Comparison of D‐dimer concentration and thromboelastography for diagnosis of cerebrovascular accidents in dogs: A retrospective study

Abstract

Background

Cerebrovascular accidents (CVAs) in dogs are diagnosed using magnetic resonance imaging (MRI). This modality is sometimes unavailable, and CVAs can resemble other lesions on MRI. D‐dimer concentration and thromboelastography (TEG) are utilized in human medicine in addition to diagnostic imaging to support diagnosis of CVAs, but their use in veterinary patients has not been assessed.

Objective

Assess utility of blood D‐dimer concentration and TEG in supporting the imaging diagnosis of CVAs in dogs.

Animals

Sixty‐eight client‐owned dogs with neurologic signs that had brain MRI and D‐dimer concentration or TEG performed.

Methods

Multicenter, retrospective study. The incidence of abnormal D‐dimer concentration or TEG was compared between patients with MRI evidence of CVA and a control population. Analysis methods included Fisher's exact test or Chi‐squared test for association and comparison of independent proportions.

Results

Neither D‐dimer concentration nor TEG was significantly associated with a CVA (P = .38 and .2, respectively). D‐dimer testing was performed in a low‐risk population and showed low sensitivity (30.8%; 95% confidence interval [CI], 10%‐61%) and high specificity (86.4%; 95% CI, 64%‐96%) for CVA diagnosis. Thromboelastography was performed in a high‐risk population and showed moderate sensitivity (64.3%; 95% CI, 44%‐81%) and specificity (66.7%; 95% CI, 24%‐94%) for CVA diagnosis. Abnormal D‐dimer concentration or TEG were not helpful in differentiating hemorrhagic from ischemic stroke (P = .43 and .41, respectively).

Conclusions

Although blood D‐dimer concentration or TEG alone are not diagnostic of CVAs in dogs, a positive D‐dimer result supports additional testing for CVA.

Abbreviations

ADC

apparent diffusion coefficient

CSF

cerebrospinal fluid

CVA

cerebrovascular accident

DWI

diffusion weighted imaging

FLAIR

fluid attenuated inversion recovery

GRE

gradient echo

MRI

magnetic resonance imaging

T1W

T1‐weighted

T2W

T2‐weighted

TEG

thromboelastography

TNCC

total nucleated cell count

1. INTRODUCTION

A cerebrovascular accident (CVA), or stroke, can be ischemic or hemorrhagic in nature. ¹ , ² Cerebrovascular accidents are suspected in veterinary patients based on history, physical, and neurologic examination findings and are presumptively diagnosed using advanced imaging, particularly magnetic resonance imaging (MRI). Characteristics on MRI for ischemic strokes include a well demarcated, T1‐weighted (T1W) hypointense, T2‐weighted (T2W) and fluid attenuated inversion recovery (FLAIR) hyperintense lesion and weak to absent contrast enhancement on post‐contrast T1W images within the first 7 to 10 days of onset of clinical signs. ² , ³ , ⁴ , ⁵ Ischemic strokes older than 24 hours appear hyperintense on diffusion weighted imaging (DWI), and hypointense on the corresponding apparent diffusion coefficient (ADC) map. ² , ³ , ⁴ , ⁵ The lesion appearance of hemorrhagic strokes on MRI changes over time, as hemorrhage progresses from oxygenated to deoxygenated blood. ³ , ⁵ Hemorrhage will appear as a signal void on T2W gradient echo (GRE or T2*) sequences. ² , ³ , ⁴ , ⁵ Another form of hemorrhagic CVAs, known as cerebral microbleeds, appear as round, small, T2*W signal voids (hypointense) within the brain parenchyma. ⁶ Many of these imaging features are non‐specific, making differentiation of CVAs from other diseases challenging. Other intra‐axial disease processes, such as meningoencephalitis or neoplasms such as glioma, also may have T2W and FLAIR hyperintensities, variable contrast enhancement, and occasional T2*W signal voids, making a misdiagnosis of CVAs possible. ⁷ , ⁸ , ⁹ Although CVA is a differential diagnosis for patients showing acute neurologic signs with subsequent improvement, other diseases may present similarly. In a recent prospective study of 20 dogs presenting for acute, non‐progressive, intracranial signs, 2 were diagnosed with ischemic stroke based on MRI features. ¹⁰ The remaining 18 dogs were diagnosed with idiopathic vestibular syndrome without evidence of CVA, neoplasia, inflammatory disease, or the diagnosis was undetermined. ¹⁰

