NEUROSURGERY

By K. Gireesh

RationaleforImmunotherapy Experimental and Clinical Gene Therapies Neuroprotection in Brain and Spinal Cord Injury Neuroprotection in Subarachnoid Hemorrhage Carpal Tunnel Syndrome Craniosynostosis Intracranial Pressure Monitoring : Operative Procedure Surgery for Movement Disorders

Rationale for Immunotherapy of Brain Glioma

There has been no change in the management of primary malignant tumors of the brain (Glioblastoma Multiforme and anaplastic astrocytomas – astrocytomas Grades 3 and 4) for 20 years.

Radiation therapy can double life expectancy, and chemotherapy may add a few more months to median survival. The modern use of Gamma Knife adjuvant radiosurgery has further pushed the survival curve with it’s high response rate, but still there is no cure for this dreaded disease. Survival may be only 1 year from the time of diagnosis, with fewer than 5% loving 5 years. Thus, a new approach had to be undertaken.

Brain tumor cells, being foreign to the body, should be able to be managed by our normal immunologic defenses as they would if a bacterial or viral infection occurred. Furthermore, there should be a "memory" of the foreign pathogen such that if the immune system came in contact with a tumor cell again, it could recognize and destroy it much like a second exposure to chicken pox.

Why doesn’t this occur with primary malignant tumors of the brain? Could it be that brain tumor have evolved a way to evade our own defenses and produce a "wall" of immunosuppression. The cure for this cancer may well lie in the breakdown of this wall, or indeed the overwhelming of this wall.

Because of the infiltrative nature of

Grade 4 astrocytomas versus the
compact nature of Grade 1 astrocytomas,
and the presence of their hypothetical
"immunosuppressive barrier", the cure
for this tumor must be a biological
one such as immunotherapy.

Approximately 8 years ago, Dr Gale (Morrie) Granger experimented with Interleukin-2 (IL-2) activated lymphocytes and their in vivo effect on human brain tumors. Some results were excellent, but toxicities could be high. IL-2 is a "cytokine" – a type of chemical messenger which can enhance the immune response toward a foreign antigen (such as a tumor cell).

It was discovered that many cytokines may be automatically produced when too unlike, foreign, white blood cells come in contact. Thus, mixing the patients white blood cells with those of an unrelated donor, and allowing them to incubate in the lab for three days causes enhanced production of at least 7 different cytokines.

"Mixture of host and donor
lymphocytes causing the production
of cytokines and the targeting of a tumor cell".

If theses are placed in the tumor bed at the time of a tumor resection, presumably this "angry" white cell mixture can break the immunosuppressive barrier that Glioblastoma Multiforme and Anaplastic Astrocytomas inherently have, and enable the hosts own immune system to "seek and destroy" the abnormal tumor cells. This is called a Mixed Lymphocyte Culture or MLC.

Results of the phase 1 MLC immunotherapy trial 1996 – 1998 (Dam et al)

A "Phase 1" study is performed to detect toxicity of a new experimental model. Of the 19 patients in the Phase 1 trial, there was no toxicity within a certain dose range of activated white blood cells, and 25% of the patietns enrolled (4 patients with glioblastoma multiforme and 1 with anaplastic astrocytoma) are currently in complete remission with no evidence of active disease up to 4 years from implantation.

BACK

Experimental and Clinical Gene Therapies
for Brain Tumor

BACK

AN HISTORICAL PERSPECTIVE

Experimental treatments employing viruses as antitumor agents were initially described in 1912, when vaccination with rabies virus was thought to have produced the disappearance of a large cervical carcinoma. A resurgence in the clinical use of viruses as a treatment of cancer occurred in the 1950s, subsequent of reports demonstrating the in vitro oncolytic capacity of viruses. These viruses include the Far East Russian encephalitis virus, Newcastle disease virus, polio virus, Egypt 101 virus, as well as herpesvirus. In one clinical study, an attenuated live rabies virus was used in 30 patients suffering from melanomatosis, with tumor regression seen in eight. One of the larger studies was performed at the National Institutes of Health and it involved the use of one of 10 human adenovirus serotypes. Thirty patients with epidermoid carcinoma of the cervix were enrolled and the adenovirus was injected by intratumoral, intra-arterial, or intravenous routes. An antitumor response, defined by the development of visible tumor necrosis at the site of injection, was observed in 26 of 30 patients. The most vigorous antitumor responses occurred in patients without pre-existing neutralizing antibodies to adenovirus. Unfortunately, antitumor responses were short-lived in the majority of treated cases. In 1974, wild-type live mumps virus was injected into the tumors of 90 terminally ill patients. Partial, but short-lived, regression was reported in 79 patients. In both of these large studies, no ill effects were observed in the normal tissues of treated patients. Despite these relatively interesting results, lack of enthusiasm for additional trials may have been produced by the uncertainly regarding mechanisms of action of virus-mediated oncolysis, lack of adequate manufacture of the viruses, and fear of virus-mediated oncogenicity. In 1970, Wright and Smith postulated that witnessed antitumor effects were probably caused by a combination of viral lysis, local production of cytokines, and infiltration of immune effector cells. It is also possible that enhanced tumor antigen presentation might have contributed to this antitumor action, in a manner similar to the antiself immune response exhibited by mice transgenic for a viral antigen. Although our current state of knowledge permits us to better define the molecular events involved in virus-mediated antitumor effects, the prescient hypotheses of Wright and Smith regarding mechanisms of virus-mediated oncolysis remain current.

With the advent of modern virology, the molecular events involved in viral-mediated cell entry and gene expression are being elucidated. This, coupled with the interest in developing technologies for the transfer of genes into mammalian cells, has led to the increasing use of genetically altered viruses as vectors. The current ability to engineer viruses with a desired phenotype, and to manufacture them reliably at high titers, can bow be combined with well-designed and controlled clinical trials, in order to determine whether virus-based gene therapy will truly become a useful adjunct in the battle against tumors.

ANTICANCER GENES

Genes that Activate Prodrugs

The first example of such a gene was described by Moolten in 1986 and it consists of the thymidine kinase (tk) gene from herpes simplex virus type 1 (HSV-1). Its encoded enzyme converts nucleoside analgos, such as acyclovir and ganciclovir, into their phosphorylated metabolites. These then act as competitive inhibitors of endogenous nucleotides for incorporation into the DNA chains of proliferating cells, leading to their death. TK-gene transfer can generate ganciclovir-mediated killing of tumor cells both in vitro an din animal models. This cytotoxicity is augmented by the "bystander" effect, which is thought to occur through gap-function-mediated transfer of the ganciclovir metabolites, at least in vitro. Most gene therapy clinical trials of brain tumors have employed this strategy.

A variation of this approach is to combine gene transfer with radiotherpay. The incorporated "false" nucleotide can sensitize DNA to radiation. When combined with a radiation-inducible promoter, this method provides a "switch" that will turn on tk-gene expression upon radiation. The tk enzyme activates BvUDR, which enhances the radiosensitivity of the tumor cell's DNA.

Several other prodrug-activating genes have been described. These include the Escherichia coli guanine phosphoribosyl transferase (gpt) gene that activates 6-thioxanthine and 6-thioguanine, the E.coli cytosine deaminase (cd) gene that activates 5-fluorocytosine, the rat cytochorme P450 2B1 gene that activates cyclophosphamide/ifosphamide, the E.coli nitroreductase (nfnB) gene that activated CB1954, the mammalian deoxycytidine kinase gene that activates cytosine arabinoside, and the bacterial carboxypeptidase G2 gene that activates 4-[ (2-chloroethyl) 2-mesyloxyethyl) amino] benzoyl-L-glutamic acid (CMDA). This list is by no means complete, and it is likely to include multiple other genes.

Synergistic interactions between different prodrug-activating gene therapies may lead to even more pronounced anticancer effects. For example, we have found that the interaction between the tk/ganciclovir and the cytosine deaminase/5-fluorocytosine deaminase or the tk/ganciclovir and P450/cyclophosphamide therapies can be very synergistic. If these results hold true, the enhance antitumor effect achieved by the use of combined gene therapies may be a means of circumventing the current limitations in efficient gene transfer.
Gene that Activate Immune Responses

The lack of effective immune responses against glial tumors of the brain has been thought to be a result of the immune-privileged status of this organ, as well to the production of immune-suppressive factors by the neoplasm. From a physical stand-point, the blood-brain and blood-tumor barriers are thought to contribute to this relative lack of immune response. Another potential contribution may be provided by the presence of the Fas ligand on the surface of numerous tumor cells. Generation of TGF- from tumor cells has also been shown to inhibit immune responses. Therefore, strategies aimed at enhancing tumor infiltration by cytotoxic lymphocytes, natural killer cells, and other antitumor effectors should prove valuable.

One of the first immunogene strategies employed an antisense insulin growth factor 1 (IGF-1) construct to provide rejection of C6 glioma cells grown in BDX rats. Results from this study provided the basis for a clinical trial in patients suffering from glioblastoma. One note of caution concerns the animal model selected to show anti-tumor efficacy: a recent report revealed the possibility of unintentional MHC mismatches between C6 cells and the BDX rat. This would imply that observed therapeutic effects may have been caused by allogenic graft rejection.

Other strategies have used transfer of the interleukin-4 gene, with evidence of tumor regression. On the contrary, the interleukin-2 gene did not provide evidence of antitumor efficacy, possibly because of side effects related ot the development of intracrerebral edema. One of the more popular genes has been the one encoding for granulocyte/macrophage colony stimulating factor (GM-CSF). Its mode of action probably involves facilitation of antigen presentation by antigen-presenting cells. Transfer of this gene has been shown to produce rejection of tumor cells. Another strategy has employed transfer of an antisense TGF- gene to produce rejection of 9L gliosarcoma tumors implanted in Fischer 344 rats. These result have also been used as the basis for a clinical trial in humans suffering from glioblastoma multiforme.

It is likely that a combination of approaches or the discovery of new immune-enhancing molecules will provide further evidence of therapeutic benefits for these tumors.

Genes that Modulate Angiogenesis

The process of neovascularization of expanding brain tumors offers a target of experimental gene therapy approaches. The molecular characterization of receptors and ligands that stimulate and modulate the ingrowth of blood vessels into tumors now permits the employment of strategies to disrupt these processes. Molecules are corresponding receptors that have been postulated, and, at least in the case of VEGF, shown to be necessary for tumor growth, include basic FGF and its receptor, VEGF and its receptor (VEGFR-1 or flt-1, and VEGFR - 2 pr flk-1), TGF- and the EGF receptor, TGF- , and human platelet factor 4 (PF4) and its receptor.

At least three reports have provided evidence that genetic methods can be used to disrupt angiogenesis in brain tumors. When a dominant version of the flk (VEGF-R2) receptor was introduced into C6 glioma cells by means of a retroviral vector, tumor involution was observed in a nude mouse model. A different strategy was employed by Saleh et al. who employed an antisense VEGF cDNA. Stable transfectants of C6 glioma cells were selected that expressed this cDNA and thus produced markedly reduced quantities of VEGF. These tumor cells were markedly inhibited in their ability to form neoplasm in nude mice when compared to parental cells, and this correlated with a statistically significant decrease in blood vessel formation. Finally, expression of antisense bFGF cDNA correlated with reduced proliferation of C6 cells in culture. In this study, in vivo experiments were not performed, and thus the extent of blood vessel formation in the antisense bFGF vs parental C6 tumors was not determined. Taken in conjunction, these experiments show that either antisense or dominant negative types of gene therapies can be applied to inhibit tumor growth. So far, the VEGF/flk receptor has provided the most fruitful target for the therapeutic exploitation.

Selectively in the angiogenesis strategies is provided by at least two factors : the molecules targeted are preferentially found in blood vessels within brain tumors and not in normal brain; and inhibition of these ligand - receptor interactions will primarily affect rapidly dividing cells, which require continuous nourishment. It is not clear whether the inability to target all tumor cells or all neoendothelial cells will limit antiangiogenesis gene therapies, since the experiments described above were performed on stably transfected cell lines and did not employ in vivo method. It is also not clear whether blocking a particular angiogenenic pathway would end up selecting for blood vessels and tumor cells that are dependent on another pathway. Nevertheless, disruption of angiogenesis remains a promising avenue for therapeutic intervention.

Genes that Are Involved in Apoptosis and Tumor suppression

As the pathways leading to tumorigenesis are being elucidated and are being shown to involve genes that control progression through the cell cycle and DNA repair, strategies to replace these defective genes are also being pursued. In general, the most successful therapeutic results have been achieved with adenovirus vectors injected into established tumors at a relatively high ration of vector to tumor cell. Furthermore, few reports have applied these technologies to the treatment of experimental brain tumors. Therefore, we will review the most interesting experimental results, even if they were obtained in tumor models that do not directly affect the CNS.

Replacement of defective p53 genes has been described in several tumor models, including medulloblastoma, lung cancer, head and neck squamous cell carcinoma, and prostate tumors. Recent work has shown that functions of the p53 gene involve induction of arrest at cell cycle checkpoints to allow for repair of DNA damage, as well as promoting apoptosis of cells that are not able to achieve such repair functions. P53 accomplishes these functions through multiple pathways: transcriptional induction of p21, an inhibitor of cyclin-dependent kinase (CDK), needed for cell-cycle progression : transcriptional induction of Gadd45, as well as regulatory interactions with ERCC3 and other factors involved in repair of DNA damage and nucleotide excision repair pathways; and induction of apoptosis when DNA repair cannot be accomplished. Mutant p53 genes thus allow cells to proceed through the cell cycle and propagate errors in their DNAs. In the overwhelming majority of cases, these errors are fatal to the cell, but, in a few select instances, these errors confer some type of growth-selective advantage and allow for uncontrolled proliferation. Replacement of the defective p53 should thus provide a molecular cure to this problem.

Two studies have employed p53 gene replacement strategies for the treatment of experimental brain tumor cells in culture. In the first, an HSV amplicon was employed to deliver a wild-type p53 gene into DAOY cells, a medulloblastoma line that bears a mutant p53 gene. Findings, suggestive of successful genetic correction, included increased expression of a gene regulated by p53 and loss of immunocytochemical staining for cyclin E, a protein that accumulates when the parental DAOY cells do not arrest type p53 gene into six glioblastoma cell lines that either expressed a wild-type (n=3) or a mutant (n=3) p53 gene. Different effects were observed after gene transfer : the former cells exhibited inhibition in their proliferation while the latter underwent apoptotic death. In both instances, this may lead to significant anticancer effects, although in vivo studies have not yet confirmed this hypothesis.

Transfer of other transgenes involved in DNA repair and control of brain tumor cell cycle have not been reported. However, in a recent report, transfer of a gene linked to apoptosis has been shown to lead to experimental brain tumor regression. As in the p53 gene transfer experiments, the ability to transfer these genes in all tumor cells remains to be determined.

ONCOLYTIC VIRUSES

As decribed in Section 1., interest in oncolytic viruses is not recent. Over the past few years there has been a resurgence in experimental anticancer therapies employing mutants of HSV-1, Newcastle disease virus and adenovirus. With HSV-1, deletions in viral genes allow selective replication of the virus gene whose function may be linked to cellular apoptosis. The first studies that described use of herpes viruses for destruction of cultured tumor cells were by Levaditi and Nicolau. In 1949. Moore employed a herpes virus as a treatment for rat sarcomas transplanted in rats. Unfortunately, these experiments were not successful, presumably because of the poor infectivity of human herpes virus for rodent tumor cells. The first recent report providing new impetus in this field was by Martuza et al. They employed a deletion mutant of the viral tk gene to show selective killing of human glioma cells both in vitro and in animal models. The observed selectivity was probably generated by the mammalian tk enzyme, which is upregulated during the G1/S phase transition of the cell cycle. This enzyme could complement the function of the deleted viral thymidine kinase, thus allowing for viral growth in dividing cells, but not in postmitotic cells.

Another strategy has employed deletion of the viral ribonucleotide reductase gene, whose mammalian counterpart is also upregulated during the G1/S phase of the cell cycle and can also complement the deleted viral function. We and others have shown that this virus can also selectively replicate in and kill glioma tumor cells in culture and in vivo. Viral-mediated gene expression could also be monitored by expression of the marker lacZ gene. One advantage provided by this virus is that it retains the endogeneous tk gene, allowing for granciclovir sensitivity. In fact, we have shown that the combination of the HSV mutant and ganciclovir can lead to an amplification of the viral killing effect, presumably by a bystander-type mechanism. Obviously, the prodrug's metabolities will also terminate replication of the virus and further tumor lysis.

Another gene that has been frequently deleted from the HSV-1 genome is the one that encodes for the 34.5 gene. This gene may be involved in the regulation of cellular apoptosis in a neuroblastoma cell line. Its carboxyl-terminal domain possesses homology with domains of MyD116 and Gadd34, which are involved in cell cycle processes and repair of DNA damage. Several different deletion mutants in 34.5 have shown anticancer efficacy both in vitro and in animal models. The basis for tumor selectivity of these mutants has not been elucidated. Another viral mutant that has attracted considerable interest possesses deletions of both the 34.5 and the ribonucleotide reductase (RR) genes. This double mutant is clearly being considered for a clinical trial, since it has been introduced into the brain of primates without evidence of ill effects. It is evident that this construct may become the first herpesvirus mutant to be injected into recurrent human glioblastomas.

Newcastle disease virus (NDV) is an avian paramyxovirus. One NDV strain has been used against human fibrosarcoma and human neuroblastoma xenografts in athymic mice. In both studies, complete tumor regression was evident in the large majority of treated tumors. The virus has been injected into rodent brains without apparent ill effects. The mechanism of selective tumor toxicity is not known. One potential contributor may relate to the induction of TNF- . It is evident that this virus seems a promising one to explore for brain tumor therapy.

A recent report has attracted considerable interest, because it employs an adenovirus mutant in the E1B 55 kD a protein. Replication of the virus is directly tied to the ability of this protein to blind and inactivated wild-type p53 in cells. However, if a cell is a p53 mutant, viral replication occurs readily. In this study, it was clearly established that the mutant adenovirus replicated in tumor cells with p53 mutations, but it replicated poorly in tumor and nontumor cells with a wild-type p53 gene. This translated into an effective antitumor response in p53-deficinet human tumor xenografts, without evidence of viral replication in normal tissues.

In conclusion, the oncolytic virus approach appears to show benefit, and clinical studies should reveal whether its potential will be realized.

CLINICAL STUDIES

The announcement of the first gene therapy clinical trial for recurrent glioblastoma multiforme was met with incredible excitement and patient requests for participation. Newspaper articles heralded the science that had allowed this breakthrough to occur. Naturally, this has produced unrealistic and probably premature expectations for success. The definition of "success" is also critical, in that it is unlikely that any single treatment approach will result in the "cure" of glioblastomas, but it is likely that some approaches may result in partial antitumor responses, with stabilization of the disease. Most practitioners of neuro-oncology would agree that such a response would be considered a "success."

In the initial phase I trial, nine patients afflicted with recurrent glioblastoma multiforme, three patients afflicted with malignant glioma, and three patients afflicted with metastatic brain tumors received stereotactic inoculation of tk-retrovirus producer cells into their brain tumor. Fourteen days later, patients were treated with a 14 d course of ganciclovir. Adverse events include neurologic dysfunctions in one patient and intratumoral hemorrhage in another. There were no complete tumor regressions by magnetic resonance imaging studies, although the authors did report evidence of partial regression in the early posttreatment period. In a few cases, biopsies were obtained after inoculation of the producer cells. These revealed that the extent of tk-gene expression was less than 1% in the biopsied tumor sample. The lack of gene transfer led the authors to conclude that "bystander" mechanisms were primarily responsible for the observed partial tumor responses.

A subsequent phase II clinical trial was then conducted in multiple centres throughout the United States. The treatment scheme involved resection of the malignant brain tumor with injection of the tk-retrovirus producer cells into multiple sites within the wall of the cavity. This was then followed by placement of a catheter within the cavity connected to an Ommaya reservoir, to allow for reinjection of procedure cells every 40 - 45 d. Ganciclovir was administered after each injection of producer cells. Thirty patients were treated. Four patients died of causes unrelated to injection of the producer cells. The average survival time of 18 evaluable patients was 25 wk from the initial injection of producer cells. Adverse events were related to inadvertent injection of producer cells into the ventricle, producing a severe aseptic meningeal reaction. MRI findings were consistent with a transient inflammatory response and evidence of partial antitumor responses. The authors concluded that this approach demonstrates safety and modest antitumor efficacy.

Based on these results, a prospective, randomized trial involving 240 patients in three countries (United States, Germany, and Canada), comparing the standard treatment of a newly diagnosed glioblastoma with surgical excision, followed by injection into the tumor margin with tk/retrovirus producer cells and subsequent treatment with ganciclovir and radiation therapy. Results from this trial are pending. These trials have been sponsored by Genetic Therapy Incorporated. Similar trials employing tk-retrovirus producer cells and ganciclovir as a treatment for malignant brain tumors have been performed in France.

Two phase I clinical trials are currently being conducted, these employ a replication-defective adenovirus that bears the tk gene to provide ganciclovir susceptibility to malignant brain tumors. In the Pennsylvania trial, the tk-adenovirus is stereotactically inoculated into the tumor. Two days later, ganciclovir is administered. Seven days after injection of the adenovirus, the tumor is resected en block and the tumor cavity is injected with the tk-adenovirus. Ganciclovir is then administered to the patient for an additional 14d. The resected specimen will then be assayed for extent for gene transfer, immune responses, and tumor necrosis.

CONCLUSIONS

The promise of cancer gene therapy is immense. However, it is also evident that the fulfillment of this promise is still unrealized. Phase I clinical trials in humans afflicted with malignant brain tumors are justified for strategies that are relatively safe and that show evidence of anticancer efficacy in appropriated models. However, since these therapies are in their infancy, they should be designed to maximize the amount of useful scientific information that can be derived. The information ideally should provide an estimate of the extent of gene transfer, vector spread, and transgene activity within the tumor, the degree of immune response within the tumor, and the temporal kinetics of anticancer gene action within the tumor. The acquisition of this information in animal models is only partially relevant to the clinical arena. Knowledge of these variables during the course of a phase I clinical trial might provide the means to effectively optimize, refine, and alter the gene therapy in use. Absence of this knowledge will only lead to guesses related to causes for the observed success of failure of the gene therapy in use and subsequent inability to circumvent limitations.

Rationale for Immunotherapy of Brain Glioma

BACK

There has been no change in the management of primary malignant tumors of the brain (Glioblastoma Multiforme and anaplastic astrocytomas - astrocytomas Grades 3 and 4) for 20 years.

Radiation therapy can double life expectancy, and chemotherapy may add a few more months to median survival. The modern use of Gamma Knife adjuvant radiosurgery has further pushed the survival curve with it's high response rate, but still there is no cure for this dreaded disease. Survival may be only 1 year from the time of diagnosis, with fewer than 5% loving 5 years. Thus, a new approach had to be undertaken.

Brain tumor cells, being foreign to the body, should be able to be managed by our normal immunologic defenses as they would if a bacterial or viral infection occurred. Furthermore, there should be a "memory" of the foreign pathogen such that if the immune system came in contact with a tumor cell again, it could recognize and destroy it much like a second exposure to chicken pox.

Why doesn't this occur with primary malignant tumors of the brain? Could it be that brain tumor have evolved a way to evade our own defenses and produce a "wall" of immunosuppression. The cure for this cancer may well lie in the breakdown of this wall, or indeed the overwhelming of this wall.

Because of the infiltrative nature of
Grade 4 astrocytomas versus the
compact nature of Grade 1 astrocytomas,
and the presence of their hypothetical
"immunosuppressive barrier", the cure
for this tumor must be a biological
one such as immunotherapy.

Approximately 8 years ago, Dr Gale (Morrie) Granger experimented with Interleukin-2 (IL-2) activated lymphocytes and their in vivo effect on human brain tumors. Some results were excellent, but toxicities could be high. IL-2 is a "cytokine" - a type of chemical messenger which can enhance the immune response toward a foreign antigen (such as a tumor cell).

It was discovered that many cytokines may be automatically produced when too unlike, foreign, white blood cells come in contact. Thus, mixing the patients white blood cells with those of an unrelated donor, and allowing them to incubate in the lab for three days causes enhanced production of at least 7 different cytokines.

"Mixture of host and donor
lymphocytes causing the production
of cytokines and the targeting of a tumor cell".

If theses are placed in the tumor bed at the time of a tumor resection, presumably this "angry" white cell mixture can break the immunosuppressive barrier that Glioblastoma Multiforme and Anaplastic Astrocytomas inherently have, and enable the hosts own immune system to "seek and destroy" the abnormal tumor cells. This is called a Mixed Lymphocyte Culture or MLC.

Results of the phase 1 MLC immunotherapy trial 1996 - 1998 (Dam et al)

A "Phase 1" study is performed to detect toxicity of a new experimental model. Of the 19 patients in the Phase 1 trial, there was no toxicity within a certain dose range of activated white blood cells, and 25% of the patietns enrolled (4 patients with glioblastoma multiforme and 1 with anaplastic astrocytoma) are currently in complete remission with no evidence of active disease up to 4 years from implantation.

Neuroprotection in Brain and Spinal Cord Injury

BACK

I. INTRODUCTION

Traumatic injury to the brain and spinal cord is responsible for mortality and lifelong morbidity, far out of proportion to the anatomical extent of damage. Central nervous system (CNS) injury is accordingly the leading cause of death in trauma [1] and one of the leading causes of lost productivity because most trauma victims are young. The brain and spinal cord have very limited ability to regenerate once injured, making it especially important to limit early damage. Mechanisms of the traumatic injury share many common pathways for cell death, with other processes such as stroke, hemorrhage, and even aging and degenerative diseases. The exciting prospect in this similarity is the some posttraumatic damage may be prevented by the same "neuroprotective" agents and strategies, as in ischemia and neurodegeneration.
Neuroprotection has been well demonstrated at the cellular level and in vivo laboratory animal studies of experimental trauma. This chapter will first outline the relevant mechanisms that damage the brain and spinal cord after human trauma. W will also review the laboratory data that support the mechanisms and agents that have entered clinical trials. Finally, we will discuss current and previous trials, as well as future trends. Clearly injury prevention is the best form of neuorprotection! There is now epidemiological evidence in the United States that improvement in motor vehicle safety is correlated with a decrease in the number of injuries and deaths due to motor vehicle head injuries. Unfortunately, this is offset by an increase in the number of gunshot wounds to the head, leaving the overall number of head trauma fatalities unchanged [2].

II. HEAD TRAUMA
A. Closed Head Injury

Brain damage after head injury is a complex process that is initiated by the traumatic impact damage and followed by a chain of secondary events that results in the final damage pattern. Nearly one third of patients who die from head injury had spoken some time after their initial injury [3,4], indicating the importance of these secondary events. This suggests that a clinically significant "window" exits for improving the outcome in this chain of secondary events. Neuroprotection therapies may thus be targeted at this window and will be discussed below in detail. If the initial traumatic insult is very severe, little hope can be placed in pharmacological intervention. Such patients usually present in deep coma, with a Glasgow Coma Score of 3 (GSC-3), and have a very poor outcome. For this reason, these patients have not been considered in many clinical neuroprotection trials.

