Predicting the need for operation in the patient with an occult traumatic intracranial hematoma

Transcription

1 J Neurosurg 55:75-81, 1981 Predicting the need for operation in the patient with an occult traumatic intracranial hematoma SAM GALBRAITH, M.D., F.R.C.S., AND GRAHAM TEASDALE, M.R.C.P., F.R.C.S. Department of Neurosurgery, Institute of Neurological Sciences, Southern General Hospital, Glasgow, Scotland ~/ Computerized tomography scanning has shown that acute traumatic intracranial hematomas are more common than was previously realized, but whether all hematomas must be removed remains controversial. About half of this series of 26 patients who were not clinically deteriorating and who were initially managed without operation had to undergo hematoma removal because they subsequently deteriorated. Features present at the time of diagnosis (age, type and site of hematoma, presence of focal signs, level of responsiveness, and degree of midline shift) were not helpful in predicting that operation would be needed. The only discriminatory factor was the level of intracranial pressure (ICP). All the patients with ICP greater than 30 mm Hg deteriorated and required operation, but only one patient whose ICP was less than 20 mm Hg deteriorated. Half the patients with ICP between 20 and 30 mm Hg did not require an operation. Intracranial pressure monitoring can, therefore, be useful in managing patients with an occult intracranial hematoma. KEY WORDS 9 head injury 9 intracranial hematoma 9 intracraniai pressure monitoring 9 outcome W 'HEN a head-injured patient deteriorates, an intracranial hematoma may be present. In the past, the patient had to deteriorate before such a patient was investigated with either angiography or exploratory burr holes. Computerized tomography (CT) scanning has made investigation much easier and more informative, and as a result it is now being used in many head-injured patients, including those who are clinically stable and who would not previously have been investigated. When these patients are found to have an intracranial hematoma, it can be difficult to decide what to do: some recover without operation, and serial scans show absorption of the hematoma, but others go on to deteriorate, sometimes rapidly. Should the surgeon evacuate the clot forthwith or await events? 1'5 This difficult question arises in only a minority of patients with traumatic intracranial hematoma, but if those who are going to deteriorate could be identified soon after the hematoma has been detected they could be operated on immediately without incurring the risks of delay, and the remainder would be spared unnecessary operation. We became aware of this problem shortly after the introduction of CT scanning in Glasgow, and initiated studies to determine which patients required operation, and whether they could be recognized at an early stage from clinical or investigative features. Our initial experience suggested that up to half of the patients could be successfully managed without operation and that, although this group was not distinguished by characteristic clinical or radiological features, intracranial pressure (ICP) was higher in those who eventually required operation. 1~ These findings have now been substantiated by the more extensive study which forms the basis for this report. During the period covered by this study, attitudes concerning the management of such patients have varied, both as to the advisability of a non-operative expectant policy and also the need for ICP monitoring. The patients reported here, therefore, are only a minority of those in whom a trial of non-operative management was considered. Clinical Material Patients in this study had a traumatic intradural hematoma (subdural, intracerebral, or burst lobe) diagnosed by CT scanning within 14 days of injury, but J. Neurosurg. / Volume 55 / July,