The D‐dimers are measurable products of hemostasis that are only detectable after both thrombosis and fibrinolysis have occurred. ¹¹ , ¹² Alternatively, thromboelastography (TEG) traces changes in the viscoelastic properties of blood during the clotting process. Abnormalities in TEG can provide evidence for hypercoagulable states that may predispose to thromboembolic events, as well as for hypocoagulable states or platelet dysfunction. ¹³ Although D‐dimers and TEG have been used to aid in diagnosing CVAs in human patients, ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ , ²¹ , ²² , ²³ studies evaluating D‐dimers and TEG for aiding CVA diagnosis in animals are limited. ²⁴ , ²⁵

The diagnosis of CVA in dogs is challenging because of lack of pathognomonic test results and limited accessibility to MRI. Our primary aim was to assess the sensitivity and specificity of blood D‐dimer concentration and TEG for supporting an MRI diagnosis of CVA in dogs. A secondary aim was to evaluate if these tests are helpful in differentiating ischemic from hemorrhagic strokes. We hypothesized that D‐dimer concentration and TEG will be abnormal in dogs diagnosed with CVA and are more likely to be abnormal in cases of ischemic stroke.

2. MATERIALS AND METHODS

2.1. Case selection and cerebrovascular accident definition

Ours was a multicenter, retrospective case‐control study evaluating 2 groups of dogs. Medical records at the University of Wisconsin—Madison Veterinary Medicine Teaching Hospital and Oradell Animal Hospital (Paramus, New Jersey), were searched to identify all dogs that had brain MRI performed, as well as testing for blood D‐dimers or TEG between August 2015 and May 2021. If D‐dimer or TEG testing was not performed at the same visit as the MRI, the medical record associated with the blood testing had to reference the neurologic complaint or MRI findings of that visit. Search terms included: canine, dog, brain MRI, TEG, and D‐dimer.

To be included in the study, dogs from each institution had to have: (a) neurologic abnormalities at presentation or a history of neurologic signs such as seizures or suspected transient vestibular syndrome; (b) a complete MRI study of the brain including both transverse and sagittal T1W (pre‐ and postcontrast) and T2W sequences, transverse T2W FLAIR, transverse T2*, and transverse DWI with an associated ADC map; and (c) either TEG, blood D‐dimer concentration or both performed. The MRI criteria for a suspected ischemic stroke included a focal or well‐demarcated T1W hypointense, T2W and FLAIR hyperintense lesion with variable contrast enhancement. ³ , ⁴ , ⁵ Evidence of restricted diffusion on DWI and ADC map was considered a supporting feature of ischemic CVA, but was not required for imaging diagnosis. ³ , ⁴ , ⁵ The MRI criteria for a suspected hemorrhagic stroke included signal void on T2* sequences; T1W, T2W, FLAIR, and T1W post‐contrast characteristics that varied based on time elapsed between onset of clinical signs and imaging. ³ , ⁴ , ⁵ , ⁶ Dogs of any age, breed, or sex were considered for inclusion. Dogs with spinal cord imaging were included if this procedure was performed in addition to a complete brain MRI study. Patients without evidence of CVA, with or without spinal cord pathology, were included as neurologic controls, whereas patients with a CVA were included in the CVA group regardless of spinal cord imaging findings.

Dogs were included into either the D‐dimer or TEG cohort based on the testing that was performed. Dogs then were included in 1 of 2 subgroups based on their imaging findings. Group 1 were dogs diagnosed with presumptive CVA (ischemic or hemorrhagic) based on MRI findings and neurologic signs. Cases with multiple intracranial lesions were included in this group, as long as 1 of the lesions was diagnosed as a CVA. Cerebral microbleeds also were included in this group. Group 2 (control group) consisted of dogs diagnosed with an encephalopathy other than CVA based on MRI findings, progressive neurological signs, and cerebrospinal fluid (CSF) analysis. The MRI findings were reviewed by a board‐certified radiologist at the time of diagnosis.

