotidine, metronidazole, ampicillin, gentamicin and
corticosteroids. After the biopsy results, all remaining
drugs were withdrawn and the dog was given intravenous
followed by oral enrofloxacin.
Follow up
Follow-up information was available on 16 cases. Dog 2
was euthanized a few days after the onset of skin lesions.
This dog had a history of chronic active hepatitis and a low
grade pulmonary carcinoma that was excised 8 months
prior. Of the 16 remaining cases, the average follow-up
time was 1.7 years (median: 0.9 years, range: 0.2 4.5 years).
Time to resolution was difficult to assess as many dogs
were not rechecked within the month following hospitalization. In three dogs, skin lesions resolved within 1 week
and in one dog within 4 weeks. Dog 13 was unusual in that
it had chronic pruritus and skin lesions for 6 months following the initial bout of severe vomiting and diarrhoea, then
the pruritus and skin lesions resolved completely and did
not recur during the follow-up period of 2.5 years. This dog
was maintained on Purina HA diet (Nestl Purina Petcare
Co., St. Louis, MO, USA). None of the 16 dogs had a recurrence of the skin lesions. Dog 1 had recurrent diarrhoea
and pruritus but did not develop any similar cutaneous
lesions. Dog 5 had a repeat episode of diarrhoea 5 months
later. Dog 7 had one episode of vomiting and diarrhoea
7 months following hospitalization. Six weeks after initial
presentation, dog 10 developed recurrent and severe
haemorrhagic vomiting and diarrhoea and was treated
with metronidazole with no recurrence of skin lesions.
This dog was euthanized for uncontrollable gastrointestinal
disease 2 weeks following relapse.
Category 2 Gastrointestinal disease associated with
skin lesions
Signalment
Five dogs presenting for treatment of gastrointestinal
disease also had skin lesions. This category included one

2006 The Authors. Journal compilation 2006 European Society of Veterinary Dermatology.

Canine eosinophilic dermatitis

intact and two castrated male dogs and two spayed

females. The median age of the dogs was 6 years (range:
49 years). Their breeds included Bichon frise (1), Pug (1),
Boston terrier (1) and two mixed breeds.
Clinical presentation
Four of the five dogs had a clinical presentation similar to
category 1 dogs, although the skin lesions or pruritus was
reported to develop with the gastrointestinal signs. One
dog had skin lesions just prior to developing diarrhoea. A
9-year-old mixed breed dog (dog 19) had pruritus at the
time of presentation and developed the skin lesions 2 days
after being treated with amoxicillin. One dog (dog 20) had
concurrent allergic skin disease, and a 6-year-old Bichon
frise (dog 18) also had urolithiasis.
Skin lesions
Four of the five dogs had cutaneous lesions located on
the ventral abdomen and thorax that were described as
severe erythema and wheals (2), erythematous papules
and wheals (1) and coalescing erythematous plaques (1).
One of the four dogs also had plaques on the head, generalized erythema and pinnal oedema. The remaining dog
had generalized erythema with erythematous ear canals.
Two dogs were reported to be pruritic while one dog had
painful lesions.
Laboratory data
Four of five dogs were hypoalbuminaemic, and one dog
each had mild leukocytosis, mildly increased ALT, mildly
increased ALP, and thrombocytopenia.
Drug score
A 7-year-old mixed breed dog (dog 20) was the only case
having a positive (+1) drug score. This dog was being
treated with enrofloxacin for a urinary tract infection prior
to the onset of gastrointestinal disturbances and skin lesions.
The four remaining dogs had an equivocal score (0).
Treatment
Treatment consisted of drug withdrawal (3/5), diphenhydramine (2/5), prednisone (3/5), famotidine (2/5), pentoxifylline (1/5) and topical chlorhexidine (1/5).
Follow up
None of the skin lesions recurred in any dog (median:
2 years, range: 0.75 6 years).
Category 3 Skin lesions without gastrointestinal
signs
Signalment
Seven dogs presented solely for evaluation of skin disease.
This group included four spayed female and three male
castrated dogs. Their breeds consisted of Labrador
retriever (2), Vizsla (1), Basenji (1), Schnauzer (1) and two
mixed breed dogs. The mean age of the dogs was
6.2 years (range: 112 years).
Clinical presentation
Lesions in the seven dogs were reported to develop
abruptly and progressed over a period of several days

