ClinicalRadiology(1990) 42, 85-87
CT Scanning and Surgical Treatment of 1551 Head Injured patients Admitted to a Regional Neurosurgical Unit p. MACPHERSON*, B. JENNETT'~ and E. ANDERSON'~
Departments of *Neuroradiology and tNeurosurgery, Institute of Neurological Sciences, Glasgow Computed tomographic (CT) scans of 1551 recently head injured patients transferred to a regional neurosurgical unit (NSU) were reviewed. Some 90% of first scans were done outside normal working hours. More than a third of first scans were normal including a fifth of the patients who were in deep coma. Haematoma was found in 50%, contusion 28%, shearing injuries 13% and general swelling in 9%. In 22% the first scan led to urgent surgical evacuation of an intracranial haematoma. More than one scan was done in 41% of patients and more than two in 10%, making 2608 scans in all. Repeat scans were more often done when the first scan was abnormal. In only five of 554 patients (1%) whose first scan had been normal were contusions or haematomas seen on subsequent scans, and in none of these was surgery required. Of 997 patients whose first scan had been abnormal a new lesion (contusion, haematoma and/or infarction) was seen on a subsequent scan in 103 cases (10%). Surgery was required (for the first time), in 57 patients whose abnormal first scans had not indicated the necessity for surgery at that time. The implications of these and other findings for the scanning of recently head injured patients in general hospitals, as scanners become more widely available, are discussed in our accompanying paper on p. 88. Macpherson, P., Jennett, B. & Anderson, E. (1990). Clinical
Radiology 42, 85-87. CT Scanning and Surgical Treatment of 1551 Head Injured Patients Admitted to a Regional Neurosurgical Unit
One of the most useful applications of CT is its contribution to the management of recently head injured patients. The most obvious benefit is for patients suspected of developing an acute intracranial haematoma, whilst the practical value of discovering other lesions is less certain. In the UK, experience of such cases has hitherto been largely confined to regional NSUs, and most reports have been about specific types of clinical case or CT abnormalities. This report is of a series of 1551 patients transferred to a regional NSU over a 5-year period when there were no other facilities in the region for scanning acute head injuries. During the latter period of the study, scanners were in operation in three other hospitals in the West of Scotland, but none of these was used to investigate head injuries in the acute stage. The data from this study are relevant to decision-making in the European situation, where specialist surgeons, radiologists and equipment are each more limited than in the USA, from which most reports on CT head scanning for trauma have come.
FACILITIES AND P O L I C I E S IN T H E WEST OF SCOTLAND The Institute of Neurological Sciences in Glasgow provides a regional service for c.2.7 million population. It receives head injuries only by secondary transfer from 21 general hospitals in its catchment area, and is thus similar to most British and European centres. The Institute is sited in the Southern General Hospital where transfers from the primary surgical wards to neurosurgery are governed by the same criteria as for other hospitals. CT Correspondenceto: Dr P. Macpherson,Department of Neuroradiology,Institute of NeurologicalSciences,Glasgow G51 4TF.
scanning began in the Institute late in 1973 but 4 years later there was no obvious improvement in outcome in patients with head injury. When a second scanner was installed in 1978, the criteria for transfer were altered to ensure that more patients were transferred for scanning and that they reached the NSU sooner after injury than previously. This policy change brought about improved results for patients operated on for acute intracranial haematoma (Teasdale et al., 1982). The number of head injured patients transferred each year to the Institute doubled after the policy changed and it has remained at this new level; this represents about 8% of patients admitted after head injury to general hospitals in the catchment area (Bryden and Jennett, 1983). This compares with an average transfer rate at that time of 5% for England and Wales, where in some places only 1% were transferred (Jennett et al., 1979). DATA ON SCANNED P A T I E N T S The Department of Neurosurgery has been collecting data about head injured patients on computer since 1968. An additional form was designed on which detailed data about CT appearances were recorded by one author (PM). The original purpose was to discover the anatomical relationship of intraeranial lesions to the site of skull fracture, and this investigation was, therefore, limited to the 58 % of patients scanned for whom skull films from the referring hospital were available in the Institute. In order to ensure that this limitation had not produced a biased sample, the clinical features and scan findings of patients for whom plain films were available were compared with those for whom they were not, during the fifth year of the study. No significant differences were found between the two groups. This survey therefore included all 1383 cases
CLINICAL RADIOLOGY
86
for whom the data had been collected over five years, plus 168 additional cases from the final year - a total of 1551.
