• Open Access

Fibrinolytic Activity in Cerebrospinal Fluid of Dogs with Different Neurological Disorders

Authors

  • C. de la Fuente,

    1. Departament de Medicina i Cirurgia Animal, Facultat de Veterinària, Universitat Autònoma de Barcelona, Bellaterra, Barcelona, Spain
    2. Fundació Hospital Clínic Veterinari, Facultat de Veterinària, Universitat Autònoma de Barcelona, Bellaterra, Barcelona, Spain
    Search for more papers by this author
  • L. Monreal,

    1. Departament de Medicina i Cirurgia Animal, Facultat de Veterinària, Universitat Autònoma de Barcelona, Bellaterra, Barcelona, Spain
    Search for more papers by this author
  • J. Cerón,

    1. Department of Animal Medicine and Surgery, Veterinary School, CMN “Campus Mare Nostrum”, University of Murcia, Murcia, Spain
    Search for more papers by this author
  • J. Pastor,

    1. Departament de Medicina i Cirurgia Animal, Facultat de Veterinària, Universitat Autònoma de Barcelona, Bellaterra, Barcelona, Spain
    2. Fundació Hospital Clínic Veterinari, Facultat de Veterinària, Universitat Autònoma de Barcelona, Bellaterra, Barcelona, Spain
    Search for more papers by this author
  • J. Viu,

    1. Departament de Medicina i Cirurgia Animal, Facultat de Veterinària, Universitat Autònoma de Barcelona, Bellaterra, Barcelona, Spain
    Search for more papers by this author
  • S. Añor

    Corresponding author
    1. Fundació Hospital Clínic Veterinari, Facultat de Veterinària, Universitat Autònoma de Barcelona, Bellaterra, Barcelona, Spain
    • Departament de Medicina i Cirurgia Animal, Facultat de Veterinària, Universitat Autònoma de Barcelona, Bellaterra, Barcelona, Spain
    Search for more papers by this author

Corresponding author: S. Añor, Universitat Autònoma de Barcelona, Facultat de Veterinària, 08193 Bellaterra (Barcelona), Spain; e-mail: sonia.anor@uab.es.

Abstract

Background

Fibrinolytic activity in cerebrospinal fluid (CSF) is activated in humans by different pathologic processes.

Objectives

To investigate fibrinolytic activity in the CSF of dogs with neurological disorders by measuring CSF D-dimer concentrations.

Animals

One hundred and sixty-nine dogs with neurological disorders, 7 dogs with systemic inflammatory diseases without central nervous system involvement (SID), and 7 healthy Beagles were included in the study. Dogs with neurological disorders included 11 with steroid-responsive meningitis-arteritis (SRMA), 37 with other inflammatory neurological diseases (INF), 38 with neoplasia affecting the central nervous system (NEO), 28 with spinal compressive disorders (SCC), 15 with idiopathic epilepsy (IE), and 40 with noninflammatory neurological disorders (NON-INF).

Methods

Prospective observational study. D-dimers and C-reactive protein (CRP) were simultaneously measured in paired CSF and blood samples.

Results

D-dimers and CRP were detected in 79/183 (43%) and in 182/183 (99.5%) CSF samples, respectively. All dogs with IE, SID, and controls had undetectable concentrations of D-dimers in the CSF. CSF D-dimer concentrations were significantly (< .001) higher in dogs with SRMA than in dogs with other diseases and controls. CSF CRP concentration in dogs with SRMA was significantly (< .001) higher than in dogs of other groups and controls, except for the SID group. No correlation was found between blood and CSF D-dimer concentrations.

Conclusions and Clinical Importance

Intrathecal fibrinolytic activity seems to be activated in some canine neurological disorders, and it is high in severe meningeal inflammatory diseases. CSF D-dimer concentrations may be considered a diagnostic marker for SRMA.

Abbreviations
CNS

central nervous system

CRP

C-reactive protein

CSF

cerebrospinal fluid

IE

idiopathic epilepsy

INF

other inflammatory neurological diseases

NEO

neoplasia affecting the central nervous system

NON-INF

noninflammatory neurological disorders

SCC

spinal cord compressive disorders

SID

systemic inflammatory diseases without CNS involvement

SRMA

steroid-responsive meningitis-arteritis

TNCC

total nucleated cell count

TR-IFMA

time-resolved immunofluorimetric assay

Fibrinolysis is the removal process of intravascular fibrin clots and it is an essential component of normal hemostasis. D-dimers are specific indicators of fibrinolysis and sensitive markers of systemic fibrinolytic activity in humans[1] and animals.[2]

Inflammation and coagulation are closely related processes, and there is an extensive crosstalk between these 2 systems.[3-8] Activation of coagulation by inflammatory processes can result in intravascular fibrin formation that could lead to organ failure. In addition, activation of the coagulation system yields proteases that induce signaling pathways to modulate the inflammatory response through specific cell receptors.[6] Thus, inflammation activates coagulation, and coagulation also affects the inflammatory response.

