- Email: [email protected]
Evaluation of assays for troponin I in healthy horses and horses with cardiac disease
Accepted Manuscript Title: Evaluation of assays for troponin I in healthy horses and horses with cardiac disease Author: N. Van Der Vekens, A. Decloedt, S. Sys, S. Ven, D. De Clercq, G. van Loon PII: DOI: Reference:
S1090-0233(14)00492-4 http://dx.doi.org/doi: 10.1016/j.tvjl.2014.11.015 YTVJL 4344
To appear in:
The Veterinary Journal
Accepted date:
27-11-2014
Please cite this article as: N. Van Der Vekens, A. Decloedt, S. Sys, S. Ven, D. De Clercq, G. van Loon, Evaluation of assays for troponin I in healthy horses and horses with cardiac disease, The Veterinary Journal (2014), http://dx.doi.org/doi: 10.1016/j.tvjl.2014.11.015. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
1 2 3 4 5 6 7 8 9 10 11 12 13
Evaluation of assays for troponin I in healthy horses and horses with cardiac disease
N. Van Der Vekens *, A. Decloedt, S. Sys, S. Ven, D. De Clercq, G. van Loon Department of Large Animal Internal Medicine, Faculty of Veterinary Medicine, Ghent, Salisburylaan 133, B-9820 Merelbeke, Belgium
* Corresponding author. Tel.: +32 9 2647584. E-mail address: [email protected] (N. Van Der Vekens).
1 Page 1 of 20
Highlights
14 15
Two troponin I assays were compared in horses, including 23 healthy horses and 72 horses with cardiac disease.
16 17
Large quantitative differences between assays were demonstrated.
18
A cut-off value for cardiac disease was identified for only one of the assays.
19 20 21
Abstract Cardiac troponin I (cTnI) is a marker for detection of myocardial damage in horses.
22
Many cTnI assays exist and medical studies have shown that the clinical performance of
23
assays differs. The aim of this study was to compare two different cTnI assays in horses.
24
Serum samples were taken from 23 healthy horses (group 1) and 72 horses with cardiac
25
disease (group 2). Cardiac troponin I was determined using assay 1 in laboratory A (limit of
26
detection, LOD, 0.03 ng/mL) and assay 2 in laboratories B and C (LOD 0.01 ng/mL). In
27
group 1, a median cTnI <0.03 (<0.03-0.04) ng/mL and <0.01 (<0.01-0.15) ng/mL was found
28
with assays 1 and 2, respectively. A higher median value was demonstrated in group 2 for
29
both assays (assay 1: 0.11 ng/mL, range 0.03-58.27 ng/mL, P <0.001; assay 2: 0.02 ng/mL,
30
range 0.01-22.87 ng/mL, P = 0.044). Although a significant correlation between assays
31
existed, large mean differences that could be important for clinical interpretation of test
32
results were found. A small mean difference was found between laboratories B and C. A
33
significant optimal (P <0.001) cut-off value for detection of cardiac disease could only be
34
determined for the assay 1 (0.035 ng/mL, sensitivity 70%, specificity 91%). Assay 1
35
performed better for detection of cardiac disease in horses in this study.
36 37
Keywords: Equine; Cardiac biomarker; Troponin T; Atypical myopathy; Valvular
38
regurgitation
2 Page 2 of 20
39
3 Page 3 of 20
40 41
Introduction Troponin I, T and C form a regulatory complex and are part of the contractile apparatus
42
of skeletal and cardiac muscle tissue (Wells and Sleeper, 2008). Specific cardiac isoforms
43
exist for troponin I and T. As a result, cardiac troponin I (cTnI) and T (cTnT) are used for
44
detection of acute myocardial injury in humans and have replaced the less specific lactate
45
dehydrogenase and creatine kinase myocardial band isoenzymes (Maynard et al., 2000). Both
46
cTnI and cTnT are part of the sarcomere component and bound to myofibrils; 6-8% of cTnI
47
and 2.8-4% of cTnT form a cytoplasmic component (Maynard et al., 2000).
48 49
Damage to cardiac myocytes causes primarily leakage of the cytosolic troponin pool.
50
With prolonged cellular stress, the structurally bound troponins are also released. This
51
explains the sustained high troponin concentrations in human patients with severe cardiac
52
disease, despite the short half-life (± 2 h; White, 2011). In human patients, the diagnostic
53
power of cTnI and cTnT are comparable (Maynard et al., 2000). However, because of the
54
wide variation in cTnI assays (Bock et al., 2008; Venge et al., 2009; La'ulu and Roberts,
55
2010; Reiter et al., 2011; Apple and Saenger, 2013) in human clinical practice, cTnT is often
56
the molecule of choice (Maynard et al., 2000).
57 58
In equine medicine, cTnI has been reported to be the best biomarker to detect
59
myocardial disease (Schwarzwald et al., 2003; Divers et al., 2009). Apart from a unique
60
equine 6-amino-acid N-terminal deletion, human and equine cTnI are largely similar. Since
61
this deletion lies outside the epitope region of most commercial assays, human cTnI assays
62
can be used for detection of myocardial damage in horses (Rishniw and Simpson, 2005). A
63
number of equine cTnI studies have established different reference values (Phillips et al.,
64
2003; Nostell and Hägström, 2008; Kraus et al., 2010; Nath et al., 2012; Slack et al., 2012;
4 Page 4 of 20
65
Buhl et al., 2013) and an equine specific half-life (± 47 min) has been determined (Kraus et
66
al., 2013).
67 68
An equine reference range of 0-0.35 ng/mL was reported in pastured and race-trained
69
Thoroughbred horses using an immunoassay (Dimension Heterogeneous Immunoassay
70
Module, Siemens) (Phillips et al., 2003). A lower reference range (<0.15 ng/mL) was found
71
with the ADVIA Centaur cTnI assay (Bayer Corporation) (Begg et al., 2006). In the most
72
recent study, a high sensitivity assay was used (ADVIA Centaur Assays TnI-Ultra, Bayer
73
Healthycare), in which all healthy horses showed a cTnI <0.03 ng/mL (Nath et al., 2012).
74 75
Cardiac troponin I assays have been thoroughly compared in human medicine (Lindahl
76
et al., 2013; Petersmann et al., 2013) and small animal veterinary medicine (Adin et al.,
77
2006). However, little is known about differences between cTnI assays in horses. Therefore,
78
the first aim of this study was to compare results of two different cTnI assays when used in
79
healthy horses and in horses with cardiac disease. A second aim was to determine a
80
significant cut-off value with these two cTnI assays for detection of cardiac disease in horses.
81 82
Materials and methods
83
Study population
84
The study was approved by the ethical committee of the Faculty of Veterinary Medicine
85
and Bioscience Engineering (approval number EC2012_57; date of approval 26 February
86
2013). The study population consisted of 23 healthy horses (group 1: age 8 ± 4 years, height
87
168 ± 5 cm, weight 567 ± 55 kg) and 72 horses with cardiac disease (group 2: age 11 ± 7
88
years, height 165 ± 10 cm, weight 503 ± 117 kg).
89
5 Page 5 of 20
90
Group 1 included 10 mares, 12 geldings and one stallion. These horses were trained 2-4
91
times a week for 30-60 min for dressage or show jumping. Their medical history was
92
determined and a thorough clinical examination was performed, along with echocardiography
93
and electrocardiography (ECG) at rest and during exercise (Televet 100, Engel Engineering
94
Services). A standardised lungeing exercise test was performed and consisted of 5 min at
95
walk, followed by 10 min at trot, 4 min at canter and 1 min at gallop (Verheyen et al., 2013).
96 97
Horses in group 2 (n = 72) were presented over a 2 year period at the Faculty of
98
Veterinary Medicine (Ghent University) and diagnosed with cardiac disease. Fifty-two horses
99
were diagnosed with cardiac disease based on echocardiography and electrocardiography. The
100
other 20 horses were diagnosed with atypical myopathy (AM) based on clinical signs, blood
101
examination and seasonality of the disease. Atypical myopathy is a seasonal, acute, frequently
102
fatal rhabdomyolysis that affects skeletal muscles and myocardium (Gehlen et al., 2006;
103
Cassart et al., 2007; Votion et al., 2007, 2014; Verheyen et al., 2012; van Galen and Votion,
104
2013; Unger et al., 2014).
105 106 107
Blood sampling and laboratory analysis A serum sample (10 mL) was collected from each horse before exercise using a
108
Vacutainer blood collection tube (Becton-Dickson) by puncture of the jugular vein. The tubes
109
were centrifuged 30 min after collection for 10 min at 2576 g and serum was transferred into
110
three cryovial tubes (2 mL, VWR International) and stored at -20 °C until analysis.
111 112
Three laboratories participated in sample analysis; for each laboratory, a separate
113
cryovial was used to avoid thawing-freezing cycles. Samples were transported on ice to each
114
laboratory; the sample transportation time was 15-90 min. In laboratory A, all samples (n =
6 Page 6 of 20
115
95) were analysed using the Access Accu cTnI assay (Beckman Coulter, lower limit of
116
detection, LOD, 0.03 ng/mL). Of these 95 samples, 67 were re-analysed with a STAT-I assay
117
(Abbott Architect, LOD 0.01 ng/mL) at laboratory B. Forty-three of these 67 samples were
118
analysed again with the STAT-I assay at laboratory C to study the effect of laboratories on the
119
test results.
