Sie sind auf Seite 1von 10

ULTRASONOGRAPHIC EVALUATION OF ADRENAL GLAND SIZE

COMPARED TO BODY WEIGHT IN NORMAL DOGS

STACY N. SOULSBY, MERRILEE HOLLAND, JUDITH A. HUDSON, ELLEN N. BEHREND

The accepted cut-off value for adrenal gland maximum diameter of 0.74 cm to distinguish adrenal gland
enlargement in dogs regardless of body weight may not be appropriate for small to medium breed dogs. The
purpose of the current retrospective study was to examine adrenal gland dimensions as a function of body weight
in healthy dogs in three weight categories (< 10 kg, 10–30 kg, and > 30 kg) representing small, medium, and
large breeds, respectively, to establish greater confidence in determining if adrenal gland size is abnormal. The
measurements of length (sagittal plane), cranial and caudal pole thickness (sagittal and transverse planes), and
caudal pole width (transverse plane) of both adrenal glands were obtained ultrasonographically in clinically
healthy dogs (n = 45) with 15 dogs in each weight group. Findings support our hypothesis that adrenal gland size
correlates with body weight in normal dogs, and more precise reference intervals should be created for adrenal
gland size by categorizing dogs as small, medium, or large breed. The caudal pole thickness of either adrenal
gland in a sagittal plane was the best dimension for evaluating adrenal gland size based on low variability, ease,
and reliability in measurement.  C 2014 American College of Veterinary Radiology.

Key words: adrenal gland, body weight, dog, ultrasound.

Introduction have been reported to vary from 0.91 to 5.02 cm for


length,2–15 from 0.19 to 1.74 cm for width,2–12 and from 0.17
A PRELIMINARY DIAGNOSIS OF hyperadrenocorticism
is made based on the history, physical examination
findings, and analysis of routine blood tests and urinaly-
to 1.07 cm for thickness/height.3–6,8,9,13–18 Additionally, al-
though dogs with hyperadrenocorticism have adrenal gland
measurements significantly greater than those of normal
sis, while specific endocrine tests are used to confirm the
dogs4,5,10,11,15–17 and dogs with hypoadrenocorticism have
diagnosis. Although no endocrine test used to screen for
adrenal gland measurements significantly less than those of
hyperadrenocorticism is perfectly accurate, the sensitivi-
normal dogs,13,14 an overlap in the range of measurements
ties and specificities are quite high.1 Thus, ultrasonog-
of dogs with either disease and normal dogs is present.
raphy is commonly used as a means of supporting a
The overlap may be a result of a body weight-associated
diagnosis of hyperadrenocorticism and in aiding in
variation in measurements of normal adrenal glands.
differentiation of pituitary-dependent hyperadrenocorti-
Whether or not a relationship exists between ultra-
cism (PDH) from adrenal-dependent hyperadrenocorti-
sonographically obtained adrenal gland measurements and
cism (ADH). Ultrasonography is not flawless, as dogs
body weight has not been clearly documented in pre-
with hyperadrenocorticism may have adrenal glands which
vious reports. Discrepancies exist among earlier stud-
are symmetrically or asymmetrically enlarged or normal-
ies that attempted to document an association between
sized.1 However, it is the preferred imaging modality in
adrenal gland measurements and body weight using
evaluating adrenal glands in veterinary medicine.
ultrasonography,3–5,7,8,11,12,16,17 but Bertolini et al. found no
A wide range of values have been documented by previ-
relationship between the volume of the adrenal glands and
ous studies with regard to normal ultrasonographic size
body weight using computed tomography.19
(length, width, and thickness) of the adrenal glands in
The terminology used in prior studies to describe the
healthy dogs. Adrenal gland measurements in normal dogs
maximum diameter of the caudal pole of the adrenal gland
as measured in a sagittal plane ultrasonographically can
From the Department of Clinical Sciences, Auburn University Col-
lege of Veterinary Medicine, Auburn, AL, 36849-5540. be confusing since previous researchers have chosen to
Address correspondence and reprint requests to Stacy N. Soulsby, define this dimension as either height,3 width,5,7,11,12 or
at Diagnostic Imaging, P.C., 22020 E. Arbor Drive, Aurora, CO 80016. thickness.8,16,17 Table 1 summarizes earlier studies which
E-mail: snsoulsby@gmail.com.
have used ultrasonography to evaluate adrenal gland size
Received January 23, 2014; accepted for publication November 4,
2014. and shows what each study labeled as the maximum
doi: 10.1111/vru.12236 dorsoventral dimension of the caudal pole of the adrenal
Vet Radiol Ultrasound, Vol. 00, No. 0, 2014, pp 1–10. gland in a sagittal plane, if a correlation with body weight

1
2 SOULSBY ET AL. 2014

TABLE 1. Summary of Findings from Previous Ultrasonographic Studies Examining the Maximum Dorsoventral Thickness of the Caudal Pole of the
Adrenal Gland in a Sagittal Plane and its Correlation with Body Weight in Healthy Dogs

Name Given to Maximum Suggested Upper Threshold for Maximum Correlation Found with
Dorsoventral Dimension of the Dorsoventral Measurement of the Maximum Dorsoventral
Previous Caudal Pole of the Adrenal Caudal Pole of the Adrenal Diameter and Body
Study Gland in a Sagittal Plane Gland in a Sagittal Plane (cm) Weight
de Chalus et al.3 Height Left – 0.54 Right – 0.67 Yes
(23 dogs 1.9 – 4.1 kg) (23 dogs 1.9 – 4.1 kg)
Left – 0.79 Right – 0.95
(17 dogs 26.8 – 38.6 kg) (17 dogs 26.8 – 38.6 kg)

