Resorption of hydroxyapatite and fluorapatite ceramiccoatings on weight-bearing implants: A quantitative andmorphological study in dogs

Søren Overgaard,1,2,3 Martin Lind,1,2 Kaj Josephsen,4 Arvid B. Maunsbach,5 Cody Bunger,1,2 Kjeld Søballe1,2

1Department of Orthopaedics, Biomechanics Laboratory, Aarhus University Hospital, Nørrebrogade 44, Building 01A,DK-8000 Aarhus C, Denmark2Institute of Experimental Clinical Research, Skejby Hospital, Aarhus University Hospital, Brendstrupgaardsvej, 8200Aarhus N, Denmark3Stereological Research Laboratory, Aarhus University Hospital, Nørrebrogade 44, 8000 Aarhus C, Denmark4Department of Dental Pathology, Operative Dentistry and Endodontics, Royal Dental College, Faculty of HealthSciences, University of Aarhus, 8000 Aarhus C, Denmark5Department of Cell Biology, Institute of Anatomy, University of Aarhus, 8000 Aarhus C, Denmark

Received 4 April 1996; accepted 11 December 1996

Abstract: Resorption (defined as loss of ceramic coatingbecause of cellular activity or dissolution) of ceramic coat-ings is a matter of concern for the long-term performance ofceramic-coated implants. A new fluorine-containing coating,fluorapatite (FA), has been shown to be more stable thanhydroxyapatite (HA) in unloaded models. In a weight-bearing model in trabecular bone, we evaluated loss (de-fined as reduction of coating irrespective of type of mecha-nism) of HA and FA coatings during 25 weeks of implanta-tion. Eight mature dogs had HA- or FA-coated implantsinserted bilaterally into the weight-bearing region of the me-dial femoral condyle. Quantified loss of ceramic coating wasestimated at the light microscopic level using stereologicalmethods. The experiment showed significant loss of bothtypes of coatings. However, no statistical difference in loss ofceramic coating was found regarding surface area, implant

coverage, volume, and thickness (p = 0.77, p = 0.13, p = 0.56,p = 0.23, respectively). Completely resorbed HA coating wasreplaced by 36 ± 6.0% (range: 26–42) bone in direct contactwith the implant surface compared with 29 ± 16.0% (range:12–59) for FA (p = 0.40), suggesting that the implant wasfirmly fixed despite loss of the ceramic coating. Transmis-sion electron microscopy in combination with electron en-ergy spectroscopy and electron spectroscopic imagingshowed that osteclast-like cells, osteocytes, macrophage-likecells, and fibroblasts had phagocytosed calcium-containingfragments, indicating cell-mediated resorption of the ce-ramic coating. © 1998 John Wiley & Sons, Inc. J Biomed MaterRes, 39, 141–152, 1998.

Key words: fluorapatite; hydroxyapatite; implant; resorp-tion; stereology

INTRODUCTION

Experimental studies have proven that hydroxyapa-tite (HA) coating is able to enhance implant fixationand bone ingrowth during different loading condi-tions when inserted into the bone of canines and hu-mans.1–12 As a result of disappointing survival rates ofuncemented components in younger patients, plasma-sprayed hydroxyapatite (HA) coating has been intro-

duced to improve the early implant fixation.13–15

Clinical short-term studies on cementless componentshave shown that HA-coated femoral stems have lowermigration early postoperatively compared with un-coated stems, which might be important for the long-term survival of prostheses.13,15–18

However, clinical experience with HA coating islimited and questions about long-term performancehave been raised. Resorption of HA coating may occurwhen implanted for prolonged periods and it has beensuggested that it is accelerated by micromotion.19 Thismight reduce mechanical fixation of the implant andlead to delamination of the coating, resulting in thirdbody wear debris and, ultimately, failure of the im-plant.20–24

A new fluorine-containing ceramic coating, fluor-apatite (FA), has been developed to reduce coatingresorption.2,25–29 FA has the same apatite structure asHA, apart from a more compact lattice, and has com-

This paper was read in part at the 2nd Combined Ortho-paedic Research Society Meeting in San Diego, November5–8, 1995

Correspondence to: S. OvergaardContract grant sponsor: Danish Rheumatism AssociationContract grant sponsor: Danish Medical Research CouncilContract grant sponsor: Velux FoundationContract grant sponsor: The Karen Elise Jensen FoundationContract grant sponsor: The University of Aarhus, Denmark

© 1998 John Wiley & Sons, Inc. CCC 0021-9304/98/010141-12

parable biocompatibility properties.30 FA coating ismore thermostable than HA during plasma-spraying,which is an advantage during manufacturing.29 Invitro studies have demonstrated lower solubility of FAcompared with HA coating, and in vivo studies usingnon-weight-bearing models have shown that FA coat-ing is more stable.2,25–28,31–34

However, the question remains as to how the FAand HA coating will perform in a more clinically rel-evant intraarticular weight-bearing model in trabecu-lar bone. In a previous study we found no differencebetween HA- and FA-coated implants regarding me-chanical fixation, total bone coverage, and bone re-modeling around the implants.35 The purpose of thepresent study was to investigate loss of HA and FAcoatings on weight-bearing implants and to quantifycoating replaced by bone ingrowth using stereologicalmethods. Another purpose was to describe cells ableto phagocytose ceramic crystals close to the implantsurface using transmission electron microscopy (TEM)in combination with electron energy loss spectroscopyfor element identification.

