Clinical Radiology (1990) 42, 88-90

Implications of Scanning Recently Head Injured Patients in General Hospitals B. J E N N E T T * and P. M A C P H E R S O N t

Departments of *Neurosurgery and t Neuroradiology, Institute of Neurological Sciences, Glasgow Increasing availability of computed tomography (CT) in general hospitals makes it appropriate to scan certain categories of acutely head injured patients in these hospitals. Policies should be devised locally indicating which types of patient should be scanned there, and what circumstances require transfer to the neurosurgical unit (NSU). Consideration must be given to the implications for training, staffing and other CT scanner commitments. The paper discusses these issues in the light of a study of the CT scans and surgical treatment of 1551 patients admitted to an N S U over a five year period, and provides models for discussion. Jennett, B. & Macpherson, P. (1990). Clinical Radiology 42, 88-90. Implications of Scanning Recently Head Injured Patients in General Hospitals

In most countries other than North America and Japan, neurosurgical facilities are restricted to large regional centres. In some countries, including Britain, CT for acute head injuries has until recently been readily available only in these centres. Long before the coming of CT neurosurgeons had been concerned that many of the limited beds in such centres might be filled with head injured patients unless transfer was controlled by strict triage criteria. Previously only seriously ill patients likely to have an intracranial haematoma were transferred. When CT revealed that a significant haematoma can be present before clinical deterioration has occurred, policies were changed in Glasgow, and twice as many patients were transferred to the NSU each year for scanning and assessment (Bryden and Jennett, 1983). More hospitals are now acquiring CT, making it necessary to develop policies about which patients should be scanned locally, and which of these should then be transferred; and to identify patients who should be sent without delay to the regional NSU, where scanning would be done. Unless a clear policy is evolved, the wider availability of scanning could lead to inappropriate management, putting some patients with serious injury or complications at greater risk than at present. There are also organizational implications for those who propose to scan recently head injured patients in a general hospital. An analysis of the CT scans and surgical treatment of 1551 patients admitted to the Glasgow NSU over a fiveyear period (Macpherson et al., 1990), together with many years' audit of head injured patients, forms the basis of this discussion paper. O B J E C T I V E S OF SCANNING ACUTELY HEAD I N J U R E D P A T I E N T S IN A GENERAL H O S P I T A L The first and most important objective is the early diagnosis of an acute haematoma. This should involve scanning patients who are not yet showing clear clinical signs of this complication, but who have risk factors which increase the probability that this complication will Correspondence to: Professor Bryan Jennett, Department of Neurosurgery, Institute of Neurological Sciences, Glasgow G51 4TF

develop. Extensive investigations of patients in A and E departments and of those who develop surgically significant haematomas have identified the statistical significance of combinations of the two risk factors (skull fracture and impaired consciousness) (Mendelow et al., 1983; Teasdale et al., 1990). This makes it possible to select which patients who are not in coma should be scanned. Scanning may reveal that a clot has already formed although it has not yet produced definite clinical signs of compression. There would be no doubt about the need for urgent transfer of such a case to the NSU. However, our study in Glasgow has shown that a number of patients whose first scan was abnormal but did not show a lesion of surgical significance, do later need surgery because of a lesion that was revealed by a second or third scan (Macpherson et al., 1990). This figure would undoubtedly be higher if the first scan was done sooner after injury, as would be possible at a district hospital. A second objective of scanning is to reduce the number of hospital admissions and transfers. Knowledge that the CT is normal, in conjunction with the clinical state, should lead to fewer patients being admitted locally for observation. Knowledge that there is no lesion requiring immediate ~urgical intervention should also reduce the necessity for urgent transfer to the NSU. In the Glasgow NSU series more than a third of first scans were normal, including those of a fifth of the patients who were in deep coma. Some of these patients were only moderately injured but had been transferred because they had risk factors for a haematoma. For those in coma the probability of their having a haematoma was much higher than for conscious patients, but these patients are at special risk during transportation (Gentleman and Jennett, 1990). The development of intensive care units in many hospitals may make it feasible to manage locally some patients who are in coma but whb do not require surgical intervention. However, even if not needing surgery for evacuation of a clot, many comatose patients could benefit from the skills of specialized medical, nursing and technical staff who are available in an NSU. They are familiar with the monitoring of intracranial pressure and of other aspects of patients in coma, as well as with appropriate medical and supportive measures. A third objective of CT would be its contribution to the