1. Mechanisms of Brain Damage After Head Injury

Figure shows four major mechanisms of brain damage and their overlap. This is based on neuropathological studies in large series of patients who died after severe head injury [5-7]. By far the most common type of damage is pyknotic neuronal shrinkage and death of the hypoxic/ischemic type, which is seen in 85-90% of cases who die (Fig. 2). These pathological changes are also seen after stroke, status epilepticus, and adjacent to focal contusions. Identical damage may also be caused by exposure to "intrinsic" neurotoxins, such as free radicals and glutamate, in high concentrations [8]. Many "neuroprotection strategies" have been shown to prevent such damage in various trauma models [8-16] (see blow).
Understanding the cellular mechanisms of the primary and secondary events is crucial in developing effective treatment. Insights into the basic mechanisms of neural tissue damage, such as the demonstration of calcium-dependent cell death, asrocyte swelling, and excitatory amino acid neurotoxicity, have provided the fundamental basis for therapies directed at saving neurons [17,18]. Although these insights have come from animal and in vitro models, the extrapolation from the Petri dish to the human situation requires evidence that these same events and processes are occurring in the human brain after head injury. This leap is easier to make in other CNS injury, such as stroke where ischemia is the single initiating cause of cell death. Recently, increased levels of excitatory amino acids have been demonstrated in human head injury in vivo using the technique of microdialysis, leading support to this mechanism of brain damage [19,20]. However, free radicals and their breakdown products have been much more difficult to demonstrate in animal models and have not been shown in humans.

Contusion and Intracerebral Hematoma

Shear forces applied stretch and tear axons, pia mater, and blood vessels, causing axonal injury (see blow), contusions, and hematomas, both within (intracerebral) and outside (extra axial: subdural and epideral hematomas) the brain. These are a major cause of hypoxic/ischemic damage, due to local and whole-brain compression, and resulting regional blood flow reduction (see blow).
Intracerebral hematomas and contusions clearly disrupt grey nd whie matter by tissue tearing, but far more damage is done by release of harmful cell damage mediators (e.g., K+, free radicals, cytokines, gluamate) and secondary, cytotoxic edema, which exacerbates ischemic neuronal damage and reduced microcirculatory blood flow.

Relationship Between Cerebral Blood Flow and Brain Damage

Figure 3 shows the relationship between cerebral blood flow (CBF) and tissue infarction time, as determined in healthy, unanesthetized squirrel monkeys [22]. It is seen that levels of CBF around 10-18 ml/100 g were tolerated for up to 2hr in these studies, without producing infarction. This window of opportunity May by much shorter in head-injured humans; nevertheless, the central hypothesis behind neuoprotection strategies for hypoxic/ischemic damage revolves around this relationship. These neuroprotective strategies are aimed at shifting the curve to the right to increase the tolerance of nervous tissue to low CBF and increased levels of harmful mediators.

Diffuse Axonal Injury

Shear forces cause development of axonal disruptions 12-36hr after impact in human (3-12hr in rats). This progresses to widespread development of "retraction balls" - blobs of axoplasm (Fig. 4) in about 25-30% of severe head injuries who die [23]. Clearly, less severe degrees of transident or partial axonal damage occur in all severely head-injured patients. Recent animal studies have shown that transient, reversible mitochondial swelling occurs first, even in mildly injured (deformed) axons [24]. More severe injury causes progressive concatenation of the cytoskeletal proteins of the axon, resulting in less of axonal transport and, lter, interruption of axons. This process may be calpain-calmodulinmediated [24]. Thus far, how3ever, no therapeutic strategies have been able to influence this process in the laboratory. Patients who show the radiological hallmarks (on Ct or MRI) clinical features of diffuse axonal injury alone may thus need to be excluded from clinical trials with drugs directed at ischemia and contusion mechanisms.

B. Gunshot Injuries

Missile injuries to the CNS are characterized by the caliber and velocity of the impacting projectile. The degree of injury is largely determined by the transfer of kinetic energy (given by the Newtonian equation KE=1/2mv2, where m is the mass of the projectile and v is its velocity) and by the relationship of the size and shape of the missile along with the KE (known as the relative stopping power [25] ). Cranial bullet injuries cause primary damage by three major mechanisms: contusion, hemorrhage, and axonal disruption. These may be magnified by secondary projectles, such as bone chips and foreign bodies, driven in the brain [26]. Primary axoinal disruption is usually minor and is overshadowed by the consequences of secondary "blast effects" of the kinetic injury, which are usually the cause of early death.
Blast effects - reverberating shock waves - create brief, massive deformation of the brain, with subsequent membrane and ion channel disruption and massive intracellular swelling. This mechanism is also seen in blunt head injury. When ionic flux, loss of resting membrane potential, and consequent astrocyte swelling are massive, raised intracranial pressure and death are likely. This usually occurs with transhemispheric or bihemispheric injuries, producing the "ground glass" appearance, with loss of gray - white differentiation and sulcal effacement, on a cranial CT scan. Abnormalities in cerebral hemodynamics, such as decreased CBF and vasospasm, have been demostrated in both experimental and clinical missile injury [27-29]
Since these processes are similar in both penetrating and closed head injury, the principles of neuroprotection apply in both conditions, although the outcome is worse overall for penetrating (gunshot) injuries.

C. Basic Principles of Neuroprotection

The search for "neuroprotective" treatments to improve outcomes is not a new idea. Initiation of clinical trials to test neuroprotective drugs began in the 1950s with a clinical trial of atropine in severe head injury [30]. Hypothermia was also first employed in the 1950s and 1960s, and is now enjoying a revival of fashion [31,32]. To date, no clinical phase III trial in head injury of a neuroprotective agents or strategy has been successful, despite more than 15 attempts at a cumulative cost of hundreds of millions of dollars. Several phase III trials have been based no minimal evidence in animal models, and thus may be an important factor in the failure of these phase III trials (Table 1).

Table 1 Completed Neuroprotection Trials in Human Severe Head Injury
(July 1996)

Invesigator, and Agent No. of Patients/ Outcome/
Year Country Comments

1. Ward et al., Scopolamine 940/U.S. Uncontrolled, no
1954 benefit
2. Six authors Corticosteroids 365 (summary Controlled, no
of 6 trials) benefit
3. Schwartz et al., Mannitol versus 59/Canada Randomized
1984 Pentobarbital crossover
permitted,
Mannitol group
had better
outcome
4. Ward et al., Barbituates 53/U.S. Controlled,
1985 (prophlaxis) Prophylactic, on
benefit
5. Eisenberg et al., Barbituates 73/U.S. Controlled, ICP
1988 (therapeutic) lowered (high
ICP patients only), outcome no better.
6. Wolf et al., 1993 THAM
(tromethamine 149/U.S. Controlled,
buffer) ameliorates
danger of
hyperventilation,
ICP control better
7. Teasdale et al., Nimodipine 255/U.K./ Double-bind
1992 (Hit I) Calcium Finland Controlled, no
antagonist benefit
8. Beaakman 1993 Nimodipine 840/Europe Double-bind
(Hir II) Controlled, no
benefit
9. Muizelaar et al., PEG-SOD-free 94/U.S. Double-bind
1993 radical scaven- Controlled, ICP
ger (phase II) lower, outcome
better

10. Muizelaar et al., PEG-SOD-free 463/U.S Double-blind
1994 controlled,
outcome was 9%
better in drug
than placebo
(p=0.15)
11. Bullock et al., CGS-1975 113/U.S./ Double-blind
1994 (glutamate U.K. controlled, ICP
NMDA lower, CPP better
antagonist)
12. Alves and Jane, Tirilazad 1170/U.S./ Double-blind
1995 (aminosteroid Canada Controlled, no
antioxidant) benefit
13. Marshall, 1995 Tiilazad 1128/Euope, Double-blind
Australia Controlled, no
Benefit
14. Bullock et al., Selfotel 266/U.S., Both terminated
1995 CGS-197 Israel 426/ due to excess
(phase III) Australia mortality in
concomitant
stroke trials
15. Cohadon et al., Synthelabo France Double-blind,
1996 eliperodil Placebo, control,
(SL-82) no benefit
Phase II
16. Harders et al., Nimodipine (L 123/Germany Double-blind,
1996 Channel (only traumatic placebo control,
1997 Calcium subarachnoid 55% relative
antagonist), hemorrhage reduction in bad
phase II patients outcome at
selected) 6months P<0.002
Subgroup analysis indicated significant benefit in patients with traumatic SAH.

Probably future phase III trials will need to fulfill more vigorous requirements prior to initiation. Ideally, these should include the following:

1. The mechanism has been demonstrated in animal models of trauma.
2. The mechanism/damage is reversed in animal models by the trial agent/drug.
3. The mechanism has been demonstrated in human neurotrauma.
4. The drug/agent is sage in humans with neurotrauma.
5. The drug/agent penetrates the brain/cord in therapeutic amounts.
6. "Endpoints" of the trial are appropriate to the drug/agent and are sufficiently sensitive.

Table 2 shows how the current mechanisms of interest may be ranked on an arbitary 0-5 scale for head injury. These current mechanisms are reviewed below.

D. Types of Neuroprotective Strategies in Head Injury
1. Free Radical Scavengers

Several neuroprotective agents are thought to act primarily by scavenging free radicals, but their actions may be multiple and may go beyond their function as free radical scavengers. The free radical scavengers tromethamine (THAM), superoxide dismutase (SOD) [coupled to the carrier molecule polyethylene glycol (PEG), and the steroid dexamethasone have been studied in prospective randomized trials of severe head injury [33-35]. THAM and

Table 2 Requirements for TBI Trials: How Do Mechanisms Compare?

Occurs in Blocked in Occurs in Drugs Drugs
Mechanisms animal models animal human available penetrate
of TBI models TBI

Free radicals 3 4 0 4 2?
Glutamate 5 5 3-4 4 3
Calcium 4 3 ? 2 2
Cytokines 3 2 2 1 1?
Inflammatory
mediators 3 3 2-3 2 3
Anticholinergic 3 4 1-2 4 4

dexamethasone both failed to improve outcome when compared to controls. THAM did affect intracranial pressure (ICP), with drug-treated patients having significantly less time under 20mm Hg when compared with controls [33]. However, renal toxicity has limited further application.
PEG-SOD (Sterling-Winthrop Pharmaceuticals) has shown more promose. In a phase II clinical trial, PEG-SOD demonstrated a trend toward improved outcome and easier control of ICP in severely head-injured patients [35]. A phase III trial was recently conducted that failed to show statistical significance in terms of improved outcome or lower ICP, although outcome was 8% better in the treatment group, and overall outcome, even in the control patients, was much better than historical trials of severe head injury.
Dexamethsone has been used for years for its entiedema effects, especially in tumor patients. Great interest in its effects on human head injury prompted many clinical trials with conflicting results [34,36-39]. However, in a randomized prospective, double-blind study of 161 comatose head-injured patients, megadoes decadron (100-mg bolus within 6hr of injury followed by 100mg/day for 4 days and then by a tapering does until day 10) failed to demonstrate clinically significant efficacy [34]. A recent study with another steroid, triamcinolone, has demonstrated a significant benefit only in a subgroup of head-injured patients, i.e., those with focal contusions [39].
The interest in steroids hs been rekindled with the advent of the nonglucocorticoid 21-aminoseroids, the Lazaroids. Lazaroids, specifically tirilazad mesylate (U74006F, Upjohn), have shown promise in laboratory studies with impact-injured mice [40,41], and preliminary data have shown trends for improved outcome in human subarachnoid hemorrhage [42]. Two major phase III trials with tirilazad in human head injury have been conducted. The first of these, the North American trial, was negative, and the second (Euopean) trial is being analyzed at the time of writing. Naturally occurring scavengers, such as vitamin C and E, are also bein evaluated for trials in human head injury.
Two major problems with all these free radical scavengers remain. First it has been impossible to demonstrate effects of free radicals in humans after head injury, so that the "therapeutic window" and "injury subtype specificity" can only be inferred from animal models; and even in models, these issues are not yet resolved [41,43]. Second, these large "scavenger" molecules have never been shown o adequately penetrate the brain, although it is speculated that they act at microvascular endothlial level only.

2. Calcium Channel Blockade

Inspired by the extensive in vitro and in vivo laboratory evidence of calcium-mediated neuronal death in models of trauma and ischemia and subsequent neuroprotection in experimental studies using calcium channel blockade, clinical trials with calcium with calcium channel blockers have been conducted in human head injury [44,45]. As with tirilazad, the calcium channel blocker nimodipine has shown clear benefit on patient outcome after SAH. To study its effects on severe head injury, nimodipine was studied in two large, prospective, multicanter trials in Europe. Nimodipine was started within 12hr of injury and outcome at 6 months was assessed. The overall results failed to show significant differences when compared to placebo [46]. However, careful analysis of subgroups revealed that those patients who had traumatics SAH on CT scan faired better than controls. This finding was tested in a subsequent small, randomized, prospective study focused only on patients with traumatic SAH (tSAH). All head-injured patients with tSAH on their initial CT scan were included (N = 123 patients). Roughly 25% had "moderate" head injuries. The results confirmed a 55% lower incidence of bad outcome in the nimodipine group compared to placebo at 6 months (P < 0.001) [47,48]. Other calcium channel blockers, particularly those that cross the blood-brain barier more readily than nimodipine, are being evaluated in animal models of brain injury [49].
Magnesium also functions as a blocker of calcium by obstructing the N-methyl-Daspartate (NMDA) ion channel and is cytoprotective by this mechanism. Its efects are well known in humans having been used extensively and effectively to prevent seizures in preeclampsia. Protection in animal models of ischemia has been shown [50]. The role of magnesium treatment in human head injury is currently being investigated.

3. Excitatory Amino Acid Antagonists

The excitatory amino acids, glutamate and aspartate, are known to cause cell death, especially in the context of compromised energy supply such as ischemia or hypoxia. Glutamate is an especially important neurotransmitter in the mammalian brain. Its actions are mediated via three ion channel receptors - the NMDA receptor, the kainate receptor, and the a-amino-3-hydroxy-5-methy14-isoxazole propionate (AMPE) receptor - and via a second-messenger-linked metabotropic receptor. Metabolism of glutamate is initiated by reuptake mechanisms at the synapse and involves astrocytes as well as neurons, which keep the extracellular concentration low, around 2uM [51]. In hundreds of laboratory studies, including several in models of traumatic injury, antagonism of glutamate has been shown to ameliorate neural injury [8,12-14,16,18,52-69]. Using microdialysis, we recently also demonstrated tremendous increases in the extracellular concentration of glutamate in patients with severe head injury [19,70]. These increases can be on the order of 10- to 50- fold over baseline and may persist for ays after injury. The presence of such high glutamate is evidence for a significant window of intervention in human head injury. Clinical studies targeting this excitotoxic mechanism have begun and involve a wide variety of related compounds. Their clinical application thus far has been limited by neuroactive side effects. Several categories of these drugs will be discussed. The majority are designed to act at the glutamate-activated NMDA receptor, and a schematic representation of this receptor is provided in Fig. 5.

Competitive NMDA Antagonists

Drugs that competitively bind the glutamate site of the NMDA receptor, e.g., CGS-19755 (Selfotel, Ciba-Geigy) and D-CPP-ene (Sandoz), have been used in phase II and III clinical studies in human head injury. In one phase II trial with CGS-19755, ICP was lowered and cerebral perfusion pressure (CPP) was improved in the treatment group [71]. However, four phase III trials in both stroke and head injury with CGS-19755 have recently been halted, after recruitment of about 1200 patients, because of excess mortality in the drug groups for stroke. The reason for the mortality difference is not yet known.

Noncompetitive Antagonists

Noncompetitive NMDA receptor antagonists, of which MK-801 is the most well known, have excellent brain penetration compared to the competitive antagonists. This is because these agents exhibit "use dependency" i.e., enhanced binding, in tissue where the glutamate content is high. When glutamate binds to the NMDA receptor, I causes a conformation change that exposes the MK-801 binding site (Fig. 5). Thus, in many respects, these agents are the "ideal" neuroprotectant for ischemia/hypoxia-mediated posttraumatic damage - the closest practical agent to the "magic bullet" concept. Mk-801 is the gold standard neuroprotectant in laboratory models of excitatory neurotoxicity. However, human neuroprotection trials with this agents were never considered due to its severe side efects in animals, i.e., sedation, agitation, and neuronal vacuolation. Another related agent, CNS-1102, is currently being evaluated in phase III trials in human head injury and stroke. The short half-life and apparent tolerable side effect profile of this agent make it, in our view, the most interesting compound of this category at the present.

Receptor Modulators

Agents that modify the NMDA receptor are also being investigated. These agents antagonize the binding of endogenous modulatory compounds such as glycine, spermine, spermidine, and putrescine. These drugs are designed to have fewer psychomotor effects, such as agitation and hallucinations, than the competitive and noncompetitive NMDA antagonists. The glycine site antagonists, such as ACEA-1021, have demonstrated efficacy in the laboratory [72,73] and phase I trials have now begun.
Antagonists of the other glutamate ion receptors, the AMPA and kainate receptors, have been developed and have demonstrated neuroprotective efficacy in both global and focal cerebral ischemia. This is in contrast to the NMDA receptor antagonists that have not consistently shown protection against global ischemia. This is an important aspect when considering the patient with severe head injury because a substantial component of the secondary ischemic injury occurs as a result of global ischemia due to compromised CPP. Several of these compounds, e.g., LY-293558, may be suitable for clinical use in the near future[69].

Presynaptic Release Inhibitors

Another strategy involves preventing the release of glutamate via presynaptic inhibition. One such compound Enadoline (Parke-Davis), is currently in phase II clinical investigation in head injury at the Medical College of Virginia. BW-619C89 (Burroughs-Wellcome/Glaxo) is another neuroprotectant that blocks sodium channels and glutamate release but has unfortunately been put on hold. The experimental glutamate antagonists are summarized in Table 3.

4. Hypothermia and Metabolic Agents

Observaions made in cold-water near-drowning patients, and in patients with overdoses of metabolic depressants (e.g., barbiturates), have shown that protection of the brain is possible by use of metabolic depression[74]. This has been confirmed in many laboratory studies. Hypothermia and hypnotic drugs act to decrease the metabolic demand of the brain. In the setting of compromised blood supply, this can be protective.

Hypothermia

The efficacy of hypothermia was championed by Rosomoff in the 1950s. Trauma trials never went beyond primate studies complications [75]. Recently, substantial experimental neuroprotection was demonstrated by hypothermia,

Table 3 Requirements for TBI Trials: How Do Glutamate Antagonists Compare

Efficacy Brain
Agent in models penetration Safety Tolerability

Selfotel (Ciga-Geigy) 4 1 3? 3
EAA-494 (Sandoz) 5 2 4 2
Cerestat-CNS-1102 4 4 4 2
ACEA-1021 3 _ _ _
Eliprodil (Synthelabo) _ 3? 4 4
CP101-606 (Pfizer) 3-4 4 3-4 4
BW619 (BW/Glaxo) 4 _ _ _
Enadoline (Parke-Davis) 4 _ _ 4
MK-801 (nonclinical) 5 5 1-2 0

leading to renewed clinical interest [76]. Human trials in severe head injury using moderate hypothermia (32° C) are underway [31]. Premliminary results indicate that systemic complications do not seem to be a significant factor in moderate hyopthermia (32° C) and that hypothermia may improve outcome. A large prospective, multicenter trial with an enrollment goal of 500 patients, is underway in the United States [32].

Hypnotics
Hypnotics such as the barbiturates have been used to produce electrical silence on the electroencephalogram (EEG), reducing the cerebral metabolic expenditure by roughly 50%. This effect has proven to be beneficial in the laboratory setting of focal or temporary ischemia [77-80]. Their use in refractory intracranial hypertension is also well known [81]. However, as a "blanket neuroprotectant," the use of barbiturates in severe head injury was not beneficial. In a prospective randomized trial of severely head injury patients, pentobarbital was given to achieve burst suppression on EEG for 3 days and then was slowly weaned off. This "prophylactic" use of pentobarbital failed to improve ICP or outcome, and caused significant hypotension [82]. Etomidate is another hypnotics, it has been shown to be protective in models of ischemia [83]. The use of etomidate coma in severe head injury is not indicated, due to a significantly increased incidence of pulmonary sepsis in ICU patients sedated with etomidate [84]. This is thought to be due to adrenal suppression. Short-term use in the setting of a known or planned ischemic event, such as temporary arterial occlusion in aneurysm surgery, seems to be the best current indication for its use [85-87].

Other

Another compound, nizonfenone, also thought to exert neuroprotection by decreasing the metabolic rate has shown encouraging results in a preliminary study in severe head injury. When compared to barbiturate-treated patients, those treated with nizofecone had significantly better outcome at 1 and 6 months. Unfortunately, there were no control patients with whom to compare [88].
Metabolic agents may play a role in selected cases or subgroups of head injury patients, such as those with refractory intracranial hypertension, rather than as a "blanket" treatment.

5. Antiinflammatory Agents and Cytokine Antagonists

These protection strategies are aimed at preventing damage due to the inflammatory cascade.

Recently, the role of polymorphonuclear leukocytes (PMNs) and the monocytemacrophage system has been studied after trauma and ischemia. In both clinical situations, these cells are activated and rapidly enter damaged tissue to exacerbate secondary damage, from 4-6 hr to several days after the event. Whole-body radiation therapy, antimeatbolite durgs, and antibodies against interleukins have been shown to reduce brain edema and ischemic damage in animal models [89].
Currently, a phase II trial with a bradykinin antagonist in underway (focused on patients with focal contusions only) to test the hypotesis that bradykinin, released in contested tissue, exacerbates pericontusional edema and thus ICP (Table 4). Indomethacin has been shown to decrease ICP and increase CPP when given as a bolus infusion to patients with severe head injury and elevated ICP, but it has not been fully evaluated in a phase III trial [90].

Table 4 Ongoing Traumatic Brain Injury Clinical Trials, July 1996

Company Drugs Phase N, country Mechanism/class

Cambridge Aptiganel II/III 600-800 NMDA glutamate
Neuroscience Cerestat U.S./E.U. antagonist
Sandoz SDZEAA- III ~400 E.U. NMDA antagonist
494 (U.S. later?)
Cortect/SB Bradycor II 160U.S. Bradykinin (BK2)
antagonist
Neurex SNX-III I/II U.S. Presynaptic Ca
Antagonist
Pharmos HU-211,
dexanabinol I ? "Neuroprotective"
cannabinoid
Astra/ICI Popofol II ~200 U.S. Steroidal
anesthetic agent
Pfizer CP101-606 II ~100 U.S. Polyamine site
NMDA
antagonist
Parke-Davis Enadoline II 80 U.S./ K opioid agonist,
CI977 E.U. glutamate release
antagonist
Cocensys/CIBA ACEA-1021 II ~120 U.S. Glycine site
NMDAantagonist
Bayer BAY-X3702 II E.U. 5-HTIA agonist/
ion channel block

6. Other Agents

Some common agents used in the routine care of head-injured patients also have neuroprotective properties. The anticonvulsants, especially phenytoin, and osmotic diurectic mannitol have been shown to exert protection in some experimental models.

E. Gunshot and Other Penetrating Injuries

Despite its growing prominence in the epidemiology of traumatic brain injury in the United State and worldwide, penetrating injury has been ignored with respect to neuroprotective drug studies. This is unfortunate because there is considerable overlap in the pathophysiology of the initial injury and secondary insults of missile wounds and of blunt trauma. Phamaceutical companies, however, feel that the pathophysiological basis for these injuries is not sufficiently well documented, and social issues make these patients less attractive for trials.

III. SPINAL CORD INJURY

Traumatic spinal cord injury (SCI) occurs when there is disruption of the normal protective elements of the spinal column, resulting in compression, shearing, rotatory, vascular, or concussive injury to the cord. As in head trauma, mechanical disruption of axonal tracts and vascular elements due to stretch and impact creates an injury characterized by hemorrhage, ischemia, and cytotoxic edema. Many of these mechanisms are now being elucidated in animal models and by the study of human postmotem material. The increasing use of acute MRI scanning for these patients now allows us to define an early radiological hallmark - the "T2 bright lesion" in the cord of the spinal injury patient, although it is currently difficult, and sometimes impossible, to differentiate between "cytotoxic cord edema" and hemorrhagic contusion in vivo using MRI.
Just as in the brain, secondary insults and the initiation of the cascade of harmful events allows for potential neuroprotection in SCI. clinical neurological status helps identify those patients who are most likely to benefit from pharmacological intervention. A motor a d sensory complete SCI (Frankel grade A) is unlikely to be helped with any intervention. This is analogous to the most severe head injury with a GCS score of 3.

A. General Principles

Prompt triage and medical stabilization cannot be overemphasized. Associated hypotension from spinal shock should be managed aggressively. The principle of hypotension as the major factor associated with poor outcome in head injury [91] applies also to spinal injury. The concept of spinal perfusion pressure must be considered as an important determinant of outcome in the patient with an isolated SCI. in the supine patient the intraspinal pressure in normally low, but in the seated patient it will equal the pressure of the column of CSF above it as determined by the height of the column, and in patients with lower cervical and thoracic cord injuries, this may be significant and compromise perfusion. In a patient with a concomitant head injury, raised intracranial pressure will be translated to the intraspinal compartment that may also affect perfusion in this setting. The use of vasopressors and aggressive monitoring including indwelling arterial transduction, central venous access and monitoring, and ven Swann - Ganz monitoring has been advocated to detect and prevent hypotension. Prompt (within hours), thorough, radiographic evaluation should be performed to rural out ongoing neural compression and to asses the stability of the spine. Decompression of the cord and involved nerve roots and stabilization of the spinal column will prevent further mechanical injury and provide the proper milieu for healing. Currently, a National Institutes of Health (NIH) sponsored trial is about to commence to evaluate the role of early spinal cord decompression and stabilization in SCI.

B. Neuroprotective Agnets

Administration of acute high-dose steroids cord trauma have shown a clear benefit in two clinical trials. Specifically, two separate prospective, randomized, doubleblind studies demonstrated an imp[rovement in neurological outcome after spinal cord injury [92,93]. The first of these studies, in 1990, was the Second National Acute Spinal Cord Injury Study (NASCIS 2), a phase III study that demonstrated benefit from high-dose methylprednisolone.

1. Methylprednisolone

Based on encouraging result in animal models, Bracken et al. studied the effects of high-dose solumedrol (methylprednisolone), 30 mg/kg i.v. bolus followed by 5.4 mg/kg/hr for 23 hr on acute SCI from nonpenetrating trauma in a multicenter, placebo-controlled, double-blind, randomized, prospective trial, i.e., the North American spinal cord injury study, NASCIS 2.A third arm of this trial was designed to assess the efficacy of the opiate antagonist, naloxone. The results of 6-month and 1-year follow-up studies revealed significant improvement in motor and sensory function with methylprednisolone when compared to placebo in those patients who received methylprednisolone within 8 hr of injury [92,94,95]. Improvement in some patients with "complete" SCI was also seen. The minimal improvement seen in the naloxone-treated patients did not achieve statistical significance in the same trial.

Methylprednisolone in SCI has now become a standard of care. It should be pointed out, however, that the dose and duration of therapy have not been fully worked out in humans. The dosages used for NASCIS 2 were based on animal studies. The administration duration of 24 hr was selected to prevent the secondary infectious complications seen in NASCIS 1 where methylprednisolone was given for 10 days. A third trial, NASCIS 3 is designed to test the duration of therapy, comparing 23 versus 47 hr of continuous infusion at 5.4 mg/kg/h after the initial bolus of 30 mg/kg, and also to test the 21-aminostroid tirilazad against methylprednisolone alone. Further clinical studies are needed o determine the optimum does and duraion of therapy. The mechanism of protection from methylprednisolone is thought to be via a free radical scavenger effect, specially as a lipid peroxidaion inhibitor. Moreover, the inhibition of the arachidonic acid cascade may also play a role. This protective ability is independent of its glucocorticoid properties [40]. In fact, elevated blood glucose may be harmful to injured neurons, and a glucocorticoid effect is undesirable. Steroids without glucocorticoid properites, namely the 21-aminosteroids (lazaroids), show preservation of the neuroprotective properties and may be more potent.