MATERIALS AND METHODS

Study design

Eight skeletally mature mongrel dogs had a stableweight-bearing implant system inserted bilaterallyinto the weight-bearing region of the medial femoralcondyles of the knees (Fig. 1).6,7 The dogs were bredfor scientific purposes and were handled according tothe Danish law on animal experimentation. Porous-coated titanium (Ti) implants were plasma-sprayedwith HA and FA coatings and randomly allocated tothe right or left knee. Postoperatively, fully weight-bearing and free activities were allowed. The dogswere terminated after 25 weeks. Six noninserted im-plants of each type served as controls for evaluation ofceramic coatings. Control implants were left in sterilepackages until dehydration and embedding. Datafrom push-out test and total bone coverage was pub-lished earlier.35

Implants

The implants were manufactured of Ti-alloy (Ti-6AI-4V) by Biomett Inc, Warsaw, IN. They were madeof a solid core and a plasma-sprayed porous coatingwith a mean pore size of 300 mm. Implants had a finalmaximal diameter of 6 mm and a length of 10 mm.They were sterilized after the coating process bygamma irradiation. Average surface roughness (Ra)

was 29.1 ± 7 mm for HA-coated implants and 28.3 ± 4mm for FA-coated implants (NS) and was measured asreported in detail previously.35

Ceramic coatings

Coatings were applied by BioInterfaces, Inc, SanDiego, CA, using plasma-spraying technique. Charac-teristics of the ceramic coatings are presented in Ta-ble I.

Implant device

The stable implant device is made of Ti-alloy (Ti-6AI-4V) except for a polyethylene plug, as described

Figure 1. Drawing of the stable implant device implantedinto the medial femoral condyle. The anchorage screw (1)has self-tapping threads to ensure firm fixation in the bone.A piston (2) is fixed to the anchorage screw, on which thetest implant (3) and the polyethylene (PE) plug (5) aremounted. A ring (4) serves as bearing and centralizer for thetest implant. The gap around the implant communicateswith the knee joint space. Note protrusion of the PE plug,which transfers the load through the implant system. Modi-fied from Søballe et al.6

142 OVERGAARD ET AL.

previously (Fig. 1).6,7 It consists of an anchorage screwon which the test implant and a polyethylene (PE)plug are mounted. The plug transfers load through theimplant system without micromovements.

Surgery

Implants were inserted under general anesthesiaand prophylactic antibiotics (Ampicillin).6,7 A guidewire was inserted into the medial femoral condyle un-der fluoroscopic control. The cavity in the condyle wasdrilled by hand and a gap of 0.75 mm around theimplant was created before the implant system wasinserted. Postoperatively, analgesics (Buprenorphine)were administered intramuscularly for 2–3 days.

Preparation

Immediately after termination, the distal femurwas harvested and cut orthogonally to the long axisof the implant on a water-cooled diamond band saw(EXAKT-Cutting Grinding System Standard, Appa-ratebau GMBH, Norderstedt, Germany). A specimenof 3 mm was fixed in 70% ethanol for histomorphom-etry. Another specimen of 300 mm was fixed in a mix-ture of 2% paraformaldehyde and 2.5% glutaralde-hyde in 0.08M sodium cacodylate buffer for TEManalysis.36

Sections for light microscopy

Control implants and bone-implant specimens weredehydrated in graded ethanol (70 to 100%) containing0.4% basic fuchsin and embedded in methylmethacry-late (Technovitt 7200 VLC, Exact, ApparatebauGMBH, Norderstedt, Germany). Due to the aniso-tropic orientation of trabecular bone, vertical sections

were performed according to Baddeley et al. to haveunbiased estimates.37,38 The vertical axis was chosenas the one parallel to the long axis of the implant [Fig.2(a)] Random rotation of the sections was achieved byrotating the embedded tissue block around this axis.Serial vertical sections were cut parallel to the definedvertical axis, with a mean distance between sections of735 mm [Fig. 2(b)]. On average, seven sections wereobtained from each implant. Every second section,with a random start, was used for histological evalu-ation at the light microscopic level [Fig. 2(b)]. By mis-

TABLE ICharacterization of Hydroxyapatite and Fluorapatite

Coatings

Parameter Hydroxyapatite Fluorapatite

*Purity (%) 97 97*Trace of TCP* (%) 3 3*Calcium/phosphate 1.69 1.61*Crystallinity (%) 68 74*Thickness (mm) 50 50Porosity (%) 15 15

TCP = tricalcium phosphate; *determined by manufac-turer. Crystallinity and purity of the coatings were mea-sured using X-ray diffraction analysis, porosity was deter-mined by scanning electron microscopy of the non-insertedcontrol implants using image analysis.

Figure 2. Schematic drawing illustrating the concept ofvertical sectioning through the bone–implant specimen. (a)Random section plane is given by rotating the embeddedbone–implant block around the vertical axis. Exhaustiveparallel sections are performed. (b) in this example sectionnumbers 2, 4, and 6 were chosen for histomorphometricevaluation.

143RESORPTION OF HA AND FA CERAMIC COATINGS

take, four control implants of each type were not cutvertically. They were randomly rotated in space beforeexhaustive serial sectioning. However, the sections ob-tained were isotropic and uniform, and random andunbiased estimates were achieved.

Each section was ground and polished to a thick-ness of approximately 50 mm using a micro grindingsystem (EXAKT-Micro Grinding System, Exact, Appa-ratebau GMBH, Norderstedt, Germany). The surfacewas counter-stained with 2% light green for fifteenmin.39 For blinded histomorphometric evaluation ofthe entire implant surface, the microscope wasequipped with a 10 × 0.40 lens. The field of vision wastransferred from the microscope to a computer moni-tor, resulting in a total magnification of X250.

