Diffuse Axonal Injury

Histopathology 1989, 15,49-59

Diffuse axonal injury in head injury: definition, diagnosis and grading

J.HUME ADAMS*, D.DOYLE*, I.FORDt, T.A.GENNARELLI$., D.I.GRAHAM* & D.R.MCLELLAN* *Department of Neuropathology and ?Department of Statistics, University of Glasgow, Glasgow, Scotland and $.Department of Neurosurgery, University of Pennsylvania, Philadelphia, USA

Accepted for publication 7 November 1988

HUME ADAMS J., DOYLE D., FORD I., GENNARELLI T.A., GRAHAM D.I. & MCLELLAN D.R. (1989) Histopathology 15,49-59

Diffuse axonal injury in head injury: definition, diagnosis and grading

Diffuse axonal injury is one of the most important types of brain damage that can occur as a result of non-missile head injury, and it may be very difficult to diagnose postmortem unless the pathologist knows precisely what he is looking for. Increasing experience with fatal non-missile head injury in man has allowed the identification or three grades of diffuse axonal injury. In grade 1 there is histological evidence of axonal injury in the white matter of the cerebral hemispheres, the corpus callosum, the brain stem and, less commonly, the cerebellum; in grade 2 there is also a focal lesion in the corpus callosum; and in grade 3 there is in addition a focal lesion in the dorsolateral quadrant or quadrants of the rostra1 brain stem. The focal lesions can often only be identified microscopically. Diffuse axonal injury was identified in 122 of a series of 434 fatal non-missile head injuries-I 0 grade I , 29 grade 2 and 83 grade 3. In 24 of these cases the diagnosis could not have been made without microscopical examination, while in a further 31 microscopical examination was required to establish its severity.

Keywords: head injury, axonal injury

Introduction

Diffuse axonal injury is an important factor contributing to a patients clinical state and governing the outcome in anyone who sustains a non-missile head injury, and it is also the commonest cause of the vegetative state and severe disability after injury (McLellan et al. 1986). Yet not much more than a decade ago, there was still considerable controversy as to the nature of the brain damage and its pathogenesis. It had been referred to by several names, viz. shearing injury (Strich 1970, Peerless &

Address for correspondence: Professor J.H.Adams, Department of Neuropathology, Institute of Neurological Sciences, Southern General Hospital, Glasgow G5 1 4TF, Scotland.

49

50 J.Hume Adams et al.

Rewcastle 1967), diffuse damage to white matter of immediate impact type (Adams et al. 1977), diffuse white matter shearing injury (Zimmerman, Bilaniuk & Gennarelli 1978) and inner cerebral trauma (Grcevic 1982). These authors took the view that the damage to white matter occurred at the moment of injury, but other experienced neuropathologists were of the opinion that it was secondary to hypoxic brain damage, to cerebral oedema or to secondary damage in the brain stem resulting from an intracranial expanding lesion (Jellinger & Seitelberger 1970, Jellinger 1977, Peters & Rothemund 1977).

In our earlier studies we questioned whether focal damage could occur in the brain stem as a result of a head injury as an isolated event (Mitchell & Adams 1973) and we suggested at that time that the structural basis of the clinical syndrome of primary brain stem injury was the type of brain damage now referred to widely as diffuse axonal injury. In our initial attempts to identify diffuse axonal injury as a distinctive clinicopathological entity (Mitchell & Adams 1973, Adams et al. 1977, 1982), it is now clear that we concentrated on the severe end of a spectrum. With increasing experience of fatal head injuries in man, it has become clear that there are less severe forms of diffuse axonal injury, and in this paper we wish to define three grades. As a result of studies in man and of experimental work (Gennarelli et al. 1982, 1987), there is now no reasonable doubt that this type of brain damage is a primary event that occurs at the moment of injury.

Attention has also been drawn to milder degrees of diffuse axonal injury by Oppenheimer (1 968) and Pilz (1 983). Since the focal lesions in the corpus callosum and in the brain stem can often only be identified microscopically, it is of the utmost importance that pathologists dealing with fatal head injury should be acutely aware of this type of brain damage lest they fail to interpret the type of brain damage that has led to the patients clinical state. This is particularly so if there is no evidence of an intracranial haematoma, severe cerebral contusions, or increased intracranial pressure.

