Journal ofNeurology, Neurosurgery, and Psychiatry, 1977, 40, 1203-1210

Histopathological variability in 'standardised' spinalcord trauma1G. KOENIG AND G. J. DOHRMANN 2

From the Sections of Neurosurgery and Neuropathology, Yale University School of Medicine,New Haven, Connecticut, USA

S U MM ARY Feline spinal cords were traumatised by the weight dropping technique. Fivetrauma groups were studied (5 gX80 cm, 10 gX40 cm, 20 gX20 cm, 40 gXl0 cm, and 80 gX5 cm), each having a 'standardised' injury of '400 g-cm.' The spinal cords were sectionedserially two hours after contusion and examined by light microscopy. Relative to the largerweights falling from lesser heights, the smaller weights falling from greater heights wereassociated with less haemorrhage, oedema, axonal disruption, and myelin fragmentation as wellas a smaller volume of grey matter containing altered anterior horn cells. In all trauma groupsthe cortical evoked responses disappeared at the time of the injury and did not reappear. Eventhough each trauma group received a '400 g-cm' contusion, each weight-height combination wasassociated with differing degrees of histopathological alterations. A plea is made for moreaccurate quantitation of experimental spinal cord trauma than the 'g-cm' unit.

Since the experiments of Allen in 1911, experi-mental spinal cord contusion has been quantitatedin gram-centimetre (g-cm) units. Trauma is pro-duced by a weight (g) falling a specified distance(cm) through a vertical tube and striking an im-pounder that is resting on the posterior surfaceof the spinal cord. The product of the weight andthe height from which it falls was expressed asg-cm. The g-cm unit is not precise as variouscombinations of weights and heights can give thesame product; nevertheless, the g-cm unit con-tinues to be used. Certain investigators have usedvarious weight-height combinations interchange-ably in their experiments.The purpose of the present study was to describe

the histopathology of the spinal cord and thecortical evoked responses after 400 g-cm'standardised' trauma, an injury produced by dif-ferent weights falling from various heights buteach giving the same g-cm value.

l Supported in part by Grant NS 10174 from the National Institute ofNeurological and Communicative Disorders and Stroke.

2Address for reprint requests: George J. Dohrmann, MD, PhD.Sections of Neurosurgery and Neuropathology, Yale UniversitySchool of Medicine, New Haven, Connecticut 06510, USA.Accepted 30 March 1977

Materials and methods

Cats were anaesthetised with an intraperitonealinjection of sodium pentobarbitone (35 mg/kg).A tracheotomy was performed and muscularparalysis was obtained with gallium triethiodide.The animals were maintained on a Harvard smallanimal respirator with the end-tidal pCO2 main-tained between 2% and 4%. Blood pressure wasrecorded continuously via a catheter in the femoralartery. Temperature was monitored by a rectalprobe and kept at approximately 370C with thehelp of a heating pad. A screw electrode wasplaced in the skull over the right somatosensorycortex and the left posterior tibial nerve wasstimulated electrically (0.1 ms pulse; 1 V/s). ABiomac 1000 computer was used to analyse thetransient cortical potentials. The computer-averaged somatosensory cortical evoked response(CER) was displayed on an oscilloscope and photo-graphed. Cortical evoked responses were obtainedbefore trauma, immediately after trauma, and atone and two hours after the trauma. Laminec-tomies were performed to expose the T5-6 levelof the spinal cord with dura mater intact. Fifteencats were divided into five trauma groups of threecats each, and the spinal cords were contused witha 400 g-cm injury using the weight dropping

