Surgery of cerebral trauma and associated critical care


Published on

Published in: Health & Medicine
  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Surgery of cerebral trauma and associated critical care

  1. 1. TRAUMA SURGERY OF CEREBRAL TRAUMA AND ASSOCIATED CRITICAL CAREAlex B. Valadka, M.D. The last 30 years have been both exciting and frustrating for those in the field of trau-Department of Neurosurgery, matic brain injury (TBI). Much has been learned, but no new treatment has been shownUniversity of Texas MedicalSchool at Houston, to improve patient outcomes despite the execution of many clinical trials. The overallHouston, Texas incidence of TBI has decreased, probably because of intensive efforts toward preven- tion and education. Rigorous assessment of available research has produced severalClaudia S. Robertson, M.D. evidence-based guidelines for the management of neurotrauma patients. The creationDepartment of Neurosurgery, of organized emergency medical services systems in many regions has improved pre-Baylor College of Medicine, hospital care. Computed tomographic scans have become the gold standard for obtain-Houston, Texas ing immediate images of patients with TBI, and ongoing advances in visualizing cere-Reprint requests: bral metabolism continue to be remarkable. The major current question regardingAlex B. Valadka, M.D., surgical treatment for TBI involves the role of decompressive craniectomy, an opera-Department of Neurosurgery,University of Texas Medical School tion that first fell out of favor and has since (in the last three decades) enjoyed a resur-at Houston, gence of interest. Growing interest in the intensive care management of TBI patients6410 Fannin Street, Suite 1020, helped to establish the new field of neurocritical care. Prophylactic hyperventilation isHouston, TX 77030.Email: no longer recommended, and earlier recommendations for aggressive elevation of blood pressure have been softened to endorsement of a cerebral perfusion pressure of 60Received, December 12, 2006. mmHg. Recombinant factor VIIa is increasingly used for minimizing complicationsAccepted, March 22, 2007. related to coagulopathy. Intracranial pressure monitoring is now recommended for the majority of TBI patients. At present, available technologies allow measurement of other aspects of cerebral metabolism including cerebral blood flow, brain oxygen tension, biochemistry, and electrical activity. Therapeutic interventions that are growing in pop- ularity or are presently under investigation include administration of hypertonic saline, hyperoxygenation, decompressive craniectomy, and hypothermia. Rehabilitation has become accepted as an important part of the TBI recovery process, and additional work is needed to identify optimal interventions in this area. Socioeconomic factors will play a growing role in our treatment of TBI patients. Although much progress has been made in the last 30 years, the challenge now is to find ways to translate that progress into improved care and outcomes for TBI patients. KEY WORDS: Computed tomography, Guidelines, Neurocritical care, Neurotrauma, Prehospital, Traumatic brain injury Neurosurgery 61[SHC Suppl 1]:SHC-203–SHC-221, 2007 DOI: 10.1227/01.NEU.0000255497.26440.01S urgical treatment of head injuries was almost certainly the greatly improved our understanding of posttraumatic cerebral first type of neurosurgical procedure performed by ancient pathophysiology (Fig. 1). At the same time, these new insights peoples several millennia ago. Even as recently as last cen- have failed to make the transition to clinically useful therapies.tury, as neurosurgery was rapidly growing as a specialty and as Many of the major clinical trials of the last decades have beennew practitioners began to settle in more and more cities across “negative” studies that have shown us what does not work. Itthe United States, trauma patients continued to be a major part has been extraordinarily difficult to demonstrate the efficacy ofof the clinical volume of many of these newly established prac- new treatments.tices. Identifying new treatments and proving their utility have The last three decades have been alternately exhilarating and been seemingly insurmountable obstacles. Progress has beenfrustrating for those with clinical and research interests in trau- made, however, in other arenas. Careful reevaluation of exist-matic brain injury (TBI). Laboratory and bedside research has ing data and publications can help in identifying which prac-NEUROSURGERY VOLUME 61 | NUMBER 1 | JULY 2007 SUPPLEMENT | SHC-203
  2. 2. VALADKA AND ROBERTSON wounds to the head (GSWH) (59). Age Another recent develop- ment has been the apparent increase in brain injuries among the elderly. Some of this increase is thought to reflect the growing number of elderly citizens in the pop- ulation at large. Another likely contributing factor is the concurrent growth in the number of elderly patients taking anticoagulant or antiplatelet drugs. Outcome Some authors report a sig- nificant decrease in mortality rates attributed to TBI over the last few decades (33). However, such reports often FIGURE 1. Findings from a patient with fatal brain injury. A, computed tomographic scans were obtained at hospi- exclude the most severely tal admission and 12 hours later, showing local cerebral ischemia progressing to global cerebral ischemia. Location of injured patients or those in brain tissue oxygen probe is shown (yellow circle). B, graphs showing progressive increase in ICP, decrease in brain whom a significant amount of oxygen level, and decreases in glucose and pyruvate levels in cerebral microdialysate, with a rapid increase in lactate and a later increase in glutamate levels. time passed between injury and arrival at a hospital (19, 31, 38, 43, 71). Such exclusionstices are supported by the best available evidence and which are understandable because those studies were often designedcan be discarded or at least placed in an appropriate context to evaluate the efficacy of potential new therapies, which arethat recognizes their limitations. Finally, emergency neurosur- unlikely to be effective in patients with devastating trauma. Atgical care, like emergency care in general, has benefited from the same time, however, the outcomes in these reports cannotsustained efforts at systems planning, integration, and cooper- be viewed as representative of the outcomes that would beation. Prehospital and interhospital coordination are much expected when all patients with TBI are included. An addi-more effective when protocols are established before they are tional concern about some studies is that the only outcomeneeded, and are far better than the alternative of waiting until captured is survival versus death at hospital discharge.a patient deteriorates before deciding where to send and how Detailed long-term outcome assessments by trained, blindedto treat him or her. examiners are generally not performed. These successes and disappointments are summarized in the The Traumatic Coma Data Bank (TCDB) was created whenfollowing sections. They begin with discussions of epidemiol- four academic centers with interest and expertise in TBI partic-ogy, prevention, and guidelines, and continue with a sequence ipated in a data collection project that began almost 30 yearsof steps that a trauma patient would experience as he or she ago. At that time, prehospital care systems and improvementsprogresses through the emergency care system, including ini- in emergency medical systems had evolved sufficiently totial prehospital and emergency department care, imaging, sur- allow rapid transport of injured patients to hospitals.gery, intensive care, and rehabilitation. Thus, it seems reasonable to use TCDB data as a reliable starting point for discussions of outcome after TBI (37). Like most such studies, however, these data suffer from at least two EPIDEMIOLOGY limitations. The first is that the data were gathered from aca- demic trauma centers, which may not necessarily have compa-Mechanism of Injury rable outcomes to nonacademic facilities or to facilities that are A welcome change in the epidemiology and demographics of not designated trauma centers. The second is that patientsTBI has been the steady decrease in incidence of severe closed- enrolled in clinical trials—even the placebo group—tend tohead injury in recent years. Unfortunately, this decrease has have better outcomes than other patients because of the extrabeen paralleled by an increase in the numbers of gunshot attention that is given to all patients in a study. For these rea-SHC-204 | VOLUME 61 | NUMBER 1 | JULY 2007 SUPPLEMENT