2.2. Data collection

The medical records were reviewed, and collected data included age at time of MRI, breed, sex, neuroanatomic localization at the time of admission based on neurologic examination, onset of clinical signs, MRI findings, CSF analysis results when available, and D‐dimer or TEG results or both.

All MRIs were performed under general anesthesia using 1 of 3 magnets: 1.5 T Siemens Avanto (Munich, Germany), 1.5 T GE Healthcare (Milwaukee, Wisconsin), and 1.0 T GE Medical Systems Signa Advantage (Milwaukee, Wisconsin). Plasma D‐dimer concentrations were measured by a commercial laboratory (Antech Diagnostics) by latex agglutination, and abnormal results were concentrations measuring >250 ng/mL (reference range, < 250 ng/mL). The TEG analysis was performed at the University of Wisconsin—Madison, Veterinary Medicine Teaching Hospital. Whole blood was collected into a sodium citrate containing tube, and kaolin‐activated TEG was performed. The TEG results generated and interpreted were R (reaction time including the time it takes blood to begin forming a clot), K (representative of clot formation time), α angle (representative of rate of fibrin cross linking), and MA (measure of clot firmness; Figure 1). Results consistent with hypercoagulability were shortened R (<3 minutes) or K (<1 minute), or increased α angle (>74°) or MA (>70 mm; Figure 1). Results consistent with hypocoagulability included prolonged R (>9 minutes) or K (>6 minutes), or decreased α angle (<34°) or MA (<40 mm). Previous studies have validated the use of D‐dimer concentration and TEG in dogs. ²⁶ , ²⁷ , ²⁸ The duration of time between onset of clinical signs and blood collection for D‐dimers or TEG also was recorded, as was the duration of time between onset of clinical signs and MRI.

FIGURE 1.

Representative thromboelastography (TEG) tracings from dogs with neurologic signs undergoing brain MRI and TEG testing. (A) TEG results consistent with normocoagulability, including normal R, K, α angle, and MA. (B) TEG results consistent with hypercoagulability, including shortened R, shortened K, and increased α angle.

The reference range for normal CSF total nucleated cell count (TNCC) was <5 cells/μL, and normal total protein concentration was <30 mg/dL for cerebellomedullary cistern collection and <45 mg/dL for lumbar collection. ²⁹ Pleocytosis was characterized as mild if the TNCC was 6‐25 cells/μL, moderate if the TNCC was 26‐100/μL, and marked if the total nucleated cell count was >100/μL. ²⁹ Albuminocytological dissociation was defined as an increase in CSF total protein concentration with a normal TNCC. ²⁹ , ³⁰

2.3. Statistical analysis

A t test, Chi‐squared, or Fisher's exact test were used to compare signalment and history data among the dogs with normal and abnormal D‐dimer results, as well as between the dogs with normal and abnormal TEG results. The association between diagnosis of a CVA and abnormal D‐dimer or TEG results was analyzed using Fisher's exact or Chi‐squared tests in the 2 cohorts, depending on sufficient sample size for the Chi‐squared test. The sensitivity and specificity for abnormal TEG or abnormal D‐dimer concentration for diagnosis of a CVA were calculated, and 95% confidence intervals (CIs) were determined. All analyses were conducted using R version 4.0. A P value < .05 was considered significant for all tests.

3. RESULTS

A total of 68 client‐owned dogs met the inclusion criteria. Forty dogs were included in Group 1 (CVA) and 28 dogs in Group 2 (neurologic control). Thirty‐five dogs had blood D‐dimer concentrations measured, and 34 dogs had TEG performed. One dog in Group 1 had both D‐dimer concentration and TEG performed. Demographic information is summarized in Table 1.

TABLE 1.

Signalment and timing of test collection in dogs with CVA and neurologic controls.

