spine volume 30, number 6, pp 613–620 ©2005, lippincott ... · pdf filewas...

SPINE Volume 30, Number 6, pp 613–620©2005, Lippincott Williams & Wilkins, Inc.

Influence of Acute Shortening on the Spinal Cord: AnExperimental Study

Norio Kawahara, MD, PhD, Katsuro Tomita, MD, PhD, Tadayoshi Kobayashi, MD,Mohamed E. Abdel-Wanis, MD, PhD, Hideki Murakami, MD, PhD, andTomoyuki Akamaru, MD,

Study Design. Morphometric changes of the spinalcord and influence on spinal cord-evoked potentials andspinal cord blood flow and postoperative function of hindlimbs were studied in various degrees of acute spinalcolumn shortening in dogs.

Objectives. To study the morphometric and physio-logic effects of acute spinal column shortening on thespinal cord.

Summary of Background Data. The technique of acutespinal column shortening is sometimes applied for cor-rection of spinal deformity, total en bloc spondylectomyoperation, or other diseases. However, safe limits andphysiologic effects of acute spinal column shorteninghave not yet been described.

Methods. Total spondylectomy of T13 was performedin dogs after spinal instrumentation placed 2 levels aboveand 2 levels below the spondylectomy level. Spinal col-umn was gradually shortened until the lower endplate ofT12 contacted the L1 upper endplate (maximum of 20mm). When any morphologic change of the dural sac orthe spinal cord was observed, the length of shorteningwas measured. Spinal cord-evoked potentials were re-corded on the exposed dura mater following epiduralstimulation at the C7 level in 8 dogs. Spinal cord bloodflow was measured during shortening in 6 dogs. Hind-limb function was evaluated 2 weeks after operation in 10dogs.

Results. No morphometric changes occurred in thedural sac and the spinal cord until shortening of 7.2 � 1.7mm (n � 6). From 7.2 � 1.7 to 12.5 � 1.1 mm shortening,the dural sac was deformed, whereas the spinal cordmaintained its shape. Shortening more than 12.5 � 1.1mm buckled the dural sac, and the spinal cord kinkeditself and was compressed by the buckled dura in itsconcave side (n � 6). No changes could be detected inspinal cord-evoked potentials in 5 or 10 mm of shorten-ing. Spinal cord-evoked potential changes were recordedin the 2 of 6 dogs with 15 mm of shortening. At 20 mm ofshortening, spinal cord-evoked potential abnormality wasobserved in 4 of 6 dogs. At shortening of 5, 10, 15, and 20mm, spinal cord blood flow was 146 � 10%, 160 � 21%,102 � 17%, and 93 � 7% of the control (29.2 � 7.9 mL/100g/min, n � 6), respectively. All 3 dogs with 10 mm of

shortening had normal hindlimb function 2 weeks afteroperation. One of the 3 dogs with 15 mm of shorteninghad paraparesis. Three of the 4 dogs with 20 mm ofshortening had also paraparesis after operation.

Conclusions. Acute spinal column shortening can becharacterized into 3 phases. Phase 1, safe range: occurredduring shortening within one-third of the vertebral seg-ment and is characterized by no deformity of the dural sacor the spinal cord. Phase 2, warning range: occurred dur-ing spinal shortening between one-third and two-thirds ofthe vertebral segment and is characterized by shrinkingand buckling of the dural sac and no deformity of thespinal cord. Phase 3, dangerous range: occurred aftershortening in excess of two-thirds of the vertebral seg-ment and is characterized by spinal cord deformity andcompression by the buckled dura. Spinal shorteningwithin the safe range increases spinal cord blood flow.

Key words: spinal shortening, spinal correction, spinalcord function, morphometric changes, spinal cord evokedpotentials, spinal cord blood flow. Spine 2005;30:613–620

The technique of spinal column shortening is sometimesapplied for correction of spinal deformity,1–4 total enbloc spondylectomy,5–7 or other diseases. Correction ofexcess spinal kyphosis may be done through anteriorwedge opening (Smith-Peterson technique8) or posteriorwedge closing osteotomies.9,10 There were 2 merits con-sidered in closing-wedge osteotomy: superior biome-chanical stability and reduction of the size of the graftedarea as a result of minimizing the spinal column defect.This reduces the amount of bone graft to be harvested.Furthermore, as autograft incorporation is influenced bythe size of the grafted area,11 remodeling of grafted boneis expected to be more rapid. Leatherman and Dickson3

stressed that the only method that can be used for thecorrection of a rigid deformity without the risk of neu-rologic complication is shortening and straightening ofthe spine. However, several investigators reported therisk of neurologic complications associated with spinalshortening in the surgical correction of spinal deformi-ties.1,4,9,12–14 During the procedure, the spinal column isshortened and the spinal cord may become redundant inrelation to the spine.13,14 It was reported that duralbuckling may occur over approximately 35° in correc-tion at any one level in closing-wedge osteotomy.13

Gertzbein and Harris9 reported their procedure was bestsuited for corrections of approximately 30° to 40°.Kawahara et al12 reported closing-opening wedge os-teotomy to sufficiently correct severe angular kyphosis,avoiding spinal cord injury due to cord kinking. The first

