I'll

 

 

in‘vi’!

z .2 a
«mun “fir;

.
. am.
k . aav.%..
,z: . . 0 :4; E
Lighllahv ‘ I? 3.31.. ..
93$. . 3.4g .
1“ . n.
s ..
mum: 95:5... .mm
.: rm... .mmuumyw. _
_.

£3135.
. ..
~£

Vi.»
.y .1
0|!

KW .
I. 5::th
1..

x3) rivflvgtnfl .
' .

1(1):. .8;
u)!- hffi.‘
. x

.311.

I)". 1
5.1.5! \rlltun.
{-1. .I x

0‘35?-

L... .22... .n:

m
3 LIBRARY
700w Michigan State

University

This is to certify that the
thesis entitled

AN OBJECTIVE METHOD OF ASSESSING SAND
ACCUMULATION ON EQUINE ABDOMINAL RADIOGRAPHS

presented by

Nathan James Keppie

has been accepted towards fulfillment
of the requirements for the

Master of degree in Large Animal Clinical Sciences

 

 

Science
(\l c5 {WM/7?: WW

Major Professor‘s Sighatfi re

 

Date

MSU is an Affirmative Action/Equal Opportunity Institution

PLACE IN RETURN Box to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 c:/ClRC/DateDue.indd-p.15

AN OBJECTIVE METHOD OF ASSESSING SAND ACCUMULATION ON EQUINE
ABDOMINAL RADIOGRAPHS

By

Nathan James Keppie

A THESIS
Submitted to
Michigan State University

in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE
Department of Large Animal Clinical Sciences

2005

ABSTRACT

AN OBJECTIVE METHOD OF ASSESSING SAND ACCUMULATION ON EQUINE
ABDOMINAL RADIOGRAPHS

By
Nathan James Keppie

Abdominal radiography has been utilized as a primary method of diagnosing sand
colic; however, assessing such radiographs is currently highly subjective. Variations in
interpretation by this subjective method influence the diagnosis, treatment and potentially
the outcome of affected horses. In order to improve the diagnostic utility of radiography
in cases of sand colic, a more objective assessment method was designed and tested by
retrospectively examining previous clinical cases of sand colic. The objective method
was determined to be more reproducible and accurate for diagnosing and assessing sand
colic than current subjective methods. This objective method was then tested on 10
horses that were experimentally administered known quantities of sand over a period of
time and radiographed. The newly developed objective method of assessing sand
accumulation again proved to be accurate and was able to differentiate different
quantities of sand whereas subjective means could not. The objective assessment method
provides an ordinal scale ranging from 0-12 that attributes a relative quantity/severity of
sand accumulation with a score of 7 or higher having an 83% likelihood of being
associated with a diagnosis of sand colic. This method is recommended to be used when
assessing equine abdominal radiographs upon initial examination for sand colic and for

re-evaluation of the patient following treatment.

ACKNOWLEDGMENTS

I gratefully acknowledge the following people who helped me throughout this
research project. First and foremost, I acknowledge my advisor, Diana Rosenstein, who
helped me everywhere along the way and always provided much needed enthusiasm and
mental and physical support. I extend my appreciation for my committee members Sue
Holcombe and Hal Schott for their interest, expertise and guidance throughout. Dr.
Carleton provided me with the eight mares for the experimental portion of the study and
is gratefully appreciated. Dr. Schott and Dr. Stick provided me with 2 more horses to
complete the study and are acknowledged. Hammond Farms graciously donated the
much needed sand for the research. The Department of Plant and Soil Sciences analyzed
the sand samples and are thanked. I could not have completed this project without the
help fiom the Stats Help Desk, Department of Plant and Soil Science for the statistical
analysis. I deeply thank everybody in the Radiology Department and the Large Animal
Clinic who put up with me and helped me out by answering to my beckoning call. And
Danielle Eifler, who put in tremendous time helping me with the experimental procedure

and supported and lived with me throughout the entire time with a smile.

iii

TABLE OF CONTENTS

LIST OF TABLES ................................................................................... vi

LIST OF FIGURES ................................................................................. vii

CHAPTER 1- Retrospective study and development of an objective assessment

for sand accumulation on equine abdominal radiographs

Introduction ................................................................................... 1
Materials and Methods ..................................................................... 12
Results ....................................................................................... 1 8
Discussion ................................................................................... 3 1
Conclusion .................................................................................. 42

CHAPTER 2- Prospective study and testing of the developed objective assessment

on experimental horses

Introduction ................................................................................. 43
Materials and Methods ..................................................................... 44
Results ....................................................................................... 52
Discussion ................................................................................... 80
Conclusion ................................................................................. 1 18
REFERENCES .................................................................................... 121
APPENDIX A- Tabulated sand parameters ................................................... 125

APPENDIX B- Individual and group average graphical representation of sand

output over time ................................................................ 130

APPENDIX C- Subjective and objective scores (with breakdown for each

parameter) for each horse ...................................................... 136

APPENDIX D- Individual horse and group average objective score trends

over time ......................................................................... 146

APPENDIX E- Individual horse objective scores based on weight of intra-

abdominal sand ................................................................. 152

APPENDIX F- Combined Group 1 and Group 2 objective scores based on

weight of intra-abdominal sand on a given day ........................... 157

APPENDIX G- Individual group objective scores based on weight of intra-

abdominal sand on a given day ............................................... 163

iv

TABLE OF CONTENTS (CON’T).

APPENDIX H- Individual objective parameter scores based on weight
of intra-abdominal sand ....................................................... 175

APPENDIX I- Scatter plots comparing scores of each objective parameter
to another ........................................................................ 178

LIST OF TABLES

TABLE 1-1: Signalment and diagnosis (sand colic or other) distribution
of horses used for the retrospective study ...................................... 25

TABLE 1-2: Correlation of abdominal radiographic parameters to a
diagnosis of sand colic ............................................................ 27

TABLE 1-3: Subjective and objective scores assigned to abdominal
radiographs for sand accumulation as observed by four
reviewers ............................................................................ 28

TABLE 2-1: Percentage of sand passed over a six-day period following

nasogastric administration of 4 g/Kg for the first two days.

(Based on a 94% recovery rate) ................................................ 117
TABLE 2-2: Tabulated sand parameters ...................................................... 125

TABLE 2-3: Subjective and objective scores (with breakdown for each
parameter) for each horse ....................................................... 136

vi

FIGURE l-l :

FIGURE l-2:

FIGURE l-3:

FIGURE 1-4:

FIGURE l-5:

FIGURE 2-1:

FIGURE 2-2:

FIGURE 2-3:

FIGURE 2-4:

FIGURE 2-5:

FIGURE 2-6:

FIGURE 2-7:

LIST OF FIGURES

An example of an abdominal radiograph showing a sand
accumulation in the colon of a horse ........................................... l 1

Objective radiographic assessment scoring sheet for sand
accumulation ....................................................................... 1 7

An example of an abdominal radiograph showing a small sand
accumulation in the diaphragmatic flexure of a horse ........................ 22

An example of an abdominal radiograph showing a moderate
sand accumulation in the ventral colon of a horse ............................ 23

Two consecutive abdominal radiographic images demonstrating
an example of a large accumulation of sand in the ventral colon of a
horse ................................................................................. 24

Schematic representation of the 12 areas radiographed from
each side of the horse’s abdomen using hi gh-intensifrcation digital
fluoroscopy ........................................................................ 51

One of three mid-abdominal accumulations of sand in Horse #10
(radiograph view #6) on Day 5 .................................................. 60

Radiograph (View #4) of sand within the stomach (Day 0.5) following
nasogastric intubation and administration of 4 g/Kg sand ................... 61

Typical appearance of the craniodorsal and mid-dorsal abdomen
(radiographs # 7, 10, and l l) demonstrating gas-filled intestines,
ribs, and vertebrae ................................................................. 62

Sand accumulation within the ventral colon (radiograph view
#2) illustrating haustra as vertical radiolucent bands ......................... 63

Typical appearance of the mid ventral colon (radiograph
number 3) ........................................................................... 64

Radiograph of sand within the cranioventral abdomen (radiograph

view #1) illustrating the diaphragmatic flexure and the stemal

flexure with the latter being slightly more caudal and ventral than

the former ........................................................................... 65

vii

LIST OF FIGURES (CON’T).

FIGURE 2-8: Typical appearance for sand accumulation within the mid-ventral

FIGURE 2-9a:

FIGURE 2-9b:

FIGURE 2-10:

FIGURE 2-1 1:

FIGURE 2-12:

FIGURE 2-13:

FIGURE 2-14:

FIGURE 2-15:

FIGURE 2-16:

FIGURE 2-17:

FIGURE 2-18:

FIGURE 2-19:

FIGURE 2-20:

FIGURE 2-21:

colon (radiograph view #2) of Group 2 horses illustrating a large,
well-defined opaque and homogeneous accumulation ....................... 66

Typical appearance for sand accumulation within the cranioventral
colon (radiograph view #1) of Group 1 horses illustrating dispersed

opaque sand particles within digesta and no focal accumulation... . . .....67
Typical appearance for sand accumulation within the mid-

ventral colon (radiograph view #2) of Group 1 horses illustrating

small, poorly-defined heterogeneous accumulations ........................ 68
Radiograph of sand seen within the small colon (radiograph

view #12) illustrating opaque feces just cranial to the pelvis .............. 69
Group 1 and Group 2 average fecal sand output over time ................ 70
Group 1 and Group 2 average objective score trends over time ........... 71
Group 1 and Group 2 average sand location score over time .............. 72
Group 1 and Group 2 average score for number of sand

accumulations over time ........................................................ 73
Group 1 and Group 2 average sand opacity score over time ............... 74
Group 1 and Group 2 average sand homogeneity score over time ........ 75
Group 1 and Group 2 average sand length score over time ............... 76
Group I and Group 2 average sand thickness score over time ............ 77
Group 1 and Group 2 objective scores based on weight of intra-
abdominal sand .................................................................. 78
Individual and group average graphical representation of sand

output over time ................................................................ 130

Individual horse and group average objective score trends
over time ......................................................................... 146

viii

FIGURE 2-22:

FIGURE 2-23:

FIGURE 2-24:

FIGURE 2-25:

FIGURE 2-26:

LIST OF FIGURES (CON’T).

Individual horse objective scores based on weight of
intra-abdominal sand ........................................................... 152

Combined Group 1 and Group 2 objective scores based on
weight of intra-abdominal sand on a given day ............................ 157

Individual group objective scores based on weight of intra-
abdominal sand on a given day ............................................... 163

Individual objective parameter scores based on weight of
intra-abdominal sand ........................................................... 175

Scatter plots comparing scores of each objective parameter
to another ........................................................................ 178

CHAPTER 1
Retrospective study and development of an objective assessment for sand

accumulation on equine abdominal radiographs

Introduction

Chronic ingestion of sand can cause non-strangulating large intestinal
intraluminal obstruction (colic) or chronic diarrhea in horses. Sand ingestion occurs most
commonly in areas with loose sandy soil, with California, Florida, Michigan, Arizona,
Colorado, and other coastal regions reported as having the highest incidence. Sand
impactions alone account for 2.4% of all colics reported (White and Lessard, 1986).
Another epidemiological study in 1997 (Kaneene et al., 1997) found that in the state of
Michigan alone, sand colic represented 5% of the total colics investigated. The sand
found in clinical cases of colic and/or diarrhea is fine beach sand or clay, but gave] or
bluestone shale has also been implicated as a cause of sand-related disease. The volume
of sand required to produce clinical signs is not known but it has been reported that
greater than 20 kg is needed to cause clinical problems in horses of small size (Ragle,
Meaghner, Schrader and Honnas, 1989). However, sand found at surgery in cases of
sand colic weighed between 8-70 kg (Ragle, Meagher, Lacroix, and Honnas, 1989; Ford
and Lokai, 1979).

Sand may be present in the hay or may be ingested from the environment.
Environmental sand ingestion may be due to ground feeding if stabled in a sandy pasture,
grazing of grass covered by silt afier flooding, grazing of roots and attached soil of plants
uprooted in short or overly grazed pastures, or drinking from shallow or muddy pools at

times of fresh water unavailability. Horses, especially young foals, may deliberately

consume sand (pica) from boredom or salt deficiency. Horses may also scavenge more
frequently and ingest sand if their diet is too low in roughage.

It is debatable whether or not sand of different particle size accumulates in
different regions in the gastro-intestinal tract. Coarse sand has been reported to
accumulate in the right dorsal colon, transverse colon and pelvic flexure while fine sand
accumulates in the left and right ventral colon, and associated stemal and diaphragmatic
flexures. Hammock et al., (1998) reported that both fine and coarse grain sand was found
equally in all areas of the large intestine. Sand may collect in the large colon because it is
a fermentative area of reduced flow and the sand consequently settles out of the ingesta.
The pelvic flexure and transverse colon become obstructed when masses of accumulated
sand become lodged in these narrowed segments. Pathophysiologically, the sand
obstruction results in gas, secretions, fluid and ingesta accumulation in the intestine
proximal to the obstruction (Freeman, 2002). This is followed by secondary bowel
distension and mesenteric traction causing pain (colic). Obstruction induced spasm
(increased motility) occurs in the intestine proximal to the obstructed site, immediately
after it occurs, and is caused by localized intestinal irritation rather than distension, and
also causes pain (Allen, 1990). The intestinal wall veins are compressed as luminal
pressure increases, causing a rise in the capillary hydrostatic pressure and filtration rates.
This results in tissue edema and a net secretion of fluid into the intestine, leading to
mucosal ischemia, cellular necrosis, loss of epithelium and villi, and mucosal
hemorrhage. As the ischemia and inflammation destroys the integrity of the intestinal
epithelium, the outer wall of enterie gram negative micro-organisms gain access to the

circulation causing an acute phase response and fever (endotoxemia). Prolonged

ischemia, edema and distension atony can cause loss of intestinal contractility. Paralytic
ileus may also result from pain and over-stimulation from the sympathetic nervous
system that complicates the proximal intestinal distension. Long standing sand
impactions may result in a weakened and devitalized colonic wall that may predispose to
rupture of the intestine. The weight of the impacted mass can also displace and/or twist
the impacted segment of bowel resulting in a strangulating large intestinal obstruction.

Clinical signs of horses presenting with sand impactions are usually mild to
moderate in severity and may be intermittent. The obstructions are usually incomplete
and allow the passage of small amounts of dry, scant fecal material that is dehydrated
fecal balls covered in mucous due to delayed passage. Dehydration is mild at first
because water is still passed and absorbed in the cecum. The obstruction may become
complete resulting in rapid accumulation of gas and ingesta with subsequent distension
and more severe signs of colic.

Sand ingestion may also cause diarrhea where the horses present with signs or a
history of inappetance, weight loss, mild dehydration, and chronic-recurrent colic.
Intestinal motility may be increased or decreased. Chronic diarrhea in adult horses is
almost always associated with large intestinal disease caused by either physical damage
to the colonic wall or physiological disturbances of colonic function (Mair, 2002). Sand
enteropathy (enteritis) is caused by sand-induced irritation to the colonic mucosa
disrupting the normal motility leading to malabsoptive diarrhea. The abrasiveness of sand
can cause mucosal damage leading to bacterial and endotoxin absorption. Damaged

mucosa also results in a decreased amount of fluid absorption, so fluid then accumulates

in the lumen. There may also be increased vascular permeability that causes fluid to leak
into the lumen. The subsequent fluid distension results in pain and diarrhea.

Sand impaction can be difficult to differentiate clinically from feed impaction,
and the diagnostic tests for sand impaction are prone to a high incidence of false positives
and false negatives (Snyder, 1992). To aid in the diagnosis for sand impaction, any
historical findings that suggest ground-feeding practices in an environment that has sandy
soil or historical signs of diarrhea followed by mild—moderate intermittent colic may
suggest current sand enteropathy.

Rectal palpation may reveal firm sand impactions in the colon, however, sand
accumulation in the dorsal right colon or transverse colon may be difficult to detect with
palpation. The weight of the impaction may drop the site of accumulation out of reach,
so rectal palpation is often not conclusive (Doran, 1993). As well, it is difficult to
differentiate sand impaction from firm digesta impaction via rectal palpation.

The history or presence of sand in the feces can indicate exposure to sand, but
does not indicate clinical disease as small amounts of sand can be found frequently in the
feces of healthy horses. Coarse sand may be visible or felt in the feces, however, fine
sand often requires sedimentation to be identified. Tests for the presence or volume of
fecal sand do not correlate well with the presence or volume of sand in the colon (Snyder,
1992). Presence of sand does not indicate sand obstruction, and absence of fecal sand
does not necessarily indicate that there is no colonic sand.

Auscultation of the ventral abdomen may reveal characteristic sand sounds. This
is described as a friction-like rub that is compatible with sand borborygmi. However, in

order to produce these sounds, there must be intestinal motility, and contact of sand-filled

bowel with the ventral abdominal body wall. Therefore, if sand is not in the ventral
colon, or if there is mechanical or paralytic ileus, these friction rubs will not be ausculted
(Ragle, Meaghner, Schrader and Honnas, 1989).

Abdominal radiography has proven to be a valuable tool in the diagnosis of sand
impaction (Korolainen and Ruohonierni 2002). Sand accumulations are often seen in the
ventral colon as opaque masses that lie in the cranioventral abdomen that tend to conform
in shape to the dependent surface of the colon (Figure l-l). Therefore, the ventral margin
of the mass may show sacculations, and the dorsal margin of the mass may be horizontal
if the sand has settled out from the ingesta, irregular if it is still mixed with ingesta, or if
severe, filling the entire colon (Fischer, 1997). Sand accumulations can be difficult to
distinguish from enteroliths if the shape of the accumulation is spherical, such as occurs
in the transverse colon (Fischer, 1997). There is currently, however, no established
criteria for assessment of visualized sand accumulations and correlating with a clinical
diagnosis. Presence and/or volume of visualized sand does not correlate well with
presenting clinical signs, presence of fecal sand, abdominal auscultation, or with surgical
findings from exploratory laparotomies (Bertone et al., 1988). Even large enteroliths
have been missed on abdominal radiography (Rose, Rose and Sande, 1980) as obstructive
changes may obscure radiopaque material. It is difficult to identify the exact location of
the sand accumulation (Hammock et al., 1998) and to differentiate mechanical from
functional ileus based on abdominal radiographs (Butler et al., 2000). Serial radiography
may be utilized to monitor the passing of intestinal sand following therapy, however, the
radiographic findings do not correlate well with clinical findings (Ruohonierni et al.,

2001).

In order to produce abdominal radiographs of diagnostic quality in adult horses, there is a
requirement for large stationary radiographic equipment with an exposure range of 90-
140 KVp and 180-600 mAs. High speed rare-earth screen/film combinations should be
used with a grid that has a grid ratio of 8:1 — 10:1 and 80-1301ines/cm. It is difficult, if
not impossible to obtain diagnostic abdominal radiographs from horses with an
abdominal width greater than 70 cm, and significant scatter radiation is produced from
any adult horse abdomen. So, the size of the patient and the capabilities of the
radiographic equipment available will determine what radiographs may be obtained. It is
recommended that the entire abdomen be radiographed as sand accumulation is not
always in the cranioventral region. If right to left lateral radiographs were taken, left to
right lateral radiographs are recommended in the specific area of concern as lesions may
become more or less prominent. In gesta may obscure sand accumulation so it is
recommended that horses be fasted at least 6 hours prior to abdominal radiography to
avoid false negative conclusions.

Abdominal ultrasonography may aid in the diagnosis of sand accumulation,
however, no structures deep to a gas-bowel wall interface can be imaged because of the
resultant acoustic shadowing. It may be difficult to differentiate intraluminal gas from
sand accumulation with ultrasonography as both cause attenuation of the soundwave.
There is also limited penetration of the sound beam to about 20-25 cm, and there is loss
of resolution at increasing depths.

A transducer with a low frequency range (5 MHz) may reveal sand impactions in
the ventral colon. The colon appears flattened against the ventral body wall with loss of

the normal sacculations (Reef, 1998). There is little peristaltic activity of the large colon

detected because it is weighed to the floor of the abdomen by the accumulation of sand.
The large colon appears as if it is adhered to the ventral body wall owing to the lack of
motility between these two structures. The sand grains on the mucosal surface of the
large intestine appear as small, pinpoint, hyperechoic structures casting small acoustic
shadows in varying directions. They may also be seen as starburst spicules that appear to
be floating due to their suspension in the ingesta. Korolainen and Ruohonierni (2002)
concluded that abdominal ultrasound was a good screening method for horses with signs
suggestive of sand accumulation with a specificity and sensitivity of 87.5%.

Once diagnosed, sand impaction can be treated medically where therapy is aimed
at controlling pain, maintaining hydration, and administering laxatives such as mineral
oil, dioctyl sodium sulfosuccinate, psyllium, and magnesium sulfate. These therapies
may also stimulate colonic motility via the gastrocolic response when instilled in the
stomach. Surgical intervention is indicated if systemic deterioration or peritoneal fluid
changes occur, or if pain cannot be controlled. Surgical intervention involves exploratory
laparotomy with enterotomies and removal of the impacted sand.

To prevent sand impactions (or diarrhea) from occurring, intermittent
administration of psyllium or bran to the feed has been advocated as a method of
reducing the risk. Feeding off of the ground or away from sandy soil and grazing in
pastures where there is adequate growth that would prevent horses from picking up sand
would also help reduce the risk. Horses that actively seek out sand may require a muzzle
or stall confinement.

Despite frequent occurrence of sand colic in certain geographic areas, there is a

paucity of published manuscripts that provide scientific information of its clinical signs,

diagnosis, treatment and prognosis. Most of the information available is anecdotal or
based on a limited number of cases (Ragle, Meagher, Lacroix, and Honnas, 1989).

Upon review of the literature, it is evident that there is poor correlation between
clinical, diagnostic and surgical findings, which causes problems in diagnosing sand colic
(Specht and Calahan, 1988). A prospective study that examined surgical findings of sand
accumulation found that intestinal sand was diagnosed before surgery in only 58% of
cases, and accidental enetroeentesis was the most common method of detecting it Ragle,
Meagher, Lacroix, and Honnas, 1989. Doran (1993) also states that because of the
“insidious onset and vague clinical signs associated with the problem, many cases are
diagnosed and treated surgically”.

There also appears to be a discrepancy in the reports as to the most common
locations of sand accumulation. Most reports state only examining radiographs of the
cranioventral abdomen for sand, where Ragle, Meagher, Lacroix, and Honnas, (1989)
found a higher percentage of sand impactions in the right dorsal colon, transverse colon,
and pelvic flexures as compared to the left or right ventral colons or associated cranial
flexures (stemal or diaphragmatic). This discrepancy would affect the sensitivity of
radiography in the diagnosis of intestinal sand.

It is important to diagnose sand colic accurately in order to prompt the correct
therapeutic regimen (medical versus surgical). Therefore, it should be a goal of clinicians
to obtain more accurate methods of diagnosis that provides better correlation with the
disease.

The objective of this study is to improve the sensitivity and specificity of

abdominal radiography as a diagnostic tool for intestinal sand accumulation in horses.

More specifically, the goal of this study is to develop an objective method of
radiographic interpretation of intestinal sand accumulation on equine abdominal
radiographs. Currently, abdominal radiographs are reviewed and a purely subjective
assessment is made as to whether the opaque material seen within the gastro-intestinal
tract is sufficient to cause colic, and as to the approximate quantity or severity of sand
accumulated. This method appears highly variable between reviewers and as such, can
result in differing opinions and different recommendations for therapy. A more objective
method of assessment would be more reliable, reproducible, and accurate. The
approximated quantity of sand accumulation should be more representative of the actual
intestinal contents that will be more useful in assessing severity of disease and selecting
appropriate methods of treatment. This goal will be accomplished in two studies:

1. Identify criteria for assessing or measuring sand accumulations. This
will be a retrospective study of equine abdominal radiographs with
known diagnoses. Horses with both positive and negative diagnosis for
sand colic will have their abdominal radiographs reviewed and the
criteria that have the highest correlation with a positive diagnosis will
be utilized to create an objective scoring scheme that will be tested
between four different reviewers. The results of this objective
radiographic assessment will then be compared to the results of the
same four reviewers interpreting the same radiographs in a more
subjective manner that is currently utilized in abdominal radiographic
interpretation. It is hypothesized that subjective scoring is inaccurate

and variable, depending on the interpreter, and that the objective

scoring would provide a more reproducible, repeatable, and more
accurate assessment of the relative size and severity of the sand

accumulation as compared to current subjective assessment methods.

