Journal of the Neurological Sciences, 1978, 37:215-225 © Elsevier/North-Holland Biomedical Press
S P E C T R O P H O T O M E T R Y OF T H E CSF (CSF-SPE) A N D T O M O G R A P H Y (CT) IN T R A U M A T I C H E A D INJURIES
215
COMPUTER
N. G. WAHLGREN*, K. G. KJELLIN and C. E. SODERSTROM Department of Neurology, Karolinska Hospital, S-104 O1 Stockholm 60 (Sweden)
(Received 12 January, 1978) (Accepted 3 March, 1978)
SUMMARY Combined examinations with quantitative CSF spectrophotometry (CSF-SPE) and computer tomography (CT) were performed on 53 patients with traumatic head injuries. In cerebral concussion the results were mainly normal in both examinations. In cerebral contusion bleeding patterns were found by CSF-SPE in all subjects, with a special bleeding pattern ($2 pattern) occurring in 86 70. CT showed findings described as typical for contusion in 8 of 14 examined patients, the remaining CT scans showing questionable or normal signs. In extra- and intracerebral haematomas, all patients had bleeding patterns on the CSF-SPE. A special bleeding component (H factor) was found in about 72 ~ . The H component was not observed during the first 3 to 4 days after the trauma. All but one patient examined later than the 4th day had an H component with or without an S-pattern. CT demonstrated a haematoma in 14 of 18 verified haematoma patients, while 4 subjects with subdural haematoma (e.g. one third of this patient group) had questionable CT findings. The combined examinations with CT and CSF-SPE, being complementary to each other, are of great value in the differential diagnosis of traumatic head injuries.
INTRODUCTION It is often difficult to distinguish between different traumatic head disorders, including cerebral concussion, cerebral contusion and extracerebral or intracerebral haematomas by usual clinical criteria. The high diagnostic significance of quantitative CSF spectrophotometry (CSF-SPE) in cerebrovascular disease has been established This investigation was supported by a grant from the Trygg-Hansa Research Fund for Bodily Injury in Personal Accident Insurance. * To whom correspondence should be sent.
216 by Kjellin and S6derstr6m in several papers since 1972 (see for instance Kjellin and S6derstr6m 1974) and in subacute and chronic subdural haematomas by Kjellin and Steiner (1974). The diagnostic possibilities of computer tomography (CT) in traumatic head injuries were postulated by Paxton and Ambrose (1974) and have been confirmed by several authors, e.g. Dublin, French and Renick (1977, incl. ref.). Combined examinations with CSF-SPE and CT in cerebrovascular disease make it possible to obtain a specific diagnosis in almost all cases (Kjellin, S6derstr6m and Cronqvist 1975; S6derstr6m, Kjetlin and Cronqvist 1975; S6derstr6m 1977). In traumatic head injuries, the diagnostic significance of combined examinations with CSF-SPE and CT has been reported in preliminary studies (Levander, Kjellin, S6derstr6m and Bergvall 1975; S6derstr6m et al. 1975; Kjellin and S6derstr6m 1976; Wahlgren, Kjellin and S6derstr6m 1977a; Bergvall, Kjellin, Levander, Svendsen and S6derstr6m 1978). The purpose of the present study was to evaluate, on an extended material, the diagnostic significance of combined examinations with CSF-SPE and CT in traumatic head injuries. The results of these two diagnostic methods were compared with clinical data including the findings by conventional neurological investigations, especially angiography, as well as operation. MATERIALAND METHODS Yearly, about 130 patients with traumatic head injuries are admitted to the emergency ward of the Department of Neurology, Karolinska Hospital, i.e. about 2 0 ~ of all hospitalized emergency patients (Wahlgren, Kjellin and S6derstr6m 1977b). CSF-SPE and CT were performed on 53 patients with traumatic head injuries. Ten cases were clinically classified as cerebral concussion. In these patients the unconsciousness was of less than 15 min duration and no patient demonstrated focal neurological signs. Fourteen patients were diagnosed as having cerebral contusion. In these subjects the unconsciousness had a longer duration than 1 hr and some patients had focal neurological signs. One intermediate group comprising 9 patients was clinically reported as serious concussion or contusion, the reason mostly being insufficient information regarding the time of disturbed consciousness. In 20 patients the diagnoses were confirmed by cerebral angiography and/or operation. The disorders found were intracerebral haematoma (3 cases), acute subdural haematoma (3), subacute and chronic subdural haematoma (12) and hygroma (2). CSF spectrophotometry was carried out as described by Kjellin (1969, 1971). The spectrophotometric findings were corrected for traumatic lumbar puncture bleeding (Hellstr6m and Kjellin 1971; Kjellin 1971) and for serum xanthochromia (Kjellin 1969, 1971). The CSF examination also included cell examination and determination of protein concentration. Different bleeding patterns found by CSF-SPE were described by KjeUin and S6derstr6m (1974). The patterns have been labelled with the following symbols: " H "
217
1
SUBARAEHNOID HAE MORRRAGE
NORMAL J
I
I
J
SUBARCHNO[D HAE MORRHAGE
( RUPTURED ANEURYSMA
( RUPTURED ANEURYSMA A F T E R 9 DAYS )
CERE BRAL CONTUSION
CERE BRAL IIAE MA'rOMA
FTER 2 DAYS )
- - × 0
, r
350
~
i
450
i
i
550
I
i
r
650 350
r
1
450
I
r
550
1
1~ 0
G
Fig. 1. CSF spectrophotometric patterns found in healthy subjects (curve 1) and in patients with intracranial bleeding (curves 2-5). An S1 bleeding pattern in a patient with subarachnoid haemorrhage is seen in curve 2. Curve 4 demonstrates an $2 bleeding pattern as seen in cerebral contusion. A cerebral haematoma with an H pattern is seen in curve 5. Wavelength in nm on the abscissa.
