The Spine Journal 5 (2005) 489–493

ntl

Biomechanical evaluation of kyphoplasty with calcium sulfate cemein a cadaveric osteoporotic vertebral compression fracture mode

Andrew Perry, MDa, Andrew Mahar, MScb, Jennifer Massie, MSa, Noemi Arrieta, BSa,Steven Garfin, MDa, Choll Kim, MD, PhDa,*

aVeterans Administration and University of California, San Diego, 3350 La Jolla Village Drive, San Diego, CA 92161, USAbBiomechanics Laboratory, Children’s Hospital, University of California, San Diego, 3020 Children’s Way, San Diego, CA 92123, USA

Received 20 October 2004; accepted 15 March 2005

Abstract BACKGROUND CONTEXT: Vertebral compression fractures can cause deformity, pain, anddisability. Kyphoplasty involves percutaneous insertion of an inflatable balloon tamp into afractured vertebra followed by injection of polymethylmethacrylate (PMMA) bone cement. PMMAhas several disadvantages such as potential thermal necrosis and monomer toxicity. Calcium sulfatecement (CSC) is nontoxic, osteoconductive, and bioabsorbable.PURPOSE: To evaluate the biomechanical performance of CSC for kyphoplasty in cadavericosteoporotic vertebral bodies.STUDY DESIGN: Destructive biomechanical tests using fresh cadaveric thoracolumbar verte-bral bodies.METHODS: Thirty-three vertebral bodies (T9 to L4) from osteoporotic cadaveric spines weredisarticulated, stripped of soft tissue, and measured for height and volume. Each vertebral bodywas compressed at 0.5 mm/s using a hinged plating system on a materials testing machine to createan anterior wedge fracture and reduce the anterior height by 25%. Pretreatment strength and stiffnesswere measured. Two KyphX inflatable balloon tamps were used to reexpand each vertebral body.After randomization, three groups were created: Group A–no cement; Group B–PMMA; Group C–calcium sulfate cement. Groups B and C were filled with the corresponding cement to 25% of thevertebral body volume. All vertebral bodies were then recompressed by 25% of the post-kyphoplastyanterior height to obtain posttreatment strength and stiffness.RESULTS: Treatment with PMMA restored vertebral strength to 127% of the intact level(4168.2 N�2288.7) and stiffness to 70% of the intact level (810.0 N/mm�380.6). Treatment withCSC restored strength to 108% of the intact level (3429.6 N�2440.7) and stiffness to 46% of theintact level (597.7 N/mm�317.5). CSC and PMMA were not significantly different for strengthrestoration (p�.4). Significantly greater strength restoration was obtained with either PMMA orCSC, compared with the control group (p�.003 and .03, respectively). Stiffness restoration tendedto be greater with PMMA than for CSC, but this difference was not statistically significant (p�.1).Both cements had significantly greater stiffness when compared with the control group (p�.001 andp�.04, respectively).CONCLUSIONS: Use of CSC for kyphoplasty yields similar vertebral body strength and stiffnessas compared with PMMA. It may be a useful alternative bone cement for kyphoplasty. Furtherstudies are required to assess the bioabsorption of CSCs after kyphoplasty in vivo.� 2005Elsevier Inc. All rights reserved.

Keywords: Calcium sulfate cement; Polymethylmethacrylate; Spine biomechanics; Kyphoplasty; Vertebroplasty; Vertebralcompression fracture

rt-

FDA device/drug status: not approved for this indication (MIIGX3,Wright Medical Inc., Arlington, TN).

Nothing of value received from a commercial entity related to thismanuscript.

1529-9430/05/$ – see front matter� 2005 Elsevier Inc. All rights reserved.doi:10.1016/j.spinee.2005.03.011

* Corresponding author. Assistant Professor-Spine Surgery, Depament of Orthopaedic Surgery, University of California, San Diego VAMedical Center, La Jolla, California, 3350 La Jolla Village Drive #112D,San Diego, CA 92161. Tel.: (858) 657-8248; fax: (858) 657-8260.