Barthez et al.5 Maximum Diameter / Left – 0.74∗ Right – 0.81∗ No (Correlation with
Width ∗ (20 dogs 4.4 – 38.8 kg) (20 dogs 4.4 – 38.8 kg) Length and Body
Weight)
Douglass et al.7 Width No upper threshold reported No (Correlation with
Length and Body
Weight)

Grooters et al.8 Thickness Left – 0.50 Right – 0.50 Yes


(14 dogs 13.6 – 20.3 kg) (14 dogs 13.6 – 20.3 kg)

Choi et al.11 Width Left – 0.60 Right – 0.60 Yes


(175 dogs 0.84 – 10.0 kg) (80 dogs 0.84 – 10.0 kg)

Mogicato et al.12 Width No upper threshold reported No (Correlation with


Length and Body
Weight)

Grooters et al.16 Thickness Left – 0.70 Right – 0.70 Yes


(10 dogs 9.1 – 38.6 kg) (10 dogs 9.1 – 38.6 kg)

De Marco et al.17 Thickness Left or Right – 0.59 (109 dogs ࣘ 10 kg) Yes


Measured in a transverse plane.
cm, centimeters; kg, kilograms.

was identified with this dimension, and if body weight from adrenal hyperplasia due to PDH. A third study by
played a role. To minimize readers’ misunderstanding of de Chalus et al. found a correlation between thickness of
the terminology used in this and previous manuscripts, the caudal pole of the adrenal gland in a sagittal plane and
the authors have decided to define “length” as the max- breed. They suggested an upper threshold of 0.54 cm and
imum craniocaudal dimension measured in the sagittal 0.67 cm for the thickness of left and right adrenal glands,
plane, “thickness” as the maximum dorsoventral dimension respectively, in Yorkshire terriers and 0.79 cm and 0.95 cm,
measured in the sagittal or transverse plane, and “width” respectively, in Labrador retrievers.3 These proposed cut-
as the maximum mediolateral dimension measured in the offs contradict the most accepted parameter for ultrasono-
transverse plane. To maintain consistency, the authors of graphic assessment of the size of canine adrenal glands. The
this manuscript have chosen to use the terms as described most commonly used parameter provides a single cut-off
above throughout the remainder of this manuscript re- value of 0.74 cm for normal adrenal gland thickness of the
gardless of how that term was used in the original pa- larger of the cranial or the caudal pole in either a sagit-
per. Table 1 shows three studies found a correlation only tal or transverse plane regardless of body weight for the
between adrenal gland length and body weight.5,7,12 How- diagnosis of PDH (sensitivity of 77% and a specificity of
ever, two other studies suggested an association may ex- 80%).5,6
ist between animal size and adrenal gland thickness, as Grooters et al. determined that ultrasonographically
adrenal gland thickness was significantly different between measured adrenal gland thickness, but not adrenal gland
normal medium and large-sized dogs.8,16 More recent stud- length or width, accurately represents gross measurements8
ies have detected a relationship between body weight and while independently, Barberet et al. noted that measure-
the thickness of the caudal pole of the adrenal gland in ments of adrenal gland thickness at the caudal pole
a sagittal plane when small breed dogs (< 10 kg) were in a sagittal plane have the least intraobserver and in-
examined.11,17 Cut-off values of 0.60 cm11 and 0.59 cm17 terobserver variability.9 Therefore, since dogs have great
were proposed for distinguishing normal adrenal glands variability in body weight and previous studies have not
VOL. 00, NO. 0 US EVALUATION OF ADRENAL GLAND SIZE IN DOGS 3