Implement the new objective interpretive scheme into a prospective
study to test its accuracy and radiographic sensitivity to known
amounts of experimentally introduced sand in healthy horses. It is
hypothesized that different known amounts of administered sand can be
detected on radiographs using this objective assessment scheme that

would not be apparent using subjective scoring.

10

Figure 1-1
An example of an abdominal radiograph showing a sand accumulation in the colon of a

horse. The accumulation is of bone opacity and has a convex ventral margin and a

horizontal dorsal margin.

 

Materials and Methods

A search through the medical records of the Large Animal Clinic of the College
of Veterinary Medicine at Michigan State University was performed through the years
1996-2003. The search included horses admitted to the clinic that had abdominal
radiographs performed during their stay. These results were further refined by searching
for the diagnoses and whether or not it was sand colic that was established as a final
diagnosis upon their discharge. A total of 60 different horses were selected, 30 with a
positive diagnosis of sand colic, and 30 that did not have a diagnosis of sand colic. Of
these 60 cases, a further 9 were discarded from the study due to poor radiographic
technique that precluded interpretation, or because the horse had multiple abdominal
disorders including sand colic.

The radiographs were initially reviewed by one observer to select different criteria
to use to construct a scoring system. Each set of abdominal radiographs was reviewed in
random order with the reviewer blinded towards the patient's history, prior radiographic
interpretation, and final diagnosis. Radiographs were reviewed in random order without
knowing whether it was the first or subsequent presentations of that horse to the hospital.
The radiographs were examined for the following criteria with the descriptors used for
evaluation of each parameter in parentheses:

- whether or not the entire abdomen was included on the radiographs (complete

study: yes/no)

- whether adequate exposure techniques were utilized to interpret the radiographs

(yes/no/partial)

- whether sand (radiopaque material) was present (yes/no)

12

- whether the entire accumulation(s) of sand was (were) visible (yes/no)

- the number of sand accumulations

- the length of the accumulations (cranial-caudal dimension- mm)

- the height of the accumulations (dorsal-ventral dimension- mm)

- opacity of the accumulation (mineral/soft tissue/mixture)

- homogeneity of the accumulation (homogeneous/heterogeneous/ areas of both)

- predominant location of accumulation (cranioventral, midventral, other)

- width (mm) of a rib (caudal rib, mid-body, side closest to film)

- rib width : length of sand accumulation (ratio)

- rib width: height of sand accumulation (ratio)

- evidence/degree of secondary changes from obstruction (intestinal distension,

significant gas accumulation, N0 = 0, Yes = 1-3)

- final diagnosis (sand colic, other)

In order to create the objective scoring method to evaluate radiographic sand
accumulation, these parameters were tested against the final diagnosis (sand colic or
other) using a Pearson correlation test (significance level set to p < 0.05) to see if there
was a linear relationship between these parameters and a positive diagnosis (ie. sand
colic). These parameters with a statistically significant (p < 0.05) correlation (r > 0.5)
were selected for use in the objective scoring assessment. Logistic regression analysis
was performed on the standardized length and height dimensions of the accumulations
(standardized to the rib width) to see if a cause-effect / prediction association could be

made with the final diagnosis. Statistical significance was set at p < 0.05.

13

The parameters that were chosen for use in the objective scoring assessment were
assigned a score that was relative to their degree of correlation to the diagnosis. For
example, the number of sand accumulations was found to have a statistically significant
strong correlation (r = 0.9) with a positive diagnosis. The location of the accumulation
was also found to have a statistically significant correlation, but not as strong (r = 0.5).
Therefore, on the objective assessment, the score assigned to the number of
accumulations ranged from 0-3, with 0 representing no sand visible, 1 representing 1-2
separate accumulations visible, and 3 representing 3 or more accumulations. The score
for the location of the accumulation, subsequently, ranged only from 0-1 due to its lower
significance, with 0 representing sand accumulating in areas other than the cranioventral
abdomen, and I for sand located in the cranioventral abdomen.

An objective scoring system radiographic assessment sheet for abdominal
radiographs was then created to provide a more accurate interpretation for sand
accumulation (Figure 1-2). It was designed so that an interpreter would evaluate the
radiographs for each criterion, assign a score for each one according to the instructions,
and then add up these scores to obtain a final total. The final total is a value out of a
possible 12 points, representing an ordinal scale of relative severity and quantity of sand
accumulation.

A subjective radiographic assessment sheet was also created. The subjective
radiographic assessment sheet was formed on the basis of utilizing current systems of
interpreting sand accumulation. The interpreter indicated whether or not they felt that the

sand accumulation was sufficient / severe enough to cause sand colic. If the interpreter

l4

felt that there was sufficient evidence for sand colic, a score was assigned ranging from
1-5 depending on the severity of sand accumulation. Therefore, a score of 0 does not
necessarily mean that no sand is present, but that it is not likely enough to cause clinical
disease.

Four reviewers independently assessed all abdominal radiographs of the 51 cases.
These reviewers included a radiology resident, a board certified radiologist, a board
certified large animal internist, and a board certified large animal surgeon and emergency
and critical care clinician. These people were selected to assess inter-observer variance
among veterinarians with different areas of expertise that frequently deal with cases of
sand colic and abdominal radiographs. Each reviewer was blind to the final diagnosis,
the prior radiographic interpretation, and the patient's Signalment and history. Each
reviewer was asked to assign a score to the sand accumulation by subjective assessment
first, followed by assigning a score by the objective assessment.

To assess intra-observer variance, one of the reviewers, the radiology resident,
reviewed all abdominal radiographs three separate times, each time separated at least by 2

week intervals to decrease the chance of recognition of particular cases.

Statistical analysis:

The results of the subjective and objective scores of all four reviewers were
tabulated and tested to see if they were normally distributed. The subjective scores
obtained from the reviewer who assessed the radiographs in triplicate were compared to
each other to assess intra-observer variance using an analysis of variance (ANOVA) that

was modified by a Tukey-Kramer adjustment. Level of significance was set at p < 0.05.

15

A similar statistical test was done on the three objective scores for each case. The
subjective scores for all three interpretations were then averaged for use in the inter-
observer variance tests stated below. A similar pooling was done for the 3 objective
scores.

The subjective scoring inter-observer variance was tested with an ANOVA
modified by a Tukey-Kramer adjustment. The level of significance was set at p < 0.05.
The subjective scores of each case were then compared to the final diagnosis in the
patient's medical record to see that if, using this method, one may accurately predict a
case of sand colic. The observed and expected results were compared to each other with
a Pearson Chi-Square test of independence with a level of significance set at p < 0.05.
The objective scores of the observed sand accumulations were tested for inter-observer
variance using an ANOVA modified by a Tukey Kramer adjustment. Level of
significance was set at p < 0.05. The objective scores of each case were then compared
to the true diagnosis using a logistic regression model to see that if, using this method,

may one accurately predict a case of sand colic and its relative severity.

16

 

 

 

 

 

 

 

 

 

 

muxcmA NH xmA NumsoocoonoE mics-Q: mu mA co m
T62: E3 W52 T52 _ um; 75:22:20
cubsmm:
Oon L v oHXm; ohmsoocowocowoz tow ouo oncoEO
am E 523 am 2 523
6.8m “ms—Em m:c=a_:E=ou< 33:52
.38. he sauce: mo $05—01; NumocouoEoz Ba 0 E SQE: Z .5339. 030

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.couflafisoom comm com “coca wctoom EoEmwomma ofiafiwoncfl 0380 .30

N; 2:5

l7

Results

A total of 51 different cases were used for the retrospective study (Table 1-1). Of
these, 42 (82%) had single admission, while 9 (18%) had multiple admissions. Twenty-
three (55%) of the horses admitted only once had a diagnosis other than sand colic, while
19 (45%) had a positive diagnosis of sand colic established. There were a total of 67
different admissions where the abdominal radiographs were reviewed. Thirty-nine (5 8%)
did not have a diagnosis of sand colic, while 28 (42%) did have a positive diagnosis of
sand colic.

Within the 51 horses, there were 6 (12%) intact males, 17 (33%) geldings and 28
(55%) females. There were 10 (20%) foals (up to 1 year of age), and 41 (80%) were
adults with age ranging from 2 to 30 years. The different breeds represented in this study
included: 13 (25%) Quarterhorses, 5 (10%) Standardbreds, 4 each (8%) of Arabians,
Thoroughbreds, Ponies, and Grades, 3 (6%) Miniatures, 2 each (4%) of Hanovarians,
Friesians, and Tennessee Walker Horses, and I each (2%) of Clydesdale, Spanish
Mustang, Donkey, Warrnblood, Haflinger, Pony of America, Appaloosa, and Paint.

Of the 24 horses that had a diagnosis of sand colic made, 3 (13%) stallions, 8
(33%) were geldings, and 13 (54%) were females. There were 8 (33%) foals, and 16
(67%) adults with age ranging from 2 to 25 years. The different breeds diagnosed with
sand colic included: 8 (33%) Quarterhorses, 2 (9%) each of Standardbreds, Arabians,
and Grades, and 1 (4%) each of Miniature, Pony, Hanovarian, Tennessee Walker Horse,
Clydesdale, Friesian, Warmblood, Haflinger, Pony of America, and Appaloosa.

Correlation coefficients of the measured radiographic parameters to the final

diagnosis are presented in Table 1-2. These determined to have a significant relatively

18

strong correlation with a positive diagnosis of sand colic (p < 0.05, r > 0.5) were selected
for use in the objective scoring assessment scheme. The parameters used included
number of sand accumulations, the opacity and homogeneity of the accumulations, the
location of the accumulation, the rib width to length of accumulation ratio, and the rib
width to height of accumulation ratio. It was determined that standardized height and
length of the sand accumulations, standardized to the width of a rib, on their own did not
have a significant cause-effect relationship with the diagnosis, but when put together in a
logistic regression analysis a significant relationship was evident. There was a significant
negative correlation with a complete study and a diagnosis, and a significant negative
correlation with the radiographic quality and a diagnosis. The simple presence of sand
and whether or not the accumulation was entirely visible on the radiographs did not have
any significant correlation with a diagnosis of sand colic. Radiographic changes that
could be attributed as secondary to obstruction were not found to have any significant
correlation with a positive diagnosis of sand colic.

Table 1-3 summarizes the results of the subjective and objective scores
determined by the 4 reviewers for each case. The data was determined to be normally
distributed. Significant differences were found between the subjective radiographic
assessments between the 4 reviewers (p = 0.004), and within the one reviewer who
examined the radiographs in triplicate (p = 0.02). Using the objective method of
radiographic assessment however, no significant differences were found between the
scores of the 4 reviewers (p = 0.127), or for the scores assessed by the reviewer who

examined the radiographs in triplicate (p = 0.24).

19

The subjective score of each case of all reviewers was compared to the final
diagnosis of each case. Using a Pearson Chi-Square test of independence, a significant
difference was found between the observed results and the final diagnosis (p = 0.05, X2 =
4.86). This means that based on radiographs alone, subjective assessment of sand
accumulation resulted in significant proportion of the cases being misdiagnosed.

Since good reproducibility was found using the objective scoring method among
the reviewers, a logistic regression model was created. This was done to determine if the
objective score could predict sand colic with accuracy. The logistic regression model
created is as follows:

In P/l-P = -4.0845 + 0.5477 x score = 0

It was found that an objective score of 7.01 or higher had an 83% chance of being
correlated with a true positive diagnosis of sand colic.

Figure 1-3 demonstrates a small accumulation of sand in the diaphragmatic
flexure of one of the horses. This particular case was scored between a 0 and l
subjectively, indicating that some observers felt that this was enough sand to cause mild
colic. Using the objective method however, scores ranged from 4-5 clearly indicating
that this was not enough sand to cause clinical disease. Figure 1-4 demonstrates a
moderate sand accumulation in the ventral colon of another horse. Subjectively, scores
ranged from 0-4 indicating some reviewers felt that this was not enough sand to cause
clinical disease, while others felt that this was enough to cause colic of moderate severity.
Figure 1-5 demonstrates a large sand accumulation in the ventral colon of another case

examined. Although both subjective and objective assessments indicated that this was

20

enough sand to cause severe clinical signs, the range of scores was much less varied

objectively compared to subjective means.

21

Figure 1-3

An example of an abdominal radiograph showing a small sand accumulation

in the diaphragmatic flexure of a horse.

 

22

Figure 1-4

An example of an abdominal radiograph showing moderate sand

accumulation in the ventral colon of a horse.

 

23

Figure 1-5
Two consecutive abdominal radiographic images demonstrating an example

of a large accumulation of sand in the ventral colon of a horse.

 

24

Table 1-1

Signalment and diagnosis (sand colic or other) distribution of horses used for the

retrospective study. A diagnosis of sand colic and other is given to those horses that

presented more than once that resolved an initial presentation of sand colic.

Horse #

\DOOQOIUI-DnUJNH

# Admissions

fiflfl----~fl_#--_fl---~—~

Breed

Standardbred
Clydesdale
Pony
Quarterhorse
Miniature
Arabian
Quarterhorse
Grade
Quarterhorse
Thoroughbred

Tennessee Walker Horse

Hanovarian
Arabian
Pony
Pony
Grade

Spanish Mustang

Quarterhorse
Donkey
Thoroughbred
Quarterhorse
Thoroughbred
Standardbred
Standardbred
Friesian
Warmblood
Pony
Quarterhorse
Arabian
Haflinger

25

.351

Male
Gelding
Gelding
Gelding
Female

Male
Female
Female
Gelding
Gelding
Female
Gelding
Female
Female
Gelding
Gelding
Female
Gelding
Female
Female
Gelding
Female
Female
Female
Female
Gelding
Female
Gelding
Gelding

Male

Ass
(years)
1 1

t—tI—II—Iill—II u—t WNH

I—A
”‘4:

Foal
2
3
2
18
17
8
l6

Foal

Diagnosis

other
sand colic
other
other
other
other
other
sand colic
sand colic
other
other
other
other
other
other
other
other
other
other
other
other
other
other
other
sand colic
sand colic
sand colic
sand colic
sand colic
sand colic

Horse #

31
32
33
34
35
36
37
38
39
40
41
42
43
44

45
46

47
48
49

50
51

Table 1-1 (con’t).

# Admissions

wm—~—————i—i—t——O—d

[\J

Breed

Quarterhorse
Grade
Quarterhorse
Miniature
Hanovarian
Arabian
POA
Appaloosa
Grade
Quarterhorse
Standardbred
Standardbred
Paint
Quarterhorse

Quarterhorse
Miniature

Tennessee Walker Horse
Thoroughbred

Quarterhorse

Quarterhorse
Friesian

26

w.

Female
Male
Male

Female

Female

Gelding

Gelding

Female

Female

Gelding

Female

Female
Male

Female

Gelding
Female

Female
Female
Female

Female
Female

Ass

(years)
3

Foal
Foal
Foal
Foal
7
25
2
1 5
1 l
Foal
Foal
Foal
2

2
6

12
18
9

Foal
2

Diaggosis

sand colic
sand colic
sand colic
sand colic
sand colic
sand colic
sand colic
sand colic
other
other
sand colic
sand colic
other
sand colic
sand colic,
other
other
sand colic,
other
other
sand colic

sand colic,
other
other

8:038:00: 00 0%:2 003030530

200 5:5 0050800 5:038:00: .00 £000: 00N_E:0:Em
042.0 5003:5000 00 50:2 0000:0055
30.0 5003:5000 .00 Ems: BEESSBm
000.0 00.0 003000 03:03:30 b03008 00 00%00\00:0E>m
30.0 wad 8:23:38: 00 :000004
50.0 00.0 8003:5000 00 b_0:0w0::0:
N000 00.0 5003:5000 00 5630
.0000 mwd 30:03:58: 0:00 00 00:5 Z
30.0 000 8:258:00: 0.55 .00 8:320:05
20.0 5.0 0:00 00 00:08::
90.0 $0. 3:030 ofimmaomcmm
30.0 5.0- >030 0:29:00
035, m 03:.» m“ a

 

 

3:230: 0.. 03% 04 00:00E:u_m E03503 .5330th 00050.8 03 0.. 0:00:
0:00 0:00 00 200E000 0 0: £80880: 0308w068 00580000 00 5:20:00

~-~ 03:0.

27

550
.550
550
:050
:050
:050
:050
.550
.550
550

0:00 0:00

0:00 0:00
.550
:050
:050
:050
:050

0:00 0:00
:050

0:00—$05

Nnvovmoxoeoooxoxoovovsv

VI

 

 

 

 

 

N N N 0. N N N o : N N N
N N N N N v o N o : _ o
o o o o o o o o o o o o
v v m m m m 0 0 0 0 0 0
N 0 0 0 0 0 0 0 0 0 0 0
m 0 m m m m N H : : 0 0
N N N N N N o o o o o o
o o o o o o o o o o o o
0 0 0 0 0 0 0 0 0 0 0 0
0 N 0 0 0 0 0 0 0 0 0 0
w w n n w R. 0 m m N m :
0 0 x. N. 0 N. 0 m : 0 0 0
N 0 0 0 0 0 0 0 0 0 0 0
0 N 0 N 0 0 o N N _ N o
v N N N N N N o _ : o N
o o o o o o o o o o o o
v m m m m 0 0 _ _ _ 0 :
0 m N. 0 n N. 0 : m N _ N
0 0 0 0 0 0 0 0 0 0 0 0
05: N00 N20 0.: as: :00 :20
m N a N N .:. w m .N. N N N
0 :03033— :0 53033—

 

 

00800 03:00.00 00:00: 03:00.55

:

gOMNOO—‘O—‘NOOOONOOO—‘N

.I

v—nl

v—‘NMV‘WONOOON

55:20:04:
0 00:0:

 

00030:: 0:000. 0: H X 00030:: 005 0: C050 :0 0:00 0:000 0:00:005 508:0 5:: 0:0. 0:30 05005050:
03:00.30 N3000> :0 0000: :000583000 0:3 :0 0830350050 8500.58 035.0: 0 0:000:00: :05? .N_-0 80:: 0:0 5:000 03:00.30
0:00 0:00 00835:: .00 00000 508:0 05 :00 5550 0350.330 05 80:05:00: m; .0:00 0:00 0300 0: :000383000 0:00 5:305
:0: 80:05:00: 0 505 6-0 800 0:0 00:08 0>000B3m 00030:: 005 0:0 _ 53030: :3 m0w05>0 05 0:0 0808000000 0:80:00

00:5 0:9. 803030: :300 :3 002030 00 :000583000 0:00 :00 0500:0050: 5880000 0: 00:55.0 5:000 03:00.30 0:0 0>000B3m

Mum 030,—.

28

3

Table 1.

(can’t),

ective scores

Ob

ective scores

Sub

Horse #

  

  

Reviewer #

Reviewer #

#

rmssnon

Ad

VI

Ml

 

other

6
6
6
0

O
10

fl
4%, v
V
A
~43 x-r
N
w
h
V
VI N
ml 6
NI '—
v—I > N
as
A
....|'9 N

 

2O
21

other
other

22
23

other

other
sand colic

24
25

10

sand colic

8
9
9

6
12

26
27

sand colic

10

10

29

sand colic

28

sand colic

29
3O
31

sand colic

10 10 10 11 12

10

sand colic

6
7
8
6
8
12
12

11

sand colic

32
33
34
35
36
37
38
39
4O
41

sand colic

sand colic

sand colic

sand colic

11

10
10

11

10
10
10

10
10
10

11
10
10

sand colic

10
10

10

11

sand colic

other

other
sand colic

5
10

11

10

10
11

11

10
12

sand colic

11

10

11

42

43(1)

other

5

2

00:00
00:00
.850
0:00 280
0:8 230
0:8 980
00:00
0050
00:00
00:00
0:00 9:3
0:00 980
8:00
0050
850
0:8 980
0:8 :50
0:00 :03
0:8 :53
00:00
00:00
00:00
0050

0:00—EEG

oommxxXovonEwhoooxfifooo

VI

oooooooohommveoxothmovghxov

MI

 

ooooooxoxooomxov-Smeoxomochoxom
:oxosomoooxoootnvo
oooonnnwovwgoocvowvwmvo

OOOOBFBOOv—VN

oooomnnwmvmeoomvowhwmmo

oooonxooooxmmmflemxoxooxxooo‘ovo

 

£3 £0 new F:

NI

_ :
n 3303013

00.800. 03000.30

 

:

_

ooooooXoMxmeoooooxv—ooov—

VI

oofiHMF-‘v—‘Vflm—‘F‘fiNN—‘v—‘NHMNNN

MI

O—‘OO—‘v—HMONOV‘imm—n—norq—afi—‘No
~OOONOONNMOVV—Mv—NOMMOON

OO—‘Ov—‘ONv—v—fiONNNNOONO—‘N—‘O

O—‘O—‘ONOVONNWVMONOM—‘MONO

 

3.3 6.0 new on:

a _
t .8335:

NI

 

 

8.800. 9500.95

_

_

A_0_m
AmVOW
Amvom
A_00m
Amvov
A_Vov
Amvwv
A_0wv
A4054
Amvnv
Amvav
A_th
Amvov
A_004
Amvmv
A_0mv
Amvvv
Amvvv
A_Vvv
Amvmv
Avvmv
Anvmv
Amvmv

Q. :2. EEIEV

 

n 00.8:

Ar=000
m4 03:.

30

Discussion

These results demonstrate that the designed objective scoring system does work in
creating a more accurate interpretation of the severity of the sand accumulation, and that
it is more reproducible than current subjective methods of assessment.

In designing the objective scoring system, it was important that the cases used
were random representative samples of cases admitted to the Michigan State University
College of Veterinary Medicine. Of the cases selected, there was a 42% sand colic to
58% “other disease” distribution. These cases do not represent all of the horses that had
abdominal radiographs performed over the 7 years searched, or all of those that had sand
colic diagnosed, so prevalence and incidence of sand colic at Michigan State University
cannot be determined from this sample. It was not part of this study to examine what
other conditions these other horses were diagnosed. It was sufficient to recognize that
they were not diagnosed with sand colic. Other indications for abdominal radiographs
include horses presenting for colic where the radiographs would help rule out sand
accumulation as the cause if the history suggests that it may be present (living in coastal
areas, Michigan, or Arizona, feeding off ground, chronic diarrhea). Abdominal
radiographs of adult horses are also indicated in horses presenting with colic to identify
enterolithiasis, pneumoperitoneum, or intra-abdominal radiopaque foreign bodies.
Abdominal radiographs are indicated in foals to identify small or large intestinal
obstruction, meconium impaction, or pneumoperitoneum from gastro-intestinal rupture
(Butler et al., 2000).

Foals were included in this study as the intention was to design a system for

application in horses of all ages presenting with sand accumulation. Foals represented

31

1/5 of the horses examined and most of these radiographs were presented on standard 14”
x 17” radiographs as compared to adult horses that are radiographed using digital
radiography with the images presented using a 6 on 1 format onto a 14” x 17” film.
Using standard radiography should actually improve the interpretation of sand
accumulation since most to all of the accumulation is visible on the radiograph. Using a
6 on 1 digital format, there is risk of overlap of the images, or there may be some portion
of the abdomen that is not radiographed which will hamper the interpretation of the
accumulated sand. However, disadvantages of routine radiographs include the radiation
exposure required, and the possibility of motion unsharpness from an unstable film
holder. A significant advantage of digital radiography is the availability of post-image
acquisition data manipulation that can improve interpretation.

There was approximately an even representation of gender in the horses used in
this portion of the study with Quarterhorses representing the breed majority at 25%. Of
the horses that were diagnosed as having sand colic, these trends continued with males
and females being equally affected and Quarterhorses representing the breed majority
with 1/3 of the cases. The breed with the highest representation in this study that did not
have a positive diagnosis of sand colic made was the Thoroughbred. Although the
number of Thoroughbreds examined is small (n=4), this may represent different
management styles compared to other breeds. A 5-year search through the medical
records of the Large Animal Clinic of horses diagnosed with colic demonstrated a similar
distribution of gender, age, and breed to that seen in these selected cases. These cases
can therefore be assumed to be a good representation of the horses admitted to the Large

Animal Clinic that could be used for the creation of this model. White and Lessard

32

(1986) examined the Signalment of horses presenting with colic and found that
Quarterhorses, Thoroughbreds, Arabians, and Standardbreds had the highest reported
incidence, which also adds credence to the use of the selected cases in this study.

In order to create a more objective method of assessing radiographic sand
accumulation, several radiographic signs and parameters were examined on the
abdominal radiographs. These were determined at random, then after comparing the
results with the final diagnosis, those radiographic findings that had the highest
correlation to a positive diagnosis of sand colic were selected for use in the final
assessment scheme.

