Amyloid beta 1-42 and tau in cerebrospinal fluid after severe traumatic brain injury
Citation Manager Formats

Abstract
Objective: To determine whether CSF amyloid beta 1-42 (Aβ-42) and tau have predictive value for prognosis after head injury.
Methods: CSF samples were collected from 29 patients with severe head trauma between 1 and 284 days post-trauma. Aβ-42 and tau levels were measured using sandwich ELISA techniques and compared with CSF levels in patients with cognitive disorders and headache.
Results: At all time points, concentrations of Aβ-42 were significantly lower in patients with traumatic brain injury (TBI) than in control groups. A significant correlation existed for Aβ-42 levels and outcome of patients. Below a cutoff of 230 pg/mL, the sensitivity of Aβ-42 to discriminate between good outcome (Glasgow Outcome Score 4 and 5) and poor outcome (Glasgow Outcome Score 1 through 3) was 100% at a specificity of 82%. CSF tau levels were significantly higher in patients with TBI than in any control group. In patients with multiple CSF samples collected at various time points between 1 and 32 days after the trauma, tau levels increased early after TBI, peaked in the second week post-trauma, and slowly decreased thereafter. Independent of outcome, all patients had normal tau levels when CSF was collected more than 43 days post-trauma.
Conclusions: Aβ-42 and tau may play a potential role in the pathophysiology of TBI. Furthermore, the results of this study suggest that Aβ-42 may be a supportive early predictor for recovery after severe head injury.
Traumatically evoked brain injury is a major cause of morbidity and mortality.1 Early assessment of recovery after traumatic brain injury (TBI) can have a major influence on decision making with regard to the level of care and services to be provided. In addition to neurodiagnostic tests, various laboratory examinations of CSF and serum samples were performed in a quest for improved prediction of recovery from severe TBI. For example, recent reports have shown that proinflammatory cytokines, excitatory amino acids, and the astroglial protein S-100B may be useful additional predictors of outcome following TBI.2-6⇓⇓⇓⇓ In the current study, we measured Aβ-42 and tau in patients with severe TBI and in controls and examined whether these biomarkers can predict clinical outcome after TBI.
Diffuse axonal injury (DAI) is a primary feature of severe head trauma that results from shear strain forces caused by rotational acceleration of the head during impact.7 DAI is one of the most frequent causes of morbidity and mortality after TBI and is difficult to detect with imaging techniques.8 Tau proteins are structural microtubulus-binding proteins primarily localized in the axonal compartment of neurons. There is evidence that tau levels are elevated more than 1,000-fold in patients with head trauma and that clinical improvement is related to decreased CSF tau levels after TBI.9
Severe head trauma is a recognized risk factor for AD, more than doubling the risk for developing AD.10 Moreover, postmortem examinations of short-term survivors of severe head injury have shown deposition of Aβ in the brain, mainly as Aβ-42.11 In addition, postmortem examinations of boxers’ brains with dementia pugilistica have demonstrated the presence of diffuse Aβ plaques,12,13⇓ and a recent report has shown that Aβ deposition occurs after repetitive mild brain trauma in a transgenic mouse model of Alzheimer amyloidosis.14 We hypothesized that similar to AD,15 amyloid beta 1-42 (Aβ-42) levels are decreased in CSF due to amyloid deposition in the extracellular compartment of the brain following TBI.
Materials and methods.
Sample criteria.
CSF samples were collected between 1998 and 2001 from 29 hospitalized patients with severe head injury {Glasgow Coma Scale [GCS] <8, TBI group} and from two control groups. GCS was determined by specially trained emergency medical physicians immediately on arrival at the scene of the accident. Patients with TBI were identified by neurologists from their clinical history, neurologic examination, and CT or MRI scans showing acute traumatic brain lesions. Several diagnostic procedures were conducted to exclude additional severe body injuries. Outcome of patients with TBI was routinely evaluated by the treating physician team using the Glasgow Outcome Scale (GOS)16 on dismissal from our hospital. One control group consisted of 19 patients with cognitive disorders (dementia group), including six patients with AD, seven with vascular dementia, two with frontotemporal dementia, and four with normal pressure hydrocephalus. The second control group consisted of 12 patients with headache who had undergone lumbar puncture (LP) to rule out subarachnoid hemorrhage (n = 5) or meningitis (n = 7) and had normal CSF cell counts and protein profiles (headache group).