Variable	D‐dimer (n = 35)			TEG (n = 34)
	Evidence of CVA (n = 13)	Neurologic controls (n = 22)	P‐value	Evidence of CVA (n = 28)	Neurologic controls (n = 6)	P‐value
Age (years), mean (SD)	9.7 (4.1)	8.5 (3.4)	.37	10.4 (4.4)	11.4 (4.1)	.61
Sex
Male	7 (53.8%)	11 (50.0%)	1.00	16 (57.1%)	3 (50.0%)	1.00
Female	6 (46.2%)	11 (50.0%)		12 (42.9%)	3 (50.0%)
Time between onset of clinical signs & D‐dimer or TEG sample collection
<7 days	10 (76.9%)	18 (81.8%)	1.00	17 (60.7%)	0 (0.0%)	.02
≥7 days	3 (23.1%)	4 (18.2%)		11 (39.3%)	6 (100.0%)
Time between onset of clinical signs & MRI
<7 days	7 (53.8%)	13 (59.1%)	1.00	22 (78.6%)	0	.001
≥7 days	6 (46.2%)	9 (40.9%)		6 (21.4%)	6 (100.0%)

3.1. D‐dimer results

3.1.1. Population data

Thirty‐five dogs had blood D‐dimer testing performed. Neither the timing of blood collection in relation to onset of clinical signs nor the time between onset of clinical signs and MRI differed significantly between neurologic controls and dogs with evidence of CVA. Patient sex and age were not significantly different between dogs with evidence of CVA and neurologic controls (Table 1). Twenty‐two dogs were purebred and included 2 each of the following breeds: Labrador Retriever, Toy Poodle, Maltese, Yorkshire terrier, and Pug. There was 1 dog each of the following breeds: Golden Retriever, English bulldog, Beagle, miniature Schnauzer, Airedale terrier, French bulldog, Soft‐coated Wheaten terrier, Cavalier King Charles spaniel, Jindo, Cairn terrier, Dutch shepherd, and miniature Pinscher. Thirteen dogs were mixed breed. Neuroanatomic lesion localization based on neurologic examination was prosencephalon (n = 20), central vestibular system (n = 11), multifocal brain (n = 2), and peripheral vestibular system (n = 1). One dog had a history of collapse episodes not consistent with cardiac disease or seizures. The prevalence of CVA in this cohort of dogs having D‐dimer testing performed was low (13/35, 37%).

Group 1 (13 dogs)

Four of 13 dogs (30.8%) with MRI evidence of a CVA had increased D‐dimer concentrations. Eight dogs (61.5%) had evidence of ischemic stroke and 5 dogs (38.5%) had evidence of hemorrhagic stroke. Of the 8 dogs diagnosed with ischemic CVA, 3 (37.5%) had an abnormal D‐dimer. Of the 5 dogs diagnosed with hemorrhagic CVA, 1 had an abnormal D‐dimer concentration. Abnormal D‐dimer concentration was not significantly associated with classification of ischemic or hemorrhagic CVA (P = .43; Table 2). An acute onset of clinical signs occurred in 92.3% (12/13) of dogs; 2 of these dogs had new onset seizures. One patient with a hemorrhagic CVA had a week‐long history of behavioral changes. Ten dogs (76.9%) had D‐dimer testing performed within 1 week of the onset of clinical signs. Cerebrospinal fluid analysis was performed in 4 dogs (Table 3).

TABLE 2.

Comparison of D‐dimer and TEG results in dogs with ischemic vs hemorrhagic CVAs.

Test performed	Test result	Ischemic CVA	Hemorrhagic CVA	Neurologic controls	P‐value
D‐dimer	Abnormal	3 (37.5%)	1 (20%)	3 (13.6%)	.43
	Normal	5 (62.5%)	4 (80%)	19 (86.4%)
TEG	Abnormal	8 (66.7%)	10 (62.5%)	2 (33.3%)	.41
	Normal	4 (33.3%)	6 (37.5%)	4 (66.7%)

TABLE 3.

Cerebrospinal fluid results in dogs with D‐dimer or TEG and brain MRI performed.

