As many as one-third of all pediatric surgical admissions are because of trauma,1 and 50 per cent of the patients have experienced head trauma.2 During the school years, one child in every 10 will suffer a significant head injury with disturbed level of consciousness. One-third of these, or 3 per cent of all children, will be hospitalized during the school years for head injury. At the Children's Hospital of Philadelphia, 15 per cent of all neurosurgical admissions for 1975 (153 patients) were because of head injury.
The mortality from head injury in children is 1 to 2 per cent, or approximately 4,000 children per year.3·4 This low percentage may explain why little new information on the pathophysiology of pediatric head injury has appeared in the literature in the past few years. The mortality from more specialized neurosurgical units is considerably higher - 10 to 13 per cent.5*6 In severe head injury with coma, the mortality remains 50 per cent.7·8
The postconcussion syndrome, mental and personality changes, and other evidence of mild to severe brain damage have often been noted in children with head injuries. However, the incidence of posttraumatic sequelae is gratifyingly low - 3 to 4 per cent of all head injuries.2·9 Following severe head trauma, the incidence of severe and prolonged neurologic, intellectual, or personality problems is high.10·11 If coma exceeds 24 ho.urs following injury, significant sequelae can be expected in 10 to 50 per cent of patients.11 Nonetheless, many patients make a remarkable recovery and are able to return to school and become selfsupporting despite neurologic deficit.12 Two to 5 per cent of patients with severe head injury will remain permanently and severely handicapped.11
Several important technologic innovations in the past two years have enabled us to significantly increase our knowledge of the pathophysiology. The application of axial transmission computerized tomography (CT scanning) and emission computerized tomography for measurements of local cerebral blood volume have added a new chapter to our knowledge of the anatomic and physiologic changes occurring in the injured brain. The use of continuous intracranial pressure (ICP) monitoring* has allowed us to define exactly the time course and magnitude of the ICP changes and to measure the effectiveness and duration of therapeutic drugs. Finally, the realization that only a small percentage of even the most severe injuries can be benefited by surgery encourages the development of a team approach to the problem of head injury. This includes the anesthesiologist-intensivist, pediatrician, and neurosurgeon as a minimum.
CLINICAL GRADING OF PATIENTS
The following discussion will be concerned mainly with the severe grades of injury, and no effort will be made to catalogue readily available facts (Table 1).
Nothing can be done to reverse the immediate effects of the primary head injury. The cellular death and disruption that occur at the time of impact are not yet amenable to therapy. The secondary events (brain swelling, brain edema, hypoxia, and shock) are all amenable to correction. Therefore, it is to this secondary head injury that all our efforts are directed.
Jennett and co-workers13 have shown that 50 per cent of patients dying from head injury have evidence of ischemic lesions in the brain, and many of these patients had elevated ICP at some point during the course of their illness. This secondary ischemia may be avoidable if adequate therapy for brain swelling and elevated ICP can be administered.
Hendrick and colleagues4 have shown that 50 per cent of children dying from head injury are awake at the time of admission to hospital, yet 70 per cent of deaths occur within the first 48 hours.14 These figures emphasize that death is due to rapid progression of some secondary events, since coma and death did not result at the moment of impact.
The pathophysiology of this secondary injury is due to alterations in the arterial gases secondary to respiratory difficulties, alterations in systemic blood pressure, and alterations in the intracranial pressure. Any obstruction of the airway, common in the unconscious patient, will result in increased arterial CO2 tension or decreased oxygen tension. CO2 is a potent cerebrovascular vasodilator. Thus, hypercapnia leads to an increase in cerebral blood volume, causing a secondary increase in intracranial pressure, and an increase in end-capillary pressure, leading to increased fluid infiltration into the tissue and edema. Hypoxemia leads to a further decrease in tissue pC^ in areas of the brain already limited in oxygen availability by edema or elevated ICP, and systemic shock leads to ischemic hypoxia of the brain if the arterial pressure is low enough. Delayed formation of an intracranial mass lesion (epidural, subdural, or intracerebral hematoma) leads to an increase in ICP and a decrease in cerebral blood flow and further ischemic damage.
