Comparative Evaluation of the Stresses on the Terminal Abutment and Edentulous Ridge in Unilateral Distal Extension Condition when restored with Different Prosthetic Options: An FEA Analysis
Corresponding Author: Bhupender Yadav, Department of Prosthodontics, Crown & Bridge and Implantology, SGT Dental College, Hospital & Research Institute, Shree Guru Gobind Singh Tricentenary University, Gurugram, Haryana, India, Phone: +91 8743019484, e-mail: firstname.lastname@example.org
Received on: 24 April 2023; Accepted on: 20 May 2023; Published on: 28 June 2023
Purpose: The purpose of this finite element analysis (FEA) was to evaluate and compare the stresses on the terminal abutment and edentulous ridge in unilateral distal extension conditions when restored with different prosthetic options.
Materials and methods: A finite element model of a unilateral maxillary edentulous arch distal to the second premolar was fabricated. The second premolar was used as a terminal abutment for three different treatment modalities, namely cast partial denture, semiprecision attachment, and flexible dentures. Three levels of loading (150, 250, and 350 N) were applied from two directions, that is, vertically (90°) and obliquely (60°), on the central fossa of the first and second molars of the prosthesis. The maximum von Mises stress distribution was recorded at two regions, that is, the terminal abutment (cervically) and the edentulous ridge. The data was divided into three groups according to the type of prosthesis. An independent t-test was used to compare the values of the three groups. Analysis of variance (ANOVA) was used to test the differences between the loading directions.
Results: The highest level of stress on the terminal abutment and edentulous ridge was noted with the flexible denture (71.39 and 93.29 MPa), followed by cast partial dentures (8.63 and 3.69 MPa), and the least stress was noted with the semi-precision attachment (3.15 and 1.57 MPa, respectively). The difference in stress levels on both abutment and ridge when the flexible denture was compared with cast partial denture and semi-precision attachment was statistically significant (p < 0.05). However, the stress levels between cast partial and semi-precision attachment were not statistically significant (p > 0.05). The stress levels increased by increasing the forces from 150 to 350 N and changing the direction of forces from vertical to oblique in all three prostheses on both the abutment and edentulous ridge.
Conclusion: Semi-precision attachment and cast partial denture did not depict any significant changes in stress levels on abutment teeth and edentulous ridge. However, flexible dentures show the highest stress levels. Hence, within the confines of the study, semi-precision attachment seems to be the most recommended treatment option, followed by cast partial dentures. Clinicians should limit flexible dentures to only provisional prostheses.
How to cite this article: Girotra M, Yadav B, Malhotra P, et al. Comparative Evaluation of the Stresses on the Terminal Abutment and Edentulous Ridge in Unilateral Distal Extension Condition when restored with Different Prosthetic Options: An FEA Analysis. Int J Prosthodont Restor Dent 2023;13(2):58-64.
Source of support: Nil
Conflict of interest: None
Keywords: Abutment, Cast partial dentures, Precision, Removable partial denture attachment
In partial edentulous conditions, the loss of a few teeth can lead to an adverse esthetic and biomechanical sequel.1 The prosthesis selection for rehabilitating a partially edentulous condition is dependent on factors like clinical need, the position of the edentulous region, the number of missing teeth, the patient’s demand, abutment teeth’ ability to resist occlusal load, mucosa, and as well as underlying residual bone.2,3 A significant challenge for any practitioner is rehabilitating Kennedy’s class I and II situation, which is present in one-third of partially edentulous patients.4 Implant-retained restoration is an option, but due to inadequate amount of bone or the existence of anatomical structures and economic reasons, this is often not feasible. Hence, rehabilitation can be done with either a removable partial denture (RPD) or a semi-fixed prosthesis.5,6
When the functional occlusal load is applied to Kennedy’s class II situation, there are differences in the displacements between the periodontal support tissues of the abutments and the residual ridge mucosa.7 The dual nature and mucosa with different resiliency generate levers around the last support tooth. Additionally, the periodontal ligament’s and mucosa’s induced deformation will likely result in exorbitant distal and medium-directional torque, which could cause the bone reconstruction of the abutment’s alveolar ridge leading to loosening of abutment and finally leading to exfoliation.8