1203F

G. Koenig and G. J. Dohrmann

technique (Albin et al., 1968; Dohrmann et al.,1976): 5 gX80 cm; 10 gX40 cm; 20 gX20 cm;40 gX10 cm; and 80 gX5 cm. As controls, twocats had laminectomies but were not traumatised.At two hours after the contusion, the cats wereperfused with 10% formalin in a retrogrademanner via the abdominal aorta, and the exposedportion of the spinal cord was excised. The speci-men was embedded in paraffin and the trau-matised area was sectioned serially (8 ,um) eithertransversely or longitudinally (two specimenstransversely and one longitudinally in eachtrauma group). Alternate sections were stainedwith haematoxylin and eosin, Bodian, orKluver-Barrera stains and examined by lightmicroscopy. All sections were examined and thesection showing maximal tissue disruption waschosen for quantitation. Haemorrhage and oedemawere graded on a 0 to 4 scale as described inTables 1 and 2. The cranio-caudal distance overwhich abnormal anterior horn cells were notedwas measured and expressed in millimetres.Quantitation of axonal disruption and myelinfragmentation was carried out by fibre counts ormyelin sheath counts at the area of maximal in-jury. These were expressed as the percentage ofabnormal fibres relative to the total number offibres noted in the section of the posteriorcolumns or of the anterior white matter.

Table 1 Haemorrhage

Trauma group

Sgx lOgx 20gx 40gx 80gx80 cm 40 cm 20 cm 10 cm S cm

Grey matter 2 3 3 4 32 3 3 4 33 2 2 3 3

White matter 2 2 2 3 20 2 2 3 32 1 2 2 2

0=no haemorrhage; 1 =few extravascular erythrocytes; 2=smallscattered haemorrhages; 3 =large discrete haemorrhages; 4 =coale-scence of haemorrhages.

Table 2 Oedema

Trauma group

Sgx lOgX 20gx 40gx 80gx80 cm 40 cm 20 cm 10 cm 5 cm

2 3 3 4 31 2 3 4 41 2 3 4 4

o=no oedema; 1=oedema involving entire area of haemorrhage;2=significant extension of oedema outside of haemorrhagic region;3 =as in 2 and some involvement of white matter; 4 =as in 2 and muchinvolvement of white matter.

Results

Although each animal sustained a 400 g-cm con-tusion of the spinal cord, numerous differences inthe histopathology between the various traumagroups were noted.

CONTROL GROUPNo alterations in the histology of the spinal cordwere noted in these animals examined two hoursafter laminectomy.

EXPERIMENTAL GROUP

HaemorrhageIn the 5 gX80 cm trauma group small areas ofhaemorrhage were noted in the central greymatter and a few scattered areas of haemorrhagewere present in the white matter immediatelysurrounding the grey matter (Fig. 1). Mainly inthe grey matter, the haemorrhages increased innumber and size in the 10 gX40 cm, 20 gX20 cmand 40 gX 10 cm trauma groups respectively (Fig.2). In the last group coalescence of the haemor-rhage was seen centrally. The amount of haemor-rhage in the 80 gX5 cm group was less than thatat 40 gX10 cm but more than in the other groups(Table 1). Little subarachnoid haemorrhage waspresent in the 5 gX80 cm and 10 gX40 cm traumagroups while much more subarachnoid haemor-rhage was noted in the remaining groups.

OedemaIn the 5 gX80 cm group oedema was present inthe region of haemorrhage; however, in the othergroups the oedema involved large areas of the

.4

Fig. 1 Photomicrograph illustrating small scatteredareas of intramedullary haemorrhage in the S gX80 cm trauma group. Bodian X 12.

1204

Histopathological variability in 'standardised' spinal cord trauma

Fig. 2 Transverse section ofspinal cord showing centralhaemorrhage in the 40 gX10 cmtrauma group. Bodian X 18.

spinal cord both proximal and distal to the haemor-rhage. The amount of oedematous tissue increasedin the 10 gX40 cm, 20 gX20 cm, 80 gX5 cm, and40 gX10 cm groups respectively. When present,the oedema in the white matter was noted tospread farther than that in the grey matter.(Table 2)

Anterior horn cellsNone of the five trauma groups showed a directrelationship between the number of pathologicallyaltered neurones and the amount of injury. How-ever, the distance over which the damagedneurones could be found did correlate somewhatwith the degree of trauma (Table 3), but thehigher incidence of altered anterior horn cells inthe 5 gX80 cm group is not explained.