Only 7 per cent of head trauma patients admitted to our service require surgery - stressing the fact that head injury, in general, is not a surgical disease but, rather, has a complex pathophysiology best treated in an intensive-care unit by team effort.
The period from the time of injury to the time of arrival in hospital is one of great risk. An improperly positioned child may develop airway obstruction and suffer secondary injury before arriving at the hospital. To avoid this, good ambulance facilities and, in particular, welltrained ambulance personnel are vital. When the injury has occurred to the freely mobile head, concern for cervical spine injury should always be entertained. This necessitates careful transport, probably best accomplished if the patient is unconscious in the lateral position with the cervical spine straight and supported.
Upon the patient's arrival in hospital, immediate clearing of the airway manually and assisted ventilation with a mask and oral airway are usually adequate to maintain low CO2 and normal pa02 . Lateral spine films can then be obtained to ensure the absence of cervical fracture or dislocation before further manipulation of the head is performed. With grade III and IV patients (Table 1), our policy is to insert a nasotracheal or oraltracheal tube for further transport. This requires the early and full participation of an anesthesiologist. All diagnostic studies are then performed under light general anesthesia with sodium thiopental (Pentothal®), nitrous oxide and oxygen, and pancuronium if necessary.
The importance of the team approach is immediately apparent. While the anesthesiologist is controlling the airway, the pediatrician is establishing an intravenous line, drawing routine blood studies, and doing a full systemic examination to identify injuries other than those to the cerebrum. At the same time, the neurosurgeon can assess the neurologic status of the patient. The early presence of all members of the team allows each one to examine the patient before any drugs have been given. If no CT scan is available and no arteriography is immediately planned in a patient who is not comatose, endotracheal intubation may be too traumatic and attempts may lead to further problems. Thus, control with an oral airway and frequent suction with the patient in a semiprone position probably constitute the best way to protect the airway and maintain normal blood gas tensions.
Plain skull films should always be obtained in the severely injured child. In children under one year of age, x-rays should be obtained in cases of minor trauma if the patient is alert. Twenty-one per cent of all patients with head injury admitted to our service have a skull fracture. Forty-six per cent of patients aged less than one year have sustained a fracture. Injury, particularly from falls, in the first year of life is frequently associated with a skull fracture but no alteration of consciousness. X-rays of the cervical spine and chest are routinely obtained in the severely injured patient. Routine blood studies - including complete blood count and differential count, electrolytes, blood urea and nitrogen, creatinine, sickle cell preparation, prothrombin time, partial thromboplastin time, and blood type - are routinely obtained. Urine is obtained for evidence of blood. When a CT scan is available, this is unquestionably the best special study. If no CT scan is available, repeated careful neurologic evaluations are the key to further investigation. Any deterioration in neurologic status should be an indication for arteriography. As we shall see later, this is almost always preferable to exploratory burr holes. The patient in deep coma should, we believe, be studied early with arteriography, since there is no way to know whether the coma is a result of primary diffuse injury or secondary injury (i.e., swelling or intracranial hematoma). Occasionally, both conditions exist. In the past, the absence of significant intracranial mass lesion has been taken as an indication that no therapy is possible. This is no longer an acceptable viewpoint. As many children require therapy for elevated ICP (10 per cent) as require surgery (7 per cent).
We obtain an immediate CT scan on all grade III and TV patients. If a surgical lesion is seen, the appropriate operation is performed and the patient returned to the intensive-care unit following surgery with an ICP monitor in place. If no mass lesion is seen, the patient is returned to the ICU, where an ICP monitor is inserted. If the patient is irritated by the nasotracheal tube and the ICP is normal, the tube is frequently removed. If the ICP is high or the patient tolerates the endotracheal tube well, it is left in place for at least 24 hours. Patients who require an endotracheal tube for airway maintenance but who are restless are usually treated either with small doses of morphine administered intravenously or with pancuronium and controlled ventilation.