E-mail address: chollkim@ucsd.edu(C. Kim)

A. Perry et al. / The Spine Journal 5 (2005) 489–493490

lolek

esto-alsale

lo

tyto

ie

).sinlyinl

cae

es

rm

tinno,

sd

e

eeow

onynalurn

G

r

cesy-s

sore

or-

n

xy

y-

ed

rg

reeb-h

.

o

Introduction

Osteoporosis is a systemic disease characterized bybone mass and microarchitectural deterioration of the sketon, leading to bone fragility and increased fracture ris[1]. Approximately 700,000 vertebral compression fracturoccur annually in the United States, mainly secondaryosteoporosis[2,3]. A patient with a single osteoporotic vertebral fracture is at higher risk of sustaining an additionvertebral fracture[4]. Painful osteoporotic vertebral fracturecan have significant effects on quality of life and physicfunction [5]. The mortality rate also increases with thnumber of vertebrae fractured[6]. Chronic pain anddecreased mobility can lead to a depressed mood andof independence.

Vertebral augmentation via vertebroplasty or kyphoplasinvolves the percutaneous injection of bone cement inthe fractured vertebral body[7]. Both vertebroplasty andkyphoplasty have been shown to provide rapid pain relwith relatively low complication rates[8]. The most com-monly used cement is polymethylmethacrylate (PMMAUnfortunately, PMMA has several potential disadvantagePMMA is not bioabsorbable and remains permanentlythe body. Its unreacted monomer is toxic, and its high pomerization temperature has resulted in temperature readas much as 113�C on the anterior cortex of treated vertebrabodies in cadaveric studies[9]. The high compressivestrength and stiffness of PMMA causes a biomechanimismatch between treated and untreated vertebral levwhich may increase the risk of adjacent-level fractur[10].

Recently, attempts have been made to develop bioabsoable bone cements that address these potential probleCements that have undergone ex vivo biomechanical tesinclude a carbonated apatite cement (Norian, CupertiCA), a bioactive cement (Orthocomp; Orthovita, MalvernPA), and a calcium phosphate cement (α-BSM; ETEX Corp.,Cambridge, MA)[11–13]. All have been shown to possescomparable biomechanical profiles as PMMA in simulatecadaveric vertebral compression fractures.

Calcium sulfate cements (CSCs) have been used succfully as bone void fillers for over 100 years[14,15]. Theosteoconductive properties of such cements have bknown for more than 40 years. These cements have lsetting temperatures compatible with living tissue[16]. Theyencourage vascular in-growth and are replaced by host bby creeping substitution[17]. Therefore, CSCs possess manfavorable properties for potential use in vertebral augmetation. Little is known, however, about the biomechanicperformance of these cements for this application. The ppose of the current study was to evaluate the biomechacal profile of a calcium sulfate cement preparation (MIIX3; Wright Medical Inc., Arlington, TN) for kyphoplastyin a simulated osteoporotic vertebral compression fractumodel.

w-

ss

f

.

-gs

lls

b-s.g,

ss-

n

e

-

-i-

e

Materials and methods

Thirty-three vertebral bodies (T9 to L4) were harvestedfrom two male and three female osteoporotic cadaverispines (range of age at death, 73 to 89 years). The spinselected had no radiographic evidence of tumors, gross kphosis, scoliosis, or previous surgery. Each spine wascanned with a Delphi W QDR Series DEXA Scanner(Hologic, Inc., Bedford, MA) to obtain the bone mineraldensity (BMD) of each vertebra. These BMDs were thencompared with the corresponding young normal mean valueand each difference expressed as a standard deviation sc(ie, T score) to estimate the degree of osteoporosis.

Specimens

The vertebrae were stripped of soft tissue and the posterielements removed to facilitate testing. The anterior, posterior, right and left lateral heights of the vertebral bodieswere measured with calipers accurate to 0.01 mm (MattooMTI Corp., Aurora, IL). Molds of the superior and inferiorend plates were made for each vertebral body using an epofiller (Bondo; Bondo Corp., Atlanta, GA) to ensure uniformand even loading during compression testing. Vertebral bodvolumes were calculated using the Archimedean water displacement principle. Volumes obtained in this manner havbeen shown to be in good agreement with that calculateusing vertebral body dimensions obtained from plain films[18,19].

Mechanical testing

Each vertebral body was compressed at 0.5 mm pesecond using a hinged-plated device in a materials testinmachine (MTS, Eden Prairie, MN) to create an anteriowedge-fracture and reduce anterior height by 25% of thintact value. Force (N) and displacement (mm) data werrecorded at 10 Hz. Stiffness data for each specimen were otained from the gradient of the load-displacement grapbetween 500 and 1500 N.