Upon examination of the abdominal radiographs, it was found that there was an
abundance of cases that did not provide complete views of the entire abdomen. The
majority of these cases only included images of the cranioventral abdomen. This
parameter was examined to see if, retrospectively, a complete abdominal series is
required to provide a relatively accurate assessment of sand accumulation, and to provide
a diagnosis of sand colic with relative severity. A significant negative correlation was
found between a complete abdominal radiographic study and the established diagnosis.
This means that as the abdominal radiographic series became more complete, a trend for
a negative diagnosis for sand colic was found. This is most likely explained as during the
radiographic examination, if sand is found, radiographs of the accumulation are taken
only of those areas, while the rest of the abdomen might not be radiographed or all
acquired images may not be printed. While if none to little sand accumulation was seen,
most to the entire abdomen was radiographed to document this fact and that the entire

abdomen was examined with digital fluoroscopy. A complete study should not therefore

33

suggest that sand colic is not present, but rather that studies are more likely to be
complete when preliminary views of the cranioventral abdomen do not show sand
accumulation. This finding allowed the use of all of these cases without complete
radiographic studies for the creation of the objective radiographic assessment scoring
scheme.

Another aspect that affects the interpretation of radiographs is the radiographic
quality. This includes the exposure technique, gray scale window and level of the printed
images, presence of artifacts such as motion unsharpness and film/screen artifacts, and
points of reference within the abdomen if many views are obtained. The radiographic
quality, focusing on exposure techniques, was critiqued while examining the radiographs.
A significant negative correlation was found between radiographic quality and the
established diagnosis. The poorer quality radiographs, such as those that were
underexposed or had poor radiographic contrast, may have hindered the interpretation
and not allowed an accurate diagnosis of sand colic if such was the case. This finding
should increase awareness of the importance of good quality radiographs, and how it may
impact interpretation, possibly providing a false negative result (Fischer, 1997).

The simple presence of sand observed on radiographs did not correlate with
clinical disease. This is not surprising since many horses may have gastro-intestinal sand
but not show any clinical signs. It has been discussed that the presence of fecal sand is a
common finding in healthy horses (Colahan, 1987). It is not known how much sand is
required to produce signs of colic or diarrhea, but the mere presence or absence of sand is
not an adequate indicator to support sand colic. This parameter was therefore not

included as a source for points in the objective assessment scheme, since this method

34

assumes that sand is present. A separate category for presence or absence should not be
used since there is no statistical correlation between this finding and sand colic.

Whether or not the full accumulation of sand was present on the radiograph did
not seem to affect the diagnosis. There was no correlation with the presence of the entire
accumulation versus part of the accumulation to a positive diagnosis. This may seem
surprising since, if part of the sand accumulation is missing from the radiograph, then the
measured size of the accumulation, its homogeneity and opacity, and its entire location
within the abdomen would be obscured. It is assumed that most of the sand of the
accumulation was radiographed, providing enough information to establish a diagnosis.

The number of sand accumulations observed was found to have a significant
linear relationship with a positive diagnosis for sand colic. This is supported by
Kaikkonen and Ruohonierni (2000) who demonstrated a significant relationship between
number of radiographic sand accumulations and signs of colic. Most cases in the present
study had 1 large accumulation, or few very small ones. However, several cases were
seen to have 3-4 fairly large size accumulations which were always associated with a
positive diagnosis. It is acknowledged that the numbers visualized may not be a true
representation of the actual number of accumulations. Sand accumulations located in the
mid-abdomen, or in areas that were not properly exposed may have been missed. Sand
typically accumulates and obstructs the ventral colon and the right dorsal colon, and these
areas are usually penetrated well enough to visualize sand accumulation. The
radiographic contrast, however, may not be adequate enough to visualize distinct but
overlying accumulations that may be seen in the stemal and diaphragmatic flexures in the

cranioventral abdomen.

35

The overall radiographic opacity and homogeneity of the accumulation was also
taken into account. It was clear that some sand accumulations appeared more opaque
than others, and some appeared more homogeneous than others. It was assumed that the
more opaque and homogeneous accumulations represented more compact, solid, and pure
sand accumulations and most likely had associated sand colic. The more heterogenous
and less opaque accumulations were considered to be accumulations of mixed contents
(sand and ingesta), with some water content that should indicate a less severe sand
accumulation and not likely causing obstruction. The size of the sand grains may also
affect the radiographic opacity, with coarse sand being more opaque than fine sand. It
has been described that fine sand settles out from the ingesta and accumulates more often
in the ventral colon, and that coarse sand does not settle out until the right dorsal colon,
but Hammock et a1. (l998) found that coarse and fine grain sand was found in various
parts of the large intestine. Both homogeneity and opacity were found to have a
significant linear relationship with a diagnosis of sand colic.

The location of the sand accumulation was examined to see it had a relationship
with a diagnosis of sand colic. Most reports state that sand causing colic is located in the
left or right ventral colon, stemal or diaphragmatic flexures, the pelvic flexure, or in the
right dorsal colon. Radiographically, left from right side is difficult to distinguish, unless
both sides of the abdomen are radiographed, which was not the situation in any of the 67
studies. Stemal and diaphragmatic flexures may be distinguished (Campbell, Ackennan
and Peyton, 1984) if the size and weight of the accumulation does not distort the normal
anatomy and position of these flexures. Both flexures are in the cranioventral abdomen,

with the stemal flexure more caudoventral than the diaphragmatic flexure that is more

36

craniodorsal. The ventral colon (left or right) can also be identified due to its ventral
location, size, and presence of sacculations (haustra) that commonly indent onto the
impacted sand. These sacculations may not be visible if they are distorted by the weight
or size of the sand. The dorsal colon is more difficult to identify specifically, but the
general dorsal location and size of the intestine (visualized by the intra-luminal sand)
help identify it as dorsal colon. The pelvic flexure may be identified by its caudoventral
location and the presence of a convex caudal aspect. Since accurate identification of sand
accumulation is difficult, it is best to describe the location based on abdominal quadrants,
unless a distinct structure such as the stomach, cecum, stemal flexure, or diaphragmatic
flexure is visualized. Upon retrospective examination of the abdominal radiographs, it
was found that there was a weak relationship (r = 0.5) between cranioventral sand
accumulations and a diagnosis of sand colic. Cranioventral distributions include both
stemal and diaphragmatic flexures, and the cranial aspects of the left and right ventral
colons. Other areas of sand accumulations were characterized by smaller or wispy
inhomogeneous accumulations, of varying opacity.

Evidence for and degree of radiographic signs that are associated as being
secondary fiom colic or obstruction were evaluated and tested to see if any relationship
existed with a diagnosis of sand colic. There was no correlation found, as many other
causes of colic can cause the same radiographic changes and are non-specific. These
include intestinal fluid distension, significant gas accumulation, and loss of serosal
margin detail. These secondary changes may also obscure the visualization of
radiopaque material resulting in a false negative diagnosis for sand colic (Rose, Rose and

Sande, 1980).

37

The length and height of the sand accumulations were directly measured and then
standardized to the measured width of a rib, which also removed any effect of
magnification, providing a relative size of the sand accumulations. It was felt that of all
of the radiographic parameters examined, the standardized length and height should have
a cause and effect association on the diagnosis, and that perhaps a prediction of sand colic
could be made based on the size of the accumulation. For this reason, logistic regression
analysis was performed on the measurements against the diagnosis as opposed to
correlation analysis. On their own, standardized accumulation length and height did not
have any significant association with a diagnosis of sand colic. When combined,
standardized length and height had a significant cause-effect relationship with a diagnosis
of sand colic. Kaikkonen et al., (2000) states that the relationship between clinical signs
of colic and the size of the radiographic sand accumulation was statistically significant.
Many of the sand accumulations visualized in this study were located primarily to be
trailing along the ventral floor of the ventral colon without any significant height. This
most likely represents some sand that has settled out from the ingesta but has not
accumulated enough to cause clinical signs. This shows that both measurements must be
considered together to provide a two-dimensional perspective of the size of the
accumulation. The standardized volume of the accumulation would be an ideal
parameter to obtain, however the third dimension cannot be measured due to the single
projection of equine abdominal radiographs and the lack of orthogonal views.

Based on the results of the preliminary retrospective examination, it was decided
that those parameters with a significant association with a diagnosis of sand colic would

be used to create the objective scoring scheme. This included: the number of sand

38

accumulations, their opacity and homogeneity, their location and standardized two-
dimensional size. A score was assigned to each category (Figure 1-2) dependent on their
statistical strength of correlation. The four reviewers then each assigned a score to the
sand accumulation that should be a reflection on its combined relative size and severity.
These scores were then compared to a purely subjective score judged based on the overall
appearance of the abdominal radiographs. The goal was to see if the current subjective
assessment of abdominal radiographs for sand colic differs than that of the objective
assessment, and to compare the accuracies of the scores to the final diagnosis.

The observed results support the proposed hypotheses. Significant differences
between the subjective scores of the four reviewers indicate that subjective scoring
provides a high degree of inter-observer variance. A high degree of intra-observer
variance was also found based on the results of the three trials that one observer
performed, indicating poor reproducibility of the assessment even within the same
person. The Pearson Chi-Square test also indicated that subjective scoring has poor
accuracy in terms of determining or predicting sand colic.

Subjective assessment is the current method of interpreting abdominal
radiographs for sand colic. These results show that this method is inaccurate, has poor
reproducibility and high inter-observer variance. This establishes a need for a more
accurate and reproducible test that will better quantify and assess the severity of sand
accumulation. If such a test can be created, a more clinically accurate assessment will be
established, providing the correct therapeutic regimen to follow. This will also improve
the current poor correlation between clinical signs, diagnostic findings and surgical

findings that plagues horses with sand colic (Specht and C olahan, 1988).

39

The objective scoring assessment scheme was created to improve the diagnosis of
sand colic and to provide clinicians a more accurate interpretation of the sand
accumulation. Results from the retrospective study indicate that the objective scores
assigned to each case from all four reviewers did not differ significantly from each other.
This means that inter-observer variance is improved and given the interpretive criteria
(Figure 1-2); others assessing the radiograph should obtain similar results. This method
also proved to be reproducible, based on the fact that the results of the three trials of the
one reviewer did not differ significantly. The objective scoring assessment is also fairly
accurate in predicting sand colic. A case with an obtained score of 7 out of 12 yielded a
diagnosis of sand colic with 83% precision. This still allows some false positives if a
score of 7 or greater is attained, however, it is unlikely. An objective assessment score of
under 7 means that there is still some sand present, but it is unlikely that the sand is
causing the clinical signs, and that other diagnostics may have to be performed to
determine the cause of the colic. The amount of sand present cannot be accurately
determined, but this ordinal scale provides a relative comparison. For example, a score
of 5 represents more sand than a score of 4, but less sand than a score of 7, and is most
likely not causing signs of colic.

Limitations of this study include:

1). Although foals were included, they were not separately examined to see
whether or not they fit the same statistical results as compared to adult horses. Foals
represented a fairly significant proportion of those cases that were diagnosed with sand
colic (33%). Although most of their radiographs were a single 14” x 17” film, this should

not affect the method of objectively examining the sand accumulation as compared to the

40

many views seen with the adult radiographs. They might actually provide a more
accurate assessment as the entire abdomen is included and there is no overlap of images
and there are no areas of the abdomen that are not radiographed.

2). Although this method of assessing sand accumulation provides a more
uniform conclusion amongst interpreters, it still may not provide additional clinical
utility. As discussed earlier, exact volumes or weights of sand accumulation cannot be
determined, so values under 7 simply provide a relative scale to compare other
accumulations. This method may have clinical application for assessing sand
accumulation for horses that re-present to the clinic with signs of chronic colic to see if a
previously observed sand accumulation is changing in size. Also, this method may be
useful in examining the success of medical therapy, where one could compare scores on
subsequent radiographs afier an initial therapeutic trial. It has been discussed that one
cannot rely on clinical signs and evidence of fecal sand to establish a conclusion of the
gastro-intestinal sand accumulation, and previous radiographic assessment post-therapy
simply provided a subjective assessment that has been proven by this study to be variable
and inaccurate.

Although this objective method has proven to be more accurate and more
reproducible than subjective assessment, it needs to be tested prospectively to determine
its sensitivity to varying quantities of sand. This would help to validate this objective

assessment tool.

41

Conclusion
Current methods of assessing radiographic sand accumulation tend to be

inaccurate and inconsistent among readers. The devised objective assessment scoring
scheme is more reproducible with improved accuracy for determining the presence of
significant sand accumulation. It provides an ordinal score ranging from 0 to 12 that
attributes a relative quantity/severity of gastro-intestinal sand accumulation, with a score
of 7 or higher having an 83% of being associated with a diagnosis of sand colic. This
objective scoring system will be helpful in applying radiographic findings to clinical

diagnosis and management of equine colic.

42

CHAPTER 2
Prospective study and testing of the developed objective assessment

on experimental horses

Introduction

Based on the results of Chapter 1 where it was determined that the developed
method of assessing sand accumulation on abdominal radiographs was more reproducible
and accurate than current subjective methods, a prospective study was designed. The
purpose of this prospective study was to utilize and test the newly developed objective
assessment method on experimental horses that have been administered known amounts
of sand into their gastro-intestinal tracts. The study was designed to evaluate whether or
not this method could differentiate 2 different quantities of gastro—intestinal sand, and to

evaluate its potential usefulness in determining clinical disease.

43

Materials and Methods

Ten experimental horses were used in this study, 8 being from a research and
teaching facility owned by Michigan State University and used for theriogenology
purposes, and 2 being donations to the Veterinary Teaching Hospital from local clients.
The 8 research mares were kept on grass pasture throughout the year that did not contain
sand. There was adequate grass coverage to prevent over-grazing. Their grass diet was
supplemented with hay for 9 months which was held in feeders that sat on the ground.
The pastures contained a holding area that was dirt-packed but did not contain any sand.
They were supplied with ad libitum water. There were no known problems related to
sand enteropathy with these mares. The 8 research horses were all intact mares with ages
ranging from 6-24 years old. Three were Quarterhorses, 3 were Thoroughbreds, 1 was a
Standardbred and 1 was a grade. One of the 2 donation horses included an 8 year-old
Thoroughbred gelding whose feeding environment history was unknown; however there
were no reports of previous sand colic. The other donation horse was a 17 year-old
Arabian mare that was stabled indoors on wood-shavings-covered cement floors and fed
grass hay and water ad libitum. There was no known previous incident of sand colic with
this mare.

The horses were divided into two groups of five for the study and were randomly
allocated into either group. The horses were numbered 1-5 in Group 1 and 6-10 in Group
2. The groups were divided based on the dose of sand that was to be administered so as
to detect sensitivity of the radiographic sand accumulation objective scoring system
devised in the retrospective study. Horses in Group 1 were administered sand at a dose

of 4 g/Kg bodyweight once a day for 2 consecutive days. Group 2 consisted of 5 horses

44

that were administered sand at a dose of 4 g/Kg bodyweight twice a day for 2 consecutive
days.

The experiment was performed working with 2 horses at a time, 1 from each
group. Each horse had a physical examination performed at the beginning of the study to
obtain baseline vital statistics (rectal temperature, heart rate, respiratory rate, packed cell
volume, total plasma solids, bodyweight, digital pulses and borborygrni) and to observe
for any underlying disease that may confound, place bias, or preclude the study from
being performed. Following the physical examination, each horse had a complete series
of abdominal radiographs taken (standing lateral views, horizontal beam) from both right
and left sides of the horse to ascertain if any intra-abdominal sand was present and
radiographically visible.

Radiographs were taken using high intensification digital fluoroscopy (Maxiray
125, GB Advantx LFX Digital Fluoroscopy System, General Electric Medical Systems,
Milwaukee, Wisconsin). A 3-phase, 12-pulse, 60-hertz generator was used that supplied
power to an x-ray tube with a 1.2 mm focal spot and 10° target angle, using exposure
settings of 100-110 kilovoltage peak (KVp) and 400-500 milliamperes (mA) with the
time of exposure (seconds) controlled by an automatic exposure controlled phototimer.
All radiographs were taken in an overlapping series consisting of 12 radiographs per side.
For purposes of radiography, the abdomen was divided into 4 rows beginning ventrally
and extending in a cranial-caudal direction, with 3 radiographs needed per row to view
the entire abdomen (Figure 2-1). Therefore, radiograph 1 was a view of the cranioventral
abdomen including the stemal and diaphragmatic flexures, radiographs 2, 3 and 6 were

the ventral colons and pelvic flexure, radiographs 4 and 7 were views of the stomach

45

region, radiographs 5, 6, and 8 were of the central abdomen including the small intestines
and dorsal colon, radiographs 10, 11 were of the dorsal abdomen, and radiographs 9 and
12 were views of the small colon at the cranial brim of the pelvis. A similar format was
utilized for the opposite side so that a total of 24 radiographs were obtained each time a
horse had abdominal radiographs performed.

Horses had to occasionally be sedated for the radiographic procedures. Sedation
was given using detomidine hydrochloride (Dorrnosedan®, 10 mg/ml, Orion
Corporation, Espoo Finland) at 0.007 — 0.015 mg/Kg intravenously, or xylazine
hydrochloride (AnaSed®, 100 mg/ml, Ben Venue Laboratories, Inc, Bedford, Ohio) at
0.5 mg/Kg intravenously.

Provided that the horses did not show any radiographic evidence of intra-
abdominal sand and were healthy, the horses were kept in individual stalls, bedded with
wood shavings overnight and for the duration of the study. The horses were fed timothy
grass hay in a controlled fashion during the first 2 days while the sand administration was
being performed. Six hours prior to any sand intubation, the horses were denied access
to food so as to empty the stomach. Following the first 2 days, the horses were then fed
ad Iibitum for the remainder of the study. They were allowed continual access to fresh
water for the entire study.

Sand was obtained in bulk quantity for the study from a local landscaping facility.
The sand used in this study was “fill” sand that was unprocessed and not consistent in its
quality. It is a type of sand that compacts well and may contain some stone and clay.

Sand particle size used in this study ranged from 0.05 mm - 6 mm. Fourteen and a half

46

percent was greater than 2 mm, and the majority (57.7 %) was less than 2 mm (0.18 - 0.5
mm). Silt and clay combined for 2.3 % of the composition.

For sand administration, both horses were sedated according to the above
description and restrained in stocks with lead ropes. A plastic nasogastric tube, V2” inner
diameter and 3%” outer diameter, was introduced following an appropriate level of
sedation had been reached. Sand (4 g/Kg bodyweight) was mixed with water that totaled
a volume of no more than 15 liters and passed through the nasogastric tube into the
stomach with the aid of a funnel and gravity-assisted passive-flow. The horses were then
radiographed according to the described procedure. The sand administration was
repeated 12 hours later for the Group 2 horse, and then both horses were radiographed
again. Day 2 of the study followed the protocol from Day 1. Day 3 through 6 involved
only twice daily abdominal radiographs that were separated by a 12-hour time span. The
study concluded for each set of horses on Day 7.

Throughout the study, the entire fecal output was collected at intervals that
coincided with the time of radiography. Feces were removed from the stall and placed in
a large plastic tub filled with water. Any sand present settled to the bottom of the tub and
was collected, dried, and weighed. This was performed so that intra-abdominal sand
content could be calculated, as the amount administered was known.

The sensitivity of sand retrieval from the tub-system was assessed by adding 500
g of sand to a tub full of sand-free feces that was then filled with water. The sand was
collected following sedimentation as was done in the experimental cases. The sand was

dried and weighed and compared to the original weight. This was repeated 2 more times.

47

The horses were monitored throughout the study for signs of colic. Daily
complete physical examinations were performed as well as abbreviated examinations for
specific colic parameters every 6 hours, and hourly walk-bys for monitoring of food and
water intake and clinical signs of pain. Hydration status and body weight were monitored
daily. Following every intubation procedure, and if there were any signs of colic, the
horses were administered flunixin meglumine (FlunixamineTM, 50 mg/ml, Phoenix
Scientific, Inc., St. Joseph, Missouri) at a dose of 1.1 mg/Kg intravenously.

All radiographs (digital fluoroscopic spot-film images) were adjusted for
grayscale window level and window width for maximal diagnostic interpretation and
digitally printed on a 4 x 1 film format (Kodak Dry View Laser lmager, model 8700,
Oakdale, Minnesota). Each horse had 26 sets of radiographs, consisting of right and left
lateral views taken from Day 0 (pre-sand administration check), and twice daily
radiographs (am and pm) from Day 1 — Day 6, with Day 1 and 2 representing the days of
consecutive sand administration. Each set of radiographs consisted of 12 images
depicting the entire abdomen and these were analyzed for sand accumulation both
subjectively and objectively using the scoring systems devised in the retrospective study.
Unless otherwise specified, all radiographic figures are orientated with the top of the
image is the dorsal aspect of the horse, the bottom is ventral, the left side is cranial and
the right side is the caudal aspect of the horse. The objective scoring system has been
established as being a relatively sensitive and accurate method of quantitatively assessing
the amount or severity of intra-abdominal sand by establishing a “sand accumulation
score” on a scale from 0-12 (0 = no sand present, 12 = a very large quantity of sand

present) with a score of 7/ 1 2 indicating a high likelihood of clinical disease. The score is

48

obtained by assessing different radiographic parameters of the sand accumulation such as
homogeneity, location, number of accumulations, opacity, height and width. During the
analysis, the interpreter was blinded to the group assignment of each case horse, and to

the time in the study the radiographs were taken.

Statistical analysis:

All statistical tests were performed using dedicated statistical software (SAS
version 9.1). All data was tested with univariate normal plots to see if they were
normally distributed. Level of significance (or) for all tests was set at a p value of 0.05.

Statistical comparisons were made between the scores obtained from right and left
lateral views, for both subjective and objective scoring methods, to see if obtaining both
views was necessary, or provided new information, to obtain a diagnosis that one view
may not accurately provide. A mix-model analysis of variance (ANOVA) for repeated
measurements was performed and modified by a Tukey-Kramer adjustment for the
multiple comparisons.

Statistical comparisons were also performed between Group 1 and Group 2 to
determine if differences existed in the subjective and objective scores. A mix-model
ANOVA for repeated measurements was performed and modified by a Tukey-Kramer
adjustment for the multiple comparisons. The power of the test was also determined to
assess the ability of the test to tell if a difference does exist.

Other comparisons made using similar testing methods include: 1) comparing the
objective scores of Group 1 versus Group 2 over time; 2) comparing the fecal-sand output

of Group 1 versus Group 2 over time; 3) comparing the objective scores for Group 1

49

versus Group 2 based on the amount of intra-abdominal sand; 4) comparing the 6
individual parameter scores from the objective scoring system of Group 1 versus Group 2
over time; 5) comparing the 6 individual parameter scores for Group 1 versus Group 2
based on the amount of intra-abdominal sand; and 6) comparing each individual
parameter to each other by combining all data. Graphs were constructed to visualize the
above comparisons and linear regression plots were calculated to ascertain if a
relationship exists between the compared parameters. Graphs and regression plots were
performed using a dedicated spreadsheet software package (Microsoft Excel for

Windows, version 9.0).

50

Figure 2-1
Schematic representation of the 12 areas radiographed fi'om each side of the horse’s

abdomen using hi gh-intensification digital fluoroscopy.

\\

- I

\\
l
.5- 0» u

51

Results

All 10 horses had a normal physical examination that was pertinent for the
procedure. Horse #10 had a lame left front leg due to a carpal chip fracture. Hydration
status, appetite, fecal output, intestinal motility and vital parameters (temperature, pulse
and respiration rate) were all within normal limits. No fecal sand was collected from any
horse prior to the experimental administration of sand. All pre-sand administration
abdominal radiographs were negative for intra-abdominal sand except for Horse #3 who
had a small quantity (subjective score = 1/5, objective score = 3/ 12) of sand visualized in
the stemal flexure. This was considered to be due to a small quantity of sand that was
nasogastrically administered 2 months previous for a trial run for this protocol. At that
time, pre-sand radiographs of this horse were negative for intra-abdominal sand.

The procedure of nasogastrically administering sand took approximately 25
minutes per horse. The sand settled quickly within the nasogastric tube and often had to
be re-siphoned or flushed with additional water. Only small quantities of sand could be
added to the funnel-tube system so as not to plug the tube. All sand particles greater than
5 mm were removed so they would not lodge within the nasogastric tube, and to mimic a
more natural setting of sand ingestion where it is considered unlikely that horses would
consume these larger particles.