  3. 3. SURGERY OF CEREBRAL TRAUMA AND CRITICAL CAREsons, TCDB data may be unduly optimistic if they are used to Thoughtful analysis of these injuries is a science that is still inanticipate the likelihoods of different outcomes in patients who its infancy. In sports such as football, neurosurgeons haveare not enrolled in clinical trials, who are treated at nonacade- assumed the lead in modifying rules of play to increase safetymic facilities, or who are managed at a hospital that is not a without diminishing the enjoyment for participants or specta-trauma center. tors (6). Improving the performance of safety equipment, such The TCDB data suggest that death can be expected in roughly as football helmets, is another area of active investigation. The30 to 36% of patients with severe closed TBI. A persistent vege- most pressing questions seem to center around when andtative state will occur in 5% or fewer of these patients. Severe whether an athlete can return to play, and whether repetitivedisability will probably be the outcome for approximately 15% minor brain injuries can have cumulative long-lasting effects. Inof patients; moderate disability can be expected in perhaps 15 to addition to clinical evaluation and neuropsychological testing,20% of patients; and a good outcome will occur in at least 25% computerized assessment tools have been developed to assistof patients. These results are comparable to those of some in answering these questions (36). Functional imaging tech-recently compiled series of unselected patients (55). niques are also under investigation in this area. In patients with GSWH, the outcome distribution is moreheavily skewed toward the extremes (62). Some studies report GUIDELINESprehospital mortality rates of almost 90% for GSWH patients,and patients who are still alive upon arrival at a hospital may The phrase “evidence-based medicine” has seen so muchhave mortality rates of 60% or higher. However, 30 to 40% of use in recent years that it has become a cliché. Basing patientpatients will achieve good outcomes or have only moderate care decisions on high-quality research data has always beendisability. The “intermediate ground” of severe disability or the mainstay of clinical practice. During the last decade,persistent vegetative state occurs less commonly in GSWH attempts to codify and label these practices have been drivenpatients than in those with closed TBI. by several factors, including busy practitioners’ desires to make A major advance over the last few decades has been the sure that they are doing the right thing according to currentlygrowing awareness that a good outcome on the Glasgow accepted practices, as well as cost-cutting attempts to reduceOutcome Scale does not necessarily mean that a patient will unwarranted variation. Of note, TBI was the subject of the firstrecover without any deficits. Problems with judgment, impulse guidelines effort in neurosurgery (8). Several other neuro-control, abstract thinking, short-term memory, and other areas trauma guidelines have been produced subsequently, includingmay be devastating for patients and families, even when a those for pediatric brain injury, penetrating brain injury, surgi-patient outwardly appears to have made a good recovery. cal management of brain injury, and prehospital management of brain injury (1, 4, 13, 26). PREVENTION These documents are quite useful as reviews of existing knowledge and categorization of the strength of evidence sup- Without question, one of the most important developments porting various management practices. For a disease as com-in neurotrauma in the last 30 years or so has been the creation plex as TBI, a common concern about applying the results of aand growth of injury prevention programs. For example, clinical trial is the generalizability of the results to a particularThinkFirst is a program that was created by organized neuro- patient or circumstance. It is often difficult to appreciate thesurgery with the goal of preventing injury via education, nuances and determine when thoughtful deviation from guide-research, and policy. Some data suggest that this educational lines is appropriate. The optimal practice of evidence-basedprogram not only increases children’s knowledge about injury medicine consists of making clinical decisions by integratingprevention, but also reduces their incidence of head and back knowledge of the available evidence with a particular patient’sinjuries (69). Other injury prevention programs, including those circumstances, a physician’s own training and experience, andat the local, regional, and statewide levels, also play important the setting in which the care is being provided. Guidelines pro-roles in injury prevention. vide an excellent place to start setting treatment goals and for- Because automobile accidents are a major cause of all types of mulating a therapeutic plan.trauma, one would expect that improvements in automobile Several retrospective reports using historical controls sug-safety would have a major impact on the incidence of TBI. Air gest that standardization of care through protocols based onbags, compulsory seatbelt laws, lower speed limits, and published guidelines may improve outcomes for TBI patients.improvements in the overall safety and crash tolerance of auto- In some cases, costs are reduced, but other reports describemobiles may all have played a role in reducing the incidence of increased costs if hospitals had not previously devoted manytraffic-related TBI during the last three decades. Greater societal resources to acute TBI management. Also, despite widespreadawareness of the dangers of driving while under the influence of dissemination, the degree to which most hospitals and practi-alcohol or other drugs is probably another major factor. A wealth tioners actually implement guidelines remains unclear.of epidemiological data indicates that motorcycle helmet laws Despite these caveats, the continuing maturation of guide-are associated with lower rates of motorcycle operator fatalities. lines efforts in neurosurgery has been a major development Sports-related neurological injuries comprise a distinct group during the last decade. Many neurosurgeons and hospitalsof traumas that occur at a predetermined time and place. have used this opportunity to review their practices and reviseNEUROSURGERY VOLUME 61 | NUMBER 1 | JULY 2007 SUPPLEMENT | SHC-205
  4. 4. VALADKA AND ROBERTSONthem when appropriate. The goals of these efforts are gency patients. However, the subsequent widespread adoptionimproved efficiency, reduced costs, and, most importantly, bet- of prehospital sedation, paralysis, and intubation of emergencyter outcomes. patients has made it impossible to perform an appropriate neu- rological evaluation on many of these individuals. Good solutions to this problem remain elusive. Reliance PREHOSPITAL CARE upon prehospital providers’ descriptions of the prehospital examination is often the only alternative, but these assess-Organization of Prehospital Care ments may be incomplete or inaccurate. Marshall et al. (39) Thirty years ago, organized networks for prehospital care created a computed tomography (CT)-based classificationwere little more than a promising idea. The subsequent years scheme that categorizes the severity of brain injury. Thehave witnessed definite improvement in the organization of Abbreviated Injury Scale used by general trauma surgeons isprehospital care systems in the United States, and such net- also based on anatomic criteria. Although useful, suchworks now exist in many regions. Mortality is significantly schemes cannot supplant clinical assessments of neurologicalreduced when injured patients receive care at a trauma center function. Instead, they are best thought of as supplementaryas opposed to a hospital that is not a trauma center (34). The sources of information.obvious implication is that organized prehospital triage andreferral systems can improve patient outcomes. However, Airwaymajor organizational gaps continue to affect the optimal pre-hospital care of many TBI patients. The need to secure