Keyhole and Key-Bur-Hole

Should Cervical Fusions Ever Be Done forCervicalgia Alone?

To the Editor:Now that I am semiretired and no longer performing operative

neurosurgery, I have been disappointed to find that some of myyounger colleagues decline to perform cervical fusions for patientswhom I refer to them if those patients have only neck pain and donot have severe radicular symptoms and/or neurological deficit. Iam told that this is because of an impression of a universally pooroutcome, although perhaps it is also in part a reaction to some ofthe excessive surgical zeal for doing cervical fusions that we allencounter in the community. My own recollection, based onmore than 40 years of performing cervical spine surgery, has notbeen so pessimistic, and I believe that it is a disservice to mypatients for them not to be able to undergo potentially quitebeneficial fusion surgery. Surgery for painful joints elsewhere inthe body is certainly widely practiced and accepted, with largenumbers of artificial hips, knees, and so on being used to treatpainful joints as our population ages. Because the cervical disc isanother joint that can be locally damaged, producing arthralgicpain, I fail to see the logic of not operating on a painful cervicaldisc when indicated.

To reinforce my memory, I did a quick chart review of my ownlast 10 patients operated on for arthralgic cervicalgia during theyears 1995 to 2002. Like most spinal neurosurgeons, the majorityof the cervical operations that I performed were for disc ruptureswith radiculitis, cervical fractures, spondylotic myelopathy, tu-mors, and other conditions. However, I did retrieve the recordsfor 10 patients on whom I had operated for neck pain as theirprincipal symptom. All these patients had severe 1- or 2-levelcervical disc degeneration, and none had associated neurologicaldeficits. All were incapacitated by intolerable pain despiteaggressive and often prolonged therapy, including anti-in-flammatory drugs, bracing, exercises, and often intradiscal steroidinjections. There were 5 men and 5 women, 8 with single-leveldisease and 2 with adjacent 2-level disease. Their ages wereyounger than those of typical patients with degenerative spon-dylopathy (mean age 45.5 years, range 36-64 years), most havinga history of remote neck injury of limited severity. All underwentanterior cervical fusion, 4 with a Cloward technique and 6 withinstrumentation. Follow-up varied from 2 to 64 months (mean 8months). Four patients reported excellent results (fully functionalwithout medication), 3 good results (functional but requiringmedication), and 2 fair results (decreased pain intensity butfunctionally limited), and only 1 reported a poor result (no painrelief).

Undoubtedly the results that can be achieved with cervicalfusion for cervicalgia will be determined chiefly by patient se-lection because most neurosurgeons and orthopedists are likely tobe technically adept at such surgery. My patients were selected

after I got to know them, all had severe and focal degenerative discdisease on imaging, and all had intolerable neck pain despitevigorous and varied alternative therapy. It should not be overlydifficult for neurosurgeons to select patients likely to benefit fromfusion surgery for cervicalgia, just as they select appropriate pa-tients for surgery for other disorders. It seems to me that ‘‘themyth of poor outcome from fusion surgery for cervicalgia’’ islikely to become self-fulfilling if proper patient selection is notcarried out, but these patients do exist, many suffer greatly, andmany of them can benefit greatly from surgery to become gratefulpatients.

Harold A. WilkinsonBoston, Massachusetts

10.1227/NEU.0b013e31820208bf

Anaplastic Foci Within Gliomas

To the Editor:The comment provided by Dr Engh1 about intraoperative vi-

sualization of anaplastic foci in gliomas points out the elegant usageof 5-aminolevulinic acid for this purpose but misses the alreadyavailable, published techniques to achieve this goal in nonenhancinggliomas. Publications back to the 1990s demonstrate radiologicaland nuclear medicine techniques (xenon-enhanced computer to-mography, positron emission tomography, especially amino acidpositron emission tomography with methionine or fluoro-ethyl-thyrosine)2-4 to investigate nonenhancing gliomas for anaplastic foci.This preoperative radiological information is used intraoperatively byimplementation into the neuronavigation to target these pre-operatively identified anaplastic areas4 (Figure). This technique isused in many centers routinely with good success and should not bewithheld from the readers of Neurosurgery as a much cheapertechnique, especially in the United States, where 5-aminolevulinicacid apparently is not available yet. This is of great interest to theneurosurgical community because all suspected low-grade gliomaswithout contrast enhancement should be considered to have ana-plastic areas as long as positron emission tomography or histologycan exclude it.

Karl RoesslerIris ZachenhoferFeldkirch, Austria

1. Engh JA. Improving intraoperative visualization of anaplastic foci within gliomas.Neurosurgery. 2010;67(2):N21–N22.

2. Roessler K, Gatterbauer B, Becherer A, et al. Surgical target selection in cerebralglioma surgery: linking methionine (MET) PET image fusion and neuronavigation.Minim Invasive Neurosurg. 2007;50(5):273-280.

3. Roessler K, Czech T, Dietrich W, et al. Frameless stereotactic-directed tissuesampling during surgery of suspected low-grade gliomas to avoid histological un-dergrading. Minim Invasive Neurosurg. 1998;41(4):183-186.

E592 | VOLUME 68 | NUMBER 2 | FEBRUARY 2011 www.neurosurgery-online.com

CORRESPONDENCE

Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited.


4. Roessler K, Nasel C, Czech T, Matula C, Lassmann H, Koos WT. Histologicalheterogeneity of neuroradiologically suspected adult low grade gliomas detected byxenon enhanced computerized tomography (CT). Acta Neurochir (Wien).1996;138(11):1341-1347.

10.1227/NEU.0b013e3182041797

Anaplastic Foci Within Gliomas

I greatly appreciate the thoughtful comments from DrsRoessler and Zachenhofer. Their group has been a leader in theevaluation of methionine positron emission tomography andxenon-enhanced computed tomography in the delineation ofanaplastic foci within gliomas. Their technique is much morewidely available than 5-aminolevulinic acid and has been used atmultiple centers with success.

The novelty of the technique proposed by Widhalm et al1 isnot the integration of metabolic imaging into neurosurgical image

guidance, which has been well demonstrated by Dr Roessler.Rather, the real-time feedback that is provided by intraoperativefluorescence of high-grade areas is further confirmation to theoperating surgeon that anaplastic foci are being sent for histo-pathological evaluation. Whether this information translates intomeaningful clinical outcomes for patients with these tumors re-mains to be seen.

Johnathan A. EnghPittsburgh, Pennsylvania

1. Widhalm G, Wolfsberger S, Minchev G, et al. 5-Aminolevulinic acid is a promisingmarker of detection of anaplastic foci in diffusely infiltrating gliomas with non-significant contrast enhancement. Cancer. 2010;116(6):1545-1552.

10.1227/NEU.0b013e31820417bd

FIGURE. Suspected low grade glioma, right hemisphere. A and B, Computed tomography and magnetic resonance imagingshowing diffuse infiltrative tumor frontal and temporo-parietal, without any contrast enhancement. C and D, Image fusion withmethionine positron emission tomography demonstrating an anaplastic focus in the deep frontal periventricular area. E,A stereotactic needle biopsie of this focus revealed highly anaplastic tumor histology with MIB-1 proliferation index of 33%,corresponding to a glioblastoma (modified from Roessler et al3).

