Finite M.
element R. Rieger,
University
M.S.,
analysis Ph.D.,’
M.
Mayberry,
of six endosseous M.S.,”
and
M.
0.
implants
Brose,
D.D.S.“’
of Texas Health Science Center, Dental Branch, Houston, Tex.
Stress magnitudes and contours in bone surrounding six endosteal post-type dental implants were calculated by using the finite element method. Comparisons were made by using Branemark, Core-Vent, Denar, Miter, Stryker, and experimental implant designs. Although certain assumptions were made that could be considered controversial, this study concluded that apical “punching stresses” with all of the implants were probably not clinically significant. Saucerization resulting from biomechanical overloads could be a possibility for three of the implants. Problems related to combinations of overloads and underloads at the same time were suggested for several more popular implants in the United States. Additional research, combining 3-D finite element models and clinical studies, was recommended for all commercially available dental implants. (J PROSTHET DENT 1990;63:671-6.)
A
lthough many clinicians have not accepted implant technology, the use of implants is becoming more popular. Presently 25 different post-type endosseous implant systems are available. The variety of implant systems has created confusion because most clinical studies on past implant successesand failures have not been well documented. This study compared the stress patterns in cortical bone surrounding six post-type endosseous implants by using finite element analysis.
LITERATURE
REVIEW
Stress or strain-induced bone remodeling is a concept proposed by Julius Wolff in 1870. Although the magnitudes and directions of stressesthat stimulate bone apposition or resorption are still topics of controversy among researchers, most agree that stresses play an important role as a feedback mechanism in bone remodeling. Several theories have been proposed concerning the reasons for stress-induced bone growth. Justus and Luftl found that altered strain in hydroxyapatite crystals caused the generation of additional calcium ions. Bone cells may react to this change by either bone apposition or resorption. The piezoelectric properites of bone have also been used to explain the stress dependency of bone remodeling.2-4 Bone produces electrical current and potentials proportional to the magnitude of applied stress with a polarity determined by the stress direct,ion. Apposition is thought to occur in regions of negatively charged bone. The stress-related
Iii
*Associate Professor and Director of Oral Biomaterials Research. “Biomedical Engineer, Biomedical Engineering Center, Ohio State University, Columbus, Ohio. *“Assistant Professor, Restorative and Prosthetic Dentistry, Ohio State University, College of Dentistry, Columbus, Ohio. 10/l/13679
THE
JOURNAL
I
Fig.
OF PROSTHETIC
DENTISTRY
5nite
1
i I I t
I
element
I
model
I
I
of Branemar
1
*k im!plant.
671
RIEGER,
Fig.
AND
BROSE
2. Finite element model of Core-Vent implant. Fig.
feedback mechanism is likely a combination of the biochemical and electrical activity. The actual magnitudes of stress that cause bone remodeling are of particular importance to this study. It has been suggested that there is an optimal value of stress for which as much bone is removed by resorption as is deposited by apposition. With stresses above this value, hypertrophy takes place; below this value, atrophy occurs. There is also a maximum stress limit above which stresses will destroy bone by pathologic resorption. Chamay and Tschantz5 concluded that compressive stressespromoted bone growth whereas tensile stresses caused resorption. In a review of functional adaptation of bone however, Fung6 reported that effects of tensile and compressive stresses were the same.
672
MAYBERRY,
3. Finite element model of Denar implant.
As important as these topics are to dental implantology, only one study has ever tried to measure the actual magnitudes of stresses promoting bone growth.7 By inserting load cells into the calvaria of rabbits, a maximum bone formation rate was noted at a compressive stress of 250 psi. The bone formation rate dropped to the control level at stresses greater than 400 psi. Most clinical research of post-type endosseous implants indicates that bone adaptation or bonding to implants is critical to the successof implants. Implant successwas defined by Schnitman and Shulmar? as mobility of less than 1 mm in any direction, bone loss no greater than one third of the vertical height of the implant, and functional service for 5 years in 75 % of all patients. These guidelines may not
JUNE
1990
VOLUME
03
NUMBER
6
FINITE
ELEMENT
ANALYSIS
OF IMPLANTS
Fig. Fig.
4. Finite element model of Miter implant.
be strict enough because implant mobility to any extent generally indicates impending failure. Extensive stress analyses of endosseous imp1ant.s is vital to the continued growth and successof dental implantology.