the airway of a severely brain-injured In 1986, the Emergency Medical Treatment and Labor Act patient has long been a basic management principle. It seems(EMTALA) became law. It was intended to prevent uninsured intuitively obvious that establishing a secure airway will facil-patients from being refused emergency assessment and treat- itate oxygenation, avoid hypercarbia, and prevent aspirationment. As with many well-intentioned government regula- (18, 42).tions, unintended consequences soon appeared. For example, Why, then, have several recent retrospective reviewsa remote hospital that is not a designated trauma center (and reported an increase in mortality in severe TBI patients whowhich does not have immediate access to operating rooms 24 were endotracheally intubated before reaching the hospital?hours per day) could be found to be in violation of EMTALA The answer seems to lie in problems with implementation,if personnel there were to deny a request for an emergency not with the basic principle itself (25). In some regions, para-patient transfer with the reasoning that the patient would be medics and other first responders may perform endotrachealmuch better served at a closer or more appropriate facility, intubation so rarely that the necessary skills cannot be main-such as a level I trauma center. Another unintended situation tained. In other situations, optimal ventilation may not beoccurs when a trauma center becomes filled with patients performed (21).who have relatively minor injuries because the center is statu- It seems safe to conclude that immediate intubation oftorily required to accept all patients in transfer. As a result, the patients with TBI is still an effective method of securing an air-trauma center may be forced to go on ambulance diversion, way, but only if the person performing the intubation is suffi-which makes the facility and its specialty attention unavail- ciently skilled. A better choice for inexperienced respondersable to the very people it was established to serve: severely may be using a bag-valve-mask device or a laryngeal mask air-injured patients who truly require immediate surgery and way to maintain the patient’s ventilation.critical care. Regional interhospital cooperation is an obvious solution. BreathingHowever, the frequently intense competition between hospitals Recommendations for the use of hyperventilation in TBIand the resulting lack of cooperation are common barriers to patients have undergone several changes over the last decade.regional coordination of services. Along with continued revi- Early observations about the effectiveness of hyperventilationsions of EMTALA, these barriers must be overcome if injured in lowering intracranial pressure (ICP) led to its widespreadpatients are to receive optimal prehospital care. The susceptibil- use. The mechanism appears to be a pH-mediated effect of con-ity of the acutely injured brain to secondary insults such as stricting cerebral resistance arterioles which thereby decreaseshypoxia and hypotension, as well as the frequency with which cerebral blood volume.such insults occur, suggests that efficient and coordinated pre- Subsequent laboratory work suggested that the vascularhospital networks are especially important for optimal care of constrictive effect of sustained hyperventilation begins to wearTBI patients. off within a few hours (45). Moreover, once a low CO2 partial pressure (PaCO2) has been maintained for more than a fewNeurological Assessment hours, any attempts to raise the PaCO2 back to normal may The first publication of Plum and Posner’s classic mono- cause the blood vessels to dilate and thereby increase bloodgraph in 1966 (50) and the description of the Glasgow Coma volume and ICP.Scale by Teasdale and Jennett in 1974 (60) had a major influence Eventually, clinical data demonstrated worse outcomes inon the standardization of neurological assessment of emer- severe TBI patients who had been routinely treated with hyper-SHC-206 | VOLUME 61 | NUMBER 1 | JULY 2007 SUPPLEMENT
  5. 5. SURGERY OF CEREBRAL TRAUMA AND CRITICAL CAREventilation (44). Thus, routine use of prophylactic hyperventi- pressures (5, 23). This protocol is based in part on the assump-lation is not recommended. tion that the hydrostatic pressures associated with elevated However, there remain at least two situations in which blood pressure may promote cerebral edema through passagehyperventilation may be appropriate. The first is as a latter- of water from the circulation to the brain, whereas increasingstage treatment for refractory intracranial hypertension. blood oncotic pressure and increasing precapillary vasocon-Protocols vary as to the exact circumstances at which this ther- striction may facilitate diffusion of water from the brain to theapy should be instituted, but most agree that it should only fol- vasculature (Fig. 2). Again, however, support for these con-low the failure of some other therapies, such as sedation, phar- cepts is based only on uncontrolled case series.macological paralysis, CSF drainage, and/or osmolar therapy. Present thinking suggests that there probably exists a floorIf advanced neurological monitoring capabilities are available, below which CPP should not decrease. However, efforts tocerebral oxygen metabolism can be tracked to ensure that increase CPP above this floor in all patients may increase com-hyperventilation does not cause a reduction in oxygen delivery plications without conferring any added benefit. A minimumto the brain. Similarly, if cerebral blood flow (CBF) monitoring CPP of 60 mmHg is probably reasonable for most patients. Asindicates that blood flow to the brain is normal or even ele- with all such recommendations, however, individual patientsvated, then mild hyperventilation may be implemented early in may fare better with either higher or lower CPPs. Ideally,the management of intracranial hypertension. detailed cerebral monitoring could be used to identify those Another indication for hyperventilation is an acute neurolog- patients who require a higher CPP and, conversely, those whoical deterioration that is known or suspected to be caused by a would not need to be subjected to the potential risks of aggres-large intracranial mass lesion. In theory, a brief period of hyper- sive efforts to increase CPP to an arbitrary level.ventilation may lower ICP long enough to allow sufficient time Of note, most of these studies were conducted in intensiveto transport a patient to an operating room for evacuation of a care units (ICUs). It is reasonable to extrapolate them to the pre-hematoma or contusion. The negative effects of a potential tem- hospital setting, but the reader should remember that specificporary compromise of CBF from vasoconstriction may be out- treatment goals may have to be adapted to that setting.weighed by the benefits of reducing ICP. In these patients, oncea clot is removed, ventilation returns to normal. In patients IMAGINGwithout focal signs suggestive of the presence of a mass lesion(such as a significantly asymmetric motor exam or asymmetry A generation ago, CT was still a new technology. Publishedof pupillary response), there is less indication to initiate imme- reports continued to explore the utility of this new imagingdiate hyperventilation. Similarly, if a patient who has been technique in brain-injured patients. Angiography and cranialhyperventilated as a preliminary treatment for a suspected radiography were beginning to fade from the scene as routineacute hematoma is found to have no such lesion upon per- tools in the evaluation of trauma patients. Magnetic resonanceformance of CT, hyperventilation is usually best stopped while imaging had not yet become available.other treatments with more acceptable risk-benefit ratios areimplemented.Circulation Similar to the breathing and ventilation strategies for treatingTBI patients, the general consensus on blood pressure goalshas undergone several changes during the last three decades.The 1980s and 1990s witnessed widespread dissemination ofthe belief that elevating blood pressure to maintain a meancerebral perfusion pressure (CPP) of at least 70 mmHgimproved outcome for patients with severe TBI. Anecdotes andcase series supported this approach, and it seemed to be consis-tent with the concept that cerebral hypoperfusion caused bylow blood pressure is a major source of secondary brain injury(53, 54). FIGURE 2. According to the Starling hypothesis, Subsequent reports, however, including a prospective, ran- fluid exchange (JV) between a capillary and the inter-domized, controlled trial, demonstrated no improvement in stitial space is determined by hydrostatic pressurepatient outcome as a result of this practice (32, 52). Robertson within the capillary (Pc), hydrostatic pressure in the interstitial space (Pi), oncotic pressure within the cap-et al. (52) found that this treatment strategy did seem to illary (πc), and oncotic pressure within the interstitialdecrease the incidence of cerebral ischemia, but it did so at the space (πi). K is a constant. Important parts of the Lundprice of increased pulmonary complications. The overall result strategy for management of patients with severe TBIfor patients was no gain. include manipulation of these parameters to attempt to Meanwhile, a management protocol from Lund, Sweden, minimize cerebral edema.called for severe TBI patients to be maintained at lower bloodNEUROSURGERY VOLUME 61 | NUMBER 1 | JULY 2007 SUPPLEMENT | SHC-207