Throughout the study, 8/10 horses maintained their hydration and body weight
(+/- 10 Kg) and did not show any signs of colic (save for mild diarrhea in 6/ 10 horses
between Day 2-3) and were considered healthy at time of study completion. Two horses
from Group 2 were the exception. Horse # 6 showed mild signs of colic on Day 2 after

the fourth and final dose of sand was administered. Colic resolved with an intravenous

52

dose (see Methods and Materials) of flunixin meglumine. Horse #10 (the 8-year old
Thoroughbred gelding donation) showed mild signs of colic on Day 2 after the fourth and
final dose of sand was administered. Intravenous flunixin meglumine was required twice
daily for 3 days to control the colic. Colic progressed to moderate on Day 6 following 2
days of no fecal output and poor appetite. Heart rate and intestinal motility remained
normal. The horse became mildly dehydrated which improved following oral fluid
therapy. The horse was euthanized on Day 6 due to progressive colic, thus radiographs
and fecal sand were not collected at this final time. Abdominal radiographs of both
Horse #6 and 10 during their episodes of colic reflected a sand accumulation appearance
compatible with sand colic, as determined by the objective scoring system. Horse #6 had
an objective score of 8/12 on Day 2 and a score of 7/12 on Day 2.5. Horse #10 had an
objective score of 8/12 on Day 4.5 and 5. Three medium-large sand accumulations were
visualized within the ventral and mid-abdomen of Horse #10 at this time (Figure 2-2).

The radiographic procedure was performed with little difficulty. Due to the
number of exposures required for the study and due to the high exposure techniques
required for the abdominal radiographs, the automatic exposure control often reached its
maximum exposure limits. This required intermittent tube cooling causing a slight delay
in the acquisition of the images. Occasionally the horses became agitated when they had
to re—enter the radiographic stocks for the opposite lateral views. Thirty-two out of 129
(25%) occasions resulted in only 1 side of the abdomen being radiographed, this being
due to lack of cooperation by the horse or x-ray tube malfunction / overheating.

Initial observations of the radiographs revealed that clear differences could be

seen between a sand-filled stomach, diaphragmatic flexure, stemal flexure, ventral colon,

53

and small colon. Sand could not be reliably identified within the small intestines or the
cecum. Intra-abdominal sand could be identified in its location and its progress through
the intestines could be subjectively monitored. Sand was consistently seen within the
stomach (radiograph number 4) following nasogastric administration. Its appearance was
a well-circumscribed homogeneous mineral-opaque mass with a convex ventral margin
and a straight horizontal dorsal margin in the mid-cranial abdomen (Figure 2-3). Only on
2 occasions was sand still visualized at the end of Day 1 in the stomach (Horse #6 and
10). Its progress or accumulation within the small intestines or dorsal colon could not be
monitored due to the underexposure of the majority of the radiographs in the mid-central
abdominal region (radiographs number 5 and 8) that corresponds to the location of the
small intestines and the dorsal colons. All other areas of the abdomen were adequately
penetrated and could be visualized on film. Sand was also never visualized in any horse
throughout the study in the craniodorsal (radiograph number 7, Figure 2-4a, and
radiograph number 10, Figure 2-4b) or mid-dorsal abdomen (radiograph number 1 1,
Figure 2-4c). These areas represented the dorsal aspect of the stomach and dorsal colons
and were consistently gas filled and contained little ingesta or feces. Ribs and vertebrae
were consistently visualized in these regions. Sand was most commonly visualized
within the ventral colons (radiograph number 2, Figure 2-5, radiograph number 3, Figure
2-6, and radiograph number 6) and within the stemal and diaphragmatic flexures
(radiograph number 1, Figure 2-7) in the cranioventral abdomen. Sand became apparent
within the ventral colon by Day 2-3 and remained within this region for the majority of
the duration of the study. lts appearance varied from homogeneous to heterogeneous,

well-circumscribed to poorly defined accumulations of sand that varied from mineral to

54

soft tissue opacity. Dorsoventrally directed radiolucent bands, interpreted as haustra,
were occasionally visualized. It was more typical for the sand to appear as interspersed
accumulations within the other colonic digesta, becoming more defined, homogeneous
and opaque in Group 2 horses (Figure 2-8) and then becoming ill-defined again towards
the end of the study, while Group 1 horses did not show any signs of subjectively
significant sand accumulation (Figure 2-9 a, b). Sand was inconsistently visualized
within the small colon by the end of the study (radiograph number 9, and radiograph
number 12, Figure 2-10) as interspersed heterogeneous accumulations mixed with feces
or as very opaque fecal balls.

The fecal sand collection procedure was performed without difficulty. Of the 3
trials to assess the accuracy of the procedure, 457 g, 483 g, and 477 g of sand were
recovered following the addition of 500 g of sand to sand-free feces. This represents an
average of a 94% recovery rate. lnitial observation of the collected sand from the trial
horses revealed that the majority of the sand was fine particles throughout the duration of
the study. Four out of ten horses (3 Group 2 horses and 1 Group 1 horse) had coarser
sand particles (2-5 mm) being passed at Day 6 of the study as the predominant type of
passed sand. The amount of passed sand over time was compared among the 10 horses
(Figure 2-11). There is a significant difference (p = 0.028) between the 2 groups for sand
output over time. Group 1 shows an increase in sand output that begins by Day 2.5,
peaks at Day 3.5 and diminishes again by Day 4.5. Group 2 shows a brief and mild
initial increase in sand output at Day 2, and then diminishes until Day 4.5 where it begins

to increase and does not decline by the end of the study.

55

Subjective and objective assessments were performed on every set of radiographs
as described in the retrospective study. The data was determined to be normally
distributed. When comparing the subjective scores for each radiograph between right and
left lateral views, no significant differences were found (p = 0.48). Similar results were
obtained in the comparison of the objective scores for each radiograph between the right
and left lateral views Q) = 0.73). Due to these findings, only 1 radiographic view was
deemed necessary to be used for the remainder of the statistical analyses. Right to left
lateral views (termed as left lateral views) were selected to be utilized because the
majority of the instances (31/32) when only 1 view was obtained, it was a left lateral
view.

Significant differences were not observed between the 2 groups of horses using
subjective scoring (p = 0.19, power 25.6%). Significant differences were present
between the 2 groups of horses using objective scoring methods (p = 0.0003, power
96.0%). Since the subjective scores did not show any statistical difference for the sand
accumulations between the 2 groups, only objective scores were used for the remainder
of the analyses.

Trends of the objective scores over time between the 2 groups are demonstrated
(Figure 2-12). Both groups show an early high score (6/ 12) on Day 0.5 which represents
the average score due to the sand within the stomach following nasogastric intubation.
Neither group reaches this score again (on average), however it is clear that Group 2 has
higher objective scores on average than does Group 1 (p = 0.0003). Both groups are seen
to have their highest average objective score assigned on Day 3 - 3.5 followed by a

gradual decrease in the score. In Group 1, the highest score obtained was a 6/12 (Horse #

56

- 3 and 4). In Group 2, the highest score obtained was an 8/ 12 (Horse # 6 and 10) and was
seen in both horses that showed signs of colic. The 3 other horses in Group 2 (Horse # 7,
8, and 9) did receive scores of 7/12 (Horse # 7 and 8 on Day 3 only, Horse # 9 on Day 4
and 4.5) but never showed any signs of colic.

When comparing the individual objective parameter score over time between the
2 groups, interesting observations are made. When looking at the location of the sand
over time (Figure 2-13), both groups show a similar distribution until Day 3 when Group
2 apparently retains sand in the cranioventral abdomen longer than Group 1. Both groups
do not attain an average score of 1 (meaning a cranioventral distribution of sand) until
Day 2. There is no statistical difference between the 2 groups, however, based only on
the location of the sand accumulation (p = 0.64).

The number of sand accumulations does show a significant difference between
the 2 groups (p = 0.033). Group 2 tends to show a consistently higher score than Group
1, especially noticed at Day 4.5 (Figure 2-14) where Group 1 on average decreases in
score and Group 2 on average increases in score.

The opacity of the sand accumulations does show a significant difference between
the 2 groups (p = 0.048). Both groups obtain the highest score possible on Day 0.5,
coinciding with the sand in the stomach (Figure 2-15). Following this, neither group
attains this score again, with Group 2 having a persistently higher opacity score than
Group 1.

Sand homogeneity also shows a significant difference between the 2 groups (p =
0.046). Both groups obtain the highest score possible on Day 0.5, coinciding with the

sand in the stomach (Figure 2-16). Following this, neither group attains this score again,

57

with Group 2 having a persistently higher sand homogeneity score than Group 1
especially following Day 4.

Although the graph of length of sand accumulation over time appears different
for Groups 1 and 2 with Group 2 having a higher mean score than Group 1 (Figure 2-17),
there is no statistical difference (p = 0.081).

The thickness of the sand accumulations also does not show any significant
differences between the 2 groups (p = 0.098), although graphically there does appear to
be a trend with Group 2 having a higher mean score for sand thickness than Group 1
(Figure 2-18). One note of interest is that Group 2 on average surpasses their Day 0.5
score on Day 3 and 3.5. The highest average score for Group 1 is on Day 0.5
corresponding to the sand present in the stomach.

When examining the objective score based on the amount of intra-abdominal
sand, (Figure 2-19), 3 significant difference is not observed (p = 0.25) when comparing
the 2 groups. There is also a poor correlation between the objective score based on the
amount of intra-abdominal sand only when combining the 2 groups together (R2 = 0.19),
or comparing as separate groups (Group 1 R2 =0.01, Group 2 R2 = 0.13). Due to the poor
correlation seen at the group level, it was questioned whether a relationship may exists at
the individual level. This was still not the case as the highest R2 value calculated was
0.44 in Horse #3.

A trend between objective score and intra-abdominal sand weight was then
questioned by examining this relationship across the different days to see if a particular
day during the experiment had a better correlation. There are weak correlations on Day

2, 5.5, and 6 with R2 values of 0.57, 0.67, and 0.66 respectively. When examining each

58

group separately at each day, some trends are noticed between the objective score and the
amount of intra-abdominal sand. A correlation is never seen in Group 1. Group 2 does
show weak-moderate correlation between objective score and intra-abdominal sand
weight on Day 2.5, 3, 3.5 and 4 with R2 values being 0.55, 0.50, 0.51 and 0.75
respectively. Group 2 has higher R2 values (stronger relationships) than Group 1 for 9/12
(75%) of the sessions radiographed.

Division of the objective scores into their individual parameters was performed to
see if any relationship exists between these 6 parameters that form the objective score and
the amount of intra-abdominal sand. Again, there was no correlation observed at this
level as the highest R2 values were obtained for the location of the sand within the
abdomen (0.26) and the length of the sand accumulation (0.24).

Comparing the individual score parameters to each other, combining both group
data, also did not show any significant correlation except a weak correlation between

sand opacity and sand homogeneity (R2 = 0.5).

59

Figure 2-2

One of three mid-abdominal accumulations of sand in Horse #10 (radiograph view #6) on
Day 5. It is more well-defined, opaque and homogeneous than other Group 2
accumulations. It is also located in a different area within the colon than most other
accumulations. Digesta within the ventral colon is seen below (ventral to) the sand. The

sand is presumed to be in the dorsal colon. Labeled image is below.

 

60

Figure 2-3
Radiograph (view #4) of sand within the stomach (Day 0.5) following nasogastric

intubation and administration of 4 g/Kg sand (top). It appears well defined, opaque and

homogeneous. Labeled image is below.

 

61

Figure 2-4 a—c
Typical appearance of the craniodorsal and mid-dorsal abdomen (radiographs # 7, 10, and

11) demonstrating gas-filled intestines, ribs, and vertebrae.

 

62

Figure 2-5
Sand accumulation within the ventral colon (radiograph view #2) illustrating haustra as

vertical radiolucent bands. Labeled image is below.

\Haustra /

 

63

Figure 2-6
Typical appearance of the mid ventral colon (radiograph number 3). Scattered small sand
accumulations are visible within the colon, presumably the pelvic flexure, alongside the

flank of the horse. Labeled image is below.

Flank

 

64

Figure 2-7
Radiograph of sand within the cranioventral abdomen (radiograph view #1) illustrating
the diaphragmatic flexure and the stemal flexure with the latter being slightly more

caudal and ventral than the former. Labeled image is below.

 

65

Figure 2-8
Typical appearance for sand accumulation within the mid-ventral colon (radiograph view

#2) of Group 2 horses illustrating a large, well-defined opaque and homogeneous

accumulation. Labeled image is below.

 

66

Figure 2-9a

Typical appearance for sand accumulation within the cranioventral colon (radiograph
view #1) of Group 1 horses illustrating dispersed opaque sand particles within digesta
and no focal accumulation. Costal cartilages are seen in the bottom left aspect of the

image, cranial and ventral to the colon. Labeled image is below.

 

Costal cartila':y s

67

Figure 2-9b
Typical appearance for sand accumulation within the mid-ventral colon (radiograph view

#2) of Group 1 horses illustrating small, poorly—defined heterogeneous accumulations.

Labeled image is below.

 

68

Figure 2-10
Radiograph of sand seen within the small colon (radiograph view #12) illustrating opaque

feces just cranial to the pelvis. Labeled image is below.

 

69

A3 08:.

 

00—
. . . . . . n can
. .. 0 00m
0 . . V 004

1 com

 

 

g 000

(3) puns passed

4 00h
. 000

2.5000 n e n

f 000
L
— 9.0.0 III. _.

e 000—

.2808 0:0 :0 00000 0020580 05 80:00 when 088% .08: 00>0 39:0 0000 :80: 092020 N 0:000 000 _ 000.5

2-” 2:3...

70

A3 05:.

O

l I | | ‘ h em”

...w... .... ... . 4 _ n

0 m...

3

. S

3

. o

_ u
_. w
«9:000 - o n __0 3

~0=000III .

_. NH

0:008 05 :0 00000 0000080 0:0 000000 000: 00b0-> .08: 00>0 0000b 0000m 03000.30 0w000>0 m 900.5 000 _ 00000

2-" 2:2...

71

A3 05:.

 

o n v m N u 0
r 8 a 8 L v O
s
Q
.. . 2.
4..
0 Nd
0
u
. 2
11 A1 41 1! n.- D
. .. v.0

 

 

'0
9
3.1035 uoynoo'] puns

 

 

 

 

 

 

 

 

 

C
At 11 1! I.- . MT 00c
. .r _
h
. 7 >6
.. a:
o Ar 0 w c
’ Ir n:- 1... 1v \ O Q 1r
0 s o 0
o O Q 0.¢
o s o 0 _
1.. o Q o Q Lr
at ' I I I ‘I I I I? I I I, I I I A“ I I I IA“ , —
N 0:000 .. O u
_ 0080+
.mCNOE

05 m0 880 080830 0:0 000000 80: 0080-> .08: 00>0 08000 80:82 0:00 0w80>0 N 0:080 000 _ 0:000

24 2:»:

72

80 05:.

 

‘ . ..
'IIIAV \.III‘IIII.IIII,III. T.—

' ‘
I ‘
IA 7‘

 

 

,ri ,,
N

V!
suoginlnumaov puns jo .raqurn N

N 9.089 n O .. fl m.N
— 0080+

.mcwoe 05 mo HOBO punt—.85 on: DHOCDU WED h0b®u> .DEC u®>0 mcomumyflsnoom @Gwm MO pong—DC MOW Duoom Omaho>w N QSOHO mgr—m — QSOu—U

:-~ 2:0...—

73

A3 052.

 

 

 

 

 

 

 

 

 

 

 

e v n N _ c
F LI 8 a. 8 I O
1.. .
1.. N...
.
.
-- m. 4 ._. 2.
. n
A? .
a l . 7 ed
a I , u
L. . . _
. . . - a...
z . .- -. .
a -
3 .. o. . . . T _
o s a .
I a _
I i... 1.. \\ l a w
-. . . 4 e. . .4 S
\ I — 3.
\ I \ a .
s I c u _
.Y .. u . av e. -- . r v.—
.. . 3
o :
NEE—U - .0. u l l r- .
— n=EU+ . ._ m.—
N

24 2:08

31033 Kigacdo puns afiuaAv

.8008 0:“ :0 8080 0.80880 08 000800 80: 80.80%. .08: 80>0 00000 56000 0:00 0888 N 0:080 080 : 900.5

74

II
I

 

 

 

 

 

N955 .. o ..

— 0:00OIII

00: :0 00000 808080 00: 000000: 000: 00000->

A3 08:.

 

 

 

 

woos .(uouafiouron puns

.mcmoe

.08: 00>0 00000 5000on0: :88 09808 N 00000 :00 _ @3000

SA 2:000

75

A3 05;

 

 

 

 

fi JvIIIIAvIIIIAIIIIAYIIIAv “ye-c

 

 

 

 

 

 

3.1039 q33u31 puns

NA
L
.. v;
z
y o.—
Nn0000 .. 0 u
N

00008 00: :0 00000 80:83 00: 0:000: «.00: 00000-> .08: 00>0 00000 :0w00_ :08 0882: N 00000 :00 _ 00000

:4 2:00....

76

A3 005,—.

 

 

 

 

 

 

 

 

 

 

 

e n v m N a o
0 Q II c
Q
hi 1.- jI 1' I.- Q 0 Q 0 .
s 4. s o 0 T N o
0 \ \ O Q o s 0
I I r \ I O Q o .0. s .
le I I IA I I IA, Q 3. I Q o Q VO
0 I I Q ILVQ o QQ _
O
.. .. .. .0 n .. .. .. .. .. 3.
o o 0 .. _ S
I s I W
IAY I I IA‘ w.° p
1.- _
m
It A! .7 M w
0 0
0 . .
. 0.0 m
0
la
II ~ 3
r v.—
N 00000 .. O u :
~00000III _, 0.—
w.—
fl
0 N
.mCmDE

00: 00 00000 :00:000m 00: 00000: 000: 00000-> .08: 00>0 00000 00000:: :000 0w000>0 N 00000 :00 0 00000

8.0 2:000

77

 

Figure 2-19

Group 1 and Group 2 objective scores based on weight of intra-abdominal sand.

 

 

 

 

 

 

 

 

Group 1
12 ~
10 ‘
g 3 4 R2 =o.013
8 .
g 6 « I I I I I
0i I I I I I I I
8 4 «
II
2 .
0 I — , I—fiw—Ifimr-—— --I—r—-— a 1
0 1000 2000 3000 4000 5000 6000
Sand Present (g)
Group 2
12 -~
10 ‘ R2 = 0.128
2 8 O O Q
0
(z O. O O
8
F
O 4 '
2 ..
o I ’t " 1* ’ T““”L—“" ""fi"’" '1‘ "“ ‘ ’ ’ ‘ ”"T l l l

0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Sand Present (g)

78

:8 00000.00 :00m

 

 

 

 

82: 88 800 SE 8% 08m 02:. 080 800 02: o
. r r I: r. ll .IIIIIIIF I I- _ 0
I I I I w
_
W. N
I
_. 0.
_ o
0.
[mo
0 m.
S
3
I I I I m
3
I I I A 0
003.07%. . 0:
2

:003800 N 00000 :00 — 00000

5.0000 3 20$..—

79

Discussion

All horses in this study began the experiment healthy and 9/10 were discharged
healthy. One of these 9 (Horse #6) experienced mild colic during the study (Day 2) but
was responsive to medical therapy. It is presumed that the colic was the result of the 4
intubations with sand administration with secondary mild enteritis / colitis. Horse #10
was euthanized on Day 6 due to progressive deterioration from colic that was non-
responsive to medical therapy. A necropsy was not performed however it is presumed
that the colic was the result of sand accumulations within the right dorsal colon causing
partial obstruction in this horse. Two medium-large accumulations of sand were seen in
areas that were not seen to contain sand in any of the other horses (Figure 2-2). There are
several confounding factors regarding this particular horse that may have predisposed it
to colic and are discussed below.

From these results, it can be assumed that administering sand at this dose and
through this method is generally safe, however, close attention should be given to the
horses receiving the twice-daily doses especially at Day 2 - 3 of the study. Close
monitoring is paramount for the health of the animals to make sure they are eating,
passing feces and maintaining adequate intestinal motility and hydration. If such adverse
clinical signs occur, prompt medical therapy is advised and further sand administration is
discouraged. The serial radiographs in this study aided in the interpretation of the colic in
Horse #10 and can provide usefiil information as to the etiology of the underlying
problem.

The technical aspects of this study were not difficult, but time consuming. The

nasogastric intubation and sand administration procedure required careful attention so as

80

not to administer too much water with the sand, which would over-distend the stomach.
The sand required a lot of water (up to 15 liters) to be mixed with to prevent settling in
the tube. Other options for sand administration were considered however the method
used was felt to be the best. Sand could have been administered through repeated
gastrotomy but this was considered too invasive. Sand could also have been administered
by nasogastric intubation and mixing with carboxymethylcellulose, which forms a gel-
like substance that would help in preventing the sand from precipitating from the water.
Carboxymethylcellulose was not used as it may affect gastro-intestinal motility and
would then not provide an accurate representation of natural clinical conditions of sand
enteropathy.

The radiographic procedure was relatively easy to perform, especially once the
horses became used to the radiographic stocks. Because some horses exhibited fear and
anxiety of the stocks and subsequently became stressed and agitated, only one lateral
radiographic view could be obtained on occasion (25%). This was found to not be a
problem in interpretation and assessing the intra-abdominal sand as the results indicate
that there is no value in obtaining radiographs from both sides of the abdomen for sand
accumulation. One view appears to provide the same assessment as the other view does.
There are slight variations in the shape and size of the accumulations, however these are
not significant.

One of the problems encountered during the radiographic procedures was the
rapid heating of the x-ray tube and the required time delay for cooling. The thick body
parts, large amount of ingesta and secondary scatter radiation required that very high

exposure factors be used so that sufficient x—rays reached the automatic exposure control.

81

Since this was done at a rapid frequency (12 views per side took about 1-2 minutes),
anode heating developed and maximum tube heat units were often reached. The limit on
the load (KV, mA, sec) that can be safely accepted by an x-ray tube is a function of the
heat energy produced during exposure, because above a certain temperature, the tungsten
anode may vaporize. The total heat produced during an exposure is a product of voltage,
current and exposure time expressed in the unit of joules (Curry et al., 1990). The ability
of an x-ray tube to make a single exposure of certain duration is termed its kilowatt rating
and is specific for each x-ray tube and should be consulted before using high exposures
such as in the current study. Using the exposure settings, tube heat units for one exposure
in this study ranged from 11,000 J — 27,500 J depending on the area of the abdomen
being radiographed. The automatic exposure control adjusted the time needed to expose
the phototimer according to the specific area of the abdomen. The maximum allowable
tube current decreased as the tube accumulated heat. This was generally noticed by the
end of the first set (left lateral view) of 12 radiographs and through the middle of the
second set (right lateral view) of 12 radiographs. Maximum exposure limits were usually
met from radiograph number 4 (central abdomen) and onwards indicating that the x-ray
tube was using the maximum allowed mA, given the selected KVp, to expose the
phototimer. This indicated that the kilowatt rating was being approached for this
particular x-ray tube. It was for this reason and due to the homogeneity of the subject
that the central abdominal radiographs were poorly exposed. Radiographs after this
point, despite containing a moderate amount of gas within the field of view, which would
not require high exposures, still resulted in maximum tube exposure limits being reached

because of the heated anode that automatically lowered the maximum allowable mA.

82

Another related factor is the anode heat storage capacity that is the ability of an x-
ray tube to withstand heat loading over an extended period of time. This is commonly
experienced during fluoroscopy used in the present study between every radiograph to
localize the desired image. The anode heat storage capacity indicates the duration of time
an anode can be functioning at a particular energy (I) level before causing excessive
anode heating. Once an anode has reached its maximum heat capacity, it must cool
before further use to prevent damage to the anode. The time required for anode cooling
is also dependent on the anode heat storage capacity and the temperature the anode
reached as heat loss is proportional to the fourth power of the temperature (Curry et al.,
1990)

The anode used in this study has a heat storage capacity of 333,000 I and the x-
ray tube has a kilowatt rating of 105 — 120 using a 1.2 mm focal spot at 0.2 - 0.5 seconds
of exposure (General Electric, 1990). Since the radiographs were producing
approximately 11,000 — 27,500 J of heat to the anode per exposure, 12 exposures
approximate a total of 132,000 — 330,000 .1 to the anode per radiographic series. Added
on top of this heat is the heat produced by the fluoroscopy that created approximately
28,800 I (using fluoroscopic techniques of 120 KVp and 4 mA and assuming an average
fluoroscopy time of 1 minute) to obtain a complete radiographic series. Total heat
produced per lateral radiographic series per horse ranged from 160,800 — 358,800 J. This
exemplifies how the anode storage capacity was reached fairly routinely for each
radiographic series and required subsequent cooling (10 minutes) before it was ready for

re-use (General Electric, 1990).

