J Periodontol • April 2005

605

* Department of Oral Medicine, Infection and Immunity, Harvard School of DentalMedicine, Boston, MA.

† Department of Periodontics, University of Texas Health Science Center at San Antonio,San Antonio, TX.

‡ Private practice, Portland OR.§ Wyeth/Genetics Institute, Cambridge, MA.� University Oral and Maxillofacial Surgery, Charlotte, NC.¶ Section of Dentistry, University of Chicago, Chicago, IL.# Institute for Advanced Dental Studies, Swampscott, MA.

In dentistry, the primary indication forthe reconstruction or preservation ofosseous structures following disease or

trauma has been the rehabilitation ofedentulous ridges to facilitate the inser-tion of endosseous implants. As bone vol-ume increases, longer and wider implantscan be utilized to enhance success rates.Currently, guided bone regeneration uti-lizing various materials has been consid-ered standard to obtain adequate bonevolume.1 During the past several years,the application of recombinant technolo-gies has included biomimetic devices thatstimulate the replacement of anatomicstructures.2-6 These promote the in vitroor in vivo development of tissues. A groupof molecules, the bone morphogeneticproteins (BMPs), members of the trans-forming growth factor-β superfamily, havebeen shown to induce heterotopic boneformation. New bone growth results fromthe differentiation of pluripotent mes-enchymal cells along osteoblastic path-ways. The purpose of this study was toevaluate the efficacy of bone induction forthe placement of dental implants by twoconcentrations of recombinant humanBMP-2 (rhBMP-2) delivered on an absorb-able collagen sponge (ACS) compared toplacebo (ACS alone) and no treatment ina human buccal wall defect model follow-ing tooth extraction.

MATERIALS AND METHODSStudy DesignThe Institutional Review Board at eachof the eight study centers approved the

Randomized Study Evaluating RecombinantHuman Bone Morphogenetic Protein-2for Extraction Socket AugmentationJoseph P. Fiorellini,* T. Howard Howell,* David Cochran,† Jay Malmquist,‡ Leslie C. Lilly,§Daniel Spagnoli,� Joseph Toljanic,¶ Archie Jones,† and Myron Nevins*#

Background: Conventional dentoalveolar osseous reconstruc-tion often involves the use of grafting materials with or withoutbarrier membranes. The purpose of this study was to evaluate theefficacy of bone induction for the placement of dental implantsby two concentrations of recombinant human bone morphogeneticprotein-2 (rhBMP-2) delivered on a bioabsorbable collagen sponge(ACS) compared to placebo (ACS alone) and no treatment in ahuman buccal wall defect model following tooth extraction.

Methods: Eighty patients requiring local alveolar ridge augmen-tation for buccal wall defects (≥50% buccal bone loss of the extrac-tion socket) of the maxillary teeth (bicuspids forward) immediatelyfollowing tooth extraction were enrolled. Two sequential cohortsof 40 patients each were randomized in a double-masked man-ner to receive 0.75 mg/ml or 1.50 mg/ml rhBMP-2/ACS, placebo(ACS alone), or no treatment in a 2:1:1 ratio. Efficacy wasassessed by evaluating the amount of bone induction, the ade-quacy of the alveolar bone volume to support an endosseousdental implant, and the need for a secondary augmentation.

Results: Assessment of the alveolar bone indicated that patientstreated with 1.50 mg/ml rhBMP-2/ACS had significantly greaterbone augmentation compared to controls (P ≤0.05). The adequacyof bone for the placement of a dental implant was approximatelytwice as great in the rhBMP-2/ACS groups compared to no treat-ment or placebo. In addition, bone density and histology revealedno differences between newly induced and native bone.

Conclusion: The data from this randomized, masked, placebo-controlled multicenter clinical study demonstrated that the novelcombination of rhBMP-2 and a commonly utilized collagen spongehad a striking effect on de novo osseous formation for the place-ment of dental implants. J Periodontol 2005;76:605-613.

KEY WORDSAlveolar bone loss/surgery; alveolar ridge augmentation;bone regeneration; clinical studies, prospective; clinical trials,controlled; clinical trials, randomized; collagen/therapeuticuse; multicenter studies; proteins, bone morphogenetic; bonemorphogenetic protein-2, recombinant human.