	CVA subtype	Test result	Normal CSF	Mild pleocytosis	Moderate pleocytosis	Marked pleocytosis	Albumino‐cytologic dissociation
MRI evidence of CVA	Ischemic	Normal D‐dimer (n = 1)	1 (100%)	0	0	0	0
		Abnormal D‐dimer (n = 2)	1 (50%)	1 (50%)	0	0	0
		Normal TEG (n = 4)	0	2 (50%)	0	0	2 (50%)
		Abnormal TEG (n = 8)	5 (62.5%)	1 (12.5%)	2 (25%)	0	0
	Hemorrhagic	Normal D‐dimer (n = 1)	0	0	0	0	1 (100%)
		Abnormal D‐dimer (n = 0)	0	0	0	0	0
		Normal TEG (n = 3)	1 (33.3%)	1 (33.3%)	0	0	1 (33.3%)
		Abnormal TEG (n = 9)	2 (22.2%)	1 (11.1%)	1 (11.1%)	1 (11.1%)	4 (44.4%)
Controls		Normal D‐dimer (n = 12)	8 (66.7%)	2 (16.7%)	1 (8.3%)	0	1 (8.3%)
		Abnormal D‐dimer (n = 2)	1 (50%)	1 (50%)	0	0	0
		Normal TEG (n = 3)	0	1 (33.3%)	0	0	2 (66.7%)
		Abnormal TEG (n = 1)	0	0	0	0	1 (100%)

Note: Not all dogs had CSF analysis performed.

Group 2 (22 dogs)

Three of 22 (13.6%) dogs in the neurologic control group had increased D‐dimer concentrations, and 19 of 22 dogs (86.4%) had normal D‐dimer concentrations. The MRI or clinical diagnoses for these patients included: neoplasia (5), idiopathic epilepsy (5), epilepsy of unknown cause (2), meningoencephalitis of unknown etiology (2), spinal fibrocartilaginous embolism (2), hydrocephalus (1), hepatic encephalopathy (1), idiopathic Horner's syndrome (1), steroid‐response meningitis‐arteritis (1), cognitive dysfunction syndrome (1), and acute non‐compressive nucleus pulposus extrusion (1). In the neurologic controls, an increased D‐dimer concentration was found in 1 dog with idiopathic epilepsy, 1 dog with neoplasia, and 1 dog with acute non‐compressive nucleus pulposus extrusion. Cerebrospinal fluid analysis was performed in 14/22 dogs (Table 3).

3.1.2. Sensitivity and specificity of D‐dimers

In this low‐risk population (CVA prevalence, 37%), abnormal blood D‐dimer concentration was not significantly associated with CVA (P = .38). The sensitivity of D‐dimer concentration for CVA diagnosis was 31.1% (95% CI, 10%‐61%), and specificity was 86.4% (95% CI, 64%‐96%; Table 4).

TABLE 4.

D‐dimer and TEG results in dogs with MRI evidence of CVA and neurologic controls.

Test performed	Group	Test result	Number of dogs	P‐value	Sensitivity	Specificity	Positive predictive value
D‐dimer (n = 35)	MRI evidence of CVA (Group 1, n = 13)	Abnormal	4 (30.8%)	.38	0.31 (0.10‐0.61)	0.86 (0.64‐0.96)	57.1% (20.0‐88.0)
		Normal	9 (69.2%)
	Neurologic controls (Group 2, n = 22)	Abnormal	3 (13.6%)
		Normal	19 (86.4%)
TEG (n = 34)	MRI evidence of CVA (Group 1, n = 28)	Abnormal	18 (64.3%)	.20	0.64 (0.44‐0.81)	0.67 (0.24‐0.94)	90.0% (67.0‐98.0)
		Normal	10 (35.7%)
	Neurologic controls (Group 2, n = 6)	Abnormal	2 (33.3%)
		Normal	4 (66.7%)

3.2. TEG results

3.2.1. Population data

Thirty‐four dogs had TEG performed. No dogs had TEG results consistent with hypocoagulability, and therefore all results defined as abnormal refer to evidence of hypercoagulability. Patient age and sex were not significantly different between dogs with evidence of CVA and neurologic controls. In the TEG group, dogs with evidence of CVA were significantly more likely to have the test performed <7 days from the onset of clinical signs (P = .02), as well as more likely to have an MRI performed <7 days from the onset of clinical signs (P = .001; Table 1). Twenty‐nine dogs were purebred and included 4 Labrador retrievers, 3 Weimaraners, 3 Shih tzus and 2 each of the following breeds: Yorkshire terrier, Miniature Schnauzer, and Welsh Corgi. There was 1 dog each of Toy Poodle, Beagle, Soft‐coated Wheaten terrier, Alaskan Klee Kai, Chihuahua, German shepherd, Golden retriever, Shetland sheepdog, Border Collie, Shiba Inu, Siberian Husky, Brittany spaniel, and Greyhound. Five dogs were mixed breeds. Neuroanatomic lesion localization based on the neurologic examination was central vestibular system (n = 13), prosencephalon (n = 10), multifocal brain (n = 5), peripheral vestibular system, cerebellum, brainstem, T3‐L3 myelopathy, C1‐C5 myelopathy, and normal neurologic examination (n = 1 each). The prevalence of CVA in this population of dogs was high (28/34, 82%).

