lourrial of Orthopaedic Xesearch lf~715-719 The Journal of Bone and Joint Surgery. Inc. 0 1998 Orthopaedic Rescarch Society

In Vitro Strength Comparison of Hydroxyapatite Cement and Polymethylmethacrylate

in Subchondral Defects in Caprine Femora

Kevin Crawford, *B. Hudson Berrey, ?William A. Pierce, and ?Robert D. Welch

Drjiurtment of Orthopczedrc Siirgery, Unzvrrs~ry of T e m ~ Southwestrun Medlcal Center, Dallas, Texas, *Departrneiit of Orthopaedic Surgrry, University of Florida, Gainesville, Florlda, and

TDepiirtmerit of Ressart h, Texizs Scottish Rite IloJpltnl for Children, D a l l i ~ , TexaA, U.S.A

Summary: Hydroxyapatite cement was investigated in sitzi for the reconstruction of juxta-articular defects. Polymethylmethacrylate is currently the most cominonly used material for the reconstruction of bone defects following the exteriorization and curettage of aggressive benign tumors. In vitro, we compared the effects of hydroxyapatite cement and polymethylmethacrylate in restoring the stiflness of the subchondral plate in a caprine femoral defect model. Ten matched pairs of caprille femora underwent nondestructive compression testing normal to the load-bearing surface. A standardized subchondral defect 12 mm in diameter was created in the medial femoral condyle. Compression testing was repeated to dctermine the reduction in stiffness caused by the defect. Each femur from each pair was randomly assigned to one of two groups (n = 9), and the defects were augmented with either polymethylmethacrylate or hydroxyapatite cement. After 12 hours, compression testing was repeated to determine the subchondral stiffness after augmentation. Compared with intact femora, the defect specimens that were later treated with either polymethylmethacrylate or hydroxy- apatite cement exhibited stiffness values of 70 (386 -+ 107 Nimm) and 59'/0 (343 5 94 Nimm) respectively, which represented a significant reduction in stiffness (p = 0.05). Augmentation with polymethylmethacrylate or hydroxyapatite cement rcstorcd stiffness by 81 (450 2 111 N/mm) and 71% (413 ? 115 Nimm), respec- tively, of the values of intact specimens. Hydroxyapatite cement restored stiffness significantly (p = 0.05) over the stiffness of the nonaugmented defect compared with the stiffness after augmentation with polymethyl- methacrylate (p = 0.1 2). Neither polymethylmethacrylate nor hydroxyapatite cement restored stiffness to that of intact femora (p = 0.05). In the current defect model, hydroxyapatite cement was comparable with poly- methylmethacrylate in restoring subchondral stiffness. Unlike polymethylmethacrylate, however, hydroxy- apatite cement has the following advantages: it is osteoconductive, is replaced by host bone, and avoids the potential for thermal necrosis. Hydroxyapatite cement may therefore provide a viable alternative to poly- methylmethacrylate for augmentation of juxta-articular and other bone defects.

The repair of juxta-articular defects created either in the treatment of musculoskeletal neoplasms or sec- ondary to trauma continues to be a challenging prob- lem facing orthopaedists. Numerous materials used to augment osseous defects have been proposed. but none have proven ideal (2,6,10,13). Conventional means of reconstructing such defects have typically utilized autogenous bone graft. However, the use of bone graft is associated with donor-site morbidity, increased operative time and blood loss, and lim-

Received May 5,1998; accepted August 14,1998. Address correspondence and reprint requcsts to R. D. Welch

at Department of Research, Tcxas Scottish Rite Hosoilal for Chil-

ited availability (2.6,20). Allograft has also been used but introduces the risk of disease transmission and immunological response (1). In addition, the pack- ing of skeletal defects with autogcnous or allograft bone provides only a marginal mechanical advantage

Both ceramics and polymers have been proposed as a means of reconstruction of bone defects that are considered at high risk of fracture (2,6,13). Histori- cally, plastcr of Paris has been utilized, and more re- cen tl y, polymethylmethacrylate has found widespread application (6,7,9,12). Disadvantages of polymethyl- methacrylate includc its highly exothermic polymer- ization reaction. which can produce thermal necrosis

(2,11) .

drcn,'2222 Welborn Street, Dallas, TX 75219, U:S.A. E-mail: rdwelch@airmail.net

Presented in part at the 43rd Annual Meeting of the Ortho- paedic Research Society, Sail Francisco, California, IJ.S.A., Fcb-

of adjacent tissues. Because polymethylmethacrylate is not biodegradable, it also lacks the capability to biomechanically integrate into the surrounding host

ruary 9, 1997. bone (6,7,9,12-14).