consistently identified a single best measurement for dis- System (Philips Medical Systems, Eindhoven, The Nether-
tinguishing normal from enlarged adrenal glands, further lands). Dogs were manually restrained in lateral recum-
research is needed to determine if a relationship exists be- bency, hair was clipped from the abdomen dorsal to the
tween adrenal gland size (length, width, and thickness) in level of the transverse processes of the lumbar spine and
small, medium, and large breed dogs. cranial to the level of the last few intercostal spaces, a
To the investigators’ knowledge, no studies have been liberal amount of coupling gel was applied to the shaved
performed in which dogs have been divided into multiple portion of the abdomen, and dogs and were scanned us-
groups based on their body weight for ultrasonographic ing a previously described ultrasound protocol6,20 to ex-
evaluation of adrenal gland dimensions in an effort to pro- amine the entire abdomen with particular emphasis being
vide reference ranges for normal adrenal gland size in vari- placed on the adrenal glands. The initial abdominal ul-
ous weight categories. Therefore, the aim of this study was trasound examination and measurements was performed
to use ultrasonography in healthy dogs to determine if a by a second-year radiology resident (S.N.S) under the su-
relationship exists between adrenal gland dimensions and pervision of a board-certified radiologist (M.H.). The ab-
body weight. We hypothesized there would be a difference dominal ultrasound examination and measurements were
in adrenal gland measurements based on body weight. The repeated by another board-certified radiologist (J.A.H.)
ultimate goal of this study was to establish more precise later in the day with each observer unaware of the other’s
reference intervals for adrenal size in three different weight measurements.
categories of dogs. The adrenal glands were imaged in two standard planes,
sagittal and transverse, preferentially positioning the
transducer caudal to the ribs and scanning subcostal rather
than intercostal. The sagittal plane was defined as the long
Materials and Methods
axis of the adrenal gland, which approximated the sagittal
This prospective study was conducted in accordance plane of the body, in which the adrenal gland length
with requirements of the Auburn University Institutional and thickness were both maximal. From this plane, the
Animal Care and Use Committee. Forty-six dogs belong- maximal length (craniocaudal dimension) of the adrenal
ing to students, staff, and faculty of the Auburn University gland and the maximal thickness of the cranial and caudal
College of Veterinary Medicine were recruited for partic- poles (dorsoventral dimension) of the adrenal gland were
ipation. Inclusion criteria were as follows: (1) age between measured with electronic calipers (Figs. 1 and 2). The
1 and 6 years; (2) body condition score of 2 to 6 out of 9; transverse plane was defined as the short axis of the adrenal
and (3) a negative occult heartworm test within the past gland obtained by rotating the probe 90o from the sagittal
year and heartworm prevention administration for at least plane at the caudal pole of the adrenal gland. From the
6 consecutive months prior to the start of the study. All transverse plane, the maximal thickness of the caudal pole
dogs were fasted for 12 h prior to participation. A physical (dorsoventral dimension) and the maximal width of the
examination, complete blood count, serum biochemistry caudal pole (mediolateral dimension) of the adrenal gland
panel, and urinalysis obtained via cystocentesis were per- were measured with electronic calipers. Each measurement
formed on each dog to establish apparent health. During was obtained three times by both sonographers (S.N.S. and
the physical examination, each dog was weighed and the J.A.H). The process was repeated for the opposite adrenal
body condition score was assessed. Exclusion criteria were gland.
as follows: (1) the presence of clinical signs, physical ex- Statistical tests were selected and performed by a board-
amination findings, or laboratory abnormalities consistent certified veterinarian in preventative medicine and profes-
with hyperadrenocorticism (e.g. polyphagia, polydipsia, sor of pathobiology at the Auburn University College of
polyuria, pot-bellied appearance, hepatomegaly, alopecia, Veterinary Medicine who has teaching and research inter-
pyoderma, or skin hyperpigmentation) or hypoadrenocor- ests in epidemiology (J.C.W.). Intraobserver, interobserver,
ticism (e.g. anorexia, vomiting, diarrhea, or weight loss); and side (left and right adrenal gland) variability were
(2) the presence of chronic medical conditions (e.g. adrenal, evaluated as random factors in a mixed model, analysis of
renal/urinary, liver, heartworm, respiratory, or cardiac variance (ANOVA) to assess repeatability and reproducibil-
disease, diabetes mellitus, or hypothyroidism); or (3) ad- ity of the measurements. The reproducibility was further
ministration of any medications (including topical or otic evaluated by calculating a coefficient of variation for
preparations containing corticosteroids, but excluding each of the adrenal gland measurements. Ultimately, each
monthly flea/tick and heartworm prevention) during the measurement between the two observers were pooled and
2 months prior to the study. averaged because there were no intraobserver differences
All ultrasonographic examinations were performed us- or side (left and right adrenal gland) differences when the
ing either an 8–5 MHz or a 5–2 MHz electric curved ar- statistics were analyzed for each observer independently of
ray transducer and a Philips iE33 xMATRIX Ultrasound the other observer. In addition, the minimal interobserver
4 SOULSBY ET AL. 2014

FIG. 1. Ultrasound images illustrating measuring the left adrenal gland with the patient in right lateral recumbency. (A) Electronic calipers measuring the
length and the thickness of the cranial and caudal poles of the adrenal gland are displayed in the sagittal plane. (B) Electronic calipers measuring the thickness
and width of the caudal pole of the adrenal gland are displayed in the transverse plane.

FIG. 2. Ultrasound images illustrating measuring the right adrenal gland with the patient in left lateral recumbency. (A) Electronic calipers measuring the
length and the thickness of the cranial and caudal poles of the adrenal gland are displayed in the sagittal plane. (B) Electronic calipers measuring the thickness
and width of the caudal pole of the adrenal gland are displayed in the transverse plane.

and side differences found when the data were pooled were Results
deemed not to be clinically significant for reasons explained
later in the text. The range, mean, and standard deviation Of the original 46 dogs that met the inclusion criteria
of each measurement (length of the adrenal gland and and were screened, one dog was found to have a urinary
thickness of the cranial and caudal poles of the adrenal tract infection and was removed from the study. The re-
gland in the sagittal plane and width and thickness of the maining 45 dogs were divided by body weight into one of
caudal pole of the adrenal gland in the transverse plane) three groups (< 10 kg, 10–30 kg, and > 30 kg), with 15
were calculated for each dog for the adrenal glands. A dogs in each group, representing small, medium, and large
mixed model, ANOVA was also used to compare measure- breed sizes of dogs, respectively. The weights (mean ± SD)
ments between the three body weight groups. Correlation of the dogs in each group were as follows: < 10 kg = 6.1 ±
between each measurement of the adrenal glands and body 2.2 kg (range 2.2–9.0 kg), 10–30 kg = 21.9 ± 4.7 kg (range
weight were assessed using a Pearson’s test for each of the 13.7–29.7 kg), and > 30 kg = 38.5 ± 7.8 kg (range 30.3–62.7
three body weight groups. The correlations were qualified kg). Breeds represented in the < 10 kg weight group were
using the following scale: r = 0 – 0.30, r = 0.31 – 0.70, Chihuahua (n = 4), Puggle (3), Dachshund (2), Jack Russell
and r ࣙ 0.71 were mild, moderate, or a strong correlation, terrier mix (2), Jack Russell terrier (1), Boston terrier (1),
respectively. Data were analyzed with statistical analysis Chinese crested (1), and Shih tzu (1). Breeds included in
systems (SASTR release version 9.1, Cary, N.C.) software. the 10–30 kg weight group were Australian shepherd mix
Significance was set at the P < 0.05 level. (2), Border collie mix (2), Boxer mix (2), Pit bull mix (2),
VOL. 00, NO. 0 US EVALUATION OF ADRENAL GLAND SIZE IN DOGS 5

