Stereology of cerebral cortex after traumatic brain injury matched to the Glasgow outcome score.

Magnetic resonance imaging provides evidence for loss of both white and grey matter, in terms of tissue volume, from the cerebral hemispheres after traumatic brain injury. However, quantitative histopathological data are lacking. From the archive of the Department of Neuropathology at Glasgow, the cerebral cortex of 48 patients was investigated using stereology. Patients had survived 3 months after traumatic brain injury and were classified using the Glasgow Outcome Scale as follows: moderately disabled (n = 13), severely disabled (n = 12) and vegetative state (n = 12); and controls. Some patients from the archive were diagnosed with diffuse axonal injury post-mortem. Comparisons of changes in cortical neuron population across Glasgow Outcome Scale groups between diffuse axonal injury and non-diffuse axonal injury patients were undertaken using effect size analyses. The hypotheses tested were that (i) thinning of the cerebral cortex occurred after traumatic brain injury; (ii) changes in thickness of cortical layers in Brodmann areas 11, 10, 24a and 4 differed; and (iii) different changes occurred for neuronal number, their size and nearest neighbour index across Glasgow Outcome Scale groups. There was a greater loss of large pyramidal and large non-pyramidal neurons with a more severe score on the Glasgow Outcome Scale from all four cortical regions, with the greatest loss of neurons from the prefrontal cortex of patients with diffuse axonal injury. There were differences in the changes of number of medium and small pyramidal and non-pyramidal neurons between different cortical regions, and between patients with and without diffuse axonal injury. Generally, a decrease in the somatic diameter of pyramidal and non-pyramidal neurons was associated with a more severe clinical outcome. However, in the motor cortex a more severe Glasgow Outcome Scale was associated with an increased diameter of medium pyramidal neurons and small non-pyramidal cells. Pyramidal and non-pyramidal neurons did not follow a Poisson distribution within the neuropil of control patients. Pyramidal neurons were usually scattered while medium and small non-pyramidal neurons were clustered. An increased spacing between remaining neurons usually occurred across Glasgow Outcome Scale groups. It is concluded that loss of neurons resulted in reduced executive and integrative capability in patients after traumatic head injury.

Thalamic nuclei after human blunt head injury.

Paraffin-embedded blocks from the thalamus of 9 control patients, 9 moderately disabled, 12 severely disabled, and 10 vegetative head-injured patients assessed using the Glasgow Outcome Scale and identified from the Department of Neuropathology archive. Neurons, astrocytes, macrophages, and activated microglia were differentiated by Luxol fast blue/cresyl violet, GFAP, CD68, and CR3/43 staining and stereological techniques used to estimate cell number in a 28-microm-thick coronal section. Counts were made in subnuclei of the mediodorsal, lateral posterior, and ventral posterior nuclei, the intralaminar nuclei, and the related internal lamina. Neuronal loss occurred from mediodorsal parvocellularis, rostral center medial, central lateral and paracentral nuclei in moderately disabled patients; and from mediodorsal magnocellularis, caudal center medial, rhomboid, and parafascicular nuclei in severely disabled patients; and all of the above and the centre median nucleus in vegetative patients. Neuronal loss occurred primarily from cognitive and executive function nuclei, a lesser loss from somatosensory nuclei and the least loss from limbic motor nuclei. There was an increase in the number of reactive astrocytes, activated microglia, and macrophages with increasing severity of injury. The study provides novel quantitative evidence for differential neuronal loss, with survival after human head injury, from thalamic nuclei associated with different aspects of cortical activation.

Slow, medium, or fast re-warming following post-traumatic hypothermia therapy? An ultrastructural perspective.

It was hypothesized that rapid rather than slow re-warming following traumatic brain injury (TBI) and short-term hypothermia results in secondary, ultrastructural pathology. After stretch injury to the right optic nerve, adult guinea pigs were randomly allocated to one of six experimental groups. Either (1) sham (all procedures but not stretch-injured; n = 4); injured and (2) maintained at normal temporalis core temperature (38.5 degrees C) for 8 hours (n = 6); (3) cooled rapidly to 32.5 degrees C (temporalis temperature), maintained for 4 h and re-warmed to 38.5 degrees C at 1 degrees C rise every 10 min (fast; n = 6); (4) cooled and re-warmed at 1 degrees C rise every 20 min (medium; n = 6); (5) cooled and rewarmed at 1 degrees C rise every 40 min (slow; n = 6) before being killed 8 h after injury; and (6) uninjured animals (n = 6) cooled to 32.5 degrees C for 4 h and then re-warmed at 1 degrees C every 10 min before killing 4 h later. Tissue was processed for light immunocytochemistry (beta-APP and RMO-14) and ultrastructural stereology. In both uninjured and injured fast re-warmed animals, there was almost total loss of axonal microtubules (MT) and an increased number of neurofilaments (NF) within the axoplasm. In the former, there was also compaction of NF. The number of MT was reduced to 40% of control values, NFs were increased but were not compacted after medium rate re-warming. Following slow re-warming the axonal cytoskeleton did not differ from that in control animals. It is concluded that re-warming faster than 1 degrees C every 40 min following mild post-traumatic hypothermia induces secondary axonal pathology.

