Cuspal Flexure and Stress in Restored Teeth Caused by Amalgam Expansion
Cuspal Flexure and Stress in Restored Teeth Caused by Amalgam Expansion
Cuspal Flexure and Stress in Restored Teeth Caused by Amalgam Expansion
Clinical Relevance
Although amalgam is being phased out, existing amalgam fillings will still be present for
many years. Clinicians should be aware that amalgam expansion may create stress
conditions that accelerate tooth cracking.
Despite the respectable life span of amalgam Cuspal Flexure and Restoration Expansion
fillings, often surpassing that of resin-based compos- Eighteen extracted human maxillary and mandibu-
ite restorations,4-6 their ultimate failure can have lar molars (approved by the Institutional Review
serious consequences for the longevity of a tooth. Board) were mounted in stainless steel rings with
Common reasons for restoration failures are second- embedded reference spheres (Figure 2). The mean
ary caries, fracture of the filling, or tooth fracture.6-8 and standard deviation of the buccal-lingual widths,
Secondary caries or fractured fillings can be treated measured at the height of contour, were 10.0 6 0.6
by replacing the filling, but a fractured tooth will mm. Nine teeth were filled with amalgam, and the
require complex or costly indirect restoration or even other nine teeth were left intact (no preparation). A
become unrestorable. sample size of nine had 95% confidence to detect a
difference of 0.65 standard deviation between
Cracks are a common sight in teeth. Dental
groups. The external enamel surfaces were etched
practitioners often observe cracks developing at a
with 37% phosphoric acid solution to obtain dull
marginal ridge or radiating from amalgam fillings,
surfaces. A mesio-occluso-distal (MOD) cavity (4-mm
as shown in Figure 1. Fractures are the final stage of
deep, 4-mm wide) was prepared in nine teeth and
crack propagation. Teeth restored with nonbonded
was restored with zinc-containing amalgam (Per-
amalgam were found to be more likely to have cracks
mite, SDI, Bayswater, Victoria, Australia). The
or fractured tooth structures than adhesive compos-
external tooth surfaces and reference spheres were
ite restorations.6,8,9 Restoration, cavity design, ex-
scanned from eight directions following restoration
cursive occlusal interference, and age have all been
using a three-dimensional optical scanner (COMET
identified as factors that predispose a tooth to
xS, Steinbichler Optotechnik GmbH, Neubeuern,
cracking.9-11 Germany). This scan was used as a baseline. After
Another factor that may also play a role is scanning, the restored teeth were immersed in
expansion of the amalgam restoration, which would normal saline solution (0.9% sodium chloride irriga-
introduce stresses in a restored tooth. Amalgam is tion USP; B. Braun Medical Inc, Bethlehem, PA,
known to expand due to phase changes and corro- USA) for eight weeks and rescanned at two, four,
sion.12,13 Expansion of amalgam in a confined cavity and eight weeks. Nine unprepared teeth were used
E302 Operative Dentistry
Figure 2. Teeth were secured in stainless steel rings with embedded reference spheres. The scans were aligned with their baseline using the
reference spheres (yellow). Cuspal flexure and expansion of the amalgam restoration on the occlusal, mesial, and distal surfaces are shown
according to the color scale.
as a control. They were scanned and stored using the preprocessor software, with 08, 58, or 108 of wall
same procedure as the restored teeth. Scanned convergence. Each cavity was 4-mm deep and 4-mm
surfaces were aligned with the baseline scans by wide, as shown in Figure 3. The outlined dentin,
means of the reference spheres on the mounting enamel, and restoration were meshed manually
rings using Cumulus software (Regents of the using quadrilateral elements, using an element
University of Minnesota; Figure 2). Buccal and relaxation option that maintained optimal element
lingual tooth surface areas (from height of cavity shapes. The root was not modeled because restora-
floor to just below the cuspal ridges) as well as tion expansion affected only the tooth crown. The
occlusal and proximal amalgam filling surfaces were tooth models were therefore fixed below the cemento-
separately selected on the baseline scans. Custom- enamel junction. Amalgam restoration surfaces were
developed software compared each of the selected free to separate or slide along the surrounding cavity
baseline surface areas with the follow-up surface surfaces with an arbitrary 0.5 coefficient of friction.