(5/7) or weeks (1/7), to 1 month (1/7). A 1-year-old, spayed

female Vizsla (dog 27) developed a few erythematous
nonpruritic papules 1 week prior to vaccination for Lyme
disease. This dog was treated with a glucocorticoid injection
at the time of vaccination in addition to cephalexin, prednisone, and chlorhexidine shampoo. The lesions dramatically worsened over the following 3 days. A 4 -year-old,
male castrated Labrador retriever (dog 26) presented to a
referral institution after being nonresponsive to numerous
medications including oral and injectable corticosteroids.
Skin lesions
The locations and characteristics of the cutaneous lesions
were highly varied among the dogs in this group. All dogs
developed erythema with papules being the primary
lesions in three of seven dogs. Five of seven dogs were
pruritic. The location of lesions varied to include the nasal
planum (5/7), ventrum (4/7), trunk and flank (1/7) and axilla
(1/7). Generalized erythroderma was noted in one dog.
Laboratory data
Clinical pathological changes included a mild increase in
ALT in one dog and ALP in another dog. Dog 26 had transient anaemia, transient hyperglycaemia, glycosuria and
hypoalbuminaemia. Two dogs did not have bloodwork
performed.
Drug score
A 7-year-old, male castrated Basenji (dog 23) was the only
dog to receive a positive drug score. This dog developed
skin lesions while being treated with paroxetine for a
behavioural disorder. The dog developed acute facial
pruritus followed by ventral erythema with targetoid
macules that resembled lesions of the dogs in category 1
(Fig. 3).
Treatment
Treatment consisted of drug withdrawal in two dogs (Vizsla
cephalexin and Basenji paroxetine). All dogs received
glucocorticoids (prednisone (6/7), methylprednisolone
(1/7)), and some of the dogs received antihistamines (2/7),
doxycycline (1/7), misoprostol (1/7), cephalexin (1/ 7), and
fluoroquinolone antibiotics (2/7).
Follow up
The mean follow-up time in this group was 1.8 years
(range: 0.64 years). A 7-year-old, female spayed Schnauzer (dog 29) was the only dog documented in this study to
have recurrence of skin lesions. This dog was followed up
for 8 months and had waxing and waning lesions since the
skin biopsy was performed.
Histopathology for categories 1, 2 and 3
Eighteen dogs had eosinophilic dermatitis without alterations
in dermal collagen. Seven dogs had subacute eosinophilic
dermatitis with early flame figures and four dogs had
eosinophilic dermatitis with well-developed flame figures.
Three dogs had eosinophilic intraepidermal pustules with
two of these dogs having gastrointestinal signs and one
having only skin signs. No correlation was found between
the presence or absence of gastrointestinal signs and
histopathological changes (Figs 4, 5).

2006 The Authors. Journal compilation 2006 European Society of Veterinary Dermatology.

343

Mauldin et al.

Figure 3. Clinical photograph: Basenji, category 3. Erythematous

macules with central clearing on the abdomen and medial hindlimbs.
(a) Closer examination of lesions from the same dog. The black circles
note the clinical progression from small macules to larger coalescing
deeply red macules with central clearing.

Discussion
The authors embarked on this retrospective study after
noticing acute skin eruptions in several dogs hospitalized
for treatment of gastrointestinal disease. Those cases
resembled the eosinophilic dermatitis syndrome described
by Holm et al.1 Broad inclusion criteria were purposely
maintained to determine whether this was a coincidence
or a general trend in the Veterinary Teaching Hospital referral population. The authors included in-house cases as well
as skin biopsy samples submitted to the dermapathology
service from veterinary practices in the USA. Three subpopulations of dogs were identified and the largest one
(17 of 29, category 1) had a clinical course similar to the
hospitalized patients.
The typical cases in category 1 were middle-age dogs that
were being treated for severe vomiting and/or diarrhoea
with or without haematochezia and haematemesis. After
approximately 4 days of symptomatic therapy, the gastrointestinal disease improved, but the dogs developed
diffuse erythroderma with large coalescing macules or
344

Figure 4. (a) Photomicrograph: category 1, dog 17. Noted the marked

cellular infiltrate within the dermis with oedema and flame figures
(arrows). Haematoxylin and eosin (H&E) 4. (b) Photomicrograph:
Higher magnification of 4a. Large clusters of eosinophils surround
hyalinized dermal collagen fibers. H&E 10.

