SCIENTIFIC ARTICLES

J Oral Maxillofac Surg 63:63-67, 2005

Biologic Mechanical Advantages of 3 Different Cranial Bone Grafting Techniques for Implant Reconstruction of the Atrophic Maxilla
Peter D. Waite, MPH, DDS, MD,* Panunn Sastravaha, DDS, MS, and Jack E. Lemons, PhD
Purpose: The purpose of this study was to test the mechanical capacities of 3 different bone grafting

techniques in the atrophic maxilla when co-stabilized with dental implants. Reconstruction of the atrophic maxilla is a difcult clinical challenge and implants cannot be placed without adequate bone. Methods: The biomechanical performance of 3 different grafting techniques was evaluated in vitro using a maxillary model, cadaveric cranial bone blocks, and dental implants. A maxillary model fabricated from polyurethane (sawbone) was selected as a substrate for this study because of consistency in shape, size, and mechanical properties. This anatomic model was more consistent than different cadaveric maxilla, where signicant variation was found to exist among atrophic specimens. Cadaveric cranial bone graft blocks were secured to the model maxilla (sandwich, ridge only, and sinus inlay) with a dental implant. The strength of the implant/bone graft complex was tested to failure in an Instron machine (Instron Inc, Canton, MA). Results: The 3 bone grafting methods showed signicantly different deformation and strength characteristics. The sandwich technique enhanced resistance to deformation under higher imposed loads. The location of the graft inuenced the overall mechanical performance (eg, the ridge onlay) and showed a signicantly higher resistance to compressive loads applied toward the alveolar ridge (mastication force). Conclusion: The ridge onlay grafting procedures created a higher biomechanical tolerance to imposed load than the sinus grafting (sinus inlay). Sinus grafting, although successful, was not the most ideal location for immediate mechanical loading resistance when compared with ridge augmentation in this in vitro model. 2005 American Association of Oral and Maxillofacial Surgeons J Oral Maxillofac Surg 63:63-67, 2005 The severely atrophic maxilla is extremely difcult to reconstruct, and the sinus lift procedure does not augment the resorbed alveolar bone. Maxillary im*Professor and Chairman, Department of Oral & Maxillofacial Surgery, University of Alabama at Birmingham, School of Dentistry, Birmingham, AL. Department of Oral & Maxillofacial Surgery, Faculty of Dentistry, Chulalongkorn University, Pthumwan, Bangkok, Thailand. Professor, Prosthodontics & Biomaterials, University of Alabama at Birmingham, School of Dentistry, Birmingham, AL. Address correspondence and reprint requests to Dr Waite: Department of Oral & Maxillofacial Surgery, University of Alabama at Birmingham, 1919 7th Ave South, School of Dentistry Building, Room 419, Birmingham, AL 35294; e-mail: pwaite@uab.edu
2005 American Association of Oral and Maxillofacial Surgeons

0278-2391/05/6301-0012$30.00/0 doi:10.1016/j.joms.2004.09.007

plant rehabilitation is possible if sufcient bone quantity and quality exist. Reconstruction of the maxillary is much different than in the mandible because of the quality of bone, resorption processes, and anatomic variations of the sinus. Longitudinal studies indicate the survival rate of mandibular and maxillary implants to be 91% to 99% and 84% to 92%, respectively.1 Alveolar augmentation with iliac bone and immediate implants has not been consistently successful. Vertical ridge augmentation with cortical blocks or veneer graft have become popular, but no studies have evaluated the biomechanical advantage.2 The purpose of this study is to measure the loading strength of dental implants when placed with cortical bone blocks. Implant stability is dependent on sufcient bone volume and quality.2 It is often impossible to place implants without bone grafting. Sailer3 has summarized several procedures: 1) sinus grafting with can63

64 cellous bone and marrow, and/or other grafting materials around implant; 2) a sinus graft with corticocancellous bone with implant placement; 3) corticocancellous onlay bone graft with immediate implant placement or as a 2-staged procedure; 4) a transplantation of a preformed composite bone graft containing implants; 5) a horizontal osteotomy and an interpositional bone graft with a delayed implant placement; 6) a Le Fort I osteotomy with an interpositional bone graft including an immediate implant placement or as a 2-staged procedure.3 In the past, endochondral bone grafts from the ilium have been the choice for grafting various bone defects, but cranial bone grafting has become more popular and has several advantages such as resistance to resorption, ease of bone harvest, same surgical eld, and a minimal donor site morbidity.4 The calvarium is an established donor area and has been extensively used in the reconstruction of trauma and congenital cranial defects.5 Maxillary implant placement with calvarial bone has been shown to give similar survival rate as the anterior mandible.6,7 Hence, the aim of this study is to investigate the biomechanical conditions of cortical block bone grafted at different locations in the posterior maxilla with simultaneous implant placement. A study of this type required a stable, consistent substrate that reproduces the maxillary anatomy to compare different block graft locations.