In the region of the haemorrhage in all groupsexcept 5 gX80 cm and 10 gX40 cm, no recognis-

Table 3 A nterior horn cells

Mean length ofgrey mattercontaining Anterior horn cellsaltered anterior (mean)horn cells

Trauma grouip (mm) Abnormal Normal

(%) (%)S gx80cm 6.0 90 10lOgx4Ocm 5.4 88 122Ogx2Ocm 6.7 85 1540 g x 10 cm 8.8 75 2580gx5cm 7.0 85 15

able tissue remained. Three concentric areasaround the haemorrhage were noted in longi-tudinal section (Fig. 3): (I) Around the haemor-rhage in the grey matter only a few neuroneswere seen. These appeared swollen with veryeosinophilc cytoplasm and no nuclei were noted.(II) Around this area was a region of neuronesthat were shrunken and angular and had deeplybasophilic cytoplasm. No nuclei were visible.(III) The outermost region of the grey matter hadneurones of normal shape with normal appearingnuclei; however, clumping of the Nissl substancewas present. Both proximally and distally beyondregion III, the neurones were without pathologicalchange.

AxonsLongitudinal sections of the spinal cords of the5 gX80 cm and the 10 gX40 cm groups showedfragmentation of the axons of the inner third of

Fig. 3 Schematic representation of spinal cord inlongitudinal section showing three grey matter regions(1, 11, and III) containing altered anterior horn cellsboth proximal and distal to the central haemorrhage.(The characteristics of cells in each of these regionsare detailed in the text.)

1205

1206

the posterior columns. In the 20 gX20 cm groupa similar number of axons appeared fragmentedin the posterior columns (Fig. 4); however, ap-proximately one-third of those of the anteriorwhite matter showed disruption as well. The mostsevere axonal injury was noted in the 40 gX 10 cmgroup where most of the axons of the posteriorcolumns were fragmented and approximately two-thirds of those in the anterior white matter werefragmented (Fig. 5). The number of disruptedaxons in the 80 gX5 cm group was slightly less

G. Koenig and G. J. Dohrmann

than that noted in the 40 gX 10 cm group(Table 4).

In transverse section a few swollen axons werenoted in the 5 gX 80 cm group in the lateral whitematter immediately adjacent to the grey matter.These increased in number in the 10 gX40 cmand 20 gX20 cm groups such that in the latterthe abnormal axons occupied the entire layer ofwhite matter around the grey matter. This layerof abnormal axons increased in width in the40 gXlO cm and 80 gX5 cm groups and was

Fig. 4 Fragmentation of theaxons of the medial third of theposterior columns is seen inlongitudinal section of spinalcord. 20 gX20 cm trauma.Bodian X35.

Fig. 5 Approximately 80% ofthe width of the longitudinally-sectioned posterior columnscontains broken axons.40 gX10 cm trauma.Bodian X34.

..............-^a|b....--- s f.w e -- ;_;s~MI

Histopathological variability in 'standardised' spinal cord trauma

Table 4 Disrupted axons and fragmented myelin

Anterior whitePosterior colmnins mnatter

Trauma group Axons Myelin Axons Myelin

Sgx8Ocm 35* 65 0 15lOgx4Ocm 25 65 0 020gx2Ocm 35 100 30 5040gxlOcm 80 100 65 10080gx5cm 75 100 55 100

* Mean percentage of fibres counted in all cats in each group.S.... ......e."

<i;N.

noted to involve approximately the medial half ofthe lateral white matter. Most of the axons inthese two trauma groups were shrunken.