  6. 6. VALADKA AND ROBERTSON Not too many years ago, performing emergency CT was atime-consuming affair. The neurosurgeons accompanying the A Cpatient had plenty of time to write their admission history andphysical and their admitting orders while in the scanner.Neurosurgery residents who had some knowledge of computerprogramming could learn to operate the scanner. Debates in theliterature and at meetings argued about the value of a single-slice computed tomographic scan through the center of thehead as a way to identify mass lesions quickly, without theneed to wait for the entire scan to be completed. Subsequent advances in CT technology have been nothingshort of remarkable. Scans of the brain can now be obtained ina matter of seconds. Touch-screen technology has made it pos-sible for virtually any physician to perform CT with only min-imal training. Image quality has improved greatly. The need toprint images on film has disappeared, as the scans are accessi-ble from any place that has Internet connectivity. A key development in medical imaging has been our abilityto visualize not only anatomic structure, but also function. CT,which remains the imaging modality of choice for patients withacute TBI, can be adapted to provide information about CBF, Bperfusion, and vascular anatomy, even in emergency settings(Fig. 3). Magnetic resonance imaging and magnetoencephalog-raphy can reveal selective activation of specific brain regions.Although positron emission tomography scanning remains themost powerful tool for acute study of CBF and metabolismafter injury, as yet its application is not as widespread as that ofCT and magnetic resonance imaging. A common problem, however, is the need to transport criti-cally ill patients to the radiology department to perform suchstudies. Patient transport has been reported by many authors tobe associated with an increase in potentially adverse events.Portable CT equipment has been available for several years, butonly recently have technological advances made these devicesmore user friendly. Important features include helical scanningcapability, low radiation exposure, wireless links to a hospital’simaging network, operation via a touch screen, ability to run onbattery power, and ability to perform perfusion and xenon-enhanced CT studies (48). Most importantly, this technology FIGURE 3. Two examples of stable xenon-enhanced computed tomo-avoids the need to transport patients to the radiology depart- graphic scans. A, scan shows left temporal ischemic area, but flows are ele-ment. Future applications may include placement of these scan- vated elsewhere in the brain. B, scan shows reduced flow throughout theners in emergency departments and even in certain types of brain, with large ischemic areas in the right temporal and frontal areas andambulances. smaller ischemic areas elsewhere (left). C, scale bar for CBF (units are mil- liliters per 100 grams per minute) is shown. SURGERY Basic principles of surgical management have not changed torn between performing an operation that may be unneces-much in the last generation. As articulated clearly by Becker sary versus having a patient undergo neurological deteriora-and others, they include prompt evacuation of contusions and tion if surgery is delayed too long.other mass lesions and use of large craniotomy openings for the Several courses of action are possible in such situations.evacuation of acute subdural hematomas (ASDHs). These include waiting a few hours to allow alcohol or other drugs to be metabolized; obtaining a follow-up computedIndications tomographic scan within several hours to determine whether Some of the most difficult decisions we make in neurotrauma the lesions have enlarged; and/or inserting a ventriculostomycare involve whether moderate-sized hematomas or contusions or parenchymal ICP monitor in search of elevated ICP, whichshould be evacuated or simply observed. A surgeon may feel might prompt an immediate craniotomy.SHC-208 | VOLUME 61 | NUMBER 1 | JULY 2007 SUPPLEMENT
  7. 7. SURGERY OF CEREBRAL TRAUMA AND CRITICAL CARE These decisions are often based on an individual physician’s be time consuming in some hospitals, and procurement ofjudgment and experience. It is often preferable to remove size- recombinant factor VIIa may require complex bureaucratic pro-able lesions early, before a patient’s condition deteriorates. cedures because of its cost. Neurologically compromised yetPrompt surgery also minimizes subsequent parenchymal treatable patients may not be able to tolerate such delays. Weedema around a contusion. Furthermore, a global measure have sometimes enjoyed success with taking such patientssuch as ICP may be normal even while CBF and metabolism promptly to surgery while the hospital’s blood bank initiatesare focally impaired near an acute traumatic lesion. the process of delivering fresh frozen plasma and/or platelets Recently published evidence-based guidelines provide some to the operating room. Blood products generally arrive in thedirection (10–12, 14, 15). Patients with Glasgow Coma Scale operating room while the neurosurgeon is in the midst of thescores of 8 or less with large mass lesions should undergo surgery, and their effect on clotting is usually immediate andprompt lesion evacuation, especially if the results of their neu- obvious. Although laboratory tests of hemostatic function arerological examinations reveal deterioration, if their pupils helpful, the surgeon can gain useful information simply byexhibit anisocoria or are fixed and dilated, or if the lesions are watching how the blood is clotting in the surgical field.causing focal neurological deficits. A midline shift of 5 mm or Recombinant factor VIIa has received a great deal of atten-more and effacement of the basal cisterns are commonly used tion recently for treatment of coagulopathy in trauma patientscomputed tomographic indicators of significant mass effect. (22). An important neurosurgical application is the immediateRegardless of the patient’s Glasgow Coma Scale score, evacu- treatment of warfarin-associated intracranial hemorrhage (9).ation is recommended for epidural hematomas larger than 30 This product may help with diffuse oozing, but brisk bleedingmm3, subdural hematomas greater than 10 mm in thickness or from a large vessel still requires direct treatment via cautery,causing more than 5 mm of midline shift, and parenchymal ligation, tamponade, etc.lesions greater than 50 mm3 in volume. Evacuation of frontal As neurosurgeons’ experience with factor VIIa has increased,and temporal contusions greater than 20 mm3 in size is recom- several questions have been raised. One is the possibility ofmended if the contusions are located frontally or temporally inducing adverse reactions, such as thrombosis of stenotic coro-and are associated with compressed cisterns or a midline shift nary arteries, in susceptible patients. The number of suchof 5 mm or more in patients with Glasgow Coma Scale scores reports will undoubtedly increase as this agent sees greaterof 6 to 8. A smaller size threshold is often used in assessing tem- use. Another concern is that the half-life of factor VIIa is only 3poral lesions because of their potentially greater risk. to 6 hours. Thus, after its period of action has passed, the hem- orrhagic diathesis may recur unless additional treatments areAnticoagulant and Antiplatelet Agents initiated, such as administration of fresh frozen plasma, vita- An increasingly common problem involves the patient who min K, or additional factor VIIa. The cost of the drug is high,develops an intracranial hemorrhage while taking Coumadin, especially for the doses required to treat bleeding in traumaPlavix, and/or aspirin medications (Fig. 4). Some neurosur- patients. To control costs, some hospitals have implementedgeons delay surgery until platelets can be administered or complex administrative-approval algorithms. The significantcoagulation parameters normalized. Vitamin K is usually read- cost also raises important questions about appropriate indica-ily available, but obtaining platelets or fresh frozen plasma can tions, especially in elderly patients with devastating hemor- FIGURE 4. Series of computed tomographic scans showing rapid ative epidural hematoma developed, requiring the patient to return to the enlargement of a contusion, which was surgically evacuated. A postoper- operating room.NEUROSURGERY VOLUME 61 | NUMBER 1 | JULY 2007 SUPPLEMENT | SHC-209