NEUROSURGERY VOLUME 68 | NUMBER 2 | FEBRUARY 2011 | E593

CORRESPONDENCE



Keyhole and Key-Bur-Hole

To the Editor:We read with great interest the anatomical study done on

MacCarty’s keyhole.1 This bur-hole is typically performedwhen an orbito-zygomatic component is added to a pure fronto-temporal or supra-orbital craniotomy as clearly illustrated in thearticle. However any burr hole in this area is often referred to as a‘‘key hole,’’ also quoted as ‘‘key-bur-hole’’.2 Hence we highlightthe technical difference in performing a bur-hole for a puresupraorbital or fronto-temporal/Pterional craniotomy.

History of Keyhole

Historically, the word ‘‘keyhole’’ was coined during the era ofthe Gigli saw when it is used to open the cranium for a fronto-temporal craniotomy. It is a keyhole because the guide for theGigli saw has to be introduced both superiorly and inferiorly fromthis hole (the latter can alternatively be rongeured).

Pure Fronto-Temporal/Supraorbital Craniotomy

However, for a pure fronto-temporal or supra-orbital crani-otomy the orbital contents should not be exposed. Perneczkyet al3 use the term ‘‘frontobasal burr hole’’ for a supraorbitalcraniotomy. For a supra-orbital craniotomy performed usinga high-speed drill, a single frontobasal burr hole is sufficient and,for cosmetic reasons, should be placed posterior to the temporal

line. Special attention must be given to the placement of this burhole, particularly with regard to its relationship to the frontalcranial base and the orbit. Even after correct placement, incorrectdirection of the drilling procedure can result in penetration ofthe orbit and not the anterior fossa. While performing a purefronto-temporal craniotomy, Michael Salcman et al4 statesthat the ‘‘crucial’’ bur hole lies on the ‘‘external orbital process’’antero-superior to the pterion. The angle of drilling is such thatthe tip of the drill points a bit more postero-superiorly so as not toenter the orbital contents (Figure).

MacCarty’s Keyhole

On the other hand, the Maccarty’s keyhole is a specific entitywhich is drilled in such an angle that the tip of the drill pointsmore antero-inferiorly. And the orbital roof divides the burr-holeinto a superior (anterior cranial fossa) and an inferior (orbitalcontent) compartment. This enables removal of the orbital rimand the zygomatic component, as rightly pointed out in thearticle.1

In the era of motorized craniotomes, the term keyhole is rarelyused for the purpose it was originally coined.

Chandramouli BalasubramanianLondon, United Kingdom

FIGURE. Key-burr-hole showing a more posteriorly angled drilling (B) and a more anteriorly angled drilling (A) for MacCarty’skeyhole.


CORRESPONDENCE



Balamurugan MangaleswararHariprakash Chakravarthy

Reginauld JohnChennai, India

1. Tubbs RS, Loukas M, Shoja MM, Cohen-Gadol AA. Refined and simplifiedsurgical landmarks for the MacCarty keyhole and orbitozygomatic craniotomy.Neurosurgery. 2010;66(6):ons230–ons233.

2. Menovsky T, De Vries J, Wurzer JA, Grotenhuis JA. Intraoperative ventricularpuncture during supraorbital craniotomy via an eyebrow incision: technical note.J Neurosurg. 2006;105(3):485-486.

3. Resich R, Perneczky A. Ten-year experience with the supraorbital subfrontal ap-proach through an eyebrow skin incision. Neurosurgery. 2005;57(4 Suppl):242-255.

4. Salcman M, Kempe LG, Heros, RC. Kempe’s Operative Neurosurgery. Vol. 1 2nd ed.New York: Springer-Verlag; 2004.

10.1227/NEU.0b013e31820417aa

Keyhole and Key-Burr-Hole

We thank Dr Balasubramanian and colleagues for theirinterest in our article and for drawing the readers’ attention tothe historical derivation of the term keyhole. As is the case withmany terms, time dulls and often changes their originalmeanings. Another good example of a structure that has beenmodified with time and an example that is even more remote inhistory is the term ‘‘torcular Herophili.’’ Originally, the term‘‘torcular’’ referred to the depression in the occipital bone thathouses the confluens of sinuses.1,2 However, over time, thisword began to be applied to the confluens of sinuses andcurrently is used without fail to describe the venous sinusand not the underlying bone. Therefore, although perhaps notthe original use of the term keyhole, current terminology nowembraces its association with the maneuver described byMacCarty.3

R. Shane TubbsBirmingham, Alabama

Aaron Afshin Cohen-GadolIndianapolis, Indiana

1. Tubbs RS, Salter G, Oakes WJ. Superficial surgical landmarks for the transversesinus and torcular herophili. J Neurosurg. 2000;93(2):279-281.

2. Tubbs RS, Oakes WJ. Letter to the Editor. Neuroanatomy. 2002;1:14.3. Tubbs RS, Loukas MM, Shoja MM, Cohen-Gadol AA. Refined and simplified

surgical landmarks for the MacCarty keyhole and orbitozygomatic craniotomy.Neurosurgery. 2010;66(6):ons230-ons233.

10.1227/NEU.0b013e31820417e3

Real-Time Facial Nerve Monitoring

To the Editor:With great interest I read the article about the strategy of

intraoperative facial nerve monitoring developed by Prell et al.1

In the abstract, the authors state that when periods of A-trains(a specific electromyographic activity of the facial muscles)are automatically detected and their length summed up(ie, ‘‘traintime’’), there was a ‘‘high correlation between traintimeas measured by real-time analysis and functional outcome im-mediately after the operation (Spearman correlation coefficient[r] = 0.664, P , .001) and in long-term outcome (r = 0.631,P , .001).’’ In the text of the article, this statement is similarlyexpressed with the exception that a 2-sided P value of ,.0001 isreported for the correlation between traintime and short-term(10 days postoperatively) clinical outcome of facial nerve functionin the given 30 patients.

Although it is common to assume that the amount of elec-tromyographic activity related to mechanical irritation of thefacial nerve during surgery electromyography is somehow relatedto facial palsy after the operation, caution should be exercisedwith the data presented here for 2 reasons. First, a nerve that hadbeen severely attached to the tumor and therefore has been se-verely irritated by the surgical preparation will take at least 6months to even show the first clinical signs of regeneration in thefacial muscles. More than a year after surgery, however, it can stillrecover to grades 2 or even 1 on the House and Brackman scale.The second consideration is a statistical issue that relates to theclinically important issue of false-positive electromyographic re-sponses in prognosticating facial nerve palsy.

The Spearman rank correlation coefficient was used to analyzethe statistical dependence between ‘‘overall traintime’’ and clin-ical outcome in this study. This coefficient is a measure to assesshow well the relationship between 2 variables can be describedwith a monotone function. A perfect Spearman correlation of +1or 21 occurs when each of the variables is a perfect monotonefunction of the other. However, it does not say anything aboutthe linearity of this function.