715

716 K. CRAWFORD ET AL.

More recently, products of hydroxyapatite have been considered as possible alternatives for the repair of metaphyseal defects because they are biocompat- able and appear to integrate into host bone slowly through cellular resorption and remodeling mecha- nisms (2,6,10). Coralline hydroxyapatite and natural coral calcium carbonate are available in both block and granular forms (2,10,15,17,19). me porous nature of products derived from coral favors osteoconduc- tion; however, it also compromises mechanical intcg- rity (1 0,15,17,19). The block form may require tedious shaping by hand or power tools to precisely fit or fill a complex defect. The granular form facilitates pack- ing of the defect, although it can be difficult to retain the granules there (2,15,17), but it lacks sufficient me- chanical strength to be used in load-bearing subchon- dral regions.

Recently, calcium phosphate cements have been developed in situ that convert isothermally into micro- porous hydroxyapatite in vivo (4,5,8,16). These prod- ucts consist of a mixture of calcium and phosphate salts that, when an aqueous solution is added, under- goes a dissolution-reprecipitation reaction that pro- duces hydroxyapatite cement. Within minutes, the mixture forms into a moldable paste that can be easily packed into defects. The material continues to harden over several minutes with complete conversion to hy- droxyapatite in vitro within 4 hours (4,5,8,16,18).

One of these products, Bone Source (OrthofixiOs- teogenics, Richardson, TX, U.S.A.), is currently being used in craniofacial reconstruction and is under clini- cal investigation for traumatic bone defects (5). An- other form of calcium phosphate cement, Norian SRS (Norian, Cupertino, CA, U.S.A.), is being investigated for the repair of Colles fractures (4) and the augmen- tation of the bone screw (16). Frankenburg et al. re- ported that the Norian SRS material, FractureGrout, has the potential to maintain mechanical integrity dur- ing its incorporation in experimental metaphyseal de- fects (8). In another report, Pierce and Welch found the in vitro compressive strength of Bone Source to be comparable with that of cancellous bone, although this material is relatively weak in tension (18).

These studies suggest that calcium phosphate ce- ment may have potential application in the recon- struction of juxta-articular defects. We compared the abilities of Bone Source and polymethylmethacrylate to restore the stiffness of the subchondral plate in a fcmoral defect model.

MATERIALS AND METHODS Ten matched pairs of skeletally mature caprine femora, wrapped

in saline-soaked gauze and stored at -2O"C, were allowed lo thaw to room temperature. The proximal portion of each femur was re- moved,leaving an 11-cm specimcn containing the femoral diaphysis and thc distal femoral condyles. The diaphysis was potted in a poly- methylincthacrylale block. Nondestructive compressive testing was

L

i P i 1 LOAD CELL 1

FIG L Schematic representation of the apparatus used for the compression testing of the medial condyle of the cdprine femora. LVDT = linear voltage displacement transformer.

performed on the normal load-bearing surfaces of the medial con- dyle of 18 femora (nine pairs) with a servohydraulic testing system (858 Bionix; MTS Systems, Minneapolis. MN. U.S.A.) operated un- der displacement control at a rate of 0.3 mdsec. The specimens were secured through a pivoting vise that provided three degrees of freedom in aligning the medial femoral condyle (Fig, 1). Com- pression was produced with a 12.7-mm-diamcter cylindrical inden- tor attached to the actuator of the testing machine. Thc diameter of the indentor was chosen on the basis of the average width of the caprine medial femoral condyle being 15-1 8 mm. For each compres- sion test, a total of three force-displacement curves were generated.

Two specimens (one pair) were loaded under compression to failure to determine the elastic limit of the medial femoral condylc. A load limit of 222 N was found to be well within the linear elastic region of the force-displacement curve of these caprine femora under the applied compression (permanent deformation began at 267 and 356 N in thc two femora).