scans. Differences between the baseline and follow-
Material properties were assigned to the dentin,
up surfaces, which geometrically constitute a vol-
enamel, and restoration elements. The applied
ume, were calculated and divided by the selected
surface areas to obtain the buccal and lingual cusp elastic moduli (84 GPa for enamel, 18 GPa for
flexure and the occlusal and proximal amalgam dentin, and 28 GPa for amalgam) as well as the
filling expansion. Total cuspal flexure was the sum Poisson’s ratios (0.30 for enamel, 0.24 for dentin, and
of the buccal and lingual flexure; proximal expansion 0.35 for amalgam) were obtained from the litera-
was the average of the mesial and distal amalgam ture.15 Transverse isotropic conditions were pre-
filling expansion. Effects of amalgam restoration on scribed for the enamel, assuming the 84 GPa elastic
cuspal flexure were compared with the control group modulus value along the principal enamel axis,
at each time interval with multivariate tests and perpendicular to the enamel-dentin junction, where-
post hoc multiple comparisons using the Games- as the value was 42 GPa in lateral directions. Since
Howell test (equal variances not assumed; IBM the objective of this analysis was stress in the tooth
SPSS Statistics, version 25.0, Armonk, NY, USA). structure for a specific experimentally measured
Effects of storage time on cuspal flexure of the cuspal deflection value, creep of the amalgam did not
amalgam-restored teeth and control teeth were need to be modeled.
analyzed using general linear model repeated mea- To produce the specific cuspal deflection, volumet-
sures (IBM SPSS Statistics). ric expansion was prescribed for the modeled
amalgam restoration using thermal analogy. This
Finite Element Analysis involved increasing notional temperature in the
Volumetric expansion was applied in a finite element amalgam to induce expansion. The resulting volume
model of an MOD-restored tooth to evaluate the change pressed the amalgam restoration against the
stress distribution associated with experimentally surrounding cavity floor and walls, generating
measured cuspal flexure. The finite element model deformation and stress in the tooth, outward flexure
was based on an image of a cross-sectioned tooth. of the cusps, and an elevation of the occlusal surface
The outlines of the dentin and enamel were traced as it was extruded. In addition, occlusal forces were
and imported into a finite element program (Marc/ applied on the lingual or buccal cuspal inclines to
Mentat, MSC Software, Palo Alto, CA, USA). Three simulate a chewing load of 20 N distributed over a
cavity shapes were created using the Mentat 1.5-mm2 surface area.
Danley & Others: Amalgam Expansion–Caused Stress E303
Figure 3. Calculated displacements and stresses in the cross section of an amalgam-filled tooth with 3-lm cuspal flexure due to amalgam
expansion. Stress concentrations are seen at the cavity line angles. Stress concentrations intensified when a masticatory load was added on the
lingual or buccal cusp (20 N over 1.5 mm2 area, indicated with arrows).
Stresses in three directions (cross-sectional plane of the occlusal, mesial, and distal amalgam surfaces
and perpendicular to the cross section according to were calculated over each affected surface (Table 1).
plane strain conditions) were collected in the dentin Positive values indicate outward movement (expan-
and enamel elements during the simulation of sion). Multivariate tests showed that the amalgam
amalgam expansion and cuspal flexure. The combina- restoration caused significant cuspal flexure at each
tion of these three stress components was expressed in time interval. The repeated-measures tests indicated
an equivalent stress value according to the modified that the storage time had no significant effect on
von Mises criterion. This criterion is based on the either the amalgam (p=0.202) or control groups
common von Mises criterion but modified such that it (p=0.069). Amalgam on proximal surfaces expanded
can account for the higher compressive and lower more than those on the occlusal surface.
