Benjamin S. Carson, M.D., Sam P. Javedan, John M. Freeman, M.D., Eileen P. G. Vining, M.D., Aaron L. Zuckerberg, M.D., Jeremy A. Lauer, M.S., and Michael Guarnieri, Ph.D.
Division of Pediatric Neurosurgery, Department of Neurological Surgery, and Pediatric Epilepsy Center, Johns Hopkins Medical Institutions, Baltimore, Maryland; and Division of Pediatric Critical Care Medicine, Sinai Hospital of Baltimore, Baltimore, Maryland
Between 1975 and 1994, 52 hemispherectomies, of which two were anatomical and 50 hemidecortications, were performed at Johns Hopkins Medical Institutions. Eighteen patients were 2 years old or less. There were three perioperative mortalities and one patient died 9 months later from causes not related to surgery. One patient developed hydrocephalus 6 years postsurgery and has been treated effectively. Seizure control and the functional status of each patient were measured as outcome variables. Forty-six (96%) of the surviving patients were seizure free or had reduced seizures as of their last follow-up examination. Twenty-one individuals (44%) were participating in age-appropriate classes or working independently, 18 were classified as semiindependent, and nine children will likely depend on a lifetime of assisted living. The relationships between the outcome variables and the patient's age at surgery, the interval to surgery, and the etiology of the disease were compared. The authors' clinical experiences strongly suggest the importance of a multidisciplinary approach to patient selection and follow-up care. Moreover, anesthetic management of infant surgery is a major component of success.
Key Words * pediatric epilepsy surgery * hemispherectomy * hemidecortication * Rasmussen's syndrome * developmental syndrome * vascular syndrome
The term "hemispherectomy" is applied to surgeries in which all or large amounts of cortical tissue, including the motor and sensory strip, are removed. The term also includes hemidecortication or hemicorticectomy (removal of cortical gray matter with preservation of the ventricle and a layer of white matter) and functional hemispherectomy (removal of a large central block of tissue and disconnection of the anterior and posterior elements). Children who have medically intractable seizures arising diffusely from lesions in a single hemisphere are candidates for the procedure.
Recent reviews of hemispherectomy illustrate the critical importance of long-term analyses of surgical interventions, especially those which initially appear to be highly successful.[10,36] Briefly summarized, in 1950 Krynauw[28] published the first hemispherectomy case series for infantile hemiplegia. Subsequent reports confirmed and extended his initial findings. The operation eliminated or reduced seizures in 70% to 80% of patients. In the majority of reported cases, surgery improved the patients' development, led to increased intelligence quotient scores, and often eliminated profoundly negative behavior. Perioperative mortality rates were 5% to 6% in case series from 1950 to 1970.[58] Long-term morbidity and mortality, however, gave rise to a much greater concern. Late complications, rarely seen before 3 to 4 years postsurgery, included signs of insidious neurological deterioration and chronic increased intracranial pressure (ICP). Oppenheimer and Griffith[35] attributed such complications to hemosiderosis secondary to minute hemorrhages. Mortality rates attributable to neurological deterioration and increased ICP reached 35%.[57,58]
Once the nature of the late complications became apparent, several neurosurgeons suggested that these complications could be controlled by postoperative monitoring and prompt intervention.[22,31,41] Others modified the classic operation (described by Dandy[14] in 1928) to prevent long-term complications.[1,25,29,30,32,36,42,43,52,54,59] Adams,[1] for example, reduced the subdural space by suturing the dura to the falx, tentorium, and floor of the middle and anterior cranial fossae and insulated the subdural cavity from the ventricular system by obstructing the ipsilateral foramen of Monro with a plug of muscle. Rasmussen[40] resected the central region and disconnected the residual frontal and parietooccipital lobes from the rest of the brain creating a "functional hemispherectomy." Although certain modifications probably have limited applications, the number of reports illustrates the growing confidence in hemispherectomy to treat intractable seizures. A 1993 survey revealed that 29 of 47 neurosurgery centers perform hemispherectomies and that, of those that do not, most are planning to initiate them.[44]
The history and evolving nature of hemispherectomy technology mandates periodic review. Recent case reports indicate that modified anatomical procedures, functional hemispherectomy, and hemidecortication provide the same seizure relief; however, more information is needed about the control of long-term complications.[10,13,36,50,52,59] A second concern is the functional status of patients. Quality of life issues have not been widely applied in epilepsy, yet such issues are emerging as important health care outcomes.[6] In addition, there has been a shift in the patient population to younger children. Surgery for epilepsy often has been delayed pending 2 years of drug trials with a major spectrum of anticonvulsant medications; however, recent publications indicate that earlier surgery may be appropriate for younger patients[16,19,36] and that earlier surgery leads to improved outcomes.[2,9,11,18 ,23,25,33,34,37-39,54,56] Nevertheless, compared to older children, surgery in infants poses far greater challenges.