Although the above problem and details are specific for this particular study and
protocol, it is a common problem in large animal abdominal radiography to have
underexposure and poor quality radiographs if the x-ray tube and generator are not
powerful enough (see Introduction, Chapter 1). Limitations in radiographic equipment
result in further limitations in the ability to correctly interpret the radiographs which is
vital in the current sand accumulation study. Underexposure may provide false negative
diagnoses, or objective scores that are also falsely low. The limitations of the x-ray tube
in the mid—central abdomen in these horses may be confounding the obtained results. The
added time required to allow for anode cooling increases the stress on the horse which
may make it more difficult to manage for additional views. These are acknowledged
concerns of the study, however the retrospective part of the study demonstrated that the
parameters used in this objective scoring system do have high correlations with the final
diagnoses and the devised test is a more sensitive and accurate method for radiographic
assessment of sand accumulation than current subjective methods (see Discussion,
Chapter 1).

Two of the main concerns of this study from the onset were the ability to
experimentally mimic naturally occurring gastro-intestinal sand accumulation, and to
avoid clinically evident colic given the experimental protocol. Several factors that were
involved in these issues were the dose and frequency of sand to be given, the method of
sand administration, the type of food and frequency / quantity of food to be consumed,
the amount and frequency of water consumption, and the housing and type of bedding for
the horses. It was not possible to mimic natural conditions of long-term grazing exposure

to sand, however, all of the above factors were considered.

84

It would have been desirable for the sand to be consumed orally with the feed;
however the amount ingested could not be adequately controlled. One report explains the
administration of sand into the cecum through a surgical typhlotomy (Hammock et al.,
1998); however this method would result in the sand bypassing the stomach and small
intestines. This method may have provided answers as to whether the objective scoring
system is sensitive to varying known amounts of colonic sand; however this current
model was designed to be useful for the typical radiographic presentations that are
commonly seen by the large animal practitioners where the sand is not necessarily all
accumulated in the colon. It was decided therefore that the sand be given through
nasogastric intubation which was also previously described with good results (Ragle,
Meagher, Schrader and Honnas, 1989). This was done knowing that nasogastric
intubation and sedation might have an effect on gastric emptying times and motility
where it has been known to delay such factors (Mahaffey and Barber, 2002; Doherty et
aL,1999)

Alpha-2 adrenergic agonists such as xylazine and detomidine exert their analgesic
effects on the central nervous system and the enteric nervous system (Merritt er al.,
1998). These drugs inhibit intestinal motility because they act by binding to the
presynaptic receptors, inhibiting the release of acetylcholine thereby inhibiting intestinal
contraction (Lester et al., 1998). The effects of xylazine or detomidine on the proximal
gastro-intestinal tract (stomach and small intestine), in particular, remains controversial
where some propose an inhibitory effect on motility (Doherty 2! al., 1999; Adams et al.,
1984) while others indicate a neutral or stimulatory effect (Merritt eta1.,1989). This

suggests that there may be a dose-dependant-response within different segments of the

85

proximal gastro-intestinal tract (Lester etal., 1998). It has been proposed that xylazine
may induce increased amplitude of mixing contractions of isolated segments of small
intestines (Stick er al., 1987). Xylazine and detomidine are also known to inhibit cecal
and colonic (especially the ventral colon) motility and cecal blood flow (Roger and
Ruckebusch, 1987; Adams et al., 1987; Rutkowski et al., 1989; Clark et al., 1988; Lester
et al., 1998) with detomidine being accepted to be a more potent and longer lasting alpha-
2 adrenergic agonist than xylazine. Xylazine typically exerts its effects for 20-40 minutes
while detomidine may last up to 1-2 hours. Mean amplitude, duration and number of
contractions of the circular and longitudinal smooth muscle are affected, and the effects
appear to be proportional to the dose (Clark et al., 1988). For these reasons, alpha-2
adrenergic agonsist are discouraged from use in cases of cecal or colonic dysfunction
such as impactions. Alpha-2 adrenergic agonists also exert systemic effects such as
bradycardia, reduced cardiac output and transient hypertension followed by hypotension
which are mediated through vascular smooth muscle receptors causing vasoconstriction
(Clark et al., 1988).

The horses in this experimental study were sedated with either xylazine or
detomidine at similar doses as reported by Merritt et al., (1998), (0.5 mg/Kg and 0.01
mg/Kg respectively). The ramifications of these previous studies would likely not have a
significant effect on the results of this study as sedation was given only for the
nasogastric intubation and sand administration procedure. An increase in gastric
emptying times and small and large intestinal transit times over the period of 1 hour
would most likely not have any effect on either the radiographic appearance of the sand

at l2-hour intervals, or in the amount of sand passed in 12-hour periods. All horses

86

 

subjectively appeared to eat well following sedation and there was no clinical evidence of
altered gastro-intestinal motility or fecal production following sedation.

The method of sand administration in this study also differs from the natural
setting because this method resulted in approximately 2 Kg of sand being administered at
once into the stomach as opposed to slow ingestion of small amounts over a long period
of time. It can be presumed that this large volume of sand would have an effect on the
gastric emptying rate and on the intestinal transit and subsequent distribution and
accumulation of sand throughout the intestines, and may be different than that from
small, chronic doses of sand. This method was chosen, with recognition of this fact,
based on the ability to know accurate quantities of sand within the gastro-intestinal
system, and to decrease multiple procedures involving sedation and nasogastric
intubation which may complicate the study and be detrimental to the horses.
Subjectively, this large volume of sand appeared to empty rapidly and consistently from
the stomach in most horses as it was observed to have moved from the stomach by Day 1
(12 hours post sand administration). As well, the sand distribution within the colon
appeared similar to distributions observed in previous clinical cases and in those
previously reported (Ruohonierni et al., 2001). The exception to the former statement
were Horses # 6 and 10 that were seen to have small amounts of sand still present in their
stomachs by Day 1. Interestingly, it was these 2 horses that experienced colic. Perhaps
this initial poor (delayed) gastric emptying may have lead to subsequent poor intestinal
motility and accumulation of sand resulting in colic. Whether this poor gastric emptying

was the result of alpha adrenergic agonist inhibition, or individual variation is not known

87

Horses that present for sand enteropathy typically have been eating normally up
until 1-2 days prior to presentation where they then become anorexic or show signs of
colic. In this experimental study, it was desired to keep the horses eating at their normal
frequency with their normal type of food since a change in diet may predispose to colic
(Cohen, 2002). For 9/10 of the horses, this was ad Iibirum grass hay feedings. We
elected to then feed all of these horses ad Iibitum grass hay starting after their last dose of
sand. This was done despite the fact that this large quantity of ingesta may obscure small
accumulations of sand from being radiographically detectable. It was desired to test the
objective scoring system on typical radiographs that may have a gastro-intestinal tract full
of ingesta. It is interesting to note that the 1 horse whose previous feeding history was
unknown was Horse #10 who experienced moderate colic. This particular horse did not
consume as much of the grass hay as the other 9 horses did during the first 2 days of the
study and seemed particularly selective in the food it did eat. Its appetite declined during
the rest of the study and whether or not the poor appetite prior to and during the sand
administration predisposed it to colic is speculative at this point. Perhaps had the horse
been fed its usual diet it would of have had a good appetite during the sand administration
and not exhibited colic afterwards.

Food was withheld for 6 hours prior to sand administration in order to be sure that
the stomach was empty because the sand required a lot (up to 15 liters) of water to be
mixed with for passage through the nasogastric tube without settling. Since such a large
volume was needed, an empty stomach was required since the equine stomach has a
capacity of only 8-20 liters (Pfeiffer and MacPherson, 1990). A previous study

demonstrated that the motility in the proximal gastro-intestinal tract is unaltered between

88

fed or 24-hour fasted horses (Merritt et al., 1989). Water was also not restricted in this
study as it was desirable for the horses to have continual access to water as they would in
a clinical setting.

The dose of sand used in this study was determined through previous references
(Hammock et al., 1998; Ragle, Meagher, Schrader and Honnas, 1989) that were deemed
to be safe with no signs of colic. It was desired that there be enough sand to be visualized
on radiographs, a volume that has not yet been determined, but yet not enough to cause
clinical disease. The protocol used in this study was modified slightly fiom the
previously described protocols. Ragle, Meagher, Schrader and Honnas, (1989),
administered 4.2 g/Kg sand through nasogastric intubation mixed with 9 ml/Kg of
carboxymethylcellulose once a day for 5 straight days with no ill effects. Hammock et
al., 1998 administered 10 g/Kg sand into the cecum once and followed the horses through
for 11 days with no ill effects. This study’s protocol was to administer by nasogastric
intubation 4 g/Kg sand mixed with water once (Group 1) or twice (Group 2) a day for 2
consecutive days. From this protocol, 2 horses showed signs of colic. This study’s dose
is similar to Ragle, Meagher, Schrader and Honnas, (1989), however
carboxymethylcellulose was not used. Carboxymethylcellulose may reduce the irritation
normally caused by sand by reducing the abrasion between the sand and the intestinal
mucosa (Ragle, Meagher, Schrader and Honnas, 1989). This may account for the fact
that 2 of the horses in the present study showed signs of colic, however
carboxymethylcellulose is not a natural food consumed by horses and its presence may
affect the appetite, motility or radiographic appearance of the sand so it was not used.

Also, in the present study, the Group 2 horses received the same dose of sand but at

89

double the frequency as Ragle, Meagher, Schrader and Honnas, (1989) study. Although
the total volume of administered sand was the same (<10 Kg), this was accomplished in
half the time as the previous study. This again may account for the colic experienced by
2 of the Group 2 horses while none of the Group 1 horses showed any adverse effects. In
comparison to Hammock et al., (1998) study, the total volume of sand administered to
Group 2 horses was similar (10 Kg) however it was spread out over a 2 day period and
was administered to the stomach and not all at once into the cecum. This study’s
protocol for sand administration was felt to be safe since it was not going above and
beyond the total dose that the horses in Hammock et al., (1998) received which all had no
ill effects. The fact that 2 of the Group 2 horses experienced colic demonstrates that it is
not necessarily the total volume of sand that is responsible for causing colic, but is also
due to its effects on passage through the entire gastro-intestinal tract and over repeated
ingestions that eventually results in disease.

The selected type of housing and bedding could have an effect on the obtained
results. It would have been preferred to keep the horses outdoors on their normal grass
pastures, which is the typical setting for sand colic cases, however since this was a
controlled experiment, it was desired that the horses be kept indoors on cement floors so
that there was no access to any sand that may occur in the outdoor soil. Commercially
packaged wood shavings were used as bedding so as not to risk any sand being present in
the bedding that may occur with cut hay. Since changing the housing environment may
. induce behavioral changes or alterations in the feeding patterns of animals that may lead
to colic (White, 1990), especially a change from a pasture to a stall (Cohen, 2002), this

was a consideration during the study. Only Horse #5 (17-year old Arabian mare

9O

donation) did not experience a change in housing, bedding, or feeding styles as she had
been housed for some time prior to the study at the Large Animal Clinic Veterinary
Teaching Hospital. Most of the horses were noticed to occasionally paw the shavings,
which was interpreted as boredom from the change of housing being removed from an
outdoor pasture setting. Aside from this, no conclusions can be made as to any adverse
effect of the change of housing or bedding on the animals’ behavior or appetite as there
was no apparent clinical change in any of these factors. Again, only Horse #10 might be
the exception as he did experience colic that may be attributable to the change in housing
with resultant anorexia.

Of the risk factors for colic, Signalment has been positively associated with
certain types of colic (Cohen, 2002); however sand colic in particular does not appear to
have any predisposing Signalment factors. In general, older horses are at an increased
risk of colic as are Arabians (Cohen, 2002; Kaneene et al., 1997). The horses used in this
study were of varying age with the oldest being 26 years old who did not show any ill
effects throughout the study. The Arabian mare (Horse #5) also did not show any
adverse effects. The age and breed dispersion of the 10 horses was not a controllable
factor in this study, however they were all adults with breeds (Quarterhorse,
Thoroughbred, Arabian, and Standardbred) typical of presenting with sand colic at the
Large Animal Clinic Veterinary Teaching Hospital and also observed in a previous study
(White and Lessard, 1986). The one variant in the Signalment of the experimental horses
was Horse #10 who was the only gelding in the study, as the rest were mares. This is
most likely not the reason for the moderate colic observed in this horse, however, it is

another confounding factor with this horse in particular. Kaneene er al., (1997),

91

concluded that geldings were at a reduced risk for colic; however this appears to be
controversial. Equid activity may also be a predisposing risk factor for colic (Kaneene et
al., 1997), as racing horses may be at a reduced risk. Horse #10 was a previous racing
Thoroughbred prior to the study however due to the carpal chip induced lameness it was
retired from such activity. All other mares were pasture horses.

It was desired that the sand used in this study be representative of the sand that is
found in Michigan soil. Michigan sand is typically a silicate type that is either glacier
deposit and is sharp and not weathered, or is a dune deposit that is rounded and
weathered. Sand particle size in Michigan ranges from 2 11m — 2 mm (Wamcke, personal
communication). The sand used in this study was “fill-type”, unprocessed sand whose
range in particle size and distribution appears to be representative of typical Michigan
soil sand. This was an important factor in the design of this project as the design was to
try and mimic, as close as possible, natural conditions of sand colic. Although not
referenced, it is assumed that different sized sand particles may be more or less likely to
result in colonic accumulation and/or irritation resulting in colic. Larger particles may
settle-out from the ingesta faster than smaller particles, which may also become
incorporated within the digesta and passed out through the feces. It has been reported
that coarse sand frequently accumulates in the right dorsal colon, transverse colon and
pelvic flexure, and fine sand typically accumulates in the left and right ventral colon, and
associated stemal and diaphragmatic flexures (Snyder and Spier, 1992). Although sand
particle size could not be evaluated on the abdominal radiographs of the current project,
3/5 of the group 2 horses and 1/5 of the Group 1 horses showed evidence of passing

predominantly large-size sand particles at the end of the study (Day 6). This does appear

92

to give credence to the fact that larger particles may be more likely to accumulate and
cause colic or to have a longer intestinal transit time as compared to smaller sized sand
particles that tend to be incorporated within the colonic digesta and have normal
intestinal transit time, providing there is good motility.

Intestinal transit time is intimately linked to intestinal motility. Transit time is
defined as the time it takes for material to travel from one portion of the gastro-intestinal
tract to another. An increase in propulsive motility decreases transit time, whereas an
increase in retentive motility increases the transit time. There are many factors that affect
gastro-intestinal emptying rates and intestinal transit times such as the composition of the
material (liquid moves through faster than solid food), moisture content (high moisture
food empties faster than dry food), fat or fiber content (low fat - high fiber meals have a
more rapid gastric emptying rate than high fat - low fiber meals), volume of ingesta, the
chemical and physical properties of the ingesta (pH, temperature), various reflex
mechanisms (cephalogastric, gastrogastric, gastrocolic, enterogastric), medications
(alpha-2 adrenergic agonists such as xylazine), and the presence of pain (enteritis / colitis,
colic), emotional stress, noise, anxiety, physical restraint and gastric intubation which can
all slow emptying rates and transit times.

Gastro-intestinal motility is carried out through both an extrinsic and intrinsic
(enterie) nervous system. The extrinsic nervous system synapses with the intrinsic
system to modify its control through parasympathetic and sympathetic pathways. The
intrinsic (enterie) nervous system consists of a myenteric plexus that regulates the
circular and longitudinal smooth muscle layers, and a submucosal plexus which has

hormonal secretory control. The regulation of gastro-intestinal motility is through slow

93

waves of basic electrical rhythm (BER) that produce muscular tone to the intestinal tract
through the pacemaker cells of the enteric nervous system. The electric pacemaker for
the small intestine is in the longitudinal muscle of the proximal duodenum (Clark, 1990).
Slow waves initiated in the longitudinal muscle are conducted to the circular muscle layer
at a rate dependent on the electrical coupling between the 2 layers. Spike potentials from
the extrinsic nervous system added onto this BER causes contraction of the smooth
muscle, resulting in motility (Adams et al., 1984). During fasting, there is a migrating
motor complex that sweeps through the gastro-intestinal tract at regular intervals and acts
as a housekeeping function to remove indigestible solids and sweep bacteria towards and
through the colon. This migrating motor complex is characterized by cyclical alterations
between no spike potentials followed by intermittent spike potentials followed by regular
spike potentials with propulsion occurring during the intermittent and regular spike
potential phases (Adams et al., 1984).

Types of intestinal contractions include peristalsis that are propulsive movements
through circular muscular contraction progressing aborally (example: migrating motor
complex), segmental contractions that are localized and slow the aboral movement of
digesta to increase exposure for absorption and is more prominent in the small intestine,
and tonic contractions that are prolonged and function to isolate one segment of intestine
from another and also slows aboral movement of digesta (Clark, 1990). Mean transit
time (includes gastric emptying and small intestinal transit) to the cecum in fed horses is
approximately 3 hours (Clark, 1990). This is the result of gastric slow waves that occur
at a frequency of 2.5/min and small intestinal slow waves at 14.5/min (Merritt et al.,

1989).

94

The functions of the stomach, small intestine and large intestine are quite
different. The stomach stores and processes food so that it can be presented to the
duodenum at a controlled rate. The small intestine allows digestion and absorption of the
nutrients. The colon functions to absorb water and electrolytes, store feces and ferment
organic matter that escape digestion and absorption of the small intestine. It must absorb
a fluid volume approximately the same as the total extracellular fluid during 24 hours
(Snyder and Spier, 1996). Both liquid and particulate matter move through the cecum at
a relatively fast rate (Snyder and Spier, 1996) as the mean transit time through the cecum
for liquid is 5 hours (Clark, 1990).

Progressive motility abruptly decreases in the large colon as transit time through
this segment is about 50 hours (Clark, 1990). The ventral and dorsal colons are the major
retention sites of ingesta. Much of the soluble and most of the insoluble carbohydrates
pass through the small intestine, and therefore are digested in the large intestine. So
microbial digestion with volatile fatty acid production in the large intestine supplies a
large percentage of the horse’s energy requirements. This function of the cecum and
colon requires that the digesta be constantly mixed and retained long enough to complete
digestion of cellulose. Consequently, resistance to flow of ingesta is apparent at the
pelvic flexure and proximal transverse colon with no retrograde movement between
dorsal and ventral colon. As particle size increases (2 cm), a selective retention occurs in
the dorsal colon.

Motor activity in the large intestine originates in the cecum, proximal right ventral
colon, and pelvic flexure (Clark, 1990). The right and left ventral colon and pelvic

flexure both have oral and aboral directed motor activity, resulting in retention of ingesta

95

enabling microbial digestion (Snyder and Spier, 1996). The proximal colon has
intermittent periods of anti-peristalsis (originating from colonic pacemaker cells) where
food can stall for long periods of time allowing mixing and preventing passage until
absorption is complete. In addition, segmental and peristaltic contractions of the large
intestine also move ingesta aborally through the small colon to be voided. Motility and
contractions of the distal colon is similar to that of the small intestine.

Within the cecum, there is a coordinated series of spike bursts beginning at the
base and progressing towards the apex that is responsible for retropulsion and retention of
ingesta for bacterial fermentation and is initiated by a pacemaker in the cecal base
(Rutkowski et al., 1989). These spike bursts occur at velocity of 3.8 cm/sec at a
frequency of 0.46 spikes bursts/min. A second type of progressive series of spike bursts
begin at the cecal apex and are conducted through the cecocolic orifice and into the right
ventral colon and are responsible for progressive motility propelling food from the cecum
into right ventral colon and are initiated by a pacemaker in the cecal apex (Rutkowski et
al., 1989). These occur with a velocity of 6 cm/sec at a frequency of 0.4 spike
bursts/min.

Intramural ganglia (intrinsic plexus) within the large colon modulate the
contractile activity of the smooth muscle so appropriate contractions occur. lmpulses are
sent through mechanoreceptors and chemoreceptors in the smooth muscle. Thus luminal
distension from ingesta initiates a stimulus. The number of propulsive and retropulsive
contractions is directly related to the fullness of the cecum and large intestine (Clark,
1990). Normal colonic motor profile during interdigestive periods consist of spike bursts

of 5-12 seconds duration grouped in complexes that migrate from right ventral colon to

96

beyond the pelvic flexure and of localized spike bursts of under 5 seconds duration
(Roger and Ruckebusch, 1987). These short spike burst are associated with mixing
contractions while the long spike bursts are associated with retropulsion and propulsion.
It is alterations in these short and long spike bursts that result in either constipation or
hypermotility (Adams et al., 1984).

The above physiological description of normal gastro-intestinal function and
motility illustrates how complex it is and how many factors may be involved in the
process. Since the large colon functions in water absorption and microbial digestion,
conditions interfering with these processes may result in severe dehydration and luminal
distension, with subsequent pain, shock and colic (Snyder and Spier, 1992). Large colon
intra-luminal obstructions in general tend to occur at locations of normal anatomical
luminal narrowing or at pacemaker centers including the cecocolic orifice, the pelvic
flexure and the transverse colon (Allen and Tyler, 1990). Sand accumulations resulting
in colic have been commonly observed to be found in the right dorsal colon, transverse
colon, left and right ventral colons, the stemal and pelvic flexures and the small colon
(Ragle, Meagher, Lacroix and Honnas, 1989) with there being some debate as to whether
sand of different particle size accumulates in different segments (Snyder and Spier, 1992;
Hammock et al., 1998).

With sand enteropathy, the sand may irritate the colonic mucosa and induce
diarrhea resulting in decreased transit times, and subsequent dehydration and weight loss.
Sand may also settle out from the digesta and accumulate in the colonic lumen. Sand
may collect in the large colon because it is a fermentative area of reduced flow and the

sand would consequently settle out in this segment of the gastro-intestinal tract as

97

opposed to the stomach or small intestine where motility is strong, fast and aborally
propulsive (Colahan, 1987). It is for this reason, that abdominal radiographs typically
demonstrate sand accumulations in the cranioventral abdomen where the left ventral
colon turns at the stemal flexure into the right dorsal colon (Bertone et al., 1988; Ford
and Lokai, 1979; Ruohonierni et al., 2001). A study by Ragle, Meagher, Schrader and
Honnas, (1989), demonstrated that horses that are non-obstructive typically show sand in
the cranioventral abdomen on radiographs (in the left ventral colon and stemal flexure).
A retrospective study of surgical cases of sand colic by Ragle, Meagher, Lacroix and
Honnas (1989), showed that the majority of sand was found be accumulating and
obstructing the right dorsal colon. There then appears to be a disparity in locations where
sand accumulates, depending if the horse is obstructive (requiring surgery) or not,
although they may be clinical for sand colic in both instances.

The results of this present study support this concept as almost all of the
accumulations of sand were visualized on abdominal radiographs in the cranioventral and
mid-ventral abdomen, corresponding to the left ventral colon and stemal flexure. These
horses either did not show any adverse clinical signs or showed mild and brief signs of
colic. The only horse that showed significant clinical signs of colic and was not passing
feces and not responding to medical therapy was Horse #10. This horse was euthanized
as it was deemed a surgical case of sand colic. Interestingly enough, 2 of the 3
accumulations of sand in Horse #10 were seen in locations not typically seen throughout
the rest of the study, presumed to be the right dorsal colon, which adds credence to the

study of Ragle, Meagher, Lacroix and Honnas (1989).

98

Altered motility is another key feature of gastro-intestinal inflammation. Both
increased and decreased motility can occur from inflammatory diseases with diarrhea
being the result of increased motility. Motility was not a specific focus in the present
study; however some discussion can be made in light of the interpretations of the
obtained data. From the above stated literature, it appears that total normal gastro-
intestinal transit time approximates 50-60 hours (2-2.5 days). Castro-intestinal motility
was assessed in each patient at regular intervals as part of a health check. All patients
had normal auscultable motility as determined by characteristic sounds (borborygrni)
heard at regular intervals. Mild diarrhea was noticed in 6/10 of the horses (all 5 Group 2
horses, 1 Group 1 horse) between Day 2 and Day 3. This was assumed to be due to the
large volume of water that was administered with the sand through nasogastric
intubation. This time frame appears to coincide with the estimated gastro-intestinal
transit time mentioned above. Fecal sand was present in significant amounts starting on
Day 2 in both groups, usually coinciding with the diarrhea. This again demonstrates
normal gastro-intestinal motility. From this point in time on, there is a divergence in the
sand output between the 2 groups. Although both groups started producing sand in their
feces by Day 2, Group 2 then declined their fecal-sand output while Group 1 showed
evidence of high fecal-sand output through Day 4 when they then started to decline to
baseline (0) levels. This timepoint (Day 4) also corresponds to 2 days following their last
dose of sand, which indicates that the gastro-intestinal motility and transit times were
unaltered in the Group 1 horses. This demonstrates that a dose of sand of 4 g/Kg once a
day for 2 days (total amount not above 5 Kg) does not appear to affect their normal

gastro-intestinal tract physiology. Group 2 horses showed a decline in fecal-sand output

99

following Day 2 and did not show a significant difference in their output pattern from
Group 1 until Day 5 when their output increased to a level equal that of Group 1 on Day
3 (their day of maximum sand output) and continued to plateau at the end of the study.
From this can be inferred that this higher dose of sand (4 g/Kg twice a day for 2 days,
total amount not above 10 Kg) can induce altered gastro-intestinal physiology resulting in
prolonged transit times from the accumulation and irritation processes of the sand. It is
interesting that the auscultable motility of the intestinal tract was always within normal
limits. Auscultation cannot differentiate between the different types of intestinal
contraction (propulsive or retentive) so it can be assumed from this study that the sand
caused a decrease in the peristaltic propulsive contractions and increased in the retentive
(“mixing”) type of contractions in an attempt to dislodge the sand accumulations.

