Outcomes of 33 patients from the wars in Iraq and Afghanistan undergoing bilateral or bicompartmental craniectomy

Transcription

1 J Neurosurg 115: , 2011 Outcomes of 33 patients from the wars in Iraq and Afghanistan undergoing bilateral or bicompartmental craniectomy Clinical article Robert D. Ecker, M.D., LCDR, USN, 1 Lisa P. Mulligan, M.D., CDR, USN, 2 Michael Dirks, M.D., CPT, USA, 3 Randy S. Bell, M.D., LCDR, USN, 2 Meryl A. Severson, M.D., CDR, USN, 2 Robin S. Howard, M.A., 4 and Rocco A. Armonda, M.D., COL, USA 2 1 Maine Medical Partners Neurosurgery & Spine, Scarborough, Maine; 2 Department of Neurosurgery, National Naval Medical Center, Bethesda, Maryland; and Departments of 3 Neurosurgery and 4 Clinical Investigation, Walter Reed Army Medical Center, Washington, DC Object. There are no published long-term data for patients with penetrating head injury treated with bilateral supratentorial craniectomy, or supra- and infratentorial craniectomy. The authors report their experience with 33 patients treated with bilateral or bicompartmental craniectomy from the ongoing conflicts in Iraq and Afghanistan. Methods. An exploratory analysis of Glasgow Outcome Scale (GOS) scores at 6 months in 33 patients was performed. Follow-up lasting a median of more than 2 years was performed in 30 (91%) of these patients. The association of GOS score with categorical variables was explored using the Wilcoxon rank-sum test or Kruskal-Wallis analysis of variance. The Spearman correlation coefficient was used for ordinal/continuous data. To provide a clinically meaningful format to present GOS scores with categorical variables, patients with GOS scores of 1 3 were categorized as having a poor outcome and those with scores of 4 and 5 as having a good outcome. This analysis does not include the patients who died in theater or in Germany who underwent bilateral decompressive craniectomy because those figures have not been released due to security concerns. Results. All patients were men with a median age of 24 years (range years) and a median initial Glasgow Coma Scale (GCS) score of 5 (range 3 14). At 6 months, 9 characteristics were statistically significant: focus of the initial injury, systemic infection, initial GCS score, initial GCS score excluding patients with a GCS score of 3, GCS score on arrival to the US, GCS score on dismissal from the medical center, Injury Severity Score, and patients with cerebrovascular injury. Six factors were significant at long-term follow-up: focus of initial injury, systemic infection, initial GCS score excluding patients with a GCS score of 3, GCS score on arrival to the US, and GCS score on dismissal from the medical center. At long-term follow-up, 7 (23%) of 30 patients had died, 5 (17%) of 30 had a GOS score of 2 or 3, and 18 (60%) of 30 had a GOS score of 4 or 5. Conclusions. In this selected group of patients who underwent bilateral or bicompartmental craniectomy, 60% are independent at long-term follow-up. Patients with bifrontal injury fared best. Systemic infection and cerebrovascular injury corresponded with a worse outcome. (DOI: / JNS101490) Key Words craniectomy penetrating head injury traumatic brain injury Abbreviations used in this paper: GCS = Glasgow Coma Scale; GOS = Glasgow Outcome Scale; ICP = intracranial pressure; ISS = Injury Severity Score; NNMC = National Naval Medical Center; WRAMC = Walter Reed Army Medical Center. Hemicraniectomy, removal of the skull to allow for brain swelling, has been used for the treatment of uncontrolled ICP in cerebral trauma, stroke, and malignant cerebral edema since its initial description for acute subdural hematoma in Three recent European trials (DECIMAL [Decompressive Craniectomy in Malignant Middle Cerebral Artery Infarction], DESTINY [Decompressive Surgery for the Treatment of Malignant Infarction of the Middle Cerebral Artery], and HAMLET [Hemicraniectomy After Middle Cerebral Artery Infarction with Life-threatening Edema Trial]) demonstrated that in patients whose hemicraniectomy was performed within 48 hours for hemispheric stroke, case mortality and poor outcome were reduced 38% and 16%, respectively. 3,4,8,9 There are no comparative randomized trials for severe head trauma and hemicraniectomy, although trials are ongoing in Australia and Europe. Two major concerns among physicians involved in the care of patients who may undergo hemicraniectomy are that hemicraniectomy increases the number of vegetative survivors, and only nondominant hemicraniectomy is appropriate. There was no increase in vegetative survivors in the European stroke trials, and cerebral dominance was not a factor in the decision tree. 9 Since the beginning of Operation Enduring Freedom 124 J Neurosurg / Volume 115 / July 2011

