Jaaha 1998 34 4

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 75

Speaking Out

Ceruminolytic Agents in these studies, it would be difficult to assume


that a 22-gauge needle would result in a perfo-
Dear Editor: rated tympanic membrane and hence, a situation
We were interested in the recent study by that is clinically similar to natural otitis externa/
Mansfield PD, et al. (The Effects of Four, Com- media.
mercial Ceruminolytic Agents on the Middle Ear. The literature indicates that most drugs/agents
J Am Anim Hosp Assoc 1997;33:479–86.) re- can cause damage if deposited directly into the
garding the use of four ceruminolytic agents in middle ear. 5 One question we have for the au-
the middle ear. While we applaud the attempt to thors is whether otoscopic examinations were
debunk dogma regarding the use of otic prepara- performed throughout the study to evaluate the
tions, and their risk of ototoxicity, we have sev- tympanic membrane. If they were and the tym-
eral concerns regarding the methodology and the panic membrane was indeed still ruptured, this
conclusions. would make the findings more clinically relevant.
Several studies have been performed with pub- It is certainly plausible and not surprising that
lished methodologies that have attempted to the changes reported would occur with these
simulate the clinical usage of otic preparations agents, as there was no drainage of these poten-
when the tympanic membrane is intact or rup- tially irritating substances. We assume that the
tured. Studies of this type provide the most clini- dog would mount an appropriate inflammatory
cal relevance and attempt to answer important response to attempt to confine or remove the
safety questions. 1,2 Unfortunately, the authors’ foreign substance. It is unclear if these agents
study did not follow the labeled guidelines re- were actually toxic to the middle ear structures
garding the recommended use of the products by or if these structures simply were “caught up” in
the various companies 3 nor does it duplicate the the inflammatory response and this resulted in
way these products often are used clinically. auditory and vestibular damage.
When evaluating agents to make recommenda- Our second concern is regarding the decision
tions regarding clinical safety/use, these issues to inject these products directly into the middle
need to be addressed. ear. Three of the four agents are labeled as clean-
Two issues preclude the practitioner from in- ers that need to be flushed well from the ear
corporating the conclusions of this study into canal. 3 Only Adams ™ Pan-Otic is labeled to re-
clinical practice. First, the instillation of main in the ear canal. ClearRx™ and Cerumenex ™
ceruminolytics via injection through the tym- are not recommended for usage when the tym-
panic membrane (i.e., transtympanic injection) panic membrane is ruptured.3 The companies pro-
into the middle ear is not representative of clini- ducing Cerumene ™ and Adams ™ Pan-Otic do not
cal circumstances found in patients with rup- make recommendations regarding the status of
tured tympanic membranes and clinical otitis the tympanic membrane. 3 Consequently, we feel
externa/media. Transtympanic administration that this study does not simulate a relevant clini-
should not result in rupture of the tympanic mem- cal situation. A prudent practitioner would copi-
brane. Results from previous studies demon- ously flush the entire canal with saline or water,
strated that dogs that underwent removal of all especially if it was later discovered that the tym-
visualized tympanic membrane via electrocau- panic membrane was ruptured. 5,6 This study
tery or curettage had substantial regrowth of the would have been more informative if these agents
tympanic membrane in as little as 10 days. 1,2,4 In were studied in dogs with experimentally induced
two of these studies (which were evaluating gen- otitis media (i.e., the tympanic membrane com-
tamicin and chlorhexidine for ototoxicity), my- pletely destroyed) which would facilitate the
ringotomies had to be repeated surgically to flushing or removal of these products from the
continue to allow complete access to the middle middle ear.
ear. Given the extensive tissue removal recorded Lastly, of additional concern was the number

273
of animals sacrificed for this study. It is unclear References
what role the guinea pigs should have in a study 1. Strain GM, Merchant SR, Neer TM, Tedford BL. Ototoxicity assessment
of a gentamicin sulfate otic preparation in dogs. Am J Vet Res
that is evaluating agents that are instilled into the 1995;56(4):532–8.
ear canals of dogs when there is much variability 2. Merchant SR, Neer TM, Tedford BL, et al. Ototoxicity assessment of a
chlorhexidine otic preparation in dogs. Prog Vet Neurol 1993;
with regard to the potential ototoxicity between 4(3):72–5.
species.6 While we recognize the need to evalu- 3. Compendium of veterinary products. 4th ed. Distributed by North
American Compendiums, Inc. Port Huron, MI, 1997–1998:123, 297,
ate agents with live animals, sacrificing 24 dogs 318.
for a study that does not follow the labeled indi- 4. Steiss JE, Bossinger TR, Wright JC, et al. Healing of experimentally
cations of these drugs nor represent the way these perforated tympanic membranes demonstrated by electrodiagnostic
testing and histopathology. J Am Anim Hosp Assoc 1992;28:307–10.
agents are used clinically seems excessive and 5. Walker EM, Fazekas-May MA, Bowen WR. Nephrotoxic and ototoxic
inappropriate. agents. Clinical Lab Med 1990;110:323–54.
6. Merchant SR. Ototoxicity. Vet Clin N Am 1994;24(5):971–9.

Respectfully,
Kimberly J. Boyanowski, DVM
Terese C. DeManuelle, DVM
Peter J. Ihrke, VMD
University of California Davis,
Davis, California

274
Endogenous Lipid (Cholesterol)
Pneumonia Associated with
Bronchogenic Carcinoma in a Cat
An 11-year-old, female domestic longhair was presented for dyspnea, vomiting, and
left forelimb lameness. A mass in the left caudal lung lobe was seen on thoracic
radiographs. The mass was resected during thoracotomy, and histopathology
confirmed a diagnosis of endogenous lipid pneumonia. The cat recovered slowly
from surgery and was euthanized 11 days following discharge because of persistent
respiratory difficulties. Necropsy findings included lipid pneumonia and bronchogenic
carcinoma, with probable tumor metastasis to the small intestine, spleen, kidney,
and left triceps muscle. J Am Anim Hosp Assoc 1998;34:275–80.

Richard M. Jerram, BVSc, MRCVS Introduction


Cheryl L. Guyer, MS, DVM Lipid pneumonia is a term used to describe the presence of lipid
substances in the lungs that produce evidence of disease either clini-
Anna Braniecki, DVM cally or radiologically. 1 Synonyms for this condition include choles-
terol pneumonia, lipoid pneumonia, and paraffinoma.
W. Kay Read, DVM, PhD
Case Report
H. Phil Hobson, DVM,
Diplomate ACVS A 2.3-kg, 11-year-old, spayed female domestic longhair was referred
to the Texas A&M University Small Animal Clinic for evaluation of
possible megaesophagus and left forelimb lameness. The cat also had
been tachypneic for a period of two weeks. Vomiting/regurgitation
C and left forelimb lameness were reported. The referring veterinarian
had performed a barium study that had demonstrated a caudal mega-
esophagus. The cat was thin and nonweight-bearing in the left fore-
limb. Body temperature was 99.7˚ F, heart rate was 180 beats per minute,
and respiratory rate was 80 breaths per minute. Heart sounds were
muffled, and breathing appeared labored.
A complete blood count (CBC) was characterized by an increased
white blood cell (WBC) count (43.9 x10 3/µl; reference range, 6 to
17 x10 3/µl). A WBC differential analysis was 92% segmented neutro-
phils, 3% band neutrophils, 2% lymphocytes, 1% monocytes, and 2%
eosinophils. Serum biochemical abnormalities included mild hyper-
glycemia (blood glucose, 133 mg/dl; reference range, 65 to 131 mg/dl);
hypophosphatemia (serum phosphorus, 3.3 mg/dl; reference range,
3.8 to 7.5 mg/dl); hypernatremia (serum sodium, 157 mmol/L; refer-
ence range, 144 to 155 mmol/L); hyperchloremia (serum chloride,
126 mmol/L; reference range, 113 to 123 mmol/L); and a decreased
alanine aminotransferase (15 IU/L; reference range, 26 to 84 IU/L).
Results of the urinalysis were normal. Serological tests for feline
leukemia virus (FeLV) and feline immunodeficiency virus (FIV) were
negative.
From the Departments of Small Animal Thoracic radiographs were obtained. A well-defined mass was
Medicine and Surgery and Veterinary
identified in the caudodorsal portion of the thoracic cavity to the left
Pathobiology,
College of Veterinary Medicine, of midline [Figure 1]. The heart was displaced from the sternum on
Texas A&M University, the left lateral radiograph [Figure 2]. A fluoroscopic, liquid barium
College Station, Texas 77843-4474. esophogram was performed; esophageal flow stopped at the level of

JOURNAL of the American Animal Hospital Association 275


276 JOURNAL of the American Animal Hospital Association July/August 1998, Vol. 34

Figure 3—Photomicrograph of a biopsy section taken from the


mass in the left caudal lung lobe. Alveoli are filled with acicular
cholesterol crystal clefts (arrow), macrophages containing fat
vacuoles of similar size, and a mixture of other inflammatory cell
types including occasional multinucleate giant cells. Inter-
alveolar septae are thickened by fibrous tissue and inflamma-
tory cells and are lined by hyperplastic and hypertrophic type II
pneumocytes (Hematoxylin and eosin stain, 25X; bar=100 µ).

left forelimb, and pain on deep palpation of the left


shoulder. Further preoperative diagnostic work-up
was declined by the cat’s owner, and treatment with
cefazolin a (22 mg/kg body weight, intravenously [IV]
qid) was instituted.
Anesthesia was induced with propofol b and main-
tained with isoflurane.c An exploratory thoracotomy
was performed at the left ninth intercostal space to
evaluate the mass and the caudal esophagus. A dis-
Figure 1—Ventrodorsal thoracic radiograph of an 11-year-old crete, yellow-to-tan, umbilicated, 3 by 2.5 by 1-cm
domestic longhair presented for possible megaesophagus and mass with thickened margins was evident on the me-
left forelimb lameness. The cat was tachypneic. A mass is
evident (arrow) in the left caudal lung lobe (left of the midline). dial aspect of the left caudal lung lobe. Adjacent lung
lobes contained multifocal, pinpoint tan lesions. A
left caudal lung lobectomy was performed. The cau-
dal esophagus appeared grossly normal. A chest tubed
was placed during closure, and intermittent thoracic
aspiration was performed for 20 hours postopera-
tively. The left caudal lung lobe was submitted for
histopathology in 10% neutral buffered formalin (NBF).
Thick (5- to 6-mm) slices of fixed tissue from the
mass and adjacent, grossly normal lung parenchyma
were submitted for routine histological processing,
sectioned at 6 µ, and stained with hematoxylin and
eosin (H&E) stain. Alveoli within the well-defined,
nonencapsulated mass contained myriads of aggre-
gates of acicular cholesterol crystal clefts that ranged
from 5- to 15-µ in width and 70- to 150-µ in length
Figure 2—Lateral thoracic radiograph of an 11-year-old domes- [Figure 3]. The clefts were surrounded by a large
tic longhair presented for possible megaesophagus and left number of epitheloid and foamy cells (presumably
forelimb lameness. The mass is seen in the dorsal lung fields
(arrow). The heart is displaced from the sternum. macrophages and hyperplastic type II pneumocytes)
that contained many small, fat vacuoles of similar
the mass, then the esophagus distended and contrast size which stained positively with oil red O stain. The
material flowed into the stomach. A neurological ex- contents of these cells stained lightly or not at all
amination showed atrophy of the left shoulder mus- with periodic acid-Schiff (PAS) stain, indicating that
culature, absence of conscious proprioception in the glycoproteins were not a prominent cellular compo-
July/August 1998, Vol. 34 Endogenous Lipid Pneumonia 277

1% band neutrophils, 4% lymphocytes, 5% mono-


cytes, and 2% eosinophils; a decreased blood urea
nitrogen (BUN, 16 mg/dl; reference range, 19 to 33
mg/dl); hypophosphatemia (serum phosphorus, 3.2
mg/dl); and hypochloremia (chloride, 112 mmol/L).
Serum thyroxine (T 4) and triglyceride levels were
within normal limits. Fluid or air in the pleural space
was not detected by thoracic radiography.
The cat improved with O 2 and antibiotic therapy
during the following three days, and at the time of the
owner’s request for discharge, the cat was able to
maintain normal O 2 saturation levels in room air.
Eleven days following discharge, the cat was eutha-
nized by the referring veterinarian due to continued
Figure 4—Photograph of the lungs taken at necropsy. Gross episodes of severe respiratory distress. A necropsy
anatomic alterations of the lungs include a missing left caudal was performed at Texas A&M University.
lung lobe (surgically removed for biopsy) and only two discern- At necropsy, a reduction of the entire body muscle
ible right lung lobes. The right lung lobes are discolored by the
neoplasm, scar tissue, and areas of lipid pneumonia. The 1- to mass was noted. This change was especially promi-
3-mm in diameter, tan nodules (arrow) scattered throughout the nent in the muscles of the left forelimb. Gross alter-
left cranial lung lobe are areas of endogenous lipid pneumonia ations in the pulmonary architecture included the
(bar=1 cm) (R=right; L=left).
absence of the left caudal lung lobe and the presence
nent. In addition to this cell population, the alveoli of only two distinct right lung lobes [Figure 4]. The
contained many histiocytic macrophages and fibroblasts. right cranial lung lobe was collapsed, pale red-brown
Neutrophils, lymphocytes, plasma cells, and multi- throughout, and contained numerous, coalescing, pale
nucleate giant cells were less numerous. Interalveolar tan, soft foci that filled approximately 50% of the
septae were thickened by fibrous connective tissue lobe. The second, caudally located lung lobe mea-
and a slightly increased number of lymphocytes and sured 4 by 3 by 1.5 cm and was firm and pale tan,
macrophages. Dense populations of lymphocytes and with delicate, white, “lacy” mottling throughout,
plasma cells were adjacent to some bronchi that were which was suggestive of fibrous tissue. The pulmo-
lined by hyperplastic, plicated epithelium. Segments nary parenchyma immediately adjacent to the large,
of the epithelium contained an increased number of primary bronchus of the lung lobe was white and
goblet cells. Neovascularization in some areas of the firm. The medial and ventral surfaces of the lobe
lung was accompanied by a moderate number of mac- were coarsely multinodular, with firm, internodular
rophages that contained hemosiderin. No evidence of adhesions. Two, 7 by 4-mm, smooth, firm plaques of
a tumor or a migrating parasite was found in the pulmonary tissue were present ventral to the right
tissue sections examined. pulmonary hilus.
Special stains for bacteria (Gram stain), fungi A single, 5-mm long, 5-mm in diameter, patent,
(Gomori methanamine silver nitrate stain), acid-fast annular stricture of the jejunal wall was observed 17
bacilli (Kinyoun stain), and partially acid-fast, bacte- cm proximal to the ileocecal junction. A strand of
ria-like fungi (Fite stain) were negative. The absence omental tissue was adhered firmly to the serosa of the
of extracellular fat globules and glycoprotein within stricture. Three (3- to 14-mm in diameter), smooth,
the mass and the presence of intracellular lipid within soft, variegated tan and purple nodules (suggestive of
the macrophages indicated the mass was a focus of nodular hyperplasia) raised above the visceral sur-
endogenous lipid pneumonia rather than exogenous face of the spleen. Portions of the affected tissues, as
lipid pneumonia or alveolar proteinosis. None of the well as representative samples of all major organ
pinpoint tan foci seen at surgery in adjacent lung systems, were collected and placed in 10% NBF.
lobes were present for evaluation in the left caudal Slices of fixed tissue were processed, sectioned, and
lung lobe. stained with H&E stain as described previously.
On the third day postoperatively, the cat became Upon microscopic analysis, plaques of tissue at the
severely dyspneic, anorectic, and depressed. Oxygen right lung hilus and the nodules on the ventral aspect
(O 2) saturation levels e were 74%; the cat was placed of the right caudal lung lobe were composed of aggre-
in an O2 -enriched Kirschner cage, f and O 2 saturation gates of neoplastic epithelial tissue that were accom-
levels were stabilized at 92% to 96%. A CBC and panied frequently by areas of cholesterol pneumonia
serum biochemistry were repeated. Significant find- [Figure 5]. The tumor cells formed glandular struc-
ings were an elevated WBC count (23.8 x10 3/µl) with tures that appeared to arise from the interstitium of
a differential analysis of 88% segmented neutrophils, bronchial walls, cross basement membranes, and ex-
278 JOURNAL of the American Animal Hospital Association July/August 1998, Vol. 34

Figure 5—Photomicrograph of tissue taken from the right caudal Figure 6—Photomicrograph of tissue taken from the right caudal
lung lobe. The area of pulmonary parenchyma is affected by lung lobe. A large bronchiole is occluded partially by scar tissue
bronchogenic carcinoma (upper left) and endogenous lipid (i.e., containing acinar clusters of tumor cells. Aggregates of tumor
cholesterol) pneumonia (upper right, arrow) (Hematoxylin and cells and connective tissue thicken the airway wall. Clusters of
eosin stain, 10X; bar=250 µ). tumor cells infiltrate adjacent lung parenchyma (arrow) (Hema-
toxylin and eosin stain, 10X; bar=250 µ).
tend into dense, peribronchial fibrous stroma or adja-
cent lung parenchyma. The tumor cells within these Tumor cells within the veins in the pelvis and
clusters varied in arrangement from single or double medulla of the kidneys slightly invaded the adjacent
layers of cuboidal cells to sheets in which the cells stroma. One of the grossly visible, tan nodules in the
piled up and lost polarity. In occasional areas, the spleen (measuring 3 by 3 by 6 mm) was a metastatic
epithelium was pseudostratified, and cytoplasmic focus of the tumor. At the intestinal stricture, clusters
blebs were seen at the apices of cells. The cells re- of glands lined by neoplastic cells were found in the
acted slightly more strongly with an antibody to high submucosa and within arterioles in the tunica muscu-
molecular weight (MW) cytokeratin than to low MW laris. Similar acini of neoplastic cells were within and
cytokeratin, suggesting bronchial epithelium as a sec- adjacent to a small arteriole within the left triceps
ond possible origin of the bronchogenic carcinoma. muscle. Reduction in the size of myofibers through-
The nuclear-to-cytoplasmic ratio of the neoplastic out this muscle bundle suggested that muscle atrophy
cells varied markedly. Round-to-oval cell nuclei were was due to interference with the nerve supply to the
moderately to deeply basophilic and varied somewhat muscle. A metastatic tumor compressing the radial
in size, with occasional giant nuclei present. The nu- nerve was sought, but none was found in multiple
clei usually had one prominent, irregularly shaped sections of the muscle examined.
nucleolus. Occasionally, cells in the periphery of
tumor nodules exhibited squamous metaplasia. The Discussion
mitotic index was zero to two mitotic figures per Lipid pneumonia can be subdivided depending on the
high-power field, and many bizarre, nonmetaphase, source of the lipid present. Exogenous lipid pneumo-
mitotic figures were seen. A few large tertiary bron- nia occurs following aspiration or inhalation of min-
chi, large arterioles, and small arteries nearly were eral, vegetable, or animal oils. Mineral oils do not
occluded by scar tissue produced by the neoplasm elicit a cough reflex but do produce a foreign body
[Figure 6]. Within some arterioles, metastatic tumor reaction in the pulmonary tissue. 1 Exogenous lipid
cells lined canalizing spaces within the fibrous tissue pneumonia is an uncommon condition in humans. It
thrombi, simulating gland acini. usually is described in elderly humans with a history
In the right cranial lung lobe, large areas of lung of chronic mineral oil use. 1 Diagnosis is made by
parenchyma that were affected by lipid pneumonia sputum examination, lung aspiration, radiography,
were near small arteries that were plugged partially computed tomography, and bronchoalveolar lavage
by tumor cells. A mild, endogenous lipid pneumonia (BAL).1–4 The granulomatous lesions produced may re-
was in scattered areas of the subpleural parenchyma semble a bronchogenic carcinoma on radiographs.1–4
of the left cranial lung lobe. In these areas, a small to Cytologically, macrophages containing lipid vacuoles
moderate number of macrophages with finely vacu- of heterogeneous size and distribution are pathognomic.5
olated cytoplasm filled alveoli. The interstitium be- Exogenous lipid pneumonia has been reported in
tween alveoli was thickened by fibrous tissue and an cattle and cats following forced administration of min-
increased number of macrophages and lymphocytes. eral oils. 6–8 Affected cats may present in varying de-
No evidence of a tumor was found in this lobe. grees of respiratory compromise, although clinical
July/August 1998, Vol. 34 Endogenous Lipid Pneumonia 279

signs usually are mild.9 Animals may be febrile, an- human workers who were exposed to oily indus-
orectic, and depressed. Diagnosis is based on histori- trial lubricants. 15 It is unknown whether lipid is
cal evidence of mineral oil administration and capable of causing tumor development in pulmo-
radiographs. Lesions seen radiographically may be nary tissue.
misidentified as neoplastic. 9 Histologically, exog- Immunohistochemistry of the tumor cells and the
enous lipid pneumonia is differentiated from endog- glandular morphology and concentration mass at the
enous lipid pneumonia by the presence of large, hilus of the right lung indicated it was a carcinoma
discrete, extracellular globules of lipid in the former that originated from bronchial glands or bronchial
condition. 7 epithelium, and suggested it arose at the hilus and
Endogenous lipid pneumonia occurs when pulmo- extended peripherally within the right lung. The den-
nary cell walls break down, causing release of choles- sity of normal lung tissue at the hilus can make radio-
terol and other lipids into the alveoli. 10 The cause of graphic detection of an early tumor at that site difficult
this cell wall breakdown is thought to be due to more or impossible. The tumor most likely was present at
proximal airway obstruction, the inhalation of irritat- the time of the left lung biopsy, since no neoplastic
ing dust particles, or a disturbance of lipid metabo- mass was found in the left caudal lung lobe; but
lism.1,7,10,11 In humans, endogenous lipid pneumonia retrospective immunohistochemistry demonstrated a
is common and usually occurs distal to an obstruction rare number of tumor cells intermingled with fat-
(i.e., neoplasia, chronic abscess, bronchiectasis). 1 The laden macrophages within alveoli affected by endog-
clinical signs seen are generally the result of the pri- enous lipid pneumonia. The cells presumably metastasized
mary obstructive process. 1 On histological examina- aerogenously to the left caudal lung lobe from the
tion, the affected area contains alveoli filled with mass at the right lung hilus. Bronchiolitis obliterans,
macrophages with foamy cytoplasm. 1,12 caused by scar tissue produced by the tumor, affected
Endogenous lipid pneumonia has been reported in one large tertiary bronchus near the hilus of the right
mice, rats, opossums, raccoons, and mongooses.10,11,13 lung in the tissue sections examined. The endogenous
The condition is reported occasionally in cats and lipid pneumonia in the peripheral right lung appar-
rarely in dogs.7,14 A case of lipid pneumonia has been ently developed secondarily to such bronchial ob-
reported in a dog; however, in that case an exogenous structions. By a similar mechanism, a microscopically
source of the lipid could not be ruled out. 14 The dis- undetected bronchus or bronchiole obstructed by scar
ease appears to occur secondary to bronchial obstruc- tissue or tumor cells presumably led to the develop-
tion, bronchial irritation, long-term inhalant exposure ment of the focus of lipid pneumonia in the left cau-
to dust particles, B vitamin-deficient diets, and hypo- dal lung lobe. In human cases of bronchogenic
physectomy. 10 Parasite infestation may be related to carcinoma, pulmonary architecture can be altered to
endogenous lipid pneumonia in opossums. 10 The sus- the point that lung lobes collapse. Extensive involve-
pected pathogenesis of this form of pneumonia is ment of the right lung field by the tumor in this cat
related to pulmonary injury which causes prolifera- may have caused the loss of distinct middle, caudal,
tion of alveolar type II cells. This results in overpro- and accessory lung lobes.
duction of cholesterol-containing surfactant that The cat displayed clinical signs of lower respira-
enters the alveoli and is phagocytosed by macro- tory tract disease. Failure of clinical signs to resolve
phages.10,14 Grossly, the lungs contain subpleural, yel- following lung lobectomy can be attributed to sud-
lowish-white, firm foci, and the pleura overlying the den, rapid growth and spread of the tumor at the
lesion often is thickened. 7 Histologically, the alveoli hilus, accompanied by the development of endog-
are distended and filled with foamy macrophages. 1,7,10 enous lipid pneumonia. The vomiting could be attrib-
Intracellular and extracellular cholesterol crystals, in- uted to the presence of the metastatic lesions in the
terstitial fibrosis, and hyperplasia of alveolar type II jejunum. The cause of the left forelimb lameness and
cells can occur in more severe cases. 1,7 muscle atrophy was suspected to be related to me-
In a small subset of human cases monitored over a tastasis of the tumor to the extensor muscles, although
period of months or years, sudden changes in the radio- a mass that directly impinged upon the radial nerve
graphic appearance and extent of the lesions of indolent roots was not found. The cause of the functional
exogenous lipid pneumonia heralded the appearance of megaesophagus was undetermined.
tumors.15 These cases provided strong evidence in sup- Both endogenous and exogenous lipid pneumonia
port of the development of neoplasms in lung tissue associated with bronchogenic carcinoma have been
altered by lipid pneumonia. This finding has led to the reported in humans. To the authors’ knowledge, this
suspicion that lipid itself, or scar tissue developing in is the first report of endogenous lipid pneumonia as-
response to it, may provide a fertile environment for sociated with bronchogenic carcinoma in a cat.
development of scirrhous tumors. Mineral oil has
been documented as a cause of scrotal cancer in (Continued on next page)
280 JOURNAL of the American Animal Hospital Association July/August 1998, Vol. 34

a
Sterile cefazolin sodium USP; Solopak Laboratories, Inc., Elk Grove 7. Dungworth DL. The respiratory system. In: Jubb KVF, Kennedy PC,
Village, IL Palmer NP, eds. Pathology of domestic animals. 3rd ed. Orlando:
b Academic Press, 1985:470–4.
Rapinovet; Mallinkrodt Veterinary, Inc., Mundelein, IL
c 8. Chalifoux A, Morin M, Lemieux R. Lipid pneumonia and severe pulmo-
AErrane; Ohmeda PPD, Inc., Liberty Corner, NJ
d nary emphysema in a Persian cat. Feline Pract 1987;17:6–10.
Trocar catheter; Deknatel, Inc., Fall River, MA
e 9. Hawkins EC. Disorders of the pulmonary parenchyma. In: Nelson RW,
Symphony N-3000; Nellcor, Inc., Pleasanton, CA Couto CG, eds. Essentials of small animal internal medicine. St Louis:
f Mosby-Year Book, 1992:221–2.
Kirschner cage; Kirschner, Inc., Timonium, MD
10. Brown CC. Endogenous lipid pneumonia in opossums from Louisiana.
J Wildlife Dis 1988;24:214–9.

References 11. Hayashi T, Stemmermann GN. Lipid pneumonia in the Hawaiian feral
mongoose. J Path 1972;108:205–10.
1. Wright BA, Jeffrey PH. Lipoid pneumonia. Sem Resp Infect 1990;5: 12. Fisher M, Roggli V, Merten D, Mulvihill D, Spock A. Coexisting
314–21. endogenous lipoid pneumonia, cholesterol granulomas, and pulmonary
2. Lee KS, Muller NL, Hale V, Newell Jr JD, Lynch DA, Im J. Lipoid alveolar proteinosis in a pediatric population: a clinical, radiographic,
pneumonia: CT findings. J Comput Assist Tomogr 1995;19:48–51. and pathologic correlation. Pediatr Path 1992;12:365–83.
3. Brechot JM, Buy JN, Laaban JP, Rochemaure J. Computed tomography 13. Hamir AN, Hanlon CA, Rupprecht CE. Endogenous lipid pneumonia
and magnetic resonance findings in lipoid pneumonia. Thorax (multifocal alveolar histiocytosis) in racoons (Procyon lotor). J Vet
1991;46:738–9. Diagn Invest 1996;8:267–9.
4. Silverman JF, Turner RC, West RL, Dillard TA. Bronchoalveolar lavage 14. Corcoran BM, Martin M, Darke PGG, et al. Lipoid pneumonia in a rough
in the diagnosis of lipoid pneumonia. Diagn Cytopathol 1989;5:3–8. collie dog. J Sm Anim Pract 1992;33:544–8.
5. Kurlandsky LE, Vaandrager V, Davy CL, Stockinger FS. Lipoid pneu- 15. Felson B, Ralaisomay G. Carcinoma of the lung complicating lipoid
monia in association with gastroesophageal reflux. Pediatr Pulmon pneumonia. Am J Roent 1983;141:901–7.
1992;13:184–8.
6. Smith BL, Alley MR, McPherson WB. Lipid pneumonia in a cow.
New Zealand Vet J 1968;16:65–7.
Cutaneous Mast Cell Tumors in Cats:
32 Cases (1991–1994)
Case records of 32 cats with cutaneous mast cell tumors (CMCTs) were reviewed.
Using the Patnaik system for grading canine mast cell tumors, the relationships
between histopathological grade and patient survival time and tumor recurrence
were examined. Tumor histopathological grade had no prognostic significance.
One-, two-, and three-year tumor recurrence rates following surgical excision were
16%, 19%, and 13%, respectively. Incomplete excision was not associated with a
higher rate of tumor recurrence. J Am Anim Hosp Assoc 1998;34:281–4.

Heather Molander-McCrary, DVM Introduction


Carolyn J. Henry, DVM, MS The reported incidence of mast cell tumors varies from 2% to 15% of
all tumors in cats. 1–3 Feline mast cell tumors are observed most
Kathleen Potter, DVM, PhD commonly on the head, neck, and trunk, but they also may occur in
the visceral organs. 4 Siamese cats are predisposed to the development
Jeff W. Tyler, DVM, PhD of mast cell tumors. 1,4 Canine mast cell tumors are graded histopatho-
Michael S. Buss, DVM, MS logically to permit prediction of patient survival time and prognosis.
In dogs, the most frequently used histopathological grading system
for mast cell tumors was proposed by Patnaik in 1984. 5 In this grad-
ing system, grade I tumors are well differentiated and confined to the
RS dermis; grade II tumors are moderate to highly cellular with neoplas-
tic cells infiltrating the lower dermal tissue with indistinct cell bor-
ders; and grade III tumors are cellular and pleomorphic with
undifferentiated boundaries and mitotic figures. Using this grading
system, Patnaik, et al. observed an association between mast cell
tumor histopathological grade and survival time in dogs. 5
Although several histopathological grading schemes have been
proposed for use in feline mast cell tumors, none has gained wide-
spread acceptance.2,6,7 Feline mast cell tumors have been classified
generally as well differentiated (i.e., compact), poorly differentiated
(i.e., anaplastic), or histiocytic. 7–9 Well-differentiated tumors are cir-
From the Departments of Veterinary cumscribed, nonencapsulated masses consisting of solid sheets of
Clinical Sciences (Molander-McCrary) uniform, round cells. 9,10 Poorly differentiated (or anaplastic) mast
and Veterinary Microbiology and cell tumors account for 5% to 10% of mast cell tumors in cats and are
Pathology (Potter), characterized by pleomorphism, increased mitotic activity, and infil-
College of Veterinary Medicine,
tration of surrounding tissues. 9,10 Histiocytic mast cell tumors are
Washington State University,
Pullman, Washington 99164-6610 and small, nonencapsulated, deep dermal or subcutaneous masses often
the Department of Veterinary Medicine grouped as papulonodular lesions on the head. 6,9,10 These masses
and Surgery (Henry, Tyler, Buss), comprise 10% to 20% of all feline mast cell tumors and are most
College of Veterinary Medicine, common in young (i.e., less than four years of age) Siamese cats. 6,9,10
University of Missouri,
To the authors’ knowledge, only one prior study with a limited
379 East Campus Drive,
Columbia, Missouri 65211. number (n=14) of cats examined the relationship between histopatho-
logical grade and patient survival time or disease-free interval (DFI).2
Doctor Molander-McCrary’s current This previous study did not examine completeness of surgical exci-
address is Pets Plus Veterinary Center, sion as a risk factor for recurrence or survival. The purpose of this
4750 East Grant,
study was to examine a larger independent population of cats with
Tuscon, Arizona 85712.
cutaneous mast cell tumors (CMCTs) in order to determine whether
Correspondence should be directed histopathological grade and completeness of surgical excision are
to Dr. Henry. useful predictors of survival and prognosis.

JOURNAL of the American Animal Hospital Association 281


282 JOURNAL of the American Animal Hospital Association July/August 1998, Vol. 34

Table
Distribution of 32 Feline Cutaneous Mast Cell Tumors (CMCTs)
Within Histopathological Criteria Categories*

Nuclear Pleomorphism Mitotic Rate Extension Beyond Dermis Eosinophils


None (0) 12 25 Absent = 25 9
Category 1 1 2 Present = 7 8
Category 2 16 3 NA† 14
Category 3 3 2 NA 1

* At least 10 fields (400X) were examined for each tumor. Classification criteria were as follows: Nuclear Pleomor-
phism, none seen (0); Category 1 for up to one cell per 400X field; Category 2 for two-to-five cells per 400X field;
Category 3 for more than five cells per 400X field. Mitotic Rate, none seen (0); Category 1 for up to one cell per
400X field; Category 2 for two-to-three cells per 400X field; Category 3 for more than three cells per 400X field.
Extension Beyond Dermis, none (0); present (1). Eosinophils, none (0); Category 1 for one-to-two cells per 400X
field; Category 2 for three-to-five cells per 400X field; Category 3 for more than five cells per 400X field

NA=not applicable

Histopathological criteria that were evaluated in-


cluded nuclear pleomorphism, mitosis, and invasion
of surrounding tissue [see Table]. Eosinophils also
were evaluated but not included in the grading
scheme. At least 10 fields at 400X magnification were
examined to categorize results. Nuclear pleomorphism
was characterized as none (0); one (1) for up to one
cell per 400X field; two (2) for two-to-five cells per
400X field; or three (3) for more than five cells per
400X field. Mitosis was defined as none (0); one (1)
for up to one mitotic figure per 400X field; two (2)
for two-to-three mitotic figures per 400X field; or
three (3) for more than three mitotic figures per 400X
field. Extension of neoplastic cells outside the dermis
Figure 1—Grade I feline cutaneous mast cell tumor. The tumor
cells are within the superficial dermis. Cells are uniform in size, was defined as absent (0) or present (1). Eosinophils
and nuclei are bland. Multinucleated cells and mitotic figures are were categorized as none (0); one (1) for one-to-two
absent (Hematoxylin and eosin stain, 80X; bar=36 µ). cells per 400X field; two (2) for three-to-five cells
per 400X field; or three (3) for more than five cells
Materials and Methods per 400X field. Histological grades were assigned
based upon the criteria described for canine CMCTs
Archived biopsy samples (from 1991 to 1994) from
by Patnaik, et al.5 Briefly, grade I was assigned to
cats with CMCTs were obtained for review. Case
tumors that were small, superficial infiltrates of well-
inclusion criteria included a histopathological diag-
differentiated mast cells [Figure 1]. Category scores
nosis of CMCT at least one year prior to the begin-
for nuclear pleomorphism and mitoses were both 0 in
ning of the study, and survival and follow-up more
grade I tumors. Grade II was assigned to larger tu-
than one month after the initial diagnosis. The histio-
mors that often extended into the deep dermis [Figure
cytic form of mast cell tumor was excluded, as were
cases with systemic mastocytosis. Practitioners who 2]. Anisokaryosis and multinucleated cells were com-
had submitted biopsy specimens were contacted by mon, but mitoses usually were rare. Grade II assign-
letter and phone to obtain information regarding tu- ment was considered appropriate when nuclear
mor location, size, lymph-node involvement, pres- pleomorphism category scores ranged from 1 to 3,
ence of distant metastasis, treatment type, local with a mitosis category score of 0 to 3. Grade III was
recurrence, date of recurrence, and the cause and date assigned to deep dermal or subcutaneous tumors com-
of death, if applicable. Complete information was posed of anaplastic cells, which often stained poorly
obtained for 32 cases. In each case, the diagnosis of with toluidine blue [Figure 3]. Category score criteria
CMCT was confirmed, and tumors were assigned a for grade III assignment included a nuclear pleomor-
histopathological grade. phism score of at least 3 and a mitosis score of at least 1.
July/August 1998, Vol. 34 Mast Cell Tumors 283

Figure 2—Grade II feline cutaneous mast cell tumor. Tumor Figure 3—Grade III feline cutaneous mast cell tumor. Pleomor-
cells infiltrate the deep dermis. Some nuclear pleomorphism phic neoplastic cells invade deep cutaneous muscle. Occa-
and numerous binucleate or trinucleate cells are present (He- sional mitotic figures were identified in this tumor (Hematoxylin
matoxylin and eosin stain, 80X; bar=36 µ). and eosin stain, 80X; bar=36 µ).

