SPINE Volume 28, Number 20S, pp S224–S233©2003, Lippincott Williams & Wilkins, Inc.

Short Segment Bone-on-Bone Instrumentation forSingle Curve Idiopathic Scoliosis

Wolfram Brodner, MD,*� Wai Mun Yue, MD,†� Hans B. Moller, MD, PhD,‡�Kelly J. Hendricks, MD,§ Timothy A. Burd, MD,§ and Robert W. Gaines, MD�

Study Design. Retrospective case series review.Objectives. To evaluate the outcomes of a new short

segment anterior scoliosis technique with complete re-moval of the discs, bone-on-bone apposition of thevertebral bodies, and dual rod instrumentation. Toevaluate a new preop planning technique for scoliosisinstrumentation.

Summary of Background Data. Scoliosis surgery tra-ditionally was performed via a posterior approach, butanterior scoliosis instrumentation has proven to be supe-rior regarding the amount of curve correction and thenumber of segments saved from instrumentation.

Methods. Thirty-one patients with single curve idio-pathic scoliosis less than 75° were operated using thebone-on-bone surgical technique with dual rod instru-mentation (Kaneda Anterior Scoliosis System, Depuy Ac-roMed, Raynham, MA from 1996 until 2001). Averagefollow-up was 40 months (range 15–77 months).

Results. Surgical correction of the major curve aver-aged 73.9% over the instrumented levels and 51.4% overthe entire curve. The average number of discs fused was4.6 for thoracic curves and 3.3 for thoracolumbar andlumbar curves. There were no implant-related complica-tions or nonunions. The compensatory curves spontane-ously improved by an average of 38.6%. Uneventful heal-ing of all fusions occurred—most within 8 to 12 weeks.One compensatory thoracic curve progressed and poste-rior instrumentation was done 28 months after correctionof the major thoracolumbar curve.

Conclusions. Surgical correction was achieved in overhalf the levels that would have been operated by standardposterior segmental fixation. Bony healing due to thebone-on-bone apposition was achieved uneventfully afterapical correction of the spinal curvature in all patients.Use of dual rod instrumentation (Kaneda Anterior Scoli-osis System) is fundamental in maintaining the correctionof the curvature achieved in the operating room. Thepreoperative planning technique worked well. [Keywords: short segment instrumentation, bone-on-bone fu-sion, dual rod instrumentation (KASS), idiopathic scolio-sis] Spine 2003;28:S224–S233

Surgical treatment of idiopathic scoliosis continues toimprove. The Harrington system of spinal implants pro-duced limited internal fixation. Healing times were from6 to 12 months, nonunion was common, and cast/braceimmobilization often had to be repeated after reinstru-mentation. Correction of 40% to 50% of the majorcurve was achieved, and the instrumentation was carriedfrom one level above to two levels below the major curve.

Creation of segmental instrumentation with pediclescrews and sublaminar wires improved the correctionrate to 55% to 65% and shortened healing time to 4 to 6months and some surgeons deleted postoperative castingor bracing. The nonunion rate was reduced but implantfracture or dislodgement still occurred. The length of thefusion was not reduced at all by the introduction of seg-mental instrumentation. Fusions were still carried be-yond the ends of the primary curve, when measured bythe Cobb method.

The Kaneda Anterior Scoliosis System (KASS; DepuyAcroMed, Raynham, MA) was developed by ProfessorKiyoshi Kaneda and introduced in 1996. The use of onestaple and two screws on each vertebral body gave anexperienced surgeon the ability to correct a major curveover only the levels indicated by the Cobb angle of themajor curve.1,2 He had no implant-related complicationsand excellent spontaneous correction of compensatorycurves, without extending the instrumentation beyondthe confines of the major curve.

This paper presents the technique and results achievedby limiting the instrumentation and fusion for idiopathicscoliosis less than 75° to only those vertebrae containedwithin the Cobb angle of the patient’s scoliosis measuredon the “stretch film.” This allowed for instrumentationone to four levels shorter than those used by ProfessorKaneda. The instrumentation included roughly half thelevels that would have been instrumented by traditionalposterior segmental instrumentation.

A new preop planning system is introduced. The re-sults of the first 31 patients operated with this approach,using the dual rod KASS implants, are presented at amean follow-up of 40 months (range 15–77 months).

Materials and Methods

Thirty-one patients were operated by the senior author(R.W.G.) between September 1996 and November 2001.There were 27 female and 4 male patients. At the time of sur-gery, patients’ age ranged from 9 to 43 years with a mean age of18.8 years. Demographic information about the patients is re-corded in Table 1.

