J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.52.7.838 on 1 July 1989. Downloaded from Journal ofNeurology, Neurosurgery, and Psychiatry 1989;52:838-841 Diffuse axonal injury in head trauma PETER C BLUMBERGS, NIGEL R JONES, JOHN B NORTH From the Neuropathology Laboratory, Institute ofMedical and Veterinary Science and Neurosurgery Department, Royal Adelaide Hospital, Adelaide, South Australia SUMMARY Diffuse axonal injury (DAI) as defined by detailed microscopic examination was found in 34 of 80 consecutive cases of head trauma surviving for a sufficient length of time to be clinically assessed by the Royal Adelaide Hospital Neurosurgery Unit. The findings indicate that there is a spectrum ofaxonal injury and that one third ofcases ofDAI recovered sufficiently to talk between the initial head injury producing coma and subsequent death. The macroscopic "marker" lesions in the corpus callosum and dorsolateral quadrants of the brainstem were present in only 15/34 of the cases and represented the most severe end of the spectrum of DAI. Diffuse axonal injury (DAI) that is, widespread superior cerebellar peduncles, (2) evidence of diffuse damage to axons in the white matter of the brain, was damage to axons. originally defined by Strich' and the concept was Patients who sustain severe DAI are unconscious expanded by Adams et al.2 Strich described diffuse from the moment ofimpact, do not experience a lucid Protected by copyright. degeneration ofthe cerebral white matter in a series of interval, and remain unconscious, vegetative or at patients with severe post traumatic dementia and in a least severely disabled until death. Their clinical state subsequent paper concluded that the axonal damage has been referred to in the past by some as primary was produced by mechanical forces shearing the fibres brainstem injury, but although there is damage to the at the moment of impact.3 Since her original observa- brainstem in DAI, it is always accompanied by tions the same pathological process has been described evidence of diffuse brain damage." DAI occurs most by others as "shearing injury',,4 "diffuse damage of commonly in patients injured in road traffic accidents2 immediate impact type"5 and "diffuse white matter and rarely in association with simple falls.'2 shearing injury".6 This concept of DAI has not remained unchallenged and some workers maintain Materials and methods that it arises as a secondary event due to hypoxic damage, post traumatic oedema or to secondary Diffuse damage to axons can only be detected micro- a tentorial scopically in appropriately stained tissue; it takes the forms of compression of the brainstem as result of axonal retraction balls (RB) in short survivors (hours to herniation.7" days),'3 microglial stars in cases of intermediate survival A major advance in this field has occurred recently (several days to weeks)2 or evidence of degeneration of fibre http://jnnp.bmj.com/ since Gennarelli and his group9'0 have shown that tracts in the white matter in patients who survive for many similar clinical and structural changes can be weeks or months.2 produced experimentally in subhuman primates using The present study was undertaken to investigate the extent non-impact controlled angular acceleration of the of axonal injury (Al) in a series of 426 fatal head injuries. head in the absence of any increase in ICP or Eighty five were patients who survived a sufficient length of hypoxaemia. Adams et al2 reported a detailed time to be managed by the Neurosurgery Department of the of45 cases and defined the Royal Adelaide Hospital. Excluded from further analysis neuropathological analysis were 341 road traffic accident victims who had not survived characteristic features ofDAI as (1) focal lesions in the long enough to reach a hospital and five patients with on September 28, 2021 by guest. corpus callosum and dorsolateral quadrant (or gunshot wounds. This left a total of80 brains ofpatients who quadrants) of the rostral brainstem adjacent to the survived for a sufficient length of time for conscious level to be assessed by the Royal Adelaide Hospital Neurosurgery Unit using the Glasgow Coma Scale.'4 These clinical data were used to define a group of 11 patients, all of whom had Address for reprint requests: Dr P C Blumbergs, Institute of Medical Box Rundle Mall SA 5000. been rendered unconscious by the initial head injury, but who and Veterinary Science, 14, PO, Adelaide, recovered and talked at some point between the time ofinjury Received 19 July 1988 and in revised form 22 February 1989. and death.5 These patients were classified as having had a Accepted 7 March 1989 lucid interval. 838 J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.52.7.838 on 1 July 1989. Downloaded from Diffuse axonal injury in head trauma 839 .i-P i .ktt .cs .F._! r W C. .