Effects of Castration on Peritoneal Fluid in the Horse

John Schumacher, DVM, MS, Jim Schumacher, DVM, MS, J. S. Spano, DVM, PhD, J. McGuire, PhD,

W. L. Scrutchfield, DVM, MS, and R. G. Feldman, DVM, MS

Twenty-four clinically normal horses were castrated by routine methods. Peritoneal fluid was collected prior to castration and at I , 3,5, and 7 days postcastration. Peritoneal fluid was collected on days 9 and 11 if nucleated cell (NC) counts were still markedly elevated on day 7. Peritonitis, defined as NC counts > IO,OOO/pl, was evident in 15 horses following castration. Mean NC counts peaked on day 5 but were <lO,OOO/pl for 74% of the horses by day 7, and 90% of the horses by day 9. One horse had a NC count > 60,00O/pl on day 11 when sampling ended. Postcastration peritoneal fluid was obviously blood-tinged in 21 horses. Peak RBC counts occurred on day 3 but markedly decreased by day 5. Elevated peritoneal RBC counts correlated well with elevated NC counts (P < 0.001). Horses with peritonitis tended to have fever (P < 0.05). Other clinical signs of peritonitis were not apparent. (Journal of Veterinary Internal Medicine 1988; 2:22-25)

BECAUSE the parietal tunic that envelops the testes is an outpouching of the peritoneum, it is possible that surgical trauma to the parietal tunic, such as castration, could result in a change in peritoneal fluid. If such change does occur, the diagnostic usefulness of abdo- minocentesis would be compromised if examination of peritoneal fluid is required in a recently castrated horse. Horses that develop clinical signs of disease such as fever, anorexia, depression, or colic following castration may undergo abdominocentesis to determine if signs are due to peritonitis. This study was undertaken to estab- lish if open castration causes any change in peritoneal fluid and, secondly, to provide data that would improve interpretation of peritoneal fluid values in horses re- cently castrated.

Materials and Methods

Routine open castration was performed on 24 normal horses with a mean age of 25 months (range, 9 months

From the Departments of Large Animal Surgery and Medicine (John Schumacher), and Pathology and Parasitology (Spano), College of Vet- erinary Medicine, Department of Research Data Analysis (McGuire), Auburn University, AL 36849; and the Departments of Large Animal Medicine and Surgery (Jim Schumacher, Scrutchfield) and Veterinary Pathology (Feldman), College of Vetennary Medicine, Texas A&M University, College Station, TX 77843.

Supported in part by grants from the C. F. Underwood Trust and L. Hemn.

Reprint requests: John Schumacher, DVM, MS, Department of Large Animal Surgery and Medicine, College of Veterinary Medicine, Auburn University, Auburn University, AL 36849.

to 1 1 years). On the day of but prior to castration, perito- neal fluid samples were collected from the horses with a sterile 5- or 7.5-cm blunt-tipped bovine teat infusion cannula inserted through a stab incision made with a No. 15 surgical scalpel blade. The entry site was the midline or several centimeters to its right on the most dependent part of the abdomen. The horses were then castrated either in lateral recumbency (14 horses) or in the standing position (10 horses) depending on surgeon preference. In all cases, open castration was aseptically performed by a technique previously described,' using White's Improved, Reimer, or Serra emasculators (Jor- gensen Laboratories Inc., Loveland, CO). The vascular cords, but not the musculofibrous cords, of two horses were doubly ligated with #1 chromic gut before ampu- tation of the cord with an emasculator. The horses were not given antibiotics or anti-inflammatory drugs, preop- eratively or postoperatively.

Peritoneal fluid was serially collected postcastration. The site of abdominocentesis was at least 2 cm from previous puncture sites.