The significant difference of fecal-sand output between the 2 groups is shown
graphically in Figure 2-11 as the 2 lines diverge from one another after Day 2 and from
that point on there are only 2 instances (Day 3 and 4.5) where the standard error of the
means overlaps between the two. Although the total amount of sand output was higher in
Group 2 horses, this was not evident until approximately Day 4.5 of the study. Before
this day, Group 1 horses had a significantly different (higher) fecal sand output than
Group 2.

There are some limitations to these interpretations of gastro-intestinal motility and
transit times based on fecal-sand output. Through the method described for fecal-sand
collection in the Chapter 2 Materials and Methods section, there is obviously some loss
of sand in the procedure as some may not settle out from the feces and would be washed

away in the process. A 94% recovery rate was determined for this procedure; however

100

this percentage may vary according to sand particle size and fecal consistency. At most,
this limitation would increase the actual amount of fecal-sand output as seen in Figure 2-
11. If a 94% recovery rate can be assumed, there is a significant difference between the 2
groups for percent of sand passed compared to that administered over the 6-day period (p
= 0.020, Table 2-1). The average percentage of fecal-sand output for Group 1 is 33.9%
while Group 2 only passed on average 18.9% of the total sand administered. This means
that the horses were discharged with roughly 66% (Group 1) and 81% (Group 2) of the
sand still within their gastro-intestinal tract. This corresponds to, on average, 1.4 Kg for
Group 1, and 3.6 Kg for Group 2. Interestingly enough, 3/5 of the Group 1 horses and
2/5 of the Group 2 horses did not show any radiographical evidence of intra-abdominal
sand accumulation at the time of discharge. This means that at least 3.6 Kg sand may be
within the gastro-intestinal tract and not evident on radiographs. This depends of course
on the distribution of the sand as less than 2 Kg of sand could be clearly seen
accumulated within the stomach, but the same quantity may not be visible if dispersed
throughout the colon with no significant accumulations present. According to Figure 2-
19, sand quantities ranging from 1800-6670 g were not radiographically visible as an
objective score of 0 was assigned. This particular score however was assigned on a day
(Day 0.5) after this same quantity was seen as an accumulation in the stomach aftei'
administration. It is most likely due to the fact that the sand was now dispersed within
the small intestines which were not visualized adequately during the study. These results
also demonstrate that radiographs may not depict gastro-intestinal sand up to a weight of
6670 g. This may be significant especially given the fact that the amount required to

cause clinical signs is not known (Bertone er al., 1988). There were no quantities of sand

101

given that were smaller than 1800 g to assess how much sand there needs to be before it
is radiographically visible.

The difference in percentage of sand output between the 2 groups is again most
likely reflective of the delayed transit time and decrease in the propulsive intestinal
contractions that was experienced by the Group 2 horses. It would be interesting to note
how long it would have taken these horses to reach the same level of sand output (33%)
as the Group 1 horses had. Also of note, is that despite there being still significant sand
within the gastro-intestinal tract of all horses at 6 days, their 12-hour output was not
indicative of this at the end of the study. Even Group 1 horses that had approximately
66% of the sand still inside their intestinal tract had a mean 12-hour output of 35 g over
the last 3 sampling periods (Day 5-6). This is in comparison to a mean l2-hour output of
412.4 g at the time of their maximum fecal—sand output (Day 3.5). This would tend to
indicate that at least in these Group 1 horses, who were interpreted to have a normal
gastro-intestinal physiology despite the sand, passed a larger percentage of the sand out
that was carried through the intestinal tract within the normal transit time (2 days) and the
remainder was slowly passing out with subsequent gastro-intestinal movements.
Although a plateau level of sand output was reached at the 6 day period for the Group 2
horses, and a decline was not observed, it can be assumed that these horses would
eventually follow the pattern of the Group 1 horses and decline in their output of sand
with still a large percentage within their intestines that would slowly pass in small
amounts over time. This variation in fecal sand output demonstrates that there is indeed
poor correlation between such diagnostic methods and establishing a diagnosis of sand

colic and that even small amounts of sand may be passing in the feces in clinically

102

normal horses. This is in agreement with other current literature (Snyder and Spier,
1992); however previous reports stated that a positive diagnosis can always be made by
rectal sand recovery (Ferraro, 1973).

Another limitation in the interpretation of this data is that it is not possible to
assume that all of the sand passed with the feces could be collected. There is obviously a
chance that not all the feces were collected however an attempt was made to be thorough.
The major problem arose from the fact that 6/10 horses experienced diarrhea to a mild
degree due to the large volume of water given to administer the sand. There is therefore
some degree of insensible loss of sand through the watery feces that could not be reliably
collected. The bedding was made of wood shavings, and if diarrhea was passed onto the
shavings, an attempt was made to dump the pile of shavings into the water bucket to
settle out the sand. Diarrhea that passed over the stall walls or onto bare floor could not
be collected reliably. This again would provide in an underestimation of the sand passed
by the horses. It is again interesting to note that the diarrhea occurred over the Day 2-3
period of the study, which was the point of increase in Group 1 fecal-sand output, and a
point of decrease in the Group 2 horses. This may mean that in fact Group 2 horses never
experienced a decline in their sand output as Figure 2-1 1 indicates, or that Group 1 horses
may have experienced an earlier rise in their sand output which may reflect an increase in
the gastro-intestinal transit and motility that may be seen with diarrhea.

Another limitation in the interpretation of the results is the lack of a complete
control in Group 1. In retrospect, it would have been more scientifically correct to
sedate, nasogastrically intubate and administer 15 liters of water to all of the Group 1

horses at the same frequency as Group 2 horses. This would have assured that the results

103

obtained would have been truly due to the difference in the quantity of sand administered
between the 2 groups only. Whether the colic seen in Group 2, its fecal sand passage, or
its objective and subjective scores could actually be due to the additional sedation, water
content or nasogastric intubations and not actually from the increased sand quantity is
speculative. Theoretically, the additional water that Group 2 horses received may have
resulted in more insensible sand loss through diarrhea and have falsely lowered their
fecal sand output. Likewise, the additional water in Group 2 horses may have altered the
radiographic appearance and structure of the sand accumulation and subsequently altered
the objective and subjective scores. Perhaps if Group 1 horses had received the same
dose of water as Group 2 horses, it would have intermixed with the sand, resulting in
smaller and more dispersed accumulations and their scores would actually be lower than
what was observed. This would theoretically strengthen the differences seen between the
2 groups. Had Group 1 horses been subject to the same amount of sedation and
intubations as Group 2 horses, they may have experienced altered (slowed) gastric
emptying and intestinal transit times. This in turn may have lead to signs of colic, and
prolonged retention of sand in the stomach as Horses #6 and 10 experienced.

Another goal of this study was to develop a technical protocol for radiographic
imaging for suspect cases of sand colic. It was determined from the retrospective part of
the study that poor radiographic quality had a negative effect on the establishment of a
diagnosis of sand colic, which was also concluded in another similar study with
enteroliths (Yarbrough et al., 1994). It was also observed that many studies did not have
a complete view of the entire abdomen to evaluate and interpret. All radiographs

evaluated in the retrospective study were in a right to left lateral view. The results of the

104

prospective part of the study indicate that high exposure factors (KVp > 90, mA >400)
are required for radiographic imaging of the equine abdomen, in agreement with current
literature (Butler et al., 2000). In some cases however, this still is not enough for
complete penetration as was seen in the central abdomen in some of the experimental
horses. However, since this region corresponds to the small intestine predominantly and
that sand does not typically accumulate in this region, the exposure factors used are
deemed satisfactory for a diagnostic study.

A mainstay in diagnostic radiology is that 2 orthogonal views are required for full
assessment of a certain structure in order to obtain a 3-dimentional image from 2-
dimentional radiographs. Since this is not possible in awake adult equines, standing
lateral views are obtained, and usually from only one side of the abdomen. Additional
benefits of opposite lateral views are obvious as a lesion may be better visualized if it
closer to the film. This shorter object-film distance results in better resolution, which
may improve the interpretation and diagnosis. It was determined from this study that 2
opposite lateral views are not necessary for a diagnosis of sand accumulation in horses.
The additional view did not significantly alter the final score or diagnosis in either
subjective or objective methods of assessment. An orthogonal view, such as a
ventrodorsal would be obviously beneficial as it may provide more information as to the
true size of the accumulation, however is not feasible in routine cases for "sand-scans".
Also, it was discovered from the retrospective part of the study that using 1 lateral view
with an objective method of assessment, the obtained score has a high correlation with a

diagnosis, especially if a score of 7 or higher is obtained.

105

Using the described 12-0verlapping-image method to radiograph the abdomen
allowed clear visualization of the individual sections of the gastro—intestinal tract.
Stomach, dorsal and ventral colon, stemal flexures, diaphragmatic flexures, pelvic
flexures and small colon could all be differentiated. Small intestine and cecum could not
be reliably seen, nor could differentiation between left or right colon be made. Sand was
seen in all locations except in the dorsal abdomen and in the mid-central abdomen. The
dorsal abdomen consisted primarily of the gas caps of the stomach and intestines, and the
mid-central abdomen consisted primarily of the small intestines and dorsal colons.
Overlapping of the images provided assurance that parts of the intestine and/or sand
accumulation were not missed on interpretation, and that the entire accumulation could
be measured.

Based on the above findings, the following recommendations are made for
radiographic imaging of an adult equine for sand colic. Using exposure factors of 100
KVp and >400 mA, good quality overlapping radiographs of the entire ventral 2/3 of the
abdomen are recommended from 1 side only. Despite that few radiographs had any
discernible findings in the mid-central abdomen, this region is still recommended to be
radiographed. This region may show sand accumulations in the dorsal colon that also
resides in this area as this area is more commonly obstructed in surgical cases, as was
seen in Horse #10. Radiographing the majority of the abdomen for assessing sand
accumulation is against many popular beliefs that state that the cranioventral abdomen is
alone sufficient (Ruohoneimi et al., 2001 ). This study clearly demonstrates that this is
insufficient and would result in a potentially non-diagnostic study. Whereas in certain

cases, such as enterolithiasis, the entire abdomen need not be radiographed if an

106

enterolith can be visualized as a diagnosis can be made (Y arbrough et al., 1994). If cases
of sand enteropathy, just visualizing l accumulation is not sufficient, and the entire
abdomen need be searched or imaged so as to provide a complete assessment of the
gastro-intestinal sand. As sand was never seen in the dorsal 1/3 of the abdomen, and that
there was no correlation of gas-distended intestines with sand colic (see Retrospective
part), the dorsal 1/3 of the abdomen is not recommended to be imaged if sand colic is the
primary or sole differential. If other causes of colic are sought, the entire abdomen need
be radiographed as this dorsal section may provide vital information on certain cases,
such as gas distention from colonic entrapment or torsion.

The radiographic appearance of the sand seen in the experimental horses closely
resembles the description of other sand accumulations from reported cases (Ruohoneimi
et al., 2001). This is despite the fact that in the current study, sand was administered as a
bolus to the stomach and still appears radiographically similar to sand that is chronically
ingested. Sand accumulations varied from distinct to dispersed within digesta, usually
tapering cranial-caudal margins, convex ventral margins, and horizontal or irregular
dorsal margins. The ventral margins were usually smooth and concave because they
were resting along the floor of the colon and usually had haustra defining their location.
The dorsal margin was often more indistinct than the ventral margins because of the
overlying or intermixed ingesta. Accumulations within the stemal flexure had rounded
cranial margins and were located more caudoventral than accumulations in the
diaphragmatic flexures, which also had rounded cranial margins. The stomach could be
differentiated based on its location being more dorsal and caudal than the colonic

flexures. The dorsal margins of the stomach sand accumulations always had straight,

107

horizontal borders due to the sand-water interface from the intubation procedure. The
size and consistency of the various accumulations could easily be appreciated; however
overlapping radiographs were essential for full assessment. Ruohoneimi et al., (2001)
also states that although large accumulations may not be entirely visible on the
cranioventral abdominal radiograph, the average size and appearance can be assumed
from the portion that is seen. The above results demonstrate that if this approach was
used, false interpretations may occur as it is clear that the accumulations may drastically
change shape, size, and consistency from 1 end to the other of the accumulation.
Whereas the cranial extent of an accumulation may appear large and homogeneous and
opaque, it may become more narrow and small and heterogeneous and less opaque
further caudally. It is also important that the entire accumulation to be visible for full
assessment for the objective scoring assessment to work properly. This again states the
importance of a complete abdominal radiographic series.

Results from the Retrospective part of the study demonstrated that current
subjective methods of assessing sand accumulation and correlating these assessments
with a diagnosis is unreliable and inaccurate. This theory needed to be tested on a
prospective model that attempted to mimic a clinical setting. With differences in known
amounts of intra-abdominal sand content between the 2 groups (Group 1 having roughly
half the quantity that Group 2 had), subjective methods of assessment could not reliably
differentiate the 2 groups as being separate. Using the devised objective scoring
assessment, Group 1 and Group 2 could be differentiated. The obtained results support
the proposed hypothesis. This indicates that current subjective methods are indeed highly

inaccurate as they may not allow radiographic differentiation between a volume of sand

108

that is half of another. Because of this, it is advised that subjective assessment not be
used when attempting to establish a diagnosis of sand colic or to interpret relative
quantity of sand seen on the radiographs and correlate these assessments with a clinical
diagnosis. Colic from sand accumulation may result from many factors such as its
overall length and height, its opacity and homogeneity and its location within the
abdomen. Only when all of these parameters are assessed can an adequate evaluation be
made as to the degree (severity, quantity) of the sand accumulation. Using the devised
objective scoring method allows each of these parameters to be individually assessed and
included into a total score for an individual horse. Further studies are recommended to
assess how sensitive this testing method is to be able to differentiate amounts of sand. By
using smaller divisions between the 2 groups, the limits of this scoring method can then
be assumed.

The objective scoring method was also touted to be relatively accurate in its
assessment for a positive diagnosis of sand colic. The Retrospective part of the study
demonstrated that an objective score of at least 7/12 was 83% precise in diagnosing colic
due to the sand accumulation. Precision values for scores above and below this value are
not known, but can simply be inferred. For example a score of 12/12 does not guarantee
100% that there is sand colic, it simply means that it is highly likely. The scores obtained
in this experimental study confirm these findings from the Retrospective part of the
study. The highest score that any of the Group 1 horses received was a 7 for Horse #3 on
Day 0.5. This day always provided the highest score for this group as 4 horses had a
score of 6 and the other horse had a score of 5. This high score reflects the entire dose of

sand within the stomach immediately following nasogastric intubation and administration

109

of the sand and therefore is misleading. The sand seen in the stomach following
administration is very opaque, homogeneous and has significant height (due to the shape
of the stomach). If such an accumulation were to be seen fi1rther distally along the
gastro-intestinal tract, this would result in a score in a similar range (5-7) but would be
more meaningful since it would represent a colonic (not gastric) accumulation. The
highest score obtained by the Group 1 horses following Day 0.5 was a 6/ 12 which means
that despite more sand present than on Day 0.5, it is distributed throughout the intestinal
tract and not significantly accumulating. To add credence to this, none of the Group 1
horses showed any signs of colic, and none had a post Day 0.5 score of 7 or higher.
Whether Horse #3 or 4 were close to experiencing colic on Day 3-4 of the study
(objective scores of 6), is speculative and again states that despite a fairly significant
accumulation of sand, clinical signs may not be present, and that other causes should be
sought if clinical signs are present.

Group 2 horses also showed a repeatedly high score at Day 0.5 following gastric
sand administration as all 5 horses had a score of 5-6. At times following this, Horse #6
had a hi gh-score of 8 on Day 8 and showed signs of colic on that day, Horse #7 had a
high-score of 7 once on Day 3 and did not colic, Horse #8 had a high-score of 7 on Day 3
and did not colic, Horse #9 had a hi gh-score of 7 on Days 4 and 4.5 and did not colic, and
Horse #10 had a high-score of 8 over Days 4.5-5 and did colic on those days. The
average Group 2 hi gh-score after Day 0.5 was still only 6, which was on Day 3. These
results demonstrate that all of Group 2 horses had scores reaching at least 7 sometime

during their study and only 2/5 (40%) of these demonstrated colic, which were those 2

110

horses that had scores of 8/ 12. Horse 9 did not colic despite having a score of 7/12 over 2
imaging periods.

These results can be explained because despite the Retrospective study claiming
that a score of 7/ 12 is 83% precise in diagnosing sand colic, there are 17% of cases that
will be misdiagnosed. These differences may also be explained because this was an
experimental setting and does not truly mimic natural sand colics that were the types of
cases that were evaluated in the Retrospective part of the study. Perhaps, the horses were
demonstrating more subtle signs of colic that were either not noticed or misinterpreted.
For example, the pawing of the bedding was interpreted as being due to boredom from
the change from pasture to a stall and may have been a sign of colic. There were
however, no additional supportive findings of colic (such as tachycardia, anorexia,
decreased intestinal motility, or lack of fecal production). This does however
demonstrate that this relative ordinal scale of 0-1 2 does work as all of the horses (2/2)
that had scores of 8/ 12 did show signs of colic.

The objective scores over the duration of the 6-day study demonstrate that the
average group score decreased following Day 0.5 due to distribution of the sand in the
intestines following gastric emptying (Figure 2-12). Scores obtained at this time were
low (1-4) because despite there being the same amount of sand in the abdomen as before,
it is now within the small intestines mixed with the chyme, not accumulating, and located
in an area prone to underexposure. As intestinal motility moved the sand further into the
large colon, scores again increased from Day 1.5 - 3.5 due to the visualization of the
sand, and its accumulation in the ventral colon. Following this time period, Group 2's

scores decline much more slowly than Group 1's scores because of the prolonged

lll

accumulation and increased amount of intestinal sand present in this group. This clearly
demonstrates the ability of the objective scoring method to differentiate between horses
with different amounts of sand.

Because one of the objectives of this scoring system was to find out if it could tell
relative quantities of sand visible on radiographs, the objective scores were plotted
against the weight of the sand that was known to be present in the abdomen at that time
(Figure 2-19). The results indicate that this method cannot predict quantity of sand
present in the horse, as the score is based on many factors, not just the weight of the sand.
The graph shows a prominent dispersion of values around the regression line indicating a
poor cause and effect relationship. Even making the above comparisons within each
group, and with each horse within each group, does not demonstrate any relationship.
Previous work states that radiographs can clearly show the amount of sand in the
cranioventral abdomen, (Ruohoneimi et al., 2001); however this study does not support
this statement. The chart in Figure 2-19 depicts how many scores of 0 were given to
cases that had relatively high amounts of intra-abdominal sand, but this sand might not
have been radiographically apparent. Since it was found that a score of 7/1 2 could
predict sand colic with 83% precision, extrapolating from Figure 2-19 shows that a score
of 7-8 requires approximately 11,000-12,000 g of sand. When extrapolating from each
individual horse's chart (Appendix E), although certain individuals showed signs of colic
with sand weights of only 6000 g, the average trend shows that approximately 10,000 g
(10 Kg) may be required to cause signs of colic, or obtain an objective score of 7/1 2.
This is in agreement with some research (Ragle, Meagher, Lacroix and Honnas, 1989),

yet contradictory to another that states that over 20 Kg is required (Ragle, Meagher,

112

Schrader and Honnas, 1989). This study was also designed on the premise that 10 Kg
(the maximum total dose) would not cause clinical signs. The above interpretation of
required weight is made in light of the acknowledged differences between this study and
naturally occurring sand colic.

The fact that Group 2 horses exemplify better relationships between their
objectives score and the amount of intra-abdominal sand on particular days as compared
to Group 1 horses is interesting. Group 2 had relatively strong relationships between
their score and the amount of sand on Day 2.5, 3, 3.5 and 4 whereas Group 1 never had a
strong relationship on any given day. This may mean that a certain threshold quantity of
sand is required in order for this objective scoring method to show relationships with the
weight of the sand and that quantities below this value are either not radiographically
visible or are not accumulating in a fashion to demonstrate a relationship. What this
threshold level may be is unknown, and further studies would be recommended to see
how sensitive this objective scoring method is to a certain baseline level of sand.

There are some significant differences between the 2 groups in terms of the
individual parameter scores over time. Group 2 had a higher location score than Group 1
in the 2“d half of the study, indicating that they often had their accumulating sand
localized to the cranioventral abdomen (Figure 2-13). This was the most standard
location for the sand to accumulate in these horses however, for reasons mentioned
previously, entire abdominal radiographs are advised so as not to miss any accumulations
for evaluation. Sand was commonly seen in the mid-ventral abdomen, which would be
missed on cranioventral abdominal views only. It is also observed that it takes

approximately 2 days before the sand is seen accumulating in the cranioventral abdomen.

113

Before this point, sand is within the small intestine and cecum or is passing through the
colon in small amounts.

Group 2 horses had a significantly higher number of accumulations score than
Group 1 horses, especially towards the end of the study (Figure 2-14). During this time
(Day 4 onward), it was common not to see any sand accumulating in the Group 1 horses,
while Group 2 horses still had a distinct accumulation visible. Horse #10 had 3 distinct
accumulations seen towards the end of the study and showed signs of colic. These results
are in agreement with another report that states that the number of accumulations has a
positive association with clinical signs (Kaikkonen et al., 2000).

Opacity and homogeneity scores were significantly higher in Group 2 horses
compared to Group 1 (Figures 2-15, 2-16). This means that Group 2 horses had more
dense accumulations than Group 1 horses due to the poor motility and clearance. This
increase in density of the accumulation will tend to lead to more focal pressure and
weight to that part of the colon, causing an increased likelihood of clinical signs.
Interestingly, it was only sand homogeneity and opacity that was found to be significantly
correlated to one another when comparing the individual parameters together. This can
be interpreted that as once an accumulation becomes more opaque, it also becomes more
homogeneous. This is inferred as a more solid and denser accumulation of pure sand.
Group 1 horses had lower score because the lower volume of sand did not cause altered
gastro-intestinal motility and the sand remained interspersed with the digesta. Also of
note, the highest opacity or homogeneity score never got as high as the score given to the

stomach accumulation that had a maximum value of 2 in both groups for both

114

parameters. So despite the increase in sand weight over time, it never accumulates to
such density as is seen in the stomach immediately following intubation.

Although sand accumulation length and height did not show any significant
difference between the 2 groups there is a visible difference (Figures 2-17, 2-18). This
again demonstrates that the overall shape of the accumulation alone does not have a
relationship with clinical signs. Accumulations may vary from long and skinny to short
and thick, but inferring a diagnosis from such appearances will be misleading. Only
Group 2 had sand height scores that were higher than the height score given for the
stomach and these scores were seen on Day 3 and 3.5 when significant accumulation
started to develop. This can create an impression as to how large some of these
accumulations were in these horses. Group 1's sand height scores never equaled their
stomach's sand height score as their sand never accumulated to such degree.

From the above, it can be seen that objective scores are based on not only the
weight, but how many accumulations it is formed in, how dense it is, where it is in the
gastro—intestinal tract, and if is a long and skinny accumulation or a thick accumulation.
Even comparing the individual parameter scores to the known amount of sand present did
not reveal any strong relationships. This demonstrates that sand colic is multifactorial in
its etiology and is not only based on 1 parameter such as weight, density, size, or location
of the sand but depends on the interrelationships of all of these parameters acting together
to cause the resultant signs of colic.

The poor relationships and correlations demonstrated in the above results for the
various comparisons may be in fact due to the small sample size used (2 groups of 5

horses) or population homogeneity and may require larger group sizes in order to

115

demonstrate significant relationships. This sample size was found to be adequate
however, for differentiating the 2 groups based on the objective scoring system (power =
96%). This method of assessment was critical and sufficiently objective enough that it
could detect differences between the 2 groups using just 5 horses per group. The
subjective method of assessment had a low power (25.6%) in its ability to test differences
between the 2 groups. Using the subjective method of assessment, approximately 25
horses per group would have been needed to see a difference. This again demonstrates
the improvement of the objective method from the subjective method of radiographic

assessment of sand accumulations.