In this study with all but 3 patients falling into 3 House andBrackman classes at 6 months (half of them were House andBrackman grade 1), the data are neither normally distributed norlinearly related to the second parameter, the intraoperativelyassessed ‘‘overall traintime.’’ In 15 of the patients, these periods ofelectromyography trains—the sum of the duration of A-trainsmeasured at the nose, at the mouth, and at the eye—amounted tojust 2 seconds.

When we reanalyzed the data from this group of patients, wefound that the Spearman correlation coefficient remains signifi-cant at the 5% level (even though it drops to 0.585). However,when ‘‘overall traintime’’ exceeds 2 seconds, the coefficient is just0.449 and no longer reaches statistical significance.

In addition, the authors themselves report 5 false-positiveresults with respect to their method of intraoperative facial nervemonitoring and facial nerve outcome.

Both these facts support the clinical observation that a numberof patients, despite having sustained intraoperative electromyo-graphic activity, including longer periods of A-trains, exhibit nofacial palsy at all after surgery. This in turn raises questions aboutthe clinical applicability of the, albeit desirable, idea of


CORRESPONDENCE



establishing a ‘‘traffic light’’ warning algorithm presented to thesurgeon based on traintime. Should we trust an automaticallycalculated variable when deciding to break up? What if we stopsurgery because the light went yellow or red and the tumor is noteven halfway out? On the other hand, what happens to surgicalconfidence if we decide to go on despite a red signal?

With the interpretational difficulties of ‘‘overall traintime’’resulting from the statistical issue described above, one wouldhesitate to implement the method in rigid decision-makingalgorithms for surgery involving the facial nerve as it stands. Still,we are confident that methods of online automated data analysiswill complement existing routine strategies for intraoperativemonitoring in the future. Therefore, we are grateful to Dr Prelland colleagues, who have repeatedly raised this issue and continueto work toward clinically applicable solutions.

Steffen Klaus RosahlWuppertahl, Germany

1. Prell J, Rachinger J, Scheller C, Alfieri A, Strauss C, Rampp S. A real-timemonitoring system for the facial nerve. Neurosurgery. 2010:66(6):1064-1073.

10.1227/NEU.0b013e318204197b

Real-Time Facial Nerve Monitoring

We appreciate the interest in our article1 and would like tothank Dr Rosahl for his insightful remarks and the effort he tookin calculating additional statistics. However, we do not fully agreewith his conclusions.

Dr Rosahl states that a compromised facial nerve will take atleast 6 months to show first signs of clinical recovery. However, inour experience, first signs will be seen after 3 to 4 months in high-grade paresis (In low-grade paresis, it may even be within weeks);further improvement will be quick over the next months. Thelion’s share of functional recovery will usually be reached after6 months, so we are sure that this is an adequate point of timeto state a definite clinical result for our patients.

Dr Rosahl is skeptical regarding the Spearman correlationcoefficients that we calculated because he has observed that ourdata were not normally distributed and did not show a linearfunction. This observation is correct; in our article, we havedescribed the correlation as resembling a logarithmic function.However, neither a linear function nor normally distributed dataare needed to calculate the Spearman coefficient because of itsnature as a rank correlation.

Dr Rosahl has recalculated the Spearman rank correlation forsubgroups of our patients. He calculated traintime correlation tofunctional outcome for a subgroup of 15 patients and the re-maining patients; he states that he used a dividing line of 2seconds of traintime for this subdivision. However, there were 14patients with , 2 seconds of traintime, not 15 patients. Althoughit would be interesting to know which patients were actuallytaken into consideration for this calculation, it is quite clear either

way that these were patients with small amounts of traintime andgood clinical results. Naturally, selective exclusion of these pa-tients from the statistical calculation will not improve correla-tions. Apart from this, direct comparison of 2 correlationcoefficients needs a certain number of values to hold statisticalsignificance. The smaller the difference between the 2 coefficientsis, the higher the needed number of values will be. In this light,comparing a coefficient of 0.664 against 0.585 or 0.449 withsubdivided groups of 15 patients is questionable.

Dr Rosahl asks if we should ‘‘trust an automatically calculablevariable when deciding to break up.’’ We should not. Releasefrom responsibility for reasonable decisions was not our in-tention. As we have discussed in a very detailed way, the operatingsurgeon still needs to integrate several crucial information sourcesto decide about breaking up. Traintime is an additional source ofinformation that might be very valuable in cerebellopontinesurgery, but it is not supposed to make us ignore other sources ofinformation.

Julian PrellHalle, Germany

1. Prell J, Rachinger J, Scheller C, Alfieri A, Strauss C, Rampp S. A real-timemonitoring system for the facial nerve. Neurosurgery. 2010:66(6):1064-1073.

10.1227/NEU.0b013e31820417d1

History and Evidence RegardingHydrostatic Shock

Without citing data in support for the claim, the otherwiseexcellent review paper, Ballistics for the Neurosurgeon,1 asserts that‘‘hydrostatic shock’’ is a ‘‘relatively recent myth.’’ However, theremote effects of ballistic pressure waves known as ‘‘hydrostaticshock’’ or ‘‘hydraulic shock’’ have considerable support and a longhistory. Reference to ‘‘hydraulic shock’’ can be found as earlyas 1898 in an article describing experiments in which fish werekilled by a remote pressure mechanism similar to underwaterdynamite explosions by firing a rifle into the water within 24inches or so of the fish.2 Upon inspection, no easily discernablewound was discoverable on the body, and death was attributedto the remote effects of the pressure wave caused by the bulletimpacting the water.

In the 1940s, Harvey and coworkers3,4 investigated inter-actions between ballistic waves and tissue in a series of experi-ments conducted at Princeton University. The fast pressuretransients caused by projectiles hitting fluids and fluid-filledtissues were detected with piezoelectric pressure transducers andspark shadowgraph photography. Without conducting sensitivehistological tests or functional tests on living test subjects, thiswork concluded that the only easily observable injury caused bythe observed pressure transients are associated with gas pockets inthe body. (The same thinking has long been associated with blastinjury.) However, hunters have long attributed instant


CORRESPONDENCE



incapacitation of game animals with the remote effects of a rapidpressure transient caused by bullet impact.5,6

In 1954, Ochsner reported results of experiments in goatscomparing high-speed projectile impacts to the thigh with a smallhigh-explosive charge taped to the same location.7 Ochsner foundit notable that blood transfusions alone could not save theseanimals and reported an average survival time of 12 hours. Heattributed these observations to the pressure changes that hadbeen documented by Harvey.