for haematoma pattern, "S" for subarachnoid haemorrhage pattern and the mixed patterns "H-kS" or " S + H " , the first symbol indicating the dominating factor. If the CSF examination is carried out during the first week, the S pattern can be further divided into S1 and $2 patterns (Kjellin and S6derstr6m 1974). The S1 pattern, as found in aneurysm bleeding, is characterized by a gradual transformation of oxyhaemoglobin, released from haemolyzed red blood cells, into bilirubin within 7-10 days if no re-bleeding occurs during that time. Considering the time after onset of the disorder, the $2 pattern looks older than the S1 pattern, i.e. the bilirubin peak is more pronounced as compared to the oxyhaemoglobin peak, which is often lacking, even during the first days of the disease. Both S 1 and $2 patterns can be combined with an H component. Figure 1 demonstrates different bleeding patterns in CSF-spectrophotometry. The findings by computer tomography are shown in Fig. 2. In cerebral contusion, the attenuation is often mixed, areas of increased attenuation merging with areas of decreased attenuation (Levander, Stattin and Svendsen 1975) and has been described as a "salt and pepper" appearance (Dublin et al. 1977). Fresh intra- and extracerebral haematomas have increased attenuation, while the density decreases with time, following a fairly uniform pattern during the first month. Furthermore, expansive haematomas will cause shift of the intracranial structures (Bergstr6m, Ericson, Levander, Svendsen and Larsson 1977; Dublin et al. 1977).
219
Fig. 2. Different findings by computer tomography. A: normal findings; B: cerebral contusion; C: intracerebral haematoma; D: subdural haematoma; E: subdural haematoma on the left side. The haematoma is isodense with the brain tissue, but it can be suspected from the ventricular compression. RESULTS In Table 1 the findings at CSF-SPE and CT have been related to clinical parameters and the results of angiography and operation. The diagnoses of cerebral concussion, serious concussion or contusion and cerebral contusion were established according to the clinical criteria reported previously. The haematomas and hygromas were confirmed by angiography and/or operation. In the cases classified as cerebral concussion all CTs were normal. CSF-SPE findings were normal in 6 of the 10 subjects and questionable in 3 cases, while a slight bleeding pattern was observed in one patient. Radiography of the skull showed fractures in two subjects of 10. Eight of the 14 patients with clinically diagnosed cerebral contusion had CT findings in accordance with this diagnosis (Levander et al. 1975; Dublin et al. 1977), while in 3 these were questionable and 3 showed normal results. Bleeding patterns were found by CSF-SPE in all these cases. Ten of the 14 patients had skull fractures. In the intermediate group "serious concussion or contusion", 6 of 9 patients had normal CT findings, one showed questionable signs, and 2 patients had CT changes in accordance with cerebral contusion. Eight patients showed a bleeding pattern on CSFSPE while 1 patient had normal findings. Skull fractures were found in 2 subjects.
220 TABLE 1 FINDINGS AT CSF-SPE AND CT RELATED TO CLINICAL PARAMETERS A N D THE RESULTS OF A N G I O G R A P H Y A N D OPERATION Diagnosis
Cases
Cerebral concussion Serious concussion or contusion Cerebral contusion lntracerebral haematoma Acute subdural haematoma Subacute/chronic subdural haematoma Hygroma
10 9 14 3 3 12 2
CSF-SPE
CT
pos.
Q
N
1 8 14 3 3 12 2
3
6 I
pos.
Q
N
2
I
6
8 3 3 8 2
3
3
10
4
Three patients with intracerebral haematoma and 3 patients with acute subdural haematoma all had pathologic CT and CSF-SPE findings as well as fractures at the skull X-ray examination. All 12 patients with subacute or chronic subdural haematomas demonstrated bleeding patterns on CSF-SPE. In 8 subjects CT showed areas of attenuation changes in accordance with a subdural haematoma while 4 of the patients had CT findings judged as questionable. In two cases with abnormal CSF-SPE and CT findings, a hygroma was found at the operation. Visible xanthochromia after centrifugation was present in 22 of the 43 cases (51%) with bleeding pattern on CSF-SPE. The visual analysis was stated as questionable in 11 cases and negative in 10. The corresponding proportions in the 26 patients with bleeding indicated by CT were 15 with visible xanthochromia (58%), 6 questionable and 5 normal, while CSF-SPE showed bleeding patterns in all cases. r~19
%
25-
S
S2
H+S
S+H
Fig. 3. Different bloc'cling patterns on CSF-SPE in cerebral concussion and contusion.
221 %
Cases1
I
5o~
n=4
n=4
25-
S
S+H
H+S
7
H
CS-S FPE
Fig. 4. Different bleeding patterns on CSF-SPE in subacute and subdural haematomas.
These results correspond very well to the results reported by Kjellin and S6derstr6m (1974). Different bleeding patterns in the cerebral concussion, intermediate and contusion groups are presented in Fig. 3. A bleeding pattern of the $2 type was found in 12 of the 14 cases with cerebral contusion. In one patient the CSF examination was performed after one week and accordingly the bleeding pattern could only be classified as of the S type. One patient had a combined $2 + H pattern. The patients included in the intermediate group who had positive CSF-SPE findings demonstrated an $2 pattern in 6 cases of 8, while one patient had an H + S pattern and one subject had an S q- H pattern. The slight bleeding found in one patient with concussion was of the $2 type. In the subacute and chronic subdural haematoma group, 10 of 12 patients had an H-factor, combined with an S pattern in 6 subjects. Two subjects had pure S patterns (Fig. 4). Two of 3 patients with intracerebral haemorrhage had combined S + H patterns, while one patient had an $2 pattern. Two of the 3 patients with acute subdural haematomas had bleeding patterns of the $2 type, while the third of these patients had an H ÷ S pattern. The two patients with hygroma both had combined S q- H patterns. DISCUSSION
The difficulties in distinguishing different traumatic head disorders by the usual clinical criteria is emphasized by this investigation. In the present material the diagnosis of cerebral concussion as opposed to contusion could not be established clinically in 9 cases, comprising about 30~o of the patients. A haemorrhage was suspected on clinical criteria in only 5 of the later verified 18 cases.
