d e n t a l m a t e r i a l s 2 8 ( 2 0 1 2 ) 961–973

Available online at www.sciencedirect.com

journal homepage: www.intl.elsevierhealth.com/journals/dema

Correlation of wear in vivo and six laboratory wear methods

S.D. Heintze a,∗ , M. Faouzi b , V. Rousson b , M. Özcan c

a R&D, Ivoclar Vivadent AG, Schaan, Liechtenstein
b Statistical Unit, Institute for Social and Preventive Medicine, University of Lausanne, Switzerland
c University of Zurich, Dental Materials Unit, Center for Dental and Oral Medicine, Clinic for Fixed and Removable Prosthodontics and

Dental Materials Science, Zurich, Switzerland

a r t i c l e i n f o a b s t r a c t

Article history: Objective. We examined the correlation between clinical wear rates of restorative materi-
Received 26 October 2009 als and enamel (TRAC Research Foundation, Provo, USA) and the results of six laboratory
Received in revised form test methods (ACTA, Alabama (generalized, localized), Ivoclar (vertical, volumetric), Munich,
21 November 2011 OHSU (abrasion, attrition), Zurich).
Accepted 16 April 2012 Methods. Individual clinical wear data were available from clinical trials that were conducted
by TRAC Research Foundation (formerly CRA) together with general practitioners. For each
of the n = 28 materials (21 composite resins for intra-coronal restorations [20 direct and 1
Keywords: indirect], 5 resin materials for crowns, 1 amalgam, enamel) a minimum of 30 restorations
Wear had been placed in posterior teeth, mainly molars. The recall intervals were up to 5 years
Laboratory test method with the majority of materials (n = 27) being monitored, however, only for up to 2 years. For
Correlation the laboratory data, the databases MEDLINE and IADR abstracts were searched for wear data
Composite on materials which were also clinically tested by TRAC Research Foundation. Only those data
Clinical for which the same test parameters (e.g. number of cycles, loading force, type of antagonist)
had been published were included in the study. A different quantity of data was available
for each laboratory method: Ivoclar (n = 22), Zurich (n = 20), Alabama (n = 17), OHSU and ACTA
(n = 12), Munich (n = 7). The clinical results were summed up in an index and a linear mixed
model was fitted to the log wear measurements including the following factors: material,
time (0.5, 1, 2 and 3 years), tooth (premolar/molar) and gender (male/female) as fixed effects,
and patient as random effect. Relative ranks were created for each material and method;
the same was performed with the clinical results.
Results. The mean age of the subjects was 40 (±12) years. The materials had been mostly
applied in molars (81%) and 95% of the intracoronal restorations were Class II restorations.
The mean number of individual wear data per material was 25 (range 14–42). The mean coef-
ficient of variation of clinical wear data was 53%. The only significant correlation was reached
by OHSU (abrasion) with a Spearman r of 0.86 (p = 0.001). Zurich, ACTA, Alabama general-
ized wear and Ivoclar (volume) had correlation coefficients between 0.3 and 0.4. For Zurich,
Alabama generalized wear and Munich, the correlation coefficient improved if only compos-
ites for direct use were taken into consideration. The combination of different laboratory
methods did not significantly improve the correlation.
Significance. The clinical wear of composite resins is mainly dependent on differences
between patients and less on the differences between materials. Laboratory methods to test
conventional resins for wear are therefore less important, especially since most of them do
not reflect the clinical wear.
© 2012 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved.