2 Outcomes after bilateral or bicompartmental craniectomy and Operation Iraqi Freedom, all severely head injured patients without significant burns have been cared for at the in NNMC Bethesda, Maryland, and WRAMC in Washington, DC. Between April 1, 2003, and October 2008, 33 patients were treated who had undergone bilateral or bicompartmental (supratentorial and infratentorial) craniectomy for severe traumatic brain injury and/ or malignant cerebral swelling in theater. This represents a sample of the most severely head injured patients seen from ongoing conflicts in Iraq and Afghanistan. This paper reviews their presentation, clinical course, and outcome with 6 months of follow-up for all patients and longterm follow-up in 30 patients. Methods After institutional review board approval, all patient consults for war-related head injury between April 1, 2003, and April 1, 2008, for WRAMC and NNMC were reviewed. Four hundred eight patients were identified, and of those 188 underwent craniectomies. Of this group, 33 patients were identified as having undergone bilateral supratentorial craniectomies or bicompartmental decompressions including both supratentorial and infratentorial craniectomies for control of ICP or removal of a traumatic mass lesion (Table 1). Specifically, patients who underwent bilateral or bicompartmental craniotomy/craniectomy for frontal sinus reconstruction or infection were not included. It should be noted that multiple coalition surgeons from the US, Britain, and Australia, with different training backgrounds, performed the craniectomies. However, since 2003 the culture of care has developed that all patients who are viewed as potentially neurologically salvageable at the initial point of neurosurgical care undergo decompression; patients with bihemispheric or bicompartmental injury receive bilateral or bicompartmental decompression. All clinic notes from initial injury to long-term follow-up were then reviewed in detail for relevant clinical data. Craniectomies were defined as bifrontal if bone was removed to the level of the coronal suture and bihemispheric if bone was removed farther back than the coronal suture. All initial CT scans were reviewed, and initial patterns of injury were broadly categorized as bifrontal, below the lateral ventricles, above the lateral ventricles, or at the level of lateral ventricles (Fig. 1). The GOS scores at 6 months were available in all patients, and long-term GOS scores (median follow-up 2.0 years, range years) were available for 30 (91%) of the 33 patients. The association of GOS scores with categorical variables was explored using the Wilcoxon rank-sum test or Kruskal-Wallis ANOVA. To provide a clinically meaningful format to present GOS scores with categorical variables, patients with GOS scores of 1 3 were categorized as having a poor outcome and those with scores of 4 or 5 as having a good outcome. Odds ratios with 95% CIs were presented for select clinical factors. The correlation of GOS scores with ordinal or continuous variables was examined using the Spearman correlation coefficient, with data presented as the median (range). All data were analyzed using SPSS for Windows (version 17, SPSS, Inc.). J Neurosurg / Volume 115 / July 2011 TABLE 1: Patient demographic and clinical characteristics Value* median age (yrs) 24 (19 46) male sex 33 (100) type of injury blast penetrating gunshot wound type of decompression bifrontal bihemispheric supra- & infratentorial location of injury bifrontal ventricular below the ventricle above the ventricle 29 (88) 4 (12) 19 (58) 8 (24) 6 (18) 16 (48) 7 (21) 7 (21) 3 (9) median initial GCS score 5 (3 14) median length of follow-up (mos) survivors deaths 34 (6 69) 2 (<1 10) * Median values are presented with their ranges in parentheses. All other values represent the number of patients with percentages in parentheses. Results The median patient age was 24 years (range years). All patients were men. The median length of follow-up of survivors was 34 months (range 6 69 months). Eighty-eight percent (29) of the injuries were blast/penetrating, of which only 12% (4) were due to gunshot wounds. The type of decompressions included bifrontal in 19 patients (58%), bihemispheric in 8 (24%), and supra- and infratentorial in 6 (18%). Injury