Vertebral augmentation

The vertebral bodies (T9-L4) were randomly assigned toone of three groups: Group A–No cement; Group B–PMMA(Simplex P; Howmedica Inc., Rutherford, NJ); Group C–calcium sulfate cement (MIIG X3; Wright Medical Inc.,Arlington, TN). Kyphoplasty was performed with each bal-loon tamp (KyphX bone tamp; Kyphon Inc., Sunnyvale, CA)inflated to a maximum of 2 cc or until 220 psi was recorded

Groups B and C were then injected bi-pedicularly withtheir respective bone cements within the cancellous void t25% volume fill. For Group B, 40 g of Simplex P powderwas combined with approximately 18 g of barium sulfateto create a powder with 32% barium sulfate by dry weight toenhance visualization of the cement under fluoroscopy[20].PMMA monomer liquid was added to this powder giving a

A. Perry et al. / The Spine Journal 5 (2005) 489–493 491

te

ed

eic%a

rirrie

r

ld

ocos

tr-

tly

in

t

pf

-

tialse

f

t-

of

-rs

e

monomer to powder ratio of 0.83 mL/g. For CSC, adequaradiographic visualization was achieved by mixing thcement according to the manufacturer’s guidelines of cemepreparation without the need for additional barium sulfat

After cement injection, all vertebral bodies were wrappein saline-soaked gauze and stored at 4�C for 24 hours. Theanterior, posterior, and lateral heights of each specimwere remeasured and anteroposterior and lateral radgraphic images were taken to confirm accurate cement plament. All vertebral bodies were then recompressed by 25of the new anterior height to obtain posttreatment data,described previously.

Statistical analysis

Statistical testing was performed using analysis of vaance to compare differences in BMD, T-scores, and vertebbody volumes between groups. A two-way analysis of vaance was used to detect the effect of cement type on pretrment and posttreatment strength and stiffness. Significanwas set as p�.05.

Results

No statistically significant differences in mean BMD oT-scores were found between any of the pretreatment grou(p�.05). In addition, no significant differences in vertebrabody volume or anterior vertebral body heights were founThe average T-score of the vertebral bodies was�2.7 (range�1.5 to�4.5). The average volume of cement injected inteach vertebral body was 7.5 cc (range 2.5 cc to 11.75 cThere was no significant difference in the mean volumePMMA or CSC that was injected into either of these groupThe average BMD of these specimens was 0.75 g/cm3 (range0.56 g/cm3 to 0.90 g/cm3). As expected, mean pretreatmenstrengths of groups A, B, and C were not significantly diffeent from each other (3152.9 N�1176.2, 3475.1 N�1559.4,and 3239.4 N�1204.6, respectively). Similarly, pretreatmenstiffness for groups A, B, and C were also not significantdifferent (1188.3 N�369.1, 1176.0 N�321.0, and1478.9N�538.4, respectively). CSC showed similar fillingcharacteristics to that in clinical practice with PMMA(Fig. 1).

e

nt.

no-e-

s

-al-at-ce

ps

.

).f.

Posttreatment strength and stiffness data are shownFigures 2 and 3. Treatment with PMMA restored vertebralstrength to 127% of the intact level (4168.2 N�288.1)and stiffness to 70% of the intact level (810.1 N/mm�380.5).Treatment with CSC restored strength to 108% of the intaclevel (3429.7 N�2440.3) and stiffness to 46% of theintact level (598.7 N/mm�318.0). There was no statisticallysignificant difference in strength between groups B and C(p�.4). Both PMMA and CSC provided significantly higherstrength restoration than bone tamp inflation only (GrouA), (p�.003 and .03, respectively). The average stiffness oGroup B (PMMA) was higher than Group C (CSC), but wasnot statistically significant (p�.1). Both PMMA and CSCprovided significantly greater stiffness when either was compared with Group A (p�.001 and p�.04, respectively).

Discussion

Although, vertebroplasty and kyphoplasty with PMMAare effective in providing pain relief of vertebral compressionfractures, there have been theoretic concerns about potenadverse effects with the use of such acrylic cements. Theconcerns include the possibility of thermal injury to sur-rounding tissue and toxicity from the unreacted monomer[9].The potential for adjacent level fractures as a result oPMMA’s exceedingly high compressive strength andstiffness profile is also of concern. The ideal bone cemenshould be biodegradable, nontoxic, have a low setting temperature, and have a biomechanical profile closer to thathuman bone[10].