IMPLANT RECONSTRUCTION OF THE ATROPHIC MAXILLA

FIGURE 1. Three different simultaneously grafted and implantation procedures. (a) the sinus-only graft (SOG), (b) the onlay graft (OG), (c) the sinus graft (SG), and (d) control (CG). Waite, Sastravaha, and Lemons. Implant Reconstruction of the Atrophic Maxilla. J Oral Maxillofac Surg 2005.

experimental model much different than the control alone. Group 1. The sinus and onlay sandwich graft (SW): Cranial bone blocks were placed at the sinus oor and at the alveolar ridge. Group 2. The onlay ridge graft (augmentation [RO]): Cranial bone block was placed only on the alveolar ridge as augmentation. Group 3. The sinus inlay graft (SI): Cranial bone block was placed on the maxillary sinus side. Group 4. The control group (CG): The polyurethane atrophic maxillary model was used without bone grafting. Each group consisted of 10 specimens. Implants were surgically placed into a total of 40 prepared models at the maxillary rst molar position. An Instron machine (Instron Inc, Canton, MA) was used to generate an axial and shear oriented force toward the apex of the implant with the displacement speed of loading at 0.125 cm/min. The force-versusdeformation relationship for vertical and 45 force applications were recorded for initial failure and nal push-out strength. This force simulates axial loading such as in masticatory and lateral shear strength. The 45 force application was rst conducted using compressive type loads of 0 to 22.24 N by direct testing along the push rod pathway (Fig 2). The vertical compressive force was performed by placing the model directly onto the compressive load cell (Fig 3). Three repeated vertical compressive loads of 0 to 88.96 N were applied for the SW and RO groups, whereas compressive oriented loads of only 0 to 44.48 N were applied to the SI and control groups. After the last load had been applied, each specimen

Materials and Methods

Using a Le Fort I osteotomy technique, cadaveric edentulous maxillae were harvested from 21 xed cadavers. An average atrophic maxilla was selected from a 54-year-old frozen male cadaver for a standardized dimension. At the rst premolar and rst molar regions, respectively, the alveolar ridge was 4 mm in width and 4 mm in height. This maxilla posterior segment was used as a template for fabrication of a polyurethane replicated model. A polyurethane resin material (Pacic Research Laboratories, Vashon, WA) was used because it could be reproducibly processed and possessed physical characteristics more similar to a natural cadaveric maxilla compared with other routinely available synthetic materials.8 Twenty split thickness cadaveric calvarial bone strips were harvested. The parietal bone area was chosen and an orthopedic sagittal saw was used to cut a full thickness from each side. The outer and inner tables were then separated, resulting in 10 mm 20 mm 3 mm standardized pieces. Three different bone grafting techniques were performed in each group and the fourth was not grafted as a control (Fig 1). Obviously, unhealed corticocancellous bone (as in iliac particulate) packed around an implant would not provide an immediate

WAITE, SASTRAVAHA, AND LEMONS

65 the axial and 45 directions, the load (N) to initial failure, the stress (N/m2) to initial failure, the stressstrain modulus (N/m2) ratios, the nal push-out forces (N), and the nal push-out forces per collar area (N/m2) of the implant. A statistic analysis system (SAS) software program was used to conduct statistical analysis on all data. At a 95% level of condence, the variables were summarized as mean standard error. Each measure was considered statistically insignicant if the P value was more than .05. F-ratios at a signicant level of P .05 were calculated to determine the presence, if any, of the differences among the groups. To conrm the differences between specied groups, least-square-mean different (LSD) multiple comparison procedures were carried out at a signicant level of P .05.