patterns were categorized as bifrontal in 16 patients (48%), at the level of lateral ventricles in 7 (21%), below the lateral ventricles in 7 (21%), and above the lateral ventricles in 3 (9%). The median initial GCS score was 5 (range 3 14). All initial GCS scores initially higher than 8 declined to less than 8 prior to decompression. At 6-month follow-up, the following 9 clinical characteristics were statistically related to long-term outcome: focus of the initial injury, systemic infection, initial GCS score, initial GCS score excluding patients with a GCS score of 3, GCS score on arrival to NNMC or WRAMC, GCS score on dismissal from the medical center, ISS, and cerebrovascular injury as defined by presence of major intracerebral arterial occlusion, pseudoaneurysm, or traumatic cerebral vasospasm. At long-term follow-up, the following 6 factors were statistically related to outcome: focus of initial injury, systemic infection, initial GCS score excluding patients with a GCS score of 3, GCS score on arrival to NNMC or WRAMC, and GCS score on dismissal from the medical center (Tables 2 and 3). Patients at 6-month and long-term follow-up had worse outcomes if they had any systemic infection defined by positive findings in a blood, urine, or sputum culture with fever. After 1 year, a patient with an initial systemic in- 125

3 R. D. Ecker et al. Fig. 1. Computed tomography scans showing injury focus above the ventricles (A), at the ventricles (B), below the ventricles (C), and bifrontal (D). fection was 4.9 times more likely to have a poor outcome (Table 4). Injury Severity Score, although not significant at long-term follow-up, was significant at 6 months (p = 0.028). Patients with bifrontal injuries (19 [58%] of 33) were more likely to have a better outcome at 6 months and at long-term follow-up than those whose injuries were not bifrontal (p = and p = , respectively). The calculated odds ratios found that after 1 year, patients with bifrontal injury were 13 times more likely to have a good outcome (Table 4). Although individually cerebral vasospasm, major cerebral blood vessel occlusion, and pseudoaneurysm did not reach the level of significance, when combined in patients with any cerebrovascular injury fared worse at 6 months and trended toward a poor outcome in the long term. The GCS score was examined at 3 different time points: at the time of injury, after arrival to a military treatment facility in the US, and at initial acute hospital dismissal. Additionally, to reduce the possible effect of patients who are sedated, intubated, and pharmacologically paralyzed, the statistical analysis was performed with the patients with GCS scores of 3 included and excluded at the first 2 time points. At 6 months, the initial GCS score was statistically predictive of outcome when patients with a GCS score of 3 were included; however, the initial GCS score was not predictive of outcome in the long term. However, with the exclusion of the patients with a GCS score of 3, the initial GCS score became predictive at both end points. The GCS on arrival to the US was predictive of outcome when patients with a GCS score of 3 were included, and became insignificant when they were excluded, likely correlating with our practice of waiting for sedation and pain medication to wear off prior to initial neurological evaluation. At dismissal from the acute hospital setting, GCS score was strongly predictive of overall outcome (p < at 6 months and at long-term follow-up). Overall, after 1 year, 18 (60%) of 30 patients achieved a GOS score of 4 or 5, 3 (10%) a GOS score of 3, 2 (7%) a GOS score of 2, and 7 (23%) a GOS score of 1. Discussion Decompressive craniectomy has been used intermittently since the early 1900s for the treatment of elevated ICP due to head trauma and a number of other causes. Its popularity has waxed and waned and while the mortality rate had been thought to be reduced, there were concerns that the numbers of severely disabled survivors made the operation unwarranted. There has been a strong resurgence of interest in recent times and there are at least 2 ongoing major randomized studies to evaluate the benefit of decompressive craniectomy: the RESCUEicp (Randomized Evaluation of Surgery with Craniectomy for Uncontrollable Elevation of ICP) trial in the United Kingdom and the DECRA (Early Decompressive Craniectomy in Patients With Severe Traumatic Brain Injury) study in Australia. During the recent conflicts in Iraq and Afghanistan, decompressive craniectomy has been used extensively. Over a period of approximately 5 years (between 2003 and 2008), 188 patients with decompressive craniectomies were treated. Of that group, 33 patients have undergone bifrontal, bihemispheric, or supra- and infratentorial decompressions. The use of bifrontal decompressions is not novel. In 1971, Kjellberg and Prieto 5 published their 126 J Neurosurg / Volume 115 / July 2011

