Histologic study of incorporation and resorption of a bonecement–collagen composite: an in vivo study in theminipigRonald Mai, MD, DMD,a Antje Reinstorf, MASc, Dr.rer.nat.,b Eckart Pilling, MD, DMD,a

Matthias Hlawitschka, MD, DMD,a Roland Jung, VetMD,c

Michael Gelinsky, MASc, Dr.rer.nat.,b Matthias Schneider, MD, DMD,a

Richard Loukota, FDSRCS, FRCS,d Wolfgang Pompe, PhD,b Uwe Eckelt, MD, DMD, PhD,a

and Bernd Stadlinger, MD, DMD,a Dresden, Germany, and Leeds, United KingdomDRESDEN UNIVERSITY OF TECHNOLOGY AND LEEDS DENTAL INSTITUTE

Objective. Calcium phosphates are clinically established as bone defect fillers. They have the capability ofosseoconduction and are characterized by a slow resorption process. The present study evaluated the suitability of anewly developed calcium phosphate cement modified with collagen type I.Study design. The modified cement paste was inserted in differently designed defects of 10 minipigs. Further, analveolar ridge augmentation was performed, applying the cement paste. The cement hardened in situ during theoperation, forming a hydroxyapatite collagen composite. Animals were sacrificed after 1, 3, 6, 12, and 18 months. Thetissue integration and resorption process was then evaluated using nondecalcified microsections. All animals wereevaluated for histology.Results.: The implanted material showed osseoconductive characteristics. Resorption started from the edge of thedefect zone, and bone substitution followed rapidly. Twelve months after placement of the cement, completeremodeling was observed.Conclusion. It can be concluded that the applied hydroxyapatite-collagen cement composite shows good resorption

and bone integration. (Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2008;105:e9-e14)

The use of alloplastic tissue replacement materials is anestablished method in reconstructive craniofacial sur-gery. Different bone replacement materials have beenexamined for their application in bone defects. Themain aim of treatment is to reconstruct these defectsanatomically and functionally.1 Therefore, a convenientbone replacement material is necessary. Materials thatcan be remodeled by bone cells should be favored.

The bridging of critical size defects is a challenge inthe reconstruction of defects in bone which may becaused by trauma or tumor.2 In the reconstruction of

Biomet Biomaterials, Darmstadt, Germany, supplied the cementpowder.aDepartment of Maxillofacial Surgery, Dresden University of Tech-nology, University Hospital Dresden.bMax-Bergmann-Center of Biomaterials, Dresden University ofTechnology.cExperimental Centre, Medical Faculty Carl Gustav Carus, DresdenUniversity of Technology.dDepartment of Maxillofacial Surgery, Leeds Dental Institute, LeedsTeaching Hospitals.Received for publication Dec 22, 2005; returned for revision Sep 3,2007; accepted for publication Sep 24, 2007.1079-2104/$ - see front matter© 2008 Mosby, Inc. All rights reserved.

doi:10.1016/j.tripleo.2007.09.016

critical size defects, it is of paramount importance togain appropriate mechanical stability, good bone integ-rity, and rapid resorption and remodeling. Therefore, toavoid malunion, resorption of the material and boneformation should happen simultaneously.

Currently, large defects have to be reconstructed byautologous bone grafts. Autologous cancellous bone isseen as the “gold standard” in the field of bone defectfilling. Any other bone replacement substances mustwithstand this comparison. Besides the advantage of anexcellent biocompatibility regarding the recipient re-gion, several disadvantages must be taken into accountin using autologous cancellous bone. Major disadvan-tages are the need for a second operative procedure,prolonged length of the operation, potential local com-plications at the donor site, and the limited quantity ofdonor bone. Autogenous bone grafts are naturally lim-ited in their amount and present a potential risk to thepatient.3,4 To avoid these disadvantages, allogenous,xenogenous bone, bone replacement substances, anddifferent mixtures with autologous cancellous bone areregularly used in critical size defects.

A clinically applicable alternative for use in cranio-facial, dentoalveolar, and preprosthetic surgery wouldbe an artificial biomaterial with an osseoconductive

potential. This bone replacement material would pref-

e9

OOOOEe10 Mai et al. March 2008

erably be a moldable paste that adjusts to the form ofthe defect, which seems to be essential for mechanicalstability.