It was well-documented that tissue, including neural tissue,could be torn or damaged by temporary cavitation.8 Docu-mentation of remote damage beyond the reach of the temporarycavity came later, in research reported by groups in Sweden andChina. In a series of articles, a Swedish research team reportedremote injury to the peripheral nerves, the spinal cord, and thebrain with use of electron microscopy of pigs shot in the thigh.9-12

These studies used high-speed pressure sensors implanted in thethigh, abdomen, neck, and brain of test animals. The pressuretransient was shown to propagate from the impact site to thebrain at close to the speed of sound. A Chinese study published in1990 also reported in vivo measurements of fast pressure tran-sients and related remote pressure wave injuries in experimentsincluding pigs and dogs.13 A later Chinese experiment usedsensitive biochemical techniques to detect remote brain injury indogs shot in the thigh.14

Analysis of data relating to rifle wounds from the VietnamWar described a number of cases of distant wounding, includingbroken bones, abdominal wounding in cases where the bulletdid not penetrate the abdominal cavity, a lung contusion re-sulting from a bullet hit to the shoulder, and distant effects onthe central nervous system.15 A case study of a World War IIsoldier sustaining a handgun wound attributed the much lateronset of epilepsy to a hydrodynamic effect.16 Another case studyof a gunshot victim attributed a spinal cord injury to the fo-cusing of shock waves remote from the bullet path.17 A 2007article reviewed this and other evidence and predicted that re-mote brain injury would be common for handgun woundscentered in the chest.18 Research published in 2009 reportedhuman autopsy results that showed ‘‘cufflike pattern haemor-rhages around small brain vessels were found in all speci-mens.’’19 This remote brain injury was attributed to the pressuretransient caused by the bullet hitting the chest. Easily visiblebrain hemorrhaging has also been correlated with instant in-capacitation in wild animals shot in the chest with much morepowerful firearms.20 Noting similarities between blast andballistic waves, a recent paper estimated mild traumatic braininjury (TBI) thresholds for the thoracic mechanism of blast-induced TBI by analyzing data from ballistic pressure wave andbehind armor blunt trauma studies.21

A myth is an assertion that has either been disproved bycareful experiment or for which there is no historical or scientificevidence in cases where it is reasonably expected. Belief in theremote effects of penetrating projectiles may have originatedwith hunters and soldiers, but their reality is now well

established in a broad body of scientific literature, even thoughthe clinical significance for the practicing neurosurgeon mightbe debatable. Perhaps the clinical significance will becomegreater with anticipated advancements in detection and treat-ment of mild TBI.

Michael CourtneyAmy Courtney

Colorado Springs, Colorado

1. Jandial R, Reichwage B, Levy M, Duenas V, Sturdivan L. Ballistics for theneurosurgeon. Neurosurgery. 2008;62(2):472-480.

2. Science and Industry. The New York Times. November 27, 1898:14. http://query.nytimes.com/mem/archive-free/pdf?_r=1&res=9F01EED61139E433A25754C2A9679D94699ED7CF. Accessed June 8, 2010.

3. Harvey EN. The mechanism of wounding by high velocity missiles. Proc Am PhilosSoc. 1948:92(4):294-304. http://www.jstor.org/stable/3143359. Accessed June26, 2008.

4. Harvey EN, McMillen JH. An experimental study of shock waves resulting from theimpact of high velocity missiles on animal tissues. J Exp Med. 1947:85(3):321-328.http://jem.rupress.org/content/85/3/321.full.pdf+html. Accessed June 8, 2010.

5. Powell EB. Killing Power, a pamphlet published by National Rifle Association,Washington, DC, 1944. As cited by Harvey EN, McMillen JH. An experimentalstudy of shock waves resulting from the impact of high velocity missiles on animaltissues. J Exp Med. 1947:85(3):321-328. http://jem.rupress.org/content/85/3/321.full.pdf+html. Accessed June 8, 2010.

6. Super speed bullets knock ’em dead. Popular Mechanics. 1942;77(4):8-10.7. Ochsner EWA Jr. Experimental wound ballistics. In: Recent Advances in Medicine

and Surgery (19-30 April 1954) Based on Professional Medical Experiences in Japanand Korea 1950-1953), vol I. US Army Medical Service Graduate School WalterReed Army Medical Center, Washington, DC. Medical Science Publication 4.http://history.amedd.army.mil/booksdocs/korea/recad1/ch4-2.html. AccessedJune 8, 2010.

8. Robinson MD, Bryant PR. Peripheral nerve injuries. In: Zajtchuk R, ed. Textbookof Military Medicine, Part IV: Surgical Combat Casualty Care: Rehabilitation of theInjured Combatant, vol 2. Washington, DC: Office of the Surgeon General, USDepartment of the Army; 1999:419-574.

9. Suneson A, Hansson HA, Seeman T. Peripheral high-energy missile hits causepressure changes and damage to the nervous system: experimental studies on pigs.J Trauma. 1987;27(7):782-789.

10. Suneson A, Hansson HA, Seeman T. Central and peripheral nervous damage followinghigh-energy missile wounds in the thigh. J Trauma. 1988;28(1 suppl):S197–S203.

11. Suneson A, Hansson HA, Seeman T. Pressure wave injuries to the nervous systemcaused by high-energy missile extremity impact: part I. local and distant effects onthe peripheral nervous system. A light and electron microscopic study on pigs.J Trauma. 1990:30(3):281-294.

12. Suneson A, Hansson HA, Seeman T. Pressure wave injuries to the nervous systemcaused by high energy missile extremity impact, part II: distant effects on thecentral nervous system. A light and electron microscopic study on pigs. J Trauma.1990;30(3):295-306.

13. Liu Y, Chen Y, Li S. Mechanism and characteristics of the remote effects ofprojectiles. J Trauma (China). 1990;6(1 suppl):16-20.

14. Wang Q, Wang Z, Zhu P, Jiang J. Alterations of the myelin basic protein andultrastructure in the limbic system and the early stage of trauma-related stressdisorder in dogs. J Trauma. 2004:56(3):604-610.

15. Bellamy RF, Zajtchuk R. The physics and biophysics of wound ballistics. In: Con-ventional Warfare: Ballistic, Blast, and Burn Injuries. Washington, DC: Office of theSurgeon General, US Department of the Army. Zajtchuk R, ed. Textbook of MilitaryMedicine, Part I: Warfare, Weaponry, and the Casualty, vol 5; 1990:107-162.

16. Treib J, Haass A, Grauer MT. High-velocity bullet causing indirect trauma to thebrain and symptomatic epilepsy. Mil Med. 1996;161(1):61-64.

17. Sturtevant B. Shock wave effects in biomechanics. Sadhana. 1998:23(5):579-596.18. Courtney A, Courtney M. Links between traumatic brain injury and ballistic

pressure waves originating in the thoracic cavity and extremities. Brain Inj.2007:21(7):657-662.


CORRESPONDENCE



19. Krajsa J. Prıciny vzniku perikapilarnıch hemoragiı v mozku pri strelnych poranenıch(Causes of pericapillar brain haemorrhages accompanying gunshot wounds). Brno,Czech Republic: Institute of Forensic Medicine, Faculty of Medicine, MasarykUniversity; 2009. http://is.muni.cz/th/132384/lf_d/. Accessed June 8, 2010.

20. Carmichael J. Knockdown power. Outdoor Life, July 31, 2003. http://www.outdoorlife.com/node/45560. Accessed June 8, 2010.