Bone ingrowth, surface area, and volume of the ce-ramic coating were evaluated by the linear intercepttechnique and by point counting.40 Bone ingrowthwas defined as bone in direct contact with the ceramiccoating surface or with the Ti implant surface. Relativesurface area of ceramic coating was defined as ceramiccoating coverage of the implant surface as a percent-age of the total implant surface. In addition, absolutevalues of ceramic coating surface area, volume, andthickness were estimated per implant length (see be-low). Estimation of surface area was performed in anunbiased manner by means of a test line system. Theline system was rotated for each new field of vision.The rotation was random and sinus-weighted in rela-tion to the vertical axis providing isotropic lines.37 An-other test line system of alternating horizontal andvertical test lines was applied to isotropic, uniform,and randomly cut sections of control implants. Esti-mation of volume on all sections was done by pointcounting. Test systems were superimposed on thefield of vision on the screen using a newly developedsoftware program (CAST-Gridt, Olympus DenmarkA/S, Glostrup, Denmark) and were calibrated to haveapproximately 200–300 interceptions or pointscounted for each parameter per specimen.41

Transmission electron microscopy (TEM)

Three randomly selected dogs were chosen forTEM. One section from each femoral condyle had theimplant gently pushed out from the surrounding bonetissue. It then was divided with radial cuts into pie-shaped pieces. All specimens were postfixed in potas-sium ferrocyanide-reduced osmium tetroxide, dehy-drated in graded ethanol (30–100%) and propylene ox-ide, and infiltrated in Epon. Thin sections (60 nm)were cut on a Reichert Ultracut E microtome andplaced on formvar-coated copper grids. The sectionswere stained with uranyl acetate and lead citrate andwere viewed at 60 kV in a Philips 301 transmissionelectron microscope.

Electron energy loss spectroscopy (EELS) andelectron spectroscopic imaging (ESI)

Thin unstained sections containing intracellular de-posits of crystals were analyzed in a Zeiss EM 912Omega energy-filtering transmission electron micro-scope. EELS spectra were recorded from crystals at amagnification of 160,000X and at an accelerating volt-age of 100 kV. Sections of Epon-embedded HA crys-tals were used as controls. The tissue distribution ofcalcium was analyzed using ESI corresponding to CaL2,3 edge at electron loss 346 eV. For comparison im-ages also were obtained at electron loss 338 eV in frontof the calcium edge and from conventional transmis-sion electron microscopy.

Stereological methods

Surface area, volume, and thickness of ceramic coat-ings were estimated using stereological methods.42

Surface area per length of the implant, SL, was esti-mated from the relationship:

Sv = 2 × S IL (mm2/mm3), (1)

where Sv is the coating surface area per unit volume,SIL is the number of intersections with the coatingsurface per length (L) of test line.37,43 Test line lengthis given by the ratio of test line (,) to test points (p) onthe monitor multiplied by the number of points thathit the implant (P): L = P (implant) X ,/p. Substitutionof the relationship V (coating) = L (implant) X A (coat-ing), where A is the section area of the coating ana-lyzed, into Equation (1) and rearranging yields:

SL = Sv × A = [2 × S I × A(coating)] / [P (implant)× ,/p] (mm2/mm) (2)

Coating volume V (coating) is calculated by using theCavaliere method.41 The area on sections was multi-plied by the distance between sections (t):

V (coating) = t × A = t × S a= t × S P × a/p (mm3), (3)

where a/p is the area per point of the test systemcorrected for magnification. Coating surface area perimplant length is calculated from Equations (2) and(3):

SL = [2 × S I × t] / [[I/a]× L(implant)] (mm2/mm) (4)

Volume of the ceramic coating per implant length wasestimated from Equation (3) divided by the length ofthe implant analyzed:

VL (coating) = [SP (coating) × [a/p] × t]/ L (implant) (mm3/mm), (5)

144 OVERGAARD ET AL.

The thickness (T) of the coating was calculated as

T = VL / SL (mm) (6)

Statistics

Data are presented as means ± standard deviations(SD). Statistical significance was determined by pairedor unpaired t test. P values less than 0.05 (two-tailed)were considered significant.

RESULTS

All dogs completed the study and were fully weightbearing within three days after surgery and through-out the study. All PE plugs and implants were in situat autopsy. Cartilage wear on the tibia plateau corre-sponding to articulation with the polyethylene plugwas observed. No clinical signs of infection were pres-ent.

Bone ingrowth

Bone ingrowth to the ceramic coating was 5–10-foldgreater compared with bone ingrowth to the Ti surface

without ceramic coating (Table II). The ceramic sur-face was covered by 65 ± 9.2% and 68 ± 11.0% bone onHA- and FA-coated implants, respectively (p = 0.52,Table II). Bone marrow covered the remaining surfaceapart from very small amounts of fibrous tissue ontwo HA-coated implants and three FA-coated im-plants. A few ceramic particles were found incorpo-rated in the surrounding bone close to the implantsurface. No foreign body reaction or osteolysis wereshown at the light microscopic level. Bone in contactwith the implant was mainly of the lamellar type.

Resorption of ceramic coating

Control implants were covered almost fully withceramic coating (Table III). The relative coating surfacearea was reduced significantly by 18 ± 4.4% and 13 ±7.5% for HA- and FA-coated implants, respectively,with no difference between coatings (p = 0.13, Fig. 3).Absolute surface area, volume, and thickness of thecoatings were reduced significantly compared withcontrol implants, but no significant differences be-tween HA and FA coatings were found (Table III).Completely resorbed HA was replaced by 36 ± 6.0%bone ingrowth compared with 29 ± 16.0% for FA coat-ings (p = 0.41) whereas bone marrow had replaced 64± 16.0% and 69 ± 19.8% of resorbed HA and FA coat-ing, respectively (Fig. 3). The difference between boneand marrow ingrowth compared to the Ti implant sur-face was significant for HA-and FA-coated implants (p= 0.002 and p = 0.02, respectively).