In materials science, many materials, such as blocksand granules of calcium phosphates, have been devel-oped and will be commercially available in the future(e.g., Osprovit, Ceros 80, and Ceros 82). However, theblock and large granule materials often have side ef-fects. Previously, only a close adaptation of newlyformed bone at the edges of the cavity has beenachieved.1,5

Another approach to bony reconstruction has beenthe use of replacement materials from a natural origin.These materials have an interconnecting porosity,which is important for the process of ossification. Thisallows the ingrowth of vessels into the implant sub-stance. Ayers showed that the amount of vasculariza-tion of the implant depends to a major extent on itsporosity.6 It was observed that with products having apore size of 200 �m (e.g., BioOss, Algipore, and In-terpore 200), ossification occurs at the periphery andthe vascularization of the inner parts of the implantdoes not take place. The material, which is not incontact with bone or periosteum, becomes surroundedby connective tissue and is not being remodeled. Ahigher ingrowth of bone was observed for products thathave a pore size of 500 �m (e.g., Interpore 500). Withthese materials, animal studies revealed that at leastpartial remodeling occured.2 Materials with a pore sizeof up to 750 �m (i.e., Biocoral, which is calciumcarbonate with an aragonite-like structure) clinicallyshow resorption of the bone replacement substance byosteoclasts and replacement with bone. This occursonly if the substance is in full contact with bone.7

Synthetic calcium phosphate cements, which aremoldable pastes applicable for use in differently shapeddefects harden in vitro by formation of hydroxyapatiteor brushite. Unfortunately they show no interconnect-ing porosity; therefore, only a slow remodeling rate isexpected.

In general, cements may serve to a limited extent asa bone replacement. However, only an animal study canshow if an applied material is able to imitate the func-tion of autologous bone.8

The aim of the present study was to verify bonehealing and remodeling of a bone replacement sub-stance based on calcium phosphate bone cement (Cal-cibon) modified with collagen type I. This compositewas developed as a fiber-reinforced material, thus moreclosely resembling native bone. To analyze the charac-teristics of remodeling and bone healing, the cementcollagen composite was used in a standardized defect

model minipig study.

MATERIALS AND METHODSImplant materials

The Biocement D (Calcibon; Biomet Merck Bioma-terials, Darmstadt, Germany) used in this study is ahydraulic fast-setting calcium phosphate bone cement.The cement precursor powder consists of tricalciumphosphate (TCP; Ca3(PO4)2), anhydrous dicalciumphosphate (DCPA; CaHPO4), and small amounts ofcalcium carbonate (CaCO3). After mixing the cementprecursor with a defined amount of disodium hydrogenphosphate solution (Na2HPO4 aq, 4% w/w), it is trans-formed into calcium-deficient carbonated hydroxyapa-tite (Ca10-x-y[(HPO4)x � (CO3)y] � (PO4)5OH,). The ce-ment precursor powder was modified by the addition of2.5% mineralized collagen I. Thus a material whichmore resembles native bone can be created9 using theprocedure outlined in the following.

The applied bovine collagen type I (Gesellschaft fürNaturextrakte, Wald-Michelbach, Germany) was puri-fied by dialysis against bidistilled water before miner-alization. The method of synchronic assembling andmineralization of collagen follows a specialized pat-tern. Briefly, the acid-soluble collagen (1 g) was dis-solved in 100 mL 0.1 mol/L HCl, and then 180 mLCaCl2 (0.1 mol/L) were added. With 22.6 mL phos-phate buffer (0.5 mol/L), hydroxyapatite was precipi-tated and the collagen fibers were assembled simulta-neously. After 24 h at 37°C, the formation of themineralized collagen was completed. The material wasseparated from the solution by centrifugation, washedwith distilled water, and freeze dried.

Animal experimentsThe minipig was chosen for this animal study be-

cause of its similarity to humans in bone metabolism.Following a standardized operational procedure, 3 drillholes were created, each with a 10 mm diameter, and 1block resection (20 � 30 mm) was prepared in thelower jaw of 10 minipigs. The hydroxyapatite/collagencement was placed in the block resection defect and 1drill hole. Furthermore, an alveolar ridge augmentationwas performed using the cement. For each pig, 2 drillholes served as reference cavities, one spontaneouslybeing filled by the pig’s own blood, the other by autol-ogous cancellous bone and fibrin glue (Fig. 1).

