|Year : 2014 | Volume
| Issue : 1 | Page : 3-10
Influence of occlusal loading on stress patterns at the bone-implant interface by angulated abutments in the anterior maxilla: A three-dimensional finite-element study
Pradeep Bholla1, Liju Jacob Jo2, Kalepu Vamsi3, Padma Ariga4
1 Department of Prosthodontics, Penang International Dental College, Malaysia
2 Department of Prosthodontics, KMCT Dental College, Calicut, Kerala, India
3 Department of Pedodontics, Penang International Dental College, Malaysia
4 Department of Prosthodontics, Saveetha Dental College and Hospital, Chennai, India
|Date of Web Publication||19-Apr-2014|
Liju Jacob Jo
Department of Prosthodontics, KMCT Dental College, Manassery PO, Mukkam, Calicut, Kerala - 673 603
Source of Support: None, Conflict of Interest: None
| Abstract|| |
Background: The long-term success of implant supported prosthesis depends on many variables. In addition to the osseointegration rules, a clinician should consider also the biomechanical factors like angulation of the abutment that may have a profound influence on the stress levels on bone for long-term function of implant-supported prosthesis.
Purpose : A 3D finite element method was used to evaluate the von Mises stresses (ΣEmax ) generated in varying thickness of cortical bone under axial and combined loading conditions by four angulated abutments in the anterior esthetic zone.
Materials and Methods: The model resembles the maxillary bone, and the material properties similar to the bone are introduced in the model and clinical loading conditions were simulated. Von mises stresses occurring for four angulated abutments (0, 15, 20, 25 degree) in (a) compact and cancellous bone, (b) thick and thin compact bone, (c) subjected to axial and combined loading, were calculated.
Results: ( a) Von Mises stresses (∑E max ) were higher in the cortical bone compared to the cancellous bone and were concentrated in the crestal (facial) region in both types of bone. (b) The ∑E max values of 0, 15 degree abutments in thin bone and 0, 15, 20 degree abutments in thick bone were within the physiological remodeling zone. (c) Stress values for a 25 degree abutment in both types of bone were above the resorption limit. (d) Stress values were higher due to combined loading compared to axial loading irrespective of the angulation or quality of bone present. (e) Clinically, within a load of 178 N angulated abutments up to 20 degrees can be placed in the anterior maxillary zone.
|How to cite this article:|
Bholla P, Jo LJ, Vamsi K, Ariga P. Influence of occlusal loading on stress patterns at the bone-implant interface by angulated abutments in the anterior maxilla: A three-dimensional finite-element study. J Dent Implant 2014;4:3-10
|How to cite this URL:|
Bholla P, Jo LJ, Vamsi K, Ariga P. Influence of occlusal loading on stress patterns at the bone-implant interface by angulated abutments in the anterior maxilla: A three-dimensional finite-element study. J Dent Implant [serial online] 2014 [cited 2022 Dec 6];4:3-10. Available from: https://www.jdionline.org/text.asp?2014/4/1/3/130944
| Introduction|| |
A key factor for the success or failure of implant is the manner in which stresses are transferred to the surrounding bone. Load transfer at the bone-implant interface depends mainly on the type of loading and the nature of the bone-implant interface.  Occlusion can be critical for implant longevity; if the occlusal force exceeds the capacity of the interface to absorb stress, the implant will fail. Axial loading is favored during the loading of implant. However on several occasions, with clinical loading restored with angled abutments, lateral occlusal forces may increase. The angulation of these implants may have a profound influence on the stress levels on bone for long-term function of implant-supported prosthesis. Angled abutments are often used to restore dental implants placed in the anterior maxilla due to esthetic or spatial needs.  Abutment angulation is one of the many biomechanical variables involved in implant dentistry that need a scientific evaluation. 
Stresses and strains generated through various angled abutments have been evaluated by methods like photo-elasticity, strain gauge analysis and finite element analysis. The finite element method is capable of providing detailed quantitative data at any location within a mathematical model.  Two- and three-dimensional finite element analysis have been used to evaluate the stresses around various dental implant systems using a model of the mandible.  Currently, a three-dimensional finite element model is considered necessary for analysis of implant design problems inherent to the maxilla. The purpose of the present study is to develop and use a 3D finite element method of the maxilla to evaluate the von Mises stresses (∑E max ) generated in varying thickness of cortical bone under axial and combined loading conditions by four angulated abutments in the anterior esthetic zone.
| Materials and Methods|| |
In the present study the model resembles the maxillary bone and the material properties similar to the bone are introduced in the model. The assumptions are made that the materials are homogenous and linear and that they have elastic material behavior characterized by two material constants of Young's modulus and Poisson's ratio.