MATERIAL
AND
METHODS
Six post-type endosseous implants were selected for this study: Branemark (Nobelpharma USA, Chicago, Ill.), CoreVent (Core-Vent Corp., Encino, Calif.), Denar (Steri-Oss, Anaheim, Calif.), Miter (Miter, Inc., Warsaw, Ind.),Driskell (Driskell Bioengineering, Galena, Ohio), and an experimental implant. All implants were assumed to have an elastic modulus of 1.59 x lo7 psi. The finite element mod-
THE
JOURNAL
5. Finite element model of Driskell implant.
OF PROSTHETIC
DENTISTRY
els of these implants are shown in Figs. 1 through 6. The axisymmetric models were made of parabolic, isoparametric elements. Most of these elements were quadrilateral, but some were triangular to accommodate specific geometries. Dimensions were taken from actual implants or drawings supplied by the manufacturers. The implants were assumed to be surrounded and bonded to cortical bone with an elastic modulus of 1.98 X lo6 psi. An axial load of 25 pounds was applied over the top surface of each implant. The Branemark, Core-Vent, and Denar implants each have a threaded portion. The models of these implants assumed a symmetry about the long axis. Although this type of symmetry does not exist for threads, the stresses would
673
RIEGER,
MAYBERRY,
AND
BROSE
Table I. Maximum bone stresses Implant
Stress
Branemark Core-Vent Denar Miter Driskell Experimental
Fig.
6. Finite element model of experimental implant.
not be significantly affected by this assumption. The CoreVent implant was difficult to model axisymmetrically because of the vents in the inverted basket portion. These holes were therefore not included in the model. The Miter, Driskell, and experimental implants have serrations. These serrations were well represented axisymmetrically. The bone core was 5 mm in radius for all implants. The finite element models were generated and solved by using a commercial computer-aided engineering program (Structural Dynamics Research Corporation, Cincinnati, Ohio). The software was mounted on a VAX 8500 computer (Digital Equipment Corporation, Maynard, Mass.). RESULTS
AND
DISCUSSION
The magnitudes of the stresses are defined in Fig. 7. Stress contour plots generated for each implant are shown 674
(X
lo6 psi)
478 661 747
794 868 417
in Figs. 8 through 13. The maximum stress found with each implant is shown in Table I. For purposes of discussion, certain assumptions will be made, based on the literature.7 First, optimal bone maintenance occurs at 250 psi. This may be found in the green area of the contour plots. Pathologic resorption of bone occurs at greater than 700 psi. This may be found in the light purple or dark purple areas. Bone atrophy occurs at less than 200 psi. This may be found in the light blue or dark blue areas. These assumptions may not be valid when poor stress distribution exists. The Branemark implant is shown in Fig. 8. Except for the maximum stress of 478 psi found near the neck, this implant transfers relatively low stressesto the bone. Thus, pathologic resorption of bone would be unlikely with this implant. What may be of most concern is the poor stress distribution and the possibility for atrophy with this implant. Stressesare concentrated at the neck and the apex with little stress transfer along the middle. In a freestanding mode with 25 pounds of force applied as modeled, virtually a fourth of the surrounding cortical bone could be hypocalcified. If placed in “group-function,” as recommended by the manufacturer, where less force is applied occlusally, almost half of the surrounding cortical bone could decalcify to a level below normal. Further, the amount of normally calcified cortical bone at the neck and apex could be reduced. The Core-Vent implant is shown in Fig. 9. Because of the elimination of the vents, this model is not accurate. However, certain features can still be discussed. A maximum stress of 661 psi may be found at the neck of this implant. The chances for pathologic resorption of bone are unlikely. As modeled, this implant is similar to the Branemark implant in that stress distribution is poor. Nearly half of the cortical bone could be hypocalcified. It may be assumed that vents would improve the situation. In actual practice, it has been noted that cortical bone does not always fill the inverted basket of this implant. Such a circumstance could seriously affect the successof this implant because a lack of “internal” support would lead to markedly increased punching stresses at the apex. The Denar implant is shown in Fig. 10. A maximum stress of 747 psi can be seen at the neck of this implant. Pathologic resorption of bone could occur in this region. Like the previous two implants, nearly half of the cortical bone along the length of this implant could be hypocalciJUNE
1990
VOLUME
63
NUMBER
6
FINITE
ELEMENT
ANALYSIS
OF IMPLANTS
Fig. 7. Key to stress magnitudes.