A subchondral defect was then created in the medial femoral condyle with a 12-mm auger-tip drill bit centered over the attach- ment of the medial collateral ligament. A high-speed burr was used to remove any remaining trabecular bone from the subchondral platc and lateral cortex of the condyle. All defects were made by a single investigator (K.C.). After crcation of the defects, the thick- ness of the subchondral plate and the depth of the defect werc documented. Each femora was again subjected to mechanical test- ing with the same protocol that was used for the intact specimens to cstablish the reduction in subchondral stillness caused by thc defcct. One femur from each pair was then randomly assigned to receive augmentation of the defect with polymethylmethacrylate (Simplcx; Howmedica, Rutherford. NJ, U.S.A.), and the opposite femur received augmentation with hydroxyapatite cement (Bone Source). Both materials were prepared according to the instructions provided by the manufacturer. The chemical composition of Bone Source is tetracalcium phosphate (CaJ[P0&O) and dicalcium phos- phate (CaHPOJ. which, when mixed with water, converts in situ to (Ca[PO,],OH) (5). The Bone Source was prcparcd with saline so- lution (2.5 mmil0 g of hydroxyapatite cement) as the mixing soh- tion. Defects were packcd by hand and then tampcd with use of a rod of a diameter similar to that of the defect. After augmentation of the defects, all femora were wrapped in gauze soakcd in saline solution and allowed to cure for 12 hours at room temperature. The stiffness of the repaired specimens was eslablished by compressive

J Orthop Res, Vol. 16, No. 6, 1998

1000

900

800

700

600

u) 500

400

300

200

100

0

z

t u) al S

v)

HYDROXYAPATITE CEMENT IN F E M O R A L DEFECTS

[ * (p = 0.05)

* (p = 0.05) I

717

PMMA (n = 9) HAC (n = 9)

FIG. 2. The mean compressive stiffness of intact caprine medial femoral condyles, the stiffness following the creation of a 12-mm subchon- dral defect, and the stiffness following augmentation of the defect with either polynicthylmethacrylate (PMMA) or hydroxyapatite cement (HAC). * 7 significantly different from the dcfcct and the augmented deCect, -r = not significantly different from the defect augmented with polymethylmethacrylate. and = significantly different from the defect augmcntcd with hydroxyapatite cement.

testing according to the protocol described previously. Values of subchondral stiffness of intact, defect, and augmented-

defect femora were statistically compared with use of a paired Student t test. The stiffness values of the femora augmented with hydroxyapatite cement and polymethylmethacrylate were also compared with use of a paired Student 1 test. The significance level was set at p 5 0.05.

RESULTS All 18 caprine femora were successfully subjected

to nondestructive compressive testing of the mcdial femoral condyle while intact, after creation of the de- fect, and following augmentation of the defect with either polymelhylmethacrylate or hydroxyapatite ce- ment (Fig. 2). The mean measured thickness of the articular cartilage and subchondral plate after burring was 3.04 5 0.87 mm for both groups. Although the thickness of the plate tended to be slightly less in the group treated with hydroxyapatite cement, the thick- ness was not significantly different from that of the specimens treated with polymethylmethacrylate (p = 0.823). The mean depth of the defects for both groups was 1.8 ? 0.3 cm.

During compression testing, a motion artifact was noted in the first stiffness curve generated with each specimen and was found to be attributable to the set- tling of the sample within the testing jig. These runs were discarded, and only the two subsequent curves per specimcn were used in the data analysis. Thc mean stiffness of the subchondral plate of the intact speci- mens was 553 & 160 and 584 -C 226 Nimm for thc groups later treated with polymethylmethacrylate or hydroxyapatite cement, respectively. Compared with

the intact femora, the defect specimens exhibited stiffness values of 70 (386 2 107 N i m m ) and 59% (343 2 94 Nimm) (p = 0.05) before treatment with polymethylmethacrylate or hydroxyapatite cement, respectively. Augmentation with polymethylmethac- rylate restored the subchondral stiffness to 81 YO (450 % 111 N/mm) of the values of the intact femora, which was not significantly different from the stiffness of the defect (p = 0.1 2). Hydroxyapatite cement restored stiff- ness to 71% (413 +- 115 N/mm) of the values of the intact specimens; this was significantly different from the stiffness of the unfilled defect (p = 0.05). The stiff- ness of the defect that was augmented with either polymethylmethacrylate or hydroxyapatite cement was still significantly less than that of the intact femora (p = 0.05).

DISCUSSION The ideal substrate for the repair of juxta-articular

defects has yet to be identified. There have been nu- merous reports in the literature describing the use of many different materials, each with advantages and disadvantages, but most have not found wide- spread application. One exception is polymethyl- methacrylate, which is often used in musculoskeletal oncology to fill defects (7,9,13,14). However, poly- methylmethacrylate has several potential disadvan- tages, including eliciting an inflammatory response and fibrous encapsulation of the material. In addition, polymethylmethacrylate is highly exothermic during polymerization and may cause thermal necrosis of sur- rounding tissues (13J4).

J Orthop Rex Vol. 16, No. 6, I998

71 8 K. CRAWFORD ET AL.