tensile strengths of enamel, dentin, and amalgam. To match the experimentally determined cuspal
The modification assigns tensile stresses higher flexure of about 3 lm, a 0.3% volumetric (=0.1%
‘‘weight’’ in the equivalent stress expression than linear) expansion was applied in the finite element
compressive stresses.16 The compressive/tensile analysis. The simulated amalgam expansion caused
strength ratios used were 384/10 for enamel, 297/99 stresses in the tooth structure, with stress concen-
for dentin, and 388/66 for amalgam.15 trations at the cavity floor and line angles (Figure 3).
The occlusal surface extruded 3.5 to 4 lm due to the
RESULTS amalgam expansion. The 58 of wall convergence
Displacements were measured across all external increased cuspal flexure by 26% compared with a
surfaces (Figure 2) after two-, four-, and eight-week cavity with straight walls and 29% when the wall
storage. Mean cuspal flexure and expansion values convergence was 108. Stress concentrations in dentin
E304 Operative Dentistry
Amalgam expansion and creep are generally Stress levels in restored teeth depend on many
considered positive effects because they close the factors, where each combination of amalgam alloy,
interfacial gap,14 thereby tacitly assuming that cavity configuration, and tooth will affect the stress
expansion is an insignificant factor in tooth frac- development and distribution. Not all variations could
tures. In a photoelastic study, cylindrical amalgam be covered in this study. The amalgam we used is a
fillings caused only minimal stresses in a block of high-copper, non-gamma-2 admixed alloy containing
photoelastic material.26 To test the significance of 0.2% zinc.28 High-copper amalgam associated with
the measured flexure values, we calculated the expansion and low creep is thought to have contrib-
stress level associated with 3-lm cuspal flexure of uted to an increased incidence and severity of cusp
a tooth with MOD filling using finite element fracture of endodontically treated posterior teeth.29
analysis. The analysis showed stress concentrations Other amalgams, for example, zinc-containing low-
at the cavity line angles. Stress concentrations are copper alloys, could have caused more expansion if
areas where stresses are higher compared with contaminated with moisture,30 or amalgams exhibit-
surrounding areas (indicated by orange and yellow ing more creep could have resulted in lower cuspal
colors in Figure 3) and therefore identify locations flexure values.14 We chose the zinc-containing high-
where maximum stress will be exceeded first and copper alloy because it allowed us to accelerate the
crack initiation is thus most likely. The analysis thus aging process needed to assess cuspal flexure before
confirmed clinical observations that line angles are reference surfaces became unstable. The controls
usually involved in cuspal fractures. Nevertheless, showed that after eight weeks in saline, the stability
there is no clinical or experimental evidence that of the reference surfaces became unreliable. Note that
amalgam expansion stress is sufficient to cause cusp moisture contamination was not an issue because the
fracture.21 In our finite element analysis, the highest fillings were placed in a dry in vitro environment.