In 15 patients from the TBI group, CSF was collected from intraventricular catheters that had been surgically placed for continuous intracranial pressure monitoring. In the remaining 14 patients from this group and in all patients from the control groups CSF was collected by LP. In patients with TBI, LP was performed to exclude a CNS infection. CSF and blood samples were collected during routine clinical care by physicians at the neurologic intensive and intermediate care units of our department (first referral hospital) and immediately forwarded to the CSF Laboratory, where samples were centrifuged at 2,000 g for 10 minutes within 4 hours of collection. Routine analysis—i.e., cell count, cytology, and albumin quotient—was performed in the course of clinical care. Remaining CSF was kept at −20 °C in polypropylene tubes for research purposes. The current study is a retrospective analysis of these CSF samples.
Analysis of CSF samples.
Kits from Innogenetics NV (Ghent, Belgium) were used to determine CSF Aβ-42 and tau. Levels of Aβ-42 and concentration of total tau were measured by sandwich ELISA techniques (Innotest hTAU-Ag and Innotest β-amyloid 1-42, Innogenetics NV) as described elsewhere.17 Interassay variability of the Aβ-42 test is less than 10%; detection limit is 50 pg/mL with a linearity range of 200 to 1,500 pg/mL. For samples with optical density values below the standard range of measurement (75 to 1,250 pg/mL for tau and 125 to 2,000 pg/mL for Aβ-42), the sample concentration was arbitrarily set at 50% of the lowest values of the standard dilution range. When tau levels exceeded the standard range, samples were re-run at a dilution of 1:5. No Aβ-42 levels above the standard range were found in this study.
Statistical methods.
Sample size was prespecified to n = 30 per group based on previous measurements of Aβ-42. Assuming a minimum difference of 100 pg/mL and a SD of 130 pg/mL, the detection of this difference will have 83% power at a significance level of 5%. Distribution of groups was analyzed by the Kolmogorov-Smirnov test. Because distribution was non-normal, Aβ-42, tau levels, and patient characteristics are given as medians and ranges (minimum, maximum) and nonparametric tests were used. The Kruskal-Wallis test was used to compare these variables among patients with TBI, dementia, and headache. Spearman correlation coefficient was used to evaluate associations between Aβ-42 and tau and GCS, GOS, CSF cell count, albumin quotient, age, and interval between TBI and CSF collection. Owing to the limited number of patients, only univariate analysis was performed. p Values <0.05 were considered to indicate significance. No corrections for multiple comparison were applied.
Results.
Patient characteristics.
Most of the patients with severe TBI were males (n = 27) aged 15 to 72 years, with a median age of 41 years. Initial GCS scores ranged from 3 to 8 (median 5). In 10 patients, the dura mater was damaged; in 3 of the cases, an otoliquorrhoe was detected; and in 2 of the cases, a rhinoliquorrhoe was detected. Most patients (n = 26) had signs of traumatic brain lesion on CT scans. Additional MRI was performed in 18 patients, 7 of them showing signs of diffuse axonal injury with diffuse white matter lesions predominantly in the corpus callosum and dorsolateral upper brainstem. Four patients underwent craniotomy for evacuation of an epidural or subdural hematoma. CSF was collected between 1 and 284 days post-trauma (median 47 days). Patients stayed at our hospital between 4 and 123 days (median 76 days) after trauma; 5 patients were re-admitted to our department for re-evaluation. Recovery was evaluated on discharge using the GOS. Four patients died during hospitalization (GOS 1). Of the 25 surviving patients, 9 remained in a vegetative state (GOS 2), 9 were severely (GOS 3) and 6 moderately (GOS 4) disabled, and 1 patient recovered completely (GOS 5).