120 121 122
Statistical analysis Samples with values less than the LOD were assigned the concentration of the limit of
123
detection so as not to underestimate the differences between groups (Dengler et al., 1998).
124
Thus, samples under the LOD were reported as 0.03 ng/mL for the Access Accu assay and
125
0.01 ng/mL for the STAT-I assay. Data were analysed using commercially available computer
126
software (SPSS version 22.0). Since data were not normally distributed, all results were
127
reported as medians, with minimum and maximal values, unless stated otherwise. A non-
128
parametric Mann-Whitney U test was used to compare cTnI values and ECG results in group
129
1 and 2. The cTnI values of healthy horses, horses with valvular heart disease and horses with
130
AM were compared using a Kruskal-Wallis rank test. The same test was used for comparison
131
of storage time in laboratories A, B and C.
132 133
A receiver operator characteristic (ROC) curve was established to determine the optimal
134
cut-off value for the Access Accu assay and the STAT-I assay (laboratory B). The two areas
135
under the ROC curve were compared with MedCalc software (version 13.2.0.0) (DeLong et
136
al., 1988). The Spearman correlation and the Lin’s concordance coefficient (Rc) were
137
calculated after log transformation. A further comparison between assays and between
138
laboratory B and C was done by Bland-Altman analysis.
139
7 Page 7 of 20
140
Results
141
Echocardiographic and electrocardiographic examination
142
All 23 horses in group 1 had normal cardiac dimensions and no or non-significant valve
143
regurgitation. Mild aortic regurgitation was present in one horse and mild mitral and tricuspid
144
regurgitation in another horse. The resting ECG did not show arrhythmias in 22 horses. One
145
healthy horse showed one atrial premature depolarisation (APD) at rest. A normal exercise
146
ECG was found in 21 horses. One horse had two and another three ventricular premature
147
depolarisations (VPDs) during lungeing.
148 149
In group 2, echocardiographic examination was performed in 52/72 horses; 49 horses
150
were diagnosed with structural heart disease; 42 horses had moderate or severe valvular
151
regurgitation, five had a ventricular septal defect (VSD) and two had an aortopulmonary
152
fistula. One of the horses with a VSD was also diagnosed with an abnormal pulmonary valve
153
and overriding aorta. The remaining three horses had suspected myocardial disease based on a
154
high number of APDs (n = 1) or VPDs (n = 2) at rest.
155 156
An exercise ECG test was performed in 36/52 horses with cardiac disease. The median
157
number of VPDs in these group 2 horses was significantly higher (0 per 30 min, range 0-400
158
per 30 min) compared to group 1 horses (0 per 30 min, 0-2 per 30 min; P = 0.01). The median
159
number of APDs was also higher in group 2 (0 per 30 min, range 0-5 per 30 min) compared to
160
group 1 (0 per 30 min, range 0-1 per 30 min; P = 0.030).
161 162
The remaining 20/72 horses were diagnosed with AM on the basis of clinical signs,
163
blood examination (creatine kinase mean ± SD 97,000 ± 84,000 IU/L; lactate dehydrogenase
164
mean ± SD 57,000 ± 39,000 IU/L), seasonality of the disease (outbreak in autumn 2013) and
8 Page 8 of 20
165
contact with the toxic seeds of the sycamore tree (Acer pseudoplatanus). A systolic murmur
166
was heard in three horses and one horse had a diastolic murmur.
167 168 169
Troponin analysis Access Accu assay (laboratory A) - All samples were analysed within 6 months after
170
collection. The median storage time until analysis was 2 days (range 0-180 days) and 69% of
171
all samples of laboratory A were analysed within 1 week. All horses in group 1 had cTnI
172
values ≤ 0.04 ng/mL (<0.03-0.04 ng/mL). cTnI values were significantly higher in group 2
173
(median 0.11 ng/mL, range <0.03-58.27 ng/mL) than in group 1 (median <0.03 ng/mL, range
174
<0.03-0.04 ng/mL; P <0.001). The optimal cut-off value for detection of cardiac disease was
175
0.035 ng/mL (Fig. 1; Table 1). Based on clinical diagnosis, horses in group 2 (n = 72) could
176
be divided into two subgroups: horses with cardiomyopathy (AM or myocarditis, n = 23) and
177
horses with moderate or severe valvular regurgitation (n = 42); horses with aortopulmonary
178
fistula (n = 2) and VSD (n = 5) were not included in these subgroups. cTnI values in horses
179
with cardiomyopathy (median 0.90 ng/mL, range <0.03-58.27 ng/mL) was significantly
180
higher than in horses with valvular regurgitation (median 0.05 ng/mL, range <0.03-30.92
181
ng/mL; P = 0.001) and healthy horses (median <0.03 ng/mL, range <0.03-0.04 ng/mL; P
182
<0.001). Horses with valvular regurgitation also showed significantly higher cTnI values than
183
healthy horses (P = 0.001; Fig. 2).
184 185
STAT-I assay (laboratories B and C) - As for laboratory A, the median sample storage
186
time was 2 days in laboratories B (range 0-34 days) and C (range 0-38 days); 91% of samples
187
in laboratory B and 86% of samples in laboratory C were analysed within 1 week. Group 1
188
had a median cTnI concentration of <0.01 ng/mL (range <0.01-0.15 ng/mL; n = 22) in
189
laboratory B; the cTnI concentration in group 2 (n = 45) was significantly higher (median
9 Page 9 of 20
190
0.02 ng/mL, range <0.01-22.87 ng/mL; P = 0.044). A cut-off value of 0.015 was determined
191
from the ROC curve (Fig. 1; Table 1); however, the sensitivity was low and the P value was
192
>0.05, which indicates that the assay did not distinguish between groups 1 and 2. No
193
significant difference was found between cTnI in group 1 (median <0.01 ng/mL, range <0.01-
194
0.13 ng/mL) and group 2 (median <0.01 ng/mL, range <0.01-12.29 ng/mL; P > 0.05) when
195
samples were analysed in laboratory C (n = 43).
196 197
Comparison between assays - No significant difference in storage time was found
198
between assays. As shown in Table 2, a significant association was demonstrated between the
199
two cTnI assays (P <0.01). A moderate agreement was found between the Access Accu and
200
the STAT-I assay from laboratory B (Rc 0.543; bias correction, Cb, 0.837; shift 0.607) and C
201
(Rc 0.496; Cb 0.752; shift 0.781). A substantial agreement was found between the STAT-I
202
assay from laboratories B and C (Rc 0.617; Cb 0.943; shift 0.320). Despite these correlations,
203
individual troponin values differed substantially and a significant difference between the areas
204
under the ROC curve was found (P <0.001). The limits of agreement (Figs. 3a-c) and the
205
mean difference were large between the two assays. The mean ± SD difference between the
206
Access Accu and the STAT-I assay was 1.18 ± 4.08 ng/mL (laboratory A vs. laboratory B)
207
and 0.92 ± 3.14 ng/mL (laboratory A vs. laboratory C). Figs. 3a and 3b demonstrate that this
208
difference becomes larger as the magnitude of the measurement increased. As expected, the
209
mean ± SD difference (0.01 ± 0.27 ng/mL) was smaller when samples were analysed with the
210
same assay in two different laboratories.
211
Discussion
212
This study demonstrates that a large quantitative difference between two cTnI assays
213
exists in horses. Cardiac troponin I reference values differed between assays and a cut-off
214
value for cardiac disease could only be determined for the Access Accu assay. A
10 Page 10 of 20
215
quantitatively smaller difference was found when results from the STAT-I assay were
216
compared between laboratories.
217 218
In agreement with previous results (Jesty et al., 2009), the median cTnI value in group
219
1 was <0.03 ng/mL with the Access Accu assay. When samples were re-analysed with the
220
STAT-I assay, median cTnI values of <0.01 ng/mL in laboratory B and <0.01 ng/mL in
221
laboratory C were obtained. A high cTnI value (>0.10 ng/mL) was found in three healthy
222
horses with the STAT-I assay in laboratory B. In a study on human beings comparing the
223
same two assays, false positive results were found for the STAT-I assay due to the presence
224
of heterophile antibodies. These are naturally occurring antibodies which have a low affinity
225
and specificity that are produced against poorly defined antigens and can bind between the
226
capture and detection antibody of an assay (Kaplan and Levinson, 1999). The presence of
227
these antibodies has previously been suggested in an equine cTnI study (Slack et al., 2012)
228
and could also explain the conflicting results in our study.
229 230
Despite the significant correlation between the Access Accu and STAT-I assay (Table
231
2), a large mean difference was found (Figs. 3a and 3b). Since no gold standard was available,
232
it could not be concluded that one cTnI assay was better than another. However, an optimal
233
cut-off value was only significant for the Access Accu assay. Therefore, the Access Accu
234
assay might perform better than the STAT-I assay in a clinical setting.