In older children, seizures are usually treated with phenytoin sodium (Dilantin®). If phénobarbital is used for the younger children, good airway maintenance is extremely important. Occasionally, because of apparent rapid deterioration, mannitol (1 gm./kg. in 20 per cent solution) is given. Following mannitol administration, it is important to complete an intracranial study to rule out a mass lesion, which may now be masked as a result of the hyperosmolar therapy. The role of steroids remains controversial in head injury, but we have generally used them. Recent evidence from West Germany15 suggests that in adult injuries a significant effect is obtained by using doses of 1:1.5 mg. /kg. of dexamethasone (Decadron®) for nine days.
In this section, a brief outline for the therapy of some specific pathologic lesions will be given. It must be remembered that any of these may exist in isolation or may exist in association with other, more serious lesions.
Simple lacerations of the scalp should be treated by shaving the area completely, debriding carefully after copious irrigation, exploring for the existence of fractures, and, after hemostasis is obtained, suturing. The profuse bleeding commonly seen in these types of wounds can be controlled by suturing the galea aponeurotica either as a separate layer or with a through-and-through suture that will not only approximate the edges of the wound but also produce the desired hemostasis.
Cephalohematomas are collections of blood between the bone and periosteum seen frequently as a complication of traumatic delivery; they generally do not require surgical excision. We strongly advise against drainage either through needle or by incision, since the blood itself is an optimum culture medium and infections of this space are extraordinarily difficult to control. On palpation many cephalohematomas feel like depressed skull fractures, because the outer rim of blood coagulates and feels hard and elevated while the center remains liquid and fluctuant. No surgery should be attempted until x-rays of the skull confirm the presence of fracture. Occasionally a cephalohematoma will calcify or even ossify, and surgical excision is then required for aesthetic purposes.
Linear fractures of the skull are documented by x-ray and may be treated conservatively. Fractures through the base of the skull can be demonstrated by plain x-rays in only 20 per cent of cases. The patient must be observed for the presence of other intracranial pathologies, but the fracture itself is of no consequence. In infants and young children, we repeat the x-rays in a few weeks to ascertain the characteristics of the fracture. A rather uncommon complication is the growing fracture of the skull that occurs in cases in which the fracture has not only broken the bone but also lacerated the dura. The arachnoid herniates and the constant pulsation separates the edges of the bone, producing a palpable mass that grows with time. The treatment of these cases requires an extensive craniotomy, with repair of the laceration of the dura after excision of the herniated portion of arachnoid (Figure 1).
Figure 1. X-rays of growing skull fracture taken six weeks apart.
Figure 2. "Ping-Pong fracture" of the right parietal region.
Figune 3. Compound comminuted skull fracture with herniating brain.
In the newborn and very young infants, depressed fractures of the skull produce no visible break in the x-rays but give the classic appearance of the so-called Ping-Pong fracture with depression (Figure 2). This is usually due to traumatic delivery and can best be demonstrated by tangential x-rays of the skull. Usually the depression is localized to a small portion of the calvarium, but it may occupy a large portion of the vault. Because of the simple procedure required to elevate such depressions, they should be surgically treated. We usually make a very small incision over one of the palpable sutures. A bone spatula is then placed under the depressed portion of the bone through the suture, and gentle pressure is brought on the bone to restore it to its normal configuration. There is evidence that some of these fractures will spontaneously elevate. If they do not, however, the delayed reconstruction of the contour of the skull is a rather difficult procedure.
Closed depressed fractures of the skull are surgically elevated if they are depressed more than the full thickness of the skull. The dura is explored for the possibility of laceration, and the fragments are replaced to follow the normal contour of the skull.