21. Courtney MW, Courtney AC, Working toward exposure thresholds for blast-induced traumatic brain injury: thoracic and acceleration mechanisms. Neuro-Image. doi:10.1016/j.neuroimage.2010.05.025. Accessed June 8, 2010.

10.1227/NEU.0b013e3182041992

Regarding ‘‘Retroclival Epidural Hematomas: AClinical Series’’

To the Editor:We read the study reported by Tubbs and collaborators1 with

great interest. Tubbs et al have performed the first prospectivestudy of retroclival epidural hematomas (REDHs) and have re-ported more than 3 cases2 of this extremely rare subset of pos-terior cranial fossa hematomas representing less than 15% of allintracranial epidural hematomas.3 In fact, retroclival epiduralhematomas are mostly misdiagnosed by standard computed to-mography (CT), multiplanar reformatted high-resolution CTand magnetic resonance imaging (MRI) being the gold standardexamination.2,4,5

However, a systematic review of all literature using PubMedhas retrieved 27 previous cases of REDHs since the first report byOrrison et al in 1986.4,6-9 All but 5 were subsequent to motorvehicle accidents. A moderate trauma by a running fall was re-sponsible for the REDH in a 12-year-old girl.10 Calli et al11

reported a fortuitous REDH at a 1-month postoperative MRI ofa posterior fossa decompressive surgery for the management ofan acute cerebellar infarction. Another case was associated witha pituitary apoplexy.12 Finally, Cho et al7 reported the only‘‘spontaneous’’ case.

REDH is not a purely pediatric entity: 5 adult cases have been

reported (range, 26-62 years), among which were 2 traumatic

cases.3,13

Craniovertebral junction injury was found in 10 cases (37%).Only 3 were instable: 1 patient died soon after injury6 and2 patients required fusion associated with evacuation of thehematoma because of tetraplegia and, in one case, respiratorydifficulties.14,15 Two other cases were evacuated16 or decompressed.12

Cranial nerve disorders have occurred in 64% of cases (14 inthe sixth, 6 in the ninth, and 8 in the twelfth cranial nerve) andwere completely resolved by 2 months postinjury as the retroclivalblood collection at control imaging. The mean Glasgow ComaScale (GCS) score was 9.5, and there effectively was no corre-lation between hematoma size and presenting symptoms.17

Exceptionally, acute hydrocephalus could be associated withREDH (2 cases, both fatal).6,18

At our institution, in 5 years, we found only 2 cases of REDH,

subsequent to high-energy road accidents, in adults. These cases

have never been reported. The first one presented with a GCS

score of 7 and delayed bilateral sixth and twelfth cranial nervepalsies with spontaneous resolution. The other patient had aninitial GCS score of 3 and remained comatose 3 months afterthe accident; this was explained by a diffuse axonal injury. Nocraniovertebral instability was found, and conservative manage-ment was performed in all cases.

Tubb et al have described well the hypothetical mechanismsof the formation of REDHs, which are still controversial. We alsobelieve that disruption of the tectorial membrane is an importantfactor for the formation of hematoma. Bleeding may arise fromthe basilar venous plexus or the neuromeningeal trunk, branch ofthe ascending pharyngeal artery, which anastomoses with the me-ningohypophyseal trunk, the inferolateral trunk, and the odontoidarterial arch system.19 Tubb et al have also written 2 excellent articleson the tectorial membrane and the basilar venous plexus.20,21

REDH could be associated with retroclival subdural hema-tomas (RSDHs), particularly in violent injuries, as in our secondpatient.6 Isolated RSDHs are rarer than REDHs: only 7 caseshave been reported,22-28 but the context, clinical presentation,and mechanisms of formation are different. Only 1 patient wasyounger than 16 years of age (range, 4-78 years). Four cases werespontaneous. The only fatal case was subsequent to a minortrauma in a patient with hemophilia.24 The 2 other traumaticcases were subsequent to an aggression25 and a fall from abuilding.26 The mean GCS score was 14 (range, 11-15).

Because subdural hematoma is not limited by the boundariesof the tectorial membrane, it is quickly redistributed from theintracranial to the subdural space.26 Frequently, there was bloodcontamination when lumbar puncture was performed.22,23 Twocases of sixth cranial nerve palsies were found.27,28 Imagingshowed complete resolution within a month and no evacuationwas necessary.

Damien PetitPhilippe Mercier

Angers, France

1. Tubbs RS, Griessenauer CJ, Hankinson T, et al. Retroclival epidural hematomas:a clinical series. Neurosurgery. 2010;67(2):404-407.

2. Yama N, Kano H, Nara S, et al. The value of multidetector row computedtomography in the diagnosis of traumatic clivus epidural hematoma in children:a three-year experience. J Trauma. 2007;62(4):898-901.

3. Ratilal B, Castanho P, Vara Luiz C, Antunes JO. Traumatic clivus epiduralhematoma: case report and review of the literature. Surg Neurol. 2006;66(2):200-202; discussion 202.

4. Calisaneller T, Ozdemir O, Altinors N. Posttraumatic acute bilateral abducensnerve palsy in a child. Childs Nerv Syst. 2006;22(7):726-728.

5. Suliman HM, Merx HL, Wesseling P, van der Sluijs B, Vos PE, Thijssen HO.Retroclival extradural hematoma is a magnetic resonance imaging diagnosis.J Neurotrauma. 2001;18(11):1289-1293.

6. Orrison WW, Rogde S, Kinard RE, et al. Clivus epidural hematoma: a case report.Neurosurgery. 1986;18(2):194-196.

7. Cho CB, Park HK, Chough CK, Lee KJ. Spontaneous bilateral supratentorialsubdural and retroclival extradural hematomas in association with cervical epiduralvenous engorgement. J Korean Neurosurg Soc. 2009;46(2):172-175.

8. Schneck MJ, Smith R, Moster M. Isolated bilateral abducens nerve palsy associatedwith traumatic prepontine hematoma. Semin Ophthalmol. 2007;22(1):21-24.


CORRESPONDENCE



9. Kim MS, Cho MS, Kim SH. Delayed bilateral abducens nerve palsy after headtrauma. J Korean Neurosurg Soc. 2008;44(6):396-398.

10. Itshayek E, Goldman J, Rosenthal G, Chikoya L, Gomori M, Segal R. Extraduralhematoma of the clivus, not limited to the severely injured patient: case report andreview of the literature. J Trauma. 2006;60(2):417-420.

11. Calli C, Katranci N, Guzelbag E, Alper H, Yunten N. Retroclival epiduralhematoma secondary to decompressive craniectomy in cerebellar infarction:MR demonstration. J Neuroradiol. 1998;25(3):229-232.

12. Goodman JM, Kuzma B, Britt P. Retroclival hematoma secondary to pituitaryapoplexy. Surg Neurol. 1997;47(1):79-80.

13. Fuentes S, Bouillot P, Dufour H, Grisoli F. Occipital condyle fractures and clivusepidural hematoma. Case report [in French]. Neurochirurgie. 2000;46(6):563-567.

14. Marks SM, Paramaraswaren RN, Johnston RA. Transoral evacuation of a clivusextradural haematoma with good recovery: a case report. Br J Neurosurg.1997;11(3):245-247.