4 Outcomes after bilateral or bicompartmental craniectomy TABLE 2: Association of clinical characteristics and GOS score at 6 months and 1 5 years GOS Score at 6 Mos GOS Score at 1 5 Yrs No. of Patients (%) No. of Patients (%) (Score 4 or 5) p Value* (Score 4 or 5) p Value* no. of patients timing delayed 0 (0) 3 (18) 0 (0) 2 (11) early 16 (100) 14 (82) 12 (100) 16 (89) focus of initial injury bifrontal 4 (25) 12 (71) 2 (17) 13 (72) all other locations 12 (75) 5 (29) 10 (83) 5 (28) type of decompression bifrontal 8 (50) 11 (65) 5 (42) 12 (67) bihemispheric 5 (31) 3 (18) 5 (42) 3 (17) supra- & infratentorial 3 (19) 3 (18) 2 (17) 3 (17) category gunshot wound 1 (6) 3 (18) 0 (0) 1 (6) penetrating head injury 15 (94) 14 (82) 12 (100) 17 (94) mechanism of injury blast 8 (50) 9 (53) 6 (50) 11 (61) projectile 8 (50) 8 (47) 6 (50) 7 (39) presence of CSF leak 1/14 (7) 5 (29) /10 (10) 4 (22) pulmonary embolism 1/15 (7) 3 (18) /11 (0) 2 (11) 0.36 seizure 6 (38) 5 (29) (50) 5 (28) 0.45 CNS infection 6/15 (40) 6 (35) /11 (36) 7 (39) 0.66 systemic infection 12/14 (86) 8/16 (50) /10 (90) 11/17 (65) ophthalmic injury 5/14 (36) 5 (29) /10 (30) 6 (33) 0.99 pseudoaneurysm 3/15 (20) 4 (24) /11 (27) 4 (22) 0.57 vasospasm 4/15 (27) 6/16 (38) /11 (45) 5/17 (29) 0.63 active cooling 8/10 (80) 10/10 (100) /8 (75) 9/9 (100) 0.29 shunt 5/12 (42) 10/13 (77) /7 (43) 11/15 (73) 0.42 major vascular occlusion 3/14 (21) 0 (0) /11 (18) 0/17 (0) 0.19 any vascular injury 10 (62) 6 (35) (83) 5 (28) * Probability values were obtained using the Wilcoxon rank-sum test or Kruskal-Wallis ANOVA to compare clinical factors with the numerical GOS outcomes at 6 months and 1 5 years. experience with 73 patients who had undergone bifrontal craniectomies for trauma, subarachnoid hemorrhage, gunshot wounds, and stroke. Fifty of those patients had suffered closed head injuries. Their data showed a very high mortality rate (82%), but of those who survived, the majority had a favorable outcome. Five of the 13 survivors returned to full function, and an additional 4, while suffering permanent neurological injury, were capable of self-care. In 1975, Venes and Collins 10 published their experience with 13 patients who also underwent bifrontal craniectomy. In their group, the mortality rate was much lower than that in the study by Kjellberg and Prieto, 5 but the percentage of severely disabled and vegetative patients was much higher. They reported a 31% mortality rate, but J Neurosurg / Volume 115 / July (38%) of 13 were severely disabled or vegetative, and none of the severely head injured group returned to their premorbid level of functioning. Little was done with bilateral craniectomy again until the publication from the group at the University of Virginia Health System where the authors performed bifrontal decompressions in 35 patients. 6 In this study, the surgeries were generally performed earlier, and outcomes were somewhat better. The authors data showed a mortality rate of 23%, and among survivors, 48% had either a good outcome or just moderate disability. However, when all patients are included, only 37% had a favorable outcome. The authors reported that there was a trend toward better outcome when surgery was performed less than 48 hours after injury. Another trend seen in this study and that of Kjellberg and 127

5 R. D. Ecker et al. TABLE 3: Median demographic and clinical variables by GOS score at 6 months and 1 5 years GOS Score at 6 Mos Median (range) (Score 4 or 5) r p Value GOS Score at 1 5 Yrs Median (range) (Score 4 or 5) r p Value age (yrs) 24 (19 46) 22 (19 36) (19 46) 24 (19 38) evacuation days 3 (2 9) 4 (2 8) (2 9) 4 (2 8) GCS (initial) 4 (3 7) 7 (3 14) (3 7) 7 (3 14) GCS (initial)* 5.5 (4 7) 8.5 (5 14) (4 7) 8 (5 14) GCS (on arrival) 4 (3 11) 8 (3 14) (3 11) 8 (3 14) GCS (on arrival)* 6 (4 11) 8 (5 14) (4 11) 8 (4 14) GCS (at discharge) 10 (3 15) 15 (11 15) 0.82 < (3 14) 15 (11 15) 0.73 < ISS 35 (25 75) 29 (21 50) (21 75) 33 (21 50) ICU days (primary injury) 10 (3 68) 19 (4 240) (3 68) 19 (3 240) * Patients with a GCS score of 3 were excluded from these analyses. Prieto 5 was that younger patients, especially