Lateral fluid percussion brain injury: a 15-year review and evaluation.

This article comprehensively reviews the lateral fluid percussion (LFP) model of traumatic brain injury (TBI) in small animal species with particular emphasis on its validity, clinical relevance and reliability. The LFP model, initially described in 1989, has become the most extensively utilized animal model of TBI (to date, 232 PubMed citations), producing both focal and diffuse (mixed) brain injury. Despite subtle variations in injury parameters between laboratories, universal findings are evident across studies, including histological, physiological, metabolic, and behavioral changes that serve to increase the reliability of the model. Moreover, demonstrable histological damage and severity-dependent behavioral deficits, which partially recover over time, validate LFP as a clinically-relevant model of human TBI. The LFP model, also has been used extensively to evaluate potential therapeutic interventions, including resuscitation, pharmacologic therapies, transplantation, and other neuroprotective and neuroregenerative strategies. Although a number of positive studies have identified promising therapies for moderate TBI, the predictive validity of the model may be compromised when findings are translated to severely injured patients. Recently, the clinical relevance of LFP has been enhanced by combining the injury with secondary insults, as well as broadening studies to incorporate issues of gender and age to better approximate the range of human TBI within study design. We conclude that the LFP brain injury model is an appropriate tool to study the cellular and mechanistic aspects of human TBI that cannot be addressed in the clinical setting, as well as for the development and characterization of novel therapeutic interventions. Continued translation of pre-clinical findings to human TBI will enhance the predictive validity of the LFP model, and allow novel neuroprotective and neuroregenerative treatment strategies developed in the laboratory to reach the appropriate TBI patients.

Evolution of GADD34 expression after focal cerebral ischaemia.

GADD34, a stress response protein associated with cell rescue, DNA repair and apoptosis, is expressed in the ischaemic brain. The C-terminal region of GADD34 has homology with the Herpes Simplex Virus protein, ICP34.5, which overcomes the protein synthesis block after viral infection by actively dephosphorylating eukaryotic translation initiation factor 2alpha (eIF2alpha). The carboxy terminus of GADD34 is also capable of dephosphorylating eIF2alpha and therefore has the capacity to restore the protein synthesis shutoff associated with ischaemia. This study examines the distribution and time course of GADD34 expression after focal cerebral ischaemia. Focal ischaemia or sham procedure was carried out on Sprague-Dawley rats with survival times of 4, 12, 24 h, 7 and 30 days. Brains were processed for histology and immunohistochemistry. Ischaemic damage was mapped onto line diagrams and GADD34 positive cells counted in selected regions of cortex and caudate. GADD34 immunopositive cells (mainly neurones), expressed as cells/mm2, were present in ischaemic brains at 4 h (e.g., peri-infarct cortex 20 +/- 5; contralateral cortex 3 +/- 1, P < 0.05). Of the time points examined, numbers of GADD34 positive cells were highest 24 h after ischaemia (peri-infarct cortex 31 +/- 7.3, contralateral cortex 0.1 +/- 0.1, P < 0.05). Immunopositive cells, following a similar time course, were identified within the peri-infarct zone in the caudate nucleus and in ipsilateral cingulate cortex (possibly as a consequence of cortical spreading depression). GADD34 positive cells did not co-localise with a marker of irreversible cell death (TUNEL). Taken together, GADD34 positive cells in key neuroanatomical locations pertinent to the evolving ischaemic lesion, the lack of co-localisation with TUNEL and the protein's known effects on restoring protein synthesis, repairing DNA and involvement in ischaemic pre-conditioning suggests that it has the potential to influence cell survival in ischaemically compromised tissue.

Apo E genotype not associated with intravascular coagulation in traumatic brain injury.