4020.qxd 4/18/05 8:58 AM Page 605

606

Effect of rhBMP-2 on Osseous Defect Healing Volume 76 • Number 4

study protocol. Participants signed a consent form andmet specific inclusion and exclusion criteria. Eightypatients requiring local alveolar ridge preservation/aug-mentation of buccal wall defects (≥50% buccal boneloss of the extraction socket) following extraction ofmaxillary teeth (bicuspids forward) were enrolled inthis study. Two sequential cohorts of 40 patients eachwere randomized in a double-masked manner toreceive 0.75 mg/ml or 1.50 mg/ml rhBMP-2/ACS,placebo (ACS alone), or no treatment in a 2:1:1 ratio.

Efficacy ParametersThe efficacy of rhBMP-2/ACS was assessed by the fol-lowing computed tomography (CT) parameters.3,4,6

Bone induction. The efficacy endpoints that wereused to measure bone induction (Fig. 1) included alve-olar bone height (one measurement) and bone width(three measurements at 25%, 50%, and 75% of theextraction socket length [ESL]). These measurementswere taken from CT scans exposed at baseline (within4 days following study treatment) and at 4 monthsfollowing study treatment. A bone height response(4 months minus baseline values) and bone widthresponse (4 months minus baseline values) werecalculated.

Adequate alveolar bone volume for dental implantplacement. The CT scan efficacy endpoints that wereused to measure whether there was adequate alveolarridge bone to support the placement of an endosseousdental implant at 4 months following treatment wereheight (one measurement) and width (one measure-ment). Adequate alveolar bone was defined as ≥6 mmin width at the narrowest point (buccal to palatal) and≥12 mm in height. In addition, the proportion of patientswho required a secondary augmentation procedure wascalculated.

Bone density. The bone density (mg/cc) of the nativebone and treatment area was assessed by CT scanusing a standard density block at baseline and at4 months. Native bone density was measured at base-line by identifying three overlapping areas within theadjacent bone. An average of these three measure-ments was calculated. For the treatment areas, twooverlapping areas were averaged.

The reliability of measurements made among threeindependent masked CT scan reviewers was examined.The intraclass correlation coefficient of the measure-ments of bone height, width, and density ranged from0.64 (bone width at 25% ESL at 4 months post-treat-ment) to 0.96 (bone height at baseline and 4 monthspost-treatment). Correlation coefficient indicated rela-tively good (0.64) to excellent (0.96) correlation.

rhBMP-2/ACSThe two components of rhBMP-2/ACS are rhBMP-2**and ACS.†† The ACS carrier component was bovine

Type I collagen and provides the matrix for the deliv-ery of the osteoinductive factor, rhBMP-2. After recon-stitution and dilution, the rhBMP-2 was evenly expressedonto the ACS. For placebo, the diluent only for rhBMP-2 was expressed onto the ACS. The total mean dose(mg) per defect site was calculated.

Surgical ProtocolAll patients in the study underwent the same surgicalprocedure, regardless of the treatment assignment.Enrolled subjects received one dose of preoperativeantibiotics and 0.12% chlohexidine rinse (15 ml). Fol-lowing local anesthesia, sulcular and vertical incisionswere utilized to release full-thickness periosteal flaps.The teeth were extracted and the extraction socket(s)debrided. Four to eight perforations of the socket wallcortical plates were made with a 1/2 round bur. Theplacebo or rhBMP-2–soaked ACS were cut into stripsand placed to fill the defect sites. A larger strip ofmaterial was then placed over the entire treatmentsite. A tension-free soft tissue wound closure wasestablished. The no treatment procedure did notinclude placement of the rhBMP-2/ACS or ACS. Post-operative care included analgesics, a 7- to 10-daycourse of oral antibiotics (penicillin or cephalosporinsfor penicillin-sensitive subjects), and twice daily 0.12%chlohexidine rinse.