(principal) stresses at the cavity line angles were in The choice of cavity configuration also affected our
the range of 10 to 25 MPa, which is well below the outcomes. A large MOD cavity significantly weakens
tensile strength of dentin and thus supports the the tooth structure, which increases the tooth
notion that amalgam expansion alone is not suffi- deformation and thus stresses in the tooth. The slot
cient to cause tooth fracture. cavity design in this study had a constant 4-mm
Cuspal fracture usually occurs during functional depth without proximal boxes to achieve maximal
or parafunctional loading, often when no excessive cuspal deformation. A small occlusal cavity or
forces are applied. Anecdotes of cuspal or tooth conventional Class II cavity is less injurious to the
fractures when chewing on soft food are very tooth stiffness, which results in less tooth deforma-
familiar. Such fracture behavior has all the charac- tion from amalgam expansion and thus lower
teristics of fatigue failure, which is a process of crack stresses in the tooth. This may explain the previ-
propagation under repetitive loading that is lower ously mentioned low stresses in a block of photoe-
than the original structural strength. Amalgam lastic material with cylindrical cavity.26 Note, how-
expansion stresses may not be sufficient to fracture ever, that stress concentrations will change
a cusp, but they may accelerate fatigue crack depending on the cavity configuration. In occlusal
initiation and propagation by raising stress levels fillings, stress concentrations due to expansion are
in a tooth. We used the finite element analysis to test likely to shift to the occlusal surface where the lower
this hypothesis and found that amalgam expansion tensile strength and fracture toughness of enamel
stresses associated with a 3-lm cuspal flexure may predispose occlusal margins to increased risk of
increased stress concentrations at the cavity line crack initiation and propagation. In this study, we
angles 44% to 178% compared with a nonexpanding considered three levels of convergence for our MOD
amalgam filling. According to the classic crack cavities (08, 58, and 108). Generally, stress concen-
growth equation known as Paris’ law, the fatigue trations intensified with increasing convergence.
crack rate is essentially a power function of the However, the value did not only depend on conver-
stress amplitude.27 Increases in the level of repeti- gence, but also the surrounding, remaining tooth
tive stress values could thus result in significant structure. For example, a smaller cavity with higher
acceleration of crack propagation and in higher conversion angle may generate lower stress concen-
incidence of fatigue fractures. The clinical signifi- trations than a large cavity with less convergence.
cance of amalgam expansion is therefore not its Similarly, stress distributions are affected by the
inability to break a cusp but rather the effect it has location and distribution of occlusal loading. Two
on increasing the stress levels. load conditions were examined in this study. They
E306 Operative Dentistry
showed that loading of unsupported cusps creates tions in the tooth structure at the lingual and buccal
significant stress concentrations at the cavity line internal line angles. These expansion stresses added
angles. Those stress concentrations caused by mas- to stresses generated by masticatory loads and were
ticatory loading can be avoided by bonding a filling to hypothesized to contribute to initiation and propa-
the cavity walls.31 However, stress concentrations at gation of tooth cracks incidentally observed around
cavity line angles that are generated by restoration amalgam restorations. Results from the finite ele-
expansion will not be prevented by bonding. Note ment analysis suggested that cavity convergence
also that expansion stresses differ from more may be a contributing factor.
familiar polymerization shrinkage stresses because
stresses caused by the expansion of Class II Acknowledgements
restorations tend to concentrate on internal tooth
This study was supported, in part, by the University of
surfaces, whereas stresses caused by shrinkage tend Tennessee Health Science Center College of Dentistry Alumni
to concentrate on external tooth surfaces.31 Restora- Endowment Fund, the Tennessee Dental Association Foun-
tion expansion stress concentrations thus coincided dation, and by the Alpha Omega Foundation Research Fund.
The authors are grateful to Dr Ralph DeLong, University of
with the location where stresses generated by Minnesota, who developed the Cumulus software.
masticatory loading also concentrated.
Clearly, predicting stresses is complex and should Regulatory Statement
not be generalized because unique factors of each tooth- This study was conducted in accordance with all the provisions
restoration complex need to be taken into account. The of the local human subjects oversight committee guidelines and
policies of the University of Tennessee Health Science Center.
results of this study should therefore not be applied The approval code for this study is 15-04302-NHSR.
blindly to other alloys or cavity configurations. The
significance of our experiment, however, is the dem- Conflict of Interest
onstration that amalgam has the potential to signifi-
The authors of this article certify that they have no
cantly raise stress levels in filled teeth. Even though proprietary, financial, or other personal interest of any nature
the life span of amalgam fillings is respectable,4,6,8 or kind in any product, service, and/or company that is
cracks in the tooth structure can lead to extensive re- presented in this article.
restoration or render a tooth unrestorable.