There is considerable evidence linking both genotype and coagulopathy to vascular complications of traumatic brain injury (TBI) and other cerebral insults. The authors explored a possible connection between the apolipoprotein E (Apo E) genotype, coagulopathy and intravascular microthombosis (IMT) in TBI. The predicted association was not confirmed.

Differential responses in three thalamic nuclei in moderately disabled, severely disabled and vegetative patients after blunt head injury.

In vivo imaging techniques have indicated for many years that there is loss of white matter after human traumatic brain injury (TBI) and that the loss is inversely related to cognitive outcome. However, correlated, quantitative evidence for loss of neurons from either the cerebral cortex or the diencephalon is largely lacking. There is some evidence in models of TBI that neuronal loss occurs within the thalamus, but no systematic studies of such loss have been undertaken in the thalamus of humans after blunt head injury. We have undertaken a stereological analysis of changes in numbers of neurons within the dorsomedial, ventral posterior and lateral posterior thalamic nuclei in patients assessed by the Glasgow Outcome Scale as moderately disabled (n = 9), severely disabled (n = 12) and vegetative (n = 10) head-injured patients who survived between 6 h and 3 years, and controls (n = 9). In histological sections at the level of the lateral geniculate body, the cross-sectional area of each nucleus and the number and the mean size of neurons within each nucleus was quantified. A statistically significant loss of cross-sectional area and number of neurons occurred in the dorsomedial nucleus in moderately disabled, and both the dorsomedial and ventral posterior thalamic nuclei in severely disabled and vegetative head-injured patients. However, there was no change in neuronal cell size. In the lateral posterior nucleus, despite a reduction in mean cell size, there was not a significant change in either nuclear area or number of neurons in cases of moderately disabled, severely disabled or vegetative patients. We posit, although detailed neuropsychological outcome for the patients included within this study was not available, that neuronal loss in the dorsomedial thalamus in moderately and severely disabled and vegetative patients may be the structural basis for the clinical assessment in the Glasgow Outcome Scale. In severely disabled and vegetative patients, loss of neurons from the ventral posterior thalamic nucleus may also reflect loss of response to afferent stimuli.

From cell death to neuronal regeneration: building a new brain after traumatic brain injury.

During the past decade, there has been accumulating evidence of the involvement of passive and active cell death mechanisms in both the clinical setting and in experimental models of traumatic brain injury (TBI). Traditionally, research for a treatment of TBI consists of strategies to prevent cell death using acute pharmacological therapy. However, to date, encouraging experimental work has not been translated into successful clinical trials. The development of cell replacement therapies may offer an alternative or a complementary strategy for the treatment of TBI. Recent experimental studies have identified a variety of candidate cell lines for transplantation into the injured CNS. Additionally, the characterization of the neurogenic potential of specific regions of the adult mammalian brain and the elucidation of the molecular controls underlying regeneration may allow for the development of neuronal replacement therapies that do not require transplantation of exogenous cells. These novel strategies may represent a new opportunity of great interest for delayed intervention in patients with TBI.

New method for the quantitative assessment of axonal damage in focal cerebral ischemia.

Quantification of damage in both grey and white matter is required for comprehensive assessment of neuroprotective drug efficacy. Although methods for quantification of neuronal perikaryal damage after ischemia are well established, assessment of axonal damage has been limited. This article describes a new method for quantitation of axonal injury after middle cerebral artery (MCA) occlusion in rats and its application to the study of the antioxidant ebselen. The methodology is based on immunohistochemical detection of amyloid precursor protein (APP) accumulation in deformed, swollen axons in zones of ischemia. Sixty-five axon-rich sites throughout the MCA territory are assessed for the presence (scored 1) or absence (scored 0) of accumulated APP in axonal swellings. Scores for individual sites are summated in predefined neuroanatomic regions (e.g. corpus callosum), stereotaxic levels, or for a total hemisphere APP score. Both intra-rater and inter-rater reproducibility were high (r = 0.87 and 0.80, respectively). Ebselen (1 mg kg(-1) hr(-1), intravenously) significantly reduced the volume of neuronal perikaryal damage (24%, P < 0.01) and axonal damage (total APP score reduced from 27 [23.9 to 35.1, 95% CI] to 21.5 [18.2 to 23.3], P = 0.002 with ebselen treatment). In conclusion, a robust and generally applicable method is described for assessing pathologic features in myelinated fiber tracts that is sensitive for detection of drug effects on axonal damage.