Methodology for Histology AnalysisAt the time of dental implant insertion, biopsies weretaken by using a trephine drill in the long axis of thetreatment site which was identified by baseline prob-ing measurements from adjacent teeth and/or a stent.The biopsies were fixed, decalcified, and embedded inparaffin. Each biopsy was evaluated for the presenceof cortical and/or trabecular bone, the thickness ofosseous trabeculae, the presence of lamellar andwoven bone, and residual collagen matrix from theACS.

Safety AssessmentOral wound examinations of the study treatment siteswere performed at baseline; days 2 and 14; and 1, 2,3, and 4 months postoperative to monitor for the occur-rence of commonly seen postoperative complicationsassociated with augmentation procedures (edema, ery-thema, exudate, hematoma, sensory loss, pain, andwound dehiscence). Periapical radiographs were takenat baseline and 2, 3, and 4 months post-treatment tomonitor for overexuberant bone formation. Serumchemistry and hematology were performed at baseline,at day 2, and 1 month postoperative. In addition, for-mation of antibiodies to rhBMP-2 or to human or bovine

** Genetics Institute, Cambridge, MA.†† Helistat, Integra LifeSciences, Plainsboro, NJ.

4020.qxd 4/18/05 8:58 AM Page 606

607

J Periodontol • April 2005 Fiorellini, Howell, Cochran, et al.

Figure 1.A) Computed tomography height and width measurements at baseline following tooth extraction (A) and 4 months post-treatment (B).

Type I collagen was evaluated by enzyme-linkedimmunosorbent assay (ELISA) at baseline and 1, 2,and 4 months postoperative.

Statistical ConsiderationsFor calculation of sample size, it was assumed 15% ofthe subjects would be lost to follow-up. The proportionof subjects with adequate bone formation for dentalimplants was evaluated. Depending on a comparisonof 20 treated subjects to 10 or 20 (pooled) no treat-ment control subjects, a difference of 43% or 31% wasdetected, respectively, in a 2-sided test, at the 5% sig-nificance level, with 90% chance of success (90%power). Categorical data were analyzed using Fisher’sexact test. The continuous variables were checked fornormality (Shapiro-Wilk). A non-parametric method(Wilcoxon rank sum) was used if the data were notnormally distributed.

RESULTSEighty subjects, 43 males and 37 females, with a meanage of 47.4 years, were enrolled in the study. A total of95 defects, 65 subjects with one defect site and 15 sub-jects with two defect sites were treated (Figs. 2 and 3). No subjects were withdrawn or lost to follow-up. Foursubjects did not receive the treatment to which they were

randomized: three were incorrectly randomized to anothertreatment group and one received a human demineral-ized freeze-dried bone allograft. Data analysis for all sub-jects including the allograft-treated patient was based onintent to treat. For some data analysis, the sample sizewas smaller due to inadequate quality of the CT scan.The mean total dose of the rhBMP-2 implanted perextraction socket was 0.9 mg in the 0.75 mg/ml groupand 1.9 mg in the 1.50 mg/ml group.

Bone InductionThe results of the bone induction data are summarizedin Tables 1 and 2. Assessment of the alveolar boneheight indicated that the subjects in the 1.50 mg/mlgroup maintained the palatal wall of the extractionsocket, whereas the other study groups experienceddecreases (P ≤0.05). Median bone width at 25% ESL(near the top of the extraction site) increased in all studygroups except the no treatment group. The increasewas statistically significant for both active rhBMP-2 treat-ment groups (P ≤0.05). The differences in the increasesbetween the 0.75 mg/ml and 1.50 mg/ml groups alsowere statistically significant (P ≤0.05). The increases inmedian width at 50% ESL (at the middle of the extrac-tion site) in the 1.50 mg/ml groups were statisticallysignificant when compared to no treatment (P ≤0.05).

4020.qxd 4/18/05 8:59 AM Page 607

608

Effect of rhBMP-2 on Osseous Defect Healing Volume 76 • Number 4

The difference in the increases between the two activerhBMP-2 treatment groups was statistically significant(P ≤0.05). The relative increase in median bone widthat 75% ESL (near the base of the extraction site) in the1.50 mg/ml group was statistically significant whencompared with that in the no treatment group (P ≤0.05).The dose-response relationship was assessed by linearregression analysis. The responses were statistically sig-nificant for height and width at 25% and 50% ESL

(Table 3). Each of the regressioncoefficients for treatment groupswas positive, indicating that thebone growth was larger with rhBMPtreatment.

