Cerebrospinal Fluid Tissue Factor and Thrombin-Antithrombin III Complex as Indicators of Tissue Injury After Subarachnoid Hemorrhage
Background and Purpose No marker that reflects and predicts brain injury due to subarachnoid hemorrhage (SAH) and cerebral vasospasm has been reported. We hypothesized that membrane-bound tissue factor (mTF) and thrombin–antithrombin III complex (TAT) in the cerebrospinal fluid (CSF) of patients with SAH become markers indicating brain injury. To evaluate the hypothesis, we correlated levels of mTF and TAT in the CSF of patients with SAH with clinical severity, the degree of SAH, and outcome.
Methods We assayed CSF mTF, TAT and myelin basic protein (MBP) in patients with SAH at intervals that included days 0 to 4 and days 5 to 9 after ictus. Classification of clinical severity of disease on admission was based on Hunt and Hess grade, degree of SAH on CT on Fisher’s grading, and outcome 3 months after SAH on the Glasgow Outcome Scale.
Results In the interval from days 0 to 4, mTF and TAT correlated with Hunt and Hess and Fisher grades, and occurrence of cerebral infarction due to vasospasm. Only mTF correlated significantly in this period with outcome. TAT, mTF, and MBP all correlated significantly with each other. From days 5 to 9, only mTF correlated with cerebral infarction, infarction volume, MBP levels, and outcome.
Conclusions Both mTF and TAT reflected brain injury from SAH and predicted vasospasm, though mTF was more sensitive and a better predictor of outcome. Unlike mTF, TAT did not correlate with vasospasm during the interval when it most commonly occurs, which raised doubt about thrombin activation as a cause.
Recently, TF has been reported in human and nonhuman primate brain.1 2 3 TF is the primary initiator of the coagulation cascade in which it activates thrombin.4 5 We therefore tried to determine whether TF and TAT, a molecular marker of thrombin formation, may be markers of brain injury due to SAH and cerebral vasospasm. We analyzed CSF samples from patients within a few days after SAH for TF and TAT and correlated these markers with severity of neurological deficit on admission, degree of SAH by CT, cerebral infarction due to vasospasm, and 3-month outcome. We also measured CSF TF and TAT concentrations during the period when cerebral vasospasm predominantly occurs. As thrombin activation is considered an etiologic factor in vasospasm,6 we also used the markers to examine that hypothesis. We used MBP, an established marker of tissue injury in various neurological diseases,7 8 to aid in evaluation of TF and TAF as markers of brain injury.
Subjects and Methods
After exclusion of SAH patients who died immediately after admission (4 patients) or underwent direct surgery for the aneurysm more than 48 hours after SAH onset because of late admission (4 patients), or demonstrated brain damage suspected of being caused by the brain trauma and/or vessel occlusion during surgery (3 patients), or who received antiplatelet drugs and/or calcium antagonists (2 patients), we studied 19 patients who suffered from SAH and underwent aneurysm surgery within 48 hours of SAH combined with continuous ventricular or cisternal drainage in the Toyama Medical and Pharmaceutical University Hospital from April 1990 to January 1992. Our study group consisted of 8 men and 11 women, aged 40 to 83 (mean, 61) years. The time of onset of SAH was determined from the detailed history. All patients included in the study had angiographically verified aneurysms. Patients were classified clinically by the system of Hunt and Hess.9 SAH on the admission CT was graded according to the method of Fisher et al.10 Clinical severity on admission was classified into the following five grades: grade 1 (H-H 1), asymptomatic, or minimal headache and slight nuchal rigidity; grade 2 (H-H 2), moderate to severe headache, nuchal rigidity, no neurological deficit other than cranial nerve palsy; grade 3 (H-H 3), drowsiness, confusion, or mild focal deficit; grade 4 (H-H 4), stupor, moderate to severe hemiparesis, possibly early decerebrate rigidity and vegetative disturbances; and grade 5 (H-H 5), deep coma, decerebrate rigidity, and moribund appearance. The degree of SAH was classified into the following four grades: group 1, no blood detected; group 2, a diffuse deposition or thin layer, with all vertical layers of blood (interhemispheric fissure, insular cistern, ambient cistern) <1 mm thick; group 3, localized clot and/or vertical layers of blood ≥1 mm in thickness; and group 4, diffuse or no subarachnoid blood, but with intracerebral or intraventricular clots.