From the Department of Orthopaedic Surgery, School of Medicine,Kanazawa University, Kanazawa, Ishikawa, Japan.Acknowledgment date: July 15, 2003. First revision date: December26, 2003. Acceptance date: April 15, 2004.The manuscript submitted does not contain information about medicaldevice(s)/drug(s).No funds were received in support of this work. No benefits in anyform have been or will be received from a commercial party relateddirectly or indirectly to the subject of this manuscript.Address correspondence and reprint requests to Norio Kawahara, MD,PhD, Department of Orthopaedic Surgery, School of Medicine, Ka-nazawa University, 13-1 Takaramachi, Kanazawa, Ishikawa 920-8641, Japan; E-mail: [email protected]

613

30° to 35° of kyphosis was corrected using the closingwedge technique with the pivot point of the anteriorlongitudinal ligament, and the spinal cord was protectedfrom the danger of deformation such as hanging down,kinking, or dural buckling. Next, the pivot point wasmoved posterior to the spinal cord, and the remainder ofthe osteotomy was performed using the opening wedgetechnique. However, the safety limits and physiologiceffects of spinal shortening on the spinal cord were notpreviously studied. The objective of this study was toclarify the effect of spinal shortening on the spinal cordshape and physiologic functions.

Materials and Methods

Surgical Preparation. Forty-six adult dogs, weighing from 10to 12 kg, were used in this experiment. Following sedation withketamine hydrochloride injection into the paravertebral mus-cles (2.5 mg/kg), a catheter was inserted into the radial vein forfluid replacement and the administration of drugs. Each animalwas orotracheally intubated with a polyethylene tube (6.5mm), paralyzed with pancuronium 0.1 mg/kg/hr and main-tained on ventilatory support with 1% halothane in 50:50N2O/O2. A cannula (2.0-mm diameter) was then inserted in aretrograde manner into the right femoral artery for repeatedblood sampling and for continuous blood pressure monitoring.The animals were then placed on a surgical table in the proneposition. Body temperature was monitored using a rectal tem-perature probe. Body temperature was maintained between 36C and 37 C with a heating pad when necessary. Arterial bloodgases were measured at 90-minute intervals. Metabolic andrespiratory acid-base balance was controlled with supplemen-tal intravenous NaHCO3 in order to maintain the pH near7.40. The PaO2 was held above 100 mm Hg with respiratoryvolume control. The concentrations of halothane and N2O/O2

were constant throughout the procedure. During the steadystate, painful stimulation did not cause any increase in bloodpressure.

Surgical Procedure. A midline longitudinal incision extend-ing from T9 to L3 was made and the laminae from T10 to L3were exposed. Texas-Scottish Rite Hospital (TSRH; SofamorDanek, Memphis, TN) screws were obliquely inserted into thevertebral body of T11, T12, L1, and L2 from the base of thetransverse processes. One screw was inserted into each vertebra.Laminectomies were performed from the caudal region of T12 tothe cranial region of L1. It was necessary to retract the T13 spinalnerves in order to make the surgical room for T13 total vertebrec-tomy. There was a fear of injuring the spinal cord by stretching theT13 spinal nerves when the spinal nerves were preserved. Accord-ingly, both nerve roots of T13 were ligated and cut. The vertebralbody of T13 as well as adjacent discs were completely resectedusing a curette and a burr with great care so as not to injure thedura or the spinal cord (Figure 1A).

Morphologic Observation. The spinal column was gradu-ally shortened by a compression maneuver between the screwsinserted into T12 and L1. The length between the lower end-plate of the T12 and the upper endplate of the L1 (an entiresegment) was 20.3 � 0.5 mm (n � 46). The maneuver wascompleted on the contact of the 2 endplates when the length ofan entire segment was 20 mm or less. The maximum shortening

was 20 mm in the case with an entire segment of more than 20mm. As the shortening procedure was performed, the morpho-logic changes of the dural sac and the spinal cord were ob-served. When morphologic change was evident, the length ofshortening was measured (n � 4). In another 5 dogs, at eachmorphologic condition displayed by the spinal canal contents(0, 5, 10, 15, and 20 mm of shortening), microfill compound(Microfil®, Flow Tech Inc., Carver, MA) was injected into thesubarachnoid space via a myelography technique. This tech-nique was employed in order to maintain the characteristicspinal cord morphologic changes for gross inspection after sac-rificing the spine segment (Figure 2).

Spinal Cord-Evoked Potentials Study. Six dogs were usedin a spinal cord-evoked potentials (SCEPs) study.15–17 Elec-trodes used for stimulation and recording consisted of 18-gauge polyethylene tubes with 2 fine platinum coil terminals(UPG-100–2P: Unique Medical, Inc., Tokyo, Japan). The elec-trodes were identical to those employed in procedures involv-ing patients. The stimulating electrode was introduced into thecervical epidural space from the laminectomy area in a rostralmanner (Figure 1C). Recording electrodes were placed on theposterior midline of the dura mater at the level of the upper endof the L1 vertebra15–17 (Figure 1C). Pulses of constant stimu-lation current (20 Hz, 0.2 msec duration) were routinely deliv-ered. A DISA 1500 4-CH Digital EMG-System (DISA Inc.) wasused to record signals. Generally, 50 consecutively recordedsignals were averaged per tracing.