Performing the operation, the subjective moldabilityof the cement paste, the initial setting time of thecement, and the adaptability to the edges of the hostbone were observed.

After the procedure, the pigs were kept in stablestypically used for this species. Daily wound inspectionwas performed and behavior, movement, and food con-sumption were monitored. The different groups, con-

sisting of 2 animals per group, were observed for 1, 3,

OOOOEVolume 105, Number 3 Mai et al. e11

6, 12, and 18 months. After the animals were sacrificed,the resected lower jaws were examined by X-ray (Si-dexis; Sirona, Bensheim, Germany). Imaging served tolocalize the drill holes, the block defect, and the aug-mented region for histologic preparations. Furthermore,the opacity of the cement and the structure of theinterface between the host bone and the cement werequalitatively observed for each group.

Histologic examinationThe samples were fixed in formaldehyde and dehy-

drated in a graded series of ethanol. The sections werethen placed into embeddeding molds containing methylmethacrylate (Technovit 9100 neu; Heraeus Kulzer,Wehrheim, Germany). One hundred–micrometer-thicksections were cut along the axis of each defect using adiamond saw microsectioning system (Exakt-Apparate-bau, Norderstedt, Germany), applying Donath’s micro-sectioning and grinding technique.10 These sectionswere reduced to 30 �m in thickness using grindingtechniques on a roll grinder containing sandpaper(Exakt-Apparatebau). Subsequently, Masson-Gold-ner staining was performed.

The sections were then imaged and histologicallyanalyzed using light microscopy (Olympus BX 61;Hamburg, Germany; and Analysis; Soft Imaging Sys-tems, Münster, Germany). This qualitative analysisserved to analyze bone healing and remodeling at thedifferent times of interest.

RESULTSFollowing application at operation, the collagen ce-

ment composite could be manually molded for about6-8 min. The material displayed good cohesion and had

Fig. 1. Microstructure of hardened hydroxyapatite-collagencomposite; scanning electron microscopic picture.

a high subjective stability after primary setting. It is not

necessary for the implant substance to be kept dryduring the hardening of the cement. Precise adaptationto the edges of the defect zone was possible (Fig. 2).

After the operation, 1 animal developed transientwound healing complications and received antibiotictreatment and analgesia. The species-specific behaviorand the food consumption of the animals remainednormal until they were sacrificed.

One month after the operation, radiographic evalua-tion revealed clearly defined interfaces at all regions ofinterest: the 3 drill holes, the block resection, and thealveolar ridge augmentation. The composite-filled re-gions were easily detected by the differences in X-rayopacity.

After 3 months of healing, the operatively createdregions of interest also could be clearly defined (Fig. 3).

Observing the cement-filled drill hole, initial bone

Fig. 2. Radiograph (right mandible, minipig) 3 months afterimplantation. A, Drill hole, empty; B, drill hole, filled withautologeous bone; C, block resection defect, filled with hy-droxyapatite-collagen (HA-C) composite; D, drill hole, filledwith HA-C composite; E, alveolar ridge augmentation, ap-plying HA-C composite.

Fig. 3. Drill hole (mandible, minipig) immediately after im-plantation. Precise adaptation of the hardened hydroxyapatite-collagen cement to the edges of the defect zone. Masson-Goldner stain, magnification �2.

formation within the cavity was demonstrated by its

OOOOEe12 Mai et al. March 2008

isodense structure at the periphery of the defect com-pared with the neighboring recipient site. In contrast, amarginal enhancement within a cavity would suggestthe fast resorption of the material implanted. The im-planted composite had a constant appearance in theradiologic evaluation. In contrast to the drill hole filledby autologous bone, the blood-filled drill hole was notmineralized at this time. The block resection showedclear signs of demarcation at the edges of the defect,indicating processes of resorption. Observing the aug-mented region, the cement texture was detectable butdisintegrated compared with the appearance after 1month.