Group 1 (28 dogs)

Twenty‐eight of 34 dogs (82.4%) with TEG performed were in Group 1. Eighteen dogs (64.3%) had abnormal TEG. Twelve dogs had evidence of ischemic CVA, and 16 dogs had evidence of hemorrhagic CVA identified on MRI. Cerebrospinal fluid analysis was performed in 24/28 dogs (Table 3).

Of the 12 dogs diagnosed with ischemic CVA, 66.7% (8/12) had abnormal TEG. Of the 16 dogs diagnosed with hemorrhagic CVA, 10 (62.5%) had abnormal TEG. Abnormal TEG was not significantly associated with classification of ischemic or hemorrhagic CVA (P = .40; Table 2).

Group 2 (6 dogs)

Two dogs (33.3%) had abnormal TEG, and 66.7% (4/6) had normal TEG. The diagnoses for these dogs were neoplasia (3), meningoencephalitis of unknown etiology (2), and otitis media interna (1). Results consistent with hypercoagulability were seen in 1 dog with meningoencephalitis of unknown etiology and 1 dog with neoplasia. Cerebrospinal fluid analysis was performed in 4 dogs (Table 3).

3.2.2. Sensitivity and specificity of TEG

In this high‐risk population (prevalence of CVA, 82%), abnormal TEG was not significantly associated with CVA (P = .2). The sensitivity of TEG for CVA diagnosis was 64.3% (95% CI, 44%‐81%), and the specificity was 67.7% (95% CI, 24%‐94%; Table 4).

4. DISCUSSION

We showed that blood D‐dimer concentration or TEG alone cannot diagnose MRI‐confirmed CVAs in dogs with neurologic disease, but a positive D‐dimer test result may be an appropriate screening test because of its specificity. A previously published study of 5 dogs with a possible hypercoagulable state suggested potential value of TEG results in diagnosing ischemic strokes, but did not find utility in blood D‐dimer concentration. ²⁴ These findings were in contrast to studies in humans, which support the use of D‐dimer concentrations as a potentially useful diagnostic marker for vascular central nervous system (CNS) disease. ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰

Despite being unable to prove their reliability as sole diagnostic tests of CVAs in dogs, our study showed that D‐dimer concentration has a reasonably high specificity in a low‐risk population. Based on our results, a CVA should be highly suspected in a dog with neurologic signs and a positive D‐dimer test result when evaluated in a low‐risk population. Therefore, D‐dimer concentration has use as a screening test in dogs with clinical signs consistent with CVA, and a positive result should prompt further diagnostic testing such as MRI.

Blood D‐dimer concentration has had been used as a supplementary diagnostic tool and prognostic factor for ischemic strokes in human medicine. In contrast to our findings, blood D‐dimer concentration has high sensitivity and high negative predictive value in human medicine, meaning a normal D‐dimer concentration helps rule out venous thromboembolism in patients with low pre‐test probability of the disease. ¹⁴ Additional diagnostic tests such as ultrasonography or computed tomography still are recommended in patients with high clinical suspicion for venous thromboembolism and in patients where D‐dimer concentrations are increased, similar to our suggestions for dogs with increased D‐dimer concentrations. ¹⁴ Increased D‐dimer concentration in humans has been shown to be a risk factor for strokes in certain diseases, such as heart failure, cancer, or atrial fibrillation, and may have utility for predicting ischemic stroke recurrence. ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ In contrast to the human medical literature, previous veterinary studies showed the unreliability of blood D‐dimer concentration for diagnosis of CVAs in dogs, but the number of dogs with CVA in these studies was low ²⁴ or dogs with CVA were analyzed together with other non‐inflammatory disease processes. ²⁵ In human medicine, D‐dimer concentrations are most commonly measured within 24 hours of the onset of clinical signs, but concentrations can remain increased up to 1 month after presentation of patients with ischemic stroke. ¹⁵ , ¹⁶ Thus, it is unlikely that the time of measurement in our patients substantially affected results.

