DENTAL IMPLANTS

J Oral Maxillofac Surg 63:1693-1707, 2005

De Novo Bone Induction by Recombinant Human Bone Morphogenetic Protein-2 (rhBMP-2) in Maxillary Sinus Floor Augmentation
Philip J. Boyne, DMD, MS, DSc,* Leslie C. Lilly, BSN, RN, Robert E. Marx, DDS, Peter K. Moy, DMD, Myron Nevins, DDS, Daniel B. Spagnoli, PhD, DDS, and R. Gilbert Triplett, DDS, PhD#
Purpose: This phase II study was designed to evaluate 2 concentrations of recombinant human bone

morphogenetic protein-2 (rhBMP-2) for safety and efcacy in inducing adequate bone for endosseous dental implant in patients requiring staged maxillary sinus oor augmentation. Materials and Methods: Patients were treated with rhBMP-2 (via an absorbable collagen sponge [ACS]), at concentrations of 0.75 mg/mL (n 18), 1.50 mg/mL (n 17), or with bone graft (n 13). Bone induction was assessed by alveolar ridge height, width, and density measurements from computed tomography scans obtained before and 4 months after treatment and 6 months post-functional loading of dental implants (density only). Results: Mean increases in alveolar ridge height at 4 months after treatment were similar among the groups; 11.3 mm, 9.5 mm, and 10.2 mm, respectively, in the bone graft, 0.75 mg/mL, and 1.50 mg/mL rhBMP-2/ACS treatment groups. Mean increases in alveolar ridge width (buccal to lingual) at the crest of the ridge were statistically different among the treatment groups; 4.7 mm, 2.0 mm, and 2.0 mm, respectively, in the bone graft, 0.75 mg/mL, and 1.50 mg/mL treatment groups (P .01 vs 0.75 mg/mL; P .01 vs 1.50 mg/mL). At 4 months postoperative new bone density was statistically different among the treatment groups; 350 mg/cc, 84 mg/cc, and 134 mg/cc for the bone graft, 0.75 mg/mL, and 1.50 mg/mL rhBMP-2/ACS treatment groups, respectively (P .003 vs 0.75 mg/mL, P .0137 vs 1.50 mg/mL, P .0188; 1.50 mg/mL vs 0.75 mg/mL). Core bone biopsies obtained at the time of dental implant placement conrmed normal bone formation. The proportion of patients who received dental implants that were functionally loaded and remained functional at 36 months post-functional loading was 62%, 67%, and 76% in the bone graft, 0.75 mg/mL, and 1.50 mg/mL rhBMP-2/ACS treatment groups, respectively. Conclusion: This study is the rst randomized controlled trial demonstrating de novo organ tissue growth in humans from a recombinant human protein. rhBMP-2/ACS safely induced adequate bone for the placement and functional loading of endosseous dental implants in patients requiring staged maxillary sinus oor augmentation. 2005 American Association of Oral and Maxillofacial Surgeons J Oral Maxillofac Surg 63:1693-1707, 2005
*Emeritus Professor, Department of Surgery, Oral and Maxillofacial Surgery Service, Loma Linda University Medical Center, Loma Linda, CA. Senior Director, Clinical Trial Operations, Wyeth Research, Cambridge, MA. Professor of Surgery and Chief, Division of Oral and Maxillofacial Surgery, Miller School of Medicine, University of Miami, Miami, FL. Director of Implant Dentistry, University of California, Los Angeles, Los Angeles, CA. Clinical Associate Professor, Harvard School of Dental Medicine, Swampscott, MA. Private Practice, University of Oral and Maxillofacial Surgery, Charlotte, NC. #Regents Professor and Chairman, Oral and Maxillofacial Surgery, Baylor College of Dentistry, Texas A&M University System, Health Science Center, Dallas, TX. Supported through research funding from Wyeth, Cambridge, MA. Address correspondence and reprint requests to Dr Lilly: 35 Cambridge Park Dr, Cambridge, MA 02140; e-mail: llilly@wyeth.com
2005 American Association of Oral and Maxillofacial Surgeons

0278-2391/05/6312-0002$30.00/0 doi:10.1016/j.joms.2005.08.018

1693

1694 Patients with edentulous posterior maxillae often present with loss of alveolar bone and increased maxillary sinus pneumatization. As a result, dental restoration is often difcult to achieve because of an inadequate volume of bone to support placement of endosseous dental implants. Various surgical procedures have been developed to correct this problem. One of these is maxillary sinus oor augmentation.1,2 In this procedure, alveolar bone height is increased by the formation of bone in the lower third of the maxillary sinus. This allows for dental restoration specically for the placement of endosseous dental implants.3 Various bone grafting materials are currently being used in maxillary sinus oor augmentation procedures with different degrees of success. These materials include autogenous bone (harvested from the patients iliac crest, tibia, mandible, or maxillary tuberosities), allogeneic bone, bone graft substitutes (eg, xenografts), and a combination of these materials. However, all of these materials have limitations, including inconvenience, cost, and morbidity associated with harvesting bone from the iliac crest, inadequate supply when harvesting bone from the oral cavity, and delayed healing. Given these limitations, the ideal agent for maxillary sinus oor augmentation procedures has not yet been identied. The bone morphogenetic proteins (BMPs) are a family of osteoinductive proteins that stimulate endochondral and intramembranous bone formation from mesenchymal cells in situ. Urist,4 who showed ectopic bone formation in rabbits from implanted bone pieces, rst described bone morphogenetic activity. Subsequently, several BMPs have been puried from bone, and many have been cloned providing tools to study the role of BMPs.5-8 The availability of recombinant BMPs has permitted denitive tests of their osteoinductive activity in a variety of experimental systems, including several animal models of clinically relevant bone defects.9,10 For example, in studies of dogs and non-human primates with large mandibular defects, rhBMP-2 induced rapid, new bone growth sufcient to heal the mandibular defects without the addition of bone graft.11-13 In addition, rhBMP-2/absorbable collagen sponge (ACS) successfully generated bone formation in goats that had undergone maxillary sinus oor augmentation procedures.14 Based on these data and further preclinical studies showing the local and systemic safety and pharmacology of rhBMP-2, a clinical pilot study to determine the feasibility of using rhBMP-2 to induce de novo bone growth in patients who required 2-stage maxillary sinus oor procedures was conducted.15 In that study, 0.43 mg/mL rhBMP-2, administered via an ACS induced new bone growth in 100% of the 11 evaluable patients treated (mean gain, 8.51 mm; range, 2.3

BONE INDUCTION WITH RHBMP-2

mm to 15.7 mm). However, the quantity of bone formed was judged adequate for 73% of these individuals, suggesting that the concentration of rhBMP-2 or the total dose administered was too low and that higher concentrations should be evaluated. Consequently, we conducted a phase II randomized, parallel group evaluation of escalating concentrations of rhBMP-2 (0.75 and 1.5 mg/mL) compared with bone graft to determine a safe and effective concentration of rhBMP-2 for maxillary sinus oor augmentation procedures.

