VDE_295.

fm Page 177 Monday, July 29, 2002 6:14 PM

Veterinary Dermatology 2002, 13, 177 186

Plasma amino acid concentrations in 36 dogs with histologically

confirmed superficial necrolytic dermatitis

Blackwell Science, Ltd

CATHERINE A. OUTERBRIDGE, STANLEY L. MARKS and QUINTON R. ROGERS

From the Department of Medicine and Epidemiology (Outerbridge and Marks), and Department of Molecular
Biosciences (Rogers), School of Veterinary Medicine, University of California, Davis, CA 95616
(Received 9 September 2001; accepted 20 November 2001)

Abstract Plasma amino acid concentrations were measured in 36 dogs diagnosed with superficial necrolytic
dermatitis (SND) via skin biopsy. The median age of the dogs was 10 years, and 27 out of 36 (75%) were male.
Twenty-two out of 36 (61%) of the dogs were accounted for by six breeds; West Highland white terriers (six),
Shetland sheepdogs (five), cocker spaniels (four), Scottish terriers (three), Lhasa apsos (two) and Border collies
(two). The mean concentration ( standard deviation) was calculated for each measured plasma amino acid and
compared to previously documented concentrations of plasma amino acids measured in dogs with acute and
chronic hepatitis. The ratio of branched chain amino acids to aromatic amino acids in the dogs with SND was
2.6, slightly lower than that in normal dogs. The mean plasma amino acid concentrations for dogs with SND
were significantly lower than for dogs with acute and chronic hepatitis. A metabolic hepatopathy in which there
is increased hepatic catabolism of amino acids is hypothesized to explain the hypoaminoacidaemia seen in SND.
Keywords: amino acids, dog, hypoaminoacidaemia, metabolic epidermal necrolysis, SND, superficial necrolytic
dermatitis.

INTRODUCTION
Superficial necrolytic dermatitis (SND) is an uncommon skin disease that has also been called metabolic
epidermal necrosis, hepatocutaneous syndrome, diabetic dermatopathy and necrolytic migratory erythema
(NME). This disease was first described in 1986, in four
dogs with diabetes mellitus1 and was thus first called
diabetic dermatopathy. The disease is typically diagnosed in older dogs, although there are rare reports of
it occurring in cats and in the black rhinoceros.25 The
most common clinical sign in SND is the development
of visually distinctive skin lesions with a characteristic
distribution.1,611 Skin lesions include erythema, crusting, exudation, ulceration and alopecia involving
footpads, peri-ocular or peri-oral regions, analgenital
regions, and pressure points on the trunk and limbs.
Lameness secondary to footpad lesions, inappetance
and weight loss are also associated with SND. Polydipsia and polyuria may be present when there is concurrent diabetes mellitus. Histopathological findings of a
marked parakeratotic epidermis with striking interand intracellular oedema, and keratinocyte degeneration in the upper epidermis along with hyperplastic
basal cells, are responsible for the characteristic red,
white and blue histological lesion that is diagnostic for
this disease.8,11
Correspondence: Dr Outerbridge, Veterinary Medical Teaching
Hospital, School of Veterinary Medicine, University of California, Davis, CA 95616. Fax: 001 530 752 9620; E-mail:
caouterbridge@vmth.ucdavis.edu
2002 Blackwell Science Ltd

Necrolytic migratory erythema (NME) is a histologically similar disease that is seen in humans. Most often
NME occurs in association with a glucagon-secreting
tumour. Glucagonoma syndrome in humans is characterized by the skin lesions of NME, hyperglycemia
resulting from carbohydrate intolerance or diabetes
mellitus, weight loss, hypoaminoacidemia and anemia.12 In human patients with NME, there is typically
a profound hypoaminoacidemia, presumed to result
from the catabolic gluconeogenic effects of glucagon.13
However, NME has been diagnosed in people with normal plasma amino acid concentrations and these have
often been patients with nonglucagonoma-associated
disease.14
Determination of plasma amino acid concentrations
in dogs with SND has been reported infrequently and
a review of the English language literature revealed
only 11 dogs in which plasma amino acid concentrations had been measured. All 11 dogs had a marked
hypoaminoacidemia,6,15,16 although only two of the
11 dogs were confirmed to have a pancreatic
tumour.15,16 Unlike in humans with NME, a documented association with a glucagonoma has not been
found in the majority of dogs with SND; instead most
dogs have an associated hepatopathy. The etiopathogenesis of the hepatic pathology seen in the majority of
dogs with SND remains unknown and it is unclear
what metabolic pathway(s) may be linking liver or pancreatic disease with the skin lesions seen in SND.
The purpose of this study was to evaluate the plasma
amino acid concentrations in a larger group of dogs
with skin biopsy confirmation of SND, and to compare
177

VDE_295.fm Page 178 Monday, July 29, 2002 6:14 PM

178

C. A. Outerbridge et al.

these values to those previously determined for dogs

with acute and chronic hepatitis.17,18

MATERIALS AND METHODS

Plasma samples submitted from veterinary hospitals
from across North America to the amino acid laboratory
in the Department of Veterinary Molecular Biosciences
at the University of California, Davis, between 1992 and
2000 were reviewed to identify the submissions from
dogs tentatively diagnosed with SND. Amino acid
concentrations were determined by the use of an automated analyser (Model 7300, Beckman Instruments,
Palo Alto, CA). The analyser utilizes cation-exchange
chromatography and spectroscopic determination of a
ninhydrin reaction with amino acids to obtain measured
values from a plasma sample. Norleucine, as an internal
standard, was used to standardize the concentrations
of amino acids across time using different batches of
reagents.
Only amino acid panels from dogs for which a skin
biopsy could be obtained and reviewed to confirm the
diagnosis of SND were included in the study. Re-cuts
of skin biopsy slides were requested from pathology
laboratories across North America. Slides were
stained with hematoxylin and eosin (H&E) and were
reviewed by a single board-certified pathologist and
one of the authors (CAO). The diagnosis of SND was
made if the characteristic pathological changes of
parakeratosis, intracellular edema and keratinoctye
degeneration in the upper epidermis and hyperplasia of
the basal cell layer were evident. Whenever adequate
information regarding the source of the plasma sample
submitted existed, additional clinical information and
follow up was requested for each dog in the study. This
included information regarding body weight, lesion
distribution, concurrent diabetes mellitus or hyperadrenocorticism, any history of recent drug administration (within 3 months of the diagnosis of SND),
results of any liver function tests, hepatic biopsies or
abdominal ultrasound, survival times and information
on any treatments given. Inclusion criteria included the
availability of a skin biopsy to confirm the diagnosis of
SND and the availability of a plasma sample for amino
acid profile determination and evaluation.
The mean standard deviation was calculated for
each measured amino acid concentration and these
values were compared to the mean plasma amino acid
concentrations measured by the same laboratory from
dogs with acute hepatitis induced by oral administration of nitrosamine (n = 12), chronic hepatitis (n = 8)
and healthy dogs (n = 13). These values were based on
previously published studies using the similar cationexchange methods with the same internal standard
that was used by the laboratory for their reference
ranges.17,18 The plasma amino acid concentrations for
dogs with chronic hepatitis were determined from dogs
with naturally occurring disease that had been diagnosed
with chronic active hepatitis based on hepatic biopsies.17,18
2002 Blackwell Science Ltd, Veterinary Dermatology, 13, 177 186