Adequate Alveolar Bone Volumefor Dental Implant PlacementA comparison of patients whorequired a secondary augmenta-tion in order to place a dental im-plant indicated that the 1.50 mg/ml-treated group had significantly(P ≤0.05) fewer procedures com-pared to the no treatment or 0.75mg/ml groups (Table 4). The ade-quacy of bone for the placementof a dental implant at the 50% and75% ESL was approximately threetimes as great in the 1.50 mg/mlrhBMP-2/ACS group as comparedto no treatment or placebo (Table5).

Bone DensityChanges in median density innative bone from baseline to 4months postoperative were 16.19mg/cc, 15.78 mg/cc, −22.77 mg/cc, and 8.61 mg/cc in the no treat-ment, placebo, 0.75 mg/ml, and1.50 mg/ml groups, respectively(data not shown). Although thedifference between baseline and4 months indicated a high level ofvariability, there were no statisti-cally significant differences amongthe treatment groups. In addition,a comparison of the density of newbone at 4 months post-treatmentrevealed no statistically significantdifferences (data not shown).

Histology of Bone CoresSixty-eight core biopsies from 57of the 80 subjects were examined;14 subjects in the no treatment

group; 11 subjects in the placebo group; 15 subjectsin the 0.75 mg/ml group; 16 subjects in the 1.50 mg/ml group; and the one subject who received the allo-graft (Fig. 4). The bone structure of approximatelytwo-thirds of the samples was exclusively trabecular,the thickness of which was graded moderate to large.Remodeling of woven bone into lamellar bone was themost common observation. Active osteoblasts in smallto moderate numbers were observed. Most specimens

Figure 2.A) Extraction site (tooth #9) exhibiting greater than 50% buccal bone height loss. B) Treatment sitewith 1.5 mg/ml rhBMP-2/ACS contour to reconstruct alveolar ridge. C) and D) Extraction sitefollowing 4 months of healing.

4020.qxd 4/18/05 8:59 AM Page 608

609

J Periodontol • April 2005 Fiorellini, Howell, Cochran, et al.

Figure 3.Baseline (A, B, and C) and 4-month post-treatment (D, E, and F) CT scans for no treatment (tooth #5), ACS alone (tooth #10), and 1.5 mg/mlrhBMP-2/ACS (tooth #4), respectively.

had few or no osteoclasts. Thevascularity of all specimens wasunremarkable, generally with asmall to moderate number ofcapillaries. No evidence of in-flammation or residual colla-gen matrix from the absorbablesponge carrier was identified inany of the specimens.

Safety AssessmentsA total of 250 adverse eventswere reported for 78 of the 80enrolled subjects. The most fre-quent reports were oral edema(75%), mouth pain (68%), andoral erythema (46%). There werea greater number of cases of oraledema and erythema in the studytreatment groups compared tothe no treatment group followingextraction. Periapical radiographsof the study subjects indicatedno exuberant bone formation orradiolucent voids. Serum chem-istry and hematology evaluationsrevealed no trends or clinicallysignificant changes.

Table 1.

Bone Height and Width Response (mm) at 4 Months

0.00 mg/ml 0.75 mg/ml 1.50 mg/mlNo Treatment rhBMP-2/ACS rhBMP-2/ACS rhBMP-2/ACS

(N = 20) (N = 17) (N = 22) (N = 21)

Height changeN (observed) 19 18 21 20Mean ± STD −1.17 ± 1.23 −1.00 ± 1.40 −0.62 ± 1.39 −0.02 ± 1.2095% CI −1.76 to −0.58 −1.74 to 0.26 −1.25 to 0.01 −0.58 to 0.54Median −1.30 −0.96 −0.63 −0.08

Width change at 25% ESLN (observed) 18 16 20 20Mean ± SD 0.57 ± 2.56 0.82 ± 1.40 1.76 ± 1.67 3.27 ± 2.5395% CI −0.71 to 1.84 0.07 to 1.56 0.98 to 2.54 2.08 to 4.45Median −0.47 0.73 1.72 3.53