Timing of Scans and Reporting In c o m m o n with most regional NSUs, the Glasgow Institute offers scanning 24 hours a day and 7 days a week for cases that justify a scan out of normal hours. O f 1551 initial head injury scans in this series, 90% were done outside normal working hours. The initial interpretation of these urgent scans was normally by the duty neurosurgeon who had to decide whether there was a significant lesion requiring emergency intervention, the neuroradiologist's routine report being written later. For the purposes of this study each scan was reviewed again by the neuroradiologist author. Because these cases were all secondarily referred from other hospitals, some at a considerable distance from the Institute, relatively few had their first scan less than 6 hours after injury. In the first 3 years of the study, 67% of first scans were within 24 hours, compared with 76% for the last 2 years. N u m b e r o f S c a n s ( T a b l e 1)
More than one scan was done in 41% of all patients. Of those whose initial scan had been abnormal, at least one repeat scan was done in 53%, compared with only 19% of those whose first scan was normal. Only two scans were done in 25%, three in 10%, and between four and 13 scans in the remaining 6%. As a consequence there were 2608 scans in 1551 patients. Repeat scans were often on patients who had been operated on for an intracranial haematoma; 30% of those having a second scan and 46% of those having three or more had had surgery. Of patients who had a repeat scan without having had an operation for haematoma, 52% had been deeply unconscious on admission to N S U (Glasgow Coma Score* < 7); 30% of those who had three or more scans but who had not had surgery had been in deep coma on admission. N o patient whose first scan was normal developed a haematoma that required operation. Of 662 patients with an abnormal first scan but who did not have immediate surgery, 57 had an operation following a subsequent scan,
Table 1 - Scans in 1551 patients - operations for haematoma
either because the original lesion had progressed, or a new lesion had developed (Table 1). S e c o n d scans in 337 patients led to 41 having an operation; and of 127 patients having a third scan, 16 then had an operation for the first time. There were 27 repeat operations among the 190 patients who had one or more post-operative scans (having had surgery immediately after their first scan).
R a d i o l o g i c a l F i n d i n g s ( T a b l e 2)
Although this was a series of patients secondarily referred to a regional NSU, 36% had a normal first scan. An abnormal scan was more often found in adults than in children, and in those with lower Glasgow C o m a Scores (Table 2). But almost a fifth of those with scores of 7 or less (i.e. in deep coma) had a normal scan. The commonest lesion reported was a haematorna; this was found in 50% of patients, but only half of these required operation. Included as haematomas detected radiologically were small, deep collections of blood such as those in the basal ganglia, and subcortical gliding contusions. These accounted for a proportion of haematomas not operated on. Most of the rest were subcortical collections of blood that were not large enough, or not causing enough shift, to justify surgery, and/0r the patient's clinical condition was static and not giving rise to concern. Of the operated haematomas as a whole, more than twice as many were intradural (subdural or intracerebral or both) as were extradural; but in children and conscious adults extradural clots were as common as intradural. Cortical contusions were found in 28 %. A contusion is a bruise of the brain which appears on CT as a superficial lesion which is initially heterogeneous; in 5 to 7 days the low attenuation part becomes predominant due to increasing oedema and resolution of the haemorrhagic component. Although these CT appearances are considered by some to be diagnostic of a contusion (Zimmerman et al., 1977), there are occasions when the differentiation from haematoma is a matter of opinion. Moreover a haematoma may develop in a region of contusion. In the present investigation, superficial lesions of mixed high and low attenuation were coded as contusion, as was the occasional patch of purely low attenuation. Where the increased-attenuation component predominated, the lesion was classified as a haematoma. So-called subcortical gliding contusions were considered as haematomas from an imaging viewpoint.