2 S. Galbraith and G. Teasdale Type of Lesion TABLE 1 Type of lesion in 26 patients Right Hemisphere Left mixed 8 6 subdural 3 4 intracerebral 3 2 frontal 5 3 temporal 3 3 parietal 1 0 multiple sites 5 6 TABLE 2 Time from injury to diagnosis and start of intracranial pressure monitoring* Group Time (hrs) < > 72 operated non-operated * The difference between the operated and non-operated groups at 24 hours was p < had not shown deterioration in responsiveness or an increase in focal neurological signs since the time of injury. Patients with severe primary brain damage, in whom an unfavorable outcome could be anticipated from the start, were excluded. In each case, the clinician responsible for the patient believed that there was a reasonable expectation of recovery without operation. Patients were closely observed by experienced staffwho regularly recorded the Glasgow Coma Scale (GSC) score, 11 and repeatedly examined for focal neurological signs, such as weakness in the limbs or pupillary changes. Intracranial pressure was measured continuously by an indwelling ventricular catheter connected to a transducer on a bedside stand, and a continuous record was displayed on a chart recorder. In the initial phase of the study, intracranial elastance was assessed by the ventricular pressure response (VPR) to the injection or withdrawal of I ml of fluid. 7 All patients were allowed to breathe spontaneously, none required intubation, and steroids, anticonvulsants, and osmotics were not given. Outcome was assessed 6 months after injury using the Glasgow Outcome Scale. 3 Significance of the resuits was tested by a Wilcoxon Mann-Whitney test or chi-square analysis. There were 26 patients who had a trial of nonoperative management and whose ICP was satisfactorily measured. Seven patients had a purely subdural and five a purely intracerebral clot, and in 14 the hematoma consisted of both subdural and intracerebral blood, with contused and edematous brain (the classic burst lobe, Table 1). In seven patients, monitoring was started within 12 hours of injury, in four at 12 to 24 hours, in six on the 2nd or 3rd day, and in the remaining nine between 4 and 14 days (Table 2). Case 1 Illustrative Cases This 43-year-old woman fell while under the influence of alcohol and was in coma for 30 minutes. She came to the hospital 2 days later; she was disorientated, but would obey commands. There were no focal signs, and CT scan showed a thin right-sided subdural hematoma with a small shift of the midline (Fig. 1). The ICP during 24 hours of recording was persistently above 20 mm Hg, episodically above 30 mm Hg, but without any upward or downward trend (Fig. 2). The VPR was 4 mm Hg/ml. The patient's condition appeared to be improving and she was kept under observation. Ten days after admission, her responsiveness deteriorated and she developed a hemiparesis. A right frontotemporal craniotomy was performed, and a subdural hematoma, 1 cm thick, was removed. After operation, there was a slow improvement in responsiveness. The patient was still moderately disabled 6 months after injury, due to psychological and personality impairment. Case 2 This 25-year-old man was a pedestrian involved in a road traffic accident. On examination at the hospital within 1 hour of the trauma, he could open his eyes and localize to pain; his verbal response was incomprehensible. A CT scan showed a large left-sided intracerebral hematoma, with shift of the midline to the right (Fig. 3 left). His ICP was never above 20 mm Hg, and there was a progressive decrease over the 72 hours during which it was measured (Fig. 4). The hematoma was not removed and he improved spontaneously. A repeat CT scan at 6 months, by which time he had made a good recovery, showed resolution of the clot with slight dilatation of the left lateral ventricle (Fig. 3 right). Summary of Cases Clinical Course and Need for Operation Twelve patients (64%) eventually required operation; 11 because of clinical deterioration and one because of failure to improve combined with raised 76 J. Neurosurg. / Volume 55 / July, 1981

3 Management of traumatic intracranial hematoma FIG. 1. Case 1. Computerized tomography scan showing a small right-sided subdural hematoma with slight compression and shift of the ventricular system to the left. FIG. 3. Case 2. Left: Computerized tomography scan showing a left-sided intracerebral hematoma with shift of the midline to the right. The ventricles are dilated. Right: Follow-up scan. The hematoma has resolved, but there is dilatation of the ventricles, more marked on the left. FIG. 2. Case 1. Part of the continuous intracranial pressure (ICP) tracing. The ICP is consistently above 20 mm Hg but only episodically above 30 mm Hg. FIG. 4. Case 2. Part of the continuous intracranial pressure (ICP) tracing. The ICP is consistently below 20 mm Hg. ICP. In the patients who did not require operation, the level of consciousness improved, any focal neurological deficits disappeared within a few days, and elevated levels of ICP fell progressively over a similar period. Comparison of Features in Operated and Non-Operated Cases Age. The patients who required operation were on average slightly younger than those who did not, but the difference was not statistically significant (Fig. 5). Time from Injury to Diagnosis. Although as a group the patients who were operated on had been diagnosed sooner (p < 0.02), in two of them the diagnosis was made 72 hours or more after the injury. On the other hand, in three of the patients not operated on, the diagnosis was made within 12 hours (Table 2). Clinical and Investigative Features. Features such as the patients' clinical state, the CT scan appearances, and the level of ICP can alter during observation. In FIG. 5. Age and need for evacuation of traumatic intracranial hematoma. The difference between the groups is not significant. J. Neurosurg. / Volume 55 / July,