The 95% confidence interval (95% CI) was calculated for the mean plasma concentration of each amino
acid from the four groups of dogs (normal, SND, acute
hepatitis and chronic hepatitis). If the intervals did not
overlap among the groups being evaluated, the difference was deemed significant (P < 0.05). A molar ratio
of branched chain amino acids (BCAA) to aromatic
amino acids (AAA) was calculated for each SND dog
in the study using the following formula:
BCAA:AAA Molar Ratio = (Valine Leucine
Isoleucine)/(Phenylalanine + tyrosine)
The mean standard deviation for the ratio of
BCAA:AAA was then calculated for the dogs with
SND.

RESULTS
Signalment
Thirty-six dogs with histological changes on skin
biopsies compatible with the diagnosis of SND were
evaluated. The diagnosis of SND in the study dogs had
been made over a 9-year period. Twenty-seven dogs
(75%) were male (19 neutered) and nine dogs (25%)
were female (eight spayed). Their ages ranged from 5 to
15 years with a mean and median age of 10 years. Dog
breeds represented included West Highland white
terrier (n = 6), Shetland sheepdog (n = 5), cocker spaniel
(n = 4), Scottish terrier (n = 3), Lhasa apso (n = 2) and
Border collie (n = 2). These six breeds comprised 61%
of the dogs. Other breeds evaluated were mixed-breed
(n = 4), terrier mix (n = 3), and one each of American
Eskimo, Australian shepherd, Cairn terrier, Maltese,
Pembroke Welsh corgi, Pomeranian and Samoyed.
One of the mixed-breed dogs was an Australian
shepherd cross and one other mixed-breed dog was a
poodle cross.
Physical examination findings and diagnostic
evaluations
Additional clinical information was obtained from the
medical records of 30 of the 36 dogs. The mean body
weight was 12.7 kg (range: 3.527.2 kg); 22 of the
30 dogs (73%) had body weights of less than 15 kg. Of
these 30 dogs skin lesions involved the footpads (30 dogs),
peri-oral region (14 dogs), elbows (10 dogs), peri-anal
region (eight dogs), nails or nail beds (six dogs), periocular region (five dogs); perivulvar region (four of the
nine female dogs), prepuce (three of the 21 male dogs),
and scrotum (two of the 21 male dogs). Concurrent
diabetes mellitus was present in eight out of 30 dogs,
with three of these eight dogs developing diabetes
mellitus after the diagnosis of SND. Three dogs had a
history of phenobarbital administration and at least six
dogs had a history of recent corticosteroid administration. Two of the dogs that had received corticosteroids
were among the dogs that developed concurrent diabetes mellitus. Three dogs had hyperadrenocorticism,

VDE_295.fm Page 179 Monday, July 29, 2002 6:14 PM

Amino acid profiles in dogs with SND

two of which were among the eight dogs with concurrent
diabetes mellitus. Two dogs had been diagnosed with
hypothyroidism and were receiving thyroid supplementation at the time of the diagnosis of SND.
Liver enzyme values were available for 16 dogs.
Alkaline phosphatase (ALP) activity was elevated in 15
of these 16 dogs, with a mean of 1080 (reference range:
1282920 IU mL1). The elevations in ALP activity
ranged from 1.2 to 10 times the upper limit of the individual laboratory reference ranges. Dogs with a history
of corticosteroid or phenobarbital administration or
with concurrent diabetes mellitus or hyperadrenocorticism had the highest elevations of ALP. Abdominal
ultrasound was performed in 15 dogs. All of these dogs
were reported to have a mottled, variably echogenic
liver and to have no evidence of a pancreatic mass.
Results of liver histopathology were available for 12 dogs.
In five of these dogs a necropsy had been performed.
In all the dogs with available liver histology, marked
vacuolar degeneration or a vacuolar hepatopathy was
described. Bile acid concentrations were available for
five dogs, and both pre- and postprandial bile acid
concentrations were normal in four of these five dogs.
In one dog an elevated postprandial bile acid concentration of 63 mol L1 (reference range: 016 mol L1),
3.2 times the upper limit of the reference range, was
documented.
Seventeen of the 30 dogs for which additional
clinical information was available were treated after
the diagnosis of SND with some form of nutritional
supplementation. The nutritional supplementation
provided included oral hyperalimentation with protein
or amino acid supplements in all 17 dogs, oral zinc
supplementation in eight of the 17 dogs and oral
supplementation with essential fatty acids in six of the
17 dogs. Intravenous solutions of amino acids were
administered in seven of the 17 dogs. These seven
dogs also received oral protein supplementation. The
remaining 13 dogs were either not treated (four dogs)
or received oral antibiotics (nine dogs) and prednisone
(four dogs) to treat the skin lesions. Many of the dogs
that received nutritional support also received oral
antibiotics (12 dogs) and/or oral prednisone (three
dogs) to treat their skin lesions.
Outcome
The mean survival time for the 30 dogs for which additional clinical information was available was 6.43 months
(range: 2 days to 32 months). Twenty-four of 30 dogs
(80%) in this study were euthanatized with a mean survival time of 3.14 months (range 2 days to 15 months).
Sixteen dogs survived less than 3 months, five dogs
survived between 3 and 6 months, two dogs survived
612 months, and one dog survived greater than
12 months. Six dogs were still alive at the end of the
study with survival times ranging from 6 to 32 months
(mean = 19.6 months). All of the dogs with survival
times greater than 8 months were receiving some form
of nutritional supplementation with increased protein
or amino acids supplements. Half of these dogs (3/6)