Materials and methods

During the 15 year period 1968-1982, full autopsies were undertaken in the Institute of Neurological Sciences, Glasgow, on 635 fatal non-missile head injuries. The brains were suspended in 10% formol saline for 3-4 weeks prior to dissection in a standard fashion (Adams & Murray 1982); the cerebral hemispheres were sliced in the coronal plane at 1 cm intervals, the cerebellum at right angles to the folia, and the brain stem horizontally. Comprehensive histological studies were undertaken in 434 of these brains. Representative blocks from all regions of the brain were embedded in celloidin and 30 pm sections were stained by Nissls method with cresyl violet and by Woelkes modification of Heidenhains technique for myelin. In 382 of these cases, representative blocks from the cerebral hemispheres, the cerebellum and the brain stem were also embedded in paraffin wax and sections stained with haematoxylin and eosin, the Luxol fast blue/cresyl violet technique and by the Palmgren technique for axons. All macroscopic and histological abnormalities were

Difluse axonal injury 51

recorded on a series of line diagrams. The abnormalities were then recorded on a proforma and the data stored on the University of Glasgows mainframe computer.

The clinical records in each case were assessed to ascertain if the patients had been able to talk immediately after their injury (Reilly et al. 1975) on the assumption that if they had, they had not sustained catastrophic brain damage at the time of injury. Patients who could talk rationally were said to have had a complete lucid interval. If they were confused, the lucid interval was said to have been partial.

In the analysis of the results, evidence of a relationship between the presence of diffuse axonal injury and the other variable studies is assessed using a chi-squared test.

Results

There were 342 males and 92 females; the age range was from 3 months to 89 years, and the duration of survival ranged from 2 h to 14 years.

Evidence of diffuse axonal injury was identified in 122 of the 434 cases, an incidence of 29%). The true incidence must have been higher since axonal damage would be difficult to identify in the 52 cases in which only celloidin sections were available, and in patients where the less severe features of diffuse axonal injury might have been obscured by other types of brain damage such as severe contusions and intracerebral haematomas, brain damage secondary to high intracranial pressure and herniation of the brain leading to haemorrhage and infarction in the brain stem, or severe hypoxic damage. Furthermore, axonal damage would also not be very obvious in patients who had survived for less than 12-18 h after their injury.

There were 10 cases of grade 1 diffuse axonal injury, there being evidence of axonal damage in the white matter of the cerebral hemispheres including the corpus callosum, in the brain stem and, occasionally, in the cerebellum: this damage can only be identified microscopically. There were 29 cases with grade 2 diffuse axonal injury, i.e. there was a focal lesion in the corpus callosum in addition to diffuse axonal damage: the focal lesion was identifiable only microscopically in 11 of these cases. The majority of the cases-83-had sustained grade 3 diffuse axonal injury since there were focal lesions both in the corpus callosum and the dorso-lateral quadrant of the rostra1 brain stem: in only 49 of these cases were both the focal lesions apparent macroscopically even in a properly fixed and dissected brain. Thus, of the 122 cases the severity of diffuse axonal injury could only be defined by histological studies in 24, whilst its presence in 31 cases would have been missed unless appropriate histological studies had been undertaken. Grades 2 and 3 can be said to be severe if the focal lesions are apparent macroscopically.

It is now clear that in our early attempts to define the clinicopathological entity of diffuse axonal injury, we concentrated on severe grade 3 cases. The focal lesion in the corpus callosum is typically haemorrhagic and tends to lie to one side of the midline (Figure l) , although it may extend to the midline and involve the interventricular septum and the pillars of the fornix. The lesions occur most frequently in the inferior part of the corpus callosum and usually extend over an

52 J.Hume Adams et al.

Figure 1. There is a haemorrhagic lesion in the corpus callosum to the left of the midline. The interventricular septum has been torn. There is a parasagittal haematoma (gliding contusion) and a small haematoma in the right basal ganglia. (From Adams & Graham 1988.)

Figure 2. There is a shrunken scar in the narrowed corpus callosum (arrow). There is also enlargement of the lateral ventricles. (From Adams & Graham 1988.)