Depressed compound fractures of the skull are a surgical emergency (Figure 3). After the fracture is documented with skull x-rays, ineluding a tangential view, a CT scan is done to rule out other intracranial pathology. Surgical exploration is then performed. The fragments of bone are elevated, the dura is visualized, and, if there is suspicion of a subdural hematoma, the dura mater is opened. If the dura is lacerated, it is imperative that it be closed in a water-tight fashion. Large cerebral hernias extruding through any dural defect are not uncommon, and they are not an acceptable surgical complication. If the dura mater itself is torn beyond repair, we use a graft from the temporalis fascia or fascia lata or, in some cases, a siliconate Dacron. The bone fragments are returned to position after soaking in gentamicin solution (1 mg./cc), and the periosteum is used to cover the fragments and hold them in place. The patient is placed on antibiotics and anticonvulsants following surgery, and ICP monitoring is often performed.
Figure 4. CT scans of intraventricular hemorrhage taken three, 23, and 123 days after injury. Scan taken on the third day shows intracerebral and intraventricular hemorrhage. The patient was awake and alert with headache. No surgery was performed. Scan taken on the 23rd day reveals resoluten of the intracerebral hemorrhage with some ventricular dilatation. Scan taken on the 123rd day shows shrinkage of the ventricles back to normal size, with residual small left occipital infarction. The patient recovered completely without any surgical intervention.
Extradural hematomas are diagnosed by clinical examination or with the help of a CT scan or angiography. Their treatment is by emergency surgery. The procedure of choice is a craniotomy over the site of the fracture, with evacuation of the hematoma and hemostasis of the bleeding vessel either by clipping the responsible branch of the meningeal arteries or by coagulation. In some children, the epidural hematoma is of venous nature and is due to the laceration of one of the dural sinuses, either the sagittal or the transverse. In these cases, after evacuation of the hematoma, we attempt to repair the dural sinus laceration. It is very seldom necessary to lígate the sinus itself, and this should be done only in life-threatening situations.
Acute subdural hematomas are generally due to cortical lacerations, and surgery should correct both problems. After evacuation of the acute subdural hematoma, an attempt should be made to find the cortical laceration and obtain hemostasis of the cortical bleeders. Immediately thereafter the patient should be placed on anticonvulsants, and postoperative care should be geared to the control of the cerebral edema that occurs in all of these cases.
The management of chronic subdural hematoma, particularly in the infant, is still a matter of debate. If the patient shows signs of increased ICP or of an enlarging head, the choices of treatment are subdural taps and drainage of the subdural collection; subdural-peritoneal or pleural shunt; and, in cases where there are thick membranes or high protein fluid, bilateral craniotomies for removal of the hematomas. There is evidence that chronic subdural hematomas that do not produce pressure or an enlarging head can be treated simply by observation and serial head measurements.
Intracerebral hematomas diagnosed by CT scan or angiography should be surgically removed only in cases in which the hematoma is acting as a space-occupying lesion, producing signs of pressure and acting as a tumor. This is particularly indicated if the hematoma is in one of the polar regions, such as the frontal or occipital poles or the temporal tip. The prognosis is far graver if the hematoma is in the ganglionic mass, and surgery is performed only in a desperate effort to save the life of the patient. This lesion carries with it major neurologic deficit if the patient survives. The surgical procedure is craniotomy with a transcortical incision and evacuation of the hematoma, with careful hemostasis of the wall of the cavity. Burr holes and drainage through a cannula, do not constitute effective treatment and are probably more dangerous than useful.
Intraventricular hemorrhage without intracerebral hematomas are not surgically explored. The patient is treated by control of the ICP and ventricular cerebrospinal fluid drainage when the clots Iy se. Although this entity was once considered fatal, most children will survive and will rarely require a shunting procedure for control of secondary hydrocephalus (Figure 4).
