A canine model was used to compare autogenous alveolar ridge augmentation bone grafting with allogeneic grafts. Defects were created by premolar extractions and measured by radiopaque markers. These markers were used for subsequent measurements made before and after grafting, and after animal sacrifice to evaluate the status of the grafted sites. The results were unexpected and disappointing.

ORAL IMPLANTOLOGYThursday May 15 2003 11:22 AMAllen Press x DTPro SystemA COMPARISON OF ALLOGENEIC ANDAUTOGENOUS ILIAC MONOCORTICAL GRAFTSTO AUGMENT THE DEFICIENT ALVEOLARRIDGE IN A CANINE MODEL.I. CLINICAL STUDYCLINICALA. Norman Cranin, DDS, DEng,Edmond Demirdjan, DDSRobert DiGregorio, PharmDKEY WORDSAutogenousAllogeneicMonocortical onlay graftDemineralized freeze-dried boneCanine modelA. Norman Cranin, DDS, DEng, is a clinicalprofessor of Oral and Maxillofacial Surgeryand Implant Dentistry, NYU College ofDentistry, New York, NY, and is the editor-inchiefof the Journal of Oral Implantology.Address correspondence to Dr Cranin at 232Kelly Boulevard, Staten Island, NY 10314(e-mail: [email protected]).Edmond Demirdjan, DDS, and RobertDiGregorio, PharmD, are with the BrookdaleUniversity Hospital and Medical Center,Brooklyn, NY.124 Vol. XXIX/No. Three/2003orim 29_301 Mp_124File # 01emA canine model was used to compare autogenous alveolar ridge augmentationbone grafting with allogeneic grafts. Defects were created by premolarextractions and measured by radiopaque markers. These markers were used forsubsequent measurements made before and after grafting, and after animalsacrifice to evaluate the status of the grafted sites. The results were unexpectedand disappointing.im results, n f o n cti nstru reco o dges r t e d in ced Aplants plaedentulousficienridge. Diental appusroaches. Althooughtesuccessfulthosseetern of bone loss of the the time of grafting or by subseqe ith e a l p are p im es, ran b e th d v haloss.eLekhescribedolm and Zarbpat-1 added benefitslantsoffered bycedbarriermureemnrae th W e r t ficien de ely v s tooth o e uenc co uralNTRODUCTION Ilveolarnseqbone loss isfa nat- aetecerhnique often usedidgto.4,5augmithent thee-tthas been criticism of these techniques,including inconsistent bony fill,8 questionablelevels of osseointegration,3 instabilityof the newly generated bone,9and infection associated with dehiscenceof the membranes.4,6,10,11Because of these potential and actualproblems, the use of autogenousmonocortical onlay grafts inserted priorto implant placement has been advocatedas one viable method toachieve predictable bony enlargementshave higher failure rates than thoseplaced in ridges with more bone.Whensuccessful, esthetic results may be unsatisfactoryor the implants may besubjected to force vectors that arethought to be deleterious to long-termimplant survival.2 To counteract thesepotentially harmful results, alveolarridge augmentation either prior to or atthe time of implant placement has beenadvocated.3-7Guided bone regeneration (GBR) isORAL IMPLANTOLOGYThursday May 15 2003 11:22 AMAllen Press x DTPro Systemorim 29_301 Mp_125File # 01emA. Norman Cranin et alFIGURES 1-5. FIGURE 1. At the time of the edentulating procedures, the tissues were opened and titanium screws were placed 6 mmbeneath the crestal levels to serve as markers for postoperative measurements. FIGURE 2. (a) and (b) Inverted cone burs were used toprepare for the insertion of amalgam markers at the cervices of the molars and canines adjacent to the edentulous study areas. FIGURE 3.Primary closures were achieved after placing the markers. FIGURE 4. Healing at all 20 sites was uneventful. FIGURE 5. (a), (b), (c), and (d)Twelve weeks later, incisions were made overlying the posterior iliac crests, and sharp and blunt dissection was employed in order toreveal the periosteum. After its reflection, burs and osteotomes were used to harvest cortico-medullary blocks. Closures were achievedusing skin staples.Journal of Oral Implantology 125possibilities that these blocks mightserve as alternatives to autogenousmonocortical blocks for ridge augmentation.AimORAL IMPLANTOLOGYThursday May 15 2003 11:22 AMAllen Press x DTPro SystemCOMPARISON OF GRAFTS. I. CLINICAL STUDYto create optimal sites for implantation.7,12 A shortcoming of this technique,however, is the need for a secondsurgical (donor) site