Blue tick hound (1), Chow chow mix (1), German shepherd cranial pole thickness of the left and right adrenal glands
mix (1), Labrador retriever (1), Labrador retriever mix (1), in the sagittal plane in 10–30 kg dogs was 0.05 cm and 0.07
Rottweiler mix (1), and Plott hound mix (1). Breeds repre- cm, respectively, and for the > 30 kg dogs was 0.06 cm and
sented in the > 30 kg weight group included Doberman pin- 0.09 cm, respectively. Similarly, the mean difference for the
scher (2), Labrador retriever (2), Labrador retriever mix (2), caudal pole thickness of the left and right adrenal gland in
Black and tan hound (1), Boxer (1), German shepherd (1), the transverse plane in 10–30 kg dogs was 0.10 cm and 0.07
Great dane (1), Golden retriever (1), Greyhound mix (1), cm, respectively, and for the > 30 kg dogs was 0.05 cm and
Red bone hound (1), Rottweiler mix (1), and Weimaraner 0.09 cm, respectively.
(1). There were seven neutered males and eight spayed fe- The mean, range, and standard deviation for each
males in the < 10 kg weight group, two intact males, four adrenal gland measurement in each body weight group are
neutered males, and nine spayed females in the 10–30 kg summarized in Table 4 while 95% confidence intervals (CI)
weight group, and 11 neutered males and four spayed fe- are provided in Table 5. Each adrenal gland parameter was
males in the > 30 kg weight group. The mean age was 3 statistically different between the three groups (P < 0.001).
years (range 1–6 years), 3 years (range 1–5 years), and 4 The correlations between body weight and the length of the
years (range 2–6 years), for the < 10 kg, 10–30 kg, and > adrenal glands measured in the sagittal plane for each of
30 kg groups, respectively. the three groups and between body weight and the cranial
Adrenal glands were visualized in both planes in all dogs pole thickness of the adrenal glands measured in the sagit-
preferentially using a subcostal approach where the adrenal tal plane for dogs 10–30 kg were significant and moderate
glands were cranially positioned. However, in some of the (r = 0.35 – 0.69; P < 0.001). The remainder of the correla-
dogs > 30 kg, an intercostal approach was necessary for tions for each adrenal gland parameter were weak (r = 0.09
visualizing the right adrenal gland as a result of its more – 0.28; P > 0.0001 – 0.21). However, all of the weak corre-
cranial anatomical position compared to the left adrenal lations were significant except for the caudal pole thickness
gland. of the adrenal glands measured in the sagittal plane for
A significant difference was found between body weight dogs > 30 kg and the caudal pole thickness of the adrenal
groups for each adrenal gland measurement regardless of glands measured in the transverse plane for dogs < 10 kg
observer or adrenal gland side (P < 0.0001). There were and > 30 kg (Table 6).
no significant differences for either observer when each
observer’s measurements were compared to their own (in-
Discussion
traobserver difference) for any of the adrenal gland pa-
rameters in each of the body weight groups regardless of In this study, we detected a relationship between ultra-
adrenal gland side (P = 0.37 – 0.99). sonographically measured adrenal gland dimensions and
A significant difference was found between observers (in- body weight in healthy dogs. Importantly, the currently ac-
terobserver difference) for all adrenal gland parameters (P cepted cut-off value for thickness of the adrenal gland 0.74
ࣘ 0.0001 – 0.03) except when the thickness of the cranial cm,5,6 which has been adopted as a means of distinguish-
pole of the left and right adrenal glands were measured in ing adrenal gland enlargement in dogs regardless of body
the sagittal plane (P = 0.41). Due to the significant differ- weight, may not be appropriate for small to medium breed
ences noted, the interobserver coefficient of variation was dogs. In this study, dogs in the < 10 kg or 10–30 kg body
calculated for each adrenal gland measurement (Table 2); weight groups all had measurements (cranial or caudal pole
all were at the upper end or slightly above what is typically thickness or caudal pole width) in any plane (sagittal or
acceptable (14–20%). Table 3 further shows the mean differ- transverse) less than 0.64 cm and 0.70 cm, respectively as
ence between the two observers’ measurements was small shown by the range of measurements obtained (Table 4).
(0.04–0.28 cm) and the caudal pole thickness of the adrenal Further, the upper threshold of the 95% confidence inter-
gland measured in a sagittal or a transverse plane had the val of the mean for the cranial or caudal pole thickness or
least difference between the two observers’ measurements caudal pole width in any plane for dogs in the < 10 kg or
(0.04–0.07 cm), which was only a 10–13% difference from 10–30 kg body weight groups was 0.54 cm and 0.64 cm, re-
the mean of the measurements. spectively (Table 5). Thus, more precise cut-off values may
Lastly, no significant difference was detected in any be able to be established for adrenal gland size by catego-
measurements between the left and right adrenal gland rizing dogs as < 10 kg, 10–30 kg, or > 30 kg in an effort
(P = 0.28–0.53) with two exceptions: (1) the thickness of to supplement a diagnosis of hyperadrenocorticism. Two
the cranial pole of the adrenal gland when measured in the earlier studies which evaluated the caudal pole thickness of
sagittal plane (P < 0.0001) and (2) the width of the caudal the adrenal glands in a sagittal plane in dogs < 10 kg found
pole of the adrenal gland when measured in the transverse the cut-off value for this dimension to be < 0.5911 cm and
plane (P = 0.0008). Both exceptions were noted in the 10– 0.6017 cm. We wanted to relate these findings in dogs < 10
30 kg and > 30 kg groups. The mean difference for the kg with dogs of other sizes. Therefore, when dividing our
6 SOULSBY ET AL. 2014

TABLE 2. Interobserver Coefficient of Variation of Body Weight Groups and Adrenal Gland Measurements

Body Weight
< 10 kg 10–30 kg > 30 kg
(n = 15) (n = 15) (n = 15)
Variable Coefficient of Variation Coefficient of Variation Coefficient of Variation
Sagittal plane
Length 18% 19% 15%
Cranial pole thickness 17% 16% 20%
Caudal pole thickness 16% 16% 17%
Transverse plane
Caudal pole thickness 18% 14% 14%
Caudal pole width 18% 15% 16%

kg, kilograms; n, number of dogs.