222 The macroscopic examination of the centrifuged CSF showed definitely visible xanthochromia in only half the patients with the clinical diagnosis of contusion and haemorrhage. It should also be stressed that the colour of the CSF gives no information about the '°patterns" of blood decomposition compounds which cause the xanthochromia. Furthermore, visual inspection cannot distinguish "bleeding xanthochromia" from xanthochromia due to other factors, e.g. barrier damage, hyperbilirubinemia and traumatic lumbar puncture bleeding. Spectrophotometry of the CSF demonstrated bleeding patterns in all patients with cerebral contusion, intracranial haemorrhage and hygroma (Table 1). The CT findings in the same groups showed signs in agreement with the diagnosis in about 70 ~ while 20 ~ showed questionable findings and the rest was normal. The spectrophotometric bleeding patterns of special importance in traumatic head injury are $2 and H, and the combination of these patterns. In cerebral contusion 86}/~ of the patients demonstrated an $2 pattern. Patients with spectrophotometric bleeding patterns in the cerebral concussion and intermediate groups all had $2 patterns except for two subjects with a mixed pattern. In the haematoma and hygroma groups an H factor was found in 75 ~ of the cases. This pattern was especially frequent in the subacute and chronic subdural haematoma group, where it was present in 10 of the 12 cases (83 ~/o). The $2 pattern differs from the S 1 pattern (as seen in aneurysm haemorrhage) by the predominance of bilirubin as opposed to haemoglobin, considering the time after onset of the disorder. In aneurysm bleeding the red blood cells haemolyze in the CSF and release oxyhaemoglobin. Oxyhaemoglobin is then transformed into bilirubin within the first 7-10 days if no re-bleeding occurs. Concerning the S2 pattern, it is reasonable to assume that the haemolyzation and most of the degradation of oxyhaemoglobin occur outside the CSF compartment, e.g. in the extra-cellular fluid of the brain. This concept would explain the spectrophotometric curve in the contusion cases, dominated by the bilirubin peak. Because of the disappearance of haemoglobin, the spectrophotometric bleeding patterns SI and S2 cannot be separated after about one week. After this time and when the time of onset is not known, the bleeding pattern can only be classified as "S". Repeated examinations were performed on one or two occasions in t0 patients with a $2 or S type of bleeding pattern (Fig. 5). The results from these examinations are consistent with that the bleeding pattern remains during the first two weeks. A successive transformation of the bleeding findings into a barrier damage (B) pattern Occurs.
The H pattern mainly consists of methaem-compounds (methaem-albumin and methaemoglobin derivatives). No H component was observed in the CSF during the first 3 days after onset of the disorder. These compounds occur in the haematoma during the decomposition of the blood clot. In the 18 patients with haematoma and the 2 patients with hygroma, a pure $2 or S pattern was present in 5 subjects (intracerebral haemorrhage 1, acute subdural haematoma 2, subacute and chronic subdural heamatoma 2). All these 5 patients were examined on the 4th day or earlier except for one patient who was punctured on the 13th day. In the other 15 patients lumbar puncture was performed later than the 4th day after the trauma.
223 Patient
No. ¢
lO~ 9-
¢
8"
II
7
•
6
¢
5
~.
4
¢
3
•
2
•
1
•
o o
tl o
¢
o o
~
o
¢
o o o
Days after
10
15
20
25
30
35
40 trauma
Fig. 5. Findings at repeated CSF-SPE in 10 cases with initial $2 (y) or S bleeding pattern. The results are related to time after trauma. • = $2 or S bleeding pattern; © = barrier damage pattern.
The most c o m m o n type of bleeding pattern, observed in 11 patients of 20 with h a e m a t o m a or hygroma, was the combination of an $2 or S pattern with an H factor, while 4 subjects had a pure H pattern. The last mentioned patients were examined rather late after onset (15, 16, 17 and 20 days, respectively). The observation that the $2 or S type of bleeding pattern generally disappeared after about two weeks may explain the pure H patterns found later. In two cases, however, combined S and H patterns were found as late as 47 and 60 days respectively, one conceivable explanation being a recent re-bleeding. Computer t o m o g r a p h y was normal in 19 cases (cerebral concussion 10, serious concussion or contusion 6, cerebral contusion 3). The results were questionable in 8 cases (serious concussion or contusion 1, cerebral contusion 3, subacute or chronic subdural h a e m a t o m a 4). Pathological changes were found in 26 patients (serious concussion or contusion 2, contusion 8, intracerebral haematoma 3, acute subdural h a e m a t o m a 3, subacute or chronic subdural h a e m a t o m a 8, hygroma 2). The patients with subacute and chronic subdural haematoma and with questionable CT scans were examined on the 7th, 8th, 1 lth and 20th day respectively after the trauma. According to Bergstr6m et al. (1977) there is a period af about 5 days between the 5th and the 20th day after the trauma when the initially increased attenuation of an intracranial h a e m a t o m a is decreased into isodensity as compared with brain tissue. The isodensity of the h a e m a t o m a is followed by a state of decreased attenuation. Dublin et al. (1977) consider the isodensity period to occur between the 1st and the 5th week. In many of the isodensity cases, however, the presence of a haematoma can be suspected from secondary effects, such as a shift of the anatomical structures of the brain and ventricular compression. The findings reported by the authors mentioned above may partly explain the questionable CT results in the 4 cases in our study. The discrepancies between the CSF-SPE and CT findings in several patients with cerebral concussion, serious concussion or contusion, and contusion, require a special comment. Ten patients in these 3 diagnostic groups had normal (Y) or questionable findings on both examinations and they had all been clinically classified as having