and the morbidityassociated with this surgery. Demineralizedallogeneic monocorticalilium blocks have become available.The ability of demineralized allogeneicbone to promote bony ingrowth hasbeen well documented.13,14 Therefore,information was sought regarding the126 Vol. XXIX/No. Three/2003orim 29_301 Mp_126File # 01emFIGURES 6-8. FIGURE 6. (a) and (b) Immediately following the harvesting, the 4 quadrants of each experimental animal were exposedsurgically. Prior to affixing the grafts to the 4 quadrants of edentulous ridges, both horizontal and vertical measurements were made inorder to record dimensional changes. FIGURE 7. (a) and (b) Both allogeneic and autogenous bone blocks were soaked in saline/blood andgentamicin solution and then were affixed with lag screws to the underlying host sites. FIGURE 8. After sacrifice the animals were perfusedwith 10% neutral formalin via the dissected carotid arteries and jugular veins.mandibular alveolar ridges in a dogmodel. Clinical and histologic examinationswere used to compare the resultsof using the 2 graft materials.MATERIALS AND METHODSFive mongrel dogs, each 25 kg or greater,were used. The 4 quadrants of eachanimal were used for 20 quadrants, 18of which were experimental; the remaining2 served as controls. ThreeThe aim of this study was to examineand compare the behavior of autogenousmonocortical blocks as comparedwith demineralized allogeneic monocorticalblocks to augment deficientORAL IMPLANTOLOGYThursday May 15 2003 11:22 AMAllen Press x DTPro Systemdistinct stages of the study were delineated:(1) dental extraction, (2) augmentation,and (3) sacrifice. Prior toeach procedure, the dogs were anesthetizedwith 30 mg/kg of intravenouspentobarbital sodium, intubated, andmaintained with oxygen and a continuousD5W drip. Boli of barbituratewere added as necessary to maintainanesthesia.In each of the 20 quadrants, a fullthicknessmucoperiosteal flap was elevatedand the 4 premolar teeth fromthe mandible and 3 from the maxillaewere extracted using the classicalhemisection technique.15 At a point approximatelymidway between the canineand molar teeth in each quadrant,a single 4 3 1.6-mm titanium bonescrew was placed 6 mm apical to thenewly created alveolar crest on thebuccal surface. The head of each screwwas countersunk in order to avoid irritationto the overlying soft tissues.Additionally, 2-mm diameter amalgamfillings were made at the cervices of thecanine and molar teeth (Figures 1, 2aand b). The titanium screws and theamalgam fillings served as referencepoints to mark changes in bone dimensionsover time. Primary closure wasthen achieved using 4-0 Vicryl (Johnsonand Johnson, New Brunswick, NJ)continuous horizontal mattress sutures(Figure 3).The dogs were observed postoperativelyin the recovery area until theywere safely reactive, at which pointthey were returned to their cages. Asingle postoperative intramusculardose of 600 000 units of Bicillin wasgiven to each in the gluteus muscle.Pain control was achieved with Buprenex0.3 mg IM every 6 hours for 3 days,followed by doses as needed. Each dogwas maintained on a pureed diet for 2weeks, followed by a regular diet ofdog chow.A 12-week period was allowed toelapse to permit initial alveolar resorptionand healing (Figure 4). After thattime, each of the 18 experimental quadrantswas grafted. Ten quadrants receivedan allogeneic graft, and 8 quadrantsreceived an autogenous graft.Each of the dogs was anesthetized aspreviously described, and harvestingof grafts from the posterior iliac crestswas performed. Each dog was securedto the operating table in a prone positionwith both legs extended caudally.Under strict sterile conditions, the areaoverlying the right posterior iliac crestwas shaved, prepared with Betadine,and draped for isolation. A 2% lidocainesolution with 1:50 000 epinephrinewas used to infiltrate the surgicalsite. Using a number 10 scalpel blade,a 6-cm incision was made over the iliaccrest. The subcutaneous tissue, glutealfat, and thoracolumbar fascia were incisedusing electrocautery. An incisionwas made through the crestal origin ofthe middle gluteal muscle on the tubersacral starting at the cranial dorsal iliacspine and extending caudally for 4 cm.An elevator was used to strip the periosteumalong with the overlying middleand deep gluteal muscles laterallyand