Differential effects of 17beta-estradiol upon stroke damage in stroke prone and normotensive rats.

We previously reported that during pro-estrus (high endogenous estrogen levels), brain damage after middle cerebral artery occlusion (MCAO) was reduced in stroke-prone spontaneously hypertensive rats (SHRSP) but not in normotensive Wistar Kyoto rat (WKY). In the present study, we examined the effect of exogenous estrogen on brain damage after MCAO in SHRSP and WKY. A 17beta-estradiol (0.025 mg or 0.25 mg, 21 day release) or matching placebo pellet was implanted into ovariectomized WKY and SHRSP (3 to 4 months old) who then underwent distal diathermy-induced MCAO 2 weeks later. Plasma 17beta-estradiol levels for placebo and 17beta-estradiol groups were as follows: WKY 0.025 mg 16.4 +/- 8.5 (pg/mL, mean +/- SD) and 25.85 +/- 12.6; WKY 0.25 mg 18.2 +/- 9.0 and 69.8 +/- 27.4; SHRSP 0.25 mg 20.7 +/- 8.8 and 81.0 +/- 16.9. In SHRSP, infarct volumes at 24 hours after MCAO were similar in placebo and 17beta-estradiol groups: SHRSP 0.025 mg 126.7 +/- 15.3 mm (n = 6) and 114.0 +/- 14.1 mm (n = 8) (not significant); SHRSP 0.25 mg 113.5 +/- 22.3 mm (n = 8) and 129.7 +/- 26.2 mm (n = 7) (not significant), respectively. In WKY, 17beta-estradiol significantly increased infarct volume by 65% with 0.025 mg dose [36.1 +/- 20.7 mm (n = 8) and 59.7 +/- 19.3 mm (n = 8) (P = 0.033, unpaired t-test)] and by 96% with 0.25 mg dose [55.9 +/- 36.4 mm (n = 8) and 109.7 +/- 6.7 mm (n = 4) (P = 0.017)]. Thus, 17beta-estradiol increased stroke damage in normotensive rats with no significant effect in stroke-prone rats. Despite being contrary to our hypothesis, our findings add substance to the recently reported negative effects of 17beta-estradiol in clinical studies.

Nucleus basalis of Meynert pathology in the human brain after fatal head injury.

Dysfunction of the basal forebrain cholinergic system has been hypothesized to contribute to deficits of memory and cognition after head injury. We have previously reported reduced levels of choline acetyltransferase activity in the cerebral cortex of patients who died after a head injury, demonstrating that there is a loss of cortical cholinergic innervation. In the present study, we examined the nucleus basalis of Meynert (NBM), which provides cortical cholinergic innervation, in fatally head-injured patients and in controls. The majority of head-injured patients had histological evidence of neuronal damage in the NBM, which was due to mechanical distortion of the tissue and/or ischemic damage. The findings demonstrate that the NBM is vulnerable after severe head injury and that secondary insults play an important role in the damage. Thus, both neuronal perikarya and terminals of the basal forebrain cholinergic system are damaged after human fatal head injury. This damage may contribute to persisting dysfunction of memory and cognition in head-injured patients who survive.

Continued in situ DNA fragmentation of microglia/macrophages in white matter weeks and months after traumatic brain injury.

Paraffin-embedded material from the pons of head-injured patients whose disability could be attributed to diffuse traumatic axonal injury, and controls, was identified from the department's archive. The cases were divided into three groups based on survival, viz Group 1 (n = 5) who survived for between 4 and 8 weeks, Group 2 (n = 5) for between 3 and 9 months, and Group 3 (n = 5) who survived for more that 12 months. Sections were stained by the TUNEL (TdT-mediated UTP nick end labelling) technique, and by H&E, LFB/CV and immunohistochemically for astrocytes (GFAP) and microglia/macrophages (CD68). Microscopic abnormalities were mapped onto line diagrams of two levels of the pons and quantitation of the response determined by an eye-piece graticule placed over the medial lemmisci, cortico-spinal and transverse fiber tracts. Data were pooled by region of interest. In the H&E and LFB/CV stained sections, there was variable pallor of staining in ascending and descending fiber tracts due to loss of myelin: within these same tracts there was an astrocytosis and increased numbers of microglia/macrophages compared with controls. In the white matter tracts of the controls, there was on average 1-2 TUNEL+ cells per unit area. In contrast, there were on average 2-16 TUNEL+ cells in the cortico-spinal tracts and in the medial lemnisci of all groups of head-injured patients. CD68+ cells co-located with the TUNEL+, and their number mirrored the TUNEL + staining with on average 16-30 cells per unit area in Group 1, 14-27 cells per unit area in Group 2, and 12-14 cells per unit area in Group 3. There was a statistical association between the TUNEL+ and CD68+ cells. Few changes were seen in the transverse fiber tracts of the pons. These findings indicate that most of the in situ DNA fragmentation occurred in microglia/macrophages in ascending and descending fiber tracts of the brain stem in which by conventional light microscopy there is Wallerian degeneration. However, in addition, a few TUNEL+ oligodendrocyte-like cells were also seen.