The CT scan is the safest, fastest, and best initial study for the child with a head injury. It readily diagnoses acute extracerebral hematoma, shows ventricular size and focal intracerebral disease, and differentiates hematomas from edema. It is also a safe method with which to obtain serial studies. The longitudinal follow-up of patients with head injuries has been invaluable to our understanding of the pathophysiology of pediatric head injury. In only one case has the scan missed a significant mass lesion. This was a one-year-old boy with a right subdural lesion that was isodense with the brain, with minimal evidence of ventrìcular shift. Other tests, including arteriography, should be performed if the CT scan fails to reveal a suspected subdural lesion. The results of 74 consecutive CT scans are shown in Table 2. Because of the usage limitations of the CT scanner, only grade lib. III, and IV patients have been studied (approximately 40 per cent of all our inpatients). Thirty-nine per cent of the scans were normal despite significant neurologic deficit. The presence of a normal scan did not necessarily signify minimal disease or a good outcome. The largest percentage of abnormal scans (24 per cent) showed diffuse brain swelling. This was diagnosed (Figure 5) by small or absent ventricles, small or absent mesencephalic cisterns, and subarachnoid space and was later verified by repeat studies showing the ventricular system to have increased in size. We initially interpreted this picture as brain edema. Careful examination of the CT scans and measurements of brain density showed that, as recovery progressed, the brain became less dense rather than more dense - the reverse of what is seen during resolution of edema. Local cerebral blood volume (LCBV) measurements were made using the Mark IV scanner, and these showed that the LCBV was increased (Figure 6) in the first two to three days after injury and had returned to normal by five to 10 days. In one patient, cerebral blood now was measured using xenon injection; a marked cerebral hyperemia, with flows 50 per cent above normal, was found.
CT DIAGNOSIS OF 74 CHILDREN WITH ACUTE HEAD TRAUMA
Figure 5. CT scans obtained between the first and 231st days after injury. Scan 1 shows small to absent ventricles with absent perimesencephalic cisterns. Repeat scan on day 9 shows enlargement of the ventricles, with an extracerebral collection overlying the left frontal area. At 16 days, the extracerebral collection has disappeared and the ventricles have further enlarged. A similar picture is seen on day 37. However, by day 231 the ventricles have returned to normal size and no evidence of atrophy or hydrocephalus is seen. The patient made a complete clinical recovery after having initially been grade IV.
The clinical picture in 56 per cent of these patients was the same. A relatively minor injury, not associated with prolonged loss of consciousness, was followed within hours by the onset of headache, pallor, vomiting, and progressive obtundity and pediatric concussion syndrome.15 In these patients, the ICP was not elevated during the first 24 hours of ICP recording.
On follow-up CT scan, 33 per cent of these patients showed, along with an increase in ventricular size and cisterns, the presence of extracerebral collections of clear-fluid CSF density over one or both frontal areas (Figure 5). In two-thirds of the cases, the collections disappeared spontaneously over one to two weeks. In one patient, a chronic subdural lesion was present that required surgical drainage; in a second, the picture went on to severe hydrocephalus, and a ventriculoperitoneal shunt was required.
As the time from injury increased, the ventricles frequently enlarged above normal size, only to return to normal again within two to three months (Figure 5). Eleven per cent of patients with diffuse swelling have required a ventriculoperitoneal shunt.
Specific findings (i.e., subdural or extracerebral hematomas) are treated in the standard fashion and are readily diagnosed with the CT scanner.
When no CT scanner is available, arteriography should be performed in all comatose patients to ensure that no mass lesion is present and in all patients who demonstrate a deteriorating Ie «'el of consciousness while under observation. There is no other way to be sure that a mass lesion is not present. We do not believe that burr holes are adequate to rule out unusually located epidural hematoma, intracerebral hematomas, or diffuse brain swelling with slowing of the cerebral circulation. Arteriography is indicated when a CT scan is difficult to interpret or reveals a picture not previously seen. This is particularly so when the suspicion of chronic subdural hematomas is entertained.
The greater the disturbance of brain function, the more difficult it is to estimate the level of ICP. Clinical signs, opisthotonos, decerebration, episodes of bradycardia, and intermittent pupillary dilatation may or may not be associated with an elevated ICP. Whereas the initial estimates of ICP are difficult, so is the evaluation of the response to therapy. Thus, the same per-kilogram dose of mannitol for two patients may have radically different effects on their intracranial pressures. Finally, when the ICP is high and difficult to control, intensive therapy (i.e., paralysis, hyperventilation, or the use of barbiturates) can be safely used only if the ICP is monitored. The important consideration is for the cerebral perfusion pressure (i.e., the difference between the systemic arterial pressure and intracranial pressure), since this represents the driving head for blood flow through the brain. Thus, all our patients who have ICP monitoring also have monitoring of systemic arterial pressure.