  8. 8. VALADKA AND ROBERTSONrhages that make meaningful recovery unlikely. These issues troublesome area. It is often helpful to place small amounts ofare important because such situations will likely occur with an absorbable hemostatic agent, such as Gelfoam (Pharmacia/increasing frequency in the future as the population ages and Upjohn; Kalamazoo, MI) or Surgicel (Johnson & Johnson;the number of patients taking anticoagulants increases. Arlington, TX) sponges or fibrillar collagen, just under the bony edge. These materials provide some extra bulk that may act asSubdural Hematomas tamponades in areas of bleeding. Several maneuvers during opening of the dura may be use- Although it has been reported (57) that reduced mortality ful for avoiding subsequent problems. The first is to open thecan be expected if ASDHs are evacuated within 4 hours, other dura in a slow, controlled manner. Going slowly at this point ofsurgeons have not been able to replicate this effect (70). Some the operation may seem counterintuitive to the need for rapidpatients may fail to improve because of coexisting parenchymal evacuation of a large clot. However, some experienced neuro-brain injury that will not improve after evacuation of a mass traumatologists speculate that a relatively slow dural openinglesion. Furthermore, even rapid progression of a patient may allow a more gradual reduction of the pressure on thethrough the emergency medical services system and to the cerebral cortex. This gradual equilibration has been assumed tooperating room may not be fast enough to reverse damage reduce the likelihood of sudden, massive herniation of braincaused by the mass effect of a clot. Nevertheless, clinical and through the craniotomy opening.laboratory evidence documents the adverse effects of a large Along similar lines, it may be helpful to open the dura in aacute mass lesion on cerebral metabolism (27). Prompt evacu- cruciate manner. Specifically, the four limbs of a cruciate open-ation is not only common sense; it is also consistent with the ing may be made without connecting them at the center of thecommon clinical observation that patients often improve after “X.” This allows evacuation of the clot from all four quadrantsevacuation of large ASDHs. These same arguments apply to of the exposure. If the brain appears to begin swelling slightly,other types of traumatic mass lesions as well. and if intradural bleeding has stopped, these four smaller inci- The use of large craniotomy flaps when performing ASDHs sions can be closed quickly. Alternatively, two of these limbsallows the surgeon to gain access to a variety of potential can be connected to create a small dural flap if access to only asources of bleeding. These include large draining veins near specific part of the opening is desired. Wider access can bethe superior sagittal sinus as well as contused tissue in the gained by connecting all four limbs. If the surgeon wishes tosubtemporal and subfrontal areas and the temporal and close quickly, a single stitch in the center can pull all four duralfrontal poles. A large craniotomy opening for evacuating an leaves together. Although this approach is especially helpful ifASDH has been a standard recommendation for many years, the surgeon is not initially planning to perform a decompres-but it is worth emphasizing in our present era of tiny incisions sive craniectomy, it is completely compatible with a decision tomade possible by image-guidance systems and endoscopic proceed with decompression if unexpected intraoperativeinstrumentation. The medial part of the scalp incision may be events indicate that it might be helpful.made on or near the midline, but it may be wise to keep the Other surgeons attempt to prevent massive brain swelling bybony opening several centimeters off the midline to avoid incising the dura with a reverse “U” incision anteriorly, inferi-arachnoid granulations and large dural and cortical veins orly, and posteriorly; the dura near the midline is left intact (3).near the midline. The midline placement of the scalp incision They report excellent results in terms of evacuating subduralpreserves the option of removing additional bone near the hematomas and parenchymal lesions, with no cases of mas-midline if subsequent uncontrollable bleeding near the supe- sive brain swelling.rior sagittal sinus necessitates more medial exposure. It alsoensures a sufficiently wide base for the scalp flap, which pre-vents necrosis of the posterior and superior edges of the flap Decompressive Craniectomyfrom inadequate blood supply. In the last few decades, interesting swings have occurred in An exception to the general practice of using larger flaps the neurosurgical community’s opinion regarding decompres-may sometimes be found in coagulopathic patients with rela- sive craniectomy. It is clear that decompressive craniectomytively focal subdural hematomas. Some authorities describe can lower ICP in many cases (Fig. 5). However, it is equallysuccessful clot evacuation in such cases without excessive dif- clear that no well-done, randomized, prospective trial has beenficulty in obtaining hemostasis. completed to answer the question of whether decompressive craniectomy truly improves outcomes for all patients, or evenSurgical Technique for certain selected subpopulations (56). Two prospective trials If epidural tack-up sutures are used in an emergency surgery, now in progress may provide important information aboutmany surgeons prefer to wait until closing to place them so as the effectiveness of decompressive craniectomy for TBInot to delay evacuation of a large mass lesion. However, dur- patients (29, 56).ing the opening, brisk epidural bleeding may be encountered Although the potential complications of decompressiveimmediately upon removal of the bone flap, especially near craniectomy are familiar to those who perform these proce-the midline. The only way to achieve control may be by plac- dures, these issues have only recently received more than aing numerous, closely spaced epidural tack-up sutures at the cursory mention in the literature. Potential problems includeSHC-210 | VOLUME 61 | NUMBER 1 | JULY 2007 SUPPLEMENT