papules that were most severe on the abdomen. The skin

reaction resembled erythema multiforme or vasculitis and
these were often noted as differential diagnoses on the
biopsy submission form. Although a drug eruption was
suspected, identifying a causative drug was problematic
since 16 of the17 category 1 dogs had been treated with
several drugs. This often included multiple classes of antibiotics, H1 blockers, and gastroprotectants. Some dogs
also received colloids for treatment of hypoproteinaemia.
The skin lesions in category 1 dogs resolved with drug
withdrawal and symptomatic treatment and did not recur.
Only six of the 17 dogs in category 1 had a positive drug
score; however, all dogs developed lesions during the
course of treatment and none had a history of similar skin
lesions before treatment. The lack of positive scoring does
not completely rule out a drug association. It is well known
that most immunologically mediated cutaneous eruptions
occur from 7 to 14 days following initiation of a new medication, and the skin reaction can occur much sooner upon
re-exposure to the same drug.17 A detailed and complete
drug history was difficult to ascertain and it is possible

2006 The Authors. Journal compilation 2006 European Society of Veterinary Dermatology.

Canine eosinophilic dermatitis

Figures 5. (a) and (b) Photomicrograph: category 1, dog 12. The

dermis contains a moderate to severe infiltrate of eosinophils with
fewer macrophages. H&E. 2 and 20.

that adult dogs were treated in the past for episodes of

diarrhoea with the same or similar drugs and thus underwent prior sensitization. In addition, the drug classification
scheme does not account for multiple drug interactions
and nonimmunological drug reactions. Furthermore, acute
reactions occurring within hours of drug administration
without previous exposure are well documented in people
and are thought to be T-cell mediated .18 These reactions
may be mediated by preactivated T cells that have crossreactivity with drugs and other peptides.19 Finally, most literature in drug eruption is obtained from human studies,
and it is possible that dogs may not react in the same timeline. In fact, if the drug-scoring scheme was altered to
include dogs that developed eruptions at 4 days after
drug administration (vs. adopted scheme of 7 days) four
additional cases would have positive scores resulting in
more than half of category 1 dogs.
The health status of the dogs should also be considered
as the dogs in category 1 were being treated for, often
severe, enteric illness when the drugs were administered.
It is well known that immunocompromised humans have
a paradoxically higher incidence of adverse drug reactions.
This may be the result of immunological or nonimmunological

reactions that occur with multiple-drug therapy, chronic

drug exposure and altered immunity and metabolism.17
An example of this is a markedly increased incidence of
trimethoprimsulfamethoxazole hypersensitivity in patients
with AIDS.20 The dogs in this study developed cutaneous
lesions similar to adverse drug eruptions in humans that are
often maculopapular or exanthematous (meaning similar
to lesions caused by viruses). Beta-lactams are the most
commonly implicated drugs in maculopapular eruptions in
humans, and the rate of skin eruptions to beta-lactams is
markedly increased if the patients have viral infections.
Almost all humans with infectious mononucleosis will
develop morbilliform eruptions to aminopenicillins.17
Similarly, the presence of an undetected gastrointestinal
pathogen could possibly explain the strong association of
gastrointestinal signs and multiple drug administration
with skin lesions in these canine cases.
Another explanation is that the skin lesions in some dogs
could have arisen from a hypersensitivity reaction unrelated to drug administration. Endoparasitism has been
documented as a triggering cause for Wells syndrome in
human patients. The literature includes a number of case
reports of Wells syndrome casually linked to toxocariasis,
giardiasis and one report of oncocerciasis.9,11,13 In this
study, none of the cases that had faecal examination
performed had intestinal parasites. One dog (category 1)
had been treated for recurrent giardiasis but was negative
on faecal flotation and Giardia antigen test. Furthermore,
category 2 dogs also had a history of acute enteric illness;
however, the skin lesions were documented when the dogs
presented for treatment of the gastrointestinal problem.
The skin eruption and pruritus in conjunction with intestinal
disturbances may reflect a systemic hypersensitivity
reaction targeting both the gastrointestinal tract and the
skin. Perhaps this is a multifactorial syndrome or part of a
generalized systemic hypersensitivity reaction to other
allergens. In this study, at least one dog was eventually
diagnosed with food allergy while three others were therapeutically placed on limited antigen or hydrolysed diets.
Twenty-four per cent of the cases reported here had no
association with gastrointestinal disease. The veterinary
literature contains one case report of a dog that developed
urticaria and eosinophilic dermatitis following administration of diethylcarbamazine.21 This dog had dermatological
and histological features similar to category 1 and 2 patients
but without gastrointestinal disease. Furthermore, one dog
in category 3 developed skin lesions similar to category 1
dogs, 9 days after beginning treatment with paroxetine
for a behaviour disorder. These two cases may further substantiate a link between the cutaneous reaction and a
causal drug association. However, in the remainder of
category 3 dogs, the inciting factors were more difficult to
document. The dogs in this group had the most varied skin
lesions and clinical findings, and most likely represented a
number of different inciting causes unlike the majority of
cases reported herein.
Finally, these data must be interpreted in light of its retrospective nature. Determining the chronology of the skin
rash relative to drug administration may not be completely
accurate as it was sometimes skewed by clinicians
assumptions, and medical records were often incomplete
and sometimes inaccurate. In one case for example, a