235 236
Many explanations exist for the differences between these assays. The Access Accu
237
assay consists of two antibodies directed against the stable NH2-region-terminus of the cTnI
238
molecule (epitopes at amino acids 24-40 and 41-49) (James et al., 2006). An additional
239
capture antibody directed against the 87-91 amino acid region is used by the STAT-I assay.
11 Page 11 of 20
240
Since these assays recognise different target amino acids of the cTnI molecule, troponin
241
values might differ. In addition, cardiac troponins can circulate as various complexes (Wu et
242
al., 1998), troponin modification products after myocardial injury exist and multiple
243
degradation products have been described in human beings. Thus, the detectability of all these
244
products and fragments might also vary according to the assay used (Labugger et al., 2000).
245 246
In our study, cTnI was measured in horses with several different diseases; 49 horses had
247
structural cardiac disease (moderate or severe valvular regurgitation, aortopulmonary fistula
248
and VSD) and 23 horses had cardiomyopathy. Significantly higher cTnI values were found
249
with the Access Accu assay in horses with cardiomyopathy than in horses with valvular
250
regurgitation and in horses with valvular regurgitation compared to healthy horses. Similar
251
findings have been reported by Nath et al. (2012) and demonstrate that the cTnI assay not
252
only detects severe myocardial damage in case of cardiomyopathy, but is also able to
253
demonstrate mild cTnI increase in case of moderate or severe valvular disease.
254 255
When results of the STAT-I assay analysed at two different laboratories were compared,
256
a smaller mean difference (0.01 ng/mL) was found than when two different assays were
257
compared. However, a large standard deviation of the mean difference (0.27 ng/mL) still
258
existed suggesting that laboratory differences should be taken into account.
259 260
As found in medical studies, it is difficult to compare troponin I assays; however, it is
261
important to be aware of assay differences, to become familiar with one assay, to develop a
262
reference range from a large number of horses, to use the same laboratory and to learn by
263
experience which horses are most likely to be positive for cardiac disease (Apple and Saenger,
264
2013). Cardiac troponin T assay differences are limited, since only one manufacturer has
12 Page 12 of 20
265
marketed a cTnT assay. The specificity of the cTnT assay has been questioned in the case of
266
severe skeletal muscle injury (Bodor et al., 1997) and kidney disease (McLaurin et al., 1997).
267
However, with newer cTnT assays, false positive results seem to be limited to primary
268
skeletal muscle disease (Jaffe et al., 2011; Apple and Collinson, 2012). Therefore, similarly to
269
human medicine, cTnT might be an alternative for future troponin measurement in horses
270
(Thygesen et al., 2010).
271 272
A limited number (n = 23) of healthy horses was used to determine the reference range
273
for individual assays. Furthermore, our study also did not determine the coefficient of
274
variation for these assays (Thygesen et al., 2007). More assay-specific research is needed in a
275
larger equine population. Not all 95 samples could be re-analysed and a lower number of
276
samples were analysed with the STAT-I assay. However, if only samples (n = 67) analysed
277
with both assays were considered, the area under the ROC-curve was still higher for the
278
Access Accu assay (0.837, P <0.001) than for the STAT-I assay (0.643, P = 0.061).
279 280
Echocardiographic and electrocardiographic examination was not performed in all
281
horses with AM. However, the presence of cardiac disease was expected, since cTnI is not
282
present in skeletal muscle (Apple, 1999; Aggarwal et al., 2009) and because ultrasound, ECG
283
and post-mortem findings from previous studies have shown that AM is associated with
284
cardiomyopathy (Cassart et al., 2007; Votion et al., 2007; Verheyen et al., 2012; van Galen
285
and Votion, 2013; Unger et al., 2014). When horses with AM were not included, results from
286
the Access Accu assay still showed a significant difference between healthy horses and horses
287
with cardiac disease.
288
13 Page 13 of 20
289
Finally, samples were not analysed at the same time for the two assays. No significant
290
difference in storage time was found between assays, or between laboratories, and most of the
291
samples were analysed within the same week. The longest storage time (180 days; n = 2) was
292
found for the Access Accu assay, where samples had the highest cTnI concentrations. Since
293
samples were stored at -20 °C according to manufacturer’s instructions and because it is
294
known that cTnI or cTnT are stable for at least 1 year in serum samples stored at -70°C (Basit
295
et al., 2007), an influence of sample storage on results is unlikely.
296 297 298
Conclusions Troponin I assays cannot be directly compared to each other, since clinically important
299
differences exist. Assay specific reference values should be used and laboratory differences
300
must be considered. Cardiac troponin I measurement not only differentiates healthy horses
301
from horses with cardiomyopathy in a population but also healthy horses from horses with
302
moderate to severe valvular regurgitation.
303 304
Conflict of interest statement
305
None of the authors of this paper have a financial or personal relationship with other
306
people or organisations that could inappropriately influence or bias the content of the paper.
307 308
Acknowledgements
309
The authors wish to acknowledge the Special Research Fund (BOF) of Ghent
310
University for the funding of the study. We also acknowledge Marta Garcia-Granero for her
311
contribution to the study.
312 313
References
14 Page 14 of 20
314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363
Adin, D.B., Oyama, M.A., Sleeper, M.M., Milner, R.J., 2006. Comparison of canine cardiac troponin I concentrations as determined by 3 different analyzers. Journal of Veterinary Internal Medicine 20, 1136-1142. Aggarwal, R., Lebiedz-Odrobina, D., Sinha, A., Manadan, A., Case, J.P., 2009. Serum cardiac troponin T, but not troponin I, is elevated in idiopathic inflammatory myopathies. Journal of Rheumatology 36, 2711-2714. Apple, F.S., 1999. Clinical and analytical standardization issues confronting cardiac troponin I. Clinical Chemistry 45, 18-20. Apple, F.S., Collinson, P.O., 2012. Analytical characteristics of high-sensitivity cardiac troponin assays. Clinical Chemistry 58, 54-61. Apple, F.S., Saenger, A.K., 2013. The state of cardiac troponin assays: Looking bright and moving in the right direction. Clinical Chemistry 59, 1014-1016. Basit, M., Bakshi, N., Hashem, M., Allebban, Z., Lawson, N., Rosman, H.S., Maciejko, J.J., 2007. The effect of freezing and long-term storage on the stability of cardiac troponin T. American Journal of Clinical Pathology 128, 164-167. Begg, L.M., Hoffmann, K.L., Begg, A.P., 2006. Serum and plasma cardiac troponin I concentrations in clinically normal Thoroughbreds in training in Australia. Australian Veterinary Journal 84, 336-337. Bock, J.L., Singer, A.J., Thode, H.C., Jr., 2008. Comparison of emergency department patient classification by point-of-care and central laboratory methods for cardiac troponin I. American Journal of Clinical Pathology 130, 132-135. Bodor, G.S., Survant, L., Voss, E.M., Smith, S., Porterfield, D., Apple, F.S., 1997. Cardiac troponin T composition in normal and regenerating human skeletal muscle. Clinical Chemistry 43, 476-484. Buhl, R., Petersen, E.E., Lindholm, M., Bak, L., Nostell, K., 2013. Cardiac arrhythmias in Standardbreds during and after racing-possible association between heart size, valvular regurgitations, and arrhythmias. Journal of Equine Veterinary Science 33, 590-596. Cassart, D., Baise, E., Cherel, Y., Delguste, C., Antoine, N., Votion, D., Amory, H., Rollin, F., Linden, A., Coignoul, F., Desmecht, D., 2007. Morphological alterations in oxidative muscles and mitochondrial structure associated with equine atypical myopathy. Equine Veterinary Journal 39, 26-32. DeLong, E.R., DeLong, D.M., Clarke-Pearson, D.L., 1988. Comparing the areas under two or more correlated receiver operating characteristic curves: A nonparametric approach. Biometrics 44, 837-845. Divers, T.J., Kraus, M.S., Jesty, S.A., Miller, A.D., Mohammed, H.O., Gelzer, A.R.M., Mitchell, L.M., Soderholm, L.V., Ducharme, N.G., 2009. Clinical findings and serum cardiac troponin I concentrations in horses after intragastric administration of sodium monensin. Journal of Veterinary Diagnostic Investigation 21, 338-343. 15 Page 15 of 20
364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413