Normal human CSF contains very low concentrations of fibrinolytic enzymes, but the fibrinolytic activity increases in pathologic conditions to minimize the detrimental effects of fibrin deposition.[9] Although the origin of physiological and disease-induced fibrinolysis in the central nervous system (CNS) remains controversial, the relationship among inflammation, coagulation, and fibrinolytic activity in CNS diseases is worth studying. CSF D-dimer concentrations have been used to assess fibrinolytic activity in the CSF of humans with CNS infectious, neoplastic, traumatic, and vascular diseases.[10-15] In addition, D-dimers in the CSF have been proposed as markers of meningeal inflammation,[16, 17] and as a useful tool to differentiate between idiopathic subarachnoid hemorrhage and traumatic spinal tap.[18-20] Furthermore, in human medicine, the role of fibrin in CNS diseases has been recently redefined not only as a blood-clotting protein but also as a factor that regulates inflammatory and regenerative cellular responses in neurodegenerative diseases.[21] Finally, some studies point to the coagulation system as being critical for tumor progression and as an effective cancer therapy target.[22, 23]

There is no information about the fibrinolytic activity in the CSF of veterinary patients, except for 1 study performed in healthy dogs about the effect of blood contamination on CSF D-dimer concentrations.1

C-reactive protein, a major acute-phase protein synthesized by hepatocytes, has been measured in the CSF and serum of dogs with neurological disorders, and its usefulness as a biomarker for meningeal inflammation has been demonstrated. CRP concentration in canine CSF, although not specific, is an adjunctive tool in the diagnosis of SRMA, and serum CRP concentrations can be used to monitor response to treatment in these patients.[24-26] Time-resolved immunofluorimetric assay (TR-IFMA) has been recently validated in canine CSF samples to measure CRP concentrations. TR-IMFA has been shown to have a high sensitivity, so very small amounts of CRP can be detected in canine CSF samples.[27]

We hypothesized that the intrathecal fibrinolytic activity is activated in disease conditions and that the activation is more marked in neurological inflammatory disorders due to the relationship between inflammation and coagulation.

The main purpose of this study was to assess the fibrinolytic activity in the CSF of dogs with different neurological disorders by measuring D-dimer concentrations in paired blood and CSF samples. In addition, the usefulness of CSF D-dimers as biomarkers of meningeal inflammation was compared with that of CRP, a molecule that is currently being studied as a biomarker of some neurological diseases, especially inflammatory disorders.

Materials and Methods

Study Population

In this prospective study, dogs examined by the Neurology Service of the Hospital Clínic Veterinari (UAB) that had a CSF analysis performed as part of their workup between October 2007 and April 2011 were included. Dogs presented for different neurological signs, and were distributed into 7 groups according to final diagnosis: (1) steroid-responsive meningitis-arteritis (SRMA), (2) other inflammatory neurological diseases (INF), (3) neoplasias primarily arising or secondarily involving the CNS (NEO), (4) spinal cord compressive disorders (SCC), (5) noninflammatory neurological disorders (NON-INF), (6) idiopathic epilepsy (IE), and (7) systemic inflammatory diseases without CNS involvement (SID).

In the SRMA group, diagnosis was reached on the basis of signalment, history, physical and neurological examinations, complete blood work, cervical spinal radiographs, presence of neutrophilic pleocytosis in the CSF with absent microorganisms, and resolving clinical signs with immunosuppressive treatment. Eight dogs were additionally sampled at 6 weeks after treatment initiation (1st monitoring). The treatment protocol used was one previously described.[28]

Diagnostic procedures in all dogs within the INF, NEO, and IE groups consisted of physical and neurological examinations, complete blood work, thoracic radiographs, abdominal ultrasound, magnetic resonance imaging, and CSF analysis. Some dogs in the INF group had also infectious agent testing and histopathology, which allowed further classification into a specific meningoencephalitis (infectious, necrotizing meningoencephalitis, granulomatous meningoencephalitis) or meningoencephalitis of unknown origin. Electrophysiology was also performed in cases of acute idiopathic polyradiculoneuritis. The NEO group included histologically confirmed neoplasms and highly suspected neoplasms on the basis of MRI findings and CSF analysis.

In the SCC group, diagnosis was reached on the basis of signalment, history, physical and neurological examinations, complete blood work, thoracic radiographs, abdominal ultrasound, myelography/magnetic resonance imaging (or both), and CSF examination. Furthermore, the diagnosis was confirmed in all dogs that were surgically treated.