Six months after implantation, only residues of thecement could be seen in the paste-filled drill hole. Incomparison, the drill hole filled by autologous bonewas fully ossified. The blood-filled drill hole was partlyossified. Analyzing the block resection, 3 differentzones could be detected. Newly formed bone was foundat the edges of the host bone. Zones of reduced opacitywere located between the edges of the host bone and theresidues of cement, found in the central part of thedefect. The cement implantated at the alveolar ridgeaugmentation site was no more detectable.

Evaluation of the X-ray photographs, taken 12-18months after the operation, showed a complete restitu-tion of all drill holes and the block resection, referred toas a “restitutio ad integrum.”

Under histologic observation, the blood-filled drillhole showed the capability of bone to perform sponta-neous regeneration. Drill holes, filled with autologouscancellous bone, occasionally showed exaggerated

Fig. 4. Drill hole (mandible, minipig) 1 month after implan-tation. Multinucleate giant cells as signs of resorption at thecomposite-bone interface. Masson-Goldner stain, magnifica-tion �40.

bone healing, which is inferior to remodeling.

One month after the operation, the cement-filled drillhole, the block resection, and the alveolar ridge aug-mentation showed numerous multinucleate giant cellsat the implant-bone interface. These cells represent firstsigns of resorption (Fig. 4).

In the marginal part of the block resection, singlecement granules could be detected, being detachedfrom the fixed position of the composite.

After 3 months, lining osteoblasts and layers of os-teoid were observed at all regions of interest. Thisdemonstrates osteoneogenesis at the composite-boneinterface (Fig. 5).

Observing the cement-filled drill hole, no signs ofconnective tissue were found.

Analyzing the block resection, the formation of newbone started over a large area at the surface. This areaof bone formation was interrupted by zones of resorp-tive cell activity and single fibrous tissue.

Six months after the operation, a major part of thecement-filled drill hole was ossified (Fig. 6).

The cement within the block resection was repre-sented by discontinuous islands of bone. Furthermore,intracavitary hydroxyapatite granules and fibrous tissuewere found within the block resection. A small area ofnonmineralized osteoid had started to form from theedges of the host bone.

Twelve months after implantation, no more cementcould be detected in the formerly cement-filled drillhole. The cement proved to be replaced by spongiousbone. Within the block resection, circular appositionalformation of bone dominated adjacent to signs of cel-

Fig. 5. Drill hole (mandible, minipig) 3 months after implan-tation. Lining osteoblasts (orange) show the beginning ofosteoneogenesis at the composite-bone interface. Masson-Goldner stain, magnification �10.

lular resorption of the cement composite. Bone healing

OOOOEVolume 105, Number 3 Mai et al. e13

was observed in the major part of the defect; however,small areas of fibrous healing could be detected.

After 18 months, only residues of hydroxyapatitegranules could be observed within a complete boneregeneration of the block resection. Pre-existent wovenbone was replaced to a major extent by lamellar bone.

DISCUSSIONThe aim of the present study was to verify bone

healing and remodeling of a bone replacement sub-stance, based on calcium phosphate bone cement (Cal-cibon) modified with collagen type I.

Based on the results, we conclude that the appliedhydroxyapatite-collagen cement is resorbable under invivo conditions and shows good biocompatibility. Re-sorption can be seen as the combination of cellularmacrophage activity and physiologic solubility.

This study confirms the suitability of the material asa bone replacement substance in pure bone cavities;however, there is a lack of statistical evidence. Incontrast to bone-bordered drill holes and block resec-tions, the opportunity to perform an alveolar ridgeaugmentation is not possible with this cement, com-pared with other hydroxyapatite-based biomaterials, es-pecially compared with sintered hydroxyapatite ceram-ics, the resorption occurred quickly.11

The paste-type properties allow a precise form-fitadaptation to the defect zone. This meets the precondi-tions for the equal distribution of forces within load-bearing bone; there is an observable difference com-pared with granular and corpuscular bone replacement

Fig. 6. Drill hole (mandible, minipig) 6 months after implan-tation. Circular appositional formation of bone next to signsof cellular resorption of the composite. Masson-Goldnerstain, magnification �4.

substances.