Survival time was defined as the time between sion, n=2) had local tumor recurrences. These recur-
tumor diagnosis and death. Cases that died due to rence rates did not differ significantly. Eleven (34%)
causes unrelated to CMCT or that were lost to follow- of the tumors were grade I, 18 (56%) were grade II,
up were classified as censored. Disease-free interval and three (9%) were grade III. Four cases died due to
was defined as the time from surgical excision to causes unrelated to CMCT, and no cases died due to
tumor recurrence or metastasis. Median survival times the tumor. Median survival times and DFI could not
and DFIs were compared between older (more than be determined because 28 of 32 cases were alive at
10 years of age) and younger (less than 10 years of the end of the study and 27 of 32 had no tumor
age) cats, sex groups (male, castrated male, female, recurrence.
spayed female), breed groups (domestic shorthair, None of the examined historical risk factors, age,
domestic medium hair, domestic longhair, Siamese, sex, breed group, histopathological indices listed in
Himalayan), nuclear pleomorphism grades, mitotic the Table, or tumor grade were associated signifi-
grades, eosinophilic grades, and histopathological cantly with either survival time or DFI. All cases
grades using the Kaplan-Meier method. Group medi- were treated with surgical excision only. The ages at
ans were deemed to differ significantly when p was time of diagnosis ranged from one to 18 years. The
less than 0.05. Proportions of cases with tumor recur- mean and median ages at the time of initial diagnosis
rence, death due to tumor, and death due to any cause were nine and 9.6 years, respectively. The number of
also were reported. The recurrence rates were com- male cases was 12 (all castrated), and the number of
pared between cases with complete and incomplete female cases was 20, 15 of which were spayed. The
excision using the chi-square test. breeds represented were domestic shorthair (66%),
domestic medium hair (3%), domestic longhair (13%),
Results Siamese (16%), and Himalayan (3%). Of cases (n=32)
The case records of 32 cats had adequate information followed for at least one year, five (16%) had tumor
available for review and complete responses to ques- recurrences, none died due to tumor complications,
tionnaires by referring veterinarians. Four (12.5%) and four (13%) died due to unrelated causes. Of cases
cases had more than one lesion at presentation, and (n=27) followed for at least two years, five (19%) had
28 (87.5%) cases had single lesions. Of the four cases tumor recurrences, none died due to tumor complica-
with multiple lesions, two had histopathological con- tions, and three (11%) died due to unrelated causes.
firmation of CMCTs for all masses and two cases had Of cases (n=16) followed for at least three years, two
only one of the masses submitted for histopathogical (13%) had tumor recurrences, none died due to tumor
examination. Local lymph-node or distant metastasis complications, and two (13%) died due to unrelated
was not detected in any case. Histopathological ex- causes.
amination of excised tissues indicated complete exci-
sion in 10 (31%) of 32 cases and incomplete excision Discussion
in 20 (63%) of 32 cases. In two cases, completeness Cutaneous mast cell tumor was not the cause of death
of surgical margins could not be determined because for any case in this study. The cases lived long lives
all excised tissue was not submitted for examination. despite tumor recurrence in 16% of cases. A correla-
Five cases (complete excision, n=3; incomplete exci- tion between histopathological grade and survival
284 JOURNAL of the American Animal Hospital Association July/August 1998, Vol. 34

time could not be established because of the apparent 2. Buerger RG, Scott DW. Cutaneous mast cell neoplasia in cats: 14 cases
(1975–1985). J Am Vet Med Assoc 1987;190:1440–4.
benign nature of this tumor. 3. Pulley LT, Stannard AA. Tumors of the skin and soft tissues. In: Moulton
Historically, investigators described feline CMCT JE, ed. Tumors in domestic animals. 3rd ed. Berkeley: Univ Calif Press,
as highly malignant and likely to metastasize to re- 1990:23–87.
4. Madewell BR, Theilen GH. Mast cell and melanocytic neoplasms. In:
gional lymph nodes and viscera within months. 11 This Theilen GH, Madewell BR, eds. Veterinary cancer medicine. 2nd ed.
belief was challenged by a retrospective study of 14 Philadelphia: Lea & Febiger, 1987:310–5.
cases in which a benign behavior was noted, 2 similar 5. Patnaik AK, Ehler WJ, MacEwen EG. Canine cutaneous mast cell
tumor: morphologic grading and survival time in 83 dogs. Vet Path
to the findings of the current study. Of 160 cats with 1984;21:469–74.
CMCTs in the literature, only four had visceral 6. Wilcock BP, Yager JA, Zink MC. The morphology and behavior of
metastasis. 6,7,12 feline cutaneous mastocytomas. Vet Path 1986;23:320–4.

Buerger and Scott reported tumor grades, survival 7. Holzinger EA. Feline cutaneous mastocytomas. Cornell Vet 1973;63:
87–93.
times, and recurrence rates for 14 feline CMCT cases.2 8. Vail DM. Mast cell tumors. In: Withrow SJ, MacEwen EG, eds. Small
Similar to this study, they reported low mortality animal clinical oncology. 2nd ed. Philadelphia: WB Saunders, 1996:
192–210.
figures and a lack of lymph-node involvement. How-
9. Gross TL, Ihrke PJ, Walder EJ. Histiocytic and mast cell tumors. In:
ever, in their study, a greater percentage (36%) of Gross TL, Ihrke PJ, Walder EJ, eds. Veterinary dermatopathology.
cats had recurrent tumors. Although completeness of St. Louis: Mosby-Year Book, 1992:467–73.
surgical excision was not evaluated as a risk factor in 10. Yager JA, Wilcock BP. Round cell tumors. In: Yager JA, Wilcock BP,
eds. Colour atlas and text of surgical pathology of the dog and cat.
Buerger and Scott’s study,2 none of the recurrences London: Mosby-Year Book Europe Limited, 1994:273–86.
they reported occurred at the site of original excision. 11. Scott EW. Feline dermatology 1900–1978: a monograph. J Am Anim
The data in the authors’ study indicates no significant Hosp Assoc 1980;16:331–459.
12. Garner FM, Lingeman CH. Mast cell neoplasms of the domestic cat.
difference in survival time or recurrence rate for cats Pathologia Veterinaria 1970;7:517–30.
with complete versus incomplete surgical excision. In 13. Turrel JM, Kitchell BE, Miller LM, et al. Prognostic factors for radiation
the previously reported study, multiple lesions oc- treatment of mast cell tumor in 85 dogs. J Am Vet Med Assoc
1988;193:936–40.
curred in 64% of cases2 as compared to 13% in the
14. Miller MA, Nelson SL, Turk JR, et al. Cutaneous neoplasia in 340 cats.
present study, and high grades (II or III) occurred in a Vet Path 1991;28:389–95.
proportion of cases (64%) similar to the present study
(66%). The small number of cases with multiple le-
sions is a limitation of this study. In the two cases
with multiple lesions confirmed to be CMCTs, one
had grade I tumors and the other had grade II tumors.
With this small number of cases, it is not possible to
predict the relationship between histopathological
grade and the presence of multiple tumors.
In dogs, recurrence, advanced stage, and high
grades are all factors that decrease survival times. 5,8,13
In this study, recurrence rates at one, two, and three
years were 16%, 19%, and 13% respectively, but
deaths due to tumor did not occur. The sex predilec-
tion of male cats that has been reported previously 2,12
was not noted in this study. This finding correlates
well with Holzinger 7 and Miller, et al. 14 who also
found no sex predilection.

Conclusion
Histopathological grade does not provide useful prog-
nostic information for solitary feline CMCTs. These
tumors should be considered benign and warrant a
good prognosis, provided visceral or systemic disease
is not identified. Incomplete tumor excision was not
associated with decreased survival time or a higher
rate of tumor recurrence.

References
1. Macy DW, MacEwen EG. Mast cell tumors. In: Withrow SJ, MacEwen
EG, eds. Clinical veterinary oncology. Philadelphia: JB Lippincott,
1989:156–66.
Gastrointestinal Adenocarcinomas
Metastatic to the Testes and
Associated Structures in Three Dogs
Primary testicular neoplasms are common in dogs, but metastases to the testes
are rare. Three dogs had enlargement of the testes and associated structures.
Upon histological examination, the enlargements were due to metastatic
adenocarcinomas. Further examination identified the gastrointestinal tract as the
primary site of the metastatic neoplasms in all three cases. The testicular
metastases reflected widespread metastatic disease. When metastatic
adenocarcinoma is found in the testes and associated structures in dogs, the
gastrointestinal tract should be examined closely for a primary tumor site.
J Am Anim Hosp Assoc 1998;34:287–90.

D. Glen Esplin, DVM, PhD, Introduction


Diplomate ACVP Pathological enlargement of the testes and associated structures can
Sharon R. Wilson, DVM, MS, be caused by inflammation or by primary or secondary testicular
Diplomate ACVP neoplasms. Primary testicular neoplasms are common in dogs, con-
sisting predominantly of seminomas, Sertoli cell tumors, and intersti-
tial (i.e., Leydig) cell tumors. These three neoplasms occur with
C about equal frequency and normally exhibit low rates of metastasis. 1,2
However, metastasis to the spermatic cord and other locations is
reported.1–3 The spermatic cord of a dog with a primary testicular
neoplasm should be examined closely for a metastatic tumor as it
frequently is difficult to determine the malignancy of a testicular
neoplasm by merely examining a section of the tumor.3 Primary
neoplasms of the spermatic cord in domestic animals occur predomi-
nantly in dogs, but they are rare. There are reports of malignant
neoplasms of adipose tissue, fibrous tissue, and striated muscle in the
canine spermatic cord. 3 Secondary neoplasms of the testes in dogs
also are rare. This report documents enlargement of the testes and
associated structures in three dogs, caused by metastasis of adenocar-
cinomas originating in the gastrointestinal tract.

Case Reports

Case No. 1
An 11-year-old, male Yorkshire terrier was presented because of
enlargement of the testes and associated structures. Castration was
recommended. At surgery, the epididymis appeared thickened, so
tissue was collected, fixed in 10% formalin, and submitted for histo-
logical examination.
Upon histological examination, one of the testes and associated
structures contained a neoplastic proliferation of round-to-flattened
From the Animal Reference epithelial cells with many of the cells forming irregular, variably
Pathology Division,
sized glands and tubules. Mucus-containing cells, including some
Associated Regional and University
Pathologists, Inc. (ARUP Laboratories), with the morphology of signet-ring cells, were observed within the
500 Chipeta Way, neoplastic cell population. Variably sized pockets of extracellular
Salt Lake City, Utah 84108. mucus were intermingled with the neoplastic cells. The cells had

JOURNAL of the American Animal Hospital Association 287


288 JOURNAL of the American Animal Hospital Association July/August 1998, Vol. 34

Figure 1—Photomicrograph of biopsy tissue from case no. 1. Figure 2—Photomicrograph of biopsy tissue from case no. 2.
Note the neoplastic proliferation of epithelial cells forming ir- Note the neoplastic glands invading the stroma that is support-
regular, variably sized glands and tubules adjacent to tubules of ing the vessels of the pampiniform plexus (arrows) of the
the epididymis (arrowheads) and the pockets of extracellular spermatic cord (Hematoxylin and eosin stain, 10X; bar=80 µm).
mucus (arrows) (Hematoxylin and eosin stain, 10X; bar=80 µm).
Case No. 2
An eight-year-old, male, shepherd mixed-breed dog
stimulated a prominent desmoplastic response. The presented for anorexia and difficulty walking. Physi-
neoplastic cells had infiltrated extensively the cal examination findings included enlargement of the
connective tissue tunics surrounding the testis, epi- testes and spermatic cord as well as a mass on the
didymis, vas deferens, and spermatic cord. Early in- right side of the rectum. The mass did not extend into
filtration into the stroma supporting the epididymis the perianal area. There was loss of proprioception in
[Figure 1] and the interstitium of the testis was seen. the right rear leg. A biochemical profile and a com-
As an incidental finding, scattered seminiferous tu- plete blood count were performed. A slightly low
bules were filled with a proliferation of round cells of serum calcium (Ca++ , 8.4 mg/dl; reference range, 8.5
spermatogenic epithelial origin, consistent with an to 11.5 mg/dl) and a mild decrease in the hematocrit
early intratubular seminoma. The diagnosis of the (31.7%; reference range, 36.0% to 52.0%) were noted.
epithelial neoplasm was adenocarcinoma. The pa- Castration was recommended.
thologist suspected the neoplasm had metastasized to At surgery, the testicular tunics and spermatic cords
this site and recommended evaluation of the patient of both testes were thickened and were biopsied. In
for evidence of a primary tumor site. addition, a biopsy of the rectal mucosa and the associ-
Ten days later, an exploratory laparotomy was per- ated mass was collected. All tissues were fixed in 10%
formed. The surgeon found a large tumor between the formalin and submitted for histological examination.
stomach and spleen, which was adherent to both or- Upon histological examination, fibrous tunics of
gans. Multiple, thickened areas were seen in the the left testis were infiltrated extensively by a neo-
mesentery. Another large mass was attached to the plastic proliferation of small, round, epithelial cells
ventral abdominal wall in the area of the left inguinal which had formed irregular, variably sized glands
canal. Tissue was collected, placed in 10% formalin, and tubules. Neoplastic glands were in close proxim-
and submitted for histological examination. The dog ity to the vas deferens and had invaded into the stroma
was euthanized. supporting the vessels of the pampiniform plexus
Histologically, the mass between the stomach and [Figure 2]. The fibrous tunics of the right testis and
spleen contained a neoplastic proliferation of epithe- spermatic cord contained a similar neoplastic cell
lial cells (some of which contained mucus) which had population which had infiltrated in a pattern similar
formed irregular, variably sized glands, tubules, and to that seen in the left spermatic cord. The right testis
some signet-ring cells. Pockets of extracellular mu- contained an interstitial (i.e., Leydig) cell tumor sur-
cus were present, and in some areas the pockets had rounded by atrophic seminiferous tubules. The bi-
coalesced, forming lakes. The findings were consis- opsy from the rectum which included both mucosa
tent with a diagnosis of gastric adenocarcinoma. The and submucosa was normal. Histologically, sections
mass had a histological appearance very similar to the of the rectal mass were infiltrated extensively by a
neoplasm previously found in the testis and associ- neoplastic proliferation of epithelial cells which had
ated structures, providing evidence that the testicular formed irregular glands and tubules having an ap-
neoplasm represented metastasis from the stomach- pearance similar to the neoplastic glands in the sper-
associated neoplasm. matic cords and testicular tunics. In one section, the
July/August 1998, Vol. 34 Adenocarcinoma 289

neoplastic glands surrounded bundles of smooth had penetrated the muscular wall completely with
muscle, indicating the tumor involved the smooth extension into mesenteric adipose tissue. Isolated foci
muscle wall of the rectum. The findings were consis- of neoplastic glands also were seen in the mesentery,
tent with a rectal adenocarcinoma metastatic to the supporting metastatic as well as infiltrating spread of
testicular tunics and spermatic cords. the neoplasm to this location. Biopsy tissue from the
In the weeks following surgery, the dog exhibited falciform ligament contained multiple, irregular, vari-
increasing problems with mobility and developed in- ably sized, and poorly defined foci of anaplastic epi-
termittent diarrhea. Within two months, the dog was thelial cells, some of which had formed small glands.
euthanized. A marked desmoplastic response was associated with
these metastatic foci. The diagnosis was mucinous
Case No. 3 intestinal adenocarcinoma with extension into the
A 10-year-old, male cocker spaniel presented after a mesentery and metastasis to the falciform ligament.
groomer reported that the dog was showing signs of The dog was euthanized 3.5 months after the ex-
abdominal pain. Upon physical examination, the right ploratory laparotomy because of anorexia and ab-
testis was larger than the left testis. Castration was dominal pain.
recommended.
At surgery, both testes and associated structures Discussion
were enlarged. A pocket of mucus was associated with In humans, carcinomas metastatic to the testes are
the right testis. Representative tissues from both testes considered rare.4,5 In one series that included over
and associated structures were collected, fixed in 10% 1,600 testicular neoplasms, only 38 cases of proven
formalin, and submitted for histological examination. metastatic carcinomas were diagnosed. The main
Upon histological examination, the tunics adjacent metastatic neoplasms included bronchogenic and pro-
to the epididymis and vas deferens from both sides static carcinomas. 4 Another report indicated that the
contained a neoplastic proliferation of epithelial cells. prostate gland was the most frequent primary site for
Neoplastic cells had formed small glands and also carcinomas metastatic to the testes. 5 Carcinomas
were present as individual cells and small clusters of metastatic to the testes appear to be even more rare in
cells. The neoplastic cells had elicited a prominent domestic animals, 3 but they have been reported in
desmoplastic response. Pockets and lakes of extracel- dogs. 6,7 This rarity may be due to the relatively low
lular mucus were present. In one testis, osseous meta- incidence of prostatic carcinomas in animals, but fail-
plasia was seen in the most prominent area of mucous ure to examine testicular tissue carefully in cases of
accumulation. In the same testis, neoplastic epithelial carcinomas of other organs also is likely a factor.
cells and pockets of extracellular mucus were present Although case nos. 1 and 2 had primary testicular
in the fibrous tunic surrounding the testis. The diag- neoplasms, the glandular carcinomas which had infil-
nosis was adenocarcinoma involving the tunics of the trated the testes and associated structures of both
epididymis, vas deferens, and testes. Metastasis from cases showed no resemblance to primary testicular
the gastrointestinal tract was suspected. neoplasms, indicating metastasis of the carcinomas
The case was referred to a veterinary surgeon who from a distant, nontesticular, primary site. Gland and
performed an exploratory laparotomy. A 20.5-cm (8- tubule formation, signet-ring cells, and lakes of
in) segment of small intestine in the area of the junc- extracellular mucus are features found in gastrointes-
tion of the jejunum and ileum was grossly abnormal tinal adenocarcinomas. 6–9 These features in the
and was resected. The tumor had spread extensively neoplasm from case no. 1 and the tumor’s close asso-
throughout the falciform ligament and involved the ciation with the stomach support a gastric origin for
area of both inguinal canals. Foci of apparent calcifi- this neoplasm. In case no. 2, origin of the pararectal
cation were seen in the pancreas. No gross evidence neoplasm from the glands of the anal sac has to be
of hepatic involvement was seen. The intestinal mass considered. Adenocarcinomas of the glands of the
and tissue samples from the falciform ligament were anal sac are very aggressive and are known to infil-
fixed in 10% formalin and submitted for histological trate into the pelvic cavity.10 No tumor mass was
examination. detected in the perianal area on presurgical physical
Histologically, the intestinal mass contained a pro- examination, but anal sac gland adenocarcinomas of-
liferation of neoplastic epithelial cells (including sig- ten are occult, lying in subcutaneous tissue adjacent
net-ring cells) originating in the intestinal mucosa. to the anus and covered by freely movable, haired
The neoplastic cells had formed irregular, variably skin. 11 However, in one study, over 90% of the dogs
sized glands which had infiltrated extensively into with adenocarcinomas of the glands of the anal sac
the submucosa and muscular wall of the organ. Lakes also had hypercalcemia, and over 90% of the tumors
of extracellular mucus were associated with the infil- occurred in female dogs.11 The glands of the anal sac
trating tumor cells. In areas, the neoplastic glands as the origin of the pararectal neoplasm in case no. 2
290 JOURNAL of the American Animal Hospital Association July/August 1998, Vol. 34

seem unlikely since the dog was a male with a slightly 6. Patnaik AK, Hurvitz AI, Johnson GF. Canine intestinal adenocarcinoma
and carcinoid. Vet Path 1980;17:149–63.
low serum Ca ++. Even though the point of origin of 7. Patnaik AK, Hurvitz AI, Johnson GF. Canine gastric adenocarcinoma.
the neoplasm was not visualized histologically, ori- Vet Path 1978;15:600–7.
gin from the rectal mucosa seems likely, as the glan- 8. Head KW. Tumours of the lower alimentary tract. Bull Wld Hlth Org
1976;53:167–86.
dular tumor was associated closely with the rectum
9. Head KW. Tumors of the alimentary tract. In: Moulton JE, ed. Tumors
and involved the smooth muscle wall of the organ. In in domestic animals. 3rd ed. Berkeley: Univ Calif Press, 1990:347–435.
case no. 3, histological examination confirmed the 10. Pulley LT, Stannard AA. Tumors of the skin and soft tissues. In: Moulton
origin of the adenocarcinoma to be from the mucosa JE, ed. Tumors in domestic animals. 3rd ed. Berkeley: Univ Calif Press,
1990:23–87.
of the small intestine. 11. Meuten DJ, Cooper BJ, Capen CC, et al. Hypercalcemia associated with
When a testicular carcinoma is secondary in an adenocarcinoma derived from the apocrine glands of the anal sac.
humans, it usually is found incidentally at autopsy of Vet Path 1981;18:454–71.

a patient with widespread metastatic disease, or at


orchiectomy for prostatic carcinoma.4 In dogs, intes-
tinal adenocarcinomas and especially gastric adeno-
carcinomas can be widely metastatic.6,7 At laparotomy,
the gastric adenocarcinoma in case no. 1 and the
intestinal adenocarcinoma in case no. 3 showed evi-
dence of widespread abdominal metastasis. The tes-
ticular metastasis appears to be a reflection of the
widespread metastatic disease in these two cases. Nei-
ther a laparotomy nor a postmortem examination was
performed for case no. 2.

Conclusion
The three cases in this report demonstrate the impor-
tance of histologically examining the testes and asso-
ciated structures in all dogs with testicular masses to
determine if a primary testicular neoplasm is present,
to check for evidence of metastasis of a primary tes-
ticular neoplasm, and to rule out metastasis from a
distant, nontesticular, primary neoplasm. Secondary
carcinomas of the testes seem to indicate widespread
metastatic disease. In all three cases of this report and
in two previous reports of metastasis to the testes, 6,7
the primary neoplasms were in the gastrointestinal tracts.
These findings would suggest that when metastatic
adenocarcinoma is found in the testes and associated
structures of a dog, close examination of the gastrointes-
tinal tract for a primary tumor site is indicated.

Acknowledgments
The authors thank Drs. D. S. Halpern, A. Grossman,
S. W. Cho, A. J. Schulman, and E. J. Walder for submit-
ting the cases and providing follow-up information.

References
1. Nielsen SW, Lein DH. Tumours of the testis. Bull Wld Hlth Org
1974;50:71–8.
2. Nielsen SW, Kennedy PC. Tumors of the genital systems. In: Moulton
JE, ed. Tumors in domestic animals. 3rd ed. Berkeley: Univ Calif Press,
1990:479–517.
3. McEntee K. Scrotum, spermatic cord, and testis: proliferative lesions.
Reproductive pathology of domestic animals. San Diego: Academic
Press, 1990:279–306.
4. Price EB Jr, Mostofi FK. Secondary carcinoma of the testis. Cancer
1957;10:592–5.
5. Meares EM Jr, Ho TL. Metastatic carcinomas involving the testis: a
review. J Urol 1973;109:653–5.
Infantile Desmoid-Type Fibromatosis
in an Akita Puppy
A 10-week-old Akita puppy was evaluated for a reported umbilical hernia. Repair of
the hernia had been attempted three times prior to referral. A defect in the ventral
abdominal wall with an associated soft-tissue mass was identified on abdominal
radiographs. Exploratory surgery was performed; the mass was resected and the
abdominal wall defect was repaired. Histopathological evaluation of the mass was
consistent with infantile desmoid-type fibromatosis.
J Am Anim Hosp Assoc 1998;34:291–4.

James L. Cook, DVM Introduction


James R. Turk, DVM, PhD, Congenital, infantile, and juvenile neoplasms occur rarely in dogs.
Diplomate ACVP Neoplasms reported to occur most frequently in young dogs include
histiocytoma, papilloma, lymphoma, osteosarcoma, and fibrosar-
Eric R. Pope, DVM, MS, coma. 1 Infantile desmoid-type fibromatosis (IDTF) has been reported
Diplomate ACVS to occur in children from birth to five years of age. 2–5 This tumor or
tumor-like lesion usually occurs as a solitary mass arising from skel-
Robert C. Jordan, DVM
etal muscle, fascia, aponeuroses, or periosteum.2 Histopathologically,
IDTF varies in its appearance, from primitive mesenchymal forms to
tumors that closely resemble adult desmoids. 2–4 To the authors’ knowl-
C edge, IDTF has not been reported in the veterinary literature.

Case Report
A 10-week-old, male Akita was presented to the University of Mis-
souri-Columbia Veterinary Medical Teaching Hospital (UMC-VMTH)
for evaluation of a reported recurrent umbilical hernia. Prior to ad-
mission, surgical repair of the hernia had been attempted three times
without success. The puppy was presented to the UMC-VMTH for
definitive surgical correction.
Upon physical examination, the puppy was alert, responsive, and
in good body condition. The puppy’s rectal temperature was 38.1˚ C,
pulse rate was 144 beats per minute, and respiratory rate was 30
breaths per minute. The most significant physical examination find-
ing was the presence of a palpable mass within the subcutaneous
tissues of the midventral abdominal wall. The mass extended from the
xiphoid cartilage to midway between the umbilicus and prepuce and
from 2-cm left to 6-cm right of midline. A palpable defect in the linea
alba was present deep to the mass at the area of the umbilicus. A
portion of the mass could be reduced within the abdominal cavity by
digital manipulation. An ulceration approximately 2.5 cm in diameter
was present in the skin immediately lateral to the umbilicus on the
From the Departments of Small Animal left side.
Medicine and Surgery (Cook, Pope, A soft-tissue mass associated with the ventral abdominal wall was
Jordan) and Pathobiology (Turk), identified on abdominal radiographs [Figure 1]. A radiographically
Veterinary Medical Teaching Hospital,
visible defect in the abdominal wall also was noted. A radiolucent,
College of Veterinary Medicine,
University of Missouri, tubular structure consistent with an intestinal loop was located out-
379 East Campus Drive, side of the abdominal cavity. The radiographs were considered to be
Columbia, Missouri 65211. consistent with umbilical herniation and an abdominal wall mass.

JOURNAL of the American Animal Hospital Association 291


292 JOURNAL of the American Animal Hospital Association July/August 1998, Vol. 34

Figure 3—Photomicrograph of a section of the tumor demon-


strating neoplastic cells stained faintly positive for Masson’s
trichrome surrounded by reticulin fibers (Masson’s trichrome
stain, 400X; bar=50 µ).

Figure 1—Lateral radiographic view of the abdomen of an Akita 11th through 13th ribs on the right) was accomplished
puppy demonstrating a soft-tissue mass associated with the
ventral abdominal wall and a defect in the midventral abdominal by sharp dissection. All grossly abnormal tissue was
musculature. removed and submitted for culture and histopatho-
logical evaluation. Omental adhesions to the affected
area were divided sharply. Adhesions of the small
intestine to the previous incision site were divided
sharply. The intestines did not appear to be affected
grossly. During the course of resection, a 1-cm com-
munication with the thoracic cavity was created. The
patient was ventilated appropriately until the defect
was closed and a negative pleural pressure was rees-
tablished. The abdominal musculature was reapposed
with the creation of only minimal tension. Subcutane-
ous tissue and skin closure was routine. The patient
recovered without incident and was discharged the
following day.
The resected tissue varied in morphology when
examined histopathologically. One section was com-
Figure 2—Photomicrograph of a section of the tumor demon-
strating epithelioid and spindle-cell infiltration of adipose tissue posed of haired skin containing a poorly defined,
and skeletal muscle (Hematoxylin and eosin stain, 400X; bar= nodular proliferation of vacuolated, epithelioid-to-
50 µ). spindle cells with generally indistinct nucleoli and no
mitotic figures. Numerous small blood vessels coursed
Packed cell volume, plasma total protein, and se- through the nodules. The cells extended into the sub-
rum urea nitrogen levels were within normal limits. jacent panniculus and skeletal muscle. Another sec-
An abdominal wrap was placed, and the puppy was tion was composed of similar epithelioid-to-spindle
monitored for deterioration in clinical status prior to cells infiltrating adipose tissue and skeletal muscle
surgery. The differential diagnoses considered in- [Figure 2]. Another section demonstrated similar cells
cluded umbilical herniation alone or in combination incarcerating portions of costal cartilage. Masson’s
with granuloma, inflammation, fibrosis, or neoplasia. trichrome, Gomori’s reticulum, and Alcian blue-peri-
Upon surgical exploration of the area, a subcutane- odic acid-Schiff stains at pH of 2.5 and 1 were per-
ous mass was associated with the external rectal formed on sections of the mass. The spindloid
sheath and it extended into the abdominal wall defect. mononuclear cells were faintly Masson’s-trichrome
Herniated tissue at the time of surgery consisted only positive and Alcianophilic at pH of 2.6 and were
of a small amount of omentum. The mass was located surrounded by reticulin fibers [Figure 3]. Histopatho-
predominantly to the right of midline. Surgical resec- logical findings were consistent with a diagnosis of
tion of the mass including overlying skin, all associ- IDTF.
ated muscle, fascia, and cartilage (including a portion The puppy was examined 17 days postoperatively,
of the xiphoid cartilage and the distal 2-to-3 cm of the at which time skin staples were removed. No compli-
July/August 1998, Vol. 34 Fibromatosis 293

cations or evidence of recurrence were reported. No based primarily on the overall diffuse growth pattern
abnormalities were found by complete physical ex- of IDTF and its tendency to contain residual muscle
amination. During telephone conversations 14 months and fat with an alternating pattern of cellular and
after surgery, the dog was reported by the owner and collagenous areas. Infantile neoplasms displaying a
referring veterinarian to have had no evidence of com- high degree of cellularity with a high mitotic index
plications or recurrence related to the surgery. that grow in an expansile, destructive manner are
more indicative of fibrosarcoma. 2–5 Further differen-
Discussion tiation based on electron microscopy or immunohis-
Infantile desmoid-type fibromatosis is a tumor or tu- tochemistry is unnecessary for definitive diagnosis
mor-like lesion in children that has not, as yet, been and has not been described in the human or veterinary
reported in the veterinary literature. This neoplasm in literature to the authors’ knowledge.
children usually presents as a firm, solitary mass that Complete surgical excision is the treatment of
is circumscribed poorly, deeply seated, and rapidly choice.2,4,5 However, resection may be inhibited by
growing. It is encountered most commonly in the first anatomical location and associated structures. Re-
two years of life. The most frequent sites of occur- sponse of IDTF to radiation therapy has not been
rence are the skeletal muscles of the head and neck, reported in the medical literature.
the shoulder and brachium, and the thigh.2 Infantile Infantile desmoid-type fibromatosis lesions appar-
desmoid-type fibromatosis lesions involving the ab- ently do not metastasize. However, incomplete surgi-
dominal musculature have been reported.3,4 In chil- cal resection may result in local recurrence. Neither
dren, males are affected slightly more often than location nor histological appearance appear to be good
females. 2 prognostic indicators of disease course. Although no
Infantile desmoid-type fibromatosis lesions may definitive etiology has been determined, trauma often
progress to incorporate and infiltrate vessels, nerves, is implicated. No hereditary role has been suggested.2,4,5
and muscle. The tumor also may involve joints, bone, The clinical and histopathological findings in this
and any other structures present, resulting in tender- case were consistent with the immature form of IDTF.
ness, pain, and dysfunction. Radiology usually Certainly, the tissue trauma associated with umbilical
demonstrates a soft-tissue mass with or without in- herniation could be implicated as a precipitating cause
volvement of associated structures. 2–5 for this neoplasm. It could not be ascertained as to
Grossly, the neoplasm consists of a firm, poorly whether the original mass palpated by the referring
defined, fibrous mass of gray-white tissue. Encapsu- veterinarian was herniated tissue or an early IDTF
lation does not occur. The tumor usually is adhered to lesion. An idiopathic etiology also is plausible. Sur-
the involved muscle and subcutaneous tissue. 2,4,5 gical excision was performed as sole treatment in this
Histopathologically, several patterns can be distin- case and appears to have been curative. Preoperative
guished that reflect sequential stages of fibroblast biopsy would have confirmed the presence of neopla-
maturation. The immature form is the most common sia but would not have altered the surgical plan. Wide
type and generally is found in infants during the first surgical excision including adjacent normal tissue was
few months of life. This form is characterized by a performed. Since the mass extended through the ab-
haphazard arrangement of small cells associated with dominal wall in some areas, seeding of the abdominal
a dense reticulin fiber meshwork and mucoid mate- cavity contents is a potential sequela. Regular follow-
rial. The appearance of the cells is consistent with a up examinations and careful observation for any signs
stage between primitive mesenchymal cells and fi- of recurrence are recommended.
broblasts. The cells infiltrate muscle diffusely and
can be associated with lymphoid cells. 2 Conclusion
The second form of IDTF is characterized by more The present case describes the first reported case of
mature-appearing, spindle-shaped fibroblasts ar- IDTF in the veterinary literature. It is important for
ranged in distinct bundles. Collagen production is the veterinary clinician to be aware of infantile neo-
seen with this type of tumor. This form of the neo- plasms when diagnosing and treating masses in young
plasm generally is considered to be more aggressive patients.
and may be difficult to distinguish from infantile
fibrosarcoma. The cellularity of this tumor may vary,
and osseous metaplasia has been reported. 2,5 References
1. Dorn CR, Priester WA. Epidemiology. In: Theilen GH, Madewell BR,
The differential diagnoses for IDTF in children eds. Veterinary cancer medicine. 2nd ed. Philadelphia: Lea & Febiger,
include myxoid liposarcoma, botryoid rhabdomyo- 1987:39–42.
sarcoma, lipoblastomatosis, fibrodysplasia ossificans 2. Enzinger FM, Weiss SW. Soft tissue tumors. St. Louis: C.V. Mosby,
1988:164–200.
progressiva, fibrous hamartoma, and fibrosarcoma.
Distinction of IDTF from other infantile neoplasms is (Continued on next page)
References (cont’d)
3. Allen PW. The fibromatoses: a clinicopatho-
logic classification based on 140 cases. Am J
Surg Path 1977;1:305–21.
4. Booher RJ, Pack GT. Desmomas of the abdomi-
nal wall in children. Cancer 1951;4:1052–65.
5. Fromowitz FB, Hurst LC, Nathan J, et al. Infan-
tile (desmoid type) fibromatosis with extensive
ossification. Am J Surg Path 1987;11:
66–75.
Malignant Mesenchymoma Associated
with an Unusual Vasoinvasive Metastasis
in a Dog
A case of a malignant mesenchymoma with an unusual, vasoinvasive, metastatic
behavior in a three-year-old, intact female basset hound is presented. Malignant
mesenchymomas are rare neoplasms in humans and in dogs. No previous reports
of a malignant mesenchymoma with vasoinvasive metastasis in the dog were found
in the literature. The constituent neoplasms are discussed in relation to reports in
the human and veterinary literature, and a potential etiology for this unique
presentation is hypothesized. J Am Anim Hosp Assoc 1998;34:295–9.

