Head Injury Guidelines

Sources: Kirkpatrick PJ. Editorial: On guidelines for the management of the severe head injury. J Neurol Neurosurg Psychiatry 1997;62:109-111; Bullock R, et al. Guidelines for the management of severe head injury. J Neurotrauma 1996;13:639-734.

Under the auspices of the charitable brain Trauma Foundation, the American Association of Neurological Surgeons and its Joint Section on Neurotrauma and Critical Care, an expert Task Force of neurosurgeons has assembled the above-titled Guidelines for the Management of Severe Head Injury.

The task force has published its Guidelines under 14 different headings, selecting treatment modalities supported by current clinical experience and well-designed, scientifically oriented reports appearing in the medical literature since 1966. The following paragraphs group together and summarize the 14 topics in three relevant sections, all aimed at the treatment of patients suffering a severe head injury.

1. Patients with severe brain trauma and fulfilling a Glasgow Coma Scale at levels of 3-8 are best treated in experienced trauma centers possessing skilled neurosurgical participation. Such centers should also have the capacity to provide telephone or other rapid assistance to help rural hospitals provide initial physiological resuscitation and, if necessary, carry out immediate removal of large intracranial hematomas. Prompt transfer from non-specialized institutions to a head trauma center is strongly recommended, using, if necessary, short-lasting neuromuscular blockade and sedation to ease the problems of transport. Protracted use of paralytic agents is not recommended.

2. First therapeutic efforts either at hospital of first treatment or trauma centers consist of assuring and maintaining mean blood pressures at or above 90 mmHg in an effort to maintain a physiological cerebral perfusion pressure (CPP) of >70 mmHg. Arterial PaO2 > 90 mmHg should be maintained if at all possible, and, unless CT signs of brain swelling or transtentorial herniation appear, PaCO2 levels should not decline below about 35 mmHg. Given CT or clinical signs of brain swelling, especially in patients older than 40 years with posturing or/and a systolic blood pressure below 90 mmHg, intracranial pressure (ICP) should be monitored via a catheter inserted in an anterior cerebral ventricle. This procedure also enables CSF drainage, if indicated. Also, pressure readings provide a moderately reliable indication of CPP (= mean BP minus ICP). Maintenance of euvolemia and nutrition are important.

The guidelines stress that dangerous levels of CPP emerge at ICP greater than 20-24 mmHg or are implied by the advent of brain swelling diagnosed by CT scan or clinical signs indicating transtentorial herniation. Efforts to reduce CPP must often be treated by both maintaining physiologic levels of systemic blood pressure and attempting to lower ICP by shrinking the brain. For the latter step, repeated small boluses of mannitol are recommended, being careful to avoid raising the serum osmols above 320 mmol/L. Passive hyperventilation producing PaCO2 levels down to but not below 25 mmHg may be lifesaving but should be applied only briefly for reasons discussed below. Some neurosurgeons apply high-dose barbiturate anesthesia to patients in whom the above supportive measures have failed to ameliorate severe increases of intracranial pressure. The possible beneficial or worsening effects of these approaches await carefully conducted late outcome studies.

3. Literature searches of three topics were sufficiently strong to recommend them unequivocally. First is that the prophylactic, sustained use of hyperventilation "should be avoided," a principle supported by a prospective, controlled study which found worse outcomes when using such techniques (Muizelaar, et al. J Neurosurg 1991;75:731-739). Second is that a large number of data indicate that prophylactic anticonvulsants are not useful or recommended unless seizures have been observed during the course of acute trauma and its immediate consequences. Third is a firm recommendation not to use glucocorticoids in brain trauma patients unless they have a specific, known depletion of internally generated adrenal hormones.


This is a useful document that summarizes current practices and recommendations of leading American neurosurgeons. The report was also reviewed and supported by representatives of several other neurological or traumatological disciplines as well as by a European Neurosurgical Advisory Committee. In addition to the lead-off reference, a 165-page loose-leaf binder and CD ROM describing the detailed Guidelines can be obtained from the American Neurosurgical Association at (847) 692-9500 ext. 39.

We emphasize and support the principle that hyperventilation should be applied only under circumstances that indicate the sudden or rapid increase of increased ICP and never as a continued process. The reason for this position is that during the continued application of hyperventilation to a low PaCO2 (e.g., 25-28 mmHg), the brain buffers its resulting alkalosis and the blood volume begins to climb. This, of course, increases the ICP, which then becomes even harder to reverse by mere hyperventilation-induced arterial constriction.

The formulators of this report deserve considerable praise and encouragement in summarizing current, literature-based approaches to the treatment of severe head trauma. As the investigators indicate, the Guidelines reflect the confirmed therapeutic practices of the most capable neurological surgeons in the United States who have a high experience with head trauma. No mention is given to the view that, as occurs in large strokes, impeded cerebral perfusion in brain trauma reflects primary cellular injury. Such injury swells cells, thereby focally squeezing capillaries in multiple damaged sites. Only late in its course (24-96 hours) does this process usually induce tissue-displacing expanding intracerebral masses which then further impede blood flow. Such a probable sequence makes your editor skeptical that changes in blood flow are an early cause of progressive brain damage. Rather, they may represent the ultimately lethal response that reflects widespread, multifocal, and presently incurable degrees of neuronal cell death. Only a research-filled future will solve these problems.

An additional comment may be deserved. The report says nothing about the much-needed detailed outcome studies of head injury patients. If controlled studies could show the non-value of prophylactic hyperventilation, why can’t collaborative outcome studies determine the value of ICP monitors and other therapies, starting with patients at Glasgow coma scales (GCS) of 6-8 and, later, proceeding to evaluate patients at worse GCS levels? Other potentially important but well-controlled studies might reflect additional ways to improve long-term knowledge about the futures of these unfortunate persons and the relationship of their success or failure to specific areas of brain damage. —fp