Transmission electron microscopy

Aggregates of apatite crystals from the ceramic coat-ing were seen in direct contact with bone trabeculae inall specimens except one. In the latter specimen, a fi-brous membrane was present at the tissue–implantinterface. At the ultrastructural level, several cell typescontaining apatite crystals were found from zero to100 mm on the implant surfaces irrespective of type ofcoating. Serially cut sections of the cells verified theintracellular content of apatite crystals. Thus in themultinucleated osteoclast-like cells adjacent to theapatite aggregates, a high number of vacuoles withcrystal fragments were located just below the cell sur-face [Fig. 4(a)]. Some vacuoles were well preservedwith a distinct limiting membrane and a granular con-tent [Fig. 4(b)]. In bone-marrow, mononuclear andmacrophage-like cells with crystal-containing phago-somes sometimes were observed [Fig. 4(c)]. Surpris-ingly, in half the specimens a few osteocytes werefound demonstrating cytoplasmic vacuoles with crys-

TABLE IITissue Ingrowth to Hydroxyapatite- and

Fluorapatite-Coated Implants in Percentage After 25Weeks’ Implantation. Mean ± SD

InterfaceHydroxyapatite

%Fluorapatite

%

Bone ingrowth to implant surfaceP = 0.43

Ceramic–bone 50.7 ± 10.2* 53.8 ± 6.4*P = 0.28

Titanium–bone 9.0 ± 3.5 5.8 ± 5.1Tissue ingrowth to ceramic surface

P = 0.52

Ceramic–bone 68.4 ± 11.0† 65.3 ± 9.2†

P = 0.70

Ceramic–bone marrow 31.3 ± 11.0 33.3 ± 8.9Ceramic–fibrous tissue 0.3 ± 0.6 1.4 ± 2.4

*Significantly more bone ingrowth to the ceramic surface(ceramic–bone) compared with the titanium surface (Ti–bone) (p < 0.0001); †significantly greater ceramic–bone areathan ceramic–bone marrow area (p = 0.002). No significantdifference between HA- and FA- coated implants for anyparameter was found. Ingrowth of fibrous tissue was onlyfound on a few implants which did not justify statisticalanalysis.

---------------

---------------

---------------

---------------

145RESORPTION OF HA AND FA CERAMIC COATINGS

tal fragments [Fig. 4(d)]. Finally, in the fibrous mem-brane a few fibroblasts with crystal-containing vacu-oles were seen [Fig. 4(e)].

Element analysis

EELS spectra of intracellular inclusions revealeddistinct peaks around 346 eV, corresponding to the CaL2,3 edge (Fig. 5). This peak was not present in thebackground EELS spectra recorded from adjacent cy-toplasmic areas but in spectra recorded from controlsections containing HA crystals. ESI at 346 eV showedcalcium associated with intracellular crystals (Fig. 6).

DISCUSSION

To our knowledge, this is the first controlled studyquantifying loss of HA and FA coatings in a weight-bearing model in trabecular bone. The study showedthat HA and FA were reduced significantly during the25-week implantation period, but no statistical differ-ence between coatings was found. Completely re-sorbed coating was replaced by approximately 33%bone on the titanium implant surface and the remain-ing part was covered by bone marrow.

Most earlier animal and human experiments on re-sorption of ceramic coatings have used non-weight-bearing models.2,9,12,44–47 In contrast, our study in-cluded a model where the implants were directlyweight bearing, with access of joint fluid to the gaparound the implant mimicking the clinical situa-tion.7,35,48 The relationship between animal and hu-man studies is uncertain. Compared with human re-

trieval studies the current study demonstrated consid-erable more HA coating loss, which might beexplained by a 2–3 times higher bone turnover rate incanines.8,49–52

Earlier studies have shown that FA coating is morestable than HA in cortical/intramedullary press fitmodels.2,12 Klein et al. found, that heat-treated FAcoatings produced with different sizes of startingpowder were significantly more stable compared withcorresponding HA coatings.12 Dhert et al. employedHA and FA coatings from the same vendor and foundsimilar results.2 In contrast, the present study showedno significant difference between coatings but foundsignificant reduction of HA and FA coatings. Thismight be explained by the use of different experimen-tal models and coating vendors, resulting in ceramiccoatings of different qualities.46,53 Moreover, differentevaluation methods were applied. Conventional his-tomorphometric evaluation of bone implants by tradi-tion has been of transverse sections without consider-ing the orientation of the surrounding tissue.1,5,54 Forestimation of volumes this raises no problems.42,43 Toavoid getting biased information on surfaces, trabec-ular bone and implant surfaces must be assumed toshow isotropic orientation.37,55 However, with respectto morphology and mechanical properties, trabecularbone has been shown to be anisotropic.56–58 Moreover,when inserting a loaded implant, the surroundingbone will adapt to the new situation by orienting thebone trabeculae during implantation.38 Therefore, tohave unbiased estimates of surfaces, either isotropicuniform random (IUR) sections or IUR test lines onvertical sections are mandatory.37,55 In the presentstudy, loss of ceramic coating was quantified by IURtest lines on vertical sections and unbiased estimateswere achieved.

TABLE IIIQuantified Parameters of Hydroxyapatite and Fluorapatite Coatings on Control Implants and on Stable

Weight-Bearing Test Implants After 25 Weeks’ Implantation. Mean Values ± SD

Quantified CeramicCoating Parameter

HA-Coated Implant FA-Coated Implant

Control Test Control Test

P = 0.13

Relative surface area (%) 98.3 ± 1.2 79.9 ± 4.4* 98.8 ± 0.4 85.8 ± 7.5†

P = 0.77

Absolute surface area (mm2/mm) 24.9 ± 2.5 12.9 ± 3.4* 21.4 ± 0.2 13.6 ± 4.2*P = 0.56

Absolute volume (mm3/mm) 1.1 ± 0.01 0.38 ± 0.1* 1.1 ± 0.1 0.43 ± 0.1*P = 0.23

Absolute thickness (mm) 44.6 ± 5.7 29.0 ± 3.4* 43.7 ± 5.8 31.3 ± 3.7†

Relative surface area of ceramic coating was defined as ceramic coating coverage of the implant surface in percentage ofthe total implant surface. Significant resorption of coating compared with control implants (*p < 0.0001, †p = 0.001)

No significant difference between HA and FA coatings on test implants was found.