In the current study, the biomechanical properties of vertebral bodies treated with kyphoplasty using either CSC oPMMA were compared. Molloy et al. used standard amountof PMMA cement (2, 4, 6, and 8 mL) for either thoracic(T6-T10), thoracolumbar (T11-L1), or lumbar (L2-L5) verte-brae during vertebroplasty[21]. Linear regression analysisof their results found a weak relation between percentagfill and restored strength and stiffness (r2�.21, r2�.27respectively). Full strength restoration was obtained withroughly a 20% fill and full stiffness restoration with a 30%

Fig. 1. Radiograph of L1 vertebral body with calcium sulfate cement. (a) axial view; (b) lateral view.

A. Perry et al. / The Spine Journal 5 (2005) 489–493492

0

500

1000

1500

2000

2500

3000

3500

4000

4500

5000

N

Pre Treatment

Post Treatment

Calcium sulfate PMMANo Cement

*NSNS

NSNS

*

Fig. 2. Failure strength of pre- and post-treatment of the No cement, Calcium sulfate and PMMA groups.* Indicates a statistically significant difference, p�0.05.

e

gaa

g

o

t

fill. In the study reported here, each vertebral body was filleto 25% of its measured volume during kyphoplasty. Thresults with PMMA were consistent to that of the aforementioned study. As such, our average posttreatment strenwith PMMA was higher than the pretreatment level, whereposttreatment stiffness was lower at this level of vertebrbody filling.

Calcium sulfate cement has a much lower native compresive strength and stiffness than that of PMMA. Comparin

d

-thsl

s-

both cements, the strength restoration with PMMA (127%)was slightly higher than with CSC (108%), although thiswas not statistically significant (p�.4). Similarly, the stiff-ness restoration with PMMA (70%) was higher than withCSC (40%), although this did not reach statistical signifi-cance (p�.1). The same results have been obtained withother cements, eg, calcium phosphate cements, which alshave much lower native compressive properties than PMMA[22]. Perhaps, the native compressive property of a cemen

0

200

400

600

800

1000

1200

1400

1600

1800

N/mm

No Cement PMMACalcium Sulfate

NS

NS

NS

NS

*

*

Pre Treatment

Post Treatment

Fig. 3. Stiffness of pre- and post-treatment of No cement, Calcium sulfate and PMMA groups.* Indicates statistically significant difference, p�0.05.

A. Perry et al. / The Spine Journal 5 (2005) 489–493 493

dr-igrtt

thoinoung

rshn

inthivc

y-Sne

th

au

n;8

-aom

nte-

te--es.

a-ine

rale

.gve

-5:

e,cticrte-

ial

ics

in

,for5;

e:

:

fic.

rte-

l4;

,ticffi-

may not always predict its performance in a vertebral boduring kyphoplasty. During cement injection, cement intedigitates into the surrounding cancellous bone. This interditation may be an important aspect of biomechanical suppoHeini et al. found that the posttreatment vertebroplasstrength and stiffness restoration is inversely proportionalthe degree of osteoporosis[23]. In their study, they founda weak correlation (r2�.131) between percentage strengrestoration and degree of osteoporosis (T-scores). Withteoporosis there is loss in trabecular bone leading tocreased intertrabecular spacing. This increased spacing callow greater cement distribution, that is, interdigitatiowithin the cancellous bone of the vertebral body durinkyphoplasty, leading to greater cement effect.

Conclusion

Calcium sulfate cement has comparable strength restotion when compared with PMMA for kyphoplasty. Stiffnesrestoration is lower when compared with PMMA, whicmay be more desirable in reducing the incident of adjacelevel fractures. CSC is bioabsorbable and has a low setttemperature, making it suitable for the addition of growfactors and heat-labile bioactive substances. Further in vstudies are needed to assess the reabsorption and interaof CSC with host bone in osteoporotic animal models.

References

[1] Riis BJ. Biochemical markers of bone turnover. II: Diagnosis, prophlaxis, and treatment of osteoporosis. Am J Med 1993;95:17S–21

[2] Melton LJ 3rd. How many women have osteoporosis now? J BoMiner Res 1995;10:175–7.