children, seemed to do better and likewise older patients did not fare well. In Kjellberg and Prieto s paper, no one older than 48 years survived. Despite the severity of their injuries and the particularly potent mechanism of injury, the outcome in our patient population compares favorably. Sixty percent had a GOS score of 4 or 5, 10% a GOS score of 3, and 30% a GOS score of 1 or 2 at long-term follow-up. Our patient population possesses favorable characteristics when compared with the civilian world in that they are a young group with a median age of 24 years and are in extremely good physical condition prior to injury. One hundred percent of the patients had penetrating head injuries, and nearly 90% were subjected to a blast as their mechanism of injury. All but 3 patients underwent early decompression in the theater of war prior to evacuation. This was in part driven by the need for extensive medical evacuation from the theater of operations to Germany and then on to the US with each leg representing 8 10 hours or longer of travel. During that time, no surgical interventions are feasible, so any intervention must be performed prior to transport. Systemic infection and not CNS infection was predictive of a poor outcome. Systemic infection was far more prevalent in patients with poor outcomes than CNS infection (90% vs 36%), and CNS infections were more evenly distributed between the good and poor outcome groups with rates less than 40% in all groups. The GCS score proved predictive of outcome in the short and long term with the exclusion of patients with a GCS score of 3 at the initial evaluation, making the finding statistically more robust despite the decrease in the number of patients evaluated. As expected, patients with a bifrontal injury pattern fared better than those with injury at the level of the ventricle or below. Additionally, the bifrontal injury pattern was the most common (50%) in patients who survived transfer to the US. Previously published data from our group have emphasized the importance of cerebrovascular injury as a cause of morbidity in blast and penetrating head injury. 1,2 This study further supports the conclusion that cerebrovascular injury is a cause of morbidity and mortality in our patients. It is impossible to fully account for the preexisting physical conditioning, young age, and proximity to first responder medical care in comparison with previously published series and, generically, civilian trauma. At a minimum, all deployed soldiers will have passed a biannual physical fitness test that places them at a higher level of fitness than their civilian counterparts. Furthermore, they must have been medically screened to deploy to Operation Enduring Freedom or Operation Iraqi Freedom; such screening excludes many orthopedic diagnoses and asthma, for example, for which a soldier may stay on active duty but cannot deploy to the theater of war. Additionally, it is unlikely that an injured soldier would not be near a first responder who could quickly establish an airway and obtain venous access for fluid resuscitation. With prolonged extrications from wrecked vehicles and the need to call for a first responder, there is often delay in the basics of civilian trauma resuscitation. Civilian severe brain injury has 2 peaks at ages years and older than 60 years, with the latter faring far worse and potentially biasing treatment data. The data from the military do not include the second peak as military members of that age are generally not in the fight. TABLE 4: Odds ratios and 95% CIs for clinical characteristics associated with poor outcome (GOS Score 1 3) at 6 Mos OR (95% CI) at 1 5 Yrs focus of initial injury (not bifrontal) 7.2 ( ) 13.0 ( ) systemic infection 6.0 ( ) 4.9 ( ) any vascular injury 3.1 ( ) 13.0 ( ) 128 J Neurosurg / Volume 115 / July 2011