From the *Department of Orthopaedics, General Hospital Vienna, Vi-enna, Austria, †Department of Orthopaedic Surgery, Singapore GeneralHospital, Singapore, ‡Department of Orthopaedics, Huddinge UniversityHospital, Huddinge, Sweden, §Department of Orthopaedic Surgery,University of Missouri, Columbia, Missouri, and the �Columbia Or-thopaedic Group, 400 Keene Street, Columbia, Missouri.The device(s)/drug(s) is/are FDA-approved or approved by correspond-ing national agency for this indication. No funds were received insupport of this work. One or more of the author(s) has/have received orwill receive benefits for personal or professional use from a commercialparty related directly or indirectly to the subject of this manuscript:e.g., royalties, stocks, stock options, decision making position.Address correspondence to Robert W. Gaines, MD, Columbia Ortho-paedic Group, 400 Keene Street, Columbia, MO 65201, USA; E-mail:rgaines@cogmds.com

S224

All patients were assessed by detailed history and physicaland neurologic examination, and standing posteroanterior(PA)and lateral radiographs of the spine were taken. Major andcompensatory curves were described by their Cobb angles. Ifthe major curve was in the thoracic spine, then the thoracolum-bar or lumbar curve was considered the compensatory curve,and if the major curve was a thoracolumbar or lumbar curve,then the thoracic curve was considered the compensatorycurve. The tilt angle was defined by the angle between a lineparallel to the endplate of the most tilted vertebra and a hori-zontal line. The tilt angle was measured at T12 or any lumbarvertebra below. Radiographic union was determined by thepresence of trabeculation between adjacent vertebrae. Spinalalignment on lateral films is dependent on the patient’s postureand especially on the position of the arms. As the radiographfilms were reviewed retrospectively and patient positioning wasnot uniform, it did not seem appropriate to report on the totallumbar lordosis or the total thoracic kyphosis. Therefore, werather measured the pre- and postoperative sagittal Cobb an-gles of the instrumented segments. The radiographic measure-ments were done by two orthopedic surgeons (W.B., W.M.Y.),who were not involved in the surgeries of this series.

In patients with any neurologic findings, unusual curve pat-terns, or any symptoms of back or leg pain and/or weakness, amagnetic resonance imaging (MRI) was done. Routine clinicalphotographs were made of each patient before and after theirsurgery.

Curve magnitude in the coronal plane ranged from 40° to72° (mean 51.8°). Patients were selected for anterior short seg-ment instrumentation using this approach when they had amajor curve with a magnitude of 75° or less, and compensatorycurves were 50° or less, and were either more than 50% flexi-ble, or could be reduced to 30° or less on their “stretch film.”Twenty of the surgeries were performed for main thoraciccurves (16 patients with Lenke 1, 2 patients with Lenke 2, and2 patients with Lenke 3 type curves).3 The thoracic instrumen-tations extended from T5-7 to T8-12. Eleven of the surgerieswere performed for curves with the apex at the thoracolumbarjunction. Nine of these patients had a Lenke type 5, and 2patients had a Lenke type 6 curve.3 Their instrumentationsextended from T10-12 to L1-3. One instrumentation was car-ried to L4 in a patient with 6 lumbar vertebrae (Table 2).

The number of vertebrae contained in the major structuralcurves averaged 7.1 vertebrae in the thoracic spine and 5.2vertebrae in the thoracolumbar and lumbar spine. The meannumber of vertebrae, which were initially part of the majorcurve but were not instrumented, was 1.5 for thoracic, and 0.9for thoracolumbar and lumbar curves (see Table 3).

Selection of Fusion Levels. A preoperative plan was madefor the surgery by performing a “stretch film” of the patient inthe office. The patient was laid supine on the radiograph tableand a 36-inch film was placed under the patient. Two otherindividuals well known to the patient were then instructed to“stretch” the patient while the patient was instructed to relax

and enjoy the experience. The three patients who were initiallyapprehensive about the process were trialed two to three timesbefore they were able to produce a relaxed radiograph (Figure 1).

