Abdominocentesis in Cattle: Technique andCriteria for Diagnosis of Peritonitis

A.D. WILSON, V.M. HIRSCH AND A.D. OSBORNE

Department of Veterinary Anesthesiology, Radiology and Surgery (Wilson),Department of Veterinary Pathology (Hirsch) and Department of VeterinaryMicrobiology (Osborne), Western College of Veterinary Medicine,University of Saskatchewan, Saskatoon, Saskatchewan S7N 0 WO

ABSTRACT

A reliable method for the collection ofperitoneal fluid from cattle using atrocar and cannula is described. Perito-neal fluid was collected from threegroups of cattle: periparturient, nor-mal and with peritonitis. The fluid wasexamined by white cell count, differen-tial cell count, total protein concentra-tion and bacteriology. The results wereanalysed to determine the best criteriafor peritonitis. Greater than 10% eosi-nophils were typical of normal perito-neal fluid. Peritoneal fluid with a rela-tive neutrophil count greater than 40%and a relative eosinophil count of lessthan 10% was frequently associatedwith the diagnosis of peritonitis. Par-turient cattle had large volumes ofperitoneal fluid with low total proteinand white cell counts. Growth ofGram-negative or anaerobic orga-nisms was associated with mortality.

Key words: Peritoneal fluid, peritoni-tis, eosinophils.

RASUMMTechnique d'abdominocentese, chezles bovins, et criteres de diagnostic dela peritoniteCet article decrit une methode fiabled'effectuer une abdominocentese, chezdes bovins, a l'aide d'un trocart etd'une canule. Cette methode permit derecolter du liquide peritoneal, chez desvaches normales, ou atteintes de peri-tonite, ou a la veille de veler. On yrechercha ensuite le nombre total deleucocytes et leur proportion respec-

tive, la quantite totale de proteines,ainsi que des bacteries. On analysa lesresultats obtenus, afin de determinerles meilleurs criteres de diagnostic dela peritonite. Une proportion d'eosino-philes superieure A 10% se revelacaract6ristique du liquide peritonealnormal; par ailleurs, une proportionrelative de neutrophiles superieure A40%, jointe a une proportion d'eosino-philes inferieure a 10%, accompagnaitsouvent une peritonite. Les vaches surle point de veler affichaient une quan-tite abondante de liquide peritoneal,pauvre en proteines et en leucocytes.L'isolement de bacteries gram-negatives signifiait une peritonitefatale.

Mots cles: liquide peritoneal, perito-nite, eosinophiles.

I N T R O D U C T IO N

There are several accounts of the analy-sis of peritoneal fluid in pathologicalconditions of the bovine abdomen(1,2,3). There are also descriptions ofnormal bovine peritoneal fluid (1,2)although no details of numbers ofcows examined were published. In pre-liminary studies at this college weexperienced difficulty in obtainingsamples from healthy cows. Fromapproximately 70% of cattle either nofluid, or fluid severely contaminatedwith blood was obtained. It is possiblethat this could lead to a biased assess-ment of normal peritoneal fluid if onlycows with high amounts of fluid yield a

sample. In our experience and that ofothers (1) cows in late pregnancy haveincreased amounts of peritoneal fluidbut detailed descriptions of its cytol-ogy have not been published. In thispaper we describe a reliable techniquefor collection of bovine peritonealfluid. Using this technique we wereable to collect and characterize sam-ples of peritoneal fluid from normalcows and compare these to samplesfrom parturient cows and cattle withperitonitis.

MATERIALS AND METHODS

Three groups of cattle were sampled.Group I consisted of 19 cows pre-

sented to the Western College of Veteri-nary Medicine, University of Sas-katchewan for alleviation of dystociaby hysterotomy. These cows wereselected for the presence of a live calf,the absence of other clinical disease,and no history of extensive fetalmanipulation. Peritoneal fluid wascollected during surgery, the skin andmuscle layers were incised, and a 5 cm14 gauge test cannula pushed throughthe peritoneum prior to opening theabdomen. The resulting free-flowingperitoneal fluid was collected into 2 mlsterile tubes for bacteriology and 2 mltubes containing 4.5 mg potassiumEDTA for cytology and total proteinassay. When possible, samples wereexamined immediately and if not,direct and spun smears were made andthe sample stored at 5°C until it couldbe examined. All samples were exam-ined within 18 hours.