-------------------------------------

-------------------------------------

-------------------------------------

-------------------------------------

146 OVERGAARD ET AL.

Recently we found that different weight-bearingconditions influenced loss of HA coating signifi-cantly.19 Continuous loading and micromotion of theimplant resulted in twofold greater reduction of theHA coating compared with a stabilized implant. In thepresent study, HA-coated implants had an 18% reduc-tion in relative surface area compared with 43% forstabilized implants and 75% for continuously loadedimplants.19 This variation might be explained by dif-ferent microenvironments around stable and unstable

implants postoperatively. Micromotion induces a fi-brous tissue membrane around the implant in contrastto a stable situation where bone ingrowth occurswithin 6 weeks after implantation.5–7 The fibrous tis-sue membrane may be more aggressive than bone inthe resorption process of ceramic coatings due to dis-solution and cell-mediated degradation.

Loss of ceramic coating is most likely caused bythree different mechanisms: (1) A dissolution may oc-cur when low pH appears, as suggested from in vitro

Figure 3. (a) Hydroxyapatite (HA)- and (b) Fluorapatite (FA)-coated implants. Ceramic coatings are almost completelyresorbed where bone marrow is in direct contact with the coatings (arrows). HA coating is partly replaced by bone ingrowthin direct contact with the titanium surface (arrow). Remaining coating is covered with bone. No signs of delamination orinflammatory reaction are seen. Light microscopy, ground sections stained with light green, and basic fuchsin. Bar = 100 mm.

147RESORPTION OF HA AND FA CERAMIC COATINGS

Figure 4. Transmission electron microscopec images demonstrating apatite crystals in specimens from hydroxyapatite- andfluorapatite-coated implants after 25 weeks’ implantation. N = nuclei, M = mitochondria. Arrows indicate crystal containingvacuoles. Sections are stained with uranyl acetate and lead citrate. Bars indicate magnification. (a) Multinucleated osteoclast-likecell adjacent to the ceramic coating. The cytoplasm contains several vacuoles with crystal fragments. (b) Higher magnificationof two vacuoles from 4(a) with granular and apatite crystal content. (c) Mononuclear cell with several apatite-containingvacuoles. (d) Osteocyte lying in a lacuna surrounded by mineralized bone matrix. Apatite crystals are shown in the cytoplasm.(e) Fibroblast surrounded by cross-cut collagen fibrils. A vacuole containing apatite crystals is shown in the cytoplasm.

148 OVERGAARD ET AL.

studies;59–61 (2) Cell-mediated resorption has beenproven. In vitro studies have shown that the osteoclastcan create resorption lacunae, probably due to low pHat the ruffle border.62–64 This study showed that os-teoclast-like cells, monocytes, and fibroblasts had

phagocytosed ceramics in accordance with earlier re-ports.62,65–69 Moreover, we found crystal fragments in-side osteocytes. The significance of this is not clear, butthe particles might have been incorporated in the pre-osteoblast or the osteoblast stage.66 The intracellular-arranged crystal fragments most likely have beenphagocytosed and not displaced into cells during theprocessing of sections because the intracellular struc-tures were well preserved on serially cut sections. Theidentification of the intracellular crystals as HA or FAwas supported by the demonstration that they containcalcium, as shown by EELS and electron spectroscopicimaging; and (3) Mechanical removal and wear duringimplantation or tissue processing may occur but wasnot likely in the present experiment.23,70

CONCLUSION

HA and FA coatings on stable weight-bearing im-plants were reduced significantly during the 25-weekstudy, but no statistical differences between coatingswere found. Completely resorbed ceramic coating was

Figure 6. (a) Conventional transmission electron micrograph of intracellular crystals. (b) Electron spectroscopic imaging(ESI) of the same area using electron energy loss 338 eV immediately adjacent to the Ca L2,3 edge. No signal is observed fromthe crystals. (c) ESI at electron loss around 346 eV, corresponding to the Ca L2,3 edge and demonstrating the presence ofcalcium in the crystals (background not subtracted). N = nuclei.

Figure 5. Electron energy loss spectra from intracellularcrystals, background from adjacent cytoplasm, and Epon-embedded hydroxyapatite crystals in control section. Theintracellular and control crystals show distinct peaks at elec-tron loss around 346 eV, corresponding to the Ca L2,3 edge.71

149RESORPTION OF HA AND FA CERAMIC COATINGS

partly replaced by bone in intimate contact with the Tiimplant surface, suggesting durable implant fixation.Osteoclast-like cells, osteocytes, macrophage-like cells,and fibroblasts had phagocytosed calcium-containingfragments, indicating cell-mediated resorption of theceramic coating.

The authors wish to acknowledge Merete Andersen, An-ette Milton, and Jane Pauli, of the Royal Dental College, andKaren Thomsen, Ole Moeskjær, and Else-Merete Løcke, ofthe Department of Cell Biology, Institute of Anatomy, Uni-versity of Aarhus, for their technical assistance. Biomet Inc.,USA, kindly provided the implants, and the ceramic coat-ings were kindly provided by BioInterfaces Inc., USA.

References

1. S. D. Cook, K. A. Thomas, and M. Jarcho, ‘‘Hydroxy-apatite-coated porous titanium for use as an orthopedicbiologic attachment system,’’ Clin. Orthop., 230, 303–312(1988).

2. W. J. A. Dhert, C. P. A. T. Klein, J. A. Jansen, E. A. vander Velde, R. C. Vriesde, P. M. Rozing, and K. de Groot,‘‘A histological and histomorphometrical investigationof fluorapatite, magnesiumwhitlockite, and hydroxyl-apatite plasma-sprayed coatings in goats,’’ J. Biomed.Mater. Res., 27, 127–138 (1993).

3. R. G. T. Geesink, K. de Groot, and C. P. A. T. Klein,‘‘Chemical implant fixation using hydroxyl-apatitecoatings. The development of a human total hip pros-thesis for chemical fixation to bone using hydroxyl-apatite coatings on titanium substrates,’’ Clin. Orthop.,225, 147–170 (1987).