Timothy M. Robinson, DVM Introduction


Richard R. Dubielzig, DVM, PhD Mesenchymal neoplasms are those tumors arising from connective
tissue, skeletal tissue, blood vessels, and lymphatics. 1 A malignant
Jonathan F. McAnulty, DVM, PhD mesenchymoma is a neoplasm of mesenchymal origin composed of
tumor cells differentiating into two or more unrelated malignant
forms.2 Mesenchymomas are extremely rare neoplasms in dogs.3 In
C addition, intravascular neoplasms generally are rare in dogs. To the
authors’ knowledge, this is the first reported case of a malignant
mesenchymoma of liposarcomatous and osteosarcomatous origin with
vasoinvasive metastasis in the dog.

Case Report
A 19.4-kg, three-year-old, intact female basset hound was referred to
the University of Wisconsin Veterinary Medical Teaching Hospital
(UWVMTH) with a diagnosis of an abdominal mass. One week prior
to presentation, the dog had become lethargic, inappetent, and had
vomited bile occasionally. The referring veterinarian performed an
exploratory celiotomy and found a large, cyst-like, intra-abdominal
mass solidly adhered to the dorsal aspect of the abdomen. The dog
was recovered from anesthesia without further dissection of the mass,
and she was referred.
On physical examination, the dog was alert, responsive, and mildly
febrile (103.6˚ F). Clinical laboratory parameters were within normal
limits except for a mature neutrophilic leukocytosis. No evidence of
metastasis was seen on thoracic radiographs. A large, circumscribed,
soft-tissue opacity displacing intestinal loops ventrally was seen on
abdominal radiographs. A 15 by 10 by 8-cm, lamelliform mass of
mixed echogenicity, which was confluent with the caudal right kid-
ney, was identified by abdominal ultrasonography. The mass ex-
tended caudally to the area of the bladder and medially ventral to the
aorta to the caudal pole of the left kidney. No evidence of splenic or
hepatic metastasis was noted. Ventral and abaxial displacement of the
From the Departments of Surgical ureters with no evidence of obstruction was discovered by an excre-
Sciences (Robinson, McAnulty) and
tory urogram. Differential diagnoses included intra-abdominal or ret-
Pathobiological Sciences (Dubielzig),
School of Veterinary Medicine, roperitoneal abscess; intravascular thrombosis; neoplasia of renal,
University of Wisconsin, ovarian, splenic, adrenal, lymphatic, or intravascular origin; teratoma;
Madison, Wisconsin 53706. nephroblastoma; and fungal or parasitic disease.

JOURNAL of the American Animal Hospital Association 295


296 JOURNAL of the American Animal Hospital Association July/August 1998, Vol. 34

Figure 1—Intraoperative photograph of the caudal abdominal Figure 2—Photograph of the mass after surgical removal from
cavity. Note the confluence of the common iliac veins into the the abdomen. Note the lamellated appearance of the interior
caudal vena cava (small arrow) which entered the caudal aspect following incision (arrow).
of the mass (large arrows).
reaching the cranial central circulation via the verte-
An exploratory laparotomy was performed, and a bral veins. After recovery, the dog was discharged.
large, cyst-like mass with a smooth wall was found Amoxicillin-clavulanic acid (22 mg/kg body weight,
extending from the left kidney to the right kidney and per os [PO] bid for seven days) was to be given.
caudally to the bifurcation of the vena cava. The Nests of neoplastic cells (some of which were vacu-
kidneys were of normal size and shape. The caudal olated, resembling fat cells) were seen upon histo-
vena cava entered the caudal aspect of the mass [Fig- logical examination of the mass. In other regions,
ure 1]. Cranial to the mass, the vena cava was very these cells formed loosely arranged vascular chan-
small. Multiple, dilated veins terminated in the wall nels. Due to the preponderance of these vascular chan-
of the mass. The mass was removed by meticulous nels, the most probable histological diagnosis was
blunt and sharp dissection. After the mass was hemangiosarcoma. The owner was given a poor prog-
resected, it was determined that approximately 10 cm nosis for the dog to be cured; however, a trial of
of the caudal vena cava was incorporated into the chemotherapy was elected.
mass. The dog recovered uneventfully from surgery. One month later, the owner returned the dog for
On cut section, the mass was granular in appear- oncologic staging of the neoplasia. Laboratory test-
ance, orange-to-brown, and exhibited a lamelliform, ing and thoracic radiographs were within normal lim-
multilayered structure on longitudinal section [Figure its. Sublumbar lymphadenopathy of possible reactive
2]. Anaerobic, aerobic, and fungal cultures were sub- or metastatic origin was seen on abdominal ultra-
mitted. A nonhemolytic Staphylococcus was isolated sonography. An intravenous adriamycin (30 mg/m 2)
on the aerobic culture and it was sensitive to penicil- and cyclophosphamide (150 mg/m 2) chemotherapy
lins and probably represented contamination from an- protocol, to be given every three weeks, was begun.
other source. The primary differential diagnoses at The dog was to be treated four times with this combi-
this time remained intravascular thrombosis, abscess nation and then be monitored for evidence of disease.
of fungal or bacterial origin, vasculitis, or neoplasia One week after the first treatment, the dog became
of vascular origin. pollakiuric, and the owner noticed an orange tint to
Since the caudal vena cava was removed from just the urine. At this time, the referring veterinarian
proximal to its caudal bifurcation to the porta hepatis, placed the dog on trimethoprim-sulfadiazine (15 mg/
a femoral venogram was performed using digital sub- kg body weight, PO bid). The dog re-presented to the
traction imaging to determine the route of venous UWVMTH for evaluation after 10 days with no im-
drainage from the caudal veins to the central circula- provement. A diagnosis of sterile hemorrhagic cysti-
tion. The contrast media flowed proximally from the tis was established following urinalysis and urine
femoral vein into the external iliac vein and common culture. Cyclophosphamide therapy was discontinued.
iliac vein where the contrast media could be seen in Treatment with oxybutinin (0.3 mg/kg body weight,
multiple, arborizing venous branches extending cra- PO tid), an anticholinergic agent with direct antispas-
nially and caudally from the external iliac and com- modic activity on smooth muscle and analgesic and
mon iliac veins. The greatest amount of venous flow local anesthetic effects, was instituted. The dog was
appeared to be through the caudal superficial epigas- discharged with instructions for the owner to monitor
tric veins and the median sacral veins, presumably urination and return the dog in three weeks for tho-
July/August 1998, Vol. 34 Mesenchymoma 297

Figure 3—Photomicrograph of a representative section of the Figure 5—Low-power photomicrograph showing neoplastic in-
abdominal mass. Note the neoplastic mesenchymal cells with filtration of the wall of a large vessel. Necrotic center (N) fills the
prominent cytoplasmic vacuolization, typical of liposarcoma vascular lumen (Hematoxylin and eosin stain, 25X; bar=160
(Hematoxylin and eosin stain, 250X; bar=20 µm). µm). High-power inset of anaplastic neoplastic cells within the
wall of this vessel (Hematoxylin and eosin stain, 400X; bar=10
µm).

the caudal vena cava. In addition, upon dissection of


the vertebral column, the body of the L 3 vertebra was
soft, friable, and partially disintegrated. The verte-
bral body was discolored, dark red, and granular.
Dark red tissue was adherent to and adjacent to the
spinal cord in the area of the L 3 vertebra, and spinal
cord compression was evident. On histological ex-
amination of the L 3 vertebral body and spinal cord,
destruction of the bony mantle and replacement by an
anaplastic, mesenchymal neoplasm was seen. The tu-
mor was composed of plump, anaplastic cells (typi-
Figure 4—Photomicrograph of the vertebral body mass showing cally with large, irregular nuclei, abundant vacuolated
malignant cells surrounded by osteoid matrix (arrows), charac- cytoplasm, and indistinct cell boundaries). Most of
teristic of osteosarcoma (Hematoxylin and eosin stain, 250X; the cells contained variably sized, clear, solitary-to-
bar=20 µm).
multiple, empty vacuolar spaces suggestive of lipid
racic and abdominal radiographs prior to the next [Figure 3]. Neoplastic infiltration extended extra-
dose of adriamycin. durally, sometimes wrapping around nerve roots. The
Five days later, the dog became urinary and fecal tumor was composed of varying sheets of poorly dif-
incontinent and was ataxic in the hind limbs. The ferentiated mesenchymal cells and abundant areas of
owner returned to the referring veterinarian, at which necrosis and fibroplasia. Tumor cells tended to be
time the dog was ataxic, paraparetic, and had patellar plump and irregular, with nuclei similar to those
hyperreflexia and conscious proprioceptive deficits within the vertebral body; however, localized areas of
in the hind limbs. Possible causes for these neuro- intracytoplasmic lipid deposition were more difficult
logical signs included adverse drug reaction, third to find. Several foci exhibited an extracellular matrix
thoracic (T 3) to third lumbar (L 3) intervertebral disk suggestive of osteoid [Figure 4]. Within the paren-
protrusion/extrusion, or other T3 -L3 lesions including chyma of the spinal cord, evidence of compression
metastatic disease. The dog remained hospitalized for with axonal swelling, axonal necrosis, and Gitter cell
six days and was treated with prednisone (0.5 mg/kg accumulation was obvious. In addition, anaplastic
body weight, PO bid), during which time her neuro- neoplastic cells were found in the wall of the mesen-
logical status worsened. At this time, the owner teric veins and a thymic vein upon examination of
elected to euthanize the dog. tissue sections taken from the mesentery between the
On gross postmortem examination, multiple vari- adrenal glands and from the thymus [Figure 5]. The
cosities were noted throughout the abdominal cavity final diagnosis was malignant mesenchymoma with
due to venous shunting secondary to interruption of an unusual vasoinvasive metastasis.
298 JOURNAL of the American Animal Hospital Association July/August 1998, Vol. 34

Discussion and locally invasive, often found invading veins,


Diagnosis of a malignant mesenchymoma requires muscle, fascia, and lymphatics. 3 Metastasis frequently
the presence of two unrelated, neoplastic cell types of occurs to the lungs, liver, and bone. 3,26 These tumors
mesenchymal origin within the tumor. 2 Malignant are found more often in the ventral subcutis and in the
mesenchymoma with vasoinvasive metastasis has not abdominal cavity, 25,29–32 with one study reporting that
been reported previously in either humans or dogs. 50% of the abdominal liposarcomas involved the peri-
Malignant mesenchymomas have been reported in renal fat caudal to the kidney. 33 Wide surgical exci-
nearly every other area of the body in humans; most sion is recommended but usually not possible due to
often reported locations are the lungs, thoracic wall, the invasive nature and locations of these tumors.3
liver, spleen, retroperitoneal, and soft tissues of the The presence of osteoid in sections of the tumor
extremities. 4–7 These tumors have included cell types presented here indicates osteosarcomatous differen-
characteristic of hemangiosarcoma, rhabdomyosar- tiation. In humans, mesenchymomas commonly have
coma, leiomyosarcoma, liposarcoma, chondrosar- either osteosarcomatous or chondrosarcomatous con-
coma, osteosarcoma, synovial cell sarcoma, and stituents or both.19–21,23,34,35 Of the canine cases of
undifferentiated sarcoma. 1 The biological behavior of primary axial osteosarcoma, tumors of the spine oc-
malignant mesenchymomas is unpredictable due to cur about 15% of the time.36 Osteosarcomas rarely
their varied constituent neoplasms. 1 Although exceed- occur extraskeletally but have been reported in the
ingly rare in dogs, mesenchymomas have been re- spleen, liver, kidney, bowel, adrenal gland, mammary
ported in the lungs, liver, spleen, long bones, muscles, tissue, vagina, testicle, and eye of dogs.36 Osteosar-
and digits. 1,8–11 One report of mesenchymomas aris- coma is an extremely aggressive neoplasm, causing
ing in the spleen found that these tumors had the osteoproliferative/osteolytic lesions which often lead
longest median survival time (12 months) after sple- to pathological fractures. Metastasis occurs early and
nectomy when compared with other primary mesen- can be assumed to have occurred in 90% of dogs at
chymal tumors of the spleen. 12 Aggressive surgical the time of presentation. 36
excision of solitary mesenchymomas often is reported The current case remains unique regarding the type
to be curative in humans. However, some reports and location of the tumor as well as its clinical pre-
present hamartomas (which are benign mixed neo- sentation. It remains unclear whether the primary tu-
plasms) as mesenchymomas, which obscures the sur- mor was in the caudal vena cava with later metastasis
vival data. 13 to the thymus and vertebral body of L 3 or if the
The malignant mesenchymoma in this dog had two primary tumor was located in the body of the L3
differentiated malignant cell types, liposarcoma and vertebra with hematogenous metastasis elsewhere.
osteosarcoma. Only one case of an osteochon- One plausible explanation is that the primary mass
droliposarcoma in the radius of a dog has been re- grew in the vena cava, with progressive obstruction
ported.14 In this report, the individual components of venous return and shunting of blood through the
metastasized independently; the osteosarcoma metas- vertebral venous veins and sinuses, ultimately seed-
tasized to the lungs, and the liposarcoma metasta- ing the vertebral body of L 3. A tumor of the caudal
sized to the axillary lymph nodes, adrenal glands, and vena cava which eventually occluded the entire
liver. No case reports of malignant mesenchymoma in venous return, causing minimal clinical signs, must
the vertebral column or vena cava of a dog have been have grown slowly. Additionally, the widespread
reported. In humans, several reports of malignant vasoinvasive metastases belie the aggressive nature
mesenchymomas in the inferior vena cava and heart of this neoplasm.
were found.15–18 Malignant mesenchymoma in bone
has been reported in a number of case reports for References
humans.19–21 Primary or metastatic malignant mesen- 1. Moore RW, Snyder SP, Houchens JW, Folk JJ. Malignant mesenchymoma
chymomas of the central nervous system in humans in a dog. J Am Anim Hosp Assoc 1983;19:187–90.
have been reported.22,23 2. Stout AP, Lattes R. Malignant mesenchymoma. In: Stout AP, Lattes R,
Liposarcomas are uncommon tumors with an over- eds. Tumors of soft tissues. Fascicle 1, second series. Armed Forces
Institute of Pathology, 1967:172–3.
all incidence reported to be between 0.2% and 0.5%
3. MacEwen EG, Withrow SJ. Soft tissue sarcomas. In: MacEwen EG,
of all canine neoplasms. The mean age of dogs at Withrow SJ, eds. Small animal clinical oncology. Philadelphia: WB
diagnosis is 9.7 years.3 Although no human or canine Saunders, 1996:211–26.
cases of liposarcoma in the vertebral column have 4. al-Kana R, Rangwala AF, Sills C, Rienzo A. Case report: malignant
mesenchymoma of the mediastinum. New Jersey Med 1992;89(11):
been published, several cases of liposarcomas invad- 851–5.
ing bone24–27 and one case of an extradural spinal 5. Mukherji SK, Rojiani AM, Younathan CM, Ros P. CT findings of
liposarcoma without primary bony invasion have been retroperitoneal malignant mesenchymoma. Abd Imaging 1994;19(1):
82–3.
reported. 28 Liposarcomas generally are firm but
poorly organized masses which tend to be aggressive
July/August 1998, Vol. 34 Mesenchymoma 299

6. Newman KD, Schisgall R, Reaman G, Guzzetta PC. Malignant 20. Reijnierse M, Kroon HM, Van der Heul RO, Mulder JD. Mesenchymoma
mesenchymoma of the liver in children. J Pediatr Surg 1989;24(8): of bone. A case report. J Bone Joint Surg 1993;75(1):112–5.
781–3. 21. Scheele PM, Von Kuster LC, Krivchenia G. Primary malignant
7. Sheffield M, Rao PR. Malignant mesenchymoma of the pleura. South mesenchymoma of bone. Arch Path & Lab Med 1990;114(6):614–7.
Med J 1996;89(5):526–8. 22. Gibson JN, Reid R, McMaster MJ. Fibrocartilaginous mesenchymoma
8. Carpenter JL, Dayal Y, King NW, Moore FM. Distinctive unclassified of the fifth lumbar vertebra treated by vertebrectomy. Spine
mesenchymal tumor of the digit of dogs. Vet Path 1991;28:396–402. 1994;19(17):1992–7.
9. Hahn KA, Richardson RC. Use of cisplatin for control of metastatic 23. Liwnicz BH, Ferreol EC. Mixed malignant mesenchymoma metastatic
malignant mesenchymoma and hypertrophic osteopathy in a dog. J Am to the central nervous system. Arch Path & Lab Med 1986;110(1):85–7.
Vet Med Assoc 1989;195(3):351–3. 24. Brodey RS, Riser WH. Liposarcoma of bone in a dog: case report. J Am
10. Watson AD, Young KM, Dubielzig RR, Biller DS. Primary mesenchy- Vet Radiolog Soc 1966;7:27–33.
mal or mixed-cell-origin lung tumors in four dogs. J Am Vet Med Assoc 25. Meierhenry EF. Metastatic liposarcoma with extensive osteolysis in the
1993;202(6):968–72. dog. J Am Anim Hosp Assoc 1974;10:478–81.
11. MacDonald RK, Helman RG. Hepatic mesenchymoma in a dog. J Am 26. Dawson EK. Liposarcoma of bone. J Path 1955;70:513–20.
Vet Med Assoc 1986;188(9):1052–3.
27. Brodey RS, Sauer RM, Medway W. Canine bone neoplasia. J Am Vet
12. Spangler WL, Culbertson MR, Kass PH. Primary mesenchymal Med Assoc 1963;143:471–95.
(nonangiomatous/nonlymphomatous) neoplasms occurring in the canine
28. Lewis DD, Kim DY, Paulsen DB, Kerwin SC. Extradural spinal lipo-
spleen: anatomic classification, immunohistochemistry, and mitotic ac-
sarcoma in a dog. J Am Vet Med Assoc 1991;199(11):1606–7.
tivity correlated with patient survival. Vet Path 1994;31:37–47.
29. Doster AR, Tomlinson MJ, Mahaffey EA, Jordan CW. Canine liposar-
13. Jain SK, Afzal M, Mathew M, Ramani SK. Malignant mesenchymoma
coma. Vet Path 1986;23:84–7.
of the chest in an adult. Thorax 1993;48(4):407–8.
30. Garvin CH, Frey DC. Liposarcoma in a dog. J Am Vet Med Assoc
14. Misdorp W, Van der Heul RO. An osteo-(chondro-)liposarcoma (“ma-
1962;140:1073–5.
lignant mesenchymoma”) of the radius in a dog, with two types of
metastases. Zbl Vet Med A 1975;22:187–92. 31. Jabara AG. Three cases of liposarcoma in dogs. J Comp Path 1964;74:
188–91.
15. Reed HM, Poppiti R, Sivina M. Malignant mesenchymoma of kidney
and inferior vena cava. Urology 1983;22(3):297–9. 32. Zwicker GM. Liposarcoma in a dog. Path Vet 1970;7:145–7.
16. Kuwashima Y, Hayashi S, Arata M, Takemura T. Rhabdomyosarcoma 33. Strafuss AC, Bozarth AJ. Liposarcoma in dogs. J Am Anim Hosp Assoc
with focal cartilaginous differentiation (malignant mesenchymoma) of 1973;9:183–7.
the inferior vena cava. Acta Path Japonica 1992;42(5):382–5. 34. Hassan MO, Gogate PA, Hampel N. Malignant mesenchymoma of the
17. McKenney PA, Moroz K, Haudenschild CC, Shemin RJ, Davidoff R. prostate: immunohistochemical and ultrastructural observations. Ultra-
Malignant mesenchymoma as a primary cardiac tumor. Am Heart J structural Path 1994;18(4):449–56.
1992;4(1):1071–5. 35. Satake I, Tari K, Nakagomi K, Ishii M, Kishi K. Retroperitoneal malig-
18. Peters P, Flachskampf FA, Hauptmann S, Lo HB, Schuster CJ. nant mesenchymoma: a case report. Internat J Urol 1994;1(3):273–4.
Bilocular atrial malignant mesenchymoma causing mitral and localized 36. Straw RC. Tumors of the skeletal system. In: MacEwen EG, Withrow SJ,
pulmonary vein flow obstruction: diagnosis by transesophageal eds. Small animal clinical oncology. Philadelphia: WB Saunders,
echocardiography. Eur Heart J 1992;13(11):1585–8. 1996:287–303.
19. Kessler S, Mirra JM, Ishii T, Thompson JC, Brien EW. Primary
malignant mesenchymoma of bone: case report, literature review, and
distinction of this entity from mesenchymal and dedifferentiated chon-
drosarcoma. Skeletal Res 1995;24(4):291–5.
Merck
4C Ad
pull June Trends, p. 5

page 300
Lymphangiosarcoma in a Dog Presenting
with Massive Head and Neck Swelling
A three-year-old, neutered male Chesapeake Bay retriever was presented for acute
onset of severe, progressive swelling of the head, neck, and cranial trunk. Diffuse
lymphangiosarcoma involving the superficial and deep dermis and subcutaneous
tissue was observed on skin biopsies. Lymphangiosarcoma is a rarely reported
tumor of the lymphatic system in dogs and cats. The importance of obtaining skin
biopsies in animals with acute edema of unknown etiology is emphasized.
Additionally, neoplasia should be considered as a potential diagnosis in a dog with
an acute onset of edema. J Am Anim Hosp Assoc 1998;34:301–4.

Theresa W. Fossum, DVM, MS, Introduction


PhD, Diplomate ACVS Lymphangiosarcomas are extremely rare tumors that arise from lym-
Matthew W. Miller, DVM, MS, phatic endothelial cells in humans and domestic animals. In humans,
Diplomate ACVIM they frequently occur secondary to chronic lymphedema.1–3 The pur-
pose of this report is to describe a case of lymphangiosarcoma in a
John T. Mackie, DVM, PhD, dog that presented for evaluation of sudden onset of diffuse swelling
Diplomate ACVP of the head, neck, and cranial trunk.

Case Report
C A three-year-old, neutered male Chesapeake Bay retriever was pre-
sented for evaluation of severe, progressive swelling of the head,
neck, and cranial trunk. The owners first had noted swelling of the
face; the swelling appeared to spread gradually over the ensuing two
months. Treatment with various antibiotics (i.e., amoxicillin, tetracy-
cline, doxycycline), diuretics, and antihistamines did not result in
obvious improvement. Respiratory difficulty was not noted; however,
the dog had become progressively more depressed and exercise intol-
erant over the past two weeks. Appetite remained good, and vomiting
or diarrhea was not noted.
On physical examination, diffuse edema of the head, neck, fore-
limbs, and cranial aspect of the trunk was noted [Figure 1]. The
edema appeared to be fluctuant, pitting, nonpainful, and confined to
the subcutaneous tissues. Differentials included cranial vena cava
obstruction due to an intraluminal mass (e.g., neoplasia, thrombus) or
extraluminal compressive lesion; arteriovenous fistula; lymphedema;
or other lymphatic obstruction.
Regenerative anemia (hematocrit, 19.1%; reference range, 37% to
55%), mild hypokalemia (potassium, 3.4 mmol/L; reference range,
3.5 to 5.0 mmol/L), hypoproteinemia (protein, 4.4 g/dl; reference
From the Departments of Small Animal
range, 5.7 to 7.8 g/dl), and mild hypoalbuminemia (albumin, 2.3 g/dl;
Medicine and Surgery (Fossum, Miller)
and Veterinary Pathobiology (Mackie), reference range, 2.4 to 3.6 g/dl) were present. Massive pleural effu-
College of Veterinary Medicine, sion was noted on thoracic radiographs, but pulmonary or mediastinal
Texas A&M University, masses were not evident. Pleural fluid analysis was consistent with a
College Station, Texas 77843. modified transudate (5,452 white blood cells/µl; total protein, 2.4
g/dl), and the cells were mainly nondegenerative neutrophils with a
Doctor Mackie’s current address is
Veterinary Pathology Services, moderate number of red blood cells (RBCs).
P.O. Box 1119, Ultrasonography was performed to evaluate the thoracic inlet and
Coorparoo DC QLD 4151, Australia. mediastinal structures. Although cavitated fluid was present within

JOURNAL of the American Animal Hospital Association 301


302 JOURNAL of the American Animal Hospital Association July/August 1998, Vol. 34

Figure 2—Photomicrograph of a section of facial skin showing


wide separation of preexisting collagen bundles and their envel-
opment by plump spindle cells (Hematoxylin and eosin stain,
860X).

nal biopsies were obtained through the endoscope.


Additionally, two punch biopsies were taken from the
skin on the ventral aspect of the neck. Copious
amounts of fluid drained from the cutaneous biopsy
sites.
Abnormalities were not observed in the gastric or
duodenal biopsies. Diffuse lymphangiosarcoma in-
volving the superficial and deep dermis and subcuta-
neous tissues was seen on skin biopsies. Thin,
irregular, preexisting collagen bundles were lined gen-
erally by a single layer of plump spindle cells and
Figure 1—Photograph of a Chesapeake Bay retriever with were separated widely by clear spaces devoid of RBCs
acute, massive swelling of the head, neck, and cranial trunk [Figure 2]. Underlying skeletal muscle fibers also
associated with cutaneous lymphangiosarcoma.
were lined by plump spindle cells. These cells were
characterized by cytoplasm which was scant and
the mediastinum, masses were not visualized. Ob- weakly eosinophilic to amphophilic and by nuclei
struction or abnormalities were not noted when the which were oval-to-elongated, slightly pleomorphic,
jugular veins were imaged. A nonselective angiogram 7-to-15 µm long, and hyperchromatic with little in-
was performed to evaluate further the patency of the ternal detail. The same cells also were present in
cranial vena cava. Contrast appeared to flow freely large numbers both individually and in clusters in the
through the vessel, and no intraluminal or extralumi- clear spaces between collagen bundles, and in some
nal mass was identified. Normal cardiac function was places they were congregated upon collagen bundles.
identified on echocardiography, and an intracardiac Mitotic figures were rare. Initially these changes were
lesion was not seen. thought to be consistent with an obstructive lym-
Trial therapy with prednisone (2 mg/kg body phatic disease process. However, after the slides were
weight, per os [PO] bid) and furosemide (4 mg/kg reviewed carefully, the neoplastic nature of the lesion
body weight, PO bid) was initiated and resulted in was identified. Lymphatic neoplasia was confirmed
moderate reduction of the edema; however, the dog by immunohistochemical examination of tissues per-
became increasingly anorectic over the next few days, formed at the Armed Forces Institute of Pathology in
necessitating potassium supplementation. The dog Washington, DC, which demonstrated positive stain-
was placed under general anesthesia, and cervical ing for the presence of Factor VIII-related antigen
radiographs were obtained which were normal. (FVIIIR:Ag) and vimentin in the cytoplasm of these
Lymphoscintigraphy was performed by injecting ra- cells.
diolabeled sulfur colloid near the nose. An apparent
absence of lymphatic flow beyond the region of the Discussion
mandibular lymph nodes was noted. Due to the pro- The diagnosis of lymphangiosarcoma was confirmed
gressive anorexia, gastroduodenoscopy was per- by evaluation of skin biopsies from the edematous
formed, and a large amount of free gastric blood was tissues; this emphasizes the importance of obtaining
found, but no obvious ulceration. Gastric and duode- biopsies in animals with unexplained subcutaneous
July/August 1998, Vol. 34 Lymphangiosarcoma 303

edema. Earlier biopsy of affected tissues may have healthy limb.19 The acute onset of lymphedema in
reduced the number of diagnostic tests performed in this dog would suggest that the tumor caused the
this dog, resulting in less expense and a more rapid lymphedema rather than arose secondary to the
diagnosis. Lymphoscintigraphy, although valuable in lymphedema.
defining the degree of lymphatic obstruction present Histologically, it can be difficult to differentiate
in this dog, did not contribute to the diagnosis of tumors that arise from lymphatic endothelium from
neoplasia and may be less important than biopsy in those of blood vessel endothelial origin. In this case,
dogs with lymphedema, regardless of cause. the histological features of envelopment of preexist-
Neoplasia should not be excluded as a potential ing collagen bundles and skeletal muscle fibers, the
diagnosis in dogs with an acute onset of swelling. lack of significant RBCs within spaces, and the promi-
Although acute head, neck, and forelimb swelling nent edema seen grossly are consistent with previous
(i.e., cranial caval syndrome) is more commonly reports of lymphangiosarcoma in humans,20–22 the
thought of as being caused by vascular occlusion, dog, 7,10,11 and the cat. 8,9,23 Human FVIIIR:Ag is
similar signs may occur due to progressive obstruc- present in both lymphatic and blood vessel endothe-
tion of lymphatics by neoplastic cells or tissue. Fur- lial cells. 24 The positive staining for FVIIIR:Ag and
thermore, in most dogs with experimental occlusion vimentin confirms that this tumor was of endothelial
of the cranial vena cava, collateral blood flow mini- origin.
mizes head and neck swelling; however, many af-
fected dogs develop chylothorax. 4 A necropsy was Conclusion
not performed in this dog, so the cause of the pleural Lymphangiosarcoma should be considered as a pos-
effusion was not identified. However, it most likely sible differential diagnosis in dogs presenting with an
was due to neoplastic obstruction of intrathoracic lym- acute onset of lymphedema. Biopsy of affected tis-
phatics, resulting in transudation of lymph into the sues may be diagnostic and should be performed in dogs
pleural space. The cause of the gastric bleeding is that have no obvious explanation for their swelling.
uncertain, but it may have been related to steroid
therapy.
Congenital lymphedema may be noted first in References
1. Tomita K, Yokogawa A, Oda Y, et al. Lymphangiosarcoma in
middle-aged dogs, and thus the age in this dog did not postmastectomy lymphedema (Stewart-Treves syndrome): ultrastruc-
preclude a diagnosis of congenital lymphedema. How- tural and immunohistologic characteristics. J Surg Oncol 1988;38:
275–82.
ever, in most dogs with congenital lymphedema, the
2. Eby CS, Brennan MJ, Fine G. Lymphangiosarcoma: a lethal complica-
limbs are affected more severely than the head and tion of chronic lymphedema. Arch Surg 1967;94:223–30.
neck. 5,6 Lymphangiosarcoma is a rarely reported tu- 3. Woodward AH, Ivins JC, Soule EH. Lymphangiosarcoma arising in
mor of the lymphatic system in dogs and cats. 7–14 It chronic lymphedematous extremities. Cancer 1972;30:562–72.

has been diagnosed at necropsy in animals presenting 4. Fossum TW, Birchard SJ. Lymphangiographic evaluation of experimen-
tally induced chylothorax after ligation of the cranial vena cava in dogs.
for evaluation of chronic lymphedema. Occasionally, Am J Vet Res 1986;47:967–71.
it has been associated with chylous effusions into the 5. Fossum TW, King LA, Miller MW, et al. Lymphedema: clinical signs,
diagnosis, and treatment. J Vet Int Med 1992;6:312–9.
thoracic or abdominal cavities or subcutaneous
6. Fossum TW, Miller MW. Lymphedema—etiopathogenesis. J Vet Int
spaces.8,15,16 A clear breed or sex predisposition has Med 1992;6:283–93.
not been identified in dogs, but medium- and large- 7. Rudd RG, Veatch JK, Whitehair JG, et al. Lymphangiosarcoma in dogs.
breed dogs may be affected more commonly.17 Al- J Am Anim Hosp Assoc 1989;25:695–8.

though too few cases have been reported to determine 8. Stobie D, Carpenter JL. Lymphangiosarcoma of the mediastinum, mes-
entery, and omentum in a cat with chylothorax. J Am Anim Hosp Assoc
the biological behavior of this tumor, metastasis has 1993;29:78–80.
been present in most dogs with lymphangiosarcoma 9. Walsh KM, Abbott DP. Lymphangiosarcoma in two cats. J Comp Path
1984;94:611–4.
at necropsy. 7,10,11,17 However, surgical excision of a
10. Kelly WR, Wilkinson GT, Allen PW. Canine angiosarcoma (lymphan-
localized mass without clinical evidence of metasta- giosarcoma): a case report. Vet Path 1981;18:224–7.
sis has been reported in a female poodle with lym- 11. Franklin RT, Robertson JJ, Thornburg LP. Lymphangiosarcoma in a
phangiosarcoma.18 In humans, this tumor occurs most dog. J Am Vet Med Assoc 1984;184:474–5.

commonly in patients with chronic lymphedema, and 12. Patnaik AK, Liu S-K. Angiosarcoma in cats. J Sm Anim Pract
1977;18:191.
some authors have suggested that the tumor develops 13. Patnaik AK, Liu S-K, Hurvitz AI, et al. Nonhematopoietic neoplasms in
because of local immunodeficiency associated with cats. J Natl Cancer Inst 1975;54:855–60.
the lymphedema; 1–3 however, the physical pressure 14. Lamb WA, Muir P. Lymphangiosarcoma associated with hyponatremia
and hyperkalemia in a dog. J Sm Anim Pract 1994;35:374–6.
imposed on the endothelium by chronic lymphedema
15. Fossum TW, Hodges CC, Scruggs DW, et al. Generalized lymphan-
may be a more plausible trigger. Stark, et al. found giectasis in a dog with subcutaneous chyle and lymphangioma. J Am Vet
that in vivo transplantation of lymphangiosarcoma to Med Assoc 1990;197:231–6.

the extremity of a lymphedematous limb was more


successful than if the tumor was transplanted to a (Continued on next page)
304 JOURNAL of the American Animal Hospital Association July/August 1998, Vol. 34

References (cont’d) 20. Lattes R. Tumors of the soft tissues. In: Hartmann WH, Cowan WR, eds.
Atlas of tumor pathology. Washington, DC: Armed Forces Institute of
16. Myers NCI, Engler SJ, Jakowski RM. Chylothorax and chylous ascites Pathology, 1982:205–9.
in a dog with mediastinal lymphangiosarcoma. J Am Anim Hosp Assoc
21. Malignant vascular tumors. In: Enzinger FM, Weiss SW, eds. Soft tissue
1996;32:263–9.
tumors. 2nd ed. St. Louis: Mosby, 1988:545–56.
17. Sagartz JE, Lairmore MD, Haines D, et al. Lymphangiosarcoma in a
22. Tumours of vasoformative tissue. In: Ashley DJB, ed. Evans’ histologi-
young dog. Vet Path 1996;33:353–6.
cal appearances of tumours. 4th ed. Edinburgh: Churchill Livingstone,
18. Shiga A, Shirota K, Une Y, et al. Lymphangiosarcoma in a dog. J Vet 1990:99–101.
Med Sci 1994;56:1199–202.
23. Swayne DE, Mahaffey EA, Haynes SG. Lymphangiosarcoma and
19. Stark RB, Dwyer EM, Forest MD. Effect of surgical ablation of regional hemangiosarcoma in a cat. J Comp Path 1989;100:91–6.
lymph nodes on survival of skin homografts. Ann N Y Acad Sci
24. Burgdorf WHC, Mukai K, Rosai J. Immunohistochemical identifica-
1960;87:140–8.
tion of factor VIII-related antigen in endothelial cells of cutaneous
lesions of alleged vascular nature. Am J Clin Path 1981;75:167–71.
Spinal Epidural Empyema in Two Dogs
Extensive, diffuse, epidural spinal cord compression was visualized myelographically
in two dogs presented for rapid development of nonambulatory tetraparesis and
paraplegia, respectively. Purulent fluid containing bacterial organisms was aspirated
percutaneously under fluoroscopic guidance from the epidural space of each dog.
One dog responded poorly to aggressive medical therapy, which included installation
of an epidural lavage and drainage system. Both dogs were euthanized due to the
severe nature of their disorder and the poor prognosis. Spinal epidural empyema
(i.e., abscess) is a rare condition in humans and has not been reported previously
in the veterinary literature. Spinal epidural empyema should be considered as a
differential diagnosis in dogs presenting with painful myelopathies, especially when
accompanied by fever. J Am Anim Hosp Assoc 1998;34:305–8.