2. GM-1 Ganglioside

A disinctly different compound, GM-1 ganglioside (GM-1), has been shown to improve the outcome in SCI in animal models. GM-1 is an extremely large molecular weight synthetic glycolipid, similar to those that are inserted into the cell membrance of spinal and brain neurons. They posses antioxidant properties and allow transmembrane communication. It is, however, uncertain as to whether the compound is taken up into injured cord tissue from blood after intravenous administration. In a prospective, randomized, placebocontrolled, double-blind phase II trial of nonpenetrating SCI known as the Marland trial, significant gains in neurological function were seen in some groups given GM-1 [93]. Specifically, useful recovery in intially paralyzed muscle groups were more likely in the GM-1 subjects compared to placebo. This compound is now being evaluated is now being evaluated in a very large, U.S. phase III trial, due for completion in 1997-1998. Although the exact mechanism of action is not know, GM-1 is though to act on the white matter tracts [96,97]. This is a novel concept because it implies glial protection rather than "neuroprotection" and the term "cytoprotective" may be more appropriate.

3. Other Agents

Many other agents are being investigated in the laboratory for use in human SCI trials. Excitatory amino acid-induced injury is also thought to occur in SCI, and therefore NMDA antagonists will be part of the Fourth NASCIS trial. Another arm of that study will target the inflammatory cascade by the use of cyclosporin.
Agents that act to decrease the metabolic activity of injured tissue have major limitations is SCI. specifically, barbiturates and etomidate depress brain function and respiration, which may be unacceptable in SCI patients with normal cognition. Hypothermia may also produce an awareness of the cold that may be unacceptable for the patient, although mild hypothermia is often observed in these patients, due to a disturbance in autonomic autoregulation of temperature. Genle correction of the hypothermia should be the goal with care not to induce hyperthermia. If hypothermia proves to be of substantial benefit in brain injury, than its use in acute SCI will need to the investigated.

C. Surgery

The role surgical decompression and stabilization in SCI remains controversial and, in particular, the timing of such procedures is a focus of heated controversy. A NIH-sponsored trial to address this question is about to begin in North America.

D. Penetrating Injury

Formal trials in neuroprotection following penetrating SCI have not been done. Future debate will concern the indications for the use of those agents tht are beneficial in nonpenetrating SCI in this population.

IV. Conclusions

Traumatic injury to the brain and spinal cord creates injury to tissue with little regenerative capacity. These injuries have traditionally been regarded as permanent. Elucidation of secondary injury mechanisms has erased the concept of completed, static damage in neurotrauma. Exciting breakthroughs in the laboratory have indicated that there is a potential for intervention in these injuries. Interventiom in the form of pharmacological agents is still in its infancy but has already proven beneficial in SAH, SCI, and, to a limited extent, in TBI.
Further understanding of the mechanisms of secondary injury after CNS trauma may provide new areas of treatment. The demonstration of programmed cell death - apoptosis - after experimental trauma has initiated interest in defining its role in human head and spinal cord injury. It is possible that gene-based therapy, designed to rescue these cells from "suicide," may be applied in the future.
Although the term "neuroprotection" is often though of in pharmacological terms, in a more general sense it includes several other important principless. First of all, the best protection is prophylaxis and prevention. Improved motor vehicle safety, air bags, and seat belt compliance have reduced the numbers of injuries. Education may help focus attention on prevention. This is the purpose of groups such s "Think First," an educational program sponsored by the Congress of Neurological Surgeons.

V. FUTURE DIRECTIONS: CNS REGENERATION AND TRANSPLANTATION

CNS regeneration will be a major focus of investigations to determine whether or not the very limited capacity of the CNS to regenerate can be improve. The first aspects of these studies have been to determine whether trophic factors, such as nerve growth factor (NGF) and brain-derived neurotrophic factor (BDNF), can improve histolofical or functional endpoints in laboratory models. Several such studies have been positive more recently. The NGF and BDNF genes have been transfected into head-injured rats using liposomes, and expression of the trophic factors has been shown [98].

Transplantation of fetal brain tissue in the setting of experimental SCI and concomitant head injury has provided some provocative and encouraging results [99,100], such studies are still far from clinical use, however, and as shown by Table 4, the current focus of trials for the next 3 years or so will be on ion channel active agents and inflammatory modulators. It is important to emphasize that improvement in the standard of care for head injury has achieved a 20% reduction in overall mortality rates for severe blunt head injury over the last two decades. It is thus imperative that the basic principles of acute trauma care be disseminated widely and applied to all patients before the newer neuroprotetant strategies are used. Only when this is done will the beneficial effects of such "potection" be clearly seen.

REFERENCES

1. A. Sauaia, F.A. Moore, E.E. Moore, K.S. Moser, R. Brennan, R.A. Read, and P.T. pons, Epidemiology of trauma deaths: a reassessment, J. Trauma 38:185-193 (1995).
2. D.M. Sosin, J.E. Sniezek, nd R.J. Waxweiler, Trends in death associated with traumatic brain injury, 1979 through 1992, JAMA 273:1778-1780 (1995).
3. P.C. Blumberg, N.R. Jones, and J.B. North, Difuse axonal injury in head trauma, J. Neurol. Nurosurg. Psychiatry 52:838-842 (1989).
4. R.P. Humphreys, E.B. Hendrick, and H.J. Hoffman, The head injured child who "talks and dies." A report of 4 cases, Childs Nerv. Syst. 6:139-142 (1990).
5. D.I. Graham, J.H. Admas, D. Doyle, Lschemic brain damage in fatal non-missile head injury, J. Neurol. Sci. 39:213-234 (1968).
6. D.I. Graham, J.H. Adams, and D. Doyle, G. M. Teasdale, A.E. Lawrence, and D.R. McLellan, Ischemic brain damage is still common in fatal non-missile head injury, J.Neurol. Neurosurg. Psychiatry 52:346-350 (1989).
7. M.J. Kotapka, D.I. graham, J.H. Adams, and T.A. Gennarelli, Hippocampal damage in fatal human hea injury: frequency and distribution, Acta Neuropathol. (Berl.) 83:530-534 (1992).
8. H. Fujisawa, D. Dawson, S.E. Browne, K.B. Mackay, R. Bullock, and
J McCulloch, Pharmacological modification of glutamate neurotoxicity in vivo, Brain Res. 629:73-78 (1993).
9. A. Marmarou, M.A.A. Foda, W. Van den Brink, J. Campbell, H. Kite, and K. Demetriadou, A new model of diffuse brain injury in rats. Part I: pathophysiology and biomechanics, J. Neurosurg. 80:291-300 (1994).
10. M.A.A Foda and A. Marmarou, A new model of diffuse brain injury in rats. Part II: morphological characterization, J. Neurosurg. 80-301-313 (1994).
11. F. Lnglis, Y. Kuroda, and R. Bullock, Glucose metabolism after acute subdural hematoma is ameliorated by a competitive NMDA antagonist, J. Neurotrauma 9:75-84 (1992).
12. H. Landholt, R. Bullock, H. Fujisawa, J. McCulloh, and S. Miller, Glutamate diffusion characteristics determine neurotoxicity in the rat brain: 14c glutamate autoadiorgraphy, J. Cereb, Blood Flow Metab. 13:S753 (1995).
13. R.L. Hayes, L.W. Jenkins, B.G. Lyeth, R.L. Balster, S.E. Robinson, G.L. Clifton, J.F. Stubbins, and H.F. Young, Pretreatment with phencyclidine, an N-Methyl-D-Aspartate antagonist, attenuates lon-term behavioral deficits in the rat produced by traumatic brain injury, J. Neurotrauma 5:259-274 (1988).
14. M.H. Chen, R. Bullock, D.I. Graham, J.D. Miller, and J. McCulloch, Ischemic Neuronal damage after acute subdural hematoma in the rat: effects of pretreatment with a glutamate antagonist, J. Neurosurg. 74:944-950 (1991)
15. Y. Kuroda, H. Fujisawa, S. Strebel, D.I. Graham, and R. Bullock, Effect of Neuroprotective N-methyl-D-aspartate antagonists on increased intracranial pressure: studies in the rat acute subdural hemaoma model, Neurosurgry 35:106-112 (1994).
16. B.G. Lyeth, L.W. Jenkins, and R.J. Hamm, Pretreatment with MK-801 reduces behavioral beficit following traumatic brain injury (TBI) in rats, Soc. Neuosci. Abstr. 15:1113 (1989).
17. B.K. Siesjo, Calcium and cell death, Magnesium 8:223-237 (1989).
18. D.W. Choi, M. Maulucci-Gedde, and A.R. Kriegstein, Glutamate neurotoxicity in cortical cell culture. J. Neurosci. 7:357-368 (1987).
19. R. Bullock, A. Stout, J.J. Woodward, et al., Excitatory amino acid release in severe human head trauma: the role of intracranial pressure and cerebral perfusion changes, Intracranial pressure, Vol.9 (H. Nagai, K. Kamiya, and S. Ishii, eds.), Springer-Verlag, Tokyo, 1994, pp. 264-267.
20. A.J. Baker, R.J. Moulton, V.H. MacMillan, and P.M. Shedden, Excitatory amino acids in cerebrospinal fluid following traumatic brain injury in humans, J. Neurosurg. 79:369-372 (1993).
21. R. Bullock, A.Zauner, J.S. Myseros, et al., Evidence for prolonged release of excitatory amino acids in severe human head trauma, Neuroprotective Agents, 1995, pp. 290-297.
22. T.H. Jones, R.B. Morawetz, R.M. Crowell, F.W. Marcoux, S.J. FitzGibbon, U.DeGirolami, and R.G. Ojemann, Thresholds of focal cerebral ischemia in awke monkeys, J. Neurosurg. 54:773-782 (1981).
23. J.H. Adams, D. Doyle, I. Ford, T.A. Gennarelli, and D.R. McLellan, Diffuse axonal injury in head injuyr: definition, diagnosis and grading, Histopathology 15:49-59 (1989).
24. E.H. Pettus, C.W. Christman, M.L. Giebel, and J.T. Povlishock, Traumatically induced altered membrane permeability: its relationship to traumatically induced reactive axonal change, J. Neurotrauma 11:507-522 (1994).
25. G.J. Wintemute, The relationship between firearm design and firearm violence, JAMA 22:1749-1753 (1996).
26. B. Karger, Pentrating Gunshots to the head and lack of immediate incapacitation. I. Wound ballistics and mechanisms of incapacitation, Int. J. Legal Med. 108:53-61 (1995).
27. D. Torbati, A. F. Jacks, M.E. Carey, J.F. Davidson, and J.B. Farrell, Cerebral cardiovascular and respiratory variables after and experimental barain missile wound, J. Neurotrauma 9(Suppl.) I:S143-155 (1992).
28. M.E. Carey, Experimental missile wounding of the brain, Neurosurg. Cline. North Am. 6:629-642(1995).
29. R.K. Kordestani, N.A. Martin, and D.Q. McBride, Cerebral Hemodynamic Disturbances following penetrating craniocerebral injury and their influence on outcome Neurosurg. Clin. North Am. 6:657-667 (1995).
30. A. Ward, Jr., Atropine in the treatment of closed head injury, J. Neurosurg. 7:398-402 (1950).
31. D.W. Marion, W.D. Obrist, P.M. Carlier, L.E. Penrod, and J.M. Darby, The use of moderate therapeutic hypothermia for patients with severe head injuries: a preliminary report, J. Neurosurg. 79:354-362 (1993).
32. G.L. Clifton, Hyponthermia in neurotrauma (abstract), J. Neurotrauma 12:365 (1995).
33. A.L. Wolf and A. Marmarou, Effect of THAM upon outcome in severe head injury: a randomized prospective clinical trial, J. Neurosurg. 78:54-59 (1993).
34. Braakman, H.J.A. Schouten, M.B. Dishoeck, and J.M. Minderhoud, Megadose steroids in severe head injury. Results of a prospective double-blind clinical trial, J. Neurosurg. 58:326-330 (1983).
35. J.P. Muizelaar, A. Marmarou, H.F. Young, S.C. Choi, A. Wolf, R.L. Schneider, and H.A. Kontos, Improving outcome of severe head injury with the oxygen radical scavenger polyethylene glycol conjugated superoxide dismutase: a phase II trial, J. Neurosurg. 78:375-382 (1993).
36. P.R. Cooper, S. Moody, and W.K. Clark, Dexamethason and severe had injury. A prospective double-blind study, J. Neurosurg, 51:307-316 (1979).
37. G. Faupel, H.J. Reulen, and D. Muller, Double-blind study on the effects of steroids on severe closed head injury, Dynamics of Brain Edema (H.M. Pappius and W. Feindel, eds.), Springer-Verlag, Berlin, 1976, pp. 337-343.
38. S.K. Gudeman, J.D. Miller, and D.P. Becker, Failure of high dose steroid therapy to influence intracranial pressure in patients with severe head injury, J. Neurosurg, 51:301-306 (1979).
39. T. Grumme, A. Baethmann, D. Kolodziejcyk, J. Krimmer, M. Fisher, B.V.E. Rothe, . R. Pelka, H. Bennefeld, E. Pollauer, H. Kostron, et al., Treatment of patients with severe head injury by triamcinolone: a prospective, conrolled clinical trial of 396 cases, Res. Exp. Med. 195:271-229 (1995).
40. E.D. Hall, Lipid antioxidants in acute central nervous system injury, Ann. Emerg. Med. 22:1022-1027 (1993).
41. J.S. Althaus, P.K. Andrus, C.M. Williams, P.F. VonVoightlander, A.R. Cazers, and E.D. Hall, The use of salicylate hydroxylation to detect hydroxyl radical generation in ischemia and traumatic brain injury. Reversal by tiilazad mesylate, Mol. Chem. Neuropathol. 20: 147-162 (1993).
42. E.C. Haley, N.F. Kassell, W.M. Alves, B.K.A. Weir, C.A. Hansen, and the Participants. Phase II trial of tirilazad in aneurysmal subarachnoid hemorrhage. A report of the Cooperative Aneurysm Study, J. Neurosurg. 82:786-790 (1995)
43. E.S. flamm, Free radicals in cerebral ischemia, Stroke 9:445-447 (1978).
44. The European Study Group in Severe Head Injury. A multicenter trial of the efficacy of nimodipine on outcome after severe head injury, J. Neurosurg. 80:797-804 (1994).
45. I. Bailey, A. Bell, and J. Gray, A trial of the effect of nimodipine on outcome after head injury. Acta Neurochir. 110:97-105 (1991).
46. J. Ohman, A. Servo, and O. Heiskanen, Long-term effects of nimodipine on cerebral infarcts and outcome after aneurysmal subarachnoid hemorrhage and surgery, J. Neurosurg. 74:8-13 (1991).
47. A. Kakarieka, A. Harders, R. Braakman,and the German tSAH Study Groups Traumatic subarachnoid hemorrhage: a clinical entity and its treatment with nimodipine (adstract), J. Neurotrauma 12:375 (1995).
48. A. Harders, A. Kakarieka, R. Braakman, and the German tSaH Study Group. Traumatic subarachnoid hemorrhage and its treatment with nimodipine, J. Neurosurg. 85:82 (1996).
49. B.W. Lin, W.D. Dietrich, R. Busto, and M.D.Ginsberg, (S)-emopamil protects against global ischemia brain injury in rats, Stroke 21:1734-1739 (1990).
50. M.B. Marinov, K.S. Harbaugh, P.J. Hoopes, H.J. Pikus, and R.E. Harbauh, Neuroprotective effects of preischemia intraarteri8al magnesium sulfate in revesible focal cerebral ischemia, J. Neurosurg. 85:117-124 (1996).
51. D. Nicholls and D. Atwell, The release and uptake of excitatory amino acids, Trends Pharmacol. Sci. 11:462-468 (1990).
52. R. Bullock and H. Fujisawa, The role of glutamate antagonists for the treatment of CNS injury, J. Neurotrauma 9:S443-S462 (1992).
53. O. Kempski, U. Von Andrian, L. Schurer, et al., Intravenous glutamate enhances edema formation after a freezing lesion, Advances in Neurology (D. Long, ed.), Raven Press, New York, 1990, pp. 219-223.
54. B.S. Meldrum, J.H. Swan, M.J. Leach, M.H. Millan, R. Gwinn, and G.M. Teasdale, Reduction of glutamate release and protection against ischemic brain damage by BW1003C87, Brain Res. 59:1-6 (1990).
55. W.L. Maxwell, R. Bullock, H. Landholt, and H. Fujisawa, Massive astrocytic swelling in response to extracellular glutamate - a possible mechanism for post-traumatic brain swelling? Acta Neurochir. (Wien) 60:122-127 (1994).
56. H. Benveniste, The excitotoxic hyponthesis in relaion to cerebral ischemia, Cerebrovasc. Brain Metab. Rev. 3:213-245 (1991).
57. R.C. Collins, B.H. Dobkin, and D.W. Choi, Selective vulnerability of the brain: new insights into the pathophysiology of stoke, Ann. Intern. Med. 110:992-1000 (1989).
58. B.G. Lyeth, L.w. Jenkins, and R.L. Hayes, The NMDA receptor. Theoretical and clinical implications of traumatic brain injury, Neuroehabilitation 50-57 (1991).
59. S.M Rothman and J.W. Olney, Excitotoxicity and the NMDA receptor, Trends Neurosci. 10:299-302 (1987).
60. Y. Katayama, D.P. Becker, T. Tamura, and D.A. Hovda, Massive increases in extracellular potassium and the indicriminate relesae of glutamate following concussive brain injury, J. Neurosug. 73:889-900(1900).
61. T. Kawamata, Y. Katayama, D.A. Hovda, A Yoshino, and D.P. Becker, Administration of excitatory amino acids antagonists via microdialysis attenuates the increases in glucose metabolism, J. Cereb. Blood Flow Metab. 12:12-24 (1992).
62. S.H. Graham, J.Chen, F.R. Sharp, and R.P. Simon, Limiting ischemic injury by inhibition of excitatory amino acid release, J. Cereb. Blood Flow Metab. 13:88-97 (1993).
63. J. Grotta, W. Clark, B. Coull, L.C. Pettigraw, B. Mackay, L.B. Goldstein, I. Meissner, D. Murphy, and L. Larue Safety and tolerability of the glutamate antagonist CGS19755 (Selotel) in patients with acute ischemic stroke. Results of a phase IIa randomized trial, Stroke 26:602-605 (1995).
64. P.A. Rozenberg and A. Aizenman, Hundred Fold increase in neuronal vulnerability to glutamate toxicity in astrocyte poor cultures of rat cerebral cortex, Neurosci. Lett. 103:162-168 (1989).
65. D.W. Choi, J.Y. Koh, and S. Petrs, Pharmacology of glutamate neurotoxicity in cortical cell culture: attenuation by NMDA antagonists. J. Neurosci. 8:185-196 (1988).
66. N. Shimada, R. Graf, G. Rosner, A. Wakayama, and W.D. Heiss, Ischemic flow threshold for extracellular glutamate increase in car cortex. J. Cereb. Blood Flow Metab. 9:603-606 (1989).
67. S.P. Butcher, R. Bullock, D.I. Graham, and J. McCulloch, Correlation between amino acid release and neuropathologic outcome in rat brain following middle cerebal artery occlusion, Stroke 21:1727-1733 (1990).
68. C.K. Park, D.G. Nehls, D.I. Graham, G.M. Teasdale, and J. McCulloch, The glutamate antagonist MK-801 reduces focal ischemic brain damage in the rat, Ann. Neurol. 24:543-551 (1988).
69. R. Bullock, D.I. Graham, S. Swanson, and J. McCulloch, Neuroprotective effect of AMPA receptor antagonist LY-293558 in focal cerebral ischemia in the cat, J. Cereb. Blood Flow Metab. 14:466-471 (1994)
70. M.R. Bullock, A. Zauner, o. Tsuji, et al., Patterns of excitatory amino acid release and ionic flux after severe human head trauma, Intracranial Pressure, Vol. 10 (H. Nagai, K. Kamiya, and S. Ishii, eds.) Springer-Verlag, Tokyo, 1995, pp. 64-71.
71. M.R. Bullock G.M. Teasdale, A. Kotake, et al., Experience wih the glutamate antagonist CGS19755 in severe human head injury: evidence for an ICP lowering effect, Intracranial Pressure, Vol. 9 (H. Nagai, K. Kamiya, and S. Ishii, eds.), Springer-Verlag, Tokyo, 1994, pp. 238-241.
72. D.W. Newell, A Barth, and A.T. Malouf, Glycine site NMDA receptor antagonists provide protection against ischemia-inuced neuronal damage in hippocampal slice cultures, Brain Res. 675:38-44 (1995).
73. D.S. Warner, H. Martin, P. Ludwig, A. McAllister, J.F.W. Keana, and E. Webar, In Vivo models of cerebral ischemia: effects of parenterally administered NMDA receptor glycine site antagonists, J. Cereb, Blood Flow Metab. 15:188-196 (1995).
74. F. Plum and A.G. Swanson, Barbiturate Poisoning treated by Physiological methods, with observations effects on beta, beta-methlethylglutarimide and electrical stimulation, JAMA 163:827-835 (1957).
75. H.L. Rosomof, Protective effects of hypothermia against pathological processes of the nervous system. Ann. N.Y. Acad. Sci. 80:475-486 (1959).
76. W.D. Dietrich, O. Alonso, R. Busto, M.Y. Globus, and M.D. Ginsberg, Post-traumatic brain hypothermia reduces histopathological damage following concussive brain injury, Acta Neuropathol. (Berl.) 87:250-258 (1994).
77. R.F. Spetzler and M.N. Hadley, Potection against cerebral ischmia: the roel of barbituraes, Cerebrovasc. Brain Metab. Rev. 1:212-229 (1989).
78. H.M. Shpiro, Barbiturates in brain ischemia, Br. J. Anaesth. 57:82-95 (1985).
79. T.M. Sundt, Jr., D.G. Piepgras, and M.J. Ebersold, Monitoring and protecting cerebral function in neurovascular surgery: rationale and techniques, Clin. Neurosurg. 29:390-416 (1982).
80. M.W. McDermott, F.A. Duriy, M. Borozny, and M.A. Mountain, Temporary vessel occlusion and barbiturate protection in cerebral aneurysm surgery, Neurosurgery 25:54-61 (1989).
81. H.M. Eisenberg, R.F. Frankowski, C.F. Contant, L.F. Marshall, and M.D. Walke, High-dose barbiturate control of elevated intracranial pressure in patients with severe head injury. J. Neurosurg. 69:15-23 (1988).

82. J.D. Ward, D.P. Becker, J.D. Miller, S.C. Choi, A. Marmarou, C.Wood, P.G. Newlon, and R. Keenan, Failure of Prophylactic barbiturate coma in the treatment of severe head injury, J. Neurosurg, 62:383-388 (1985).
83. J.C. Watson, J.C. Drummond, P.M. Patel, T. Sano, and #W.U.H.S. Akrawi, An assessment of the cerebral protective effects of etomidate in a model of incomplete forebrain ischemia in the rat, Nurosurgery 30:540-544 (1992).
84. I.M. Ledingham, and I. Watt, Influence of sedation on mortality in critically ill multiple trauma patients, Lancet 1:1270 (1983).
85. C.A.F. Tulleken, A Ven Diern, J. Jonkman, and Z. Kalenda, Clinical and experimental experience wih etomidate as a brain protective agent, J. Cereb. Blood Flow Metab. 2:S92-S97 (1982).
86. J.P. Muizelaar, The use of electroencephalography and bain protection during operations for basilar aneurysms, Neurosurgery 25:899-603 (1989).
87. H.H. Batjer, A.I. Frankfurt, P.D. Purdy, S.S. Smith, and D.S. Samson, Use of etomidate, temporary arterial occlusion, and intraoperative angiography in surgical treatment of large and giant cerebral aneurysms, J. Neurosurg. 62:234-240 (1988).
88. A. Tamura, T. Kuihar, and H. Nakayama, Treatment of severe head injury: a comparative study of nizofenone and barbiturate, Intracranial Pressure, Vol. 6 (J.D. Miller, ed.), SpringerVerlag, New York, 1986 pp. 774-778.
89. P.J. Kane, P. Modha, R.D. Strachan, S. Cook, I.R. Chambers, C.B. Clayton, and A.D. Mendelow, The effect of immunosuppression on the development of cerebral oedema in an experimental model of intracerebral haemorrhage: whole body and regional irradiation, J. Neurol. Neurosurg. Psychatry 55:781-786 (1992).
90. A.E. Biestro, R.A. Alberti, A.E. Soca, M. Cancela, C.B. Puppo, and B. Borovich, Use of indomethacin in brain-injured patients with cerebal perfusion pressure impairment: preliminary report, J. Neurosurg. 83:627-630 (1995).
91. R.M. Chestnut, S.B. Marashall, J. Piek, B.A. Blunt, M.R. Klauber, and L.F. Marashall, Early and lae systemic hypotension as a grequent and fundamental source of cerebral ischemia following severe head injury in the Traumatic Coma Data Bank, Acta Neurochir. (Suppl) 59:121-125 (1993).
92. M.B. Bracken, M.J. Shepard, W.F. Collins, T.R. Holford, W. Young, D.S. Baskin, H.M. Eisenberg, E. Flamm, L. Leo-Summers, J. Maroon, et al., A randomized, controlled trial of methylprednisolone or naloxone in the treatment of acute spinal-cord injury, N. Engl. J. Med. 322:1405-1411 (1990).
93. F.H. Geisler, F.C. Dorsey, and W.P. Coleman, Recovery of motor function after spinal-cord injury - a randomized, placebo-controlled trial with GM-1 ganlioside, N. Engl. J. Med. 324:1829-1838 (1991).
94. M.B. Bracken, M.J. Shepared, W. F. Collins, Jr., T.R. Holford, D.S. Baskin, H.M. Eisenberg, E. Flamm, L. Leo-Summers, J.C. Maroon, and L.F. Marshall, Methylprednisolone or naloxone treatment after acute spinal cord injury: 1-Year follow-up date. Results of the Second National Acute Spinal Cord Injury Study, J. Neurosurg. 76:23-31 (1992).
95. M.B. Bracken and T.R. Holford, Effects of timing of mehylprednisolone or naloxone administration on recovery of segmental and long-tract neurological function in NASCIS 2, J. Neurosurg. 79:500-507 (1993).
96. F.H. Geisler, F.C. Dorsey, and W.P. Coleman, Past and current clinical studies with GM-1 ganglioside in acute spinal cord injury, Ann. Emerg. Med. 22:1041-1047 (1993).
97. F.H. Geisler, GM-1 Ganglioside and motor recovery following human spinal cord injury, J. Emerg. Med. 11:49-55 (1993).
98. I. Mochetti and J.R. Wrathall, Neurotrophic factors in central nervous system trauma, J. Neurotrauma 12:853-870 (1995).
99. G. Sinson, M. Voddi, and T.K. McIntosh, Combined fetal neural transplantation and nerve growth factor infusion: effects on neurological outcome following fluid-percussion brain injury in the rat, J. Neurosurg. 84:655-662 (1996).
100. J.J. Bernstein and W.J. Goldberg, Experimental spinal cord transplantation as a mechanism of spinal cord regeneration, Paraplegia 33:250-253 (1995).

BACK

Neuroprotection in Subarachnoid Hemorrhage

BACK

I. Introduction

Subarachnoid hemorrhages (SAH) from rupture of an intracranial, saccular aneurysm occurs in 6 per 100,000 persons years1, and is an important cause of stroke in patients younger than 60 years2. Outcome after aneurysmal rupture is poor : in population based studies half of the patients die and half of the patients who survive the bleed remain disabled. In patients who survive the initial bleed, secondary ischemia is one of the major causes of death and disability4. Because this secondary ischemia develops in most patients 4 - 10 days after the onset of the hemorrhage, this subset of cererbal ischemia is unique in that neuroprotective treatment can be instituted before ischemia occurs.