We now describe the seizure status and functional outcome of 52 hemispherectomies, a series spanning the period 1975 to 1994. The age at operation was two years or less for 18 patients; thus, this is the largest series of infant surgeries yet reported.
Clinical Material and Methods
Patient Population
Patients were selected for hemispherectomy on an individual basis, in consultation with their family, neurologist, and neurosurgeon.[5,54,55] General characteristics of the 52 patients are given in Table 1. The pathological condition indicating surgery fell into three groups: Rasmussen's, developmental, and vascular syndromes. The last group was subdivided into patients with Sturge-Weber syndrome and those having other vascular diseases.
Rasmussen's syndrome, a "progressive unilateral encephalopathy of childhood,"[54] is defined by clinical findings such as a sudden onset of focal seizures with progressive unilateral spread and eventual hemiparesis.[51] Pathological findings include perivascular round-cell infiltration, microglial nodules, astrocytosis, and spongy degeneration. Although diagnosis is determined almost exclusively on the basis of clinical impression, pathological examination of the excised hemisphere in 22 of our patients generally revealed findings consistent with the diagnosis. Developmental syndromes found in 23 patients included congenital malformations such as hemimegalencephaly, tuberous sclerosis, hamartomas, sebaceous nevus cyst syndrome, and others. These syndromes were diagnosed using the physician's clinical impression and neuroradiological data and were confirmed by pathological analysis of excised tissue. Previously, we had placed patients with Sturge-Weber syndrome, a congenital vascular malformation diagnosed on the basis of the patient's medical history and skin findings, in the "developmental abnormalities" category.[56] In this report, we have placed the three patients with Sturge-Weber syndrome in the "vascular" category, which also includes four patients suffering from congenital hemiplegia with porencephaly.
Fourteen patients underwent previous focal resections with insufficient relief. A clear hemiplegia or handicapping hemiparesis was present in 33 patients preoperatively.
As shown in Table 1, the Sturge-Weber population had the youngest patients, with an average age at the time of surgery of 1.3 years compared with an average age of 9.9 and 3.1 years in the Rasmussen's and developmental series, respectively. The patients with Sturge-Weber and developmental syndromes had the shortest average interval from initial diagnosis to surgery: 0.6 and 2.5 years, respectively.
The two initial patients in this series, one a 14-year-old boy and the other a 16-year-old boy, both with Rasmussen's syndrome, were treated with an anatomical hemispherectomy; a hemidecortication procedure was used in the remaining 50 cases.
Surgical Methods
In general, we started the procedure with a standard temporal lobectomy, followed by a frontal lobectomy and then a parietal lobectomy, followed by an occipital lobectomy. In most cases this did not add to the length of the operation but could result in substantial blood loss, particularly if a coagulopathy developed. Because most hemispheric surface bleeding is by way of the middle cerebral circulation, there is little reason to attack the anterior cerebral circulation when using the lobectomy approach. While performing the frontal lobectomy, the branches of the anterior cerebral artery were coagulated and divided as they were encountered. In an attempt to preserve the majority of the ventricular system, we performed a decortication, which leaves a thin mantle of white matter over the ependymal surface and leaves most of the basal ganglia and thalamus intact. Because a complete temporal lobectomy has been performed, the cerebrospinal fluid (CSF) outflow from the trigonal region was occluded with Gelfoam and Surgicel, which were used to reconstruct a ventricular surface.
One further modification must be made because, occasionally, one encounters significant bleeding from the cortical veins, which for unexplained reasons are often engorged as they drain toward the sagittal sinus. An abnormal blood supply to the affected hemisphere of children with hemimegalencephaly has been noted previously.[4] To avoid this problem, we decided to create a cortical channel that was parasagittal in nature and extended from the anterior falx to the posterior extent of the occipital lobe. This channel includes not only the cerebral tissue; it also interrupts all of the draining veins, greatly decreasing their tendency to bleed as the residual wedge of cortex adjacent to the falx and sagittal sinus is removed. The resultant perisagittal wedge of cortical tissue was resected after coagulating and dividing the residual portions of the cortical venous system as they enter the sagittal sinus. This technique undoubtedly results in a triangular wedge of cortex containing the venous stubs that is very similar to that described by Dandy.[14] The insular cortex was the last area of resection because in very young patients, manipulation of this area seems to be associated with some degree of cardiovascular instability.