Can Vet J 1985; 26: 74-80

Reprint requests to Dr. A.D. Wilson, University of Bristol, Department of Veterinary Medicine, Langford House, Bristol, BS18 7DU, England.This work was supported by the Departments of Veterinary Medicine, Radiology and Surgery, Veterinary Pathology and Veterinary Microbiology,University of Saskatchewan.

74

Group 2 contained 21 apparentlyhealthy nonpregnant Holstein cows.Each cow was examined and a 5 mlblood sample collected for a completeblood count. Abdominocentesis wasperformed at a site 10 cm cranial and10 cm to the right-hand side of theumbilicus (Figure 1). The area wasclipped, washed with a surgical scrubcontaining 5% povodine iodine (Beta-dine Scrub Purdue Frederick Com-pany (Canada) Ltd., Toronto, Ontario)and swabbed with 70% alcohol. FivemL of 2% Lidocaine HCI containing1:100,000 epinephrine (Armitage Car-roll Ltd., London, Ontario) wasinjected into the subcutaneous tissueand tissues of the abdominal wall. Ableb of local anesthetic was left underthe skin so that the anesthetized sitecould be readily identified. A stab inci-sion was made through the skin using aNo. 15 scalpel blade (Figure 1), takingcare to avoid cutting large cutaneousblood vessels. If excessive bleedingoccurred it was controlled by applyingpressure to the site with a sterile gauzesponge. A sterile 80 mm logn Nelsontrocar and cannula of 4 mm internaldiameter (Aesculop-Werke, A.G.Normals Jetter and Scheerer, D-7 1,Tuttlingen, Germany) was pushed ver-tically through the abdominal wall to itsfull length (Figure 2). If the skin wasincised completely only a modest pres-sure was required to penetrate theremaining structures of the abdominalfloor. Once the cannula was in positionthe trocar was removed and a sterile80 cm long 10 French gauge infant feed-ing tube (Argyle Division of SherwoodMedical, St. Louis, Missouri) insertedinto the abdomen via the cannual (Fig-ure 3) leaving 10-20 cm outside.

Fluid was collected by allowing it toflow from the tube into containers asin group 1. Also, the tube acted as awick, fluid being drawn through thecannual and running down the outsideof the tubing to be collected into acontainer. The time taken to collectsufficient (0.5 mL each for bacteriol-ogy and cytology) varied from fiveseconds to ten minutes. All sampleswere analysed immediately followingcollection.

Subsequently all cows in this groupwere subjected to laparotomy.Group 3, 22 cattle with clinical peri-

tonitis, were sampled using a teat can-nula pushed through a skin incision in

H IURl 1. Site ol azbdominocentesis in catttle using the trocair aind cannulalil method.

ABDOMINUS

SKIN

F1I(UiRt 2. Collection of bovine peritonieail Iluid showing the tissues penietraited by trocatr aindca nnulit.

FIGU RE 3. Collection of peritoneal fluid using trocar and cannula and fenestrated tube showing sitesof collection (arrows), through and around the tube.

75

the abdomen 5 cm cranial and 5 cmmedial to the point where the left milkvein disappears. On some occasionsthe method using the trocar and can-nula described above was used whenno fluid could be obtained by teat can-nula. Traumatic reticuloperitonitis,ruptured abdominal viscus and post-operative sepsis were the most com-mon causes of peritonitis in this series.In all cases diagnosis was confirmed byexploratory laparotomy or postmor-tem. The cows were further classifiedas acute or chronic cases, based on thenature of the pathological changes.Any cow in which organized fibrousadhesions were observed was classifiedas chronic irrespective of whetherother changes, such as fibrin deposi-tion, indicative of a more acute reac-tion, were present. Those in which nofibrous adhesions were detected wereclassified as acute, regardless of theapparent time course of the disease.