In the NON-INF group, diagnosis was reached on the basis of signalment, history, physical and neurological examinations, complete blood work, thoracic radiographs, abdominal ultrasound, myelography/magnetic resonance imaging (or both), and CSF analysis. Electrophysiology was also performed in dogs with diseases involving peripheral nervous system.

The SID group included dogs with clinical signs that mimicked neurological signs (spinal pain), but that finally had a diagnosis of systemic inflammatory disease without CNS involvement. These dogs had a CSF analysis performed as part of the diagnostic protocol and also had several other diagnostic tests, including complete blood work, urinalysis, synovial fluid analysis, thoracic radiographs, abdominal ultrasound, spinal column radiographs/magnetic resonance imaging (or both), and infectious agent testing.

Dogs were excluded from the study if any anti-inflammatory or immunosuppressive treatment had been given before presentation, or the diagnostic tests were not conclusive and a definitive or highly presumptive diagnosis could not be reached.

Additionally, blood and CSF samples were obtained from 7 control healthy Beagles. Animal care and experimental procedures involving these dogs were approved by the Ethics Committee in Animal and Human Experimentation of the Universitat Autònoma de Barcelona (authorization reference number: 5,719).

Sample Collection and Handling

Blood samples were collected by venipuncture from the jugular or cephalic veins into 1-mL tubes containing 3.8% sodium citrate. Blood samples were centrifuged at 2,000 × g for 10 minutes and plasma was separated within 1 hour of sample collection. CSF was collected aseptically, under general anesthesia, by puncture at the cerebellomedullary cistern or the lumbar cistern. The puncture site was closest caudally to neurolocalization, except for dogs with suspected inflammatory cervical lesions and control dogs, which had all cisternal taps. CSF was analyzed immediately for total nucleated cell count (TNCC), red blood cell count, and protein concentration. CSF was considered normal when the TNCC was <5 cells/μL, the total protein concentration was <25 mg/dL (cerebellomedullary cistern) or <45 mg/dL (lumbar cistern), and the red blood cell count was <4,000 cells/μL.[29] Plasma and CSF samples were aliquoted, stored, and frozen at –70°C until assayed.

D-Dimers

Blood and CSF D-dimer concentrations were determined in duplicate by a quantitative immunoturbidimetric latex agglutination assay2 with commercial reagents and controls3 according to the instructions provided by the manufacturer. The assay had been previously used to assess local fibrinolytic pathway activation in different biological fluids of canine4 and equine origin.[30, 31] Results were reported in ng/mL.

CRP

C-reactive protein was measured in blood and CSF samples, as previously described,[27] using a TR-IFMA method.5-11–k Results were reported in mg/L.

Statistical Analysis

Data analysis was performed by using a commercial software (SPSS for Windows, version 18.0).12 Frequencies and percentages were used for qualitative variables. For descriptive analysis of age of dogs, TNCC, red blood cell count, and CSF total protein concentration, median and minimum and maximum values were used. D-dimer and CRP concentrations, which were not normally distributed, were expressed as median and interquartile range (25–75th percentile). For inferential analysis, a 1-way rank ANOVA with Bonferroni adjustment for multiple comparisons among groups was used. The Spearman's rank correlation coefficient was used to compare results of the main variables in CSF and blood. The Wilcoxon signed rank test was performed for SMRA pre- and posttreatment comparisons. Significance was set at < .05 for all tests.

Results

Dogs Included in the Study

One hundred and eighty-three dogs were included in the study: 11 SRMA, 37 INF, 38 NEO, 28 SCC, 15 IE, 40 NON-INF, 7 SID, and 7 control dogs. One hundred and thirteen dogs were male (62%) and 70 were female (38%). Ages ranged from 0.3 to 14 years (median 6 years). The breed distribution reflected the hospital's referral population with mixed-breeds and 38 different pure breeds represented (Table 1).

Table 1. Breed distribution of dogs included in the study
BreedSRMA (n = 11)INF (n = 37)NEO (n = 38)SCC (n = 28)IE (n = 15)NON-INF (n = 40)SID (n = 7)
  1. SRMA, steroid-responsive meningitis-arteritis; INF, other inflammatory neurological diseases; NEO, neoplasias affecting the central nervous system; SCC, spinal cord compressive disorders; IE, idiopathic epilepsy; NON-INF, noninflammatory neurological disorders; SID, systemic inflammatory diseases without CNS involvement.

  2. Breeds: Alaskan Malamute (AM), American Staffordshire Terrier (AST), Belgian Tervueren (BT), Bernese Mountain dog (BMD), Burdeaux dogue (BD), Border Collie (BC), Borzoi (BZ), Cairn Terrier (CTE), Cotton de Tulear (CT), Dobermann (D), English Bulldog (EB), Fila Brasileiro (FB), Great Dane (GD), Griffon (G), Jack Russel (JR), Labrador Retriever (LR), Maltese (M), Miniature Pinscher (PCH), Miniature Poodle (MP), Miniature Schnauzer (MS), Pitbull (PB), Portuguese water dog (PW), Pug (P), Rottweiler (RW), Samoyed (S), Sharpei (SH), Shih Tzu (SHT), Siberian Husky (SH),Weimaraner (W), West Highland White Terrier (WHWT), Whippet (WP).