Development of the mathematical model for the maxilla involved the use of a computer tomography to derive geometry and density values. A dried human maxilla was scanned along sagittal planes at 1.5 mm increments using a GE 9800 quick scanner. The scans were digitized and entered into a computer software program. Cross sections were reassembled to yield a three dimensional model of the maxilla half. Reflection of the model through the median plane produced the whole maxilla. A section of this model was taken from the premaxilla to focus the stress analysis on the region of bone that would surround the implant.
Two mathematical models were developed to simulate the D 2 and D 3 density bone using the digitized data computed from the computer tomography numbers of the scan; the isotropic cortical bone was 3.0 mm for D 2 and 1.5 mm for D 3 on the facial, lingual and occlusal aspects of the bone.  The cancellous bone had a density approximately 17% that of cortical bone around the implant in both the models. 
Bone implant interface
The FEA model assumed a state of optimal osseointegration, which means that the cortical and trabecular bone are assumed to be perfectly bonded to the implant.
Modeling of fixture
The fixture model for the study was a generic derivative of the 4.2 × 13 mm implant. ,, The implant is a self-threaded, single piece, cylindrical fixture, , made of commercially pure titanium. Fixture elements were assigned the isotropic elastic properties of pure titanium. Cementable abutments  of four angulations (0°, 15°, 20°, 25°) were evaluated in the present investigation. The four abutments had a base diameter equal to the implant diameter with a 7 mm height.
To simulate the actual clinical situation, a load of 178 N, which is the average biting forces for incisors,  was applied along the long axis of each abutment and a combined load was applied 45° to the long axis of implant for all the angulated abutments.
| Methodology|| |
FEA is a technique for obtaining a solution to a complex mechanical problem by dividing the problem domain into a collection of much smaller and simpler domains (elements) in which the field variables can be interpolated with the use of shape functions. An overall approximated solution to the original problem is determined based on variational principles. In other words, FEA is a method whereby, instead of seeking a solution function for the entire domain, one formulates the solution functions for each finite element and combines them properly to obtain the solution to the whole body. Because the components in a dental implant-bone system are extremely complex geometrically, FEA has been viewed as the most suitable tool for analyzing them. A mesh is needed in FEA to divide the whole domain into elements. The process of creating the mesh, elements, their respective nodes, and defining boundary conditions is referred to as "discretization" of the problem domain.
Finite element analysis uses a complex system of points called nodes which make a grid called mesh. This mesh is programmed to contain the material and structural properties which define how the structure will react to certain loading conditions. Nodes are assigned a certain density throughout the material depending on the anticipated stress levels of particular area. Regions which will receive larger amount of stress usually have a higher density node than those which experience little or no stress; all the nodes are converted into a mathematical load and the stress was evaluated.
ANSYS WORK BENCH 10.0 software was used for solid modeling, geometric construction, finite element mesh creation and post processing.
The material properties of the cortical bone, cancellous bone and the implant fixture were 
- Cortical bone (Young's modulus - 13700 Mpa; Poisson's ratio - 0.3)
- Cancellous bone (Young's modulus - 7930 Mpa; Poisson's ratio - 0.3)
- Implant (Young's modulus - 1.025 Mpa; Poisson's ratio - 0.35).
| Results|| |
The results of numerical analysis are shown for von mises stresses occurring for four angulated abutments in (a) compact and cancellous bone, (b) thick and thin compact bone, and (c) subjected to axial and combined loading, Sixteen different colors were selected for indication of specific values of von mises stresses [Figure 1],[Figure 2],[Figure 3] and [Figure 4].
Von mises stress distribution at bone implant interface for the thin cortical and cancellous bone at various angulations under axial loading of 178 N for a 0 degree abutment is 3.5845 Mpa. When the angulation of the abutment is 15 degree, the stresses are 4.3228 Mpa which is an increase of 20%; the stresses are further increased to 5.3509 Mpa, which is about 50% from the baseline values; the stresses are further increased to 5.9409 Mpa which is about 65% from the baseline values.