the iliocostalis and longissimuslumborum muscles medially. The crestand the wing of the ilium were thenexposed inferiorly for a distance of 3cm. Using dental drills with copioussaline irrigation, two 2 3 1.5-cm osteotomieswere made and the monocorticalblocks created were elevated by usingan osteotome as a lever (Figure 5,a-d). The harvested blocks were storedin a sterile container containing normalsaline to which was added 80 mg ofgentamicin sulfate, marrow blood, andmedullary bone that had been retrievedat operation using a gouge. Afterestablishing hemostasis at the donorsite, primary wound closure wasachieved in layers using 3-0 Vicryl suturesfor the deeper layers and staplesfor the skin. Bacitracin ointment wasapplied over the incision prior to dressingit.The Pacific Coast Tissue Bank (LosAngeles, Calif) provided canine allogeneicbone in the form of demineralizedsplit ilium blocks and demineralizedcortical powder. It had been sterilizedusing ethylene oxide gas andfreeze dried in the same manner as huorim29_301 Mp_127File # 01emA. Norman Cranin et alman allografts. Physically and dimensionally,it was clearly comparable withthe autogenous specimens that hadbeen harvested. The manufacturer'srecommended protocol for rehydrationwas used just prior to surgery. The allograftswere transferred to sterile containerscontaining normal saline with80 mg of gentamicin sulfate added atroom temperature. The solution waschanged twice to remove residues ofgas sterilization. The bone powder wasrehydrated for 30 minutes and theblocks for 3 hours.In all of the 18 experimental quadrants,midcrestal incisions were madewith 2 vertical releasing incisions onthe facial surfaces and full-thicknessmucoperiosteal flaps were elevated. Atthis point 4 measurements were madeprior to grafting: (1) the distance betweenthe base of the head of the markerscrew and the buccal plate to determinehorizontal bone loss; (2) the distancefrom the crest of the ridge to thecenter of the marker screw to determinevertical bone loss in the center ofthe edentulous segment; and (3) and(4) the distances from the bases of themarker amalgams in the canine andfirst molar to the crest of the ridge todetermine vertical bone loss adjacentto these teeth (Figure 6a and b). Uponcompletion of recording the bone lossof the residual alveolar ridge, the graftingprocesses were initiated. All recipientsites were prepared to receive thegrafts by perforating the buccal corticalplates using a one-half round bur in ahigh-speed handpiece at multiple sites(from 16 to 20 times) in order to stimulatebone marrow bleeding. The mucosalflaps then were undermined toseparate them from the underlyingperiosteum, which ensured adequateflap mobility for tension-free primaryclosures. Following this, the host siteswere grafted with either allogeneic orautogenous bone as indicated in Table1.The donor tissues were trimmedand smoothed as required so that theywould fit the host bone beds with minimalunderlying deficiencies. SingleJournal of Oral Implantology 127TABLE 1Left Maxilla26101418ORAL IMPLANTOLOGYThursday May 15 2003 11:22 AMAllen Press x DTPro SystemCOMPARISON OF GRAFTS. I. CLINICAL STUDYQuadrant number and graft type*Right Maxilla Dog15139123417 5*Experimental quadrantsautogenous graft: 1, 4, 5, 7, 11, 12, 16, 17, 20. Experimentalquadrantsallogeneic graft: 2, 3, 6, 8, 9, 10, 14, 15, 18. Control quadrantsno graft: 13,19.holes were placed through the boneblocks with burs. They were affixed tothe host bone deeply enough to assurepositive retention with 2.0 3 6.0-mmand 2.0 3 8.0-mm lag screws.16 Smallperipheral deficiencies were filled withautogenous medullary chips or allogeneicdemineralized powder as appropriate(Figure 7a and b).Measurementswere taken to record the amount ofhorizontal and vertical augmentationusing the previously placed amalgammarkers and screws, the buccal plates,and the alveolar crests as references.The flaps were closed using 4-0 Vicrylhorizontal mattress sutures.The dogs were observed in the recoveryarea until they were safely reactive,at which point they were returnedto their cages. They receivedpostoperative IM Bicillin 600 000 unitsin their gluteus muscles. Pain controlconsisted of Buprenex 0.3 mg every 6hours over the course of 3 days followedby doses as required. For the remainderof the study the dogs