German Shepherd211441
Boxer5283
Cocker Spaniel1453
French Bulldog64122
Yorkshire Terrier31311
Golden R213
Beagle312
Mixed breed19752121
Other (n = 2)BC, WHWT, SHTPB, BTD, LRLRWHWT
Other (n = 1)BMD, RWSH, RW, AST, P, CT, M, PCH, LRBD, AM, EB, AST, MP, GP, CT, GD, RW, WP, SH, FBSH, EB, MS, BDSHT, SH, S, W, MP, JR, AST, RW, BD, DCTE, BZ, GD, PW

SRMA

Five of 11 dogs were male (45%) and 6 were female (55%). Ages ranged from 0.3 to 1 year (median 0.9 years).

INF

Twenty-four of 37 dogs were male (65%) and 13 were female (35%). Ages ranged from 1 to 10 years (median 5 years). Inflammatory disorders included were 19 meningoencephalitis of unknown origin, 7 granulomatous meningoencephalitis, 3 necrotizing meningoencephalitis, 2 bacterial meningomyelitis, 2 acute idiopathic polyradiculoneuritis, and one of each of the following: idiopathic eosinophilic meningoencephalitis, Prototheca meningoencephalitis, distemper virus encephalitis, and trigeminal neuritis. In this group, 44% of the final diagnoses were confirmed by histopathologic examination.

NEO

Seventeen of 38 dogs were male (45%) and 21 were female (55%). Ages ranged from 4 to 14 years (median 10 years). Histopathologically confirmed neoplasias included 7 meningiomas (5 intracranial; 2 spinal), 10 intracranial gliomas, 3 peripheral nerve sheath tumors (1 cranial nerve; 2 spinal nerve), and one of each of the following: intracranial primary lymphoma, intracranial multicentric lymphoma, pituitary carcinoma, invasive nasal neuroblastoma, and multicentric plasma cell tumor. Suspected neoplasias on the basis of MRI findings and CSF analysis included 6 intracranial meningiomas, 4 intracranial gliomas, and one of each of these: pituitary neoplasia, medulloblastoma, and vertebral bone neoplasia. In this group, 66% of the final diagnoses were confirmed by histopathologic examination.

SCC

Twenty-one of 28 dogs were male (75%) and 7 were female (25%). Ages ranged from 0.8 to 14 years (median 7.5 years). This group included 19 dogs with Hansen type I intervertebral disk disease, and 9 dogs with caudal cervical spondylomyelopathy. In this group, 75% of the final diagnoses were surgically confirmed.

IE

Twelve of 15 dogs were male (80%) and 3 were female (20%). Ages ranged from 1.5 to 11 years (median 4 years). All dogs in this group had an onset of epileptic seizures between 1 and 5 years, normal interictal neurological examination and unremarkable bloodwork, CSF analysis, and brain MRI.

Non-INF

Twenty of 40 dogs were male (50%) and 20 were female (50%). Ages ranged from 0.5 to 13 years (median 7.5 years). This group included 10 noncompressive traumatic spinal cord injuries, 13 cerebrovascular accidents, 6 fibrocartilaginous embolisms, 4 idiopathic vestibular syndromes, 3 suspected degenerative myelopathies, and one of each of the following: congenital hydrocephalus, Jack Russell Terrier hereditary ataxia, cerebellar abiotrophy, and Dancing Doberman disease.

SID

Six of seven dogs were male (85%) and one was a female (15%). Ages ranged from 0.3 to 4 years (median 1 year). Final diagnoses in this group were 6 nonerosive immune-mediated polyarthritis and 1 immune-mediated polymyositis. All these dogs had normal CSF analysis.

Control Dogs

These were 7 male, 1-year-old Beagle dogs with blood and CSF analysis results within normal ranges.

Routine CSF Analysis

Results of complete CSF analysis were available for 163/183 dogs. Red blood cell counts were <4,000/μL in all samples, and only 11/163 samples had red blood cell counts between 501 and 3,900/μL. Partial CSF analysis results (only total protein concentration) were available for 20/183 dogs and 15 of these had increased total protein concentrations. Five dogs with partial CSF analysis that had CSF total protein concentration within normal ranges were not included in either the normal CSF or altered CSF analysis groups for statistical comparisons, because of lack of CSF cell counts. Thus, finally, 47/178 (26.4%) dogs had normal CSF analysis and 131/178 (73.6%) had altered CSF (either cell count, protein concentration, or both). Results of CSF analysis for each group of dogs are shown in Table 2.