Width change at 50% ESLN (observed) 18 16 20 20Mean ± STD 1.62 ± 2.51 1.79 ± 1.68 2.45 ± 1.38 3.97 ± 2.4895% CI 0.37 to 2.87 0.90 to 2.69 1.80 to 3.10 2.81 to 5.13Median 0.83 1.67 2.99 3.95

Width change at 75% ESLN (observed) 18 16 20 20Mean ± STD 1.65 ± 2.06 1.82 ± 1.50 2.29 ± 1.07 2.68 ± 1.3795% CI 0.63 to 2.68 1.02 to 2.62 1.79 to 2.79 2.04 to 3.32Median 1.72 1.88 2.35 2.63

4020.qxd 4/18/05 8:59 AM Page 609

610

Effect of rhBMP-2 on Osseous Defect Healing Volume 76 • Number 4

Table 2.

Pairwise Comparisons (P values) of Bone Height and Width Response

No Treatment No Treatment No Treatment rhBMP-2/ACS rhBMP-2/ACS rhBMP-2/ACSversus versus versus 0.00 mg/mL 0.00 mg/mL 0.75 mg/mL

rhBMP-2/ACS rhBMP-2/ACS rhBMP-2/ACS versus versus versus0.00 mg/mL 0.75 mg/mL 1.50 mg/mL 0.75 mg/mL 1.50 mg/ml 1.50 mg/ml

Height changeParametric tests* NS NS 0.007 NS 0.028 NSNonparametric tests† NS NS 0.008 NS NS NS

Width change at 25% ESLParametric tests NS NS 0.000 NS 0.001 0.028Nonparametric tests NS 0.025 0.003 NS 0.002 0.037

Width change at 50% ESLParametric tests NS NS 0.000 NS 0.003 0.024Nonparametric tests NS NS 0.008 NS 0.002 0.012

Width change at 75% ESLParametric tests NS NS 0.042 NS NS NSNonparametric tests NS NS NS NS NS NS

* P values of parametric tests are from ANOVA F-test.† P values of nonparametric tests are from Wilcoxon rank sum test.

The formation of antibodies torhBMP-2, bovine Type I collagen,and human Type I collagen wereassessed in the 80 subjects. Noantibodies were detected torhBMP-2 or to human Type I col-lagen. Antibodies to bovineType I collagen were detectedin 11 subjects. Three of thesesubjects, two in the no treat-ment group and one in the 1.50mg/ml group, had positive anti-body titers at baseline. Of theremaining eight, four subjectshad a transient immune re-sponse. Titers in these subjectsreturned to baseline level by4 months. The remaining foursubjects had detectable anti-

bodies at the 4-month evaluation.These findings were similar to previ-ous studies where bovine collagendevices were utilized and clinicaladverse events did not differ fromnon-treated subjects.7-9

DISCUSSIONConventional dentoalveolar osseousaugmentation procedures for creatingbone volume for dental implantsoften involve the use of graftingmaterials with or without barrier

Table 3.

rhBMP-2/ACS Bone Height and Width Dose Response

rhBMP-2/ACS rhBMP-2/ACS rhBMP-2/ACS Slope0.00 mg/mL 0.75 mg/mL 1.50 mg/mL Estimation* P Value†

Height change −1.00 −0.62 −0.02 0.66 0.029

Width change at 25% ESL 0.82 1.76 3.27 1.65 0.000

Width change at 50% ESL 1.79 2.45 3.97 1.47 0.001

Width change at 75% ESL 1.82 2.29 2.68 0.57 NS

* The slopes are from simple linear regression.† P values are from t test.

Table 4.

Dental Implant Placement With and Without a SecondaryAugmentation

0.00 mg/mL 0.75 mg/mL 1.50 mg/mLNo Treatment rhBMP-2/ACS rhBMP-2/ACS rhBMP-2/ACS

(N = 20) (N = 17) (N = 22) (N = 21)

Without augmentation 9 (45%) 10 (59%) 12 (55%) 18 (86%)*†

With augmentation 11 (55%) 7 (41%) 10 (45%) 3 (14%)

* P = 0.045 0.75 mg/ml rhBMP-2/ACS versus 1.50 mg/ml rhBMP-2/ACS.† P = 0.009 no treatment versus 1.50 mg/ml rhBMP-2/ACS.P values based on 2-sided Fisher’s exact test.