Total
Firstscan 997abnormal Operation 3i5 (First
554Normal 1551 Patients Table 2 - Frequency of various lesions (as % of different patient groups)
Secondscan190 337 103 Operation ~Second) ~(First) Thirdscan 61 187 Operation ~Second) ~First) Furtherscans55
38
0
0
630 (at least2)
248 (at least 3)
93 (4 or more)
All Age (years) GCS* patients 15 3 7 8-12 13-15
n 1551 335' 1216 Abnormal scan 64% 41% 71% Radiological haematoma 50% 27 57 Operated haematoma 26% 13 30 (any site) Contusions 28% 8 34 Blood in ventricles 5% 4 5
382 81% 67 41 37 10
515 654 65% 52% 48 40 27 15 30 8
20 0
* Glasgow Coma Score on admission to NSU (3 7 = severely brain damaged).
CT FOR HEAD INJURIES IN AN NSU Table 3 - New lesions seen on subsequent scans in 103 patients whose first scans were abnormal
Contusions Extradural H Subdural H intracerebral H Basal Ganglia H Subcortical H Cerebellar Peduncle H intraventricular H Infarct
New lesion alone
Associated other new lesion
Total
21 5 16 14 3 1 1 7 17
13 1 6 8 4 2 0 3 5
34 6 22 22 7 3 1 10 22
H ~ Haematoma.
Evidence o f shearing injuries (diffuse axonal injury [DAI]), in the absence o f either extradural or intradural surgically significant h a e m a t o m a , was detected in 13 % o f all cases. Such evidence comprised one or m o r e o f the following: gliding subcortical contusions, small haematomas in corpus callosum a n d / o r in the rostral brain stem (cerebellar peduncles). All these lesions were m o r e common in adults than in children, and they were progressively more often f o u n d in patients whose consciousness was more impaired (Table 2). General brain swelling was recorded when the third ventricle was absent in patients who had < 5 m m midline displacement o f the brain. This was f o u n d in 9%, m a n y o f whom also had obliteration of the basal subarachnoid cisterns. Blood in the ventricles was visualized in 5%.
87
DISCUSSION I n a highly selected g r o u p o f patients suspected o f having serious or complicated head injury, one third had a normal scan. Over 40% o f patients h a d m o r e than one scan and some had surgery only after the second or third scan. There is, therefore, need for caution before accepting the value o f a single scan in excluding the need for neurosurgical intervention. In addition to the demonstration o f surgically remediable conditions, C T can contribute to the assessment o f early and late prognosis. F o r example, compression o f the third ventricle in the absence o f a localized mass lesion has been shown to be associated with raised intracranial pressure and p o o r prognosis (Teasdale et al., 1984; C o l q u h o u n and Burrows, 1989). Likewise, the presence o f clear evidence o f severe diffuse axonal injury in a patient w h o remains in deep c o m a for several days after injury makes death or vegetative survival the likely outcome. These radiological features, in combination with certain clinical signs and the patient's age, might lead to a decision to limit treatment (Barlow and Teasdale, 1986; Jennett, 1987). The results o f this survey f o r m the basis for a further paper (p. 88) in which consideration is given to the potential benefits and practical implications o f C T becoming m o r e widely available for the investigation o f patients w h o have suffered recent head injury. Acknowledgement. Elaine Anderson was supported by a grant from the Chief Scientist Office, SHI-ID, as part of the Scottish Head Injury Management Study.
REFERENCES New L e s i o n s A f t e r F i r s t S c a n ( T a b l e 3)
Of 554 patients whose first scan had been considered to be unequivocally normal, a subsequent examination within 'the acute period' showed abnormalities in only eight (1%). Contusions were seen in three patients on scans taken 1, 2 and 4 days after the initial scan. Delayed acute subdural h a e m a t o m a was noted on two occasions, at 2 and 3 days; neither o f these h a e m a t o m a s required surgery. Three patients developed cerebral infarction later on the first day, and 2 and 5 days later. N o n e was related to neck t r a u m a and these patients were regarded as having had a cerebrovascular accident subsequent to their head injury. In 997 patients whose initial C T was abnormal, a subsequent scan in 'the acute period' showed the development o f a new lesion in 103 (10%). The m o s t frequent lesions were contusions, subdural h a e m a t o m a s and infarctions (Table 3). O f the 22 intracerebral h a e m a t o m a s Which developed in a part o f the brain that had been normal on the first scan, nine had surgical evacuation.