Peritoneal fluid samples were evaluated for specific gravity (sp gr) using a hand-held temperature-compen- sated refractometer ( A 0 Scientific Instruments, Buffalo, NY). Peritoneal fluid RBC counts, total and differential NC counts, as well as total and differential blood leuko- cyte counts, packed cell volume (PCV), and peripheral blood fibrinogen concentration were determined 1, 3, 5, and 7 days postcastration in all horses. Evaluations were made on peritoneal fluid and peripheral blood collected on day 9, and on day 11 for horses whose peritoneal NC

22

Vol. 2 . NO. 1 CASTRATION AND PERITONEAL FLUID IN THE HORSE 23

count was above lO,OOO/pl on the previous sample. An enterocentesis inadvertently occurred during abdomi- nocentesis of one horse on day 7 postcastration. Sam- pling of this horse’s blood and peritoneal fluid was dis- continued. Fibrinogen concentration was determined by calculating the difference between total protein values of serum and plasma. Total WBC counts on peripheral blood and NC counts on peritoneal fluid were per- formed using a particle counter (Coulter Counter, Coulter Electronics Inc., Hialeah, FL). Differential WBC counts from peripheral blood were made. Differ- ential cell counts on peritoneal fluid were made from cytocentrifuge preparations (Cytospin 2, Shandon Southern Instruments Inc., Sewickley, PA). All slides for differential counts were stained using Wrights-Leish- man stains. Monocytes, mesothelial cells, and macro- phages in the peritoneal fluid were classified collectively as mononuclear cells. The diagnosis of peritonitis was based on a peritoneal NC count exceeding 10,000/p1.2

Statistical Analyses

The relationships of peritoneal fluid constituents values and peripheral blood cell values were determined by computing Pearson correlation coefficients for all possi- ble combinations of these values3 The effects of position of castration, days after castration, and horse age were determined by analysis of variance-covariance proce- dures used for data with unequal numbers of observa- tions in the s~bce l l s .~ A repeated measures (split plot in time) model with position of castration as the whole plot treatment, days postoperation as subplot effects, and age of the horses as a covariate was used. The effect of age of horse was determined to be not significant and was re- moved from the model. The model selected to use is as follows:

Yljkl = f PI + HI] + Dk $- PDlk + e,ki where:

YIJkl = the ith observation in the kth D class in the jth

u = the population mean when equal frequencies

PI = effect of the ith position, i = 1, 2.

H class within the ith P class.

exist in all subclasses.

HI, = effect of the jth horse within the ith position, j = 1, n, and is the appropriate error term for testing position effects, with mean = 0 and being normally distributed assumed.

Dk = effect of the kth day postoperation, k = 0, 1, 3, 5, 7.

PDlk = effect of interaction of the kth day with the ith position.

el,kl = random errors associated with D and PD ef- fects, assumed to be normally distributed with mean = 0.

The assumption of normal and independent distribu- tion of the error terms with means = 0, and common variance of means is required to compute valid tests of hypotheses. The error terms noted in the model are the appropriate variances to be used in testing of hypotheses under the assumptions that must be made in order to use analysis of variance p r o c e d ~ r e s . ~ . ~

The least squares means computed by use of this model are the best estimates of means possible to com- pute from data with unequal subclass numbers and are appropriate for testing of hypotheses using the error variances obtained from the above model. Details of computing these least squares means are described by H a r ~ e y . ~

Data collected on days 0, 1, 3, 5, and 7 were used in the analyses. Too few samples were obtained for later days for use in analyses. The means reported in Tables 1 and 2 are least-square means for days 0, I , 3, 5, and 7 and raw means of days 9 and 1 1 . Tests of hypotheses about days were thus limited to days 0, 1, 3, 5, and 7.

Results

Mean values of the peripheral neutrophil count peaked 24 hours postcastration and declined to normal values by day 5. The mean values of other peripheral white blood cells did not change during the sampling period. Mean plasma fibrinogen values were abnormally ele- vated on day 3 and had not returned to normal levels by day 7 (Table I ) .

Sixty-three percent ( 1 5 of 24) of the horses developed peritonitis after castration. The mean NC count of peri- toneal fluid was abnormally elevated 1 day after castra- tion and continued to increase until day 5 postcastra- tion. By day 7, the mean peritoneal NC count had de- creased to nearly 1 O,OOO/pl (Table 2). Seventy-four percent ( 1 7 of 23) of the horses had a NC count < 1 O,OOO/pl by day 7. Two horses had NC counts above 20,OOO/p1 on day 1 1 when sampling ended. Neutrophils were the main cell type contributing to the total count, with lymphocytes and mononuclear cells making small contributions (Table 2). Neither toxic or degenerative changes in the neutrophils nor bacteria were observed in any of the peritoneal fluid samples.