The adjoining gyrus, such as the superior temporal gyrus or rectus gyrus, was not resected during surgery on the aneurysm. The diagnosis of ruptured aneurysm was confirmed intraoperatively in all cases and included the following sites: internal carotid artery (7 patients), anterior communicating artery (6), vertebrobasilar circulation (3), middle cerebral artery (2), and anterior cerebral artery (1). No patients received intrathecal fibrinolytic therapy, antiplatelet agents, or prophylactic calcium channel blockers. Of the 19 patients studied, 8 showed delayed neurological deterioration such as hemiparesis or aphasia 5 to 9 days after SAH. Although CT did not show cerebral infarction in the early stage of SAH, follow-up scanning 6 to 10 days after SAH revealed cerebral infarction in 6 patients, 2 of whom were asymptomatic. Additional studies with transcranial Doppler ultrasonography, cerebral angiography, or 99mTc-d,l-hexa- methyl-propyleneamine oxime single-photon emission CT (99mTc-HMPAO SPECT) demonstrated that infarction in all 6 patients was caused by vasospasm following SAH. Outcome was assessed 3 months after SAH using the Glasgow outcome scale.11
Eighteen and 19 CSF specimens were sampled from patients with SAH between days 0 and 4 (ie, the early stage of SAH) and days 5 and 9 (ie, time of cerebral vasospasm after ictus), respectively. (One patient’s CSF specimen could not be obtained from days 0 to 4 due to a technical problem.) All samples were obtained postoperatively. CSF specimens were sampled from two control groups. One control group (control 1) consisted of 14 patients suspected of having cervical spondylosis who underwent myelography that demonstrated no abnormalities. A second control group (control 2) included 3 patients with arteriovenous malformations and 2 with brain tumors who underwent surgery and ventricular or spinal drainage between days 0 and 4 after surgery.
Measurement of TF, TAT and MBP in CSF
Serial CSF samples were obtained from either continuous ventricular drainage (1 patient) or cisternal drainage (18). Samples were centrifuged at 1500g at 4°C. The supernatant was stored at −80°C until assayed.
CSF TF was measured according to a sensitive method described by Nakamura et al12 13 by use of murine anti-human TF monoclonal antibodies HTF-K14 and HTF-K180. Soluble TF and mTF were fractionated by phase separation with Triton X-114.12 13 The TAT concentration of each sample was measured with an enzyme-linked immunosorbent assay14 using the Enzygnost-TAT (Behring Werke). MBP in CSF was measured with a radioimmunoassay kit (Diagnostic Systems Laboratories).
CT Measurement of Cerebral Infarct Volume
Volumes of cerebral infarcts were calculated by CT, with a slice thickness of 5 mm. On serial slices in one direction, the low-density area due to cerebral infarction was outlined with the cursor, and the volume of cerebral infarction was calculated by integration of the cross-sectional surfaces of the outlined areas.
Findings were reported as mean±SD. For statistical analysis, we used Dunnett’s test14 and Wilcoxon’s U test. Relations among mTF, TAT, and MBP were evaluated with Pearson’s correlation coefficient. A value of P<.05 was accepted as statistically significant.
CSF mTF and TAT of Controls and SAH Patients
A statistically significant difference was evident between patients with SAH and not only control group 1 but also control group 2 in CSF mTF concentration between days 0 and 4 (Dunnett’s test; Table 1⇓). Levels of mTF from days 5 to 9 were similar to those in two controls (Table 1⇓).
During the interval from days 0 to 4 the CSF TAT value was greater in patients than in the two controls (Dunnett’s test) (Table 1⇑).
Correlation of CSF mTF and TAT With Clinical Severity on Admission
Concentrations of mTF from days 0 to 4 after SAH correlated with H-H grade. Although the data of the H-H 4 to 5 group showed the wide distribution, significant differences were noted in mTF between the H-H 1 to 2 and H-H 4 to 5 groups (Dunnett’s test) and between the H-H 3 and H-H 4 to 5 groups (Dunnett’s test; Table 2⇓). TAT for the same interval also correlated with H-H grade. Although the data of H-H 4 to 5 showed wide scatter, significant differences were noted in TAT between the H-H 1 to 2 and H-H 4 to 5 groups (Dunnett’s test; Table 2⇓) and between the H-H 3 and H-H 4 to 5 groups (Dunnett’s test; Table 2⇓). During the intervals from days 5 to 9 after ictus, no difference was detected in mTF and TAT levels among the three H-H groups (Table 2⇓).
Correlation of CSF mTF and TAT Concentrations With Degree of SAH on Admission CT
A correlation was found between mTF and Fisher grade for the interval of days 0 to 4 after SAH. Although Fisher group 4 showed wide scatter of data, significant differences in mTF were found between Fisher group 1 to 2 and Fisher group 4 (Dunnett’s test) and between Fisher groups 3 and 4 (Dunnett’s test; Table 2⇑). TAT for the same interval also correlated with Fisher grade. Although the data of the Fisher 4 group showed wide scatter, significant differences in TAT concentration existed between Fisher group 1 to 2 and Fisher group 4 (Dunnett’s test) and between Fisher groups 3 and 4 (Dunnett’s test; Table 2⇑). For the interval of days 5 to 9 after SAH, no difference was detected in mTF and TAT levels among the three Fisher groups (Table 2⇑).