The Cobb angle was measured on the stretch film. The ver-tebrae included in the Cobb angle on the stretch film weredefined as the vertebrae to be fused. A measurement (in mm)was then made from the top edge of the top vertebra to thebottom edge of the bottom end vertebra on the concave side ofthe curve. An identical measurement was then made from thesame vertebrae on the convex side of the curve. The thicknessesof the intervertebral discs were then measured on the concaveand convex sides of the curve, because most of the correctionachieved with this technique occurs by removing the discs thor-oughly at the apex of the curve. The thicknesses of the discswere summed together on the convex and concave sides, andthen subtracted from the longitudinal measurement made onthe concave and convex sides of the curve (Figure 2).

In a straight normal spine, such measurements would beequal; thus, if the subtracted sums were within 5 to 10 mm ofone another, it was assumed that the reconstructed spine wouldapproach “straight” after the discs were removed. If this werethe case, the levels chosen on the stretch film would be used asa preoperative plan for the discectomy and instrumentation. Ifthe difference in the subtracted measurements was more than10 mm, this indicated the need to either add another disc to thepreoperative plan or take off bony wedges from the endplatesto achieve perfect “straight” postoperative alignment of theinstrumented segments.

The preoperative plan created in this way is only a plan. Theactual correction achieved by the total discectomy and instru-mentation using this technique is monitored by the use of theimage intensifier during the operative procedure, because mostcurves can be overcorrected using this technique, and overcor-rection is not a goal of this surgical approach.

Surgical Technique. The patient is securely positioned on apeg-board lateral positioning table for this surgery. Routineexposure is accomplished by techniques for lateral spine expo-sure by many authors, including Kaneda. Extra pleural, retro-peritoneal exposure is performed when the instrumentationextends proximally to T11 or below. Thoracotomy exposure isused when the instrumentation is contained within the thoraciccage. Double thoracotomy has not been used.

The rib heads are then resected back to the level of theintervertebral foramen. Over the vertebrae to be instrumented,the foramen must be carefully visualized to properly place thestaples and remove the intervertebral discs. The segmental

Table 2. Curve Distribution According to the LenkeClassification3

Curve Type Number of Patients

Lenke 1 16Lenke 2 2Lenke 3 2Lenke 4 0Lenke 5 9Lenke 6 2Total 31

Thoracic and lumbar modifiers were determined but are not presented due tothe number of patients in this series.

Table 1. Patient Demographics

Gender (F/M)Mean Age

(Range) Mean Follow-Up (Range)

27/4 18.8 (9–34) years 40 (15–77 mos)

S225Short Segment Bone-on-Bone Instrumentation • Brodner et al

vessels are isolated and doubly clipped before they aretransected. Once all the segmental vessels are carefully cut, theentire spinal column is exposed over the segment to be oper-ated, and the great vessels are carefully identified and protectedwith a wide malleable retractor. The retractor is bent to fit theunderside of the spine that will be instrumented.

The intervertebral discs are then resected over the areaschosen by the preoperative plan. First, the intervertebral fora-men is identified by careful dissection with a curved curette.Then, a Penfield dissector is placed into the spinal canalthrough the intervertebral foramen, and the posterior side ofthe intervertebral disc is palpated. This identifies the posterioredge of the disc, which will be removed. Any epidural bleedingthat happens is controlled by thrombin-soaked Gelfoam, ap-plied through the foramen. Now, the intervertebral disc is re-sected using a Cobb elevator, sharp osteotomes, curettes, androngeurs. The disc space can be opened by using instrumentscarefully placed between the vertebrae. All the disc tissue be-tween the vertebrae must be removed until the two vertebrae,which are adjacent to the involved disc, begin to collapse to-ward one another. The concave most portion of the anulus ismost easily resected using a Kerrison rongeur, and the posteri-or-most portion of the anulus removed with a sharp straightlong handled rongeur. Once the interspace approaches bone-on-bone apposition, Gelfoam is applied to the disc space tocontrol venous oozing and attention turns to the next disc. Thediscectomies proceed from the apical one toward the ends ofthe curve. This pattern facilitates exposure of the more periph-eral discs.

If the curvature, after total disc excision, is not totallystraight, the patient is “repositioned” on the operating table;the anesthesiologist and the circulating nurse gently remove thepegs, which stabilize the patient’s torso. Then, they carefullylift up the patient’s torso, produce gentle traction, and thenreplace the patient and stabilize the trunk with the pegs. Thisprovides an intermediate level of correction of the spinal curve,which facilitates placing the implants, because the spine iscloser to straight after this maneuver for most patients.

Appropriate size vertebral body staples are then selectedand placed in the center of the vertebral bodies over the seg-ment of the spine that will receive implants.