Gehlen, H., Rohn, K., Deegen, E., Stadler, P., 2006. Labordiagnostische untersuchungen bei pferden mit herzerkrankungen: Wertigkeit verschiedener kardialer marker. Pferdeheilkunde 22, 532-541. Jaffe, A.S., Vasile, V.C., Milone, M., Saenger, A.K., Olson, K.N., Apple, F.S., 2011. Diseased skeletal muscle: A noncardiac source of increased circulating concentrations of cardiac troponin T. Journal of the American College of Cardiology. 58, 1819-1824. James, S., Flodin, M., Johnston, N., Lindahl, B., Venge, P., 2006. The antibody configurations of cardiac troponin I assays may determine their clinical performance. Clinical Chemistry 52, 832-837. Jesty, S.A., Kraus, M., Gelzer, A., Rishniw, M., Moïse, N.S., 2009. Effect of transvenous electrical cardioversion on plasma cardiac troponin I concentrations in horses with atrial fibrillation. Journal of Veterinary Internal Medicine 23, 1103-1107. Kaplan, I.V., Levinson, S.S., 1999. When is a heterophile antibody not a heterophile antibody? When it is an antibody against a specific immunogen. Clinical Chemistry 45, 616-618. Kraus, M.S., Jesty, S.A., Gelzer, A.R., Ducharme, N.G., Mohammed, H.O., Mitchell, L.M., Soderholm, L.V., Divers, T.J., 2010. Measurement of plasma cardiac troponin I concentration by use of a point-of-care analyzer in clinically normal horses and horses with experimentally induced cardiac disease. American Journal of Veterinary Research 71, 55-59. Kraus, M.S., Kaufer, B.B., Damiani, A., Osterrieder, N., Rishniw, M., Schwark, W., Gelzer, A.R., Divers, T.J., 2013. Elimination half-life of intravenously administered equine cardiac troponin I in healthy ponies. Equine Veterinary Journal 45, 56-59. La'ulu, S.L., Roberts, W.L., 2010. Performance characteristics of five cardiac Troponin I assays. Clinica Chimica Acta 411, 1095-1101. Labugger, R., Organ, L., Collier, C., Atar, D., Van Eyk, J.E., 2000. Extensive troponin I and T modification detected in serum from patients with acute myocardial infarction. Circulation 102, 1221-1226. Lindahl, B., Eggers, K.M., Venge, P., James, S., 2013. Evaluation of four sensitive troponin assays for risk assessment in acute coronary syndromes using a new clinically oriented approach for comparison of assays. Clinical Chemistry Laboratory Medicine 51, 18591864. Maynard, S.J., Menown, I.B.A., Adgey, A.A.J., 2000. Troponin T or troponin I as cardiac markers in ischaemic heart disease. Heart 83, 371-373. McLaurin, M.D., Apple, F.S., Voss, E.M., Herzog, C.A., Sharkey, S.W., 1997. Cardiac troponin I, cardiac troponin T, and creatine kinase MB in dialysis patients without ischemic heart disease: Evidence of cardiac troponin T expression in skeletal muscle. Clinical Chemistry 43, 976-982. 16 Page 16 of 20
414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463
Nath, L., Anderson, G., Hinchcliff, K., Savage, C., 2012. Serum cardiac troponin I concentrations in horses with cardiac disease. Australian Veterinary Journal 90, 351357. Nostell, K., Hägström, J., 2008. Resting concentrations of cardiac troponin I in fit horses and effect of racing. Journal of Veterinary Cardiology 10, 105-109. Petersmann, A., Ittermann, T., Fries, C., Lubenow, N., Kohlmann, T., Kallner, A., Greinacher, A., Nauck, M., 2013. Comparison of the 99th percentiles of three troponin I assays in a large reference population. Clinical Chemistry and Laboratory Medicine 51, 2181-2186. Phillips, W., Giguère, S., Franklin, R.P., Hernandez, J., Adin, D., Peloso, J., 2003. Cardiac troponin I in pastured and race-training Thoroughbred horses. Journal of Veterinary Internal Medicine 17, 597-599. Reiter, M., Twerenbold, R., Reichlin, T., Haaf, P., Peter, F., Meissner, J., Hochholzer, W., Stelzig, C., Freese, M., Heinisch, C., et al., 2011. Early diagnosis of acute myocardial infarction in the elderly using more sensitive cardiac troponin assays. European Heart Journal 32, 1379-1389. Rishniw, M., Simpson, K.W., 2005. Cloning and sequencing of equine cardiac troponin I and confirmation of its usefulness as a target analyte for commercial troponin I analyzers. Journal of Veterinary Diagnostic Investigation 17, 582-584. Schwarzwald, C.C., Hardy, J., Buccellato, M., 2003. High cardiac troponin I serum concentration in a horse with multiform ventricular tachycardia and myocardial necrosis. Journal of Veterinary Internal Medicine 17, 364-368. Slack, J., Boston, R.C., Soma, L., Reef, V.B., 2012. Cardiac troponin I in racing standardbreds. Journal of Veterinary Internal Medicine 26, 1202-1208. Thygesen, K., Alpert, J.S., White, H.D., 2007. Universal definition of myocardial infarction. Journal of the American College of Cardiology 50, 2173-2195. Thygesen, K., Mair, J., Katus, H., Plebani, M., Venge, P., Collinson, P., Lindahl, B., Giannitsis, E., Hasin, Y., Galvani, M., et al., , 2010. Recommendations for the use of cardiac troponin measurement in acute cardiac care. European Heart Journal 31, 21972204. Unger, L., Nicholson, A., Jewitt, E.M., Gerber, V., Hegeman, A., Sweetman, L., Valberg, S., 2014. Hypoglycin A concentrations in seeds of Acer pseudoplatanus trees growing on atypical myopathy-affected and control pastures. Journal of Veterinary Internal Medicine 28, 1289-1293. van Galen, G., Votion, D.M., 2013. Management of cases suffering from atypical myopathy: Interpretations of descriptive, epidemiological and pathophysiological findings. Part 2: Muscular, urinary, respiratory and hepatic care, and inflammatory/infectious status. Equine Veterinary Education 25, 308-314. 17 Page 17 of 20
464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495
Venge, P., James, S., Jansson, L., Lindahl, B., 2009. Clinical performance of two highly sensitive cardiac troponin I assays. Clinical Chemistry 55, 109-116. Verheyen, T., Decloedt, A., De Clercq, D., van Loon, G., 2012. Cardiac changes in horses with atypical myopathy. Journal of Veterinary Internal Medicine 26, 1019-1026. Verheyen, T., Decloedt, A., Van Der Vekens, N, Sys, S., De Clercq, D., van Loon, G., 2013. Ventricular response during lungeing exercise in horses with lone atrial fibrillation. Equine Veterinary Journal 45, 309-314. Votion, D.M., Linden, A., Saegerman, C., Engels, P., Erpicum, M., Thiry, E., Delguste, C., Rouxhet, S., Demoulin, V., Navet, R., et al., 2007. History and clinical features of atypical myopathy in horses in Belgium (2000-2005). Journal of Veterinary Internal Medicine 21, 1380-1391. Votion, D.M., van Galen, G., Sweetman, L., Boemer, F., de Tullio, P., Dopagne, C., Lefere, L., Mouithys-Mickalad, A., Patarin, F., Rouxhet, S., et al., 2014. Identification of methylenecyclopropyl acetic acid in serum of European horses with atypical myopathy. Equine Veterinary Journal 46, 146-149. Wells, S.M., Sleeper, M., 2008. Cardiac troponins. Journal of Veterinary Emergency and Critical Care 18, 235-245. White, H.D., 2011. Pathobiology of troponin elevations: Do elevations occur with myocardial ischemia as well as necrosis? Journal of the American College of Cardiology 57, 2406-2408. Wu, A.H.B., Feng, Y.J., Moore, R., Apple, F.S., McPherson, P.H., Buechler, K.F., Bodor, G., Standar, A.A.C.C.S.C., 1998. Characterization of cardiac troponin subunit release into serum after acute myocardial infarction and comparison of assays for troponin T and I. Clinical Chemistry 44, 1198-1208.
18 Page 18 of 20
496
Figure legends
497 498
Fig. 1. Receiver operator characteristic (ROC) curve of the Access Accu (laboratory A) and
499
STAT-I (laboratory B) assay for detection of cardiac disease ( cut-off of the Access Accu
500
assay).
501 502
Fig. 2. Box plot comparing troponin I concentration measured with the Access Accu in
503
healthy horses (group 1), horses with moderate or severe valvular regurgitation (group 2a) and
504
horses with cardiomyopathy (group 2b).
505 506
Figs. 3a-c. Bland-Altman assays comparing two different troponin I assays and measurement
507
of the STAT-I assay in two different laboratories (B and C). The mean difference between
508
assays increases with a higher troponin value (group 1: healthy horses, group 2: horses with
509
cardiac disease).
510 511
Table 1. Troponin I sensitivity and specificity to detect cardiac disease for troponin I.
512
Method
Laboratory
n
Access Accu
A
95
0.035
70%
STAT-I
B
67
0.015
STAT-I
C
43
0.015
513
Cut-off Sensitivity Specificity
P value
AUC
95% CI
91%
<0.001
0.83
0.70-0.97
55%
68%
0.11
0.64
0.47-0.81
30%
86%
0.30
0.59
0.42-0.77
AUC, area under the curve; 95% CI, 95% confidence interval.
19 Page 19 of 20
514
Table 2 Spearman correlation coefficient between two troponin assays for cardiac troponin I
515
determination in horses.
516
Access Accu Laboratory A
STAT-I Laboratory B
STAT-I Laboratory C
Access Accu Laboratory A
STAT-I Laboratory B
STAT-I Laboratory C
Correlation coefficient
1.000
X
X
n
95
Correlation coefficient
0.454
1.000
X
n
67
67
Correlation Coefficient
0.475
0.412
1.000
n
43
42
43
517 518 519
* Significant correlation at P <0.01.