Correlation of CSF mTF and TAT Concentrations With Cerebral Infarction Due to Cerebral Vasospasm
Statistically significant differences were found in mTF and TAT levels for the interval of days 0 to 4 between the cerebral infarction and noninfarction groups (Wilcoxon’s U test; Table 2⇑). For days 5 to 9, a significant difference in mTF was evident between the two groups (Wilcoxon’s U test; Table 2⇑), whereas no significant difference in TAT existed between the two groups (Table 2⇑). Furthermore, significant differences were detected in the noninfarction group between intervals of days 0 to 4 and 5 to 9 (Wilcoxon’s U test; Table 2⇑). However, no significant changes in mTF occurred over time in the cerebral infarction group (Table 2⇑). Although mTF decreased with time in the non–cerebral infarction group, high levels of mTF persisted in the cerebral infarction group. There was a significant decrease in TAT with time in both the cerebral infarction and noninfarction groups (Wilcoxon’s U test; Table 2⇑).
Correlation of CSF mTF and TAT Concentrations With Outcome
Concentrations of mTF for days 0 to 4 after SAH correlated with Glasgow Outcome Scale. Significant differences in mTF concentration existed between the good GR group (n=6) and a combined group (SD/VS/D; n=8), including SD (n=1), VS (n=1), and D (n=6) (Dunnett’s test), and between the MD (n=4) and SD/VS/D groups (Dunnett’s test; Table 2⇑). Although the SD/VS/D group appeared to show larger values of TAT, no statistically significant differences were detected among the three groups (Table 2⇑). On the other hand, for the days 5 to 9 interval, concentration of mTF correlated with outcome comparing the GR and SD/VS/D groups (Dunnett’s test), whereas no significant correlation was detected between TAT and outcome (Table 2⇑).
Correlation Among CSF mTF, TAT and MBP Concentrations
For the interval days 0 to 4 after SAH, the correlations between mTF and TAT, between mTF and MBP, and between TAT and MBP were statistically significant (Table 3⇓). For days 5 to 9, significant correlation was observed only between mTF and MBP (Table 3⇓).
When patients were divided into two groups according to CSF MBP values, the mTF concentrations for days 0 to 4 were 1240±810 pg/mL (n=6) in the group >100 ng/mL and 281±171 pg/mL (n=12) in the group <100 ng/mL. TAT concentrations for days 0 to 4 were 3960±3660 (n=6) in the group >100 ng/mL and 1190±871 ng/mL (n=12) in the group <100 ng/mL. Statistically significant differences were found in mTF and TAT levels between the two groups (P<.05 and P<.05 by Wilcoxon’s U test). For days 5 to 9, mTF values were 676±167 pg/mL (n=3) in the group >100 ng/mL and 107±66.8 pg/mL (n=16) in the group <100 ng/mL, with TAT concentrations for days 5 to 9 at 408±245 ng/mL (n=3) in the group >100 ng/mL and 299±259 ng/mL (n=16) in the group <100 ng/mL. A significant difference in mTF, then, was evident between the two groups (P<.01, Wilcoxon’s U test), whereas no significant difference in TAT existed between the two groups.
Correlation of CSF mTF With Volume of Cerebral Infarction Due to Vasospasm
The correlation coefficient between CSF mTF concentration for the interval of days 5 to 9 and the volume of cerebral infarction due to vasospasm in 6 patients was .824 (y=3.02x+69.4, P<.05). CSF mTF concentration correlated significantly with extent of cerebral infarction.
The brain is rich in TF coagulant activity,1 and TF has been localized in human and nonhuman primate brain tissue in immunohistochemical studies.1 2 3 However, until recently very little information has been available on TF concentration in the CSF of normal subjects or patients with diseases of the central nervous system. Recently, a sensitive enzyme immunoassay was developed to measure both soluble TF and mTF.12 13 Using this method, we measured both types of TF in the CSF of control subjects and patients with SAH. When we measured TF in plasma and CSF in normal subjects, the total TF concentration was five times greater in CSF than in plasma, while no difference was evident in the proportion of TF in CSF and plasma that was membrane bound (data not shown). Concentrations of mTF increased after onset of SAH and mTF values at an early stage after SAH correlated with clinical severity on admission, SAH extent on admission CT, occurrence of cerebral infarction due to vasospasm, and 3-month outcome. Furthermore, mTF correlated significantly with MBP, a marker of brain damage. Therefore, it is believed that mTF was released from the brain due to SAH injury, with its concentration paralleling the degree of tissue injury. Although no statistical difference was detected, surgery itself had a tendency to induce increased CSF mTF. Therefore, the artifactual contribution of surgery should be considered in some cases. Interestingly, mTF levels at the time of vasospasm were higher in patients with cerebral infarction due to vasospasm, and mTF decreased less with time in these patients, reflecting increased release of mTF into the CSF. CSF mTF concentration also correlated with the extent of cerebral infarction. Therefore, mTF in the CSF may be a useful marker of brain injury due to SAH and vasospasm after SAH.