Materials and Methods

PATIENT POPULATION

Patients 18 years of age with inadequate alveolar bone height (less than 6 mm conrmed on computed tomography [CT] scan) in the posterior maxilla who were candidates for staged bilateral or unilateral maxillary sinus oor augmentation were eligible. All patients provided written informed consent. Patients were excluded if they had acute or chronic sinus disease; sinus pathology in a location that would interfere with postoperative radiographic measurements; signicant untreated periodontal disease, caries, or oral infection; required onlay ridge augmentation to achieve adequate bone volume for placement of endosseous dental implants; used any nicotine-containing products within 2 weeks before surgery; were pregnant; had insulin-dependent diabetes; were taking medications or having treatments known to affect bone turnover; or had a disease that affects bone metabolism. Therapeutic agents or devices known to interfere with or to have an effect on bone induction were not permitted during the study period.
STUDY DESIGN

The primary objective for this phase II study was to evaluate the safety and efcacy of 2 concentrations of rhBMP-2 to induce adequate bone for endosseous dental implant placement (following a maxillary sinus oor augmentation procedure) to select 1 concentration for investigation in a future pivotal study. Secondarily, we wanted to estimate patient and dental implant success rates following 36 months of functional loading of dental implant(s) placed and loaded into the newly induced bone within the grafted sinus(s). Thus, this multicenter study was designed as a randomized, parallel evaluation of escalating concentrations (2) of rhBMP-2 as compared with bone graft. Bone graft was dened as the bone grafting material currently used at each investigative site. The materials that could be used were limited to autogenous bone or a combination of autogenous bone and allogeneic bone.

BOYNE AT AL

1695 plant placement. Oral examination, radiograph analysis, adverse event surveillance, measurement of vital signs, and evaluation of serum chemistry, hematology, and formation of antibodies to the study treatment monitored safety. The Institutional Review Board at each participating center approved the study protocol.
SURGICAL PROCEDURE

FIGURE 1. Medical illustrations of surgical procedure implanting rhBMP-2/ACS. Boyne at al. Bone Induction With rhBMP-2. J Oral Maxillofac Surg 2005.

Six investigative sites were to enroll a total of 48 patients into 2 cohorts consisting of 24 patients each. Within each cohort, patients were to be randomized in a 2:1 ratio to receive either rhBMP-2/ACS (16 patients) or bone grafting material (8 patients). Patients in the rst cohort were to be treated with either rhBMP-2 at a concentration of 0.75 mg/mL or the bone grafting material. Patients in the second cohort were to receive either rhBMP-2 at a concentration of 1.50 mg/mL or bone grafting material. The 2 cohorts were to be treated sequentially; the second cohort was treated after demonstration of acute safety in the rst cohort. Before randomization, patients were stratied based on dentate status (ie, totally or partially edentulous) to control for the possible confounding effect of this variable on functional loads placed onto a full arch prosthesis supported by dental implants as opposed to smaller units supported by dental implants. The patients treatment course was approximately 52 months in duration and was composed of several phases: short-term safety and bone induction (4 months), dental implant placement and osseointegration (initiated with the placement of dental implants and completed by prosthesis placement), and a 36 month functional loading phase. The patient was considered a treatment failure if he/she did not receive dental implants within 12 months postoperative and prosthetics within 24 months following dental im-

The surgical procedure and postoperative medications were standardized across the participating centers. Anesthesia was to include local inltration with a mixture of epinephrine (not to exceed 1%) and lidocaine. The lateral surface of the maxilla was to be exposed by raising a mucoperiosteal ap by making an incision on the palatal or buccal surface of the alveolar ridge. A round bur or diamond was to be used to create an adequate window on the lateral maxillary wall. Patients in the rhBMP-2/ACS treatment group were to have had the bony window removed. The sinus membrane was to be elevated up the medial wall, but not stripped superiorly beyond the ostium. The study treatment (rhBMP-2/ACS or bone graft) was to be placed anteriorly and extended to the medial wall. The study treatment was not to be implanted if the sinus membrane was perforated and could not be overlapped onto itself. No concurrent extraneous materials, such as a barrier membrane, were to be used for the rhBMP-2/ACS treatment group. However, barrier membranes (of any type) could be used in the bone graft group, as determined by the operating surgeon. Medical illustrations and surgical photographs of the surgical procedure are provided in Figures 1, 2.
STUDY MATERIAL

The rhBMP-2 was manufactured by Wyeth/Genetics Institute (Cambridge, MA) using recombinant DNA techniques. The ACS was derived from highly puried bovine tendon type I collagen and is marketed under the trade name Helistat Absorbable Collagen Hemostatic Agent (Integra Life Sciences, Plainsboro, NJ). rhBMP-2/ACS is now commercially available as Infuse

FIGURE 2. Clinical photographs of surgical procedure implanting rhBMP-2/ACS. Boyne at al. Bone Induction With rhBMP-2. J Oral Maxillofac Surg 2005.

1696

BONE INDUCTION WITH RHBMP-2

FIGURE 3. CT scan cross-section Linear measurements. Boyne at al. Bone Induction With rhBMP-2. J Oral Maxillofac Surg 2005.

Bonegraft through Medtronic (Medtronic, Memphis, TN).

EFFICACY EVALUATION

Sofamor

Danek

measurement minus the baseline height measurement. To measure bone width, 3 horizontal lines were drawn perpendicular to the vertical line at 1/4, 1/2, and 3/4 of the distance from the alveolar crest to the maxillary sinus oor along the vertical axis of the height measurement line. Width measurements were taken at positions within the sinus where bone or nonmineralized bone occurred throughout the horizontal line. Width measurement was the distance between the medial and lateral borders of the bone on each of the 3 horizontal lines (Fig 3). The density of newly induced bone was assessed with the aid of a standard density block. Density of the treatment area was measured in 2 area of interest boxes excluding cortical bone. Newly induced bone density was the average of 3 independent bone density measurements at 4 months and was averaged over the 2 area of interest boxes of all planned dental implant sites (Fig 4). Histology Full-thickness bone core biopsies of the study treatment area were obtained at the time of dental implant placement for those patients who received dental implants. Biopsies were taken by using a trephine (inner diameter 2.0, outer diameter 2.7) and drilling from the alveolar ridge toward the maxillary sinus in the intended dental implant position. The biopsies were xed in neutral buffered 10% formalin, decalcied in formic acid, and processed and embedded longitudinally in parafn. Blocks were cut at a thickness of 4 m and were stained with Mayers hematoxylin and eosin. An independent pathologist at ClinTrial BioResearch (Montreal, Canada) evaluated each biopsy, on a qualitative scale from 1 (small amount) to 5 (large amount), for the presence of cortical and/or trabecular bone, thickness of osseous trabeculae, and the presence of lamellar and woven bone. Bone cores were scored for osteoblasts and osteoclasts and were evaluated for evidence of pathologic change. Crestal Bone Loss Adjacent to Dental Implants Intraoral lms were evaluated by 3 unique radiologists. They measured the distance, in millimeters, between the shoulder of the dental implant and the rst contact of crestal bone to the implant (bone height) on both the mesial and distal sides. They also documented any abnormal ndings around the dental implants (eg, over exuberant bone formation). To ensure that differences in measurements from visit to visit were caused by actual bone changes, the length of the dental implant from the implant apex to the implant shoulder was measured on the baseline image. This value was registered on the source document for each implant. On the follow-up radiographs the same distance was