∗
Corresponding author at: Bendererstrasse 2, FL-9494 Schaan, Liechtenstein. Tel.: +423 235 3570; fax: +423 233 1279.
E-mail address: siegward.heintze@ivoclarvivadent.com (S.D. Heintze).
0109-5641/$ – see front matter © 2012 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.dental.2012.04.006
962 d e n t a l m a t e r i a l s 2 8 ( 2 0 1 2 ) 961–973
1. Introduction
In restorative dentistry, the wear resistance of restorative
materials is mainly evaluated with laboratory methods that
use an electric device that puts a counterpart object (e.g. sty-
lus) into contact with the material that is to be tested. Some
devices simulate chewing movements that occur in the mouth
[1,2] and some methods use an abrasive medium [3–5]. The
methods differ with regard to the applied force, the mate-
rial type and movement of the counterpart, the number of
cycles, etc. [1]. In 2001 the International Standard Organization
ISO published a technical specification “Wear by two- and/or
three-body contact” describing 8 laboratory methods without,
however, giving further comments on them [6]. Furthermore,
discrepancies in wear rates of the same material using the
same wear method and wear parameters were found [1]. The
variations maybe due to different material batches, light cur-
ing conditions or the wear generating device which does not
always generate reproducible results.
For contemporary resin materials wear of posterior restora-
tions seems to be no longer a clinical problem, also over a
period of ten years or more [7,8] because it is restricted to
contact areas and does not affect the entire restoration with
complete loss of the anatomy as it was the case with the early
composite resins of the eighties. Wear can be measured on
replicas but may not be detected during clinical inspection.
Therefore, the issue of wear should not be over-estimated.
However, wear is important if materials with new monomers
and/or fillers are developed and their wear behavior cannot be
deduced from similar products. In this case, laboratory meth-
ods may help to assess the wear resistance before the material
is evaluated in a clinical trial, whose results will be present
only years after the beginning of the study.
Ideally, the laboratory wear rates of specific materials
should reflect the wear that is measured in vivo with the
same materials. Some methods claim to have a high corre-
lation between in vitro and in vivo wear, but these correlations
are only based on a limited number of restorations with
mostly 2–3 materials [9–11]. In addition, they are linearly
extrapolated from 6-month data [11], the clinical wear was
measured with inadequate and inaccurate methods such as
the Moffa–Lugassy scale [3,4,10] or the correlation analysis
was carried out with pooled clinical data from studies with
different study protocols [12].
To date, no effort has been undertaken to systematically
compare data of laboratory wear methods with clinical wear
data. Such an approach has been hampered by the fact that the
peer-reviewed dental journals contain very few clinical wear
data which were generated by applying a reliable and accu-
rate quantitative wear method [9,13,14]. In the nineties, the
Technologies in Restorative and Caries (TRAC) Research Foun-
dation (formerly Clinical Research Associates CRA) clinically
evaluated many resin materials for both direct and indirect
Class II restorations and for indirect crowns and fixed dental
prostheses.
The purpose of the present study was to correlate these
data with wear data published in the literature gained from the
most common and current laboratory wear methods, which
tested the same materials. Four of these six selected methods
are also part of the ISO Technical Specification [6]. A round-
robin test that involved five of the six wear methods examined
in the present study revealed that the results from one wear
method to another were hardly comparable [15]. The reason
for that is that these laboratory wear methods follow differ-
ent tribological concepts [1]. It can therefore be expected that
the correlation to clinical wear is also different. As the clinical
wear process is a complex mechanism with different tribolog-
ical phenomena occurring at the same time, it can be assumed
that there is no single laboratory wear method capable of
showing a good correlation with clinical wear. However, if two
or three methods that come very close to the clinical wear are
combined, the correlation may be improved.
The following hypotheses were examined:
1. There is a method-dependent correlation between labora-
tory wear methods and the clinical wear rates.
2. If two or three wear methods are combined, the correlation
can be improved.
2. Materials and methods
2.1. Wear in vivo
Clinical wear data were derived from clinical studies, which
the TRAC Research Foundation (formerly CRA) in Provo, USA,
organized and carried out together with general practitioners.
Some of the results were previously reported at AADR/IADR
conferences [16–21].
Wear was mainly analyzed in big molar restorations, some-
times including the build-up of a cusp. A minimum of 30
restorations was placed in 30 subjects. The drop-out rate after
2–3 years was about 10%. For composite materials for indi-
rect use, full-coverage crowns were fabricated in especially
selected dental laboratories.
A non-surface-contact method was devised to determine
the in vivo wear of dental restorations by quantifying the focal
distance of a microscope with long focal length (Olympus
BH-2, Olympus Optical Co. Ltd., Japan) in sharp focus with a
narrow depth of sharpness (about 2 ␮m) at 200× on the sur-
face of an epoxy replica of the restored tooth [22,23]. Custom
software which was initially developed for SEM mapping [24]
provided the following parameters for each sample placed
onto the test apparatus: establishment of a zero elevation
from which all measurements were made, reference points
defining the perimeter of the surfaces to be measured, pre-
cise movement of an x–y stage, determination of the point of
sharpest focus, and acquisition of x–y–z data every 50 ␮m over
the pre-defined occlusal table of the replica of the restored
tooth. The precision of the system was determined by digi-
tizing and fitting the same specimens repeatedly at random
times over a period of one year which resulted in an outcome
of a ±3 ␮m difference over fifty runs performed over a twelve-
month period. This accuracy is comparable to measurements
carried out by Roulet using a measuring stylus [25].