-. .0 ir- If a~~~~~ t 'p. 0 - I 0 3- iP -:L ..J -F 0 AI0** i _S 7 0 U ".r iI v .1 I 1%. A i _ e C~~~~~~~~~ I/ i! I Amj ARP.- .- _.LA C4 a a Protected by copyright. 2 _ Fig Axonal retraction balls and swellings in corpus callosum. Immunochemical stainfor neurofilament proteins (NFP). ( x 200) The brains were analysed in a manner similar to that eight hours. Retraction balls and axonal swellings are termed developed by Hume Adams and his group'6 and in addition the microscopic markers of DAI. All focal lesions were serial whole brain sections were microscopically assessed. carefully noted as RBs and axonal swellings are frequently Detailed descriptions ofthe neuropathological findings were found around infarcts2 (especially in the brain stem) and made in all cases. Any lesions found (including skull and haemorrhages. A case was classified as mild DAI when at scalp injuries) were charted on a series ofline diagrams ofthe least a few unequivocal RBs were found. When RBs were skull, cerebral hemispheres, brainstem and cerebellum: numerous and widespread, AT was classified as severe. In macroscopic abnormalities were recorded photographically. addition, in all cases the "macroscopic lesions" (tears, The cerebral hemispheres were cut in 1 cm thick slices in the haemorrhages and infarcts) were looked for in the corpus coronal plane and whole brain paraffin sections prepared of callosum and dorsolateral brain stem and recorded: these are each slice and stained by Nissl's method using cresyl violet, termed macroscopic markers of DAI. Using these markers http://jnnp.bmj.com/ Weil stain for myelin, haematoxylin and eosin (H & E) and we have classified our cases into four groups. various other techniques as appropriate. The cerebellum was (1) Corpus callosum and dorsolateral quadrant of brain sectioned in the sagittal plane and representative blocks stem. stained as above. The brainstem was sliced into nine blocks (2) Corpus callosum only and two representative sections each of midbrain, pons and (3) Dorsolateral quadrant of brain stem only medulla were stained by the H & E, Nissl and Weil methods. (4) No lesions in corpus callosum or dorsolateral quadrant Axonal injury was assessed by the presence of retraction of brain stem. balls (RBs) not associated with focal lesions (fig). RBs were easily identified in H & E stained sections as eosionphilic on September 28, 2021 by guest. round, oval or elongated masses ofvarying size depending on Results the direction of sections but were particularly well demon- strated by silver methods such as the Glees and Marsland and Evidence of DAI was found in 34 of the 80 cases that Palmgren techniques for axons, which were routinely used on survived to reach the Royal Adelaide Hospital. the brainstem and corpus callosum blocks. Immunocyto- chemical techniques for neurofilament proteins which specifically identify axons were also used on selected cases Survival times and axonal swellings could be identified within 1-2 hours by The survival times for these cases compared with those this method whereas with silver stains they became visible at without DAI are seen in table 1. J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.52.7.838 on 1 July 1989. Downloaded from 840 Blumbergs, Jones, North Table 1 Survival time Table 3 Duration ofinability to speak-DAIgroup Survival time Diffuse Axonal Recovered Injury 1-8 8-24 24-48 2-7 1-4 >4 Macroscopic Never spoke to lucid N hrs hrs hrs days weeks weeks markers N after injury interval Severe Mild DAI present 34 3 3 2 13 7 6 CC + DLQ 15 0 0 12 3 DAIabsent 46 12 16 2 11 4 1 CC alone 8 4 4 7 1 DLQ alone 3 2 1 3 no lesion CC Causes ofinjury + DLQ 8 2 6 2 6 Table 2 shows the breakdown as to the type of accident. Traffic accidents were the leading cause of head injury. In five cases, the injuries resulted from Discussion falls; in four ofthese, the patient fell from a considera- ble height, greater than the person's own height. DAI is well recognised as the most common structural abnormality underlying severe neurological dis- Distribution ofDAI ability"7 and the vegetative state'8 following head The corpus callosum was affected in 94% of the cases injury. Our findings ofvariation in the severity ofDAI, with AT, the internal capsule in 86%, the cerebral and that one third of the cases with DAI were not white matter in 79%, the fornices in 76%, the mid- permanently unconscious from the time of injury brain in 78%, the pons in 88%, the medulla in 36% support the concept of a spectrum of axonal injury. and the cerebellum in 72%. The microscopic markers Gennarelli et al9"0 have hypothesised that there is a of AI were present in both the cerebrum (corpus continuum ofaxonal damage varying from functional callosum, fornices, cerebral white matter, internal abnormalities alone to increasingly severe and wide- Protected by copyright.