Bone Induction Bone induction was quantied by CT scans performed within 4 weeks before (baseline) and 4 months after implantation of rhBMP-2/ACS or bone graft. Measurement of the 4-month postoperative CT scan was performed rst to establish points of measurement. These same reference points were then used to measure the baseline CT scan. The sites to be measured were determined by the investigators from the 4-month CT scan. This was achieved by the independent radiology center reformatting the 4-month CT scan and sending it to the investigator to identify the locations where the dental implants were to be placed. All measurements were then taken at these locations, which were within the sinus. Bone height, width, and density measurements were taken by the radiologists from one 2-mm thick, cross-sectional, multiplanar, reformatted image for each proposed endosseous dental implant site identied by the investigator, as determined on the 4-month CT scan. Measurements of the corresponding sites were then obtained from the baseline CT scans. Each measurement was made by 3 independent radiologists blinded to study treatment. Bone height was measured along a vertical line drawn from the alveolar crest to the oor of the maxillary sinus parallel to the plane of the planned endosseous dental implant site. The height measurement was taken from the level of the alveolar crest to the oor of the maxillary sinus (Fig 3). A patients bone height at baseline and at 4 months was reported as the average of the measurements taken by 3 independent radiologists of all the planned dental implant sites in the patient. The patients change in bone height following treatment was the 4-month height

BOYNE AT AL

1697 post surgery, and at an additional time point for those patients who had an elevated titer ( 50) at 4 months postoperative.
STATISTICAL METHODS AND ANALYSIS

measured, and if the distance was 0.5 mm longer or shorter than the baseline, the follow-up image was corrected accordingly to match the implant length between the baseline and follow-up images.16 These corrected images were used by 3 radiologists to perform the measurements. Crestal bone loss was reported as the average of the measurements taken by 3 independent radiologists. Factors affecting patient and dental implant success. The time to dental implant and prosthesis placement was collected. Additional variables that could also affect dental implant success were recorded: topography of the alveolar ridge, whether crestal bone reduction was performed, quality of bone at the time of dental implant placement (ie, Branemark scale type I-IV), whether an additional onlay augmentation procedure was required for dental implant placement, manufacturer, rootform type, surface, and size of the dental implant placed. Patient success. A patient was dened as having had a successful maxillary sinus oor augmentation procedure if he/she met the following criteria: had a maxillary sinus oor augmentation procedure; had dental implants (at least 1) placed into the grafted sinus (newly induced bone) without an additional maxillary sinus oor augmentation procedure; achieved osseointegration of a sufcient number of dental implants placed into the grafted sinus to allow placement of an implant borne prosthetic device; and maintained use of the prosthesis following functional loading. Dental implant success. A dental implant was considered a success providing it was functionally loaded and met the following criteria: was immobile when tested clinically (providing it was unattached); did not have any evidence of continuous peri-implant radiolucency; and had no chronic symptoms of pain, infection, or neuropathies. Dental implant survival. A dental implant was considered survived if it was not removed following its placement. Safety Safety was monitored by performing oral examinations, radiographs, monitoring for the occurrence of Adverse Events, and the collection of blood samples to measure serum chemistries, hematology, and antibody formation to the study treatment. Immune response. Formation of antibodies to rhBMP-2, bovine type I collagen, and human collagen were assessed by enzyme-linked immunosorbent assay (ELISA). The samples were collected for antibody testing at baseline (before surgery), 1 and 4 months

Because this study was designed to evaluate the ability of a recombinant human protein to induce bone formation in patients following a maxillary sinus oor augmentation procedure, the patient rather than the dental implant was the primary unit of analysis. Therefore, all efcacy and safety analyses were performed for the intent-to-treat population, which included all enrolled patients in their respective randomized treatment groups. Analyses were performed using SAS (version 6.12; SAS, Chicago, IL) executed in a UNIX environment. Quantitative variables were summarized by treatment group and by visit using descriptive statistics. Qualitative variables were descriptively summarized by treatment group using counts and percentages. The denominator for calculating patient and dental implant success rates remained constant throughout the study, ie, the number of patients enrolled per treatment group, and the total number of dental implants placed (including replacements) remained unchanged. Patients (and their respective dental implants) who were discontinued from the study early because of treatment failure, their or the investigators request to terminate their participation early, or lost to follow-up where not removed from the denominator when calculating patient and dental implant success rates. For each quantitative variable, the distribution of the data was assessed using the Shapiro-Wilk test before further statistical testing. If the normality assumption was violated (P .1), then a nonparametric test (Kruskal-Wallis test) was used. All statistical testing was 2-sided and conducted at the 5% signicance level, unless otherwise specied. Quantitative demographic variables were tested for baseline comparability among the 3 treatment groups using the F-test from an analysis of variance (ANOVA), while qualitative items were tested using the Pearsons chi-square test. An overall comparison of changes in bone measurements from baseline to 4 months postoperative among the treatment groups was carried out by ANOVA. If the overall P value was signicant, then pairwise comparisons were carried out. Because of the small sample size statistical testing was not performed on the clinical outcome parameter.

Results
Forty-eight patients were enrolled between October 1996 and June 1997 at 6 centers in the United States: The Baylor College of Dentistry, Dallas, TX;

1698

BONE INDUCTION WITH RHBMP-2

Table 1. SUMMARY OF DEMOGRAPHIC VARIABLES BY TREATMENT GROUP

Variable Race (n/%) Caucasian Hispanic Other Gender (n/%) Male Female Currently use alcohol (n/%) No Yes Age (yrs) N (observed) Mean SD Weight (kg) N (observed) Mean SD Dentate Status N (observed) Partially edentulous Totally edentulous

Bone Graft (n 13) 8 (62) 4 (31) 1 (8) 5 (38) 8 (62) 7 (54) 6 (46) 13 57 11 13 74.5 23.6 9 (69) 4 (31)

rhBMP-2/ACS 0.75 mg/mL (n 18) 15 (83) 2 (11) 1 (6) 8 (44) 10 (56) 10 (56) 8 (44) 18 57 12 18 68.6 14.4 13 (72) 5 (28)

rhBMP-2/ACS 1.50 mg/mL (n 17) 15 (88) 1 (6) 1 (6) 6 (35) 11 (65) 8 (47) 9 (53) 17 52 9 17 71.2 14.7 10 (59) 7 (41)

P Value .3679*

.9310* .9345* .3158

.6485

.7386*

*P values for discrete variables are from extended Fishers exact test. P values for continuous variables are from Analysis of Variance F-test. Boyne at al. Bone Induction With rhBMP-2. J Oral Maxillofac Surg 2005.