Treatment partial dentures are fabricated in patients where changes are anticipated in the soft tissues, and the final restoration cannot be fabricated until the tissues stabilize. Due to reasons ranging from cost to psychology, many patients prefer the treatment of partial dentures as their definitive prosthesis. Recent improvements in nylon-based thermoplastic resins hold promise as a replacement material that could address the issues with conventional partial dentures.9
Since 1929, definitive cast RPDs have been made with chromium-cobalt (Cr-Co) alloy as the preferred material. It has a metallic framework along with a metallic denture base to dissipate forces.10 Forces are transmitted through rests, guiding planes, and direct retainers during functional movements to abutment teeth. Metal bases have many salient features, such as precision, durability, and resistance to deformation. The high price, esthetic problems due to clasps, and the metallic taste of the partial cast denture are major obstacles in limiting the availability of prosthodontic service.11
Another treatment modality for Kennedy’s class II clinical cases is a union of fixed partial dentures and RPDs using precision and semi-precision attachment.12 This form of prosthesis not only provides esthetics but also provides the functional advantage of a fixed prosthesis that leads to decreased compression of the edentulous ridge by stress distribution, retention, and enhanced phonetics and mastication.13
Despite the tremendous development in the field of prosthodontics, especially in the restoration of partially edentulous arches, the impact of the prosthetic treatments on the remaining tissues or teeth is little understood. Patients have returned with complaints of loose dentures, denture displacement during chewing, altered sensation, and weakening of the periodontal quality of the remaining teeth.14 Due to the absence of an epidemiological survey, the effect of the different prosthetic choices on oral health is often difficult to determine. Thus, this study was done to evaluate and compare the stresses on the terminal abutment and edentulous ridges by using different removable prosthetic options such as a partial cast denture, flexible denture, and semi-precision attachment in Kennedy’s Class II edentulous situations using FEA. The null hypothesis was that there should be no difference in the amount of stress exerted by different prostheses on the abutment and edentulous ridge in class II situations.
MATERIALS AND METHODS
This FEA study was undertaken in the Department of Prosthodontics, Crown, and Bridge and Implantology, Faculty of Dental Sciences, SGT University, Gurugram, Haryana, India, after the approval of the Ethical Committee (SGTU/FDS/MDS/24/1/519).
Construction of Model Geometry for the Calculation
The three-dimensional (3D) geometry of a partially edentulous maxilla with missing molars unilaterally, consisting of both cortical and cancellous bone, was reconstructed from computerized tomography (CT) scans (slice thickness of 0.30 mm) of a patient with missing molars unilaterally that is, Kennedy’s class II situation (Fig. 1). The data were transferred to Mimics software (Materialise NV, Leuven, Belgium) for simulation. This software processed the 3D images and converted the CT data to 3D computer-aided design (CAD) models, and the digital imaging and communications in medicine films were prepared. The models were obtained in the form of a set of stereolithography (STL) files. In order to use these 3D models, the STL files had to be converted into a format (initial graphics exchange specification file). This made it possible to exchange information digitally between CAD systems. Data were transferred for modeling that could be recognized by a processing finite element program such as SolidWorks (SolidWorks standard 2019, Dassault Systèmes, Waltham, Massachusetts, United States of America).5
In this study, three virtual models of the prosthesis for the Kennedy class II partially edentulous jaw were designed using the SolidWorks design computer program with the same abutment tooth, that is the second premolar. The prosthesis for the edentulous portion varied as described below. To describe a virtual model, the average distance between the crest of the alveolar bone and the enamel cementum junction of 2 mm was used.
Model 1 is a free-end saddle made of Co-Cr alloy cast partial denture incorporating the reverse circlet clasps with mesial occlusal rests on the second premolar of the edentulous side and two simple circlet clasps on the dentulous side and denture bases made of heat-cured polymethyl methacrylate resin (Fig. 2A).
Model 2 is a free-end saddle made of flexible denture base resin made of nylon-based super polyamide resin in the area of the first and second molar with the acrylic teeth and clasp encircling the second premolar (Fig. 2B).
Model 3 is a free-end saddle supported by semi-precision attachment on premolars and replacing the first and second molar. It consists of milled crowns made of a Co-Cr alloy covered by ceramics and attached by the mobile part. The mobile component was constructed as the simulated metal base (Co-Cr alloy) covered with acrylate with two teeth (first and second molars) (Fig. 2C).