4. R. G. T. Geesink, K. de Groot, and C. P. A. T. Klein,‘‘Bonding of bone to apatite-coated implants,’’ J. BoneJoint Surg., 70-B, 17–22 (1988).

5. K. Søballe, E. S. Hansen, H. Brockstedt–Rasmussen,and C. Bunger, ‘‘Hydroxyapatite coating converts fi-brous anchorage to bony fixation during continuousimplant loading,’’ J. Bone Joint Surg., 75B, 270–278(1993).

6. K. Søballe, H. Brockstedt–Rasmussen, E. S. Hansen,and C. Bunger, ‘‘Hydroxyapatite coating modifies im-plant membrane formation. Controlled micromotionstudied in dogs,’’ Acta Orthop. Scand., 63, 128–140(1992).

7. K. Søballe, E. S. Hansen, H. Blockstedt–Rasmussen,P. H. Jørgensen, and C. Bunger, ‘‘Tissue ingrowth intotitanium and hydroxyapatite-coated implants duringstable and unstable mechanical conditions,’’ J. Orthop.Res., 10, 285–299 (1992).

8. R. D. Bloebaum, K. N. Bachus, M. H. Rubman, andL. D. Dorr, ‘‘Postmortem comparative analysis of tita-nium and hydroxyapatite porous-coated femoral im-plants retrieved from the same patient. A case study,’’J. Arthroplasty, 8, 203–211 (1993).

9. A. A. Hofmann, K. N. Bachus, and R. D. Bloebaum,‘‘Comparative study of human cancellous bone remod-eling to titanium and hydroxyapatite-coated implants,’’J. Arthroplasty, 8, 157–166 (1993).

10. P. Ducheyne, L. L. Hench, A. Kagan, M. Martens, A.Bursens, and J. C. Mulier, ‘‘Effect of hydroxyapatite im-pregnation on skeletal bonding of porous-coated im-plants,’’ J. Biomed. Mater. Res., 14, 225–237 (1980).

11. J. E. Dalton, S. D. Cook, K. A. Thomas, and J. F. Kay,‘‘The effect of operative fit and hydroxyapatite coatingon the mechanical and biological response to porousimplants,’’ J. Bone Joint Surg., 77-A, 97–110 (1995).

12. C. P. A. T. Klein, J. G. Wolke, J. M. de Blieck–Hogervorst, and K. de Groot, ‘‘Features of calciumphosphate plasma-sprayed coatings: An in vitro study,’’J. Biomed. Mater. Res., 28, 961–967 (1994).

13. K. Søballe, S. Toksvig–Larsen, J. Gelineck, S. Fruens-gaard, E. S. Hansen, L. Ryd, U. Lucht, and C. Bunger,‘‘Migration of hydroxyapatite-coated femoral prosthe-sis. A roentgen stereophotogrammetric study,’’ J. BoneJoint Surg., 75B, 681–687 (1993).

14. H. Malchau, P. Herberts, and L. Ahnfelt, ‘‘Prognosis oftotal hip replacement in Sweden. Follow-up of 92,675operations performed 1978–1990,’’ Acta Orthop. Scand.,64, 497–506 (1993).

15. R. G. Geesink and N. H. Hoefnagels, ‘‘Six-year resultsof hydroxyapatite-coated total hip replacement,’’ J.Bone Joint Surg., 77B, 534–547 (1995).

16. J. Karrholm, H. Malchau, F. Snorrason, and P. Herberts,‘‘Micromotion of femoral stems in total hip arthro-plasty. A randomized study of cemented, hydroxyapa-tite-coated, and porous-coated stems with roentgen ste-reophotogrammetric analysis,’’ J. Bone Joint Surg., 76A,1692–1705 (1994).

17. J. Karrholm, B. Borssen, G. Lowenhielm, and F. Snor-rason, ‘‘Does early micromotion of femoral stem pros-theses matter? 4–7-year stereoradiographic follow-upof 84 cemented prostheses,’’ J. Bone Joint Surg., 76B,912–917 (1994).

18. L. Ryd, B. E. Albrektsson, L. Carlsson, F. Dansgard, P.Herberts, A. Lindstrand, L. Regner, and S. Toksvig–Larsen, ‘‘Roentgen stereophotogrammetric analysis asa predictor of mechanical loosening of knee prosthe-ses,’’ J. Bone Joint Surg., 77B, 377–383 (1995).

19. S. Overgaard, K. Søballe, E. S. Hansen, K. Josephsen,and C. Bunger, ‘‘Implant loading accelerates resorptionof hydroxyapatite coating,’’ J. Orthop. Res., 14, 888–894(1996).

20. R. D. Bloebaum and J. A. Dupont, ‘‘Osteolysis from apress-fit hydroxyapatite-coated implant. A case study,’’J. Arthroplasty, 8, 195–202 (1993).

21. M. Jarcho, ‘‘Retrospective analysis of hydroxyapatitedevelopment for oral implant applications,’’ Dent. Clin.North. Am., 36, 19–26 (1992).

22. R. D. Bloebaum, D. Beeks, L. D. Dorr, C. G. Savory, J. A.Dupont, and A. A. Hofmann, ‘‘Complications with hy-droxyapatite particulate separation in total hip arthro-plasty,’’ Clin. Orthop., 298, 19–26 (1994).

23. P. Campbell, H. McKellop, S. H. Park, and A. Malcom,‘‘Evidence of abrasive wear particles from hydroxyapa-tite-coated hip prosthesis,’’ [Abstract], Trans. Orthop.Res. Soc., 18, 224 (1993).

24. P. Ducheyne, ‘‘Bioactive ceramics,’’ [Editorial], J. BoneJoint Surg., 76-B, 861–862 (1994).

25. H. W. Denissen, W. Kalk, H. M. Nieuport, C. Mangano,and J. C. Maltha, ‘‘Preparation-induced stability of bio-active apatite coatings,’’ Int. J. Prosthodont., 4, 432–439(1991).