520
20 Page 20 of 20
S1090-0233(14)00492-4 http://dx.doi.org/doi: 10.1016/j.tvjl.2014.11.015 YTVJL 4344
To appear in:
The Veterinary Journal
Accepted date:
27-11-2014
Please cite this article as: N. Van Der Vekens, A. Decloedt, S. Sys, S. Ven, D. De Clercq, G. van Loon, Evaluation of assays for troponin I in healthy horses and horses with cardiac disease, The Veterinary Journal (2014), http://dx.doi.org/doi: 10.1016/j.tvjl.2014.11.015. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
1 2 3 4 5 6 7 8 9 10 11 12 13
Evaluation of assays for troponin I in healthy horses and horses with cardiac disease
N. Van Der Vekens *, A. Decloedt, S. Sys, S. Ven, D. De Clercq, G. van Loon Department of Large Animal Internal Medicine, Faculty of Veterinary Medicine, Ghent, Salisburylaan 133, B-9820 Merelbeke, Belgium
* Corresponding author. Tel.: +32 9 2647584. E-mail address: [email protected] (N. Van Der Vekens).
1 Page 1 of 20
Highlights
14 15
Two troponin I assays were compared in horses, including 23 healthy horses and 72 horses with cardiac disease.
16 17
Large quantitative differences between assays were demonstrated.
18
A cut-off value for cardiac disease was identified for only one of the assays.
19 20 21
Abstract Cardiac troponin I (cTnI) is a marker for detection of myocardial damage in horses.
22
Many cTnI assays exist and medical studies have shown that the clinical performance of
23
assays differs. The aim of this study was to compare two different cTnI assays in horses.
24
Serum samples were taken from 23 healthy horses (group 1) and 72 horses with cardiac
25
disease (group 2). Cardiac troponin I was determined using assay 1 in laboratory A (limit of
26
detection, LOD, 0.03 ng/mL) and assay 2 in laboratories B and C (LOD 0.01 ng/mL). In
27
group 1, a median cTnI <0.03 (<0.03-0.04) ng/mL and <0.01 (<0.01-0.15) ng/mL was found
28
with assays 1 and 2, respectively. A higher median value was demonstrated in group 2 for
29
both assays (assay 1: 0.11 ng/mL, range 0.03-58.27 ng/mL, P <0.001; assay 2: 0.02 ng/mL,
30
range 0.01-22.87 ng/mL, P = 0.044). Although a significant correlation between assays
31
existed, large mean differences that could be important for clinical interpretation of test
32
results were found. A small mean difference was found between laboratories B and C. A
33
significant optimal (P <0.001) cut-off value for detection of cardiac disease could only be
34
determined for the assay 1 (0.035 ng/mL, sensitivity 70%, specificity 91%). Assay 1
35
performed better for detection of cardiac disease in horses in this study.
36 37
Keywords: Equine; Cardiac biomarker; Troponin T; Atypical myopathy; Valvular
38
regurgitation
2 Page 2 of 20
39
3 Page 3 of 20
40 41
Introduction Troponin I, T and C form a regulatory complex and are part of the contractile apparatus
42
of skeletal and cardiac muscle tissue (Wells and Sleeper, 2008). Specific cardiac isoforms
43
exist for troponin I and T. As a result, cardiac troponin I (cTnI) and T (cTnT) are used for
44
detection of acute myocardial injury in humans and have replaced the less specific lactate
45
dehydrogenase and creatine kinase myocardial band isoenzymes (Maynard et al., 2000). Both
46
cTnI and cTnT are part of the sarcomere component and bound to myofibrils; 6-8% of cTnI
47
and 2.8-4% of cTnT form a cytoplasmic component (Maynard et al., 2000).
48 49
Damage to cardiac myocytes causes primarily leakage of the cytosolic troponin pool.
50
With prolonged cellular stress, the structurally bound troponins are also released. This
51
explains the sustained high troponin concentrations in human patients with severe cardiac
52
disease, despite the short half-life (± 2 h; White, 2011). In human patients, the diagnostic
53
power of cTnI and cTnT are comparable (Maynard et al., 2000). However, because of the
54
wide variation in cTnI assays (Bock et al., 2008; Venge et al., 2009; La'ulu and Roberts,
55
2010; Reiter et al., 2011; Apple and Saenger, 2013) in human clinical practice, cTnT is often
56
the molecule of choice (Maynard et al., 2000).
57 58
In equine medicine, cTnI has been reported to be the best biomarker to detect
59
myocardial disease (Schwarzwald et al., 2003; Divers et al., 2009). Apart from a unique
60
equine 6-amino-acid N-terminal deletion, human and equine cTnI are largely similar. Since
61
this deletion lies outside the epitope region of most commercial assays, human cTnI assays
62
can be used for detection of myocardial damage in horses (Rishniw and Simpson, 2005). A
63
number of equine cTnI studies have established different reference values (Phillips et al.,
64
2003; Nostell and Hägström, 2008; Kraus et al., 2010; Nath et al., 2012; Slack et al., 2012;
4 Page 4 of 20
65
Buhl et al., 2013) and an equine specific half-life (± 47 min) has been determined (Kraus et
66
al., 2013).
67 68
An equine reference range of 0-0.35 ng/mL was reported in pastured and race-trained
69
Thoroughbred horses using an immunoassay (Dimension Heterogeneous Immunoassay
70
Module, Siemens) (Phillips et al., 2003). A lower reference range (<0.15 ng/mL) was found
71
with the ADVIA Centaur cTnI assay (Bayer Corporation) (Begg et al., 2006). In the most
72
recent study, a high sensitivity assay was used (ADVIA Centaur Assays TnI-Ultra, Bayer
73
Healthycare), in which all healthy horses showed a cTnI <0.03 ng/mL (Nath et al., 2012).
74 75
Cardiac troponin I assays have been thoroughly compared in human medicine (Lindahl
76
et al., 2013; Petersmann et al., 2013) and small animal veterinary medicine (Adin et al.,
77
2006). However, little is known about differences between cTnI assays in horses. Therefore,
78
the first aim of this study was to compare results of two different cTnI assays when used in
79
healthy horses and in horses with cardiac disease. A second aim was to determine a
80
significant cut-off value with these two cTnI assays for detection of cardiac disease in horses.
81 82
Materials and methods
83
Study population
84
The study was approved by the ethical committee of the Faculty of Veterinary Medicine
85
and Bioscience Engineering (approval number EC2012_57; date of approval 26 February
86
2013). The study population consisted of 23 healthy horses (group 1: age 8 ± 4 years, height
87
168 ± 5 cm, weight 567 ± 55 kg) and 72 horses with cardiac disease (group 2: age 11 ± 7
88
years, height 165 ± 10 cm, weight 503 ± 117 kg).
89
5 Page 5 of 20
90
Group 1 included 10 mares, 12 geldings and one stallion. These horses were trained 2-4
91
times a week for 30-60 min for dressage or show jumping. Their medical history was
92
determined and a thorough clinical examination was performed, along with echocardiography
93
and electrocardiography (ECG) at rest and during exercise (Televet 100, Engel Engineering
94
Services). A standardised lungeing exercise test was performed and consisted of 5 min at
95
walk, followed by 10 min at trot, 4 min at canter and 1 min at gallop (Verheyen et al., 2013).
96 97
Horses in group 2 (n = 72) were presented over a 2 year period at the Faculty of
98
Veterinary Medicine (Ghent University) and diagnosed with cardiac disease. Fifty-two horses
99
were diagnosed with cardiac disease based on echocardiography and electrocardiography. The
100
other 20 horses were diagnosed with atypical myopathy (AM) based on clinical signs, blood
101
examination and seasonality of the disease. Atypical myopathy is a seasonal, acute, frequently
102
fatal rhabdomyolysis that affects skeletal muscles and myocardium (Gehlen et al., 2006;
103
Cassart et al., 2007; Votion et al., 2007, 2014; Verheyen et al., 2012; van Galen and Votion,
104
2013; Unger et al., 2014).
105 106 107
Blood sampling and laboratory analysis A serum sample (10 mL) was collected from each horse before exercise using a
108
Vacutainer blood collection tube (Becton-Dickson) by puncture of the jugular vein. The tubes
109
were centrifuged 30 min after collection for 10 min at 2576 g and serum was transferred into
110
three cryovial tubes (2 mL, VWR International) and stored at -20 °C until analysis.
111 112
Three laboratories participated in sample analysis; for each laboratory, a separate
113
cryovial was used to avoid thawing-freezing cycles. Samples were transported on ice to each
114
laboratory; the sample transportation time was 15-90 min. In laboratory A, all samples (n =
6 Page 6 of 20
115
95) were analysed using the Access Accu cTnI assay (Beckman Coulter, lower limit of
116
detection, LOD, 0.03 ng/mL). Of these 95 samples, 67 were re-analysed with a STAT-I assay
117
(Abbott Architect, LOD 0.01 ng/mL) at laboratory B. Forty-three of these 67 samples were
118
analysed again with the STAT-I assay at laboratory C to study the effect of laboratories on the
119
test results.