The image intensifier is then used to assess: 1) the quality ofcoronal plane correction of the major curve; 2) the quality ofresponse of the compensatory curves; 3) the positioning of thevertebral body staples; 4) the quality of correction of the “tilt

angle” of the major curve; and 5) the quality of sagittal planecorrection of the primary curve. Any adjustments regardingselection of fusion levels are made at this time.

After fluoroscopy assessment, the staples are adjusted untiltheir position on the vertebra is perfect. As the staples serve asa screw guide, their position on the vertebra is essential forproper screw placement. The depth gauge is then used to esti-mate proper screw length, and all the screws in the constructare placed halfway thru the vertebrae, which will receive them.The c-arm is then used again to assess their position and ad-justments are made if necessary. The screws are then drivenhome, and their position is once again reassessed with the c-arm. Bicortical purchase is essential for this technique, at leaston the end vertebrae in the construct.

Once the screws are properly placed, then the rods areadded onto the screws. The spine is held straight in the coronalplane, and proper position of the spine in the sagittal plane isestimated. The appropriate length rod is then cut to fit and bentinto the sagittal plane necessary at the segment of spine beinginstrumented. Whichever rod fits easiest is applied first. It isdropped into the open screws and secured there with the slidingcaps, which are then crimped into place. The center-most cap isthen secured by tightening the setscrew. Then, the large com-pressor is used to bring the peripheral vertebrae toward theapex. Once bone-on-bone apposition can be palpated with aninstrument between the vertebrae being compressed, the set-screw is tightened, and the procedure moved to the next adja-cent vertebra. The procedure is repeated until there is bone-on-bone throughout the construct or, if the sagittal planecorrection is not perfect, an interbody bone graft or cage isapplied. If this is a cage, the procedure is repeated to achievebone-to-cage correction. If partial bony defects exist afterbone-on-bone apposition is achieved (usually created duringthe discectomy phase of the procedure), they are filled withautogenous rib bone from the exposure rib, or from bone bor-rowed from the vertebral body at the apex of the construct.

Once the first rod has been placed and correction achieved,the c-arm is again used to assess: 1) the quality of apical andentire spine coronal plane correction; 2) the quality of sagittalplane correction both of the apical segment and the entirespine; 3) the quality of correction of the tilt angle of the spine;4) the proper positioning of the spinal implants; and 5) thequality of bone-on-bone apposition between the instrumentedvertebrae.

Table 3. Coronal Curves Magnitude and Tilt Angles Before and After Surgery

ThoracicInstrumentations

(n � 20)%

Corr.

TL/LInstrumentations

(n � 11)%

Corr.

WholeSeries

(n � 31)%

Corr.

Preop major curve 52.1° 51.2° 51.8°Final instrumented curve 16.5° 68.3% 8.1° 84.2% 13.5° 73.9%Final total curve 28.3° 45.7% 19.5° 61.9% 25.2° 51.4%Preop compensatory curve 32.6° 34.8° 33.9°Final compensatory curve 20.8° 36.2% 20.8° 40.2% 20.8° 38.6%Preop tilt angle 23.8° 29.8° 25.9°Final tilt angle 12.6° 47.1% 9.7° 67.4% 11.5° 55.6%

Corr.: correction, TL/L: thoracolumbar/lumbarThe average magnitude of the major and compensatory curves and of the tilt angles before and after surgery is given. The magnitude of the instrumented curveis also shown. If the major curve was in the thoracic spine, then the thoracolumbar or lumbar curve was considered the compensatory curve, and if the majorcurve was a thoracolumbar or lumbar curve, then the thoracic curve was considered the compensatory curve. The tilt angle was measured from the most tiltedvertebra, which was either a lumbar or the 12th thoracic vertebra.Absolute values and the percentage of correction are presented.

S226 Spine • Volume 28 • Number 20S • 2003

Once all these elements are acceptable, then the second rodis placed. No additional correction is achieved with the secondrod. The second rod is cut to length, bent into appropriatesagittal plane alignment, dropped into the open screws, and thecaps are applied. The setscrews are tightened to firmly securethe second rod.

An epidural catheter is placed into the intervertebral fora-men at the top of the construct for postoperative analgesia.4 Ifthe implants are in the chest, they are protected from the lungby either closure of the parietal pleura over them or by a Gore-Tex pericardial patch sewed to the edges of the pleural incision.

Routine muscle and skin closure is then performed. A chestdrain is used if the implants are within the chest. If thoraco-lumbar instrumentation is performed with the extra pleural-retroperitoneal approach, no chest drain or other drain isnecessary.