Barlow, P & Teasdale, G (1986). Prediction of outcome and the management of severe head injuries: the attitudes of neurosurgeons. Neurosurgery, 19, 989 991. Bryden, JS & Jennett, B (1983). Neurosurgical resources and transfer policies for head injuries. British Medical Journal, 286, 1791 1793. Colquhoun, IR & Burrows, EH (1989). The prognostic significance of the third ventricle and basal cisterns in severe closed head injury. Clinical Radiology, 40, 13 16. Jennett, B (1987). Decisions to limit treatment. Lancet, ii, 787-789. Jennett, B & Macpherson, P (1990). Implications of scanning recently head injured patients in general hospitals. ClinicalRadiology, 42, 88 90. Jennett, B, Murray, A, Carlin, J, McKean, M, MacMillan, R & Strang, I (1979). Head injuries in three SCottish neurosurgical units. Scottish Head Injury Management Study. British Medical Journal, 2, 955958. Teasdale, E, Cardoso, E, Galbraith, S & Teasdale, G (1984). CT scan in severe diffuse head injury: physiological and clinical correlations. Journal of Neurology, Neurosurgery and Psychiatry, 47, 600-603. Teasdale, G, Galbraith, S, Murray, L, Ward, P, Gentleman, D & McKean, M (1982). Management of traumatic intracranial haematomas. British Medical Journal, 285, 1695-1697. Zimmerman, RA, Bilaniuk, LT, Dolinskas, C, Gennarelli, T, Bruce, D & Uzzell, B (1977). Computed tomography of acute intracerebral haemorrhagic contusion. ComputedAxial Tomography, 1, 271-280.
CT Scanning and Surgical Treatment of 1551 Head Injured patients Admitted to a Regional Neurosurgical Unit p. MACPHERSON*, B. JENNETT'~ and E. ANDERSON'~
Departments of *Neuroradiology and tNeurosurgery, Institute of Neurological Sciences, Glasgow Computed tomographic (CT) scans of 1551 recently head injured patients transferred to a regional neurosurgical unit (NSU) were reviewed. Some 90% of first scans were done outside normal working hours. More than a third of first scans were normal including a fifth of the patients who were in deep coma. Haematoma was found in 50%, contusion 28%, shearing injuries 13% and general swelling in 9%. In 22% the first scan led to urgent surgical evacuation of an intracranial haematoma. More than one scan was done in 41% of patients and more than two in 10%, making 2608 scans in all. Repeat scans were more often done when the first scan was abnormal. In only five of 554 patients (1%) whose first scan had been normal were contusions or haematomas seen on subsequent scans, and in none of these was surgery required. Of 997 patients whose first scan had been abnormal a new lesion (contusion, haematoma and/or infarction) was seen on a subsequent scan in 103 cases (10%). Surgery was required (for the first time), in 57 patients whose abnormal first scans had not indicated the necessity for surgery at that time. The implications of these and other findings for the scanning of recently head injured patients in general hospitals, as scanners become more widely available, are discussed in our accompanying paper on p. 88. Macpherson, P., Jennett, B. & Anderson, E. (1990). Clinical
Radiology 42, 85-87. CT Scanning and Surgical Treatment of 1551 Head Injured Patients Admitted to a Regional Neurosurgical Unit
One of the most useful applications of CT is its contribution to the management of recently head injured patients. The most obvious benefit is for patients suspected of developing an acute intracranial haematoma, whilst the practical value of discovering other lesions is less certain. In the UK, experience of such cases has hitherto been largely confined to regional NSUs, and most reports have been about specific types of clinical case or CT abnormalities. This report is of a series of 1551 patients transferred to a regional NSU over a 5-year period when there were no other facilities in the region for scanning acute head injuries. During the latter period of the study, scanners were in operation in three other hospitals in the West of Scotland, but none of these was used to investigate head injuries in the acute stage. The data from this study are relevant to decision-making in the European situation, where specialist surgeons, radiologists and equipment are each more limited than in the USA, from which most reports on CT head scanning for trauma have come.