TABLE 3. Mean Difference Between Observers’ Measurements and Percentage Difference from the Mean for Adrenal Gland Measurements in Body
Weight Groups

Body Weight
< 10 kg 10 – 30 kg > 30 kg
(n = 15) (n = 15) (n = 15)
Mean Difference Between Mean Difference Between Mean Difference Between
Variable Observers’ Measurements Observers’ Measurements Observers’ Measurements
Sagittal plane
Length 0.22 cm (14%) 0.27 cm (12%) 0.28 cm (11%)
Cranial pole thickness 0.12 cm (34%) 0.07 cm (16%) 0.08 cm (16%)
Caudal pole thickness 0.04 cm (11%) 0.06 cm (13%) 0.07 cm (13%)
Transverse plane
Caudal pole thickness 0.04 cm (10%) 0.06 cm (12%) 0.07 cm (12%)
Caudal pole width 0.04 cm (11%) 0.06 cm (13%) 0.10 cm (17%)

kg, kilograms; n, number of dogs; cm, centimeters;%, percentage of the mean.

TABLE 4. Comparison of Body Weight Groups and Adrenal Gland Measurements

Body Weight
< 10 kg 10 – 30 kg > 30 kg
(n = 15) (n = 15) (n = 15)
Variable Mean∗ Range SD Mean∗ Range SD Mean∗ Range SD
Weight (kg) 6.13 2.20–9.00 2.07 21.89 13.70–29.70 4.66 38.51 30.30–62.70 7.83
Sagittal plane (cm)
Length 1.61 0.90–2.39 0.29 2.19 1.23–3.24 0.41 2.66 1.62–3.78 0.41
Cranial pole thickness 0.35 0.20–0.56 0.06 0.43 0.28–0.69 0.07 0.50 0.29–0.88 0.10
Caudal pole thickness 0.37 0.24–0.54 0.06 0.45 0.31–0.68 0.07 0.54 0.33–0.80 0.09
Transverse plane (cm)
Caudal pole thickness 0.40 0.27–0.58 0.07 0.50 0.36–0.70 0.07 0.58 0.37–0.81 0.08
Caudal pole width 0.38 0.26–0.64 0.07 0.48 0.31–0.66 0.07 0.58 0.32–0.82 0.09

The mean for all measurements is statistically different between the body weight groups (P < 0.001).
kg, kilograms; n, number of dogs; SD, standard deviation; cm, centimeters.

TABLE 5. Ninety-five Percent Confidence Intervals for Body Weight Groups and Adrenal Gland Measurements

Body Weight
< 10 kg 10 – 30 kg > 30 kg
(n = 15) (n = 15) (n = 15)
Variable Mean∗ 95% Confidence Interval† Mean∗ 95% Confidence Interval† Mean∗ 95% Confidence Interval†
Sagittal plane (cm)
Length 1.61 1.03–2.19 2.19 1.37–3.01 2.66 1.84–3.48
Cranial pole thickness 0.35 0.23–0.47 0.43 0.29–0.57 0.50 0.30–0.70
Caudal pole thickness 0.37 0.25–0.49 0.45 0.31–0.59 0.54 0.36–0.72
Transverse plane (cm)
Caudal pole thickness 0.40 0.26–0.54 0.50 0.36–0.64 0.58 0.42–0.74
Caudal pole width 0.38 0.24–0.52 0.48 0.34–0.62 0.58 0.40–0.76

The mean for all measurements is statistically different between the body weight groups (P < 0.001).

Mean ±2 standard deviations.
kg, kilograms; n, number of dogs; cm, centimeters.
VOL. 00, NO. 0 US EVALUATION OF ADRENAL GLAND SIZE IN DOGS 7

TABLE 6. Correlation of Body Weight Groups and Adrenal Gland Measurements

Body Weight
< 10 kg 10 – 30 kg > 30 kg
(n = 15) (n = 15) (n = 15)
r P- r P- r
Variable Value Correlation Value Value Correlation Value Value Correlation P- Value
Sagittal plane
Length 0.60 Moderate <0.0001 0.59 Moderate <0.0001 0.69 Moderate <0.0001
Cranial pole thickness 0.24 Weak 0.0010 0.35 Moderate <0.0001 0.17 Weak 0.0256
Caudal pole thickness 0.24 Weak 0.0011 0.18 Weak 0.0184 -0.09 Weak 0.2082
Transverse plane
Caudal pole thickness 0.10 Weak 0.2043 0.28 Weak 0.0001 0.09 Weak 0.2089
Caudal pole width 0.19 Weak 0.0122 0.24 Weak 0.0010 0.18 Weak 0.0138

Scale used to qualify the correlation was r = 0 – 0.3: weak, r = 0.31 – 0.7: moderate, and r ࣙ 0.71: strong.
kg, kilograms; n, number of dogs; cm, centimeters.