Curtis W. Dewey, DVM, MS, Introduction


Diplomate ACVIM, Bacterial infections of the central nervous system (CNS) are not
Diplomate ACVS
reported commonly in dogs and cats. Potential sources for CNS bac-
Gregg D. Kortz, DVM, terial infections include hematogenous spread and direct extension
Diplomate ACVIM from a nearby contaminated or infected area.1–3 Spinal epidural em-
pyema (i.e., abscess), an accumulation of purulent material in the
Cleta S. Bailey, DVM, PhD, epidural space of the vertebral canal, is a rare disorder in humans and
Diplomate ACVIM has not been documented in dogs. Hallmarks of the human disease
include spinal hyperesthesia and fever. Diagnosis of spinal epidural
empyema often is delayed in humans because of its rarity and nonspe-
C cific clinical signs. Unfortunately, delays in diagnosis and appropri-
ate treatment are associated with an increased likelihood of permanent
neurological dysfunction or death.4–13
Two dogs with initial clinical signs of spinal hyperesthesia and
fever each progressed to nonambulatory status within a 24-hour pe-
riod. The dogs had purulent, epidural fluid accumulations causing
spinal cord compression. The purposes of this report are to describe
the clinical and myelographic findings of these two dogs with spinal
epidural empyema and to review the current literature concerning this
disease in humans.

Case Reports

Case No. 1
A two-year-old, intact male Irish wolfhound presented to the Univer-
sity of California-Davis Veterinary Medical Teaching Hospital (UCDVMTH)
From the Veterinary Medical Teaching following neck and back pain of nine days’ duration that had pro-
Hospital (Dewey, Kortz) and the gressed to nonambulatory tetraparesis 24 hours prior to admission.
Department of Surgical and Radiological
The owner noted that the dog was experiencing pain shortly after a
Sciences (Bailey),
University of California, hike they had taken in a wooded area. The patient had been examined
Davis, California 95616. by two veterinarians prior to referral to the UCDVMTH. The dog had
been found to be febrile as well as in pain and was treated with
Doctor Dewey’s current address is the glucocorticoids, nonsteroidal anti-inflammatories, a first-generation
Department of Small Animal Medicine
cephalosporin, and enrofloxacin. Upon physical examination, the pa-
and Surgery,
College of Veterinary Medicine, tient was depressed, febrile (105˚ F), and reacted painfully to palpa-
Texas A&M University, tion of the cervical and caudal thoracic spinal regions. Upon
College Station, Texas 77843-4474. neurological examination, the patient was found to be nonambulatory

JOURNAL of the American Animal Hospital Association 305


306 JOURNAL of the American Animal Hospital Association July/August 1998, Vol. 34

A A

Figures 2A, 2B—Radiographs demonstrating the placement of


epidural catheters in the (A) cervical and (B) thoracolumbar
spinal regions of case no. 1. The catheters were to serve as an
ingress-egress system for lavage of the epidural space. (Cour-
tesy of Phil Koblik, DVM, MS, Diplomate ACVR).

A needle aspirate of the epidural space was per-


Figures 1A, 1B—Myelographic appearance of epidural empy- formed dorsally at the L 1 to second lumbar (L 2) inter-
ema in case no. 1. The arrows indicate the ventral displacement vertebral level under fluoroscopic guidance. An
of the dorsal contrast column by epidural fluid in the (A) cervical
and (B) thoracolumbar regions of the spine. (Courtesy of Phil opaque, orange-colored fluid was retrieved. Numerous
Koblik, DVM, MS, Diplomate ACVR). degenerative neutrophils and abundant gram-positive
cocci (the latter of which were arranged frequently in
chains) were identified on cytological examination of
tetraparetic with intact spinal reflexes. Due to the the epidural fluid. Bacterial organisms subsequently
multiple areas of spinal hyperesthesia, a multifocal or were cultured from the epidural fluid and identified
diffuse myelopathy (rather than a focal, cervical spi- as beta-hemolytic Streptococcus canis. Blood cul-
nal cord lesion) was suspected. Hematological and tures also were positive for the same organism. After
serum biochemical profile abnormalities included a obtaining the cytology report, two polypropylene
mature neutrophilia with lymphopenia, thrombocy- catheters were placed in the dorsal epidural space,
topenia, mild hypoalbuminemia, hypercholesterolemia, under fluoroscopic guidance. One was placed at the
and increased levels of alkaline phosphatase and cre- C 1 to second cervical (C 2 ) intervertebral level and the
atine kinase. Fibrinogen levels also were elevated. other was placed at the 13th thoracic (T 13) to the L 1
The patient was anesthetized, and survey radio- intervertebral level [Figures 2A, 2B]. These catheters
graphs of the spine were obtained. Slight narrowing were to serve as an ingress-egress system for lavage
at the sixth cervical (C 6) to seventh cervical (C7) of the epidural space.
intervertebral disk space, with dorsal tipping of the The patient was transferred to the intensive care
cranial aspect of the C 7 vertebra, was appreciated on unit (ICU) and placed on intravenous fluid and antibi-
the survey radiographs. A lumbar cerebrospinal fluid otic therapy (i.e., ticarcillin and clavulanic acid). The
(CSF) sample was analyzed prior to injection of dog experienced a respiratory arrest and was intu-
iopamidol for the myelogram. This CSF included 81 bated. The patient did not regain consciousness after
red blood cells (RBCs)/µl, three nucleated cells/µl the arrest and had to be placed on ventilatory support,
(73% neutrophils, 27% mononuclear cells), and a pro- as there were no spontaneous efforts to breathe. Due
tein level of 153 mg/dl. This was interpreted as mild to the poor prognosis, the owner elected to have the
purulent inflammation. A cerebellomedullary CSF dog euthanized. Upon necropsy examination, puru-
sample was obtained and appeared to be blood con- lent fluid was evident throughout the entire spinal
taminated. This sample was analyzed as the myelo- epidural space, with the largest accumulation in the
gram was being performed. This CSF contained area of the L 1-L 2 intervertebral space. Histopathology
22,600 RBCs/µl, 4,800 nucleated cells/µl (96% neu- results included evidence of a diffuse, severe, fibri-
trophils, 4% mononuclear cells), and a protein level nopurulent meningomyelitis and a mild, neutrophilic
of 2,106 mg/dl. This fluid was interpreted as purulent cerebral meningitis.
inflammation. No organisms were seen in either CSF
sample. Severe dorsal compression of the spinal cord Case No. 2
from the first cervical (C 1) to third cervical (C3 ) and A two-year-old, intact male Irish wolfhound presented
the third thoracic (T3) to first lumbar (L 1 ) vertebral to the UCDVMTH for anorexia and inactivity of four
segments was identified on myelography [Figures 1A, 1B]. days’ duration. On the morning of admission, the dog
July/August 1998, Vol. 34 Spinal Epidural Empyema 307

A (L3 ) to fourth lumbar (L4) intervertebral level. The


fluid retrieved was hemorrhagic and viscous. Cyto-
logical findings from this fluid included a small num-
ber of gram-positive cocci among neutrophils and
numerous RBCs. Subsequent culture of this epidural
fluid yielded Staphylococcus intermedius and Clostrid-
ium perfringens. Blood cultures also were positive
for both organisms. The hemorrhagic nature of both
the diarrhea and epidural fluid in this dog prompted
measurement of a prothrombin time and a partial
B thromboplastin time. These were within normal lim-
its. Due to the poor prognosis for recovery, the owner
elected to euthanize the patient. Necropsy results in-
cluded abundant hemorrhagic fluid in the epidural
space from the T 4 vertebral level caudally, as well as
hemorrhagic gastrointestinal contents. The mucosa of
the jejunum and ileum were reddened diffusely. Se-
vere epidural and intradural hemorrhage from the T4
Figures 3A, 3B—Myelographic appearance of epidural empy- spinal cord segment to the level of the cauda equina
ema in case no. 2. The arrows indicate the ventral displacement was evident on histopathology. Pathological changes
of the dorsal contrast column by epidural fluid in the (A) cranial
thoracic and (B) thoracolumbar regions of the spine. (Courtesy within the cord were limited to focal areas of is-
of Phil Koblik, DVM, MS, Diplomate ACVR). chemic neuronal necrosis and intraparenchymal he-
morrhage. No gross or histopathological evidence of
had become paraplegic. On physical examination, the a coagulopathy was identified.
patient was found to be depressed, moderately dehy-
drated, febrile (106.4˚ F), and tachypneic, with bloody Discussion
diarrhea and scleral hyperemia. The patient appeared Spinal epidural empyema has not been reported pre-
to be in pain when the thoracolumbar and lumbosac- viously in dogs. The authors prefer the terminology
ral regions of the spine were palpated. Findings of a “empyema” over “abscess” to describe the disorder,
neurological examination included a miotic right pu- as the former term more accurately describes the ex-
pil, purposeful motor function and intact spinal re- tent of the purulent fluid accumulation.14 Both cases
flexes in the thoracic limbs, and absence of motor had exhibited vague and nonspecific clinical signs
function, spinal reflexes, and deep pain perception in before deteriorating neurologically, a scenario de-
the pelvic limbs. A diffuse-to-multifocal myelopathy, scribed in the human counterpart of the disease. Both
from the T 3 to first sacral (S 1) spinal cord seg- cases also presented with clinical signs of spinal hy-
ments, was suspected based upon the neurological peresthesia and fever, recognized as hallmarks of spi-
examination. Hematological and serum biochemi- nal epidural empyema in humans. 4–13
cal profile abnormalities included a mature neutro- Spinal epidural empyema is a rare disease in hu-
philia with lymphopenia, mild elevations of blood mans. 4–13 Most physicians never encounter this dis-
urea nitrogen (BUN) and creatinine, mildly increased ease in their professional careers.5 The rarity of spinal
total bilirubin, increased alkaline phosphatase lev- epidural empyema, combined with the relatively non-
els, hypercholesterolemia, and hyperfibrinogenemia. specific, typical clinical signs, often results in a delay
The patient was anesthetized, and survey radio- in diagnosis. 4–13 Neurological deterioration can
graphs of the spine were obtained. No abnormalities progress rapidly. 6–8,11–13 Patients rarely regain neuro-
were evident on these radiographs. A lumbar CSF logical function if they are paralyzed for more than
sample was analyzed, and the results included 282,000 24-to-48 hours before a definitive diagnosis is reached
RBCs/µl, 340 nucleated cells/µl (86% neutrophils, and appropriate treatment instituted.7,11,12 Delayed di-
14% mononuclear cells), and a protein level of 209 agnosis and therapy also are associated with an in-
mg/dl. This was interpreted as hemorrhagic CSF. A creased mortality rate;5,7–10,12 therefore, a high index
myelogram was performed via lumbar injection of of suspicion is required in order to identify patients
iopamidol. A space-occupying lesion in the dorsal with this disease rapidly and to treat them successfully.
epidural space from the fourth thoracic (T4 ) to sixth A number of bacterial etiological agents have been
lumbar (L 6) vertebral levels was identified, compress- reported, the most common being Staphylococcus
ing the spinal cord [Figures 3A, 3B]. Under fluoro- aureus. 4–13 Species of Streptococcus also are impli-
scopic guidance, a needle aspirate of the dorsal cated frequently. 5,7,8,11,15 While the source of infec-
epidural space was obtained from the third lumbar tion not always is identified, skin infections are
308 JOURNAL of the American Animal Hospital Association July/August 1998, Vol. 34

thought to be the most common source. 5,8,11 Blood should be considered. To the authors’ knowledge,
cultures often yield the same organism found in the such an immune-deficiency disorder has not been de-
epidural space. 7 The pathophysiological mechanisms scribed in this breed.
responsible for neurological dysfunction may include As demonstrated in this report, spinal epidural em-
both direct compression of neural tissue by purulent pyema should be included on the differential diagno-
material as well as impairment of intrinsic spinal cord sis list for canine myelopathies. It would appear from
vasculature. 5,7,8,11,13 Factors that may predispose hu- the rapid deterioration and ultimate death of these
mans to spinal epidural empyema include spinal two dogs, that, similar to the analogous human dis-
trauma, diabetes mellitus, immunosuppression, intra- ease, timely diagnosis is paramount to successful
venous drug use, and alcoholism. 7,9,12,13 Evidence of therapy. A high index of suspicion should be main-
accompanying vertebral osteomyelitis is seen in less tained for spinal epidural empyema, especially in the
than 50% of the cases. 7,8,11,12 Cerebrospinal fluid dog with a painful myelopathy accompanied by fever.
analysis often yields a mixed pleocytosis and elevated
protein levels.5–8,10–12 Diagnosis of spinal epidural
empyema can be aided by myelography, computed References
1. Braund KG. Neurological diseases. In: Braund KG, ed. Clinical syn-
tomography (CT), or magnetic resonance imaging dromes in veterinary neurology. 2nd ed. St. Louis: Mosby-Year Book,
(MRI). Traditionally, myelography has been the diag- 1994:81–332.
nostic tool of choice; MRI is preferred currently due 2. Fenner WR. Bacterial infections of the central nervous system. In:
Greene CE, ed. Infectious diseases of the dog and cat. 2nd ed. Philadel-
to its noninvasive nature. 4–13,15,16–19 phia: WB Saunders, 1990:184–96.
Some controversy exists concerning medical ver- 3. Dow SW, LeCouteur RA, Henik RA, Jones RL, Poss ML. Central
sus surgical treatment of spinal epidural empyema. nervous system infection associated with anaerobic bacteria in two
dogs and two cats. J Vet Int Med 1988;2:171–6.
Most authors advocate surgical drainage in combina- 4. Lange M, Tiecks F, Schielke E, Yousry T, Haberl R, Oeckler R. Diagno-
tion with antibiotic therapy.4–13 However, consider- sis and results of different treatment regimes in patients with spinal
able success has been reported with nonsurgical abscesses. Acta Neurochir (Wien) 1993;125:105–14.
5. Maslen DR, Jones SR, Crislip MA, Bracis R, Dworkin RJ, Flemming JE.
management of this disorder. 4,5,9,11,18,20 In some situa- Spinal epidural abscess: optimizing patient care. Arch Int Med
tions, nonsurgical management has been regarded as 1993;153:1713–21.
the preferred mode of therapy.5,11,12 Indications for 6. King M. Spinal epidural abscess: an elusive diagnosis. South Med J
1994;87(2):288–9.
nonsurgical treatment include 1) patients who are poor
7. Darouiche RO, Hamill RJ, Greenberg SB, Weathers SW, Musher DM.
surgical candidates; 2) an extensively long lesion, for Bacterial spinal epidural abscess: review of 43 cases and literature
which a laminectomy would be impractical and possi- survey. Medicine 1992;71(6):369–85.

bly lead to vertebral instability; 3) lack of severe 8. Baker AS, Ojemann RG, Swartz MN, Richardson EP. Spinal epidural
abscess. N Engl J Med 1975;293(10):463–8.
neurological deficits; and 4) complete paralysis for 9. Curling OD, Gower DJ, McWhorter JM. Changing concepts in spinal
more than 72 hours. 5,11,12 epidural abscess: a report of 29 cases. Neurosurg 1990;27(2):185–92.
Both dogs described in this report had extensive 10. Danner RL, Hartman BJ. Update of spinal epidural abscess: 35 cases and
review of the literature. Rev Infect Dis 1987;9(2):265–74.
accumulations of fluid extending over multiple verte- 11. Carey ME. Infections of the spine and spinal cord. In: Youmans JR, ed.
bral segments. The reason for the hemorrhagic nature Neurological surgery: a comprehensive reference guide to the diagnosis
and management of neurosurgical problems. 4th ed. Vol. 5. Philadel-
of the epidural fluid in case no. 2 remains unknown. phia: WB Saunders, 1996:3270–304.
A potential source of infection was not identified in 12. Simpson RK, Azordegan PA, Sirbasku DM, Weathers SW, Lidsky MD,
case no. 1, but blood cultures were positive for the Baskin DS. Rapid onset of quadriplegia from a panspinal epidural
abscess. Spine 1991;16(8):1002–5.
same organism found in the epidural fluid. Case no. 2 13. McGee-Collett M, Johnston IH. Spinal epidural abscess: presentation
had enteritis and associated hemorrhagic diarrhea, and treatment—a report of 21 cases. Med J Aust 1991;155:14–7.
making the gastrointestinal tract a likely source of 14. Anderson DM, ed. Dorland’s illustrated medical dictionary. 28th ed.
Philadelphia: WB Saunders, 1994:546.
infection. The same two organisms each were cul- 15. Gelfand MS, Bakhtian BJ, Simmons BP. Spinal sepsis due to Streptococ-
tured from the epidural fluid and blood from case no. cus milleri: two cases and review. Rev Infect Dis 1991;13:559–63.
2. An epidural drainage and lavage system was placed 16. Spiegelmann R, Findler G, Faibel M, Ram Z, Shacked I, Sahar A.
Postoperative spinal epidural empyema: clinical and computed tomogra-
successfully in case no. 1, similar to a system used in phy features. Spine 1991;1(10):1146–9.
humans with spinal epidural empyema.21 Unfortu- 17. Browman MW, Lapointe JS. Spinal epidural abscess: unusual CT and
microbiologic findings. Can Assoc Radiol J 1993;44(3):215–6.
nately, this patient was euthanized due to systemic
18. Hanigan WC, Asner NG, Elwood PW. Magnetic resonance imaging and
complications before this therapy could be evaluated. the nonoperative treatment of spinal epidural abscess. Surg Neurol
Case no. 2 was euthanized without attempting therapy, 1990;34:408–13.

due to the poor prognosis. Interestingly, both patients 19. Teman AJ. Spinal epidural abscess: early detection with gadolinium
magnetic resonance imaging. Arch Neurol 1992;49:743–6.
were young, male Irish wolfhounds. While this may 20. Wheeler D, Keiser P, Rigamonti D, Keay S. Medical management of
be a simple coincidence, it is unusual for the only two spinal epidural abscesses: a case report and review. Clin Infect Dis
1992;15:22–7.
reported cases of such a reputedly rare disease to
21. Cwikiel W. Percutaneous drainage of abscess in psoas compartment and
occur in the same breed. The possibility of an im- epidural space. Acta Radiologica 1991;32:159–61.
mune-deficiency problem in Irish wolfhounds also
Surgical and Chemotherapeutic Treatment
of Alveolar Echinococcosis in a Dog
Surgical removal of macroscopically detectable metacestode tissue followed by
postoperative chemotherapy according to established human protocols resulted in
complete clinical remission and immediate normalization of hyperglobulinemia in a
dog with alveolar echinococcosis (AE). The disease is caused by the metacestode
stage of the cestode, Echinococcus multilocularis. In endemic areas, AE should be
included in the differential diagnosis of polycystic liver masses, especially if
concomitant hyperglobulinemia is present. However, the importance of AE is not
only the disease of the single dog itself but also the potential risk of infection for
humans in an endemic area. J Am Anim Hosp Assoc 1998;34:309–14.

Markus Haller, Dr. med. vet. Case Report


Peter Deplazes, Dr. med. vet. In October 1995, a 10-year-old, male dachshund was presented to the
University of Zürich Veterinary Teaching Hospital with a history of
Franco Guscetti, Dr. med. vet. gradually progressive abdominal enlargement. Earlier the dog had
developed general exercise intolerance. Food and water intake was
Juan C. Sardinas, DVM, normal, and the dog was alert with no other clinical signs of illness.
Diplomate ACVS The dog had never been used as a hunting dog, but catching mice was
Iris Reichler, Dr. med. vet. reported to be one of his favorite habits. Upon physical examination,
the dog was excited, in good general condition, and had a pendulous
Johannes Eckert, abdomen and a rectal temperature of 39.3˚ C. Upon palpation, a large,
Prof. Dr. med. vet. lobulated mass was found occupying the entire abdominal space. No
other abdominal organs could be identified on palpation. Initial diag-
nostic evaluation included a complete blood count (CBC), serum
C biochemistry, urinalysis, thoracic and abdominal radiographs, and
abdominal ultrasonography.
A lobulated mass filling the abdomen and displacing the stomach
dorsally and small intestine caudodorsally [Figure 1] was noted on
abdominal radiographs. Thoracic radiographs were within normal
limits. On ultrasonographic examination, the abdominal mass was
very heterogeneous, consisting of nodular parts with high echogeni-
city and cavernous parts with lower echogenicity [Figure 2]. Some of
the cavernous parts appeared to be filled with fluid. The liver was
presumed to be the origin of the mass, but several liver lobes of
normal texture and echogenicity also were identified.
From the Veterinary Teaching Hospital, The only hematological change found was a mild, nonregenerative
Department of Veterinary Internal anemia (packed cell volume, 36%). Biochemical changes included a
Medicine (Haller), marked hyperproteinemia (protein, 114 g/L; reference range, 59 to 67
Institute of Parasitology g/L), hypoalbuminemia (albumin, 26 g/L; reference range, 27 to 37
(Deplazes, Eckert),
g/L), hypocholesterolemia (cholesterol, 3.0 mmol/L; reference range,
Institute of Veterinary Pathology
(Guscetti), 4.2 to 8.6 mmol/L), and a mildly elevated creatine kinase activity
Department of Veterinary Surgery (255 U/L; reference range, 68 to 197 U/L). All other parameters
(Sardinas, Reichler), including hepatic enzyme activities were within normal limits. Urine
University of Zürich, protein/creatinine ratio was 0.8. A serum electrophoresis was per-
Winterthurerstr 260, CH-8057 Zürich.
formed; hypergammaglobulinemia was the cause of the hyperpro-
Doctor Haller’s current address is teinemia [Figure 3].
Kleintierklinik Dr. Bolliger, Under ultrasonographic guidance, a fine-needle aspirate of the
Kieferstrasse 2, CH-4665 Oftringen. abdominal mass was made, and several milliliters of a milky fluid

JOURNAL of the American Animal Hospital Association 309


310 JOURNAL of the American Animal Hospital Association July/August 1998, Vol. 34

Figure 1—Lateral abdominal radiograph of a male dachshund


presented for gradual abdominal enlargement. A lobulated
mass is seen displacing the stomach dorsally and the small
intestines caudodorsally.

Figure 3—Serum electrophoresis before initiation of treatment,


identifying a hypergammaglobulinemia as the cause of the
hyperproteinemia.

Figure 2—Ultrasonographic examination of the mass, showing


a heterogenous echogenicity and multiple, cavernous, fluid-
filled structures.

were removed for analysis. The specific gravity of Figure 4—Intraoperative photograph of the firm, pale yellow-to-
white, abdominal mass. Note the multiple surface lobulations.
the fluid was 1.041, protein was 68 g/L, and triglycer-
ide concentration was below the detectable range.
Approximately 1,000 cells/µl were counted; most of moved from the omentum. All other abdominal or-
them were damaged severely. Cytological evaluation gans appeared normal. The resected liver lobes and
of the fluid revealed cellular debris, with occasional omental mass were submitted for histopathology.
plasma cells and mesenchymal cells that had irregular Histologically, the hepatic mass consisted of nu-
nucleoli and a coarse chromatin pattern. Cytology did merous cystic structures, each of which was lined by
not allow a definitive diagnosis. a thin (10-to-15 µm), acellular, laminated layer which
A ventral midline exploratory laparotomy was per- was folded irregularly [Figure 5]. This layer was peri-
formed. A large, bilobar mass was found occupying odic-acid-Schiff (PAS) positive. In some of the cysts,
much of the abdominal cavity. The mass was firm, a thin and delicate layer consisting of flattened cells
pale yellow-to-white, with multiple surface lobula- with small nuclei was seen on the inner side of the
tions. It measured 13-to-15 cm in diameter [Figure 4]. acellular membrane. The cysts were filled with amor-
The mass originated from the left-medial and right- phous material, and some contained a few calcareous
lateral liver lobes; the other liver lobes appeared nor- bodies. Necrotic material with areas of calcification
mal. Complete lobectomy of the affected liver lobes surrounded the cysts and was delineated by active
was performed, and the whole mass was removed. A granulation tissue with neutrophils, plasma cells, and
small mass, 2-to-3 mm in diameter, which appeared a few lymphocytes. Focally there was a granuloma-
grossly similar to the primary mass, also was re- tous inflammation with single, multinucleated giant
July/August 1998, Vol. 34 Alveolar Echinococcosis 311

Figure 5—Hepatic cyst of Echinococcus multilocularis consist-


ing of an acellular membrane which includes few cells and
amorphous material. The cyst is embedded in necrotic tissue
surrounded by a granulomatous type of reaction with granula-
tion tissue (Hematoxylin and eosin stain, 300X; bar=50 µm).

cells. The histological findings were considered to be


typical for metacestodes of the tapeworm Echinococ-
cus multilocularis (E. multilocularis), although no
protoscoleces were identified. 1 The metacestode stage
of E. multilocularis is known as alveolar echinococ-
cosis (AE).
Following histological diagnosis, serology was per-
formed by enzyme-linked immunosorbent assay
(ELISA) using antigens (the metacestode antigens,
EmG11 2 and Em2, which both were affinity purified
and which are functionally identical, and the recom-
binant antigen II/3-10, which is functionally different Figure 6—Serum electrophoresis after three months of treat-
from EmG11 and Em22) which are highly specific for ment. Hypergammaglobulinemia has disappeared almost com-
pletely (compare to Figure 3).
E. multilocularis. A mixture of the Em2 and the II/3-10
antigens is used in the commercial Em2 plus ELISA. a,3 for metastasis formation of E. multilocularis meta-
A strong positive result was obtained in the ELISA cestodes. 6,7 Albendazole b (10 mg/kg body weight,
with the II/3-10 antigen, and weakly positive results once daily with food) was administered. The treat-
were seen with the EmG11 antigen as well as with the ment consisted of alternating cycles of four weeks’
Em2 plus-ELISA. treatment followed by two weeks without medica-
To exclude a concomitant intestinal infection with tion. 7 To exclude any undetected concomitant infec-
adult tapeworms of the same species, several fecal tion with intestinal stages of E. multilocularis, the
samples were examined within 10 days by a com- dog was treated orally with praziquantelc (10 mg/kg
bined sedimentation/flotation method for cestode eggs body weight [double the therapeutic dose]). This
and by an ELISA for specific coproantigens of E. medication was repeated once after two weeks. An
multilocularis. 4 All tests were negative for intestinal intensive follow-up monitoring program was estab-
parasites. These negative results indicated that the lished. Every three months, a CBC, serum biochemis-
dog did not carry intestinal E. multilocularis stages at try profile, serum electrophoresis, ultrasonographic
the time of examination. However, a previous infec- examination of the liver, and serological tests for E.
tion and thus a potential zoonotic risk for humans multilocularis were performed. To date (May 1997),
could not be ruled out. Therefore, several humans the dog has been treated with 15 cycles of chemo-
who had been in contact with the dog were tested by a therapy during a follow-up period of 21 months.
serological surveillance program recommended by the The dog has been clinically normal and in good
Swiss National Center for Echinococcosis; 5 no posi- general condition. Three months after surgery, the
tive reactions were found. hyperglobulinemia had resolved and the serum elec-
Postoperative chemotherapy similar to a protocol trophoresis had a normal distribution pattern [Figure
used in human medicine was begun because of the 6]. Repeated biochemical profile results remained in
tumor-like proliferation and the well-known potential the normal reference range, including hepatic enzyme
312 JOURNAL of the American Animal Hospital Association July/August 1998, Vol. 34

logical role in the parasite’s life cycle. They are de-


Table nominated as accidental hosts which may suffer from
AE, the disease caused by the proliferation of the
metacestode stage of E. multilocularis. Dogs may
acquire the E. multilocularis infection as final hosts
or as accidental hosts. Therefore, epidemiological and
clinical aspects of this dual role will be discussed.
The geographic distribution of E. multilocularis is
restricted to the northern hemisphere and includes
parts of central and eastern Europe, Asia, and North
America. 9,10 In European countries and regions, the
average prevalence rates of intestinal E. multilocularis
infection in foxes range from less than 1% to greater
than 40%.10 In an endemic part of eastern Switzer-
land, with prevalence rates of E. multilocularis infec-
tion in foxes being greater than 30%, only 0.3% of
661 dogs and 0.2% of 452 cats were infected with
intestinal stages of the parasite.10 The dog presented
in this report originated from a highly endemic area
with an average prevalence rate of 44% for E.
Table—Serial enzyme-linked immunosorbent assay (ELISA) multilocularis infection in the fox population; this is
results with various antigens specific for Echinococcus
multilocularis before and after surgery. For each serum sample, one of the unusual cases in which the dog is an acci-
relative optical densities (ODs) were calculated by subtraction dental host for the metacestode stage. Three similar
of antigen-dependent cut-off values. Values above zero are cases in dogs had been observed previously in Switzer-
therefore considered positive. For each antigen, a cut-off value
was determined as the mean OD plus three standard deviations land, and two cases have been reported from Germany,
of 13 clinically normal control dogs. but they were diagnosed postmortem or euthanized
immediately after diagnosis without treatment.10,11
activities. Ultrasonographic evaluation of the liver The eggs of E. multilocularis may reach the dog’s
revealed a displacement of the gall bladder to the left, intestine from the environment after peroral ingestion
resulting from the loss of the two resected liver lobes. or they may be released within the dog’s own intes-
The remaining liver tissue appeared to be normal in tine if the animal is infected with adult stages of the
texture and echogenicity. Serum antibody levels were parasite. It has been demonstrated for Taenia hyda-
less markedly positive 46 days after surgery in all tigena and other Taenia species that oncospheres may
three serological tests. After 109 days, only the II/3- hatch from some eggs which are released into the
10 ELISA showed low specific reaction which con- small intestine.12 Furthermore, oncosphere-specific
verted to negative in the following test dates on days serum antibodies have been detected in dogs two
221 and 326 after surgery [see Table]. months after experimental infection with Echinococ-
After almost two years of treatment, the dog did cus granulosus (E. granulosus), which suggests that
not show any signs of recurrence of the disease or oncospheres penetrated the mucosa and stimulated
side effects from the medication. The dog continues the immune system.13 Details on this early phase of
to be treated with albendazole as described previously. infection still are unknown. A similar mode of infec-
tion is assumed for E. multilocularis, resulting in
Discussion migration of hatched oncospheres from the mucosa to
Echinococcus multilocularis is a cestode for which the liver with subsequent development to the meta-
the natural life cycle typically involves rodents as cestode stage. However, it appears that such infec-
intermediate hosts and foxes as definitive hosts. In tions of dogs resulting in AE are extraordinary events.
some geographic regions, other carnivores (e.g., coy- Intestinal stages of E. multilocularis normally do
otes, wolves, domestic dogs or cats) may act as final not cause clinical signs in the definitive hosts, even if
hosts. 8 Intermediate hosts ingest parasite eggs (which a large number (greater than 10,000 per animal) of
have been excreted by the definitive hosts) from which parasites are present. Up until now, a specific diagno-
metacestodes develop. Definitive hosts acquire the sis in the living animal was very difficult to make, as
infection by the ingestion of intermediate host ani- the small proglottids of E. multilocularis excreted in
mals infected with the larval stage (i.e., metacestode) the feces often were overlooked and the eggs of the
containing protoscoleces of the parasite. Occasion- parasite could not be differentiated microscopically
ally, eggs from E. multilocularis may be ingested by from those of Taenia spp., E. granulosus, or Multi-
humans or animals which do not play an epidemio- ceps spp. By the use of newly developed techniques
July/August 1998, Vol. 34 Alveolar Echinococcosis 313

such as coproantigen detection by sandwich ELISA liver of one dog and in the omentum exclusively in
or characterization of E. multilocularis eggs by polym- the other dog. 11 In the latter case, the liver apparently
erase chain reaction (PCR), reliable diagnosis of the was not involved; however, liver lesions may be small
intestinal infection is now possible.4,14 In contrast to and grossly undetectable in the living animal. In the
the intestinal stages of E. multilocularis, the meta- case reported here, parasite masses had infiltrated
cestode can be highly pathogenic in intermediate and larger parts of the liver and also the peritoneal cavity.
accidental hosts. In humans, it may cause severe AE Recommendations for treatment of dogs carrying
which is lethal in more than 50% to nearly 100% of metacestodes of E. multilocularis could not be found
untreated patients. 6 The location of metacestodes in in the veterinary literature. Therefore, a schedule from
humans as well as in other accidental hosts and natu- human medicine was adopted for the treatment of this
ral intermediate hosts is almost exclusively the liver. case. Surgical resection of all macroscopically abnor-
This stage of the parasite is characterized by tumor- mal tissue and postoperative chemotherapy with
like proliferation and by the potential of metastasis albendazole was established. The dog has been in
formation in adjacent or distant organs. 6 clinical remission for 21 months postoperatively.
The initial clinical symptoms of hepatic AE in Treatment with albendazole is recommended indefi-
humans are epigastric pain, jaundice, or both; but nitely in this dog. Postoperative chemotherapy is es-
these occur in only about one-third of patients. Rou- sential for the treatment of parasite tissue not removed
tine laboratory tests normally do not yield specific at the time of surgery. In human patients, recurrences
findings. Jaundice might be seen in patients with in- of AE were observed in 37% of 19 cases reported
trahepatic bile duct compression or obstruction. Hy- after discontinuation of long-term (more than two
perglobulinemia is present in most patients with AE.6 years) therapy with mebendazole.6
Diagnosis of AE in humans is based on clinical find-
ings and morphological features revealed by imaging Conclusion
techniques such as radiography, ultrasonography, Surgical resection followed by postoperative chemo-
computerized tomography, or magnetic resonance im- therapy with albendazole resulted in long-term clini-
aging. Immunodiagnostic tests are important for the cal remission of AE in a dog. In endemic areas, AE
etiologic diagnosis of AE. The test sensitivity is about should be included in the differential diagnosis of
95% to 100%, and specificity also is high if purified hepatic masses, especially if concomitant hyperglobu-
and specific antigens are used. 3,15 linemia is present.
The choice of treatment modalities depends on the
stage of the disease at the time of diagnosis. Surgical a
Em2 plus ELISA; Bordier Affinity Products, CH-1023 Crissier, Switzer-
resection of the involved liver segment and of land
metacestode lesions from other affected organs is in- b
Valbazen; SmithKline Beecham Animal Health, Great Britain
dicated in all operable cases. 6,7 Due to the tumor-like c
Droncit; Bayer AG, Leverkusen, Germany
proliferation, even after macroscopically radical sur-
gery, it is difficult or impossible to predict whether or
not surgery successfully removed all metacestode tis- References
sue. Therefore, postoperative chemotherapy with 1. Chitwood M, Lichtenfels JR. Identification of parasitic metazoa in
tissue sections. Experimental Parasitol 1972;32:407–519.
mebendazole or albendazole for at least two years 2. Deplazes P, Gottstein B. A monoclonal antibody against Echinococcus
after radical surgery is recommended with careful multilocularis Em2 antigen. Parasitol 1991;103:41–9.
monitoring of the patient over a minimum of 10 years 3. Gottstein B, Jacquier P, Bresson-Hadni S, Eckert J. Improved primary
immunodiagnosis of alveolar echinococcosis in humans by an enzyme-
for possible recurrence. 6,7 After nonradical resection, linked immunosorbent assay using the Em2 plus antigen. J Clin
or if inoperable, patients require continuous chemo- Microbiol 1993;31:373–6.
therapy for many years as drug treatment normally 4. Deplazes P, Eckert J. Diagnosis of the Echinococcus multilocularis
infection in foxes and other final hosts. Appl Parasitol 1996;37:245–52.
does not kill the parasite but inhibits its proliferation.
5. Eckert J, Ewald D, Siegenthaler M, Brossard M, Zanoni RG, Kappeler A.
Alveolar echinococcosis in the dog presented here Der “Kleine Fuchsbandwurm” (Echinococcus multilocularis) in der
exhibited several common features with human AE, Schweiz: Epidemiologische Situation bei Füchsen und Bedeutung für
den menschen. Bulletin, Bundesamt fuer Gesundheitswesen, Bern, Nr 25
notably very similar morphological findings in ab- 1993;S:468–76.
dominal images, hypergammaglobulinemia, and 6. Ammann RW, Eckert J. Cestodes. Echinococcus. In: Weinstock JV, ed.
Parasitic diseases of the liver and intestines. Gastroent Clin N Am,
pathomorphological lesions. The lack of cholestatic 1996:655–89.
jaundice or other obvious signs of illness most likely 7. World Health Organization guidelines for treatment of cystic and alveo-
contributed to this dog’s advanced stage at presenta- lar echinococcosis. WHO informal working group on echinococcosis.
Bull Wrld Hlth Org 1996;74:231–42.
tion. Two cases of AE described in Germany were
8. Rausch RL. Life cycle patterns and geographic distribution of Echino-
presented for clinical examination for suspected pyo- coccus species. In: Thompson RCA, Lymbery AJ, eds. Echinococcus
metra and abdominal tumor, respectively. Exploratory and hydatid disease. Oxon: CAB International, 1995:88–134.
laparotomy revealed large metacestode masses in the (Continued on next page)
314 JOURNAL of the American Animal Hospital Association July/August 1998, Vol. 34