Once hemostasis was achieved, the hemispherectomy cavity was copiously irrigated until the fluid was crystal clear. While keeping the cavity dry, a layer of Gelfoam and Surgicel was placed over the residual raw surface. This allows these materials to adhere and, hopefully, form the basis of an adhesive layer that will preclude subsequent hemorrhage. We have discovered no adverse effects from leaving the hemispherectomy cavity unfilled with saline. The dura was closed in a water-tight fashion and was not stitched medially to the falx and tentorium or laterally to the inner table of the skull. Eventually the hemispherectomy cavity fills with CSF. Introduction of this fluid does not appear to cause the layer of Gelfoam and Surgicel to float off.
Outcome Determination
Seizure frequency was ascertained either from reports of stays in the inpatient epilepsy monitoring unit, doctor referral notes, or family reports. Follow-up information was obtained from each patient's hospital record, clinical visits to the departments of neurology and neurosurgery, and telephone interviews with patients and their families. The length of follow-up review was calculated in years as the lowest whole number difference between the year of last recorded contact with the patient and the year of the operation. Thus a patient who underwent surgery in June 1992 would be assigned 2 years of follow-up review if contacted in May 1995.
Scatterplots of outcome variables were prepared using a commercially available software program (KaleidaGraph for Macintosh, version 4.1; Synergy Software, Reading, PA).
RESULTS
Three patients died perioperatively. A patient with Sturge-Weber syndrome (a 1-year-old boy) had massive diploic veins and uncontrollable bleeding from burr holes. His surgery had to be completed under cardiac standstill, and the patient died of complications. Two female patients in the developmental group (one girl aged 1 year and the other 2 years) died from what appears to have been autonomic instability. These cases have been discussed briefly.[8] A fourth mortality, not related to surgery, was that of a 2-year-old girl with developmental dysplasia who developed seizures in the remaining hemisphere and died 9 months posthemispherectomy.
The data provided in Table 2 show that 46 (96%) of the remaining patients were seizure free or had reduced seizures as of their last follow-up examination; 19 of these had been weaned from anticonvulsant medications. Two patients (4%) experienced essentially no change in seizure frequency. Of the three categories, patients in the Rasmussen's group exhibited the best seizure outcomes. Overall, 16 of 22 experienced no seizures postoperatively, and the majority of these patients no longer took anticonvulsant medications. Outcomes among the patients with developmental diseases were less successful, due perhaps to the variety of underlying conditions included in this category.
Excluding the four fatalities previously discussed, significant intraoperative bleeding and perioperative fluid loss and management problems were encountered in six cases of developmental syndromes, five of Rasmussen's syndrome, and one case of Sturge-Weber syndrome. Eight of these cases have already been described.[8] Additional perioperative findings are listed in Table 3. Radiologically detectable bleeding into the subdural cavity occurred in 21 patients (44%) in the several days immediately after surgery. The bleeding was active and required intervention in only three cases; in the remaining cases it spontaneously resolved within several days to weeks. Fever frequently was encountered in the days immediately following the operation; however, only nine patients (19%) had blood or CSF cultures that tested positive for infection. Hydrocephalus was encountered in 16 patients (33%) who later required ventriculoperitoneal shunt placement. More numerous perioperative findings were identified among the developmental group; we believe this reflects the younger age range of the group, an age range placing the patients at greater risk for anesthetic and surgical complications.
Other side effects including osteomyelitis or skull deformities leading to subsequent cranioplastic corrections (boneflap removal) were encountered in two patients. Signs of precocious puberty were found in six girls in our series.
Contralateral hemianopsia and hemiparesis were encountered in all of the patients, although the disability caused by the latter varied. Almost all parents, even those whose children are likely to remain dependent, stated in follow-up interviews that the quality of their children's lives is dramatically improved. The one exception is a parent whose child had been treated at another institution by a functional hemispherectomy with incomplete seizure relief. Shortly after returning from surgery, the child, a 10-year-old girl with Rasmussen's syndrome, suffered what appeared to be a seizure. A computerized tomography (CT) scan revealed massive edema of the remaining hemisphere. A magnetic resonance (MR) image confirmed the edema with no evidence of subdural or intracranial bleeding. Repeated imaging showed that the edema resolved over several days; however, the child lapsed into a partial coma and has never fully recovered.