179

received intravenous administration of amino acid

solutions.
Amino acids
Plasma amino acid concentrations were significantly
decreased (P < 0.05) in all the dogs with SND compared
with the normal dogs and dogs with acute and
chronic hepatitis. The mean for total plasma amino
acid concentration was 895 nmols mL1 (range: 490
1516 nmols mL1), 30% of the value for total plasma
amino acid concentration reported in normal dogs
without metabolic disease (mean: 3230 nmol mL1,
range: 28014031 nmols mL1). The total plasma amino
acid concentrations for the acute hepatitis and chronic
hepatitis dogs were calculated using the reference values
for the mean for each measured plasma amino acid.
Dogs with acute hepatitis had a mean total plasma
amino acid concentration of 5246 nmol mL1 (range:
36586834 nmols mL1) and dogs with chronic hepatitis
had a mean total plasma amino acid concentration of
3079 nmol mL1 (range: 26203539 nmols mL1). The
values for the dogs with acute or chronic hepatitis were
3 or more times the values found in the dogs with SND.
Glutamic acid, phenylalanine, tryptophan and ornithine
were the only four amino acids whose mean values in
dogs with SND were not significantly different (P <
0.05) from the mean reference values for normal dogs.
The mean values of all of the other amino acids in the
dogs with SND were 60% or less than the mean reference values for the normal dogs (Fig. 1). The mean
values for glutamine, arginine, threonine and proline were
less than 20% of the mean reference values for normal
dogs (Fig. 1). Although there was variation in the values
of each amino acid among the individual dogs with
SND, in 26 of the 36 dogs (72%) either glutamine (19/
36) or arginine (7/36) was the most diminished plasma
amino acid. The mean value for the molar ratio of
BCAA:AAA for the dogs with SND dogs in the study
group was 2.6 (95% CI 2.32, 2.88).
The mean plasma amino acid concentrations for dogs
with SND were lower than the mean concentrations for
dogs with chronic hepatitis except for the amino acid
tryptophan (Fig. 2). The difference between the mean
plasma concentration for each amino acid in these two
groups was statistically significant (P < 0.05) for 15 of
the 22 amino acids originally measured in dogs with
chronic hepatitis (Table 1). Mean plasma concentrations
for asparagine, citrulline, isoleucine, leucine, ornithine,
tryptophan and valine were not significantly different
between the dogs with SND or chronic hepatitis
(Table 1). Three of these amino acids, isoleucine, leucine
and valine, are branched chain amino acids and are not
initially metabolized in the liver. The difference between
the mean plasma concentrations for amino acids in
the dogs with SND and the dogs with acute hepatitis
was statistically significant (P < 0.05) for 12 of the 19
amino acids originally measured in the dogs with acute
hepatitis (Table 1).
The dogs with survival times of greater than
8 months had significantly lower (P < 0.05) mean
2002 Blackwell Science Ltd, Veterinary Dermatology, 13, 177186

VDE_295.fm Page 180 Monday, July 29, 2002 6:14 PM

180

C. A. Outerbridge et al.

Amino Acid Concentrations (% of Normal Dogs)

Figure 1. Plasma amino acid concentrations in dogs with superficial necrolytic dermatitis (SND) as a percentage of the mean for normal dogs.
The broken line represents the mean plasma amino acid concentrations for the normal dogs. The bars represent the maximum concentrations
for each amino acid.

SND

Acute hepatitis
Chronic hepatitis

900

1714

504

717

300
250
200
150
100
50
0
valine

tyrosine

tryptophan

threonine

taurine

serine

proline

phenylalanine

ornithine

methionine

lysine

leucine

isoleucine

histidine

glycine

glutamine

glutamic acid

citrulline

aspartic acid

asparagine

arginine

alanine

Figure 2. Plasma amino acid concentrations in dogs with superficial necrolytic dermatitis (SND), acute hepatitis and chronic hepatitis as a
percentage of the mean for normal dogs. The broken line represents the mean plasma amino acid concentrations for the normal dogs.

plasma concentrations for alanine, glutamine, glycine,

proline and threonine than did the dogs with survival
times of less than 8 months (Table 2). Because alanine
and glutamine are the two amino acids found in the
highest concentrations in plasma, they contribute
2002 Blackwell Science Ltd, Veterinary Dermatology, 13, 177 186

most to the total plasma amino acid concentrations,

resulting in significantly lower (P < 0.05) total plasma
amino acid concentrations in the dogs with survival
times greater than 8 months. Dogs with survival times
< 8 months had a mean total plasma amino acid

VDE_295.fm Page 181 Monday, July 29, 2002 6:14 PM

Amino acid profiles in dogs with SND

181

Table 1. Plasma amino acid concentrations in healthy dogs and dogs with superficial necrolytic dermatitis (SND) or liver disease. Values
expressed as mean in nmols mL1 plasma S.E.M
Amino acid

Normal
n = 13

SND
n = 36

Acute hepatitis
n = 12

Chronic hepatitis
n=8

Alanine
Arginine
Asparagine
Aspartic acid
Citrulline
Glutamic acid
Glutamine
Glycine
Histidine
Isoleucine
Leucine
Lysine
Methionine
Ornithine
Phenylalanine
Proline
Serine
Taurine
Threonine
Tryptophan
Tyrosine
Valine

436 39.4
138 11.6
25.8 3.2
10.7 0.8
39.2 0.6
27.7 4.3
967 53.1
191 15.2
82.5 5.2
79.8 11.4
156 19.0
190 21.6
58.2 5.6
18.6 2.9
59.6 5.9
172 31.3
117 8.2
128 23.9
192 19.2
65.3 6.6
48.4 4.2
212 22.1

110 9.0*
19.5 1.9*
12.7 1.6
4.1 0.5*
10.3 0.9
33.7 5.3*
101.6 9.3*
65.5 3.5*
46.9 2.7*
35.7 3.2
68.9 5.3
42.2 3.8*
13.4 1.4*
17.4 2.7
57.5 2.8*
27.2 2.3*
47.5 3.3*
25.4 2.5
32.0 2.6*
61.9 4.7
19.0 1.4*
95.4 6.1

480 102
250 117
unavailable
10 2.4
unavailable
290 68.3
516 58
540 115
260 126
90 35
130 33.8
450 194
170 114
360 121
330 72.2
60 2.1
180 29.8
210 94
180 77.1
150 73.8
380 117
210 35.8

253 42.6*
87.7 13.2*
25.5 5.0
18.7 3.7*
18.1 3.7
86.5 12.0*
977 53.1*
173 21.8*
101 14.1*
48 4.7
95.8 9.6
197 21.0*
44.1 5.3*
38.8 12.0
110 16.5*
105 13.2*
173 20.9*
128 26*
133 14.7*
53.4 8.7
91.6 15.6*
121 13.0

Means of plasma amino acids in nmol mL1 standard error of the mean. If there was no overlap for the 95% confidence intervals between two
groups, the difference between the means was deemed significant at P < 0.05.
*Significant difference between the means for chronic hepatitis and SND.
Significant difference between the means for acute hepatitis and SND.
Amino acid concentrations in the dogs with acute hepatitis were determined 23 weeks after the oral administration of dimethylnitrosamine.