Results

For the entire series, the apical correction was 73.9%,and the total curve correction was 51.4%. Thoracic andthoracolumbar curves behaved differently.

Thoracic primary curves were improved in the coro-nal plane from an average 52.1° (range 40°–72°) to 28.3°(range 10°–51°). The mean correction rate was 68.3%for the instrumented levels and 45.7% for the totalcurves.

Thoracolumbar and lumbar curves were improvedfrom an average 51.2° (range 40°–70°) to 19.5° (range0°–38°). The mean correction was 84.2% for the instru-mented levels and 61.9% for the total curves (Table 3).

The average number of levels (discs) fused was 4.6 forthoracic curves and 3.3 for thoracolumbar and lumbarcurves.

After surgery, the mean number of vertebrae includedin the operated curves increased by 1.1 for thoracic andby 1.6 for thoracolumbar and lumbar curves because thecurves were less angular (Table 4).

The lumbar compensatory curves improved by an av-erage 36.2% after thoracic instrumentations (from

Figure 1. “Stretch film” setting. The patient is laid supine on theradiograph table and a 36-inch film is placed under the patient.Two persons then “stretch” the patient gently while the patient isinstructed to relax and enjoy the experience.

Figure 2. Preoperative planning. From the Cobb angle on the“stretch film,” a measurement (in mm) is made from the top edgeof the top vertebra to the bottom edge of the bottom end vertebraon the concave side of the curve. An identical measurement ismade from the same vertebrae on the convex side of the curve.The thicknesses of the intervertebral discs were then measuredon the concave and convex sides of the curve. The thicknesses ofthe discs were summed together on the concave and convexsides, and then subtracted from the longitudinal measurementmade on the concave and convex sides of the curve. If thesubtracted sums were within 5 to 10 mm of one another, then itwas assumed that the reconstructed spine would approach“straight” after the discs were removed. If the difference in thesubtracted measurements was more than 10 mm, this indicatedthe need to either add another disc to the preoperative plan ortake off bony wedges from the endplates to achieve “straight”alignment.

Table 4. Number of Vertebrae Included in the ScolioticCurves and Proposed Number of Vertebrae Instrumentedif Posterior Instrumentation Would Have Been Performed

Number ofVertebrae

ThoracicInstrumentations

TL/LInstrumentations

WholeSeries

Preop curve 7.1 5.2 6.4Final curve 8.2 6.8 7.7Instrumented 5.6 4.3 5.1Proposed posterior

instrumentation10.8 9.4 10.3

TL/L: thoracolumbar/lumbarThe number of vertebrae in the preoperative major curve, in the instrumentedcurve, and in the final total curve is given. The data are split into thoracic andthoracolumbar/lumbar instrumentations. Note that after apical curve correc-tion the number of vertebrae included in the curve is increased. This indicatesa smoother or less acute shape of the curve. The number of vertebrae, whichwould have been included into posterior fusions, was about twice the numberof vertebrae included in anterior instrumentations.

S227Short Segment Bone-on-Bone Instrumentation • Brodner et al

32.6°–20.8° mean), and the thoracic compensatorycurves improved by an average 40.2% after thoracolum-bar and lumbar instrumentations (from 34.8°–20.8°mean).

The mean correction of the tilt angle was 47.1%(from 23.8°–12.6°) after thoracic instrumentations and67.4% (from 29.8°–9.7°) after thoracolumbar and lum-bar instrumentations (Table 3 and Figures 3, 4, and 5).

The kyphosis of the instrumented segments increasedfrom an average 14.1° before to 23.2° after thoracic in-strumentations. The kyphosis of the instrumented seg-ments was improved from an average 6.4° before to 3.7°after thoracolumbar and lumbar instrumentations (Ta-bles 5 and 6).

Loss of correction of the instrumented levels betweenthe first and the last postoperative radiograph was 1.1° inthe coronal plane and 1.0° in the sagittal plane.

One hundred twenty-eight interspaces were operatedin this series. Of this total, 121 were closed by bone-on-bone apposition. At seven levels, cages or structural bonegrafts were used to restore or preserve normal sagittalplane profile. The level distribution was: 1 at T12–L1, 2at L1–L2, and 4 at the L2–L3 level. All operated levelswere fused at the latest follow-up (Figure 6).

There were no implant fractures, no dislocated sta-ples, and no broken vertebral bodies.