MyelinIn both longitudinal and transverse sections, dis-ruption of the myelin was noted to correspond tothe areas of axonal damage described above; how-ever, the fragmentation of the myelin extendedmore anteriorly and posteriorly than the areas ofinjured axons (Fig. 6, Table 4). In transversesections of the spinal cord from the 40 gYX10 cmand 80 gX5 cm groups, areas of cavitation wereseen (Figs. 7 and 8).

Cortical evoked responsesThe cortical evoked responses were present in allanimals both before and after laminectomy. Im-mediately after trauma the CERs of all traumagroups disappeared and had not returned twohours after contusion.

Fig. 6 Longitudinal section of contused spinal cordshowing disrupted myelin in the medial two-thirds ofthe posterior columns. 5 gX80 cm trauma.Kluver-Barrera X 16.

Discussion

Schmaus (1890) conducted the first experiments onspinal cord trauma by striking the backs ofrabbits and studying the spinal cord. Over theensuing years various investigators studied experi-mental spinal cord trauma (Dohrmann, 1972). Thefirst standardisation and quantitation of spinalcord trauma was done in the experiments of Allen

Fig. 7 Myelin sheaths of thelateral white matter are largelyintact as seen in this transversesection of spinal cord.5 gX80 cm trauma.Kluver-Barrera X 130.

1207

G. Koenig and G. J. Dohrmann

Fig. 8 Disruption of myelin ofthe lateral white matter withearly cavitation in transversesection of spinal cord.80 gXS cm trauma.Kluver-Barrera X 130.

(191 1) where a weight was dropped through a tubeto strike the spinal cord. The trauma was des-cribed in g-cm units, the product of the weight(g) and the height (cm). Although g-cm is not a

true physical unit of measure, it was used todefine the degree of trauma delivered to the spinalcord. With the rekindling of interest in experi-mental spinal cord trauma in the late 1960s theg-cm unit of Allen was again used to describe theinjury (Osterholm, 1974). More recently variouscombinations of weights and heights have beenused to produce the same g-cm trauma. Muchvariability can be seen between the results ofcertain laboratories. Although some of this may

be because different trauma devices and differenttypes of animals are used, some of the variabilitymay well be due to the use of varying combina-tions of weights and heights which give a super-

ficial appearance of the same amount of trauma.The light microscopy (Goodkin and Campbell,

1969; Wagner et al., 1969; White et al. 1969;White and Albin, 1970; Ducker et al., 1971;Wagner et al., 1971; Wagner and Dohrmann, 1975;Yeo et al., 1975) and electron microscopy(Dohrmann et al., 1971; Dohrmann et al., 1972) ofexperimental spinal cord trauma have beendescribed. The central haemorrhage and thealterations in the neurones and myelinated fibresappear to increase with time after trauma. In thepresent experiment it has been shown that thehistopathology of 400 g-cm spinal cord trauma isquite variable. In general the changes are lessmarked in the trauma groups where a small weighthas fallen from a greater height-that is, 5 gX80 cm; 10 gX40 cm-than where a larger weight

has fallen from a lesser height-that is 40 gX10 cm; 80 gX5 cm. After trauma there is tearingof the thin-walled muscular blood vessels withinthe central grey matter which accounts for thehaemorrhagic areas noted therein (Dohrmann etal., 1971; Fairholm and Turnbull, 1971; Dohrmannand Allen, 1975). These blood vessels are prone todisruption by a postero-anterior force, such asused in this experiment, because they are orientedperpendicular to that force (Turnbull, 1972) andbecause the grey matter offers less support thanthe white matter. In this experiment the amountof haemorrhage increased from the 5 gX80 cmgroup to that at 40 gX10 cm. The change inmomentum (impulse of the deformation force) inthe 40 gX10 cm group would be over 20 timesthat of the 5 g X 80 cm group and a linear relation-ship exists between impulse and amount of intra-medullary haemorrhage (Dohrmann and Panjabi,1976). This increase in the impulse could accountfor the variation in the degree of vascular dis-ruption noted in these 400 g-cm trauma groups.