2006 The Authors. Journal compilation 2006 European Society of Veterinary Dermatology.

345

Mauldin et al.

referring clinician strongly suspected that amoxicillin had

incited the skin eruption; however, careful review of the
primary care veterinarians record indicated that the skin
rash developed prior to amoxicillin administration.
The authors believe that this cutaneous reaction is
unique to the dog as it does not fit the clinical criteria for
Wells syndrome in human beings or a typical type 1
hypersensitivity-urticarial reaction. These cases were hallmarked by acute and persistent maculopapular eruptions,
and none had urticarial wheals as the sole presentation.
Furthermore, the dogs lacked the large indurated plaques
that are seen in people with Wells syndrome. Seventy-five
per cent of the dogs in this study developed skin lesions
in association with gastrointestinal signs, with 58% occurring during or after treatment for enteric disease and a
drug association was suspected in many cases. Veterinary
dermatopathologists should be prompted to ask about
drug exposure and enteric disease if the histopathology
shows severe eosinophilic dermatitis. This eosinophilic
dermatitis should also be considered in the differential
diagnosis by dermatologists and veterinary generalists if
a dog presents with an acute onset of clinical lesions
(erythema, erythroderma, papules, and macules) that
resemble erythema multiforme or vasculitis. Skin biopsy is
critical to rule out more severe diseases with similar clinical lesions as the outcome of the eosinophilic dermatitis
reported here is typically favourable.

Acknowledgements
The authors would like to thank Drs James Jeffers, Jean
Greek, Robert Schick and Norma White-Withers for their
dermatological expertise and case submission, and the
referring veterinarians who graciously supplied medical
record information. The authors also thank Dr Ken Drobatz
for statistical support.

References
1. Holm KS, Morris DO, Gomez SM et al. Eosinophilic dermatitis in
nine dogs compared with eosinophilic cellulitis in humans. Journal
of the American Veterinary Medical Association 1999; 215: 649
53.
2. Wells GC, Smith NP. Eosinophilic cellulitis. British Journal of
Dermatology 1979; 100: 1019.
3. Fisher GB, Greer KE, Cooper PH. Eosinophilic cellulitis (Wells
syndrome). International Journal of Dermatology 1985; 24: 1017.