Post-acute alterations in the axonal cytoskeleton after traumatic axonal injury.

All previous analyses of axonal responses to traumatic axonal injury (TAI) have described the ultrastructure of changes in the cytoskeleton and axolemma within 6 h of injury. In the present study we tested the hypothesis that there are, in addition, ultrastructural pathological changes up to 1 week after injury. TAI was induced in the adult guinea pig optic nerve of nine animals. Three animals were killed at either 4 h, 24 h, or 7 days (d) after injury. Quantitative analysis of the number or proportion of axons within 0.5-micro m-wide bins showed an increase in the number of axons with a diameter of less than 0.5 micro m at 4 h, 24 h, and 7 d, the presence of lucent axons at 24 h and 7 d and that the highest number of injured axons occurred about half way along the length of the nerve. A spectrum of pathological changes occurred in injured fibers-pathology of mitochondria; dissociation of myelin lamellae but little damage to the axon; loss of linear register of the axonal cytoskeleton; differential responses between microtubules (MT) and neurofilaments (NF) in different sizes of axon; two different sites of compaction of NF; loss of both NF (with an increase in their spacing) and MT (with a reduction in their spacing); replacement of the axoplasm by a flocculent precipitate; and an increased length of the nodal gap. These provide the first ultrastructural evidence for Wallerian degeneration of nerve fibers in an animal model of TAI.

Simple morphometry of axonal swellings cannot be used in isolation for dating lesions after traumatic brain injury.

Disruption of fast axonal transport as a result of traumatic brain injury is characterized by the accumulation of beta-amyloid precursor protein (APP) in axonal swellings. A recent report has suggested a correlation between the size of axonal swellings and survival time up to about 85 h after blunt head injury. The authors of the report concluded that this correlation, in conjunction with other evidence, might be useful in forensic science for timing injuries. To test this hypothesis we have used image analysis software to measure a number of different morphological parameters of axonal swellings. Paraffin sections from 63 cases of fatal head injury were stained with an antibody raised against the N-terminus of APP and counterstained with haematoxylin. Three different measurements were made of the APP-immunoreactive axonal swellings from the corpus callosum: (i) minimum and (ii) maximum Feret diameters, and (iii) area. Linear regression revealed a significant correlation between survival time and the minimum Feret diameter (p < 0.0001) and the area (p < 0.001) of axonal swellings. Our findings are in agreement with the previous study in that there is a significant correlation between axonal swelling size and survival time. However, we would suggest that the large variability in swelling size within individual cases and the heterogeneity of the original trauma seriously compromise the utility of such information in the timing of lesions.

Association between intravascular microthrombosis and cerebral ischemia in traumatic brain injury.

OBJECTIVE: To determine the association between traumatic cerebral ischemia and intravascular thrombosis, a common finding after traumatic brain injury (TBI). METHODS: We reviewed samples of the frontal cortex and hippocampus from individuals who had sustained a fatal TBI. Sections stained with hematoxylin and eosin were reviewed and rated for severity of selective neuronal necrosis (SNN). Because intravascular fibrin microthrombi may lyse within a few days of TBI, we restricted our analysis to patients who had died within 48 hours of injury. Medical records in all cases were reviewed to rule out severe or prolonged hypotension or hypoxemia. Eleven patients with severe or global SNN were compared with 11 patients in whom SNN was mild or absent. Slides adjacent to the hematoxylin and eosin sections were stained with an immunofluorescent antibody to antithrombin III and were reviewed for intravascular microthrombosis. The number of microthrombi on each slide was counted by an investigator blinded to the hematoxylin and eosin findings, and density of intravascular microthrombi was calculated. RESULTS: Intravascular microthrombi were noted in every section, excluding control (non-TBI) brain tissue. However, the density of microthrombi varied with the degree of SNN. We found a highly significant difference in the mean density of microthrombi between patients with severe SNN (7.74 +/- 3.7/cm(2)) and those with little or no SNN (2.58 +/- 1.0/cm(2)). Furthermore, a good correlation was noted between the location of intravascular microthrombi and that of SNN. CONCLUSION: These data support a strong link between intravascular microthrombosis and neuronal death after brain trauma in humans and may have important implications for new therapeutic approaches.