The peritoneal fluid was blood-tinged in 21 of 24 horses the day after castration. By day 5, the peritoneal fluid was blood-tinged in 1 1 horses, and by day 7 , 7 of 24 horses still had blood-tinged peritoneal fluid. The perito- neal fluid of the two horses whose vascular cords were ligated was markedly blood-tinged the day after castra- tion. The mean values of the peritoneal RBC count peaked on day 3 (Table 2).

Mean specific gravity of abdominal fluid was elevated above the baseline values on every sampling day (Table 2) but did not rise above established normal values for our laboratories. RBC count values of peritoneal fluid

N

P

TA

BL

E I.

Mea

ns o

f Pac

ked

Cel

l Vol

umes

. Per

iphe

ral W

hite

Blo

od C

ell C

ount

s. F

ibri

noge

n V

alue

s. an

d Bo

dy T

empe

ratu

res

of 2

4 H

orse

s Sub

ject

ed to

Cas

trat

ion

Fi hr

i n og

e n

Tem

pera

ture

D

ay

N

PCV

W

BC

Neu

trop

hils

L

ymph

ocyt

es

Mon

ocyt

es

Eosin

ophi

ls (m

g/dl

) (“0

0 24

3

4f

I 10

.432

f 4

49

5,84

5 f 3

41

4.18

4f 2

56

212 f 3

9 18

2 It 4

2 38

9 t 2

7 38

.4

I 24

3

4f

I 13

.952

f 7

61

9.37

9 f 5

62

4.02

5 f 3

56

160 f 6

3 37

8 f 3

3 38

.6

383 f 7

1 (2

6-45

) (5

.600

-15.

800)

(4

.256

- 10

.902

) ( I

. 120

-7.3

92)

(0-8

54)

(0-7

50)

20

0-60

0 .

37.5

-39.

2

(27-

44)

(8.9

00-2

3.40

0)

(4.7

17-1

5.21

0)

(525

-7.8

44)

(0-1

.208

) (0

-1.1

84)

100-

700

37.2

-40.

I 3

24

34

f 1

12.2

34 f 6

66

7.53

1 f 4

64

4.20

0 f 3

64

255 It 4

3 21

2 t 4

2 44

9 f 2

8 38

.4

(25-

43)

(4.1

00-1

7.00

0)

(2.6

65-

I 1.0

50)

( I. 1

89-8

.692

) (0

-710

) (0

-756

) 20

0-70

0 36

.7-4

0.0

5 24

3

4?

1 11

.265

? 5

33

6.37

3 f 5

21

4.36

7 f 2

77

284 f 6

9 21

6 f 5

3 46

8 2 4

8 38

.4

7 23

3

3f

I 11

.558

f 5

01

6.38

8 f 3

90

4.53

0 f 4

09

260 f 5

6 34

9 f 6

5 45

0 f 3

1 38

.4

9 7

32 f 2

12

.842

f 6

92

6.7

I7 f 9

54

5.26

8 f 8

23

407 f 8

1 43

1 f 1

78

200k

71

37.9

II

5 3

4f

I 12

,740

f 8

45

6.70

3 f 5

78

5.32

8 f 6

57

467 f 1

29

213 t 5

4 56

0 ~

t 98

38. I

(26-

48)

(6,7

00-1

7.80

0)

( I .4

62-

12.3

88)

( I .4

7 1-

6.70

8)

(0- I

.424

) (0

-954

) 30

0- I

.000

37

.2-3

9.4

(27-

5 I)

(6

.200

- 16

.100

) (2

.666

- 10

.725

) (1

.617

-9.1

77)

(0-8

58)

(0- 1

,298

) 10

0-70

0 37

.2-3

9.7

(28-

39)

( 10.

800-

15.

600)

(4

.104

- I 1

.622

) (2

.384

-9.2

04)

(129

-780

) 10

0-40

0 37

.8-3

8. I

(32-

36)

( 10.