120 121 122
Statistical analysis Samples with values less than the LOD were assigned the concentration of the limit of
123
detection so as not to underestimate the differences between groups (Dengler et al., 1998).
124
Thus, samples under the LOD were reported as 0.03 ng/mL for the Access Accu assay and
125
0.01 ng/mL for the STAT-I assay. Data were analysed using commercially available computer
126
software (SPSS version 22.0). Since data were not normally distributed, all results were
127
reported as medians, with minimum and maximal values, unless stated otherwise. A non-
128
parametric Mann-Whitney U test was used to compare cTnI values and ECG results in group
129
1 and 2. The cTnI values of healthy horses, horses with valvular heart disease and horses with
130
AM were compared using a Kruskal-Wallis rank test. The same test was used for comparison
131
of storage time in laboratories A, B and C.
132 133
A receiver operator characteristic (ROC) curve was established to determine the optimal
134
cut-off value for the Access Accu assay and the STAT-I assay (laboratory B). The two areas
135
under the ROC curve were compared with MedCalc software (version 13.2.0.0) (DeLong et
136
al., 1988). The Spearman correlation and the Lin’s concordance coefficient (Rc) were
137
calculated after log transformation. A further comparison between assays and between
138
laboratory B and C was done by Bland-Altman analysis.
139
7 Page 7 of 20
140
Results
141
Echocardiographic and electrocardiographic examination
142
All 23 horses in group 1 had normal cardiac dimensions and no or non-significant valve
143
regurgitation. Mild aortic regurgitation was present in one horse and mild mitral and tricuspid
144
regurgitation in another horse. The resting ECG did not show arrhythmias in 22 horses. One
145
healthy horse showed one atrial premature depolarisation (APD) at rest. A normal exercise
146
ECG was found in 21 horses. One horse had two and another three ventricular premature
147
depolarisations (VPDs) during lungeing.
148 149
In group 2, echocardiographic examination was performed in 52/72 horses; 49 horses
150
were diagnosed with structural heart disease; 42 horses had moderate or severe valvular
151
regurgitation, five had a ventricular septal defect (VSD) and two had an aortopulmonary
152
fistula. One of the horses with a VSD was also diagnosed with an abnormal pulmonary valve
153
and overriding aorta. The remaining three horses had suspected myocardial disease based on a
154
high number of APDs (n = 1) or VPDs (n = 2) at rest.
155 156
An exercise ECG test was performed in 36/52 horses with cardiac disease. The median
157
number of VPDs in these group 2 horses was significantly higher (0 per 30 min, range 0-400
158
per 30 min) compared to group 1 horses (0 per 30 min, 0-2 per 30 min; P = 0.01). The median
159
number of APDs was also higher in group 2 (0 per 30 min, range 0-5 per 30 min) compared to
160
group 1 (0 per 30 min, range 0-1 per 30 min; P = 0.030).
161 162
The remaining 20/72 horses were diagnosed with AM on the basis of clinical signs,
163
blood examination (creatine kinase mean ± SD 97,000 ± 84,000 IU/L; lactate dehydrogenase
164
mean ± SD 57,000 ± 39,000 IU/L), seasonality of the disease (outbreak in autumn 2013) and
8 Page 8 of 20
165
contact with the toxic seeds of the sycamore tree (Acer pseudoplatanus). A systolic murmur
166
was heard in three horses and one horse had a diastolic murmur.
167 168 169
Troponin analysis Access Accu assay (laboratory A) - All samples were analysed within 6 months after
170
collection. The median storage time until analysis was 2 days (range 0-180 days) and 69% of
171
all samples of laboratory A were analysed within 1 week. All horses in group 1 had cTnI
172
values ≤ 0.04 ng/mL (<0.03-0.04 ng/mL). cTnI values were significantly higher in group 2
173
(median 0.11 ng/mL, range <0.03-58.27 ng/mL) than in group 1 (median <0.03 ng/mL, range
174
<0.03-0.04 ng/mL; P <0.001). The optimal cut-off value for detection of cardiac disease was
175
0.035 ng/mL (Fig. 1; Table 1). Based on clinical diagnosis, horses in group 2 (n = 72) could
176
be divided into two subgroups: horses with cardiomyopathy (AM or myocarditis, n = 23) and
177
horses with moderate or severe valvular regurgitation (n = 42); horses with aortopulmonary
178
fistula (n = 2) and VSD (n = 5) were not included in these subgroups. cTnI values in horses
179
with cardiomyopathy (median 0.90 ng/mL, range <0.03-58.27 ng/mL) was significantly
180
higher than in horses with valvular regurgitation (median 0.05 ng/mL, range <0.03-30.92
181
ng/mL; P = 0.001) and healthy horses (median <0.03 ng/mL, range <0.03-0.04 ng/mL; P
182
<0.001). Horses with valvular regurgitation also showed significantly higher cTnI values than
183
healthy horses (P = 0.001; Fig. 2).
184 185
STAT-I assay (laboratories B and C) - As for laboratory A, the median sample storage
186
time was 2 days in laboratories B (range 0-34 days) and C (range 0-38 days); 91% of samples
187
in laboratory B and 86% of samples in laboratory C were analysed within 1 week. Group 1
188
had a median cTnI concentration of <0.01 ng/mL (range <0.01-0.15 ng/mL; n = 22) in
189
laboratory B; the cTnI concentration in group 2 (n = 45) was significantly higher (median
9 Page 9 of 20
190
0.02 ng/mL, range <0.01-22.87 ng/mL; P = 0.044). A cut-off value of 0.015 was determined
191
from the ROC curve (Fig. 1; Table 1); however, the sensitivity was low and the P value was
192
>0.05, which indicates that the assay did not distinguish between groups 1 and 2. No
193
significant difference was found between cTnI in group 1 (median <0.01 ng/mL, range <0.01-
194
0.13 ng/mL) and group 2 (median <0.01 ng/mL, range <0.01-12.29 ng/mL; P > 0.05) when
195
samples were analysed in laboratory C (n = 43).
196 197
Comparison between assays - No significant difference in storage time was found
198
between assays. As shown in Table 2, a significant association was demonstrated between the
199
two cTnI assays (P <0.01). A moderate agreement was found between the Access Accu and
200
the STAT-I assay from laboratory B (Rc 0.543; bias correction, Cb, 0.837; shift 0.607) and C
201
(Rc 0.496; Cb 0.752; shift 0.781). A substantial agreement was found between the STAT-I
202
assay from laboratories B and C (Rc 0.617; Cb 0.943; shift 0.320). Despite these correlations,
203
individual troponin values differed substantially and a significant difference between the areas
204
under the ROC curve was found (P <0.001). The limits of agreement (Figs. 3a-c) and the
205
mean difference were large between the two assays. The mean ± SD difference between the
206
Access Accu and the STAT-I assay was 1.18 ± 4.08 ng/mL (laboratory A vs. laboratory B)
207
and 0.92 ± 3.14 ng/mL (laboratory A vs. laboratory C). Figs. 3a and 3b demonstrate that this
208
difference becomes larger as the magnitude of the measurement increased. As expected, the
209
mean ± SD difference (0.01 ± 0.27 ng/mL) was smaller when samples were analysed with the
210
same assay in two different laboratories.
211
Discussion
212
This study demonstrates that a large quantitative difference between two cTnI assays
213
exists in horses. Cardiac troponin I reference values differed between assays and a cut-off
214
value for cardiac disease could only be determined for the Access Accu assay. A
10 Page 10 of 20
215
quantitatively smaller difference was found when results from the STAT-I assay were
216
compared between laboratories.
217 218
In agreement with previous results (Jesty et al., 2009), the median cTnI value in group
219
1 was <0.03 ng/mL with the Access Accu assay. When samples were re-analysed with the
220
STAT-I assay, median cTnI values of <0.01 ng/mL in laboratory B and <0.01 ng/mL in
221
laboratory C were obtained. A high cTnI value (>0.10 ng/mL) was found in three healthy
222
horses with the STAT-I assay in laboratory B. In a study on human beings comparing the
223
same two assays, false positive results were found for the STAT-I assay due to the presence
224
of heterophile antibodies. These are naturally occurring antibodies which have a low affinity
225
and specificity that are produced against poorly defined antigens and can bind between the
226
capture and detection antibody of an assay (Kaplan and Levinson, 1999). The presence of
227
these antibodies has previously been suggested in an equine cTnI study (Slack et al., 2012)
228
and could also explain the conflicting results in our study.
229 230
Despite the significant correlation between the Access Accu and STAT-I assay (Table
231
2), a large mean difference was found (Figs. 3a and 3b). Since no gold standard was available,
232
it could not be concluded that one cTnI assay was better than another. However, an optimal
233
cut-off value was only significant for the Access Accu assay. Therefore, the Access Accu
234
assay might perform better than the STAT-I assay in a clinical setting.