werekept on soft diets. The skin stapleswere removed after 14 days.After 6 months of healing the dogswere sacrificed. Under general anesthesiathe common carotid and anteriorjugular veins were catheterized bilaterally.Sacrifice was completed by theadministration of an overdose of pentobarbitaland potassium chloride. Thiswas followed by saline flush exsanguinationof the carotid arteries using 110mm of Hg pressure with a return viathe jugular veins and finally with theperfusion of 10% neutral formalin untila return was noted, at which time thevessels were ligated (Figure 8).15128 Vol. XXIX/No. Three/2003Right Mandible Left Mandible4812371115 1620 19The 4 jaw quadrants of each dogwere resected and intact ridge measurementswere recorded; then, midcrestalincisions were made with 2 verticalreleasing incisions on the facialsurfaces, and full thickness mucoperiostealflaps were reflected. Measurementswere taken using amalgammarkers and screws as well as the buccalplates and the alveolar crests as referencesin the same manner as prior tografting. Examination was done forquantitative evaluations between theimmediate preaugmentation and postsacrificelevels of mineralization andosseous integration of all the graftedsites (Figure 9a and b). Each of thespecimens then was fixed in 10% neutralformalin, demineralized with formicacid, embedded in nitrocellulose,and sectioned serially in a buccolingualdirection. Histologic examinationswere done with light microscopy afterH E staining; these results will befound in Part II of this article.RESULTSDifferences in the outcome of thisstudy based on the 2 selected graft materialswere determined by comparingthe height and width of the grafts in18 grafted experimental quadrants and2 nongrafted control quadrants and byanalyses of histologic specimens. Measurementswere made (1) prior tografting, (2) immediately after grafting,and (3) upon sacrifice of each hostanimal. Calculations were made to determine(1) the volume of augmentationsites (immediate postgraft), (2) theamount of graft resorption (postsacri-fice), and (3) the percent of grafted maorim29_301 Mp_128File # 01emterial that had become integrated tothe host bone in regard to four references:(a) width, (b) height at center ofedentulous section, and (c) and (d)height at terminal ends of the edentuloussection adjacent to the canine andmolar teeth. Measurements of resorptionat the time of sacrifice comparedthe experimental quadrants with thecontrol quadrants. Graft integrationpercentages were recorded among theexperimental quadrants. Statisticallysignificant differences were determinedusing a repeated measures analysisof variance (ANOVA; SPSS Win7.0) with a significance level set at 0.05.Prior to grafting there were no statisticallysignificant differences in theamount of alveolar resorption betweenany of the 20 quadrants at any of the 4reference points. Resorption of bonewas measured clinically after sacrificeas well and it was determined thatagain after eliminating the extremesthere were no statistically significantdifferences among the control and experimentalquadrants, which demonstratedthat there was no significantvariation of augmentation levels regardlessof the graft material used (Table2).In a subgroup analysis, it was notedthat the quantity of surviving graftmaterial varied in respect to the differentreference points. The center of theedentulous segments had retainedmore of the graft material than the peripheralareas immediately adjacent tothe teeth. Upon sacrifice, the percentageof graft volume, however, was notstatistically significantly different betweenthe types of grafts concerning(1) mandible vs maxilla, (2) width augmentationvs height augmentation, and(3) edentulous section of ridge vs areasadjacent to teeth (Table 2).In addition to the measurementsdescribed, the gross and subjective appearanceof each graft was rated on a4-point categorical scale: (1) ''good,''(2) ''fair,'' (3) ''mostly resorbed,'' and(4) ''completely resorbed.'' A graftedsite was considered good if there wasclear evidence of a substantial amountORAL IMPLANTOLOGYThursday May 15 2003 11:22 AMAllen Press x DTPro Systemof surviving graft material at sacrifice,and fair if there was clear evidence ofgraft material but not in substantialamounts. Mostly resorbed describedthose quadrants where only a