400-

I 5,

300)

(5

,520

-8.8

74)

(3.0

16-7

.038

) (2

50-9

66)

(125

-416

) 40

0-80

0 38

.0-3

8.3

(0-1

.152

)

(I) 0 I

c

0 I

m

D

rn

--I D

!-

Dat

a ex

pres

sed

as m

ean f S

EM o

f cel

ls/Fl

(m

in - m

ax)

N =

Num

ber

of h

orse

s. Le

ast S

quar

es m

ean

for

days

0-7

and

num

eric

al a

vera

ges f

or d

ays 9

and

1 I.

5

TA

BL

E 2. M

eans

of R

BC a

nd N

ucle

ated

Cel

l Diff

eren

tial C

ount

s of P

erito

neal

Flu

id F

rom

24

Hor

ses

Subj

ecte

d to

Cas

trat

ion

Day

N

R

BC

N

CC

N

eutr

ophi

ls

Lym

phoc

ytes

M

onoc

ytes

Eo

sinop

hils

Sp G

r

0 24

7.

903 f 2

.484

2.

728 f 3

96

1.78

5 f 3

37

46 f 3

2 9

69

f 16

6 6

+-5

1.

012 2 ,0

006

(0-5

5.1

00)

(4

52-8

.600

) (3

6-6.

622)

(0

-750

) (0

-750

) (0

-1 1

7)

1 24

49

1.13

7 f 3

69.5

84

18,6

44 f 8

.147

15

.2 10

f 7

.323

26

6 f 2

49

3.12

7 f 8

35

33

+ 1

8 1.

015 f ,0

01 I

3 24

84

6.20

5 f 4

78.0

22

25.6

63 f 8

.84

I 22

.757

f 8

. I32

67

8 f 5

36

2,68

1 f 8

45

I1 f

6

1.01

6 k .0

010

(300

-9 X

lo6)

(2

.422

-200

.000

) (2

9 1 -

180.

000)

(0

-6.0

00)

(0- 1

4.00

0)

(0-3

94)

(0-9

X lo6

) (2

,400

-165

.000

) ( 1

,800

- 157

,320

) (0

-1 2

.680

) (0

-14.

274)

(0

-131

) 5

24

169,

546 f 5

0.02

8 30

, I77

f 1

8.36

3 26

.576

f 1

6.51

0 73

0 f 7

28

2.84

7 f 1

.348

3

7f

19

1.01

6 2 ,0

010

(0-8

40.5

00)

(2.0

00-4

20.5

00)

( 1.2

00-3

78.4

50)

(0-1

6,82

0)

(0-2

5.23

0)

(0-3

10)

7 23

18

1.10

2 f 1

49,4

40

12,2

59 f 4.

002

10.2

76 k

3.7

27

302 f 1

66

1.31

I f 1

89

105 f 8

1 1.

015t

.001

0

9 6

244,

967 f 1

74.7

97

21,6

67 f 1

0,78

0 16

,743

f 10

.154

53

7 f 4

07

3.90

3 f 1

,280

45

6 k 3

12

1.01

6 t .0

019

(0-3

.46

X lo

6)

(1.0

2 1-

92.7

00)

(36 1

45,2

84)

(0-3

.708

) (5

2-2.

889)

(0

-1,8

54)

(0-1

.1 x

106

) (2

.800

-74.

500)

(2

.240

-67.

050)

(0

-2.4

80)

(560

-8.3

30)

(0- I

.900

) b

II

5 33

.797

f 3

I .23

1 20

.912

f 1

1,44

9 17

.560

f 1

1,28

7 70

f 7

0 3.

052 f 6

98

229 f 2

14

1.01

5 * .0

018

- c,

(0

- 158

.600

) (4

.440

-65.

0 19

) (2

.220

-6 1

.768

) (0

-350

) (1

.216

-5.4

25)

(0-

1.08

5)

g[u

3

0,

%<

$3

E

Z

2.2

33

Dat

a ex

pres

sed

as m

ean f S

EM o

f cel

ls/pl

(M

in - M

ax).

N =

Num

ber

of h

orse

s. Le

ast S

quar

es m

ean

for

days

0-7

and

num

eric

al a

vera

ges f

or d

ays 9

and

I I.

a 2

.

Vol. 2 . NO. 1 CASTRATION AND PERITONEAL FLUID IN THE HORSE 25

were directly correlated to NC counts ( P < 0.00 1). There was no correlation between peritoneal neutrophil counts and the peripheral neutrophil count ( P > 0.05). The position of castration or age of the horse did not signifi- cantly ( P > 0.05) influence the peritoneal fluid NC or RBC counts.