Note
Clinicians should therefore be aware that cracks
This study was presented in part at the International
associated with amalgam restorations are likely to Association for Dental Research annual meeting in San
propagate in a fatigue process that eventually leads to Francisco, March 2017.
catastrophic fracture. Viewing teeth with amalgam
filling up close with an intraoral camera helps to verify (Accepted 30 May 2018)
the presence of microcracks, especially when using a
REFERENCES
bright light to transilluminate the tooth. Moreover,
clinicians should be aware that higher convergent 1. US Public Health Service, Committee to Coordinate
Environmental Health and Related Programs (1993)
cavity walls as well as sharp cavity line angles Dental Amalgam: A Scientific Review and Recommended
intensify the local stress concentrations, which are Public Health Service Strategy for Research, Education
likely to accelerate crack propagation, and that and Regulation: Final report of the Subcommittee on Risk
bonding an amalgam filling may restore structural Management (PHS publication No. 342-322/60025) US
integrity but will not eliminate the effect of amalgam Government Printing Office, Washington, DC.
expansion stresses. Amalgam fillings have a good 2. Beazoglou T, Eklund S, Heffley D, Meiers J, Brown LJ, &
track record for longevity and technique insensitivity, Bailit H (2007) Economic impact of regulating the use of
amalgam restorations Public Health Reports 122(5)
but if clinicians are concerned about maintaining
657-663.
tooth integrity, amalgam may not be the best choice.
Amalgam fillings may create stress conditions that 3. US Food and Drug Administration (2009) Department of
Health and Human Services, 21 CFR Part 872 Docket No.
accelerate tooth cracking. Amalgam fills a cavity but FDA-2008-N-0163 Dental Devices: Classification of Den-
does not restore the tooth.31 tal Amalgam, Reclassification of Dental Mercury, Desig-
nation of Special Controls for Dental Amalgam, Mercury,
CONCLUSION and Amalgam Alloy; Final Rule. Federal Register Vol. 74,
No. 148 August 2009. Retrieved online April 10, 2018
During the eight-week storage in normal saline from: https://www.gpo.gov/fdsys/pkg/FR-2009-08-04/pdf/
solution, amalgam fillings expanded and pushed E9-18447.pdf
cavity walls outward. Finite element analysis 4. Mjör IA, Dahl JE, & Moorhead JE (2000) Age of
showed that the expansion caused stress concentra- restorations at replacement in permanent teeth in
Danley & Others: Amalgam Expansion–Caused Stress E307
general dental practice Acta Odontologica Scandinavica Nigerian teaching hospital Journal of Endodontics 35(3)
58(3) 97-101. 334-336.
5. Forss H & Widström E (2001) From amalgam to 19. Sheth JJ, Fuller JL, & Jensen ME (1988) Cuspal
composite: selection of restorative materials and restora- deformation and fracture resistance of teeth with dentin
tion longevity in Finland Acta Odontologica Scandinavica adhesives and composites Journal of Prosthetic Dentistry
59(2) 57-62. 60(5) 560-569.
6. Van Nieuwenhuysen JP, D’Hoore W, Carvalho J, & Qvist 20. Paffenbarger GC, Rupp NW, & Patel PR (1979) Dimen-
V (2003) Long-term evaluation of extensive restorations sional change of dental amalgam and a suggested
in permanent teeth Journal of Dentistry 31(6) 395-405. correlation between marginal integrity and creep Journal
7. Forss H & Widström E (2004) Reasons for restorative of the American Dental Association 99(1) 31-37.
therapy and the longevity of restorations in adults Acta 21. Jokstad A (1991) Influence of cavity depth on marginal
Odontologica Scandinavica 62(2) 82-86. degradation of amalgam restorations Acta Odontologica