Table 2. Plasma amino acid concentrations in dogs with superficial

necrolytic dermatitis (SND) with survival times (ST) less than and
greater than 8 months. Values expressed as mean in nmols mL1
plasma SEM
Amino acid

Normal

ST < 8 months

ST > 8 months

Alanine
Arginine
Asparagine
Aspartic acid
Citrulline
Glutamic acid
Glutamine
Glycine
Histidine
Isoleucine
Leucine
Lysine
Methionine
Ornithine
Phenylalanine
Proline
Serine
Taurine
Threonine
Tryptophan
Tyrosine
Valine

436 39.4
138 11.6
25.8 3.2
10.7 0.8
39.2 0.6
27.7 4.3
967 53.1
191 15.2
82.5 5.2
79.8 11.4
156 19.0
190 21.6
58.2 5.6
18.6 2.9
59.6 5.9
172 31.3
117 8.2
128 23.9
192 19.2
65.3 6.6
48.4 4.2
212 22.1

119.7 11.2*
20.1 8
13.1 2.0
4.3 0.6
10.8 1.2
36.7 6.6
114.7 11.2*
69.5 3.7*
48.4 3.3*
38.4 3.8
74.2 5.8
46.8 4.6
14.0 1.7
19.2 3.4
61.8 3
29.2 2.8*
49.4 3.7
27.5 3.3
35.9 2.8*
63.1 5.6
20.2 1.5
101.3 6.7

72.5 7.1*
14.1 2.1
10.1 1.3
3.4 0.4
7.1 1.6
21.2 3.5
64.7 10.1*
46.5 5.4*
37.7 5.3*
24.8 5.3
18.7 10.4
26.4 5.4
11.9 2.6
10.9 3
45.2 5
17.8 2.1*
37.2 3.2
21 3.14
18 1.7*
57.5 8.2
15.9 2.6
74.6 14.6

Means of plasma amino acids in nmol mL1 standard error of the

mean. If there was no overlap for the 95% confidence intervals for the
mean between two groups, the difference between the means was
deemed significant at P < 0.05.
*Significant difference between the means for dogs with survival
times > 8 months and survival times < 8 months.

concentration of 1020 nmol mL1 (95% CI 918, 1122),

while the dogs with survival times > 8 months had a
mean total plasma amino acid concentration of
669 nmol mL1 (95% CI 528, 811).

DISCUSSION
SND is a disease that affects primarily older dogs, and
the mean age of the dogs in this study was 10 years,
which is similar to that in other reports.6,9,10 The age
range was 515 years, which is similar to the 416 years
age range reported in a recent review of all reported
cases in the literature.10 Male dogs were over represented in this study which is similar to other studies.6,7,9
As dogs in this study came from throughout North
America it was not possible to case-match to one specific patient population. However, Shetland sheepdogs,
West Highland white terriers, cocker spaniels and
Scottish terriers may have a predisposition to SND as
they have been reported previously to be affected
breeds.6,7,9,10,1922
The skin lesions in SND include erythema, crusting,
exudation, ulceration and alopecia, and may precede
any other clinical signs. A marked crusting, fissuring
and ulceration of the footpads is clinically very suggestive of SND and was seen in all of the dogs in this study
for which examination findings were available. Cutaneous lesions may also involve the face (often peri-ocular
and peri-oral), analgenital regions, and pressure points
2002 Blackwell Science Ltd, Veterinary Dermatology, 13, 177186

VDE_295.fm Page 182 Monday, July 29, 2002 6:14 PM

182

C. A. Outerbridge et al.

on the trunk and limbs. Lesions were documented in all

of these areas in the study group. Secondary cutaneous
infections with bacteria, yeast or dermatophytes,
particularly involving the feet, are often present in dogs
with SND.811,23 No information was available about
the frequency of secondary infections in the affected
dogs in this study.
The prognosis for dogs with SND is generally poor
and the majority of dogs in this study had survival
times of less than 6 months. However, 20% of the dogs
in this study were maintained for over 12 months. Most
of these dogs were receiving oral hyperalimentation
with protein supplements and intravenous amino acid
infusions.
All of the dogs with SND in this study had marked
hypoaminoacidemia. The pattern of the plasma amino
acid profiles was significantly different from that seen
in dogs with acute or chronic hepatitis (Table 1 and
Fig. 2) or for any other known disease or nutrient
deficiency, including severe deficiency of that amino
acid in the diet. In a comparative sense, several amino
acids (arginine, leucine, lysine, methionine, proline,
threonine and valine) were lower than ever seen in any
other condition except that of animals with a glucagonoma. The profound differences observed in the
amino acid profiles of the dogs with SND versus acute
and chronic hepatitis cannot be explained by differences
in the assays or equipment utilized over the 20-year
period. Both analysers utilized in the study measure
ninhydrin reaction products, and both utilize cationexchange chromatography. In addition, all amino acids
are run with a known (preweighed) internal standard
(e.g. norleucine) to minimize interassay variation. The
insignificant differences in amino acid profiles that one
would expect when repeated profiles are performed in
the same animal over time do not nullify any of the
findings observed in the amino acid profiles of the SND
dogs, in which the mean values of all but four amino
acids were 60% those of the healthy control dogs.
Most plasma amino acid concentrations increase
in dogs with nitrosamine-induced acute liver disease
(Fig. 2). This hyperaminoacidemia is proposed to be
a result of several contributing factors, including compromised hepatic metabolism resulting in a decreased
uptake of amino acids and lysis of necrotic liver tissue
with release of amino acids into plasma.18,24 Hormonal
imbalances among glucagon, insulin and adrenal hormones occur in acute hepatitis, and favour increased
protein breakdown in peripheral tissues.18 This further
increases the plasma amino acid concentrations that
are then unable to be metabolized by the diseased liver.
In chronic liver disease, some of the plasma amino acid
concentrations remain elevated above normal while
others decrease towards or below normal baseline
(Fig. 2). However, none of the mean plasma amino
acid concentrations in dogs with chronic hepatitis
decreased below 60% of the normal reference range
(Fig. 2). The changes in plasma amino acid concentrations seen in dogs with chronic hepatitis result from
compromised hepatic metabolism and hormonal
2002 Blackwell Science Ltd, Veterinary Dermatology, 13, 177 186