4020.qxd 4/18/05 8:59 AM Page 610

611

J Periodontol • April 2005 Fiorellini, Howell, Cochran, et al.

Figure 4.Histological section at 32 weeks 0.75 mg/ml rhBMP-2/ACS (originalmagnification ×10; hematoxylin and eosin).

regenerate. Several proteinsincluding BMPs have been iden-tified which are potent biologicmediators that promote eventsin wound healing. BMPs havebeen shown to be responsiblefor post-fetal bone induction,including normal bone remod-eling healing and repair.14-16

The potential therapeutic uti-lity of rhBMP-2 in orthopedicand craniofacial reconstructionhas been investigated. Preclin-ical studies evaluated inductionand repair of bony defects in a

variety of indications.3-5,17-20 The increases in thequantity and rate of bone induction by rhBMP-2 wereaffected by both the delivery system and by the animalspecies used.21 Reports utilizing rhBMP-2 in humanshave examined safety and technical feasibility. Howellet al.4 reported that in local ridge preservation andaugmentation, clinical results with 0.43 mg/ml rhBMP/ACS suggested that the product was well toleratedlocally and systemically, with no adverse events. TherhBMP/ACS was easy to handle and could be adaptedto the ridge or extraction site. Overall, bone fill wasobserved and maintained in the rhBMP/ACS-treatedextraction sites.4 In maxillary sinus augmentationsindication, Boyne et al. reported that 12 subjectstreated with 0.43 mg/ml rhBMP/ACS had no seriousevents and no clinically significant changes in bloodcounts, blood chemistries, or urinalysis.3 The CTresults indicated an efficacious alternative to the con-ventional bone graft methodologies for achieving ade-quate bone volume for dental implants in the posteriormaxilla.

The results from this present study indicate that anovel method to recreate the alveolar ridge in order tosupport a dental implant was efficacious. Our findingsindicated that the buccal wall extraction defect modelutilized to assess rhBMP-2 was effective. Partial healingwas observed in sites treated with the ACS placeboand no spontaneous bone healing was observed inthe no treatment group. Previous studies involvingpost-extraction healing and histology following uncom-plicated extractions indicate similar patterns to thedata presented in this investigation.22,23 Subjectstreated with the 1.50 mg/ml rhBMP-2 were restored tofunction when compared to the other treatment groupsaccording to bone quantity efficacy parameters. Qual-ity measures indicated that newly formed bone wasno different to native structures.

The rhBMP-2/ACS data in this investigation alsoindicate a significant enhancement in patient outcomesand several advantages over current therapies. First,not only was the volume required to place a dental

Table 5.

Percent of Implant Positions with Adequate Bone Formation*

rhBMP-2/ACS rhBMP-2/ACS rhBMP-2/ACSNo Treatment 0.00 mg/mL 0.75 mg/mL 1.50 mg/mL P Value

25% ESL 12.5 6.25 25 56.25 0.008,† 0.034,‡ 0.009§

50% ESL 9.09 20.45 29.54 40.9 <0.001,† ‡ 0.012§

75% ESL 14.28 21.42 30.35 32.14 0.023,† 0.005‡

* Width at narrowest ≥6 mm and height ≥12 mm.† Comparability among all treatment groups; chi-square test.‡ 1.50 mg/mL versus no treatment; 2-sided Fisher’s exact test.§ 1.50 mg/mL versus 0.00 mg/mL; 2-sided Fisher’s exact test.

membranes to foster selective cell and tissue repopu-lation. The clinical and histologic efficacy seems to bedependent on the type, source, biocompatiblity, andability to maintain volume.10,11 The rationale for uti-lizing several of these materials to promote osseousregeneration of the alveolus has been the presence ofosteoinductive proteins.12,13 To achieve tissue-specificregeneration with proteins or peptides, several strate-gies have been utilized. First, biomaterials have beendeveloped, such as the ACS, which provide a matrixto facilitate cell/tissue ingrowth. Secondly, ligands havebeen identified that induce existing adult tissues to