Eleven of the horses had an elevated body tempera- ture (range, 38.4 to 40. I "C) on at least one of the sam- pling days postcastration. There were significant posi- tive correlations ( P < 0.05) between peritoneal RBC and NC counts with body temperature.

All horses maintained a normal appetite and attitude. Even though the horses were confined to stalls postcas- tration, preputial edema was not considered excessive in any horse. No clinical abnormalities were observed in any of the horses during the 2 months immediately postcastration.

Discussion

Clinical peritonitis is purportedly an infrequent compli- cation of castration of the horse.5 In this study, peritoni- tis commonly occurred postcastration, but clinical signs other than fever were not observed. Because necropsies ofthe horses were not performed, it is uncertain whether the peritonitis was localized or diffuse. Although bacte- rial culture of peritoneal fluid was not done, neither bacteria nor degenerate neutrophils were noted in any postcastration peritoneal fluid analysis. This suggests that the influx of neutrophils was not a response to an intraperitoneal bacterial inoculum.6 A previous study has shown that serial collection of peritoneal fluid by the technique used in this study causes no significant in- crease in the peritoneal NC count.'

Nearly all the horses had blood-tinged peritoneal fluid postcastration. The close correlation between peritoneal RBC and NC counts suggests that hemorrhage may, at least in part, be responsible for the influx of white blood cells into the abdominal cavity. In man, free blood in contact with the peritoneum can cause pe r i t~n i t i s .~ ,~

This study and others indicate that the equine perito- neum is capable of tremendous inflammatory reaction to seemingly minor insult^.^^'^ Therefore, a high perito-

neal NC count may not indicate a severe peritonitis or the need for treatment with antimicrobials or peritoneal lavage.

Peritonitis likely occurs in many horses as a result of castration. Since peritonitis may be related to postopera- tive intra-abdominal hemorrhage, improperly adjusted emasculators or poor surgical technique may increase the likelihood of peritoneal irritation. The decision to treat horses with postcastration peritonitis should be in- fluenced more by the severity of clinical signs and the character of the peritoneal fluid than by the RBC or NC count. The finding of toxic or degenerative neutrophils or bacteria in the peritoneal fluid indicates the need for antimicrobial therapy and possibly peritoneal lavage. This study demonstrated that increased peritoneal NC count postcastration is not necessarily an indication of clinically significant peritonitis.

References

1. Vaughan JT. Surgery of the testes. In: Walker DF. Vaughan JT. Bovine and Equine Urogenital Surgery. Philadelphia: Lea & Febiger, 1980: 145- 169.

2. Ricketts SW. Peritonitis. In: Robinson NE. ed. Current Therapy in Equine Medicine 2. Philadelphia: WB Saunders, 1987: 79-8 1.

3. Steel GD. Torrie JH. Principles and Procedures of Statistics. New York: McGraw-Hill, 1960: 183-276.

4. Harvey WR. Least squares analysis of data with unequal subclass numbers. Washington, DC: Agricultural Research Service-

5. Vaughan JT. Peritonitis and acute abdominal diseases. In: Ander- son NV, ed. Veterinary Gastroenterology. Philadelphia: Lea & Febiger, 1980: 651-573.

6. Adams SB. Fessler JF, Rebar AH. Cytologic interpretation of peri- toneal fluid in the evaluation of equine abdominal crises. Cor- nell Vet 1980; 70:232-246.

7. Schumacher J, Spano JS, Moll HD. Effects of enterocentesis on peritoneal fluid constituents in the horse. J Am Vet Med Assoc 1985; 186: 1301-1 303.

8. Shearman DJC, Finlayson NDC. Diseases of the Gastrointestinal Tract and Liver. New York: Churchill Livingstone, 1982; 365.

9. Aach RD. Abdominal pain. In: Blacklow RS. ed. MacBryde's Signs and Symptoms: Applied Pathologic Physiology and Clinical Interpretation, 6th ed. Philadelphia: JB Lippincott.

10. Blackford JT, Schneiter HL, VanSteelhouse JL, et al. Equine peri- toneal fluid analysis following celiotomy. In: Proceedings of the Equine Colic Research Symposium, Vol 2. Lawrenceville. NJ: Veterinary Learning Systems, 1986; 130-1 33.

USDA, ARS H-4. 1975: 85-1 12.

1983: 165-179.