Increase or decrease more than 30% of the amplitude fromthe baseline of the first positive or negative wave of the SCEPswas evaluated to be abnormal.15,16

Measurement of Spinal Cord Blood Flow. Spinal cordblood flow (SCBF) was measured by a hydrogen clearance tech-nique in 6 dogs.18,19 Maximum change of the blood flow wasexpected in the center of the spinal cord deformation. There-fore, SCBF was measured at the T13 nerve root branching offlevel of the spinal cord, which is the site of maximum kinking ofthe spinal cord with the 15 or 20 mm of spinal shortening. Theanteroposterior diameter of the spinal cord at the T13 was 6.03� 0.06 mm (n � 6). The gray matter exists from 1.58 � 0.07mm to 3.25 � 0.07 mm (n � 6) in depth from the midline of theposterior surface of the spinal cord at T13 level in the dogs. Theplatinum electrode with a diameter of 200 ı̀m was inserted 2.5mm in depth from the midline of the posterior surface of thespinal cord through a small hole of the dural sac at the T13

Figure 1. Schematic illustration of the experimental study. A, totalspondylectomy of the 13th vertebra. B, measurement of spinalcord blood flow. C, recording of spinal cord evoked potentials.

614 Spine • Volume 30 • Number 6 • 2005

nerve root branching off level of the spinal cord (Figure 1B). Itwas certain that the apex of the platinum electrode was insertedin the gray matter of the spinal cord. A 5-mV, 40-second po-tential was applied between the platinum and silver/silver chlo-ride reference electrodes inserted into the paravertebral musclesadjacent to the spondylectomy level. Spinal cord blood flowwas measured at 0, 5, 10, 15, and 20 mm (or an entire segment)spinal shortenings. The time course of hydrogen concentrationwas recorded on an RBF-2 recorder (Biomedical Science, Inc.,Kanazawa, Japan). The data were plotted in monoexponentialform and SCBF was calculated as previously described.18,19

Spinal cord blood flow of each spinal shortening phase waspresented in comparison to the baseline SCBF observed in eachanimal after laminectomy and corpectomy (total spondylec-tomy of T13) before spinal shortening.

The data were presented as a mean � standard deviation. Sta-tistical analysis was performed using the paired t test. Values wereconsidered significant when the P value was 0.05 or less.

Spinal Angiography. Spinal angiography was performed 2weeks after each shortening operation of the spine. After ani-mals were given general anesthesia and a catheter was insertedinto the right femoral artery, they were placed in a right recum-bent position. After the left sixth rib was excised and the tho-racic cavity was opened, the aorta above the seventh thoracicvertebra was marked with 1–0 silk to prepare for ligation of theaorta. Subsequently, both common iliac arteries were delin-eated by the incision of the peritoneum, and the left commoniliac artery was ligated. The right common iliac artery waspreserved for measurement of the intra-arterial pressure. A16G indwelling needle was placed into the thoracic aorta thatwas marked previously, and heparinized physiologic saline, thebottle of which was hung from the ceiling for irrigation, was

infused through it. Then, the aorta was ligated with the threadthat was used previously for marking. The celiac artery and themesenteric artery were identified, and they were clamped withhemostatic forceps so that blood flow was interrupted. Afterlocal irrigation and lavage with 3 L of heparinized physiologicsaline, 100 mL of 10% formalin was infused into the vessel forfixation as the intra-arterial pressure was being monitored toensure the same as the average blood pressure before clampingof the aorta. Then, 100 mL of silicon compound was infusedmanually. After 1 hour waiting for the compound to be scle-rosed, the animals were placed in a prone position and thespinal cord between the 10th thoracic vertebra and the thirdlumber vertebra was excised, and the vessels around the surfaceof the spinal cord were examined. The maximum and mini-mum diameters of the anterior spinal artery were selected, andthey were measured on the spinal cord from the midpoint be-tween the T12 and T13 nerve branching off level to the mid-point between the T13 and L1 nerve branching off level, exceptthe site of maximum kinking, with a stereomicroscope. Themean value of the maximum and minimum diameters was eval-uated as the representative of the diameters of the anteriorspinal artery in the shortened area of the spinal cord. In addi-tion, the diameter of the anterior spinal artery was also mea-sured at the site of maximum kinking in the group with the 15or 20 mm of spinal shortening operations. The data were pre-sented as a mean � standard deviation. Statistical analysis wasperformed using the Bonferroni/Dunn analysis. Values wereconsidered significant when the P value was 0.05 or less.

Postoperative Neurologic Examination. The spine wasshortened by 10 mm in 3 dogs, 15 mm in 3 dogs, and 20 mm (oran entire segment) in 4 dogs. Neurologic function was evalu-

Figure 2. Morphologic changesof the dural sac and the spinalcord in spinal shortening. A, mor-phologic changes of the duralsac during operation. The spinalcanal contents breached the spi-nal canal in the case of greaterthan 12.5 mm of spinal shorteninglength. At 20 mm of spinal short-ening, severe spinal cord kinkingoccurred. B, sagittal cross-section of the spinal cord. Thespinal cord maintained its linearorientation in 5 and 10 mm short-ening. At spinal shortening of 15mm and 20 mm, the spinal cordkinking was observed.