To ensure accurate results, each component used in the fabrication of the prosthesis, like the virtual model of the fixed part of the restoration with appropriate supporting structures (abutment teeth with milled crowns), virtual model of the mobile free end saddle part base covered with acrylate with two acrylic molar teeth, virtual model of the ”Servo-Dental snap-in-latch” attachment in case of semi-precision attachment were scanned separately and assembled to convert into the 3D solid model of maxilla using solid works modeling software. To survey the legitimacy of the outcomes obtained, the components listed above are consolidated to establish the genuine size system for assessment (1:1 proportion) than natural teeth.5
The virtual model analysis was done by importing the results of the geometries into FEA preprocessing software [Analysis of Systems (ANSYS) Workbench, Canonsburg, Pennsylvania, United States of America] using the translators. To compute the results, a nonlinear analysis was carried out. The imported geometries in ANSYS were checked for geometric continuities and errors. The error/twisted surface geometries were repaired, if any.5
Generation of the Finite element Mesh
The next step was to create a mesh that required the selection of appropriate element types and nodes to obtain an efficient and precise model. The 3D 10-node tetrahedral sort of finite elements (the alternative of 20 nodal, purported Brick element) mesh was utilized. All the boundaries of the materials used were isotropic in the design, each having similar properties in all directions, so there were just two independent material constants.5
Selection of the Material Properties
|Material||Young’s modulus of elasticity (Mpa)2||Poisson’s ratio (γ)4|
|Natural teeth||0.002 × 104||0.30|
|Cortical bone||1.37 × 104||0.30|
|Spongy bone||0.137 × 104||0.30|
|Cr-Co framework||18.5 × 104||0.35|
|Ceramics||6.9 × 104||0.33|
|PMMA||0.29 x 104||0.36|
|Nylon polyamide||0.24 x 104||0.39|
|Stainless steel||19 x 104||0.265|
|Artificial teeth||0.3 x 104||0.30|
Defining the Boundary Conditions
Both the teeth and the denture were demonstrated as linear, homogeneous, and isotropic materials. Since the analysis was static, it was independent of loading time, and all contact border conditions between all parts were considered equal.5
Load Input Determination
During the FEA, each component of the free-end saddle (the first and second molar) of the partially edentulous Kennedy Class II jaw was loaded to observe the behavior of the entire framework and its components under various conditions. It was examined at the same points under the load of the same biting forces by converting the force applied into pressure (as per the equation P = F/S, where P is pressure, F is force, and S represents the region of the tooth on which the force was applied). Three levels of loading (150, 250, and 350 N) were applied from two directions, that is, vertically (90°) and obliquely (60°), on the central fossa of the first and second molars of the prosthesis.5
Execution of Analysis and Interpretation of Results
Visual outcomes were described by levels of color grades, varying between red, green, and blue, with the highest stress values being presented in red. The color gradient table was standardized; subsequently, the colors found in all the analyzed models represented similar amounts of stress. The results of the simulation were analyzed in terms of von Mises equivalent stress levels. The data was divided into three groups according to the type of prosthesis. The data was collected and analyzed using Statistical Package for the Social Sciences (SPSS) software (IBM Corp. Released 2015. IBM SPSS Statistics for Windows, Version 23.0. Armonk, New York, United States of America; IBM Corp.). An independent t-test was used to compare the values of the three groups. ANOVA was used to test the differences between the loading directions.
The result for each loading was obtained as stress distribution-colored images (Fig. 3), and numerical values were recorded. It was observed that on increasing the forces from 150 to 350 N, the stresses exerted by all three prostheses increased both on the abutment and edentulous ridge (Table 2). Also, when the direction of forces changed from vertical to oblique again, the forces were increased (Table 2).
|S. no.||Design of prosthesis||Loading forces||Angulation||Abutment tooth cervically (MPa)||Edentulous ridge (MPa)|
|1.||Cast partial denture||150 N||90°||7.7||3.3|
|2.||Flexible denture||150 N||90°||43.55||58.13|
|3.||Semi-precision attachment||150 N||90°||1.82||0.9105|
On comparing the force exerted by the three prostheses on the terminal abutment and edentulous ridge, it was observed that in the flexible denture, the forces were significantly (p = 0.171) more on the ridge (93.29 MPa) as compared to the abutment (71.39 MPa) (Fig. 4) whereas in both cast partial (8.63MPa and 3.69 MPa; p = 0.001) (Fig. 3) and semi-precision attachment (3.15 and 1.57 MPa; p = 0.011) (Fig. 5) it was significantly more on the abutment teeth (Table 3).