References (cont’d) 12. Coman BJ, Rickard MD. The location of Taenia pisiformis, Taenia ovis
and Taenia hydatigena in the gut of the dog and its effect on net
9. Schantz PM, Chai J, Craig PS, et al. Epidemiology and control of hydatid environmental contamination with ova. Z Parasitenkd 1975;47:237–48.
disease. In: Thompson RCA, Lymbery AJ, eds. Echinococcus and hy-
13. Gasser RB, Lightowlers MW, Obendorf DL, Jenkins DJ, Rickard MD.
datid disease. Oxon: CAB International, 1995:233–331.
Evaluation of a serological test system for the diagnosis of natural
10. Eckert J. Der “gefährliche Fuchsbandwurm” (Echinococcus Echinococcus granulosus infection in dogs using E. granulosus
multilocularis) und die alveoläre Echinokokkose des Menschen in protoscolex and oncosphere antigens. Aust Vet J 1988;65:369–73.
Mitteleuropa. Berl Münch Tierärztl Wschr 1996;109:202–10.
14. Mathis A, Deplazes P, Eckert J. An improved test system for PCR-based
11. Geisel O, Barutzki D, Minkus G, Hermanns W, Löscher T. Hunde als specific detection of Echinococcus multilocularis eggs. J Helminthol
Finnenträger (Intermediärwirt) von Echinococcus multilocularis. 1996;70:219–22.
Kleintierpraxis 1990;35:275–80.
15. Gottstein B. Echinococcus multilocularis infection: immunology and
immunodiagnosis. Advances in Parasitology 1992;31:321–80.
Uroperitoneum in Cats:
26 Cases (1986–1995)
Uroperitoneum (UP) was diagnosed in 26 cats. Trauma was the most common
cause (84.6%), including blunt abdominal trauma (59.1%), urethral catheterization
(31.8%), and bladder expression (9.1%). The bladder was the most frequent site of
urine leakage following blunt abdominal trauma (84.6%), while the urethra was the
most common site following catheterization (71.4%). Common historical complaints
were anuria (53.8%) and vomiting (50%).
On physical examination, the bladders were palpable in nine (69%) of 13 cases;
four of the nine had ruptured bladders. The ability to urinate did not exclude a
diagnosis of UP since four noncatheterized cases reportedly urinated. Twenty-five
cases were azotemic on presentation. The creatinine or potassium (K+)
concentration in the serum compared to that in the peritoneal effusion (mean ratio,
1:2 and 1:1.9, respectively) was a useful indicator for UP. When performed, positive
contrast radiography was diagnostic. Drainage of urine from the peritoneal cavity
appeared to improve patient stabilization. Morbidity and mortality depended largely
on the severity of associated injuries. J Am Anim Hosp Assoc 1998;34:315–24.

Marcel Aumann, Dr. med. vet. Introduction


Leila T. Worth, VMD, PhD Uroperitoneum (UP) is defined as an accumulation of urine in the
peritoneal cavity following leakage from the kidneys, ureter, bladder,
Kenneth J. Drobatz, DVM or urethra. The etiology of UP differs with species, gender, and age.1
As in humans, UP in dogs and cats most commonly results from
trauma to the abdomen or pelvis. 2–8 Between 39% and 46.2% of dogs
RS with pelvic fractures had associated urinary tract injuries. 5,6 Urinary
tract injuries were diagnosed in only 0.5% of cats with pelvic frac-
tures. 7 Less common etiologies of UP include transitional cell carci-
noma of the bladder in dogs 9 and urethral calculi in cats. 10
A mortality rate of 11% to 44% is reported in humans with UP3 and
a mortality rate of 42.3% is reported in dogs with UP. 5 Although
associated injuries are responsible for most deaths, a delay in diagno-
sis of UP increases mortality rate.5,11 The diagnosis of UP represents
a challenge, 5 but a tentative diagnosis often can be made on the basis
of history, physical examination, clinicopathological findings, and
abdominal radiography. 6 However, radiographic contrast studies, ex-
ploratory surgery, or necropsy are necessary to confirm the diagnosis
and locate the source of urine leakage. 5,6
Treatment for UP depends largely on the etiology as well as the
From Veterinary Centers of America,
extent and location of urine leakage. 6,10,12 Although surgical treat-
Veterinary Referral Associates (Aumann), ment is most common, 4,12 conservative therapy has been reported for
15021 Dufief Mill Road, ruptured bladder and urethral tears in humans and animals.1,3,10,13,14
Gaithersburg, Maryland 20878 and the Despite the reports of UP following abdominal trauma 7 and ure-
Department of Clinical Studies thral calculi 10 in cats, little is known about this condition in cats. The
(Worth, Drobatz),
School of Veterinary Medicine,
goals of this retrospective study were to identify typical clinical
University of Pennsylvania, signs, clinicopathological findings, etiologies, and methods of diag-
Philadelphia, Pennsylvania 19104–6010. nosis and to assess treatment and outcome in cats with UP.

JOURNAL of the American Animal Hospital Association 315


316 JOURNAL of the American Animal Hospital Association July/August 1998, Vol. 34

The causes for UP were divided into trauma


Table 1 (including blunt abdominal trauma [i.e., hit by car,
Historical Complaints in 26 Cats unknown trauma]), catheterization, and bladder ex-
with Uroperitoneum pression; urethral obstruction; and neoplasia. If avail-
able, the site of urine leakage was recorded and
correlated with the etiology of UP.
Historical Complaint No. of Cats
Therapy and case management were at the discre-
Anuria 14 tion of the clinician. Drugs, fluid therapy, peritoneal
Vomiting 13 drainage or dialysis, urinary catheterization, and sur-
Lethargy 12 gery were recorded for each case. Urine output was
Trauma 10 recorded in each case that had an indwelling urinary
Anorexia 10 catheter placed. The ability to urinate was determined
Dysuria 8 from the medical record progress notes for the re-
Lameness 4 maining cases. Surgical findings were compared to
Pain 3 those obtained by radiology, at necropsy, or both.
Shock 3
Outcomes were recorded and, where applicable, gross
pathological and histological findings were reported.
Coma 2
The Mann-Whitney test was used to describe dif-
Diarrhea 2
ferences in median hospitalization time for cases with
Hematuria 1 UP with or without associated injuries. The chi-square
Hematemesis 1 (χ 2) test was used to assess the differences between
Renal calculi 1 males and females with UP or bladder rupture, re-
Urethral tear 1 spectively. A p of less than 0.05 was considered sig-
Polydipsia 1 nificant. Analyses were performed using a statistical
software program. b Data is presented as mean±
standard deviation (SD) or median.

Results
Materials and Methods Thirty-four cases were diagnosed with UP during the
Medical records of all cats diagnosed with UP at the study period. Eight of the 34 cases were excluded
Veterinary Hospital of the University of Pennsylva- because the diagnosis was not confirmed via radio-
nia between April 1986 and December 1995 were graphic contrast study, exploratory surgery, or
reviewed. Cases were included in the study if the necropsy, leaving 26 cases for evaluation. The num-
diagnosis was confirmed by radiographic contrast ber of cases for each parameter assessed was variable
study of the urinary tract, exploratory surgery, or because of differences in case management. A defini-
necropsy, or a combination. tive diagnosis of urinary tract leakage was made in 12
The medical records were reviewed for signalment, (46.2%) cases through radiographic contrast studies,
historical and clinical signs, and clinicopathological in 12 (46.2%) cases at surgery, and in two (7.6%)
values at admission including packed cell volume cases at necropsy. The median time period from ad-
(PCV); white blood cell (WBC) count; serum concen- mission to final diagnosis was 13 hours (range, one to
trations of urea nitrogen, a creatinine, phosphorus, so- 62 hrs; n=17).
dium (Na +), potassium (K +), chloride (Cl -), and total Breeds included domestic shorthair (n=23), domes-
protein; and urinalysis results. When available, se- tic longhair (n=1), Siamese (n=1), and Russian blue
rum concentrations of urea nitrogen and creatinine (n=1). There were 13 castrated males, eight intact
that were collected after peritoneal drainage or dialy- males, three spayed females, and two intact females.
sis and again after surgery or before discharge also Males were diagnosed with UP significantly more
were reviewed. The macroscopic appearance and cy- often than females (p less than 0.02). Median age was
tology of the peritoneal effusion were evaluated. The 2.1 years (range, five months to 17.5 years; n=25).
concentrations of total solids, urea nitrogen, creati- Historical complaints are summarized in Table 1. Me-
nine, and K + in the peritoneal effusion also were dian duration of illness prior to presentation was 24
evaluated. Values were compared to those of serum. hours (range, one to 240 hrs; n=16).
The sources of the samples and the bacterial isolates Upon physical examination, median rectal tempera-
were recorded in those cases that had microbiological ture on admission was 36.2˚ C (range, 33.8˚ C to 39.7˚ C;
cultures performed. Radiographic findings, including n=22). Five of 23 cases were bradycardic (heart rate
contrast studies, were reviewed by a radiologist less than 140 beats per min), three of which were
(Worth). hyperkalemic (serum K +, greater than 5.1 mEq/L).
July/August 1998, Vol. 34 Uroperitoneum 317

Blood urea nitrogen (BUN) and creatinine concen-


Table 2 trations were recorded in 17 cases after surgery or
Physical Examination Findings before discharge. Three cases remained azotemic af-
in 26 Cats with Uroperitoneum ter surgery, two of which had the highest creatinine
values (22.1 mg/dl and 20.0 mg/dl, respectively) on
presentation. The time required for resolution of
Physical Examination Findings No. of Cats
azotemia tended to be longer in cases with multiple
Dehydration 21 injuries unrelated to the urinary tract (median, 58 hrs;
Altered mentation (i.e., depressed, 18 range, nine to 96 hrs; n=7) compared to cases with
stuporous, comatose) urinary tract injuries only (median, 16.5 hrs; range,
Hypothermia (<37.8˚ C) 13 nine to 72 hrs; n=6).
Tachypnea (>40 breaths/min) 12 Gross hematuria was noted in six of the 26 cases.
Poor perfusion 11 Urinalysis was recorded in seven cases, and findings
Distended abdomen 9 included proteinuria (n=7), hematuria (n=7), glucos-
Bladder palpable 9 uria (n=1), and bilirubinuria (n=1). Urine specific
Pain on abdominal palpation 8 gravity ranged from 1.015 to 1.024. Urine sediment
Signs of recent trauma (excluding 6 analysis was evaluated in three cases and was charac-
pelvic fractures) terized by numerous red blood cells (RBCs) (n=2)
Bradycardia (<140 beats/min) 5 and tubular casts (n=1).
Bladder not palpable 4 Abdominocentesis was performed in 14 cases, and
Fluid wave on abdominal palpation 3 fluid was obtained in 13 cases. The fluid was de-
Pelvic fractures 3
scribed as serosanguineous in all samples for which it
was recorded (n=8). Values for total solids (TS) (n=4)
Tachycardia (>220 beats/min) 2
were less than 2.5 g/dl in three cases and greater than
Heart murmur 1
2.5 g/dl in one case. In the latter case, toxic and
degenerative neutrophils with intracellular cocci and
Two cases were tachycardic (heart rate greater than rods were seen upon cytology of the fluid. The mean
220 beats per min). Abdomens were distended in nine creatinine concentration of the abdominal fluid±SD
cases, and fluid waves were palpable in three cases. was 12.5±4.9 mg/dl (n=5), and the mean K+ concen-
Eight cases were painful on abdominal palpation. tration±SD was 9.3±7.4 mEq/L (n=5). In four cases,
Signs of recent trauma, most commonly associated the mean serum creatinine-to-abdominal fluid creati-
with the caudal abdomen and rear limbs, were noted nine ratio was 1:2 (range, 1:1.1 to 1:4.1), and the
in six of the 26 cases, three of which had pelvic corresponding mean K + ratio was 1:1.9 (range, 1:1.2
fractures. Bladders were palpable in nine of 13 cases, to 1:2.4). Both the creatinine ratio and the K + ratio
not palpable in four of these 13 cases, and not as- were reversed in the remaining case (1:0.8 and 1:0.9,
sessed in the remaining 13 cases. Other abnormalities respectively). Abdominal fluid urea nitrogen concen-
detected on physical examination are summarized in trations a (n=5) were higher than the serum concentra-
Table 2. tions in four cases and were equal to the serum
Clinicopathological findings are summarized in concentration in the remaining case.
Table 3. All except one case were azotemic on admis- Three of five aerobic bacteriological cultures ob-
sion. Peritoneal drainage only was utilized in four tained from the peritoneum (n=4) and the bladder (n=1)
cases, and serum creatinine decreased from greater at surgery were positive. Organisms isolated included
than 14 mg/dl to 10.4 mg/dl in one case, increased Enterococcus spp. (n=3), Staphylococcus epidermidis
from 3.0 mg/dl to 3.6 mg/dl in another case, and was (n=1), and alpha-Streptococcus (n=1).
not assessed in the remaining two cases. Serum crea- Table 4 summarizes the abdominal radiographic
tinine decreased from 13 mg/dl to 10.4 mg/dl, 22.1 findings. Radiographic contrast studies were per-
mg/dl to 7.7 mg/dl, and 5.1 mg/dl to 0.6 mg/dl, re- formed on 12 cases and included eight positive con-
spectively, in the three cases in which peritoneal di- trast cystograms, seven retrograde urethrograms, and
alysis was performed. Thirteen of 24 cases were three excretory urograms. When performed, radio-
hyperkalemic (serum K +, greater than 5.1 mEq/L), graphic contrast studies were diagnostic in all cases.
and one case was hypokalemic (serum K +, less than Cystograms demonstrated ruptured bladders in five
3.5 mEq/L) on presentation. Thirteen of 15 cases were cases, while urethrograms identified five urethral tears
hyperphosphatemic (phosphorus, greater than 6.0 mg/ and one bladder avulsion. An excretory urogram re-
dl). Hyperkalemia and hyperphosphatemia resolved vealed urine leakage in a case with chronic renal
following surgery and before discharge in those cases failure and UP postnephrotomy, but it failed to differ-
for which K + and phosphorus levels were recorded. entiate between ureteral and bladder leakage.
318 JOURNAL of the American Animal Hospital Association July/August 1998, Vol. 34

Table 3
Clinicopathological Findings in 26 Cats with Uroperitoneum

Percentage (Number Percentage (Number


of Cats Below of Cats Above Number Reference
Mean±SD* Range Reference Range) Reference Range) of Cats Range
Packed cell volume (%) 41.6±10.3 15.0–56.0 3.8 (n=1) 30.8 (n=8) 26 24–45
Total solids (g/dl) 8.1±1.3 5.5–10.5 — 80.8 (n=21) 26 5.5–7.1
White blood cell 16,980±7,800 4,200–29,300 6.3 (n=1) 31.3 (n=5) 16 5,500–19,500
count (cells/µl)
Sodium (mEq/L) 147.3±9.5 125.0–170.0 45.8 (n=11) 12.5 (n=3) 24 148–157
Chloride (mEq/L) 116.1±7.1 104.0–125.0 50.0 (n=6) 8.3 (n=1) 12 115–128
Potassium (mEq/L)
Admission 5.8±2.1 3.2–9.8 4.2 (n=1) 54.2 (n=13) 24 3.5–5.1
Abdominal fluid 9.3±7.4 4.1–22.2 — — 5
Phosphorus (mg/dl) 10.4±5.6 3.6–26.5 — 86.7 (n=13) 15 2.5–6.0
Urea nitrogen (mg/dl)
Admission 115.3±77.1 19.0–318.0 — 93.3 (n=14) 15 15–29
Postsurgery/ 20.3±16.0 5.0–56.0 — — 14 —
predischarge
Creatinine (mg/dl)
Admission 8.0±5.7 1.0–22.1 — 95 (n=19) 20 0.5–2.0
Abdominal fluid 12.5±4.9 5.5–15.6 — — 5 —
Postdrainage/ 6.6±4.4 0.6–10.6 — — 5 —
dialysis
Postsurgery/ 1.8±2.2 0.6–9.5 — — 15 —
predischarge

* SD=standard deviation

Exploratory laparotomies were performed in 19 fluid accumulation (n=2) and acute peritonitis (n=1).
cases. Urinary tract defects were diagnosed before Necropsies were performed on three of the five cases
surgery using contrast radiography or were suspected after cardiac arrest postsurgery.
based on clinical signs and clinicopathological and Table 5 summarizes the underlying causes for UP
radiographic findings in 17 cases. Ruptured bladders and the sites of urine leakage. Causes for UP were
were incidental findings in two cases that went to trauma (n=22), including blunt abdominal trauma
surgery for confirmed septic peritonitis and an ab- (n=13), catheterization (n=7), and bladder expression
dominal wall hernia with possible bowel rupture, re- (n=2); urethral obstruction (n=2); urinary bladder neo-
spectively. Urological abnormalities noted at surgery plasia (n=1); and nephrotomy (n=1). The bladder was
included ruptured bladder (n=14), urethral tear (n=3), the most common site of urine leakage following
avulsion of the ureter (n=1), urethral obstruction blunt abdominal trauma (n=11), and the urethra was
(n=1), uroretroperitoneum (n=1), bilateral subcapsu- the most common site following catheterization (n=5).
lar renal hematoma (n=1), bladder tumor (n=1), and All cases that sustained bladder or urethral trauma
partial-thickness bladder tear (n=1). Locations of the following catheterization were catheterized because
bladder ruptures were recorded in eight cases and of urethral obstruction. There was no significant dif-
occurred at the apex (n=3), at the trigone (n=2), dor- ference between the number of males and females
sally (n=2), and ventrally (n=1). Multiple urological with bladder rupture.
abnormalities were present in one case. Hyperemic Definitive treatment was initiated in 19 cases.
peritoneum was noted in two cases, one of which had Treatment included surgical repair (n=18) and con-
septic peritonitis. servative therapy (n=1). Conservative therapy was
Necropsy (n=5) findings pertaining to the urinary utilized in one case with a urethral tear because of a
tract included hemorrhagic cystitis (n=3), acute blad- minor urological defect and financial limitations.
der rupture (n=2), chronic bladder rupture (n=1), Medical management included intravenous admin-
chronic cystitis (n=1), and avulsion of the urethra istration of a balanced electrolyte solution (n=24),
(n=1). More than one urological abnormality was whole blood (n=5), antibiotics (n=22), corticoster-
present in one case. Other findings were peritoneal oids (n=7), opiate analgesics (n=5), low-dose infu-
July/August 1998, Vol. 34 Uroperitoneum 319

Table 4
Radiographic Findings in 26 Cats with Uroperitoneum

No. of
Radiographic Study Radiographic Findings Cats
Survey radiographs (n=18)* Decreased intraperitoneal detail 18
Bladder not visible 14
Decreased retroperitoneal detail 7
Pelvic fractures 4
Kidneys not visible 2
Abdominal hernia 1
Cranially displaced bladder 1
Renal calculi 1
Small/irregular kidneys 1
Free intra-abdominal air 1

Cystogram (n=8) Ruptured bladder 5

Retrograde urethrogram (n=7) Urethral tear 5


Avulsion of the urethra 1

Excretory urogram (n=3) Cranially displaced bladder 1


Small kidneys 1
Distended renal pelvis 1
Intraperitoneal contrast material 1

* The number in the parentheses indicates the number of cats in which the radiographic study was performed

Table 5
Etiology, Sex Distribution, Site of Urine Leakage, and Outcome in 26 Cats with Uroperitoneum

No. Females/
Etiology of Cats Males Site of Urine Leakage Outcome
Trauma 22 3/19 Bladder (n=15); urethra (n=7); Survived (n=15);
ureter (n=1) euthanized/died (n=7)
Blunt abdominal 13 3/10 Bladder (n=11); urethra (n=2); Survived (n=10);
trauma ureter (n=1) euthanized/died (n=3)
Catheterization 7 0/7 Bladder (n=2); urethra (n=5) Survived (n=4);
euthanized (n=3)
Bladder expression 2 0/2 Bladder (n=2) Survived (n=1);
euthanized (n=1)

Urethral obstruction 2 0/2 Bladder (n=2) Survived (n=1); died (n=1)

Bladder neoplasia 1 1/0 Bladder (n=1) Euthanized (n=1)

Nephrotomy 1 1/0 Undetermined Euthanized (n=1)

Total 26 5/21 Bladder (n=18); urethra (n=7); Survived (n=16);


ureter (n=1) euthanized/died (n=10)

sion of dopamine (n=4), furosemide (n=4), systemic Urinary catheters were placed in 14 cases for vari-
vasodilators (n=2), and mannitol (n=1). Cases with ous reasons, including relief of a urethral obstruction
hyperkalemia received dextrose and regular insulin (n=7), assessment of urine output (n=6), and for the
intravenously (n=5), calcium gluconate (n=2), and management of a urethral tear (n=1). All of the uri-
sodium bicarbonate (n=2). nary catheters were placed prior to surgery and re-
320 JOURNAL of the American Animal Hospital Association July/August 1998, Vol. 34

mained in place after surgery in eight cases to pro- nal trauma, traumatic catheterization of cats with ure-
mote healing of urethral tears or to assess urine out- thral obstruction, and bladder expression. The blad-
put. Urine output was described subjectively as der was the most frequent site of urine leakage
adequate in eight catheterized and noncatheterized following blunt abdominal trauma. Blunt abdominal
cases, and output (range, 3.9 to 19 ml/kg per hr) was trauma was reported to be the most common cause for
recorded in four cases. Three cases with urinary cath- bladder rupture in humans 13 and dogs.5,6 Urethral tears
eters remained anuric. Urine output was not assessed are uncommon in small animals and may result from
in 11 cases. pelvic fractures, urethral obstruction, or improper
Peritoneal drainage was performed in four cases, catheterization. 10 In five of seven cats, urethral tears
and peritoneal dialysis was performed in three cases resulted from traumatic catheterization for urethral
for a duration of five-to-16 hours. Peritoneal drainage obstruction. One case in this study had a bladder
was performed in the conservatively treated case to tumor leading to rupture. To the authors’ knowledge,
promote resolution of azotemia. In the remaining bladder tumor resulting in UP has not been reported
cases (n=6), peritoneal drainage/dialysis was insti- previously in cats, although it has been described in
tuted as a means of presurgical stabilization. the dog. 9
Surgical treatment for ruptured bladder consisted Males were diagnosed with UP significantly (p
of routine closure of the defect (n=8). Temporary less than 0.02) more often than females. Traumatic
cystostomies were performed in four cases with ure- bladder rupture is reported to occur more frequently
thral tears or disruption of the urethra before defini- in male dogs because of the limited ability of the
tive surgical repairs through prepubic urethrostomies long, narrow urethra to dilate rapidly in response to
(n=3) or routine closure of the urethra (n=1). A ne- increased intravesicular pressures. 10 Although a simi-
phrectomy was performed in the case with the avulsed lar mechanism seems likely in cats, there was no
ureter. significant difference between the frequency of blad-
Seven of 26 cases with UP were euthanized before der rupture in males and the frequency in females in
definitive treatment could be initiated, because of this study. Feline lower urinary tract disease (FLUTD)
financial concerns (n=4), severity of sustained inju- was diagnosed in 52.4% of male cats with UP. It was
ries (n=1), or severity of medical problems including the most common cause associated with urethral tears
the development of acute renal failure postnephrot- and the second most common cause associated with
omy (n=1) and an unresectable bladder tumor found bladder rupture, both of which were induced iatro-
at surgery (n=1). Three cases suffered cardiac arrest genically in the majority of these cases. Urethral ob-
seven-to-28 hours postsurgery; endomyocarditis was struction with FLUTD occurs more frequently in male
diagnosed in one case on necropsy. No underlying cats 10 and can lead to bladder distention and rupture.
cause could be identified in the other two cases. The Catheterization of the inflamed urethra in cats with
remaining 16 cases were discharged. Median hospi- FLUTD also could result in urethral tears.
talization time for cases with multiple injuries unre- Median duration of illness prior to presentation
lated to the urinary tract (231 hrs; range, 147 to 504 was 24 hours (range, one to 240 hrs). Death might
hrs; n=6) was significantly longer (p less than 0.005) occur as early as 24 hours after rupture of the bladder
than for cases with urinary tract abnormalities only in clinical cases 10 and has been reported to occur
(92 hrs; range, 60 to 360 hrs; n=10). A correlation within 47 to 90 hours in dogs with experimentally
between the time of urinary tract injury to presenta- induced bladder rupture. 5
tion and outcome could not be established because of Anuria and dysuria were by far the most common
the small number of cases (n=5) for which the time of historical complaints (in 84.6% of cases). Almost 50%
onset of clinical signs related to UP could be deter- of these cats were males with urethral obstruction.
mined with certainty. Anuria and dysuria were associated consistently with
urinary tract trauma in dogs with pelvic fractures 6
Discussion and were reported in almost 50% of the cases in
Uroperitoneum has been described previously in cats which urine leakage occurred from sites other than
as a result of pelvic fractures 7 and urethral calculi. 10 the urethra in the present study. Urine output in cath-
Although the overall incidence of UP was not re- eterized and noncatheterized cats was adequate in
ported, urinary tract injuries were diagnosed in 0.5% 46.2% of the cases, and only 21.4% of the cases
of cats with pelvic fractures. 7 In the authors’ study, remained anuric after catheterization. The authors’
UP accounted for 0.1% of the hospital feline popula- findings confirm reports in dogs and other species
tion. Trauma was the most common cause for UP in that the ability to urinate does not exclude a diagnosis
cats evaluated in this study and was reported previ- of urinary tract rupture. 1,5,6
ously as being the most common cause for UP in Depressed mentation was the most common physi-
humans 13 and dogs. 5,6 Trauma included blunt abdomi- cal examination finding and might be attributable to
July/August 1998, Vol. 34 Uroperitoneum 321

uremia, dehydration, electrolyte and acid/base abnor- unrelated to the urinary tract. Prerenal factors were
malities, pain, or trauma. Median rectal temperature responsible most often for this difference, although
was 36.2˚ C. Uremia can result in hypothermia 10 as renal factors could not be ruled out in some cases.
well as decreased rectal perfusion. The median heart Pulmonary contusions prevented aggressive fluid re-
rate was 157 beats per minute, and bradycardia oc- suscitation and correction of prerenal factors of
curred in a minority of the cats. Hyperkalemia (serum azotemia in some of the cases where UP was caused
K + range, 5.15 to 9.84 mEq/L), which was noted in by blunt abdominal trauma.
50% of the cases, was the most likely cause for the Electrolyte disturbances including hyperkalemia,
bradycardia; however, hypothermia and acid/base dis- hyponatremia, and hypochloremia have been reported
turbances could have been contributing factors. Physi- in humans, 15 dogs, 5 and foals 1 with UP. Urine is high
cal examination findings suggestive of urinary tract in K + and low in Na + and Cl -. These electrolytes
injury in this study included distended abdomen, ab- equilibrate rapidly across the peritoneum. 5 Vomiting
dominal pain, fluid wave on abdominal palpation, leads to further decreases in total body Na + and Cl -. 5
hematuria, and lack of a palpable bladder. These find- Hyperkalemia, hyponatremia, and hypochloremia
ings, however, are not specific for urinary tract in- were not consistent findings in this study as hypokale-
jury, nor does their absence rule out urinary tract mia, hypernatremia, and hyperchloremia were ob-
leakage. 6 The incidence of clinically undetected uri- served in some cats. Preexisting renal disease,
nary tract trauma in this study was 15.4%, and it had duration of illness, concomitant systemic disease, and
been reported to be as high as 20% to 50% in dogs previous treatment might influence the electrolyte
with pelvic fractures. 6 A palpable bladder does not changes found in UP.
rule out bladder rupture, because the bladder was Gross hematuria has been reported in humans and
palpable in 20% of the cats with ruptured bladders. dogs with urinary tract injury. 2,3,5,6,16 In humans, gross
Clinical and clinicopathological signs (i.e., dry hematuria was a consistent finding in patients with
mucous membranes, decreased skin turgor, increased bladder rupture. 17 In dogs with pelvic fractures, hem-
PCV and TS) of dehydration were noted in the major- aturia was the most frequent clinical sign suggestive
ity of the cats in this study. Dehydration and hemo- of urinary tract injury. 6 It occurred most often with
concentration can be related to decreased fluid intake bladder mucosal irregularities/clots but generally was
and excess fluid loss due to vomiting. The osmotic not associated with a ruptured bladder. 6 In the au-
effect of the hypertonic urine, possibly together with thors’ study, gross/microscopic hematuria was noted
peritonitis, leads to marked fluid shifts into the peri- in 42.3% of cats and was observed with avulsion of
toneal cavity. 5 Neutrophilia with a left shift occurred the ureter, bladder rupture, and urethral tears. How-
in 25% of the cases. Inflammation due to trauma or ever, 63.6% of cats with hematuria had concomitant
peritonitis is the most likely cause, although in a lower urinary tract disease. In these cases, hematuria
study of experimentally induced bladder rupture, the might have been attributable to lower urinary tract
WBC count could not be correlated with the degree of inflammation rather than injury. Hematuria was noted
peritoneal inflammation. 5 in 30% of cats with ruptured bladder unrelated to
Azotemia was noted in 96.2% of the cats at presen- lower urinary tract disease. The absence of hema-
tation. A direct correlation between the duration of turia, therefore, does not rule out severe urinary tract
UP and the severity of azotemia has been reported in injury.
dogs 5 and foals. 1 Such a correlation could not be Serosanguineous fluid was obtained via abdomi-
established in this study because the exact onset of nocentesis in 13 of 14 cases. Fluid analysis was per-
UP remained undetermined in the majority of the formed in four of these cases and suggested a modified
cases. Prerenal factors (e.g., dehydration) may con- transudate in three cases and an exudate in one case
tribute substantially to this azotemia. Blood urea ni- with a ruptured bladder tumor and septic peritonitis.
trogen, creatinine, and phosphorus concentrations The gross appearance of peritoneal effusion might be
were higher than those reported in dogs and foals misleading in the diagnosis of UP in cats, because it
with UP, 1,5 despite comparable durations of illness. does not resemble urine.
Cats with UP due to urethral obstruction and those The results of this study confirm findings in other
previously diagnosed with renal disease had the high- species that serum and abdominal fluid creatinine
est serum concentrations of BUN and creatinine. concentrations differ, and therefore they are useful in
Decreased renal function subsequent to urethral ob- the diagnosis of UP. 1,5,15 The large creatinine mol-
struction may potentiate the effects of UP. Median ecule does not diffuse readily across the peritoneal
duration for resolution of the azotemia after surgical membrane into the blood, leading to a gradient
correction of the urological defect, or institution of between creatinine concentration in serum and ab-
conservative treatment, was 36 hours. Azotemia dominal fluid. 15 Richardson suggested a serum creati-
tended to be prolonged in cats with multiple injuries nine-to-abdominal fluid creatinine ratio of 1:2 or
322 JOURNAL of the American Animal Hospital Association July/August 1998, Vol. 34