imbalances.18,24 Catecholamines may be the most

important hormones involved in producing the plasma
amino acid profile seen in chronic liver disease. Neither
the increase in glucagon nor the decrease in insulin
concentrations documented to occur in chronic hepatitis was associated with the changes in plasma amino
acids concentrations.17,18
The liver plays a critical role in adjusting amino acid
balance. Hepatic transaminases allow the liver to synthesize dispensable amino acids, to metabolize the
amino group from multiple amino acids into urea via
glutamate synthesis, and to catabolize the carbon
skeleton of amino acids into intermediates of the
Krebs cycle that can be used for energy or can enter
into gluconeogenesis. In chronic and acute hepatitis,
compromised hepatic metabolism results in increased
concentrations of many plasma amino acids. However,
this was not seen in the dogs with SND in this study, as
the majority of plasma amino acid concentrations were
less than 60% of normal. The difference between the
mean concentrations for the plasma amino acids of the
dogs with SND versus the dogs with acute and chronic
hepatitis was significant for the majority of plasma
amino acids. These differences suggest that the pathogenesis of the hypoaminoacidemia found in dogs with
SND cannot be explained by compromised hepatic
metabolism.
There was no significant difference in the mean
concentrations of the branched chain amino acids
(isoleucine, leucine and valine) between the dogs with
chronic hepatitis and the SND dogs (Fig. 2). The initial
step (transamination) of the branched chain amino
acids (BCAA) catabolism occurs in a variety of tissues,
especially skeletal muscle, but not in the liver, while the
aromatic amino acids (AAA) are catabolized entirely
in the liver. The ratio of BCAA to AAA has been recognized as an indicator of hepatic insufficiency.24,25
The BCAA:AAA ratio decreases with the severity of
hepatic dysfunction or portal-systemic shunting.24 The
mean BCAA:AAA ratio in the dogs with SND in this
study was 2.6:1.0, which is not indicative of severe hepatic
dysfunction. The normal ratio is approximately 3.0
4.0:1.0, and most dogs with severe hepatic insufficiency
have ratios of less than 1.5:1.0.26
The total amino acid concentrations documented in
the dogs with SND in this study were 30% of the total
amino acid concentrations seen in normal dogs or dogs
with acute or chronic hepatitis. Although there was no
information available in this study about diet or food
intake in individual dogs, neither anorexia nor protein
malnutrition could account for the degree of hypoaminoacidemia seen in this study. In starvation, the
BCAA in the peripheral circulation initially increase as
protein is catabolized and fewer BCAA are incorporated into muscle tissue. Although there is a generalized decrease in amino acid concentrations during
starvation, over time, the decrease in alanine is the
most prominent as it is a required substrate for gluconeogenesis.27 Neither of the above changes was seen in
the dogs in this study. It seems probable that an as yet

VDE_295.fm Page 183 Monday, July 29, 2002 6:14 PM

Amino acid profiles in dogs with SND

unexplained increase in hepatic catabolism of amino
acids might account for the severity of the hypoaminoacidemia documented in the dogs with SND, as
it is doubtful that compromised hepatic metabolism or
protein malnutrition are the cause.
The fact that plasma amino acid concentrations
were lower in the dogs with survival times greater than
8 months is difficult to explain. Perhaps these dogs had
a more gradual decrease in plasma amino acid concentrations over a longer period of time, permitting better
tolerance of the deranged amino acid metabolism and
hypoaminoacidemia. Another possibility is that the
nutritional support provided for the dogs with longer
survival times allowed for an acceptable quality of
life despite an ongoing disturbance in amino acid
metabolism. Another unexplained result is the fact
that glutamic acid, phenylalanine, ornithine and tryptophan concentrations were not significantly decreased
in the study. However, other reports evaluating plasma
amino acid concentrations in dogs with SND have also
cited these four plasma amino acids to be least likely to
decrease.6,15,16 If generalized amino acid catabolism
were occurring glutamic acid might not become as
depleted, as it is produced by the transamination of
other amino acids. Glutamic acid is produced when the
amino group from other amino acids undergoes
transamination with -ketoglutarate.28 The catabolic
pathway for the carbon skeleton of tryptophan is the
most complex of all the pathways of amino acid catabolism.28 A number of endocrine regulatory factors
influence the rate-limiting reactions of amino acid
catabolism. Tryptophan catabolism is not influenced
by glucagon while the catabolism of almost all of the
other amino acids is increased by glucagon.29
The actual etiopathogenesis of the skin lesions
seen in dogs with SND or in humans with NME is not
known. Human patients with NME have had resolution
of their lesions with surgical excision of a glucagonoma
or the use of somatostatin analogues, suggesting that
hyperglucagonemia plays a direct role in the development of the skin lesions.13,14 It has been suggested that
in NME, hyperglucagonemia results in increased epidermal arachidonic acid, and the resultant metabolites
are responsible for the inflammatory changes in the
skin.30 However, evidence exists that glucagon alone
cannot explain the development of NME lesions in all
human patients, as some patients have had resolution
of skin lesions in the face of elevated glucagon concentrations. Patients who received somatostatin had
persistent elevations in glucagon after the infusion,
yet their skin lesions resolved completely. In addition,
approximately 50% of people with nonglucagonomaassociated NME have normal plasma glucagon concentrations.14,31 Nonglucagonoma-associated NME
has been reported in people with celiac disease, chronic
malabsorption, cirrhosis, nonalpha islet cell tumours,
pancreatitis, hepatitis, inflammatory bowel disease and
in heroin abusers.14
Some people with nonresectable pancreatic tumours
have had skin lesions resolve with administration of