Table 2. Cerebrospinal fluid results for dogs in each group (n = 183)
VariableCD (n = 7)SRMA (n = 11)INF (n = 37)NEO (n = 38)SCC (n = 28)IE (n = 15)NON-INF (n = 40)SID (n = 7)
  1. C, cerebellomedullary cistern; CD, control dogs; IE, idiopathic epilepsy; INF, other inflammatory neurological diseases; L, lumbar cistern; NEO, neoplasia affecting central nervous system; NON-INF, noninflammatory neurological disorders; RBC, red blood cells; SCC, spinal cord compressive disorders; SID, systemic inflammatory disease without CNS involvement; SRMA, steroid-responsive meningitis-arteritis; TNCC, total nucleated cell count.

  2. Data are expressed as median (range). Cerebrospinal fluid cell counts available from 163/183 dogs. Cerebrospinal fluid total protein concentrations available from all dogs included in the study.

TNCC (cells/μL)0 (0–3)649 (60–7290)12 (0–2,900)1 (0–129)0 (0–9)0 (0–4)0 (0–46)0 (0–3)
Total protein (mg/dL)20.2 (17.9–23.4)130.7 (39.4–573)62 (18.2–1,500)41.4 (14.4–392)71 (21–226)22.7 (13–30.8)50.2 (13.5–293.6)17.5 (13–44)
RBC (cells/μL)0 (0–0)45 (0–732)5 (0–1,600)2 (0–3,900)0 (0–400)0 (0–105)5 (0–3,600)1 (0–122)
Puncture siteC (n = 7)C (n = 11)C (n = 29) L (n = 8)C (n = 32) L (n = 6)L (n = 28)C (n = 15)C (n = 24) L (n = 16)C (n = 4) L (n = 3)

Correlations between CSF D-Dimer and CRP Concentrations with CSF Analysis Results

Cerebrospinal fluid D-dimer and CRP concentrations in dogs with altered CSF analysis were significantly (< .001 and = .002, respectively) higher than those of dogs with normal CSF analysis. There were significant, but weak correlations between D-dimer concentrations and TNCC (rho = 0.352, < .01), and D-dimers and red blood cell count (rho = 0.184, = .019). A moderate correlation was found between D-dimers and total protein content (rho = 0.552, < .001). CSF CRP concentrations were significantly but weakly correlated with TNCC (rho = 0.255, = .01), red blood cell count (rho = 0.221, = .005), and total protein content (rho = 0.232, = .002).

Comparison of CSF and Blood D-Dimer Concentrations

All dogs in the control, IE, and SID groups had undetectable D-dimer concentrations in the CSF. Twenty of 37 dogs in the INF group, 23/38 dogs in the NEO group, and 20/40 in the NON-INF group had detectable concentrations of D-dimers in the CSF. In the SCC group, 5/28 dogs had detectable concentrations of D-dimers in the CSF, and the two with the highest values (210 ng/mL and 65 ng/mL, respectively) had a xanthochromic CSF. All dogs in the SRMA group had detectable concentrations of D-Dimers in the CSF, and the values for this group were significantly (< .001) higher than those of other groups and control dogs (Table 3). Dogs in the SID group had the highest blood D-dimer concentrations, but no significant differences were found between groups and controls (Table 3). No correlation was found between blood and CSF D-dimer concentrations.

Table 3. D-dimer and C-reactive protein concentrations in cerebrospinal fluid and blood
VariableCD (n = 7)SRMA (n = 11)INF (n = 37)NEO (n = 38)SCC (n = 28)IE (n = 15)NON-INF (n = 40)SID (n = 7)
  1. CD, control dogs; CRP, C-reactive protein; CSF, cerebrospinal fluid; IE, idiopathic epilepsy; INF, other inflammatory neurological diseases; NEO, neoplasias affecting central nervous system; NON-INF, noninflammatory neurological disorders; SCC, spinal cord compressive disorders; SID, systemic inflammatory diseases without CNS involvement; SRMA, steroid-responsive meningitis-arteritis.

  2. Values expressed as: median (interquartile range). Analyses were performed by the 1-way rank-ANOVA with Bonferroni's adjustments.

  3. a

    Significantly different from all groups and control dogs in the same row (< .001).

  4. b

    Significantly different from all groups and controls except for SID group in the same row (< .001).

  5. c

    Significantly different from controls in the same row (< .05).