C T OF H E A D I N J U R I E S I N G E N E R A L H O S P I T A L S

monitoring of patients who remain at the base hospital and who have impaired consciousness or coma. Repeat scans may indicate the development or resolution of various intracranial complications. Some haematomas in the Glasgow series were first operated on only after a second or third scan had shown a surgically significant lesion. In all these cases the initial CT scan had shown some abnormality, but on occasions this was only a minor lesion (e.g. small contusion). The surgically significant lesion had developed either as an extension of the original abnormality or of a new lesion. However, although eight patients with a normal first scan developed a lesion, including two with a subdural haematoma, no patient whose initial scan had been normal required surgery. This may not always apply when initial scans are performed sooner after injury. A fourth objective is to identify patients with lesions other than contusions and haematomas. In our series evidence of shearing injuries in the absence of surgically significant intracranial haematoma was detected in 13 % and general brain swelling in 9% of patients. In some clinical settings these are associated with a poor prognosis with currently available treatment (Colquhoun and Burrows, 1989; Teasdale et al., 1984). Such patients might qualify for entry into trials of new therapies and their identification could, therefore, become important. PRACTICALITIES From cases in which a clear-cut haematoma is revealed, it is easy to assume that CT is a simply performed and easily interpreted investigation, which should be widely available. However, without appropriate training, the interpretation of radiological findings by non-specialist staff could lead to false positive and false negative diagnoses of intracranial haematoma, whilst other more subtle abnormalities could easily be overlooked. Further, the continued management of seriously injured patients cannot be conducted on the basis of a single scan. In the Glasgow study more than one scan was performed in over 40% of patients, and as already explained some had surgery only after the second or third scan. Repeat scans were most often requested because of neurological deterioration or continuing coma. Clinicians looking after head injured patients in general units have to be aware of the possible development of such remediable complications. Only a few of the many head injuries seen at a large general hospital each week wiUjustify scanning. That may make it difficult for radiographers to develop and maintain the necessary skills to obtain good images, and for local radiologists and clinicians to learn to interpret them reliably. Even with high-resolution CT performed by radiographers experienced in scanning restless and unconscious patients, we have found that 7% showed appreciable movement artefact. There would likely be more such instances at a primary hospital, where scans are performed earlier and the patients unsedated. Radiographers could be sent to the regional NSU for experience of dealing with head injured patients, and of the setting up of optimum scanning cuts and hard-copy factors. Initial tuition and revision seminars in the interpretation of SCans for radiologists and clinicians could be held from time to time by regional neuroradiologists. It is, however, now possible to transfer data from a local scanner by

89

telephone line to a specialized unit, where the image can be scrutinized by a neurosurgeon or neuroradiologist (Marsh et al., 1989). We recommend that such links be developed between the regional NSU and its catchment general hospitals. This would be useful not only for head injuries but for other diagnostic scanning problems. Local discussions would need to determine how general radiologists and neuroradiologists would collaborate, and whether consultation outside normal hours would be different (e.g., directly with the neurosurgeon on call). Another problem is the frequency with which scans after acute injury are required outside normal working hours; this applied to 90% of first scans in the Glasgow study. Whilst regional units have arrangements for dealing with this, other hospitals may need to call out radiographers and possibly radiologists. If there is to be 24-hour cover, then a rota of suitably trained radiographers would be required. When emergency scans are required for acute head injuries during the working day, there may be conflict with commitments already booked for the scanner. PATIENT S E L E C T I O N This paper is concerned primarily with the implications for general hospitals of undertaking CT for recently head injured patients. This begs the question of who should be scanned and where? There has been much emphasis in recent years on the use of guidelines to select patients who should have skull radiography after recent head injury. For these guidelines to be simply adapted for the use of CT would be wasteful of resources. A scanner with booked commitments could not image all the head injured patients who at present have skull films taken in this country. In some circumstances, skull radiographs should still be the initial investigation, as part of the triage for scanning, and when a vault fracture that may be depressed is suspected under a scalp laceration. In those who have had, or continue to have, alteration in conscious level, CT should be performed locally without recourse to skull films. Other categories of patients should be referred to the NSU without being imaged at the general hospital. Whatever the initial course of action taken in an individual case, the two main aims should be to expedite the transfer to the NSU of patients likely to require surgical intervention; and to identify patients for whom urgent admission to the N S U is not necessary. This latter category could cover a wide range of clinical states from the trivial head injury not requiring admission, to the comatose patient who might be cared for locally if ICU facilities were adequate. Avoiding unnecessary transfers would reduce the hazards associated with ambulance transfer of seriously ill patients, and the need for medical and nursing escorts to abandon their other duties. This should also make more beds available in NSUs for the admission of more appropriate patients. The preliminary transfer of a patient in coma from one Accident and Emergency Department to another general hospital for CT is a policy to be discouraged. It can only result in more delay for patients who need specialist care, as well as an increased risk o f complications during transportation for those who have subsequently to be transferred to the NSU. We recommend that comatose patients requiring C T who have been admitted to a