In spinal cord trauma the oedema is mainlyvasogenic. Plasma travels from the intravascularspace to the extracellular space of the spinal cordvia tears in the walls of blood vessels in the greymatter. Green and Wagner (1973) and Griffithsand Miller (1974) noted that the oedema spreadcentrifugally from the central grey matter to in-volve the remainder of the grey matter and the sur-rounding white matter. In this study the amountof oedema appeared to be related to the degree ofhaemorrhage and, therefore, the amount ofvascular disruption.

Neuronal changes were probably due to the

1208

Histopathological variability in 'standardised' spinal cord trauma

direct effect of the trauma as well as secondaryalterations such as ischaemia after the trauma andthe ensuing haemorrhage. With increase in themass of the falling weight, the length of greymatter over which neuronal alterations were notedincreased such that it was maximal in the 40 gX10 cm group.

Alterations in the axons and myelin were seen

first at the junction of the grey matter and thewhite matter. This can be explained in part byshearing or tensile stresses acting at tissue-tissueinterfaces. The axonal and myelin disruption in-creased in severity from the 5 gX 80 cm groupthrough those at 40 gX 10 cm. The energy trans-ferred to the spinal cord in the latter group isapproximately 100 times that transferred to thespinal cord in the former group (Dohrmann andPanjabi, 1976). This great differential in theamount of energy absorbed could explain thedifference seen in the amount of disruption ofaxons and myelin. It is interesting that the myelindisruption extends over a larger area than theaxonal fragmentation.The reason that the pathological changes seen

in the 80 gX5 cm trauma group are usually lessthan those of the 40 gX 10 cm group but greaterthan the other trauma groups is probably relatedto mechanical factors such as the ratio of theweight to the impounder (Dohrmann and Panjabi,1976).

Cortical evoked responses have been used toassess the function of the white matter of thespinal cord after trauma (Donaghy and Numoto,1969; D'Angelo et al., 1973). Absence of the CERcorrelates with injury to the posterior columns(Allen et al., 1974). Interestingly, in each of thetrauma groups in this study, the injury deliveredto the spinal cord was sufficient to abolish theCER for at least two hours after contusion.

In general the histopathological variability in400 g-cm trauma described here indicates that theg-cm unit for quantitating spinal cord trauma isimprecise. Certainly the 5 gX80 cm injury is dif-ferent from that at 10 gX40 cm, and so on. Thedegree of spinal cord trauma should be describedin terms of weight and height and not merely interms of g-cm. Impounder mass should also benoted if an impounder is used. Before comparisonscan be made between groups of animals sustainingdifferent amounts of spinal cord trauma, thistrauma must be more precisely defined.

ReferencesAlbin, M. S., White. R. J., Acosta-Rua, G., andYashon, D. (1968). Study of functional recovery

produced by delayed localized cooling after spinal

cord injury in primates. Journal of Neurosurgery,29, 113-120.

Allen, A. R. (1911). Surgery of experimental lesion ofspinal cord equivalent to crush injury of fracturedislocation of spinal column: a preliminary report.Journal of the American Medical Association, 57,878-880.

Allen, W. E., D'Angelo, C. M., and Kier, E. L. (1974).Correlation of microangiographic and electrophysi-ological changes in experimental spinal cord trauma.Radiology, 111, 107-115.

D'Angelo, C. M., VanGilder, J. C., and Taub, A.(1973). Evoked cortical potentials in experimentalspinal cord trauma. Journal of Neurosurgery, 38,332-336.

Dohrmann, G. J. (1972). Experimental spinal cordtrauma: a historical review. Archives of Neurology(Chicago), 28, 468-473.

Dohrmann, G. J., and Allen, W. E. (1975). Micro-circulation of traumatizecl spinal cord: a correla-tion of microangiography and blood flow patternsin transitory and permanent paraplegia. Journal ofTrauma, 15, 1003-1013.