235 236
Many explanations exist for the differences between these assays. The Access Accu
237
assay consists of two antibodies directed against the stable NH2-region-terminus of the cTnI
238
molecule (epitopes at amino acids 24-40 and 41-49) (James et al., 2006). An additional
239
capture antibody directed against the 87-91 amino acid region is used by the STAT-I assay.
11 Page 11 of 20
240
Since these assays recognise different target amino acids of the cTnI molecule, troponin
241
values might differ. In addition, cardiac troponins can circulate as various complexes (Wu et
242
al., 1998), troponin modification products after myocardial injury exist and multiple
243
degradation products have been described in human beings. Thus, the detectability of all these
244
products and fragments might also vary according to the assay used (Labugger et al., 2000).
245 246
In our study, cTnI was measured in horses with several different diseases; 49 horses had
247
structural cardiac disease (moderate or severe valvular regurgitation, aortopulmonary fistula
248
and VSD) and 23 horses had cardiomyopathy. Significantly higher cTnI values were found
249
with the Access Accu assay in horses with cardiomyopathy than in horses with valvular
250
regurgitation and in horses with valvular regurgitation compared to healthy horses. Similar
251
findings have been reported by Nath et al. (2012) and demonstrate that the cTnI assay not
252
only detects severe myocardial damage in case of cardiomyopathy, but is also able to
253
demonstrate mild cTnI increase in case of moderate or severe valvular disease.
254 255
When results of the STAT-I assay analysed at two different laboratories were compared,
256
a smaller mean difference (0.01 ng/mL) was found than when two different assays were
257
compared. However, a large standard deviation of the mean difference (0.27 ng/mL) still
258
existed suggesting that laboratory differences should be taken into account.
259 260
As found in medical studies, it is difficult to compare troponin I assays; however, it is
261
important to be aware of assay differences, to become familiar with one assay, to develop a
262
reference range from a large number of horses, to use the same laboratory and to learn by
263
experience which horses are most likely to be positive for cardiac disease (Apple and Saenger,
264
2013). Cardiac troponin T assay differences are limited, since only one manufacturer has
12 Page 12 of 20
265
marketed a cTnT assay. The specificity of the cTnT assay has been questioned in the case of
266
severe skeletal muscle injury (Bodor et al., 1997) and kidney disease (McLaurin et al., 1997).
267
However, with newer cTnT assays, false positive results seem to be limited to primary
268
skeletal muscle disease (Jaffe et al., 2011; Apple and Collinson, 2012). Therefore, similarly to
269
human medicine, cTnT might be an alternative for future troponin measurement in horses
270
(Thygesen et al., 2010).
271 272
A limited number (n = 23) of healthy horses was used to determine the reference range
273
for individual assays. Furthermore, our study also did not determine the coefficient of
274
variation for these assays (Thygesen et al., 2007). More assay-specific research is needed in a
275
larger equine population. Not all 95 samples could be re-analysed and a lower number of
276
samples were analysed with the STAT-I assay. However, if only samples (n = 67) analysed
277
with both assays were considered, the area under the ROC-curve was still higher for the
278
Access Accu assay (0.837, P <0.001) than for the STAT-I assay (0.643, P = 0.061).
279 280
Echocardiographic and electrocardiographic examination was not performed in all
281
horses with AM. However, the presence of cardiac disease was expected, since cTnI is not
282
present in skeletal muscle (Apple, 1999; Aggarwal et al., 2009) and because ultrasound, ECG
283
and post-mortem findings from previous studies have shown that AM is associated with
284
cardiomyopathy (Cassart et al., 2007; Votion et al., 2007; Verheyen et al., 2012; van Galen
285
and Votion, 2013; Unger et al., 2014). When horses with AM were not included, results from
286
the Access Accu assay still showed a significant difference between healthy horses and horses
287
with cardiac disease.
288
13 Page 13 of 20
289
Finally, samples were not analysed at the same time for the two assays. No significant
290
difference in storage time was found between assays, or between laboratories, and most of the
291
samples were analysed within the same week. The longest storage time (180 days; n = 2) was
292
found for the Access Accu assay, where samples had the highest cTnI concentrations. Since
293
samples were stored at -20 °C according to manufacturer’s instructions and because it is
294
known that cTnI or cTnT are stable for at least 1 year in serum samples stored at -70°C (Basit
295
et al., 2007), an influence of sample storage on results is unlikely.
296 297 298
Conclusions Troponin I assays cannot be directly compared to each other, since clinically important
299
differences exist. Assay specific reference values should be used and laboratory differences
300
must be considered. Cardiac troponin I measurement not only differentiates healthy horses
301
from horses with cardiomyopathy in a population but also healthy horses from horses with
302
moderate to severe valvular regurgitation.
303 304
Conflict of interest statement
305
None of the authors of this paper have a financial or personal relationship with other
306
people or organisations that could inappropriately influence or bias the content of the paper.
307 308
Acknowledgements
309
The authors wish to acknowledge the Special Research Fund (BOF) of Ghent
310
University for the funding of the study. We also acknowledge Marta Garcia-Granero for her
311
contribution to the study.
312 313
References
14 Page 14 of 20
314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363
Adin, D.B., Oyama, M.A., Sleeper, M.M., Milner, R.J., 2006. Comparison of canine cardiac troponin I concentrations as determined by 3 different analyzers. Journal of Veterinary Internal Medicine 20, 1136-1142. Aggarwal, R., Lebiedz-Odrobina, D., Sinha, A., Manadan, A., Case, J.P., 2009. Serum cardiac troponin T, but not troponin I, is elevated in idiopathic inflammatory myopathies. Journal of Rheumatology 36, 2711-2714. Apple, F.S., 1999. Clinical and analytical standardization issues confronting cardiac troponin I. Clinical Chemistry 45, 18-20. Apple, F.S., Collinson, P.O., 2012. Analytical characteristics of high-sensitivity cardiac troponin assays. Clinical Chemistry 58, 54-61. Apple, F.S., Saenger, A.K., 2013. The state of cardiac troponin assays: Looking bright and moving in the right direction. Clinical Chemistry 59, 1014-1016. Basit, M., Bakshi, N., Hashem, M., Allebban, Z., Lawson, N., Rosman, H.S., Maciejko, J.J., 2007. The effect of freezing and long-term storage on the stability of cardiac troponin T. American Journal of Clinical Pathology 128, 164-167. Begg, L.M., Hoffmann, K.L., Begg, A.P., 2006. Serum and plasma cardiac troponin I concentrations in clinically normal Thoroughbreds in training in Australia. Australian Veterinary Journal 84, 336-337. Bock, J.L., Singer, A.J., Thode, H.C., Jr., 2008. Comparison of emergency department patient classification by point-of-care and central laboratory methods for cardiac troponin I. American Journal of Clinical Pathology 130, 132-135. Bodor, G.S., Survant, L., Voss, E.M., Smith, S., Porterfield, D., Apple, F.S., 1997. Cardiac troponin T composition in normal and regenerating human skeletal muscle. Clinical Chemistry 43, 476-484. Buhl, R., Petersen, E.E., Lindholm, M., Bak, L., Nostell, K., 2013. Cardiac arrhythmias in Standardbreds during and after racing-possible association between heart size, valvular regurgitations, and arrhythmias. Journal of Equine Veterinary Science 33, 590-596. Cassart, D., Baise, E., Cherel, Y., Delguste, C., Antoine, N., Votion, D., Amory, H., Rollin, F., Linden, A., Coignoul, F., Desmecht, D., 2007. Morphological alterations in oxidative muscles and mitochondrial structure associated with equine atypical myopathy. Equine Veterinary Journal 39, 26-32. DeLong, E.R., DeLong, D.M., Clarke-Pearson, D.L., 1988. Comparing the areas under two or more correlated receiver operating characteristic curves: A nonparametric approach. Biometrics 44, 837-845. Divers, T.J., Kraus, M.S., Jesty, S.A., Miller, A.D., Mohammed, H.O., Gelzer, A.R.M., Mitchell, L.M., Soderholm, L.V., Ducharme, N.G., 2009. Clinical findings and serum cardiac troponin I concentrations in horses after intragastric administration of sodium monensin. Journal of Veterinary Diagnostic Investigation 21, 338-343. 15 Page 15 of 20
364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413