4 S. Galbraith and G. Teasdale FIG. 6. Responsiveness and need for evacuation of traumatic intracranial hematoma. Flo. 7. Midline shift and need for evacuation of traumatic intracranial hematoma. The difference between the groups is not significant. analyzing these features, we have considered only the findings at either the time of initial diagnosis of the hematoma (clinical state and CT appearances) or, in the case of ICP, after a defined interval, namely 6 hours after the scan. Level of Responsiveness. In the surgical group, the initial level of responsiveness, as reflected by the G CS, was lower than the non-operated group, but this did not usefully discriminate between them. Thus, there were patients in coma in both groups, and one of the patients eventually operated on was only mildly confused when first diagnosed (Fig. 6). Focal Signs. Four of those operated on had focal signs, such as a hemiparesis, but so did three who were managed without operation. Computerized Tomography Findings. On CT scanning, the hematomas were found to be equally distributed between the right and left sides, and were of TABLE 3 Type of hematoma in 26 patients Type of Operated Non-Operated Hematoma Group Group fight 6 8 left 6 6 mixed 7 7 subdural 3 4 intracerebral 2 3 frontal 4 4 temporal 2 4 parietal 1 0 multiple sites 5 6 similar pathological type and anatomical distribution (Table 3). There was no difference in the extent of midline shift in the two groups; some patients operated on had no shift and other patients were successfully managed without operation despite having large shifts (Fig. 7). There was no relationship between the ICP and midline shift (Fig. 8). Intracranial Pressure. The mean ICP 6 hours after the start of recording was mm Hg in patients who were operated on, as compared with 16 6 mm Hg in patients who did not require an operation (p < 0.001) (Fig. 9). Only one of 10 patients with an ICP persistently below 20 mm Hg required an operation, and this was several days later; all patients with an ICP sustained above 30 mm Hg required an operation. The hematoma had to be removed in half of the patients whose ICP was between 20 and 30 mm Hg (Table 4). The ventricular pressure response was measured in nine cases; in each it was under 4 mm Hg/ ml, and did not predict the need for operation. TABLE 4 Mean intracranial pressure (ICP) at 6 hours ICP (nun Hg) Operated Group Non-Operated Group < > J. Neurosurg. / Volume 55 / July, 1981

5 Management of traumatic intracranial hematoma FIG. 8. Relationship between intracranial pressure (ICP) and midline shift in traumatic intracranial hematoma. Outcome. Two of the patients who required an operation eventually died and another two remain severely disabled (Table 5). Neither patient died as a result of brain-stem compression; one died 3 weeks after injury from a pulmonary embolus, and the other died after remaining severely disabled for 6 months. One patient, aged 80 years, who was managed without operation also died later, but this was from an unrelated cause 4 months after diagnosis of the hematoma. A follow-up scan had shown complete resolution of the hematoma. Another patient from this group is severely disabled. There was no consistent relationship between the patients' clinical state at the time of diagnosis and outcome, however the clot was treated. Two of the patients who died or remained severely disabled after operation had only minor alterations of the level of responsiveness at the time of diagnosis. By contrast, two of the patients who did not undergo operation made a good recovery, despite having been in coma (Table 6). In both groups of patients, a bad outcome TABLE 5 Outcome at 6 months in 26 patients Good/ Severe Vegetative/ Group Moderate Disability Dead operated non-operated FIG. 9. Intracranial pressure (ICP) and need for evacuation of traumatic intracranial hematoma. The difference between the groups was significant at p < generally reflected the ICP having been greater than 20 mm Hg, but two of the patients who died, one from each group, both had initial ICP's of less than 20 mm Hg (Table 7). Discussion It is now two centuries since Percivall Pott first pointed out that an intracranial hematoma could be present in the absence of clinical deterioration, and that it could be difficult to decide what to do? When a hematoma shown by CT scanning is large, discrete, and accessible, it seems natural to remove it, but not to do so if the hematoma is small, indistinct, or inaccessible. Between these extremes, and perhaps also-depending on the patient's responsiveness, lie a TABLE 6 Level of responsiveness and outcome Operated Group Non-Operated Group Glasgow Coma Good/ Severe Vege- Good/ Severe Vegeta- Scale Score Moder- Dis- tative/ Moder- Dis- tive/ ate ability Dead ate ability Dead < l 0 J. Neurosurg. / Volume 55 / July,