|Design of prosthesis||Location||Mean stress values (MPa)||Standard deviation||p-value|
|Cast partial denture||Abutment tooth||8.63300||0.740512||0.001*|
|Flexible denture||Abutment tooth||71.39367||22.655944||0.171|
|Semi-precision attachment||Abutment tooth||3.155833||1.1004265||0.011*|
*Indicates p-value of <0.05 as significant
On comparing mean stress among all the three prosthesis designs on the terminal abutment, maximum stress was seen in flexible denture followed by cast partial and semi-precision attachment at both vertical and oblique forces, and the difference was statistically significant with p = 0.001 (Table 4). Similar results were obtained by comparing mean stress among all the three prosthesis designs on the edentulous ridge (p = 0.001) (Table 5).
|Location||Prosthesis type||Mean stress at 90° (MPa)||Standard deviation at 90°||Mean stress at 60° (MPa)||Standard deviation at 60°|
|Terminal abutment||Cast partial denture||8.01600||0.315005||9.25000||0.360000|
|p = 0.001*|
*Indicates p-value of <0.05 as significant
|Location||Prosthesis type||Mean stress at 90° (MPa)||Standard deviation at 90°||Mean stress at 60° (MPa)||Standard deviation at 60|
|Edentulous ridge||Cast partial denture||3.430000||0.1300000||3.96000||0.150000|
|p = 0.001*|
*Indicates p-value of <0.05 as significant
The principles involved in the science of RPD fabrication have grown to the next level and are way past the mechanical stage. This form of prosthesis encompasses both hard and soft tissues, as well as an amalgamation of material science, lever principles, and physics.15
According to Jain et al.,16 due to poor denture stabilization, retention, and rotational motions, the usage of unilateral distal extension partial dentures is restricted. Since it lacks the effect of cross-arch stabilization, a unilateral prosthesis is often less secure. Giffin,17 in his research, stated that when the functional occlusal load is applied, it creates levers that lead to a rotary movement that typically takes place near the fulcrum of the terminal abutments. This condition not only impairs the patient’s comfort and denture performance but also traumatizes the tissues that support dentures and damages the abutments.2,5
Therefore, for the proper functioning of a prosthesis, it is essential to take into consideration not only the retention, stability, and support but also the distribution of stresses to the abutment and supporting tissues. Poor designing of the prosthesis may lead to uneven stress distribution resulting in excessive resorption of the alveolar bone.18,19 The null hypothesis proposed in the present study was rejected. In the present study, observations were made in the form of von Mises stress. Different prosthesis designs permit different movements in three planes which affect dental biomechanics. On comparing the mean stress values, it was found to be significantly higher in the edentulous ridge (93.297 MPa) as compared to abutment teeth (71.393 MPa) in flexible dentures. This could be because of the inherent flexibility of thermoplastic nylon that acted as a stress breaker, inducing less stress on abutment teeth or maybe less prone to fracture and deformity.20 The flexible base snugly adapts over the underlying tissues leading to the broad stress distribution and therefore transmits less stress on the abutment teeth than on the edentulous ridge.21
As noticed, on comparing the mean stress values at the cervical aspect of terminal abutment among all the three prosthesis designs at 90 and 60° masticatory loading, it was found to be significantly higher among flexible partial dentures, that is, 66.27 MPa at 90°, 22.72 MPa at 60° followed by cast partial and semi-precision attachment under both loading conditions. In the present study, while evaluating flexible dentures, the stress concentration was located on the buccal side of the abutment and in the middle third of its root. The reason for this lies in the fact that due to the rotational movement of the tooth around a center point in the root toward the buccal, which agrees with studies done by Goodkind22 and Tebrock et al.23
The maximum von Mises stress values on abutment teeth and the ridge at both 90 and 60° were found to be nonsignificant when cast partial denture was compared with semi-precision attachment. In both cases, the stress values were more on the abutment than the ridge, but again the difference was not statistically significant. In the partial cast denture on applying the vertical and oblique forces of 150, 250, and 350 N, it was noticed that the deflections were maximum at the buccal slope of the edentulous ridge, and the occlusal rests of terminal abutment of the edentulous side, at the entire crest, and on the first premolar and second molar on the dentulous side. On applying a vertical force of 150 N, von Mises Stress of 7.7 MPa was present occlusal on abutment teeth, while von Mises stress of 3.3 MPa was present on the edentulous ridge, which increased with oblique forces and increasing load. This result shows that stresses were comparatively higher on abutment teeth than on edentulous ridge. This could possibly be due to cross arch stability provided by the major connector and tripod configuration to equally distribute the forces among all the involved abutments to prevent damage to the supporting structures and thus transmit fewer stresses on the edentulous ridge.24
Thus, cast partial denture not only provides adequate retention and stability but also increases the masticatory efficiency and preserves the remaining oral structures.25 The design of clasps is such that the reciprocal arm counteracts the forces of the retentive arm during function.