Analysis of Peritoneal FluidThe white blood cell count (WBC)

was determined using an automatedcell counter (Coulter Model S, CoulterElectronics Inc., Hialeah, Florida).Total protein was determined on thesupernatant using a refractometer(T.S. Meter, American Optical Co.,Buffalo, New York). Fibrinogen wasdetermined by the method describedby Kaneko and Smith (4) on all 19samples obtained in group 2 and tensamples in group 3. Direct and spunsmears were stained with Wright's-Giemsa for cytological examination.One hundred cells were counted froma representative area of each sample toyield a differential cell count. Largemononuclear cells, including macro-phages, activated macrophages andmesothelial cells were all classified asmonocytes. Samples for bacteriologywere cultured aerobically and anae-robically on blood agar at 37°C.

Statistical AnalysisDue to the asymmetrical distribu-

tion of many of the sample variablesand the small sample sizes, mediansand observed ranges were used to de-scribe the location and distribution ofthe samples. Comparisons betweentwo samples were performed using theMann-Whitney U test (5). Also, mul-tiple comparisons between groupswere performed by the Kruskal-Wallis

one-way analysis of variance (6) andDunn's Multiple Range Test (7). Two-by-two tables were constructed bystandard methods (8).

RESULTS

Group 1Peritoneal fluid was collected from

every cow in the parturient group.Large volumes were present in every

20

15

10

5

case. The fluid was yellow in color,slightly opaque and clotted on stand-ing. No lesions were detected duringsurgery in any of these cows.

Group 2 (normal nonpregnant)Twenty-one attempts were made to

collect peritoneal fluid. In one case(5%) no fluid was collected and inanother (5%) the fluid was severely

R =18.3M 08.1

C.)

zw

0

w

LL.

W. B.C. x 109/L

FIGURE 4. Histograms of white blood cell count in peritoneal fluid of normal, parturient cattle andcattle with peritonitis.

10

C.)zuJ

aw

LI-

5.

x =38.9M=41

0 KU. * _ lm . a

10PARTURIENT x =19

5 M-17

10NORMAL I = 29.4

5 M=30

o 10 20 30 40 50 60lb io 30 40 5~0 60

g/L

FIGURE 5. Histograms of total protein in peritoneal fluid of normal and parturient cattle withperitonitis.

76

PERITONITIS

*i i . p.sPARTURIENT t =02.9

5 M=02.5

i8:. 1NORMAL t = 12.2

5- M=08.8

0 10 20 30 40 50 70 80 90 100 110 120130

PERITONITIS

contaminated with blood and couldnot be analyzed. In the remaining 19(90%) sufficient (> 0.5 mL) fluid foreach of bacteriology and cytology wascollected. The fluid clotted on stand-ing. The color varied from clear to yel-low and the turbidity also varied.Hematological abnormalities were notdemonstrated in this group. No lesionswere detected at laparotomy.

Group 3The volume of fluid obtained from

cattle with peritonitis varied from lessthan I mL to more than 10 L. Colorand turbidity were highly variable -

some samples contained pieces offibrin. Not all samples clotted in thisgroup.

Frequency histograms of the totalwhite cell counts, the differentialcytology and total protein assays per-formed on peritoneal samples from allthree groups are shown in Figures 4-9.From these and Table I, it can be seenthat the observed ranges overlapextensively. Relative neutrophil andeosinophil numbers had the clearestseparation.

Fibrinogen levels in peritoneal fluidwere < I g/L ' in 15 out of 19 samplestaken from normal cows. The meanvalue was 1.42 g/ L-', median I g/ L-'and range 1-4g/L-'. Fibrinogenvalues in peritoneal fluid of cattle withclinical peritonitis had a mean value of3.5 g/L-', a median value of 3g/L-'and a range from 1-8 g/ L-'. The differ-ences between the group medians wassignificant (P < 0.05).