Von mises stress distribution at bone implant interface for the thick cortical and cancellous bone at various angulations under axial loading of 178 N for a 0 degree abutment was 2.986 Mpa. When the inclination of the abutment was 15 degrees, the stresses were 3.8264 Mpa, which is an increase of about 28%; the stresses are further increased to 4.9759 Mpa for an abutment inclination of 20 degrees, which is about 66% from the baseline values. When the abutment angulation was increased to 25 degrees the stresses were 5.6259 Mpa, which was about 88% more than the baseline values.
When von mises stresses for thin and thick bone under similar axial loading and angulations were compared, the stresses were 3.5845 Mpa and 2.986 Mpa (0 degree abutment), 4.3228 Mpa and 3.8264 Mpa (15 degree abutment), 5.3509 Mpa and 4.9759 Mpa (20 degree abutment), 5.9409 Mpa and 5.6259 Mpa (25 degree abutment) respectively. When the thickness of the cortical bone is increased there is a decrease of 17%, 11%, 7%, 3% of stresses for 0, 15, 20 25 degree abutments [Table 1] and [Graph 1] [Additional file 1].
|Table 1: Axial loading – comparison of von mises stresses (thin bone and thick bone)|
Click here to view
Von mises stress distribution at bone implant interface for the thin cortical and cancellous bone at various angulations under combined loading of 178 N for a 0 degree abutment is 23.408 Mpa. When the angulation of the abutment is 15 degree the stresses are 24.474 Mpa, which is an increase of 5%; the stresses are further increased to 26.395 Mpa, which is about 13% from the baseline values; the stresses are further increased to 30.144 Mpa, which is about 28% from the baseline values.
Von mises stress distribution at bone implant interface for the thick cortical and cancellous bone at various angulations under combined loading of 178 N for a 0 degree abutment was 21.806 Mpa. When the inclination of the abutment was 15 degrees, the stresses were 23.377 Mpa, which is an increase of about 7%; the stresses are further increased to 26.368 Mpa for an abutment inclination of 20 degrees, which is about 21% from the baseline values. When the abutment angulation was increased to 25 degrees the stresses were 27.943 Mpa, which was about 28% more than the baseline values.
When von mises stresses for thin and thick bone under similar combined loading and angulations were compared, the stresses were 23.408 Mpa and 21.806 Mpa (0 degree abutment), 24.474 Mpa and 23.377 Mpa (15 degree abutment), 26.395 Mpa and 26.368 Mpa (20 degree abutment), 30.144 Mpa and 27.943 Mpa (25 degree abutment), respectively. When the thickness of the cortical bone is increased, there is a decrease of 7%, 5%, 0.1%, 7% of stresses for 0, 15, 20 25 degree abutments [Table 2] and [Graph 2] [Additional file 2].
When the stresses for thin bone with abutment angulations 0, 15, 20, 25 degrees subjected to axial and combined loading were compared, the values of stresses for combined loading (23.408 to 30.144) far exceeded that of axial loading (3.5845 to 5.9409).
When the stresses for thick bone with abutment angulations 0, 15, 20, 25 degrees subjected to axial and combined loading were compared, the values of stresses for combined loading (21.806 to 27.943) far exceeded that of axial loading (2.986 to 5.6259).
From the results it is evident that the load transferred to the compact bone is always higher than the spongious bone irrespective of the abutment angulation or type of loading. However, the stresses were less in the thick compact bone compared to the thin compact bone. As the inclination of the abutment was increased there was a corresponding increase in the stress patterns under both loading conditions. There was a 5- to 7-fold increasing stresses due to combined loading than axial loading, because of the load transfer mechanisms and the increase in the inclination angle of applied load. The von mises stresses are mainly concentrated around the neck of the implant in the cortical bone with a gradual decrease of stresses towards the apical region.
| Discussion|| |
An angled abutment allows the placement of implants in the most favorable quantity and quality of available bone in patients with compromised osseous anatomy, especially in the maxilla in the esthetic zone. Good quality bone is dependent upon appropriate level of bone remodeling necessary to maintain the bone density and the avoidance of the bone microfracture and failure. These processes are in turn governed by the stress distribution in the bone which might depend on the location, angle of abutment, type of loading and quality of the bone.