6 Outcomes after bilateral or bicompartmental craniectomy This analysis does not include the patients who died in theater or in Germany who underwent bilateral decompressive craniectomy because those figures have not been released due to security concerns. The number of patients treated may well be higher. Furthermore, although this is the largest series of patients of its kind, the granularity and power of the analysis that can be achieved in a series of 33 patients is limited. Areas of analysis that have not been found to be significant, or even considered, may well impact overall outcome and survival. All surviving patients in this series have had their skulls reconstructed using a variety of techniques. Conclusions Bilateral and/or bicompartmental decompressive craniectomy can be an effective treatment strategy for uncontrolled ICP in select patients with bilateral or supratentorial and infratentorial severe traumatic brain injury with swelling. In this necessarily selected sample of soldiers with severe penetrating brain injury from the conflicts in Iraq and Afghanistan, 60% achieved a good clinical outcome, 10% remain dependant but interactive, 7% are vegetative, and 23% are dead at long-term followup. This does not represent an increase in the number of vegetative survivors compared with historical series. Systemic infection portended a poor outcome. Initial GCS score, GCS score on transfer, and GCS score on dismissal from acute hospitalization were predictive of outcome. Patients with bifrontal injuries fared best, and high ISS and presence of cerebrovascular injury were predictive of poor outcome at 6 months. Disclosure The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper. The views expressed in this talk are those of the authors and do not reflect the official policy or position of the Department of the Navy, Department of the Army, the Department of Defense, or the United States Government. The authors acknowledge that research protocol NNMC , A Retrospective Review of Craniospinal Vascular Injury from Operation Iraqi and Enduring Freedom received applicable NNMC Institutional Review Board review and approval. Author contributions to the study include the following. Conception and design: Ecker, Dirks. Acquisition of data: Ecker, Dirks. Analysis and interpretation of data: Ecker, Mulligan, Dirks, Severson, Howard, Armonda. Drafting the article: Ecker. Critically revising the article: Ecker, Mulligan, Bell, Severson, Armonda. Reviewed final version of the manuscript and approved it for submission: all authors. Statistical analysis: Howard. Administrative/ technical/material support: Mulligan, Severson. Study supervision: Ecker. Acknowledgments The authors wish to thank all the service men and women that have served in Operation Iraqi Freedom and Operation Enduring Freedom. Their dedication and commitment to service inspire us all. References 1. Armonda RA, Bell RS, Vo AH, Ling G, DeGraba TJ, Crandall B, et al: Wartime traumatic cerebral vasospasm: recent review of combat casualties. Neurosurgery 59: , Bell RS, Vo AH, Roberts R, Wanebo J, Armonda RA: Wartime traumatic aneurysms: acute presentation, diagnosis, and multimodal treatment of 64 craniocervical arterial injuries. Neurosurgery 66:66 79, Hofmeijer J, Kappelle LJ, Algra A, Amelink GJ, van Gijn J, van der Worp HB: Surgical decompression for space-occupying cerebral infarction (the Hemicraniectomy After Middle Cerebral Artery infarction with Life-threatening Edema Trial [HAMLET]): a multicentre, open, randomised trial. Lancet Neurol 8: , Jüttler E, Schwab S, Schmiedek P, Unterberg A, Hennerici M, Woitzik J, et al: Decompressive Surgery for the Treatment of Malignant Infarction of the Middle Cerebral Artery (DES- TINY): a randomized, controlled trial. Stroke 38: , Kjellberg RN, Prieto A Jr: Bifrontal decompressive craniotomy for massive cerebral edema. J Neurosurg 34: , Polin RS, Shaffrey ME, Bogaev CA, Tisdale N, Germanson T, Bocchicchio B, et al: Decompressive bifrontal craniectomy in the treatment of severe refractory posttraumatic cerebral edema. Neurosurgery 41:84 94, Ransohoff J, Benjamin V: Hemicraniectomy in the treatment of acute subdural haematoma. J Neurol Neurosurg Psychiatry 34:106, Vahedi K: Decompressive hemicraniectomy for malignant hemispheric infarction. Curr Treat Options Neurol 11: , Vahedi K, Hofmeijer J, Juettler E, Vicaut E, George B, Algra A, et al: Early decompressive surgery in malignant infarction of the middle cerebral artery: a pooled analysis of three randomised controlled trials. Lancet Neurol 6: , Venes JL, Collins WF: Bifrontal decompressive craniectomy in the management of head trauma. J Neurosurg 42: , 1975 Manuscript submitted August 29, Accepted February 23, The contents of this manuscript were presented during the Neurotrauma and Critical Care scientific session at the AANS 78th An nual Meeting, May 3, 2010, Philadelphia, Pennsylvania. Please include this information when citing this paper: published online March 25, 2011; DOI: / JNS Address correspondence to: Robert D. Ecker, M.D., Maine Medical Partners Neurosurgery & Spine, 49 Spring Street, Scarborough, Maine robertecker@me.com. J Neurosurg / Volume 115 / July