The functional status of the patients, determined through interviews with patients in the adult cases and with parents of minors, is described in Table 4. "Dependent" status was assigned to individuals who require assistance with daily living functions such as eating, dressing, and bathing, and who will probably continue to be dependent in adult life. "Semiindependent" status denotes individuals who perform near-age-appropriate daily living functions, who often attended special education programs, and who work in sheltered employment. It is possible that people, currently in the semiindependent group, with sufficient training and education could in the future become independent. "Independent" status denotes individuals who are age appropriate in school or who work independently despite physical disabilities. If there were questions about an individual's score, the more dependent score was assigned. Overall, nine (19%) of the surviving patients will likely depend on a lifetime of assistance with feeding, dressing, and personal hygiene. In this group, we included four infants (> 1 year of age) with clinical evidence of permanent disabilities. Eighteen (37%) of the surviving patients were classified as semiindependent; 21 (44%) were in age-appropriate classes or working independently.
Formal intellegence quotient testing in this population has major limits and constraints. The age of the patient, frequency of seizures, and use of medication are all factors that limit accurate assessment prior to surgery. Follow-up data are being accumulated.[7] It is the impression of the families and physicians involved that the patients' general intellectual functions usually have improved, often dramatically, and sometimes have become normal.
Table 5 summarizes outcome data from 26 patients who have been followed for at least 4 years. The postoperative seizure status and the functional status of this subgroup are similar to those of the entire group (Tables 2 and 4, respectively). We have followed each patient in this series through clinic visits and phone consultations. A 14-year-old boy, 6 years postsurgery for Rasmussen's syndrome, recently received a shunt for hydrocephalus. This patient had been seizure free without drugs and was functioning independently. Except for this case, we have not encountered late complications of the type previously described.[35,57,58]
Discussion
Surgical Observations
The choice of initial surgical methods was influenced by the successes that have been described with en bloc resection of the hemisphere after surgically controlling the major feeding vessels.[21,25] Initially, when this lobe-by-lobe hemispherectomy was undertaken, the main branches of both the anterior cerebral and middle cerebral arteries would be clipped distal to the perforating arteries. Later, we noted that after controlling the middle cerebral artery circulation, the vast majority of bleeding was controlled and there was no need to clip the anterior cerebral artery, a position that was strongly confirmed after we observed two cases of crossing anterior cerebral arteries on the corpus callosum dorsal surface. Although it is unlikely that a surgeon would clip the wrong anterior cerebral artery, the risk of such a mistake in the case of crossing anterior cerebral arteries appeared to outweigh the benefit. Initially, vascular control included the anterior, middle, and posterior cerebral arteries. This was particularly easy to accomplish in cases in which the hemisphere was atrophic. In the mid to late 1980s, our focus began to shift more toward younger patients because we believed that these individuals had the greatest potential for functional recovery after hemispherectomy.[54] We found that younger patients' hemispheres were less atrophic because the disease process, which most frequently was Rasmussen's syndrome, had not yet matured. In some cases, rather than encountering atrophy, we found an actual enlargement of the affected hemisphere with delicate and friable surfaces. This made it extremely difficult to elevate the posterior temporal lobe and to trace the posterior cerebral artery toward the posterior midbrain and clip it. After completing a number of hemispherectomies without controlling the posterior cerebral circulation, we noted that omission of this maneuver did not significantly affect the amount of bleeding or the outcome of surgery.
Three of our early patients experienced a postoperative decline in consciousness, which in most cases lasted only a few days. In one case, the patient remained in a coma for over a month before emerging to achieve a good outcome. There was no obvious explanation for this and we suspect that the manipulation of the brainstem associated with en bloc resections might be responsible. Although it has been suggested that the decision to use the fragmentation or en bloc technique is a matter of technical preference,[51] we found that by performing hemispherectomy piecemeal, that is, progressing lobe by lobe, the necessity for lifting and manipulating large sections of the brain is obviated, and the likelihood of edema occurring in the opposite lobe or brainstem is much smaller. For that reason, we performed all subsequent hemispherectomies one lobe at a time. During the development of this strategy, we postulated that the possibilities of late hemorrhagic complications might be decreased by preserving the majority of the ventricular system; thus the decortication approach was adopted.