6 S. Galbraith and G. Teasdale TABLE 7 lntracranial pressure (1CP) and outcome in 26 patients Operated Group Non-Operated Group ICP Good/ Severe Vegeta- Good/ Severe Vegeta- (mm Hg) Moder- Dis- rive/ Moder- Dis- rive/ ate ability Dead ate ability Dead < l > l group of patients about whom most neurosurgeons would be uncertain, although the size of this group and its characteristics will vary from surgeon to surgeon. We estimate that such a group represents about 10% of the head-injured patients who we deal with in our institute. At present, we operate on about 150 patients with traumatic intracranial hematoma from a total population of three million, of whom approximately 10,000 are admitted to Scottish hospitals each year with a head injury. 4 We found that, in the group we selected, a trial of non-operative management in patients with acute hematoma failed in half the cases because the patient eventually required an operation, usually because of clinical deterioration. Moreover, although all the patients were anticipated to recover with an expectant policy, two died and two remained severely disabled. Neither of the deaths was directly the consequence of compression, but it is nevertheless possible that evacuation of the clot immediately after diagnosis might have led to a more rapid recovery in one case, and avoided the pulmonary embolus in another. A similar argument might apply to the patients who remain severely disabled, neither of whom was in coma at the time of diagnosis. It would, therefore, be helpful to be able to predict which patients will eventually require operation, and remove the hematoma before it produces secondary brain damage. Although the patients who were operated on tended to be slightly younger and less responsive, neither of these features, nor the level of responsiveness, nor even the extent of midline shift were helpful predictors. The poor correlation between the extent of midline shift and the need for operation was surprising, especially as the magnitude of the ventricular pressure response, which has been held to be a predictor of the reserve available to compensate for an increase in the size of a mass lesion, correlates more closely with the extent of the midfine shift than with the level of ICP. 7 The level of ICP has seldom been measured preoperatively in patients with a traumatic intracranial hematoma. Papo, et al., ~ found that the ICP was high before operation in each of 13 patients with traumatic hematoma, and that it was usually lowered by removal of the clot. In three of four patients who did not require an operation, the ICP ranged from 25 to 30 mm Hg. In our patients, the ICP was higher in those who required an operation (30 _+ 9 mm Hg) than in those who did not ( mm Hg). All patients with an ICP greater than 30 mm Hg eventually required an operation, but of those with an ICP less than 20 mm Hg only one required an operation. Based on these findings, we suggest that the hematoma should be removed if the ICP is persistently above 30 mm Hg within the first few hours after the start of monitoring, whether or not there is clinical deterioration. If it is less than 20 mm Hg, ICP measuring can be stopped and monitoring continued by clinical observation. If the ICP lies between 20 and 30 mm Hg, ICP monitoring should be continued, and it is our experience that its trend is usually declared within 24 hours. Intracranial pressure monitoring has been used in head injury to predict outcome, to determine the efficacy of measures to control intracranial hypertension, and to assist in the diagnosis of complications. Although there are reports of its use in each of these circumstances, 6 its value is still debated. By contrast, the results of the present study show that it is undoubtedly of value in the management of patients with a traumatic intracranial hematoma who have not (as yet) deteriorated. Measurement of ICP gives information not available from clinical or radiological investigations, and its results provide reliable grounds for deciding whether or not to remove the hematoma. We are not proposing that a period of ICP monitoring is required before operation in every patient with a traumatic intracranial hematoma, rather that it is useful when the surgeon is in doubt about management. Early removal of the clot in patients who will otherwise eventually deteriorate ought to improve the outcome in this small but important group of headinjured patients. References 1. Becker DP, Miller JD, Ward JD, et al: The outcome from severe head injury with early diagnosis and intensive management. J Neurosurg 47: , Galbraith S, Teasdale G, Blaiklock C: Computerised tomography of acute traumatic intracranial haematoma: reliability of neurosurgeons' interpretations. Br Med J 2: , Jennett B, Bond M: Assessment of outcome after severe brain damage. A practical scale. Lancet 1: , Jennett B, Murray A, Carlin J, et al: Head injuries in Scottish hospitals. Lancet 2: , Kalbag RM: Head injury management. J Neurosurg 48: , 1978 (Letter) 80 J. Neurosurg. / Volume 55 / July, 1981

7 Management of traumatic intracranial hematoma 6. Miller JD, Becker DP, Ward JD, et al: Significance of intracranial hypertension in severe head injury. 3 Neurosurg 47: , Miller JD, Pickard JD: Intracranial volume/pressure studies in patients with head injury. Injury 5: , Papo I, Caruselli G, Scarpelli M: Time course in intracerebral traumatic hematomas, in Shulman K, Marmarou A, Miller JD, et al (eds): Intracranial Pressure IV. Berlin/Heidelberg/New York: Springer-Verlag, 1980, pp Pott P: Observations on the Nature and Consequences of Those Injuries to Which the Head is Liable from External Violence. London: Hitch C, Hawes L, 1768, pp Teasdale G, Galbraith S, Jennett B: Operate or observe? ICP and the management of the "silent" traumatic intracranial haematoma, in Shulman K, Marmarou A, Miller JD, et al (eds): lntracranial Pressure IV. Berlin/ Heidelberg/New York: Springer-Verlag, 1980, pp Teasdale G, Jennett B: Assessment of coma and impaired consciousness. A practical scale. Lancet 2:81-83, 1974 Received November 20, Accepted in final form February 24, These studies were supported by the Medical Research Council. Address reprint requests to: Sam Galbraith, M.D., Department of Neurosurgery, Institute of Neurological Sciences, Southern General Hospital, Glasgow, G51 4TF, Scotland. J. Neurosurg. / Volume 55 / July,