greater as diagnostic for UP in foals. 1 The mean se- Prolonged renal excretion of dye has caused renal
rum-to-abdominal fluid creatinine ratio in this study damage in humans, and excretory urography should
was 1:2 (range, 1:1.1 to 1:4.1). The corresponding K + not be performed on dehydrated animals. 18 Excretory
ratio appeared to be an equally useful indicator of urography might have contributed to the deterioration
urine leakage, with a mean value of 1:1.9 (range, of renal function in one cat that was presented with
1:1.2 to 1:2.4). However, the serum-to-abdominal acute or chronic renal failure (serum creatinine, 13
fluid creatinine and K + ratios (1:1.1 and 1:1.2, re- mg/dl), renal calculi, and UP postnephrotomy. In an-
spectively) may be very small in cats with UP. Values other cat that had an excretory urogram performed,
for the serum-to-abdominal fluid creatinine and K + serum creatinine (9.5 mg/dl) remained markedly el-
ratios in cats with peritoneal effusion of different evated 19 hours postsurgery. This cat had severe
etiology have not been reported in the literature. Ab- azotemia (serum creatinine, 20 mg/dl) after urethral
dominal fluid K + concentrations in dogs with perito- obstruction and a urethral tear. The excretory uro-
neal effusion from heart failure, septic peritonitis, grams on these two cats demonstrated poor renal func-
hemorrhage, pancreatitis, or hypoalbuminemia were tion at the time of the radiographic studies.
similar to those of serum. c The creatinine as well as Exploratory surgeries were performed in 19 cats.
the K + ratio were reversed in one cat with urethral Urinary tract defects were diagnosed before surgery
obstruction and a ruptured bladder. In this cat, samples using contrast radiography or they were suspected
for serum and abdominal fluid analysis most likely based on clinical signs and clinicopathological and
were taken at different times. The serum creatinine radiographic findings in 17 cats. Ruptured bladders
might have been measured when the cat’s urethra still were incidental findings at surgery in two cats that
was obstructed and the bladder wasn’t ruptured, yet went to surgery for confirmed septic peritonitis and
the abdominal sample was obtained after relief of the for an abdominal wall hernia with possible bowel
urethral obstruction and after dilution of the serum rupture, respectively.
creatinine with intravenous fluids. Accurate serum- In humans, bladder rupture frequently leads to
to-abdominal fluid urea nitrogen concentrations were uroretroperitoneum because the bladder is located ret-
not evaluated in this study. The small urea molecules roperitoneally. 2,3 Bladder rupture in dogs and cats
diffuse rapidly across the peritoneal membrane into usually results in UP because of the intraperitoneal
the blood, not allowing for a gradient to be estab- location of the urinary bladder in these species. 10 In
lished. 15 In dogs with UP, concentrations of urea ni- this study, blunt abdominal trauma commonly resulted
trogen in abdominal fluid were found to be equal to in bladder rupture at the apex, while defects in the
concentrations in blood. 5 Azostick readings, however, dorsal and ventral aspects of the trigone were associ-
were lower for blood than for the abdominal fluid in ated more often with urethral obstruction and trau-
this study, suggesting this parameter may be a diag- matic catheterization.
nostic tool in cats with UP. Signs of peritonitis, including hyperemic perito-
The results of this study are similar to previous neum and petechial hemorrhage in the parietal and
reports that survey radiographs may be suggestive of visceral peritoneum, were described at surgery or
urinary tract injury but are rarely diagnostic. 6 In the necropsy in three of 24 cats. Urine is irritating to the
majority of cases, survey radiographs showed a de- tissues and causes a chemical peritonitis. 10 If a uri-
crease in abdominal detail, supportive of peritoneal nary tract infection is present before urinary tract
effusion. One cat with a transected urethra had a rupture, a septic peritonitis may develop. 10 Although
cranially displaced bladder. peritonitis has been reported as a consistent finding in
Positive contrast radiography, including excretory dogs with ruptured bladders,5 the degree of peritonitis
urography, cystography, and retrograde urethrog- might differ between species, duration and amount of
raphy, was diagnostic in all cases in which it was urine leakage, and concurrent bacterial contamination.
performed, making it the method of choice for the Uroperitoneum might be a source of sepsis. 10 Bac-
diagnosis of urinary tract injuries, as previously re- teriological culture was positive in more than 50% of
ported in other species. 2,3,5,6 In humans, the accuracy the samples. Clinically normal dogs often have posi-
of retrograde cystography ranges from 85% to 100%2,3 tive urine cultures. 5 Antibiotic therapy is recom-
and depends largely on the amount of contrast me- mended in human patients with renal trauma or
dium injected. 2 False negative results may result from urethral injuries. 19 The majority of cases in this study
inadequate distension of the bladder with contrast received antibiotics. Broad-spectrum antibiotic therapy
medium. 2 In two studies in dogs, all urinary tract should be considered in cats with UP, especially in
lesions were demonstrated by contrast radiography. 5,6 severely traumatized patients.
Sites of urine leakage were identified in two of the three Fluid therapy and drainage of urine from the peri-
excretory urograms performed, while the third failed to toneal cavity appear to be the most important factors
distinguish between ureteral and bladder damage. in the stabilization of patients with UP. In patients
July/August 1998, Vol. 34 Uroperitoneum 323

with urethral disruption or obstruction, a temporary mortality rate in humans and dogs with ruptured blad-
cystostomy can be performed. 4,12 Effective urine ders. 5,11 This seems logical given the severe fluid,
drainage from the peritoneal cavity can be accom- electrolyte, and metabolic imbalances resulting from
plished with simple intra-abdominal catheter or di- UP. However, no correlation was found between the
alysis catheter placement. 12 Peritoneal dialysis in duration of illness prior to treatment and mortality.
combination with fluid therapy led to a marked de-
crease in serum creatinine in all cats in which it was a
Blood urea nitrogen or Azostix; Azostix; Miles Inc., Elkhart, IN
instituted. Although peritoneal dialysis or temporary b
GraphPad InStat; GraphPad Software, San Diego, CA
cystostomy appeared to be superior in this study, c
Personal communication, Dr. D. Hughes, Department of Clinical Studies,
adequate urine drainage likely can be achieved by School of Veterinary Medicine, University of Pennsylvania, Philadelphia,
PA
intra-abdominal catheter placement alone.
For humans, the decision between surgical or con- Acknowledgments
servative treatment for urinary tract leakage is based
on the site and the extent of the lesion as well as The authors would like to thank Dr. Chantal Parent
associated injuries. 2–4,13,17,19 Although surgical repair for her help in preparing the manuscript.
is the treatment of choice, urethral catheter drainage
alone is successful in patients with extra-peritoneal References
bladder rupture or partial urethral tears, 3,4,13,17 and it 1. Richardson DW, Kohn CW. Uroperitoneum in the foal. J Am Vet Med
seems to be an appealing option in multiple injured Assoc 1983;182:267–70.
patients. 3 Surgical treatment of urinary tract injuries 2. Carroll PR, McAninch JW. Major bladder trauma: mechanisms of injury
and a unified method of diagnosis and repair. J Urol 1984;132:254–7.
in small animals has been described in detail. 20 Con-
3. Cass AS, Luxemberg M. Features of 164 bladder ruptures. J Urol
servative management has been reported in animals 1987;138:743–5.
with bladder rupture or partial urethral tears.14,21,22 In 4. Spirnak JP. Pelvic fractures and injury to the lower urinary tract.
Surg Clin Am 1988;68:1057–69.
this study, one cat with a urethral tear was treated
5. Burrows CF, Bovee KC. Metabolic changes due to experimentally
successfully with urethral catheter drainage alone. In induced rupture of the canine bladder. J Am Vet Res 1974;35:1083–8.
humans and animals, urethral tears can be managed 6. Selcer BA. Urinary tract trauma associated with pelvic trauma. J Am
by catheter drainage for two-to-four weeks, 4,22 after Anim Hosp Assoc 1982;19:785–93.
which a urethrogram should be performed before cath- 7. Böhmer E. Beckenfrakturen und - luxationen bei der Katze in den Jahren
1975–1982. Inaugurations-Dissertation, Ludwig Maximilians-Universität
eter removal. 4 Treatment options for disruption of the München, 1985.
urethra are immediate surgical repair, primary ure- 8. Rawlings CA, Wingfield WE. Urethral reconstruction in dogs and cats.
J Am Anim Hosp Assoc 1976;12:850–60.
throplasty, or cystostomy drainage followed by
9. Grognet J. Transitional cell carcinoma and subsequent rupture of the
delayed urethroplasty. 4 In this study, temporary cys- canine bladder: a case report and review of the literature. Can Vet J
tostomies were performed in four cats with urethral 1983;24:338–40.
tears before definitive surgical repairs of the urethras. 10. Osborne CA, Low DG, Finco DR. Canine and feline urology. Philadel-
phia: WB Saunders, 1972:343–9.
Cystostomy catheter placement promotes patient sta-
11. Prather GC, Kaiser TF. The bladder in fracture of the bony pelvis:
bilization through urine drainage. 4,8 Since uremic pa- significance of a “tear drop bladder” as shown by cystogram. J Urol
tients are at high risk to develop complications during 1950;63:1019.
12. Bjorling DE. Traumatic injuries of the urogenital system. Vet Clin N Am
anesthesia, 23 surgical correction of urinary tract leak- Sm Anim Pract 1984;14:61–76.
age should be delayed until metabolic, fluid, and elec- 13. Corriere JN, Sandler CM. Management of ruptured bladder: seven years
trolyte disturbances have been corrected. 5 of experience with 111 cases. J Trauma 1986;26:830–3.
Although only 61.5% of cats with UP were dis- 14. Lavoie J-P, Harnagel SH. Nonsurgical management of ruptured urinary
bladder in a critically ill foal. J Am Vet Med Assoc 1988;192:1577–9.
charged, the prognosis appears to be good for patients
15. Sulivan MJ, Lackner LH, Banowsky LHW. Intraperitoneal extravasa-
with a treatable underlying cause and no concurrent tion of urine. J Am Med Assoc 1972;221:491–2.
injuries. Those cases (n=7) that were euthanized be- 16. Fallon B, Wendt JC, Hawtrey CE. Urological injury and assessment in
patients with fractured pelvis. J Urol 1984;131:712.
fore definitive treatment could be initiated might have
17. Palmer JK, Benson GS, Corriere JN Jr. Diagnosis and initial manage-
survived with aggressive therapy. Cases that died or ment of urological injuries associated with 200 consecutive pelvic
were euthanized after surgery generally had severe fractures. J Urol 1983;130:712–4.
concurrent injuries. In humans, concurrent injuries 18. Feeney DA, Barber DL, Osborne CA. Advances in canine excretory
urography. Proceed, 30th Gaines Veterinary Symposium. White Plains,
are responsible for most of the deaths in patients with New York, Gaines Dog Research Center; 1981:35–41.
ruptured bladders. 11 The authors noted a significant 19. Whitaker RH. Urological trauma. Ann Aca Med 1992;21:258–62.
longer median hospitalization time for cats with mul- 20. Christie BA, Bjorling DE. Principles of urinary tract surgery. In: Slatter
D, ed. Textbook of small animal surgery. 2nd ed. Philadelphia: WB
tiple injuries unrelated to the urinary tract compared Saunders, 1993:1415–73.
to those with urinary tract abnormalities only. It is 21. Osborne CA, Polzin DJ, Feeney DA, et al. The urinary system: patho-
possible that patients that did not survive after sur- physiology, diagnosis, treatment. In: Gourley IM, Vasseur PB, eds.
General small animal surgery. Philadelphia: JB Lippincott, 1985:
gery were stabilized inadequately before surgery. A 479–659.
delay in diagnosis and treatment greatly increases the (Continued on next page)
324 JOURNAL of the American Animal Hospital Association July/August 1998, Vol. 34

References (cont’d) 23. Deutsch S. Anesthetic management in acute and chronic renal failure.
Vet Clin N Am 1973;3:57–64.
22. Weaver RG, Schulz JW. Experimental and clinical studies of urethral
regeneration. Surg Gynecol Obstet 1962;115:729–36.
Morbidity and Mortality Associated with
Anesthetic Management in Small Animal
Veterinary Practice in Ontario
During 1993, 66 small animal practices participated in a prospective study to
evaluate the incidence and details of anesthetic-related morbidity and mortality.
Considering a total of 8,087 dogs and 8,702 cats undergoing anesthesia, the
incidences of complications were 2.1% and 1.3%, respectively. Death occurred in
0.11% and 0.1% of cases, respectively. Logistic regression models were developed
and showed that a significant odds ratio (OR) of complications in dogs was
associated with xylazine (OR, 91.5); heart rate monitoring (OR, 3.2); American
Society of Anesthesiologists (ASA) 3, 4, or 5 classification (OR, 2.5); isoflurane (OR,
2.4); butorphanol (OR, 0.35); technician presence (OR, 0.26); acepromazine (OR,
0.24); ketamine (OR, 0.21); and mask induction (OR, 0.2). Complications in cats
were associated with ASA 3, 4, or 5 classification (OR, 5.3); diazepam (OR, 4.1);
intubation (OR, 1.7); butorphanol (OR, 0.45); and ketamine (OR, 0.17). Cardiac
arrest in dogs was associated with xylazine (OR, 43.6) and ASA 3, 4, or 5
classification (OR, 7.1). Cardiac arrest in cats was associated with ASA 3, 4, or 5
classification (OR, 21.6) and technician presence (OR, 0.19). This paper reports the
incidences of complications and cardiac arrest in small animal practice and identifies
common complications and factors that may influence anesthetic morbidity and
mortality. This information may be useful in comparing anesthetic management
practices. J Am Anim Hosp Assoc 1998;34:325–35.

Doris H. Dyson, DVM, DVSc, Introduction


Diplomate ACVA
The death rate related to general anesthesia in humans has been
M. Grant Maxie, DVM, PhD, estimated to be as low as 1:10,000.1–3 This may be a significant
Diplomate ACVP underestimation of the risk, as several studies (even from the 1990s)
indicate mortality rates as high as 0.5 to 1.5 per 1,000.4–6 Human
Dan Schnurr, MSc anesthesia is performed by specially trained practitioners using moni-
toring techniques able to detect low blood pressure, poor oxygen-
ation, arrhythmias, esophageal intubation, hypoventilation, and
O accurate end-tidal inhalant concentration. Veterinarians perform an-
esthesia with much less information, and their patients often are
monitored intermittently by technicians with limited training in anes-
thesia. The risk of mortality associated with anesthesia in animals
may be much higher than in humans.
In a prospective study of anesthetic accidents in Great Britain in
From the Department of Clinical Studies,
the mid 1980s, one in 679 healthy dogs and cats died.7 Veterinary
Ontario Veterinary College,
University of Guelph, patients with recognized clinical disorders carried an enhanced risk
Guelph, Ontario, Canada. of death; one in 31 animals died. The only drug identified as influenc-
ing risk was xylazine. In another retrospective survey, Vermont vet-
All correspondence should be addressed erinarians indicated that mortality occurred in one in 1,000 dogs and
to Dr. Doris H. Dyson,
one in 1,500 cats. 8 No specific drug was identified as problematic.
Department of Clinical Studies,
Ontario Veterinary College, This difference in mortality may be related to the differences in
University of Guelph, anesthetic management and drug use in Great Britain and the United
Guelph, Ontario, Canada, N1G 2W1. States, or possibly it may be due to the retrospective, recollection

JOURNAL of the American Animal Hospital Association 325


326 JOURNAL of the American Animal Hospital Association July/August 1998, Vol. 34

method used in the second study. The use of The forms for special anesthetics and complica-
isoflurane would not have been considered in Great tions were standard test score forms which were
Britain at the time of the study, and isoflurane was scanned and transferred to ASCII format. The forms
increasing in popularity during the time of the contained the answers to 50 multiple choice catego-
study in Vermont. These authors felt that another ries. The first 30 categories were answered for each
prospective study with a computerized data base is standard protocol. These categories included details
needed to evaluate the risk of morbidity and mor- on preoperative management and evaluation, premedi-
tality related to anesthetic management in the cation, induction, maintenance, monitoring, and post-
1990s. operative drugs. Complications (both minor and
During 1993, veterinarians in Ontario small ani- major), including cardiac arrest, were pursued in 20
mal practices maintained records on animals anesthe- categories [Appendix 1]. Veterinarians were encour-
tized during a six-month period. These records aged to write additional details on any animal that
included details on drugs, techniques, and occurrences arrested and, when possible, postmortem examination
perceived by the veterinarian as unexpected or com- reports were to be included. Each practice also was
plicating anesthetic management. The purposes of this required to fill out a detailed form on a random num-
study were to determine the incidence and details of ber of cases which had a standard protocol assigned
anesthetic-related morbidity and mortality and to to determine the accuracy of this method of recording.
identify drugs or management practices which en- The analysis was limited to dogs and cats due to
hanced or reduced risk. the small number and great variety of other species.
The anesthetic protocols and patient data were en-
Materials and Methods tered into the computer. All scanned forms were in-
From January to July of 1993, 76 practices in Ontario cluded in the data base. Accuracy of data entry was
agreed to participate in a survey about complications assessed roughly by computing descriptive statistics
associated with anesthetic management in small ani- (including minimum and maximum values) on weight,
mals. The majority (65) of practices were selected age, and regimen and by comparing a random sample
from a randomly chosen group which proved to be of records. All statistical analyses were carried out
compliant and interested in anesthesia as determined using SAS.a The variables that were considered ini-
from a preliminary survey on anesthetic agent use. tially are listed in Appendix 2. Simple associations of
The remainder were selected randomly from the list- complications and cardiac arrests with drug use, tech-
ing of small animal practices accredited by the Col- niques, and ASA classification were evaluated using
lege of Veterinarians of Ontario. Veterinarians were chi-square (χ2 ) analysis. Separate analysis by species
asked to record on an anesthetic log the details re- was carried out due to differences noted in the χ 2 tests
lated to all small animals undergoing anesthesia dur- between dogs and cats. Logistic regression was used
ing the six-month period and the drugs and techniques to consider the influence of age and weight on these
used on each animal. A rank of health status based on variables. All variables with a p less than 0.25 ini-
the American Society of Anesthesiologists’ (ASA) tially were considered in development of logistic re-
risk categories was requested. Healthy animals were gression models to predict the risk of complications
considered to be those within ASA 1 and 2 classifica- and cardiac arrest. A combination of backward elimi-
tions (i.e., elective procedures in animals with no nation and manual comparison of possible models
known abnormalities or where the type of procedure was used to derive the final models which included
or clinical condition was considered of minor signifi- variables with a significance of p less than 0.05. Sig-
cance). Classifications of ASA 3, 4, and 5 were used nificant odds ratios (ORs) were calculated, indicative
for animals in which the clinical condition (i.e., obvious of an association more (OR greater than one) or less
disease or dysfunction diagnosed, ASA 3; signifi- (OR less than one) frequently than expected (OR
cantly compromised by the disease or condition, ASA equals one representing the overall incidence of com-
4; moribund, ASA 5) resulted in a progressive impact plications).
on anesthetic management. In an effort to simplify
the recording process, the practices listed and num- Results
bered their standard protocols and assigned the spe- Sixty-six veterinary practices (87% of those that were
cific protocol number on the anesthetic log. For contacted initially) participated in the study. Indi-
special cases failing to fit a standard protocol, practi- vidual practices reported on 25 to 841 animals. In-
tioners filled out a separate, detailed form. Whenever cluded in the survey were 16,878 animals (including
a complication was noted, the same detailed form was 8,087 dogs and 8,702 cats).
used to describe the anesthetic protocol and to ex- Duplicate, detailed forms randomly were requested
plain the problems noted. Drug doses were not con- in order to assess the accuracy of the standard proto-
sidered in this study. col assignment in the study; the forms were com-
July/August 1998, Vol. 34 Anesthesia 327

Appendix 1
Complication Guidelines
Classification of Complication
Assessment of severity Minor since easily treated
Minor since no significant consequence
Major concern created
Definition of Minor Incidents
1. Heart Rate Bradycardia (60–70 beats/min) and irregular
Bradycardia (50–70 beats/min) and regular
Normal rate but irregular
Tachycardia (clinician defined)
2. Apnea Short period causing some concern (good surgical plane)
Short period when deep plane suspected
3. Intraoperative respiratory Slow rate
concerns making stable Shallow
anesthesia difficult Panting
4. Mucous membrane color change Gray
Injected
Pale
5. Intubation difficulty Small larynx
Long soft palate
Visualization problem
Other
6. Recovery extended up to 30 minutes more than expected
7. Significant hangover
8. Excitable recovery Requiring restraint
Requiring sedation/analgesia
9. Other minor concern Explanation requested
Definition of Major Incident
1. Bradycardia <60 beats/min and irregular
<50 beats/min and regular

2. Irregular heart rate with a rapid pulse (clinician defined)


3. Cyanosis With no obvious respiratory depression
With obvious respiratory depression
4. Prolonged apnea requiring control During a surgical plane
During a deep plane
5. Collapse (loss of conciousness After premedication
or excessive sedation) After apparent recovery
6. Laryngeal problems Spasm
Edema
Bleeding/trauma
7. Impossible intubation Tracheostomy performed
Animal recovered without intubation
8. Prolonged recovery >30 min but <4 hr
>4 hr
9. Cardiac arrest Successful resuscitation
Unsuccessful resuscitation
10. Other major concern Explanation requested

pleted on 420 animals. Differences existed between premedication selected (often the use of atropine or
what the practitioner said was done on the detailed acepromazine differed) and to the technician moni-
form and the assigned protocol information in 20% of toring. The detailed forms, when available, were con-
these animals. Most of the differences related to the sidered to be the more accurate assessment except
328 JOURNAL of the American Animal Hospital Association July/August 1998, Vol. 34

In 78% of dogs, anesthesia was continued with an


Appendix 2 inhalant (38% with halothane, 36% with isoflurane,
Variables Determined From the Data Base and 4% with methoxyflurane) in 100% oxygen,
that Were Considered Initially in Analysis whereas 2% of the remaining dogs received oxygen
during injectable maintenance. The Bain circuit was
Drugs Management Details Other used in a third of the animals, and the circle system in
the remaining dogs was connected to a breathing sys-
Acepromazine Apnea monitor Age
tem. Technicians monitored dogs intermittently or
Atropine Bain circuit ASA* continuously throughout anesthesia in 86% of the
Butorphanol Blood pressure monitor (1, 2 cases. Mechanical apnea monitors were used com-
Diazepam Blood tests versus monly in dogs (35%), while some method of heart rate
Glycopyrrolate Heart monitor 3, 4, 5) detection (i.e., simple stethoscope, esophageal stetho-
Halothane Intubation Breed scope, electrocardiogram, or other) was used in 66% of
Innovar Laryngeal spray Weight the cases. Blood pressure measurement was unusual
Isoflurane Mask/tank induction (0.2%). Postoperative analgesia (primarily butorphanol)
Ketamine Oxygen administration was used in 16% of the dogs, and the use of nonste-
Local anesthetic Postoperative analgesia
roidal anti-inflammatories was rare (less than 1%).
Meperidine Stylet use at intubation Feline Anesthetic Management
Methohexital Technician present
Most cats (71%) were weighed, and preanesthetic evalu-
Methoxyflurane Weighed (versus estimate) ations were done without blood tests in the majority
Morphine (83%) of cats. An ASA classification was assigned in
Oxymorphone 99% of the cats, with 95% of the cats being ASA 1 or 2.
Nitrous oxide Acepromazine (65%), atropine (44%), and butor-
NSAIDs† phanol (37%) were used most commonly for pre-
Thiobarbiturate medication. Meperidine (19%) and glycopyrrolate
Xylazine (9%) were chosen less frequently. Combinations of
these drugs were common.
* American Society of Anesthesiologists’ (ASA) Anesthesia was induced with ketamine or ketamine
classification of health status combinations (including diazepam/ketamine [11%]

Nonsteroidal, anti-inflammatory drugs and xylazine/ketamine [7%]) in 47% of cats, inhalant
by mask in 27% of cats, and thiobarbiturate in 23% of
cats. Some incomplete records existed. Veterinarians
when related to technician monitoring. It was as- intubated 42% of anesthetized cats, with some form
sumed that technician monitoring was limited or ab- of stylet used during tube placement in 24% of these
sent if so indicated on either form. cats. Lidocaine spray or drops were applied to the
larynx in 90% of these animals.
Canine Anesthetic Management Maintenance of anesthesia was extended or supple-
Anesthetic management in dogs varied. Dogs (82%) mented by an inhalant anesthetic (28% with halothane
were weighed, and 71% were evaluated without in- and 27% with isoflurane) in 100% oxygen in 55% of
formation from blood tests. American Society of An- the anesthetized animals, whereas only 3% of cats
esthesiologists’ classifications were recorded for 99% during injectable maintenance received oxygen
of the cases (92% were ASA 1 or 2). supplementation. Almost all of the cats (98%) that
Acepromazine (68%), atropine (59%), and butor- received oxygen were connected to a Bain circuit.
phanol (37%) were the most common preanesthetic Technicians monitored cats at least intermittently dur-
medications. Meperidine (28%) and glycopyrrolate ing anesthesia in 83% of the cases. Apnea monitors
(10%) were selected occasionally. Combinations of these were used occasionally (23%), while some type of
drugs were common. Xylazine use was uncommon (1%). heart rate monitor was common (65%). Blood pres-
Anesthesia was induced with thiobarbiturates sure detection by means other than manual palpation
(52%), diazepam/ketamine (26%), and inhalant by of pulses was rare (0.1%). Postoperative analgesia
mask (11%). The remainder of recorded inductions consisted of opioids (primarily butorphanol) in 21%
consisted of fentanyl/droperidol, methohexital, and of cats and nonsteroidal anti-inflammatories in 4%.
xylazine/ketamine (less than 0.2%). Intubation was
common (77%), lidocaine laryngeal spray was used Anesthetic Complications
occasionally (12%), and use of a stylet to stiffen the The incidence of anesthetic complications was greater
tube was rare (less than 2%). in dogs than in cats (2.1% and 1.3%, respectively).
July/August 1998, Vol. 34 Anesthesia 329

Table 1 Table 2
The Incidence of Complications Related to The Frequency of Surgeries Performed
Anesthesia According to Breed* According to Species

Number Incidence Frequency


Species Breed Affected (%) Species Surgery (%)
Canine Airedale terrier 1 4 Canine Ovariohysterectomy 24*
Bulldog 3 25 Castration 20
Bull terrier 1 5 Dentistry 19
Boston terrier 1 5 Orthopedics 3
Bouvier des Flandres 1 4
Feline Ovariohysterectomy 34
Boxer 4 6
Castration 33†
Brachycephalic breeds† 12 4
Declaw‡ 27
Brittany 1 7
Dentistry 11
Fox terrier 1 8
Orthopedics 0.5
German shorthaired 1 5
pointer
* A significantly higher level of anesthetic complica-
Jack Russell terrier 5 10 tions (p of 0.016) than overall rate
Mastiff 1 7 †
A significantly lower level of anesthetic complications
Pomeranian 3 4 (p of 0.001) than overall rate

Terrier (all breeds)‡ 17 4 Ovariohysterectomy or castration also was
Shih tzu 8 5 performed in almost all cases
Weimaraner 2 25
West Highland white 4 8 vided in approximately 60% of the canine cases and
terrier
in 40% of the feline cases in which complications
were noted on the anesthetic log. The problem was
Feline Himalayan 6 4
ranked as serious about 22% of the time in each
* Where the incidence within the breed was greater species. The types of problems reported were very
than 3.5% similar in both species.

Includes bulldog, Boston terrier, chow chow, The most commonly reported complication was
Pekingese, pug, Chinese shar pei, shih tzu related to respiration. Respiratory depression or ap-

Includes all unspecified terriers, Airedale terrier, nea was noted intraoperatively in 43 dogs and 30 cats,
Australian terrier, Bedlington terrier, border terrier,
while five cats showed signs of respiratory distress
Boston terrier, bull terrier, Cairn terrier, fox terrier,
Jack Russell terrier, Kerry blue terrier, Scottish postoperatively. Difficulty with intubation was re-
terrier, Sealyham terrier, soft-coated wheaten terrier, corded in seven dogs. Eleven cats either were diffi-
West Highland white terrier, Yorkshire terrier cult to intubate or demonstrated laryngeal spasm or
edema. One endotracheal tube in a cat was plugged
with mucus. One dog and one cat aspirated their re-
The specific breeds of dogs and cats with greater than spective endotracheal tubes in recovery; the cat re-
3.5% incidence of complications are listed in Table 1. quired bronchoscopy as a result.
The most commonly performed surgeries are listed Cardiac arrhythmias also were fairly common.
in Table 2 as well as the incidence of major orthope- Tachycardia was recorded in five dogs and eight cats.
dic procedures (i.e., fracture fixation, cruciate liga- Bradycardia was found in 50 dogs compared to 12
ment repair, arthroscopy, excision arthroplasty, and cats. A greater percentage of bradycardic cats had
amputation). The surgery performed was specified in irregular rhythms accompanying the slow heart rate
94% of the cases. Chi-square analysis on the most (n=4), while only five dogs demonstrated irregular
common surgeries showed a significant difference in rhythms.
complications during canine ovariohysterectomies Undesirable recoveries were observed in many ani-
(higher, p of 0.016) and feline castrations (lower, p of mals. Recovery times were much longer than expected
0.001). in 12 dogs and 16 cats. Five dogs and seven cats
The drugs and techniques which were associated exhibited excitement at recovery which required re-
with anesthetic-related complications are listed in straint, sedation, or analgesia. One dog collapsed in
Table 3. Explanations on computer forms were pro- recovery. Other problems during recovery in cats in-
330 JOURNAL of the American Animal Hospital Association July/August 1998, Vol. 34

Table 3
Multiple Logistic Regression Models Related to Complications and Cardiac Arrest

Odds Standard p
Species Consequence Factors Ratio* Error Value
Canine Complication Xylazine 91.51 0.35 0.00
Heart monitor 3.21 0.29 0.00
ASA† 3, 4, or 5 2.51 0.26 0.00
Isoflurane 2.44 0.22 0.00
Butorphanol 0.35 0.26 0.00
Technician present 0.26 0.32 0.00
Acepromazine 0.24 0.21 0.00
Ketamine 0.21 0.26 0.00
Mask induction 0.20 0.48 0.00
Cardiac arrest Xylazine 43.61 0.83 0.00
ASA 3, 4, or 5 7.08 0.72 0.01

Feline Complication ASA 3, 4, or 5 5.28 0.23 0.00


Diazepam 4.08 0.44 0.00
Intubation 1.69 0.22 0.02
Butorphanol 0.45 0.22 0.00
Ketamine 0.17 0.37 0.00
Cardiac arrest ASA 3, 4, or 5 21.60 0.71 0.00
Technician present 0.19 0.27 0.02

* Association with lower incidence of consequence if less than 1 and higher incidence if greater than 1

American Society of Anesthesiologists’ (ASA) classification of health status

cluded prolonged hypothermia (temperature less than with cardiomyopathy, another with laryngospasm, and
37˚ C), collapse (reduced consciousness following the third cat was suspected to have died as a result of
normal recovery), and renal failure. the duration of the urethral blockage and the accom-
Other recorded problems were rare. Subcutaneous panying high potassium and metabolic acidosis.
thiopental injections were reported in four dogs and Veterinarians blamed the anatomical airway confor-
four cats. Urticaria was present in two dogs. Two mations of two dogs for the arrests, and in one of
dogs vomited after xylazine administration, and one these cases, the respiratory obstruction was not diag-
seizured following administration of acepromazine, nosed until too late because of blankets placed around
butorphanol, and glycopyrrolate. Three dogs remained the dog for rewarming. Trauma-induced respiratory
in a light plane of anesthesia, while two dogs and one complications were considered to be the causes of
cat required much more than the usual dose of thio- cardiac arrest in a dog following chest trauma. Of the
pental for induction. Unusually deep anesthetic re- remaining animals, clinicians felt that deep anesthe-
sponses to standard drug doses were mentioned in the sia was responsible for the arrests in two cases and an
comments about three cats. antibiotic reaction in another. Suspected reasons for
the arrests were unavailable in 50% of the cases.
Cardiac Arrest The drugs and techniques which may have influ-
The overall incidences of cardiac arrest were similar enced the incidence of cardiac arrest are listed in
in dogs and cats (1:899 or 0.11% and 1:967 or 0.10%, Table 3.
respectively). Details about these animals are sum-
marized in Table 4. In healthy animals (ASA 1 and 2), Discussion
the incidences of death were 1:1,483 in dogs and This investigation provides practitioners with a real-
1:2,065 in cats. Only four animals were necropsied. istic estimate of the risks related to anesthesia of dogs
The postmortem examination of one dog provided no and cats. An estimate is helpful for adequately in-
explanation for the arrest. One cat was diagnosed forming and preparing clients. It also provides a mea-
Table 4
The Details on all Animals with Cardiac Arrest

Diagnosis
Age Weight (Clinician/
Species Breed (yrs) (kg) ASA* Procedure Anesthetic Routine Prearrest Complications Lived Post Mortem)
July/August 1998, Vol. 34

Canine Bulldog 1 25 — Castration Incomplete record Respiratory depression/ No Brachycephalic


cyanosis syndrome
Mixed- 4 11 4 Chest wounds Atropine/mask induction/ Excessive anesthetic No Trauma-induced
breed dog isoflurane requirements/shallow respiratory
respiration/apnea/ complications
subcutaneous emphysema
Jack Russell 3 6.5 1 Ovariohyster- Xylazine/thiopental/ Arrest in recovery No No explanation
terrier ectomy halothane
Shih tzu 5 5.5 1 Ovariohyster- Xylazine/thiopental/ Bradycardia No No explanation
ectomy halothane
Bouvier 0.1 5 1 Ear crop Acepromazine/atropine/ Collapse in recovery No Allergic reaction to
des Flandres butorphanol/thiopental antibiotic
Yorkshire terrier 0.5 1.5 1 Dentistry Acepromazine/butor- Rapid, irregular heart rate No No findings on post
phanol/atropine/ mortem
thiopental/isoflurane
Shih tzu 0.6 4 1 Castration Acepromazine/butorpha- Respiratory obstruction during No Brachycephalic
nol/ketamine recovery syndrome
Pomeranian 8 2.5 3 Dentistry Glycopyrrolate/meperi- Apnea No Deep anesthesia
dine/thiopental
Golden 12 30 4 Examination Atropine/thiopental/ Rapid, irregular heart rate/ No No explanation
retriever halothane apnea

Feline DSH† — — 2 Mass excision Incomplete record Respiratory depression/ Yes No explanation
cyanosis
DSH 8 5 3 Catheterization Mask induction/ Respiratory depression/ No High serum
for FUS‡ isoflurane apnea/cyanosis potassium/
acidosis
DSH 1 4.5 1 Castration Ketamine No warning signs No Cardiomyopathy
DSH 0.5 3 1 Ovariohyster- Acepromazine/butor- Cyanosis during recovery No Laryngospasm
ectomy/ phanol/glycopyrro-
declaw late/mask induction/
halothane
Anesthesia

(Continued on next page)


331
332

Table 4 (cont’d)
The Details on all Animals with Cardiac Arrest

Diagnosis
Age Weight (Clinician/
Species Breed (yrs) (kg) ASA* Procedure Anesthetic Routine Prearrest Complications Lived Post Mortem)
Feline DSH — 4 1 Castration Acepromazine/butor- Apnea/cyanosis No Deep anesthesia
phanol/atropine/
thiopental
DSH 4 4.5 1 Castration/ Acepromazine/butor- No warning signs No No explanation
declaw phanol/atropine/
thiopental/halothane
DSH 13 3 4 Examination Diazepam/ketamine Rapid, irregular heart No No explanation
JOURNAL of the American Animal Hospital Association

rate/cyanosis
DSH 18 2.5 5 Dentistry Acepromazine/butor- Low heart rate/cyanosis No No explanation
phanol/atropine/mask
induction/isoflurane
DSH 9 5.5 5 Ovariohyster- Acepromazine/meperidine/ No warning signs No No explanation
ectomy/ atropine/mask induction/
pyometra isoflurane