Axial loading/combine loading - thin/thick bone
When the stress in the thin (1.5 mm) and the thick (3.0 mm) compact bone due to four different angulated abutments subject to axial loading (178N) were evaluated, it was observed that von mises stress (∑E max ) values for the thin cortical shell with 0, 15, 20, 25 degree abutments under axial loading were 3.5845-5.9409 Mpa, and during combined loading they were 23.4-30.144 Mpa. The stress values for cancellous bone under similar angulations and loading conditions were 1.6-4.7 Mpa (axial loading) and 9.8-17.99 Mpa (combined loading). Von mises stress (∑E max ) values for the thick cortical shell with 0, 15, 20, 25 degree abutments under axial loading were 2.9-5.6 Mpa and in combined loading they ranged from 21 to 27 Mpa. The stress values for cancellous bone under similar circumstances range from 1.3 to 3.4 Mpa for axial loading and 9.4 to 13.1 Mpa for combined loading.
In both the situations with thin and thick compact bone, the ∑E max stress values were higher for the cortical bone than the cancellous bone and most of these stresses were concentrated in the crestal region, , where there was variation in the thickness of cortical bone irrespective of the angulation. However, stresses were more at the crestal region with combined loading (24 Mpa) compared to axial loading (4 Mpa), which is in accordance with a study by Borchers and Reichart.  This is because, the implant fixture is completely supported by a rigid layer of cortical bone in the crestal region, which extends approximately to one-fourth of its implant length in the thick bone and probably even more less in the thinner bone. 
For a 0 degree angle in thin bone, peak stresses occurred in crestal and lingual shell of the cortical bone with a magnitude of 3.584 Mpa but as angulation increased to 15, 20 and 25 degrees, the concentration of the compressive stresses shifted to the cortical layer of bone on the facial side of the fixture (4.3-5.9Mpa) under axial loading. However under combined loading the stresses on all the abutments were concentrated on the facial aspect of the bone (23-30 Mpa). As the density of the bone increased, the stresses were concentrated on the facial aspect for all the abutments under all loading conditions (2.9-5.6 Mpa under axial loading and 21.8-27.9 Mpa under combined loading). These values are in accordance with the study done by Clelland and Martin D Grass et al. ,, Hence, clinically preservation of buccal supporting bone is desirable to obtain physiologic bone modeling response and to enhance facial plate. Insufficient bone volume may result in buccal fenestration or dehiscence, which can precipitate mucosal irritation, decreased support and potential implant failure.
The magnitude of these stresses reduced as the thickness of the cortical shell was doubled and progressively towards the apical region which is in accordance with the studies done by Clelland and Senay et al. , It is predicted that placement of implants in bone with greater thickness of cortical shell and greater density of core will result less micromovement and reduced stress concentrations there by increasing the likelihood of fixture stabilization and tissue integration. 
As the inclination of the abutment increased there was a corresponding increase in the stress values under axial and combined loading conditions. Hasser et al.  established a quantitative relationship between compressive stresses and bone remodeling and found that the physiological zone for bone remodeling is in the range of 0 to 5 Mpa. The ∑E max stresses under axial loading for thin and thick cortical bone for 0 and 15 degree abutments were within the physiologic zone (5 Mpa) indicating bone remodeling and lamellar bone formation. ∑E max values are within the physiological zone for a 20 degree abutment with a thick bone (4.9 Mpa) but marginally higher for the thin bone (5.3 Mpa), which approach this potential resorption limit. The stress values for a 25 degree angulated abutment (both thick and thin bone) were above the described limit (5.6 and 5.9 Mpa), which might initiate bone resorption which is in accordance with a study done by Clelland et al.  However Celletti and coworkers  found no adverse effects on surrounding bone with straight or pre-angled abutments in monkeys. In yet another study by Ashok Sethi et al.,  it was observed that the survival rate for all the angulated abutments (0-45 degrees) were the same clinically when certain protocols outlined were followed. A study by Dorothy Eger  revealed no significant differences between angled and standard abutments for any of the clinical variables and there were no clinically evident physiologic disadvantages in the use of angled abutments, although conclusions about attachment levels need to be evaluated.
