1. Original Study Journal of Veterinary Emergency and Critical Care 24(4) 2014, pp 398–402
doi: 10.1111/vec.12203
Evaluation of two point-of-care ethylene glycol
tests for dogs
Karina J. Creighton, BVSc; Amy M. Koenigshof, DVM, MS, DACVECC; Christian D. Weder, DVM
and L. Ari Jutkowitz, VMD, DACVECC
Abstract
Objective – To evaluate 2 point-of-care ethylene glycol (EG) tests in dogs.
Design – Prospective, randomized, blinded laboratory evaluation.
Setting – University teaching hospital.
Animals – Ten healthy adult dogs.
Interventions – Jugular venipuncture and in vitro evaluation for detection of EG in canine blood.
Measurements – Whole blood samples were centrifuged and separated, and the plasma was divided into 30
aliquots. The aliquots were mixed with EG to provide EG concentrations ranging from 0 to 100 mg/dL. The
EG concentration of each sample was confirmed using gas chromatography. For the VetSpec EG Qualitative
Reagent Test Kit, 100 L of each sample was added to test vials and compared with 20 and 50 mg/dL reference
vials. For the Kacey EG Test Strips, 20 L of each sample was added to the test circle and compared with the
color chart provided by the manufacturer. For each test, samples were prepared in groups of 5 and presented
in randomized order to 2 readers who were blinded to the presumed EG concentration. Samples were scored
as negative, 20–50 mg/dL, or greater than 50 mg/dL. For each test, the sensitivity and specificity for detecting
EG was calculated. Cohen’s unweighted kappa coefficient was calculated to determine the degree of agreement
between readers.
Main Results – For detecting EG, the Kacey EG Test Strips had excellent sensitivity and specificity (both 100%)
and good agreement between readers. The VetSpec EG Qualitative Reagent Test Kit was less sensitive and
specific (65% and 70% for the first reader, 95% and 40% for the second) with less agreement.
Conclusions – Of the 2 systems evaluated, the Kacey EG Test Strips displayed greater accuracy and ease of use.
(J Vet Emerg Crit Care 2014; 24(4): 398–402) doi: 10.1111/vec.12203
Keywords: antifreeze, canine, point-of-care systems, toxicology
Introduction
Despite widespread knowledge about its toxicity, ethy-
lene glycol (EG) continues to be a common toxic
exposure in dogs.1
EG is most commonly found in
automotive antifreeze and coolants, but may also be
an ingredient in solvents, paints, and other household
From the Department of Small Animal Clinical Sciences, College of Veteri-
nary Medicine, Michigan State University, East Lansing, MI, 48824. Present
addresses: Karina Creighton, Red Bank Veterinary Hospital, Tinton Falls, NJ
07724; Christian Weder, Department of Veterinary Clinical Sciences, College
of Veterinary Medicine and Biomedical Sciences, Colorado State University,
Fort Collins, CO 80523, USA.
Supported by the Department of Small Animal Clinical Sciences, College of
Veterinary Medicine, Michigan State University, East Lansing, MI 48824.
Presented in abstract form at the 18th International Veterinary Emergency
and Critical Care Symposium, San Antonio, TX, September 2012.
The authors declare no conflict of interests.
Address correspondence and offprint requests to
Dr. Creighton, Red Bank Veterinary Hospital, 197 Hance Avenue, Tinton
Falls, NJ 07724, USA. Email: karina.creighton@rbvh.net
Submitted December 23, 2012; Accepted May 26, 2014.
Abbreviations
EG ethylene glycol
GC gas chromatography
and industrial products.2
Like other alcohols such as
ethanol and methanol, ingestion of EG initially causes
central nervous system signs such as ataxia, disorien-
tation, lethargy, stupor, or coma.3,4
EG intoxication can
lead to fatal nephrotoxicity in dogs if not recognized
and treated in the hours following ingestion.5–7
While
EG is not nephrotoxic, metabolites of EG, primarily
glycolic acid, cause cytotoxic damage to renal tubular
cells leading to acute tubular necrosis and acute re-
nal failure.8,9
Accumulation of calcium oxalate crystals
within the tubular lumens may also cause an obstruc-
tive nephropathy.2
Administration of an inhibitor of al-
cohol dehydrogenase, the enzyme responsible for the
first step in the metabolism, allows EG to be excreted
398 C
Veterinary Emergency and Critical Care Society 2014
2. Two point-of-care ethylene glycol tests for dogs
unchanged in urine and prevents the forma-
tion of nephrotoxic metabolites.8,10
Ethanol and 4-
methylpyrazole have been shown to be effective anti-
dotes to EG intoxication if administered to dogs within
6–8 hours of ingestion.4,8,11
Beyond this therapeutic win-
dow dialysis is indicated, as this will remove both EG
and its toxic metabolites.12,13
EG intoxication is suspected when there is a history of
possible exposure and compatible clinical signs or lab-
oratory abnormalities. Common early biochemical ab-
normalities include metabolic acidosis, increased anion
gap, increased serum osmolar gap, hyperglycemia, hy-
perphosphatemia, and hypocalcemia.3,4,6
Urinalysis may
show calcium oxalate monohydrate crystalluria.3,4,6
Gas
chromatography (GC) is considered the gold standard
method for quantitative measurement of EG levels in
serum or plasma.14,15
However, the availability of GC
is limited and results may not be obtainable within a
clinically useful timeframe.