224 cerebral concussion in 9 cases and serious concussion in 1 case. Thirteen patients had normal (7) or questionable CT scans and pathologic CSF-SPE. Clinically, one was classified as cerebral concussion, 6 as serious concussion or contusion and 6 as cerebral contusion. The remaining l0 of the 23 subjects had pathologic CT as well as CSF-SPE findings. Two of these 10 patients belonged to the intermediate group, and 8 to the cerebral contusion group. These findings suggest that CSF spectrophotometry is more sensitive in detecting intracranial diffuse bleeding than computer tomography. It is probable that extra- or intracerebral haematomas of limited magnitude as welt as pathological changes near bone structures are, in some cases, not detected by computer tomography. In conclusion, quantitative CSF spectrophotometry and computer tomography are found to be complementary to each other in the examination of traumatic head injuries. Spectrophotometry is more sensitive in detecting intracranial diffuse bleeding than CT. However, it cannot localize the site of the lesion. Furthermore, the differentiation between contusion and h a e m a t o m a seems not to be possible during the first 3-4 days by CSF-SPE, corresponding to the time required for the H-factor to be detectable. Our investigation also indicates that an $2 type of bleeding pattern disappears after about two weeks. The H-factor, however, can remain for a long time. By CT it is possible to localize the lesion and to detect a haematoma during the first days. However, this method has been found less sensitive than CSF-SPE in diffuse bleeding as in subarachnoid haemorrhage and cerebral contusion, and also in cases with haematomas of limited size, especially situated near bone structures. Previous studies as mentioned above have shown that for about one week within the period from the 5th day up to possibly 5 weeks subdural haematomas have the same attenuation as the surrounding brain tissue and that the CT diagnosis in these cases may be difficult even when secondary effects such as a shift of anatomical structures are present. By combined examination with quantitative CSF spectrophotometry and computer tomography it is possible to differentiate between uncomplicated cerebral concussion and cerebral contusion. It also makes it possible to detect haematomas that might require surgical intervention. The relevance of abnormal findings at CSF-SPE and CT to late complications, such as postconcussion symptoms, epilepsy and hydrocephalus is being investigated in a prognostic study. The diagnostic significance of methods other than CSF-SPE, e.g. isoelectric focusing of the CSF, to detect an H factor during the first days after traumatic head injury, is also under investigation.
REFERENCES
Bergstr6m, M., C. Ericson, B. Levander, P. Svendsen and S. Larsson (1977) Variation with time of the attenuation values of intracranial haematomas. J. comp. ass. Tomogr., 1 : 57. Bergvall, U., K. G. Kjellin, B. Levander, P. Svendsen and C. E. S0derstr6m (1978) Computed tomography of the brain and spectrophotometry of the CSF in cerebral concussion and contusion, Acta radiol. Diagn., To be published. Dublin, A. B.,B. N. French and J. M. Renick (1977) Computed tomography in head trauma, Radiology, 122: 365-369.
225 Hellstr6m, B. and K. G. Kjellin (1971) The diagnostic value of spectrophotometry of the cerebrospinal fluid in the newborn period, Develop. Med. Child Neurol., 6: 789-797. Kjellin, K. G. (1969) The binding of xanthochromic compounds in the cerebrospinal fluid, J. neurol. Sci., 9 : 597-601. Kjellin, K. G. (1971) Bilirubin compounds in the cerebrospinal fluid, J. neurol. Sci., 13 : 161-173. Kjellin, K. G. and C. E. S6derstrOm (1973) Abstract in the lOth International Congress of Neurology, Barcelona, 8-15 Sept. 1973 (International Congress Series, No. 296), Excerpta Medica, Amsterdam, p. 167. Kjellin, K. G. and C. E. S6derstr6m (1974) Diagnostic significance of CSF spectrophotometry in cerebrovascular diseases, J. neurol. Sei., 23 : 359-369. Kjellin, K. G. and C. E. S6derstr6m (1975) Cerebral haemorrhages with atypical clinical patterns, J. neurol. Sci., 25 : 211-226. Kjellin, K. G, and C. E. S6derstr6m (1976) Abstract in the Proceedings of the 7th Congress of Neurology and Neurochemistry with International Participation, Bratislava, p. 162. Kjellin, K. G. and L. Steiner (1974) Spectrophotometry of cerebrospinal fluid in subacute and chronic subdural haematomas, J. Neurol. Neurosurg. Psychiat., 37 : 1121-1127. Kjellin, K. G., C. E. S6derstr6m and S. Cronqvist (1975) Cerebrospinal fluid spectrophotometry and computerized transverse axial tomography (EMl-scanning) in cerebrovascular diseases, Europ. Neurol., 13 : 315-331. Levander, B., S. Stattin and P. Svendsen (1975) Computer tomography of traumatic intra- and extracerebral lesions, Acta radiol. Diagn., Suppl. 346, p. 107. Levander, B., K. G. Kjellin, C. E. S6derstr6m and U. Bergvall (1975) Abstract in the Proceedings of the Annual General Meeting of the Swedish Society of Medical Sciences, Stockholm, p. 318. Paxton, R. and J. Ambrose (1974) The EMI scanner - - A brief review of the first 650 patients, Brit. J. Radiol., 47: 530-565. S6derstr6m, C. E. (1977) Diagnostic significance of CSF spectrophotometry and computer tomography in cerebrovascular disease, Stroke, 5:606-612. S6derstr6m, C. E. and K. G. Kjellin (1972) Abstract in the Proceedings of the Annual General Meeting of the Swedish Society of Medical Sciences, Stockholm, p. 301. S6derstr6m, C. E., K. G. Kjellin and S. Cronqvist (1975) Computer tomography compared with spectrophotometry of the cerebrospinal fluid in cerebrovascular diseases, Acta radiol. Diagn., Suppl. 346, pp. 130-142. Svendsen, P. (1976) Computer tomography of traumatic extracerebral lesions, Brit. J. Radiok, 49 : 1004. Wahlgren, N. G., K. G. Kjellin and C. E. S6derstr6m (1977a) Abstract in the llth Worm Congress of Neurology, Amsterdam, 11-16 Sept. 1977 (International Congress Series, No. 427), Excerpta Medica, Amsterdam, p. 335. Wahlgren, N. G., K. G. Kjellin and C. E. S6derstr6m (1977b) The work of the acute section of a neurology department, Liikartidningen, 74: 33-36.