CSF sample characteristics.
Samples were stored for a mean of 1.9 years. There was no significant difference in storing time between TBI and control groups (TBI group mean 1.85 years, control groups mean 1.95 years).
Aβ-42.
Concentrations of Aβ-42 were significantly lower in patients with TBI than were those in any control group (figure 1). No significant difference was found between Aβ-42 levels of patients with dementia or headache. Aβ-42 levels correlated neither with the interval between head trauma and CSF collection date (R = −0.22; p = 0.91) nor with initial GCS (R = 0.24; p = 0.08). Correlation between Aβ-42 and the albumin quotient was observed, however (R = −0.23; p = 0.03). Longitudinal Aβ-42 determination performed in two patients with TBI showed no significant trends for Aβ-42 levels (figure 2A). Aβ-42 levels correlated with the age of patients in the TBI group (R = 0.34; p = 0.008). However, a significant correlation existed for Aβ-42 levels and outcome of patients (R = 0.82; p = 0.001; figure 3A). Below a cutoff of 230 pg/mL, the sensitivity of Aβ-42 for poor outcome (GOS 1 through 3) was 100% at a specificity of 82%.
Figure 1. CSF Aβ-42 levels (pg/mL) for traumatic brain injury (TBI), dementia, and headache diagnostic groups.
Figure 2. CSF Aβ-42 (A) and tau levels (B) in two patients with traumatic brain injury (TBI) with consecutive CSF collections. Similar Aβ-42 levels were observed at all time points (A). Tau levels increased rapidly after TBI, peaked within the second week post-trauma, and slowly decreased thereafter (B). Outcome was poor in both patients. ⋄ = Patient 1; ▪ = Patient 2.
Figure 3. Levels of CSF Aβ-42 (A) and tau (B) are correlated with clinical outcome (Glasgow Outcome Scale [GOS]). A correlation was observed between Aβ-42 levels and GOS (p < 0.001). Tau levels did not correlate with outcome (p > 0.05).
Tau.
Concentrations of tau were significantly higher in the TBI than in the dementia and headache groups (figure 4). No significant difference was found between tau levels of patients with dementia or headache. A significant negative correlation was seen with the interval between head trauma and CSF collection (R = −0.42; p = 0.02). In addition, longitudinal tau protein determination was performed in two patients with TBI with poor outcome (GOS 2 and 3). In both patients tau levels increased rapidly after TBI, peaking between days 5 and 15 post-trauma and slowly decreasing thereafter (figure 2B). All patients with TBI had normal tau levels when CSF was collected more than 43 days post-trauma (median 153 pg/mL). No correlations were found between CSF tau and initial GCS (R = 0.01; p = 0.98) or patient age (R = −0.1; p = 0.45), whereas CSF cell counts (R = 0.43; p = 0.001) and protein levels (R = 0.46; p = 0.001) correlated significantly with tau levels. Furthermore, no significant correlation was observed with outcome after TBI (R = −0.25; p = 0.19; figure 3B), even when only those patients (n = 8) with CSF collection between days 5 and 15 post-trauma, when tau levels were supposed to be highest, were analyzed (p = 0.07).
Figure 4. CSF tau levels (pg/mL) for traumatic brain injury (TBI), dementia, and headache diagnostic groups.
Discussion.
The results of the current study show that Aβ-42 is significantly decreased in CSF after severe TBI, and also demonstrate an association between low Aβ-42 and poor early outcome after TBI. In contrast to Aβ-42, CSF tau is highly elevated when measured at early time points post-trauma. In our series, tau decreased rapidly after trauma, and normal CSF levels were seen when samples were taken more than 6 weeks post-trauma. Severity and outcome of TBI did not significantly correlate with tau levels.