* American Society of Anesthesiologists’ (ASA) classification of health status



DSH=domestic shorthair or mixed-breed cat

FUS=urinary tract blockage due to feline urologic syndrome
July/August 1998, Vol. 34
July/August 1998, Vol. 34 Anesthesia 333

sure by which one can evaluate the quality of anes- cation incidence. Persians have significant brachyce-
thetic management within a practice. As new drugs phalic anatomy; however, an increased incidence of
and techniques become popular, comparisons can be respiratory complications was not determined.
made. Self-evaluation is important, but discipline and Possibly the risk that this study evaluates is related
quality control committees within licensing organiza- to human observational and judgment skills. The fre-
tions may turn to these studies in veterinary practice quent error determined in human anesthetic manage-
assessment. ment was inadequate monitoring of the patient.10,13
It is difficult to say exactly what is being measured Monitoring by technicians (at least intermittently) was
by these studies. The association of a drug or tech- associated with a significant difference in reported
nique with respiratory or cardiac arrest does not nec- canine complications and feline cardiac arrests. Rec-
essarily implicate cause and effect. The causes of ognizing complications other than cardiac arrest re-
anesthetic complications and mortality are quite com- quires monitoring; therefore, one might think that
plex; multiple factors often are involved. The specific more complications would be detected with vigilant
anesthetic regimen or technique employed may cause monitoring. However, complications were reduced.
death in a small percentage (12%) of cases.9 Exami- Technicians monitor depth of anesthesia and may re-
nation of anesthesia-related deaths in humans, which spond to minor changes (i.e., shallow respiration,
resulted in litigation, identified principle events (e.g., heart rate trends) before a complication arises. This
failed intubation, drug-related problems, and equip- fact emphasizes the need to promote technician train-
ment difficulties) and several contributing factors in- ing (especially in monitoring the depth of anesthesia)
cluding inadequate supervision, poor preoperative and to respect the importance of monitoring in a prac-
assessment, and failure to communicate. 10 tice. Two anesthetic deaths were attributed to deep
Preexisting clinical disease or organ dysfunction anesthesia which was not corrected or detected in
plays a significant role in morbidity and mortality. time. Use of monitors for detection of respiratory or
The authors’ study supports this observation in both heart rate appeared to have little impact on reducing
dogs and cats. Others have reported similar find- morbidity or mortality. Interpretation and assessment
ings. 4,6,7 The original intent of the ASA classification of information from monitors and other related signs
was prediction of risk. 11 The limitation of this system of anesthetic depth are required to make accurate
lies with the inability to categorize patients consis- judgments. The incidence of complications actually
tently. American Society of Anesthesiologists’ clas- increased with the use of heart rate monitors, and this
sification of 2 melds with ASA 3 when blood work may be related to improved detection. Quite likely
indicates problems not apparent clinically, and it is the monitors were not instrumental in promoting fine
obvious that all animals are not given the same work- adjustments in the anesthetic management due to the
up. Emergency anesthesia is associated with a much limited correlation between anesthetic depth and heart
greater risk than elective anesthesia, 4 probably due to rate. 14 It may be that actions were not taken until
the time limitation for stabilization of the patient. significant changes were noted that would be re-
Adequate preoperative preparation of patients may portable in this study. Four animals arrested dur-
shift the ASA classification to a less risky category. ing recovery in this study, and many recovery
Certain breeds have problems (i.e., higher inci- complications could have been prevented with en-
dence of disease or anatomical abnormalities) which hanced vigilance; in the study from Britain,
may influence anesthetic management. Breed sensi- approximately 25% of the fatalities occurred post-
tivity to anesthetics is mentioned commonly by cli- operatively. 7 Monitors are unlikely to replace the
ents and breed associations. Pekingese dogs have been technician for assessment of patients postopera-
noted to have a high incidence of mortality.7 In the tively, but they provide additional information
authors’ study, three of the nine dogs experiencing which enhances the technician’s ability to make
cardiac arrest were brachycephalic, and two of these valid judgments intraoperatively.
developed cardiac arrest secondary to obvious respi- A significant aspect of this research is the defini-
ratory obstruction. Respiratory complications were tion of drugs and techniques which could be replaced
high in anesthetized bulldogs. It is well known that or promoted to improve anesthetic outcome. The au-
brachycephalic breeds present a greater challenge for thors’ findings agree with those of Clarke and Hall7
anesthesia.12 It is surprising that Jack Russell terriers about the increased risk associated with xylazine in
had such a high incidence of complications; evalua- dogs. Xylazine was part of the protocol in 25% of the
tion of individual cases provided no clues for the dogs with reported complications. The complication
increased risk. Only the Himalayan stood out within reported most commonly with xylazine was bradycar-
the feline group for complications, and the incidence dia, as expected with this alpha-2 agonist. Two of the
was not very high. It is possible that the brachycephalic animals given xylazine, thiopental, and halothane
anatomy of some Himalayans enhanced their compli- went on to arrest. The selection of dosages, monitor-
334 JOURNAL of the American Animal Hospital Association July/August 1998, Vol. 34

ing techniques, and other circumstances could influ- complications. Clarke suspected that acepromazine
ence outcome, so it is difficult to conclude a specific was advantageous in anesthetic management. 7 The
cause; but caution in using xylazine during general authors did not show that acepromazine influenced
anesthesia is recommended. If reasonable alternatives mortality, but acepromazine may reduce the chance
exist, it would be feasible to use them. The authors of complications related to anesthesia in dogs. The
did not consider the use of xylazine for chemical antiarrhythmic effect, the reduction in minimum al-
restraint alone in this study. Xylazine was used only veolar concentration (MAC) of inhalant anesthetic, or
in association with ketamine in cats in this study and the influence on smoothing recovery could be the expla-
was found to be comparable in safety to other regi- nation. Since blood pressure was not measured in this
mens. No evaluation of the interaction of xylazine study, no comments regarding hypotension can be made.
with other general anesthetic agents in cats was possible. Ketamine was much more common in this study
No other drugs or techniques stood out as remark- than in Clarke and Hall’s study. 7 Ketamine was asso-
ably as xylazine. Isoflurane was not used in the study ciated with a lower-than-expected complication rate
done in Britain, 7 and its use was more popular in this in both dogs and cats in this study. The sympathetic
study than at the time of the Vermont study (36% stimulation associated with ketamine may reduce the
versus 13% in dogs and 27% versus 10% in cats, incidence of bradyarrhythmias and significant hy-
respectively). 8 Isoflurane was identified as a variable potension. Clarke and Hall showed an increase in
which was associated (2.6 times more often than ex- mortality in cats given ketamine with xylazine, but
pected) with anesthetic complications in dogs. The not in cats given ketamine alone.7 Myocardial dam-
most common complication reported with isoflurane age has been detected in cats that died following
was respiratory depression, assumed by the practi- anesthesia with xylazine/ketamine, which may have
tioner to be related to a brief period of deep anesthe- been related to myocardial hypoxia. 16 The use of
sia. Isoflurane also may cause respiratory depression ketamine with xylazine was lower in this study than
in patients at an appropriate anesthetic depth. Due to in the Vermont study (7% versus 25% in cats), where
the rapid changes in anesthetic depth with this rela- the combination did not appear to contribute to mor-
tively insoluble agent, it is easy to deepen patients tality. An earlier, well-publicized report in Ontario 17
quickly, and respiratory depression is expected with and the teaching strategy at the Ontario Veterinary
deeper planes of anesthesia. 15 College likely are responsible for this difference.
The OR reported with diazepam must be consid- Early work with the combination reported safety
ered in context. Diazepam always was used in associa- which was comparable to that determined in this study
tion with ketamine in cats. Examination of the incidence with xylazine/ketamine; approximately 1:2,300 healthy
of complications with this combination of drugs was cats died. 18 This similarity and the difference noted
not significantly different than the overall rate. from the study of Clarke and Hall may be related to
Intubation was associated with an increased risk of the selection and limitation of use to healthy animals
complications in cats. Ten percent of complications in this study.
were related to intubation in cats. The suggestion of Inclusion of butorphanol in the anesthetic regimen
an increase in risk with intubation is so minor that was associated with a smaller OR for complications.
intubation should not be avoided in cats if the dura- This may be related to its analgesic and sedative
tion or type of procedure warrants it. Being prepared effects which can reduce the need for deeper anesthe-
to intubate and monitoring for adequate respiration sia and can smooth recovery. Although postoperative
may be sufficient for short procedures in cats. analgesia was not noted to be influential, the presence
Duration of anesthesia was not considered in this of butorphanol as part of premedication may extend
study, but one might question its impact. The major- sufficiently into recovery, considering the short dura-
ity of cats being castrated were not intubated, while a tion of most practice surgeries.
greater percentage of cats undergoing ovariohyster- Mask induction of dogs was associated with re-
ectomy were intubated. Duration of anesthesia corre- duced complications in this study. The advantage of
lates with mortality in humans.4 A tenfold increase in this technique may relate to use in critical patients,
mortality was reported as anesthesia duration in- where cautious titration of anesthetic administration
creased from less than 30 minutes to between 30 and is required.
179 minutes. The results related to castration in cats This study on anesthetic morbidity and mortality
and intubation may be secondary to the duration of in small animals indicates that the incidence of car-
anesthesia. Longer duration of anesthesia may con- diac arrest in dogs and cats is less than reported in
found the interpretation of negative impact of other Britain 7 and not remarkably different than that re-
drugs as well. ported in Vermont 8 or in recent studies in humans. 4–6
This study identified a number of drugs that ap- One might speculate on the improvement observed
peared to be associated with a reduced incidence of with time. Within the last 15 years the authors have
July/August 1998, Vol. 34 Anesthesia 335

seen increased undergraduate veterinary education in 3. Lunn JN, Mushin WW. Mortality associated with anaesthesia.
Anaesthesia 1982;37:856.
anesthesia, more specific technician training in the 4. Pedersen T, Eliasen K, Hendrksen E. A prospective study of mortality
area of anesthetic monitoring (as well as greater tech- associated with anaesthesia and surgery: risk indicators of mortality in
nician employment), and continuing education in in- hospital. Acta Anaesthesiol Scand 1990;34:176–82.
5. Schwilk B, Friess L, Friesdorf W, Ahnefeld FW, Georgieff M. Preopera-
tensive care and anesthesia; all could be making an tive risk factors and intraoperative and postoperative risk management
impact. Veterinarians apply this knowledge intelli- in 11,890 anesthesias. Initial results of a prospective study. Anaesthesiol
Intensivmed Notfallmed Schmerzther 1993;28:484–92.
gently to select the more appropriate agents (or to
6. Forrest JB, Rehder K, Cahalan MK, Goldsmith CH. Multicenter study of
eliminate those with a greater potential for adverse general anesthesia. III. Predictors of severe perioperative adverse out-
effects in the specific individual case). Continued comes. Anesthesiol 1992;76:3–15.
advances in training could improve the statistics further. 7. Clarke KW, Hall LW. A survey of anaesthesia in small animal practice:
Assoc Vet Anaesthesiol/Brit Sm Anim Vet Assoc report. J Assoc Vet
Veterinarians must be cautious about the interpre- Anaesthesiol 1990;17:4–10.
tation of drug safety from these types of studies. The 8. Dodman NH, Lamb LA. Survey of small animal anesthetic practice in
difference shown between dogs and cats with xylazine Vermont. J Am Anim Hosp Assoc 1992;28:439–45.
9. Duncan PG. Editorial: who’s to blame? Can J Anaesthesiol 1992;
may be related to the fact that xylazine in dogs was 39:110–3.
associated with several other anesthetic agents while 10. Gannon K. Mortality associated with aneaesthesia. A case review study.
xylazine only was used in cats in association with Anaesthesia 1991;46:962–6.
ketamine in this study. Use of acepromazine in 11. Ross AF, Tinker JH. Anesthesia risk. In: Miller RD, ed. Anesthesia. 3rd
ed. New York: Churchill Livingston, 1990:715–42.
healthy, young patients may reduce the induction and 12. Short CE. Congenital problems in canine breeds. In: Short CE, ed.
maintenance dose of anesthesia and smooth recovery, Principles and practice of veterinary anesthesia. Baltimore: Williams &
but it could result in severe hypotension in very old, Wilkins, 1987:303–7.
13. Memery HN. Anesthesia mortality in private practice. A ten-year study.
debilitated, or dehydrated animals, especially if anes- J Am Med Assoc 1965;194:1185–9.
thetic dose adjustment is not made properly. 19 14. Dyson DH. Assessment of three audible monitors during hypotension in
Isoflurane use could be limited to critical patients for anesthetized dogs. Washington: Proceed, Ann Mtg Am Coll Vet
Anesthesiol, 1993:25.
the advantages of better cardiac output, reduced
15. Dohoo SE. Isoflurane as an inhalational anesthetic agent in clinical
arrhythmias, and rapid dose adjustments; 15 but it also practice. Can Vet J 1990;31:847–50.
can be used satisfactorily in any patient with proper 16. Van der Linde-Sipman JS, Hellebrekers LJ, Lagerwey E. Myocardial
monitoring and management. This study cannot pro- damage in cats that died after anaesthesia. Vet Quart 1992;15:91–4.
17. Gillick A. High-frequency complaints. Ontario Vet Assoc Newsletter
duce absolute conclusions on drug selection. Proper 1981;5:12.
preoperative stabilization, intelligent use of drugs 18. Arnbjerg J. Clinical use of ketamine-xylazine in the cat. Nord Vet Med
(dose and technique possibly being more important 1979;31:145–54.
than choice), careful observation of the patient until 19. Boyd CJ, McDonell WN, Valliant A. Comparative hemodynamic effects
of halothane and halothane-acepromazine at equipotent doses in dogs.
recovery is complete, and appropriate responses to Can J Vet Res 1991;55:107–12.
early signs of deep anesthesia, bradycardia, hypo-
ventilation, or hypotension are likely to make a more
significant difference in morbidity and mortality than
either the elimination of specific drugs or the addition
of any new agent. Whenever an anesthetic-related death
occurs, it is essential to evaluate the details to deter-
mine if there is some factor that requires improve-
ment which may help to avoid another incident.

a
Statistical Analysis Software package; SAS Institute, Inc., Cary, NC

Acknowledgments
This research was supported by a grant from Pet Trust.
The authors thank the veterinarians and techni-
cians who not only were willing to contribute a lot of
time and effort to this study but were patient with the
time it has taken for these results to be made available.

References
1. Maaloe R, Hansen CL, Pedersen T. Death under anesthesia. Definition,
causes, risk factors and prevention. Ugeskr Laeger 1995;157:6561–5.
2. Kubota Y, Toyoda Y, Kubota H, et al. Frequency of anesthetic cardiac
arrest and death in the operating room at a single general hospital over
a 30-year period. J Clin Anesthesiol 1994;6:227–8.
Early Detection of Canine Hip Dysplasia:
Comparison of Two Palpation and
Five Radiographic Methods
Hip joint laxity was evaluated in four breeds (i.e., greyhound, Labrador retriever, Irish
setter, hound mixed-breed) of puppies (n=32) by Ortolani’s and Bardens’ maneuvers,
by subjective assessment of radiographs (Orthopedic Foundation for Animals [OFA]
method), and by four radiographic measurement indices. Puppies were studied at
four, six-to-10, 16-to-18, and 52 weeks of age. The purpose of this study was to
compare palpation and radiographic methods of hip laxity detection in puppies for
predicting the development of degenerative joint disease (DJD) by one year of age.
Twenty-seven (42%) hips developed DJD. Ortolani’s method was not a reliable
predictor of hip dysplasia at six-to-10 weeks; it was significantly predictive at 16-to-
18 weeks but had a high incidence of false negatives. Bardens’ and subjective (OFA)
assessment methods were not reliable at six-to-10 or 16-to-18 weeks. Radiographic
measurements taken with femurs in a neutral position and hips distracted (distraction
index [DI] and Norberg angle) and measurements taken with femurs extended in
OFA position (Norberg angle) of six- to 10-week-old puppies accurately predicted
DJD occurrence by one year of age (p less than 0.01). Distraction index
measurement (PennHIP method) was the most accurate in predicting the
development of DJD (p less than 0.001). Distraction index radiography in puppies
six-to-10 and 16-to-18 weeks of age was the most reliable predictor of hip dysplasia.
Norberg angle measurement was more reliable during hip distraction than when hips
were measured in the OFA position in 16- to 18-week-old puppies, but had similar
reliability in six- to 10-week-old puppies. J Am Anim Hosp Assoc 1998;34:339–47.

William M. Adams, DVM Introduction


R. Tass Dueland, DVM, MS Canine hip dysplasia (CHD), first reported by Schnelle over 60 years
ago, 1 may be defined as hip laxity of genetic origin. Hip laxity
Jeff Meinen, DVM frequently results in radiographic signs of degenerative joint disease
(DJD) and clinical lameness. Canine hip dysplasia continues to plague
Robert T. O’Brien, DVM, MS many breeds despite 25 years of selective breeding based on evalua-
Elizabeth Giuliano, DVM tion of hip conformation using a standardized radiographic view
recommended by the Orthopedic Foundation for Animals (OFA). 2,3
Erik V. Nordheim, PhD Reported incidence of CHD in the four most-frequently OFA-evalu-
ated breeds (Labrador retriever, golden retriever, rottweiler, and Ger-
man shepherd dog) ranges from 14% to 23% (mean, 21%).3 The OFA
O recently reported significant reduction in the incidence of CHD in
only one (i.e., rottweiler) of these four breeds between 1972 and 1990.3
Success in reducing CHD incidence is limited by the insensitivity
and timing of detection methods which result in failure to remove

From the Department of Surgical Sciences (Adams, Dueland, Meinen, O’Brien,


Giuliano), School of Veterinary Medicine, University of Wisconsin, 2015 Linden
Drive West, Madison, Wisconsin 53706 and the Department of Statistics
(Nordheim), College of Letters and Science and Department of Forestry,
College of Agricultural and Life Sciences, University of Wisconsin, 1630 Linden
Drive, Madison, Wisconsin 53706.

JOURNAL of the American Animal Hospital Association 339


340 JOURNAL of the American Animal Hospital Association July/August 1998, Vol. 34

tions were followed by sequential hip examinations


Table 1 through 52 weeks of age (n=32). Greyhounds were
chosen because of the very low breed incidence of hip
dysplasia; the LRs intentionally were bred from par-
ents known to have hip dysplasia; the ISs were bred
from OFA-certified “good” parents; and the HM pup-
pies were from unknown parentage.
Puppies from four-to-16 weeks of age were anes-
thetized by mask induction using isoflurane for all
palpation and radiographic studies. Older puppies re-
ceived subcutaneous preanesthetic (i.e., butorphanol
[0.2 mg/kg body weight] plus acepromazine [0.1 mg/kg
body weight]) prior to mask induction. All puppies
were placed in homes for the duration of the study
and kept as pets.
Three subjective hip evaluation methods (i.e.,
Table 1—Comparison of distraction index (DI) variation over Ortolani’s maneuver,9 Bardens’ maneuver,6 and ex-
time for four breeds (greyhound, Labrador retrievers, Irish tended-hip radiography [OFA method 2]) were tested.
setters, hound mixed-breed dogs) of puppies. Indices were In addition, four radiographic measurement methods
measured from hip radiographs taken under anesthesia with the
hind limbs in a neutral position (PennHIP method). Means and (i.e., distraction index [PennHIP method4 ], Norberg
standard deviations at four-to-seven ages are shown. Note the angle with hind legs extended, 5,11 Norberg angle with
relative consistency between six-to-10 weeks of age and 52 hind legs neutral and distracted, and Norberg angle
weeks of age. Greyhounds and Labrador retrievers were mea-
sured at four weeks of age; note the much greater laxity with hind legs neutral and compressed) of hip evalua-
compared to later measurements of the same hips. Greyhounds tion were tested.
had significantly tighter hips than the other three breeds. A positive Ortolani’s maneuver consisted of a pal-
pable “clunk” when a subluxated hip was reduced
affected animals from the breeding pool. Alternative while the puppy was positioned in lateral recumbency.
methods espousing earlier and more sensitive detec- A positive Bardens’ maneuver consisted of a 2 mm or
tion of CHD have been investigated.4–7 Radiographic greater estimation of palpable hip laxity. All palpa-
detection of hip joint instability and histological evi- tion maneuvers were performed by one of the authors
dence of acetabular remodeling have been documented (Dueland). Extended-hip radiographs were analyzed
in puppies as young as seven weeks.8 Progression of subjectively by one of the authors (Adams) and a
radiographic subluxation and histological remodeling reader for the OFA, a in a blinded fashion, using crite-
of dysplastic hips by 12 weeks of age was described ria established by the OFA.10 Distraction indices (DIs)
as dramatic.8 The potential of a screening method for were measured as previously reported. 4
very early detection of CHD motivated the investiga- These examinations were performed as soon as
tion reported here. puppies were obtained, and they were repeated at six-
The purposes of this study were to compare the to-eight, 10-to-12, 16-to-18, 24-to-26, and 52 weeks
predictability of seven methods of hip laxity detec- of age for a total of three-to-seven examinations per
tion with occurrence of DJD by one year of age and to hip. Four GH puppies (eight hips) were examined
determine how early each method could be applied to only three times (at four, 26, and 52 weeks of age) as
puppies accurately as a screening process for CHD. controls to assess any deleterious effects the seven
Five of the methods have been reported by others as hip manipulation episodes might have on their litter
effective ways of detecting CHD in puppies as early mates. All other puppies were examined at six-to-10
as four months of age. 3,4,6,7 In addition, two radio- weeks of age, all were evaluated at 52 weeks of age,
graphic methods which utilized Norberg angle mea- and all but the four control GHs were examined at 16-
surements (in a distraction view and in a compression to-18 weeks of age—for a total of 348 individual hip
view) were evaluated. evaluations using the seven previously described
methods. Greyhound and LR puppy hips also were
Materials and Methods examined at four weeks of age to compare consis-
Two litters of greyhounds (GHs; n=12) and a litter of tency of measurements as the puppies matured [Table
Labrador retrievers (LRs; n=4) first were examined at 1].
four weeks; two litters of Irish setters (ISs; n=12) On the extended-hip view taken at 52 weeks, three
first were examined at six-to-seven weeks; and a lit- board-certified radiologists (Adams, LJ Forrest, and
ter of hound mixed-breed puppies (HMs; n=4) first O’Brien) and a board-certified surgeon specializing
were examined at 10 weeks of age. These examina- in orthopedics (Dueland) independently evaluated, in
July/August 1998, Vol. 34 Canine Hip Dysplasia 341

Figure 3—Irish setter #A2, 52 weeks old, has a Morgan line


Figure 1—Irish setter #A7 at 52 weeks of age has degenerative (enthesophyte) on the right femoral neck (open arrow), but has
joint disease (DJD) of both hips. Morgan lines (enthesophytes) no other stress lines and no coxofemoral subluxation. These
are present bilaterally (open arrows), and other stress lines also hips were recorded as negative for degenerative joint changes.
are present bilaterally (solid black arrows). In addition, sublux- Slight rotation of the pelvis is causing asymmetry in the appear-
ation of the left femoral head and mild remodeling of the cranial ance of the two hips. Ortolani’s sign was negative for both hips
margin of the left acetabulum are present. Ortolani’s and Bardens’ at seven and 52 weeks of age. Bardens’ sign was positive for the
signs were negative in both hips at seven and 17 weeks of age, left hip at seven weeks but was negative for both hips at 52
but the Ortolani’s sign was positive in both hips at 52 weeks. A weeks. Distraction indices were 0.31 (right hip) and 0.28 (left
positive Bardens’ sign was recorded in the right hip at 52 weeks. hip). Norberg angles were 110˚ (right hip) and 105 ˚ (left hip).
Distraction indices are 0.53 (right hip) and 0.69 (left hip).
Norberg angles are 107˚ (right hip) and 90 ˚ (left hip). eral to the femoral head (i.e., Morgan line) 12 and no
other stress lines/proliferations on the femoral neck
or acetabulum but with femoral head subluxation or
suboptimal alignment was not regarded positive for
DJD [Figure 2]. Conversely, if only a Morgan line
was evident, and no subluxation was observed on the
extended-hip view, this was not regarded as being
positive for DJD [Figure 3]. Consensus of three out of
four readers that DJD had developed was recorded as
being positive.
For statistical analysis, hips examined at six-to-10
weeks of age were grouped, those examined at 16-to-
18 weeks of age were grouped, and those examined at
50-to-52 weeks of age were grouped. The graph in
Table 1 is reported at actual age at time of measurement
for each group of puppies. Methods used to analyze data
included logistic regression and analysis of variance
(ANOVA). Significance was set as p less than 0.01.
The primary statistical tool used to determine
which of the six methods best predicted DJD was
logit analysis or logistic regression. 13 For each of 18
combinations of age groups and methods statistically
Figure 2—Labrador retriever #5, 52 weeks old, has no degen- analyzed [Table 2], a logistic regression was per-
erative joint changes. The right femoral head does not appear formed with occurrence or nonoccurrence of DJD as
well aligned to the acetabulum, but there is slight rotation of that the response variable. The p value for each analysis
femur, potentially contributing to the less than ideal appear-
ance. Ortolani’s sign was negative for both hips at eight weeks, was compared with a threshold value of 0.01 in order
but was positive at 12, 16, and 52 weeks of age. Bardens’ sign to minimize the probability of predicting a hip devel-
was negative at eight, 12, and 16 weeks, but was positive at 52 oping DJD when in fact it did not. The results [Table
weeks. Distraction indices were 0.41 (right hip) and 0.43 (left
hip). Norberg angles were 103˚ (right hip) and 103 ˚ (left hip). 2] provide the actual p value so that the reader may
compare with some alternative threshold. In addition
a blinded fashion, all hips for evidence of DJD. De- to p value, the authors calculated the percentage of
generative joint disease was defined as two or more correct prediction, number of false negatives, and num-
stress lines at the joint capsule attachments or one or ber of false positives for each of the 18 combinations.
more stress lines and femoral head malalignment (i.e., For the radiographic predictors (all measured on a
subluxation) [Figure 1]. Absence of a stress line lat- continuous scale), a hip was predicted to have DJD if
342 JOURNAL of the American Animal Hospital Association July/August 1998, Vol. 34

Table 2
Accuracy of Degenerative Joint Disease (DJD) Prediction Methods at Three Ages

Method Data Six-to-10 Wks of Age 16-to-18 Wks of Age 50-to-52 Wks of Age
Ortolani’s p value 0.9714* 0.0023 0.0001
maneuver n 51 47 59
% correct 54.9 72.3 88.1
No. of false negatives 23 10 2
No. of false positives 0 3 5

Bardens’ p value 0.9679* 0.9627* 0.0194


maneuver n 51 47 59
% correct 56.9 57.4 66.1
No. of false negatives 22 20 19
No. of false positives 0 0 1

Norberg p value 0.0022 0.0012 0.0027


angle n 51 50 59
OFA† view % correct 66.7 70 79.7
No. of false negatives 8 5 10
No. of false positives 9 10 2

Norberg p value 0.0122 0.2343 0.0252


angle n 51 50 59
compression % correct 64.7 50 64.4
view No. of false negatives 7 10 15
No. of false positives 11 15 6

Norberg p value 0.0007 0.0015 0.0002


angle n 51 50 59
distraction % correct 68.6 80 86.4
view No. of false negatives 8 3 3
No. of false positives 8 7 5

Distraction p value 0.0001 0.0002 0.0001


index n 51 50 59
% correct 78.4 82 88.1
No. of false negatives 5 2 2
No. of false positives 6 7 5

* Due to data structure, calculated p value is larger than it should be



OFA=Orthopedic Foundation for Animals

the resulting fitted logistic equation led to a probabil- incidence was still quite high in 16- to 18-week-old
ity greater than 0.5 that the hip would develop DJD. puppies. Of the four radiographic measurement meth-
The one-way ANOVA was used to compare radio- ods used at six-to-10 weeks and 16-to-18 weeks, the
graphic results at different time points. Finally, one- DI and Norberg angle measured in neutral-positioned
way ANOVA was used to compare two groups of hips during distraction were the most promising pre-
GHs (one group [i.e., the control] examined only three dictors of DJD at one year of age. A significant rela-
times and the other group examined seven times). tionship for both time points also was noted between
the Norberg angle measured in the extended-hip
Results (OFA) view to subsequent development of DJD.
Both palpation methods resulted in very high num-
bers of false-negative determinations in six- to 10- Palpation Methods
week-old puppies [Table 2]. This false-negative Ortolani’s and Bardens’ maneuvers were performed
incidence remained high in 16- to 18-week-old pup- at 340 hip evaluation times (98% of evaluations). All
pies. Similarly, radiographic subjective assessment 27 hips diagnosed as having DJD at 52 weeks did
of hip conformation (using the OFA method) in six- have Ortolani’s assessment at six-to-10, 16-to-18, and
to 10-week-old puppies resulted in a very high inci- 52 weeks; false-negative Ortolani’s signs (negative
dence of false-negative readings. This false-negative Ortolani’s sign, but did develop DJD) recorded at
July/August 1998, Vol. 34 Canine Hip Dysplasia 343

ers disagreed in 57 (21%) of 270 hip evaluations of


Table 3 immature puppies six-to-18 weeks of age. Range of
disagreement by breed was 5.6% (GH) to 38% (LR).
Comparing these results with consensus diagnosis of
DJD at 52 weeks of age, both readers had a high
incidence of false-negative assessments. Of hips
which had DJD at 52 weeks, 24 (89%) of 27 were
assessed negative for hip dysplasia at six-to-10 weeks
of age and 15 (56%) of these same 27 hips still were
assessed negative for hip dysplasia at 16-to-18
weeks of age. Of hip evaluations on six- to 10-
week-old puppies which had no DJD at 52 weeks,
three (6%) of 52 were assessed positive for hip
dysplasia (false positives). No false positives were
recorded on 16- to 18-week-old puppies.
Table 3—Distraction index (DI) comparison between grey-
hound hips and hips of three other breeds (Labrador retriever,
Irish setter, and hound mixed-breed puppies) combined. Means
Distraction Index
and standard deviations at five ages (four ages for Irish setters) The DI range for all hips, including all ages, was 0.05
are shown. Although there was a significant difference in DI of to 1.0 and varied by breed [Table 1]. From eight
the greyhounds between eight and 16 weeks, the difference in
DI for the other breeds combined and for DJD-positive puppies weeks on, all greyhounds had DIs less than 0.28,
only (not shown) was not significant between eight and 16 despite ranging from 0.36 to 0.77 at four weeks of
weeks nor between eight and 52 weeks. age. Measurements were difficult at four weeks of
age due to the large amount of unossified bone; sub-
these three ages were 85%, 37%, and 7%, respec- chondral ossification by six-to-eight weeks of age
tively. Conversely, false-positive Ortolani’s signs was adequate for measurement. Mean DIs at six-to-
were recorded in 0% (zero of 24 hips), 15% (three of eight and 52 weeks were 0.18 and 0.10, 0.64 and 0.53,
20 hips), and 16% (five of 32 hips) at these three and 0.48 and 0.46 for GHs, LRs, and ISs, respec-
ages, respectively. Only four (15%) of 27 hips which tively. Mean DIs at 10 and 52 weeks for HMs were
developed DJD had positive Ortolani’s signs on as- 0.65 and 0.62. When data was combined for the three
sessment at six-to-10 weeks of age, although no false breeds at risk for hip dysplasia and was compared to
positives were recorded at this age. As with the the data for GHs, despite large standard deviations in
Ortolani’s maneuver, results of the Bardens’ maneu- the combined breeds group, there was no overlap in
ver were recorded only as positive or negative. Find- DI range at any age between these two groupings
ings were very similar to Ortolani’s maneuver results [Table 3].
at six-to-10 weeks. False-negative Bardens’ signs re- At six-to-10 week and 16-to-18 week measure-
corded at six-to-10, 16-to-18, and 52 weeks were ments, hips with DIs of 0.31 to 0.41 comprised 19%
81%, 74%, and 70%, respectively. Only 19% of hips (five of 27 hips) and 7% (two of 27 hips), respec-
which developed DJD were Bardens’ sign positive on tively, of hips which later developed DJD. The lowest
assessment at six-to-10 weeks of age, and again, no DI in a hip which developed DJD by one year was
false positives were recorded. 0.26 (measured at 52 weeks of age). Distraction index
in this hip was 0.35 at eight weeks. A DI less than
Subjective Radiographic Assessment (Orthopedic 0.26 was measured in 24% (12 of 51 hips), 30% (15
Foundation for Animals [OFA] Method) of 50 hips), and 44% (26 of 59 hips) of all hips
Degenerative joint disease was diagnosed in 27 (42%) analyzed from six-to-10 weeks, 16-to-18 weeks, and
of 64 hips at 52 weeks of age. No DJD (0%) was one year of age, respectively.
found in the 24 GH hips. Degenerative joint disease At six-to-10 week and 16-to-18 week measure-
was diagnosed in three (38%) of eight LR hips, 16 ments, most hips (90% [nine of 10 hips] and 87% [13
(67%) of 24 IS hips, and eight (100%) of eight HM of 15 hips], respectively) with DIs greater than 0.60
hips. Five (8%) of 64 hips had abnormal conforma- developed DJD by one year. The five hips with DIs
tion (i.e., malalignment of femoral head with acetabu- greater than 0.70 at six-to-eight weeks developed DJD
lar margin), 50% or less of the femoral head beneath by one year. Five (83%) of six hips with DIs greater
the acetabulum, or both; but they had no clear evi- than 0.70 at 16-to-18 weeks developed DJD by one
dence of DJD at 52 weeks. These five were deleted year. Of all hips that developed DJD, their respective
from statistical analysis and are not included in Table DI means and medians by age grouping were 0.56 and
2. On subjective (OFA method) evaluation for hip 0.60 (range, 0.31 to 0.77) at six-to-10 weeks; 0.61
dysplasia using the extended-hip view, the two read- and 0.57 (range, 0.33 to 0.92) at 16-to-18 weeks; and
344 JOURNAL of the American Animal Hospital Association July/August 1998, Vol. 34