15. Papadopoulos SM, Dickman CA, Sonntag VK, Rekate HL, Spetzler RF. Traumaticatlantooccipital dislocation with survival. Neurosurgery. 1991;28(4):574-579.

16. Muller JU, Piek J, Kallwellis G, Stenger RD. Prepontine epidural hemorrhage [inGerman]. Zentralbl Neurochir. 1998;59(3):185-188.

17. Kwon TH, Joy H, Park YK, Chung HS. Traumatic retroclival epidural hematomain a child: case report. Neurol Med Chir (Tokyo). 2008;48(8):347-350.

18. Vera M, Navarro R, Esteban E, Costa JM. Association of atlanto-occipitaldislocation and retroclival haematoma in a child. Childs Nerv Syst. 2007;23(8):913-916.

19. Lasjaunias P, Moret J, Theron J. The so-called anterior meningeal artery of thecervical vertebral artery. Normal radioanatomy and anastomoses. Neuroradiology.1978;17(1):51-55.

20. Tubbs RS, Kelly DR, Humphrey ER, et al. The tectorial membrane: anatomical,biomechanical, and histological analysis. Clin Anat. 2007;20(4):382-386.

21. Tubbs RS, Hansasuta A, Loukas M, et al. The basilar venous plexus. Clin Anat.2007;20(7):755-759.

22. Tomaras C, Horowitz BL, Harper RL. Spontaneous clivus hematoma: case reportand literature review. Neurosurgery. 1995;37(1):123-124.

23. Schievink WI, Thompson RC, Loh CT, Maya MM. Spontaneous retroclivalhematoma presenting as a thunderclap headache. Case report. J Neurosurg.2001;95(3):522-524.

24. Myers DJ, Moossy JJ, Ragni MV. Fatal clival subdural hematoma in a hemo-philiac. Ann Emerg Med. 1995;25(2):249-252.

25. Casey D, Chaudhary BR, Leach PA, Herwadkar A, Karabatsou K. Traumatic clivalsubdural hematoma in an adult. J Neurosurg. 2009;110(6):1238-1241.

26. Ahn ES, Smith ER. Acute clival and spinal subdural hematoma with spontaneousresolution: clinical and radiographic correlation in support of a proposed patho-physiological mechanism. Case report. J Neurosurg. 2005;103(2 suppl):175-179.

27. Guilloton L, Godon P, Drouet A, Guerard S, Aczel F, Ribot C. Retroclivalhematoma in a patient taking oral anticoagulants [in French]. Rev Neurol (Paris).2000;156(4):392-394.

28. van Rijn RR, Flach HZ, Tanghe HL. Spontaneous retroclival subdural hematoma.JBR-BTR. 2003;86(3):174-175.

10.1227/NEU.0b013e31820419a7

Response to Letter to the Editor

We thank the authors for their interest in our article and fortheir additional comments with which we agree. We would alsoadd that other cases of retroclival hematomas may exist in theliterature and may be reported using different nomenclature(skull base hematoma, etc), thereby making a complete captureof these cases problematic. Conservative treatment appears tobe appropriate in the overwhelming number of cases. Withimproved imaging modalities, we believe these pathologic entitieswill become increasingly diagnosed.

R. Shane TubbsBirmingham, Alabama

Aaron Afshin Cohen-GadolIndianapolis, Indiana

10.1227/NEU.0b013e31820419bd

Surgical Approach and Safety of Spinal Cord StemCell Transplantation

The October 2009 edition of Neurosurgery featured the timelyreport of our team’s preclinical work to develop safe techniquesfor ventral horn spinal cord stem cell transplantation.1 The reportcoincided with Food and Drug Administration (FDA) approvalof the first trial to examine the safety of spinal cord stem celltransplantation for motor neuron disease. We anticipate that thistrial will be followed by a series of trials in North America,Europe, and Asia. These trials will coincide with similar ap-proaches applied to traumatic and demyelinating spinal corddisease.

The FDA approved protocol is entitled ‘‘A Phase I, Open–label, First–in–human, Feasibility and Safety Study of HumanSpinal Cord–Derived Cell Transplantation for the Treatment ofAmyotrophic Lateral Sclerosis.’’ As alluded to in the title, thetherapeutic product is derived from NIH–banked human fetalspinal cord. Technology developed originally in Ron McKay’slaboratory was used as the intellectual property platform ofa company called NeuralStem, Inc. (Rockville, Maryland). Un-like many protocols for the propagation of stem cells, Neural-Stem’s cells are propagated on laminar surface rather than as free–floating neurospheres.2-4 Hopes for amyotrophic lateral sclerosis(ALS) therapy rest on experiments in the SOD1 mutant rodentmodel of familial ALS conducted published in 2006. The SOD1gene has been found to have a variety of point mutations ina subset of patients with familial ALS. When mutant humanSOD1 is expressed in transgenic animals (rodents and pigs), theydevelop a phenotype that closely resembles human ALS. Xu et aldemonstrated that spinal cord grafts of the NeuralStem cells hadthe ability to preserve motor neuron numbers in the spinal cordsof SOD1 rodents and also prolonged their survival. Severalmechanisms are postulated to explain the efficacy of the grafts.First, a fraction of the cells are found to develop a gabaergicneuronal phenotype. These inhibitory cells form synapses withsurrounding cells providing a means to suppress excitotoxicitythought to play a role in the etiology of degenerative motorneuron death. The remaining cells develop an astrocytic phe-notype. Excitotoxicity in ALS has been ascribed to defects in glialexcitatory amino acid scavenging. Thus, it is possible that theremaining cells prevent toxic build up of excitatory transmitters.Finally, the cells are demonstrated to secrete a variety of neuralgrowth factors that may contribute to neural protection. It is clearthat these cells do not replace lost motor neurons and neuro-muscular junctions.

As mentioned, a variety of competing approaches exist for bothmolecular and cellular spinal cord therapies of motor neurondiseases, spinal cord injury, and demyelinating diseases. Our team


CORRESPONDENCE



began work on the techniques for safe and accurate ventralhorn targeting in collaboration with Clive Svendsen, PhD (Ce-dars–Sinai, Los Angeles, California; and Madison, Wisconsin),coauthor on our October Neurosurgery manuscript in 2004.Dr Svendsen’s team had documented the neuroprotective prop-erties of human fetal cortically derived cells.

Unlike the NeuralStem’s cells, these cells are grown as neu-rospheres. In addition, these cells do not form neurons ontransplantation, but rather all differentiate into astrocytes. Toaugment the protective capacity of these cells, Dr Svendsen’steam used lentiviral vectors to induce the expression and secretionof glial cell-derived neurotrophic factor (GDNF). Thus, theSvendsen cells act as organic minipumps for growth factors inaddition to scavenging excitatory amino acids. His team hasdemonstrated the ability of these grafts to preserve spinal cordmotor neurons in the SOD1 rat model.5 In 2005, Dr Svendsenand I submitted a PreIND application for transplantation of thesecells into humans. The manuscript published in October 2009reports on some of the critical preclinical work conducted tosupport this application. The master cell bank has now beencompleted, and vector production has been funded by the Na-tional Gene Vector Laboratories.