S P E C T R O P H O T O M E T R Y OF T H E CSF (CSF-SPE) A N D T O M O G R A P H Y (CT) IN T R A U M A T I C H E A D INJURIES
215
COMPUTER
N. G. WAHLGREN*, K. G. KJELLIN and C. E. SODERSTROM Department of Neurology, Karolinska Hospital, S-104 O1 Stockholm 60 (Sweden)
(Received 12 January, 1978) (Accepted 3 March, 1978)
SUMMARY Combined examinations with quantitative CSF spectrophotometry (CSF-SPE) and computer tomography (CT) were performed on 53 patients with traumatic head injuries. In cerebral concussion the results were mainly normal in both examinations. In cerebral contusion bleeding patterns were found by CSF-SPE in all subjects, with a special bleeding pattern ($2 pattern) occurring in 86 70. CT showed findings described as typical for contusion in 8 of 14 examined patients, the remaining CT scans showing questionable or normal signs. In extra- and intracerebral haematomas, all patients had bleeding patterns on the CSF-SPE. A special bleeding component (H factor) was found in about 72 ~ . The H component was not observed during the first 3 to 4 days after the trauma. All but one patient examined later than the 4th day had an H component with or without an S-pattern. CT demonstrated a haematoma in 14 of 18 verified haematoma patients, while 4 subjects with subdural haematoma (e.g. one third of this patient group) had questionable CT findings. The combined examinations with CT and CSF-SPE, being complementary to each other, are of great value in the differential diagnosis of traumatic head injuries.
INTRODUCTION It is often difficult to distinguish between different traumatic head disorders, including cerebral concussion, cerebral contusion and extracerebral or intracerebral haematomas by usual clinical criteria. The high diagnostic significance of quantitative CSF spectrophotometry (CSF-SPE) in cerebrovascular disease has been established This investigation was supported by a grant from the Trygg-Hansa Research Fund for Bodily Injury in Personal Accident Insurance. * To whom correspondence should be sent.
216 by Kjellin and S6derstr6m in several papers since 1972 (see for instance Kjellin and S6derstr6m 1974) and in subacute and chronic subdural haematomas by Kjellin and Steiner (1974). The diagnostic possibilities of computer tomography (CT) in traumatic head injuries were postulated by Paxton and Ambrose (1974) and have been confirmed by several authors, e.g. Dublin, French and Renick (1977, incl. ref.). Combined examinations with CSF-SPE and CT in cerebrovascular disease make it possible to obtain a specific diagnosis in almost all cases (Kjellin, S6derstr6m and Cronqvist 1975; S6derstr6m, Kjetlin and Cronqvist 1975; S6derstr6m 1977). In traumatic head injuries, the diagnostic significance of combined examinations with CSF-SPE and CT has been reported in preliminary studies (Levander, Kjellin, S6derstr6m and Bergvall 1975; S6derstr6m et al. 1975; Kjellin and S6derstr6m 1976; Wahlgren, Kjellin and S6derstr6m 1977a; Bergvall, Kjellin, Levander, Svendsen and S6derstr6m 1978). The purpose of the present study was to evaluate, on an extended material, the diagnostic significance of combined examinations with CSF-SPE and CT in traumatic head injuries. The results of these two diagnostic methods were compared with clinical data including the findings by conventional neurological investigations, especially angiography, as well as operation. MATERIALAND METHODS Yearly, about 130 patients with traumatic head injuries are admitted to the emergency ward of the Department of Neurology, Karolinska Hospital, i.e. about 2 0 ~ of all hospitalized emergency patients (Wahlgren, Kjellin and S6derstr6m 1977b). CSF-SPE and CT were performed on 53 patients with traumatic head injuries. Ten cases were clinically classified as cerebral concussion. In these patients the unconsciousness was of less than 15 min duration and no patient demonstrated focal neurological signs. Fourteen patients were diagnosed as having cerebral contusion. In these subjects the unconsciousness had a longer duration than 1 hr and some patients had focal neurological signs. One intermediate group comprising 9 patients was clinically reported as serious concussion or contusion, the reason mostly being insufficient information regarding the time of disturbed consciousness. In 20 patients the diagnoses were confirmed by cerebral angiography and/or operation. The disorders found were intracerebral haematoma (3 cases), acute subdural haematoma (3), subacute and chronic subdural haematoma (12) and hygroma (2). CSF spectrophotometry was carried out as described by Kjellin (1969, 1971). The spectrophotometric findings were corrected for traumatic lumbar puncture bleeding (Hellstr6m and Kjellin 1971; Kjellin 1971) and for serum xanthochromia (Kjellin 1969, 1971). The CSF examination also included cell examination and determination of protein concentration. Different bleeding patterns found by CSF-SPE were described by KjeUin and S6derstr6m (1974). The patterns have been labelled with the following symbols: " H "
217
1
SUBARAEHNOID HAE MORRRAGE
NORMAL J
I
I
J
SUBARCHNO[D HAE MORRHAGE
( RUPTURED ANEURYSMA
( RUPTURED ANEURYSMA A F T E R 9 DAYS )
CERE BRAL CONTUSION
CERE BRAL IIAE MA'rOMA
FTER 2 DAYS )
- - × 0
, r
350
~
i
450
i
i
550
I
i
r
650 350
r
1
450
I
r
550
1
1~ 0
G
Fig. 1. CSF spectrophotometric patterns found in healthy subjects (curve 1) and in patients with intracranial bleeding (curves 2-5). An S1 bleeding pattern in a patient with subarachnoid haemorrhage is seen in curve 2. Curve 4 demonstrates an $2 bleeding pattern as seen in cerebral contusion. A cerebral haematoma with an H pattern is seen in curve 5. Wavelength in nm on the abscissa.