(CSF D-dimer) (ng/mL)0 (0–0)283a (217–872)4 (0–39)6.3 (0–41)0 (0–0)0 (0–0)2 (0–15.9)0 (0–0)
(Blood D-dimer) (ng/mL)0 (0–67)83 (16–244)64.5 (6.5–146.2)26 (0–135)110 (19.6–168.1)61 (0–120)52 (8–146.4)122.5 (61–2,064)
(CSF CRP) ×10−3(mg/L)1.6 (0.6–9.6)645.1b (256.9–3,146)13 (4.7–82.1)13.3 (6–98.3)6.3 (2.5–17)7 (2.7–8.8)11.2 (4–56.2)43.5c (30.1–193.5)
(Blood CRP) (mg/L)4 (3–11.3)220.8 b (85–327.1)5.2 (4–42.1)7.7 (4–27.4)5.7 (3.9–15.2)10 (4.7–16.9)11.7 (3.6–45.1)179c (46.9–392.2)

Comparison of CSF and Blood CRP Concentrations

C-reactive protein concentrations were detected in the CSF of 182/183 dogs. CRP concentration was significantly (< .001) higher in the CSF of dogs with SRMA than in dogs from other groups and controls, except for dogs with SID. In fact, the CSF CRP concentration in the SID group was significantly higher than in the SCC (= .038) and the control groups (= .012). No significant differences were observed between any other group and controls (Table 3). The CRP blood concentration was significantly higher in the SRMA group than in all other groups and controls, except for the SID group. The blood CRP concentration in the SID group was significantly higher than that in the SCC (= .013), NON-INF (= .036), and control groups (= .009). No significant differences were observed between any other group and controls (Table 3). There was a significant correlation between the CRP concentrations in blood and CSF (rho = 0.652, = .01).

Correlation between D-Dimer and CRP Concentrations in CSF

There was a significant, but moderate correlation between the concentrations of D-dimers and CRP in CSF (rho = 0.457, < .001).

Follow-up CSF and Blood Samples in the SRMA Group

Cerebrospinal fluid and blood samples from 8/11 SRMA dogs taken 6 weeks after treatment initiation were available. All dogs were in clinical remission and had a CSF analysis within normal range (median TNCC = 2 cells/μL, range: 0–4 cells/μL; median total protein = 21.2 mg/dL, range: 15.4–24.4 mg/dL). Six weeks after treatment initiation, CSF D-dimer concentrations were undetectable in all dogs. Concentrations of D-dimers and CRP in CSF and blood were significantly (< .05) lower than those observed in the same dogs before treatment (Table 4).

Table 4. C-reactive protein and D-dimer concentrations in dogs with steroid-responsive meningitis-arteritis at the moment of diagnosis and 6 weeks after treatment initiation
VariableSRMAdiagnosis (n = 11)SRMA6 weeks (n = 8)
  1. CRP, C-reactive protein; CSF, cerebrospinal fluid; SRMA, steroid-responsive meningitis-arteritis.

  2. Values expressed as: median (interquartile range). Analyses were performed by Wilcoxon signed rank test.

  3. a

    Significantly different in the same row (< .05).

(CSF D-dimer) (ng/mL)283 (217–872)0a (0–0)
(Blood D-dimer) (ng/mL)83 (16–244)4.25a (0–50.3)
(CSF CRP) ×10−3 (mg/L)645.1 (256.9–3,146)0.6a (0.2–2.9)
(Blood CRP) (mg/L)220.8 (85–327.1)3.5a (3–5.1)

Discussion

The results of this study demonstrate that an intrathecal fibrinolytic pathway is activated in the CSF of dogs with some neurological disorders, especially in dogs with SRMA. Our study also demonstrates that the CSF fibrinolytic activity can be measured, and suggests that the local activation of fibrinolysis is not linked to a concurrent systemic hyperfibrinolysis. The results support the use of CSF D-dimer concentrations as a potentially useful diagnostic marker of SRMA.

D-dimers are specific degradation products of cross-linked fibrin generated by the action of plasmin, thus any increase in CSF D-dimer concentration should be a specific indicator of active intrathecal fibrinolysis. In fact, the D-dimer test is being used as a sensitive marker of local fibrinolytic activity in humans and animals, and it is suggested that the fibrinolytic pathways are locally activated to avoid the detrimental effects of fibrin deposition in different body compartments.4,[9-13, 30, 31] Although the kinetics of D-dimers in the CSF of humans and animals are unknown, due to the short half-life of these products in serum and the fast turnover of CSF, a raised CSF D-dimer concentration might be indicative of recent hypercoagulation and hyperfibrinolysis.

Cerebrospinal fluid endogenous fibrinolysis has been studied in humans under normal and pathologic conditions and it is reported that normal human CSF contains very low to undetectable concentrations of fibrinolytic enzymes and D-dimers.[9] Our study suggests that the situation is similar in dogs, because no D-dimers were detected in the CSF of healthy control dogs or in dogs with SID.