We anticipate a final FDA application with Dr Svendsen’s cellssometime in the next year. We are also supporting the preclinicaldevelopment of stem cells intended for use in spinal cordtransplantation for the treatment of ALS and transverse myelitisby Q Therapeutics (Salt Lake City, Utah), and the academic teamof Angelo Vescovi in Italy. Other teams are pursuing the idea ofembryonic stem cell transplantation for spinal muscular atrophyin infants.

Extensive work has been done in the Far East with humanapplication of spinal cord grafting. Most notably Dr Huang(Beijing) has reported a large series of olfactory ensheathing cellgrafts into the spinal cords of patients with chronic spinal cordinjuries.6-8 He has also transplanted these cells into the brains ofALS patients, a limited group of which received free hand cervicalinjections as well. The results of this work have not, to ourknowledge, been published as yet. In my personal correspondencewith Dr Huang, it has become clear that he has abandoned thisprocedure, though it is not clear why.

Anecdotal evidence supports efficacy of the transplants forspinal cord injury, and it is clear that this approach bears furtherscrutiny. However, based on our work in over a hundred pigs, weare emphatically opposed to free hand cord injections. This ap-proach has no reproducible targeting accuracy, and more im-portantly leaves the patient vulnerable to sheer injuries, pressureinjury, and graft reflux.

The FDA approved NeuralStem trial is aimed at establishingsafety and feasibility. We have developed the concept of ‘‘riskescalation’’ to replace the common ‘‘dose escalation’’ used inpharmacological trials. We envision the ultimate therapy in-volving staged lumbar and cervical transplants using multiplebilateral lumbar injections to preserve ambulation, and mul-tiple unilateral cervical injections (C3–C5) to preserve

diaphragmatic and proximal upper extremity function. Tobegin with the least risk possible, we will initially enroll non–ambulatory patients for lumbar unilateral injections,proceeding to lumbar bilateral injections in non–ambulatorypatients. The main risk of these first cohorts is pain and boweland bladder dysfunction. We will then proceed to ambulatorypatients, starting with unilateral multiple injections andproceeding to bilateral multiple injections. The increased riskto these cohorts involves potential loss of ambulation. Next, wewill move to unilateral cervical injection with the risk ofquadriplegia, followed finally by staged bilateral lumbar fol-lowed by unilateral cervical injection. The device described inthe October 2009 manuscript has been modified in a variety ofways to optimize safety and accuracy. Subsequent reports de-scribing this development are in production or review atpresent.

While preclinical studies address the safety of a specific bi-ological product, very little work has been done in large animalsthat models spinal cord transplantation into humans. That is,when we transplant human cells into pigs, they are, by definition,xenografts. While this is the required safety data requested for anIND, it is not a good model for human allografts proposed fortrials. As such, our team at Emory has recently submitted an RO1application for continued study of surgical techniques, graft re-jection, imaging, and graft control in the pig model. It is not clearwhich cell line will prove the most beneficial for ALS patients.However, we have great hopes that much will be learned fromthese initial trials about the best way to conduct human spinalcord transplantation.

Nicholas BoulisThais Federici

Atlanta, Georgia

1. Riley J, Federici T, Park J, et al. Cervical spinal cord therapeutic delivery: preclinicalsafety validation of a stabilized microinjection platform. Neurosurgery.2009;65(4):754–762.

2. Xu L, Yan J, Chen D, et al. Human neural stem cell grafts ameliorate motor neurondisease in SOD–1 transgenic rats. Transplantation. 2006;82(7):865–875.

3. Xu L, Ryugo DK, Pongstaporn T, Johe K, Koliatsos VE. Human neural stem cellgrafts in the spinal cord of SOD1 transgenic rats: differentiation and structural in-tegration into the segmental motor circuitry. J Comp Neurol. 2009;514 (4):297–309.

4. Yan J, Xu L, Welsh AM, et al. Combined immunosuppressive agents or CD4 anti-bodies prolong survival of human neural stem cell grafts and improve disease outcomesin amyotrophic lateral sclerosis transgenic mice. Stem Cells. 2006;24(8):1976–1985.

5. Suzuki M, McHugh J, Tork C, et al. GDNF secreting human neural progenitorcells protect dying motor neurons, but not their projection to muscle, in a rat modelof familial ALS. PLoS ONE. 2007;2(1):e689.

6. Chen L, Huang H, Zhang J, et al. Short–term outcome of olfactory ensheathingcells transplantation for treatment of amyotrophic lateral sclerosis. Zhongguo Xiu FuChong Jian Wai Ke Za Zhi. 2007;21(9):961–966.

7. Huang H, Chen L, Wang H, et al. Safety of fetal olfactory ensheathing celltransplantation in patients with chronic spinal cord injury. A 38–month followupwith MRI. Zhongguo Xiu Fu Chong Jian Wai Ke Za Zhi. 2006;20(4):439–443.

8. Huang H, Chen L, Wang H, et al. Influence of patients’ age on functional recoveryafter transplantation of olfactory ensheathing cells into injured spinal cord injury.Chin Med J (Engl). 2003;116(10):1488–1491.

10.1227/NEU.0b013e3182095e2e


CORRESPONDENCE



Direct Posterior Reduction and Fixation

To the Editor:Having read the article by Dr Feng-Zeng Jian et al1 entitled

‘‘Direct posterior reduction and fixation for the treatment of basilarinvagination with atlantoaxial dislocation,’’ we are attracted by thesurgical technique and its excellent efficacy which they reported.However, some issues still confused us and we wish to bring themto the authors and readers of Neurosurgery.

The authors described a method using C2 pedicle screws andoccipital instruments, which was claimed to provide reductiveforce to correct the cervico-medullary angle (CMA) and de-compress the craniovertebral junction.1 After the reduction theyperformed posterior fixation and fusion. The authors stated theirprocedure was ‘‘novel.’’ To our knowledge, however, this tech-nique of posterior reduction and fixation used by Jian et al wasfirst introduced by Abumi et al2 in 1999. Calling it ‘‘a novelposterior approach’’ without referring to the original author isunacceptable and unfair for the originality.

It is widely accepted that the reducible atlantoaxial dislocationcan be corrected by posterior reduction alone. It is noticed that 5(17.2%) of Jian’s 29 cases were reducible atlantoaxial dislocation(AAD), even though the author did not specify the cases. Thesecases should not be counted in calculation of reduction rate. Toour experience, however, favorable outcome for the irreducibleAAD (IAAD) cannot be achieved by posterior procedure withoutadditional anterior AA release.3–5 In the setting of IAAD, the‘‘single posterior longitudinal distraction’’1 between C2 and theocciput will be resisted by the contracted tissues at the front(longus capitis, longus colli, anterior longitudinal ligament, alar,

and apical ligaments). As a result, the effect of reduction islimited. In fact, the 2 cases whom Jian et al claimed to achieveanatomic reduction in their Figure 4 and 5 are not completelyreduced. In their Figure 6C, where the authors also claimed toachieve anatomic reduction, it is clear that the shape of theodontoid, clivus and C2 spinous process is inconsistent withthose in Jian’s Figure 6A (Figure 1), hence they must be takenfrom different sections. The reductive effect should be evaluatedin the same mid-sagittal computed tomographic (CT) scan.Moreover, if we draw a horizontal line via the center point of C1arch, we can find that the subdental synchondrosis (pointed bythe broken arrow) was not pulled down enough to be defined ascomplete reduction (Figure 1).