After hemispherectomy, if hydrocephalus developed, it was characterized by low or normal pressure and, therefore, a low-pressure shunting system was used. The shunt was placed into the contralateral ventricle. The disadvantage of placing the shunt in the hemispherectomy cavity is the possibility of its being occluded by debris external to the adhesive coating created by the application of Gelfoam and Surgicel. Also, shunting of the cavity might encourage shifting of the residual hemisphere toward the hemispherectomy cavity.
Perioperative Findings and Late Complications
Significant intraoperative bleeding has been described in several series. Griffiths and coworkers[24] described catastrophic intraoperative bleeding in two infants, one of whom died. Others have recommended staged operations to control intraoperative blood loss[45] and rapid surgical procedures to decrease the patient's exposure.[43,52] More information is needed concerning the cause of this bleeding, especially if the tendency toward surgery in infants continues.
Rasmussen[40] has noted that low-grade, increased ICP is commonly present for 7 to 10 days in patients who have extensive cranial resections and it usually resolves spontaneously without specific treatment. A persistent increase in ICP beyond the usual postoperative period suggests inadequate absorption of the CSF. Prompt shunting should be considered when there is clear evidence of elevated ICP that persists past the usual postoperative period because with prompt shunting, the increased ICP has no long-term effect.[40]
The incidence of postoperative shunting has varied among hemispherectomy series. Rasmussen[40] reported three shunts in 31 patients with anatomical (classic) hemispherectomy, one in 14 patients with functional hemispherectomy, and five in 60 patients with subtotal hemispherectomy. Ogunmekan and coworkers[34] used three shunts in a series of 12 patients with Sturge-Weber-Dimitri syndrome. Taha and coworkers[45] used two shunts in a series of three hemimegalencephalic cases. Winston and coworkers[59] used one shunt in a series of 11 hemispherectomies for various causes. Peacock[12] and others[50] examined the potential of routine shunts to prevent late complications following hemispherectomies in 21 patients with an average age of 6 years and an age range of 3 months to 14 years. The follow-up duration in that series is now approaching 10 years in the earliest patient, and no hemorrhagic complications have been observed. However, there have been a number of nonhemorrhagic complications, including CSF infections (two patients) and shunt obstructions (three patients). More recently, patients have received shunts only when they displayed features of raised ICP (W. J. Peacock, personal communication, 1995).
Long-term complications of hemispherectomy, including mental slowing, somnolence, tremor, ataxia, and evidence of chronic increased ICP, have been attributed to slow seepage of blood into the subdural hemispherectomy cavity and subsequent superficial hemosiderosis.[35,58] In most recent series, late complications have been monitored by newer imaging techniques such as CT and MR imaging. The low rate of late complications in this series (Table 5) is consistent with the low rate described in other post-1970 reports (Table 6).
Seizure Status
Postoperative seizure status has dominated outcome assessment for epilepsy surgery since Horsley[26] reported the first successful resection in 1886.[48] Krynauw[28] reported that 10 of 12 patients in his series were seizure free without medication. Similar outcomes have been reported in all subsequent series, although by current standards the follow-up time in many series was short.[10] In the present study, hemispherectomy reduced or eliminated seizures in 22 of 22 patients with Rasmussen's syndrome, 19 of 20 patients in the developmental group, and five of six vascular patients.
In a series of functionally hemispherectomized patients with an average age of 7 years (16 months to 15 years), Villemure and Rasmussen[53] observed that etiology and age at surgery did not seem to play decisive roles in the effectiveness of the operation in controlling seizures. They found a similar lack of correlation with respect to the interval between the onset of seizures and surgery. Hemispherectomy provided better seizure control for patients with Rasmussen's syndrome in our series compared to patients in the developmental category (Table 2). When these data were plotted (KaleidaGraph software), we found little or no correlation between age at surgery or interval to surgery and the effectiveness of hemispherectomy in controlling seizures in developmental patients. However, we observed a mild but positive correlation between seizure control in the patients with Rasmussen's syndrome and their age at surgery. The plots for the Rasmussen's group illustrated a similar inverse relationship between seizure control and interval to surgery. However, it is difficult to assess the significance of such trends because the seizure status recorded at discharge changed over a period of years in several cases. Children who were seizure free, with or without medication, occasionally developed seizures. Conversely, others who experienced auras or small seizures eventually improved. Engel and coworkers[20] reviewed the problems of obtaining objective measures of seizure outcome. They found that seizure patterns change with time, patient population, drug therapy, and as yet poorly understood psychosocial factors. Our findings agree with these observations and suggest that seizure control scores must be interpreted in the overall context of the patient's functional abilities, including abilities that reflect the handicapping nature of the seizures. Pre- and posthemispherectomy evaluation of the good hemisphere with emerging functional techniques may allow better evaluation of the relative roles of plasticity, anticonvulsant medications, and electrical interference.