Plasma fibrinogen data was alsoavailable for the normal cows and 20out of 21 of the cattle with peritonitis.A cut-off point of 6 g/ L, the uppernormal reference value for the clinicallaboratory, was used and a 2 x 2 tableconstructed (Table II). For compari-son, Table III shows a 2 x 2 table inwhich peritoneal fluids were classifiedas suggestive of peritonitis if both rela-tive eosinophil counts were < 10% andrelative neutrophil counts were> 40%.The relative neutrophil and eosinophilcounts were also useful in distinguish-ing parturient (Group 1) cows fromcattle with peritonitis. Using the samecriteria as above true +ve rate = 100%,false -ve rate = 0%, true -ve rate = 90%and false +ve rate = 10%.The fluid from parturient (Group 1)

cows exhibited interesting features.

TABLE IRESULTS OF KRUSKAL-WALLIS 1-WAY ANALYSIS OF VARIANCE ON SIX PARAMETERS OF

BOVINE PERITONEAL FLUID IN NORMAL AND PARTURIEN-I CAITTEAND CATILE WITH PERITONIIIS

Mean Ranks'Group I Group 2 Group 3

(Parturient) (Normal) (Peritonitis) Chi-Square SignificanceWhite Blood Cells 14.55 42.50 32.70 25.94 < 0.001Total Protein 14.95 30.68 43.00 26.65 < 0.001Monocytes 46.05 19.11 25.33 25.83 < 0.001Neutrophils 20.11 19.97 48.02 35.91 < 0.001Eosinophils 31.13 47.50 13.14 40.69 < 0.001Lymphocytes 32.53 37.58 20.86 11.14 < 0.001

aValues underlined indicate no significant differences was demonstrated between mean ranks ofgroups at the a = 0.5 level using Dunn's Method of Comparisons.

20

15-

101

C.)z

0LLI

5.

PERITONITIS x -1.4M=O

O-*10.

PARTURIENT x = 21.55- M-20

,8 * - I **s * .NORMAL A = 51.1

5- M=560 0 A. iO i0

PERCENTFIGURE 6. Histograms of relative eosinophil count in peritoneal fluid of normal and parturient cattleand cattle with peritonitis.

20

15110

C.)zuJ

LL

5

PERITONITIS x =78M= 85

O ~~~.._ . . . _ . L L . L10

PARTURIENT x = 30.75 M 27

O. m m. _m _..10

NORMAL x = 29.95 M=26

I .-L .1 L . - .1b 20 30 40 50 60 70 80 90 160

PERCENTFIGURE 7. Histograms of relative neutrophil count in peritoneal fluid of normal and parturient cattleand cattle with peritonitis.

77

I

z

LI

10

5

PERITONITIS x =19.2M= 15

0m* - - mm- -

10-PARTURIENT x = 43.6

5 M=37

18.NORMAL x = 12.8

5- M :10'lp ..0. s . . . . .0 10 20 30 40 50 60 70 80 90

PERCENTFIGiURE 8. Histograms of relative monocyte count in peritoneal fluid of normal and parturiand cattle with peritonitis.

20-

151

z

LLI.

10-

5.I PERITONITIS x =1.3M=O

18 - 4.

5 PARTURIENT x =4.1M=3

18NORMAL x =6.1

5 M=5

0 10 20 30 40 50 60 70 80 90

PERCENTFIGURE 9. Histograms of relative lymphocyte count in peritoneal fluid of normal and parturiand cattle with peritonitis.

TABLE II2 x 2 TABLE OF PLASMA FIBRINOGEN AS A TEST FOR PERITONITIS

Peritonitisa Normal b TotTest Positivec 16 1 1,Test Negatived 4 18 2,Total 20 19 35

True Positive Rate 80%False Negative Rate 20%False Positive Rate 5%True Negative Rate 95%L20 out of 21 cattle in peritonitis group (3).bAll cows in normal group (2).cFibrinogen > 6 g/ L-'dFibrinogen < 6 g/ L

From Table I it can be seen that WBCand TP were significantly lower thanin either normal (Group 2) or peritoni-tis (Group 3) samples and relativemonocyte counts were higher.The median plasma fibrinogen levels

between the acute and chronic cases ofperitonitis were not significantly differ-ent; however, in cows which had peri-tonitis the median relative monocytecount in peritoneal fluid was signifi-cantly higher in the chronic group.Using a cut-off point of > 15% mono-cytes, 18 out of 21 samples (86%) werecorrectly classified as acute or chronic