Figure 3. There is a recent haemorrhagic lesion in the dorsolateral quadrant of the rostral pons. (From Adams & Graham 1988.)

Figure 4. There is a granular and slightly shrunken lesion in the dorsolaterdl quadrant of the rostral pons. (From Adams & Graham 1988.)

54 J.Hume Adams et ul.

antero-posterior distance of several centimetres. Occasionally, abnormalities are restricted to the splenium. If the patient survives for some weeks, the lesion becomes rather granular and brown in colour, whilst with a survival of months it is represented by a shrunken cystic scar (Figure 2). The lesions in the dorsolaterdl quadrants of the rostra1 brain stem follow a similar time course (Figures 3 & 4), but occasionally there is no haemorrhage into these focal lesions, there simply being histological evidence of vacuolation and rarefaction of the affected tissue. The diffuse injury to axons can only be seen microscopically and takes three forms depending on the duration of survival. In patients of short survival (days) there are large numbers of axonal bulbs throughout the white matter of the cerebral hemispheres, the cerebellum and the brain stem, as well as adjacent to the focal lesions. These can be seen in sections stained by haematoxylin and eosin (Figure 5a), but they are more readily apparent in silver impregnations (Figure 5b). In patients of intermediate survival (weeks), there are large numbers of small clusters of microglia throughout the white matter of the cerebral hemispheres, the cerebellum and the brain stem (Figure 6) . There is also a diffuse non-specific astrocytosis. In patients who survive vegetative for a longer period (months), there is Wallerian-type degeneration of the white matter in the cerebral hemispheres, the brain stem and the spinal cord (Figure 7). By this time the ventricles are usually enlarged as a result of a reduction in volume of the white matter (Figure 2).

The relationship between diffuse axonal injury and fracture of the skull, type of accident and lucid interval are given in Table 1 . As in our earlier study (Adams et ul. 1982), there is a significantly reduced incidence of fracture of the skull and a lucid interval, and a significantly increased incidence of road traffic accidents and a decreased incidence of falls in patients who sustain diffuse axonal injury, even when less severe degrees of diffuse axonal injury are included. In our earlier studies on patients with the most severe type of diffuse axonal injury, we established that it occurred only as a result of a fall from greater than ones own height (Adams et ul. 1984). Analysis of the 22 cases with all grades of diffuse axonal injury resulting from a fall discloses that at least 19 fell from a height-six down stairs, three from a 2nd or 3rd storey window, three from a ladder, two from a lorry and one each into the hold of a ship, down a lift shaft, from a railway bridge, from the top of a wall and down a mountain! In the remaining three cases there was some doubt as to the nature of the injury: two were found unconscious in the street under the influence of alcohol and one was found unconscious in an office. One of the former had grade I diffuse axonal injury, while the other two had microscopic grade 2 diffuse axonal injury. Thus, we have still not encountered a case that could be classified as severe diffuse axonal injury in someone who has sustained a simple fall from his or her own height.

In contrast to our previous studies on the most severe type of diffuse axonal injury when all of the patients had been unconscious from the moment of injury, 17 of the 122 patients with diffuse axonal injury experienced a lucid interval. Only two cases had a total lucid interval, i.e. they were able to talk clearly and rationally, and both of these had grade I diffuse axonal injury. Of the 15 cases who had a partial lucid interval, not one had the most severe type as shown by the presence of macroscopic focal lesions in the corpus callosum and in the dorsolateral quadrant of

Diffuse axonal injury 5 5

Table 1. The relationship between diffuse axonal injury (DAI) and fracture of skull, type of injury and the occurrence of a lucid interval as defined in the text

Cases with DAI Cases without DAI P (n= 122) (n=312) value

Fracture of skull 70 (57%) 254 (8 I %)