II. Frequency And Clinical Characteristics of Secondary Ischemia After SAH

Secondary ischemia occurs in approximately 25% of patients who are admitted alive and in 40% of patients in whom the aneurysm is operated as soon as possible after admission5. In most patients the first signs of secondary ischemia occur 4 - 10 days after the initial hemorrhage. Secondary ischemia differs in several aspects from what is known as ischemic stroke, which is usually caused by thrombembolic events. First, in most patients secondary ischemia develops gradually whereas thromboembolic ishcemia is characterized by a sudden onset of symptoms. Second, in secondary ischemia after SAH focal signs are usually accompained by a decrease in level of consciousness, and in 25% of instance of secondary ischemia a decreased level of consciousness is the only sign. This impaired consciousness indicates that secondary ischemia is a multifocal or diffuse disorder. The multifocal or diffuse nature of the ischemia is confirmed by CT scanning or postmortem studies that show multivascular lesions in the brain of the vast majority of patients6.

III. Cause and Predictive Factors For The Development Of Secondary Ischemia

Despite many years of research the pathogenesis of secondary ischemia following SAH has still not been elucidated. Risk factors for the development of secondary ischemia are an impaired consciousness on admission, large amounts of extravasated blood, the presence of hydrocephalus, the restriction of fluid intake in case of hyponatremia, and treatment with antihypertensive drugs7-9. The sequence of events leading to secondary ischemia has not yet been elucidated, but it is generally held belief that after the hemorrhage a thus far undentified factor is released cerebral blood flow and thereby secondary ischemia. Experimental research has focused on identifying this factor derived from the blood that causes vasospasm.

Several observations in patients with SAH are not in accordance with this scheme. First, the presence of subarachnoid blood per se is not sufficient for the development of secondary ischemia : in patients with a perimesencephalic SAH, secodnary ischemia does not occur10. Second, in other causes of SAH, such as primary intracerebral hematoma or a ruptured arteriovenous malformation, the typical characteristics for secondary ischemia do not occur. Third, many patients with vasospasm never develop secondary ischemia.

These observations collectively suggest that not only the presence of subarachnoid blood per se but also the origin of the blood determines whether or not secondary ischemia will develop.

IV. Treatment Of Secondary Ischemia

Despite the lack of pathophisological insight, progress has been made in the prevention and treatment of secondary ischemia after aneurysmal SAH. This progress has been made by changes in general medical care as well as by specific drug treatment.

A. Calcium Antagonists

Initially, the rationale for the use of calcium antagonists in the prevention of treatment of secondary ischemia was based on the assumption that these drugs reduce the frequency of vasospasm by counteracting the influx of calcium in the vascular smooth muscle cell. This antivasospastic effect of calcium antagonists was confirmed by many in vitro studies with intracranial arteries and also by in vivo assessments of arterial lumen changes after experimental SAH.

Clinical trials have been undertaken with three types of calcium antagonists - nimodipine, nicardipine and AT877, of which nimodipine is the most extensively studied and used5. The methodologically best conducted trial on nimodipine was reported at the end of the 1980s11. This trial found a statistically significant, 40% reduction in the frequency of "poor outcome", but no reduction in the frequency of vasopasm. This lack effect on vasospasm despite improvement of outcome suggested that the beneficial effect of calcium antagonists might result from neuroprotection, not from prevention of vasoconstriction. Results from a North American trial on nicardipine showed somewhat conflicting results; in this trial a significant reduction in the frequency of vasospasm was observed, but not in the frequency of poor outcome12. This increased the confusion on the mode of action of calcium antagonists and again underlined the weak relation between vasospasm and outcome.

We recently performed a systematic reviews of all randomized controlled trials on calcium antagnoists in patients with SAH5. In this meta-analysis pooled data from trials on all three calcium antagonists, totaling 2434 randomized patients, showed a significant reduction in frequency of poor outcome, which resulted from a reduction in the frequency of secondary ischemia. When analysed separately, the nimodipine trials showed a significant reduction in the frequency of poor outcome, but the nicardipine and AT877 trials did not. On the other hand, nicardipine an dAT877 significantly reduce the frequency of vasospasm, whereas the nimodipine trials showed only a trend toward reduction of vasospasm despite a larger number of patients included in nimodipine trials than in nicardipine of AT877 trials.

From these data can be concluded that administration of nimodipine improves outcome in patients with SAH but that uncertainty remains as to whether nimodipine acts through neuroprotection, through reducing the frequency of vasospasm, or through both. Nicardipine and At877 definitely reduce the frequency of vasospasm, but the effect on overall outcome is unproven.

B. Neuroprotective Drugs Other Than Calcium Antagonists

Recently, tirilazad has been studied in two randomized, controlled trials. This drug belongs to the category of 21 aminosteroids that inhibit iron-dependent lipid peroxidation. In one of these trials, performed in European and Australian centres, the highest dose used significantly reduced the frequency of death and also, although to a much lesser extent, of poor outcome13. This beneficial effect was, however, observed only in men. The lack of effect in women was tentaively explained by the difference in steroid metabolism between men and women. However, the beneficial effect on poor outcome was not confirmed in the other, concurrently conducted, trial performed in North America. A formal meta-analysis did not reveal a beneficial effect on poor outcome in any dosage used, not even when men and women were analysed separately. The lack of effect in the North American trial might be related to the much higher proportion of patient treated with phenytoin, which induces liver enzymes and therefore steroid metabolism. Presently there is no convincing evidence supporting the use of tirilazad in patients with SAH, and further clinical trials after clarification of the pharmacokinetics in women and phenytoin users should be awaited.

Another hydroxyl radical savenger, nicaraven, was also recently tested in double blind, placebo-controlled trial14. This Japanese, multicenter study found that the drug decreased the frequency of secondary ischemia but did not affect overall outcome at 3 months after the hemorrhage.

Other neuroprotective drugs, such as glutamate antagonists, have not yet been studied in clinical trials of patients with SAH.

C. Other Interventions

As described earlier, treatment with antihypertensive drugs and fluid restriction were found to increase the risk of secondary ischemia. In a large consecutive series of patients admitted to a single institution, management with increased fluid intake and withholding antihypertensive drugs and fluid restriction resulted in a 50% reduction in the frequency of secondary ischemia, compared to a historical control group that was otherwise treated similarly15. Besides withholding antihypertensive drugs and fluid restriction, treatment with hypervolemia, hemodiultion, and induced hypertension, the so-called triple-H therapy, has become widely used, although evidence from clinical trials is still lacking. Presently, one randomized controlled trial is investigating whether aggressive induction of hemodilution and hypertension improves outcome in these patients.

Several other drugs or management strategies have been suggested, but none of these has thus far proven to be effective in properly designed, randomized clinical trials. All of these intervention aim to decrease the frequency of vasospasm after SAH, on the assumption that a decrease in vasospasm results in improved outcome. As described earlier, this assumption is at best weak, because many patients with vasospasm never develop secondary ischemia, and some patients with secondary ischemia have no vasospasm. A first type of drug is the endothelin antagonists. Endothelin is a potent vasoconstrictor, and its production is stimulated by oxyhemoglobin. In animal studies endothelin receptor antagonists underwent reverse vasoconstriction16. A second drug is calcitonin gene-related peptide, a potent vasodilator. In a randomized clinical trial, no effect of this drug was found17. A third strategy aiming to reduce the frequency of vasospasm is lysis of the intracisternal blood clot with intrahecally administered recombinant tissue plasminogen activator18,19 but the effectiveness of this strategy is not proven.
A last strategy, now being tested in a pliot study, is salicyclic acid in suppositories for patients who were operated within 3 days of the hemorrhage. The rationale of this therapy is that platelets are activated in all patients after SAH and that the activation is stronger in patients who develop secondary ischemia20. Moreover, Jvela found that patients who had taken aspirin before the hemorrhage occurred had a lower risk for secondary ischemia than patients who had not been taken aspirin21.

V. Conclusion

Secondary ischemia after subarachnoid hemorrhage constitutes an intriguing clinical subset of cerebral ischemia because neuroprotective treatment can be instituted before ischemia occurs. Despite the lack of understanding of its pathophysiology, secondary ischemia after subarachnoid hemorrhage is the first and thus far the only type of cerebri ischemia for which treatment with clinically significant effects has become available. The beneficial effect of the calcium antagonist nimodipine has been confirmed in systematic meta-analyses5,22,23. The pathways through which this beneficial effect is achieved are still unknown. Whether management with avoidance of antihypertensive drugs and fluid restriction will result in a significant decrease in the frequency of secondary ischemia remains to be proven ; evidence so far is based on observentional studies that used historical controls.

References :

1. F.H. H. Linn, G.J. E. Rinkel, A. Algra, and J. van Gijn, Incidence of subarachnoid hemorrhage : role of region, year and rate of computed tomography : a meta-analysis, stroke 27:625 - 629 (1996).

2. J.M. Bamford, P.A.G. Sandercock, M.S. Dennis, J.P.S. Burn, and C.P. Warlow, A prospective study a acute cerebrovascular disease in the community : the Oxfordshire community stroke project 1981 - 1986. 2. Incidence, case fatality rates and overall outcome at one year of cerebral infarction, primary intracerebral and subarachnoid hemorrhage, J. Neurol. Neurosurg. Psychiatry 53: 16 - 22 (1990)/

3. J.W.Hop, G.J.E. Rinkel, A. Algra, and J.van Gijn, Case fatality rates and functional outcome after subarachnoid hemorrhage : a systemic review. Stroke 1997

4. W.T. Longstreth, Jr., L. M. Nelson, T.D. Koepsell, and G. van Belle, Clinical course of spontaneous subarachnoid hemorrhage : a population - based study in King County, Washington, Neurology 43 : 712 - 718.

5. V. L. Feigin, G.J. E. Rinkel, A. Algra, M. Vermeulen, and J. van Ginjn, Calcium antagonists in patients with subarachnoid haemorrhage : a systemic review.

6. A. Hijdra, J. Van Fijn, S. Stefanko, K.J. Van Dongen, M. Vermeulen, and H. van Crevel, Delayed cerebral ischemia after aneurysmal subarachnoid hemorrhage : clinicoanatomic correlations, Neurology 36:329 - 333. (1986)

7. A. Hijdra, J. Van Gijn, N. Nagelkerke, M. Vermeulen, and H. van Crevel. Prediction of delayed cerebral ischemia, rebleeding and outcome after aneurysmal subarachnoid hemorrhage, Stroke 19: 1250 - 1256. (1988)

8. E. F. M. Wijdicks, M. Vermeulen, G. D. Murray, A. Hijdra, and J. Van Gijn, Volume depletion and natriuresis in patients with a ruptured intracranial aneurysm, Ann. Neurol. 18 : 211 - 216 (1985).

9. E. F. M. Wijdicks, M. Vermeulen, G. D. Murray, A. Hijdra, and J. van Gijn, The effects of treating hypertension following aneurysmal subarachnoid hemorrhage, Clin. Neurol. Neurosurg. 92:111 - 117 (1990).

10. G. J. E. Rinkel, E. F. M. Wijdicks, M. Vermeulen, D. Hasan, P. J. A. M. Brouwers, and J. van Gijn. The clinical course of perimesencephalic nonaneurysmal subarachnoid hemorrhage. Ann. Neurol. 29463 - 468 (1991).

11. J. D. Pickard, G. D. Murray, R. Illingworth, et al., Effect of oral nimodipine on cerebral infarction and outcome after subarachnoid hemorrhage : British aneurysm nimodipine trial, Br. Med. J. 298: 636 - 642 (1989).

12. E. C. Haley, Jr., N. F. Kassell, and J. C. Torner, A randomized controlled trial of high-dose intravenous nicardipine in aneurysmal subarachnoid hemorrhage. A report of the Cooperative Aneurysm Study, J. Neurosurg. 78:537 - 547 (1993).

13. N. F. Kassell, E. C. Haley, Jr., C. Apperson - Hansen, and W. M. Alves, Randomized, double blind, vehicle-controlled trial of tirilazad mesylate in patients with aneurysmal subarachnoid hemorrhage : a cooperative study in Europe, Australia and New zealand, J. Neurosurg. 84: 221 -228 (1996).

14. T. Asano, K. Takakura, K. Sano, et al., Effects of a hydroxyl radical scavanger on delayed ischemic neurological deficits following aneurysmal subarachnoid hemorrhage : results of a multicenter, placebo-controlled double-blind trial, J. Neurosurg. 84:792 - 803 (1996).

15. D. Hadan, M. Vermeulen, E. F. M. Wijdicks, A. Hijdra, and J. van Gijn, Effect of fluid intake and antihypertensive treatment on cerebral ischemia after subarachnoid hemorrhage, stroke 20: 1511 - 1515 (1989)

16. P.L. Foley, H.H. Caner, N. F. Kassell, and K. S. Lee, Reversal of subarachnoid hemorrhage induced vasoconstriction with an endothelin receptor antagonist, Neurosurgery 34 : 108 - 113 (1994).

17. European CGRP in Subarachnoid Hemorrhage Study Group, Effect of calcitonine - gene- related peptide in patients with delayed postoperative cerebral ischemia after aneurysmal subarachnoid hemorrhage, Lancet 339: 831 - 834 (1992).

18. J. Ohman, A. Servo, and O. Heikanen, Effect of intrathecal fibrinolytic therapy on clot lysis and vasospasm in patients with aneurysmal subarachnoid hemorrhage, J. Neurosurg. 75:197 - 201 (1991).

19. D. Stolke and V. Seifer, Single intracisternal bolus of recombinant tissue plasminogen activator in patietns with aneurysmal subarachnoid hemorrhage : preliminary assessment of efficacy and safety in a n open clinical study, Neurosurgery 30:877 - 881 (1992).

20. Y. Hirashima, S. Endo, M. Kurimoto, E. Tsukamoto, and A. Takaku, platelet - activating factor and antiphospholipid antibodies in subarachnoid haemorrhage, Acta Neurochir. Wien 128 : 144-149 (1994).

21. S. Juvela, Aspirin and delayed cerebral ischemia after aneurysmal subarachnoid hemorrhage, J. Neurosurg. 82 : 945 - 952 (1995).

22. R. Di Mascio, R. Marchioli, and G. Tognoni, From pharmacological promises to controlled clinical trials to meta-analysis and back : the case of nimodipine in cerebrovascular disorders, clin. Trials Meta-Anal. 29:57 - 79 (1994).

23. F. G. Barker, II, and C. S. Ogilvy, Efficacy of prophylactic nimodipine for delayed ischemic deficit after subarachnoid hemorrhage : a metaanalysis, J. Neurosurg. 84:405 - 414 (1996).

Carpal Tunnel Syndrome

BACK

Definition

A compression neuropathy of the median nerve at the wrist.

Epidemiology

• Incidence : 0.1% per year
• Prevalence : at least 2.1% of the population:
  - symptoms : 14.4% of the general population.
  - Clinically certain CTS: 3.8%
  - Neurophysiologic CTS: 4.9%
  - Clinical and neurophysiologic CTS: 27%
• Age : most common in middle age.
• Gender : M<F

Pathology

Focal demyelination of the median nerve in the carpal tunnel beneath the flexor retinaculum at the wrist with deformation of the myelin lamellae by the mechanical stress.

Pathophysiology

Chronic increased tissue pressure within the carpal tunnel, causing chronic focal compression of the medial nerve, with deformation of the myelin lamellae, and probably also nerve ischemia.

Etiology

• Idiopathic : 45%of cases of CTS.
• Obesity : present in 45 - 70% of cases of CTS
• Pregnancy, hormonal agents, and oophorectomy: present in 10% of cases of CTS.
• Active blue collar workers; hard work with hands and repetitive wrist movements and vibrations.
• Diabetes : present in 3 - 6% of cases of CTS.
• Rheumatoid arthritis and other inflammatory joint diseases of the wrist: 2- 5%
• Thyroid disorders: 3%
• Wrist fracture or sprain: 10%
• Carpal tunnel/hand surgery:1%
• Miscellaneous others:
  - Carpal bone hypertrophic osteoarthropathy.
  - Acromegaly
  - Flexor tenosynovitis
  - Anomalous flexor tendons.
  - Proximal origin of lumbrical muscles.
  - Aberrant muscles
  - Bifid median nerve.
  - Persistence of median artery and other vascular variants.
  - Cysts and tumors in the carpal tunnel.
  - Amyloidosis
  - Hemodialysis

Clinical features

Symptoms

• Paresthesia, with or without numbness and pain, involving the palmar surface of the hand and often extending proximally into the forearm, arm and even neck.
• The symptoms are frequently worse at night and after use of the hand, and characteristically are relieved by rapid flexion and extension movements of the wrist.
• Often bilateral, but symptoms are usually markedly worse on one side.

Signs

Median nerve provocative tests

• Tinel's sign
• Phalen's sign
• In most cases, these maneuvers are negative.

Severe nerve compression

• Sensory loss over some or all of the digits innervated by the median nerve
• Wasting and weakness of abductor pollicis brevis, flexor pollicis brevis, and opponens pollicis.

Differential Diagnosis

• Cervical radiculopathy
• Brachial plexopathy
• Proximal median neuropathy
• Diffuse peripheral neuropathy
• Motor neuron disease
• Spondylotic myelopathy
• Syringomyelia
• Multiple sclerosis.

Investigations

Nerve conduction studies
Ensure the temperature of the hands is at least 30 C

• Median nerve distal sensory latency
• Wrist-palm sensory conduction velocity
  
• Ulnar nerve distal sensory latency
  
• Sensitivity only about 70% in diagnosing CTS based on history and examination, and yet also a high rate of false positives.

MRI of the median nerve in the carpal tunnel

• May image the compressed median nerve.

Diagnosis

The key to diagnosis is a careful history and examination, including a search for the underlying cause. There is no consensus as to whether CTS is a clinical or electrophysiologic diagnosis, and no standard diagnostic criteria have been established. However, normal electrophysiologic studies do not rule out CTs. Indeed, carpal tunnel syndrome should be suspected on the history because it is not until the condition is moderate to severe that relevant sensory and motor deficits are detectable on physical examination.

Clinically certain CTS

• Recurring nocturnal and/or activity -related pain, numbness or tingling involving palmar aspects of at least two of the first four fingers at least twice weekly during the preceding 4 weeks.
• Positive nerve percussion and/or wrist flexion tests
• Median nerve sensory and/or motor deficit is supportive but not necessary for the diagnosis.

Electrophysiologic CTS
Median - ulnar peak sensory latency difference >0.8 ms.

Treatment

• Avoid or treat any precipitating/causal factors
• Limit repetitive wrist and hand movements.
  
• Achieve ideal body weight.
  
• Treat underlying disease
  
• Diuretics may be helpful, particularly for patients who develop CTS late in pregnancy.

Immobilization with splints
Plamar wrist splints, maintaining the wrist in a neutral or slightly extended position, and worn at night may be appropriate and helpful, particularly when symptoms are mainly nocturnal.

Local injection of corticosteroid and anesthetic in to the carpal tunnel
A single injection of 1 ml of fluid containing a long-acting corticosteroid and local anesthetic, between the tendons of the radial flexor muscle and the long palmar muscle, and angled acutely toward the carpal tunnel improves the symptoms of CTS in about three quarters of patients at 1 month after treatment, and is still effective at 1 year in half of patients. It is also safe; there are no adverse effects. The purpose of the lignocaine is to aford a painless injection and to ensure that the injection was carried out properly by demonstrating diminished sensation in the median nerve territory after the injection.

Ultrasound
20 sessions of ultrasound applied to the area over the carpal tunnel, on a once daily basis for the first 2 weeks, and then twice weekly for the next 5 weeks, may result in satisfying short and medium term benefits for patients with mild to moderate idiopathic CTS. If effective, the mechanism may be an anti-inflammatory effect of such treatment, or stimulation of nerve regeneration or nerve conduction.

Yoga-based intervention
A yoga-based intervention consisting of 11 yoga-postures designed for strengthening, stretching, and balanching each joint in the upper body, along with relaxation, given twice weekly for 8 weeks may be associated with pain reduction and improvement in grip strength and the Phalen sign.

Surgical decompression
Rationale
Aims to create the appropriate circumstances under which the nerve can recover, and thus the symptoms resolve. It does not aim to improve nerve function itself. The capacity of the nerve to recover also depends on the age of the patient, co-existing disease, and the severity and duration of the deficit.

Possible indications

• No improvement with conservative management
• Deterioration clinically or on serial electrodiagnostic studies
  
• Nerve conduction block localized on nerve conduction studies

Possible outcomes

• Immediate benefit should be realized if 'positive' symptoms dominate in a patient with neurophysiologically confirmed CTS.
  
• Improvement will be slow, or not at all, if the main clinical features are 'negative': axons regrow at only 1-2 mm / day. The main goal of surgery in these patients is prevention of further loss of function.

Adverse effects
Some patients complain of a vague lack of strength in the wrist following section of the flexor retinaculum.

Prognosis

• Depends on the cause and outcome of treatment
• An increasingly recognized cause of work disability.

BACK

Craniosynostosis

BACK

INTRODUCTION

Craniosynostosis is the early closure or the congenital absence of one or several of the cranial sutures. The simple concept that the suture closes and that this results in the cranial deformity has over the years been questioned, and with the new findings of genetic alteration in a number of children with both isolated and syndromic synostosis, this simple mechanistic explanation does not explain the underlying pathology. The progressive cranial deformity is explainable by restrictions on growth in the affected area, but the underlying cause of this growth restriction remains unknown. The absence of the suture is probably related to inherent growth problems in the specific tissues of that area of the clavarium. The most frequent genetic alterations that have been identified in synostosis affect the fibroblastic growth factor receptors in children with cranial facial syndromes, but other genes have been identified in single suture synostosis. The incidence of craniosynostosis is 1 in 3000 births and the most frequently involved suture is the saggital. Craniofacial syndromes occur in only 1 in 30 - 40 000 live births. The metopic suture is often closed or not present at birth, yet is rarely a cause of significant cranial deformity, and rigidity of the metopic suture is often found in one parent of an affected child. In familial cases the ridge is usually found behind the hairline and rarely requires therapy.

As a result of the current pediatric recommendations that babies should not sleep on their stomachs for the first 6 months of life, skull deformity related to postural influences has become almost endemic in the USA. The recommendations about sleeping position were based on studies showing a decrease in SIDS when front sleeping was avoided, and indeed the incidence of SIDS has decreased. Unfortunately, the aversion to front sleeping often results in the child lying on the back of the head until they develop to the point to being able to turn over independently. Many children have a preference for the side to which the head is turned, and this results in flattening of the occipital region on that side and a progressive parallelogram deformity of the cranium, with anterior displacement of the ipsilateral ear, bulging of the ipsilateral forehead and bulging of the ipsilateral cheek. The clinical appearance is unique and should not be confused with that due to craniosynostosis. The sutures are open and do not seem to close as a result of 'disuse'. This deformity can be prevented or minimized by constantly turning the baby from side to side and allowing the child to lie on the abdomen when a carer is present. In children with a tight sternocleidomastoid muscle physical therapy can also help to increase neck mobility and avoid deformity. Surgical correction is rarely required if the condition is correctly identified. The use of a molding orthosis has produced satisfactory improvement in the most affected children.

In cases of craniosynostosis surgery is indicated to correct the morphological deformity of the skull and to prevent or correct functional problems related to the synostosis. The morphological abnormalities relate to the distortion of the skull and face, and surgery is most successful when performed in the first year of life. The skull is still growing and the bone is quite mallelable, making correction of most of the deformity possible at a single operation. This should not be considered cosmetic, but has a clear functional component to maximize the social interaction of the child and to permit the use of safety devices such as bicycle and other protective helmets. Whether single-suture synostosis has any effects on cerebral development remains unknown. There is no clear differentiation between the morphological and the functional benefits of the surgery. A final reason to operate is the concern for elevated intracranial pressure, either in infancy or later in childhood. Several studies have reported elevated ICP in children with single - suture synostosis. The pressure is usually only moderately elevated and there is no evidence that this is harmful. It is common to see either an increased beaten-silver appearance or the absence of CSF over the cortex on CT scan or both in the area of the suture that is closed. It is logical to assume that the local pressure against the bone that is required to produce growth is going to be higher in areas of greater resistance. This does not imply that the pressure is in any way pathological or producing physiological changes. A few children with single-suture synostosis have been documented with papilledema, and therefore pathological elevations of pressure can occur but are rare. The primary reason for surgery in the first year of life in children with single-suture synostosis is because a better morphological result, i.e. a more normal head shape, can be obtained with less intervention. The risks of elevated ICP need to be discussed with the parents, as do the unknown risks of locally increased pressure, but these should not be used to force the parents towards surgery. The functional benefits of less testing, easier social interactions and easier helmet fitting should be stressed.

The children with cranofacial syndromes have a variety of functional problems, many of which require surgical correction. Thus in the children with craniofacial syndromes Carpenter's. Apert, Crouzon, Pfeiffer, Jackson - Weis and Saethre - Chotzen bring the commonest ones - multiple surgical proceudre are required and it takes a team approach to plan the timing and priorities. If surgery is carried out on the cranium in the first 6 months of life at least one, and often several more operations, are required. Although the evidence from Renier et al. suggests that cognitive function is better if surgery is performed in the first year life, the evidence is strong, and the current trend is to delay the primary cranial surgery as long as possible in the hope that, ultimately, fewer operations will be required and the cosmetic appearance will be better. Thus the decision to operate on the cranium is based on a drop off in cranial growth, increasing morphological abnormalities, such as turricephaly, or functional problems such as papilledema or corneal exposure. The initial cranial expansion may involve the anterior or the posterior cranium. Early surgery may be required to alleviate functional problems, e.g. tracheostomy to insure an adequate airway, or tarsorrhaphy to protect the cornea in cases of exorbitism. If there is hydrocephalus then a shunt is usually performed prior to any other cranial surgery.

DIAGNOSIS

The diagnosis is made on the basis of the child's history, any family history of similar facial features or surgery the synostosis in any family member, and the clinical examination. The primary factor in the latter is the shape of the calvarium and the facial bones. Total body examination is required to identify any other anomalies: finger or the fusions or alternations in shape: long bone anormalities; or skin abnormalities.

Suspicion of closed sutures can be confirmed by plan skull X-ray. Prior to surgery a CT scan is usually obtained, often with 3D reconstruction. The CT has a dual function. The first is to evaluate the intracranial contents, to insure that there are no major intracranial anomalies, to evaluate ventricular size and to insure which sutures are fused. The 3D CT adds information on the presence of hindbrain herniation and the thickness of the bone for splitting it also gives a better appreciation of the extent of the cranial deformity and an ability to better evaluate the degree of midfacial retrusion and orbital abnormality, but in children with sagittal synostosis it rarely adds useful information. Whenever there is a disturbance of facial growth the 3D CT is of value for planning the operation, and so in anterior plagiocephaly, bicoronal synostosis and the craniofacial syndromes it is recommended.

MRI is rarley used for presurgical evaluation. It is of value in the children at highest risk for intracranial anomalies. Saethre- Chotzen and Apert syndromes. Also, in children identified by reconstructed CT to have a hind brain anomaly, the MRI with CSF flow study across the foramen magnum and MR venography are the best ways to define and evaluate the degree of abnormality. MR venography is also important when there is any question of abnormal venous drainage, to insure that vital venous pathways are not interfered with during the surgical procedure.

Radiological studies of other areas of the body are performed depending on the clinical findings and diagnosis. These may include X-rays of the chest, spine, long bones, hands or feet.