population, we kept dogs < 10 kg as representative group Distance measurement errors in ultrasonography are due
for small breed dogs and arbitrarily divided dogs into two to inherent image pixelation, which are random in nature
additional weight categories of 10–30 kg and > 30 kg to rep- and present in all digital images, and modality-specific reg-
resent groups of medium and large breed dogs, respectively. istration errors, which result from the image acquisition
In our study, as in earlier studies,5,7,12 the length of the process.21 Digital caliper measurements in ultrasound are
adrenal glands measured in the sagittal plane had the high- dependent on image blurring and the caliper placement skill
est correlation, which was moderate, with body weight of the operator. Vertical digital caliper measurements are
(Table 6). The finding of only weak to moderate correla- assumed not to have error since these measurements only
tions (r = 0.09 – 0.35) between the remaining adrenal gland have to take into account the axial blurring of the target’s
measurements and body weight may have been due to the image, while horizontal digital caliper measurements take
small sample size in each body weight group (n = 15). In into consideration beam width and result in a wider lat-
addition, the weakest correlations, i.e. caudal pole thick- eral blurring of the target’s image.21 Therefore, it has been
ness of the adrenal gland in the sagittal plane for dogs > suggested that with careful technique, horizontal measure-
30 kg and the caudal pole thickness of the adrenal gland in ments have the likelihood of random ± 1 pixel error in each
the transverse plane for dogs < 10 kg and > 30 kg, were not cursor’s placement.21 Pixelation errors are small in clinical
statistically significant (P > 0.05). If the sample size were images, so for a typical pixel with of 0.30 mm, the maxi-
larger for each of the body weight groups, the correlations mum error would be ± 0.64 mm, ± 1.20 mm, and ± 1.80
may have been stronger and reached significance for all. mm for cursor placement uncertainties of ± 1, ± 2, and ± 3
The finding of significant interobserver differences and pixel widths, respectively.21 Ultrasound-specific errors are
a high normal to above normal coefficient of variation for due to target misregistration and edge shifting, which are
each adrenal gland measurement (Table 2) was unexpected. unavoidable in clinical ultrasound images but adds most of
At the time the study was completed, one of the observers the unpredictability of error in distance measurements.21
(S.N.S.) was beginning the second year of a radiology resi- The authors attempted to minimize error and improve ac-
dency. The other observer (J.A.H.) is a board-certified radi- curacy by following published recommendations including
ologist with over 20 years of experience and whose primary (1) performing many measurements (three in this study for
modality of interest is ultrasound. It is well understood each adrenal gland measurement by each observer) and av-
that the findings of ultrasonography are highly observer- eraging their results; (2) using a separate ultrasound image
dependent based on skill level. Ultrasonographic measure- for each of the parameters measured using digital calipers,
ments of any organ, but particularly the adrenals, depend using a standard transducer; (3) selecting a standard fre-
on the expertise of the observer and the measured length quency range and multizone focus scheme for the trans-
and width of both adrenal glands increases with greater ducer; (4) placing the structure to be measured in the center
training of the observer.12 Thus, the significant interob- of the slice thickness; (5) orienting the measured structure
server differences in our study could be due to a difference in the scan plane so the distance to be measured is as close
in experience. However, the mean differences between the to possible to being parallel to the image lines; (6) magni-
two observers’ measurements were minimal, particularly fying the structure to be imaged so it occupies at least one
for the measurements of the thickness of the caudal pole of half of the image field of view, and (7) adjusting the gain to
the adrenal glands in both a sagittal and transverse plane obtain a standard gray-scale of the anatomy.21 One study
(Table 3). Therefore, the interobserver variability was likely in humans determined there was an error of less than 20%
statistically, but not clinically significant. for ultrasonographic measurements in 77% of normal and
8 SOULSBY ET AL. 2014