116

0.: CON» SN: :0: o:
EBN ooow ENN EON 0
Wm 00me :mn cox. m
c: ooow 02— N2: 0
N.0N GONG 00ch mNmN :
:._N oocm w: :0 m
m.mm 000m me: 092 v
mim comm 3: x:— m
:0 000m :OMN mEN N
odm CON: :02 w _ N_ _

 

 

 

 

:0800 :00m 00 000003. .0000<

8000—.00 :00m 00 0000:

 

:00 0 02:. 08:0 0.5.0 0.. .x.

.000 000.0800 £000 0 00 8.05 .000:

030 000: 00: 00.: mv<w v .00 0000000008060 0::m0w0000 “00:50:00 :00000 0:000 0 00>0 :00000 :000 .00 0w00000000

TN 030:.

117

Conclusion

The objective of this study was to improve the sensitivity and specificity of
abdominal radiography as a diagnostic tool for intestinal sand accumulation in horses.
The goal was to develop an objective method of radiographic interpretation of intestinal
sand accumulation on equine abdominal radiographs. It is felt that this study
demonstrates that this goal has been achieved and the hypothesis proven correct.
Through the creation of an objective scoring system, sand accumulation on equine
abdominal radiographs can now be assessed in a more reliable, reproducible and accurate
fashion than current subjective methods of assessment. Objective assessments can
reliably differentiate different amounts of sand on radiographs that subjective methods
could not.

The devised objective scoring system uses the combination of a subset of scores
from 6 parameters (location, number of accumulations, opacity, homogeneity, length and
height) to provide a total score ranging from 0-12. It was found that a score of 7 or
higher had an 83% chance of correctly predicting sand colic. This objective method of
assessment was then tested on experimental horses that were nasogastrically administered
sand in 2 different doses, Group 1 receiving half of the dose that Group 2 received. By
examining the radiographs following sand administration, objective methods of
assessment could differentiate the 2 groups whereas subjective methods could not. It was
also relatively accurate method of assessment as all (2/2) horses that demonstrated
clinical signs of colic had objective scores of 8/ 12 at the times of colic. There was no

score given higher than 7 for horses that did not experience colic.

118

Radiographic interpretation of sand accumulation goes beyond trying to estimate
the approximate amount of sand in the gastro-intestinal tract. Clinical signs of colic
depend not only on the total weight of the offending sand, but also on where it is located,
how many accumulations are present, its density and its combined length and height. It is
all of these interrelationships that need to be considered when evaluating abdominal
radiographs.

It was determined that this protocol (4 g/Kg sand bid for 2 days) for the Group 2
horses may induce altered gastro-intestinal motility and result in relatively significant
sand accumulation and cause colic. In the one horse that experienced signs of colic
sufficient to warrant surgery, sand was visualized in the dorsal colon as opposed to the
ventral colon that was the most common sight of accumulation in asymptomatic or
medically responsive colic cases.

It was determined that radiographing from 1 side of the abdomen only is
satisfactory as opposite lateral views do not significantly alter the interpretation. The
ventral 2/ 3 of the abdomen should be radiographed however and not just the
cranioventral abdomen if searching for signs of sand accumulation.

Results from this study should aid in establishing a protocol for abdominal
radiography for sand colic, and provide a more accurate method of assessing the sand
accumulation. This in turn will provide a better assessment of the severity of the disease
and subsequently a better and more accurate therapeutic regimen. Results from this
research should aid in establishing better correlations between the diagnostics, clinical

examination and surgical findings, which is often a problem with sand colic.

119

Future tangents of this research may involve establishing a minimum quantity of
sand that is radiographically visible within the colon utilizing the objective scoring
system. This would help in then knowing how much sand there has to be in order for a
score to be assigned. As well, future work may also involve establishing a more strict
sensitivity of this objective scoring system. The current project demonstrated that this
method can differentiate 2 doses, one being double the other. Performing this research
with more groups and with smaller differences in the doses between the 2 groups may
establish the limits of the benefits of the objective scoring system. Correlating all of the
above results with surgical or necropsy findings will also help in validating this proposed
new objective assessment system for sand accumulation on equine abdominal

radiographs.

120

REFERENCES

Adams, S., Lamar, C., Masty, J. 1984. Motility of the distal portion of the jejunum and
pelvic flexure in ponies: Effects of six drugs. Am J Vet Res. 45. 795-800.

Allen D, Tyler D. Pathophysiology of acute abdominal disease. In: White, N, ed. 1990.
The Equine Acute Abdomen. Lea and Febiger. Philadelphia, USA. 65-87.

Bertone, J .J ., Traub-Dargatz, J .L., Wrigley, R.W., Bennett, D.G., Williams, R.J. 1988.
Diarrhea associated with sand in the gastrointestinal tract of horses. I Am Vet Med
Assoc. 193 (11). 1409-1412.

Butler, J.A., Colles, C.M., Dyson, S.J., Kold, 8.13., Poulos, P.W. 2000. Clinical
Radiology of the Horse. Second edition. Blackwell Science. Oxford, United Kingdom.

Campbell, M.L., Ackerman, N., Peyton, LC. 1984. Radiographic gastrointestinal
anatomy of the foal. Vet Radiol. 25 (5). 194-204.

Clark E. Intestinal motility. In: White N, ed. 1990. The Equine Acute Abdomen. Lea
and Febiger. Philadelphia, USA. 36-48.

Clark, E., Thompson, S., Becht, J ., Moore, J. 1988. Effects of xylazine on cecal
mechanical activity and cecal blood flow in healthy horses. Am J Vet Res. 49 (5). 720-
723.

Cohen, N. Etiology, risk factors, and pathophysiology of colic. In: Mair, T., Divers, T.,
Ducharrne, N., eds. 2002. Manual of Equine Gastroenterology. W.B. Saunders. London.
101-106

Colahan P. Sand colic. In: Robinson, N, ed. Current Therapy in Equine Medicine, 2nd
edition. 1987. W.B. Saunders. Philadelphia, USA. 55-58.

Curry T, Dowdy J, Murry R. 1990. Christensen's Physics of Diagnostic Radiology, 4th
edition. Lippincott Williams and Wilkins. Philadelphia, USA.

Doherty T, Andrews F, Provenza M, Frazier D. 1999. The effect of sedation on gastric
emptying of a liquid marker in ponies. Vet Surg. 28(5). 375-379.

Doran, R. 1993. Field management of simple intestinal obstruction in horses.
Compendium of Continuing Education of Practicing Veterinarians. 15 (3). 463-471.

Ferraro, G.L. 1973. Diagnosis and treatment of sand colic in the horse. Vet Med Small
Anim Clin. 68. 736.

121

Fischer, A.T. 1997. Advances in diagnostic techniques for horses with colic. Vet Clin
North Am Equine Pract. 13 (2). 203-219.

Ford, J., Lokai, MD. 1979. Complications of sand impaction colic. Vet Med Small Anim
Clin. 74. 573-578.

Freeman D. Pathophysiology of intestinal obstruction. In: Mair, T., Divers, T.,
Ducharrne, N., eds. 2002. Manual of Equine Gastroenterology. W.B. Saunders. London.
103-106

General Electric. Direction 46-014428, revision 9. Tube ratings Maxiray 125, 50 and 60

Hz, 10° (175 mRad) target angle; Product operating documentation. 1990. General
Electric Medical Systems. Milwaukee, Wisconsin, USA.

Hammock, P.D., Freeman, D.E., Baker, GJ. 1998. Failure of psyllium mucilloid to
hasten evacuation of sand from the equine large intestine. Vet Surg. 27. 547-554.

Kaikkonen, R., Raekallio, M., Luukkanen, L., Ruohonierni, M. 2000. Radiographic
appearance of sand accumulations in the large colon and their relationship with age,
breed and clinical signs of the horse. Suomen Elainlaakarilehti. 106 (1). 16-22.

Kaneene, J.B., Miler, R., Ross, W.A., Gallagher, K., Marteniuk, J., Rook, J. 1997. Risk
factors for colic in the Michigan (USA) equine population. Prev Vet Med. 30. 23-36.

Korolainen, R., Ruohonierni, M. 2002. Reliability of ultrasonography compared to
radiography in revealing intestinal sand accumulations in horses. Equine Vet J. 34 (5).
499-504.

Lester, G., Merritt, A., Neuwirth, L., Vetro-Widenhouse, T., Steible, C., Rice, B. 1998.
Effect of org-adrenergic, cholinergic, and nonsteroidal anti-inflammatory drugs on
myoelectrical activity of ileum, cecum, and right ventral colon and on cecal emptying of
radiolabeled markers in clinically normal ponies. Am J Vet Res. 59 (3). 320-327.

Mahaffey M, Barber D. The stomach. In: Thrall D, ed. Textbook of Diagnostic
Veterinary radiology, 4th edition. 2002. W.B. Saunders Company. Philadelphia, USA.
615-638.

Mair T. Differential diagnosis and evaluation of chronic diarrhea in the adult horse. In:
Mair, T., Divers, T., Ducharrne, N., eds. 2002. Manual of Equine Gastroenterology.
W.B. Saunders. London. 427-430.

Merritt, A., Cambell-Thompson, M., Lowrey, S. 1989. Effect of xylazine treatment on
equine proximal gastrointestinal tract myoelectrical activity. Am J Vet Res. 50(6). 945-
949.

122

Merritt A, Burrow J, Hartless C. Effect of xylazine, detomidine, and a combination of
xylazine and butorphanol on equine duodenal motility. 1998. Am J Vet Res. 59(5).
619-623.

Pfeiffer C, MacPherson B. Anatomy of the gastrointestinal tract and peritoneal cavity.
In: White N, ed. 1990. The Equine Acute Abdomen. Lea and Febiger. Philadelphia,
USA. 2-24.

Ragle, C.A., Meagher, D.M., Lacroix, C.A., Honnas, CM. 1989. Surgical treatment of
sand colic — results in 40 horses. Vet Surg. 18 (1). 48-51.

Ragle, C.A., Meagher, D.M., Schrader, J.L., Honnas, OM. 1989. Abdominal auscultation
in the detection of experimentally induced gastrointestinal sand accumulation. J Vet
Intern Med. 3. 12-14.

Reef, V.B. Adult abdominal ultrasonography. 1n: Equine Diagnostic Ultrasound. W.B.
Saunders. Philadelphia. USA. 1998. 273-363.

Roger, T., Ruckebusch, Y. 1987. Colonic az-adrenoreceptor-mediated responses in the
pony. J Vet Pharrnacol Therap. 10. 310-318.

Rose, J.A., Rose, E.M., Sande, RD. 1980. Radiography in the diagnosis of equine
enterolithiasis. Proc Am Assoc Equine Pract. 26. 211-220.

Ruohonierni, M., Kaikkonen, R., Raekallio, M., Luukkanen, L. 2001. Abdominal
radiography in monitoring the resolution of sand accumulations from the large colon of
horses treated medically. Equine Vet J. 33 (1). 59-64.

Rutkowski, J., Ross, M., Cullen, K. 1989. Effects of xylazine and/or butorphanol or
neostigrnine on myoelectric activity of the cecum and right ventral colon in female
ponies. Am J Vet Res. 50(7). 1096-1101.

Snyder J, Spier S. Disorders of the large intestine associated with acute abdominal pain.
In: Smith, B, ed. 1996. Large Animal Internal Medicine, 2"d edition. Mosby-Year
Book, Inc. St. Louis, USA. 765-774.

Snyder J, Spier S. Nonstrangulating and strangulating obstruction of the ascending
colon. In: Robinson, N, ed. 1992. Current Therapy in Equine Medicine, 3rd edition.
W.B. Saunders. Philadelphia, USA. 218-223.

Specht, T.E., Colahan, PT. 1988. Surgical treatment of sand colic in equids: 48 cases
(1978-1985). J Am Vet Med Assoc. 193 (12). 1560-1564.

Stick JA, Chou CC, Derksen FJ, Arden WA. 1987. Effects of xylazine on equine

intestinal vascular resistance, motility, compliance, and oxygen consumption. Am J Vet
Res. 48(2). 198-203.

123

Warncke, D. Soil Testing Lab, Crop and Soil Sciences. Plant and Soil Sciences.
Michigan State University, East Lansing, Michigan, USA. Personal communication.

White, N. Epidemiology and etiology of colic. In: White N, ed. 1990. The Equine
Acute Abdomen. Lea and F ebi ger. Philadelphia, USA. 51-64.

White, NA., Lessard, P. 1986. Risk factors and clinical signs associated with cases of
equine colic. Proc AAEP Ann Conv. 637-644.

Yarbrough, T.B., Langer, D.L., Snyder, J .R., Gardner, I.A., O’Brien, TR. 1994.

Abdominal radiography for diagnosis of enterolithiasis in horses: 141 cases (1990-1992).
J Am Vet Med Assoc. 205 (4). 592-595.

124

APPENDIX A

 

Table 2-2
Tabulated sand parameters. X = no data available.
am §2n_d M M 91—1“.
Group # Horse # I_)_ax Administered Administered Passed Passed Present
(g) Total jg) (g) Total jg) Within
Abdomen (g)

1 1 O 0 0 trace trace 0
1 1 0. 5 2100 2100 trace trace 2100
l 1 l 0 2 100 trace trace 2 100
1 1 1.5 2100 4200 17 17 4183
1 l 2 0 4200 300 317 3883
1 1 2.5 0 4200 88 405 3795
l l 3 0 4200 287 692 3505
1 l 3.5 0 4200 333 1025 3175
l l 4 0 4200 193 1218 2982
1 1 4.5 0 4200 trace 1218 2982
1 1 5 0 4200 trace 1218 2982
l l 5.5 O 4200 trace 1218 2982
1 l 6 0 4200 trace 1218 2982
l 2 0 0 0 trace trace 0
1 2 0.5 2500 2500 trace trace 2500
1 2 1 0 2500 O trace 2500
1 2 1 5 2500 5000 4 4 4996
1 2 2 0 5000 1 13 117 4883
l 2 2.5 0 5000 586 703 4297
l 2 3 0 5000 600 1303 3697
1 2 3.5 0 5000 670 1973 3027
1 2 4 0 5000 150 2123 2877
1 2 4.5 0 5000 40 2163 2837
l 2 5 0 5000 trace 2163 2837
l 2 5.5 0 5000 trace 2163 2837
1 2 6 0 5000 trace 2163 2837

125

Table 2-2 (con’t).

m

--H--——fl#‘_~fl~—~fl--~

 

0
(15

L5

25

215

‘15

55

(15

L5

25

.15

415

55

 

M M. Sal—lid 08.!!!
Horse # gay Administered Administered Passed Passed
(g) Total (g) (g) Total (g)
0 0 trace trace
1900 1900 nace nace
0 1900 0 uace
1900 3800 nace uace
0 3800 trace trace
0 3800 17 17
0 3800 164 181
O 3800 59 240
0 3800 283 523
0 3800 358 881
0 3800 90 971
0 3800 132 1103
0 3800 15 1118
0 0 0 0
2500 2500 0 0
0 2500 "ace nace
2500 5000 nace nace
0 5000 65 65
0 5000 179 244
0 5000 142 386
O 5000 800 1186
0 5000 220 1406
0 5000 239 1645
0 5000 44 1689
0 5000 30 1659
0 5000 "ace 1659

AAA-1:-AAAAAAAAAwwwwwwwwwwwww

126

 

 

Within
Abdomen (g)
0
1900
1900
3800
3800
3783
3619
3560
3277
2919
2829
2697
2682
O
2500
2500
5000
4935
4756
4614
3814
3594
3355
3311
3281
3281

 

Table 2-2 (con’t).

Group # Horse # 2g! Administered Administered Passed

NNNNNNNNNNNNN—tI—Iu—DI—II—I—t—tI—I—II—tu—lI—tu—l

 

O5QOOO‘O‘OO‘O‘C‘O‘ONONUIMUIUtUtU’tUtUtkIIMMUIK/t

O
()5

L5

25

35

‘15

55

(15

L5

25

115

4L5

55

mt! .SJM Sand
jg) Total jg) jg)
0 0 0
1800 1800 0
0 1800 O
1800 3600 3
0 3600 69
O 3600 26
0 3600 21
0 3600 200
0 3600 158
0 3600 45
O 3600 89
O 3600 84
0 3600 36
0 0 0
2300 2300 0
2300 4600 0
2300 6900 2
2300 9200 19
0 9200 2
O 9200 276
O 9200 27
O 9200 186
0 9200 607
0 9200 361
0 9200 665
0 9200 378

127

 

 

Sand
Pasnul

Total jg)

299
326
512
1119
1480
2145
2523

M
[hwsent
VVflmdn
Abdomen jg)
0
1800
1800
3597
3528
3502
3481
3281
3123
3078
2989
2905
2869
0
2300
4600
6898
9179
9177
8901
8874
8688
8081
7720
7055
6677

Table 2-2 (con’t).

 

$9.4. §a_nd. SJ!!! $.20! £1!!!
Group # Horse # M Administered Administered Passed Passed Present
jg) Total jg) jg) Total jg) Within
Abdomen jg)

2 7 O 0 0 trace trace 0
2 7 0.5 2000 2000 trace trace 2000
2 7 1 2000 4000 0 trace 4000
2 7 l .5 2000 6000 trace trace 6000
2 7 2 2000 8000 28 28 7972
2 7 2.5 0 8000 12 40 7960
2 7 3 0 8000 122 162 7838
2 7 3.5 0 8000 15 177 7823
2 7 4 0 8000 27 204 7796
2 7 4.5 0 8000 79 283 7717
2 7 5 0 8000 439 722 7278
2 7 5.5 0 8000 429 1 151 6849
2 7 6 0 8000 51 1202 6798
2 8 0 0 0 0 0 0
2 8 0.5 2200 2200 0 0 2200
2 8 1 2200 4400 trace trace 4400
2 8 l .5 2200 6600 trace trace 6600
2 8 2 2200 8800 trace trace 8800
2 8 2.5 0 8800 trace trace 8800
2 8 3 0 8800 trace trace 8800
2 8 3.5 0 8800 84 84 8716
2 8 4 0 8800 56 140 8660
2 8 4.5 0 8800 88 228 8572
2 8 5 0 8800 26 254 8546
2 8 5.5 0 8800 260 514 8286
2 8 6 0 8800 192 706 8094

128

Table 2-2 (con’t).

Group# Horse# Day

NNNNNNNNNNNNNNNNNNNNNNNNNN

0000000000000

u—nu—nn—nu—au—sn—nu—nn—nbu—nu—nu—nu—n
OOOOOOOOOOOOO

O
()5

L5

2.5

155

4&5

545

(15

L5

225

115

‘15

505

129

 

$104 M1 M
W Ail—milli—SRLCA Passed
jg) Total jg) jg)
0 0 0
2000 2000 0
2000 4000 0
2000 6000 nace
2000 8000 nace
0 8000 nace
0 8000 hace
0 8000 65
0 8000 7
0 8000 92
0 8000 97
0 8000 820
0 8000 1000
0 0 0
1800 1800 0
1800 3600 0
1800 5400 235
1800 7200 220
0 7200 52
0 7200 264
0 7200 50
O 7200 68
0 7200 50
0 7200 200
0 7200 52
IX IX IX

Sand
Pasmui

Total jg)

0

164

1081
2081

235
455
507
771
821
889
939
1139
1191

 

 

 

Ming!
0

2000
4000
6000
8000
8000
8000
7935
7928
7836
7739
6919
5919
0
1800
3600
5165
6745
6693
6429
6379
6311
6261
6061
6009
IX

APPENDIX B
Figure 2-20

Individual and group average graphical representation of sand output over

time.

Horse 1

Passed Sand (g)

 

 

T

0 l 2 3 4
Time (d)

{I
I
4|

Horse 2

Passed Sand (g)

 

 

 

 

Time ((1)

130

Figure 2-20 (con’t).

Horse 3

Passed Sand (g)

 

 

Passed Sand (g)

 

 

 

131

 

 

 

 

 

 

r5
14
)
w m.
r38 r
5 .m w
n T A
M. l
P
H u
0
r.
G
12
.
=1 .7
\.}
t
’n _ .r
o
c
(
m addd—q_qqfi 0 -dqdqq_««
7.. OM00 MWOOOO 00000000M00
M 00 000 mwoowooo 0
m 1987 54321 1 87 54321
We A3 65m @825 A3156 comma.—
0'
F

Time (d)

132

Figure 2-20 (con’t).

1000-

800—
700‘
600‘
500*
400*
300‘
200‘
100*

Passed Sand (g)

Horse 6

 

 

 

Q
0

Passed Sand (g)

 

 

   

 

v W’r“ " ' 1 T 1 1

2 3 4 5 6
Time (d)

Horse 7

 

Time (d)

133

Figure 2-20 (con’t).

1000-
900*
800“
700-
600*

400 4
300 s
200 s
100 j

Passed sand (g)

 

1000 -
900 r
800 s
700 a
600 a

400 ~
300 a
200 ~
100 s

Passed Sand (g)

 

4}

0

41

Horse 8

3 4 5 6
Time (d)

Horse 9

 

v Y T l 1

3 4 5 6
Time (d)

134

Figure 2-20 (con’t).

Horse 10

1000 a
900 ~

Passed Sand (g)

 

 

Group 2 Average

Passed Sand (g)

 

 

 

 

Time (d)

135

Wm
Wm

mé
WV

mam
Wm

ON
ON.

m.—
m._

md
We

oo~~oooooooXomox—ooooooooo

oooooooooooXoooxoooooooooo

oommooooo~~><~~o><~~oo~ooooo

OO00NNv—v—OVMXVOMXWVVMVMNNNN

gas—imamAmqmqmqmqchzqmqm—Jm—Jm
MM

o o o
3.398% 3.326% 3-3985 3-398% 3-398% 3 #73985 3.398%
lull—3 54 as. am 3253.; gum flax—oucficl H.936 n 5525533.. .8339— ?équO 95335 33> g u 8.8: mdmclamw

63282“ 8% o: n X
.88: :08 no» £80883 :03 n8 c30335 53¢ 88% 038%? was 338.35

 

 

 

 

 

m..~ «EFF
U VAR—297:4

136

Table 2-3 (con’t).

Homogeneity Sand Thickness Sand Len th

0 aci

Score j0-2)

Location Accumulation #

Ob ective

Sub ective

View

Group # Horse # M

Score j0—2) Score j0-2) Score j0-2)

Score j0—§) Score j0-12) Score j0-l) Scorej0-3)

0.5

0.5

1.5
1.5

2.5

2.5

137

3.5
3.5

4.5

4.5

5.5

5.5

OOOOOOOOOF‘~~~OOOOOOOOOOOO
ONNOOOOOOOO~~~~OOOOOOOOOO
ONNOO~~OO~~~~~OOOOOOOOO~~
FINN—lv—O—iv-iv—fl—Q—lI—O—O—lv—l—iv—‘v—QI—OI—QI—t—tv—‘I-‘NN
u—OOo—n—nw—uu—nu—nu—n—nv—n—v—‘u—u—nc—nu—ou—nv—u—n—a—n—av—aw—
MM!‘FNNVQ’MMWWO0WV’MMNNNNMMWW
I—IVM—d—Iv—lv—‘u—i—‘u—II—GMN—‘Nv—I—I—flI—lU—II—i—‘I—‘NN
AM—JM—JM—Jnfi—JM—JMAM-Jnfi—JMAM—Jafiu—Jofi—JCZ
MM
MM”)MMMMMMMMMMMMMMMMMMMMMMM
v—‘v—l—n—Iu—nc—a—‘v—Iu—I—O—tu—Iu—n—u-ou—I—n—OU—tu—Ov—tv—iu—GO-O-‘I—

o _ _ _

o o _
3.6. 3 9.89. m 3.8. u o. .85 3-3 9.8. X An-3 9.8. om. 3-3 0.85 3% 2.: 985 $6. 3 «.8. am
|u|5 51— 656 $95.25- .:8m HH-onaddoE 8:. dwnlnloa O a 8:2:8:8< 8:88-— o_>c8_.b_ O.- o. >_.8~_..a..=m.. 33> a a 8.8: a a... 9.0

o

 

 

 

 

 

8.83 E «3.8

138

Table 2-3 (con’t).

Homogeneity Sand Thickness Sand Length

Rel

0

Score j0-2)

Accumulation #

Location

Objective
Score 0-5 Score j0-12) Score (0-1)

View Sub ective

 

Group # Horse # flax

Score j0-2) Score go-g; Score j0-2)

Score j0-3)

0.5

0.5

1.5
1.5

2.5

2.5

139

3.5
3.5

4.5

4.5

5.5
5.5

Table 2-3 (con’t).