4020.qxd 4/18/05 8:59 AM Page 611

612

Effect of rhBMP-2 on Osseous Defect Healing Volume 76 • Number 4

implant restored, but the quantity of bone regenera-tion at the top of the alveolar crest (75% ESL) alsooptimizes dental implant position. When the implantlocation approaches or is in the natural tooth rootposition, fabrication of the prosthesis becomes lessdifficult which usually results in lower laboratory andpatient costs. Second, currently utilized augmentationprocedures require advanced surgical skills.24 Theimplantation of the rhBMP-2/ACS added no additionalcomplexity to the surgical procedure greater than thetooth extraction. Next, the patients in this investiga-tion did not experience post-surgical exposures, whichcan occur with a traditional membrane augmentation.The absence of such complications maximizes therestored bone volume and reduces the number ofpatient visits. In addition, the patient did not requireharvesting of bone from a secondary site such as theiliac crest. This clearly reduced patient morbidity.Lastly, clinical safety profiles were similar for treat-ment groups indicating that the rhBMP/ACS implan-tation procedure was not different than a toothextraction.

In conclusion, the stimulation of host healingresponses with rhBMP-2 versus dependence on theosteoconductive properties of a carrier materialenhanced predictability and provided a substantialpatient clinical benefit. This randomized, masked,placebo-controlled multi-center clinical study demon-strated that the novel combination of a rhBMP-2 anda commonly utilized collagen sponge had a strik-ing effect on de novo osseous formation, whichallowed the placement of dental implants. As investi-gators continue to characterize function of bone stim-ulating molecules and consider cofactors includingnutrients, hormones, age, and gender requirements,clinicians will be better able to provide patient-spe-cific therapies.

ACKNOWLEDGMENTSThis study was supported by a grant from Wyeth/Genetics Institute, Cambridge, Massachusetts. Ms. Lillyis Director of Clinical Research and Development atWyeth/Genetics Institute. The authors thank Drs. MardenAlder, Ray Williams, Michael Mastromarino, Julio Sekler,and Marc Nevins for their participation in this study andDr. XiaoJian Li for the histology preparations.

REFERENCES1. Buser D, Ingimarsson S, Dula K, et al. Long-term sta-

bility of osseointegrated implants in augmented bone:A 5-year prospective study in partially edentulouspatients. Int J Periodontics Restorative Dent 2002;22:109-117.

2. Howell TH, Fiorellini JP, Paquette DW, et al. A phase I/IIclinical trial to evaluate a combination of platelet-derivedgrowth factor-BB and recombinant insulin-like growth

factor-I in patients with periodontal disease. J Periodontol1997;68:1186-1193.

3. Boyne PJ, Marx RE, Nevins M, et al. A feasibility studyevaluating rhBMP-2/absorbable collagen sponge formaxillary sinus floor augmentation. Int J PeriodonticsRestorative Dent 1997;17:11-25.

4. Howell TH, Fiorellini J, Jones A, et al. A feasibilitystudy evaluating rhBMP-2/absorbable collagen spongedevice for local alveolar ridge preservation or aug-mentation. Int J Periodontics Restorative Dent 1997;17:125-140.

5. Cochran DL, Jones AA, Lilly-Clark L, Fiorellini JP,Howell TH. Evaluation of recombinant human bone mor-phogenetic protein-2 in oral applications including theuse of endosseous implants: Up to 3-year results inhumans. J Periodontol 2000;71:1241-1257.

6. Bianchi J, Fiorellini J, Howell TH, Sekler J, Curtin H,Friedland B. Measuring the efficacy of rhBMP-2 toregenerate bone – A radiographic study using theSimplant program. Int J Periodontics Restorative Dent2004;24:579-587.

7. van den Bergh JP, ten Bruggenkate CM, Groeneveld HH,et al. Recombinant human bone morphogenetic protein-7 in sinus floor elevation surgery in 3 patients comparedto autogenous bone grafts. A clinical pilot study. J ClinPeriodontol 2000;9:627-637.