100 peritonitis.The frequency with which different

organisms were isolated from perito-ent cattle neal fluid in each of the three groups of

cows in shown (Table IV). The samplesfrom group I taken aseptically duringsurgery were all sterile. The growth oforganisms in group 2 samples mostlikely represents contamination duringcollection. The same organisms arefound in samples from the peritonitisgroup with similar frequency. In addi-tion, Corynebacterium pyogenes,Escherichia coli, Actinomyces andanaerobic Gram-negative rods werecultured. Growth of Gram-negativeorganisms was always associated withdeath or the necessity for euthanasia ofcases.

D I S C U S S O NIn our experience, the trocar and can-nula method of collecting peritonealfluid has proved reliable. The high

100 success rate (90%) in clinically normalcows should prevent bias as a result ofselecting only cows in which there is a

ent cattle large volume of peritoneal fluid. It alsoproved useful in collecting peritonealfluid from some cases where priorattempts with a teat cannula hadfailed. No complications were encoun-tered during the study. However, sub-sequently, penetration of a viscus has

tal occurred on two occasions. In one the7 rumen was penetrated; exploratory2 laparotomy showed this was due to the

rumen having been totally adhered tothe floor of the abdomen. In a secondcase the abomasum was penetrated. Asubsequent diagnosis of impactedabomasum was made and surgeryrevealed a grossly distended aboma-sum laying on the floor of the abdo-men. Both cases were able to returnhome after treatment and in neither

TABLE Ill2 x 2 TABLE OF PERITONEAL TAP DAIA AS A TEST FOR PERIT-ONITIS

Peritonitisa Normalb Total

Test Positivec 21 0 21Test Negatived 0 19 19

Total 21 19 40

True Positive Rate 100%;False Negative Rate 0C%False Positive Rate 0';cTrue Negative Rate 100%'All cows in peritonitis group (3).hAll cows in normal group (2).'Positive test < 10% eosinophils and > 40%ic neutrophils.dNegative test > 10%, eosinophils and < 40% neutrophils.

TABLE IVFREQUENCY 0F ISOLATION OF BACTERIA SPECIES IN PERITFONEAL FLUID SAMPLES

Group I Group 2 Group 3Organism (Parturient) (Normal) (Peritonitis)

Bacillus sp. 0 5 3Staphllococcus epidlernfidis 0 4 6Streptococcus sp. 0 2 2Cor,vnehacterium sp. 0 0 ICori nehacterium pyogenes 0 0 1Escherichia (oli 0 0 4Enterohacter sp. 0 I IClostridlium sp. 0 2 2Actinomyces sp. 0 0 IGram-negative anaerobes 0 0 3No growth 19 8 5

case did we detect side effects directlyattributable to the penetration. Thesmall amount of available informationsuggests that this procedure is notassociated with undue complications.However, it is the authors' opinionthat this method of collecting abdomi-nal fluid should only be used if moreconservative methods have failed toyield an adequate sample, and grossdistension of abdominal organs can-not be detected by rectal palpation.There were few gross differences

between samples of peritoneal fluidfrom healthy and diseased cows. Allnormal and most diseased samplesclotted readily. This could not berelated to the total protein content. Allthe cows in the parturient group had ahigh volume of peritoneal fluid whichalso clotted readily, despite having asignificantly lower total protein. Aprevious report (1) indicated thatnormal peritoneal fluid is clear andwatery with only a small amount pres-ent; pathological fluids were describedas increased in volume, cloudy and yel-low to bloody in color. The findings ofthis study do not support these obser-

vations. It has been stated that a perito-neal fluid sample with greater than6 x 109 WBC/L and total proteingreater than 3 g/dL was consistentwith the diagnosis of peritonealinflammation in 86% of cases (2).Although this statement is supportedby the findings in this paper, it wouldalso be true of 50% of the normalgroup in this series, making it a poordiagnostic criterion. The WBC countsin this series have mean values of 18.3 x109/ L in the peritonitis group and12.2 x 109/L in the normal group.However, median values were verysimilar (8.8 and 8.1 respectively, Fig-ure 4). Therefore, only in extremecases is WBC count of peritoneal fluida clear diagnostic sign of peritonitis.Similarly, total protein in peritonealfluid is not a useful indicator of perito-nitis in the cow. Fibrinogen levelsmeasured on peritoneal fluid arehigher in cows with peritonitis but thecomplete overlap of the ranges pre-vents its use as a good diagnostic testfor peritonitis.The most useful criteria for classify-