183

intravenous amino acids.13,14 This finding, along with

the hypoaminoacidemia seen in the majority of
glucagonoma patients, has led to the hypothesis that
increased gluconeogenesis triggered by hyperglucagonemia results in low plasma amino acid concentrations
and epidermal protein depletion with the resultant
skin lesions of NME.13,14,32 Hypoaminoacidemia was
proven to result from increased hepatic clearance in
one human patient with a glucagonoma.33 Essential
fatty acid and zinc deficiency have both been proposed
to contribute to the etiology of NME, because the
histological appearance of skin lesions in patients
with zinc deficiency or essential fatty acid deficiency
share some similarities to that seen in NME.13,14 Some
patients have had improvement of their skin lesions
after zinc or EFA supplementation but neither has
proven to be helpful for long-term resolution of the
skin lesions in NME.34
In dogs with SND, both surgical excision of a pancreatic tumour or treatment with somatostatin have been
reported to result in resolution of the skin lesions.11,16
However, a glucagon-producing pancreatic tumour or
hyperglucagonaemia has only been documented in a
small minority of dogs with SND.15,16,3538 Glucagon is
not a hormone that is routinely measured in dogs and
there are also biologically active metabolites that may
not be detected by all assays.6 Based on the available
clinical information, none of the dogs in this study
were diagnosed as having a pancreatic tumour, and
glucagon concentrations were normal in the two study
dogs in which it was evaluated.
The majority of dogs with SND have nonglucagonomaassociated disease and often there is a documented
association with characteristic hepatic changes.
Abdominal ultrasound in affected dogs can reveal an
almost pathognomonic honeycomb pattern to the
liver, consisting of variably sized hypoechoic regions
surrounded by hyperechoic borders.39,40 The livers
from the dogs with SND without associated pancreatic
neoplasia appear irregular, with multiple nodules that
have a gross appearance that resembles cirrhosis.6,20
Hepatic histopathology usually reveals a distinctive
vacuolar hepatopathy with parenchymal collapse.
There is some debate as to whether or not the hepatic
lesions seen in SND reflect true cirrhosis. Some studies
report that the hepatic lesions are consistent with
miconodular cirrhosis20,22 but this was not supported
by a study that documented minimal increase in collagen within portal areas histologically.6 These hepatic
changes have not been reported in those dogs with
SND that had associated pancreatic neoplasia.
In addition to the association with hepatic pathology,
there are also reports of dogs with a history of phenobarbital or primidone administration developing
SND,10,41,42 and a recent reported association in four
dogs with gastrointestinal signs and malabsorption.43
Three dogs in this study had a history of phenobarbital
administration. However, there was no resolution of
the cutaneous lesions when the drug was discontinued,
and none of these dogs had survival times greater than
2002 Blackwell Science Ltd, Veterinary Dermatology, 13, 177186

VDE_295.fm Page 184 Monday, July 29, 2002 6:14 PM

184

C. A. Outerbridge et al.

8 months. There were no dogs in the current study

reported to have vomiting, diarrhea or signs compatible
with a malabsorptive disorder. An association between
diabetes mellitus and SND has been reported previously,1,6 and diabetes mellitus was seen in 25% of the
dogs in this study for which clinical information was
available. Hyperglucagonaemia, if it were present, could
explain the risk of development of diabetes mellitus
seen in dogs with SND. The severe vacuolar liver disease
seen in the majority of dogs with SND and the association in some dogs with diabetes mellitus suggests that
an underlying hormonal or metabolic disturbance is
occurring in dogs with SND.
The metabolic pathways that link hepatic and
pancreatic disease with the skin lesions seen in SND
are unknown. The fact that severe hypoaminoacidaemia
is documented to occur in all cases of SND in which
plasma amino acids have been measured makes it likely
that this metabolic derangement is directly contributing to the cutaneous lesions seen in affected dogs. This
study suggests that the plasma amino acid profiles in
dogs with SND cannot be explained by compromised
or insufficient hepatic metabolism. Increased hepatic
catabolism of amino acids is proposed as the causal
explanation of the marked hypoaminoacidaemia seen
in dogs with SND. Intravenous administration of
amino acids will initially bypass the portal circulation,
resulting in the delivery of amino acids to peripheral
tissues before hepatic uptake and catabolism can occur.
The fact that some dogs respond better to therapy with
intravenous amino acid infusions rather than oral
protein hyperalimentation supports the proposed
hypothesis of increased hepatic catabolism of amino
acids in dogs with SND.

ACKNOWLEDGEMENTS
We thank Dr Thelma Lee Gross, California Veterinary
Dermatopathology and IDEXX, Sacramento, CA,
USA, for reviewing the skin biopsies in this study;
Daniel Wong, Department of Molecular Biological
Sciences, School of Veterinary Medicine, University of
California, Davis, CA, USA for his assistance in measuring the amino acid panels; and Dr Philip H. Kass,
Department of Population Health and Reproduction,
School of Veterinary Medicine, University of California,
Davis, CA, USA, for advice regarding statistical evaluations in this study.

REFERENCES
1. Walton, D.K., Center, C.A., Scott, D.W., Collins, K.
Ulcerative dermatosis associated with diabetes mellitus
in the dog; a report of four cases. Journal of the American
Animal Hospital Association 1986; 22: 7988.
2. Patel, A., Whitbread, T.J., McNeil, P.E. A case of metabolic epidermal necrosis in a cat. Veterinary Dermatology
1996; 7: 221 6.
2002 Blackwell Science Ltd, Veterinary Dermatology, 13, 177 186