56 J.Hume Adams et al.

There is no doubt that at least the most severe grades of diffuse axonal injury are a cause of clinical evidence of severe brain damage in head injury, and that it is the commonest cause of the vegetative state and severe disability after a non-missile head injury (McLellan et al. 1986). Whether or not, in the absence of other complications of head injury, it is a cause of death remains unclear. Even in the present series there was evidence of a high intracranial pressure in 70% of the cases. It is our impression, therefore, that in patients with diffuse axonal injury and no other complications leading to a high intracranial pressure, they either die as a result of other factors such as multiple injuries, hypoxic brain damage or cerebral fat embolism, or exist in coma, vegetative or severely disabled until they die as a result of a respiratory infection. Axonal bulbs are brought about by the extrusion of axoplasm resulting from anterograde axoplasmic flow, and classic swellings take about 12-18 h to develop. Hence the difficulty in diagnosing beyond reasonable doubt the occurrence of diffuse axonal injury in patients who survive for a shorter time than this after their injury.

Thus, since most patients who sustain diffuse axonal injury die within a week or so of their injury, rather than weeks or months, the diagnosis is based on the presence of axonal bulbs rather than the occurrence of clusters of microglia and long tract degeneration in longer survivors. Axonal bulbs, although they can be seen in sections stained by haematoxylin and eosin (Figure 5a), are best seen by the silver impregnation techniques (Figure 5b), or by immunocytological techniques for neurofilaments. According to Vowles et al. (1 987) the Nauomenko-Feigin technique is required to demonstrate these swellings in early infancy. A major problem in this field is the interpretation of sinusoidally enlarged axons which, when present on their own, may be indistinguishable from post-mortem artifact. It has been suggested, however, that when associated with an astrocytic response, such sinusoidal changes are likely to represent the earliest stage in the time course of the development of axonal bulbs and therefore should correlate with the depth of coma (Vanezis et al. 1987). However, in our experience such axons are often seen in control specimens and we are therefore reluctant to diagnose diffuse axonal injury unless we can identify at least occasional classic axonal bulbs in a patient surviving some 15- 18 h after head injury. There is no doubt that genuine axonal injury can be identified earlier in perfusion-fixed experimental material (Graham et al. 1985, Erb & Povlishock 1988, Maxwell et al. 1988), but it is difficult to identify these changes in the brains of head-injured individuals who live for only a short time. The immunohistochemical identification of an astrocytic response, while helpful, is by no means conclusive because of pre-existing disease and the vagaries of immunohis- tochemical staining in post-mortem material.

The identification and definition of the less severe grades of diffuse axonal injury have considerable implications for pathologists, particularly those involved in forensic work. Since the diffuse injury to axons can only be identified microscopi- cally, appropriate histological screening therefore has to be undertaken in any fatal head injury to ascertain whether or not diffuse axonal injury is present. This is most readily seen in the parasagittal white matter, in the corpus callosum, in the midbrain and in the pons. Of particular medicolegal significance is the fact that we have still to

Difluse uxonal injury 57

encounter a case of diffuse axonal injury in anyone who has simply fallen from his or her own height, the inference being that the change in velocity of the skull in a fall of this type is insufficient to produce the severe shear and tensile strains in the brain required to produce axonal disruption. Thus, if diffuse axonal injury is found in a patient who appears to have died as a result of such a fall, the pathologist must be alerted to the fact that the original injury must have been much more severe, e.g. as a result of an assault when the head may have been forcibly propelled onto the ground or a solid wall.

Throughout the long period that we have had an interest in brain damage resulting from head injury, it has become increasingly apparent that there are certain misconceptions about the interpretation of some of the lesions on which we have placed particular importance, viz. focal lesions in the corpus callosum, axonal bulbs, clusters of microglia and long tract degeneration. Thus, a lesion in the corpus callosum is not diagnostic of diffuse axonal injury since focal infarcts (pale or haemorrhagic) are frequently brought about by distortion and shift of the brain caused by an intracranial haematoma leading to a compromised circulation through the pericallosal arteries. Axonal bulbs are not restricted to head injury since they occur in any situation where axons have been disrupted; they will inevitably occur adjacent to focal lesions such as contusions, infarcts or haematomas. There are also other causes of multiple microscopic collections of microglia in the white matter, the most important of which are severe hypoxic brain damage, cerebral fat embolism and any type of meningoencephalitis. Wallerian degeneration, too, is not restricted to patients who have survived for several months after a head injury since this is simply indicative of degeneration of a tract brought about by any type of brain damage, particularly infarction.