There are currently 2 commercially available point-of-
care EG assays that may assist in the early diagnosis of
EG intoxication.a,b
Both of these tests are colorimetric as-
says designed to detect EG at levels as low as 20 mg/dL.
However, to the authors’ knowledge there have been no
published studies evaluating the accuracy of these tests
for detecting EG in canine plasma. The purpose of this
study was to evaluate 2 point-of-care EG tests using ca-
nine plasma containing EG concentrations ranging from
0 to 100 mg/dL, and compare the results with those ob-
tained using GC.
Materials and Methods
This investigation was performed with approval from
the Michigan State University Institutional Animal Care
and Use Committee. Whole blood samples (10mL) were
collected via jugular venipuncture, with informed con-
sent, from 10 healthy dogs owned by staff and students of
the Michigan State University Veterinary Teaching Hos-
pital or by clients associated with the hospital blood
donor program. Health status was determined by his-
tory and physical examination. Dogs had not received
any medications other than flea and heartworm preven-
tative for at least 2 weeks prior to sample collection.
After collection the plasma was separated, divided into
3 aliquots of at least 700 L, and frozen at −40°C until
the study date.
Pure EGc
(density 1.113 g/mL) was diluted with
phosphate-buffered saline to a concentration of 10
g/L. Plasma samples were thawed and warmed to
room temperature. The aliquots of each patient sample
were divided into 3 treatment groups (10 samples in each
group). For group A, 42 L of the dilute EG was added to
658 L of the plasma sample and mixed by rapid pipet-
ting to obtain a test sample with an EG concentration of
approximately 60 mg/dL. For group B, 17 L of the di-
lute EG was added to 683 L of the plasma sample and
mixed by rapid pipetting to obtain a test sample with
an EG concentration of approximately 25 mg/dL. For
group C, 700 L of unadulterated plasma was used as a
negative control. One investigator prepared all samples.
Each sample was evaluated by Test 1a
and Test 2b
as well
as GC.
Evaluation of Test 1
The tests were prepared in batches of 5 and presented
to 2 readers in the same order. All 30 samples were
randomizedd
in the order they were presented to the
readers. The readers were blinded to the concentration
of the samples used for each test, and to the results of
the other reader. For each batch of 5 tests, 5 test vials
and 2 reference vials were prepared by adding 1 mL of
reagent diluent to each vial. Working simultaneously, 1
investigator then added 0.1 mL of plasma to each of the
test vials, while a second investigator added 0.1 mL of
the 20 mg/dL control to 1 reference vial and 0.1 mL of
the 50 mg/dL control to the second reference vial. After 5
minutes, the samples were handed to 1 reader who had
1 minute to evaluate the 5 samples. All samples were
then passed to the second reader. The color change in
the test samples was compared with the reference vials
and scored as negative, greater than 20 mg/dL but less
than 50 mg/dL, or greater than 50 mg/dL.
Evaluation of Test 2
Using the pipette provided by the manufacturer, 20 L
of each test sample was pipetted onto a test strip. Tests
were prepared in batches of 5, with the randomization of
all 30 samples being repeated so that the samples were
analyzed in a different order from Test 1. After 8 minutes,
the samples were handed to 1 reader who had 1 minute
to evaluate the 5 samples. All samples were then passed
to the second reader. Using the color chart provided by
the manufacturer, the test strips were scored as negative,
greater than 20 mg/dL but less than 50 mg/dL, or greater
than 50 mg/ dL.