On evaluating the stresses generated on the terminal abutment and edentulous ridge when restored with semi-precision attachment partial denture, it was found that on applying 150 N vertical force, von Mises stress was found to be 1.82 MPa, on the abutment teeth, whereas it was 0.91 MPa on the edentulous ridge, which increased on applying oblique forces. This result shows that these stresses were almost the same on the edentulous ridge and abutment teeth and increased with increasing force from 150 to 350 N. The advent of precision attachments has neglected the need for clasps. Therefore, the maximum von Mises stress on the attachment itself is 11.33 MPa. This could be due to a certain amount of movement between the two sections of the prosthesis. Attachment being resilient acts as a nonrigid stress breaker and helps distribute the occlusal load without breaking the prosthesis.26 It offsets the leverage forces exerted on the crown effectively. The attachment here accepts most of the loads and may undergo plastic deformation and fatigue, which protects the abutment teeth and edentulous ridge.27 Therefore, the attachment itself acts as a stress breaker, causing the equal distribution of stresses on the abutment and edentulous ridge, which is in agreement with the research of Todorovic et al.,24 who concluded that the when the abutment is in the partial edentulous situation and is subjected to vertical and oblique stresses the attachment absorbs the stress and thus saving the abutment from any damage.
According to Owall,26 extra coronal precision attachments are resilient and allow the prosthesis to move freely and distribute the harmful forces away from the abutments. However, the result of the present study was not in agreement with Wang et al.12 who found that the maximum stress was concentrated on the residual ridge with stress-breaking design in distal extension cases.
The present FEA study demonstrated that flexible partial denture exerts maximum von Mises stresses on the abutment tooth as well as on the residual ridge, and hence its use as a definitive prosthesis should be avoided. On the contrary, the cast partial denture fabrication process is complex but occlusal forces are distributed well among abutment teeth and edentulous ridge. However, semi-precision attachment exerts the least stress on the ridge and abutment tooth as maximum stress is accepted by the attachment itself, thus acting as a stress breaker and preventing stress transfer on the residual ridge. Therefore, considered the best treatment option, followed by a partial cast denture.
The present study has certain limitations, like the essential anisotropic tissues were considered isotropic. The static load loads that were different from the dynamic loads seen during the function were added. Finite element modeling is an extremely accurate and precise method for structure analysis. However, living structures are more than mere artifacts, which are beyond the confines of set parameters and values.
Within the limitations of this FEA study, it can be concluded that semi-precision attachment and cast partial denture did not show significant changes in stress levels of abutment teeth and edentulous ridge. However, flexible dentures show the highest stress levels on the abutment and edentulous ridge. Hence within the confines of the study, it can be suggested that semi-precision is the most recommended treatment option, followed by cast partial dentures. Clinicians should use flexible dentures as only provisional prostheses and should not be used as definitive prostheses.
Bhupender Yadav https://orcid.org/0000-0002-7432-0918
1. Suwal P, Singh R, Ayer A, et al. Cast partial denture versus acrylic partial denture for replacement of missing teeth in partially edentulous patients. J Dent Mater Tech 2017;6(1):27–34. DOI: 10.22038/JDMT.2016.7841
2. Ahmed B, Alkhiary Y, Amin H, et al. Stress analysis of different designs of distal extension partial dentures with pier abutment: a finite element analysis. Egyptian Dent J 2015;61(2):1451–1457.