In our earlier studies on severe diffuse axonal injury, not one of the patients was able to talk after their injury (Adams et ul. 1982). It is of interest, therefore, that 17 (14%) of the 122 cases with diffuse axonal injury in the present series did experience a complete or partial lucid interval as defined above. The two with a complete lucid interval had grade I injury, while the remaining 15 had grade 2, in only one of whom was the focal lesion identifiable macroscopically. Thus, even in this enlarged series, none of the 17 patients who talked had the most severe type of diffuse axonal injury. Furthermore, all died as a result of some other type of pathology-raised intracranial pressure brought about by oedema related to contusions, diffuse brain swelling or intracranial haematoma in 15, fat embolism in one, and massive gastro- intestinal haemorrhage in one.

It is probably unfortunate that so much emphasis has to be placed on focal lesions in defining the severity of a diffuse process. However, identical structural changes have been produced in non-human primates subjected to angular acceleration of the head, and there is a close correlation between the severity of the brain damage and the grade of diffuse axonal injury (Gennarelli et ul. 1982, 1987, Gennarelli, Adams & Graham 1986). In this model, all three grades of diffuse axonal injury have been produced, the severity of the axonal injury depending principally on the magnitude and duration of the acceleration and on the direction of head motion. With sagittal acceleration, grade 1 but not grades 2 and 3 can be produced.

58 J.Hume Adams et al.

Acceleration of the same magnitude in the coronal plane produces more severe diffuse axonal injury, usually grade 3, whilst rotation in the horizontal plane produces diffuse axonal injury of intermediate grade similar to that seen in man. As the grade of diffuse axonal injury increases, there is deeper and more prolonged coma and the residual neurological deficits in survivors are more severe. Focal axonal injury remote from the site of impact has been produced in other animal models (Povlishock et al. 1983, Povlishock & Becker 1985, Lighthall 1988) but this type of damage is not the same as that seen in diffuse axonal injury in man.

Acknowledgements

Figures 1, 2, 3 ,4 and 7 are reproduced from An Introduction to Neuropathology by J.H.Adams and D.I.Graham by permission of Churchill Livingstone, Edinburgh.

References

ADAMS J H , MITCHELL DE, GRAHAM DI, DOYLE D. Diffuse brain damage of immediate impact type. Its relationship to primary brain stem damage in head injury. Brain 1977; 100; 489-502.

ADAMS JH, GRAHAM DI , MURRAY LS, SCOTT G. Diffuse axonal injury due to nonmissile head injury in humans: an analysis of 45 cases. Ann. Neurol. 1982; 2; 557-563.

ADAMS JH, MURRAY M F. Atlas of Post-Mortem Techniques in Neuropathology. Cambridge: Cambridge University Press, 1982.

ADAMS JH, DOYLE D, GRAHAM DI , LAWRENCE AE, MCLELLAN DR. Diffuse axonal injury in head injuries caused by a fall. Lancet 1984; ii; 1420-1422.

ADAMS J H , DOYLE D, GRAHAM DI, LAWRENCE AE, MCLELLAN DR. Gliding contusions in nonmissile head injury in humans. Arch. Pathol. Lab. Med. 1986a; 110; 485-488.

ADAMS JH, DOYLE D, GRAHAM DI , LAWRENCE AE, MCLELLAN DR. Deep intracerebral (basal ganglia) haematomas in fatal non-missile head injury in man. J . Neurol. Neurosurg. Psychiat.

ADAMS J H , GRAHAM DI. An Introduction to Neuropathology. Edinburgh: Churchill Livingstone, 1988.

ERB DE, POVLISHOCK JT. Axonal damage in severe traumatic brain injury: an experimental study in cat. Acta Neuropathol. 1988; 76; 347-358.

GENNARELLI TA, THIBAULT LE, ADAMS JH, GRAHAM DI, THOMPSON CJ, MARCINCIN RP. Diffuse axonal injury and traumatic coma in the primate. Ann. Neurol. 1982; 12; 564-574.