ing peritoneal fluid samples in this

study were the relative numbers ofeosinophils and neutrophils. By con-sidering two variables, i.e. both eosino-phils < 10% and neutrophils . 40%,simultaneously, a better separation ofperitonitis cases and normal cows wasachieved than by the use of either vari-able on its own. This same classifica-tion was also useful in differentiatingbetween parturient cows and thosewith peritonitis.From Tables 1I and III it can be seen

that peritoneal fluid analysis is slightlybetter than the measurement ofplasma fibrinogen as a diagnostic testfor peritonitis. In addition, highplasma fibrinogen may be associatedwith many other diseases, including fatnecrosis, hepatitis, mastitis, pericardi-tis, pneumonia, metritis, displacedabomasum and abscesses (9). Thisseems to indicate that peritoneal fluidanalysis would be a better diagnostictest of peritonitis than plasma fibrino-gen determination.The classification procedure used in

this study relies heavily on the absenceof eosinophils; previous reports havestated that eosinophils were uncommonin bovine peritoneal fluid (1,2,3). Thealmost ubiquitous occurrence of eosino-phils in this study is contrary to theseobservations. None of the blood sam-ples examined had eosinophil countsoutside the normal range. It is possiblethat the eosinophils may be present as aresult of some pathological process.For example, nematodes of the genusSetaria are commonly encountered inthe bovine abdomen. Setaria have beenreported in cattle from Colorado, Con-necticut, Florida, Georgia, Illinois,Maryland, Minnesota, Montana, NewYork, North Carolina, Pennsylvania,South Dakota, and Texas in the UnitedStates ( 12) and are also encountered inSaskatchewan.

Alternatively, eosinophils may bepresent in the peritoneal fluid of nor-mal, nonparasitized cattle. Studies inrats show that eosinophils are pri-marily distributed in subepithelialtissues of the digestive, urinary andrespiratory tracts (13). However,repeated washings of the peritoneumwith normal saline yielded an increas-ingly large percentage of eosinophils inthe fluid (14). Similar observationshave been made in guinea pigs and, inaddition, eosinophils were noted to bepresent in some cases before stimula-

79

tion with saline (15). Also, the study inguinea pigs showed increased neutro-phil numbers and decreased eosino-phil numbers with injection of manyantigens, including an Ascaris extract.A third possibility is that the eosino-

phils are a result of the samplingprocedure. The presence of a foreignbody in the peritoneal cavity for peri-ods up to ten minutes could inducechanges of cell type in the sample. Thistheory is not supported by the obser-vation of large numbers of eosinophilsin samples collected in less than oneminute, and by a different methodfrom parturient cows.

Generally used criteria, such asWBC count and total protein, provedunreliable for diagnosis of peritonitis.A larger body of information will berequired before definite statements ondiagnosis and prognosis based on peri-toneal fluid analysis can be made.

REFERENCES

1. OEHME FW. Cytologic examination of theperitoneal fluid in the diagnosis of cattledisease. J Am Vet Med Assoc 1969; 155:1923- 1927.

2. OEHME FW. NOORDSLY JL. Examination ofperitoneal fluid in differential diagnosis ofbovine diseases. Vet Med Small Anim Clin1970; 65: 54-59.

3. HIRSCH VM. TOWNSEND HGG. Peritoneal fluidanalysis in the diagnosis of abdominal dis-orders in cattle: A retrospective study. CanVet J 1982; 23: 348-354.

4. KANEKO JJ. SMITI'H R. The estimation ofplasma fibrinogen and its clinical signifi-cance in the dog. California Veterinarian1967; 21: 21-24.

5. SARD DM. Dealing with data: The practicaluse of numerical information (10) Analysisof rank. Vet Rec 1979; 104: 160-162.