8. Opdam NJM, Bronkhorst EM, Loomans BAC, & Huys- Scandinavica 49(2) 65-71.
mans MCDNJM (2010) 12-year survival of composite vs 22. Versluis A, Tantbirojn D, Lee MS, Tu LS, & DeLong R
amalgam restorations Journal of Dental Research 89(10) (2011) Can hygroscopic expansion compensate polymeri-
1063-1067. zation shrinkage? Part I: deformation of restored teeth
9. Ratcliff S, Becker IM, & Quinn L (2001) Type and Dental Materials 27(2) 126-133.
incidence of cracks in posterior teeth Journal of Prosthetic 23. Tantbirojn D, Pfeifer CS, Braga RR, & Versluis A (2011)
Dentistry 86(2) 168-172. Do low-shrink composites reduce polymerization shrink-
10. Kahler B, Kotousov A, & Melkoumian N (2006) On age effects? Journal of Dental Research 90(5) 596-601.
material choice and fracture susceptibility of restored 24. Suiter EA, Watson LE, Tantbirojn D, Lou JSB, & Versluis
teeth: an asymptotic stress analysis approach Dental A (2016) Effective expansion: balance between shrinkage
Materials 22(12) 1109-1114. and hygroscopic expansion Journal of Dental Research
11. Lubisich EB, Hilton TJ, & Ferracane J (2010) Cracked 95(5) 543-549.
teeth: a review of the literature Journal of Esthetic and 25. Francis AV, Verı́ssimo C, Braxton AD, Tantbirojn D,
Restorative Dentistry 22(3) 158-167. Soares CJ, & Versluis A (2014) Cusp flexure caused by
12. Jensen SJ & Jörgensen KD (1985) Dimensional and phase cavity preparation Journal of Dental Research
changes of dental amalgams Scandinavian Journal of 93(Special Issue B) Abstract 949.
Dental Research 93(4) 351-356. 26. Osborne JW (1999) Expansion of contaminated amalgams
13. Okabe T & Mitchell RJ (1996) Setting reactions in dental assessed by photoelastic resin Quintessence International
amalgam: part 2. The kinetics of amalgamation Critical 30(10) 673-681.
Reviews in Oral Biology and Medicine 7(1) 23-35. 27. Paris P & Erdogan F (1963) A critical analysis of crack
14. Osborne JW (2006) Creep as a mechanism for sealing propagation laws Journal of Basic Engineering 85(4)
amalgams Operative Dentistry 31(2) 161-164. 528-533.
15. Craig RG & Powers JM (2002) Restorative Dental 28. SDI (2011) Pre-dosed Amalgam Capsules Permite, Lojic þ
Materials Mosby, St. Louis MO. & GS-80; Instructions for Use Retrieved online October
11, 2017 from: http://www.webcitation.org/6sxsXNlRR
16. Versluis A, Tantbirojn D, & Douglas WH (1997) Why do
shear bond tests pull out dentin? Journal of Dental 29. Hansen EK & Asmussen E (1993) Cusp fracture of
Research 76(6) 1298-1307. endodontically treated posterior teeth restored with
amalgam: teeth restored in Denmark before 1975 versus
17. World Health Organization (2010) Future Use of Materi-
after 1979 Acta Odontologica Scandinavica 51(2) 73-77.
als for Dental Restoration: Report of the Meeting Con-
vened at WHO HQ, Geneva, Switzerland 16th to 17th 30. Yamada T & Fusayama T (1981) Effect of moisture
November 2009 WHO Document Production Services, contamination on high-copper amalgam Journal of
Geneva, Switzerland. Retrieved online October 11, 2017 Dental Research 60(3) 716-723.
from: http://www.webcitation.org/6sxseUsC7
31. Versluis A & Tantbirojn D (2011) Filling cavities or
18. Udoye CI & Jafarzadeh H (2009) Cracked tooth syn- restoring teeth? Journal of the Tennessee Dental Associ-
drome: characteristics and distribution among adults in a ation 91(2) 36-42.