FACILITIES AND P O L I C I E S IN T H E WEST OF SCOTLAND The Institute of Neurological Sciences in Glasgow provides a regional service for c.2.7 million population. It receives head injuries only by secondary transfer from 21 general hospitals in its catchment area, and is thus similar to most British and European centres. The Institute is sited in the Southern General Hospital where transfers from the primary surgical wards to neurosurgery are governed by the same criteria as for other hospitals. CT Correspondenceto: Dr P. Macpherson,Department of Neuroradiology,Institute of NeurologicalSciences,Glasgow G51 4TF.
scanning began in the Institute late in 1973 but 4 years later there was no obvious improvement in outcome in patients with head injury. When a second scanner was installed in 1978, the criteria for transfer were altered to ensure that more patients were transferred for scanning and that they reached the NSU sooner after injury than previously. This policy change brought about improved results for patients operated on for acute intracranial haematoma (Teasdale et al., 1982). The number of head injured patients transferred each year to the Institute doubled after the policy changed and it has remained at this new level; this represents about 8% of patients admitted after head injury to general hospitals in the catchment area (Bryden and Jennett, 1983). This compares with an average transfer rate at that time of 5% for England and Wales, where in some places only 1% were transferred (Jennett et al., 1979). DATA ON SCANNED P A T I E N T S The Department of Neurosurgery has been collecting data about head injured patients on computer since 1968. An additional form was designed on which detailed data about CT appearances were recorded by one author (PM). The original purpose was to discover the anatomical relationship of intraeranial lesions to the site of skull fracture, and this investigation was, therefore, limited to the 58 % of patients scanned for whom skull films from the referring hospital were available in the Institute. In order to ensure that this limitation had not produced a biased sample, the clinical features and scan findings of patients for whom plain films were available were compared with those for whom they were not, during the fifth year of the study. No significant differences were found between the two groups. This survey therefore included all 1383 cases
CLINICAL RADIOLOGY
86
for whom the data had been collected over five years, plus 168 additional cases from the final year - a total of 1551.
Timing of Scans and Reporting In c o m m o n with most regional NSUs, the Glasgow Institute offers scanning 24 hours a day and 7 days a week for cases that justify a scan out of normal hours. O f 1551 initial head injury scans in this series, 90% were done outside normal working hours. The initial interpretation of these urgent scans was normally by the duty neurosurgeon who had to decide whether there was a significant lesion requiring emergency intervention, the neuroradiologist's routine report being written later. For the purposes of this study each scan was reviewed again by the neuroradiologist author. Because these cases were all secondarily referred from other hospitals, some at a considerable distance from the Institute, relatively few had their first scan less than 6 hours after injury. In the first 3 years of the study, 67% of first scans were within 24 hours, compared with 76% for the last 2 years. N u m b e r o f S c a n s ( T a b l e 1)
More than one scan was done in 41% of all patients. Of those whose initial scan had been abnormal, at least one repeat scan was done in 53%, compared with only 19% of those whose first scan was normal. Only two scans were done in 25%, three in 10%, and between four and 13 scans in the remaining 6%. As a consequence there were 2608 scans in 1551 patients. Repeat scans were often on patients who had been operated on for an intracranial haematoma; 30% of those having a second scan and 46% of those having three or more had had surgery. Of patients who had a repeat scan without having had an operation for haematoma, 52% had been deeply unconscious on admission to N S U (Glasgow Coma Score* < 7); 30% of those who had three or more scans but who had not had surgery had been in deep coma on admission. N o patient whose first scan was normal developed a haematoma that required operation. Of 662 patients with an abnormal first scan but who did not have immediate surgery, 57 had an operation following a subsequent scan,
Table 1 - Scans in 1551 patients - operations for haematoma
either because the original lesion had progressed, or a new lesion had developed (Table 1). S e c o n d scans in 337 patients led to 41 having an operation; and of 127 patients having a third scan, 16 then had an operation for the first time. There were 27 repeat operations among the 190 patients who had one or more post-operative scans (having had surgery immediately after their first scan).