615Influence of Acute Shortening on the Spinal Cord • Kawahara et al

ated with a modified Tarlov grading system 2 days and 14 daysafter operation.20,21

Grade 0—complete paraplegia with no hind extremity mo-tion

Grade 1—minor joint movements

Grade 2—major joint movements

Grade 3—animal can stand

Grade 4—animal can walk

Grade 5—animal can climb a 20° inclined plane

Results

Morphologic ObservationThe dural sac and spinal cord did not exhibit any mor-phologic changes until a shortening length of 7.2 � 1.7mm (n � 4, 36.0 � 8.5% of an entire segment) (Figure2). From 7.2 � 1.7 to 12.5 � 1.1 mm (n � 4, 36.0 � 8.5to 62.5 � 5.5% of an entire segment), the dura materwas deformed in the manner of an accordion. However,the spinal cord maintained its linear orientation. At spi-nal shortening of more than 12.5 � 1.1 mm (n � 4,62.5 � 5.5% of an entire segment), the dura mater buck-led and spinal cord kinking was observed. The spinalcanal contents extruded from the spinal canal in the caseof greater than 12.5 mm of spinal shortening. At 20 mm(or an entire segment) of spinal shortening (n � 4), severespinal cord deformity occurred (Figure 2A).

In another 5 dogs, at each morphologic condition dis-played by the spinal canal contents (0, 5, 10, 15, and 20mm shortenings), microfill compound (Microfil®, FlowTech Inc., Carver, MA) was injected into the subarach-noid space via a myelography technique. The spinal cordmaintained its linear orientation in 5 and 10 mm short-ening. At spinal shortening of 15 mm and 20 mm, thespinal cord kinking was observed (Figure 2B).

Spinal Cord-Evoked Potentials StudySix dogs were used for SCEP study (Table 1, Figure3).15–17 No significant changes occurred in SCEPs at 5mm or 10 mm of shortening. Increase of the amplitude ofthe P1 wave more than 30% of the control were recordedin 2 of the 6 dogs with 15 mm of shortening. At 20 mmof shortening (or an entire segment), which demon-strated severe kinking of the spinal cord, abnormality ofthe SCEPs was observed in 4 of the 6 dogs (Table 1,Figure 3).

Measurement of Spinal Cord Blood FlowBaseline SCBF (after laminectomy and before shorten-ing) was 29.9 � 7.91 mL/100 g/min (n � 6). At 5 mm ofspinal shortening, SCBF increased to 146 � 10% (n � 6)in comparison to the SCBF occurring before spinal short-ening. The difference is statistically significant (P �0.007). At 10 mm of shortening, SCBF increased to160 � 21% (n � 6) in comparison to baseline SCBF. Thedifference was also statistically significant (P � 0.03).There was no statistically significant difference betweenSCBF at 5 mm and 10 mm shortening (P � 0.39). At 15mm of shortening, following spinal cord kinking, SCBF

decreased to 102 � 17% (n � 6) relative to the levelobserved before the procedure. At 20 mm (or an entiresegment) shortening, characterized by severe spinal cordkinking, SCBF was 93 � 7% (n � 6) relative to thereading obtained at the preshortening stage (Figure 4).

Spinal AngiographyIn the dogs without vertebral shortening operations, theanterior spinal artery ran straight longitudinally in par-allel to the spinal cord. The diameter of the anteriorspinal artery was 0.575 � 0.066 mm (n � 3). In thegroup with the 5 mm of shortening operation of thevertebra, the anterior spinal artery mildly meandered,and the diameter of the artery was 0.769 � 0.017 mm(n � 3), significantly larger than that in the group with-out the shortening operations (P � 0.0078). In the groupwith the 10 mm of spinal shortening operations, the an-

Table 1. Changes of Spinal Cord-Evoked Potentials inSpinal Shortening

Spinal Shortening 5 mm 10 mm 15 mm 20 mm

Dog 1N1 — — — DecreaseP1 — — — —

Dog 2N1 — — — —P1 — — — —

Dog 3N1 — — — —P1 — — Increase Increase

Dog 4N1 — — Decrease DecreaseP1 — — Increase Increase

Dog 5N1 — — — DecreaseP1 — — — —

Dog 6N1 — — — —P1 — — — —

N1 � the first negative wave; P1 � the first positive wave; — � no remarkablechanges; Decrease � decrease more than 30% of N1 or P1 amplitude of thecontrol; Increase � increase more than 30% of N1 or P1 amplitude of thecontrol.