hint ofsurviving graft volume existed, andthose quadrants that exhibited no signsof any surviving graft material wereincluded in the completely resorbedgroup. A good or fair outcome wasconsidered a success; those areas thatwere mostly or completely resorbedwere considered as failures (Figures10a and b).Four grafts were determined to besuccessful; 3 of them were allogeneicand were rated as good, and the fourthwas autogenous and was rated as fair.Five of the grafts were completely resorbedand 9 were mostly resorbed, fora total of 14 grafts being consideredfailures. Although there were moreFIGURES 9-10. FIGURE 9. (a) and (b) These postoperative radiographs were exposed at the sixth month. Both maxilla and mandible revealthe presence of the amalgam and titanium screw markers as well as the lag screws that had been used to affix the grafts. Poor mineralizationis reflected by these films; this is a consistent early phenomenon found to be associated with humans who have undergone similarprocedures. FIGURE 10. (a) and (b) These specimens, studied after they were resected from the jaws, reveal both success and failure. Themandibular quadrant has all of the screws and markers in place and little to indicate loss of dimensions. The maxillary quadrant showsa nonviable separating graft in a seriously resorbed site. Both were allogeneic sites.orim 29_301 Mp_129File # 01emA. Norman Cranin et alfore we had considered autogenousbone to be the gold standard whenused in human grafting procedures,and retrospective reviews have indicatedhigh percentages of success, someaspects of this study require closerevaluation. The demineralized freezedriedbone grafts did only slightly betterthan the autogenous specimensand, with few exceptions, the gross appearancesof the host site interfaces ofboth showed disappointing results.This is not typical of our clinical observationson the use of these materialsin uncontrolled and random humanstudies.16 A review of the clinical resultsin this study indicated a slightpreference for operative site: 3 werefound to be successful in maxillae(where the vascularity was found to bemore abundant), and 1 was consideredsuccessful in mandibles.Journal of Oral Implantology 129successful outcomes in gross appearancein the allogeneic group, a greaternumber of these grafts were determinedto have completely resorbed ascompared with the autogenous group.When compared statistically in the -i2test (SPSS Win. 7.0), there were no statisticallysignificant differences in thegross appearances of the grafts.DISCUSSIONThe most interesting aspect of thisstudy is the high percentage of partlyand completely resorbed grafts. Despitethe fact that prior to mucoperiostealelevation all but 2 quadrants appearedto have presented favorableridge forms, only 4 of 18 study siteswere adjudged to be successful on abasis of clinical evaluation of the exposedridges. Of these, 75% representedthe allogeneic group. Since hereto-TABLE 2Mucosanormalnormalnormalnormalknife-edgenormalnormalnormalnormalnormalnormalnormalnormalknife-edgenormalnormalnormalnormalnormalnormalavailable to the Brookdale ImplantGroup from the John and FlorenceLawrence Foundation and the BaraschFamily Foundation.1. Lekholm U, Zarb GA. Patient selectionand preparation. In: BranemarkP-I, Zarb GA, Albrektsson T, eds. Tissue-Integrated Prostheses: Osseointegrationin Clinical Dentistry. Chicago:Quintessence; 1985:199-209.2. Jemt T, Lekholm U. Implant treatmentin edentulous maxillae: a five-yearfollow-up report on patients with differentdegrees of jaw resorption. Int JOral Maxillofac Implants. 1995;10:303.3. Becker W, Schenk R, Higuchi K,et al. Variations in bone regeneration adjacentto implants augmented with barriermembranes alone or with demineralizedfreeze-dried bone or autologousgrafts: a study in dogs. Int J OralMaxillofac Implants. 