Dohrmann, G. J., and Panjabi, M. M. (1976). Spinalcord deformation velocity, impulse and energy re-lated to lesion volume in "standardized" trauma.Surgical Forum, 27, 466-468.

Dohrmann, G. J.. Panjabi, M. M., and Wagner, F. C.(1976). An apparatus for quantitating experimentalspinal cord trauma. Surgical Neurology, 5, 315-318.

Dohrmann, G. J., Wagner, F. C., and Bucy, P. C.(1971). The microvasculature in transitory trau-matic paraplegia. An electron microscopic study inthe monkey. Journal of Neurosurgery, 35, 263-271.

Dohrmann, G. J., Wagner, F. C., and Bucy, P. C.(1972). Transitory traumatic paraplegia: electronmicroscopy of early alterations in myelinated nervefibers. Journal of Neurosurgery, 36, 407-415.

Donaghy, R. M. P., and Numoto, M. (1969). Prog-nostic significance of sensory evoked potential inspinal cord injury. Proceedings of the Spinal CordInjury Conference, 17, 251-257.

Ducker, T. B., Kindt, G. W., and Kempe, L. G. (1971).Pathological findings in acute experimental spinalcord trauma. Journal of Neurosurgery, 35, 700-708.

Fairholm, D. J., and Turnbull, I. M. (1971). Micro-angiographic study of experimental spinal cordinjuries. Journal of Neurosurgery, 35, 277-286.

Goodkin, R., and Campbell, J. B. (1969). Sequentialpathologic changes in spinal cord injury: pre-liminary report. Surgical Forum, 20, 430-432.

Green, B., and Wagner, F. C. (1973). Evolution ofedema in the acutely injured spinal cord: a fluore-scence microscopic study. Surgical Neurology, 1,98-1 01.

Griffiths, I. R.. and Miller, R. (1974). Vascularpermeability to protein and vasogenic oedema inexperimental concussive injuries to the canine spinalcord. Journal of the Neurological Sciences, 22,291-304.

Osterholm, J. L. (1974). The pathophysiological re-sponse to spinal cord injury: the current status of

1 209

G. Koenig and G. J. Dohrmann

related research. Journal of Neurosurgery, 40, 5-33.Schmaus, H. (1890). Beitraege zur pathologischenAnatomie der Rueckenmarkserschuetterung. Vir-chows Archiv fur Pathologische Anatomie undPhysiologie, 122, 470-495.

Turnbull, I. M. (1972). Microvasculature of the humanspinal cord. Journal of Neurosurgery, 35, 141-147.

Wagner, F. C., and Dohrmann, G. J. (1975). Altera-tions in nerve cells and myelinated fibers in spinalcord injury. Surgical Neurology, 3, 125-131.

Wagner, F. C., Dohrmann, G. J., and Bucy, P. C.(1971). Histopathology of transitory traumatic para-plegia in the monkey. Journal of Neurosurgery, 35,272-276.

Wagner, F. C., Dohrmann, G. J., Taslitz, N., Albin,

M. S., and White, R. J. (1969). Histopathology ofexperimental spinal cord trauma. Proceedings of theSpinal Cord Injury Conference, 17, 8-10.

White, R. J., and Albin, M. S. (1970). Spine and spinalcord injury. In Impact Injury and Crash Protection.Edited by E. S. Gurdjian, W. A. Lange, L. M.Patrick, and L. M. Thomas, pp. 63-85. Charles C.Thomas: Springfield, Illinois.

White, R. J., Albin, M. S., Harris, L. S., and Yashon,D. (1969). Spinal cord injury: sequential morph-ology and hypothermic stabilization. Surgical Forum,20, 432-434.

Yeo, J. D., Payne, W., and Hinwood, B. (1975). Theexperimental contusion of the spinal cord in sheep.Paraplegia, 12, 275-296.

1210