The results indicate that Aβ-42 plays a possible pathophysiologic role in severe brain injury. Postmortem examinations of short-term survivors of severe brain injury showed Aβ-42 deposition.11 Cerebral accumulation of Aβ-42 occurs in all patients with AD.15 Similar to our results in patients with TBI, CSF Aβ-42 levels are lower in AD as compared to control subjects.18-20⇓⇓ Low CSF levels may reflect an increased recruitment of Aβ-42 from the CSF and the brain interstitial fluid to cerebral deposits.18 However, our results stand in contrast to a previous study, which found increased CSF Aβ-42 levels in six patients within the first week after severe TBI.21 This discrepancy may be due to different determination techniques and different outcomes after TBI.
Interestingly, Aβ-42 levels significantly correlated with early outcome after TBI in our study. This observation raises the question whether Aβ-42 deposition in the brain is involved in neuronal damage after TBI. In this study, which was based on residual CSF samples frozen for research purposes, outcome was assessed using the GOS, which does not permit discrimination between motor function disturbances and cognitive disorders. Future studies with a prospective study design including cognitive tests must be conducted to solve this question.
The positive correlation between CSF Aβ-42 and patient age may be explained by the fact that older patients had less severe trauma and therefore had higher Aβ-42 levels as compared to younger patients in our study (median GCS score 7 in patients >50 years vs median GCS 5 in patients <50 years), because CSF Aβ-42 values do not correlate with age in healthy adults.22 In this study we found a negative correlation between the albumin quotient and CSF Aβ-42 levels. High CSF protein levels are a consequence of disruption of the blood–brain barrier or reduced CSF circulation occurring after TBI. It seems likely that the severity of TBI influences both the degree of blood–brain barrier disruption and deposition of Aβ-42. This fact most likely explains the correlation between these markers.
Tau is a normal human brain phosphoprotein that binds to microtubules in the neuronal axons, thereby promoting microtubule assembly and stability.23 Increased levels of CSF tau are probably a consequence of axonal damage and have been reported in AD,15 Creutzfeldt-Jakob disease,24 Guillain-Barré syndrome,25 and TBI.9 Loss of axonal microtubules is a common feature of TBI,26 and TBI is expected to release intracellular microtubule binding proteins, such as tau, into the extracellular space where they are transported by convective bulk flow to CSF.9 Importantly, axonal injury is one of the most common forms of injury in patients with TBI, accounting for up to 48% of all primary lesions,27 and is one of the most frequent causes of morbidity after TBI.8 Several lines of evidence suggest that measuring CSF tau levels may provide an alternative means of assessing axonal injury in TBI.9 In our study, tau levels were more than tenfold higher in patients with TBI than in the control groups. Highest tau levels were observed when CSF was collected during the second week after trauma. Furthermore, we found a time-dependent decrease in CSF tau levels in patients with serial CSF collections and normal tau levels in all patients when CSF was collected more than 43 days post-trauma. Importantly, CSF tau did not correlate with severity (initial GCS) or outcome (GOS) of brain injury in our study. To further elucidate whether CSF tau reflects axonal damage after TBI, future studies could compare CSF tau levels at different time points post-trauma with brain lesions, particularly with localization, on MRI scans, because the wide range of tau levels might be due to the distance of white matter lesions to the ventricles.
In addition, tau is quickly cleared from CSF. In contrast to a previous study,9 tau levels did not correlate with clinical improvement or outcome of patients with TBI. Different intervals between head trauma and CSF collection in our study and the time-dependent removal from CSF disqualify tau as a predictor of outcome after TBI. Also, tau did not correlate with outcome in patients with early CSF collection.
It remains unclear whether hyperphosphorylated tau plays a role in TBI. For example, in AD, hyperphosphorylated tau, which is probably initiated by amyloid deposition, causes neurofibrillary tangles.28 Interestingly, a recent report also suggests that tau hyperphosphorylation occurs after experimental closed head injury in mice.29 In the current study we measured only the concentration of total tau. Future studies are needed to investigate the role of hyperphosphorylated tau in TBI.