  9. 9. SURGERY OF CEREBRAL TRAUMA AND CRITICAL CARE trauma operations. Unless the brain is markedly swollen, this practice may not be necessary in many cases. Mildly protruding brain can often tolerate watertight closure of the overlying dura, even if the brain has to be carefully pro- tected and depressed with a malleable retractor during dural closure. Postoperative ICP is often lower than one might expect, which perhaps serves as clinical corrobora- tion of experimental data that suggest decompressive craniectomy may actually increase edema of the under- lying brain (20). A third area that is less controversial is the size of the decompression. The gen- eral rule is “the bigger, the better.” This must be kept in mind during the planning of the initial craniotomy. If the surgeon anticipates even a FIGURE 5. Preoperative and postoperative conventional and xenon-enhanced computed tomographic scans from a slight possibility of leaving patient who underwent decompressive craniectomy. ICP decreased, CPP and global CBF increased, and regional CBF the bone flap off, it is helpful increased dramatically. to position the patient with the head turned as far later-postoperative epidural/subgaleal hematomas, subdural fluid ally as possible. This may require significant turning of theaccumulation remote from the craniectomy site, brain injury head (if the cervical spine has been cleared) or positioning theat the edges of the craniectomy from impingement during patient in the lateral position (if cervical spine injury is knownoutward brain swelling, lack of brain protection under the or has not been excluded). This extensive turning of the headcraniectomy site, possible impairment of cerebral metabolism allows access to the posterior and inferior regions of the cranialfrom lack of overlying cranium, the need for subsequent sur- vault, which are often not well decompressed even by a stan-gery to replace bone, and potential resorption of the bone dard large trauma flap. Another important point is to take theflap, among others. decompression to the floor of the temporal fossa. Extending the inferior margin of the craniectomy down to the temporalControversies cranial base is recommended for achieving optimal decompres- Several questions exist about the way these operations are sion of the perimesencephalic cisterns (46). This recommenda-performed. The first concerns timing. Some data suggest that tion is consistent with other results indicating that largepatients fare poorly if they receive a decompressive craniec- craniectomies are more effective than smaller ones (31).tomy early in their course of treatment (2). It might be better totreat those patients with medical interventions that are as Techniqueaggressive as possible; if these measures fail to improve out- The importance of performing a sufficiently large decom-come, then surgical decompression seems unlikely to have any pression cannot be emphasized too strongly (Fig. 6). Crani-better results. Other practitioners, however, feel that patients ectomies or dural openings that are too small may causewho arrive in the emergency department with poor neurolog- swollen brain tissue to herniate through the bony defect,ical examinations and diffuse brain swelling as evidenced by thereby causing strangulation, infarction, and creation of addi-CT have little to lose by undergoing prompt decompression. tional brain swelling from inward tracking of the resultant cere- A second controversy involves performing these operations bral edema (Fig. 7).routinely, whenever a patient undergoes a trauma craniotomy. If a patient who has had a recent craniotomy is returnedSome surgeons leave bone flaps off as a routine part of their to the operating room for removal of the bone flap, the sur-NEUROSURGERY VOLUME 61 | NUMBER 1 | JULY 2007 SUPPLEMENT | SHC-211
  10. 10. VALADKA AND ROBERTSON INTENSIVE CARE UNIT MANAGEMENT Evolution of Neurocritical Care The field of neurocritical care is rapidly becoming a distinct discipline. Traditionally, critical care medicine has been domi- nated by a focus on the heart and lungs. More recently, improved knowledge of central nervous system pathophysiol- ogy, as well as the availability of therapies for previously untreatable conditions such as ischemic stroke, have driven interest in the management of these patients within ICUs. The Neurocritical Care Society was created only a few years ago as a multidisciplinary group with the goal of improving the care FIGURE 6. Bifrontal decompres- FIGURE 7. Contrast-enhanced of patients with life-threatening neurological illnesses (47). sion. The frontal craniectomy computed tomographic scan of a In the last few decades, our knowledge and technology per- could have been extended inferi- patient in whom a bony decom- taining to ICUs have mushroomed (Fig. 8). Unfortunately, this orly, but the surgeon deliberately pression was performed, but the wealth of information sometimes makes it difficult to filter chose not to do so to avoid the dura was not opened widely. important bits of data from the background noise. Knowledge frontal sinus. Necrotic brain tissue can be seen does not necessarily equal wisdom. More information in and of herniating through the defect. itself is not automatically associated with a coordinated direc- tion of patient care, and in fact, it may cause distraction andgeon may wish to use a “T” incision to facilitate removal of unnecessary confusion. An intensivist who concentrates on dis-more bone. In general, these “T” incisions begin at the pos- eases of the nervous system may be uniquely able to balanceterosuperior part of the existing incision and curve gently management of the nervous system with cardiac, pulmonary,toward the inion. The scalp on the two sides of the “T” can renal, and other critical care reflected to expose additional bone medially, posteriorly, The growth of knowledge, procedures, and diagnostic testsand inferiorly. This bone can be removed as a separate piece available to intensivists has been intimidating for some neuro-that can be secured to the original bone flap at the time of surgeons. However, a neurosurgeon’s unique expertise andreplantation. perspective on the management of his or her patients are more The closing of a decompressive craniectomy is the ideal time than adequate to offset a lack of intimate familiarity with cur-to make preparations for subsequent replacement of the bone rent trends in ventilator or pressor management. The latterflap. Most surgeons perform a duraplasty or dural augmenta- issues can be learned, but a neurosurgeon’s judgment and accu-tion as part of these procedures. It may be wise to lay another mulated wisdom cannot be duplicated by other physicians.sheet of dural substitute over the entire area of exposed dura. Neurosurgeons should be confident that they bring a uniqueThis maneuver prevents adhesions from forming between the and much-needed perspective to the management of theirdura and the scalp. These adhesions can make it difficult to patients, no matter how sick those patients may be. After all,expose the craniectomy site during subsequent surgery for most intensivists will not see patients several weeks or monthsbone flap replacement. A similar precaution may be used to after they leave the ICU, but the neurosurgeon may be in fre-minimize scarring and adhesion of the temporalis muscle to quent contact with such patients and their families for years.the scalp. Careful attention should be paid to the handling This type of long-term feedback is invaluable as a means forand preservation of the temporalis to optimize long-term cos- reevaluating one’s judgment and management practices.metic results. The critical care community has benefited from several arti- In general, bone flaps should be replaced as soon as possible. cles and abstracts that report improved care when ICUs areThe determining factor may be the persistence of cerebral administered by board-certified intensivists as opposed to ICUsedema, which may require weeks to subside sufficiently to that are run by other physicians (64). However, some of theseallow cranioplasty. For other patients, bone flap replacement studies are methodologically suspect, and the anticipatedmay be delayed by the presence of intradural CSF collection, advantages of hiring intensive care physicians to manage ICUswhich sometimes causes protrusion of the brain through the may not always materialize. For example, it is sometimescraniectomy defect. Extra-axial CSF collections along the mid- assumed that the presence of an intensivist will lead to greaterline or at the craniectomy site may respond to CSF drainage via efficiency and thus decreased costs of ICU care. However, thea lumbar drain inserted the day before surgery. Another option exact opposite may occur, as costs may increase significantly atfor treating patients with hydrocephalus is insertion of a ven- hospitals that have previously not devoted appropriatetriculoperitoneal shunt with a programmable valve. The resist- resources to these types of patients (49). Another uncertainty isance can initially be set low to facilitate CSF drainage. After whether such reported improvements as decreased costs andcranioplasty, the resistance can be set to normal to prevent shorter lengths of ICU stay are direct results of an intensivist’sexcessive CSF drainage from creating an epidural “dead space” presence or whether they instead represent greater institu-under the bone flap. tional adherence to standardized patient management prac-SHC-212 | VOLUME 61 | NUMBER 1 | JULY 2007 SUPPLEMENT