LABORATORY INVESTIGATIONS

Simple tests are usually all that are required in infants. CBC and platelet count plus electrolytes are usually the only tests necessary. Up to 3 - 4 months of age PT and PTT are not very reliable and are rarely obtained unless there is a suspicion of abnormal clotting by history or in the family.

In children under 1 year of age the need for blood transfusion is 50 - 80%, depending on the surgical procedure, and appropriately typed and matched packed red cells should be available in 60 - 150 ml packs. Average tranfusion volumes in children under 1 year are 100 - 150 ml, either in the operating theater or in the pediatric intensive care unit. Although many parents feel inclined to use designated donor blood this is not usually advised by the blood bank, as there is little evidence that is safer and there is always the concern about sensitization, should interfamily organ transplantation become necessary. Transfusions given to children under 6 months of age usually have the cells irradiated.

In older children a more complete laboratory investigation may be required, including blood clotting studies, renal evaluation, nutritional evaluation and serum proteins, calcium and magnesium.

PEROPERATIVE DISCUSSION

The greatest threat to the patient's life and cerebral function is blood loss and shock. This, combined with intraoperative hypothermia, is the commonest event associated with poor outcome. In operations for sagittal synostosis the time when the patient shows signs of shock and hypovolemia is usually 1 - 2 hours after surgery, in the recovery room or the ICU. The infant should be observed in a monitored setting, with pulse oximeter and blood pressure recordings, for at least the first 2 postoperative hours. Many craniofacial teams will have all patients spend at least the night of surgery in the ICU. The risk, and the precautions taken to avoid it, need to be discussed with the parents. The risk of infection in the first operation is less than 1%. The risk of dural laceration is low and this is rarely a problem if the dura is repaired properly at the time of surgery. The risk of injury to the brain as a result of the use of cranitome or retraction should be very small, but each surgeon should be aware of his or her own percentage and use this figure when discussing this risk. Subdural bleeding can occur and, if recognized intraoperatively by dural discoloration, can be easily drained with no resultant problem. If not identified intraoperatively it can result in focal cerebral compression or elevated ICP, which is usually identified because the child a slow to wake up from the anesthesia. Other complications are rare, but include injury to the eyeball or optic other than to nerve II is rare. There is a risk of injury to the facial nerve if the lambdoid suture is opened too low into the mastoid, as can occur in saggital and lambdoid synostosis surgery. Finally, depending on the type of synostosis, the risk of a need for further surgery should be explained.

OPERATIVE PROCEDURES

ANESTHESIA

All the operations for craniosynostosis are performed under general anesthesia. Preoperative sedation is usually with oral Versed. Anesthesia is then induced with an inhlational agent and an i.v. agent started once the child in anesthetized. Usually two intravenous and one arterial line are placed. Surgery almost always exceeds 2 hours, and a Foley catheter is placed which also permits monitoring of the urine output. In children under 5 years of age the inspiratory gas is warmed and humidified and the i.v. fluids are warmed. Blood loss is often replaced with crystalloid until the major blood loss is complete. If there is nay sign of instability, e.g. blood pressure variation with respiration, then packed red cells are given. Despite an apparently normal temperature and pH in the OR it is common for the first blood gas measurement obtained in the ICU or recovery room to show a metabolic acidosis. In addition, the child's temperature can drop quite quickly when the surgical drapes and warming blankets are removed, which may add to the acidosis. Every effort should be made to keep the temperature up during extubation and transport to the ICU.

SAGITTAL SYNOSTOSIS

There are many operations for the correction of sagittal synostosis, ranging from simple strip craniectomy of the sagittal suture to total cranial vault reconstruction. There may be good results from strip craniectomy if it is done very early - i.e. within 2 months of birth - but in many children the craniectomy site will rapidly close and require further surgery, or the head will still be obviously scaphocephalic. As the major function of the surgery is to correct the morphology of the head, any surgery that does not achieve this goal has failed. There is no support for the interposition of Silastic or other material between the bone edges: this does not delay healing in most children.

The ideal age for correlation of sagittal synostosis is before 6 months of age. Up to this point the width of the skull base in the parietal area is normal; after 6 months this distance also becomes relatively narrow and, as the skull base cannot be expanded by surgery, the results are less good. Because of the weight of the child, concern for blood loss and the changing from fetal to adult hemogobin, surgery may be safest and the results excellent between 3 and 6 months of age. In children aged 5 or older it requires a very extensive operation to achieve what is usually a fairly modest improvement, and prolonged discussion should take place with the parents and the craniofacial team before surgery is suggested.

Most current operations are designed to acutely change the shape of the cranium by shortening it and widening the biparietal area. Despite the appearance of forehead bossing and bifrontal narrowing, it is rare that anthropological measurements support that the forehead is outside the normal range. Thus there is no need to do two operations, one on the posterior aspect of the cranium and one on the anterior aspect, in succession in children under 1 year of age. With the current cranial remodeling operations most children have a forehead that falls within the limits of normal measurements. In children 2 years of age or older correction of the forehead may be necessary either as the primary operation or as a second procedure, depending on the degree of posterior abnormality. Late surgery, i.e. after 2 years of age, usually focuses on correcting the frontal bossing.

The operative procedure should always be adapted to the particular patient because closure of the sagittal suture can result in a variety of cranial deformities. The most frequent is severe occipital beaking, with marked narrowing of the biparietal area. In this case surgery should remove the beaking, widen the biparietal region and shorten the calvarium. In other cases the beaking may be less marked but a saddle deformity occurs just posterior to the coronal sutures, in association with the lengthening of the cranium and parietal narrowing. In these cases it is necessary to alter the shape from the coronal suture posteriorly, and a slightly different technique may be required. Finally, if the major deformity is frontal bossing and parietal narrowing without occipital beaking, the operation may focus on the anterior two-thirds of the skull rather than the posterior two-thirds.

TECHNIQUE

Depending on the selected operation the patient may be positioned in the prone position in a horseshoe head rest, or in the sphinx position. The author rarely uses the sphinx position, but if it is to be used in children with syndromic disease the surgeon must be sure that there is no hindbrain herniation or bony instability that could lead to compromise of the medulla or upper spinal cord and resultant neurological damage. If the prone position is used there must be no pressure on the eyeballs, most of the pressure being taken on the forehead. It is not possible to remodel the forehead adequately in the prone position.

A coronal incision is made at approximately the interaural line, using low settings on the cautery with the Colorado needle. The use of the Colorado needle makes the incision almost bloodless, and no skin sutures or clips are required. There is also minimal blood loss during skin closure, as there are no clips to remove from the skin edges. The skin incision is made as a zig-zag rather than straight, as this markedly improves the cosmetic result by minimizing the visibility of the scar. The scalp flaps are reflected anteriorly and posteriorly in the subgaleal plane until the coronal sutures are exposed anteriorly and the posterior exposure is below the occipital bulge at the level of the torcula. In children under 6 months there is less bone bleeding if the pericranium is left attached and only freed along the proposed craniotmy cuts. In older children the pericranium can be reflected as flaps, to be replaced at the end of the procedure. In children with an open anterior fontanelle the epidural space is entered by freeing the pericranium at the margin of the fontanelle and doing the epidural dissection of the anterior region through this opening. This avoids the needs for anterior burr holes. In children with a closed fontanelle the burr holes are made at the level of the coronal sutures, 3 - 4 cm from the midline. The posterior burr hole is made over lambda, allowing access to free the dura from the sagittal and the lambdoidal sutures. The anterior and posterior burr holes are connected with the craniotome and a 6 - 8 cm wide portion of midline skull removed. Occasionally the sagittal sinus is partially enveloped by the bony closure of the suture and can be torn if the bone is not removed carefully while visualizing the suture and its contact with the bone.

The next step is to free the lambdoid sutures from the dural attachment and place burr holes bilaterally at the junction of the squamosal and lambdoid sutures. There is a large emissary vein just inferior to this point and the dura should not be dissected off the bone lower than the otherwise significant hemorrhage can occur. If bleeding is encountered the quickest and best way to control it is to place Avitene in the epidural space between the dura and bone with gentle pressure form a cottonoid. Attempts to coagulate this with bipolar cautery can enlarge the hole in the sigmoid sinus and produce more bleeding. The dura under the occipital bone is now freed down to the lateral sinuses. Although it is not necessary all the time, it is safer if the whole occipital bone is to be removed and replaced, to place a burr hole just superior to the torcula, below the occipital bulge, and to free the dura from medial to lateral. This lessens the risk of damaging the lateral sinuses and the torcula. Saw cuts are then made from the burr holes at the inferior margins of the lambdoid sutures, toward the torcula. If the occipital bone is to be removed and replaced the craniotomy incision is carried from the midline to each side and the bone removed. If the occipital bone is to be reshaped in situ then the cuts are begun laterally and brought to a point 1.5 - 2 cm lateral to the torcula. Saw cuts are made to remove the posterior margin of the parterial bones and the residual bone trimmed to remove any sharp edges that would be noticeable once the parietal bones are widened. Finally, cuts are made anteriorly from the inferior lambdoid burr hole just superior to or inferior tot he squamosal suture, to allow widening of the inferior portion of the parietal bone. These cuts extend 3 - 6cm.

The removed bone is then used to support the new cranial shape. In children under 1 year of age the bone is used to pull the reshaped occipital bone anteriorly to shorten the head. Because of the elasticity of the cranial bone this results in significant widening of the biparietal area. In older children the bone is usually placed at right - angles to the direction of the saggital suture, to maintain the width of the skull. Once hemostasis is second the scalp is closed with 4/0 Vicryl in the galeal layer and 5/0 Monocryl in the skin. No drains are left and no dressings applied other than Polysporin ointment.

There are now reports of the endoscope being used to perform sagittal craniectomies in younger patients, but results and benefits from this type of surgery are not yet available.

The time of greatest risk appears to be the first 2 hours postoperatively. We have looked at the pH and base excess in children in the OR just prior to closure and again on arrival in the pediatric ICU immediately at the end of the operation, and there is invariably a metabolic acidosis that occurs between the intraoperative blood gas and the ICU sample. This can worsen over the first 2 hours, probably because of clearance of acid and potassium from the muscles that occurs with warming. It is also possible that this relates to the release of cytokines and other inflammatory mediators as a response to the surgical trauma. Suffice to say that because of this delayed changed it is better to have the patient to either to the pediatric ICU or at least to stay in the recovery room for the first 2 postoperative hours. They should be monitored with BP, pulse rate and pulse oximetry. In addition, blood gas values to measure the degree of acidosis should be taken once per hour during this period. Any base excess >-3 is best treated with intravenous sodium bicardonate. Hemoglobin and hematocrit values should also be obtained on arrival and 2 hours prior to discharge from the recovery room, if the child is in that locale. Sudden onset of hypotension, hyperpnea and unresponsiveness can occur, and is usually the result of the combination of continued blood loss into the scalp plus metabolic acidosis. If this occurs in an unsupervised area the results can be death or severe hypoxic encephalopathy. This appears to occur in 1 - 12% of cases.

All the author's patietns go to the pediatric ICU overnight. A small percentage - 15 - 25% - will require a further blood transfusion in the ICU. In general, if the child's vital signs are stable and there is no acidosis we usually consider transfusion at a hemoglobin of 7 g or less. If there is continuing postoperative acidosis and instability of vital signs then transfusion of packed cells is given, regardless of the hemoglobin. In this latter case another reason for the continuing instability of the patient needs to be found and corrected.

The ICU stay is usually overnight and the child is usually ready for discharge by the fourth postoperative day, provided they are feeding well. The swelling does not subside for 8 - 12 days, and we no longer keep them in hospital until the eyes are open. Wound care consists of cleaning the incision with 0.5% hydrogen peroxide and applying Polysporin ointment every 8 hours for 48 hours, then simply washing the scalp with regular shampoo thereafter. The sutures dissolve in 10 days to 3 weeks. It is important to teach the parents to wash the incision to prevent the formation of scabs as the sutures dissolve. If scabs from the incision beneath them closes less well and leaves a wider scar. In families where the physician feels postoperative would care will not be adequate it is better to use staples or removable sutures.

ANTERIOR PLAGIOCEPHALY

The diagnosis is made by clinical examination. The findings are quite specific: the eyebrow is elevated on the side of the lesion, the lateral orbital rim is pulled posteriorly and superiorly, and the base of the nose is deviated to the side of the synostosis. There are a number of operative procedures reported for the correction of this morphological abnormality. Because of the high incidence of ophthalmological abnormalities in these children the surgery must address the shape of the orbit and the location of the eyeball, as well as cranial deformity. Moderate elevations of ICP have been reported in anterior plagiocephaly, but the operation is performed primarily to correct the functional problems related to eye movement and the morphological distortions of the skull. As in all forms of cranial synostosis the operation need to be modified for the individual patient. The decision to perform a unilateral or bilateral orbital advancement will depend on the amount of compensatory bulging that has occurred in the contralateral frontal region. In many children the forehead on the contralateral side and the position of the orbital bandeau are normal, and can be used as a template for the abnormal side. In these cases unilateral orbital advancement is all that is required.

SURGICAL PROCEDURE

This patient is positioned supine with the head in a head-holder that allows access to the whole calvarium, back to the lambdoid suture. A coronal zig - zag incision is made at the level of the interaural line. The scalp flap is peeled forward in the subgaleal plane until the orbital rims are visible. The temporalis muscle on one or both sides - depending on whether a unilateral or a bilateral bandeau is to be removed - is freed from the temporal bone in a subperiosteal fashion until the lateral orbit to below the orbitofrontal suture is exposed. It is at this point that a final decision about unilateral bandeau is required and the normal one can be used as a template.

The craniotomy flap is then outlined. This will usually cross the midline and is often bifrontal. In the child with an open fontanelle, which is rare in plagiocephaly, the periosteum is freed at the margin of the frontanelle and entrance to the epidural space obtained. Usually a burr hole is made at the inferior margin of the closed coronal suture, at the level of the lesser wing of the sphenoid bone. This allows the dura to be freed from the high - riding sphenoid wing and avoids dural lacerations in this area. If a bilateral flap is to be made then a second burr hole is placed at the inferior margin of the coronal suture on the normal side. If a central frontal burr hole is made it should be placed superior to the upper margin of the proposed bandeau. The dura is always tightly attached along the sutures and these areas are best freed by blunt dissection in the epidural space, prior of the use of the craniotome. This reduce the risk of a dural tear. The intraorbital contents are dissected free using a subperiosteal technique and endeavoring to keep the periorbita intact.

Once the frontal flap is removed an epidural dissection is made to expose the orbital roof, anterior and middle fossae and intervening sphenoid wing. The dura entering the foramen cecum is freed to allow access to the inferior potion of the frontonasal suture. Once this is freed the bandeau cuts can be right at the level of the nasion. The posterior limb of the bandeau must be taken back into the middle fossa for several centimeters, so that after the advancement there is still adequate bone on the bandeau to reattach to the parietal bone. The rigidity of the tongue-in-groove type advancement is improved by making a V-shaped cut into the parietal bone; this avoids the need for any plates or screws. The bandeau is freed by making cuts in the middle fossa, through the lesser wing of the sphenoid into the orbit, and then across the anterior aspect of the orbital roof to the midline. The lateral orbital cut is now made straight, rather than as a Z-plasty.

Once the bandeau has been removed the high sphenoid wing and the enlarged orbit can be addressed. Some of the thickened and elongated lesser wing is either rongeured down or cut with the saw and the bone preserved for use during reconstruction. The lateral aspect of the roof of the orbit is freed with a saw cut and the roof of the orbit lowered to decrease to add some thin bone to the intraorbital space to obtain a projection of the global that is symmetrical with the opposite side.

Before beginning the reconstruction the lateral canthus is picked up with a suture, as this needs to be reattached after the advancement is stabilized to prevent drooping of the lateral aspect of the orbital contents. The reconstruction is begun by recurving and shaping the bandeau prior to reinsertion in its new postion. When only one side is advanced this is easier than when both sides are involved, as in the latter case the ipsilateral orbital rim needs to be advanced and the curvature increased, whereas the contralateral side needs the curvature of the superior orbital rim to be reduced. To obtain the exact correction on both sides can be very trying and difficult. In children under 1 year of age the involved superior orbital rim should be advanced so that it projects slightly more than the contralateral side, as the growth on the involved side is likely to be slower than on the other. Once the orbital rim and orbital volume are corrected the forehead bone is replaced. The flattening of the ipsilateral frontal bone is such that the frontal flaps need to be switched, with the more normal contralateral bone being replaced on the side of the synostosis. In addition, it is almost always necessary to reshape the frontal bones and rotate them to obtain the best possible morphological result. The reshaping can be done using a variety of techniques, from radical cuts to simply bending with a Tessier bone bender. In older children multiple tongue -in groove cuts will allow some expansion of the bone, as well as reshaping, and permit complete calvarial closure. In children less than 1 year of age a bony gap is usually left between the replaced forehead bone and the residual parietal bone. Bone grafts, if available. Are used at a inferior margin of the advancement to support it, and the bone gap are left superiorly in the region of the anterior frontanelle.

Once the bone reconstruction is complete the lateral canthus is reattached using drill holes through the lateral aspect of the superior orbital rim, and sutured using 4/0 Monocryl suture. The temporalis muscle is now freed posteriorly by dividing the fascia and advanced anteriorly so that it can be reattached to the now advanced lateral orbital wall. If this is not done a cosmetically disfiguring temporal dent is left. The bone dust is used to fill in any irregularitis and the pericranium is approximated. The scalp is closed in two layers using 3/0 or 4/0 vicryl in the galea and 4/0 or 5/0 Monocryl in the skin. The hair is washed and Polysporin ointment applied. No drains are placed and no dressing are used.

BICORONAL SYNOSTOSIS

Most patients with bicoronal synostosis have a cranifacial syndrome. This affects the timing of the first surgery, as the assumption must be made that there is very little growth potential in the cranial vault and therefore the earlier the first operation the greater the number of operations that may be required. In addition, children with syndromic disease have abnormal facial growth, and therefore the cranial surgery has to be planned with a view to the later facial surgery and an integrated plan developed. The timing of the first operation depends on a number of factors. The functional problems are the major driving factors for early surgery: airway obstruction, severe exophthalamos or exorbitism. Kleiblastschaddel, pansynostosis with intracranial hypertension, or hydrocephalus. In the absence of any of these factors surgery is delayed as long as possible, at least to the end of the first year if feasible. The child is regularly checked for rate and shape of head growth, and papilledema. Our current approach is to postpone the first surgery in to the second or third year of life if possible, but this rarely occurs because either the head growth rate falls off, the head becomes too severely misshapen, or signs of increased intracranial pressure occur, usually manifest as papilledema. Ranier et al. have published evidence that supports performing the first cranial expansion by the end of the first year of life, based on cognitive test data comparing the outcome from surgery performed before and after 1 year of age. It is not clear that the results are purely the result of surgical timing, and their study was a retrospective review not a randomized trial. Further studies are required to help establish trial. Further studies are required to help establish whether their results truly support surgery prior to 1 year of age.

Current most children with syndromic synostosis still have their first cranial surgery performed in their first year. In the infant with syndromic synostosis it is increasingly common that the first operation is to expand the posterior cranial vault. This insures that the brain has adequate room to grow, relieves any intracranial hypertension, and still preserves the frontal bone and bandeau for later advancement. Cerebellar tonsillar herniation is common in Crouzon's disease, but can also occur in the other syndromes, including Saethre-Chotzen. It is still not clear when the herniation occurs. Cinalli et al.have postulated that it is a result of the closure of the lambdoid sutures, common in Crouzon syndrome, and is progressive, beginning in the first year of life. It is at least theoretically possible that posterior decompression may prevent the development of cerebellar tonsillar herniation.

TECHNIQUE

The bifrontal craniotomy and orbital advancement is similar to the techniques described for unicoronal synostosis. The major differences are that the surgery is always bilateral, the advancement is usually greater, and there is usually a need to bone-graft the nose. The coronal graft the nose. The coronal zig-zag incision is positioned well behind the hairline and the inferior limbs are directed to the top of the ear. The size of the bifrontal bone flap is dependent on the shape of the forehead, the height of the skull and the presumed need for split cranial bone for reconstruction. The child usually has a tall brachycephalic skull, and to maximize the morphological improvement the skull vault must be lowered as well as the bifrontal advancement performed. In re-operations the bandeau is often quite deformed and deficient, so that a complete new one has to be made. The best curvature for a new bandeau is usually at the level of the midparietal area. In the young children this is usually full-thickness bone; by age 5 and older split cranial bone can sometimes be used, depending on the thickness of the calvarium.

After removing the frontal bone flap an epidural dissection is performed over the roofs of the orbit and into the middle fossa. The frontal fossa is shortened in its AP diameter and the lesser wings of the sphenoid are elevated. In the frontal fossa it is important to free dura from the foramen cecum to allow the nasal cut to be made as low as possible in the frontal fossa. The frontal fossa is deep at the level of the crista galli, and care must be taken not to tear the very thin dura that covers the olfactory rootlets. Once the dura of the foramen cecum is free the crista galli can be exposed to obtain more retraction before the frontal saw cut is made. The dura must be freed from the lesser wing of the sphenoid bone and the frontal aspect of the middle fossa to prevent damage to the dura and/or the temporal tip, as the saw cuts to free the bandeau are made. This is especially true in the syndromic diseases, where the temporal tip is often at the level of, or anterior to, the lateral orbit. The bandeau can then be cut with the least amount of frontal lobe retraction necessary to permit the entrance of the saw blade. The anterior inferior cut of the bandeau enters the posterior lateral aspect of the orbit just at the junction of the lesser wing of the sphenoid bone and the lateral wall of the orbit. As this cut is made the orbital contents must be seen and protected by a malleable retractor to prevent injury. The anterior midline cut is made at the level of the foramen cecum and angled superiorly to enter the most inferior aspect of the frontal fossa. The superior cut for the bandeau is then made across the anterior aspect of the roof of the orbit, across the sphenoid wing and posteriorly in the middle fossa. The classic incision in the lateral orbital wall has been Z-plasty, but we now use just a straight incision, as low on the lateral wall as necessary to achieve the desired advancement. Once the bandeau is reshaped and replaced. Usually with 2/0 PDS suture, the frontal bone is then contoured to fit and reattached. Whenever the periorbita is freed below the level of the orbitofrontal suture, the lateral canthi need to be reattached to prevent orbital dystopia. This is done as described in the section on plagiocephaly.

The degree of lowering of the vault depends on the extent of the frontal advancement. When the second, more posterior, craniotomy flap is raised the burr holes are placed in the midline over the sagittal sinus to insure that the sinus is not encased in bone and to prevent it being torn. It is safer to make two lateral burr holes as well, to be sure the dura is freed and not accidentally cut by the craniotome. Troublesome bleeding in this area is usually not from the sinus but from the dural granulations, and is best controlled with Avitene and gentle pressure, as excessive coagulation of the granulations can result in increased rather than decreased bleeding. The vault is lowered by removing bone from the inferior margins of the parietal flap and then suturing the bone flap to the inferior portion of the parietal bone. It is usually possible to split some of the cranial bone, thereby obtaining additional bone for reconstruction. In children over 1 year of age every effort is made to completely reconstruct the calvarium using whatever bone is available. The whole surgical site is then irrigated with saline an dbacitracin solution. The bone dust from the opening is used to smooth out any defects. The scalp is closed with 3/0 or 4/0 Vicryl sutures in the galeal layer and usually 5/0 Monocryl for the skin. After washing the hair polysporin ointment is applied. No drains are left and no dressings applied.

TRIGONENCEPHALY

This severe abnormality of the anterior cranial vault is associated with metopic synostosis, but has also been reported in a number of chromosomal abnormalities, and so careful examination of the whole child is required. The incidence of developmental delay is greater in children with trigonencephaly then with other single-suture synostoses. The association of intracranial abnormalities, especially absence of the corpus callosum, should be recognized on CT scan, and in children where there is any question of intracranial anomaly an MRI should be performed.

SURGERY

The surgical technique is very similar to that for bicoronal synostosis, with removal of a bifrontal craniotomy flap followed by removal of the bandeau. In this case the triangular - shaped bandeau has to be reshaped by flattering and widening. This usually requires drilling of the inner table of the bandeau at a number of sites to permit the expansion. The degree of advancement is great because of the widening, and it is important to take the limbs of the bandeau well into the parietal area so that adequate contact can be made with the residual cranium to obtain rigid fixation. If the bandeau is weak centrally then a bone graft can be used to reinforce it, rather than a plate. The frontal bones require to be rotated and expanded by relaxing cuts, to widen them sufficiently to cover the widened bandeau.

POSTERIOR DECOMPRESSION

In infants with craniofacial dysostosis and restricted cranial growth it is increasingly common to perform a posterior decompression as the first operation. This is also the case in infants with kleiblatschaddel. This posterior correction allows the brain to expand, relieves any intracranial hypertension and still preserves the anterior bone for a frontal advancement at a later date. The technique is similar to that described for the frontal procedure, and is performed with the child in the prone position. The stealth incision is made at the intra-aural line and the whole of the midline occipital musculature. If the hindbrain hernia is to be decompressed at the same session the muscle incision is opened in the midline to expose the foramen magnum and the arch of C1. When this is the first procedure the child frequently has severe cranial lacunae and its very difficult to dissect the posterior dura from the bone.

SUMMARY

The main objective is to establish the correct diagnosis before any surgical intervention. Positional deformities rarely require surgery. Single-suture synostosis has a different outcome from multiple-suture synostosis in terms of both the esthetics of the correlation and the need for repeat surgery. Children with craniofacial dysostosis require a whole team approach in preparation for the multiple operations, including facial surgery, that will be required. The surgical technique is designed to maximize the morphological and functional benefits of the operation with the least blood loss and without injury to cerebral structures. Blood loss and hypothermia are the two major intraoperative risks, and avoidance or preparedness to correct these should be part of the surgical planning. Postoperatively the early risks are acidosis and hypotension, continued blood loss under the scalp flap, blood clotting problems such as diffuse intravascular coagulopathy, and brain swelling. The later problem are CSF leakage, infection and continued morphological abnormalities.

BACK

Intracranial Pressure Monitoring : Operative Procedure
and Technical Considerations

BACK

INTRODUCTION

Since the initial measurement of intracranial pressure using lumbar puncture in the 1900s by Queckenstedt, and more recent manometric methods of continuous ICP monitoring by Guillaume and Janny, routine monitoring of ICP has been performed both in the operating theatre and postoperatively in the intensive care unit. Lundberg was the first to make continuous recordings of intracranial pressure in the normal and pathologic stage using ventriculostomy, and is credited for much of our understanding about ICP waveform patterns. It was recognized that ICP can be monitored prospectively and used to manage patients effectively. Miller et al identified that ICP monitoring and management of elevated ICP in neurosurgical intensive care was important in brain injury. Today ICP monitoring via pressure transducer or fluid - coupled techniques is routine.

The intention of this chapter is to briefly review the pathophysiology of ICP, to enumerate the types of monitoring techniques and their unique properties and pitfalls, the indications for monitoring, and to discuss recent controversies regarding ICP monitoring.