4. Weiss G, Shemer A, Yitzchak C et al. Wells syndrome: report

of a case and review of the literature. International Journal of
Dermatology 2001; 40: 13352.
5. Aberer W, Konrad K, Wolff K. Wells syndrome is a distinctive
disease entity and not a histologic diagnosis. Journal of the Academy
of Dermatology 1988; 18: 10514.
6. Consigny S, Courville P, Young P et al. Histological and clinical
forms of the eosinophilic cellulitis. Annales de Dermatologie et de
Vnrologie 2001; 128 (3 Part 1): 2136.
7. Newton JA, Greaves MW. Eosinophilic cellulites (Wells syndrome)
with florid histologic changes. Clinical and Experimental Dermatology 1988; 13: 31820.
8. Moossavi M, Mehregan DR. Wells syndrome: a clinical and
histopathologic review of seven cases. International Journal of
Dermatology 2003; 42: 627.
9. van den Hoogenband HM. Eosinophilic cellulitis as a result of
onchocerciasis. Clinical and Experimental Dermatology 1983; 8:
4058.
10. Melski JW. Wells syndrome, insect bites, and eosinophils.
Dermatology Clinics 1990; 8: 28793.
11. Huni MA, Gerbig AW, Braathen LR et al. Toxocariasis and Wells
syndrome: a causal relationship? Dermatology 1997; 195: 3258.
12. Moreno M, Luelmo J, Monteagudo M et al. Wells syndrome
related to tetanus vaccine. International Journal of Dermatology
1997; 36: 51836.
13. Canonne d Dubost-Brama A, Segard M, Piette F et al. Wells
syndrome associated with recurrent giardiasis. British Journal of
Dermatology 2000; 143: 4257.
14. Sekin D, Demirhan B. Drugs and Wells syndrome: a causal relationship? International Journal of Dermatology 2001; 40: 13352.
15. Hinn AC, Olivry T, Luther PB et al. Erythema multiforme, Stevens
Johnson syndrome and toxic epidermal necrolysis in the dog:
clinical classification, drug exposure, and histopathologic
correlations. Veterinary Allergy and Clinical Immunology 1998; 6:
1320.
16. Prez T, Dayer E, Girard J-P. Hypersensitivity reactions to drugs:
correlation between clinical probability score and laboratory diagnostic procedure. Journal of Investigative Allergy and Clinical
Immunology 1995; 5: 27682.
17. McKenna JK, Leiferman KM. Dermatologic drug reactions. Immunology and Allergy Clinics of North America 2004; 24: 399423.
18. Gerber BO, Pichler WJ. Cellular mechanisms of T cell mediated
drug hypersensitivity. Current Opinion in Immunology 2004; 16:
7327.
19. Gerber BO, Pichler WJ. Noncovalent interactions of drugs with
immune receptors may mediate drug-induced hypersensitivity
reactions. American Association of Pharmaceutical Scientists
Journal 2006; 8: E160E165.
20. van der Ven AJ, Koopmans PP, Vree TB et al. Drug intolerance in
HIV disease. Journal of Antimicrobial Therapy 1994; 34: 15.
21. Vitale CB, Irke PJ, Gross TL. Putative diethylcarbamazine-induced
urticaria with eosinophilic dermatitis in a dog. Veterinary Dermatology, 1994; 5: 197203.

Rsum Les rapports mdicaux et les sections histopathologiques de 29 chiens prsentant une dermatite
osinophilique particulire ressemblant au syndrome de Well ont t revus dans le but dlucider la
pathognie de cette maladie. Les donnes mdicales tudies taient laspect des lsions cutanes, la
prsence de signes systmiques ou dans dautres organes, les anomalies biochimiques et les traitements.
Les sections histologiques des chiens avec une dermatite osinophilique modre svre sans folliculite
ou furonculose ont t revues pour la prsence de figures de collagne en flamme. Trois catgories de
malades ont t observes. La catgorie 1 consistait en 17 chiens traits pour vomissements et/ou diarrhe
(souvent avec hmatochzie ou hmatmse), avant (moyenne 4.6 jours) lapparition des lsions cutanes.
Quatorze chiens de ce groupe prsentaient des lsions rythmateuses (macules, papules ou plaques), qui
taient marques sur labdomen. Seize des 17 chiens ont reu de nombreux mdicaments et 59% taient
hypoalbuminmiques. La catgorie 2 consistait en cinq chiens qui prsentaient des lsions cutanes et des
symptmes digestifs initialement, et quatre de ces chiens taient hypoalbuminmiques. La catgorie 3
incluait sept chiens sans trouble digestif. Une imputabilit mdicamenteuse tait retrouve chez six chiens
du groupe 1 et chez un chien des groupes 2 et 3. Dix huit chiens prsentaient une dermatite osinophilique
sans figure en flamme, sept chiens avec des figures en flamme dbutantes, et 4 avec des figures en flamme
346