D‐dimer concentrations in the CSF also have been assessed in human and veterinary medicine for evaluation of CNS disease. Increased concentrations have been associated with immune‐mediated encephalitis, infectious meningitis, neoplasia, and cerebrovascular disease. ²⁵ , ³¹ , ³² , ³³ , ³⁴ , ³⁵ We did not assess CSF D‐dimer concentrations in our study. Instead, we focused on blood D‐dimer concentrations in dogs with MRI evidence of cerebrovascular disease. The most studied and proven use of D‐dimers in human medicine is associated with venous thromboembolism and uses a blood sample. Additionally, blood sampling is less invasive and does not require anesthesia, unlike CSF collection.

The human medical literature indicates that changes in TEG are consistent with hypercoagulability in patients with ischemic stroke and may have some predictive value for classifying ischemic vs hemorrhagic strokes. ²¹ , ²² , ²³ Patients diagnosed with ischemic CVAs were more likely to have TEG changes consistent with hypercoagulability, and certain parameters (K value, angle) reflected severity of cerebral infarction. ²¹ Additionally, parameters such as an R value <5 minutes were associated with increased risk of hemorrhagic transformation of acute ischemic stroke. ²² People diagnosed with cerebral hemorrhage most often had evidence of a hypocoagulable state, although some individuals had results suggesting a hypercoagulable state. ²¹ Thromboelastography also may be used to guide and monitor response to thrombolytic, antiplatelet, and anti‐coagulant treatment. ²³ We were unable to assess for an association between hemorrhagic CVAs and hypocoagulability because no dogs had TEG results consistent with hypocoagulability, and TEG changes in response to anti‐stroke treatment also were not assessed. Unlike in human medicine, a hypercoagulable state based on TEG in dogs was not useful in differentiating between CVA types in this small population of dogs.

Although blood D‐dimer concentrations and TEG have been evaluated in veterinary medicine, their use has been assessed predominantly in systemic disease and thromboembolic conditions other than CVAs. D‐dimer concentrations are used to look for evidence that a thrombotic event has occurred, and have been shown to be increased secondary to pulmonary thromboembolism, disseminated intravascular coagulation, orthopedic surgery, and neoplasia in dogs. ¹¹ , ¹² In contrast, TEG is most useful for identification of hypercoagulability, and abnormalities have been associated with sepsis, hyperadrenocorticism, immune‐mediate hemolytic anemia, and neoplasia. ¹³ , ³⁶ , ³⁷ In these patients, the potential increases in D‐dimer concentration and TEG associated with a variety of pathologic conditions indicate that blood D‐dimer or TEG alone may not be specific enough to be used in diagnosing CVA. We did not control for the effect of systemic disease on test results and although an effect on our patient results cannot be ruled out, it is considered less likely because the included patients were presented for evaluation of neurologic rather than systemic signs.

When comparing ischemic to hemorrhagic strokes, neither abnormal blood D‐dimer concentration nor TEG was found to be significantly associated with a CVA diagnosis. Of note, more hemorrhagic strokes than ischemic strokes were present in the TEG group in our study. This observation was unusual, because ischemic strokes are more common than hemorrhagic strokes in dogs. ¹ , ³⁸ This finding is suspected to be secondary to inclusion of cerebral microbleeds in the hemorrhagic group. Necropsy examinations in both humans and dogs have shown hemosiderin deposits, as well as evidence of both acute and chronic hemorrhage in the region of the microbleeds. ⁷ , ³⁹ Although cerebral microbleeds sometimes are considered an incidental finding, a previous retrospective study showed that dogs with cerebral microbleeds were slightly more likely to present with vestibular signs, in addition to proteinuria and cortical atrophy. ⁷ These dogs also had shorter median survival time compared with age‐ and breed‐matched controls. ⁷ In human medicine, cerebral microbleeds are considered a component of cerebral small vessel disease and often are seen secondary to cerebral amyloid angiopathy, a common cause of cerebral hemorrhage, but are not exclusive to this disease process. ⁴⁰ Furthermore, cerebral microbleeds are considered an imaging biomarker for risk of hemorrhagic and ischemic stroke in people, and have been found to be associated with a higher risk for subsequent intracranial hemorrhage compared to ischemic strokes. ⁴¹