More recently, preparations of calcium phosphate have been considered as possible alternatives for re- constructing bone defects in the subchondral region. One such compound is hydroxyapatite derived from coral, but despite excellent tissue compatibility, its usefulness has been limited by less than desirable structural and physical properties (2,10,15,17). Prepa- rations of hydroxyapatite derived from coral are brit- tle and undergo biodegradation at a very slow rate. if at all (2,15,17). Preparations of natural coral calcium carbonate are reported to biodegrade at a relatively rapid rate but are also brittle (19).

The hydroxyapatite cement investigated in the cur- rent study is composed of tetracalcium phosphate and anhydrous dicalcium phosphate that undergo an iso- thermic dissolution-reprccipitation in the presence of water to form hydroxyapatite (5) . During the setting phase, the cement may be sculptured to fit complicated surgical defects as necessary and becomes fully crys- tallized to hydroxyapatite within 4 hours (5,18). The resulting compound is osteoconductive and undergoes remodeling and replacement with host bone (5 ) .

Other calcium-phosphate cements have been dem- onstrated experimentally to maintain the mechanical integrity of metaphyseal tibia1 defects during incor- poration into host bone (8). Kecently,Pierce and Welch reported that mixing solutions and environmental con- ditions may influence the final compressive and tensile strength of hydroxyapatite cement in vitro (18). De- ionized water and saline solution were considered thc optimal mixing solutions in that report. For this reason, saline solution was chosen for the current study. The acute compressive strength of hydroxyapatite cement in v i m has been shown to be comparable with that of cancellous bone (26.2-33.75 MPa) (3); however, it is relatively weak in tension (1.09-1.21 MPa) (18). These studies suggest that hydroxyapatite cement may have potential for the reconstruction of juxta-articular dc- fects that require support of the subchondral plate.

In a previous report, Frassica et al. compared poly- methylmethacrylate with autologous bone graft in a canine femoral defect model (9). Polymethylmethac- rylate was found to significantly restore subchondral stiffness to 98% of the values of the intact femora. We found that polymcthylmethacrylate restored stiffness to 81 96, but stiffness remained significantly less than that of intact femora. The size of the defect, which was considerably larger in the canine model, likely contributed to the difference between the studies. In addition, the defects created in the current study ex- tended to within 3-4 mm of the articular surface, whereas in the report by Frassica et al., the defects were 6-9 mm beneath the articular surface.

A potential limitation to using a defect model is the possible difficulty in creating a standardized and rcpro- ducible defect. We attempted to minimize variability

by having a single investigator create all defects. The uniformity of our compression results suggests we were SUCCessfU~ in creating a consistent defect model. An- other potential limitation may be that each specimen was used as its own control and, after successive com- pressions. may have influenced the forcc-displacement properties. To prevent any plastic deformation, we compresscd to a load limit of 222 N, which is well within the linearly elastic portion of the force-displacement curve for these femora. The current study evaluated only static compressive testing of augmented femora. Fatigue testing may have been a more appropriate method of analysis: however. it is not without potential limitations. During the curing phase for the hydroxy- apatite cement, the femora were stored at room tem- perature; this does not accurately represent the in vivo situation. However, Pierce and Welch reported that curing hydroxyapatite cement in a water bath at 37°C and 97% relative humidity actually increased its com- pressive strength compared with those specimens cured at room temperature and relative humidity (18).

In the current study, defects augmented with hy- droxyapatite cement were found to be significantly stiffer than the nonaugmented controls, whereas the augmentation with polymethylmethacrylate did not result in a significant increase in stiffness compared with the nonaugmented controls. Both materials failed to return the stiffness to the values of the intact fem- ora. Although there was no significant difference be- tween the two groups regarding the final thickness of the subchondral plate, there was a trend toward a slightly thinner plate (i.e., a more severe defect) in the specimens treated with hydroxyapatite cement.

Hydroxyapatite cement may provide a valuable al- ternative for the reconstruction of bone defects. The isothermic setting reaction that occurs at or near phys- iological pH prevents the possibility of thermal injury to adjacent tissue. The ease of preparation and surgical handling compares favorably with polymethylmethac- rylate. The osteoconductive properties of hydroxyapa- tite cement and the slow replacement with host bone are additional advantages.

Acknowledgment: This study was supported by the Rcscarch Fund, Texas Scottish Rite Hospilal lor Children, Dallas, Texas. The Bone Source was graciously providcd by OrrhofixiOsteogenics. Richardson, Texas, U.S.A. The authors are grateful h r the stalisti- cal analysis conducted by Dr. Richard Browne and for the assis- tance provided by Dwight Bronson. Mark Zobitz, Stacy Carpenter, and Tracy Wassell.