Institute for Advanced Dental Studies, Swampscott, MA; Loma Linda School of Dentistry, Loma Linda, CA; Doctors Hospital at University of Miami, Coral Gables, FL; West Coast Center for Osseointegration, Los Angeles, CA; and University Oral and Maxillofacial Surgery, Charlotte, NC. The last patient completed the study in February 2002. Eighteen patients were treated with 0.75 mg/mL rhBMP-2/ACS, 17 patients were treated with 1.50 mg/mL rhBMP-2/ACS, and 13 patients were treated with standard bone graft material. Forty-one (85%) patients had bilateral and 7 (15%) patients had unilateral sinus procedures performed. At baseline, the mean alveolar ridge height was similar among the treatment groups and ranged from 5.0 to 6.2 mm. One patient from the bone graft group was enrolled in violation of the bone height requirement ( 6 mm). The demographic and baseline characteristics of the patients enrolled were similar among the 3 treatment groups (Table 1).
STUDY TREATMENT

graft, 0.75 mg/mL, and 1.50 mg/mL rhBMP-2/ACS treatment groups, respectively. The mean total dose of rhBMP-2 implanted per sinus was 8.9 mg (range, 5.2 to 12.0 mg) for patients treated with 0.75 mg/mL and 20.8 mg (range, 10.8 to 24.0 mg) for patients treated with 1.50 mg/mL. The total rhBMP-2 dose implanted ranged from 6 to 24 mg and from 15 to 48 mg in the 0.75 and 1.50 mg/mL groups, respectively. In the standard bone graft group, 7 patients received autograft alone and 6 patients received autograft plus allograft. Autograft was harvested from the iliac crest (n 4), the tibial plateau (n 6), the chin (n 1), or the maxilla (n 2), and yielded 1 to 21 cc of bone.
BONE INDUCTION

The volume of study treatment implanted depended on the estimated size of the patients sinus, the procedure type (ie, unilateral vs bilateral), and the surgical antral void created by the operating surgeon. The mean total volume of study treatment implanted per sinus was 6.9 cc, 11.9 cc, and 13.8 cc in the bone

At 4 months postoperative the mean bone height changes from baseline were 11.29 mm, 9.47 mm, and 10.16 mm in the 0.75 mg/mL rhBMP-2/ACS, 1.50 mg/mL rhBMP-2/ACS, and bone graft treatment groups, respectively. Patients in the bone graft group showed a slightly greater gain in bone height compared with the patients in the 2 rhBMP-2/ACS treatment groups; however, there were no signicant differences between any of the treatment groups (P .6074) (Table 2).

BOYNE AT AL

1699

Table 2. BONE HEIGHT AND WIDTH CHANGE AT 4 MONTHS POSTOPERATIVE BY TREATMENT GROUP

Bone Graft (n 13) Height change (mm) N (observed) Mean SD 95% CI upper 95% CI lower Width change at 1/4 N (observed) Mean SD 95% CI upper 95% CI lower Width change at 1/2 N (observed) Mean SD 95% CI upper 95% CI lower Width change at 3/4 N (observed) Mean SD 95% CI upper 95% CI lower 13 11.29 4.12 8.8 13.78 13 4.66 2.75 3 6.32 13 10.17 2.98 8.37 11.97 13 10.56 3.17 8.64 12.47

rhBMP-2/ACS 0.75 mg/mL (n 18) 18 9.47 5.72 6.63 12.32 18 2.02 2.73 0.67 3.38 18 8.54 5.47 5.82 11.26 18 11.86 5.15 9.29 14.42

rhBMP-2/ACS 1.50 mg/mL (n 17) 16 10.16 4.7 7.65 12.66 16 1.98 2.41 0.69 3.27 16 7.8 3.87 5.74 9.86 16 10.78 4.63 8.31 13.25

P Value .6074*

.0130*

.3468*

.6822*

*P value is from parametric ANOVA F-test. Indicates the P value is less than .05. Boyne at al. Bone Induction With rhBMP-2. J Oral Maxillofac Surg 2005.

Bone width at the crest of the ridge (1/4) increased signicantly in the bone graft group (mean 4.66 mm) compared with the 2 rhBMP-2/ACS groups (mean 2.02 mm and 1.98 mm in the 0.75 mg/mL and 1.50 mg/mL rhBMP-2/ACS treatment groups, respectively (P .01 vs 0.75 mg/mL; P .01 vs 1.50 mg/mL) (Table 2). The bone width at the midpoint of the treated sinus (1/2) also increased most notably in the bone graft group (mean, 10.17 mm) compared with the 2 rhBMP-2/ACS groups (mean, 8.54 mm, 7.80 mm, respectively). However the difference among the treatment groups was not statistically signicant (P .3468) (Table 2). Bone width at the most apical end of the treated sinus (3/4) increased similarly among the treatment groups; mean values were 10.56 mm, 11.86 mm, and 10.78 in the bone graft, 0.75 mg/mL, and 1.50 mg/mL groups, respectively (P .6822) (Table 2). Representative CT scan images of rhBMP-2/ACS and bone graft induced bone are shown in Figures 5, 6. Bone Density As expected a signicant difference in new bone density was observed in favor of the bone graft group compared with the 0.75 and 1.50 mg/mL rhBMP-2/

ACS treatment groups (P .0003 and P .137, respectively). Surprisingly the 1.50 mg/mL treatment group also showed a signicantly greater bone density compared with the 0.75 mg/mL treatment group (P .0188). The hypothesis for this nding is that the

FIGURE 4. CT scan cross-section: Density measurements. Boyne at al. Bone Induction With rhBMP-2. J Oral Maxillofac Surg 2005.

1700

BONE INDUCTION WITH RHBMP-2

FIGURE 5. CT scan cross-section of a patient treated with 0.75 mg/mL rhBMP-2. Baseline 4 months postoperative. Boyne at al. Bone Induction With rhBMP-2. J Oral Maxillofac Surg 2005.

FIGURE 6. CT scan cross-section of patient # 22 treated with autograph. Baseline 4 months postoperative. Boyne at al. Bone Induction With rhBMP-2. J Oral Maxillofac Surg 2005.

higher of the 2 rhBMP-2 concentrations resulted in faster bone formation, which was manifested by greater radiographic density, ie, mineralized bone, at 4 months postoperative. However, following 6 months of functional loading the density of the newly induced bone increased signicantly for the 0.75 and 1.50 mg/mL treatment groups and became comparable to the bone graft group (P .6452) (Table 3). Histology Core bone biopsy specimens were obtained at the time of dental implant placement on average 6.9 months (range, 5.9 to 11.5 months) postoperative from the newly induced bone within the grafted sinus. The histologic analysis from these specimens indicated unambiguous bone induction by rhBMP-2. Further there were no differences in the histologic parameters that were evaluated among the

treatment groups. A moderate to large amount of trabecular bone had formed which contained woven bone (small to moderate amounts) and lamellar bone (moderate to large amounts). Small to moderate amounts of osteoblasts and small amounts of osteoclasts were present. The bone marrow contained small amounts of brotic material, was moderately vascular, and had very little mononuclear cell inltration. No evidence of mixed inammatory cell inltration was observed in any of the specimens. There was no evidence of residual collagen sponge matrix in any of the rhBMP-2/ACS treated patients. However, specimens obtained from the bone graft treated patients contained acellular remnants of the autograft and or allograft grafting material. Representative examples of the rhBMP-2/ACS induced bone and bone graft induced bone are shown in Figures 7, 8.

Table 3. NEWLY INDUCED BONE DENSITY (mg/cc)

Variable Density at 4 mos

Statistic N (observed) Mean SD Max Median Min N (observed) Mean SD Max Median Min

Bone Graft (n 13) 13 350 243 41 337 797 8 448 213 81 494 689

rhBMP-2/ACS 0.75 mg/mL (n 18) 18 84 50 23 68 240 11 456 131 244 532 615

rhBMP-2/ACS 1.5 mg/mL (n 17) 15 137 77 33 121 335 11 508 126 353 456 705

P Value .0003*

Density at 6 mos post-functional loading

.6452

*P value is from non-parametric Kruskal-Wallis test. P value is from ANOVA F-test. Indicates the P value is less than .05. Boyne at al. Bone Induction With rhBMP-2. J Oral Maxillofac Surg 2005.