8. Hyder P, Sigh G, Adam S. Humoral responses to type Icollagen after surgical curettage procedures employingbovine collagen implants. Biomaterials 1992;13:693-696.

9. The Canadian Multicenter Hemashield Study Group.Immunologic response to collagen-impregnated vascu-lar grafts: A randomized prospective study. J Vasc Surg1990;12:741-746.

10. McClelland M, Delustro P. Evaluation of antibody classin response to bovine collagen treatment in patients withurinary incontinence. J Urol 1996;155:2068-2073.

11. Nevins M, Mellonig JT, Clem DS, Reiser GM, Buser DA.Implants in regenerated bone: Long-term survival. IntJ Periodontics Restorative Dent 1998;18:34-45.

12. Nevins M, Mellonig JT. The advantages of localizedridge augmentation prior to implant placement. A stagedevent. Int J Periodontics Restorative Dent 1994;14:97-111.

13. Feuille F, Knapp CI, Brunsvold MA, et al. Clinical andhistologic evaluation of bone replacement grafts in thetreatment of localized alveolar ridge defects. Part 1: Min-eralized freeze-dried bone allograft. Int J PeriodonticsRestorative Dent 2003;23:29-35.

14. Schwartz Z, Mellonig JT, Carnes DL, et al. Ability ofcommercial deminerialized freeze-dried bone allograftto induce new bone formation. J Periodontol 1996;66:822-828.

15. Urist MR. Bone formation by autoinduction. Science1965;165:893-899.

16. Urist MR, Iwata H, Ceccotti PL, et al. Bone morphogen-esis in implants of insoluble bone gelatin. Proc Natl AcadSci (USA) 1973;70:3511-3515.

17. Wang EA, Rosen V, Cordes P, et al. Purification and char-acterization of distinct bone-inducing factors. Proc NatlAcad Sci (USA) 1988;85:9484-9488.

18. Toriumi DM, Kolter HS, Luxenberg DP, et al. Mandibularreconstruction with a recombinant bone-inducing fac-tor. Functional, histologic, and biomechanical evalua-tion. Arch Otolarynogol Head Neck Surg 1991;117:1101-1112.

4020.qxd 4/18/05 8:59 AM Page 612

613

J Periodontol • April 2005 Fiorellini, Howell, Cochran, et al.

19. Nevins M, Kirker-Head C, Nevins M, et al. Bone forma-tion in the goat maxillary sinus induced by absorbablecollagen sponge implants impregnated with recombinanthuman bone morphogenetic protein-2. Int J PeriodonticsRestorative Dent 1996;16:8-19.

20. Einhorn TA, Majeska RJ, Mohaideen A, et al. A singlepercutaneous injection of recombinant human bone mor-phogenetic protein-2 accelerates fracture repair. J BoneJoint Surg Am 2003;85:1425-1435.

21. Govender S, Csimma C, Genant HK, et al. Recombinanthuman bone morphogenetic protein-2 for treatment ofopen tibial fractures; a prospective, controlled, ran-domized study of four hundred and fifty patients. J BoneJoint Surg Am 2002;84-A:2123-2134.

22. Wikesjö U, Sorensen R, Wozney J. Augmentation of alve-olar bone and dental implant osseointegration: Clinicalimplications of studies with rhBMP-2. J Bone Joint SurgAm 2001;83-A:136-145.

23. Evian CI, Rosenberg ES, Coslet JG, et al. The osteogenicactivity of bone removed from healing extraction sock-ets in humans. J Periodontol 1982;53:81-85.

24. Lekovic V, Camargo PM, Klokkevold PR, et al. Preser-vation of alveolar bone in extraction sockets usingbioabsorbable membranes. J Periodontol 1998;69:1044-1049.

Correspondence: Dr. Joseph P. Fiorellini, Harvard School ofDental Medicine, Department of Oral Medicine, Infection andImmunity, 188 Longwood Ave., Boston, MA 02115. Fax:617/432-1897; e-mail: joseph_fiorellini@hsdm.harvard.edu.

Accepted for publication August 13, 2004.

4020.qxd 4/18/05 8:59 AM Page 613