REFERENCES 1. Buck BE, Malinin TT. Brown MD: Bone transplantation and

human immunodeficiency virus: an estimate of risk of ac- quired immunodcficiency syndrome (AIDS). Clin Orthap 240:129-136, 1989

2. Bucholz RW, Carlton A, Holmes RE: Hydroxyapatite and tri- calcium phosphate bone graft substitutes. Orfhop Clin Nordr Am 18:323-334,1987

J Ortliop Res, Vol. 16, No. 4 1999

HYDROXYAPATITE CEMENT I N FEMORAL DEFECTS 719

3. Carter DR. Hayes W C The compressive behavior of bone as a two-phase porous structure. J R o w Joint Surg [Am/ 59:954- 962. 1977

4. Constanz RR, Ison IC, Fulmer MT, Poser RD. Smith ST, Van- Wagoner M, Ross J. Goldstein SA, Jupiter JB, Rosenthal DI: Skeletal repair by in situ formation of the mineral phase of bone. Science 267:1796-1799, 1995

5. Costantino PD, Fricdman CD, Joncs K, Chow LC, Sissoii GA: Experimental hydroxyapatite cement cranioplasty. Plast Re- constr Surg 90:174-185, 1992

6. Uaniien CJ, Parsons JR: Bone graft and bone graft substi- lules: a review of current technology and applications. JAppI

7. Enderle A, Willerl IIG, Zichner L: Morphology of the tissue surrounding a tcmporary bonc ccmcnt plug. In: Limb Sdvrzge in Musculvskeletul Oncology, pp 470-479. Ed by WF Ennek- ing. New York, Churchill Livingstone, 1987

8. Frankenburg EP, Bauer TW. Jiang M. Poser RD. Coldstein SA: Mechanical integrity of calcium phosphate cement in in-vivo metaphyseal models over time. Truns Orthop Res Soc 21:224,1996

9. Frassica FJ, Gorski JP. Pritchard DJ, Sim FH. Chao EY A comparative analysis of subchondral replacement with poly- methylmethacrylate or autologous bonc graft in dogs. Clin Orthop 293:378-390. 1993

10. Holmes RE, Bucholz RW. Mooncy V: Porous hydroxyapatite as a bone-graft substitute in metaphyseal defects.JBone Joint

11. Hopp SG, Dahners LE; Gilbert JA: A study of the mechanical strength of long bone defects treated with various bone au- tograft substitutes: an experimental investigation in thc rab- bit. J Orthop Res 7579.584. 1989

BioMLater 2:187-208, 1991

S U ~ R /Am] 68~904-911, 1986

12. Huiskies R: Some fundamental aspects of human joint rc- placement: analyses of stresses and heat conduction in bone- prosthesis structures. Acta Orthop Scand suppl 185:l-208, 1980

13. Johnston JO: Treatment of giant cell tumor of bone by ag- gressive curettagc and packing with bone cement. In: Limb Salvage in Musculvskelelul Oncology, pp 512-515. Ed by WF Enneking. New York, Churchill Livingstone, 1987

14. Leeson MC, Lippitt SB: Thermal aspects of the use 01 poly- niethylniethacrylate in large nietaphyscal defects in bonc: a clinical review and laboratory study. Clin Orthop 295:239-245, 1993

15. Marlin RB. Chapman MW, Sharkey NA, Zissimos SL, Bay B, Shors EC: Bone ingrowth and mechanical propertics of cor- alline hydroxyapatite 1 yr after implantation. Biomuterzrrls 14341 -348,1993

16. Moore DC, Frankenburg EP, Goulet JA, Goldstein SA: Hip screw augmentation with an in situ-setting calcium phosphate cement: an in vitro biomechanical analysis. J Orrhop Truuma

17. Patat JL. Guillemin G: Natural coral as bone graft substitute biomaterial: 13 years of experience. In: Sixth Proceedings oj the International Congress on Cotrel-Dubousset Instrurnerita- lion, pp 113-122. Montpellier, Sauramps Medical, 1989

18. Picrcc WA. Welch RD: Acute conipressive strength of hy- droxyapatite cement. Trans Orthop Res SOC 22:755, 1997

19. Sciadini MF. Dawson JM. Johnson KD: Evaluation of bone- derived hone protein with a natural coral carrier as a bone- graft substitute in a canine segmental defect model. J Orthop Rev 15:844-857, 1997

20. Younger EM. Chapman M W Morbidity at bone graft donor sites. J Orthop Trauma 3:192-195, 1989

11:577-583,1997

J Orthop Rcs, Vul. 16, No. 6, 1998