Figure 3. The spinal cord evoked potentials of dog 4 in spinalshortening. More than 30% decrease of the N1 amplitude (**) andmore than 30% increase of the P1 amplitude (*) were recorded inthe SCEPs study in 15 and 20 mm of shortening.


terior spinal artery meandered even more evidently thanin the group with the 5 mm shortening operations, andthe diameter of the artery was 0.750 � 0.050 mm (n �3). In the group with the 15 mm of spinal shorteningoperations, the degree of meandering of the anterior spi-nal artery was higher than that in the group with the 10mm of shortening operations, and the diameter was0.775 � 0.090 mm (n � 3). The artery was constrictedand the diameter of it was 0.629 � 0.16 mm (n � 3) atthe site where kinking of the spinal cord was observed. Inthe group with the 20 mm (or an entire segment) ofspinal shortening operations, the diameter of the anteriorspinal artery was 0.792 � 0.029 mm (n � 3). The arterywas obstructed and the diameter of it was 0.132 � 0.229mm (0, 0, 0.396 mm, n � 3) at the site of spinal cord

kinking (Figure 5). Compared to the group without thespinal shortening operations, the diameter of the anteriorspinal artery was significantly increased in the groupswith the 5, 10, 15, and 20 mm (or an entire segment) ofspinal shortening operations (P � 0.0049), without sig-nificant differences in diameter between the groups withthe shortening operations (Figures 5 and 6).

Postoperative Neurologic ExaminationThree dogs with 10 mm of shortening had a modifiedTarlov Grade 5 two days after operation. Two of the 3dogs with 15 mm of shortening had a modified TarlovGrade 5 two days after operation, whereas another had amodified Tarlov Grade 4 two days after operation. Oneof the 4 dogs with 20 mm of shortening had a modifiedTarlov Grade 5 two days after operation, whereas the

Figure 4. Changes of spinal cord blood flow in spinal shortening.At 5 mm of spinal shortening, the reading was 146% on average incomparison to the SCBF occurring before spinal shortening. At 10mm of shortening, SCBF increased to 160% of the control. At 15mm of shortening, following spinal cord kinking, SCBF decreasedto 102% of the control. At 20 mm shortening, SCBF was 93% of thecontrol.

Figure 5. Morphologic changesof the anterior spinal artery in thespinal shortening. The anteriorspinal artery was in a linear ori-entation in the preshorteningmodel. The anterior spinal arter-ies meandered, and their diame-ter was increased in the speci-mens of more than 5 mm spinalshortening. The artery was con-stricted at the site where kinkingof the spinal cord was observedin the 15 mm of shortening. At 20mm of shortening, the anteriorspinal artery was obstructed atthe bottom of the spinal cordkinking.

Figure 6. Changes of the diameter of the anterior spinal artery inthe spinal shortening. The diameter of the anterior spinal arterywas 0.575 mm on average (n � 3). The diameter of the artery ineach group with 5 mm, 10 mm, 15 mm, and 20 mm of spinalshortening was 0.769 mm, 0.750 mm, 0.775 mm, and 0.792 mm,respectively (n � 3).


other 3 had a modified Tarlov Grade 4 two days afteroperation. The neurologic condition did not change 14days after operation in all dogs with postoperative para-paresis except 1, which had partially improved neurolog-ically.

Discussion

It is clear that the safety of spinal shortening might de-pend on the spinal level of shortening. The safe range ofspinal shortening might be more at the cauda equinalevels than at the spinal cord level. This may be related tothe fact that the nerves in the cauda equina may be ca-pable of responding to spinal shortening by increasingtheir redundancy. The spinal cord terminates in dogsnear the L6/7 spine level, whereas in humans it termi-nates near the L1 level. The T13 level was used in ourspinal shortening experiment. This level may correspondto the middle lower thoracic spine in humans.

An entire segment of the T13 had a mean of 20.3 mm.The dural sac and spinal cord did not exhibit any mor-phologic changes until a shortening length of 7.2 mm.From 7.2 to 12.5 mm, the dura mater was deformed inthe manner of an accordion. However, the spinal cordmaintained its linear orientation. At spinal shortening ofmore than 12.5 mm, the dura mater buckled and spinalcord kinking was observed. From our morphologicstudy, 3 phases may be defined during spinal shorteningof 1 vertebra (Table 2). Phase 1 involved approximatelyone-third of the vertebral segment; within which therewas no deformity of the dura or the spinal cord. Phase 2involved less than two-thirds of the vertebral segment.No deformity of the spinal cord occurred; however,shrinkage of the dura was evident. Phase 3 involvedgreater than two-thirds of the vertebral segment. Kinkingof the spinal cord and dural buckling occurred.

The morphologic changes were a result of 2 mecha-nisms by which the spinal cord accommodated the spinalshortening. First, the dural sac slid by telescoping in thespinal canal. Second, the spinal cord slid by telescopingin the dural sac. When the range of tolerance of the spinalcord to spinal shortening was exceeded, kinking of thespinal cord occurred.

In this study, SCEPs did not change during 5 or 10 mmof shortening that is in Phase 1 or 2 in any dog. All 3 dogswith 10 mm of shortening had a modified Tarlov Grade5 after operation. Spinal cord-evoked potential changeswere recorded in the 2 of 6 dogs with 15 mm of short-ening, which is in Phase 3. One of the 3 dogs with 15 mm

of shortening had paraparesis in the hindlimbs (modifiedTarlov Grade 4) after surgery. At 20 mm of shortening(or an entire segment), which demonstrated severe kink-ing of the spinal cord, SCEP abnormality was observed in4 of 6 dogs. Three of the 4 dogs with 20 mm of shorten-ing had paraparesis in the hindlimbs (modified TarlovGrade 4) after the operation. These findings revealed thatspinal cord function was not compromised in the Phase 1or 2 (no deformity in the spinal cord) in spinal shorten-ing. However, there is a risk of spinal cord damage due tospinal cord deformation or kinking in Phase 3. In anactual operation, it is impossible to observe the conditionof the spinal cord from outside of the dura. Kostuik22

described that he performed closing wedge osteotomyfor correction of kyphosis safely, observing the spinalcord in direct vision by durotomy in a revision patient,and did a plasty of the scarred dura. If spinal shorteningup to Phase 2 is required, direct observation of the spinalcord by durotomy during correction may also be safe. Itwould be followed by dural plasty.