Figure 6. CT and blood volume scans at four to 20 days after injury. On day 4 the CT scan shows small ventricles, with diminution in size of perimesencephalic cisterns. Blood volume at that time in the white matter is markedly elevated. By day 8 the CT scan has returned to normal except for some diminished density on the right side. By day 20 the regional blood volumes have returned to a normal level. The child made a complete clinical recovery.
Figure 7. Typical volume pressure curve for the intercranial space.
Ten per cent of our head-injury patients (grade IV and some grade III) are considered sick enough to require ICP monitoring. Because of the frequent finding of small or absent ventricles on the CT scan, we use a subarachnoid hollow screw.17 This is inserted just anterior to the coronal suture via a twist drill hole, and the procedure can safely be accomplished in the intensive-care unit. The monitor is inserted immediately following surgery, if necessary, or immediately following CT scanning.
Within the cranium there are three compartments: blood, brain, and CSF. If one enlarges, it must be at the expense of the other two. The brain contributes much less to the compliance of the system than the blood and CSF. Such factors as the rate of increase and the site of increase of volume within the cranium are important. If focal increase in volume occurs (i.e., epidural hematoma), brain shift may occur and may cause interference with function and signs of uncal herniation even when the ICP is not extremely high. The volume pressure curve for the CSF is an exponential curve (Figure 7). The steep portion of the curve at which the ICP increases rapidly with further small increases in volume varìes, depending on the rate of addition of volume. Thus, a single measure of pressure does not inform us where on the curve the patient is and, therefore, how much at risk he or she is of developing pressure waves. The concept of the volume pressure response (VPR) was introduced to give additional information about the compliance.18 Using a ventricular cannula, a small amount of fluid can be removed or added and the value ΔV/ΔP (change of volume/change of pressure) calculated. The smaller the value, the lower the intracranial compliance. The subarachnoid screw does not lend itself to VPR testing, since the area of dispersion of pressure is quite varìable. Despite these limitations, an estimate of the compliance can be made using the subarachnoid screw. The patients undergo suction and are turned every hour. This is accompanied by variations in ICP (Figure 8). In patients with a high compliance, the spikes of pressure return to baseline immediately after the suction is completed. In patients with a low compliance, small-plateau waves are triggered (Figure 8). If these are seen, the chances are high that the ICP will rise and require therapy; indeed, we believe that therapy can usefully be instigated at this time.
We have not found the ICP to be elevated in the first 24 hours following injury, even in patients with a CT picture of diffuse swelling. We have, however, seen small-plateau or A waves triggered by suction.19 Fifty per cent of the patients we have monitored have gone on to develop an increase in ICP over 25 mm. Hg, requiring therapy. Seventy-six per cent of the patients with diffuse brain swelling we have monitored have developed elevated ICP. This increase in ICP is delayed to the second or third day after injury. Once elevated ICP occurs, therapy is required for a considerably longer period than has previously been realized. For our patients, the mean length of therapy was 6.5 days; the longest time was 11 days.
The beneficial effects of corticosteroids in children with head injury are still unproven. Recent data concerning head injury in adults suggest that if steroids are used they should be given in doses equivalent to 1.5 mg. /kg. of dexamethasone (Decadron®) per day. We have employed rather arbitrary criteria for the use of corticosteroids in injured children. However, all the children with ICP monitors received dexamethasone in a dose of 1 mg./kg./day for at least seven days.