pathologic aortic walls.22 Using the data from the study by differences were deemed clinically insignificant. This was
Grooters et al., we calculated a similar error of an average decided based on the mean differences for each adrenal
of 21% where ultrasonographic measurements underesti- gland measurement between the two observers being
mated gross measurements of the adrenal glands.8 minimal (0.04–0.28 cm) for each of the three body weight
An additional unexpected finding was detection of a sig- groups as shown in Table 3. The minor disparities noted
nificant difference between the left and right adrenal glands in the measurements between the two observers were most
for the thickness of the cranial pole measured in the sagit- likely a result in the difference in experience level between
tal plane and the width of the caudal pole measured in the the two sonographers performing the measurements. De-
transverse plane. The difference in the thickness of the cra- spite the variation in experience level, the mean differences
nial pole measured in the sagittal plane between the left and between the two observers when measuring the caudal pole
right adrenal glands could be explained by the variation in of the adrenal gland in a sagittal plane was only 0.04–0.07
the gross structure of the cranial pole of the adrenal glands. cm. Third, a significant difference in the thickness of the
Both observers subjectively reported occasional difficulty caudal pole between the left and right adrenal gland was
in their ability to completely distinguish the cranial pole not detected. Fourth, both observers were always able to
of the adrenal gland regardless of side or body weight of easily and reliably visualize the caudal pole of the adrenal
the dog. The inability to fully evaluate the cranial poles gland in the sagittal plane; the caudal pole is often more
may be due to the slight dorsoventral flattening of the left clearly visually defined and consistently has a similar shape
adrenal gland while the cranial pole of the right adrenal than the cranial pole and the adrenal glands are typically
gland has an acute angular bend with the end of the cra- evaluated in a sagittal plane rather than a transverse plane
nial pole projected cranially.23 This angular bend can result ultrasonographically. Similarly, previous studies found that
in the cranial pole of the right adrenal gland having an the thickness of the caudal pole of the adrenal gland mea-
“arrowhead” shaped appearance2 and make measurement sured in the sagittal plane best correlated with grossly mea-
of this portion of the adrenal gland difficult. The “arrow- sured adrenal glands5,6,8 and had the least intraobserver
head” shape of the cranial pole was noted by both observers and interobserver variability.9,12
in some dogs in the current study. In such cases, both ob- Dogs have great variability in body weight. Our hypoth-
servers attempted to measure the cranial pole of the adrenal esis was that adrenal gland measurements varied with body
gland toward the tip of the “arrowhead.” The difference in weight. Our objective was to see if body weight size made a
shape between the cranial poles of the right and left adrenal difference in this pilot study to determine if eventually more
glands may also account for the difference in measurements precise reference intervals for adrenal size in three different
of the cranial pole of the left and right adrenal glands. Thus, weight categories of dogs could be established. Based on
delineation and measurement of the cranial poles is more previous research using a single cut-off value of 0.74 cm
challenging than that of the caudal poles of each adrenal for distinguishing a normal adrenal gland from one that
gland. A similar conclusion about the difficulty in visual- is enlarged for all dogs regardless of their body size does
izing and measuring the cranial pole of the adrenal glands not appear to be appropriate. Thus, in the populations of
has been reported.12 In addition, numerous studies have dogs used in this study, the authors of this paper found the
documented that the right adrenal gland is more difficult following values for the maximum thickness of the caudal
to image as compared to the left adrenal gland.5,7,16,20 The pole of the adrenal gland in the sagittal plane for either the
difference in the width of the caudal pole measured in the left or the right adrenal gland in normal, healthy dogs: (1) ࣘ
transverse plane can be explained by the lack of skill by 0.54 cm for dogs < 10 kg, (2) ࣘ 0.68 cm for dogs 10–30 kg,
both observers in evaluating the adrenal gland in this man- and (3) ࣘ 0.80 cm for dogs > 30 kg. However, additional
ner. The transverse plane is not typically used to image the studies evaluating at risk populations of dogs (e.g., older
adrenal glands as part of the ultrasound examination in our dogs) and dogs with known hyperadrenocorticism for fine
clinical patients. Also, it has been previously reported that tuning of these suggested cut-off values.
if the measurement for the width of the adrenal gland in There were a few factors which were limitations in this
the transverse plane was taken at an oblique angle by one study. First, the overall sample size was only 45 dogs,
or both observers, the width could be overestimated.5–7 with 15 dogs in each body weight group and creating
Therefore, for several reasons, the best measurement for authoritative reference intervals require a larger population
assessing adrenal gland size appears to be the thickness of dogs. However, this pilot study showed that even with
of the caudal pole measured in the sagittal plane. First, only 15 dogs in each body weight group, a significant
no intraobserver differences were detected for either ob- difference was seen with regard to the upper limit of
server with regard to this dimension regardless of whether what is considered normal for the caudal pole thickness
the data were examined separately for each observer or of the adrenal gland in a sagittal plane in dogs divided
pooled for both observers. Second, although interobserver into different body weights. In addition, we chose our
differences were detected when the data were pooled, these new cut-off values for the thickness of the caudal pole
VOL. 00, NO. 0 US EVALUATION OF ADRENAL GLAND SIZE IN DOGS 9

of the adrenal gland measured in a sagittal plane based the animals’ normal history, physical examination, blood
on the upper threshold of the recorded range for each work, and urinalysis, and therefore, specific endocrine test-
body weight group rather than 95% confidence interval ing was not performed. For future studies, comparing nor-
of the mean ± two standard deviations, because prior mal, healthy dogs to dogs with confirmed hypoadrenocor-
studies3,5,8,11,16,17 have historically used this same method. ticism or hyperadrenocorticism and/or dogs with other
The difference in the proposed new cut-off values between medical conditions or dogs on medications such as insulin,
dogs weighing < 10 kg and dogs weighing 10–30 kg was steroids, or antibiotics, would allow sensitivity and speci-
0.14 cm and the difference in the proposed cut-off values ficity as well as false positive and false negative rates to be
between dogs weighing 10–30 kg and dogs weighing > 30 calculated for our proposed new cut-off values.
kg was similar at 0.12 cm. Additional research is needed In conclusion, we found that adrenal gland size varies
to acquire a larger sample size in each weight group and significantly between dogs divided into < 10 kg, 10–30 kg,
determine whether the proposed cut-off values from our and > 30 kg groups. We propose that an adrenal gland is
study are accurate. Second, only dogs between 1 and 6 likely to be normal in size if the thickness of the caudal
years of age and with a body condition score of 2 to 6 out pole of the adrenal gland in the sagittal plane is (1) ࣘ 0.54
of 9 were used. Including dogs of all ages and body weight cm for dogs < 10 kg, (2) ࣘ 0.68 cm for dogs 10–30 kg,
condition could have effects on adrenal gland size as well and (3) ࣘ 0.80 cm for dogs > 30 kg. These proposed new
as the ability to image and measure them accurately. cut-off values should be used cautiously until additional
Specific endocrine testing (e.g., an ACTH stimulation studies involving a larger number of dogs, older dogs (who
test, a low-dose dexamethasone suppression test, or a urine are more likely to be at risk for hyperadrenocorticism) and
cortisol to creatinine ratio) to definitively rule out under- dogs with known hyperadrenocorticism divided into similar
lying hyperadrenocorticism was not done. The most re- weight groups can be performed to support or refute our
cent consensus statement from the American College of conclusions.
Veterinary Internal Medicine on diagnosing canine hyper-
adrenocorticism states that hyperadrenocorticism is a clin-
ACKNOWLEDGMENTS
ical diagnosis and testing for this disease is not warranted
The authors would like to thank James C. Wright, DVM, PhD,
in dogs without clinical signs of hyperadrenocorticism.1 In DACVPM for his help with the statistical analysis and are grateful to
our population of dogs, it was determined that none of the the Department of Clinical Sciences of the Auburn University College
dogs were likely to have hyperadrenocorticism based on of Veterinary Medicine for financial support of this project.