Functional Outcomes
Behavioral changes have been used to monitor the effectiveness of hemispherectomy.[28] Cairns questioned this outcome measurement; others noted that many psychological factors, including parents' anxiety, played a role in the child's behavior.[27] Wilson[58] was the first to present social and economic factors as outcome measurements for hemispherectomy. In his series of 50 patients who underwent "classic" surgery, a "favorable" social movement from institutional to family life or from unemployment to full earning occurred in 10 cases. Thirty patients showed no "vertical" social movement, "but the quality of their lives and the lives of their families was almost always improved"; three experienced an "unfavorable" movement. Lindsay and coworkers[29] found a similar distribution in patients who had several types of surgery. Davies and coworkers[15] reported that 10 patients in a series of 17 who underwent classic hemispherectomy between 1950 and 1971 became employable after surgery, usually in a sheltered setting. More recent series also have started to monitor social and economic outcomes.[40,46]
We examined the relationship between the current functional status of each individual, their age at surgery, and their interval to surgery. There was no significant correlation between the patient's current function and age at surgery or interval to surgery in patients with Rasmussen's or developmental syndromes. The absence of this correlation in patients with Rasmussen's syndrome may relate to the overall higher functional scores in this group compared to the developmental cases: 15 patients in the Rasmussen's series are in the independent category, whereas only four members of the developmental group are independent. Additional data points may shed light on this issue. We have performed ten additional hemispherectomies with good results since January 1995, and the future addition of these results to our case reviews may offer more insight into the roles of plasticity (determined by age at surgery) and accumulated injury (determined by the interval to surgery) in the outcome of hemispherectomy.
Finally, our clinical experience strongly suggests that when very small children and infants undergo hemispherectomy, it is extremely important that the surgical and anesthetic teams be in close contact to ensure prompt and adequate replacement of blood loss. In an infant, it does not take long to lose an entire blood volume; one must constantly be aware of the possibility of coagulopathy and replace clotting factors with fresh frozen plasma if whole blood is not being used for the replacement. In addition, temperature in infants must be monitored and maintained in a relatively normothermic range. Instability of vital signs seems to be associated with significant hypothermia and this instability is more difficult to control in the presence of hypothermia.
References
1. Adams CBT: Hemispherectomy--a modification. J Neurol Neurosurg Psychiatry 46:617-619, 1983
2. Adler J, Erba G, Winston KR, et al: Results of surgery for extratemporal partial epilepsy that began in childhood. Arch Neurol 48:133-140, 1991
3. Ameli NO: Hemispherectomy for the treatment of epilepsy and behavior disturbance. Can J Neurol Sci 7:33-37, 1980
4. Andermann F: Functional hemispherectomy: clinical indications and outcome, in Wyllie E (ed): The Treatment of Epilepsy: Principles and Practices. Philadelphia: Lea & Febiger, 1993, pp 1110-1115
5. Arroyo S, Freeman JM: Epilepsy surgery in children: state of the art. Adv Pediatr 41:53-81, 1994
6. Baker GA: Health-related quality-of-life issues: optimizing patient outcomes. Neurology 45 (Suppl 2):S29-S34, 1995
7. Brandt J, Vining EPG, Stark RE, et al: Hemispherectomy for intractable epilepsy in childhood: preliminary report on neuropsychological and psychosocial sequelae. J Epilepsy 3 (Suppl):261-270, 1990
8. Brian JE Jr, Deshpande JK, McPherson RW: Management of cerebral hemispherectomy in children. Clin Anesthesiol 2:91-95, 1990