GENNARELLI TA, ADAMS JH, GRAHAM DI. Diffuse axonal injury-a new conceptual approach to an old problem. In Baethmann A, Go KG, Unterberg A eds. Mechanisms ofSecondary Brain Damage. New York: Plenum Press, 1986; 15-28.

GENNARELLI TA, THIBAULT LE, TOMEI G, WISER R, GRAHAM DI , ADAMS J H . Directional dependence of axonal brain injury due to centroidal and non-centroidal acceleration. 31st Stapp Car Crash Conference. Society of Automotive Engineers, 1987; 49-53.

GRAHAM DI , ADAMS J H , LOGAN S, GENNARELLI TA, THIBAULT L. The distribution, nature and time course of diffuse axonal injury. Neuropathol. Appl. Neurobiol. 1985; 11; 319.

GRCEVIC N. Topography and pathogenic mechanisms of lesions in inner cerebral trauma. Rad. jug. Akad. Znan. Umj. 08 med. Nauke. 1982; 42/18; 265-331.

JELLINGER K . Pathology and pathogenesis of apallic syndrome following closed head injuries. In Ore GD, Gerstenbrand F, Lucking CH, Peters G, Peters UH eds. The Apaflic Syndrome. Berlin: Springer, 1977; 88-103.

1986b; 49; 1038-1043.

Diffuse axonal injury 59

JELLINGER K , SEITELBERGER G. Protracted post-traumatic encephalopathy: pathology, pathogene-

LIGHTHALL J W. Controlled cortical impact: a new experimental brain injury model. J . Neurotrauma

MAXWELL W L , KANSAGRA A M , GRAHAM D I , ADAMS J H , GENNARELLI T A . Freeze fracture studies of reactive myelinated nerve fibres after diffuse axonal injury. Acta Niwropafhol. 1988; 76;

MCLELLAN D R , ADAMS J H , GRAHAM D I , KERR AE, TEASDALE G M . The structural basis of the vegetative state and prolonged coma after non-missile head injury. In Pap0 I, Cohadon F, Massarotti M eds. Le Coma Traumafique. Padova: Liviana Editrice, 1986; 165- 185.

MITCHELL D E , ADAMS J H . Primary focal impact damage to the brain stem in blunt head injuries- does it exist? Lancer 1973; 2; 215-218.

OPPENHEIMER D R . Microscopic lesions in the brain following head injury. J. Neurol. Neurosurg. Psychiat. 1968; 31; 299-306.

PEERLESS SJ , REWCASTLE NB. Shear injuries of the brain. Can. Med. Assoc J . 1967; 96; 577-582. PEmRs G, ROTHEMUND E. Neuropathology of the traumatic apallic syndrome. In Ore G D ,

Gerstenbrand F, Lucking CH, Peters G, Peters U H eds. The Apallic Syndrome. Berlin: Springcr, 1977; 78.

sis and clinical implications. J. Neurol. Sci. 1970; 10; 51-94.

1988; 5; 1-15.

395-406.

PILZ P. Axonal injury in head injury. Acta Neurochir. (Wien) Suppl. 32; 1983; 119 123. POVLISHOCK J T , BECKER D P . Fate of reactive axonal swellings induced by head injury. Lab. Invest.

POVLISHOCK J T , BECKER D P , CHENG C L Y , VAUGHAN G W . Axonal change in minor hcad injury.

REILLY P L , ADAMS J H , GRAHAM D I , JENNETT B. Patients with head injury who talk and die.

STRICH SJ . Lesions in the cerebral hemispheres after blunt head injury. J. Clin. Puthol. 1970; 23 (suppl

VANEZIS P, CHAN K K , SCHOLTZ C L . White matter damage following acute head injury. Forensic

VOWLES GH, SCHOLTZ C L , CAMERON J M . Diffuse axonal injury in early infancy. J . Clin. Pathol.

Z I M M E R M A N R A , BILANIUK LT, GENNARELLI T A . Computed tomography of shearing injuries of

1985; 52; 540-552.

J . Neuropathol. Exp. Neurol. 1983; 42; 225-242.

Lancet 1975; 2; 375-377.

4); 166-171.

Sci. Inter. 1987; 35; 1-10.

1987; 40; 185-189.

the cerebral white matter. Radiology 1978; 127; 393-396.