The time in the operating room averaged 6 hours and18 minutes (range 4 hours and 20 minutes–8 hours and17 minutes). The blood loss averaged 910 mL (range 300mL–2500 mL).

No infections, no vascular problems, and no neuro-logic complications with the spinal cord or cauda equinaoccurred.

The patients were in the intensive care unit (ICU) for 2to 5 days (mean 3.2 days). Their chest tubes were in anaverage of 3.4 days. The epidural catheters were weanedon the second to fourth day and removed on the fifth day.One pneumothorax and one pleural effusion, both re-quiring the replacement of a chest tube, one pneumonia,and one urinary tract infection occurred. Three patientsdeveloped a narcotic ileus. All of these postsurgical com-plications resolved uneventfully.

The patients were ambulated on the second or thirdday and a thoracic lumbar sacral orthosis (TLSO) ap-plied on the fourth or fifth day. The patients wore theTLSO when out of bed generally for 6 to 8 weeks. Theydid not wear it in bed or sleep in it.

The patients were in the hospital for an average of 9days (range 6–14 days).

Four patients had numbness on the anterolateral thighof the exposure side due to retraction of the ilioinguinaland/or genitofemoral nerves during the procedure. Theywere a nuisance-grade annoyance and merely needed re-assurance. Five patients mentioned a “sympathectomyeffect” following the surgery. It was also a nuisance onlyand did not need treatment for any of the patients. Many

of the patients mentioned numbness around the incisionfor several months following the surgery.

The functional rehabilitation of the patients in thisseries proceeded much more promptly than with poste-rior segmental instrumentation patients because of theexcellent internal fixation, excellent apposition affordedby the bone-on-bone apposition, and lack of iatrogenicback muscle violation. All the schoolchildren were backin school by 3 months following the surgery, and all theworking adults were fully employed by 4 months aftersurgery. Most were back to work part-time within 6 to 8weeks of their surgery.

No patients took narcotic medication for more than 7weeks following the surgery, and none required narcoticmedication at final follow-up.

Five adult patients operated for curve correction toalleviate lumbar disc symptoms from disc degenerationbelow the primary thoracic or thoracolumbar curve havehad prompt spontaneous relief, so far, after their pri-mary curve correction, and none has had to have primarytreatment, yet, for his or her primary preoperativecomplaint.

Discussion

The anterior approach for the correction of idiopathicscoliosis has been already shown to provide equal orbetter curve correction,1,2,5–9 less implant problems,more rapid healing, and a lower pseudarthrosis rate thanposterior segmental surgery, even with segmental sys-tems using pedicle screws. Kaneda et al’s reports usingthe KASS system demonstrated these benefits whensurgery was done over levels, which included all thevertebrae contained in the patient’s Cobb angle on thepreoperative standing radiograph. Their recommen-dations were to use the “rod-rotation maneuver” togain correction, to perform only a limited discectomyover the operated motion segments, and to place bonegraft, obtained from the exposure rib, into the discspace.1,2

This patient group demonstrates that the combinationof a more complete diskectomy allows for bone-on-bonecorrection of the apical levels of single curve idiopathicscoliosis. This provides the ability to limit the surgicalprocedure to only those vertebrae contained within themeasured Cobb angle on the patient’s preoperativestretch film. For the curves in the series, which were ini-tially between 40° and 72°, the number of instrumentedvertebrae averaged 5.6 for the thoracic cases and 4.3 forthe thoracolumbar and lumbar cases. This restriction ofoperated vertebrae limited the surgery to half the levelsthat would have been operated with traditional posteriorsegmental instrumentation and pedicle screws. A greaternumber of noninstrumented vertebrae in the lumbarspine will leave the remaining lumbar spine more mobile.By maintaining mobility, and correcting the “tilt angle”

S228 Spine • Volume 28 • Number 20S • 2003

Figure 3. A, 17-year-old femalewith a 56° Lenke type 1AN rightthoracic curve from T5 to L1. Thecurve is reduced to 25° on the“stretch film.” B, 9 months afteranterior KASS instrumentationfrom T6 to T11, the instrumentedsegments were straightened outcompletely (0°) in the coronalplane, and the entire curve mea-sured 18°. The sagittal profile inthe instrumented segments wasimproved from 16° to 25° of ky-phosis. C, Detail of the instru-mented segments 9 months aftersurgery: partial trabeculation atT8 –T9, bony fusion of the otherdisc spaces. D, Clinical view be-fore and 6 weeks after surgery;the previous trunk shift to theright is nicely balanced now. E,Range of motion and trunk bal-ance 6 weeks after surgery.