Results
The load-versus-deformation stiffness ratios were dened as the amount of load that caused 1 unit of dimensional change. These stiffnesses were calcu-

FIGURE 2. The prepared model under compressive imposed load in 45-oriented direction. Waite, Sastravaha, and Lemons. Implant Reconstruction of the Atrophic Maxilla. J Oral Maxillofac Surg 2005.

was loaded until the point of initial deformation, determined as the highest point just before the observable declination of force versus deformation curve. The nal testing procedure was a vertical compressive oriented load to the implant disrupting the model interface. A xture was designed to support the model and allow the implant to be pushed through the model. The highest load level was determined as the nal push-out force magnitude. The implant thread and collar contact areas in the model and/or bone graft was calculated using direct visualization (Micro-Vu Instrumentation; Micro-Vu Corp, Winsor, CA). The number of the implant threads in the thickness of bone was determined. The implant thread surface contact area was determined by the formula Ac (r22 r12). In this formula, Ac was the surface contact of the implant, r2 was the radius of the implant collar (2.0 mm) and r1 was the radius of the implant body (1.62 mm). This measured the amount of implant surface engaged in the bone. The measurements evaluated in this study were the load-versus-deformation stiffness ratios performed in

FIGURE 3. The prepared model under compressive imposed load in vertically oriented direction. Waite, Sastravaha, and Lemons. Implant Reconstruction of the Atrophic Maxilla. J Oral Maxillofac Surg 2005.

66

IMPLANT RECONSTRUCTION OF THE ATROPHIC MAXILLA

FIGURE 4. Load-deformation stiffness ratio in axial direction. SW, sinus and onlay graft; RO, onlay graft; SI, sinus graft; CG, control. Waite, Sastravaha, and Lemons. Implant Reconstruction of the Atrophic Maxilla. J Oral Maxillofac Surg 2005.

FIGURE 6. Load to initial failure. SW, sinus and onlay graft; RO, onlay graft, SI, sinus graft; CG, control. Waite, Sastravaha, and Lemons. Implant Reconstruction of the Atrophic Maxilla. J Oral Maxillofac Surg 2005.

lated directly from each deformation curve. In the axial direction, the intergroup average of ratios was statistically different (P .0001). The results are shown in Figure 4. The LSD multiple comparisons conrmed the signicant differences, a P value less than .05. The resistance of SW to the deformation under an imposed load was higher. In the 45 direction, the intergroup average of the ratios also showed signicant differences (P .0001; Fig 5). However, LSD multiple comparisons showed no signicant differences between SW and RO groups (P .11) nor between RO and SI groups (P .11). There was a statistically signicant difference between the SW and the SI groups (P .004). The load to initial failure (Fig 6) was signicantly different (P .0001). LSD multiple comparisons showed statistically signicant differences between each pair (P .0001), except between SW and RO (P .06), which had calvarial bone placed on the alveolar ridges. Compared with the control, the load to initial failure in all models was higher (P .0001). LSD multiple comparisons did not show a signicant difference between SI and CG groups (P .0504), whereas the others were signicantly different (P .0001). Before the interface disruptions be-

tween the implant and the models, RO was shown to possess the highest internal force to resist the external imposed load per unit area. A deformation magnitude of each model under an imposed load in this study was determined by a direct calculation from a linear proportion of load-deformation curve. The original contact zone was set as the depth that the implant was engaged. It was a summary of the thickness of the calvarial bone as sinus and/or onlay graft (3 mm), and the thickness of the model (5.1 mm) was dependent on the group. The original dimension of the SW group was 11.1 mm and that of the SI and RO group was 8.1 mm. LSD multiple comparisons indicated that all grafted models, regardless of the grafting methods, possessed a higher stiffness than the controls. There were intergroup significant differences between the groups on which the onlay grafts were placed. No signicant differences were observed between SW and RO.

Discussion
Reconstruction of severely atrophic maxillae remains a major challenge for oral and maxillofacial surgeons. This study investigated the biomechanical characteristics of 3 different block grafting procedures (sinus and onlay, onlay, and sinus grafts alone) with simultaneous implant placement. A split-thickness cranial bone was chosen as a grafting material but cortical bone from the mandible could have been used if the quantity could have been harvested and standardized. The synthetic model was selected to use in the testing procedure instead of a cadaveric maxillary bone to eliminate anatomic variations found for human specimens. This standardized the amount of support from the maxilla per se and allowed more accurate testing of the bone grafting technique. Even though it was impossible to fabricate a synthetic model from a material that had exactly the same

FIGURE 5. Load-deformation stiffness ratio in 45 direction. SW, sinus and onlay graft; RO, onlay graft; SI, sinus graft, CG, control. Waite, Sastravaha, and Lemons. Implant Reconstruction of the Atrophic Maxilla. J Oral Maxillofac Surg 2005.