Cerebrospinal fluid collection and analysis were not part of the inclusion criteria for our study and not performed in all dogs. Previous studies have shown variable CSF results in dogs with cerebrovascular disease. Increased protein concentration, especially in cases with multiple strokes, as well as a mild mononuclear or neutrophilic pleocytosis, is common; however normal CSF results are also common. ³⁶ , ⁴² In our study, most cases with MRI evidence of stroke had either normal CSF results, mild pleocytosis, or albuminocytological dissociation, consistent with previous reports. ³⁰

Currently, multiple challenges exist for diagnosing CVAs in dogs with neurologic disease. Magnetic resonance imaging is not always a diagnostic option because of client financial constraints, comorbidities precluding general anesthesia, or lack of availability. Even when MRI is performed, it may be difficult to distinguish between strokes and other intracranial lesions. Some dogs included in our study could have been classified incorrectly as having a CVA or as a control. Although MRI has high sensitivity and specificity for detecting intracranial lesions, differentiating CVA from neoplasia, specifically gliomas, or other parenchymal diseases is not always possible. ⁸ , ⁹ Despite the well‐described characteristics of CVAs on MRI, a retrospective study showed that MRI was only 38.9% sensitive for classifying cerebrovascular disease in veterinary patients, and unique MRI features to distinguish this disease process from other parenchymal lesions have not yet been identified. ² , ⁴ , ⁵ , ⁶ , ⁹ , ¹⁰ Being able to distinguish among intracranial disease processes is important, because treatment options and prognosis vary markedly among different pathologies and is likely to influence owners' decisions for continued treatment vs euthanasia.

Our study had several limitations, including its retrospective nature, different study populations with different CVA prevalence for each diagnostic test, small sample size (especially in neurologic controls for TEG [Group 2]), and the fact that only 1 dog had both D‐dimer concentration and TEG performed. The test selected was based on availability at the participating institutions. The timing of the tests in relation to the onset of clinical signs varied, which could have affected the test results. Moreover, no dogs had histopathologic confirmation of the diagnosis. The inclusion criteria precluded enrollment of all patients with suspected strokes, and thus our population did not reflect the true stroke population. This feature not only limited our population size but also may have selected for patients that were more severely affected neurologically. Furthermore, the control population consisted of dogs with a diagnosis other than CVA, and it is possible that the underlying CNS disease processes in the control population could have affected the TEG or blood D‐dimer concentration results.

In conclusion, blood D‐dimer concentration and TEG alone cannot be used for diagnosis of CVAs in dogs, but D‐dimer concentration may have use as screening test. Future prospective studies evaluating MRI, TEG and D‐dimer concentration using blood and CSF samples in patients with suspect stroke are recommended to further evaluate these diagnostic tests. Biomarkers are wanted to increase accuracy of CVA diagnosis and provide alternative options to advanced imaging for diagnosis of CVAs in veterinary patients.

CONFLICT OF INTEREST DECLARATION

Authors declare no conflict of interest.

OFF‐LABEL ANTIMICROBIAL DECLARATION

Authors declare no off‐label use of antimicrobials.

INSTITUTIONAL ANIMAL CARE AND USE COMMITTEE (IACUC) OR OTHER APPROVAL DECLARATION

Authors declare no IACUC or other approval was needed.

HUMAN ETHICS APPROVAL DECLARATION

Authors declare human ethics approval was not needed for this study.

DiPaola E, Cameron S, Rylander H, Zidan N, Hetzel S. Comparison of D‐dimer concentration and thromboelastography for diagnosis of cerebrovascular accidents in dogs: A retrospective study. J Vet Intern Med. 2024;38(2):1083‐1091. doi: 10.1111/jvim.17000