3. Day, M.J. Review of thymic pathology in 30 cats and

36 dogs. Journal of Small Animal Practice 1997; 38: 393
403.
4. Munson, L., Koehler, J.W., Wilkinson, J.E., Miller, R.E.
Vesicular and ulcerative dermatopathy resembling superficial necrolytic dermatitis in captive black rhinoceroses
(Diceros bicornis). Veterinary Pathology 1998; 35: 3142.
5. Runge-Harms, U., Beardi, L. A case of metabolic epidermal necrosis in a cat. Proceedings of the 15th Annual Congress of the ESVD-ECVD Maastricht, The Netherlands:
European Society of Veterinary Dermatology 1998, 175.
6. Gross, T.L., Song, M.D., Havel, P.J., Ihrke, P.J. Superficial
necrolytic dermatitis (necrolytic migratory erythema) in
dogs. Veterinary Pathology 1993; 30: 7581.
7. Miller, W.H. Jr, Scott, D.W., Buerger, R.G., Shanley, K.J.,
Paradis, M. et al. Necrolytic migratory erythema in dogs:
a hepatocutaneous syndrome. Journal of the American
Animal Hospital Association 1990; 26: 57381.
8. Gross, T.L., Ihrke, P.J., Walder, E.J. Veterinary Dermatopathology: a Macroscopic and Microscopic Evaluation
of Canine and Feline Skin Disease. St. Louis: Mosby Year
Book 1992, 468.
9. Angarano, D.W. Metabolic epidermal necrosis. In:
Griffen, D.E., Kwochka, K.W., Macdonald, J.M., eds.
Current Veterinary Dermatology. St. Louis: Mosby-Year
Book 1993, 3025.
10. Byrne, K.P. Metabolic epidermal necrosishepatocutaneous
syndrome. Veterinary Clinics of North America, Small
Animal Practice 1999; 29: 133755.
11. Scott, D.W., Miller, W.H. Griffen, D.E. Muller and
Kirks Small Animal Dermatology, 6th edn. Philadelphia:
W.B. Saunders 2000, 444448.
12. Edney, J.A., Hofmann, S., Thompson, J.S., Kessinger, A.
Glucagonoma syndrome is an underdiagnosed clinical
entity. American Journal of Surgery 1990; 160: 6259.
13. Marinkovich, M.P., Botella, R., Sangueza, O.P.
Necrolytic migratory erythema without glucagonoma
in patients with liver disease. Journal of the American
Academy of Dermatology 1995; 32: 6049.
14. Mullans, E.A., Cohen, P.R. Iatrogenic necrolytic
migratory erythema: a case report and review of
nonglucagonoma-associated necrolytic migratory erythema. Journal of the American Academy of Dermatology
1998; 38: 86673.
15. Bond, R., McNeil, P.E., Evans, H., Srebernik, N.
Metabolic epidermal necrosis in two dogs with different
underlying diseases. Veterinary Record 1995; 136: 466
71.
16. Torres, S.M.F., Caywood, D.D., OBrien, T.D.,
OLeary, T.P., McKeeever, P.J. Resolution of superficial
necrolytic dermatitis following excision of a glucagonsecreting pancreatic neoplasm in a dog. Journal of the
American Animal Hospital Association 1997; 33: 13319.
17. Strombeck, D.R., Rogers, Q. Plasma amino acid concentrations in dogs with hepatic disease. Journal of the
American Veterinary Medical Association 1978; 173: 93
6.
18. Strombeck, D.R., Harrold, D., Rogers, Q., Wheeldon, E.,
Stern, J., Schaeffer, M. Plasma amino acid, glucagon and
insulin concentrations in dogs with nitrosamine-induced
hepatic disease. American Journal of Veterinary Research
1983; 44: 202836.
19. Turnwald, G.H., Foil, C.S., Wolfsheimer, M.D.,
Williams, M.D., Rougeau, B.L. Failure to document
hyperglucagonemia in a dog with diabetic dermatopathy

VDE_295.fm Page 185 Monday, July 29, 2002 6:14 PM

Amino acid profiles in dogs with SND

20.

21.

22.

23.

24.

25.

26.

27.
28.
29.

30.

31.

resembling necrolytic migratory erythema. Journal of the

American Animal Hospital Association 1989; 25: 3639.
McNeil, P.E. The underlying pathology of the hepatocutaneous syndrome. a report of 18 cases. In: Ihrke, P.J.,
Mason, I.S., White, S.D., eds. Advances in Veterinary
Dermatology, Vol. III. Edinburgh: Pergamon Press 1998,
113 29.
Foster, A.P., Panciera, D.L., Cooley, A.J. Recognizing
canine hepatocutaneous syndrome. Veterinary Medicine
1997; 92: 1050 5.
Hill, P.B., Auxilia, S.T., Munro, E., Genovese, L.,
Silkstone, M.A., Kirby, B. Resolution of skin lesions and
long-term survival in a dog with superficial necrolytic
dermatitis and liver cirrhosis. Journal of Small Animal
Practice 2000; 41: 51923.
Campbell, K.L., Matousek, J.L., Lichtensteiger, C.A.
Cutaneous markers of hepatic and pancreatic disease in
dogs and cats. Veterinary Medicine 2000; 95: 30614.
Rutgers, C., Stradley, R.P., Rogers, W.A. Plasma amino
acid analysis in dogs with experimentally induced hepatocellular and obstructive jaundice. American Journal of
Veterinary Research 1987; 48: 696702.
Gulick, B.A., Liu, I.K.M., Qualls, C.W., Gribble, D.H.,
Rogers, Q.R. Effect of Pyrrolizidine alkaloid-induced
hepatic disease on plasma amino acid patterns in the
horse. American Journal of Veterinary Research 1980; 41:
1894 8.
Guilford, G., Center, S.A., Strombeck, D.R. Williams,
D.A., Meyer, D.J. Strombecks Small Animal Gastroenterology, 3rd edn. Philadelphia: W.B. Saunders Co. 1996,
569 73.
Felig, P. Amino Acid Metabolism in Man. Annual
Review Biochemistry 1975; 44: 93355.
Lehninger, A.L. Principles of Biochemistry. New York:
Worth Publishers 1982, 53155.
Shepartz, B. Regulation of Amino Acid Metabolism in
Mammals. Philadelphia: W.B. Saunders Company 1973,
117 8.
Peterson, L.L., Shaw, J.C., Acott, K.M., Mueggler, P.A.,
Parker, F. Glucagonoma syndrome: in vitro evidence
that glucagon increases epidermal arachidonic acid.
Journal of the American Academy of Dermatology 1984;
11: 468 73.
Sohier, J., Jeanmougin, M., Lombrail, P., Passa, P. Rapid
improvement of skin lesions in glucagonoma with intravenous somatostatin infusion. Lancet 1980; 1: 40.