TF is a cytokine receptor-like glycoprotein composed of extracellular, membrane, and cytoplasmic domains and has a molecular weight of 46 kDal.16 It has been known that a free form of receptor exists in plasma or other body fluids.17 A proteolytic shedding or release of vesicles from live or damaged cells is the probable source of these free (soluble) receptors. Phase separation using Triton X-114 was used to separate membrane-bound and soluble proteins.18 This method was used to fractionate a recombinant TF in which the membrane domain was deleted (soluble) and a recombinant wild-type TF (mTF), and applied to analyses of plasma samples.12 13 TF complexes tightly with coagulation factor VIIa and serves as the cofactor for factor VIIa–dependent factors IX and X.4 5 Soluble TF lacks protein procoagulant activity, while mTF has potent procoagulant activity; release of mTF from brain injury after SAH may initiate coagulation disorders such as disseminated intravascular coagulation (DIC).12 13
Recently, Itoyama et al19 reported that plasma values of TAT, a molecular marker of thrombin activation, correlated in the early stage with clinical severity at onset and with outcome.19 Therefore, we assessed the relation between CSF TAT and clinical severity at onset, occurrence of cerebral vasospasm, and outcome. Although no significant correlation was detected between CSF TAT values and outcome 3 months after SAH, CSF TAT concentration in the early period after SAH correlated significantly with clinical grade on admission, SAH on admission CT, and occurrence of cerebral infarction due to cerebral vasospasm. Furthermore, TAT correlated with CSF mTF and MBP. As mTF has potent procoagulant activity, increased TAT soon after SAH may be secondary to mTF.
Some investigators have reported hypercoagulability and disseminated intravascular coagulation associated with SAH.20 21 Disseminated intravascular coagulation associated with SAH has been thought to predispose to cerebral vasospasm. Thrombin promotes endothelin-1 gene expression,22 release of serotonin and platelet-derived growth factor from platelets, and chemotaxis and aggregation of neutrophils.23 Endothelin-1, serotonin, and platelet-derived growth factor are potent vasoconstrictors, and the role of inflammation has been emphasized in the pathogenesis of vasospasm after SAH.24 25 Therefore, thrombin activation has been hypothesized to be the initial cause of vasospasm. However, CSF TAT values during the interval when vasospasm usually occurs had no evident correlation with the occurrence of cerebral infarction due to vasospasm.
In conclusion, these results suggest that CSF mTF and TAT values in early stages after SAH may predict severity of brain injury due to SAH and occurrence of cerebral infarction due to vasospasm. CSF mTF during the period when cerebral vasospasm usually occurs also may reflect the severity of cerebral infarction due to vasospasm. Therefore, these CSF markers will contribute to the supplement of clinical features known to predict the occurrence of cerebral vasospasm and the prognosis. Further studies are required to better assess the relation between thrombin activation and the occurrence of cerebral vasospasm after SAH.
Selected Abbreviations and Acronyms
|H-H||=||Hunt and Hess grade|
|MBP||=||myelin basic protein|
|mTF||=||membrane-bound tissue factor|
|TAT||=||thrombin-antithrombin III complex|
© 1997 American Heart Association, Inc.
- Received January 7, 1997.
- Revision received June 18, 1997.
- Accepted June 18, 1997.
- Copyright © 1997 by American Heart Association
Suzuki K, Ogawa A, Sakurai Y, Nishino A, Uenohara K, Mizoi K, Yoshimoto T. Thrombin activity in cerebrospinal fluid after subarachnoid hemorrhage. Stroke.. 1992;23:1181-1182.
Nakamura S, Kamikubo Y, Okajima K, Asakura H, Nakamura K. Plasma tissue factor: its sensitive assay and characterization. Thromb Haemost.. 1993;69:744. Abstract.
Nakamura S, Kamikubo Y. Tissue factor [in Japanese]. J Med Technol.. 1994;38:925-932.
Fernandez-Bortran R. Soluble cytokine receptors: their role in immunoregulation. FASEB J.. 1991;5:2567-2574.
Bordier C. Phase separation of integral membrane proteins in Triton X-114 solution. J Biol Chem.. 1981;256:1604-1607.
Ettinger MG. Coagulation abnormalities in subarachnoid hemorrhage. Stroke.. 1970;1:139-142.
Kassell NF, Sasaki T, Colohan ART, Nazar G. Cerebral vasospasm following aneurysmal subarachnoid hemorrhage. Stroke.. 1985;16:562-572.