Gas chromatography
The remainder of each sample was refrigerated and sub-
mitted to the Michigan State University Diagnostic Cen-
ter for Population and Animal Health for quantitative
analysis by GC. Each serum sample was filtered using a
micropartition systeme
and centrifuged at 2,000 × g for
10 minutes. Fluid recovery was 100 L per sample, and
1 L was analyzed directly by GC. EG concentrations
were determined on a gas chromatographf
equipped
C
Veterinary Emergency and Critical Care Society 2014, doi: 10.1111/vec.12203 399
3. K. J. Creighton et al.
Table 1: Agreement between gas chromatography and Test 1 for each reader
Gas chromatography: Negative
(n = 10)
20–50 mg/dL
(n = 10)
50 mg/dL
(n = 10)
Negative
(n = 10)
20–50 mg/dL
(n = 10)
50 mg/dL
(n = 10)
Reader 1 Negative 7 3 4 Reader 2 Negative 4 1 0
20–50 mg/dL 1 4 1 20–50 mg/dL 3 4 1
50 mg/dL 2 3 5 50 mg/dL 3 5 9
with dual flame ionization detectors. Chromatographic
columns included one 30 m × 0.53 mm × 1 m DB-
WAX column and one 30 m × 0.53 × 1 m DB-FFAP
column. GC temperature settings were column 120°C,
injector 235°C, and detector 250°C. Sample results were
compared to 55 ppm (5.5 mg/dL), 110 ppm (11 mg/dL),
220 ppm (22 mg/dL), 550 ppm (55 mg/dL), and 1,100
ppm (110 mg/dL) EG standards, and concentrations de-
termined by interpolation on a standard curve. All pos-
itives determined on the first chromatographic column
were confirmed by retention time match on the second
column.
Statistical Methods
Each test was evaluated as a qualitative and semiquan-
titative test with GC being used as the gold standard.
The sensitivity and specificity for detecting EG were
calculated for each reader. Interrater agreement and
agreement as a semiquantitative test when compared to
GC was evaluated by calculating the kappa statistic, with
the strength of agreement based on the following kappa
values: ࣘ0.20 poor, 0.21–0.40 fair, 0.41–0.60 moderate,
0.61–0.80 substantial, and 0.81–1.00 good agreement.16
Results
Gas chromatography
Samples from group A had a median EG concentration
of 100.2 mg/dL (range 52–108). Samples from group B
had a median EG concentration of 35.4 mg/dL (range
32.3–37.6). No EG was detected in any of the control
samples.
Test 1 (Table 1)
For the first reader, this test had a sensitivity of 65% and
a specificity of 70% for detecting the presence of EG. For
the second reader, this test had a sensitivity of 95% and
a specificity of 40% for detecting EG. When evaluated as
a qualitative test only, there was fair agreement between
the readers (kappa value of 0.37).
When evaluated as a semiquantitative test for deter-
mining whether EG concentrations were 20 mg/dL,
20–50 mg/dL or 50 mg/dL, there was fair agreement
between Test 1 and the results obtained by GC (kappa
values of 0.3 and 0.35 for reader 1 and reader 2, re-
spectively). There was moderate agreement between the
readers (kappa value of 0.41).
Both readers reported difficulty distinguishing color
differences between test and reference vials and between
20 mg/dL and 50 mg/dL reference vials within batches.
The sensitivity for both readers improved for the last 15
tests read (100% for both readers) compared with the
first 15 tests (41% for reader 1 and 91% for reader 2).
Specificity remained similar for the first reader (68% for
the first 15 tests, 71% for the last 15), but improved for
the second reader (0% for the first 15 tests, 57% for the
last 15).
Test 2 (Table 2)
For detecting the presence of EG, this test had a sensi-
tivity of 100% and a specificity of 100% for both readers,
with good agreement between readers (kappa value of
1.00). When evaluated as a semiquantitative test there
was substantial agreement between Test 2 and the results
obtained by GC (kappa value of 0.8 for both readers).
There was substantial agreement between the readers
(kappa value of 0.68).
Discussion
This study evaluated 2 currently available point-of-care
EG tests. Test 1 is an enzymatic and colorimetric assay.
The plasma or serum is added to a reagent solution con-
taining bacterial dehydrogenase. If EG is present, NAD
is reduced to form NADH in an amount proportional to
the EG concentration. The NADH then reacts with tetra-
zolium salt and diaphorase to form a purple dye. The
color change is compared with a reference vial that has
a control serum sample, spiked with either 20 mg/dL or
50 mg/dL of EG, added to the reagent solution. Test 2 is a
colorimetric assay, where plasma is added to the reagent
circle on the test strip and the resulting color change is
compared with a chart provided by the manufacturer.
The manufacturer does not make specific information
regarding the methodology available. Both assays need
to be read within specific timeframes recommended
by the manufacturers, as colorimetric reactions will
continue beyond this time and possibly give different
results.