Fig. Fig. Fig. Fig.
8. Stress distribution for Branemark implant. 9. Stress distribution for Core-Vent implant. 10. Stress distribution for Denar implant. 11. Stress distribution for Miter implant.
fied. Punching stress at the apex could also be a problem. The Miter implant is shown in Fig. 11. A maximum stress of 794 psi can be seen at the neck. Pathologic resorption of bone could occur to the first serration; and indeed, there is clinical evidence of saucerization in the absence of inflammation with this implant. Unlike the previous implants, the stress distribution for this implant is good. Although such is not a part of this study, this implant could function well as a free-standing implant. The Driskell implant is shown in Fig. 12. A maximum THE
JOURNAL
OF PROSTHETIC
DENTISTRY
stress of 868 psi may be seen in Fig. 12, A at the neck of this implant. This is the greatest stress found with any of the six implants. Pathologic resorption of bone could occur at the neck of this implant. Approximately a third of the surrounding cortical bone could be hypocalcified. The experimental implant, coded RBT411, is shown in Fig. 13. The maximum stress of 417 psi may be found at the apex of this implant. This is the lowest maximum stress found with any of the six implants. Consequently, there is little chance that pathologic resorption of bone could occur 675
RIEGER,
MAYBERRY,
AND
BROSE
Fig. 12. A, Stress distribution for Driskell implant. B, Close-up of Driskell implant. Fig. 13. Stress distribution for experimental implant.
with this implant. Although there is some evidence that bone hypocalcification could occur in approximately a fifth of the supporting cortical bone, there is also evidence that normal remodeling could occur in the same region.
CONCLUSION Adams9 illustrated that a cylindrical implant design would direct most of an applied axial load to the apical bone. Tapered designs were recommended by Adams for better stress distribution. All of the designs modeled show some punching stress; but the highest of these stresses, found with the Denar implant, does not seem to be of a clinically significant magnitude. Pathologic resorption of bone at the crest was indicated for the Denar, Miter, and Driskell implants. Bone in the crestal region may be critical to the long-term success of dental implants. Its loss, increasing the crown-to-root ratio, could endanger the implant’s stability and provide a safe haven for microorganisms. Of the implants studied, the Miter and experimental RBT411 implants had the best bone stress distributions. Additional research, combining three-dimensional finite element analyses and clinical measurement of bone activity around implants, is recommended for all dental implants presently available.
REFERENCES 1. Justus R, Luft JH. A mechanicochemical hypothesis for bone rsmodeling induced by mechanical stress. Calcif Tissue Res 1970;5:222-35. 2. Fukada E, Yasuda J. On the peizoelectric effect of bone. J Physiol Sot Jap 1957;12:1158-65. 3. Shamos MH, Lavine LS, Sbamos MI. Piezoelectric effect in bone. Nature 1963;197:81-5. 4. Bassett CAL, Becker RO. Generation of electric potentials by bone in response to mechanical stress. Science 1962;137:1663-‘70. 5. Chamay A, Tschants P. Mechanical influences in bone remodeling. Experimental research on Wolff’s law. J Biomech 1972;5:1’73-80. 6. Fung YC. Biomechanics: mechanical properties of living tissues. New York: Springer-Verlag, 1981. 7. Hassler CR, Rybicki EF, Cummings, KD, Clark LC. Quantitation of compressive stresses and its effects on bone remodeling. Bull Hosp Bone Joint Res 1977;38:90-3. 8. Schnitman PA, Shulman LB. Recommendations of the consensus development conference on dental implants. J Am Dent Assoc 1979; 98:373-7. 9. Adams WK. Employing finite element modelling in the design optimization of prosthodontic implants [Master of Science Thesis]. Columbus, Ohio: The Ohio State University, 1985. Reprint
requests
to:
DR. M. R. RIEGER DENTAL BRANCH UNIVFJWTY OF TEXAS HEALTH HOUSTON. TX 77030
SCIENCE CENTER
JUNE
1990
VOLUME
63
NUMBER
6
element R. Rieger,
University
M.S.,
analysis Ph.D.,’
M.