R a d i o l o g i c a l F i n d i n g s ( T a b l e 2)
Although this was a series of patients secondarily referred to a regional NSU, 36% had a normal first scan. An abnormal scan was more often found in adults than in children, and in those with lower Glasgow C o m a Scores (Table 2). But almost a fifth of those with scores of 7 or less (i.e. in deep coma) had a normal scan. The commonest lesion reported was a haematorna; this was found in 50% of patients, but only half of these required operation. Included as haematomas detected radiologically were small, deep collections of blood such as those in the basal ganglia, and subcortical gliding contusions. These accounted for a proportion of haematomas not operated on. Most of the rest were subcortical collections of blood that were not large enough, or not causing enough shift, to justify surgery, and/0r the patient's clinical condition was static and not giving rise to concern. Of the operated haematomas as a whole, more than twice as many were intradural (subdural or intracerebral or both) as were extradural; but in children and conscious adults extradural clots were as common as intradural. Cortical contusions were found in 28 %. A contusion is a bruise of the brain which appears on CT as a superficial lesion which is initially heterogeneous; in 5 to 7 days the low attenuation part becomes predominant due to increasing oedema and resolution of the haemorrhagic component. Although these CT appearances are considered by some to be diagnostic of a contusion (Zimmerman et al., 1977), there are occasions when the differentiation from haematoma is a matter of opinion. Moreover a haematoma may develop in a region of contusion. In the present investigation, superficial lesions of mixed high and low attenuation were coded as contusion, as was the occasional patch of purely low attenuation. Where the increased-attenuation component predominated, the lesion was classified as a haematoma. So-called subcortical gliding contusions were considered as haematomas from an imaging viewpoint.
Total
Firstscan 997abnormal Operation 3i5 (First
554Normal 1551 Patients Table 2 - Frequency of various lesions (as % of different patient groups)
Secondscan190 337 103 Operation ~Second) ~(First) Thirdscan 61 187 Operation ~Second) ~First) Furtherscans55
38
0
0
630 (at least2)
248 (at least 3)
93 (4 or more)
All Age (years) GCS* patients 15 3 7 8-12 13-15
n 1551 335' 1216 Abnormal scan 64% 41% 71% Radiological haematoma 50% 27 57 Operated haematoma 26% 13 30 (any site) Contusions 28% 8 34 Blood in ventricles 5% 4 5
382 81% 67 41 37 10
515 654 65% 52% 48 40 27 15 30 8
20 0
* Glasgow Coma Score on admission to NSU (3 7 = severely brain damaged).
CT FOR HEAD INJURIES IN AN NSU Table 3 - New lesions seen on subsequent scans in 103 patients whose first scans were abnormal
Contusions Extradural H Subdural H intracerebral H Basal Ganglia H Subcortical H Cerebellar Peduncle H intraventricular H Infarct
New lesion alone
Associated other new lesion
Total
21 5 16 14 3 1 1 7 17
13 1 6 8 4 2 0 3 5
34 6 22 22 7 3 1 10 22
H ~ Haematoma.
Evidence o f shearing injuries (diffuse axonal injury [DAI]), in the absence o f either extradural or intradural surgically significant h a e m a t o m a , was detected in 13 % o f all cases. Such evidence comprised one or m o r e o f the following: gliding subcortical contusions, small haematomas in corpus callosum a n d / o r in the rostral brain stem (cerebellar peduncles). All these lesions were m o r e common in adults than in children, and they were progressively more often f o u n d in patients whose consciousness was more impaired (Table 2). General brain swelling was recorded when the third ventricle was absent in patients who had < 5 m m midline displacement o f the brain. This was f o u n d in 9%, m a n y o f whom also had obliteration of the basal subarachnoid cisterns. Blood in the ventricles was visualized in 5%.