26. C. P. A. T. Klein, J. G. Wolke, J. M. de Blieck–Hogervorst, and K. de Groot, ‘‘Calcium phosphateplasma-sprayed coatings and their stability: An in vivostudy,’’ J. Biomed. Mater. Res., 28, 909–917 (1994).

27. I. Heling, R. Heindel, and B. Merin, ‘‘Calcium-fluorapa-tite. A new material for bone implants,’’ J. Oral Implan-tol., 9, 548–555 (1981).

28. F. C. Driessens, ‘‘Relation between apatite solubility

150 OVERGAARD ET AL.

and anti-cariogenic effect of fluoride,’’ Nature, 243, 420–421 (1973).

29. E. Lugschneider, T. Weber, and M. Knepper, ‘‘Produc-tion of biocompatible coatings of hydroxyapatite andfluorapatite,’’ [Abstract], National Thermal Spray Confer-ence, Cincinnati, Ohio, 1988, pp. 332–343.

30. A. S. Posner, ‘‘The mineral of bone,’’ Clin. Orthop., 200,87–99 (1985).

31. M. Okazaki, T. Aoba, Y. Doi, J. Takahashi, and Y. Mori-waki, ‘‘Solubility and crystallinity in relation to fluo-ride content of fluoridated hydroxyapatites,’’ J. Dent.Res., 60, 845–849 (1981).

32. E. J. Duff and A. A. Grant, ‘‘Apatite ceramics for use inimplantation,’’ in Mechanical Properties of Biomaterials,’’G. W. Hastings and D. F. Williams (eds.), Wiley, NewYork, 1980, pp. 465–475.

33. S. D. Davis, D. F. Gibbons, R. L. Martin, S. R. Levitt, J.Smith, and R. V. Harrington, ‘‘Biocompatibility of ce-ramic implants in soft tissue,’’ J. Biomed. Mater. Res., 6,425–449 (1972).

34. T. W. Bauer, ‘‘Hydroxyapatite: Coating controversies,’’Orthopedics., 18, 885–888 (1995).

35. S. Overgaard, M. Lind, H. Glerup, S. Grundvig, C.Bunger, and S. Søballe, ‘‘Hydroxyapatite and fluorapa-tite ceramic coatings as fixation for weight-loaded im-plants,’’ Clin. Orthop., 336, 286–296 (1997).

36. M. J. Karnovsky, ‘‘A formaldehyde-glutaraldehydefixative of high osmolarity for use in electron micros-copy,’’ [Abstract], J. Cell Biol., 27, 137–138 (1965).

37. A. J. Baddeley, H. J. G. Gundersen, and L. M. Cruz–Orive, ‘‘Estimation of surface area from vertical sec-tions,’’ J. Microsc., 142, 259–276 (1986).

38. S. A. Goldstein, L. S. Matthews, J. L. Kuhn, and S. J.Hollister, ‘‘Trabecular bone remodeling: An experi-mental model’’ J. Biomech., 24, 135–150 (1991), [Pub-lished erratum appears in J. Biomech., 26, 367 (1993)].

39. K. Gotfredsen, E. Budtz–Jorgensen, and L. N. Jensen,‘‘A method for preparing and staining histological sec-tions containing titanium implants for light micros-copy,’’ Stain Technol., 64, 121–127 (1989).

40. D. B. Kimmel and S. S. J. Webster, ‘‘Measurements ofarea, perimeter, and distance: Details of data collectionin bone histomorphometry,’’ in Bone Histomorphometry:Techniques and Interpretation,’’ R. R. Recker (ed.), CRCPress, Inc., Boca Raton, Florida, 1983, pp. 89–108.

41. H. J. G. Gundersen and E. B. Jensen, ‘‘The efficiency ofsystematic sampling in stereology and its prediction,’’J. Microsc., 147, 229–263 (1987).

42. H. J. G. Gundersen, F. T. Bendtsen, L. Korbo, N. Mar-cussen, A. Møller, K. Nielsen, J. R. Nyengaard, B. Pak-kenberg, F. B. Sørensen, A. Vesterby, and M. J. West,‘‘Some new, simple and efficient stereological methodsand their use in pathological research and diagnosis,’’APMIS, 96, 379–394 (1988).

43. E. R. Weibel, Stereological Methods. Practical Methods forBiological Morphometry, Academic Press, Inc., NewYork, 1979.

44. C. P. A. T. Klein, P. Patka, J. G. C. Wolke, J. M. deBlieck–Hogervorst, and K. de Groot, ‘‘Long-term in vivostudy of plasma-sprayed coatings on titanium alloys oftetracalcium phosphate, hydroxyapatite and alpha-tricalcium phosphate,’’ Biomaterials, 15, 146–150 (1994).

45. S. H. Maxian, J. P. Zawadsky, and M. G. Dunn, ‘‘Effectof Ca/P coating resorption and surgical fit on thebone/implant interface,’’ J. Biomed. Mater. Res., 28,1311–1319 (1994).

46. J. E. Dalton and S. D. Cook, ‘‘In vivo mechanical andhistological characteristics of HA-coated implants vary

with coating vendor,’’ J. Biomed. Mater. Res., 29, 239–245(1995).

47. H. Caulier, J. P. van der Waerden, Y. C. Paquay, J. G.Wolke, W. Kalk, I. Naert, and J. A. Jansen, ‘‘Effect ofcalcium phosphate (Ca-P) coatings on trabecular boneresponse: A histological study,’’ J. Biomed. Mater. Res,29, 1061–1069 (1995).

48. T. P. Schmalzried, M. Jasty, and W. H. Harris, ‘‘Peri-prosthetic bone loss in total hip arthroplasty. Polyeth-ylene wear debris and the concept of the effective jointspace,’’ J. Bone Joint Surg., 74A, 849–863 (1992).