It was reported that blood flow in the tissue of only2.0 mm3 immediately surrounding the apex of the plati-num electrode was measured in the hydrogen clearancemethod.23 Although, the spinal cord blood flow can bemeasured in any part of any one segment or in all seg-ments at the same time and the location of the measure-ment is precisely known in the autoradiographic tech-nique.24–27 However, if it is desirable to observe changesin SCBF produced by physiologic or pathologic eventsover a period of time, the disadvantage of having to killthe animal at the time of measurement may be overcomeby performing experiments on groups of animals at dif-fering time intervals after changing the particular param-eter, for instance, spinal injury.27 On the other hand, it ispossible to observe serial changes in SCBF produced byphysiologic or pathologic events over a period of time inthe same animal in the hydrogen clearance method.28–31

We desired to observe serial changes in SCBF by spinalshortening in the same dog and employed the hydrogenclearance method to measure SCBF in this series.

In the literature, dog gray matter blood flow has beenreported in the range of 40.6 to 63 mL/100g/min,whereas white matter blood flow has been reported inthe range of 10.3 to 21.7 mL/100g/min.18,32 The baselineSCBF (in the condition of postlaminectomy and ligation ofbilateral T13 nerve roots) in our study was 29.9 � 7.91mL/100 g/min using a hydrogen clearance method. Torib-atake33 noted that SCBF decreased to approximately 78%of baseline as a result of laminectomy. Also, Anderson etal34 reported that SCBF depressed to 22% to 45% of base-line following laminectomy. The baseline SCBF in our ex-periment was relatively low in comparison to that in theliterature, as it was measured following laminectomy andligation of both segmental arteries of T13.

At 5 mm of spinal shortening, SCBF increased signif-icantly to 146% in comparison to the SCBF occurringbefore spinal shortening. At 10 mm of shortening, SCBFincreased to 160% compared to baseline SCBF. This in-

Table 2. Morphological Phases Occurred During SpinalShortening of a Single Vertebral Segment

Shortening Length* Dura Spinal Cord

Phase 1: less than one-third No deformity No deformityPhase 2: one-third to two-thirds Shrinkage No deformityPhase 3: more than two-thirds Buckling Kinking

*Shortening length is shown in proportion to a single vertebral segment.


crease could not be attributed to an elevation in bloodpressure, as blood pressure was maintained at a constantlevel throughout the spinal shortening procedure. Youngand Flamm31 suggested high-dose corticosteroid therapyimmediately increases SCBF by a local vasodilatory ef-fect. Ohashi et al35 reported that SCBF diminished as aresult of decrease of the diameter of the spinal arteries inspinal cord injury. The SCBF increased within minutesfollowing spinal shortening in this study. The rapid in-crease of SCBF during spinal shortening strongly sug-gests a local vasodilatory effect.31 Spinal angiogram inthis study revealed that in the group with the 5 mm ofshortening operation of the vertebra, the anterior spinalartery mildly meandered, and the diameter of the arterywas 0.77 mm, significantly larger than 0.58 mm in thegroup without the shortening operations. In the groupwith the 10 mm of spinal shortening operations, the an-terior spinal artery meandered even more evidently thanin the group with the 5 mm of shortening operations, andthe diameter of the artery was 0.75 mm. The spinal cordtermination is tethered to the coccygeal bone by the filumterminale. The spinal cord is also tethered segmentally tothe spinal column by the nerve roots. As the spinal cordtether is relaxed by spinal shortening, tension of the spi-nal cord and the spinal arteries decreases. Consequently,vasodilation of the arteries of the spinal cord might oc-cur, resulting in a concomitant reduction of resistance toblood flow. An increase in SCBF resulted immediately.This phenomenon is strongly in contrast to the decreaseof SCBF when the spinal cord is stretched.25,36–38 Fol-lowing kinking of the spinal cord, SCBF decreased sig-nificantly due to constriction of the spinal arteries in theventral aspect. However, the arteries present in the dor-sal aspect were not constricted. Therefore, SCBF main-tained a mean of 93% of baseline levels despite severekinking of the spinal cord even at 20 mm or an entiresegment of shortening.

Spinal cord blood flow changes have been described inmany different animal and mechanical injury models.24–

26,28–31,39–42 Severe injury has been associated with im-pairment of normal autoregulation and persistent reduc-tion in SCBF.30,39,40,42,43 On the other hand, it has beenalso reported that hyperemia was observed in the injuredspinal cord,28,39 especially in the white matter of thespinal cord in the incomplete injury.29,30,41 It was sug-gested that increase or decrease of SCBF may depend onthe severity of the spinal cord injury.29,30,39 In our series,at 5 mm and 10 mm of spinal shortening, SCBF mea-sured in the gray matter increased significantly to 146%and 160%, respectively, compared to baseline SCBF. Al-though, no significant changes occurred in SCEPs at 5mm or 10 mm of shortening, and 3 dogs with 10 mm ofshortening had a modified Tarlov Grade 5 (normal limbfunction) 2 days and 2 weeks after operation. It is sus-pected that the increase of SCBF associated with increaseof the diameter of the anterior spinal artery in spinal

shortening in our series may not indicate hyperemia ob-served in the injured spinal cord in the other reports.