BOYNE AT AL

1701

FIGURE 7. Histology from patient # 10 treated with 0.75 mg/mL rhBMP-2. Twenty-eight weeks postoperative (Goldner stain; magnication 10). Boyne at al. Bone Induction With rhBMP-2. J Oral Maxillofac Surg 2005.

FIGURE 8. Histology from patient # 50 treated with combination of demineralized, freeze dried allograft and autograft 34 weeks postoperative (Goldner stain, magnication 10). Boyne at al. Bone Induction With rhBMP-2. J Oral Maxillofac Surg 2005.

Exploratory Analysis In an exploratory analysis of responding patients, a rhBMP-2/ACS responder was dened as any patient who had at 4 months postoperative an increase of greater than 2 mm of alveolar ridge height. Based on this criteria, 94.4% (17/18) and 88.2% (16/17) of the patients treated with 0.75 mg/mL and 1.5 mg/mL, respectively, were considered responders to rhBMP-2/ACS treatment.
PATIENT OUTCOME

Factors Which Could Affect Functional Loading Success Several factors that could affect patient and dental implant success analyses were evaluated for comparability among the treatment groups. Most of the variables evaluated were comparable among the treatment groups; time to dental implant placement (mean, 6.9 1.0 months), crestal bone reduction

before dental implant placement (mean, 79%), dental implant stability on placement (mean, 97%), rootform type (43% press-t, 57% threaded), implant surface (37% rough, 63% smooth), rootform diameter (mean, 4 mm), and rootform length (mean, 12.5 mm). However, 3 differences were detected; bone quality at the time of dental implant placement as assessed by the surgeon, the time interval between dental implant placement and prostheses placement, and the shape of the alveolar ridge. The investigators assessed twice as many patients with type IV bone in the 0.75 mg/mL group as compared with the 1.50 mg/mL rhBMP-2/ACS group at the time of dental implant placement. This rating was consistent with the surgeons anecdotal comments that the bone in the 1.50 mg/mL group had greater resistance to drilling than the 0.75 mg/mL group and felt more like the autograft induced bone (Table 4).

Table 4. BRANEMARK CRITERIA BONE QUALITY ASSESSMENT AT TIME OF DENTAL IMPLANT PLACEMENT

Category N (observed) I* I-II II-III III-IV

Bone Graft (n 13) 12 0 1 7 4

rhBMP-2/ACS 0.75 mg/mL (n 18) 15 0 1 8 6

rhBMP-2/ACS 1.50 mg/mL (n 17) 15 0 3 9 3

Total (n 48) 42 0 5 24 13

*Only dental implants that were placed into newly induced bone were presented. For patients with multiple dental implant scores, the mean score value was used. Branemarks criteria (1 highest quality to 4 lowest quality). Boyne at al. Bone Induction With rhBMP-2. J Oral Maxillofac Surg 2005.

1702

BONE INDUCTION WITH RHBMP-2

Table 5. NUMBER (%) OF PATIENTS WHO RECEIVED PROSTHESIS (FUNCTIONALLY LOADED) AND MAINTAINED FUNCTIONAL LOADING (SUCCESS)

Variable No. (%) received dental implants without additional augmentation (visit 50)* No. (%) received prosthesis (functionally loaded) (visit 60) Remained functionally loaded at 6 mos N (observed) Success (% of observed) Success (% of enrolled) Remained functionally loaded at 12 mos N (observed) Success (% of observed) Success (% of enrolled) Remained functionally loaded at 18 mos N (observed) Success (% of observed) Success (% of enrolled) Remained functionally loaded at 24 mos N (observed) Success (% of observed) Success (% of enrolled) Remained functionally loaded at 30 mos N (observed) Success (% of observed) Success (% of enrolled) Remained functionally loaded at 36 mos N (observed) Success (% of observed) Success (% of enrolled)

Bone Graft (n 13)

rhBMP-2/ACS 0.75 mg/mL (n 18)

rhBMP-2/ACS 1.50 mg/mL (n 17)

13 (100) 13 (100) 13 11 (85) 11 (85) 11 8 (73) 8 (62) 10 9 (90) 9 (69) 10 9 (90) 9 (69) 9 9 (100) 9 (69) 8 8 (100) 8 (62)

15 (83) 13 (72) 12 11 (92) 11 (61) 12 11 (92) 11 (61) 13 12 (92) 12 (67) 12 12 (100) 12 (67) 12 12 (100) 12 (67) 12 12 (100) 12 (67)

15 (88) 14 (82) 14 14 (100) 14 (82) 14 14 (100) 14 (82) 14 14 (100) 14 (82) 14 13 (93) 13 (76) 13 13 (100) 13 (76) 13 13 (100) 13 (76)

*Only patients who received dental implants placed into newly induced bone without additional maxillary sinus oor augmentation were counted as functionally loaded or as successes in subsequent sections of this table. Boyne at al. Bone Induction With rhBMP-2. J Oral Maxillofac Surg 2005.

The time from dental implant placement to prosthesis placement was approximately 2 months longer in the 0.75 mg/mL treatment group (mean, 11.1 3.7 months) compared with bone graft (mean, 8.9 2.4 months) and 1.50 mg/mL rhBMP-2/ACS (mean, 8.3 3.5) treatment groups. In addition, the shape of the alveolar ridge was different among the treatment groups at the time of dental implant placement. A larger proportion of patients in the rhBMP-2/ACS treatment groups had knife edge ridges, 31% and 37% in the 0.75 and 1.50 mg/mL rhBMP-2/ACS treatment groups compared with the bone graft group (11%). This nding is consistent with the CT scan width measurement taken at the crest of the alveolar ridge at 4 months postoperative.
PATIENT AND DENTAL IMPLANT SUCCESS

Bone Graft All 13 patients (100%) in the bone graft treatment group received a total of 63 dental implants into the

newly induced bone within the grafted sinus. These same patients, accounting for 53 of the dental implants (84%), were functionally loaded with prostheses. Eleven of the 13 patients (85%), accounting for 46 dental implants (73%), were functionally loaded at 6 months postloading. Eight of 13 patients (62%), accounting for 32 dental implants (51%), were functionally loaded at 12 months postloading. Nine of 13 patients (69%), accounting for 37 dental implants (59%), were functionally loaded at 18, 24, and 30 months postloading. The count increased from 8 to 9 patients because of 1 patient who regained his/her loading status. Eight of 13 patients (62%), accounting for 33 dental implants (52%), were functionally loaded at 36 months postloading (study completion) (Tables 5 and 6). The dental implant survival rate (regardless of loading status) was 81% (51 of 63 implants). The majority of the failures (67%) occurred before functional loading. The etiology of dental implant failures was pri-

BOYNE AT AL

1703

Table 6. NUMBER (%) OF DENTAL IMPLANTS THAT RECEIVED PROSTHESIS (FUNCTIONALLY LOADED) AND MAINTAINED FUNCTIONAL LOADING (SUCCESS)

Variable Dental implant placed into newly induced bone* Number (%) received prosthesis (functionally loaded) (visit 60) Remained functionally loaded at 6 mos N (observed) Success (% of observed) Success (% of placed) Remained functionally loaded at 12 mos N (observed) Success (% of observed) Success (% of placed) Remained functionally loaded at 18 mos N (observed) Success (% of observed) Success (% of placed) Remained functionally loaded at 24 mos N (observed) Success (% of observed) Success (% of placed) Remained functionally loaded at 30 mos N (observed) Success (% of observed) Success (% of placed) Remained functionally loaded at 36 mos N (observed) Success (% of observed) Success (% of placed)