87
DISCUSSION I n a highly selected g r o u p o f patients suspected o f having serious or complicated head injury, one third had a normal scan. Over 40% o f patients h a d m o r e than one scan and some had surgery only after the second or third scan. There is, therefore, need for caution before accepting the value o f a single scan in excluding the need for neurosurgical intervention. In addition to the demonstration o f surgically remediable conditions, C T can contribute to the assessment o f early and late prognosis. F o r example, compression o f the third ventricle in the absence o f a localized mass lesion has been shown to be associated with raised intracranial pressure and p o o r prognosis (Teasdale et al., 1984; C o l q u h o u n and Burrows, 1989). Likewise, the presence o f clear evidence o f severe diffuse axonal injury in a patient w h o remains in deep c o m a for several days after injury makes death or vegetative survival the likely outcome. These radiological features, in combination with certain clinical signs and the patient's age, might lead to a decision to limit treatment (Barlow and Teasdale, 1986; Jennett, 1987). The results o f this survey f o r m the basis for a further paper (p. 88) in which consideration is given to the potential benefits and practical implications o f C T becoming m o r e widely available for the investigation o f patients w h o have suffered recent head injury. Acknowledgement. Elaine Anderson was supported by a grant from the Chief Scientist Office, SHI-ID, as part of the Scottish Head Injury Management Study.
REFERENCES New L e s i o n s A f t e r F i r s t S c a n ( T a b l e 3)
Of 554 patients whose first scan had been considered to be unequivocally normal, a subsequent examination within 'the acute period' showed abnormalities in only eight (1%). Contusions were seen in three patients on scans taken 1, 2 and 4 days after the initial scan. Delayed acute subdural h a e m a t o m a was noted on two occasions, at 2 and 3 days; neither o f these h a e m a t o m a s required surgery. Three patients developed cerebral infarction later on the first day, and 2 and 5 days later. N o n e was related to neck t r a u m a and these patients were regarded as having had a cerebrovascular accident subsequent to their head injury. In 997 patients whose initial C T was abnormal, a subsequent scan in 'the acute period' showed the development o f a new lesion in 103 (10%). The m o s t frequent lesions were contusions, subdural h a e m a t o m a s and infarctions (Table 3). O f the 22 intracerebral h a e m a t o m a s Which developed in a part o f the brain that had been normal on the first scan, nine had surgical evacuation.
Barlow, P & Teasdale, G (1986). Prediction of outcome and the management of severe head injuries: the attitudes of neurosurgeons. Neurosurgery, 19, 989 991. Bryden, JS & Jennett, B (1983). Neurosurgical resources and transfer policies for head injuries. British Medical Journal, 286, 1791 1793. Colquhoun, IR & Burrows, EH (1989). The prognostic significance of the third ventricle and basal cisterns in severe closed head injury. Clinical Radiology, 40, 13 16. Jennett, B (1987). Decisions to limit treatment. Lancet, ii, 787-789. Jennett, B & Macpherson, P (1990). Implications of scanning recently head injured patients in general hospitals. ClinicalRadiology, 42, 88 90. Jennett, B, Murray, A, Carlin, J, McKean, M, MacMillan, R & Strang, I (1979). Head injuries in three SCottish neurosurgical units. Scottish Head Injury Management Study. British Medical Journal, 2, 955958. Teasdale, E, Cardoso, E, Galbraith, S & Teasdale, G (1984). CT scan in severe diffuse head injury: physiological and clinical correlations. Journal of Neurology, Neurosurgery and Psychiatry, 47, 600-603. Teasdale, G, Galbraith, S, Murray, L, Ward, P, Gentleman, D & McKean, M (1982). Management of traumatic intracranial haematomas. British Medical Journal, 285, 1695-1697. Zimmerman, RA, Bilaniuk, LT, Dolinskas, C, Gennarelli, T, Bruce, D & Uzzell, B (1977). Computed tomography of acute intracerebral haemorrhagic contusion. ComputedAxial Tomography, 1, 271-280.