In contrast to other forms of cement, a dry intraop-erative environment is not necessary at the time ofapplication. Being a hydraulic cement, it reaches itsinitial hardness within an acceptable 15-20 min in awater- or blood-containing environment. The cementcomposition of calcium-deficient hydroxyapatite andcollagen type I gives this biomaterial an osseoconduc-tive potential. Its application as a matrix for the definedrelease of immobilized proteins is a major focus ofinterest.12,13

This easily formable bone replacement substanceshows good resorption.

Observing the literature, special membrane systemshave been applied in “large bone defects” to avoid theingrowth of fibrous tissue.14 Owing to of the lack of aninterporous connectivity and the stable surface structureof the cement, the application of a membrane was notnecessary in this animal study.

The properties shown by the applied hydroxyapatite-collagen cement describe a bone replacement substancethat is stable in volume, biodegradable, and osteocon-ductive. In contrast to established polymethyl methac-rylate (PMMA)–based cements, the application of thehydroxyapatite-collagen cement enables ossificationthroughout an augmented volume. This occurs withouttoxic, immunologic, and thermal interactions with thebone.

Hydroxyapatite is a biocompatible substance thatdoes not cause any chronic, inflammatory, allergenic,or toxic reactions.

Nonceramic hydroxyapatite is also referred to ashydroxyapatite cement or calcium phosphate cement. Itis produced through direct crystallization of hydroxy-apatite at physiologic temperatures and pH levels with-out any external heating.15 This cement consists ofpartially hydrated tetracalcium phosphates and bical-cium phosphates. With the addition of water, hydroxy-apatite cements dissolve and form a paste that can bemolded and easily applied in bone defects. The pastehardens to cement after 15 min and is transformed intohydroxyapatite within the following 4 hours. At thatpoint, it is no longer water soluble.16 Similar to ceramichydroxyapatite, hydroxyapatite cement has an osseo-conductive potential. In contrast to ceramic hydroxy-apatite, hydroxyapatite cement ossifies without any lossin volume.16 The slow replacement of bone minimizesthe danger of changes in the outer surface structure ofthe reconstructed area.

After the process of final hardening, a slow butcontinuous dissolving process due to the activity ofmacrophages that induce resorption begins in the sub-sequent weeks and months. This continuous long-termdissolving process in liquid environments is shown by

calcium phosphate cements in contrast to sintered hy-

OOOOEe14 Mai et al. March 2008

droxyapatite; body fluids represent such a liquid envi-ronment. Blood and other body fluids contain calciumand phosphate ions. The behavior of calcium phosphatecements after implantation depends primarily onwhether body fluids in contact with the cement areoversaturated, saturated, or undersaturated.17 At a pHlevel of 7.4 these fluids are oversaturated with hydroxy-apatite, saturated with calcium-deficient apatite (CDA)and carbonate apatite (CA,) and undersaturated withcalcium carbonate (CaCO3) and bicalcium phosphate(DCP). The latter substances especially tend to dissolvevery slowly in the absence of cellular activity. How-ever, a pH level of 7.4 can only be taken as an averagevalue, because living cells constantly change the pHlevel of their natural surrounding. In the proximity ofosteoclasts, the pH level can decrease to 5, whereasaround osteoblasts, it can rise up to 8.5. In general itcan be observed that calcium phosphate cements have ahigher solubility at low pH levels and a lower solubilityat high pH levels.18 Therefore, osteoclasts dissolvecalcium carbonate, bicalcium phosphate, carbonateapatite, and even hydroxyapatite. Osteoblasts, however,stop this process and stimulate the crystal growth ofcarbonate apatite and calcium-deficient apatite.

Biocement D is a calcium phosphate cement which,through the addition of a liquid hardening solution, canbe implanted as a paste into a bone defect zone. Usinga hardening accelerator, it is possible to implant thecement into the operating area within adequate timewithout an increase in temperature and with just subtlechanges in the pH level.9 Nevertheless, the high rigidityof the material does not allow implantation into areaswhere mechanical stability is essential. Therefore 2.5mass-% freeze-dried mineralized collagen type I wasadded to Biocement D in in vitro experiments.

In conclusion, the opportunity to achieve sufficientbone healing within bone defects is possible by apply-ing this collagen-enriched calcium phosphate cement.The cement displayed a process of resorption that oc-curred simultaneously with bone formation. An easyintraoperative handling and a good clinical applicationwere observed.