for haematoma pattern, "S" for subarachnoid haemorrhage pattern and the mixed patterns "H-kS" or " S + H " , the first symbol indicating the dominating factor. If the CSF examination is carried out during the first week, the S pattern can be further divided into S1 and $2 patterns (Kjellin and S6derstr6m 1974). The S1 pattern, as found in aneurysm bleeding, is characterized by a gradual transformation of oxyhaemoglobin, released from haemolyzed red blood cells, into bilirubin within 7-10 days if no re-bleeding occurs during that time. Considering the time after onset of the disorder, the $2 pattern looks older than the S1 pattern, i.e. the bilirubin peak is more pronounced as compared to the oxyhaemoglobin peak, which is often lacking, even during the first days of the disease. Both S 1 and $2 patterns can be combined with an H component. Figure 1 demonstrates different bleeding patterns in CSF-spectrophotometry. The findings by computer tomography are shown in Fig. 2. In cerebral contusion, the attenuation is often mixed, areas of increased attenuation merging with areas of decreased attenuation (Levander, Stattin and Svendsen 1975) and has been described as a "salt and pepper" appearance (Dublin et al. 1977). Fresh intra- and extracerebral haematomas have increased attenuation, while the density decreases with time, following a fairly uniform pattern during the first month. Furthermore, expansive haematomas will cause shift of the intracranial structures (Bergstr6m, Ericson, Levander, Svendsen and Larsson 1977; Dublin et al. 1977).
219
Fig. 2. Different findings by computer tomography. A: normal findings; B: cerebral contusion; C: intracerebral haematoma; D: subdural haematoma; E: subdural haematoma on the left side. The haematoma is isodense with the brain tissue, but it can be suspected from the ventricular compression. RESULTS In Table 1 the findings at CSF-SPE and CT have been related to clinical parameters and the results of angiography and operation. The diagnoses of cerebral concussion, serious concussion or contusion and cerebral contusion were established according to the clinical criteria reported previously. The haematomas and hygromas were confirmed by angiography and/or operation. In the cases classified as cerebral concussion all CTs were normal. CSF-SPE findings were normal in 6 of the 10 subjects and questionable in 3 cases, while a slight bleeding pattern was observed in one patient. Radiography of the skull showed fractures in two subjects of 10. Eight of the 14 patients with clinically diagnosed cerebral contusion had CT findings in accordance with this diagnosis (Levander et al. 1975; Dublin et al. 1977), while in 3 these were questionable and 3 showed normal results. Bleeding patterns were found by CSF-SPE in all these cases. Ten of the 14 patients had skull fractures. In the intermediate group "serious concussion or contusion", 6 of 9 patients had normal CT findings, one showed questionable signs, and 2 patients had CT changes in accordance with cerebral contusion. Eight patients showed a bleeding pattern on CSFSPE while 1 patient had normal findings. Skull fractures were found in 2 subjects.
220 TABLE 1 FINDINGS AT CSF-SPE AND CT RELATED TO CLINICAL PARAMETERS A N D THE RESULTS OF A N G I O G R A P H Y A N D OPERATION Diagnosis
Cases
Cerebral concussion Serious concussion or contusion Cerebral contusion lntracerebral haematoma Acute subdural haematoma Subacute/chronic subdural haematoma Hygroma
10 9 14 3 3 12 2
CSF-SPE
CT
pos.
Q
N
1 8 14 3 3 12 2
3
6 I
pos.
Q
N
2
I
6
8 3 3 8 2
3
3
10
4
Three patients with intracerebral haematoma and 3 patients with acute subdural haematoma all had pathologic CT and CSF-SPE findings as well as fractures at the skull X-ray examination. All 12 patients with subacute or chronic subdural haematomas demonstrated bleeding patterns on CSF-SPE. In 8 subjects CT showed areas of attenuation changes in accordance with a subdural haematoma while 4 of the patients had CT findings judged as questionable. In two cases with abnormal CSF-SPE and CT findings, a hygroma was found at the operation. Visible xanthochromia after centrifugation was present in 22 of the 43 cases (51%) with bleeding pattern on CSF-SPE. The visual analysis was stated as questionable in 11 cases and negative in 10. The corresponding proportions in the 26 patients with bleeding indicated by CT were 15 with visible xanthochromia (58%), 6 questionable and 5 normal, while CSF-SPE showed bleeding patterns in all cases. r~19
%
25-
S
S2
H+S
S+H
Fig. 3. Different bloc'cling patterns on CSF-SPE in cerebral concussion and contusion.
221 %
Cases1
I
5o~
n=4
n=4
25-
S
S+H
H+S
7
H
CS-S FPE
Fig. 4. Different bleeding patterns on CSF-SPE in subacute and subdural haematomas.
These results correspond very well to the results reported by Kjellin and S6derstr6m (1974). Different bleeding patterns in the cerebral concussion, intermediate and contusion groups are presented in Fig. 3. A bleeding pattern of the $2 type was found in 12 of the 14 cases with cerebral contusion. In one patient the CSF examination was performed after one week and accordingly the bleeding pattern could only be classified as of the S type. One patient had a combined $2 + H pattern. The patients included in the intermediate group who had positive CSF-SPE findings demonstrated an $2 pattern in 6 cases of 8, while one patient had an H + S pattern and one subject had an S q- H pattern. The slight bleeding found in one patient with concussion was of the $2 type. In the subacute and chronic subdural haematoma group, 10 of 12 patients had an H-factor, combined with an S pattern in 6 subjects. Two subjects had pure S patterns (Fig. 4). Two of 3 patients with intracerebral haemorrhage had combined S + H patterns, while one patient had an $2 pattern. Two of the 3 patients with acute subdural haematomas had bleeding patterns of the $2 type, while the third of these patients had an H ÷ S pattern. The two patients with hygroma both had combined S q- H patterns. DISCUSSION
The difficulties in distinguishing different traumatic head disorders by the usual clinical criteria is emphasized by this investigation. In the present material the diagnosis of cerebral concussion as opposed to contusion could not be established clinically in 9 cases, comprising about 30~o of the patients. A haemorrhage was suspected on clinical criteria in only 5 of the later verified 18 cases.