Numerous authors demonstrated that reperfusion ofSCBF was of paramount importance in the recovery of spi-nal cord function after spinal cord injuries.24,26,31,35,38,39 Itis possible that this mechanism can also be reproduced byshortening of the spinal cord in Phase 1 and 2, as shown inour study. In our clinical experience, neurologically com-promised patients had a significant recovery after decom-pression and shortening of the spinal column. This wasobserved in patients undergoing total en bloc spondylec-tomy for spinal tumors.6,7 It is possible that 1 of the mech-anisms responsible for this is an increase of SCBF achievedby limited (Phase 1 and 2) shortening.

Conclusion

As shown in our study, 3 morphologic phases occurredduring spinal shortening. Phase 1 occurred within one-thirdof the single vertebral segment and exhibited no deformityof the dura or the spinal cord. Phase 2 involved spinal short-ening of less than two-thirds of the vertebral segment. Nodeformity of the spinal cord was evident; however, shrink-age of the dura occurred. Finally, Phase 3 occurred aftershortening in excess of two-thirds of the vertebral segmentand demonstrated kinking of the spinal cord. The SCBFwas markedly increased in Phase 1 and 2, but was mini-mally decreased in Phase 3 in comparison to the baselineSCBF. Phase 1 includes the safe range of spinal shortening.Phase 2 is safe; however, it is a range in which cautionshould be used. Phase 3 includes the dangerous range. Spi-nal shortening, up to the threshold of spinal cord kinking indogs, caused a meaningful and rapid increase of SCBF.These findings should prove beneficial in various surgicalprocedures employing spinal shortening.

Key Points

● Spinal column shortening after spondylectomyof T13 vertebra was performed in 46 dogs.● During acute spinal shortening, the spinal cordand dura pass through 3 phases: Phase 1, safetyrange: occurred during shortening within one-thirdof the vertebral segment and is characterized by nodeformity of the dural sac neither the spinal cord;Phase 2, warning range: occurred during spinalshortening between one-third and two-thirds of thevertebral segment and is characterized by shrinkingand buckling of the dural sac and no deformity ofthe spinal cord; and Phase 3, dangerous range: oc-curred after shortening in excess of two-thirds ofthe vertebral segment and is characterized by spinalcord deformity and compression by the buckleddura. In this phase, spinal cord-evoked potentialchanges could be detected.● Spinal cord blood flow was markedly increasedin phases 1 and 2.


References

1. Bradford DS, Clifford BT. Vertebral column resection for the treatment ofrigid coronal decompensation. Spine 1997;22:1590–9.

2. Hodgson AR. Correction of fixed spinal curves. J Bone Joint Surg Am 1965;47:1221–7.

3. Leatherman KD, Dickson RA. Two-stage corrective surgery for congenitaldeformities of the spine. J Bone Joint Surg Br 1979;61:324–8.

4. Sharrard WJW, Drennan JC. Osteotomy-excision of the spine for lumbarkyphosis in older children with myelomeningocele. J Bone Joint Surg Br1972;54B:50–60.

5. Shimizu K, Ido K, Fujio K, et al. Total spondylectomy and spinal shorteningfor giant-cell tumour of spine. Lancet 1996;348:342.

6. Tomita K, Kawahara N, Baba H, et al. Total en bloc spondylectomy. Spine1997;22:324–33.

7. Tomita K, Kawahara N, Baba H, et al. Total en bloc spondylectomy forsolitary spinal metastasis. Int Orthop 1994;18:291–8.

8. Smith-Petersen MN, Larson CB, Aufranc OE. Osteotomy of the spine forcorrection of flexion deformity in rheumatoid arthritis. J Bone Joint Surg1945;27:1–11.

9. Gertzbein SD, Harris MB. Wedge osteotomy for the correction of post-traumatic kyphosis. Spine 1992;17:374–9.

10. Heinig JA. Egg shell procedure. In: Luque ER, ed. Segmental Spinal Instru-mentations. Thorofare, NJ: Slack; 1984:221–30.

11. Bauer TW, Muschler GF. Bone graft materials. Clin Orthop 2000;371:10–27.12. Kawahara N, Tomita K, Baba H, et al. Closing-opening wedge osteotomy to

correct angular kyphotic deformity by a single posterior approach. Spine2001;26:391–402.

13. Lehmer SM, Keppler L, Buscup RS, et al. Posterior transvertebral osteotomyfor adult thoracolumbar kyphosis. Spine 1994;19:2060–7.

14. Pedrals JG. Spinal shortening in scoliosis surgery. Spine 1996;21:2515–9.15. Baba H, Kawahara N, Tomita K, et al. Spinal cord evoked potentials in

cervical and thoracic myelopathy. Int Orthop 1993;17:82–6.16. Baba H, Tomita K, Kawahara N, et al. Spinal cord evoked potentials in

thoracic myelopathy with multisegmental vertebral involvement. Spine1992;17:1291–5.