  11. 11. SURGERY OF CEREBRAL TRAUMA AND CRITICAL CARE Another important variable in this system is the degree of neu- rological sophistication possessed by the eICU doctor. Anticipation versus Reaction Physicians are trained to prevent problems and also to antic- ipate them before they occur. Unfortunately, this goal of being proactive has not always resulted in improved patient out- comes. During the last few decades, well-conducted clinical trials have shown that attempts to prevent intracranial hyper- tension by immediate initiation of therapies that are often effec- tive for treating established intracranial hypertension not only do not improve outcome; they may sometimes actually worsen outcome. The list includes hyperventilation, barbiturates, phar- macological paralysis, hypothermia, and artificial elevation of CPP (19, 28, 44, 52, 67). Decompressive craniectomy might also find its way onto that list if a prospective trial is conducted. FIGURE 8. Complex relationships have been discov- Thus, despite our natural inclination to control a situation by ered between CBF, cerebral metabolic rate of oxygen playing offense and initiating treatments before complications (CMRO2), cerebral arteriovenous difference of oxygen get out of hand, it might be best for us to sit back and play content (AVDO 2 ), hyperemia, hypoperfusion, defense. The best that we can do may be to react immediately ischemia, and infarction. (From, Robertson CS, when circumstances change. Prompt treatment of established Narayan RK, Gokaslan ZL, Pahwa R, Grossman RG, problems may be better for our patients than initiation of treat- Caram P Jr, Allen E: Cerebral arteriovenous oxygen ments for complications that have yet to happen. Of course, difference as an estimate of cerebral blood flow in intensive efforts continue in the search for interventions that comatose patients. J Neurosurg 70:222–230, 1989). are truly effective when given prophylactically.tices, with a resultant reduction in the sometimes expensive Tracheostomy and Ventilator Weaningidiosyncratic variation that exists among practitioners. If the Management of the ABCs (airway, breathing, and circula-latter is true, then the presence of an intensivist per se may not tion) was discussed in the context of prehospital care, but mostbe as important as the commitment of a hospital and medical of our understanding of these topics comes from research con-staff to reevaluate and refine existing patient management ducted in ICUs. An additional issue that arises in ICUs sur-practices as necessary. rounds the timing of tracheostomy. Many trauma surgeons These external pressures are presently causing difficulty for and intensivists push for early tracheostomy. Many of theirsome neurosurgeons who have served for years as directors of patients, however, have significant pulmonary problems andtheir local neurosurgery ICUs. In some cases, they are being do not suffer from brain injuries. It is usually best not to per-displaced as medical directors and even prevented from serv- form any procedures on patients with acute and severe TBIing as primary physicians for their patients while the patients unless those procedures are essential. Furthermore, TBIare in the ICU. patients will often remain in ICUs longer than other patients, Another practical issue facing ICUs relates to available per- because they require intracranial monitoring. Thus, perform-sonnel. There are at present far too few intensivists to fill cur- ing early tracheostomy may not significantly shorten therent and projected needs. It remains unclear how recommenda- length of ICU stay in these patients.tions to increase intensivist presence in ICUs will be reconciled Patients who appear to be “waking up” may be given somewith the relatively small supply of intensivists available. One extra time to see if they can be given a trial of extubation. Thepotential solution is an electronic ICU, or “eICU.” In this sys- traditional teaching that a tracheostomy should be performedtem, each bed from several different ICUs is electronically within 2 weeks is based more on dogma than fact. On the otherlinked to a central area, which may be geographically remote, hand, if it appears that a patient will likely not experiencethat has continuously available bedside monitoring data from enough neurological recovery to protect his or her airway, it iseach linked patient. Thus, a nurse or physician in this central reasonable to proceed expeditiously with a tracheostomy oncearea can simultaneously monitor many patients in different the patient’s cerebral metabolism appears to have stabilized.ICUs. Each bed may also have a dedicated video camera,microphone, and speaker, and progress notes and orders are Monitoringavailable via fax or electronic medical record. Different institu- Although detailed monitoring of cerebral pathophysiology istions report varying degrees of success and satisfaction with not yet performed at many hospitals that treat TBI patients,this model. Obviously, this system cannot handle emergencies such monitoring represents the best way for evaluating a par-that require a physician’s presence at the bedside, such as intu- ticular patient’s metabolic pattern and, when necessary, forbations, central venous access, chest-tube placement, etc. intervening in an appropriate manner (Fig. 9).NEUROSURGERY VOLUME 61 | NUMBER 1 | JULY 2007 SUPPLEMENT | SHC-213
  12. 12. VALADKA AND ROBERTSON A concept that is gaining increasing recognition is theimportance of regional heterogeneity of cerebral metabolismand the difference between regional and global measures ofcerebral metabolism. Even relatively large areas of focal abnor-mality may not affect a global measure of cerebral metabo-lism, such as jugular venous oxygen saturation. Similarly, afocal monitor such as a brain-tissue oxygen sensor may notreveal the presence of a large abnormal area if the focal mon-itor lies within normal tissue. Those who use cerebral monitor-ing devices must be aware of their limitations as well as theirpotential usefulness. Continuous electroencephalographic monitoring has beenreported to detect seizures in more than 20% of patients withmoderate and severe TBI during the first 2 weeks postinjury(66). These results suggest that adverse electrophysiologicalevents may often be missed and may represent an importantand underappreciated mechanism of secondary brain injury. FIGURE 9. A patient undergoing many different types of cerebral andIntracranial Pressure metabolic monitoring. The most widely used monitoring techniques measure ICP.Parenchymal devices have become popular over the last fewdecades, but ventriculostomy remains the recommended form oxygen administered via ventilator, transfusion of packed redof ICP monitoring technology. The interested reader is referred cells, and reduction of ICP via evacuation of mass lesions orto the excellent summaries contained in Guidelines for the other interventions. Although it is difficult to conduct high-Management of Severe Traumatic Brain Injury (7), which reviews quality studies that demonstrate improvement in patient out-both the different types of ICP monitoring technology and the come as a result of PbtO2 monitoring, many neurosurgeonsbasis for using 20 mmHg as the threshold value for treating have experience with at least a few patients for whom thispatients with elevated ICP. seemed to be the case.Brain Tissue Oxygenation Cerebral Blood Flow For many years, measurement of cerebral oxygen metabo- The role that CBF plays in influencing outcome from TBI haslism could be performed only via intermittently sampling been intensely scrutinized for many years (Fig. 12). The collec-blood from the jugular bulb and comparing its oxygen content tive efforts of many investigators over many years indicate thatto that of arterial blood. Subsequently, jugular venous oximetry CBF passes through several changes after a severe TBI. It isallowed for continuous measurement of the oxygen saturation often quite low during the first few hours after injury. After aof blood flowing out of the brain (58). few hours or days, CBF subsequently increases, often to supra- Although direct monitoring of brain-tissue oxygenation was normal levels. It then gradually decreases and may even passnot even a dream 30 years ago, it is presently a widespread through a phase of vasospasm before it finally normalizes (40).