0.56 and 0.61 (range, 0.26 to 0.71) at 52 weeks. The Comparison of Measurement Techniques and Age
most remarkable change in DI was in one dog which For the 27 hips which developed DJD by one year, the
had two hips that were luxated (DI, 1.0) at 26 weeks; respective ranges for the three most promising mea-
but by 52 weeks, the DIs had decreased to 0.63 and surement techniques (DI, NorOFA, and NorDis) at
0.67, respectively. age six-to-10 weeks were 0.31 to 0.77, 79˚ to 99˚, and
63˚ to 84˚. Distraction index measurements varied
Norberg Angles somewhat between six-to-10 weeks and 16-to-18
weeks (p of 0.06, ANOVA), but DI measurements did
Extended-Hip (OFA) View not vary between six-to-10 weeks and 52 weeks (p of
The Norberg/extended-hip (NorOFA) angles for all 0.9, ANOVA). There was a trend toward hips becom-
breeds at six-to-52 weeks of age ranged from 74˚ to ing tighter between 16 and 52 weeks (p of 0.05,
116˚. No hip developed DJD that had a NorOFA angle ANOVA). Both NorOFA and NorDis measurements
greater than 107˚. However, none of the 51 hips mea- varied significantly between six-to-10 weeks and 52
sured at six-to-10 weeks and only three (6%) of the weeks (p of 0.0001, ANOVA). For all evaluated hips
50 hips measured at 16-to-18 weeks of age had (n=23) which had no DJD by one year, DI did not
NorOFA angles greater than 107˚. vary between eight and 16 weeks, but significant
For hips with angles less than 95˚ at six-to-10 variation occurred for NorOFA and NorDis measure-
weeks of age, 19 (68%) of 28 developed DJD. Hips ments between these ages (p less than 0.001, ANOVA).
with angles less than 95˚ at 16-to-18 weeks (n=8) and For 16 GH hips at eight weeks of age, the ranges
one year of age (n=11) all developed DJD. However, for the three measurements were DI of 0.13 to 0.27,
seven (26%) of the 27 DJD hips had angles of 105˚ to NorOFA of 90˚ to 103˚, and NorDis of 75˚ to 95˚. These
107˚ at one year of age. measurements improved significantly at 16-to-18
weeks and 52 weeks (p less than 0.001, ANOVA);
Neutral-Positioned Hips Measured During however, the most dramatic improvement occurred
Distraction (NorDis) between four and six weeks of age (p less than 0.001,
Norberg/distraction angles (NorDis) were decreased ANOVA) [Table 1].
predictably for all breeds except the GHs (range, 58˚ Eight GH hips (i.e., the controls) were evaluated
to 104˚ for the other three breeds and 75˚ to 109˚ for only at four, 24, and 52 weeks of age. Sixteen GH
GHs at six-to-52 weeks). No hip developed DJD that hips were measured at four, six, eight, 12, 16, 26, and
had a NorDis angle greater than 91˚. Fourteen percent 52 weeks of age. Comparisons of DI, NorOFA, and
of 51 hips measured at six-to-10 weeks and 41% of NorDis measurements between these two groups were
hips measured at 16-to-18 weeks of age had NorDis made to determine if repeated hip manipulation under
angles greater than 91˚. Of angles less than 85˚ at six- anesthesia had any effect on results. There were no
to-10, 16-to-18, and 52 weeks of age, 69%, 92%, and differences in the three ages at which both groups
100%, respectively, developed DJD by one year of were measured (p of 0.19 to 0.92, ANOVA).
age. However, specificity of the 85˚ cut-off varied
between age brackets; at six-to-10 weeks, none of the Comparison of Measurement Techniques and
12 hips that measured at greater than 84˚ developed Degenerative Joint Disease
DJD, whereas 15 (41%) of 37 hips and 11 (30%) of 37 At six-to-10 weeks of age, no hip with a DI value less
hips measured at greater than 84˚ at 16-to-18 weeks than 0.31 (n=16 hips) or a NorDis angle measurement
and 52 weeks, respectively, developed DJD. For hips greater than 84˚ (n=12 hips) later developed DJD. For
with NorDis angles less than 85˚ at six-to-10 weeks, statistical ANOVA, a 50% probability of developing
27 (69%) of 39 developed DJD. For angles less than DJD was predicted at DI values of 0.40 and 0.41 for
85˚ at 16-to-18 weeks, 12 (92%) of 13 hips developed ages six-to-10 weeks and 16-to-18 weeks, respec-
DJD, and at one year 15 (100%) of 15 hips had DJD. tively. Distraction index was somewhat better at pre-
dicting DJD than NorDis, as seen by p values (0.0001
Neutral-Positioned Hips Measured During versus 0.0007) and percentage of correct prediction
Compression (78.4% versus 68.6%) in Table 2. At 16-to-18 weeks
Norberg/compression (NorCom) angles (range, 89˚ to of age, the difference between these two methods was
120˚ for all breeds at six-to-52 weeks) were increased minimal. Nine of 10 hips with DI values greater than
consistently when compared to NorOFA, although 0.61 and all seven hips with NorDis angle measure-
the change was minimal for GHs. No hips developed ments less than 75˚ at six-to-10 weeks of age later
DJD that had a NorCom angle greater than 113˚. developed DJD. A similar trend was found at 16-to-
However, only 4% of hips measured at six-to-10 and 18 weeks.
16-to-18 weeks of age had NorCom angles greater A backward elimination procedure with logistic
than 113˚. regression was performed to determine if data from
July/August 1998, Vol. 34 Canine Hip Dysplasia 345

more than one of the six methods (Ortolani’s maneu- sured below the 0.26 DI threshold for DJD than above
ver, Bardens’ maneuver, distraction index, Norberg the 107˚ NorOFA threshold for DJD (24% versus
angle, Norberg angle during distraction, and Norberg 0%). Norberg angle measurement during hip distrac-
angle during compression) of measuring laxity was tion (NorDis), as first reported here, also appears
useful in predicting DJD at eight, 16, and 52 weeks of accurate in predicting DJD, though not with quite as
age. Distraction index alone was selected by the back- great a sensitivity as the DI (68.6% and 78.4% cor-
ward elimination procedure. Thus, once DI was rect, respectively).
known, none of the other methods added significantly Results of this study do not support the concern
to prediction of DJD occurrence. that repeated hip palpations or stress radiography may
lead to joint instability. Despite repeating hip palpa-
Discussion tion and stress views seven times in GH puppies, no
Both palpation methods and all four radiographic mea- evidence of DJD or increased laxity was detected
surement methods used to determine laxity (and there- compared to control hips. In fact, GH hips became
fore likelihood of DJD) when applied to one-year-old significantly tighter over time. This finding in a tight-
hips, were significantly diagnostic as determined by hipped breed does not address the possibility that
consensus opinion of radiographic DJD findings at 52 hips at risk (i.e., lax hips) may be traumatized by
weeks. Both palpation methods (i.e., Ortolani’s ma- repeated stress radiography, but a similar tendency
neuver and Bardens’ maneuver) were quite insensi- for DI scores to decrease (i.e., decreased laxity) be-
tive at six-to-10 weeks (false negatives, 85% and tween 16 and 52 weeks of age also was observed in
81%, respectively) and at 16-to-18 weeks (false nega- hips that developed DJD during this study. As all
tives, 37% and 74%, respectively) for predicting DJD, examinations were performed under fluoroscopic con-
even in the hands of an experienced orthopedic sur- trol, only enough force was applied to visualize sub-
geon. Likewise, subjective assessment of extended- luxation. In cases of significant joint laxity, the
hip (OFA) radiographs was quite insensitive at femoral head was observed fluoroscopically to “pop
six-to-10 weeks (89% false negative) and at 16-to-18 out” of acetabular alignment with very minor force.
weeks (56% false negative) for predicting DJD, in Applying more force resulted in little, if any, addi-
agreement with previous reports. 14,15 It is notable that tional subluxation.
although palpation and subjective radiographic meth- A limitation of this study was the statistical neces-
ods were insensitive at these two age brackets, speci- sity to pool data from several breeds, due to the small
ficity of these methods was high, as very few number of puppies. As illustrated in Table 1, breeds
false-positive results were recorded at either six-to- at risk for hip dysplasia may fall into a DI range quite
10 weeks or 16-to-18 weeks of age. When all seven apart from the range measured from a breed (e.g., the
methods were applied at six-to-10 weeks of age, three GH) with tight hips. Clinical application of a laxity
radiographic measurement methods (i.e., DI, NorOFA, index such as DI will require assessment of indi-
NorDis) were related significantly to the incidence of vidual breeds in sufficient numbers for statistical
DJD at 52 weeks. analysis.
Distraction index was statistically the most predic- It is of interest to note that the mean DI of one-
tive method for DJD when applied at either six-to-10 year-old LR (0.53) and IS (0.46) hips here reported
weeks or 16-to-18 weeks of age. At six-to-10 weeks are almost identical to median DI scores for these
of age, all hips with DIs less than 0.31 were normal at breeds in the PennHIP data base (0.52 and 0.47, re-
one year and 81% of hips with DIs in the range of spectively). 20 It also is notable that these LR puppies
0.31 to 0.41 were normal at one year. At 16-to-18 were from hip dysplasia-positive parents, and the IS
weeks, predictability of DI values (in the range of puppies were progeny of OFA-certified “good” par-
0.31 to 0.41) for normal hips at one year increased to ents. Similarity in the mean DIs of these LR and IS
93%. This result is similar to the previously reported hips is consistent with the report that there has been
88% correct prediction for normal hips in 16-week- no significant decrease in the incidence of hip dyspla-
old LRs at DI less than 0.40.14 Prediction of DJD by sia in these two breeds between 1972 and 1990. 3
DI for puppies 16 weeks of age in this report was The presence of DJD was based on the consensus
found to be more accurate than previously reported. 14 of four expert readers. Credibility of this assessment
At an earlier age (six-to-10 weeks) in this study, a DI was strengthened by disqualification of any hips upon
greater than 0.60 predicted DJD in 90% of hips, while which the four readers did not agree and by retrospec-
a DI greater than 0.70 predicted DJD in 100% of hips. tive verification of the occurrence of enthesophytes
As an early screening test to detect puppies at not visualized on earlier films of the same joints.
minimal risk for developing DJD, DI appears more Sensitivity of detecting DJD was limited by the fact
useful than NorOFA, as suggested by others. 14,15 At that hips were not radiographed beyond 52 weeks of
six-to-10 weeks of age, many more hips were mea- age. Historically, radiographic signs of hip dysplasia
346 JOURNAL of the American Animal Hospital Association July/August 1998, Vol. 34

are reported present in 62% to 80% of affected dogs laxity threshold (DI greater than 0.70) above which
at one year of age. 10,16 In a more recent study, DJD DJD appears nearly certain. Even in very tight-hipped
present by three years of age was evident in 32 (94%) puppies, there is a significant difference in hip radio-
of 34 dogs when radiographed at one year of age. 17 graphic appearance between four and eight weeks of
Specificity of enthesophyte findings resulting in a age, resulting in unacceptable measurement inaccu-
diagnosis of DJD was limited by absence of histo- racy at four weeks of age.
pathological verification. Enthesophytes seen in this
series were typical in appearance to those associated a
E.A. Corley, DVM, PhD, President and Executive Director, Orthopedic
with coxofemoral DJD 8,12 and were documented as Foundation for Animals, Columbia, MO
b
acquired changes. G.K. Smith, VMD, PhD, School of Veterinary Medicine, University of
Pennsylvania, 3800 Spruce Street, Philadelphia, PA
This study supports the conclusions of others that
measurement of joint laxity is a sensitive predictor of
Acknowledgments
DJD, 7,14,17 that a DI less than 0.30 indicates minimal
risk of CHD, and that a DI greater than 0.70 when This research was supported by a grant from the
applied to puppies at four months 15 or six months 18 of American Animal Hospital Association.
age makes DJD all but certain. This study also con- The authors are grateful to Drs. E. A. Corley, G. K.
firms the predictability of increased risk of CHD as Smith, and L. J. Forrest for their excellent assistance
the DI increases between 0.30 and 0.70. Results pre- and consultations and to James Fialkowski and Peter
sented here are in disagreement with a study of 16 Crump for their detailed technical assistance.
German shepherd dog puppies radiographed from
eight weeks of age. In that study, hip laxity was not References
sufficiently reliable to use in a predictive sense at 1. Schnelle GB. Some new diseases in the dog. Am Kennel Gaz 1935;52:
eight weeks of age. 19 This differing result underscores 25–6.
the opinion that there is breed variation between a 2. Rendano VT, Ryan G. Canine hip dysplasia evaluation: a positioning and
labelling guide for radiographs to be submitted to the Orthopedic Foun-
given degree of hip laxity and prediction of DJD.b dation for Animals. Vet Radiol 1985;26:170–6.
In this study, over 50% of hips which developed 3. Corley EA, Keller GG. Hip dysplasia: a progress report and update.
DJD by one year were assessed as not being dysplastic Orthopedic Foundation for Animals, 1993:3.
at 16-to-18 weeks of age on subjective radiographic 4. Smith GK, Biery DN, Gregor TP. New concepts of coxofemoral joint
stability and the development of a clinical stress-radiographic method
assessment in the extended-hip (OFA) position. This for quantitating hip joint laxity in the dog. J Am Vet Med Assoc
finding does not support the OFA report of 91% reli- 1990;196:59–70.
5. Henricson B, Norberg I, Olsson S-E. On the etiology and pathogenesis
ability of that method in puppies three-to-six months of hip dysplasia: a comparative review. J Sm Anim Pract 1966;7:673–87.
of age. 21 Additional imaging studies in puppies be- 6. Bardens JW, Hardwick H. New observations on the diagnosis and cause
ginning at eight weeks of age presently are underway of hip dysplasia. Vet Med Sm Anim Clin 1968;63:238–45.
using other CHD-susceptible breeds, in order to ad- 7. Madsen JS, Svalastoga E. Early diagnosis of hip dysplasia—a stress-
radiographic study. Vet Clin Orthop Traum 1995;8:114–7.
dress these differing reports.
8. Riser WH. The dysplastic hip joint: its radiographic and histological
development. J Am Vet Radiol Soc 1973;14:35–50.
Conclusion 9. Chalman JA, Butler HC. Coxofemoral joint laxity and the Ortolani sign.
J Am Anim Hosp Assoc 1985;21:671–6.
In the hands of a trained clinician, a positive
10. Corley EA, Keller GG. Hip dysplasia: a guide for dog breeders and
Ortolani’s or Bardens’ sign at six-to-10 weeks or 16- owners. 2nd ed. Columbia, MO: Orthopedic Foundation for Animals,
to-18 weeks of age was an accurate predictor for later 1989:4.
development of DJD. However, false-negative inci- 11. Douglas SW. The Norberg method of assessing hip dysplasia. In:
Williamson HD, Douglas SW, eds. Veterinary radiological interpreta-
dence with these palpation methods at both ages was tion. 1st ed. Philadelphia: Lea & Febiger, 1970:109.
unacceptable. Prediction of DJD by subjective radio- 12. Morgan JP. Canine hip dysplasia: significance of early bone spurring.
graphic assessment of extended-hip (OFA) views at Vet Radiol 1987;28:2–5.
13. Hosmer DW, Lemeshow S. Applied logistic regression. New York:
six-to-10 weeks and at 16 weeks of age also resulted J Wiley and Sons, 1989:64–9.
in a high number of false negatives and cannot be 14. Lust G, Williams AJ, Burton-Wurster N, et al. Joint laxity and its
recommended as a routine screening method. association with hip dysplasia in Labrador retrievers. Am J Vet Res
1993;54:1990–9.
Development of coxofemoral DJD by 52 weeks of
15. Smith GK, Gregor TP, Rhodes WH, Biery DN. Coxofemoral joint laxity
age was predicted accurately in six- to 10-week-old from distraction radiography and its contemporaneous and prospective
LR, IS, and HM puppies by three radiographic mea- correlation with laxity, subjective score, and evidence of degenerative
joint disease from conventional hip-extended radiographs in dogs.
surement methods. Radiographic assessment with hips Am J Vet Res 1993;54:1021–42.
in neutral position using a distraction index (PennHIP 16. Jessen CR, Spurrell FA. Radiographic detection of canine hip dysplasia
method) was the most accurate of two palpation and in known age groups. Proceed, Am Vet Med Assoc Symposium on Hip
Dysplasia, 1972:93–8.
five radiographic methods tested. A previously sug-
17. Smith GK, Popovitch CA, Gregor TP, Shofer FS. Evaluation of risk
gested joint-laxity threshold (DI greater than 0.30) factors for degenerative joint disease associated with hip dysplasia in
for development of DJD was substantiated, as was a dogs. J Am Vet Med Assoc 1995;206:642–7.
July/August 1998, Vol. 34 Canine Hip Dysplasia 347

18. Flückiger M. Stress radiography for the detection of hip joint laxity in 20. Smith GK. PennHip director. Quarterly breed update. Malvern, PA:
the dog. Proceed, Euro Assoc Vet Diag Imaging Meeting, Berlin, Synbiotics, October 1997:4.
September 1995:30. 21. Corley EA, Keller G, Lattimer JC, Ellersieck MR. Reliability of early
19. Smith GK. Diagnosis of canine hip dysplasia. Proceed, Western Vet radiographic evaluations for canine hip dysplasia obtained from the
Conf, 1995:12–9. standard ventrodorsal radiographic projection. J Am Vet Med Assoc
1997;211:1142–6.
The Effects of Stains and Investigators
on Assessment of Morphology of
Canine Spermatozoa
Percentage and types of morphological abnormalities found in canine spermatozoa
were evaluated by three investigators using three stains (Giemsa-Wright stain
[Diff-Quik], eosin Y/nigrosin [Hancock], and eosin B/nigrosin [Society for
Theriogenology morphology stain] with conventional light microscopy, compared to
phase contrast microscopy on unstained samples. The percentage of spermatozoa
with abnormal heads, midpieces, and tails varied by technique and by investigator.
Average percentages of morphologically normal spermatozoa were significantly
higher in samples stained with Diff-Quik and samples examined by phase contrast
microscopy than in samples stained with Hancock or Society for Theriogenology
morphology stains. No effect of investigator on the percentage of morphologically
normal spermatozoa was assessed. Results suggest that staining or preparation
technique may alter the morphology of canine spermatozoa artifactually.
J Am Anim Hosp Assoc 1998;34:348–52.

Margaret V. Root Kustritz, Introduction


DVM, PhD Morphological defects of spermatozoa may be predictors of fertility.
Patricia N. Olson, DVM, PhD A positive correlation has been reported between normal morphol-
ogy, acrosomal function, fertilizing capacity of spermatozoa, and
Shirley D. Johnston, DVM, PhD success of in vitro fertilization in humans. 1–5 Morphological abnor-
malities of spermatozoa are associated with decreased fertility in the
Teresa K. Root, BS bull, 6–11 stallion,12–14 boar, 8,15 ram, 8 and dog. 16–19 Renton, et al.19 and
Plummer, et al.18 described midpiece abnormalities associated with
infertility in the dog. Oettle 17 and Mickelsen, et al. 16 demonstrated
O positive correlations in the dog between percentage of morphologi-
cally normal spermatozoa and fertility rate, and the total number of
morphologically normal spermatozoa and pregnancy rate, respectively.
Morphology of spermatozoa can be assessed using conventional
light microscopy or phase contrast microscopy, with or without stain-
ing. Staining has been shown either to increase or decrease percent-
ages of morphologically normal spermatozoa in the bull, 20–24
stallion, 24,25 boar, 24 ram, 24 buck,24 rabbit, 20 human, 20,26–30 and dog. 31
In one study, eosin B/nigrosin stain (Society for Theriogenology
morphology stain), with osmolality of 135 mOsm and pH of 2.27,
caused a significant increase in the percentage of canine spermatozoa
with bent tails.31
The objective of the study reported herein was to determine whether
the percentages of morphologically normal and abnormal spermato-
zoa evaluated by three investigators were different in stained samples
(using one of three stains) examined with conventional light micros-
From the Department of copy or in unstained samples examined with phase contrast microscopy.
Small Animal Clinical Sciences,
College of Veterinary Medicine, Materials and Methods
University of Minnesota,
1352 Boyd Avenue, Breeder clients were solicited for inclusion of intact male dogs in this
St. Paul, Minnesota 55108. study. Twenty-three dogs were presented for breeding soundness

348 JOURNAL of the American Animal Hospital Association


July/August 1998, Vol. 34 Canine Spermatozoa 349

examination and semen collection; all were included also were evaluated using phase contrast microscopy.
in the study. The dogs ranged in age from 10 months Spermatozoa lying on their sides were not evaluated.
to 14 years (mean±standard deviation [SD], 6.2±3.8 Statistical analysis was performed using multivari-
yrs) and represented 12 breeds. ate analysis of variance (MANOVA). The data as-
Canine semen was collected into a latex artificial sumed a normal distribution. The influence of
vagina, as described previously, in the absence of a technique, investigator, or both and interaction be-
teaser bitch. 32 Samples and equipment were held at tween technique and investigator on percentage of
room temperature. Volume was recorded for each morphologically normal spermatozoa, percentage of
sample, and then aliquots were removed for evalua- spermatozoa with abnormal heads, percentage with
tion of progressive motility and total number of sper- abnormal midpieces, and percentage with abnormal
matozoa in the ejaculate. Progressive motility was tails were assessed, with a p value of less than 0.05
evaluated immediately; total number of spermatozoa considered significant. Pair-wise comparisons were
in each ejaculate was evaluated, and the remainder of made using the least significant difference test.
each sample was processed for morphology assess-
ment within four hours of collection. Semen evalua- Results
tion has been described previously. 32 Semen was collected from the 23 dogs. Mean volume
From each sample, an aliquot of semen was diluted ±SD was 3.2±2.3 ml (range, 1.0 to 10.0 ml). The
1:9 with formol-buffered saline and refrigerated for percentage of progressively motile sperm averaged
less than one week at 4˚ C until evaluated by phase 69%±24% (range, 10% to 90%). Total number of
contrast microscopy. The remainder of the undiluted spermatozoa in the ejaculate ranged from 10 million
semen sample then was used to make nine slides, to 1.7 billion (mean±SD, 334±375 million).
three of which were stained with a rapid Giemsa- The percentage of spermatozoa that was morpho-
Wright stain (Diff-Quik),a three of which were stained logically normal and the percentage that had various
with an eosin Y/nigrosin stain (Hancock), b and three morphological abnormalities of the head, midpiece,
of which were stained with an eosin B/nigrosin stain and tail were assessed by three investigators using the
(Society for Theriogenology morphology stain).c Mor- four described techniques [Tables 1–4]. The most
phology specimens were prepared for staining with Diff- common morphological abnormalities of the heads of
Quik by placing a drop of semen on a glass slide, drawing canine spermatozoa were detached heads, knobbed
it out with a second slide (as for a blood smear), and acrosomes, and detached acrosomes. Proximal and
allowing it to air dry. The slides were immersed serially distal cytoplasmic droplets and reflex (i.e., bent)
in each of the three Diff-Quik solutions for five minutes, midpieces were the most frequently identified mor-
rinsed, and allowed to air dry. Slides were prepared phological abnormalities of the midpiece. The three
for morphology assessment with Hancock or Society most common morphological abnormalities of the tail
for Theriogenology morphology stain by placing a were bent tails, tails tightly coiled over the midpiece,
drop of semen and a drop of either Hancock or Soci- and proximally coiled tails.
ety for Theriogenology morphology stain on one end The stain technique used had an effect on the
of each glass slide. The drops were mixed, and the percentage of morphologically normal spermatozoa
resultant mixture was drawn out with another glass assessed (p of 0.02). When compared to Diff-Quik-
slide. Each slide was allowed to air dry. Stained stained samples and unstained samples examined
slides were held at room temperature pending using phase contrast microscopy, the percentage of
evaluation. morphologically normal spermatozoa was signifi-
A factorial design was used; three investigators cantly lower in Hancock- and Society for Therio-
evaluated all samples using each of the four techniques genology morphology-stained samples [Table 1].
(i.e., three staining techniques for conventional light No effect of the investigator on the percentage of
microscopy and phase contrast microscopy). The in- morphologically normal spermatozoa was assessed
vestigators were veterinarians with experience in (p of 0.25), and no interaction between investiga-
evaluating canine semen. Stained slides were coded tor and technique (p greater than 0.25) was
with random numbers and were examined in random assessed.
order by the investigators using oil immersion at A significant interaction of technique and investi-
1,000X magnification. One hundred spermatozoa gator on the percentage of spermatozoa with head
were examined on each slide. Percentages of morpho- abnormalities was identified (p less than 0.01). One
logically normal sperm and spermatozoa with defects investigator evaluated a significantly larger number
of the head, midpiece, tail, or combinations were of spermatozoa as having head abnormalities on only
recorded. All investigators classified morphological the Diff-Quik-stained samples compared to the other
abnormalities according to a descriptive chart pre- investigators [Table 2]. The presence of the interac-
pared jointly. One hundred unstained spermatozoa tion precluded further evaluation of the effects of
350 JOURNAL of the American Animal Hospital Association July/August 1998, Vol. 34

Table 1
Percentage of Morphologically Normal Spermatozoa* Assessed by
Four Techniques and Three Investigators in 23 Canine Samples

Techniques
Society for
Theriogenology
Diff-Quik Stain† Hancock Stain Morphology Stain Phase Contrast†
Investigator 1 62 (21) 56 (19) 52 (20) 63 (20)
Investigator 2 70 (19) 60 (22) 54 (23) 64 (19)
Investigator 3 53 (34) 59 (24) 51 (22) 63 (23)

* Numbers displayed are mean percentages with standard deviations in parentheses



Percentages of morphologically normal spermatozoa are higher using these techniques (Diff-Quik and phase
contrast) than others (Hancock and Society for Theriogenology morphology stains)

Table 2
Percentages of Spermatozoa with Head Abnormalities* Assessed by
Four Techniques and Three Investigators in 23 Canine Samples

Techniques
Society for
Theriogenology
Diff-Quik Stain Hancock Stain Morphology Stain Phase Contrast
Investigator 1 10 (11) 10 (9) 7 (9) 5 (5)
Investigator 2 8 (12) 14 (16) 14 (11) 6 (7)
Investigator 3† 28 (37) 13 (13) 11 (15) 4 (6)

* Numbers displayed are mean percentages with standard deviations in parentheses



Percentages of Diff-Quik-stained spermatozoa with head abnormalities are higher for this investigator; this interaction
precludes further statistical assessment

investigator and technique on the percentage of sper- compared to the evaluations of the other two investi-
matozoa with head abnormalities. gators [Table 4]. When compared to Diff-Quik- and
The percentage of spermatozoa with midpiece ab- Hancock-stained samples and unstained samples ex-
normalities was affected significantly by technique, amined using phase contrast microscopy, all three
with a lower percentage assessed in Diff-Quik-stained investigators evaluated the percentage of spermato-
samples and a higher percentage identified in un- zoa assessed with tail abnormalities as being increased
stained samples examined using phase contrast mi- in Society for Theriogenology morphology-stained
croscopy compared to Hancock- and Society for samples. No interaction between technique and in-
Theriogenology morphology-stained samples (p less vestigator was recognized (p of 0.25).
than 0.01) [Table 3]. Neither an effect of investigator
on the percentage of midpiece abnormalities (p greater Discussion
than 0.25) nor an interaction between investigator Although percentages of normal spermatozoa and of
and technique was assessed (p greater than 0.25). specific abnormalities of the head, midpiece, and tail
Investigator (p less than 0.01) and technique (p differed significantly by technique, a much less sig-
less than 0.01) had an effect on the percentage of nificant effect on the percentage of normal spermato-
spermatozoa assessed as having tail abnormalities. zoa was assessed. All techniques used may increase
One investigator classified a significantly larger per- some abnormalities while decreasing others. The ap-
centage of spermatozoa as having tail abnormalities parent increase in a given abnormality could be due
July/August 1998, Vol. 34 Canine Spermatozoa 351

Table 3
Percentages of Spermatozoa with Midpiece Abnormalities* Assessed by
Four Techniques and Three Investigators in 23 Canine Samples

Techniques
Society for
Theriogenology
Diff-Quik Stain† Hancock Stain Morphology Stain Phase Contrast‡
Investigator 1 14 (15) 22 (15) 21 (19) 24 (19)
Investigator 2 13 (9) 21 (13) 17 (16) 25 (18)
Investigator 3 10 (12) 17 (18) 17 (18) 25 (19)

* Numbers displayed are mean percentages with standard deviations in parentheses



Percentages of spermatozoa with midpiece abnormalities are lower using this stain (Diff-Quik) than with others
(Hancock and Society for Theriogenology morphology stains)

Percentages of spermatozoa with midpiece abnormalities are higher using this technique (phase contrast) than with
others (Hancock and Society for Theriogenology morphology stains)

Table 4
Percentages of Spermatozoa with Tail Abnormalities* Assessed by
Four Techniques and Three Investigators in 23 Canine Samples

Techniques
Society for
Theriogenology
Diff-Quik Stain Hancock Stain Morphology Stain† Phase Contrast

Investigator 1 14 (36) 16 (13) 23 (14) 10 (7)
Investigator 2 10 (10) 9 (10) 18 (14) 7 (5)
Investigator 3 8 (8) 10 (13) 20 (18) 8 (9)

* Numbers displayed are mean percentages with standard deviations in parentheses



Percentages of spermatozoa with tail abnormalities are higher using this technique (Society for Theriogenology
morphology stain) than others (Diff-Quik and Hancock stains and phase contrast microscopy)

Percentages of spermatozoa with tail abnormalities are higher for this investigator compared to the other
investigators

either to accentuation or creation of that defect by nificantly [Table 3]. Some samples contained many
that technique as evidenced in this study by the sig- spermatozoa with cytoplasmic droplets which were
nificantly higher percentage of spermatozoa with tail visible as transparent bubbles on the midpiece when
abnormalities on Society for Theriogenology mor- using phase contrast microscopy but which were de-
phology-stained samples [Table 4]. The majority of flated or absent on the stained samples, presumably
these tail abnormalities were bent tails, suggesting a due to processing of the sample.
mechanism similar to that described by Johnson, et The investigators agreed on the percentage of mor-
al., 31 whereby changes in the chemical composition phologically normal spermatozoa more uniformly than
of the stain induced morphological changes in the on abnormal spermatozoa. In a study comparing vis-
spermatozoa. An apparent decrease in a given abnor- ual assessment with computer-assisted assessment of
mality could be due to either a masking of that defect morphology of human spermatozoa, the percentage of
or the removal of the defect by that technique. For morphologically normal spermatozoa was the only
example, the percentage of midpiece abnormalities parameter classified by both methods with equal pre-
on unstained samples evaluated by phase contrast mi- cision. 33 In this study, differences in observed per-
croscopy compared to stained samples increased sig- centages of morphological abnormalities were great,
352 JOURNAL of the American Animal Hospital Association July/August 1998, Vol. 34

despite review of and agreement on definition of ab- 16. Mickelsen WD, Memon MA, Anderson PB, Freeman DA. The relation-
ship of semen quality to pregnancy rate and litter size following artificial
normal forms. The significant effect of investigator insemination in the bitch. Theriogenology 1993;39:553–60.
on the percentage of abnormalities noted highlights 17. Oettle EE. Sperm morphology and fertility in the dog. J Rep Fert
the need for a standardized nomenclature for mor- 1993;Suppl. 47:257–60.
18. Plummer JM, Watson PF, Allen WE. A spermatozoal midpiece abnor-
phology assessment. mality associated with infertility in a Lhasa apso dog. J Sm Anim Pract
1987;28:743–51.
Conclusion 19. Renton JP, Harvey MJA, Harker S. A spermatozoal abnormality in dogs
related to infertility. Vet Rec 1986;118:429–30.
The most reliable measure of morphology of canine
20. Drevius LO. Spiralization in tails of mammalian spermatozoa in hypo-
spermatozoa between investigators in this study is tonic media. Nature 1963;197:1123–4.
percentage of morphologically normal spermatozoa. 21. Harasymowycz J, Ball L, Seidel GE. Evaluation of bovine spermatozoal
Clinicians should be aware that the staining or prepa- morphologic features after staining or fixation. Am J Vet Res
1976;37:1053–7.
ration technique used may alter the morphology of 22. Salisbury GW, Willett EL, Seligman J. The effect of the method of
canine spermatozoa artifactually and that consistency making semen smears upon the number of morphologically abnormal
spermatozoa. J Anim Sci 1942;1:199–205.
of technique is crucial for comparison of morphology
23. Sekoni VO, Gustafsson BK, Mather EC. Influence of wet fixation,
samples within or between animals. staining techniques, and storage time on bull sperm morphology.
Nord Vet Med 1981;33:161–6.
a 24. Wurgau T. The significance of tertiary anomalies for the evaluation of
Diff-Quik; Scientific Products, Columbia, MD sperm morphology. Thesis: Tierarztliche Fakultat, Ludwig Maximilians
b Universitat Munchen, German Federal Republic, 1986.
Hancock; Lane Manufacturing, Denver, CO
c 25. Oetjen M. Use of different acrosome stains for evaluating the quality of
Modified Blom’s; Society for Theriogenology, Lane Manufacturing,
Denver, CO fresh and frozen horse semen. Thesis: Tierarztliche Hochschule
Hannover, German Federal Republic, 1988.
26. Ali JI, Grimes EM. Sperm morphology: unstained and stained smears in
fertile and infertile men. Arch Androl 1989;22:191–5.
References 27. Bornman MS, Sevenster CB, DeMilander C, Otto BS, Olivier I. The
1. Enginsu ME, Dumoulin JCM, Pieters MHEC, Evers JLH, Geraedts PM. effect of semen processing on sperm morphology. Andrologia
Predictive value of morphologically normal sperm concentration in the 1989;21:117–9.
medium for in-vitro fertilization. Int J Androl 1993;16:113–20. 28. Cameron RDA. Semen collection and evaluation in the ram. The prepa-
2. Fukuda M, Morales P, Overstreet JW. Acrosomal function of human ration of spermatozoa for morphological examination. Aust Vet J
spermatozoa with normal and abnormal head morphology. Gamete Res 1977;53:384–6.
1989;24:59–65. 29. Meschede D, Keck C, Zander M, Cooper TG, Yeung CH, Nieschlag E.
3. Grow DR, Oehninger S, Seltman HJ, et al. Sperm morphology as Influence of three different preparation techniques on the results of
diagnosed by strict criteria: probing the impact of teratozoospermia human sperm morphology analysis. Int J Androl 1993;16:362–9.
on fertilization rate and pregnancy outcome in a large in vitro fertiliza- 30. Michael AY, Drejer JO, Bagger PV, Detlefsen GU, Stakemann G.
tion population. Fert Steril 1994;62:559–67. Complete staining of human spermatozoa and immature germ cells
4. Kruger TF, DuToit TC, Franken DR, et al. A new computerized method combined with phase contrast microscopy. Arch Androl 1987;
of reading sperm morphology (strict criteria) is as efficient as 19:217–21.
technician reading. Fert Steril 1993;59:202–9. 31. Johnson C, Jacobs J, Walker R. Diagnosis and control of Brucella canis
5. Kruger TF, Menkveld R, Stander FSH, et al. Sperm morphologic fea- in kennel situations. Morphology-stain induced spermatozoal abnor-
tures as a prognostic factor in in vitro fertilization. Fert Steril malities. In: Proceed, Society for Theriogenology, 1991:236–9.
1986;46:1118–23. 32. Root MV, Johnston SD. Basics for a complete reproductive examination
6. Andersson M, Vierula M, Alanko M. Three types of acrosomal aberra- of the male dog. Vet Med and Surg 1994;9:41–5.
tions of bull spermatozoa and their relation to fertility. Acta Vet Scand 33. Wang C, Leung A, Tsoi W, et al. Computer-assisted assessment of
1990;31:175–9. human sperm morphology: comparison with visual assessment.
7. Foote RH, Hough SR, Johnson LA, Kaproth M. Electron microscopy and Fert Steril 1991;55:983–8.
pedigree study in an Aryshire bull with tail-stump sperm defects.
Vet Rec 1992;130:578–9.
8. Hancock JL. The morphologic characteristics of spermatozoa and
fertility. Int J Fertil 1959;4:347–59.
9. Peet RL, Mullins KR. Sterility in a poll Hereford bull associated with
the “tail stump” sperm defect. Aust Vet J 1991;68:245.
10. Peet RL, Kluck P, McCarthy M. Infertility in 2 Murray Grey bulls
associated with abaxial and swollen midpiece sperm defects. Aust Vet J
1988;65:359–60.
11. Williams WW, Savage A. Observations on the seminal micropathology
of bulls. Cornell Vet 1925;15:353–74.
12. Held JP, Prater P, Stettler M. Spermatozoal head defect as a cause of
infertility in a stallion. J Am Vet Med Assoc 1991;199:1760–1.
13. Jasko DJ, Lein DH, Foote RH. Determination of the relationship be-
tween sperm morphologic classifications and fertility in stallions: 66
cases (1987–1988). J Am Vet Med Assoc 1990;197:389–93.
14. Jasko DJ, Little TV, Lein DH, Foote RH. Comparison of spermatozoal
movement and semen characteristics with fertility in stallions: 64 cases
(1987–1988). J Am Vet Med Assoc 1992;200:979–85.
15. Bach VS, Neundorf P, Stemmler KH, Mudra K, Ueckert H. Amount and
variation of abnormal sperms of boar. Monatshefte fur Vet Med
1982;37:463–7.

You might also like