S229Short Segment Bone-on-Bone Instrumentation • Brodner et al

as well, the degenerative changes with chronic low backpain as seen after Harrington fusions to the lower lumbarspine10 hopefully can be prevented.

Correction rates of anterior instrumentations in idio-pathic scoliosis were reported to be 47%, 58%, and 71%for thoracic curves2,7,9 and 70%, 72%, 83% for thora-

Figure 4. A, 15-year-old girl with a 50° Lenke type 1A-right thoracic curve with the trunk decompensated to the right, the left lumbarcompensatory curve was 33°. The curve is reduced to 15° on the “stretch film,” and the instrumented segments were corrected to 8° aftera T8 –T12 KASS procedure. B, Balanced spine in the coronal plane 14 months after surgery; the main thoracic curve was improved to 20°and the compensatory curve spontaneously corrected to 16°. C, Lateral standing radiographs: the instrumented levels in the sagittal profilemeasured 8° of lordosis before and 15° of kyphosis after surgery. D, Presurgical view from the back. There is trunk decompensation tothe right and a moderate rib hump on the right side. E, Bending forward: the rib hump is clearly visible, when observed from the side. F,The trunk is balanced 18 months postinstrumentation; the minimal residual rib hump is hardly visible.

S230 Spine • Volume 28 • Number 20S • 2003

Figure 5. A, Posteroanterior andlateral preoperative standing ra-diographs of a 13-year-old girlwith a 59° Lenke 5CN left thora-columbar curve. The curve is re-duced to 37° on the “stretchfilm.” There is thoracolumbarjunctional kyphosis of 7° in thelater instrumented segments T12to L3. B, A carbon fiber cage isinserted into the L2–L3 discspace to improve the sagittalalignment. C, 2 months after an-terior instrumentation, the entirethoracolumbar curve is reducedto 13° with the instrumented lev-els T12 to L3 improved to 4°. D,Presurgical trunk decompensa-tion to the left with trunk rotation.E, View 2 months after surgicalcorrection: the trunk is balancedover the pelvis; there is no moretrunk rotation.

S231Short Segment Bone-on-Bone Instrumentation • Brodner et al

columbar and lumbar curves.1,9,11 However, most ofthese authors performed their fusions from end-to-endvertebra of the major structural curve. An average of 1.3vertebrae was saved in our patient series compared to“end-to-end vertebra” instrumentation. Fusion levels re-cently were reported to include 4.3 segments in the tho-racolumbar junction and 6.8 segments in the thoracicspine.12 Compared to this report, an average of 1 verte-bra in thoracolumbar and lumbar surgeries and 2.2 ver-

tebrae in thoracic procedures would have been saved inour series.

Loss of initial curve correction occurs with single rodconstructs to a greater degree than with dual rod con-structs after anterior scoliosis surgery. An average 6° lossof correction in the coronal and 8° in the sagittal planewas reported when using single solid 6.4 mm diameterrods and interbody bone grafting.11 This is far morethan it was described for a dual rod system. Kaneda etal reported 1.2° for the coronal and 1.0° for sagittalplane in the thoracic spine,2 and 1.5° for both coronaland sagittal plane in the thoracolumbar and lumbarspine.1 The correction loss in our patients—1.1° in thecoronal plane and 1.0° in the sagittal plane— confirmsm i n i m a l l o s s o f c o r r e c t i o n f o r d u a l r o dinstrumentations.

One patient, a young man with a 72° primary thora-columbar curve and a structural thoracic curve of 45°,required reoperation for his thoracic curve. This curvepattern was retrospectively classified as a Lenke Type 6curve. After the primary thoracolumbar curve was ini-tially straightened to 23°, the thoracic curve spontane-ously corrected to 34°. Subsequently, his thoracic curveprogressed over the next 25 months to 48° and requiredan additional posterior instrumentation.

Recommendations on the type of athletic activitiesafter scoliosis surgery are controversial.13 However, webelieve that anterior short segment correction offers thebest chance for an unrestricted active sports life, if scoli-osis surgery has to be performed.

We are continuing to follow this group of patientsclosely and will update this report. Even longer-term fol-

Figure 6. Radiologic assessment of fusion after bone-on-bone apposition from T7 to T11 on PA and oblique radiographs. A and B,Trabecular bony bridging at all instrumented levels 2 months after surgery. C and D, The bony fusion is more mature 5 months aftersurgery.