Homogeneity Sand Thickness Sand Len th

8C]

0

$2519.11

Accumulation #

Location

Objective
Score j0-§) Score j0-12) Score (0-1)

View Sub'ective

 

Group # Horse # m1

Score j0-2) Score j0-2) Score j0-2)

Score j0-3)

0.5

0.5

1.5
1.5

2.5
2.5

140

3.5
3.5

4.5

4.5

5.5

5.5

Table 2-3 (con’t).

Homogeneity Sand Thickness Sand Length

8C1

O

$942192)

Accumulation #

Location

Objective

Subjective
Score j0-§) Score j0-12) Score j0-l)

 

View

Group # Horse # pg!

Score (0-2) Score j0-2) Score j0-2)

Score j0-3)

0.5

0.5

1.5
1.5

2.5
2.5

141

3.5
3.5

4.5

4.5

5.5
5.5

Table 2-3 (con’t).

Homogeneity Sand Thickness Sand Len th

mm

8C|

0

Score j0-2)

Location Accumulation #

Objective

View Sub ective

 

Group # Horse # 2;!

WM.)

Score jO-S) Score j0-12) Score j0-1) Score j0-3)

0.5

0.5

1.5
1.5

2.5
2.5

142

3.5
3.5

4.5

4.5

5.5

5.5

X X X X X X X X M o m N
o o o o o M M M -M o m N
M M M M M M o N M Wm m N
M M M M M M o N -M Wm m N
X X X X X X X X M m m N
o o o M o M N M -M m m N
X X X X X X X X M We. m N
o o o M M M m N -M m6 m N
X X X X X X X X M v m N
o o M M M M v M -M v m N
X X X X X X X X M Wm m N
o M M M M M m N -M Wm m N
X X X X X X X X M m m N
M M N M M M h m 1M m m N
M M M M M M o N M wN m N
M M M M M M o N -M mN m N
M M M M M M o m M N x N
M o M M M M m M -M N m N
o o M M M M v N M WM m N
o M M M M M m m -M WM m N
o o M M o M m N M M m N
o o M M o M m M M M m N
o o N N M o m N M We m N
o o N N M o m N \M md m N
o o o o o o o o M o m N
o o o o -M o x N

o o o c
3.3 0.85 MN.-3 28m Mn-3 98%. MN. -3 28w Mm.-3 0.89m. 29...... «Mo-low Mum—f3 989m. Mm_-_=|M.o....8m
sauna-M M83 8223.; 28w awe—8883M flfiamc u 8M8M=E=uo< 8388M 0380....5. o>M8~M...M=m. 32> Hala- u 8.8: n 98.5

 

 

 

 

 

8.53 3 2...:-

143

Table 2-3 (con’t).

Homogeneity Sand Thickness Sand Length

0 aci

Score j0-2)

Location Accumulation #

Objective

View Sub'ective

 

Group # Horse # 231

Score j0-2) Score j0-2) Score j0—2)

Score j0-3)

Score (0-5) Score (0-12) Score 0-1

 

0.5

0.5

1.5
1.5

2.5

(‘3

144

3.5
3.5

4.5

4.5

5.5
5.5

x x x x x x x 22 ,22 2 222 N
x x 22 22 x x x x 2 2 222 N
22 22 N N N 2 22 2 m 2 222 N
22 22 2 2 2 2 v 22 2 3 222 N
22 22 N N N 2 N N a N 222 N
22 22 N N N 2 22 2 2 N 222 N
22 22 N N N 2 22 2 m 3 222 N
22 22 N N N 2 22 2 2 Ne 22. N
22 22 _ N N 2 2 2 N ”2 v 222 N
22 22 N N 2 2 2 2 2 v 222 N
22 22 2 2 2 2 v 2 m NN 222 N
22 22 2 2 2 2 v 2 2 N.N 222 N
22 22 2 2 2 22 N 2 m N 222 N
22 22 2 N 2 22 v 22 2 N 222 N
22 22 2 2 2 2 22 2 m NN 222 N
22 22 2 2 2 2 v N 2 NN 222 N
22 22 22 22 22 22 22 22 22 N 222 N
22 22 22 2 2 2 N 22 2 N 222 N
22 2 2 2 2 22 N 2 2 N2 222 N
22 2 2 2 2 22 v 22 2 N2 222 N
x 22 x x x x x 22 m 2 222 N
22 22 22 22 22 22 22 22 2 2 222 N
22 2 N N 2 22 22 N 2 N22 222 N
22 2 N N 2 22 c N 2 N22 222 N
22 22 22 22 22 22 22 22 m 22 222 N
22 22 22 22 22 22 22 22 2 22 222 N

 

MN-3989m MN__-3988w 3.398% 3.3985 Mm-3989m M73 «.85 373985 3.3 985
lung. .84.. madm- 83335. M255 flEEJc-E. 22:... Hula-amO n 28332283322 2832.825 2.134% a 33> a 22232.8: 22298.5

 

 

 

 

 

6.288 MN 323,—-

145

APPENDIX D

Figure 2-21

Individual horse and group average objective score trends over time.

Horse 1

122

10*

Objective Score

 

 

 

Time (d)
Horse 2
12 a
10
Q)
‘5 8 “
U
m
0
.3
E
:3,
O

 

 

Time (d)

146

Figure 2-21 (con’t).

Horse 3
12 1

10 “

Objective Score

 

 

 

Time (d)

Horse 4

12‘

10

 

Objective Score
a-

&

N

 

147

Figure 2-21 (con’t).

Objective Score

Objective Score

1

12

 

Horse 5

 

12*

10*

 

I

2 3
Time (d)

Group 1 Average

 

 

 

148

Figure 2-21 (con’t).

1

12

10“

Objective Score

 

 

Objective Score

 

 

Horse 6
2 3 4 5 6
Time (d)
Horse 7

 

 

 

 

149

Figure 2-21 (con’t).

Horse 8
12 2

10%

Objective Score

 

Time ((1)

Horse 9

Objective Score

 

 

 

 

09 ——--
0 1 2 3 4 5 6

Time (d)

150

Figure 2-21 (con’t).
Horse 10

12a

10“

Objective Score

 

 

o v T r
o 1 2 3

Time (d)

Group 2 Average

12-

10 '*

Objective Score
a

 

 

 

 

Time (d)

151

 

 

APPENDIX E

Figure 2-22

Individual horse objective scores based on weight of intra-abdomina] sand.

Horse 1

1

12

1

10

Objective Score

 

0 7 I j 'LI I I I I
0 1000 2000 3000 4000 5000 6000 7000 8000
Sand Present (g)

Horse 2

12 ~
R2 = 0.086

10*

Objective Score

 

R2 = 0.1739

1

9000 10000

 

 

 

' "—_"" ’ "' fi‘ " " "m" I "“" j T

Sand Present (g)

152

T 1

0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000

Figure 2-22 (con’t).

 

 

 

 

Horse3
12 2
R2 =0.4449
10 a
‘s’ 8 *
5 -
é 6*
i a n
8 4 ~ -
II I I I
2 2 I I
0 I T I I I I T I I I
0 1000 2000 3000 4000 S000 6000 7000 8000 9000 10000
Sand Present (g)
Horse4
12 ~ 2
R =0.l894
10 2
E 8*
U
m
'3 6 2 I I I
8 .
8 4*
I
2 .2 I I
I I
0 I I I = I I I T I I I
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Sand Present (g)

153

Figure 2-22 (con’t).

 

 

 

 

Horse 5
12 -
R2 =0.0056
10 a
g 8 2
O
In
E a
8 4 4 I
I
2 2 I
0 I = T * I T T I T I 1
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Sand Present (g)
Horse6
12 2
R2 =03372
l0 2
g s— .
J’- .
.2 6 ~ ..
g .
s 4 -«
O O O
2 ~ 0
0 I I T‘ 2 I I " I 3 I I I 1
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Sand Present (g)

154

Figure 2-22 (con’t).

12*

10*

Objective Score

 

Horse 7

R2 = 0.2601

 

0 -
0 1000

llj
10“

8'1

Objective Score
a

r

2000

1

3000

I l I I I 7 1

4000 5000 6000 7000 8000 9000 10000
Sand Present (g)

Horse 8

R2 = 0.0889

 

 

0 1000

I

2000

3000

r 1 I I I 1

4000 5000 6000 7000 8000 9000 10000
Sand Present (g)

155

Figure 2-22 (con’t).

12*

10*

Objective Score

 

Horse 9
R2 = 0.1229

 

0 I f I

0 1000 2000 3000

12*

10*

Objective Score

 

0 I --——M—H—NNN—~—T -

3000

0 1000 2000

T I T 7 I I 1

4000 5000 6000 7000 8000 9000 10000
Sand Present (g)

Horse 10

R2 = 0.175

 

 

""“fl‘” "T "”" "—'—' I—___1

8000 9000 10000

"I I T T

4000 5000 6000 7000
Sand Present (g)

156

APPENDIX F

Figure 2—23

Combined Group 1 and Group 2 objective scores based on weight of intra-abdominal

sand on a given day.

Day 0.5

12 1
R2 = 0.319

10*

Objective Score

 

 

 

0 ‘*" I" I "'I— I I ”' I I I I Fl

0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Sand Present (g)

Day 1

12]

10 I R2 = 0.326

Objective Score
a-

O

 

 

 

0 i I H22,_2,2., - ——- --—w .._.M I‘—“'" - -

0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Sand Present (g)

157

Figure 2-23 (con’t).

 

 

 

 

 

Day 1.5
12 ~ R2 =0.l674
10 4
0
s 8 **
O
V)
g 6 4
8 O O
8 4- e o
O O
2 2
O
0 I I I I 3 I I I I 1
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Sand Present (g)
Day 2
12 _ R2=0.5673
10 *
g s I .
U
m
.2 6 1
8
3.
s 4 «
2 .
0'11 .' ‘ ' I ‘ 'I" -2_22.2r__ I I " I I I I
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Sand Present (g)

158

Figure 2-23 (con’t).

12*

10*

Objective Score
:-

 

Day 2.5

R2 = 0.4615

 

 

 

 

 

 

 

4 2
2 * o
0 “I I I If I I I I I I I
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Sand Present (g)
Day 3
'2 I R2 = 0.2569
10 j
a s _-
O
m
9 6 ..
'5
.32.
8 4 .
2 N o e
0 TH“ '___ "‘"I'" I I "‘"I" "__"’ ' "“‘" " I I I I
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Sand Present (g)

159

Figure 2-23 (con’t).

Objective Score

Objective Score

12*

10*

 

12*

10*

 

1000

I

1000

Day 3.5

R2 = 0.3088

 

I f I f r I I I I

2000 3000 4000 5000 6000 7000 8000 9000 10000
Sand Present (g)

Day 4

R2 = 0.3773

 

 

 

I . I.“_‘—I—~"" "1'” T"‘—‘— I I I I

2000 3000 4000 5000 6000 7000 8000 9000 10000
Sand Present (g)

160

Figure 2-23 (con’t).

12*

10*

Objective Score

 

AA
'1'

3000

I I

2000

r

0 1000

4000

Day 4.5

R2 = 0.3424

I

10000

I

9000

I

8000

I

7000

I

6000

I

5000

Sand Present (g)

12 w

10 '1

Objective Score

 

* "“_T""" —— " I———’*

0 1000 2000 3000

_r._.

4000

Day 5

R2 = 0.2414

 

I

9000

I

10000

 

I I

8000

I

7000

_,_T_ . ._ I

5000 6000
Sand Present (g)

161

Figure 2-23 (con’t).

Day 5.5

12*

10*

Objective Score

R2 = 0.671

 

 

 

0 I I :3, I I I
0 1000 2000 3000 4000 5000 6000
Sand Present (g)

Day 6

12*

104

Objective Score

7000

I I I

8000 9000

R2 = 0.6621

0

 

 

 

2000 3000 4000 5000 6000 7000
Sand Present (g)

0 1000

162

I I I

8000 9000 10000

APPENDIX G
Figure 2-24
Individual group objective scores based on weight of intra-abdominal sand on a given

day.

Group 1, Day 0.5
12 ~ R2 = 0.3634

10*

.. /-

Objective Score

 

 

0“ I I I I I I“ I I I —I

0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Sand Present (g)

Group 2, Day 0.5
12 *

1 _
10 1 R —0.2332

6 4 00
/.

Objective Score

 

 

_‘W

 

0 2+2_. ' I " ’ "fi.’_ "" ’i—f—‘_ “ Ti—""—" I '—' T f ' ' "T

0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Sand Present (g)

163

Figure 2-24 (con’t).

 

Group 1, Day 1

 

 

 

 

‘2 I R2 =0.09
10 *
E 8*
O
m
.2 6 N
E
8 44
2 N I I
a”
0 I = I g I I I I I I I I
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Sand Present (g)
Group 2, Day]
12 7
1 R’=0.119
10
“'5’ 8 *
O
m
0
.3. 6 *
8 e
8 4 -
/fl
2 2
0 “— ’T' ““"‘"T""""' 'I' '4' I I“' I I I I 1
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000

Sand Present (g)

164

Figure 2-24 (con’t).
Group 1, Day 1.5

 

 

 

12 *
R2 =0.2483
10 *
0
"6 3 ‘
O
m
E.
8 4 -— -
I I
2 .4 /
I
0 I I I I '4 I I I I I
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Sand Present (g)
Group 2, Day 1.5
12 ‘ R2 = 0.1486
10 *

Objective Score
a

O I

4‘4 /
O
I

 

""T‘" 7 I 1

 

 

0 _1)____ _ I "_"”-'"' "I '" " ' ""— 7 I I

0 1000 2000 3000 4000 S000 6000 7000 8000 9000 10000
Sand Present (g)

165

 

Figure 2-24 (con’t).
Group 1, Day 2
12 0

10 « R2 = 0.1554

Objective Score

 

 

-
0 T I T '1 I I I T I 1

0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Sand Present (g)

Group 2, Day 2

12 — R2 = 0.4418

10“

Objectlve Score
a
\

 

 

 

4 a o
O
2 4
o+————-—-——I— I --I . I I I I I I
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Sand Present (g)

166

Figure 2-24 (con’t).

 

 

 

 

 

Group 1, Day 2.5
12 — R2 =0.2619
10 4
E 8 -
O
m
3 6 -
'4:
8 I I
E
o 4 *
/
2 — I
0 T I I I I I I T I I
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
0 Sand Present (g)
Group 2, Day 2.5
‘2 I R2 =0.5466
l0 4
E 8 ~
fi .
“a:
'4: 6 " °
53. /
8 4 o o
2
0 I I _.._ I I I T I I I I
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Sand Present (g)

167

Figure 2-24 (con’t).

12—

10*

Objective Score

 

Group 1, Day 3
R2 = 0.2176

 

12-

10*

Objective Score

I I I I I I T r I I

1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Sand Present (g)

Group 2, Day 3

R2 = 0.5002

 

 

 

I I

10000

I I j -_' I I I I I

1000 2000 3000 4000 5000 6000 7000 8000 9000
Sand Present (g)

168

Figure 2-24 (con’t).

Objective Score

Objective Score

12*

10“

 

Group 1, Day 3.5

R2 = 0.2244

/

 

12-

10 '*

-
I I W— I I t I I I I

1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Sand Present (g)

Group 2, Day 3.5

R2 = 0.5088

/

'I‘ ___ I I I

 

 

1. ._

 

 

 

T T" " .._, I'—_ WI—

1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Sand Present (g)

169

Figure 2-24 (con’t).

Objective Score

Objective Score

122

10 '2

Group 1, Day 4

R2 = 0.09

2’

 

 

 

- -
I I 'I' r I I I I I 7

1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Sand Present (g)

Group 2, Day 4

R2 = 0.7449

/

 

1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Sand Present (g)

170

Figure 2-24 (con’t).

Objective Score

Objective Score

12*

107

 

12

1

1

10

 

Group 1, Day 4.5

R2 = 0.0151

I I
/

--
I I ‘I- I I I I I I I

1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Sand Present (g)

Group 2, Day 4.5

R2 = 0.4793

 

I I I I I I I I I I

1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Sand Present (g)

171

Figure 2-24 (con’t).

Objective Score

Objective Score

 

Group 1, Day 5

 

 

 

 

12 1
10 7
R2 = 0.0055

8 _

6 a

4 ~ I

2 q I

/
I
0 I I =9! I I I I I I I
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Sand Present (g)
Group 2, Day 5

12 ‘ R2 =0.3974

10 J

8 4 o

6 4 o o

4 p /

2 ~ 0 o

0 “—“_" I" "*‘"" ' “T """"' I "T "' I I ' “‘"" I I 1

 

0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000

Sand Present (g)

172

Figure 2-24 (con’t).

 

 

 

 

 

Group 1, Day 5.5
12]
R2 =0.0409

10 4
§ 8 J
U
In
3 6—
U
.2.
8 4 ~

I
2 7 I
——’1
0 I I flI I I I I I f n
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Sand Present (g)
Group 2, Day 5.5
12 —
R2 = 0.3579

10 7
0
§ 8 "
In
3 6 — o
3
E o
8 4 I /

0
2J
0 I _I" I I I I I I I I
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Sand Present (g)

173

Figure 2-24 (con’t).

Objective Score

Objective Score

12*

107

 

12-

107

Group 1, Day 6

R2 = 0.2215

I I I V I I I I I I I

1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Sand Present (g)

Group 2, Day 6

R2 = 0.1296

 

I " I I “m 'T—“"’ I I 9 I I I I

 

 

1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Sand Present (g)

I74

APPENDIX H
Figure 2-25

Individual objective parameter scores based on weight of intra-abdominal sand.

Combined Group Sand Location Score per Sand Weight

 

 

 

11D .. O-..” O...“ 0.-.... “-0.0”.
g R2 =0.2641
8
U)
E
O
'5
a
O
O
u-I
'6
E
a
CI)

yo—~I~++-+T
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000

Sand Weight (g)

Combined Group Number of Sand Accumulations Score per Sand Weight

 
       
   

 

u 3 I o .
In
8
‘2 2.5 .
I:
e
E
E 2 I
S
:3
3 15 —< 2
4g - R =0.ll46
'8
a 14’ 9000000000»-.. ..... m 0
"5
:3
.1: 0.5 7
E
2
0 I W ow———Q— --¢————«oT-v———o——To—~—+—m—I—————I

0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Sand Weight (g)

175

Figure 2-25 (con’t).

Combined Group Sand Opacity Score per Sand Weight

        
  

 

2— mo... 9 coo oo oo o
1.87
1.6 ~ R2 = 0.1855
0
‘5 1,4.
5
£1.2“
g 14) oo cocoon-coo «on». 00.... “co...
a.
O 0.8w
'8
I 0.6J
In
0J3“
OJ!“
0‘ '9 I °°' " ' '

I"
Q N
1 A

fl —
I I
N

1

Sand Homogeneity Score

 

 

0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000

Sand Weight (g)

Combined Group Sand Homogeneity Score per Sand Weight

....... . .0 . . .
2

R=0.0464

....-O....... O O... . O“.

 

-————T— W-Q———— o, ——~p———uqo—o—— -I--—————--1
1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Sand Weight (g)

176

Figure 2-25 (con’t).
Combined Group Sand Thickness Score per Sand Weight

2.. g C. .

.... .0 . .0. 0000

Sand Thickness Score
8
.
.
0
.

 

 

0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Sand Weight (g)

Combined Group Sand Length Score per Sand Weight

] ~ 0. o o I o .0 0 on o
0.9 I
0-8 7 1?.2 = 0.2436

"I'" “I

0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Sand Weight (g)

 

 

I77

APPENDIX 1
Figure 2-26

Scatter plots comparing scores of each objective parameter to another.

Sand Location versus Number of Accumulations

R’ = 0.1136

Number of Accumulations Score
H
m

 

 

0 I r I I I
0 0.2 0.4 0.6 0.8

Sand Location Score

Sand Location versus Sand Opacity

2 "I R’=0.0991
1.8 ~
1.6 I
1.4 4
1.2 4

Sand Opacity Score
a c
54 '00 n-t
i 1 F
I

0.4 7

P
N
1

- _____.__ _.

 

G

1.. ___..._ . "’I—TI" , - _.___..-—rfi_... . ___..____ ‘T—" --_

0.2 0.4 0.6 0.8 1

Sand Location Score

9

I78

Figure 2-26 (con’t).

Sand Location versus Sand Homogeneity

.__-

1?.2 = 0.0025

F???
NbOfiflN
1111.1.

 

Sand Homogeneity Score
a

 

9???
GIG-BIN”
ii 1

 

I I I I

0.2 0.4 0.6 0.8

Sand Location Score

G

Sand Location versus Sand Thickness

2 1!
1.8 I R2=0.02
1.6 7
1.4
1.2 1

0.8 a
0.6 7
0.4 ~

Sand Thickness Score
a

H

 

0.2 7

 

 

 

 

 

0 I "_— I I _. I “I'—
0 0.2 0.4 0.6 0.8

Sand Location Score

179

Figure 2-26 (con’t).

Sand Location versus Sand Length

1.8 7 R2 = 0.1236
1.6 ~

1.4 7
1.2 7

0.8 7
0.6 7
0.4 7
0.2 7

Sand Length Score

 

 

 

0 I I I I
0 0.2 0.4 0.6 0.8

Sand Location Score

Number of Accumulations versus Sand Opacity

2 1' -
1.8 7
1.6 7
1.4 7
1.2 7

R2 = 0.4219

0.8 7
0.6 7
0.4 7
0.2 7

Sand Opacity Score
a
I

 

 

 

 

0 1" ” I I fl “'_I—"—‘ I
0 0.5 l 1.5 2 2.5

Number of Accumulations Score

180

Figure 2-26 (con’t).

1.8 7
1.6 7
1.4 7
1.2 7

1 a
0.8 7
0.6 7
0.4 7
0.2 7

0 9—

Sand Homogeneity Score

 

1.8 7
1.6 7
1.4 7
1.2 7

0.8 7
0.6 7
0.4 7
0.2 7

Sand Thickness Score
fit

Number of Accumulations versus Sand Homogeneity

R2 = 0.3805

J

I I T I I
0.5 l 1.5 2 2.5

Number of Accumulations Score

(a)

Number of Accumulations versus Sand Thickness

R2 = 0.1362

I

 

 

 

I +"' I I I
0.5 l 1.5 2 2.5

Number of Accumulations Score

b)

181

Figure 2-26 (con’t).

Number of Accumulations versus Sand Length

1.8 7
1.6 7
1.4 7
1.2 7

l 7} 0
0.8 7
0.6 7
0.4 7
0.2 7

R2 = 0.0875

Sand Length Score

 

 

0 I I I I r
0 0.5 l 1.5 2 2.5 3

Number of Accumulations Score

4)

Sand Opacity versus Sand Homogeneity

1-6 7 R2 = 0.5009

Sand Homogeneity Score

 

 

 

Sand Opacity Score

182

Figure 2-26 (con’t).

Sand Opacity versus Sand Thickness

2 7 I
1.8 7
1.6 7 R2 =0.0948
1.4 7
1.2 7

1 7 I
0.8 7
0.6 7
0.4 7
0.2 7

Sand Thickness Score

 

 

 

0 * T I I
0 0.5 l 1.5

Sand Opacity Score

Sand Opacity versus Sand Length

1.8 7
1.6 7
1.4 7
1.2 7

R2 = 0.0282

0.8 7
0.6 7
0.4 7

Sand Length Score
I

 

0.2 7

 

 

 

0 ¥———-—- I - -—— I. — - 7

Sand Opacity Score

183

Figure 2-26 (con’t).

Sand Homogeneity versus Sand Length

1.8 7
1.6 7
1.4 7
1.2 7

R2 = 0.019

0.8 7
0.6 7
0.4 7
0.2 7

Sand Length Score
a
I

 

 

0 *fi‘ " I I? I
0 0.5 l 1.5

Sand Homogeneity Score

Sand Homogeneity versus Sand Thickness

1.8 7
1.6 7
1.4 7
1.2 7

R2 = 0.1782

0.8 7
0.6 7
0.4 7
0.2 7

Sand Thickness Score
I

 

 

 

 

o 41—————— ,-

Sand Homogeneity Score

184

Figure 2-26 (con’t).

Sand Thickness versus Sand Length

1.8 7
1.6 7 R2 =0.2538
1.4 7
1.2 7

1 £ I I
0.8 7
0.6 7
0.4 7
0.2 7

0 ‘1'—“"' I ’ ' * I I

Sand Length Score

 

 

Sand Thickness Score

185

 

I"IIIIIIIIIIIIII‘LIIIIII