For IAAD, in order to achieve the reduction procedure, theodontoid should have two motions: descending and tilting for-ward (Figure 2). To our experience, cervical traction and transoralatlantoaxial release lead the migrated odontoid to descend. For-ward-bending of the locked plates and C2 pedicle screws cancause the odontoid to tilt forward and achieve the anatomicreduction (Figure 2 and 3). After the anatomic reduction, thefacet joint should be wedge-shaped and opened anteriorly (Figure3I and 3J). However, Jian’s posterior distraction only providedextension force for the odontoid and caused the facet joint toopen posteriorly (Figure 4). Consequently, it leads to an aggra-vated kyphosis of the craniovertebral junction (shown by Jian’sFigure 1) and the potential compression by the odontoid ex-tension. Unfortunately, we believe the technique Jian et al usedfor IAAD and basilar invagination is reducing the AAD ina wrong direction.

FIGURE 1. Jian’s Figure 6C and 6A had very different shape of the odontoid. A horizontal line via the center point of C1 archwas drawn. The subdental synchondrosis (pointed by the broken arrow) was not pulled down enough to be defined as completereduction.


CORRESPONDENCE



FIGURE 2. During the reduction procedure, the odontoid has 2 motions: descending and tilting forward. Forward-bending ofthe locked plates and C2 pedicle screws can cause the odontoid to tilt forward and achieve the anatomic reduction.

FIGURE 3. A, a 20-year-old male had basilar invagination, AAD and C1 occipitalization. B, reconstructive CT showed upward migration ofthe odontoid. C, preoperative MRI revealed the ventral compression and Chiari malformation. D, preoperative CT revealed the facet joint slideanteriorly and inferiorly. E, the contralateral facet joint. F, the patient underwent transoral release and posterior occiput-C2 fixation and fusion.Postoperative lateral X-ray showed an anatomic reduction. G, postoperative MRI obtained 5 days after surgery showed complete decompression.H, at the 4 months follow-up, CT showed anatomic reduction and solid fusion. I, after the anatomic reduction, the facet joint should be wedge-shaped and opened anteriorly. J, the contralateral facet joint was also opened anteriorly. K, MRI obtained 6 months after surgery showed thereduction of Chiari malformation.


CORRESPONDENCE



FIGURE 4. Jian’s posterior distraction only provided extension force for the odontoid and caused the facet joint to openposteriorly. Consequently, it leads to an aggravated kyphosis of the craniovertebral junction.

FIGURE 5. For case 14 in Jian’s report (the only 38-year-old female), the preoperative and postoperative CMA was all 135 degrees.


CORRESPONDENCE



For the patients with AAD and basilar invagination, the ventralcompression from the upward migrated odontoid is the primarypathology. The complete reduction of the odontoid is the mostimportant factor to improve the CMA and restore the herniatedcerebellar tonsils5 (Figure 3K). In doing so, the posterior de-compression is not reasonable. The authors removed part of theposterior margin of the foramen magnum for all 29 cases.1

However, only 7 of them had Chiari malformations. The in-dications for the remaining 22 cases were unreasonable. In ad-dition, the postoperative evaluation lost its anatomic landmarkbecause of removal of the osseous margin6 and the evaluation ofChamberlain’s line and McRae’s line was trustless.

Some information in Jian’s report was unbelievable. Jian’sFigure 4 was from case 14 (the only 38-year-old female in Jian’sTable 2), and the postoperative CMA was actually 135 degrees(Figure 5). However, the authors reported it as 150 degrees (fromJian’s Table 3). Jian’s Figure 5 was from case 7 (the only 44-year-old female), and they reported the CMA was 142 degrees. But wecan find the actual angle was only 132 degrees (Figure 6). Theauthors confirmed the fusion using a lateral X-ray in their Figure5. It is not an objective confirmation. The fusion status should bejudged on the reconstructive CT scan like their Figure 4D.Furthermore, readers can find the images in Jian’s Figure 6 werenot from the same patient because: 1) the odontoid in Jian’sFigure 6A, 6C, and 6E had different shape. 2) the pedicle screw inJian’s Figure 6D have different trajectory with that in 6F. 3) theimages in Jian’s Figure 6E and 6F had uncommon small C2spinous process and large C3 process, which differed with those in6A and 6C. 4) subdental synchondrosis of C2 found in Jian’s

Figure 6C was near the level of C1 anterior arch, while that in 6Ewas far underneath the level of C1 anterior arch (both of 6C and6E were postoperative and should have the same position of thesubdental synchondrosis).

Again we thank the authors’ report and extremely expect toparticipate in the discussion on this issue.

Chao WangShenglin Wang

Beijing, China

1. Jian FZ, Chen Z, Wrede KH, Samii M, Ling F. Direct posterior reduction andfixation for the treatment of basilar invagination with atlantoaxial dislocation.Neurosurgery. 2010;66(4):678-687; discussion 687.

2. Abumi K, Takada T, Shono Y, Kaneda K, Fujiya M. Posterior occipitocervicalreconstruction using cervical pedicle screws and plate-rod systems. Spine (Phila Pa1976). 1999;24(14):1425-1434.

3. Wang C, Yan M, Zhou HT, Wang SL, Dang GT. Open reduction of irreducibleatlantoaxial dislocation by transoral anterior atlantoaxial release and posterior in-ternal fixation. Spine (Phila Pa 1976). 2006;31(11):E306-E313.

4. Wang C, Wang S. Visocchi M, Pietrini D, Tufo T, Fernandez E, Di Rocco C,(2009). Pre-operative irreducible C1-C2 dislocations: intra-operative reduction andposterior fixation. The ‘‘always posterior strategy’’. Acta Neurochir (Wien).2009;151(5), (10):551-560;discussion;1329-1331; author reply 1333-1336.

5. Wang S, Wang C, Yan M, Zhou H, Jiang L. Syringomyelia with irreducibleatlantoaxial dislocation, basilar invagination and Chiari i malformation. Eur Spine J.2010;19(3):361-366.

6. Wang S, Wang C, Passias PG, Li G, Yan M, Zhou H. Interobserver and intra-observer reliability of the cervicomedullary angle in a normal adult population. EurSpine J. 2009;18(9):1349-1354.

10.1227/NEU.0b013e3181f3586a

FIGURE 6. For Jian’s case 7 (the only 44-year-old female), the preoperative CMA was 127 degree, while the postoperative anglewas 132 degrees.


CORRESPONDENCE



Documents

Keyhole and Key-Bur-Hole