Amyloid beta accumulation in axons after traumatic brain injury in humans.

OBJECT: Although plaques composed of amyloid beta (AD) have been found shortly after traumatic brain injury (TBI) in humans, the source for this Abeta has not been identified. In the present study, the authors explored the potential relationship between Abeta accumulation in damaged axons and associated Abeta plaque formation. METHODS: The authors performed an immunohistochemical analysis of paraffin-embedded sections of brain from 12 patients who died after TBI and from two control patients by using antibodies selective for Abeta peptides, amyloid precursor protein (APP), and neurofilament (NF) proteins. In nine brain-injured patients, extensive colocalizations of Abeta, APP, and NF protein were found in swollen axons. Many of these immunoreactive axonal profiles were present close to Abeta plaques or were surrounded by Abeta staining, which spread out into the tissue. Immunoreactive profiles were not found in the brains of the control patients. CONCLUSIONS: The results of this study indicate that damaged axons can serve as a large reservoir of Abeta, which may contribute to Abeta plaque formation after TBI in humans.

Neuropathological findings in disabled survivors of a head injury.

We investigated how the occurrence and severity of the main neuropathological types of traumatic brain injury (TBI) influenced the severity of disability after a head injury. Eighty-five victims, each of whom had lived at least a month after a head injury but then died, were studied. Judged by the Glasgow Outcome Scale (GOS), before death 35 were vegetative, 30 were severely and 20 were moderately disabled. Neuropathological assessment showed that 71 (84%) victims had sustained cerebral contusions, 49 (58%) had diffuse axonal injury (DAI), 57 (67%), had ischemic brain damage (IBD), 58 (68%) had symmetrical ventricular enlargement, and in 47 (55%) intracranial pressure (ICP) had been increased. Thirty-five (41%) had undergone evacuation of an intracranial hematoma. Brainstem damage was seen in only 11 (13%). Analysis (χ(2) test for trends) of the relationship between these features and outcome showed that findings of DAI, raised ICP, thalamic damage, or ventricular enlargement (all p<0.005), and IBD (p=0.04) were associated with an increasingly worse outcome. Conversely, moderate or severe contusions (p=0.001) were increasingly associated with better outcomes, and evacuation of a hematoma was associated (p=0.001) with outcomes likely to be better than vegetative. We conclude that diffuse or multifocal neuropathological patterns of TBI from primary axonal injury or secondary ischemic damage are most likely to be associated with the most severely impaired outcomes after a head injury.

Stereology and ultrastructure of chronic phase axonal and cell soma pathology in stretch-injured central nerve fibers.

Magnetic resonance imaging (MRI) suggests that with survival after human traumatic brain injury (TBI), there is ongoing loss of white and grey matter from the injured brain during the chronic phase. However; direct quantitative experimental evidence in support of this observation is lacking. Using the guinea pig stretch-injury optic nerve model, quantitative evidence by stereology of damage to the optic nerve and retina was sought. Stretch injury was applied to the right optic nerve of 15 adult male guinea pigs. Three animals each at 1, 2, 3, 8, or 12 weeks' survival were killed and prepared for transmission electron microscopy (TEM). The estimated number of intact and injured axons within bins of transverse diameters 0-0.5, 0.51-1.0, 1.01-1.5, 1.51-2.0, 2.01-2.5, and 2.51-3.0 µm in the middle segment of each injured optic nerve and from 5 control animals were compared across all survival time points. The estimated numbers of intact and pyknotic retinal ganglion cells from the same animals were also compared. Loss of myelinated fibers continued throughout the experimental period. The most rapid loss was of the largest fibers; loss of intermediate-sized fibers continued, but the numbers of the smallest fibers increased from 3 weeks onward. There was hypertrophy and proliferation of glial cells within the surrounding neuropil. A relatively low-grade loss of retinal ganglion cells occurred throughout the experiment, with about 60% remaining at 12 weeks' survival. We provide quantitative evidence that after traumatic axonal injury (TAI) there is a continuing loss of nerve fibers and their cell bodies from a CNS tract over a 3-month post-traumatic interval.