[3] Cohen LD. Fractures of the osteoporotic spine. Orthop Clin NorAm 1990;21:143–50.

[4] Kado DM, Browner WS, Palermo L, Nevitt MC, Genant HK,Cummings SR. Vertebral fractures and mortality in older women:prospective study. Study of Osteoporotic Fractures Research GroArch Intern Med 1999;159:1215–20.

[5] Schlaich C, Minne HW, Bruckner T, et al. Reduced pulmonary functioin patients with spinal osteoporotic fractures. Osteoporos Int 1998261–267.

[6] Silverman SL, Minshall ME, Shen W, Harper KD, Xie S. The relationship of health-related quality of life to prevalent and incident vertebrfractures in postmenopausal women with osteoporosis: results fr

y

-t.yo

s--ld

a-

tg

otion

.

p.

:

l

the Multiple Outcomes of Raloxifene Evaluation Study. ArthritisRheum 2001;44:2611–9.

[7] Galibert P, Deramond H, Rosat P, Le Gars D. [Preliminary note othe treatment of vertebral angioma by percutaneous acrylic verbroplasty]. Neurochirurgie 1987;33:166–8.

[8] Heini PF, Walchli B, Berlemann U. Percutaneous transpedicular verbroplasty with PMMA: operative technique and early results. A prospective study for the treatment of osteoporotic compression fracturEur Spine J 2000;9:445–50.

[9] Belkoff SM, Molloy S. Temperature measurement during polymeriztion of polymethylmethacrylate cement used for vertebroplasty. Sp2003;28:1555–9.

[10] Berlemann U, Ferguson SJ, Nolte LP, Heini PF. Adjacent vertebfailure after vertebroplasty. A biomechanical investigation. J BonJoint Surg [Br] 2002;84:748–52.

[11] Schildhauer TA, Bennett AP, Wright TM, Lane JM, O’Leary PFIntravertebral body reconstruction with an injectable in situ-settincarbonated apatite: biomechanical evaluation of a minimally invasitechnique. J Orthop Res 1999;17:67–72.

[12] Belkoff SM, Mathis JM, Erbe EM, Fenton DC. Biomechanical evaluation of a new bone cement for use in vertebroplasty. Spine 2000;21061–4.

[13] Bai B, Jazrawi LM, Kummer FJ, Spivak JM. The use of an injectablbiodegradable calcium phosphate bone substitute for the prophylaaugmentation of osteoporotic vertebrae and the management of vebral compression fractures. Spine 1999;24:1521–6.

[14] Watson JT. The use of an injectable bone graft substitute in tibmetaphyseal fractures. Orthopedics 2004;27:s103–7.

[15] Kelly CM, Wilkins RM. Treatment of benign bone lesions with aninjectable calcium sulfate-based bone graft substitute. Orthoped2004;27:s131–5.

[16] Peltier LF. The use of plaster of Paris to fill defects in bone. ClOrthop 1961;21:1–31.

[17] Hadjipavlou AG, Simmons JW, Yang J, Nicodemus CL, Esch OSimmons DJ. Plaster of Paris as an osteoconductive materialinterbody vertebral fusion in mature sheep. Spine 2000;25:10–discussion 16.

[18] Panjabi MM, Takata K, Goel V, et al. Thoracic human vertebraquantitative three-dimensional anatomy. Spine 1991;16:888–901.

[19] Panjabi MM, Goel V, Oxland T, et al. Human lumbar vertebraequantitative three-dimensional anatomy. Spine 1992;17:299–306.

[20] Jasper LE, Deramond H, Mathis JM, Belkoff SM. The effect omonomer-to-powder ratio on the material properties of cranioplastBone 1999;25:27S–9S.

[21] Molloy S, Mathis JM, Belkoff SM. The effect of vertebral bodypercentage fill on mechanical behavior during percutaneous vebroplasty. Spine 2003;28:1549–54.

[22] Tomita S, Molloy S, Jasper LE, Abe M, Belkoff SM. Biomechanicacomparison of kyphoplasty with different bone cements. Spine 20029:1203–7.

[23] Heini PF, Berlemann U, Kaufmann M, Lippuner K, Fankhauser Cvan Landuyt P. Augmentation of mechanical properties in osteoporovertebral bones: a biomechanical investigation of vertebroplasty ecacy with different bone cements. Eur Spine J 2001;10:164–71.