REFERENCES
1. Behrend EN, Kooistra HS, Nelson R, et al. Diagnosis of spontaneous 11. Choi J, Kim H, Yoon J. Ultrasonographic adrenal gland mea-
canine hyperadrenocorticism: 2012 ACVIM consensus statement (small an- surements in clinically normal small breed dogs and comparison with
imal). J Vet Intern Med 2013;27:1292–1304. pituitary-dependent hyperadrenocorticism. J Vet Med Sci 2011;73:985–
2. Graham J. Adrenal glands. In: Pennick D, d’Anjou M-A (eds): Atlas 989.
of small animal ultrasonography. Ames: Blackwell, 2008;385–396. 12. Mogicato G, Layssol-Lamour C, Conchou, F, et al. Ultrasono-
3. de Chalus T, Combes A, Bedu A-S, et al. Ultrasonographic adrenal graphic evaluation of the adrenal glands in healthy dogs: repeatability, re-
gland measurements in healthy Yorkshire terriers and Labrador retrievers. producibility, observer-dependent variability, and the effect of bodyweight,
Anat Histol Embryol 2013;42:57–64. age and sex. Vet Rec 2011;168:130–135.
4. Widmer WR, Guptill L. Imaging techniques for facilitating diag- 13. Wenger M, Mueller C, Kook PH, Reusch CE. Ultrasonographic
nosis of hyperadrenocorticism in dogs and cats. J Am Vet Med Assoc evaluation of adrenal glands in dogs with primary hypoadrenocorticism or
1995;206:1857–1864. mimicking disease. Vet Rec 2010;167:207–210.
5. Barthez PY, Nyland TG, Feldman EC. Ultrasonographic evaluation 14. Hoerauf A, Reusch C. Ultrasonographic evaluation of the adrenal
of the adrenal glands in dogs. J Am Vet Med Assoc 1995;207:1180–1183. glands in six dogs with hypoadrenocorticism. J Am Anim Hosp Assoc
6. Barthez PY, Nyland TG, Feldman EC. Ultrasonography of the 1999;35:214–218.
adrenal glands in the dog, cat, and ferret. Vet Clin North Am Small Anim 15. Hoerauf A, Reusch C. Ultrasonographic characteristics of both
Pract 1998;28:869–885. adrenal glands in 15 dogs with functional adrenocortical tumors. J Am Anim
7. Douglass JP, Berry CR, James S. Ultrasonographic adrenal gland Hosp Assoc 1999;35:193–199.
measurements in dogs without evidence of adrenal disease. Vet Radiol Ul- 16. Grooters AM, Biller DS, Theisen SK, Miyabayashi T. Ultrasono-
trasound 1997;38:124–130. graphic characteristics of the adrenal glands in dogs with pituitary-dependent
8. Grooters AM, Biller DS, Merryman J. Ultrasonographic parameters hyperadrenocorticism: comparison with normal dogs. J Vet Intern Med
of normal canine adrenal glands: comparison to necropsy findings. Vet Ra- 1996;10:110–115.
diol Ultrasound 1995;36:126–130. 17. De Marco V, Pereira RS, Kage NK, et al. Ultrasonographic adrenal
9. Barberet V, Pey P, Duchateau L, et al. Intra- and interobserver vari- glands thickness measurement in dogs with pituitary-dependent hypera-
ability of ultrasonographic measurements of the adrenal glands in healthy drenocorticism in comparison with normal dogs matched by weight body. J
beagles. Vet Radiol Ultrasound 2010;51:656–660. Vet Intern Med 2010;24:746–747 (abstr.).
10. Hoerauf A, Reusch C. Ultrasonographic evaluation of the adrenal 18. Benchekroun G, de Fornel-Thibaud P, Rodriguez Piñeiro MI,
glands in healthy dogs, dogs with no evidence of endocrine disease and dogs et al. Ultrasonography criteria for differentiating ACTH dependency
with Cushing’s disease. Vet Radiol Ultrasound 1995;36:434 (abstr.). from ACTH independency in 47 dogs with hyperadrenocorticism
10 SOULSBY ET AL. 2014

and equivocal adrenal asymmetry. J Vet Intern Med 2010;24:1077– 21. Goldstein A. Errors in ultrasound digital image distance measure-
1085. ments. Ultrasound Med Biol 2000;26:1125–1132.
19. Bertolini G, Furlanello T, De Lorenzi D, Caldin M. Computed to- 22. Pignoli P, Tremoli E, Poli A, et al. Intimal plus medial thickness of
mographic quantification of canine adrenal gland volume and attenuation. the arterial wall: a direct measurement with ultrasound imaging. Circulation
Vet Radiol Ultrasound 2006;5:444–448. 1986;74:1399–1406.
20. Nyland TG, Mattoon JS, Herrgesell EJ, Wisner ER. Adrenal glands. 23. Hullinger RL. The endocrine system. In: Evans HE (ed): Miller’s
In: Nyland TG, Mattoon JS (eds): Small animal diagnostic ultrasound, 2nd anatomy of the dog, 3rd ed. Philadelphia: W.B. Saunders, 1993;559–
ed. Philadelphia: Saunders, 2002;196–206. 585.