6. DANIEL WW. Applied non-parametric statis-tics. Boston: Houghton Mifflin Co, 1978.

7. DUNN OJ. Multiple comparisons using ranksums. Technometrics 1964; 6: 241-252.

8. WEINSTEIN MC. FINEBERG HV. Clinical deci-

sion analysis. Philadelphia, London,Toronto: WB Saunders Co, 1980.

9. McSHERRY BJ, HORNEY Ft. DE GROOT JJ.

Plasma fibrinogen values in normal and sickcows. Can J Comp Med 1970; 34: 191-197.

10. BARRIGA oo. The immunology of parasiticinfections. Baltimore: University ParkPress, 1981.

I1. KAY AB. The role of the eosinophil. J AllergyClin Immunol 1979; 64: 90-104.

12. BECKLUND WW. WALKER ML. Taxonomy,hosts and geographic distribution of theSetaria (hematoda:filarioidea) in the UnitedStates and Canada. J Parasitol 1969; 55:359-368.

13. TAVASSOLI M. Eosinophil eosinophilia andeosinophilic disorders. CRC Crit Rev ClinLab Sci 1981; 16: 35-83.

14. BOSWORITH N, ARCHER GT. The eosinophilcontent of the peritoneal cavity of the rat.Aust J Exp Biol 1961; 39: 165-170.

15. CLEIGH GJ. LOEGERING D. Selective stimula-tion and purification of eosinophils andneutrophils from guinea pig peritonealfluids. J Lab Clin Med 1973; 82: 522-528.

BOOK REVIEW

Atlas of Skin Diseases of the Horse:Diagnosis and Treatment in EquineDermatology. L. F. M ontes and J.T.Vaughan. Published by W.B.Saunders Company, Toronto. 1983.202 pages. Price $85.80.

The Atlas of Skin Diseases of theHorse is an attempt to present somecommon equine dermatoses in a clearand concise manner, according to theauthors experience, and represents acollaborative effort between LeopoldMontes, a human dermatologist, andT. Thomas Vaughan, a recognizedveterinary clinician. In their introduc-tion the authors state clearly that itwas not their intention to produceeither a textbook or a treatise onequine dermatology. The text on thewhole conforms well to the looselyaccepted definition of an "Atlas",being replete with illustrations andcontaining limited text, however thecomprehensive though superficial,treatment which one may expect of apublication in this style is not fullyachieved.The presentation ofthe book is ofan

extremely high quality, being wellbound and printed on a high quality,glossy bond. The layout is very tidyand the liberal use of space makes thebook very easy to scan. The organiza-tion follows mainly an etiologic cate-gorization of skin diseases with somearrangement by topography/ anatomi-cal structure. A brief review of struc-ture and function of the skin as anorgan, and glossory of terms lead off,and the text concludes with a brief dis-cussion oftreatment. Within each etio-logic category, diseases and their man-agement are described by presentationof one or more case examples, eachlavishly illustrated by full-colour glossand histopathologic photographs.With the exception of some of thehigh-power photomicrographs whichlack definition, the photography is of ahigh quality.The principal strengths of the book

lie in the quality of presentation, thequality of the photographic reproduc-tions and the succinctness of the text.Detracting from these strengths areshortcomings in content and oftensome limitations in the use of English.

In reviewing structure and functionof the skin the text uses some excellentand fascinating scanning electronmicrographs. Since most veterinariansview the structure of skin histologi-cally however, more emphasis in thisarea would have been valuable, espe-cially with regard to the detailed struc-ture of the epidemis, which reflects somany of the changes occurring in theskin disease. The glossary is very use-ful for the uninformed, though it con-tains mistakes. Before proceeding tocase presentation, a brief review of thelimited responses which skin can maketo insult and of their significancewould have helped greatly in promot-ing understanding of the specificdermatoses.

In the case presentations some con-ditions, especially those for which asurgical correction is frequently ap-plied appear to be over-emphasized,while little material is presented onparasitic skin diseases, and immune-mediated conditions are not covered atall. While references are provided ineach section, they primarily support

Cont'd on page 89

80