Under pathologic conditions, the CSF fibrinolytic activity can be strongly activated in humans, and raised CSF D-dimer concentrations have been observed in some inflammatory,[32] infectious,[12] traumatic,[15] neoplastic,[13] and vascular CNS diseases.[14, 33] In our study, D-dimers were detected in the CSF of all dogs in the SRMA group, and in some dogs of the INF, NEO, NON-INF, and SCC groups. No D-dimers were detected in the CSF of idiopathic epileptic dogs. Our results indicate that intrathecal fibrinolysis is activated in some dogs with different pathologic conditions affecting the CNS, as it happens in humans.

The highest CSF D-dimer concentrations were found in dogs with CNS inflammatory and neoplastic disorders (Table 3). In inflammatory diseases, the relationship between inflammation and coagulation could explain the local activation of the fibrinolytic pathways, as it occurs in systemic inflammatory disorders, which cause increased blood D-dimer concentrations. In dogs with SRMA, the elevated CSF D-dimer concentration could also be related to blood vessel inflammation, increased risk of subarachnoid bleeding, and clot formation. Interestingly, 17/20 dogs with inflammatory disorders (including some dogs with histologically confirmed diagnosis) did not have detectable CSF D-dimer concentrations, which suggests that intrathecal fibrinolysis was not activated in these animals. A potential explanation for this could be that some inflammatory CNS diseases in dogs affect mainly deep parenchymal regions, without or with minimal meningeal involvement. Thus, it is possible that lesion localization has an influence on D-dimer concentrations in the CSF.

In neoplastic disorders, either the inflammatory response induced by the tumor or associated hemorrhages could trigger local fibrinolytic activity. Interestingly, tissue factor, an enzyme cofactor involved in activation of coagulation is overexpressed in a wide variety of human tumors including CNS neoplasias,[22, 23] and has been recently reported in canine tumor cell lines.[34] Although there is no information about the overexpression of tissue factor in canine CNS neoplasms, tissue factor-mediated coagulation activation may play a role in local fibrinolytic activity in these cases.

Half of the dogs in the NON-INF group had detectable concentrations of D-dimers in the CSF. Dogs in this group were mainly animals with noncompressive traumatic spinal cord injuries and cerebrovascular accidents, and most of these had detectable CSF D-dimers, as it occurs in humans with acute cerebrovascular disease and traumatic brain injury.[14, 15]

Only 5/23 dogs in the SCC group had detectable D-dimers in the CSF. Of these, the two with the highest values had a xanthochromic CSF. Thus, local activation of coagulation and fibrinolysis resulting from subarachnoid hemorrhage was thought to be the most likely explanation.

A traumatic spinal tap could also cause an increase in the CSF D-dimer concentration. A previous study performed on healthy dogs did not find statistically significant variations in CSF D-Dimer concentrations before and after contamination with peripheral blood, and found that CSF red blood cell counts <4,000 cells/μL did not have any influence on CSF D-dimer concentration.1 None of the samples included in the study had a red blood cell count above the upper limit of contamination reported previously, thus it seems unlikely that iatrogenic blood contamination could influence our results. In fact, the D-dimer test has been described as a useful tool to differentiate traumatic spinal tap from subarachnoid hemorrhage in humans.[19, 20]

Blood D-dimer concentrations were measured in all dogs, and no statistically significant differences were found between groups. No correlation was detected between blood and CSF D-dimer concentrations either. Only 15/183 dogs included in the study had blood D-Dimer concentrations above the reference ranges for healthy dogs,[2] and only 5 of these 15 dogs had detectable D-dimer concentrations in the CSF. Furthermore, dogs in the SID group had the highest blood D-dimer values, but no fibrinolytic activity in the CSF. These findings suggest that there is intrathecal fibrinolytic activity, and that this is independent from the blood fibrinolytic activity, as it occurs in humans.[9]

When multiple comparisons between groups were performed, dogs in the SRMA group had marked and significantly higher CSF D-dimer values than dogs from all other groups. The source of fibrinolytic enzymes in the CNS remains controversial, and 2 main mechanisms have been proposed: direct leakage through a damaged blood–brain barrier, and existence of an endogenous fibrinolytic system in which endothelial cells of the brain and meninges appear to be the source of fibrinolytic activity.[9] Both mechanisms could play a role in SRMA: direct damage to the meninges and meningeal arteries can alter the blood-CSF barrier allowing direct leakage of fibrinolytic enzymes into the CSF, and severe meningeal inflammation could trigger the endogenous fibrinolytic system. The findings of our study are in agreement with those from human studies, in which fibrin degradation products in the CSF are significantly higher in patients with neurological disorders involving the meninges, and they are correlated with the severity of meningeal inflammation.[16, 17]