Bone Graft (n 13) 63 53 (84) 53 46 (87) 46 (73) 39 32 (82) 32 (51) 43 37 (86) 37 (59) 43 37 (86) 37 (59) 39 37 (95) 37 (59) 34 33 (97) 33 (52)

rhBMP-2/ACS 0.75 mg/mL (n 18) 83 65 (78) 59 56 (95) 56 (67) 59 57 (97) 57 (69) 52 50 (96) 50 (60) 59 59 (100) 59 (71) 59 59 (100) 59 (71) 59 59 (100) 59 (71)

rhBMP-2/ACSqa 1.50 mg/mL (n 17) 73 60 (82) 58 55 (95) 55 (75) 47 44 (94) 44 (60) 55 54 (98) 54 (74) 53 52 (98) 52 (71) 50 49 (98) 49 (67) 52 52 (100) 52 (71)

*Only patients who received dental implants placed into newly induced bone without additional maxillary sinus oor augmentation were counted as functionally loaded or as successes in subsequent sections of this table. The percentage is calculated as 100 times the number of implants, which received prosthesis, divided by the number of dental implant placed. Boyne at al. Bone Induction With rhBMP-2. J Oral Maxillofac Surg 2005.

marily inadequate bone quality (92%), with 1 dental implant failing because of infection. 0.75 mg/mL rhBMP-2/ACS Group Fifteen of the 18 patients (83%) randomized to the 0.75 mg/mL treatment group received a total of 83 dental implants into the newly induced bone in the grafted sinus. Two of the 3 patients who did not receive dental implants were treatment failures following their maxillary sinus oor augmentation procedures. The third patient withdrew from the study 16 weeks postsurgery despite a good bone induction response. Thirteen of the 18 patients (72%), accounting for 65 dental implants (78%), were functionally loaded with dental implant-borne prostheses. Eleven of the 18 patients (61%), accounting for 56 dental implants (67%), remained loaded at 6 months and 12 months post-functional loading. Twelve of the 18 patients (67%), accounting for 50 dental implants (60%), were in function at 18 months postloading. One patient regained his/her functional loading status at 18

months post-functional loading. Twelve of 18 patients (67%), accounting for 59 dental implants (71%), were in function at 24 through 36 months post-functional loading (Tables 5 and 6). The dental implant survival rate (regardless of loading status) was 88% (73 of 83 implants). Eighty percent of the failures occurred before functional loading. The etiology of dental implant failures was primarily because of inadequate bone quality (90%). 1.50 mg/mL rhBMP-2/ACS Fifteen of the 17 patients (88%) in the 1.50 mg/mL treatment group received a total of 73 dental implants in the newly induced bone within the grafted sinus. Two patients were treatment failures following their maxillary sinus oor augmentation procedures and required reaugmentation procedures for dental implant placement. Fourteen of the 15 patients (82%), accounting for 60 dental implants (82%), were functionally loaded with dental implant-borne prostheses and remained functionally loaded at 6, 12, and 18

1704

BONE INDUCTION WITH RHBMP-2

Table 7. NUMBER (%) OF PATIENTS WITH FREQUENT ADVERSE EVENTS (>10% OF PATIENTS) BY TREATMENT GROUP, BODY SYSTEM, AND COSTART TERM: FOUR-MONTH POSTOPERATIVE PERIOD

Body System COSTART Term Body as a whole Dehiscence Edema Face edema Headache Pain Digestive system Mouth pain Oral edema Oral erythema Hemic and lymphatic system Ecchymosis Musculoskeletal system Arthralgia Bone disorder Respiratory system Sinusitis Skin and appendages Rash

Bone Graft (n 13) (%) 2 (15) 6 (46) 5 (38) 5 (38) 8 (62) 8 (62) 6 (46) 3 (23) 3 (23) 1 (8) 4 (31) 6 (46)

rhBMP-2/ACS 0.75 mg/mL (n 18) (%) 2 (11) 7 (39) 2 (11) 1 (6) 14 (78) 10 (56) 3 (17) 6 (33) 4 (22) 3 (17) 2 (11)

rhBMP-2/ACS 1.50 mg/mL (n 17) (%) 1 (6) 14 (82) 3 (18) 3 (18) 15 (88) 8 (47) 4 (24) 3 (18) 2 (12) 2 (12) 1 (6) 1 (6)

(n

Total 48) (%) 5 (10) 6 (13) 26 (54) 5 (10) 9 (19) 37 (77) 26 (54) 13 (27) 12 (25) 5 (10) 7 (15) 8 (17) 9 (19)

P Value* Overall .842 .0001 .0132 .349 .0609 .2184 .7652 .2306 .6447 .0648 .5934 .2083 .0243

*P value is from 2-sided Fishers exact test. Indicates P value is less than .05. Boyne at al. Bone Induction With rhBMP-2. J Oral Maxillofac Surg 2005.

months post-functional loading. Thirteen of the 17 patients (76%), accounting for 52 of the 73 dental implants (71%) placed, were loaded at 24 through 36 months postloading (Tables 5 and 6). The dental implant survival rate (regardless of loading status) was 79% (58 of 73 dental implants). Eight of the 15 dental implant failures occurred at either the time of placement (2) or before loading. The etiology of dental implant failures was primarily because of inadequate bone quality (57%) and periimplantitis (27%).
CRESTAL BONE LOSS ADJACENT TO DENTAL IMPLANTS

SAFETY

As expected an average of 1.5 mm of crestal bone adjacent to the dental implants was lost from the time of dental implant placement to prosthesis placement. There appeared to be slightly greater bone loss in the 1.50 mg/mL treatment group (2.0 mm) as compared with the bone graft and 0.75 mg/mL groups; 1.3 mm and 1.4 mm, respectively. Following prosthesis placement (functional loading) to 1 year post-functional loading, crestal bone loss averaged between 0.6 mm and 0.4 mm on the mesial and distal side of the dental implants, respectively. This loss was comparable across the 3 treatment groups. Following 1 to 3 years of functional loading crestal bone loss stabilized to less than 0.2 mm per year. These data are consistent with that reported in the dental implant literature.