2006 The Authors. Journal compilation 2006 European Society of Veterinary Dermatology.

Canine eosinophilic dermatitis

importantes. Ces modifications histologiques ne corrlaient pas avec la prsentation clinique. Plus de 50%
des chiens ont dvelopp une dermatite osinophilique suivant un traitement pour trouble digestif svre.
Les auteurs proposent quil sagisse dun syndrome unique qui pourrait tre associ avec une cause
mdicamenteuse.
Resumen Los historiales mdicos as como secciones histolgicas de 29 perros diagnosticados con una
peculiar dermatitis eosinoflica parecida al sndrome de Well fueron revisados en un intento de desvelar la
patognesis del sndrome. Los historiales mdicos se revisaron buscando informacin referente a la
apariencia de las lesiones dermatolgicas, la presencia o no de signos sistmicos en otros aparatos,
anormalidades en los anlisis clnicos y la historia de tratamientos farmacolgicos. Las secciones histolgicas
de perros con dermatitis eosinoflica de moderada a severa, sin foliculitis ni furunculosis se revisaron y
evaluaron para la presencia de figuras en llama en el colgeno. Se encontraron tres categorias de pacientes:
la categoria 1 correspondi a 17 perros tratados para el vmito y/o diarrea (a menudo con hematoquezia
o hematemesis) previos a la aparicin de lesiones en la piel (media 4,6 dias). Catorce de estos perros presentaron lesiones eritematosas (mculas, ppulas o placas) que fueron ms pronunciadas en el abdomen.
Dieciseis de los 17 perros recibieron variados tratamientos farmacolgicos, y un 59% fueron hipoalbuminmicos.
La categoria 2 consisti en cinco perros que manifestaron lesiones de la piel y signos gastrointestinales en
la presentacin siendo cuatro de los perros hipoalbuminmicos. La categoria nmero 3 incluy siete perros
sin enfermedad gastrointestinal. Un valor positivo para tratamiento farmacolgico se encontr en seis perros
de la categoria 1 y en un perro de cada una de las otras dos categorias. Un total de 18 casos presentaron
dermatitis eosinoflica sin figuras en llama, siete casos tuvieron figuras en llama en estados tempranos y
cuatro manifestaron figuras en llama totalmente desarrolladas. Estos cambios no se asociaron con las
categorias de presentacin clnica. Ms de un 50% de los perros desarrollaron dermatitis eosinoflica tras
el tratamiento contra una enfermedad gastrointestinal severa. Los autores proponen que este proceso
representa un sndrome unico que podra tener asociacion causal con frmacos.
Zusammenfassung Die Krankengeschichten und die histopathologischen Schnitte von 29 Hunden, die
mit einer einzigartigen eosinophilen Dermatitis diagnostiziert worden waren, welche dem Wells Syndrom
hnlich ist, wurden als Versuch, die Pathogenese dieses Syndroms zu erklren, berarbeitet. Die Krankengeschichten wurden durchgesehen, um Informationen ber das Aussehen der dermatologischen
Vernderungen, ber systemische Anzeichen in anderen Organsystemen, ber klinische Abnormalitten der
Analyten und ber die Behandlung mit Medikamenten zu erhalten. Die histologischen Schnitte von Hunden
mit moderater bis ausgeprgter eosinophiler Dermatitis ohne Follikulitis und Furunkulose wurden berarbeitet und auf das Vorhandensein von kollagenen Flammenfollikeln berprft. Die Patienten wurden in drei
Kategorien eingeteilt. Kategorie 1 bestand aus 17 Hunden, die vor dem Auftreten von Hautlsionen (durchschnittlich 4,6 Tage) auf Erbrechen und/oder Durchfall (oft mit Blut im Kot oder Bluterbrechen) behandelt
wurden. Vierzehn der Hunde aus Kategorie 1 hatten erythematse Vernderungen (Maculae, Papeln und
Plaques), welche am Abdomen besonders ausgeprgt waren. Sechzehn der 17 Hunde erhielten verschiedene
Klassen von Medikamenten, und 59% zeigten Hypoalbuminmie. Kategorie 2 bestand aus fnf Hunden mit
Hautvernderungen und gastrointestinalen Symptomen bei der Erstuntersuchung. Vier dieser Hunde
zeigten Hypoalbuminmie. Kategorie 3 bestand aus sieben Hunden ohne intestinale Erkrankung. Eine
positive Medikamentenbeurteilung wurde bei sechs Hunden der Kategorie 1 gefunden und bei je einem
Hund aus den Kategorien 2 und 3. Achtzehn Flle hatten eine eosinophile Dermatitis ohne Flammenfiguren,
sieben Flle zeigten Flammenfiguren im Frhstadium und vier zeigten gut entwickelte Flammenfiguren.
Diese Vernderungen korrelierten nicht mit den Kategorien der klinischen Prsentation. Mehr als 50% der
Hunde entwickelten eosinophile Dermatitis nach der Behandlung einer schweren gastrointestinalen
Erkrankung. Die Autoren schlagen vor, dass es sich hierbei um ein einzigartiges Syndrom handelt, welches
mit einer Reaktion auf Medikamente im Zusammenhang steht.

2006 The Authors. Journal compilation 2006 European Society of Veterinary Dermatology.

347