(and increasingly used) monitoring technique. New publica- Xenon-enhanced CT is a powerful technique for obtainingtions and ongoing discussions continue to inform us of the both global and regional quantitative CBF measurements, butproper interpretation of these data, as well as treatment thresh- it is presently not approved by the United States Food andolds and optimal interventions for increasing brain-tissue oxy- Drug Administration. It is hoped that such approval will begen tension (PbtO2). A PbtO2 reading of 10 mmHg probably forthcoming. Parenchymal or surface monitors use a variety ofrepresents a minimally acceptable value, although some would techniques to assess CBF, such as laser Doppler flowmetry,prefer a higher threshold, such as 15 mmHg or even 20 mmHg. thermal diffusion, and others. These are useful technologies, Although placement of these monitors in uninjured brain but many of them are invasive. Transcranial Doppler sonogra-has been recommended, others have suggested that a more phy is a helpful measure of CBF velocity through the largeappropriate strategy is placement in brain that is at risk because arteries of the circle of Willis. Its limitations include a highof adjacent contusions, hematomas, or infarcts (Figs. 10 and degree of dependence on the operator’s skill, and a lack of11). Some neurosurgeons have taken these monitors to the information about regional metabolic activity.operating room to monitor tissue oxygenation during cere-brovascular procedures or other operations that may require Microdialysistemporary vessel occlusion. Cerebral microdialysis is an invasive Food and Drug Depending upon an individual patient’s situation, appropri- Administration-approved method of measuring biochemicalate ways to address a low PbtO2 value might include elevation changes in brain tissue (Figs. 13 and 14). For example, manyof arterial blood pressure, increase in the fraction of inspired investigators have reported that patients experience increases inSHC-214 | VOLUME 61 | NUMBER 1 | JULY 2007 SUPPLEMENT
  13. 13. SURGERY OF CEREBRAL TRAUMA AND CRITICAL CARE include the logistics of organ- izing, processing, and storing the many samples; for exam- ple, collecting dialysate every 30 minutes for 5 days gener- ates 240 samples that must be tracked. These can be frozen and stored for later analysis. In general, most ICUs that perform cerebral microdialy- sis use the information not as a primary monitoring tool, but rather as a supplemental source of information that corroborates the impression provided by other monitors and may occasionally serve as an early warning that something is beginning to go wrong. Future applications could include detection of intraparenchymal drug lev- els, indirect measurement of activity of stem cells or cloned genes, and targeted delivery of therapeutic agents. FIGURE 10. PbtO2 catheter (yellow arrow) was placed in a region that appeared to be uninjured. PbtO2 values are Treatment often low immediately after injury, but then begin to increase. When placed in uninjured tissue, these monitors reflect From one perspective, lit- global secondary insults. tle has changed in our man- agement of patients with TBI over the last generation. The most important parts of manage- ment remain prompt detection and immediate correction of secondary insults. Our present approach to the treatment of patients with elevated ICP uses some of the same tools that we used 30 years ago, such as initiation of sedation, induction of pharmacological paralysis, administration of mannitol, drainage of CSF, hyperventilation, and induction of barbiturate coma. Despite intensive and very expensive efforts, no “magic bullet” has been discovered for the “cure” of brain injury. Although this situation may sound bleak, nothing could be further from the truth. A great deal has been learned over the last three decades. Laboratory studies have deepened our understanding of the cellular and molecular events that follow injury. Appropriate preclinical testing has become even more recognized as an essential step to take before new therapies can be brought to clinical trials. Important principles of clinical trial design, execution, and analysis have been identified and FIGURE 11. Placement of a PbtO2 catheter near a contused area. PbtO2 accepted, and these will affect the design of future studies. level decreases as the contusion enlarges. Such focal metabolic changes may Also, the shrinking availability of healthcare dollars in both not be detected by global monitors of cerebral metabolism such as ICP or the clinical and research arenas has led to greater awareness of SjvO2 monitoring. possible conflicts of interest and establishment of appropriate guidelines for interaction of clinical and research physicianslactate, excitatory amino acids, and glycerol and decreases in with industry, while simultaneously emphasizing the impor-glucose and pyruvate during periods of metabolic stress. tant role that industry funding plays in moving forward withTechnical difficulties associated with performing microdialysis new technologies.NEUROSURGERY VOLUME 61 | NUMBER 1 | JULY 2007 SUPPLEMENT | SHC-215
  14. 14. VALADKA AND ROBERTSON Although we have gained a deeper understanding of the precise roles for many new potential treatments, we have also eliminated a few things that were found not to work. Steroids are no longer recommended as part of the treatment of patients with TBI. Similarly, prophylactic hyperventilation is not rec- ommended. The practice of deliberately dehydrating patients to prevent brain swelling has been replaced by an emphasis on maintaining normal intravascular volumes. On the other hand, we have also verified the benefit of some practices, such as administration of prophylactic anticonvulsants during the first week after a patient experiences TBI. Perhaps most importantly, it has become clear over the last few decades that different TBI patients may vary in their patho- FIGURE 12. Relationships between patient outcome, physiological profiles. Three patients who were riding in the cerebral blood flow, and cerebral metabolic rate of oxy- same automobile during a crash can have very different types of gen (CMRO2). (From, Robertson CS, Contant CF, head injury. One may have severe generalized cerebral edema, Gokaslan ZL, Narayan RK, Grossman RG: Cerebral another may have primarily diffuse axonal injury with low ICP, blood flow, arteriovenous oxygen difference, and out- come in head injured patients. J Neurol Neurosurg and a third may have a large mass lesion that requires immedi- Psychiatry 55:594–603, 1992. Reproduced with per- ate surgery. The metabolic picture of the same patient may even mission from the BMJ Publishing Group.) fluctuate from day to day or hour to hour. Ideally, these patients might benefit from different approaches to treatment and from ongoing reassessment and changes in the therapeu- tic plan. However, the sophis- ticated monitoring tech- niques that can guide such physiologically targeted therapy are available in only a few hospitals. We hope that these techniques will spread to more facilities and thereby add to our knowl- edge of how to tailor man- agement to an individual patient’s pathophysiological profile. In the near future, treatment decisions may also be based on a patient’s geno- typically determined antici- pated responses to specific interventions. Hypertonic Saline Recent years have wit- nessed a growing use of interventions that continue to be investigated and debated. Many of these are older treatments that have undergone a resurgence of interest. Hypertonic saline is FIGURE 13. Example of microdialysate changes during global cerebral ischemia. As cerebral perfusion pressure and receiving increasing atten- PbtO2 values decrease, pyruvate and glucose levels also decrease, whereas increases are seen in lactate and glutamate tion for its role in the treat- levels and the lactate-pyruvate ratio. (From, Hlatky R, Valadka AB, Goodman JC, Robertson CS: Patterns of cerebral ment of patients (especially energy substrates during ischemia measured in the brain by microdialysis. J Neurotrauma 21:894–906, 2004.) children) with elevated ICP.SHC-216 | VOLUME 61 | NUMBER 1 | JULY 2007 SUPPLEMENT