There was a significant correlation between CSF protein levels, cell counts, and tau. Increased cell counts occur in subarachnoid hemorrhage and in infections of the CNS. Increased protein levels may indicate a disruption of the blood–brain barrier and reduced CSF circulation. One consequence of TBI is increased permeability of the blood–brain barrier, permitting the passage of proteins among the brain, CSF, and peripheral circulation. It seems unlikely that a breakdown of the blood–brain barrier may contribute to the increase in tau in CSF, as tau is a neuronal protein and not detectable in plasma. CSF cell counts and protein levels also inversely correlated with the interval between TBI and CSF collection. This is most likely explained by CSF circulation and reconstitution of the blood–brain barrier, which parallels normalization of tau levels. Neither CSF protein nor CSF cell counts correlated with GOS. Correlations between Aβ-42, tau, and outcome are unadjusted and probably confounded by other variables that correlated significantly with Aβ-42 and tau in the univariate analysis. Owing to the limited number of patients a multiple linear regression analysis including all univariately significant variables was not powerful enough. This limits the results of the study.
Acknowledgments
Supported by a grant from the Austrian Science Fund (FWF; P15308) to A.K.
- Received July 9, 2002.
- Accepted January 30, 2003.
References
- ↵
- ↵
- ↵
- ↵
Zhang H, Zhang X, Zhang T, Chen L. Excitatory amino acids in cerebrospinal fluid of patients with acute head injuries. Clin Chem . 2001; 47: 1458–1462.
- ↵
Pleines UE, Morganti-Kossmann MC, Rancan M, Joller H, Trentz O, Kossmann T. S-100 beta reflects the extent of injury and outcome, whereas neuronal specific enolase is a better indicator of neuroinflammation in patients with severe traumatic brain injury. J Neurotrauma . 2001; 18: 491–498.
- ↵
Rothoerl RD, Woertgen C, Brawanski A. S-100 serum levels and outcome after severe head injury. Acta Neurochir . 2000; 76 (suppl): 97–100.
- ↵
- ↵
- ↵
- ↵
Schofield PW, Tang M, Marder K, et al. Alzheimer’s disease after remote head injury: an incidence study. J Neurol Neurosurg Psychiatry . 1997; 62: 119–124.
- ↵
- ↵
Roberts GW, Allsop D, Bruton C. The occult aftermath of boxing. J Neurol Neurosurg Psychiatry . 1990; 53: 373–378.
- ↵
- ↵
Uryu K, Laurer H, McIntosh T, et al. Repetitive mild brain trauma accelerates Abeta deposition, lipid peroxidation, and cognitive impairment in a transgenic mouse model of Alzheimer amyloidosis. J Neurosci . 2002; 22: 446–454.
- ↵
Hulstaert F, Blennow K, Ivanoiu A, et al. Improved discrimination of AD patients using beta-amyloid(1–42) and tau levels in CSF. Neurology . 1999; 52: 1555–1562.
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
Sjogren M, Vanderstichele H, Agren H, et al. Tau and Abeta42 in cerebrospinal fluid from healthy adults 21–93 years of age: establishment of reference values. Clin Chem . 2001; 47: 1776–1781.
- ↵
Goedert M, Jakes R, Spillantini MG, et al. Tau protein in Alzheimer’s disease. Biochem Soc Trans . 1995; 23: 80–85.
- ↵
- ↵
Vandermeeren M, Mercken M, Vanmechelen E, et al. Detection of tau proteins in normal and Alzheimer’s disease cerebrospinal fluid with a sensitive sandwich enzyme-linked immunosorbent assay. J Neurochem . 1993; 61: 1828–1834.
- ↵
- ↵
- ↵
- ↵
Disputes & Debates: Rapid online correspondence
NOTE: All contributors' disclosures must be entered and current in our database before comments can be posted. Enter and update disclosures at http://submit.neurology.org. Exception: replies to comments concerning an article you originally authored do not require updated disclosures.
- Stay timely. Submit only on articles published within the last 8 weeks.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- 200 words maximum.
- 5 references maximum. Reference 1 must be the article on which you are commenting.
- 5 authors maximum. Exception: replies can include all original authors of the article.
- Submitted comments are subject to editing and editor review prior to posting.
You May Also be Interested in
Related Articles
- No related articles found.