400 C
Veterinary Emergency and Critical Care Society 2014, doi: 10.1111/vec.12203
4. Two point-of-care ethylene glycol tests for dogs
Table 2: Agreement between gas chromatography and Test 2 for each reader
Gas chromatography: Negative
(n = 10)
20–50 mg/dL
(n = 10)
50 mg/dL
(n = 10)
Negative
(n = 10)
20–50 mg/dL
(n = 10)
50 mg/dL
(n = 10)
Reader 1 Negative 10 0 0 Reader 2 Negative 10 0 0
20–50 mg/dL 0 6 0 20–50 mg/dL 0 6 0
50 mg/dL 0 4 10 50 mg/dL 0 4 10
The results of this study suggest that Test 2 is a sen-
sitive and specific method of detecting the presence of
EG in canine plasma. When used as a semiquantitative
test there was also substantial agreement with results
obtained by GC, considered to be the gold standard for
measuring EG concentrations in plasma, with any dis-
agreement due to a positive bias.
Test 2 has previously been evaluated using cat
plasma.17
That study found a lower sensitivity at the
lowest level of detection (20 mg/dL) and lower speci-
ficity at higher EG concentrations (80 mg/dL) than the
current study. The reasons for this difference are unclear
but may relate to changes in the test methodology. The
authors of the previous study report that the version of
the test used at that time was specific for EG, whereas the
current manufacturer’s instructions indicate that false
positives may occur with plasma samples that contain
other alcohols or sugar alcohols. As the test is designed
to be used in cats and dogs, species difference is a less
likely cause but reevaluation of this test in cats may be
warranted.
Test 1 was less sensitive and less specific for detecting
EG in canine plasma. There was also only fair agreement
between readers, with a kappa statistic of 0.37 indicating
that the agreement was less than would have been ob-
tained by chance. Test 1 was perceived by both readers
to be less user-friendly, particularly with regard to dif-
ficulties in distinguishing color differences between test
and reference vials. The consistency of reference vials
from batch to batch was also questioned. Although all
control and test vials were prepared in a standardized
procedure, the possibility of nonuniform reconstitution
of the reference vials cannot be excluded. If this was the
case, it should still be considered a limitation of this test,
as all samples and reference vials were prepared accord-
ing to manufacturer’s specifications. The sensitivity and
specificity of this test showed some improvement be-
tween the first and last 15 samples read, suggesting the
possibility of a learning curve for this test. However, this
may not reflect a true learning curve, but rather an artifi-
cial advantage conferred by the study design for readers
to compare test vials to a mental catalog of recently ob-
served vials. The average veterinary practice may not
see enough clinical cases within a suitable time frame to
benefit from this potential advantage.
When Test 1 was evaluated as a semiquantitative assay
there was only fair agreement with the results obtained
by GC. However, it should be noted that this is not de-
signed to be a semiquantitative test. Although both 20
mg/dL and a 50 mg/dL control vials are provided, the
manufacturer’s recommendation is that canine samples
be compared with the 50 mg/dL control only. The 20
mg/dL control is provided for comparison with feline
samples, as the reported lethal dose is lower in cats than
in dogs.18,19
In this study, samples were compared to both
controls for 2 reasons: to evaluate both tests in a similar
manner, and to determine the sensitivity and specificity
of the assay at the limit of detection reported by the
manufacturer. Furthermore, although peak plasma EG
greater than 50 mg/dL is considered to be the lethal
dose in dogs, in the clinical setting it would be unwise
to ignore a detectable plasma concentration that is 50
mg/dL as variable rates of absorption and metabolism
would make it difficult to determine whether a single
measurement represented peak plasma concentration.
The median EG concentration for groups A and B as
determined by GC was higher than intended. This is
likely to be the result of imprecise measurement and
mixing of the plasma and diluted EG. However, as the
results of the 2 point-of-care tests were compared with
the EG levels obtained by GC rather than the presumed
concentration, this lack of precision does not affect the
results of this study.
One limitation of this study is that it was an in vitro
evaluation and may not reflect the sensitivity and speci-
ficity of these tests in clinical practice. In particular, this
study could not evaluate the effect of EG metabolites
or other compounds that may be contained in commer-
cial EG products on the performance of these tests. Also,
both of these tests are also likely to give false positives
in dogs exposed to other alcohols and sugar alcohols.
The possibility of false positive results should be consid-
ered when making treatment decisions. However, given
the serious consequences of a false negative result, the
high sensitivity for detecting EG makes Test 2 a clinically
useful test.
In conclusion, of the 2 tests evaluated, Test 2 appears to
have the greatest accuracy for confirming EG exposure.