17. Tamaki T, Noguchi T, Takano H, Spinal cord monitoring as a clinical uti-lization of the spinal evoked potential. Clin Orthop 1984;184:58–64.

18. Harakawa I, Yano T, Sakurai T, et al. Measurement of spinal cord bloodflow by an inhalation method and intraarterial injection of hydrogen gas. JVasc Surg 1997;26:623–8.

19. Koshu K, Kamiyama K, Oka N, et al. Measurement of regional blood flowusing hydrogen gas generated by electrolysis. Stroke 1982;13:483–7.

20. Bohlman HH, Bahniuk E, Raskulinecz G, et al. Mechanical factors affectingrecovery from incomplete cervical spinal cord injury: a preliminary report.Johns Hopkins Med J 1979;145:115–25.

21. Tarlov IM, Klinger H. Spinal cord compression studies: II. Time limits forrecovery after acute compression in dogs. Arch Neurol Psychiatr 1954;71:271–90.

22. Kostuik JP. Functional outcome and radiographic correction after spinalosteotomy. Presented at: 10th Annual meeting of Japanese Spinal Instrumen-tation Society; October 26–27, 2001; Tokyo, Japan.

23. Stosseck K, Lubbers DW, Cottin N. Determination of local blood flow byelectrochemically generated hydrogen. Construction and application of themeasuring probe. Pflugers Arch 1974;348:225–38.

24. Carlson GD, Warden KE, Barbeau JM, et al. Viscoelastic relaxation andregional blood flow response to spinal cord compression and decompression.Spine 1997;22:1285–91.

25. Dolan EJ, Transfeldt EE, Tator CH, et al. The effect of distraction on regionalspinal cord blood flow in cats. J Neurosurg 1980;53:756–64.

26. Holtz A, Nystrom B, Gerdin B. Relationship between spinal cord blood flowand functional recovery after blocking weight-induced spinal cord injury inrats. Neurosurgery 1990;26:952–7.

27. Sandler AN, Tator CH. Regional spinal cord blood flow in primates. J Neu-rosurg 1976;45:647–59.

28. Kobrine AI, Doyle TF, Martins AN. Local spinal cord blood flow in exper-imental traumatic myelopathy. J Neurosurg 1975;42:144–9.

29. Lohse DC, Senter HJ, Kauer JS, et al. Spinal cord blood flow in experimentaltransient traumatic paraplegia. J Neurosurg 1980;52:335–45.

30. Nishijima Y, Sasaki M, Yamazaki Y, et al. Variation of spinal cord bloodflow upon stepwise spinal cord impaction injury. J Jpn Orthop Assoc 1984;58:803–12.

31. Young W, Flamm ES. Effect of high-dose corticosteroid therapy on bloodflow, evoked potentials, and extracellular calcium in experimental spinalinjury. J Neurosurg 1982;57:667–73.

32. Sandler AN, Tator CH. Review of the measurement of normal spinal cordblood flow. Brain Res 1976;118:181–98.

33. Toribatake Y. The effect of total en bloc spondylectomy on spinal cordcirculation. J Jpn Orthop Assoc 1993;67:1070–80.

34. Anderson DK, Nicolosi GR, Means ED, et al. Effect of laminectomy onspinal cord blood flow. J Neurosurg 1978;48:232–8.

35. Ohashi T, Morimoto T, Kawata K, et al. Correlation between spinal cordblood flow and arterial diameter following acute spinal cord injury in rats.Acta Neurochir (Wien) 1996;138:322–9.

36. Harada Y. Study on experimental spinal cord injuries by distraction usingspinal cord evoked potentials. J Jpn Orthop Assoc 1983;57:685–701.

37. Nordwall A, Axelgaad J, Harada Y, et al. Spinal cord monitoring using evokedpotentials recorded from feline vertebral bone. Spine 1979;4:486–94.

38. Winter RB. Spine update: neurologic safety in spinal deformity surgery.Spine 1997;22:1527–33.

39. Ducker TB, Salcman M, Lucas JT, et al. Experimental spinal cord trauma, II:blood flow, tissue oxygen, evoked potentials in both paretic and plegic mon-keys. Surg Neurol 1978;10:64–70.

40. Fehlings MG, Tator CH, Linden RD. The relationships among the severity ofspinal cord injury, motor and somatosensory evoked potentials and spinalcord blood flow. Electroencephalogr Clin Neulophisiol 1989;74:241–59.

41. Sandler AN, Tator CH. Effect of spinal cord compression injury on regionalspinal cord blood flow in primates. J Neurosurg 1976;660–76.

42. Senter HJ, Venes JL. Loss of autoregulation and post-traumatic ischemiafollowing experimental spinal cord trauma. J Neurosurg 1979;50:198–206.

43. Wallace MC, Tator CH. Successful improvement of blood pressure, cardiacoutput, and spinal cord blood flow after experimental spinal cord injury.Neurosurgery 1987;20:710–5.


spine volume 30, number 6, pp 613–620 ©2005, lippincott ... · pdf filewas...

Documents