PHYSIOLOGY OF INTRACRANIAL PRESSURE

Intracranial pressure is a unique physiologic variable owing to the rigid properties of the skull and the incompressible nature of the three intracranial components. Any change in the volume of one of these components necessitates a change in volume of one or more of the others. This is the basic tenet of the Monro-Kellie doctrine, which is applicable in most neurosurgical cases, with the exception of situations in which the skull has been removed and the dura is open. The intracranial compartment volume consists of 100 - 150ml CSF, 100 ml blood, and 1000 and 1200 ml brain. The modes of altering CSF volume entail drainage or limitation of production. Cerebral blood volume is contained in a variety of vessels, including capillaries, arterioles, venules and large arteries. Typically cerebral blood volume is contained largely in the venous system, with less than 20% contained within the arteriolar compartment. Changes in PCO2 and blood pressure and brain metabolism can lead to changes in cerebral blood volume, and hence ICP. Brain parenchymal volume can vary as a result of tissue edema. Surgical excision, the presence of a mass or underlying atrophy. As tissue edema develops intracellular and extracellular volume may increase the volume of tissue within the calvarium, and thus increase intracranial pressure. The two types of edema that are often active in pathophysiologic situations include cytotoxic edema and vasgenic edema. These two types coexist, however, the relative contribution of this tissue edema to increased intracranial pressure has not been fully evaluated. It is important to recognize that tissue edema may create ICP gradients, globally increased ICP, or both, but that the volume of edematous tissue necessary to raise intracranial pressure is unknown.

The derived value of cerebral perfusion pressure - is an important variable to introduce. The CPP, derived from the ICP, is often used as a goal of management. In situations of preserved pressure autoregulation, actual cerebral blood flow will remain unchanged at CPP from 50 to 150 mmHg. CPP values below 60 mmHg are associated with global brain ischemia in humans. In situations of altered pressure regulation, cerebral blood flow is somewhat linearly related to CPP, and thus maintaining a desired CPP may avoid unwanted reductions in cerebral blood flow.

AIMS OF THE OPERATION

The aim of the operation is to insert a reliable, accurate ICP monitor into the central nervous system without causing tissue injury. A second aim is to introduce a therapeutic option for increased ICP, namely CSF drainage. With newer monitors capable of measuring tissue temperature or tissue oxygen content, monitoring of these two parameters in another aim.

INDICATIONS

There are several indications of monitoring ICP, many of which are not mutually exclusive

• The monitor parenchymal integrity in the comatose patient suspected of increased ICP due to the presence of a mass lesion of diffuse brain edema. Increased ICP may be the only sign that tissue is at risk for herniation or ischemia when the clinical examination is very poor and difficult to quantify. Examples of such conditions are intracranial hemorrhage, subarachnoid hemorrhage, hydrocephalus, brain trauma, meningitis/encephilitis, hepatic encephalopathy, brain tumors and status epilepticus.
• To determine the optimal hemodynamic state of the patient in coma. In essence, to find the optimal cerebral perfusion pressure one needs to measure the ICP.
• To determine prognosis in comatose states in which ICP and CPP are suspected to be abnormal, but in which the degree of abnormality is unknown and malignant intracranial hypertension would after ICU or surgical treatment strategy. Examples of this include hepatic encephalopathy, in which ICP is used to determine candidacy for liver transplantation.
• To assess low ICP syndromes, such as CSF leaks after craniotomy, or over functioning ventriculoperitoneal shunts with siphoning effects.
• To monitor for adverse events in brain-injured patients undergoing general anesthesia. General anesthetics may decrease mean arterial pressure and hence CPP, alter cerebral blood volume, or mask intracranial hemorrhage or brain edema that may increase ICP. Thus for a brain-injured patient requiring an exploratory laparotomy an ICP monitor is required.
• To facilitate brain relaxation during craniotomy. Such as for clipping of a ruptured cerebral aneurysm. The careful use of ventriculostomy to lower ICP, allowing exposure of the aneurysm, permits less direct brain retraction. Care must be taken not to lower ICP dramatically, thus precipitating re-rupture of the aneurysm.

PREOPERATIVE TESTS AND SPECIAL CONSIDERATIONS

Informed consent from the patient or next of kin is required in all cases, except for life-threatening situations of impending herniation where a delay in the opeartion may cause irreversible brain injury. Prior to ventriculostomy, but not necessarily before parenchymal monitor insertion, it is required that an image of the brain and CSF space be reviewed to determine the size, shape and shift of ventricles; the presence of midline shift; possible entrapment of ventricles; the presence of a mass lesion that may obstruct the route of access of the ventriculostomy, or may bleed during insertion. It is also important to determine whether CSF drainage is therapeutically required. If not, one may consider placing a parenchymal monitor instead. Coagulation studies, including platelet count, prothrombin time, international normal ration and partial tissue thromboplastin are normal. In circumstances of hepatic failure with altered PT/INR, a bleeding time may be necessary. If coagulation parameters are abnormal, correction of these abnormalities with influsions of platelets, fresh frozen plasma and/or cryoprecipitate is needed before beginning the procedure. In addition, an abnormality in coagulation function may require the insertion of the monitor to be performed in the operating room under direct visualization. The patient must also be given an anticonvulsants, although no clear studies have found that boluses of an anticonvulsant prior to the operation are necessary. We would not recommended delaying the operation of bolus an anticonvulsant owing to the minimal risk of seizure during insertion, the time delay during the bolus, and the hypotensive effect of anti-convulsant, which may adversely affect CPP.

Selection of the site to monitor is a crucial step in preparation. Usually the right frontal approach is chosen to avoid trauma to the left dominant hemisphere. This is true to ventriculostomy and parenchymal monitors alike. However, other considerations, such as the site of pathology, direction of midline shift, entrapment of one lateral ventricle, prior or future craniotomy site, location of mass lesion or ICH are important, and may affect the side selected. In the setting of a trapped ventricle with lateral shift there are no clear data to assist in selecting the side to enter, although theoretically reduction of pressure on the side towards which the brain is shifting could increase rather than decrease the herniation. Thus one should plan the degree and extent of CSF drainage and pressure reduction prior to selecting the side, and be cognisant that the ventriculostomy may worsen the condition.

ANESTHETIC CONSIDERATIONS

Most patients requiring ICP monitor placement will be in coma as a result of their disease, or will be under sedation as a prerequisite for general anesthesia. In patients who are awake, or who, despite being in coma, are moving, sedation with a short-acting benzodiazepine, such as 4 mg midazolam i.v., is sufficient to suppress unwanted head movement and anxiety. The use of analgesics, such as 2 - 4 mg morphine sulfate i.v. may augment the sedative effect of midazolam and the local anesthetic effect. Rarely. In the intubated patient neuromuscular blockade is required to completely stop head movement. In this situation, vecuronium 10 mg. I.v. is very useful in providing 30 min of neuromuscular junction blockade, and allows for insertion of the monitor in the best state. Vecuronium does not alter ICP or MAP and is a safer alternative to pancuronium and succinylcholine.

OPERATIVE TECHNIQUE

VENTRICULOSTOMY

The operation to place a pressure monitor has several identical technical steps that are common to both pressure transducer and fluid-coupled techniques. We will begin with placement of the ventiruclostomy. Positioning the patient's head is critical to success. (1) Place the patient squarely on his or her back and elevate the bed 20 - 30. (2) Position the head in the midline so that both ears are at equal height above each shoulder. (3) The patient's nose should be exactly perpendicular to the bed. After selecting the appropriate side, an area 2 cm anterior to the coronal suture and 3 cm lateral to the sagittal suture is manually palpated. This should correspond to the midpupillary line, or 13cm posterior tot he eyebrow. The site of insertion is determined and the hair overlying an 8 cm diameter area of the site is shaved and the site washed with chlorhexidine. It is important to shave the area used to tunnel the ventricular catheter later in the operation. Subsequently the area is prepared in a sterile fashion with Betadine and draped with three drapes aligned in a triangular array, with the center drape perpendicular to the midline. Injection of the target scalp and periosteum with 1% lidocaine containing epinephrine in a 3 cm diameter circle is performed, and the lidocaine allowed to take effect. Using a no. 11 scalpel a 1.5 cm incision, aligned peripendicular to the sagittal suture, is performed and the overlying fascia bluntly dissected using a hemostat. A hand drill with an 11/64 - in bit is selected and, using a drill guard sleeve, a burr hole is drilled aiming peripendicular to the cranium towards the inner canthus and towards the tragus of the ipsilateral ear, which will be referred to as 'the approach' from here on. Alternatively, the Ghajar guide may be used to facilitate the approach.

Drilling is performed while an assistant holds the head to keep it steady. Once resistance to the drill increases the outer skull cortical table is perforated and drilling should be slowed and performed in short turns until through the outer table. The speed can then be increased until the inner table is encountered, and again drilling is slowed until the inner table is perforated. The drill is withdrawn and the dura pierced using a ventricular cone needle or a no. 15 scalpel. Alternatively this may be performed with an 18 guage needle. Saline is used to wash away bone fragments and avoid pushing them into parenchyma during catheter insertion. The monitor of choice is then inserted. For ventricular catheters, the catheter, with stylet inserted, is passed in the same x, y and z planes used to drill the burr hole, to a depth 5 cm below the scalp. This corresponds to the first marking on the catheter being just below scalp surface. The catheter should not be advanced beyond this 7 cm if the ventricle is not entered. If a second pass is required, a more medical approach is often necessary and one should aim for the bridge of the nose.

Selecting a more posterior approach should be the next step. More than three passes should not be attempted and one should confirm the landmarks and then drill a new for 5 - 8 cm to exit through a second 1.5 cm incision. One may use a special trocar to tunnel the catheter, or simply a kelly clamp. The catheter is secured to the entry site with an anchoring stitch and also at the exit incision. The catheter is then secured to the fluid collection system and a security stitch placed to couple to the Luer lock. The distal end of the catheter is then secured to the scalp by a third anchoring stitch. The two incision sites are dressed to avoid colonization of the operative site. The catheter is zeroed after insertion, and then periodically every 4 - 6 hours.

PARENCHYMAL MONITORING

Parenchymal monitor placement entails all the initial steps of drilling and perforating of the dura. However, the next step involves placing the skull bolt system, which is inserted by manually twisting the bolt into the skull up to its base. Prior to insertion the probe is then sterilely zeroed by gingerly twisting the set screw. The probe is inserted 2 - 3 cm until there is a steep rise in ICP and then withdrawn slightly until a stable reading is obtained. On some types of probes there are marking indicating the proper depth of insertion. The O - ring locking screw is tightened to prevent catheter movement and then the covering housing is snapped over the beveled end of the bolt. Some types of pressure transducer parenchymal monitors need to be zeroed prior to insertion whereas others do not. Specific instruction from the manufacture is required prior to insertion.

After completing placement of the monitor the bolt is wrapped in multiple layers of sterile Vasline gauze, followed by a sterile dressing. A cone of sterile dressing is made that ascends the catheter in order to reinforce the pliable parenchymal monitor and prevent damage caused by lateral torsion. No further zeroing may take place while the monitoring is in place.

SUBDURAL MONITORING

This technique involves placing a fluid-coupled or pressure transducer monitor under the dura, and is used in the setting of elective craniotomy or for hepatic encephalopathy. If an existing craniotomy is present, select an exit site in the scalp and use o trocar to puncture the skin. Next the monitor is zeroed and slipped gently in through the tunnel created by the trocar: one must place the ICP monitor, typically a pressure transducer type, under the dura prior to closure. A 3/0 nylon stitch is used to close the dura surrounding the monitor and the monitor is anchored to the scalp at an oblique angle. Often the skull must be fashioned to create a beveled edge, allowing the catheter to exit the skull at an oblique angle. This is accomplished by using a Midas Rex to bevel the inner table of the skull. The monitor is then secured to the scalp with an anchoring stitch after wrapping the probe four times with the distal end of the anchoring stitch.

If a craniotomy has not been performed, a 3 cm ovoid burr hole must be fashioned using the Midas Rex. Drilling should be continued until a thin rim of inner table remains; this is then removed. The posterior edge of the outer table is fashioned with the Midas to create a notch through which the monitor may pass with a shallow approach angle. The dura is then incised using a no. 15 scalpel and gently reflected. The pressure transducer is zeroed if necessary and the dura elevated with a no. 3 Penfield and the pressure transducer advanced under direct vision of the cortical veins. Care must be taken not to lacerate these during insertion. The dura and scalp are then closed in succession. Anchoring stitches are used to secure the monitor.

SPECIAL CONSIDERATION

WAVEFORM ANALYSIS

Once the monitor is placed careful waveform analysis is required to insure that the ICP is both valid and reliable. The typical ICP waveform has three wavelet components, with the first being the largest in amplitude. Damping of the wave form, the absence of wavelets and lack of reaction to measures known to increase ICP are indications of monitor malfunction. An increase in the amplitude of the second wavelet greater than that of the first suggests decreased intracranial compliance, despite the overall magnitude of the ICP. In fluid-coupled devices there may be occlusion by thrombus, tissue fragments or air in the system; the stopcock may be positioned in the off or in the draining position; or the ventricular ependymal surface may have collapsed against the ventriculostomy holes. It is also possible that the catheter is no longer in CSF, owing to inadverent withdrawal during tunneling or suturing, or post procedure due to traction on the draining system. Proper zeroing and leveling of the transducer at the level of the tragus also will influence the ICP value and the waveform. The validity of the ICP measurement can also be affected by improper zeroing of the pressure transducer prior to insertion. Care must be taken to zero these types of monitors properly prior to insertion. Once insertion is complete there may be a drift from baseline, giving rise to falsely low or high readings. This drift may be as much as - 10 + 30 mmHg, and may be associated with a damped or normal waveform.

COMPLICATIONS

The major complication of ICP monitor insertion is intracerebral hemorrhage. The risk of hemorrhage for parenchymal monitors is 0.5%, whereas that for ventricular catheters is higher, between 1 and 2%. Ventriculostomy-associated hemorrhages comprise the largest risk, yet among six studies the mean incidence of hemorrhage was 1.1%. Recently, Eddy and colleagues found no occurrences of hemorrhage in 297 trauma patients undergoing parenchymal ICP monitoring. Hematomas requiring surgical evacuation occur in 0.5% of patients. In patients with fulminant hepatic failure intraparenchymal monitors had a hemorrhage rate of 4% despite attention to coagulation function. The overall incidence of hemorrhagic complications of ICP monitoring in children is in the same range of 0.3 - 1.0%.

Non-hemorrhagic complications include infection, drift of the pressure transducer, obstruction of fluid-coupled devices and breakage requiring replacement. Ventricular catheters have a 6.3% incidence of obstruction. Higher rates are seen more commonly when ICPs>50 mmHg are seen. Replacement of the monitor is required in 10 - 20% of cases in large prospective series. Drift of some of the pressure transducer devices may be significant, requiring replacement of the probe or placement of a fluid-coupled device. New probes with improved sensors are available or under development that minimize this drift complication.

CSF colonization is the second most feared complication of ICP monitoring, although clinically significant intracranial/ CSF infection has not been reported in published trials. The range of CSF colonization in studies of antibiotic prophylaxis is 4.5 - 9.0%. In a comparison analysis. Ghajar found no significant difference in mean colonization rates between ventricular devices and subdural devices, but found a much greater mean incidence in parenchymal placed pressure transducer devices. Given that the ventricular and subdural devices are tunneled subcutaneously, the decreased incidence of colonization associated with these devices may imply an independent anti-infective effect of tunneling. In addition to the type of device, the duration of ICP monitoring affects the colonization/ infection rate. Winfield et al. found that the rate of colonization/ infection rose from 2% to 8% during the first 14 d of monitoring. However, the actual rate of infection per day remained stable. Older studies performed without use of contemporary self-contained CSF collection systems found that the risk of infection/ colonization rose from 8% at 5 d to 40% at 12 d, which is much higher than recent safety studies which found infection/ contamination rates of 0.01 - 0.3%. Aucoin et al, also found that irrigation of fluid-coupied systems increased the infection/ contamination rate from 6 to 19%. Thus irrigation should be performed infrequently and exquisite sterile technique, including Betadine preparation of the collection portal, used to aspirate and irrigate. The difficulty in applying this older literature is that closed CSF collection systems are now being used, which has seemingly decreased the infection/ colonization rates considerably and makes it difficult to justifiably recommend limiting the duration of monitoring and restricting irrigation. Table. 1.1 summarizes the complications of ICP monitoring.

CONTROVERSIES

There are several controversial questions to be considered when applying ICP monitoring to patient care. The first is whether ICP gradients exist, which entails several practical issues of the operation, namely, should ICP monitoring be performed in the hemisphere exhibiting pathology?. Should only one ICP monitor be used, or should more than one be used to assess for tissue pressure gradients? Is there a risk of herniation when a ventriculostomy device is inserted and CSF drained? ICP gradients have been shown to exist in both clinical and experimental settings. Pressure gradients of 10 mmHg or more may be present between the left and right hemispheres, and 20 mmHg between the supratentorial and infratentorial compartments. Moreover, Schickner and Young found that the ICP recorded from a parenchymal monitor exceeded simultaneous ventricular ICP measurements in over 66% of cases, suggesting that pressure gradients exist between the parenchyma and the ventricle as well. Despite the evidence for the presence of tissue gradients, no prospective trial has shown a benefit from dual monitoring, nor can it be recommended. However, the risk of triggering herniation should be kept in mind and caution used in draining CSF when this, including upward herniation from a cerebellar mass effect, is possible.

The second question is whether prophylactic antibiotics should be used during the monitoring period. Wyler and Kelly found that the rate of CSF infection tripled from 9% to 27% when patients monitored by ventriculostomy were not given antibiotics. Stenager et al. found a similar increase in CSF infections in patients without antibiotic prophylaxis compared to those receiving prophylactic antibiotics. However, three studies have found no difference in infection rates between patients with or without prophylactic antibiotics, although the rates of infection are higher than in more recent studies. Thus there is no clear consensus as to whether to use prophylactic antibiotics or not, and further studies using contemporary devices are needed.

Table 1.1 : Complications of ICP monitoring

NEWER DEVELOPMENTS IN INTRACRANIAL MONITORING

The ability to perform specialized monitoring through the burr hole site has broadened the range of parameters that may be monitored. Some of these include brain temperature, brain parenchymal oxygen tension, brain neurochemistry and cerebral blood flow. Each monitor may be used via a burr hole and a stabilizing bolt system. The various bolt systems entail up to three entry sites for the desired instruments. These bolts are somewhat larger than those used for intracranial pressure monitors alone, but have a similar feel and adaptability. No increase in morbidity has been reported when using the larger double or triple bolt systems. The methods of insertion can be adapted to these other monitoring techniques, and the information obtained is complementary to that of ICP. Although not widely used at the time of writing, these techniques have been extensively studied experimentally and have reliable characteristics. The reader should consider the articles by the following authors: Bullock, Zauner, Goodman and Roberston, Hillered, Persson and Vespa.

CONCLUSION

The technique of ICP monitoring has become more widespread over the last decade and remains an important facet in the critical care of neurosurgical patients. Once mastered, the techniques may be applied to a variety of monitors that are inserted into the brain or CSF through the skull. Close attention to sterility, monitor function and the technical limitations of the devices used are as important as the insertion technique itself.

References :

1. Queckenstedt H, Zur diagnoe der ruckenmarkkompression. Dtsch Zeitschr Nervenheilkd 1916;55:325 - 333.

2. Guillaume J, Janny P. Manometrie intracranienne cotinue: Interet de la methode et premiers resultats. Rev Neurol (paris) 1951;84:131 - 142.

3. Lundberg, N. Continuous recording and control of Ventricular fluid pressure of neruosurgical practice. Acta Psychiatr Neurol Scand 1960;149(Suppl):1 - 193.

4. Miller JD, Becker DP, Ward JD. Significance of intracranial hypertension in severe head injury. J Neurosurg 1977;47:503 - 516.

5. Ghajar J. Intracranial pressure monitoring techniques. New Horizons (Society of Critical Care Medicine) 1995;3:395 - 399.

6. Kellie.In: Weed LH. (ed) some limitations of the Monro- Kellie hypothesis. Arch Surg 1929;18:1049 - 1068.

7. Chan KH, Miller JD, Dearden NM. The effect of changes in cerebral perfusion pressure upon middle cerebral artery blood flow velocity and jugular venous oxygen saturation after severe brain injury. J Neurosurg 1992;77:555 - 61.

8. Robertson CS, Narayan RK et al. Clinical experience with a continuous monitor of intracranial compliance. J Neurosurg 1989;71:673 - 680.

9. Narayan RK, Kishore PR, Becker DP, Ward JD et al. Intracranial pressure: to monitor or not to monitor? A review of our experience with severe head injury. J Neurosurg 1982;56:650 - 659.

10. Paramore CG, Turner DA. Relative risks of ventriculostomy infection and morbidity. Acto Neurochir 1994;127 (1 - 2):79 - 84.

11. North B, Reilly P. Comparison among three methods of intracranial pressure recording. Neurosurgery 1986;18:730 - 732.

12. Sunbarg G, Nordstrom CH, Soderstrom S. Complications due to prolonged ventricular fluid pressure recording. Br J Neurosurg 1988;2:485 - 495.

13. Shapiro S, Bowman R, Callahan J, Wolfla C. The fiberoptic intraparenchymal cerebral pressure monitor in 244 patients. Surg Neurol 1996;45 (3):278 - 282.

14. Eddy VA, Vitsky JL, Rutherford EJ, Morris JA Jr. Aggressive use of ICP monitoring is safe and alters patient care. Am Surg 1995;61(1):24 - 29.

15. Blei AT, Olafsson S, Webster S, Levy R. Complications of intracranial pressure monitoring in fulminant hepatic failure. Lancet 1993;341(8838):157 - 158

16. Pople IK, Muhlbauer MS, Sanford RA, Kirk E Results and complications of intracranial pressure monitor in 244 patients. Surg Neurol 1996;45(3):278 - 282.

17. Piek J, Bock WJ. Continuous monitoring of cerebral tissue pressure in neurosurgical practice - experience with 100 patients. Intense Care Med 1990;16:184 - 188.

18. Wyler AR, Kelly WA. Use of antibiotics with external ventriculostomies. J Neurosurg 1972;37:185 - 187.

19. Smith RW, Alksne JF, Infections complicating the use of external ventirculostomy. J Neurosurg 1976;44:567 - 570.

20. Winfield JA, Rosenthal P, Kanter RK, Casella G. Duration of intracranial pressure monitoring does not predict daily risk of infectious complications. Neurosurgery 1993;33:424 - 431.

21. Aucoin P, Kotilainen HR, Gantz NM, Davidson R et al. Intracranial pressure monitors: epidemiologic study of risk factors and infections. Am J Med 1986;80:369 - 376.

22. Kanter RK, Weiner LB, Patti AM. Infectious complications and duration of intracranial pressure monitoring. Crit Care Med 1985;13:837 - 839.

23. Mayhall CG, Archer NH, Lamb A et al. Ventriculostomy-related infections: a prospective epidemiologic study. N Engl J Med 1984;310:553 - 559.

24. Weaver DD, Winn, HR, Jane JA. Differential intracranial pressure in patients with unilateral mass lesions. J Neurosurg 1982;56:660 - 665.

25. Broaddus WC, Pendelton GA, Delashaw JB. Differential intracranial pressure in patients with unilateral mass lesions. J Neurosurg 1982;56:660 - 665.

26. Wolfla CE, Luerssen TG, Bowman RM, Putty TK. Brain tissue gradients created by expanding frontal epidural mass lesion. J Neurosurg 1996;84:642 - 647.

27. Schinckner DJ, Young RF, Intracranial pressure monitoring: fiberoptic monitor compared to ventircular catheter. Surg Neurol 1992;37:251 - 254.

28. Stenager E, Gerner - smidt P, Kock - Jensen C. Venticulostomy - related infections:an epidemiological study. Acta Neurochir (Wien) 1986;83:20 - 23.

29. Rosner MJ, Becker DP. ICP monitoring: complications and associated factors. Clin Neurosurg 1976;23:494 - 519.

30. Bullock R, Zauner A, Myseros JS et al. Evidence for prolonged release of excitatory amino acids in severe human head trauma: relationship to clinical events. Ann NY Acad Sci 1995;59:290 - 297.

31. Zauner A, Bobtenberg E, Woodward J et al. Multiparametric continuous monitoring of brain metabolism and substrate delivery in neurological patients. Neurol Res 1997;19:265 - 273.

32. Goodman JC, Valadka AB, Gopinath SP, Cormio M et al. Lactate and excitatory amino acids measured by microdialysis are decreased by pentobarbital coma in head injured patietns. J Neurotrauma Oct 1996;13 (10):549 - 556.

33. Hillered L, Persson L, Ponten U et al. Neurometabolic monitoring of the ischaemic human brain using microdialysis. Acta Neurochir 1990;102(3-4):91 - 97.

34. Persson L, Valtysson J, Enblad P et al. Neurochemical monitoring using intracerebral microdiaysis in patients with subarachnoid hemorrhage. J Neurosrug 1996;84:606 - 616.

BACK

Surgery for Movement Disorders

BACK

INTRODUCTION

There are currently three major areas of focus in the surgery of movement disorders: ablative procedures, deep-brain stimulation, and 'restorative' therapies using cell transplantation or intracerebral delivery of growth factors. Procedures to be discussed, and disorders for which they may be indicated. All procedures on this rapidly expanding list alter function to the basal ganglia and/or cerebellothalamic pathways. Ablative surgery and deep-brain stimulation procedures attempt to compensate for, rather than to correct, the cellular to biochemical brain abnormality causing disordered movement. The 'restroative' therapies attempt to correct the cellular or biochemical brain defect that is at the root of the disorder, by replacing or promoting the survival of degenerating cell populations.

Movement disorders that may be considered for central nervous system surgery include parkinson's disease, non-parkinsonian tremor, dystonia, hemiballismus, and Huntington's disease. Motor abnormalities in these conditions are qualified using standard rating scales, such as the Unified Parkinson's Disease Rating Scale. Unified Hungtington's Disease Rating Scale and Fahn-Marsden Dystonia Rating Scales. Rigorous use of these rating scales is essential to follow the progression of disease or to measure the effect of treatment.

The major motor signs and known pathophysiologic mechanisms of these diseases summarized below.
Ÿ Parkinson's disease. This is a progressive, degenerative disease involving the basal ganglia. The cardinal clinical signs of the disease are testing tremor, rigidity, bradykinesia, and postural instability. The fundamental biochemical defect is loss of dopamine in the striatum and in other basal ganglia nuclei, due to loss of the dopaminergic cells of the substantia nigra, pars compacta.

• Non-parkinsonian tremor. Non-parkinsonian tremors include essential tremor and the cerebellar outflow tremors, which are most commonly due to trauma, infarction, or multiple sclerosis. Essential tremor is the most common of these, and the most amenable to surgical treatment: it mainly affects elderly persons. Essential tremor is a postural and action tremor, predominantly in the hands. Its pathophysiologic basis is unknown, but presumably involves cerebellothalamic pathways. In many cases is appears to be inherited, with an autosomal dominant pattern.
• Dystonia. Dystonia refers to a symptom rather than to a particular disease, and a heterogeneous group of disorders may have dystonia as the primary manifestation. Dystonia is characterized by sustained muscle contractions leading to twisting, repetitive movements and abnormal postures. Primary forms of dystonia are often due to single-gene defects, and their genetic bases are under active investigation. Dystonia may also be secondary to traumatic, neoplastic, or infectious lesions involving the basal ganglia or, less commonly, the motor thalamus.
• Hemiballismus. Hemiballismus is characterized by violent, uncontrollable dyskinesias of the proximal extremities. It is usually due to ischemic lesions involving the subthalamic nucleus and/or its efferent pathways.
• Hungtington's disease. This is an autosomal dominant degenerative brain disease. In early stages its cardinal motor sign in chorea, whereas later stages are often characterized by an akinetic rigid state. Non-motor signs such as depression, psychotic behavior, and dementia often result in greater disability than the motor signs. Its pathologic hallmark is degeneration of striatal output neurons. Huntington's disease is due to expansion of a trinucleotide repeat region on chromosome 4 that encodes for the protein Huntingtin, but it is not known how this gene defect produces the observed pathology and clinical signs.