Since we are convinced that the initial phase of diffuse swelling is due to increased blood volume and increased blood flow, we instigate hyperventilation to a paC02 of 25 torr* as soon as the patient is stabilized in the ICU. We have shown that this is effective in decreasing the hyperemia and that it may be possible to avoid the secondary elevations in ICP that occur on the second or third day. Mannitol itself causes cerebral hyperemia; it should be used cautiously in the early stages of injury, when the ICP is not elevated but a cerebral hyperemia, frequently with a low compliance, is present. Later, when ICP waves are occurring, mannitol (in a dose of 1 gm. /kg. given intravenously) is effective in abolishing the waves and lowering the ICP. We have found, however, that the effects in children (lasting about three hours) are shorter lived than those in adults. Mannitol becomes less effective with repeated use, and larger amounts are required. After 48 hours of administration, it may become ineffective. If the ICP is still a problem, we cool the patients to 32 degrees. Frequently, the hypothermia itself will control the waves for a time. If they recur, mannitol may again be effective.
The use of barbiturates for the treatment of increased ICP remains somewhat anecdotal, yet we are convinced that these compounds have a role to play in the treatment of headinjury patients. The use of small amounts of thiopental just before suction will avoid the plateau waves we have discussed above. Also, when the ICP is not responding to other therapy, the use of pentobarbital in hourly doses to maintain a blood level of 3 mg./100 ml. will frequently bring the pressure under control. Therapeutic procedures, such as paralysis and hyperventilation or the use of barbiturates, cannot be used in the absence of ICP monitoring, since there is no way of ensuring that the therapy is effective and that the cerebral perfusion pressure is increased and not decreased by the therapy. The use of such potentially hazardous therapy again emphasizes the importance of a team approach to the treatment of severe head injuries.
Figure 8. ICP tracing in head injury. Tracing A shows suction ICP spikes that immediately fall to the baseline. Tracing B shows small-plateau waves triggered by suction, suggesting low compliance.
We have found that if therapy is needed beyond three to four days, a repeat CT scan will frequently show ventricular enlargement with increase in the subarachnoid spaces. If no mass is present and this picture is seen, we use repeated lumbar punctures or ventricular drainage as an ancillary measure to maintain a normal ICP.
Among patients in whom therapy for ICP is required, 37 per cent have received dexamethasone, hyperventilation, and mannitol; 25 per cent, these three plus hypothermia to 32 degrees; and 37 per cent, these four plus barbiturates. Thus, when the ICP is elevated it stays elevated for a long time and can be particularly difficult to keep under control. The mortality in this group of monitored patients is 6.6 per cent. We believe this is encouragingly low compared with the 50 per cent mortality reported for children in coma following head injury.11 The morbidity is major, however. Eighty-six per cent of patients in whom the ICP did not rise above 25 mm. Hg and in whom no therapy other than dexamethasone was required recovered in a few weeks without major residual neurologic difficulties. One of the patients in whom the ICP was elevated and required therapy for 11 days died because of progressive pulmonary interstitial fibrosis some two weeks after the ICP problems had resolved. Eighty-seven per cent of this group had significant neurologic deficit that required intensive rehabilitation at the time of discharge from the acute-care hospital. One-third were in vegetative state, and one-third had severe focal neurologic deficits. AU were discharged to rehabilitation hospitals. At the present time, 70 per cent of these children are performing school work at their grade level despite severe hemiparesis in one. (One of these children had no caloric responses for two weeks following the injury.) One child remains in a severe diffusely damaged neurologic state and is just beginning to smile and move spontaneously. The other 20 per cent are slowly recovering from their injury and will probably return to a self-supportive state in time. The realization that 10 per cent of headinjury patients require ICP monitoring and only 7 per cent require surgical intervention following their head injury emphasizes the importance of ICP monitoring.
The mortality for the past 150 patients with head injuries admitted to our hospital is 1.3 per cent. Considering that 30 per cent of the patients were in grade III or IV coma, this is an unusually low rate. Rehabilitation begins as soon as the immediately life-threatening period is over. Every effort to stimulate the brain with visits from the parents and the use of tape recorders to continuously supply the sound of parental voices is begun in the ICU. With intensive postinjury rehabilitation, we are always surprised to see how well even the most damaged children will do with time. If death can be avoided, most of the children who suffer from major head injuries can be expected to eventually make a useful recovery.