Axial loading induces a more uniform, histologically quiescent remodeling response that gradually decreases from the coronal aspect to the apex of the implant. Non-axial loading elicits a more dynamic remodeling of the surrounding cortical and especially trabecular bone tissue.  It is predicted that reduction in the elastic modulus of the bone around the neck of the implant by a factor of 16 produces only two-fold reduction in the peak stress. This results in stress levels capable of inducing fatigue failure in a much weaker bone. Hence it is extremely important to have good quality dense bone around the neck of the implant to withstand the predicted peak stresses. Failure to achieve this after implantation and subsequent healing may result in local fatigue failure and resorption at the neck upon resumption of physiological loading. 
Angulated abutments in the esthetic zone may subject to axial and combined loading clinical conditions. In these situations, the amount of lateral forces may increase,  subjecting the implant to withstand more stresses. In our study the ∑E max values for combined loading for all angulated abutments in both types of bone were 5-7 fold higher than that of axial loading and were concentrated around the neck of the implant which is in accordance with the studies done by Clift et al.  and Holmes et al.  Hence clinically, it may be advisable to avoid eccentric contacts in anterior abutments to avoid lateral loading components, which may contribute to increase in stresses. Within a load of 178 N, angulated abutments up to 20 degrees can be placed in the anterior maxilla zone but further clinical scientific evaluation needs to be done.
Future studies can be done with in vivo behavior of oral implants and this necessitates the creation of patient-dependent finite element models.
| Summary|| |
The need for angulated abutments has been recognized as a result of the difference in angle between the bone available for implant placement and the long axis of the planned restoration. An angled abutment is frequently used to achieve the desired position to improve esthetics and function, however there are concerns regarding possible adverse effects of axial and non-axial forces on the survival of implants due to angled abutments.
The results showed that the Von mises stresses (ΣEmax ) were located in the crestal (facial) region of the cortical bone and the magnitude of stresses decreased with the increase in the thickness of the cortical bone and towards the apical region. The ΣEmax stress values for the 0 and 15 degree abutments in thin bone and 0, 15 and 20 degree abutments in thick bone were within the physiological zone capable of bone remodeling and lamellar bone formation. But the stress values for a 25 degree abutment in both types of bone were above this zone.
Stress values continued to be higher when the loading was combined compared to axial forces irrespective of the angulation of the abutment or the type of bone present. Hence, axial loading is always preferable for all the abutment angulations as the stress values are lower but in reality there is a transverse component in addition to the vertical component and stresses were 5-7 fold more in the combined loading conditions and this signifies that the eccentric contacts should definitely be eliminated in the anterior esthetic zone with angulated abutments.
| Conclusion|| |
Within the limitations of the study, the following conclusions can be drawn.
- Von Mises stresses (∑E max ) were higher in the cortical bone compared to the cancellous bone
- These were concentrated in the crestal (facial) region in both types of bone
- The magnitude of stresses decreases with a increase in thickness of the cortical bone and towards the apical region
- The ∑E max values of 0, 15 degree abutments in thin bone and 0, 15, 20 degree abutments in thick bone were within the physiological remodeling zone. Stress values for a 25 degree abutment in both types of bone were above the resorption limit
- Stress values were higher due to combined loading compared to axial loading irrespective of the angulation or quality of bone present
- Clinically, within a load of 178 N angulated abutments, up to 20 degrees can be placed in the anterior maxillary zone. However it is advisable to avoid eccentric contacts to eliminate lateral forces which enhance the stress concentration several fold at the crestal region.