All of the patients who participated in the study experienced adverse events. Nearly half of the events (261 of 546) occurred during the rst 4 month postoperative. The majority of the adverse events during the rst 4 months following surgery were mild (56%) or moderate (38%) in severity, and adverse event severity were comparable across treatment groups. The most frequent adverse events, those that occurred in at least 10% of the patients during the rst 4 months postoperative, are summarized in Table 7. These events were transient and consistent with the surgical procedures performed (maxillary sinus oor augmentation procedure, or bone graft harvest procedure). The majority of the events were equally distributed among the treatment groups with the following exceptions. There was a signicantly greater proportion of patients in the bone graft group with edema (P .0023, vs 0.75 mg/mL; and P .0028, vs 1.50 mg/mL) and rash (erythema) (P .0429, vs 0.75 mg/mL; and P .0247, vs 1.50 mg/mL) than in the 2 rhBMP-2 groups. In addition, there was a greater proportion of patients in the bone graft group with pain versus the 2 rhBMP-2 groups. These complaints of edema, rash (erythema), and pain were experienced from the autograft harvest site. Less frequent events, sensory loss (1 patient) and gait disturbance (2 patients), were also experienced by patients in the bone

BOYNE AT AL

1705 tration of rhBMP-2 using the CT scan density data (as opposed to the height and width gain data), 2) the difference in CT scan ridge width data seen between the bone graft group and rhBMP-2/ACS treatment groups, 3) the increased density of the rhBMP-2 induced bone following 6 months of functional loading, and 4) the patient success and dental implant survival rates at 36 months following functional loading. It was anticipated that 1 of the 2 concentrations of rhBMP-2 would generate more bone than the other and have higher patient and dental implant success rates. This was not the case. In this study, both concentrations were equally effective. However, the difference between the 2 doses appeared to be in the rate of bone formation as evidenced by the CT scan density data and the investigators assessment of the bone quality at the time of dental implant placement. This nding led the investigators to delay prosthesis placement by approximately 2 months in the 0.75 mg/mL group, as compared with the bone graft and 1.50 mg/mL treatment groups, to allow the soft quality bone a longer time interval to mature. This increased time interval is important in that it prolonged the patients treatment course. The gain in alveolar ridge width from baseline measurements was to be expected. Width measurements were taken at positions within the sinus where bone or nonmineralized bone occurred throughout the horizontal width measurement lines, which began at the base of the sinus moving superiorly. As a result where bone previously did not exist (ie, before surgery), there was no value for bone width. Once the lower aspect of the sinus had lled with bone, bone width could then be measured and was seen again. Regarding the signicant difference in width gain between the groups, the most notable difference occurred at the base of the sinus (1/4 width measurement). The etiology of this nding is not known. However, following the removal of the bony window at the osteotomy site, it may be possible that the bone graft packed into the sinus maintained or even expanded the lateral dimensions of the sinus cavity, and/or there was some degree of collapse of the lateral wall of the sinus cavity following the implantation of the rhBMP2/ACS. Regardless, in this study alveolar ridge width did not impact the ability to place dental implants. While the results of the CT scan density data at 6 months following functional loading were welcomed, they were also somewhat expected. This same phenomenon had been seen in the rst rhBMP-2 pilot study. One would expect to see an increase in density in newly induced bone following the functional loading of dental implants placed into that bone. The increase in density suggests that the bone is responding to the mechanical forces placed on it and is consistent with Wolffs law.17

graft treatment group; these events resolved by 1 and 2 months, respectively. The 1.50 mg/mL rhBMP-2/ ACS treatment group had a signicantly greater amount of facial edema during the rst 4 months following surgery than did the bone graft group (P .0227) and the 0.75 mg/mL rhBMP-2/ACS group (P .0152). Immune Response Prevalence. None of the patients had antibodies to rhBMP-2 before surgery. Four percent of the patients had antibodies to bovine type I collagen before surgery; 0%, 0%, and 12% (2 patients) in the bone graft, 0.75 mg/mL, and 1.50 mg/mL treatment groups, respectively. Incidence. Four percent of the patients on study had an immune response to rhBMP-2 following study treatment; 0%, 0%, and 12% (2 patients) in the bone graft, 0.75 mg/mL, and 1.50 mg/mL treatment groups, respectively. Both of the responses were transient. Nineteen percent of the patients had an immune response to bovine collagen following study treatment; 23% (3 patients), 11% (2 patients), and 24% (4 patients) in the bone graft, 0.75 mg/mL, and 1.50 mg/mL treatment groups, respectively. The immune responses were considered transient in 6 of 9 patients and positive titers were still present at the follow-up visit in 3 patients. None of the samples that were positive for antibodies to bovine type I collagen had detectable levels of antibodies to human type I collagen. No clinical manifestations of an immune response or a neutralizing effect of the biological activity of rhBMP-2 have been identied in any of the patients described above.

Discussion
This study was designed to determine the most safe and effective concentration of rhBMP-2/ACS for maxillary sinus oor augmentation procedures and to estimate the efcacy rates (patient and dental implant) of rhBMP-2/ACS and bone graft in these procedures. The randomization schedule was effective in evenly distributing among treatment groups demographic and other confounding variables associated with patients. As a result, the safety and efcacy data generated from this study are valid. The higher of the 2 rhBMP-2 concentrations was deemed the most effective for maxillary sinus oor augmentation procedures. Furthermore, the patient and dental implant success rates were estimated and found comparable among the 3 treatment groups following 36 months of functional loading. However, several results from this study require discussion; 1) determining the most effective concen-

1706 The results from this study are based on intent to treat and the patient, as opposed to dental implants, as the primary unit of analysis. As a result it is difcult, if not impossible, to compare the data generated from this study with the nonrandomized case studies reported in the literature. In addition, the true estimation of the success of the maxillary sinus oor augmentation procedure is lacking because no randomized controlled trials have been published in the literature. This paucity of data led to the organization of a sinus graft consensus conference in 1998.18 This consensus conference attempted to determine the efcacy of the sinus oor augmentation bone graft procedure. However, the method by which they obtained the retrospective data to make this assessment was awed in that only cases in which endosseous implants were placed and functionally loaded for at least 1 year were accepted. This inclusion criteria prevented the calculation of patient success because the total number of patients who received the grafting procedure was not collected. It also prevented a correct calculation of dental implant survival because the total number of dental implants placed could not be tabulated. At best we can compare the results from this study with retrospective and prospective studies conducted in patients with 2-stage maxillary sinus oor augmentation procedures using autograft alone or autograft combined with allograft as the bone grafting material. Fifteen such studies were reviewed and found generally not to report success in terms of the number of patients with successful grafting procedures. However, a study reported by Blomqvist et al19 was a notable exception. This study evaluated autograft bone harvested from the iliac crest in 2-stage maxillary sinus oor augmentation. The patient success rate (the number of patients who received prosthesis) was 76% (38 of 50 patients) and the dental implant survival rate was 84% after an average of 16 months (range, 0 to 34 months) following prosthesis placement. The remaining studies consistently reported dental implant survival rate (meaning the number of implants not removed) from 57% to 100%.20-33 The dental implant survival rate in these studies is comparable to the survival rate reported in this study after 36 months of functional loading; 81%, 88%, and 79% in the bone graft, 0.75 mg/mL, and 1.50 mg/mL treatment groups, respectively. Finally, the safety data from this study supported not only the safety of rhBMP-2/ACS for maxillary sinus oor augmentation procedures, but captured the morbidity incurred by patients who undergo an autograft procedure for maxillary sinus oor augmentation. The primary objective of this study was to evaluate the safety and efcacy of 2 concentrations of rhBMP2/ACS, as compared with bone graft, to induce bone

BONE INDUCTION WITH RHBMP-2

in maxillary sinus oor augmentation procedures to select 1 concentration for investigation in a future pivotal study. Secondarily, we wanted to estimate patient and dental implant success rates in this procedure. Based on these data, we conclude that: 1) both concentrations of rhBMP-2 are safe and have a safety prole similar to bone graft; 2) both concentrations of rhBMP-2 induce a similar amount of bone which is similar to that induced by bone graft; 3) the higher concentration of rhBMP-2 induces bone formation more rapidly as compared with the lower concentration; and 4) rhBMP-2/ACS is effective in inducing bone following a maxillary sinus oor augmentation procedure which can support the placement and long-term functional loading of dental implants in approximately 75% to 80% of the patients treated. These study results support the use of rhBMP2/ACS at a concentration of 1.50 mg/mL in future studies of maxillary sinus oor augmentation. Acknowledgments
The authors acknowledge the nurses, study coordinators, and patients for their hard work and participation in this study. In addition, they would like to acknowledge X. Jian Li, MD, PhD (Wyeth) for evaluating and providing the histology photomicrographs and Hugh Xiao, PhD for his statistical knowledge in the analysis of this study. Lastly, they would like to acknowledge the pioneering efforts of Dr Marden Alder, Chris Peach, Dr Nummikoski, and their colleagues at the University of Texas Health Science Center at San Antonio for their collaborative approach in developing the CT scan and periapical measurement methodology.