PHYSIOLOGY OF MOVEMENT DISORDERS:
BASAL GANGLIA MODEL

THE BASAL GANGLIA - THALAMOCORTICAL MOTOR CIRCUIT

The cortex, basal ganglia, and thalamus participate in multiple loop circuits controlling motor, limbic, and associative functions. The basal ganglia-thalamocortical motor circuit plays a key role in regulating motor behavior. The cortical regions participating in the motor circuit are the motor, premotor, and somatosensory cortices. The basal ganglia structures that participate in this circuit include portions of the striatum, the internal and external segments of the globus pallidus, the subthalamic nucleus, and the substantia nigra, pars compacta and pars reticulata. The portions of these nuclei that participate in the motor circuit are called the sensorimotor portions, and contain neurons whose discharge rates are modulated by joint movements. Sensorimotor regions are anatomically separate from the regions that regulate non-motor functions. Because of this, it is possible to design surgical interventions that alter function in the motor circuit without affecting non-motor circuits.

The current model of the basal ganglia - thalamocortical motor circuit is illustrated schematically. Cortical efferent projections enter the basal ganglia via the striatum. The major basal ganglia output nuclei are GPi and SNr, which project to subdivision of the motor thalamus. Within the basal ganglia motor function is thought to be regulated by two major pathways, both of which are modulated by striatal deopamine. The 'direct' pathway is a monosynaptic projection from the putamen to GPi/SNr. In the 'indirect' pathway, the putamen projects of GPi/SNr via intermediate nuclei, GPe and STN. Dopaminergic innervation of the putamen by SNc is inhibitory to striatopallidal fibers in the indirect pathway, but excitatory to striatopalidal fibers in the direct pathway. The non-dopaminergic projections between the basal ganglia nuclei are GABAergic, except for the projection from STN to GPi/SNr, which is glutamatergic.

In this scheme, the indirect pathway provides negative feedback to cortical regions, while the direct pathway provides positive feedback. Balance between the direct and indirect pathways is critical for maintaining the appropriate modulation of cortical regions concerned with motor control.

MODELS OF MOVEMENT DISORDERS

Abnormal activity in the basal ganglia - thalamocortical motor circuit is thought to be responsible for many of the motor abnormalities associated with the above diseases. Greater understanding of these alterations has allowed movement disorder specialists to begin building a theoretical basis for surgical intervention.

Movement disorders are often classified as hyperkinetic or hypokinetic, although symptom complexes of a given disorder may not fit neatly into these categories. In early Huntington's disease there is selective degeneration of neurons that give rise to the indirect pathway. This results in hypoactivity of STN, hypoactivity of the basal ganglia output nuclei, and hyperactivity of the thalamocortical pathway. The resulting cortical hyperactivity is manifested by hyperkinetic model, since damage to STN also results in decreased activity of the basal ganglia output nuclei and disinhibition of the thalamocortical pathway. In idiopathic torsion dystonia discharge rates in GPi are decreased, suggesting that dystonia may also be grouped with the hyperkinetic disorders. However, the pathophysiology of dystonia remains poorly understood.

Abnormalities in hypokinetic movement disorders, such as Parkinson's disease. While the loss of dopaminergic cells in the SNc is the fundamental defect in Parkinson's diseases, this loss results in many other abnormalities in the basal ganglia - thalamocortical motor circuit. Loss of striatal dopamine, in both direct and indirect pathways, leads to increased activity in the output nuclei, GPi and SNr, which in turn leads to excessive inhibition of the thalamocortical pathway. The deficiency in cortical excitation is manifested by the hypokinetic motor signs of parkinsonism. This model of Parkinson's disease is supported by metabolic and electrophysiologic studies of the basal ganglia of the parkinsonian monkey. In this model, metabolic activity and neuronal discharge frequencies in GPi and STN are increased compared with normals.

Not all movement disorders can be described in terms of alterations in the basal ganglia - thalamocortical motor circuit. Projections from the basal ganglia output nuclei to brain - stem nuclei, such as the pedunculopontine nucleus, may also be important in movement disorders, though the function of these projections is less well understood. Furthermore, many forms of tremor may reflect derangements in cerebellothalamocortical pathways, rather than in the basal ganglia

This chapter will discuss surgery for movement disorders according to the following headings:

• Ablative surgery
  Pallidotomy
  Thalamotomy
  Deep-brain stimulation
  Surgical localization of basal ganglia nuclei
• Restorative therapies
  Cell transplantation
  Intracerebral delivery of neurotrophic factors

PALLIDOTOMY

PALLIDOTOMY FOR PARKINSON'S DISEASE

Gpi pallidotomy is the surgical destruction of the motor Gpi. Pallidotomy has a long history in the treatment of Parkinson's disease and other movement disorders. Russell Meyers, in the 1940s, was the first to describe surgery in the region of the Gpi for PD, showing that lesions of the Gpi or its outflow tracts could reduce rigidity and bradykinesia without causing paresis. In the1950s several others confirmed this. The introduction of L-DOPA for Parkinson's disease in the 1960s resulted in an effective medical treatment for parkinsonian motor signs, and a temporary decrease in enthusiasm for surgical treatment of Parkinson's disease.

The recent resurgence of interest in pallidotomy as a treatment for Parkinson's disease is due to mulitple factors: (1) recognition that long-term medical therapy for Parkinson's disease is often unsatisfactory, with patients eventually suffering from drug-induced dyskinesis, motor fluctuations, and variable responses to medication;

(2) greater understanding of the pathophysiology of Parkinson's disease, providing a rational theoretical basis for pallidotomy; and

(3) use of newer techniques, including CT - and MRI - guided sterotaxis and single unit microelectrode recording, making surgical intervention in the basal ganglia more precise and safe.

Theory

The scientific basis of Gpi pallidotomy, in the case of Parkinson's disease, can be understood from the pathophysiologic model of Parkinson's disease. in parkinsonism, STN and GPi are hyperactive, and the thalamocortical projection is hypoactive. The model predicts that lesions in the motor controlling region of GPi should decrease the inhibitory influence of basal ganglia output nuclei on the motor thalamus, and restore thalamocrotical activity. Normalization of thalamocortical activity should, in turn, lead to amelioration of the motor signs associated with parkinsonism. In support of this theory, pallidotomy in Parkinson's disease patients is associated with increased activation of cortical motor areas, as determined by positron emission tomography. Following pallidotomy, the supplementary motor area is more active during performance of timed motor tasks, compared with the preoperative state. The activity of motor and premotor cortex in the resulting state is increased as well. The model illustrates how lesioning certain parts of the circuit can compensate for the effects of dopamine loss on cortical activity, even though lesioning clearly does not restore dopaminergic function.

The appropriate target location within Gpi was originally discovered by trial and error. Leksell, in the 1950s, systematically varied the target in Gpi from the anterodorsal region to the posterolateral region, with improved symptom control in the latter. This postero-lateral target was recently repopularized by Laitinen et al. Electrophysiologic studies have justified the choice of posterolateral Gpi as the appropriate pallidotomy target, since in both humans and in non-human primates the posterolateral part of Gpi is the part containing neurons whose discharge is modulated by joint movements. Thus, posterolateral Gpi is the sensorimotor part, and is therefore the part participating in the basal ganglia - thalamocortical motor circuit. Lesions in other non-motor parts of Gpi should primarily affect non-motor functions. Lesions outside the motor territory of Gpi may temporarily relieve parkinsonian motor signs, presumably by extension of edema into, or partial lesioning of, the motor territory, but the benefit of such lesions is transient.

Indications

Candidates for pallidotomy should have a diagnosis of idiopathic Parkinson's disease, which requires the presence of at least two of the four cardinal signs of the disease, and a history of beneficial response to L-deoxyphenylalanine. Most candidates are moderately disabled, with Hohn and Yahr scores of 3 - 4 when off, have disabling drug-induced dyskinesias and severe on-off fluctuations. The majority of patients have rigidity and bradykinesia as predominant symptoms, although patients with tremor-dominant Parkinson's disease are not excluded. There is no specific age-cut off.

Contraindications

Contraindications to pallidotomy include dementia, extensive brain atrophy, and 'Parkinson's plus' syndromes.

Outcome

Although Gpi pallidotomy has been performed for forty years. Older studies are inadequate to predict the efficacy of contemporary pallidotomy. Symptoms in contemporary Parkinson's disease patients are different from those prior to the introduction of L-DOPA, due to the side effects of exposure to dopamine-replacing medications. Surgical techniques have substantially improved. Older studies lack documentation of consistent lesion location and outcome assessment based on standard parkinsonian rating scales. Thus, pallidotomy warrants re-evaluation from a modern perspective.

Most contemporary studies cite significant benefit in parkinsonian motor signs following Gpi pallidotomy, with the benefits sustained over the period of evaluation. In one study of microelectrode-guided pallidotomy using blinded investigators, UPDRS motor scores in the off state improved by 30% at 6 months follow-up. The total akinesia score improved by 33%, while the gait score improved by 15%. There was a nearly complete elimination of 'on' dyskinesias on the side opposite to the lesion, although the patients continued to take the same doses of dopaminergic medications as before the surgery. In our experience results for 'off' period improvements and 'on' dyskinesias at 12, 18 and 24 months from surgery were similar to these. In addition, we found that tremor improved in the majority of patients and 'on' periods were prolonged. The outcome from pallidotomy beyond two years is not known. Although it is clear that some patients experience some loss of benefit, this may be related to inadequate lesion size or suboptimal location, or to the possibility that disease progression overcomes the observed benefit from surgery.

Lesion location clearly affects clinical outcome. Most modern series attempt to target the posterolateral sensoriomotor portion of Gpi. One of our early patients presented to us following a pallidotomy elsewhere, with a lesion anterior to the sensorimotor part of Gpi. The patient reported immediate resolution of all parkinsonian motor signs, with a complete relapse over the ensuing week. Repeat pallidotomy in this patient, with lesion placement posterior and lateral to the first lesion, has provided sustained benefit at 3 years. Evaluating 11 patients into two groups based upon lesion location determined on thin cut, volumetrically acquired postoperative MRI. The patients with lesions clearly centered in the sensorimotor portion of Gpi had sustained improvement in UPDRS scores at 18 months. Patients with lesions which involved only a portion of this area, or which encroached upon Gpe, sustained significantly less benefit. The best size and exact location of the lesion for optimal long-term benefit, however, remain uncertain.

Complications

It one of the firs pallidotomy series of the modern era, there was a 14% incidence of visual field deficits, presumably due to injury to the optic tract. In studies utilizing microelectrode recording, this proportion has been substantially lower. In our own series, 4 of 170 patients, all from early in the series, developed partial visual field cuts. Nine patients developed weakness in the face or limb, all but one of which resolved completely. Other major complications from our series included symptomatic intraparenchymal hemorrhages along the lesioning tract in two patients, a subdural empyema requiring drainage, a large subdural hematoma at a site distant from the surgical intervention, and an unexplained postoperative encephalopathy in an elderly patient. Although there were no immediate deaths, the patients with the latter two complications died several months after surgery. Other reported complications of pallidotomy include hypophonia or increased difficulty with speech articulation.

Based upon these studies, we can conclude that unilateral microelectrode - guided pallidotomy is relatively safe and effective for at least two years in the majority of patients with moderate - advanced idiopathic Parkinson's disease.

Bilateral pallidotomy

Following simultaneous or staged bilateral pallidotomy, the likelihood of significant speech deterioration has been reported to be as high as 53 - 60%. One group has reported that bilateral pallidotomy produced major cognitive deficits in 2 of 2 patients treated. We have performed six staged bilateral pallidotomies, and the majority are hypophonic to varying degrees. We have not observed significant cognitive changes in any of these patients, although only 2 to 6 have undergone formal neuropsychologic evaluation. The safety of bilateral pallidotomy needs closer scrutiny before it can be recommended.

PALLIDOTOMY FOR OTHER MOVEMENT DISORDERS

While medically intractable Parkinson's disease is the predominant indication for Gpi pallidotomy, the operation has been applied to other movement disorders as well. The first use of stereotactic pallidotomy for a movement disorder was in fact for Hungtington's disease. The procedure was effective for decreasing choreiform movements contralateral to the lesion. It is unclear whether this could provide functional benefit to patients, and the brain atrophy associated with Huntington's diseases increases the risk of postsurgical hemorrhage. Isolated case reports indicate the pallidotomy may be effective for alleviating the involuntary movements associated with hemiballismus. Recently, we and others have observed significant improvement in patients with idiopathic generalized dystonia. One study with short follow up has also demonstrated efficacy in secondary dystonias. Careful clinical follow up, however, is needed to document long-term effectiveness, before use of pallidotomy in the treatment of dystonia can be recommended.

Understanding the clinical effects of pallidotomy

According to the model, pallidotomy relieves hypokinetic parkinsonian motor signs by removing excessive inhibitory activity from the pallidum, leading to a normalization of activity in the thalamocortical projection. Pallidotomy not only relieves parkinsonian bradykinesia, however, but also relieves tremor and L-DOPA-induced dyskinesias. Hyperkinetic signs of other movement disorders can also be relieved by palidotomy. These latter effects of Gpi pallidotomy are difficult to explain in light of the current model of hypo and hyperkinetic disorders, probably because the model considers abnormal motor function to result from alterations in mean discharge rates within the basal ganglia- thalamocortical circuit. In the pathophysiology of movement disorders, however, alterations in the pattern of action potential firing may be much more important than alterations in mean rates. While the model is a useful starting point for understanding the effects of surgical interventions in movement disorders, it will certainly be refined in the future.

STN as a potential lesioning target

Clinical work on ablative surgery for Parkinson's disease has focused on Gpi as a lesioning target. While Gpi is clearly hyperactive in parkinsonism, experiments in parkinsonian monkeys show that the major source of excitatory input to Gpi, the subthalamic nucleus, is also hyperactive. STN lesions in parkinsonian monkeys decrease spontaneous and movement - evoked neuronal discharges in Gpi, and alleviate all parkinsonian motor signs. These findings have generated interest in STN ablation as a surgical procedure for human parkinsonism. Because STN projections affect both basal ganglia output nuclei, SNr as well as Gpi, lesions of STN may be more effective against parkinsonism than lesions of Gpi alone.

Surgical lesioning of the STN has the potential to lead to serious complications. In humans, chronic intractable hemibalism is often associated with ischemic lesions that involve STN and neighboring structures. In monkeys treated with methylphenyltetrahydropyridine, isolated STN lesions can produce hemiballistic movements, but these usually resolve over time. In some older series on thalamotomy for parkinson's disease, lesions sometimes included small parts of the subthalamic nucleus, and intractable hemiballismus from such lesions was not routinely observed. The exact structures responsible for permanent hemiballism have not been precisely localized. Thus, the possible side effects of isolated STN lesions in human are not certain.
THALAMOTOMY

Thalamic ablations have been performed in the treatment of many neurologic conditions, including pain and psychiatric derangments. When performed for movement disorders the target for thalamotomy is within the motor thalamus. Thalamotomy for PD was reported 40 years ago byHassler and Reichert who experimented with lesions of a variety of sizes and shapes within this area. The motor thalamus is subdivided into several nuclei. From posterior, and ventralis oralis anterior. Voa and Vop are the basal ganglia receiving areas, while vim is the cerebellar receiving area. For many years, thalamotomy was the mainstay of surgical treatment for a variety of symptom complexes in several movement disorders, including parkinsonian tremor and rigidity, non-parkinsonian tremor, and a variety of a dyskinesias and dystonias. These attempts were largely justified on a trial-and-error basis.

VIM THALAMOTOMY FOR TREMOR
Theory

The most accepted contemporary use of thalamotomy in movement disorders is Vim thalamotomy for parkinsonian and essential tremor. The efficacy of this procedure was discovered empirically, and its theoretical basis is still uncertain. Vim receives its major afferent projections from deep cerebellar nuclei and projects to motor cortex. Microelectrode recording in Vim in patients with parkinsonian and essential tremors identifies cells discharging in bursts that are time-locked to the patient's tremor, suggesting that tremor of several etiologies is associated with abnormal discharge in the cerebellothalamic pathway. While it is reasonable that interruption of this pathway alleviates tremor, a more exact explanation for the efficacy of Vim thalamotomy for tremor is lacking.

Indications and results

Appropriate candidates for Vim thalamotomy are patients with essential tremor, or tremor - dominant Parkinson's disease, whose tremor has become debilitating despite optimal medical management. Several recent reviews have documented long-term efficacy of 60 - 90% for alleviation of tremor in these disease. cerebellar outflow tremors, such as those due to multiple sclerosis or to cerebellar trauma, are more resistant to treatment. Long-term memory deficits or worsening of dysarthria are reported complications of unilateral thalamotomy, particularly on the dominant side, but these complications are more common following bilateral thalamotomy. Bilateral thalamotomy is, therefore, not frequently recommended.

THALAMOTOMY FOR CONDITIONS OTHER THAN TREMOR
Thalamotomy for parkinsonian rigidity

During the era when thalamotomy was the predominant surgical treatment for Parkinson's disease, investigators searched for thalamic lesion targets that might relieve symptoms other than tremor. Some found that extending the thalamotomy lesion anteriorly into Vop/Voa improved rigidity. However, rigidity, bradykinesia, and motor fluctuations appear to be better controlled by Gpi pallidotomy than by thalamotomy. Thus, for Parkinson's disease, pallidotomy is preferable to thalamotomy in all but the minority of Parkinson's disease patients for whom tremor is the dominant disease feature. Since pallidotomy also appears to be effective for parkinsonian tremor in 85% of cases, pallidotomy may be performed even for tremor - dominant Parkinson's disease.

Thalamotomy for dystonia

In the 1950s, in the course of performing thalamotomy for Parkinson's disease, it was noted that Parkinson's disease associated dystonia was often relieved by these lesions. Since then several groups have performed thalamotomy for non-parkinsonian dystonia. The results have been variable, which may reflect heterogeneous patient populations and variations in the site of thalamic lesioning. Published series have included various forms of secondary dystonia as well as idiopathic dystonias. The largest series of thalamotomy for idiopathic dystonias. The largest series of thalamotomy for idiopathic torsion dystonia is from Cooper who reported marked improvement in 24.5% and mild improvement in 45.2%, with a mean follow-up of 7.9 years. In Andrew's series only 4 of 16 patients with generalized dystonia had sustained benefit, while 62% of patients with focal or segmental dystonia had sustained benefit. The appropriate thalamic target for best relief of dystonia is unclear. Intended lesion location has been decided empirically, since the understanding of thalamic and basal ganglia dysfunction in dystonia is poor and thus sound theoretical bases for target localization are lacking. Yamashiro and Tasker's most successful results in thalamotomy for dystonia were from lesions involving both Vim and Vop. In other series, lesions were placed in Vop, and if ineffective, were extended further anteriorly into Voa with a second surgery. It would appear that lesions to treat dystonia must be much larger than those that are effective for distal tremor.

No published study has been performed prospectively, with blinded evaluation, with documentation of consistent lesion location, or with use of detailed quantitative rating scales of motor disability. Thus, the true long-term outcome of thalamotomy for dystonia is unknown. The neurosurgical treatment of dystonia is currently being re-evaluated and is likely to evolve rapidly.

Thalamotomy for hemiballismus

Since hemiballismus is rare condition and may resolve spontaneously. Surgery for intractable hemiballismus is seldom performed. However, several recently reported causes of Vim thalamotomy for hemiballsmus showed excellent control, although follow-up was short.

DEEP-BRAIN STIMULATION

Intraoperative stimulation of subcortical nuclei has long been performed during ablative surgery for movement disorders, to provide confirmation of target location prior to lesioning. Investigators found empirically that deep brain stimulation in the motor thalamus at high frequency arrested tremor similar to permanent lesioning, but that the effect reverses when stimulation is stopped. It appeared as if DBS created a reversible inactivation of a focal brain region near the electrode. This observation has motivated surgeons to implant permanent indwelling stimulators in the same target nuclei utilized or considered for lesioning surgery. The rationales for selecting motor thalamus, Gpi, or STN as targets for stimulator placement are similar to the rationales for selecting these sites for lesioning.

MECHANISMS

The effects of DBS are poorly understood. The evidence that chronic high-frequency stimulation inactivates neighboring neurons is circumstantial; DBS in several subcortical nuclei can produce similar physiologic effects as lesioning in experimental animals as well as in humans. DBS may inactivate a region by depolarization blockade of neurons within a certain radius around the active contact. However, it is also possible that cells are activated, rather than inactivated, by DBS, depending upon the stimulation parameters and upon cell morphology, membrane properties, distance from the stimulating electrode, and baseline discharge from the stimulating electrode, and baseline discharge rate; PET studies of patients undergoing Vim stimulation for tremor support this possibility. Those studies show changes in cerebellar metabolism during clinically effective, but not during ineffective, stimulation. Since Vim has no known direct projection to the cerebellum, these changes in cerebellar metabolism appear to be due to antidromic activation of cerebellar afferents to Vim. Stimulation - induced activation of neuronal discharge could have similar effects on motor function as lesioning, by replacing or 'jamming' abnormally oscillating patterns of discharge with a single constant frequency discharge pattern.

However, the effects of DBS are not always identical to the effects of lesioning. In Gpi we have found a lack of improvement in parkinsonian signs in response to macro-stimulation at the same location where a lesion leads to complete amelioration of parkinsonian motor signs. DBS cannot be considered as equivalent to lesioning, and the exact effects of DBS in a particular nucleus must be established empirically.

Although an autopsy report on a patient eight years after implantation of a Vim stimulator showed no ill effects, other than a thin rim of gliosis around the electrode track, long-term effects of DBS in the brain are not clear.

INDICATIONS AND RESULTS
Thalamic Stimulation

Chronic Vim thalamic stimulation is effective for medically intractable parkinsonian or essential tremor. In a series of 100 patients who received 147 stimulators for tremor. 88% of parkinson's disease patients and 61% of essential tremor patients had complete or near complete tremor resolution at the longest follow-up, ranging from six months at eight years. These are comparable to the best results for thalamotomy, and Vim stimulation may be performed bilaterally with fewer side effects than thalamotomy. At present, Vim stimulation is the procedure of choice when contralateral thalamotomy has been performed or is anticipated. Whether it should completely replace unilateral Vim thalamotomy requires further study.

Thalamic stimulation is potentially useful for dystonia, although experience is very limited. In five reported cases, posttrumatic or postinfarction hemidystonia had a favorable response to stimulation of the thalamic sensory relay nucleus ventralis caudalis. Benabid et al. implanted Vim stimulators in five dystonia patients, of whom two had familial generalized dystonia. There were modest improvements in limb rigidity, aspiration of saliva, and flexed postures, although these were not quantified.

Gpi Stimulation

The technique of chronic stimulation of Gpi is in its infancy. A small number of reported cases suggest that chronic Gpi stimulation has similar effects as pallidotomy for rigid/bradykinetic PD, and may be safe for bilateral use. It is possible that pallidal stimulation contralteral to a previous pallidotomy would benefit patients with bilateral disease, while avoiding the potential complications associated with bilteral pallidotomy, but this remains to be confirmed in clinical trials. By analogy with Gpi lesioning for dystonia, investigation of Gpi stimulation for dystonia also seems warranted.

STN stimulation

Based on work in parkinsonian monkeys showing substantial benefit of STN lesions, Benabid and colleagues have implanted bilateral stimulators in the STN in several Parkinson's disease patients who are predominantly rigid/bradykinetic. In the first three patients, UPDRS motor scores were improved by 42 - 82% three months after surgery. Akinesia and rigidity were the motor signs that improved the most. Dyskinesias could be induced by STN stimulation at higher current intensities but were reversible when the current intensity was reduced. There appeared to be a therapeutic window of stimulation parameters within which parkinsonian motor signs could be ameliorated without inducing involuntary movements. Thus, this technique appears promising, and is now being tested at multiple centers.

TECHNICAL CONSIDERATIONS

Target localization for placement of chronic stimulators utilizes the same principles and techniques as lesioning surgery. However, placement of a stimulator may be done in two stages. The first surgery is to place the stimulator itself, with an externalized contact for in hospital testing. Once the efficacy of stimulation is confirmed and the parameters optimized, the second procedure may be performed several days later to internalize the lead and attach it to a battery and programming device implanted in the infraclavicular region.

COMPARISON WITH LESIOING

The major advantages of deep brain stimulation compared to creating a lesion are that DBS is:

• Reversible
• Permits adjustment in the magnitude of its effect at any time during treatment
• May be safer than lesioning techniques for bilateral use. Also, DBS lends itself readily to objective evaluation since, by varying the status of the stimulator, its effects can be easily studied in a double - blinded fashion.

This advantages of DBS, as compared with lesioning, are :

• The technique requires implantation of a foreign body with concomitant risks of infection and migration.
• Batteries must be replaced at regular intervals, or an external power source must be worn against the skin.
• A single stimulator is limited in the volume of tissue it can affect, thus rendering it potentially less practical in regions where the physiologic target is large.

SURGICAL LOCALIZATION OF BASAL GANGLIA NUCLEI
LOCALIZATION OF SENSORIMOTOR GPI

The target for pallidotomy or pallidal stimulation is within the sensorimotor region of Gpi. Physiologically, sensorimotor Gpi is defined as that part of the nucleus containing neurons whose discharge rates are modulated by passive or active movements. Anatomically, sensorimotor Gpi is located in the posterolateral part of the nucleus. Three method are used, ideally together, to determine target location prior to lesioning or stimulator placement:

• Image -guided stereotactic localization:
• Microelectrode mapping
• Macrostimulation

Even the use of extensive microelectrode mapping does not negate the need for precise stereotactic targetting since precise stereotaxis will maximize the amount of information obtained from the initial microelectrode pass, and minimize the number of passes required.

Image - guided localization

For stereotactic localization, we use the Leksell series G headframe. With this system images must be obtained in planes that are orthogonal to the frame axes, unless software is used to correct the angle. Frame placement is illustrated. The frame is placed symmetrically on the head with the aid of earplugs that align the frame with the external auditory canals. The earplugs are placed in the middle of three possible holes on the earplug carrier, and the carrier is adjusted to a Y-position of 95. The earplugs are simultaneously advanced, slowly and under intravenous sedation, until they fit snugly in the external auditory canals. The anteroposterior axis of the frame is then angled such that the inferior margin of the U-shaped anterior intubation ring aligns with the tip of the nose. The skull pins are advanced, and when all four lightly touch the skin, the earplugs are retracted. This algorithm places the Y-axis parallel to the orbitomeatal line, which is itself parallel to the AC-PC line. Precise and reproducible frame placement is important to insure that the preoperative images and the intraoperative instrument trajectories occur in standard anatomical planes. Also, with this algorithm, the eyes and mouth remain unencumbered by the frame, so as to allow visual field examination and airway access.

Most groups target a standard anatomical landmark which is defined relative to the line between the anterior and the posterior comissures. Leksell's Gpi pallidotomy target, as repopularized by Laitinen, locates the inferior edge of the anatomical target at 21 - 22 mm lateral, 2 - 3 mm anterior and 2 - 6 mm below the mid AC-PC point.

Out targeting algorithm is based both on identification of the AC-PC line and on direct visualization of the borders of Gpi on MR images. Axial and coronal MR images are obtained orthogonal to the frame axes, with 2 mm slice thickness, using an inversion recovery sequence and it shows an example of a set of images used for targeting. The lateral and anteroposterior coordinates of the target are determined on the axial scan at the level of the AC-PC line, from a point due lateral to the mid AC-PC point, and approximately 2 mm medial to the external accessory lamina. This point is superior to the actual target at the base of Gpi. The Z-coordinate is determined on the coronal scan closest to the mid-AC-PC point, from the superolateral edge of optic tract. The approach to this target is parasagittal at an angle of 60 from the AC-PC line.