Further studies are needed to determine whether similar
results would be obtained in clinical cases.
C
Veterinary Emergency and Critical Care Society 2014, doi: 10.1111/vec.12203 401
5. K. J. Creighton et al.
Footnotes
a
VetSpec Ethylene Glycol Qualitative Reagent Test Kit, Catachem Inc, Ox-
ford, CT.
b
Kacey Ethylene Glycol Test Strips, Kacey Inc, Asheville, NC.
c
Ethylene Glycol ReagentPlus, Sigma-Aldrich, St. Louis, MO.
d
Excel 2010, Microsoft Corporation, Redmond, WA.
e
MPS-1, EMD Millipore, Billerica, MA.
f
Varian Star 3800, Agilent Technologies Inc, Santa Clara, CA.
References
1. McLean MK, Hansen SR. An overview of trends in animal poisoning
cases in the United States: 2002–2010. Vet Clin North Am Small Anim
Pract 2012; 42(2):219–228.
2. Kruse JA. Methanol and ethylene glycol intoxication. Crit Care Clin
2012; 28(4):661–711.
3. Grauer GF, Thrall MA, Henre BA, et al. Early clinicopathologic
findings in dogs ingesting ethylene glycol. Am J Vet Res 1984;
45(11):2299–2303.
4. Connally HE, Thrall MA, Forney SD, et al. Safety and efficacy of
4-methylpyrazole for treatment of suspected or confirmed ethylene
glycol intoxication in dogs: 107 cases (1983–1995). J Am Vet Med
Assoc 1996; 209(11):1880–1883.
5. Rowland J. Incidence of ethylene glycol intoxication in dogs and cats
seen at Colorado State University Veterinary Teaching Hospital. Vet
Hum Toxicol 1987; 29(1):41–44.
6. Thrall MA, Grauer GF, Mero KN. Clinicopathologic findings in dogs
and cats with ethylene glycol intoxication. J Am Vet Med Assoc 1984;
184(1):37–41.
7. Beckett SD, Shields RP. Treatment of acute ethylene glycol (an-
tifreeze) toxicosis in the dog. J Am Vet Med Assoc 1971; 158(4):472–
476.
8. Grauer GF, Thrall MA, Henre BA, et al. Comparison of the effects of
ethanol and 4-methylpyrazole on the pharmacokinetics and toxicity
of ethylene in the dog. Toxicol Lett 1987; 35(2–3):307–314.
9. Smith BJ, Anderson BG, Smith SA, et al. Early effects of ethylene
glycol on the ultrastructure of the renal cortex in dogs. Am J Vet Res
1990; 51(1):89–96.
10. Li TK, Theorell H. Human liver alcohol dehydrogenase: inhibi-
tion by pyrazole and pyrazole analogs. Acta Chem Scand 1969;
23(3):892–902.
11. Dial SM, Thrall MA, Hamar DW. Efficacy of 4-methylpyrazole for
treatment of ethylene glycol intoxication in dogs. Am J Vet Res 1994;
55(12):1762–1770.
12. Jacobsen D, McMartin KE. Methanol and ethylene glycol poison-
ings. Mechanism of toxicity, clinical course, diagnosis and treatment.
Med Toxicol 1986; 1(5):309–334.
13. Gabow PA, Clay K, Sullivan JB, et al. Organic acids in ethylene
glycol intoxication. Ann Intern Med 1986; 105(1):799–800.
14. Aarstad K, Dale O, Aakervik O, et al. A rapid gas chromatographic
method for determination of ethylene glycol in serum and urine. J
Anal Toxicol 1993; 17(4):218–221.
15. Jacobsen D, McMartin KE. Antidotes for methanol and ethy-
lene glycol poisoning. J Toxicol Clin Toxicol 1997; 35(2):127–
143.
16. Petrie A, Watson P. Statistics for Veterinary and Animal Science.
Oxford: Blackwell Science; 1999, pp. 173–175.
17. Acierno MJ, Serra VF, Johnson ME, et al. Preliminary validation of
a point-of-care ethylene glycol test for cats. J Vet Emerg Crit Care
2008; 18(5):477–479.
18. Grauer GF, Thrall MA. Ethylene glycol (antifreeze) poisoning in the
dog and cat. J Am Anim Hosp Assoc 1982; 18(3):492–497.
19. Milles G. Ethylene glycol poisoning with suggestions for its
treatment as oxalate poisoning. Arch Pathol 1946; 41(6):631–
638.
402 C
Veterinary Emergency and Critical Care Society 2014, doi: 10.1111/vec.12203