C-reactive protein, a major acute phase reactant, has recently been described as a biomarker of neurological diseases, especially inflammatory CNS disorders. CRP has been measured in the CSF of dogs with different neurological disorders and has been proposed as a useful marker for the diagnosis of SRMA.[24-26] For this reason, CRP was also measured in CSF and blood samples of the same patients to assess the relationship between inflammation and coagulation. In our study, CRP was detected in 182/183 CSF samples, which demonstrates the high sensitivity of the TR-IFMA method compared with that of other techniques previously reported.[27] Our study showed that CRP is present in the CSF of healthy dogs, a fact that has not been previously reported. In addition, dogs with SID had increased CSF CRP concentrations. Although the CRP concentration in the CSF of SRMA dogs was significantly higher than that of dogs in other neurological disease groups and controls, there were no significant differences between the CSF concentrations of CRP in SRMA and SID dogs. There was a good and significant correlation between CSF and blood CRP concentrations, which is in agreement with previous reports.[27, 35] In the SID group, 6/7 dogs were diagnosed of nonerosive immune-mediated polyarthritis, a disease that can occur occasionally in conjunction with SRMA.[28, 36] In these animals (that had normal CSF analyses), the increased CSF CRP concentrations could be thought to be due to occult SRMA, because of the waxing and waning course of the disease. In the authors' opinion, a more likely explanation is that CRP passes easily from blood to CSF, so that CSF CRP concentration is directly affected by the CRP blood concentration. Previous studies have found increased CSF CRP concentrations in dogs with sepsis without CNS involvement,[24] and the same findings have been described in people with bacterial infections (patients with pneumonia and urinary tract infections).[35] Thus, the CRP concentration in the CSF seems to be directly affected by the blood CRP concentration, and this would limit its diagnostic specificity for neurological diseases affecting the CNS.[24, 35]

Cerebrospinal fluid D-dimer concentrations had a significant but moderate correlation with CSF CRP concentrations. In addition, when CSF D-dimer and CRP concentrations were measured in 8/11 dogs with SRMA 6 weeks after treatment initiation (dogs free of clinical signs), no fibrinolytic activity was detected in any posttreatment initiation CSF sample, and significantly lower concentrations of CRP were detected in the CSF and blood of these dogs. Our findings suggest that CSF D-dimer concentration can be a reliable marker of meningeal inflammation that is not affected by systemic disease, as it occurs with CRP, so CSF D-dimer concentration might be an adjunctive diagnostic marker for SRMA. Further studies are necessary to evaluate its potential value and to compare it with that of currently used parameters, such as CSF TNCC, total protein, and IgA. In addition, the potential diagnostic value of CSF D-dimers for meningitides of different etiologies remains unknown, but deserves further study.

Our study had some limitations, namely the low number of samples and the heterogeneity of diseases in some groups that could influence some statistical results. Nevertheless, the data presented demonstrate that the intrathecal fibrinolytic system is activated in some canine neurological disorders, especially in SRMA. Future investigations should be carried to elucidate the relevance of the coagulation and fibrinolytic systems, their role as potential therapeutic targets and the value of D-dimers as biomarkers of certain CNS diseases.

Acknowledgments

The authors acknowledge Lluís Gaitero, Sergio Ródenas, Arianna Negrin, Roberto José-López, and Maria Oliveira for their help in sample collection, and Lara Armengou, Mª Ángeles Delgado, Josefa Hernández, and Silvia Martínez for their technical support in sample analysis.

Conflict of Interest: This work was not supported by any grant, and none of the authors of this paper have conflicts of interest that could influence the content of the paper. Part of this study was presented at the American College of Veterinary Internal Medicine Forum, Anaheim, California, June 2010.

Footnotes

  1. 1

    Rossmeisl JH, Troy GC, Inzana KD, et al. Normal and blood-contaminated canine cerebrospinal fluid D-dimer concentrations. J Vet Intern Med 2002;16:369 (abstract)

  2. 2

    Miniquant, Biopool, Trinity Biotech, Wicklow, Ireland

  3. 3

    Miniquant-1, Biopool, Trinity Biotech

  4. 4

    Escanilla N, Leiva M, Peña MT, Monreal L. Aqueous humor fibrinolytic activity in dogs with ocular disease. Vet Ophthalmol 2010;13:354 (abstract)

  5. 5

    Sulfo-NHS-Biotin, Pierce Biotechnology Inc, Rockford, IL

  6. 6

    DELFIA streptavidin microtitration strips, PerkinElmer, Wallac Oy, Turku, Finland

  7. 7

    DELFIA/AutoDELFIA wash concentrate, PerkinElmer

  8. 8

    DELFIA assay buffer, PerkinElmer

  9. 9

    DELFIA Eu-labeling kit, PerkinElmer

  10. 10

    DELFIA enhancement solution, PerkinElmer

  11. 11

    VICTOR2 1420 multilabel counter, PerkinElmer

  12. 12

    SPSS Inc, Chicago, IL

Ancillary