9. Carlson J, Netley C, Hendrick EB, et al: A reexamination of intellectual disabilities in hemispherectomized patients. Trans Am Neurol Assoc 93:198-201, 1968
10. Carson BS, Lauer JA, Guarnieri M: Hemispherectomy: a review. Neurosurg Q (In press, 1996)
11. Chugani HT, Jacobs B: Metabolic recovery in caudate nucleus of children following cerebral hemispherectomy. Ann Neurol 36:794-797, 1994
12. Chugani HT, Shewmon DA, Peacock WJ, et al: Surgical treatment of intractable neonatal-onset seizures: the role of positron emission tomography. Neurology 38:1178-1188, 1988
13. Cosgrove GR, Villemure JG: Cerebral hemicorticectomy for epilepsy. J Neurosurg 79:473, 1993 (Letter)
14. Dandy WE: Removal of right cerebral hemispheres for certain tumors with hemiplegia. Preliminary report. JAMA 90:823-825, 1928
15. Davies KG, Maxwell RE, French LA: Hemispherectomy for intractable seizures: long-term results in 17 patients followed for up to 38 years. J Neurosurg 78:733-740, 1993
16. Devinsky O, Pacia S: Epilepsy surgery. Neurol Clin 11:951-971, 1993
17. Di Rocco C, Iannelli A, Marchese E, et al: Terapia chirurgica dell' emimegalencefalia epilettogena. Minerva Pediatr 46:231-237, 1994
18. Duchowny M: Identification of surgical candidates and timing of operation: an overview, in Wyllie E (ed): The Treatment of Epilepsy: Principles and Practices. Philadelphia: Lea & Febiger, 1993, pp 999-1008
19. Duchowny MS: Pediatric epilepsy surgery: special considerations, in Wyler AR, Hermann BP (eds): The Surgical Management of Epilepsy. Boston: Butterworth-Heinemann, 1994, pp 171-188
20. Engel JJ, Van Ness PC, Rasmussen TB, et al: Outcome with respect to epileptic seizures, in Engel JJ (ed): Surgical Treatment of the Epilepsies. New York: Raven Press, 1993, pp 609-621
21. Falconer MA, Rushworth RG: Treatment of encephalotrigeminal angiomatosis (Sturge-Weber Disease) by hemispherectomy. Arch Dis Child 35:433-447, 1960
22. Falconer MA, Wilson PJE: Complications related to delayed hemorrhage after hemispherectomy. J Neurosurg 30:413-426, 1969
23. Freeman JM, Vining EPG: Hemispherectomy--the ultimate focal resection, in Engel JJ (ed): Surgical Treatment of the Epilepsies. New York: Raven Press, 1993, pp 111-118
24. Griffiths PD, Welch RJ, Gardner-Medwin D, et al: The radiological features of hemimegalencephaly including three cases associated with proteus syndrome. Neuropediatrics 25:140-144, 1994
25. Hoffman HJ, Hendrick EB, Dennis M, et al: Hemispherectomy for Sturge-Weber syndrome. Childs Brain 5:233-248, 1979
26. Horsley V: Brain-surgery. Br Med J 2:670-675, 1886
27. Howells JG, Kay EI: Cerebral hemispherectomy. Lancet 2:683, 1951 (Letter)
28. Krynauw RA: Infantile hemiplegia treated by removing one cerebral hemisphere. J Neurol Neurosurg Psychiatry 13:243-267, 1950
29. Lindsay J, Ounsted C, Richards P: Hemispherectomy for childhood epilepsy: a 36-year study. Dev Med Child Neurol 29:592-600, 1987
30. Matheson JM, Truskett P, Davies MA, et al: Hemispherectomy: a further modification using omentum vascularized free flaps. Aust NZ J Surg 63:646-650, 1993
31. Mathew NT, Abraham J, Chandy J: Late complications of hemispherectomy: report of a case relieved by surgery. J Neurol Neurosurg Psychiatry 33:372-375, 1970
32. Mathis JM, Barr JD, Albright AL, et al: Hemimegalencephaly and intractable epilepsy treated with embolic hemispherectomy. AJNR 16:1076-1079, 1995
33. Moshe SL, Shinnar S: Postscript: early intervention, in Engel J Jr (ed): Surgical Treatment of the Epilepsies, ed 2. New York: Raven Press, 1993, pp 123-132
34. Ogunmekan AO, Hwang PA, Hoffman HJ: Sturge-Weber-Dimitri disease: role of hemispherectomy in prognosis. Can J Neurol Sci 16:78-80, 1989
35. Oppenheimer DR, Griffith HB: Persistant intracranial bleeding as a complication of hemispherectomy. J Neurol Neurosurg Psychiatry 29:229-240, 1966
36. Peacock WJ: Hemispherectomy for the treatment of intractable seizures in childhood. Neurosurg Clin North Am 6:549-563, 1995