4. Patrnogić V, Todorović A, Sćepanović M, et al. Free-end saddle length influence on stress level in unilateral complex partial denture abutment teeth and retention elements. Vojnosanit Pregl 2013;70(11):1015–1022. DOI: 10.2298/vsp110603028p
5. Sabri LA, Abdulkareem JF, Salloomi KN, et al. Finite element analysis of class II mandibular unilateral distal extension partial dentures. Proc Insti Mech Eng, Part C: J Mech Eng Sci 2022;236(17):9407–9418. DOI: 10.1177/0954406222109663
7. Igarashi Y, Ogata A, Kuroiwa A, et al. Stress distribution and abutment tooth mobility of distal-extension removable partial dentures with different retainers: an in vivo study. J Oral Rehabil 1999;26(2):111–116. DOI: 10.1046/j.1365-2842.1999.00345.x.
8. Chen X, Mao B, Zhu Z, et al. A three-dimensional finite element analysis of mechanical function for 4 removable partial denture designs with 3 framework materials: CoCr, Ti-6Al-4V alloy and PEEK. Sci Rep 2019;9(1):13975. DOI: 10.1038/s41598-019-50363-1
9. Hundal M, Madan R. Comparative clinical evaluation of removable partial dentures made of two different materials in Kennedy Applegate class II partially edentulous situation. Med J Armed Forces India 2015;71(Suppl 2):S306–S312. DOI: 10.1016/j.mjafi.2012.08.020
11. Ramakrishnan H, Singh RG. Three-dimensional finite element analysis of the stress distribution pattern in the design modifications of U-shaped palatal major connector. Indian J Dent Res 2010;21(4):506–511. DOI: 10.4103/0970-9290.74219
12. Wang HY, Zhang YM, Yao D, et al. Effects of rigid and nonrigid extracoronal attachments on supporting tissues in extension base partial removable dental prostheses: a nonlinear finite element study. J Prosthet Dent 2011;105(5):338–346. DOI: 10.1016/S0022-3913(11)60066-8
13. Patel H, Patel K, Thummer S, et al. Use of precision attachment and cast partial denture for long-span partially edentulous mouth-a case report. Int J Appl Dent Sci 2014;1(1):22–25.
15. Petridis H, Hempton TJ. Periodontal considerations in removable partial denture treatment: a review of the literature. Int J Prosthodont 2001;14(2):164–172.
16. Jain AR, Philip JM, Ariga P. Attachment-retained unilateral distal extension (Kennedy’s class II modification I) cast partial denture. Int J Prosthodont Restor Dent 2012;2(3):101–107. DOI: 10.5005/jp-journals-10019-1057
19. Xiao W, Li Z, Shen S, et al. Influence of connection type on the biomechanical behavior of distal extension mandibular removable partial dentures supported by implants and natural teeth. Comput Methods Biomech Biomed Engin 2016;19(3):240–247. DOI: 10.1080/10255842.2015.1009450
21. Kumar N, Koli DK, Jain V, et al. Stress distribution and patient satisfaction in flexible and cast metal removable partial dentures: finite element analysis and randomized pilot study. J Oral Biol Craniofac Res 2021;11(4):478–485. DOI: 10.1016/j.jobcr.2021.06.004
23. Tebrock OC, Rohen RM, Fenster RK, et al. The effect of various clasping systems on the mobility of abutment teeth for distal-extension removable partial dentures. J Prosthet Dent 1979;41(5):511–516. DOI: 10.1016/0022-3913(79)90082-9
24. Todorović A, Radovic K, Grbovic A, et al. Stress analysis of a unilateral complex partial denture using the finite-element method. MaterTech 2010;44(1):41–47.
25. Bhathal M, Batra J, Attresh G, Sambyal S. A review on stresses-induced by removable partial dentures. Int J Contemp Dent Med Rev 2015.
26. Owall B. Precision attachment-retained removable partial dentures: Part 2. Long-term study of ball attachments. Int J Prosthodont 1995;8(1):21–28.
27. Leo HE, Wu JH, Wang CH. Biomechanical analysis of distal extension removable partial dentures with different retainers. J Dent Sci 2008;3(3):133–139.
© The Author(s). 2023 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (https://creativecommons.org/licenses/by-nc/4.0/), which permits unrestricted use, distribution, and non-commercial reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.