222 The macroscopic examination of the centrifuged CSF showed definitely visible xanthochromia in only half the patients with the clinical diagnosis of contusion and haemorrhage. It should also be stressed that the colour of the CSF gives no information about the '°patterns" of blood decomposition compounds which cause the xanthochromia. Furthermore, visual inspection cannot distinguish "bleeding xanthochromia" from xanthochromia due to other factors, e.g. barrier damage, hyperbilirubinemia and traumatic lumbar puncture bleeding. Spectrophotometry of the CSF demonstrated bleeding patterns in all patients with cerebral contusion, intracranial haemorrhage and hygroma (Table 1). The CT findings in the same groups showed signs in agreement with the diagnosis in about 70 ~ while 20 ~ showed questionable findings and the rest was normal. The spectrophotometric bleeding patterns of special importance in traumatic head injury are $2 and H, and the combination of these patterns. In cerebral contusion 86}/~ of the patients demonstrated an $2 pattern. Patients with spectrophotometric bleeding patterns in the cerebral concussion and intermediate groups all had $2 patterns except for two subjects with a mixed pattern. In the haematoma and hygroma groups an H factor was found in 75 ~ of the cases. This pattern was especially frequent in the subacute and chronic subdural haematoma group, where it was present in 10 of the 12 cases (83 ~/o). The $2 pattern differs from the S 1 pattern (as seen in aneurysm haemorrhage) by the predominance of bilirubin as opposed to haemoglobin, considering the time after onset of the disorder. In aneurysm bleeding the red blood cells haemolyze in the CSF and release oxyhaemoglobin. Oxyhaemoglobin is then transformed into bilirubin within the first 7-10 days if no re-bleeding occurs. Concerning the S2 pattern, it is reasonable to assume that the haemolyzation and most of the degradation of oxyhaemoglobin occur outside the CSF compartment, e.g. in the extra-cellular fluid of the brain. This concept would explain the spectrophotometric curve in the contusion cases, dominated by the bilirubin peak. Because of the disappearance of haemoglobin, the spectrophotometric bleeding patterns SI and S2 cannot be separated after about one week. After this time and when the time of onset is not known, the bleeding pattern can only be classified as "S". Repeated examinations were performed on one or two occasions in t0 patients with a $2 or S type of bleeding pattern (Fig. 5). The results from these examinations are consistent with that the bleeding pattern remains during the first two weeks. A successive transformation of the bleeding findings into a barrier damage (B) pattern Occurs.
The H pattern mainly consists of methaem-compounds (methaem-albumin and methaemoglobin derivatives). No H component was observed in the CSF during the first 3 days after onset of the disorder. These compounds occur in the haematoma during the decomposition of the blood clot. In the 18 patients with haematoma and the 2 patients with hygroma, a pure $2 or S pattern was present in 5 subjects (intracerebral haemorrhage 1, acute subdural haematoma 2, subacute and chronic subdural heamatoma 2). All these 5 patients were examined on the 4th day or earlier except for one patient who was punctured on the 13th day. In the other 15 patients lumbar puncture was performed later than the 4th day after the trauma.
223 Patient
No. ¢
lO~ 9-
¢
8"
II
7
•
6
¢
5
~.
4
¢
3
•
2
•
1
•
o o
tl o
¢
o o
~
o
¢
o o o
Days after
10
15
20
25
30
35
40 trauma
Fig. 5. Findings at repeated CSF-SPE in 10 cases with initial $2 (y) or S bleeding pattern. The results are related to time after trauma. • = $2 or S bleeding pattern; © = barrier damage pattern.
The most c o m m o n type of bleeding pattern, observed in 11 patients of 20 with h a e m a t o m a or hygroma, was the combination of an $2 or S pattern with an H factor, while 4 subjects had a pure H pattern. The last mentioned patients were examined rather late after onset (15, 16, 17 and 20 days, respectively). The observation that the $2 or S type of bleeding pattern generally disappeared after about two weeks may explain the pure H patterns found later. In two cases, however, combined S and H patterns were found as late as 47 and 60 days respectively, one conceivable explanation being a recent re-bleeding. Computer t o m o g r a p h y was normal in 19 cases (cerebral concussion 10, serious concussion or contusion 6, cerebral contusion 3). The results were questionable in 8 cases (serious concussion or contusion 1, cerebral contusion 3, subacute or chronic subdural h a e m a t o m a 4). Pathological changes were found in 26 patients (serious concussion or contusion 2, contusion 8, intracerebral haematoma 3, acute subdural h a e m a t o m a 3, subacute or chronic subdural h a e m a t o m a 8, hygroma 2). The patients with subacute and chronic subdural haematoma and with questionable CT scans were examined on the 7th, 8th, 1 lth and 20th day respectively after the trauma. According to Bergstr6m et al. (1977) there is a period af about 5 days between the 5th and the 20th day after the trauma when the initially increased attenuation of an intracranial h a e m a t o m a is decreased into isodensity as compared with brain tissue. The isodensity of the h a e m a t o m a is followed by a state of decreased attenuation. Dublin et al. (1977) consider the isodensity period to occur between the 1st and the 5th week. In many of the isodensity cases, however, the presence of a haematoma can be suspected from secondary effects, such as a shift of the anatomical structures of the brain and ventricular compression. The findings reported by the authors mentioned above may partly explain the questionable CT results in the 4 cases in our study. The discrepancies between the CSF-SPE and CT findings in several patients with cerebral concussion, serious concussion or contusion, and contusion, require a special comment. Ten patients in these 3 diagnostic groups had normal (Y) or questionable findings on both examinations and they had all been clinically classified as having