37. Peacock WJ, Comair YG, Hovda DA: Hemispherectomy for intractable seizures of childhood, in Apuzzo MLJ (ed): Neurosurgical Aspects of Epilepsy. Park Ridge, Ill: American Association of Neurological Surgeons, 1991, pp 187-198
38. Poirier P, Lassonde M, Villemure JG, et al: Sound localization in hemispherectomized patients. Neuropsychologia 32:541-543, 1994
39. Polkey CE, Binnie CD: Assessment and selection of candidates for surgical treatment of epilepsy. Epilepsia 36 (Suppl 1):S41-S45, 1995
40. Rasmussen T: Cerebral hemispherectomy: indications, methods, and results, in Schmidek HH, Sweet WH (eds): Operative Neurosurgical Techniques: Indications, Methods, and Results, ed 2. New York: WB Saunders, 1988, Vol 2, pp 1235-1241
41. Rasmussen T: Hemispherectomy for seizures revisited. Can J Neurol Sci 10:71-78, 1983
42. Rasmussen T, Villemure JG: Cerebral hemispherectomy for seizures with hemiplegia. Cleve Clin J Med 56 (Suppl Pt 1):S62-S83, 1989
43. Schramm J, Behrens E, Entzian W: Hemispherical deafferentation: an alternative to functional hemispherectomy. Neurosurgery 36:509-516, 1995
44. Spencer DD, Ojemann GA: Overview of therapeutic procedures, in Engel J Jr (ed): Surgical Treatment of the Epilepsies, ed 2. New York: Raven Press, 1993, pp 455-471
45. Taha JM, Crone KR, Berger TS: The role of hemispherectomy in the treatment of holohemispheric hemimegaloencephaly. J Neurosurg 81:37-42, 1994
46. Tinuper P, Andermann F, Villemure JG, et al: Functional hemispherectomy for treatment of epilepsy associated with hemiplegia: rationale, indications, results, and comparison with callosotomy. Ann Neurol 24:27-34, 1988
47. Verity CM, Strauss EH, Moyes PD, et al: Long-term follow-up after cerebral hemispherectomy: neurophysiologic, radiologic, and psychological findings. Neurology 32:629-639, 1982
48. Vickrey BG, Hays RD, Hermann BP, et al: Outcomes with respect to quality of life, in Engel J Jr (ed): Surgical Treatment of the Epilepsies, ed 2. New York: Raven Press, 1993, pp 623-635
49. Vigevano F, Di Rocco C: Effectiveness of hemispherectomy in hemimegalencephaly with intractable seizures. Neuropediatrics 21:222-223, 1990
50. Villemure JG, Adams CBT, Hoffman HJ, et al: Hemispherectomy, in Engel J Jr (ed): Surgical Treatment of the Epilepsies, ed 2. New York: Raven Press, 1993, pp 511-518
51. Villemure JG, Andermann F, Rasmussen TB: Hemispherectomy for the treatment of epilepsy due to chronic encephalitis, in Andermann F (ed): Chronic Encephalitis. Boston: Butterworth-Heinemann, 1991, pp 235-241
52. Villemure JG, Mascott CR: Peri-insular hemispherotomy: surgical principles and anatomy. Neurosurgery 37:975-981, 1995
53. Villemure JG, Rasmussen T: Functional hemispherectomy in children. Neuropediatrics 24:53-55, 1993
54. Vining EPG, Freeman JM, Brandt J, et al: Progressive unilateral encephalopathy of childhood (Rasmussen's syndrome): a reappraisal. Epilepsia 34:639-650, 1993
55. Vining EPG, Freeman JM, Carson BS: Hemispherectomy, in Wyler AR, Hermann BP (eds): The Surgical Management of Epilepsy. Boston: Butterworth-Heinemann, 1994, pp 146-154
56. Vining EPG, Freeman JM, Carson BS, et al: Hemispherectomy in children: the Hopkins experience, 1968-1988, a preliminary report. J Epilepsy 3 (Suppl):169-176, 1990
57. White HH: Cerebral hemispherectomy in the treatment of infantile hemiplegia. Review of the literature and report of two cases. Confin Neurol 21:1-50, 1961
58. Wilson PJE: Cerebral hemispherectomy for infantile hemiplegia. A report of 50 cases. Brain 93:147-180, 1970
59. Winston KR, Welch K, Adler JR, et al: Cerebral hemicorticectomy for epilepsy. J Neurosurg 77:889-895, 1992
Manuscript received October 23, 1995.
Accepted in final form January 15, 1996.
Address for Mr. Lauer: World Health Organization, Geneva, Switzerland.
Address reprint requests to: Benjamin S. Carson, M.D., Johns Hopkins Neurological Surgery, 600 North Wolfe Street, Harvey 811, Baltimore, Maryland 21287-8811.