1995;10:143-154.4. Buser D, Bragger U, Lang NP, etORAL IMPLANTOLOGYThursday May 15 2003 11:22 AMAllen Press x DTPro SystemCOMPARISON OF GRAFTS. I. CLINICAL STUDYClinical resultsmucosa-covered and denuded*Dog Quadrant Graft TypeautogenousallogeneicallogeneicautogenousautogenousallogeneicURULLLLRURULautogenousallogeneicallogeneicallogeneicautogenousautogenousLLLRURULLLLRcontrolallogeneicallogeneicautogenousautogenousallogeneicURULLLLRURULcontrolautogenousLLLR11112222333344445555*Mucosal refers to ridge appearance prior to stripping the mucoperiosteum. Denudedrefers to clinical appearance of ridge after soft tissue reflection. Percent change of size fromoriginal healed edentulous ridge to same site at the end of the study (6 10%). UR indicatesupper right; UL, upper left; LL, lower left, LR, lower right. The rating was assigned to eachquadrant using criteria cited in Methods: good, fair (successful); partly resorbed (poor);completely resorbed (failed).Percent change of graft after denuding, as compared with pregraft volume, as measuredwith calipers.Integration of graft to host bone was observed at sites marked ''good'' as 50%-80%; atsites marked ''fair,'' as 20%-49%.Dogs heal well and about 3 timesmore rapidly than humans do,17 andthey have served well in other implantrelatedstudies.18 The low levels of successhere may be attributed to cage biting,which was noted among the animalsin this group despite the extensiveuse of effective analgesics for 72postoperative hours. The use of theirpaws appeared to be another deleteriousfactor that may have contributed tofailure. Further Discussion, Summary,and Conclusions will follow Part II ofthis article.ACKNOWLEDGMENTSDr John Ley prepared this researchprotocol and was a member of the earlyoperating team. The participation ofDrs Michael Katzap and Colleen Watsonis acknowledged. Appreciation isextended to Effem Moghilevsky, DVM.This study was made possible in partby a generous grant from the PacificCoast Tissue Bank and from funds130 Vol. XXIX/No. Three/2003%Change Comment Denudedbonyresorbedbonyfibrousfriablebonyfairfailedgoodpoorpoorgood130210190220220140friableresorbedbonymixedfibrousmixedfailedpoorgoodfailedpoorpoor1201201500110120bonyfriablefibrousmixedknife-edgefibrousgoodfailedpoorpoorpoorpoor0250220130220120bonyfibrousgoodfailed0280REFERENCESorim 29_301 Mp_130File # 01emal. Regeneration and enlargement of jawbone using guided tissue regeneration.Clin Oral Implant Res. 1990;1:22-32.5. Buser D, Dula K, Belser U, et al.Localized ridge augmentation usingguided bone regeneration. I. Surgicalprocedure in the maxilla. Int J PeriodontRest Dent. 1993;13:29-45.6. Jovanovic SA, Spiekermann H,Richter EJ. Bone regeneration aroundtitanium dental implants in dehisceddefect sites: a clinical study. Int J OralMaxillofac Implants. 1992;7:233-245.7. Misch CM. Comparison of intraoraldonor sites for onlay grafting priorto implant placement. Int J Oral MaxillofacImplants. 1997;12:767-776.8. Palmer RM, Floyd PD, Palmer PJ,et al. Healing of implant dehiscence defectswith and without expanded polytetrafluorethylenemembranes: a controlledclinical and histological study.Clin Oral Implant Res. 1994;5:98-104.9. Becker W, Lekholm U, Dahlin C,et al. The effect of clinical loading onbone regenerated by GTAM barriers: astudy in dogs. Int J Oral Maxillofac Implants.1994;9:305-313.10. Jovanovic SA, Schenk RK, OrsiniM, et al. Supracrestal bone formationaround dental implants: an experimentaldog study. Int J Oral MaxillofacImplants. 1995;10:23-31.11. Simon M, Baldoni M, Rossi P,et al. A comparative study of the effectivenessof e-PTFE membranes withand without early exposure during thehealing period. Int J Periodont Rest Dent.1994;14:167-180.12. Misch CM, Misch CE. The repairof localized severe ridge defectsfor implant placement using mandibularbone grafts. Implant Dent. 1996;4(4):261-267.13. Narang R, Ruben MP, HarrisMH, Wells H. Improved healing of experimentaldefects in the canine mandibleby grafts of decalcified allogeneicbone. Oral Surg Oral Med Oral Pathol.1970;30:142-150.14. Narang R, Wells H. Stimulationof new bone formation on intact bonesby decalcified allogeneic bone matrix.ORAL IMPLANTOLOGYThursday May 15 2003 11:22 AMAllen Press x DTPro SystemOral Surg Oral Med Oral Pathol. 1971;32:668-676.15. Cranin AN, Sirakian A, RussellD, Klein M. The role of incision designand location in the healing of alveolarridges and implant host sites. Int J OralMaxillofac Implants. 1998;3:483-491.16. Cranin AN, Klein M, SimonsA. Atlas of Oral Implantology. 2nd ed. St.Louis: Mosby; 1999:160-162.orim 29_301 Mp_131File # 01emA. Norman Cranin et al17. Roberts WE. Bone tissue interface.J Dent Ed. 1988;52:804-809.18. Cranin AN. Dog v man. Presentationmade at NYU College of Dentistry:October 14, 2001; New York, NY.Journal of Oral Implantology 131