224 cerebral concussion in 9 cases and serious concussion in 1 case. Thirteen patients had normal (7) or questionable CT scans and pathologic CSF-SPE. Clinically, one was classified as cerebral concussion, 6 as serious concussion or contusion and 6 as cerebral contusion. The remaining l0 of the 23 subjects had pathologic CT as well as CSF-SPE findings. Two of these 10 patients belonged to the intermediate group, and 8 to the cerebral contusion group. These findings suggest that CSF spectrophotometry is more sensitive in detecting intracranial diffuse bleeding than computer tomography. It is probable that extra- or intracerebral haematomas of limited magnitude as welt as pathological changes near bone structures are, in some cases, not detected by computer tomography. In conclusion, quantitative CSF spectrophotometry and computer tomography are found to be complementary to each other in the examination of traumatic head injuries. Spectrophotometry is more sensitive in detecting intracranial diffuse bleeding than CT. However, it cannot localize the site of the lesion. Furthermore, the differentiation between contusion and h a e m a t o m a seems not to be possible during the first 3-4 days by CSF-SPE, corresponding to the time required for the H-factor to be detectable. Our investigation also indicates that an $2 type of bleeding pattern disappears after about two weeks. The H-factor, however, can remain for a long time. By CT it is possible to localize the lesion and to detect a haematoma during the first days. However, this method has been found less sensitive than CSF-SPE in diffuse bleeding as in subarachnoid haemorrhage and cerebral contusion, and also in cases with haematomas of limited size, especially situated near bone structures. Previous studies as mentioned above have shown that for about one week within the period from the 5th day up to possibly 5 weeks subdural haematomas have the same attenuation as the surrounding brain tissue and that the CT diagnosis in these cases may be difficult even when secondary effects such as a shift of anatomical structures are present. By combined examination with quantitative CSF spectrophotometry and computer tomography it is possible to differentiate between uncomplicated cerebral concussion and cerebral contusion. It also makes it possible to detect haematomas that might require surgical intervention. The relevance of abnormal findings at CSF-SPE and CT to late complications, such as postconcussion symptoms, epilepsy and hydrocephalus is being investigated in a prognostic study. The diagnostic significance of methods other than CSF-SPE, e.g. isoelectric focusing of the CSF, to detect an H factor during the first days after traumatic head injury, is also under investigation.
REFERENCES
Bergstr6m, M., C. Ericson, B. Levander, P. Svendsen and S. Larsson (1977) Variation with time of the attenuation values of intracranial haematomas. J. comp. ass. Tomogr., 1 : 57. Bergvall, U., K. G. Kjellin, B. Levander, P. Svendsen and C. E. S0derstr6m (1978) Computed tomography of the brain and spectrophotometry of the CSF in cerebral concussion and contusion, Acta radiol. Diagn., To be published. Dublin, A. B.,B. N. French and J. M. Renick (1977) Computed tomography in head trauma, Radiology, 122: 365-369.
225 Hellstr6m, B. and K. G. Kjellin (1971) The diagnostic value of spectrophotometry of the cerebrospinal fluid in the newborn period, Develop. Med. Child Neurol., 6: 789-797. Kjellin, K. G. (1969) The binding of xanthochromic compounds in the cerebrospinal fluid, J. neurol. Sci., 9 : 597-601. Kjellin, K. G. (1971) Bilirubin compounds in the cerebrospinal fluid, J. neurol. Sci., 13 : 161-173. Kjellin, K. G. and C. E. S6derstrOm (1973) Abstract in the lOth International Congress of Neurology, Barcelona, 8-15 Sept. 1973 (International Congress Series, No. 296), Excerpta Medica, Amsterdam, p. 167. Kjellin, K. G. and C. E. S6derstr6m (1974) Diagnostic significance of CSF spectrophotometry in cerebrovascular diseases, J. neurol. Sei., 23 : 359-369. Kjellin, K. G. and C. E. S6derstr6m (1975) Cerebral haemorrhages with atypical clinical patterns, J. neurol. Sci., 25 : 211-226. Kjellin, K. G, and C. E. S6derstr6m (1976) Abstract in the Proceedings of the 7th Congress of Neurology and Neurochemistry with International Participation, Bratislava, p. 162. Kjellin, K. G. and L. Steiner (1974) Spectrophotometry of cerebrospinal fluid in subacute and chronic subdural haematomas, J. Neurol. Neurosurg. Psychiat., 37 : 1121-1127. Kjellin, K. G., C. E. S6derstr6m and S. Cronqvist (1975) Cerebrospinal fluid spectrophotometry and computerized transverse axial tomography (EMl-scanning) in cerebrovascular diseases, Europ. Neurol., 13 : 315-331. Levander, B., S. Stattin and P. Svendsen (1975) Computer tomography of traumatic intra- and extracerebral lesions, Acta radiol. Diagn., Suppl. 346, p. 107. Levander, B., K. G. Kjellin, C. E. S6derstr6m and U. Bergvall (1975) Abstract in the Proceedings of the Annual General Meeting of the Swedish Society of Medical Sciences, Stockholm, p. 318. Paxton, R. and J. Ambrose (1974) The EMI scanner - - A brief review of the first 650 patients, Brit. J. Radiol., 47: 530-565. S6derstr6m, C. E. (1977) Diagnostic significance of CSF spectrophotometry and computer tomography in cerebrovascular disease, Stroke, 5:606-612. S6derstr6m, C. E. and K. G. Kjellin (1972) Abstract in the Proceedings of the Annual General Meeting of the Swedish Society of Medical Sciences, Stockholm, p. 301. S6derstr6m, C. E., K. G. Kjellin and S. Cronqvist (1975) Computer tomography compared with spectrophotometry of the cerebrospinal fluid in cerebrovascular diseases, Acta radiol. Diagn., Suppl. 346, pp. 130-142. Svendsen, P. (1976) Computer tomography of traumatic extracerebral lesions, Brit. J. Radiok, 49 : 1004. Wahlgren, N. G., K. G. Kjellin and C. E. S6derstr6m (1977a) Abstract in the llth Worm Congress of Neurology, Amsterdam, 11-16 Sept. 1977 (International Congress Series, No. 427), Excerpta Medica, Amsterdam, p. 335. Wahlgren, N. G., K. G. Kjellin and C. E. S6derstr6m (1977b) The work of the acute section of a neurology department, Liikartidningen, 74: 33-36.