Mayberry,
of six endosseous M.S.,”
and
M.
0.
implants
Brose,
D.D.S.“’
of Texas Health Science Center, Dental Branch, Houston, Tex.
Stress magnitudes and contours in bone surrounding six endosteal post-type dental implants were calculated by using the finite element method. Comparisons were made by using Branemark, Core-Vent, Denar, Miter, Stryker, and experimental implant designs. Although certain assumptions were made that could be considered controversial, this study concluded that apical “punching stresses” with all of the implants were probably not clinically significant. Saucerization resulting from biomechanical overloads could be a possibility for three of the implants. Problems related to combinations of overloads and underloads at the same time were suggested for several more popular implants in the United States. Additional research, combining 3-D finite element models and clinical studies, was recommended for all commercially available dental implants. (J PROSTHET DENT 1990;63:671-6.)
A
lthough many clinicians have not accepted implant technology, the use of implants is becoming more popular. Presently 25 different post-type endosseous implant systems are available. The variety of implant systems has created confusion because most clinical studies on past implant successesand failures have not been well documented. This study compared the stress patterns in cortical bone surrounding six post-type endosseous implants by using finite element analysis.
LITERATURE
REVIEW
Stress or strain-induced bone remodeling is a concept proposed by Julius Wolff in 1870. Although the magnitudes and directions of stressesthat stimulate bone apposition or resorption are still topics of controversy among researchers, most agree that stresses play an important role as a feedback mechanism in bone remodeling. Several theories have been proposed concerning the reasons for stress-induced bone growth. Justus and Luftl found that altered strain in hydroxyapatite crystals caused the generation of additional calcium ions. Bone cells may react to this change by either bone apposition or resorption. The piezoelectric properites of bone have also been used to explain the stress dependency of bone remodeling.2-4 Bone produces electrical current and potentials proportional to the magnitude of applied stress with a polarity determined by the stress direct,ion. Apposition is thought to occur in regions of negatively charged bone. The stress-related
Iii
*Associate Professor and Director of Oral Biomaterials Research. “Biomedical Engineer, Biomedical Engineering Center, Ohio State University, Columbus, Ohio. *“Assistant Professor, Restorative and Prosthetic Dentistry, Ohio State University, College of Dentistry, Columbus, Ohio. 10/l/13679
THE
JOURNAL
I
Fig.
OF PROSTHETIC
DENTISTRY
5nite
1
i I I t
I
element
I
model
I
I
of Branemar
1
*k im!plant.
671
RIEGER,
Fig.
AND
BROSE
2. Finite element model of Core-Vent implant. Fig.
feedback mechanism is likely a combination of the biochemical and electrical activity. The actual magnitudes of stress that cause bone remodeling are of particular importance to this study. It has been suggested that there is an optimal value of stress for which as much bone is removed by resorption as is deposited by apposition. With stresses above this value, hypertrophy takes place; below this value, atrophy occurs. There is also a maximum stress limit above which stresses will destroy bone by pathologic resorption. Chamay and Tschantz5 concluded that compressive stressespromoted bone growth whereas tensile stresses caused resorption. In a review of functional adaptation of bone however, Fung6 reported that effects of tensile and compressive stresses were the same.
672
MAYBERRY,
3. Finite element model of Denar implant.
As important as these topics are to dental implantology, only one study has ever tried to measure the actual magnitudes of stresses promoting bone growth.7 By inserting load cells into the calvaria of rabbits, a maximum bone formation rate was noted at a compressive stress of 250 psi. The bone formation rate dropped to the control level at stresses greater than 400 psi. Most clinical research of post-type endosseous implants indicates that bone adaptation or bonding to implants is critical to the successof implants. Implant successwas defined by Schnitman and Shulmar? as mobility of less than 1 mm in any direction, bone loss no greater than one third of the vertical height of the implant, and functional service for 5 years in 75 % of all patients. These guidelines may not
JUNE
1990
VOLUME
03
NUMBER
6
FINITE
ELEMENT
ANALYSIS
OF IMPLANTS
Fig. Fig.
4. Finite element model of Miter implant.
be strict enough because implant mobility to any extent generally indicates impending failure. Extensive stress analyses of endosseous imp1ant.s is vital to the continued growth and successof dental implantology.