Gehlen, H., Rohn, K., Deegen, E., Stadler, P., 2006. Labordiagnostische untersuchungen bei pferden mit herzerkrankungen: Wertigkeit verschiedener kardialer marker. Pferdeheilkunde 22, 532-541. Jaffe, A.S., Vasile, V.C., Milone, M., Saenger, A.K., Olson, K.N., Apple, F.S., 2011. Diseased skeletal muscle: A noncardiac source of increased circulating concentrations of cardiac troponin T. Journal of the American College of Cardiology. 58, 1819-1824. James, S., Flodin, M., Johnston, N., Lindahl, B., Venge, P., 2006. The antibody configurations of cardiac troponin I assays may determine their clinical performance. Clinical Chemistry 52, 832-837. Jesty, S.A., Kraus, M., Gelzer, A., Rishniw, M., Moïse, N.S., 2009. Effect of transvenous electrical cardioversion on plasma cardiac troponin I concentrations in horses with atrial fibrillation. Journal of Veterinary Internal Medicine 23, 1103-1107. Kaplan, I.V., Levinson, S.S., 1999. When is a heterophile antibody not a heterophile antibody? When it is an antibody against a specific immunogen. Clinical Chemistry 45, 616-618. Kraus, M.S., Jesty, S.A., Gelzer, A.R., Ducharme, N.G., Mohammed, H.O., Mitchell, L.M., Soderholm, L.V., Divers, T.J., 2010. Measurement of plasma cardiac troponin I concentration by use of a point-of-care analyzer in clinically normal horses and horses with experimentally induced cardiac disease. American Journal of Veterinary Research 71, 55-59. Kraus, M.S., Kaufer, B.B., Damiani, A., Osterrieder, N., Rishniw, M., Schwark, W., Gelzer, A.R., Divers, T.J., 2013. Elimination half-life of intravenously administered equine cardiac troponin I in healthy ponies. Equine Veterinary Journal 45, 56-59. La'ulu, S.L., Roberts, W.L., 2010. Performance characteristics of five cardiac Troponin I assays. Clinica Chimica Acta 411, 1095-1101. Labugger, R., Organ, L., Collier, C., Atar, D., Van Eyk, J.E., 2000. Extensive troponin I and T modification detected in serum from patients with acute myocardial infarction. Circulation 102, 1221-1226. Lindahl, B., Eggers, K.M., Venge, P., James, S., 2013. Evaluation of four sensitive troponin assays for risk assessment in acute coronary syndromes using a new clinically oriented approach for comparison of assays. Clinical Chemistry Laboratory Medicine 51, 18591864. Maynard, S.J., Menown, I.B.A., Adgey, A.A.J., 2000. Troponin T or troponin I as cardiac markers in ischaemic heart disease. Heart 83, 371-373. McLaurin, M.D., Apple, F.S., Voss, E.M., Herzog, C.A., Sharkey, S.W., 1997. Cardiac troponin I, cardiac troponin T, and creatine kinase MB in dialysis patients without ischemic heart disease: Evidence of cardiac troponin T expression in skeletal muscle. Clinical Chemistry 43, 976-982. 16 Page 16 of 20
414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463
Nath, L., Anderson, G., Hinchcliff, K., Savage, C., 2012. Serum cardiac troponin I concentrations in horses with cardiac disease. Australian Veterinary Journal 90, 351357. Nostell, K., Hägström, J., 2008. Resting concentrations of cardiac troponin I in fit horses and effect of racing. Journal of Veterinary Cardiology 10, 105-109. Petersmann, A., Ittermann, T., Fries, C., Lubenow, N., Kohlmann, T., Kallner, A., Greinacher, A., Nauck, M., 2013. Comparison of the 99th percentiles of three troponin I assays in a large reference population. Clinical Chemistry and Laboratory Medicine 51, 2181-2186. Phillips, W., Giguère, S., Franklin, R.P., Hernandez, J., Adin, D., Peloso, J., 2003. Cardiac troponin I in pastured and race-training Thoroughbred horses. Journal of Veterinary Internal Medicine 17, 597-599. Reiter, M., Twerenbold, R., Reichlin, T., Haaf, P., Peter, F., Meissner, J., Hochholzer, W., Stelzig, C., Freese, M., Heinisch, C., et al., 2011. Early diagnosis of acute myocardial infarction in the elderly using more sensitive cardiac troponin assays. European Heart Journal 32, 1379-1389. Rishniw, M., Simpson, K.W., 2005. Cloning and sequencing of equine cardiac troponin I and confirmation of its usefulness as a target analyte for commercial troponin I analyzers. Journal of Veterinary Diagnostic Investigation 17, 582-584. Schwarzwald, C.C., Hardy, J., Buccellato, M., 2003. High cardiac troponin I serum concentration in a horse with multiform ventricular tachycardia and myocardial necrosis. Journal of Veterinary Internal Medicine 17, 364-368. Slack, J., Boston, R.C., Soma, L., Reef, V.B., 2012. Cardiac troponin I in racing standardbreds. Journal of Veterinary Internal Medicine 26, 1202-1208. Thygesen, K., Alpert, J.S., White, H.D., 2007. Universal definition of myocardial infarction. Journal of the American College of Cardiology 50, 2173-2195. Thygesen, K., Mair, J., Katus, H., Plebani, M., Venge, P., Collinson, P., Lindahl, B., Giannitsis, E., Hasin, Y., Galvani, M., et al., , 2010. Recommendations for the use of cardiac troponin measurement in acute cardiac care. European Heart Journal 31, 21972204. Unger, L., Nicholson, A., Jewitt, E.M., Gerber, V., Hegeman, A., Sweetman, L., Valberg, S., 2014. Hypoglycin A concentrations in seeds of Acer pseudoplatanus trees growing on atypical myopathy-affected and control pastures. Journal of Veterinary Internal Medicine 28, 1289-1293. van Galen, G., Votion, D.M., 2013. Management of cases suffering from atypical myopathy: Interpretations of descriptive, epidemiological and pathophysiological findings. Part 2: Muscular, urinary, respiratory and hepatic care, and inflammatory/infectious status. Equine Veterinary Education 25, 308-314. 17 Page 17 of 20
464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495
Venge, P., James, S., Jansson, L., Lindahl, B., 2009. Clinical performance of two highly sensitive cardiac troponin I assays. Clinical Chemistry 55, 109-116. Verheyen, T., Decloedt, A., De Clercq, D., van Loon, G., 2012. Cardiac changes in horses with atypical myopathy. Journal of Veterinary Internal Medicine 26, 1019-1026. Verheyen, T., Decloedt, A., Van Der Vekens, N, Sys, S., De Clercq, D., van Loon, G., 2013. Ventricular response during lungeing exercise in horses with lone atrial fibrillation. Equine Veterinary Journal 45, 309-314. Votion, D.M., Linden, A., Saegerman, C., Engels, P., Erpicum, M., Thiry, E., Delguste, C., Rouxhet, S., Demoulin, V., Navet, R., et al., 2007. History and clinical features of atypical myopathy in horses in Belgium (2000-2005). Journal of Veterinary Internal Medicine 21, 1380-1391. Votion, D.M., van Galen, G., Sweetman, L., Boemer, F., de Tullio, P., Dopagne, C., Lefere, L., Mouithys-Mickalad, A., Patarin, F., Rouxhet, S., et al., 2014. Identification of methylenecyclopropyl acetic acid in serum of European horses with atypical myopathy. Equine Veterinary Journal 46, 146-149. Wells, S.M., Sleeper, M., 2008. Cardiac troponins. Journal of Veterinary Emergency and Critical Care 18, 235-245. White, H.D., 2011. Pathobiology of troponin elevations: Do elevations occur with myocardial ischemia as well as necrosis? Journal of the American College of Cardiology 57, 2406-2408. Wu, A.H.B., Feng, Y.J., Moore, R., Apple, F.S., McPherson, P.H., Buechler, K.F., Bodor, G., Standar, A.A.C.C.S.C., 1998. Characterization of cardiac troponin subunit release into serum after acute myocardial infarction and comparison of assays for troponin T and I. Clinical Chemistry 44, 1198-1208.
18 Page 18 of 20
496
Figure legends
497 498
Fig. 1. Receiver operator characteristic (ROC) curve of the Access Accu (laboratory A) and
499
STAT-I (laboratory B) assay for detection of cardiac disease ( cut-off of the Access Accu
500
assay).
501 502
Fig. 2. Box plot comparing troponin I concentration measured with the Access Accu in
503
healthy horses (group 1), horses with moderate or severe valvular regurgitation (group 2a) and
504
horses with cardiomyopathy (group 2b).
505 506
Figs. 3a-c. Bland-Altman assays comparing two different troponin I assays and measurement
507
of the STAT-I assay in two different laboratories (B and C). The mean difference between
508
assays increases with a higher troponin value (group 1: healthy horses, group 2: horses with
509
cardiac disease).
510 511
Table 1. Troponin I sensitivity and specificity to detect cardiac disease for troponin I.
512
Method
Laboratory
n
Access Accu
A
95
0.035
70%
STAT-I
B
67
0.015
STAT-I
C
43
0.015
513
Cut-off Sensitivity Specificity
P value
AUC
95% CI
91%
<0.001
0.83
0.70-0.97
55%
68%
0.11
0.64
0.47-0.81
30%
86%
0.30
0.59
0.42-0.77
AUC, area under the curve; 95% CI, 95% confidence interval.
19 Page 19 of 20
514
Table 2 Spearman correlation coefficient between two troponin assays for cardiac troponin I
515
determination in horses.
516
Access Accu Laboratory A
STAT-I Laboratory B
STAT-I Laboratory C
Access Accu Laboratory A
STAT-I Laboratory B
STAT-I Laboratory C
Correlation coefficient
1.000
X
X
n
95
Correlation coefficient
0.454
1.000
X
n
67
67
Correlation Coefficient
0.475
0.412
1.000
n
43
42
43
517 518 519
* Significant correlation at P <0.01.
520
20 Page 20 of 20