References
1. Boyne PJ, James RA: Grafting of the maxillary sinus oor with autogenous marrow and bone. J Oral Surg 38:613, 1980 2. Tatum H: Maxillary sinus and implant reconstruction. Dental Clin North Am 30:107, 1986 3. Smiler DG, Johnson PW, Lozado JL, et al: Sinus lift grafts and endosseous implants. Treatment of the atrophic posterior maxilla. Dent Clin North Am 36:151, 1992 4. Urist MR: Bone: Formation by autoinduction. Science 150:893, 1965 5. Urist MR, Mikulski A, Lietze A: Solubilized and insolubilized bone morphogenetic protein. Proc Natl Acad Sci U S A 76: 1828, 1979 6. Bentz H, Nathan RM, Rosen DM, et al: Purication and characterization of a unique osteoinductive factor from bovine bone. J Biol Chem 264:20805, 1989 7. Wozney JM, Rosen V, Celeste AJ, et al: Novel regulators of bone formation: Molecular clones and activities. Science 242:1528, 1988 8. Wozney JM: Bone morphogenetic proteins and their gene expression, in Noda M (ed): Cellular and Molecular Biology of Bone. San Diego, CA, Academic Press, 1993, pp 131-167 9. Wozney JM: BMPs: Roles in bone development and repair, in Proceedings of the Portland Bone Symposium, Aug 2-5, 1995; Portland, OR 10. Wang EA, Rosen V, DAlessandro JS, et al: Recombinant human bone morphogenetic protein induces bone formation. Proc Natl Acad Sci U S A 87:2220, 1990 11. Cochran DL, Nummikoski PV, Jones AA, et al: Radiographic analysis of regenerated bone around endosseous implants in the canine using recombinant human bone morphogenetic protein-2. Int J Oral Maxillofac Implant 12:739, 1997 12. Hanisch O, Tatakis DN, Boskovic MM, et al: Bone formation and reosseointegration in peri-implantitis defects following sur-

BOYNE AT AL
gical implantation of rhBMP-2. Int J Oral Maxillofac Implant 12:604, 1997 Zellin G, Linde A: Importance of delivery systems for growthstimulatory factors in combination with osteopromotive membranes. An experimental study using rhBMP-2 in rat mandibular defects. J Biomed Mater Res 35:181, 1997 Nevins M, Kirker-Head C, Nevins M, et al: Bone formation in the goat maxillary sinus induced by absorbable collagen sponge implants impregnated with recombinant human bone morphogenetic protein-2. Int J Peridont Rest Dent 16:9, 1996 Boyne PJ, Marx RE, Nevins M, et al: A feasibility study evaluating rhBMP-2/absorbable collagen sponge for maxillary sinus oor augmentation. Int J Periodont Rest Dent 17:11, 1997 Webber RL, Ruttimann UE, Groenhuis RAJ: Computer correction of projective distortions in dental radiographs. J Dent Res 63:1032, 1984 Wolff J: Das gaetz der Transformation. Transformation der Knocken, 1892 Jensen O, Shulman L, Block M, et al: Report of the Sinus Consensus Conference of 1996. Int J Oral Maxillofac Implants 13:(suppl), 1993 Blomqvist JE, Alberius P, Isaksson S: Two-stage maxillary sinus reconstruction with endosseous implants: A prospective study. Int J Oral Maxillofac Implants 13:758, 1998 Jensen J, Simonsen, EK, Sindet-Pedersen S: Reconstruction of the severely resorbed maxilla with bone grafting and osseointegrated implants: A preliminary report. J Oral Maxillofac Surg 48:27; discussion 33, 1990 Tidwell JK, Blijdorp PA, Stoelinga PJ, et al: Composite grafting of the maxillary sinus for placement of endosteal implants. A preliminary report of 48 patients. Int J Oral Maxillofac Surg 21:204, 1992 Lundgren S, Moy P, Johansson C, et al: Augmentation of the maxillary sinus oor with particulated mandible: A histologic and histomorphometric study. Int J Oral Maxillofac Implants 11:760, 1996 Lundgren S, Nystrom E, Nilson H, et al: Bone grafting to the maxillary sinuses, nasal oor and anterior maxilla in the atro-

1707
phic edentulous maxilla. A two-stage technique. Int J Oral Maxillofac Surg 26:428, 1997 Watzek G, Weber R, Bernhart T, et al: Treatment of patients with extreme maxillary atrophy using sinus oor augmentation and implants: Preliminary results. Int J Oral Maxillofac Surg 27:428, 1998 van den Bergh JP, ten Bruggenkate CM, Krekeler G, et al: Sinus oor elevation and grafting with autogenous iliac crest bone. Clin Oral Implants Res 9:429, 1998 Reinert S, Konig S, Eunger H, et al: Follow-up studies of 3-dimensional osteoplastic reconstruction of the extremely atrophied maxilla combined with implants. Mund Kiefer Gesichtschir 3:S30, 1999 (suppl 1) Majorana C, Redemagni M, Rabagliati M, et al: Treatment of maxillary ridge resorption by sinus augmentation with iliac cancellous bone, anorganic bovine bone, and endosseous implants: A clinical and histologic report. Int J Oral Maxillofac Implants 15:8973, 2000 Krekmanov L, Heimdahl A: Bone grafting to the maxillary sinus from the lateral side of the mandible. Br J Oral Maxillofac Surg 38:617, 2000 Krekmanov L: Placement of posterior mandibular and maxillary implants in patients with severe bone deciency: A clinical report of procedure. Int J Oral Maxillofac Surg 15:722, 2000 van den Bergh JPA, ten Bruggenkate CM, Groeneveld HHJ, et al: Recombinant human bone morphogeneic protein-7 in maxillary sinus oor elevation in 3 patients compared to autologous bone grafts: A clinical pilot study. J Clin Periodontal 27:627, 2000 Wannfors K, Johansson B, Hallman M, et al: A prospective randomized study of 1- and 2-stage sinus inlay bone grafts: 1-Year follow-up. Int J Oral Maxillofac Surg 27:627, 2000 Widmark G, Andersson B, Carlson GE, et al: Rehabilitation of patients with severely resorbed maxillae by means of implants with or without bone grafts: A 3-to 5-year follow-up clinical report. Int J Oral Maxillofac Implants 16:73, 2001 Raghoebar GM, Timmenga N, Reintsema H, et al: Maxillary bone grafting for insertion of endosseous implants: Results after 12-124 months. Clin Oral Implant Res 12:279, 2001

13.

24.

14.

25. 26.

15.

16.

27.

17. 18.

28. 29. 30.

19.

20.

21.

31. 32.

22.

33.

23.