MATERIAL
AND
METHODS
Six post-type endosseous implants were selected for this study: Branemark (Nobelpharma USA, Chicago, Ill.), CoreVent (Core-Vent Corp., Encino, Calif.), Denar (Steri-Oss, Anaheim, Calif.), Miter (Miter, Inc., Warsaw, Ind.),Driskell (Driskell Bioengineering, Galena, Ohio), and an experimental implant. All implants were assumed to have an elastic modulus of 1.59 x lo7 psi. The finite element mod-
THE
JOURNAL
5. Finite element model of Driskell implant.
OF PROSTHETIC
DENTISTRY
els of these implants are shown in Figs. 1 through 6. The axisymmetric models were made of parabolic, isoparametric elements. Most of these elements were quadrilateral, but some were triangular to accommodate specific geometries. Dimensions were taken from actual implants or drawings supplied by the manufacturers. The implants were assumed to be surrounded and bonded to cortical bone with an elastic modulus of 1.98 X lo6 psi. An axial load of 25 pounds was applied over the top surface of each implant. The Branemark, Core-Vent, and Denar implants each have a threaded portion. The models of these implants assumed a symmetry about the long axis. Although this type of symmetry does not exist for threads, the stresses would
673
RIEGER,
MAYBERRY,
AND
BROSE
Table I. Maximum bone stresses Implant
Stress
Branemark Core-Vent Denar Miter Driskell Experimental
Fig.
6. Finite element model of experimental implant.
not be significantly affected by this assumption. The CoreVent implant was difficult to model axisymmetrically because of the vents in the inverted basket portion. These holes were therefore not included in the model. The Miter, Driskell, and experimental implants have serrations. These serrations were well represented axisymmetrically. The bone core was 5 mm in radius for all implants. The finite element models were generated and solved by using a commercial computer-aided engineering program (Structural Dynamics Research Corporation, Cincinnati, Ohio). The software was mounted on a VAX 8500 computer (Digital Equipment Corporation, Maynard, Mass.). RESULTS
AND
DISCUSSION
The magnitudes of the stresses are defined in Fig. 7. Stress contour plots generated for each implant are shown 674
(X
lo6 psi)
478 661 747
794 868 417
in Figs. 8 through 13. The maximum stress found with each implant is shown in Table I. For purposes of discussion, certain assumptions will be made, based on the literature.7 First, optimal bone maintenance occurs at 250 psi. This may be found in the green area of the contour plots. Pathologic resorption of bone occurs at greater than 700 psi. This may be found in the light purple or dark purple areas. Bone atrophy occurs at less than 200 psi. This may be found in the light blue or dark blue areas. These assumptions may not be valid when poor stress distribution exists. The Branemark implant is shown in Fig. 8. Except for the maximum stress of 478 psi found near the neck, this implant transfers relatively low stressesto the bone. Thus, pathologic resorption of bone would be unlikely with this implant. What may be of most concern is the poor stress distribution and the possibility for atrophy with this implant. Stressesare concentrated at the neck and the apex with little stress transfer along the middle. In a freestanding mode with 25 pounds of force applied as modeled, virtually a fourth of the surrounding cortical bone could be hypocalcified. If placed in “group-function,” as recommended by the manufacturer, where less force is applied occlusally, almost half of the surrounding cortical bone could decalcify to a level below normal. Further, the amount of normally calcified cortical bone at the neck and apex could be reduced. The Core-Vent implant is shown in Fig. 9. Because of the elimination of the vents, this model is not accurate. However, certain features can still be discussed. A maximum stress of 661 psi may be found at the neck of this implant. The chances for pathologic resorption of bone are unlikely. As modeled, this implant is similar to the Branemark implant in that stress distribution is poor. Nearly half of the cortical bone could be hypocalcified. It may be assumed that vents would improve the situation. In actual practice, it has been noted that cortical bone does not always fill the inverted basket of this implant. Such a circumstance could seriously affect the successof this implant because a lack of “internal” support would lead to markedly increased punching stresses at the apex. The Denar implant is shown in Fig. 10. A maximum stress of 747 psi can be seen at the neck of this implant. Pathologic resorption of bone could occur in this region. Like the previous two implants, nearly half of the cortical bone along the length of this implant could be hypocalciJUNE
1990
VOLUME
63
NUMBER
6
FINITE
ELEMENT
ANALYSIS
OF IMPLANTS
Fig. 7. Key to stress magnitudes.