| References|| |
|1.||Chun HJ, Shin HS, Han CH, Lee SH. Influence of implant abutment type on stress distribution in bone under various loading conditions using finite element analysis. Int J Oral Maxillofac Implants 2006;21:195-202. |
|2.||Saab XE, Griggs JA, Powers JM, Engelmeier RL. Effect of abutment angulation on the strain on the bone around an implant in the anterior maxilla. A finite element study. J Prosthet Dent 2007;97:85-92. |
|3.||Clelland NL, Lee JK, Bimbenet OC, Brantley WA. A three dimensional finite element stress analysis of angled abutments for an implant placed in the anterior maxilla. J Prosthodont 1995;4:95-100. |
|4.||Ismail YH, Pahountis LN, Fleming JF. Comparison of two dimensional and three dimensional finite element analysis of a blade implant. Int J Oral Implantol 1987;4:25-31. |
|5.||Clelland NL, Lee JK, Bimbenet OC, Gilat A. Use of an axisymmetric finite element method to compare maxillary bone variables for a loaded implant. J Prosthodont 1993;2:183-9. |
|6.||Lum LB, Osier JF. Load transfer from endosteal implants to supporting bone: An analysis using statics. Part two: Axial loading. J Oral Implantol 1992;18:349-53. |
|7.||Himmlová L, Dostálová T, Kácovský A, Konvicková S. Influence of implant length and diameter on stress distribution: A finite element analysis. J prosthet Dent 2004;91:20-5. |
|8.||Baggi L, Cappelloni I, Di Girolamo M, Maceri F, Vairo G. The influence of implant diameter and length on stress distribution of osseointegrated implants related to crestal bone geometry: A three dimensional finite element analysis. J Prosthet Dent 2008;100:422-31. |
|9.||Mailath G, Stoiber B, Watzek G, Matejka M. Bone resorption at the entry of osseointegrated implants a biomechanical phenomenon. Finite element study. Z Stomatol 1989;86:207-16. |
|10.||Siegele D, Soltesz U. Numerical investigations of the influence of implant shape distribution in the jaw bone. Int J Oral Maxillofac Implants 1989;4:333-40. |
|11.||Papavasiliou G, Tripodakis AP, Kamposiora P, Strub JR, Bayne SC. Finite element analysis of ceramic abutment-restoration combinations for osseointegrated implants. Int J Prosthodont 1996;9:254-60. |
|12.||Helkimo E, Carlsson GE, Hclkimo M. Bite force and state of dentition. Acta Odontol Scand 1977;35:297-303. |
|13.||Sevimay M, Turhan F, Kiliçarslan MA, Eskitascioglu G. Three dimensional finite element analysis of different bone quality on stress distribution in an implant supported crown. J Prosthet Dent 2005;93:227-34. |
|14.||Clelland NL, Ismail YH, Zaki HS. Three-dimensional finite element stress analysis in and around the Screw-Vent Implant. Int J Oral Maxillofac Implants 1991;6:391-8. |
|15.||Clift SE, Fisher J, Watson CJ. Finite element stress and strain analysis of the bone surrounding a dental implant: Effect of variations in bone modulus. Proc Inst Mech Eng H 1992;206:233-41. |
|16.||Borchers L, Reichart P. Three-dimensional stress distribution around a dental implant at different stages of interface development. J Dent Res 1983;62:155-9. |
|17.||Gross MD, Nissan J, Samuel R. Stress distribution around maxillary implants in anatomic photoelastic models of various geometry. Part 1. J Prosthet Dent 2001;85:442-9. |
|18.||Gross MD, Nissan J. Stress distribution around maxillary implants in anatomic photoelastic models of various geometry. Part 2. J Prosthet Dent 2001;85:450-4. |
|19.||Canay S, Hersek N, Akpinar I, Asik Z. Comparison of stress distribution around vertical and angled implants with finite-element analysis. Quintessence Int 1996;27:591-8. |
|20.||Holmes DC, Loftus JT. Influence of bone quality on stress distribution for endosseous implants. J Oral Implantol 1997;23:104-11. |
|21.||Hassler CR, Rybicki EF, Cummings KD, Clark LC. Quantification of bone stresses during remodeling. J Biomech 1980;13:185-90. |
|22.||Celletti R, Pameijer CH, Bracchetti G, Donath K, Persichetti G, Visani I. Histologic evaluation of osseointegrated implants restored in nonaxial functional occlusion with Preangled abutments. Int J Periodontics Restorative Dent 1995;15:563-73. |
|23.||Sethi A, Kaus T, Sochor P. The use of angulated abutments in implant dentistry: Five-year clinical results of an ongoing prospective study. Int J Oral Maxillofac Implants 2000;15:801-10. |
|24.||Eger D, Gunsolley JC, Feldman S. Comparison of angled and standard abutments and their effect on clinical outcomes: A preliminary report. Int J Oral Maxillofac Implants 2000;15:819-23. |
|25.||Barbier L, Vander Sloten J, Krzesinski G, Schepers E, Van der Perre G. Finite element analysis of non-axial versus axial loading of oral implants in the mandible of the dog. J Oral Rehabil 1998;25:847-58. |
[Figure 1], [Figure 2], [Figure 3], [Figure 4]
[Table 1], [Table 2]