The algorithm is designed to place the initial micro-electrode track on a parasagittal plane corresponding to the lateral 20 plane in the Schaltenbrand and Bailey atlas, to provide the longest possible trajectory within Gpi on this plane, and to terminate at the optic tract. In any given individual, the plane through the basal ganglia that best corresponds to S & B 20 may not actually be 20mm from the midline; we have found that the parasagittal plane that best fits the S&B 20 plane, as varified by microelectrode mapping of nuclear boundaries, may vary from 18 to 24 mm lateral to the midline. Therefore, determining the lateral coordinate based on direct visualization of the boundaries of Gpi is more accurate than fixing the lateral coordinate at a given distance from the midline without consideration for individual anatomic variations.

Commercially available software may be used to enhance the accuracy of image-based targeting. Image fusion algorithms allow an MRI to be computationally fused to a stereotactic CT, providing a hybrid image that combines the nuclear resolution of MRI with the spatial accuracy of CT. Three-dimensional surgical planning packages are available for the major stereotactic systems, which allow simultaneous visualization of a single intended trajectory on multiple - imaging planes.

Microelectrode recording and microstimulation

For extracellular microelectrode recording and microstimulation, we use glass-coated platimum - iridium electrodes of impendance 0.5 - 1.0 M at 100 Hz, tip diameter 2 - 4 m. The electrode is mounted on a hydraulic microdrive, which is itself mounted on a coarse drive, attached to the stereotatic are system. The element of the hydraulic microdrive that holds the electrode is coupled to a potentiometer which is used to generate a digital readout of electrode depth. Action potentials are amplified, filtered, displayed on an oscilloscope, and played on an audio monitor. For reduction of unwanted electrical noise in the electrically challenging operating room environment, we utilize a high-impedance preambplifer, whose headstage is mounted on the arc system near the electrode itself. Fluorescent lights produce too much electrical noise and need to be extinguished. Most other pieces of equipment may be kept on, provided that they are adequately grounded. The micro-electrode is lowered along the stereotactically determined trajectory, through a 1.5 cm burr hole.

Based on the physiologic properties of neurons in the basal ganglia, microelectrode recording can be used to precisely localize the boundaries of Gpi and to define the sensorimotor territory. The action potential discharge characteristics of neuronal cells encountered on a trajectory to Gpi. The majority of striatal neurons, likely corresponding to the medium spiny neurons, have very low spontaneous discharge rates. Gpe neurons have higher spontaneous rates and typically discharge in 'burst' or 'pausing' patterns. Gpi neurons have still higher, and more regular, mean spontaneous discharge rates than Gpe neurons. Cells with slower, very regular discharge rates, 'border' cells, are typically found in the white matter laminae surrounding Gpe and Gpi. Cells of the nucleus basalis also have very regular discharge rates. Transitions from gray matter to white matter are easily distinguished by the reduction in background noise, and by the fact that the rare action potentials recorded from axons are of opposite polarity to those recorded from neuronal cell bodies.

Beginning with the entry of the microelectrode into subcortical gray matter, the depths of neuronal cell bodies encountered are recorded as tickmarks along a scaled reconstruction of the electrode track. Segments of the track reconstruction are shaded to indicate striatal cells, Gpe cells or Gpi cells. The optic tract, identified by light-evoked fiber activity, is often found below the inferior margin of Gpi. A single microelectrode track mainly provides information about the depth of the target, the track determines the depth of the sterotactic target in relation to the inferior and superior borders of Gpi and to the optic tract.

Utilization of microelectrode - derived information for target localization. To optimize localization in three dimensions, we make multiple miroelectrode tracks along parallel parasagittal trajectories separated by 2- 4 mm. The scaled, shaded microelectrode track reconstructions are superimposed on drawings of parasagittal sections adapted from the Schaltenbrand and Bailey atlas, using the surgical team's judgment of 'best fit' of the microelectrode tracks to the atlas. Since the stereotactic coordinates of the microelectrode tacks are known, this technique registers the physiologically determined boundaries of Gpi and surrounding structures, in reference to the stereotactic coordinate system.

Receptive fields of neurons can be used to determine location within the boundaries of a nucleus. Cells in Gpi that are responsive to joint movements usually respond to movement of one or of a small number of joints in a restricted region on the contralateral side of the body. Identification of such neurons in Gpi provides the valuable information that the electrode tip is located in the sensorimotor subdivision of the nucleus, rather than in a non-motor subdivision. Also, sensorimotor Gpi is somatotopically organized, with leg representations tending to be more dorsal and more medial than am representations. Thus, determining the somatotopy of a group of neurons provides localizing information even within the sensorimotor subdivision of Gpi.

Microstimulation, e.g. passing current through the microelectrode, is used to localize both the corticospinal tract and the optic tract. We normally stimulate with biphasic pulses, 0.2 ms, 300 Hz, 5 - 40 A. The CST in the internal capsule is identified by the current threshold for producing muscle contractions. The optic tract is identified by the current threshold required to evoke visual phenomena. A threshold for muscle contractions or for visual activation of <10 A indicates that the electrode tip is within, or is very close to, the corticospinal tract or optic tract, respectively. Some representative microstimulation thresholds are marked on the microelectrode track reconstructions. Microstimulation - evoked activation of optic tract and of CST are sought when the microelectrode up exists the Gpe - Gpi complex. CST is encountered only on trajectories that pass posterior of Gpi of that exist Gpi near its posterior margin.

Localization of the optic tract by microstimulationn evoked phosphenes requires that the patient be alert and cooperative, to consistently report subjective visual phenomena. In a less-than - cooperative patient it is probably more reliable to localize the optic tract by light-evoked action - potential discharges.

Stereotactic targeting routinely places our initial microelectrode track in sensorimotor Gpi within 1.5 mm of the S& B lateral 20 plane. In most cases, this track has a lengthy segment within Gpi and optic tract is identified at the bottom. Usually, only two additional microelectrode tracks are necessary for definition of the boundaries of Gpi: one tracks 3 - 4 mm posterior to the first, and one tracks 2 - 4 mm lateral to the first. Occasionally, the first tracks maps onto posterior Gpi; in this case the second track is made 3 - 4 mm anterior to the first one.

Macrostimulation

'Macrostimulation' refers to passing current through the lesioning probe or deep-brain stimulator. Macrostimulation provides the final check on target localization just prior to lesioning or to permanent stimulator placement. As with microstimulation, current thresholds for activating optic tract and internal capsule indicate the proximity of the lesioning probe to these structures. Using a radionics radiofrequency lesioning probe with an exposed tip of 1.2 x 3.0 mm, we stimulate with cathodal pulses, 0.2 ms, 300 Hz, o.1 - 2.0 mA. We have avoided damage to optic nerve and CST and requiring that, at an intended lesions site, the threshold for activating these structures by >1.0 mA or 0.5 mA. Respectively. If this criterion is not met, the lesioning electrode is moved.

Relation of microelectrode recording to other localization methods
Image - guided stereotaxis alone has fundamental limitations in its ability to guide the surgeon precisely to a predetermined target. These limitations include the application accuracy of the stereotactic system, which is at least 15mm, and the face that the accuracy of imaging is limited either by non-linear distortion of images or by inability to visualize nuclear boundaries. Furthermore, the targets for movement disorder surgery are ultimately defined by physiologic, not by anatomic, criteria. In the case of pallidal localization, the region that should be targeted is the sensorimotor portion of Gpi. It is only identified with certainly by detecting neurons whose discharge is modulated by joint movements.

Microelectrode recording adds precision to target localization by providing three types of information.

1. Position of the boundaries of nuclei, using spontaneous discharge patterns and identification of white matter laminae between nuclei.
2. Identification of the motor territory of a nucleus, and localizatioon within the territory, by identifying cell responses to joint movement and comparing those with the known somatotopic organization of the nucleus.
3. Localization of surrounding structures, using micro-stimulation-evoked motor or sensory responses, or by recording action potentials evoked by sensory stimuli.

In addition, microelectrode recording is performed intraoperatively rather than preoperatively, and can thus account for brain shifts that may occur after head positioning and burr hole placement. Since it is likely that long-term outcome from ablative surgery is critically dependent on lesion location, we except that the use of microelectrode techniques well lead to improved long-term outcome.

Use of macrostimulation techniques, while useful, does not obviate the need for microelectrode techniques. In the case of lesioning of Gpi for PD, for example, relief of parkinsonian motor signs, or failure to relieve these signs, by macrostimualtion at a particular site within Gpi, does not accurately predict the efficacy of lesioning at the same site. Therefore, macrostimulation thresholds are primarily used to avoid damage to structures outside of Gpi, whereas optimizing lesion size and location within Gpi depends on defining the boundaries of sensorimotor Gpi with microelectrode recording.

Disadvantages of extensive microelectrode recording are that it adds time and expense to the surgery. Multiple passes of the electrode theoretically increase the risk of hemorrhage. However, by optimizing sterotactic targeting for the initial microelectrode track as described earlier for pallidotomy, the number of passes necessary for complete mapping of the target structure is minimized.

Lesioning technique for Gpi
Our lesioning goal in pallidotomy is to safely lesion the sensorimotor portion of Gpi. We usually find that the sensorimotor portion of Gpi maps to the S&B atlas on parasagittal planes 18.5 mm through 23.0 mm. We utilize a lesioning probe with a 1.2 x 3 mm exposed tip and lesion times of 60s. when lesion characteristics with this probe are studied in egg white, the diameter of the coagulum varies nearly linearly from 1 to 4 mm for temperatures from 60 to 80 C. The exact X-,Y- and Z- coordinates of each lesion are then dictated by the microelectrode-derived map of the stereotactic coordinates of sensorimotor Gpi, combined with the predicted spread of 60-s lesions at a given temperature. Since the target region is geometrically complex, we use lesions at multiple depths along three or four parallel trajectories to 'shape' the lesion to the boundaries of sensorimotor along a parasagittal plane corresponding approximately S&B 21.5 - 22, and a third lesion tract 2 - 2.5 mm medial to the first two. The mean size of the lesion by early postoperative MRI is approximately 120 mm. An example of a reasonable lesioning strategy. However, the exact strategy varies, depending upon an individual patient's Gpi boundaries, grouping of movement related cells within Gpi, or body location of worst motor disability. MRI of a pallidotomy lesion 4 hours after surgery.

LOCALIZATION OF VENTRALIS INTERMEDIUS

As with pallidal localization, the techniques for localization of Vim and surrounding structures include image-guided stereotaxis, microelectrode recording and microstimulation and macrostimulation.

Direct stereotactic targeting of Vim is hampered by the face that the borders of this small nucleus, other than the lateral border, are not normally visualized on MRI. Thus, the target for the initial microelectrode exploration is selected largely based on standard measurements from the AC - PC line. The centre of Vim is approximately in the same axial plane as the AC - PC line, and approximately 5 mm anterior to the posterior commissure. The lateral coordinate is the most variable, depending on the width of the third ventricle, and is approximately half the width of the third ventricle plus 11.5 mm, or 2 mm medial to the border between thalamus and internal capsule.

Identification of Vim by microrecording and microstimulation. As with localization of Gpi, multiple parallel, parasagittal microelectrode tracks are made, and scaled, shaded track reconstructions are superimposed on parasagittal cuts from the S&B atlas in order to register the atlas in the sterotactic coordinate system. As the electrode traverses the borders of thalamic nuclei, changes in spotnaneous discharge rate and pattern are not as obvious as for the basal ganglia, and unlike Gpi, there are not clearly identifiable white matter laminae demarcating nuclear borders. Localization of the microelectrode tip within motor thalamus, therefore, depends strongly on identification of neuronal responses to passive and active limb movement, cutaneous sensory responses and microstimulation - induced paresthesias.

On the approach to thalamus, the first neuronal structure encountered may be the caudate nucleus, characterized by low spontaneous discharge rates. The first thalamic neurons to be encountered may be those in the nucleus reticularis, a thin nucleus that envelops much of the anterior thalamus. These neurons have spontaneous discharge rates that vary with the patient's state of attention and evoked discharges in response to diffuse stimuli. Other non-motor nuclie are then traversed. Entry into motor thalamus is indicated by the identification of neurons whose discharge frequencies are modulated by joint movements. Within motor thalamus, a subset of Vop and Vim neurons have discharge rates that oscillate in phase with the patient's tremor. Vop neurons tend to be modulated by active movement, and Vim by passive movement. On posterior tracks, portions of sensory thalamus may be entered. Vc contains neurons whose discharges are modulated by cutaneous stimuli, and have small well demarcated receptive fields. The somatotopic organization of Vim and Vc provides further localizing information. Face representations tend to be medial, leg is lateral, and arm is between the two. However, due to the 'onion skin' arrangement of somatotopic subdivisions, a single microelectrode track may enter a given somatotopic given at multiple points.

Microstimulation identifies the posterior and lateral borders of Vim. Laterally, stimulation thresholds of 5 - 10 A at 300 Hz for evoked muscle contractions indicate that the electrode is at or has transversed the border between Vim and internal capsule. Posteriorly, similar thresholds for evoked paresthesias indicate that the electrode has transversed the border between Vim and Vc.

Prior to the placement of a lesion or permanent stimulating electrode, macrostimulation provides the final check on target location. Suppression of tremor by high frequency stimulation predicts that a lesion or chronic stimulation in the same location will be effective. Thresholds for muscle contractions and cutaneous sensation must be greater than - 0.5 mA, otherwise the electrode is too close to CST or to Vc, respectively.

For lesioning, lesions are made along several parasagittal trajectories, as with pallidotomy. Exact lesion size and location is influenced by the patient's symptoms. Isolated distal arm tremor warrants a fairly small lesion in the arm territory of Vim. Leg involvement as well as arm mandates extension of the lesion laterally within Vim. Proximal involvement mandates a larger lesions, probably because the somatotopic representation of proximal musculature within Vim is anatomically more dispersed than that of distal musculature. Thus, the exact boundaries of the lesion are determined mainly by identifying the region of Vim where neuronal discharges are modulated by passive movement in the body distribution affected by the tremor. An example of a lesioning strategy for a patient with essential tremor.

RESTORATIVE THERAPIES

All of the previously discussed procedure attempt to alleviate motor signs by creating a compensatory blockade in neuronal circuits whose activity is abnormal. Another set of therapies, referred to here as 'restorative', seek to directly repair portions of the missing circuitry that have been lost through a degenerative process. These therapies fall into two categories.

• Intracerebral cell transplantation; and
• Intracerebral delivery of neuropathic factors

Restorative therapy is highly experimental but is currently being studied in
patients with Parkinson's disease and Hungtington's disease. theory, results, and techniques will be briefly reviewed.

CELL TRANSPLANTATION
Cell transplantation for Parkinson's disease

Idiopathic Parkinson's disease is the neurodegenerative disorder most amenable to a transplantation paradigm because it a relatively focal disease, which in early or mid-stages involves mainly the loss of one cell type, the dopaminergic cells of the SNc. Since the major projection from SNc innervates the striatum, the strategy of all clinical transplantation efforts thus far has been to place dopamine-secreting tissue within the striatum.

Adrenal chromaffin autografts
Intrastriatal autografting of adrenal chromaffin tissue was the first technique attempted in clinical transplantation for Parkinson's disease. Dennervated adrenal medulla secretes dopamine, and thus offered the theoretical promise of serving as an autologous 'dopamine pump' within the brain. This technique avoided the immunogloic and ethical issues associated with transplantation of non-autologous fetal tissues. Several studies showed only very modest improvement along with significant surgical morbidity. Autopsy studies of transplanted patterns revealed minimal survival of chromaffin cells.

Adrenal chromaffin graft survival, and differentiation into a neuronal phenotype, is enhanced by treating the grafts with nerve growth factor after implantation. Based upon this observation, one reported patient has undergone intrastriatal placement of an adrenal autograft, followed by a 23-day intraparenchymal fusion of nerve growth factor into the region of the graft. Clinical improvements in this patient were longer - lasting than in the same group's original effort with adrenal medulla tissue with peripheral nerve as a source of NGF. Laboratory work continues in an effort to find clinically applicable means of enhancing the survival of adrenal chromaffin grafts.

Fetal ventral mesencephalic allografts
Theory

In the developing mammalian fetus, the ventral mesencephalon contains the dopaminergic cell bodies that are the precursors of the mature SNc. In both rodent and primate model of Parkinson's disease, grafts of fetal VM tissue, placed heterotopically into striatum, have been shown to survive, reinnervate portions of the striatum, and correct abnormal motor behavior. Anatomic and physiologic studies of intrastriatal fetal VM grafts suggest that synaptically mediated host-graft interactions occur and result in a more complex level of function than would be expected from implantation of an unregulated dopamine pump.

Patient Characteristics and clinical outcomes
As of 1997, at least eleven groups have reported at least 200 patients who have undergone transplantation of human fetal VM cells to the striatum for parkinsonism. Most transplanted patients suffered from idiopathic parkinson's disease, although two reported patients suffered from MPTP - induced parkinsonism. All patients had relatively advanced disease, with Hoehn and Yahr stages of about 3 when on and 4 - 5 when off. Most patients were 40 - 60 years old and had suffered from Parkinson's disease for 5 - 20 years prior to surgery. Patients were primarily symptomatic from bradykinesia and rigidity rather than from tremor.

Outcomes have been highly variable. Observations common to many studies were improvements being rarely immediate but taking several months, a worsening in many patients motor symptoms during the first 4 - 6 weeks after surgery, and sometimes transiently increased dyskinesias. There were several transient psychiatric complications, including hallucinations, panic attacks and obessive - compulsive disorder. There were several complications of immunosuppressive therapy requiring its cessation. In most studies there was modest benefit in most but not all patients, particularly in reduction of amount of time spent off, reduction in drug-induced dyskinesia and some improvement in fine motor tasks. The majority of patients reported did not have a significant change in their H & Y functional status, and very few are off medication.

Types of transplant protocols used
All groups used human fetal tissue dissected from the ventral mesencephalon. Most groups used tissue from elective abortion, rather than from spontaneous abortions. Treatment protocols differed widely with respect to several important variables: exact location of transplant, surgical tenchnique used, unilateral versus bilateral implantation, age of fetal tissue used, preparation of tissue, number of fetuses used to provide donor tissue, interval between harvest and implantation, and use of immunosuppressive therapy. Inadequate attention to technical considerations has produced disastrous results.

Positron emission tomography (PET) scanning of graft survival
Survival of grafted fetal dopaminergic cells can be assessed non-invasively using PET. A tracer of dopaminergic metabolic activity, such as 6-fluoro-L-DOPA, is used as the positron-emitting source. The concentration of that positron source in the brain following intravenous administration reflects metabolic activity in dopaminergic nerve cells and terminals. In Parkinson's disease, uptake of the tracer in the striatum is abnormally decreased compared with normal controls. Several groups have used PET to follow survival of fetal transplants in the striatum. In the patients followed for the longest time, there is evidence for graft survival at three years. Tracer uptake by PET scanning cannot distinguished between survival of transplanted tissue versus sprouting of host dopaminergic tissue, which is known to occur with lesioning of the striatum. However, shot of autopsy studies, PET is currently the best method to follow graft survival.

Autopsy studies of graft survival
One autopsy report, from a patient transplanted at the University of South Florida, demonstrated excellent graft survival. This patient died 18 months after grafting from a pulmonary embolism following orthopedic surgery. He had received bilateral grafts into the putamen had been immunosuppressed for 6 months after transplantation. Many tyrosline hydroxylase positive cells were observed in the grafts, up to 1000 per section, and they extended neuronal processes up to 7 mm into surrounding normal brain. There was no evidence of sprouting of TH-positive fibers from the host brain. A section through the grafted region of this patient's brain. This study demonstrated that human fetal tissue implanted in a human can survive and robustly reinervate host tissues, and that long-term immunosuppression may not be needed for survival of human neural tissue transplants, even if multiple donors are used.

Which transplant protocols were most effective?
A small number of very well-documented patients have had dramatic clinical benefit with good evidence of graft survival, and it is worth considering them individually. One patient from the Swedish series of idiopathic Parkinson's disease patients is the only reported transplant patient now off all dopamimetic drugs at three years follow-up. This patient's off periods have operated side. Symptoms continued to improve during the second year after surgery, then remained stable. Two patients with MPTP - induced parkinsonism, also opearted by the Lund group, have also had great benefit, with 50-point decreases in UPDRS scores two years post-transplant, reductions in medications, elimination of drug -induced dyskinesias, and continued improvement during the second year. The four patients in the University of South Florida group also had greater and better documented improvement than most patients in other series, with a 22 point decrease in total UPDRS when off, reduction in percentage off times from 30% to 12%, and reduction in the percentage time on with dyskinesia from 44% to 3.8%.

Technical parameters common to the best documented patients showing best results are: use of younger tissue, implantation of tissue from multiple fetal sources as multiple sites; implantation into the putament rather than just the caudate alone; and surgery using stereotactically guided injection rather than open craniotomy. Although long-term immunosuppression may not be necessary for graft survival, most patients with the best clinical results had at least short term immunosuppression. In order to better assess the benefit of fetal transplantation for Parkinson's disease, two randomized, double-blind studies have been sponsored by the National Institutes of Health.

Alternative tissue sources for cell transplantation
Harvesting enough human fetal cells for a successful transplantation, all at the same stage of development, is logistically difficult and raises ethical questions. Therefore, a variety of alternative tissue sources are being explored, including xenogeneic tissues, and genetically engineered autografts that produce growth factors or L-DOPA/dopamine.

Dopamine-secreting xenografts
The use of porcine xenografts for neuronal cell transplantation, as an alternative to human fetal allografts, is under active investigation. In rats systemically immunosuppressed with cyclosporin, porcine fetal VM cells can survive, extend processes and integrate into surrounding brain tissue.

Immune-mediated rejection of xenografts is, in general even more vigorous than that of allografts; this could limit the clinical utility of xenografts. Transplant rejection is mediated mainly by major histocompatibility antigens. One strategy employed to circumvent rejection is to used an antibody fragment directed against MHC antigens to induce tolerance of the foreign tissue in the host environment. Treating pancreatic and hepatic xenografts with the variable region fragment of a monoclonal antibody against MHC antigens, just prior to transplantation, induces tolerance in the host. This tolerance is long-lasting even though the transplant is treated only once. By using antibody fragments from which the constant region has been removed, activation of the complement cascade and antibody mediated destruction is avoided while cellular immunity is modified.

This MHC masking technique has been applied to porcine neural xenografts. When pretreated with fragments of a monoclonal antibody against antigens, intrastriatal xenografts of porcine fetal VM survive in non-immunosuppressed host rats as well as untreated xenografts in immunosuppressed rats. In 1995 a clinical trial of porcine fetal neuronal transplantation for Parkinson's disease began, using - treated cells in one arm of the study.

Another strategy to shelter xenografts from the host immune system is to encapsulate xenograft tissue within a synthetic, biologically compatible polymer coating that allows diffusion of xenografts - produced molecules but prevents immune attack of the xenograft. In the hemiparkinsonian MPTP-monkey, implantation of polymer-encapsulated PC12 cells, a dopamine - secreting cell line derived from a rat pheochromocytoma, is effective in ameliorating parkinsonian symptoms. a disadvantage of implanting encapsulated tissue is that graft - host synaptic interactions, which may be an important mechanism of behavioral recovery in fetal cell transplantation, are prevented.

Genetically engineered autologous tissue
To completely circumvent the immunologic complexity of xeno-or allotransplantation, a patient's own tissues could be removed, genetically engineered to adopt a useful neuronal phenotype, then reimplanted into the host brain. Primary fibroblasts have been engineered to express tryosine hydroxylase using retrovirus - mediated transfection; these cells then secrete L-DOPA. Implantation of autologous L-DOPA secreting fibroblasts in the 6-hydroxydopamine model of parkinson's disease results in behavioral improvement for at least 8 weeks post-implantation, but the behavioral improvement declines after 2 weeks and the implanted fibroblasts show some signs of neoplasia. The use of plasmid-transfected primary cultured muscle cells, in the same experimental paradigm, has resulted in more stable behavioral amelioration, with evidence of continued L-DOPA secretion up to 6 months post-transplant. As with polymer-encapsulated tissue, a disadvantage of using genetically engineered non-neuronal tissues is that it is not possible for such tissues to form functional synapses with host tissue. In addition, long-term expression of foreign genes in transfected cell is difficult to achieve with present gene transfer methods.

Cell transplantation for Huntington's disease
Fetal cell allografting has also been attempted for Huntington's disease. Tissue from the lateral ganglionic eminence, the precursor of the mature striatum, has been placed stereotactically along multiple trajectories in the caudate nucleus and putamen.

The theoretical rationale for cell transplantation in Huntington's disease is less developed than is the case for Parkinson's disease. while striatal degeneration is the earliest hallmark of Huntington's disease, the disease ultimately produces widespread brain pathology. Nevertheless, intrastriatal fetal cell transplantationn in animal models of Huntington's disease has shown promise, at least for alleviating motor signs. Given that there is currently no effective treatment for Huntington's disease, and that the disease inexorably progresses towards total disability and death, careful clinical trials seem justifiable. There is currently inadequate follow-up time, and too few patients, to assess outcome.

NEUROTROPHIN THERAPY FOR PARKINSONS'S DISEASE

In Parkinson's disease intracerebral transplantation of dopamine-secreting cells can correct dopamine depletion in host tissue. However, the transplantation paradigm does not necessarily reverse the fundamental pathology of Parkinson's disease, that is progressive loss of dopaminergic cells of the SNc. It is possible that transplanted cells are subject to the same degenerative process as host dopaminergic cells. As a result, there has been interest in identifying a neurotrophic factor which could halt or reverse the loss of SNc neurons. Such a factor could provide a more direct and effect treatment for Parkinson's disease than any of the procedures discussed above, and could be particularly useful for early Parkinson's disease prior to extreme loss of SNc cells.

Currently, the most potent dopaminergic neurotrophin is glial cell - derived neurotrophic factor. GDNF is a member of the transforming growth factor superfamily, isolated in 1993 on the basis of its ability to support the survival an differentiation of embryonic midbrain dopaminergic neurons in culture. GDNF has been shown to protect dopaminergic neurons from toxic insults in vivo. In the MPTP-treated primate model of Parkinson's disease, intracerebral administration of GDNF results in improvements in tremor, rigidity, and bradykinesia, as well as increases in the number of dopaminergic neurons within the lesioned SNc. Interestingly, striatal dopamine remained depleted after GDNF treatment, raising the possibility that amelioration of parkinsonian motor signs may not depend strictly on restoration of striatal dopamine. An initial clinical trial of intraventricular administration of GDNF in Parkinson's disease patients of in progress.

CONCLUSIONS

Surgery for movement disorders is a rapidly advancing area. While ablative surgery has been performed for over forty years, changes in the patient population, surgical techniques, and standards for quantitative motor assessment have changed, mandating a thorough reassessment of ablative techniques progresses, non-ablative and restorative technique promise to drastically alter clinical treatments.

At this time, there is little rigorous long-term efficacy data for any of the surgical options for movement disorders. At evaluation up to two years, unilateral posterolateral Gpi pallidotomy for Parkinson's disease offers significant benefit for most motor signs. Thalamotomy and thalamic stimulation have proven efficacy for Parkinsonian and essential tremor. Beyond this, the many other procedures and indications discussed here are highly experimental.

The technical armamentarium for localization in movement disorders surgery includes image - based stereotaxis, microelectrode recording, microstimulation, and macrostimulation. All of these techniques should be used in concert for optimal results. For all types of movement disorder surgery, extreme precision in localizing specific subdivision of subcortical nuclei will probably prove essential for optimal long-term outcome and complication avoidance.

BACK