49. D. B. Kimmel and W. S. S. Jee, ‘‘A quantitative histo-logic study of bone turnover in young adult beagles,’’Anat. Rec., 203, 31–45 (1982).

50. K. Søballe, K. Gotfredsen, H. Brockstedt–Rasmussen,P. T. Nielsen, and K. Rechnagel, ‘‘Histologic analysis ofa retrieved hydroxyapatite-coated femoral prosthesis,’’Clin. Orthop., 272, 255–258 (1991).

51. D. C. Hardy, P. Frayssinet, A. Guilhem, M. A. Lafon-taine, and P. E. Delince, ‘‘Bonding of hydroxyapatite-coated femoral prostheses. Histopathology of speci-mens from four cases,’’ J. Bone Joint Surg., 73B, 732–740(1991).

52. F. Lintner, G. Bohm, M. Huber, and R. Scholz, ‘‘Histol-ogy of tissue adjacent to an HAC-coated femoral pros-thesis,’’ J. Bone Joint Surg., 76B, 824–830 (1994).

53. J. C. Dean, C. L. Tisdel, V. M. Goldberg, J. Parr, D.Davy, and S. Stevenson, ‘‘Effects of hydroxyapatite tri-calcium coating and intracancellous placement on boneingrowth in titanium fibermetal implants,’’ J. Arthro-plasty, 10, 830–838 (1995).

54. D. R. Sumner, T. M. Turner, R. M. Urban, and J. O. Gal-ante, ‘‘Remodeling and ingrowth of bone at two yearsin a canine cementless total hip-arthroplasty model,’’ J.Bone Joint Surg., 74A, 239–250 (1992). [Published erra-tum appears in J. Bone Joint Surg., 74A, 793 (1992).]

55. A. Vesterby, J. Kragstrup, H. J. Gundersen, and F.Melsen, ‘‘Unbiased stereologic estimation of surfacedensity in bone using vertical sections,’’ Bone, 8, 13–17(1987).

56. J. Galante, W. Rostoker, and R. D. Ray, ‘‘Physical prop-erties of trabecular bone,’’ Calcif. Tissue Res., 5, 236–246(1970).

57. W. J. Whitehouse, ‘‘The quantitative morphology of an-isotropic trabecular bone,’’ J. Microsc., 101, 153–168(1974).

58. J. W. Vahey, J. L. Lewis, and R. J. Vanderby, ‘‘Elasticmoduli, yield stress, and ultimate stress of cancellousbone in the canine proximal femur,’’ J. Biomech., 20,29–33 (1987).

59. S. H. Maxian, J. P. Zawadsky, and M. G. Dunn, ‘‘In vitroevaluation of amorphous calcium phosphate andpoorly crystallized hydroxyapatite coatings on tita-nium implants,’’ J. Biomed. Mater. Res., 27, 111–117(1993).

60. C. P. A. T. Klein, J. M. de Blieck–Hogervorst, J. G. C.Wolke, and K. de Groot, ‘‘Studies of the solubility ofdifferent calcium phosphate ceramic particles in vitro,’’Biomaterials, 11, 509–512 (1990).

61. G. Daculsi, R. Z. LeGeros, and D. Mitre, ‘‘Crystal dis-solution of biological and ceramic apatites,’’ Calcif. Tis-sue Int., 45, 95–103 (1989).

62. J. D. de Bruijn, Calcium Phosphate Biomaterials: Bone-Bonding and Biodegradation Properties, Ph.D. Thesis, Cip-Data Koninklijke Bibliotheek, Den Haag, 1993.

63. U. Fallon, ‘‘Alterations in the pH of osteoclast resorbingfluid reflects changes in bone degradative activity,’’[Abstract], Calcif. Tissue Int., 36, 458 (1984).

151RESORPTION OF HA AND FA CERAMIC COATINGS

64. G. Vaes, ‘‘Cellular biology and biomechanical mecha-nism of bone resorption,’’ Clin. Orthop., 231, 239–271(1988).

65. M. Ogura and J. E. Davies, ‘‘Resorption of calcium hy-droxyapatite substrata by osteoclast-like cells in vitro,’’in Bioceramics, Vol. 4, Proceedings of the 4th InternationalSymposium on Ceramics in Medicine, W. Bonfield, G. W.Hastings, K. E. Tanner (eds.), Butterworth–HeinemannLtd, London, 1991, pp. 121–126.

66. C. M. Muller–Mai, C. Voigt, and U. Gross, ‘‘Incorpora-tion and degradation of hydroxyapatite implants of dif-ferent surface roughness and surface structure inbone,’’ Scanning Microsc., 4, 613–622 (1990).

67. I. Orly, M. Gregoire, J. Menanteau, M. Heughebaert,and B. Kerebel, ‘‘Chemical changes in hydroxyapatite

biomaterial under in vivo and in vitro biological condi-tions,’’ Calcif. Tissue Int., 45, 20–26 (1989).

68. M. Gregoire, I. Orly, L. M. Kerebel, and B. Kerebel, ‘‘Invitro effects of calcium phosphate biomaterials on fibro-blastic cell behavior,’’ Biol. Cell, 59, 255–260 (1987).

69. K. Gomi, B. Lowenberg, G. Shapiro, and J. E. Davies,‘‘Resorption of sintered synthetic hydroxyapatite by os-teoclasts in vitro,’’ Biomaterials, 14, 91–96 (1993).

70. T. W. Bauer, ‘‘The histology of HA-coated implants,’’ inHydroxylapatite Coatings in Orthopaedic Surgery, R. T. G.Geesink and M. T. Manley (eds.), Raven Press, NewYork, 1993, pp. 305–318.

71. C. C. Ahn and O. L. Krivanek, eds., EELS Atlas, GantanInc., Warrendale, PA, 1983.

152 OVERGAARD ET AL.