The authors are grateful to Mrs. Kleiber for excellenttechnical assistance in histologic examination.

REFERENCES1. Gunther KP, Scharf HP, Pesch W, Puhl HJ. [Integration proper-

ties of bone substitute materials. Experimental studies on ani-mals]. Orthopade 1998;27:105-17.

2. Gosain AK, Song L, Yu P, Mehrara BJ, Maeda CY, Gold MT,Longaker LI. Osteogenesis in cranial defects: reassessment of theconcept of critical size and the expression of TGF-beta isoforms.

Plast Reconstr Surg 2000;106:360-71.

3. Grob D. [Problems at the donor site in autologous bone trans-plantation]. Unfallchirurg 1986;89:339-45.

4. Younger MW, Chapman EM. Morbidity at bone graft donorsites. J Orthop Trauma 1989;3:192-5.

5. Rueger JM. [Bone substitutes. State of the art, and what liesahead? ] Unfallchirurg 1996;99:228-36.

6. Ayers RA, Simske SJ, Nunes LM, Wolford CR. Long-term boneingrowth and residual microhardness of porous block hydroxy-apatite implants in humans. J Oral Maxillofac Surg 1998;56:1297-301.

7. Nolan P, Templeton P, Mollan DJ, Wilson RA. Osteoinductivepotential of human demineralised bone and a bioceramic inabdominal musculature of the rat. J Anat 1991;174:97-102.

8. Shirakata Y, Oda S, Kinoshita A, Kikuchi S, Tsuchioka I, Ish-ikawa H. Histocompatible healing of periodontal defects afterapplication of an injectable calcium phosphate bone cement. Apreliminary study in dogs. J Periodontol 2002;73:1043-53.

9. Knepper-Nicolai B, Reinstorf A, Hofinger I, Flade K, Wenz W,Pompe R. Influence of osteocalcin and collagen I on the mechan-ical and biological properties of Biocement. Biomol Eng2002;19D:227-31.

10. Donath K. The diagnostic value of the new method for the studyof undecalcified bones and teeth with attached soft tissue (Sage-Schliff [sawing and grinding] technique). Pathol Res Pract1985;179:631-3.

11. Oreffo RO, Driessens FC, Planell JT, Triffitt JA. Effects of novelcalcium phosphate cements on human bone marrow fibroblasticcells. Tissue Eng 1998;4:293-303.

12. Knabe C, Driessens FC, Planell JA, Gildenhaar R, Berger G, ReifD, et al. Evaluation of calcium phosphates and experimentalcalcium phosphate bone cements using osteogenic cultures.J Biomed Mater Res 2000;52:498-508.

13. Oreffo RO, Driessens FC, Planell JT, Triffitt JA. Growth anddifferentiation of human bone marrow osteoprogenitors on novelcalcium phosphate cements. Biomaterials 1998;19:1845-54.

14. Chang YL, Stanford JC, Keller CM. Calcium and phosphatesupplementation promotes bone cell mineralization: implicationsfor hydroxyapatite (HA)-enhanced bone formation. J BiomedMater Res 2000;52:270-8.

15. Takagi S, Chow LC, Markovic M, Friedman PD, Costantino CD.Morphological and phase characterizations of retrieved calciumphosphate cement implants. J Biomed Mater Res 2001;58:36-41.

16. Bifano CA, Edgin WA, Colleton C, Bifano PD, Constantino SL.Preliminary evaluation of hydroxyapatite cement as an augmen-tation device in the edentulous atrophic canine mandible. OralSurg Oral Med Oral Pathol Oral Radiol Endod 1998;85:512-6.

17. Ramselaar MM, Van Mullem JR, De Wijn PJ. Porous rootreplacements: reactions of the surrounding tissues. J Dent 1986;14:202-8.

18. Khairoun I, Boltong MG, Driessens JA, Planell FC. Effect ofcalcium carbonate on clinical compliance of apatitic calciumphosphate bone cement. J Biomed Mater Res 1997;38:356-60.

Reprint requests:

Ronald Mai, MD, DMDDepartment of Maxillofacial SurgeryFaculty of MedicineDresden University of TechnologyFetscherstr. 74D-01307 DresdenGermany

ronald.mai@uniklinikum-dresden.de