Table 6. Changes of the Sagittal Profile of theInstrumented Levels Before and After Surgery

Number of Patients

Better 20Unchanged 6Worse 5

The sagittal profile was classified as unchanged, when the sagittal alignmentwas in a range of �/� 3 degrees after surgery when compared to beforesurgery, it was classified as better, when the change was �3 degrees to-wards normal, and it was classified as worse, when the change was �3degrees towards abnormal. In the thoracic spine a change towards kyphosiswas regarded as normal and a change toward lordosis was regarded asabnormal, whereas in the lumbar spine the opposite was true.

Table 5. Mean Sagittal Profile of the InstrumentedLevels Before and After Surgery

ThoracicInstrumentations

TL/LInstrumentations

Preop sagittal profile 14.1° kyphosis 6.4° kyphosisFinal sagittal profile 23.2° kyphosis 3.7° kyphosis

TL/L: thoracolumbar/lumbar

S232 Spine • Volume 28 • Number 20S • 2003

low-up will further document the functional benefits forour patients using this short-segment approach.

Key Points

● Short segment bone-on-bone instrumentationpermits 73.9% correction of the apical portion and51.4% correction of idiopathic scoliosis less than75° in adolescents and adults in over half the levelsroutinely performed by posterior segmentalinstrumentation.● Bony healing due to the bone-on-bone apposi-tion was achieved uneventfully during apical cor-rection of the spinal curvature in all patients.● Spontaneous correction of compensatory curvesoccurred by 38.6%.● A new preoperative planning system wasdeveloped.● Bone-on-bone apposition through the apical seg-ments of spinal curvature is principally achieved bythorough discectomy, particularly including theanulus fibrosus, which lies directly anterior to thespinal canal.● Use of dual rod instrumentation (KASS) is fun-damental in maintaining the correction of the cur-vature achieved in the operating room, thoughmost of the achieved correction is accomplishedbefore any spinal implants are applied.

References

1. Kaneda K, Shono Y, Satoh S, et al. New anterior instrumentation for themanagement of thoracolumbar and lumbar scoliosis. Application of theKaneda Two-Rod System. Spine 1996;21:1250–61.

2. Kaneda K, Shono Y, Satoh S, et al. Anterior correction of thoracic scoliosiswith Kaneda anterior spinal system. A preliminary report. Spine 1997;22:1358–68.

3. Lenke LG, Betz RR, Harms J, et al. Adolescent idiopathic scoliosis. A newclassification to determine extent of spinal arthrodesis. J Bone Joint Surg Am2001;83:1169–81.

4. Lowry KJ, Tobias J, Kittle D, et al. Postoperative pain control using epiduralcatheters after anterior spinal fusion for adolescent scoliosis. Spine 2001;26:1290–3.

5. Zielke K. Ventral derotation spondylodesis. Results of treatment of cases ofidiopathic scoliosis. Z Orthop Ihre Grenzgeb 1982;120:320–9.

6. Bernstein RM, Hall JE. Solid rod short segment anterior fusion in thoraco-lumbar scoliosis. J Pediatr Orthop B 1998;7:124–31.

7. Betz RR, Harms J, Clements DH, et al. Comparison of anterior and posteriorinstrumentation for correction of adolescent thoracic scoliosis. Spine 1999;24:225–39.

8. Majd ME, Castro FP, Holt RT. Anterior fusion for idiopathic scoliosis. Spine2000;25:696–702.

9. Sweet FA, Lenke LG, Bridwell KH, et al. Prospective radiographic and clin-ical outcomes and complications of single solid rod instrumented anteriorspinal fusion in adolescent idiopathic scoliosis. Spine 2001;26:1956–65.

10. Conolly PJ, Von Schroeder HP, Johnson GE, et al. Adolescent idiopathicscoliosis. Long-term effect of instrumentation extending to the lumbar spine.J Bone Joint Surg Am 1995;77:1210–16.

11. Ouellet JA, Johnston CE. Effect of grafting technique on the maintenance ofcoronal and sagittal correction in anterior treatment of scoliosis. Spine 2002;27:2129–35.

12. Rhee JM, Bridwell KH, Won DS, et al. Sagittal plane analysis of adolescentidiopathic scoliosis. The effect of anterior versus posterior instrumentation.Spine 2002;27:2350–6.

13. Rubery PT, Bradford DS. Athletic activity after spine surgery in children andadolescents. Results of a survey. Spine 2002;27:423–7.

S233Short Segment Bone-on-Bone Instrumentation • Brodner et al