Fig. Fig. Fig. Fig.
8. Stress distribution for Branemark implant. 9. Stress distribution for Core-Vent implant. 10. Stress distribution for Denar implant. 11. Stress distribution for Miter implant.
fied. Punching stress at the apex could also be a problem. The Miter implant is shown in Fig. 11. A maximum stress of 794 psi can be seen at the neck. Pathologic resorption of bone could occur to the first serration; and indeed, there is clinical evidence of saucerization in the absence of inflammation with this implant. Unlike the previous implants, the stress distribution for this implant is good. Although such is not a part of this study, this implant could function well as a free-standing implant. The Driskell implant is shown in Fig. 12. A maximum THE
JOURNAL
OF PROSTHETIC
DENTISTRY
stress of 868 psi may be seen in Fig. 12, A at the neck of this implant. This is the greatest stress found with any of the six implants. Pathologic resorption of bone could occur at the neck of this implant. Approximately a third of the surrounding cortical bone could be hypocalcified. The experimental implant, coded RBT411, is shown in Fig. 13. The maximum stress of 417 psi may be found at the apex of this implant. This is the lowest maximum stress found with any of the six implants. Consequently, there is little chance that pathologic resorption of bone could occur 675
RIEGER,
MAYBERRY,
AND
BROSE
Fig. 12. A, Stress distribution for Driskell implant. B, Close-up of Driskell implant. Fig. 13. Stress distribution for experimental implant.
with this implant. Although there is some evidence that bone hypocalcification could occur in approximately a fifth of the supporting cortical bone, there is also evidence that normal remodeling could occur in the same region.
CONCLUSION Adams9 illustrated that a cylindrical implant design would direct most of an applied axial load to the apical bone. Tapered designs were recommended by Adams for better stress distribution. All of the designs modeled show some punching stress; but the highest of these stresses, found with the Denar implant, does not seem to be of a clinically significant magnitude. Pathologic resorption of bone at the crest was indicated for the Denar, Miter, and Driskell implants. Bone in the crestal region may be critical to the long-term success of dental implants. Its loss, increasing the crown-to-root ratio, could endanger the implant’s stability and provide a safe haven for microorganisms. Of the implants studied, the Miter and experimental RBT411 implants had the best bone stress distributions. Additional research, combining three-dimensional finite element analyses and clinical measurement of bone activity around implants, is recommended for all dental implants presently available.
REFERENCES 1. Justus R, Luft JH. A mechanicochemical hypothesis for bone rsmodeling induced by mechanical stress. Calcif Tissue Res 1970;5:222-35. 2. Fukada E, Yasuda J. On the peizoelectric effect of bone. J Physiol Sot Jap 1957;12:1158-65. 3. Shamos MH, Lavine LS, Sbamos MI. Piezoelectric effect in bone. Nature 1963;197:81-5. 4. Bassett CAL, Becker RO. Generation of electric potentials by bone in response to mechanical stress. Science 1962;137:1663-‘70. 5. Chamay A, Tschants P. Mechanical influences in bone remodeling. Experimental research on Wolff’s law. J Biomech 1972;5:1’73-80. 6. Fung YC. Biomechanics: mechanical properties of living tissues. New York: Springer-Verlag, 1981. 7. Hassler CR, Rybicki EF, Cummings, KD, Clark LC. Quantitation of compressive stresses and its effects on bone remodeling. Bull Hosp Bone Joint Res 1977;38:90-3. 8. Schnitman PA, Shulman LB. Recommendations of the consensus development conference on dental implants. J Am Dent Assoc 1979; 98:373-7. 9. Adams WK. Employing finite element modelling in the design optimization of prosthodontic implants [Master of Science Thesis]. Columbus, Ohio: The Ohio State University, 1985. Reprint
requests
to:
DR. M. R. RIEGER DENTAL BRANCH UNIVFJWTY OF TEXAS HEALTH HOUSTON. TX 77030
SCIENCE CENTER
JUNE
1990
VOLUME
63
NUMBER
6