90

CLINICAL RADIOLOGY

hospital without CT facilities should be transferred directly to the NSU. Regional trauma centres equipped to deal with patients who have major multiple injuries, including those with head injuries, have recently been recommended (Yates, 1988). These would not help with most head injuries, however, because only a minority of patients have multiple injuries, while many who develop intracranial haematoma were not initially severely injured, and would not, therefore, have been sent to a trauma unit. In 1984 a representative group of British neurosurgeons published guidelines including criteria for neurosurgical consultation in the initial management of head injured adults (Group of Neurosurgeons, 1984). The main goal was to identify patients in whom there was a substantial risk of their already having, or of their developing, an acute intracranial haematoma. The statistical risks on which this advice was based (Group of Neurosurgeons, 1984) have recently been updated and extended to include children (Teasdale et aL, 1990). With the increasing availability of CT the situation has now altered. Although the basic principles of scanning are likely to be similar in all general hospitals, what is undertaken will vary according to local circumstances. To ensure that optimal use is made of both local scanning facilities and of regional NSUs, local clinicians (including anaesthetists and radiologists) should agree with neurosurgeons and neuroradiologists what arrangements seem most appropriate where they work. As a basis for such

discussions, a checklist on the implications of scanning is given in Appendix 1, and some suggested criteria about when and where to scan in Appendix 2. REFERENCES

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. Gentleman, D & Jennett, B (1990). Audit of transfer of unconscious head-injured patients to a neurosurgical unit. Lancet, i, 330-334. Group of Neurosurgeons (1984). Guidelines for initial management after head injury in adults. British Medical Journal, 288, 983-985. Macpherson, P, J e n n e t t i ~ - ~ Ariderson, E (1990). CT scanning and surgical treatment of 1551 patients admitted to a regional neurosur. gical unit. Clinical Radiology, 42, 85-87. Marsh, H, Maurice-Williams, RS & Hatfield, R (1989). Closed head injuries: where does delay occur in the process of transfer to neurosurgical care? British Journal of Neurosurgery, 3, 13-20. Mendelow, AD, Teasdale, G, Jennett, B, Bryden, J, Hessett, C & Murray, G (1983). Risks ofintracranial haematoma in head injured adults. British Medical Journal, 287, 1173 1176. Teasdale, E, Cardoso, E, Galbraith, S & Teasdale, G (1984). CT scan in severe diffuse head injury: physiological and clinical correlations. Journal of Neurology, Neurosurgery & Psychiatry, 47, 600-603. Teasdale, GM, Murray, G, Anderson, E, Mendelow, AD, Macmillan, R, Jennett, B & Brookes, M (1990). Risks of traumatic intracranial haematoma in children and adults: implications for managing head injuries. British Medical Journal, 300, 363-367. Yates, DW (1988). Action for accident victims: plans to stop patients dying unnecessarily from major trauma. British Medical Journal, 297, 1419-1420.

APPENDIX 1 Items for Consideration Before Embarking on the Use of CT for Head Injury in a General Hospital

Period of cover: (a) Normal working hours - f i t in emergencies between booked cases; fewer radiographers involved, therefore better trained; interpretation by radiologists. (b) Out-of-hours cover - rota of radiographers required; interpretation by radiologists or clinicians? Training: Radiographers - technique Radiologists - interpretation Clinicians - interpretation (rotating doctors less likely to be accurate)

Aspects of interpretation: (a) Realizing that reliance cannot always be placed on a single examination- delayed haematomas do develop and existing lesions enlarge; (b) Recognizing the features of shearing injuries and general swelling. Transfer of images: (a) Working hours - by radiologists to neuroradiologists or neurosurgeons; (b) 0utwith working hours - by clinicians or radiologists to neurosurgeons; (c) Necessity for a formal neuroradiological report, especially if the patient is not transferred. Keeping of coma patients in the local ICU: Availability of beds and expertise.

APPENDIX 2 Criteria to Form Basis of Discussions for CT Scanning and for Transfer to N S U after Recent Head Injury ]based on Guidelines (Group of Neurosurgeons, 1984) and on Risk factors for Intracranial Haematoma (Mendelow et al., 1983; Teasdale et al., 1990)]

For CT at first hospital: Confusion or more marked impairment of consciousness. Skull fracture. Difficulty in assessing the patient, e.g. alcohol intoxication, epilepsy. Transfer to NSU precluded for medical or practical reasons. Transfer to NSU regardless of CT findings- better to postpone CT:

Compound fracture, or penetrating injury (inc. gunshot). Fractured skull base suspected (a) clinically: CSF from nose or ear, periorbital or mastoid haematoma; (b) radiologically: sinus fluid level, IC gas. Deteriorating level of consciousness. Focal neurological signs. Transfer to NSU because of CT findings: Significant haematoma. Consultation with NSU: General swelling. Local swelling/shift/contusion. Diffuse axonal injury. Coma persisting after resuscitation.

Implications of scanning recently head injured patients in general hospitals.

Increasing availability of computed tomography (CT) in general hospitals makes it appropriate to scan certain categories of acutely head injured patie...
384KB Sizes 0 Downloads 0 Views