185

32. Kasper, C.S., McMurry, K. Necrolytic migratory erythema without glucagonoma versus canine superficial
necrolytic dermatitis: is hepatic impairment a clue to
pathogenesis? Journal of the American Academy of Dermatology 1991; 25: 53441.
33. Almdal, T.P., Heindorff, H., Bardram, L., Vilstrup, H.
Increased amino acid clearance and urea synthesis in a
patient with glucagonoma. Gut 1990; 8: 95648.
34. Kasper, C.S. Necrolytic migratory erythema. Unresolved problems in diagnosis and pathogenesis. A case
report and literature review. Cutis 1992; 49: 1208.
35. Cerundolo, R., McEvoy, F., McNeil, P.E., Lloyd, D.H.
Ultrasonographic detection of a pancreatic glucagonsecreting multihormonal islet cell tumour in a Dachshund with metabolic epidermal necrosis. Veterinary
Record 1999; 145: 6626.
36. Torres, S.M.F., Johnson, S., McKeever, P.J., Hardy, R.
Superficial necrolytic dermatitis and a pancreatic endocrine tumour in a dog. Journal of Small Animal Practice
1997; 38: 24650.
37. Miller, W.H., Anderson, W.I., McCann, J. Necrolytic
migratory erythema in a dog with a glucagon-secreting
endocrine tumor. Veterinary Dermatology 1991; 2: 179 82.
38. Gross, T.L., OBrien, T.D., Davies, A.P., Long, R.E.
Glucagon-producing pancreatic endocrine tumors in
two dogs with superficial necrolytic dermatitis. Journal of
the American Veterinary Medical Association 1990; 197:
161922.
39. Jacobson, L.S., Kirberger, R.M., Nesbit, J.W. Hepatic
ultrasonography and pathological findings in dogs with
hepatocutaneous syndrome: new concepts. Journal of
Veterinary Internal Medicine 1995; 9: 399404.
40. Nyland, T.G., Barthez, P.Y., Ortega, T.M., Davis, C.R.
Hepatic ultrasonographic and pathologic findings in
dogs with canine superficial necrolytic dermatitis. Veterinary Radiology and Ultrasound 1996; 37: 2005.
41. Heinricks, P.M. Dermatitis associated with the use of
primidone in a dog. Journal of the American Veterinary
Medical Association 1987; 191: 2378.
42. Bloom, P., Rosser, E.J., Dunstan, R. Anti-convulsant
hepatitis-induced necrolytic migratory erythema. Proceedings of the Second World Congress of Veterinary
Dermatology Montreal, Quebec: 1992, 56.
43. Florant, E., Guillot, J., Degorce-Rubiales, F., Mailot, M.
Four cases of canine metabolic epidermal necrosis.
Veterinary Dermatology 2000; 11(Suppl. 1): 18.

Rsum Les concentrations en acides amins plasmatiques ont t mesures chez 36 chiens souffrant de
dermatite ncrolytique superficielle (DNS), diagnostique par biopsie cutane. Lge moyen des chiens tait de
10 ans, et 27 sur 36 (75%) taient des males. Six races reprsentaient 22/36 chiens 36 (61%): West Highland white
terriers (six), Shetland (cinq), cocker spaniels (quatre), Scottish terriers (trois), Lhasa apsos (deux) et Border
collies (deux). Les concentrations moyennes ( dviation standard) de chaque acide amin ont t mesures et
compares aux valeurs mesures chez des chiens souffrant dhpatite aige ou chronique. Le rapport des acides
amins chaine longue et aromatiques tait 2.6 fois plus faible pour les chiens souffrant de SND par rapport aux
chiens normaux. Les concentrations moyennes dacides amins taient plus faibles chez les chiens SND par
rapport ceux prsentant une hpatite aige ou chronique. Lhypothse dhpatopathie mtabolique associe
une augmentation du catabolisme hpatique des acides amins est propose pour expliquer lhypoaminoacidmie
recontre dans la SND.
Resumen Se midieron las concentraciones plasmticas de aminocidos en 36 perros diagnosticados de dermatitis
necroltica superficial (SND) mediante biopsia cutnea. La edad media de los perros fue de 10 aos, y 27 de 36
(75%) eran machos. Veintids de 36 (61%) de los perros pertenecan a seis razas; West Highland white terriers
(seis), pastores de Shetland (cinco), cocker spaniels (cuatro), Scottish terriers (tres), Lhasa apsos (dos) y Border
collies (dos). La concentracin media ( desviacin estandard) fue calculada para cada aminocido plasmtico
2002 Blackwell Science Ltd, Veterinary Dermatology, 13, 177186

VDE_295.fm Page 186 Monday, July 29, 2002 6:14 PM

186

C. A. Outerbridge et al.
medido y comparada con las concentraciones de aminocidos plasmticos medidos, documentadas previamente
en perros con hepatitis aguda y crnica. El ndice de aminocidos con cadena ramificada sobre aminocidos
aromticos en perros con SND fue 2.6, levemente inferior al de perros normales. La concentracin plasmtica
media de aminocidos para perros con SND fue significativamente inferior a la de perros con hepatitis aguda y
crnica. Para explicar la hipoaminoacidemia observada en la SND se emite la hiptesis de una hepatopata
metablica en la que existe un catabolismo heptico de aminocidos aumentado.
Zusammenfassung Bei 36 Hunden, die mit Hilfe von Hautbiopsien mit superfizieller nekrolytischer Dermatitis
(SND) diagnostiziert wurden, wurden die Aminosurekonzentrationen im Plasma gemessen. Das mediane
Alter der Hunde betrug 10 Jahre, und 27 der 36 Hunde (75%) waren mnnlich. Zweiundzwanzig der 36 Hunde
(61%) gehrten den folgenden sechs Rassen an: West Highland White Terrier (sechs), Shetland Sheepdog (fnf ),
Cocker Spaniel (vier), Scottish Terrier (drei), Lhasa Apso (zwei) und Border Collie (zwei). Fr jede gemessene
Plasma-Aminosure wurde die mittlere Konzentration ( Standardabweichung) berechnet und mit zuvor
dokumentierten Plasma-Aminosurekonzentrationen von Hunden mit akuter und chronischer Hepatitis
verglichen. Das Verhltnis verzweigter zu aromatischer Aminosuren lag bei Hunden mit SND mit einem Wert
von 2.6 etwas niedriger als bei normalen Hunden. Die mittleren Plasma-Aminosurekonzentrationen bei
Hunden mit SND waren bedeutend niedriger als bei Hunden mit akuter oder chronischer Hepatitis. Als Ursache
des beobachteten Mangels an Plasma-Aminosuren bei SND wird eine metabolische Hepatopathie diskutiert,
die mit erhhtem hepatischen Katabolismus von Aminosuren einhergeht.

2002 Blackwell Science Ltd, Veterinary Dermatology, 13, 177 186