WAITE, SASTRAVAHA, AND LEMONS

67 on the sinus side and compressive force on the alveolar ridge side. Hence, the timing of implant incorporation, osseointegration healing, and functional loading are important clinical considerations. The parameters that directly demonstrated the ability of cranial block bone to increase the resistance of the host bone-graft implant matrix under the imposed load were the values of the nal push-out force and the nal push-out force per area. The magnitude of these forces was high enough to cause the system failure by pushing implant through each model. Our outcome showed that the value of the nal push-out force was much higher than the load levels that created the disruption of the implant thread interface. Logically, the value of the nal push-out is dependent on the amount of cranial bone used. However, our results showed that the SW and RO groups were able to tolerate an approximately equal magnitude of imposed load before system failure. The overall outcome in this study supports the onlay block graft procedure. However, it cannot be stated from a clinical perspective that an application of grafting materials on the alveolar ridge is superior to that of the sinus oor in spite of the result that the onlay graft was able to tolerate higher compressive load.

physical properties as those of cadaveric maxillary bone, it was important to select a model whose stiffness was most similar to the maxillary bone but less than the cranial implant interface.8 Unless this could be achieved, the biomechanical properties of different grafting procedures under the compressive load might be misrepresented. The maxillary bone segment represented a dimensionally and anatomically average structure for our simulated 1-stage horizontal implant/bone grafting procedure. When performing this procedure, the residual alveolar ridge height should be at least 3 to 6 mm to achieve an initial stability of the implant.9 The replicated maxilla also eliminated practical difculties for the testing procedure by excluding the maxilla with any impacted tooth, rough, irregular, knife-edged ridges, and those with a perforated or septated maxillary antrum. Our results showed that each group could withstand a deformation under the vertical compressive load that was approximately 4 times greater than that for the 45-oriented load. This conrmed that the implant system was not able to withstand a rotating type shear-oriented force compared with the directly axial compressive load.10 It is also shown that grafting procedures increased the relative bone mass of the host alveolar bone. Consequently, the ability of the system to tolerate an imposed load was increased. The RO group and the SW group were able to withstand twice the magnitude of force as the SI group. This implies that not only the amount but also the location of the grafting material applied to the residual alveolar ridge is able to inuence the biomechnical responses of the system. Clinically, this may indicate block grafting has greater mechanical advantage than particulate sinus grafting. The bone graft implant matrix tolerated a much higher magnitude of imposed compressive load for the onlay grafting procedures than the SI grafting method. It can be explained by the compressive force applied to the occlusal surface, which pushed the onlay graft toward the alveolar ridge but pushed the sinus graft and maxillary bone apart. Thus, the system lost the contribution related to mass, resulting in reduction of total strength. The amount of load (74.78 to 199.14 N) that was able to initiate the disruption of an implant thread interface was within the magnitude range of bite force (42 to 1,245 N). It is worth noting that if premature load is implied, the sinus graft implant matrix is much more susceptible to postoperative failure compared with the onlay implant matrix. This may be because of the presence of tensile force

References
1. Nystrom E, Legrell PE, Frossell A, et al: Combined use of bone grafts and implants in the severely resorbed maxilla: A 2 year longitudinal study. Int J Oral Maxillofac Surg 24:20, 1995 2. Widmark G, Anderson B, Ivanoff CJ: Mandibular bone graft in the anterior maxilla for single-tooth implants: Presentation of a surgical method. Int J Oral Maxillofac Surg 26:106, 1997 3. Sailer HF: A new method of inserting endosseous implants in totally atrophic maxilla. J Craniomaxillofac Surg 17:299, 1989 4. Zins JE, Whitaker LA: Membranous versus endochondral bone: Implications for craniofacial reconstruction. Plast Reconstr Surg 72:778, 1983 5. Kline RM, Wolfe SA: Complications associated with harvesting cranial bone grafts. Plast Reconstr Surg 95:5, 1995 6. Branemark PI, Zarb GA, Albrektsson T: Tissue-Integrated Prostheses: Osseointegration in Clinical Dentistry. Chicago, IL, Quintessence, 1985 7. Donovan MG, Dickerson NC, Hanson LJ, et al: Maxillary and mandibular reconstruction using calvarial bone grafts and Branemark implants: A preliminary report. J Oral Maxillofac Surg 52:588, 1994 8. Bredbenner TL, Haug RH: Substitutes for human cadaveric bone in maxillofacial rigid xation research. Oral Surg Oral Med Oral Path Oral Radiol Endod 90:574, 2000 9. Kent J, Block M: Simultaneous maxillary sinus oor bone grafting and placement of hydroxyapatite-coated implants. J Oral Maxillofac Surg 47:238, 1989 10. Bidez MW, Mich CE: Clinical Biomechanics in Implant Dentistry: Contemporary Implant Dentistry. St Louis, MO, Mosby, 1999