The major advances in the treatment of children with head injuries are related to the introduction of the CT scan - making rapid, noninvasive diagnoses and serial studies readily available - and to the realization that only a small part of the treatment of head injury is surgical. The latter point, coupled with the concept of the second head injury, has contributed to the development of a team approach to the intensive care of these patients and to the development of safe, reliable ICP monitoring techniques in children. With this approach, few, if any, children should die because of secondary brain swelling or edema. If life is preserved, what is its quality? We and others11 have shown that even in the worst cases, if the child does not die and an intensive rehabilitation program is instigated, recovery to a self-sufficient state can be expected in all but a very small percentage of children.
1. Rickham, P. P. Head injuries in childhood. HeIv, Chit. Acta 28 (1961). 560-575.
2. Craft, A. W., Shaw, D. A., and Cortjidge. N. E. F. Head injuries in children. Br. Med. J. 4 (1972), 200203.
3. Allen. P. D. Head injuries in children. Surg. Clin. North Am. 21 (1941). 323-330.
4. Hendrick, E. B., Harwood-Nash, D. C1 and Hudson, A. R. Head injuries in children; A survey ot 4,465 consecutive cases at the Sick Children's Hospital, Toronto, Canada. Clin. Neurosurg. 11 (1964), 46,
5. Harris, P. Head injuries in childhood. Arch. Dis. Child. 32 (1957), 488.
6. Scarcella. G., and Fields, W. S. Recovery from coma and decerebrate rigidity of young patients following head injury. Acta Neurochir. (Wein) 10 (1962). 134-144.
7. Graszkiewicz, J., Doron. Y.. and Peyser, E. Recovery from severe craniocerebral injury in brain stem lesions in childhood. Surg. Neurol. 1 (1973). 197-201.
8. Robertson. R. C. L, and Pollard. L., Jr. Decerebrate state in children and adolescents. J. Neurosurg. 12(1955). 13-17.
9. Hjem, B.. and Nylander, I. Acute head injuries in children: Traumatology, therapy and prognosis. Acta Pediatr. Suppl. 152, 1963).
10. Hjem, B., and Nylander, I. Late prognosis of severe head injuries in children, Aren. Dis. Child 37 (1962), 113-116.
11. Pazzaglia, P.. et al. Clinical course and prognosis of acute post-traumatic coma. J. Neurol. Neurosurg. Psychiatry 38 (1975). 149-154.
12. Richardson, F. Some effects of severe head injury: A follow-up study of children and adolescents after prolonged coma. Dev. Med. Child Neurol. 5 (1963), 471-482.
13. Jennet, B., et al. Ischemic brain damage after fatal blunt head injury. In McDowell, F. H" and Brennen, R. W, (eds.). Cerebral Vascular Diseases, Eighth Conference. New York: Grune & Stratton, 1973, pp. 163-170.
14. Jamieson, D. L., and Kaye, H. H. Accidental head injury in childhood. Arch. Dis. Child. 49 (1974), 376-381.
15. Faupel, G.. et al. Ooubie-Wind study on the effects of steroids on severe closed head injury. In Dynamic Aspects of Cerebral Edema. Berlin: Springer. (In press.)
16. Schnitker, M. T, A syndrome of cerebral concussion in children. J. Pediatr. 35 (1949), 557-560.
17. James, H. E-. Bruno, L A., and Schut, L. Intracranial subarachnoid pressure monitoring in children. Surg. Neurol. 3 (1975), 313-315.
18. Miller. J. D. Volume and pressure in the craniospinal axis. Clin. Neurosurg. 22 (1975), 76-105.
1 9 . Lundberg, N . Continuous recording and control of ventricular fluid pressure in neurosurgical practice. Acta Psychiatr. Scand. 36 (Suppl. 149. 1960).
CLINICAL GRADING OF PATIENTS
CT DIAGNOSIS OF 74 CHILDREN WITH ACUTE HEAD TRAUMA