A human check was performed on each separate profile
within each matrix to locate illogical outliers that could form
if the system suffered any unusual vibrations, impact, or
software glitches during its unsupervised runs at night and
 d e n t a l m a t e r i a l s 2 8 ( 2 0 1 2 ) 961–973
over weekends. Six axes of rotation were used along with a
best fit algorithm for the fitting of the two matrices digitized
representing two different points in time. A color-coded dis-
play obtained after digitization was compared very carefully
to the high resolution dies as well as the scanning electron
microscope photographs of the test teeth to make sure the
topography matched perfectly between the three representa-
tions. Two operators worked separately and independently to
perform final adjustments of all the best-fit analyses, and were
required to compare and reach agreement if there were differ-
ences. Knowns were constructed by a firm recommended by
the National Bureau of Standards, and these knowns were dig-
itized with each die set every time samples were set up and
the system was operated to check for system errors. The data
from the knowns were subjected to the fit program as well
as the digitizing programs to validate all steps in the process.
For final analyses, two operators worked independently with
every die set fitted. It was required they achieve a final solution
within ±5 ␮m of each other.
The individual data of each subject were available, includ-
ing age, gender, type of tooth restored, number of restored
surfaces, and wear after each recall interval.
2.2. Wear in vitro
The following laboratory wear methods were selected accord-
ing to their significance and frequency of citation in the
literature [1].
2.3. ACTA-Method
The ACTA-Method has been developed at the Dental College in
Amsterdam (The Netherlands) (ACTA = Academisch Centrum
Tandheelkunde Amsterdam) [26]. Two metal wheels rotate in
different directions with about 15% difference in the circum-
ferential speed while having near contact. The test specimens
(n = 24 and 28, resp.) are placed on the circumference of the
wheel while the other wheel serves as antagonist. The force
with which the two wheels come together is adjusted to 15 N.
The wheels are placed in a slurry of white millet seeds in a
buffer solution. After 50,000, 100,000 and 200,000 cycles the
maximal vertical loss of the test specimens is measured with
a profilometry device.
2.4. Alabama-Method
The Alabama wear method was developed at the University of
Alabama (USA) by Leinfelder and Suzuki and is therefore also
called Leinfelder wear method. The method has been first pub-
lished in 1989 [27] and is a modification of a device that has
been originally designed by Roulet [25]. Several major modi-
fications were made over the years. In the first publication in
1989, a polyethylene tape was used as intermediate substance,
driven by a tape advancing system [27]. The tape was replaced
by a slurry of PMMA beads ten years later [4]. The original force
was 55 N, which was increased to 75 N ten years later. In the
first publication, a stainless steel stylus with 2 mm radius hit
the specimen without rotation. In the new method an addi-
tional 30◦ clockwise rotation was integrated as soon as the
stylus hits the surface of the specimen which was then called
963
“localized wear” (Alabama loc). Additionally, with a new spec-
imen a flat stylus made of polyacetal is brought into contact
with the flat specimen; the wear produced by this approach is
called “generalized wear” (Alabama gen). The materials speci-
mens are incorporated into extracted molars that are trimmed
flat. The stylus for generalized wear is made of polyacetal, the
one for localized wear is stainless steel. The original publica-
tion states that each spring is calibrated with a 200 kg load
cell in conjunction with a universal testing machine prior to
testing, but no data have been reported with regard to the
deviations, the scattering of results and the time intervals
of force measurements or the replacement frequency of the
spring.
2.5. Ivoclar-Method
After processing and before testing, the specimens (n = 8) are
kept dry at a temperature of 37 ◦ C for 24 h. The specimens are
mounted in a chewing simulator that is commercially avail-
able (CS-4, SD Mechatronik, Germany). Antagonists are made
by pressed IPS Empress ceramic (Ivoclar Vivadent) which is
glazed two times at a temperature of 870 ◦ C. The diameter of
the antagonist is 2.4 mm at a height of 600 ␮m. A weight of 5 kg
is put on each vertical bar. The sliding movement is fixed at
0.7 mm. The frequency of the antagonist movement is 1.6 Hz.
A total of 120,000 cycles of unidirectional antagonist move-
ments are carried out. Thermocycling with a frequency of
320/120,000 cycles is included in the wear setting with a tem-
perature difference between 5 ◦ C and 55 ◦ C. After completing
the wear generating procedure, replicas are made with white
super hard plaster (Fuji Superhard Rock, GC Corporation,
Japan). The plaster replicas are analyzed with a commercially
available laser scanner device (Laserscan 3D or etkon es 1,
Willytec, Germany) and the appropriate Match-3-D-software
[28]. The volumetric (Ivoclar vol) as well as the maximal ver-
tical loss (Ivoclar vert) (1% percentile) are calculated by the
software.
2.6. Munich-Method
The Munich-Method has been developed at the Maximilian
University in Munich (Germany) [29,30]. After processing the
specimens (Ø 7.5 mm, n = 8) were kept in Ringer’s solution
for 24 h. After that period the specimens are mounted in
a sliding wear tester (Munich Artificial Mouth) with pneu-
matic cylinders as actuator. The wear tester is configured as
a pin-on-block-design allowing the test specimens to slide
under permanent contact to the spherical antagonist (Degus-
sit aluminum oxide, 5 mm diameter) at a linear distance of
8 mm and a vertical load of 50 N. During the chewing sim-
ulation the specimens are rinsed with distilled water with
a temperature of 37 ◦ C. At 6000, 10,000, 30,000, and 50,000
double cycles (bidirectional forth-and back-movement), repli-
cas are made with white plaster (Fuji Superhard Rock, GC
Corporation, Japan). The volumetric loss of the wear facet is
quantified on the plaster models with the Laserscan 3D device
(Willytec, Germany) described above and converted to vertical
loss.
964 d e n t a l m a t e r i a l s 2 8 ( 2 0 1 2 ) 961–973
2.7. OHSU-Method
This method has been developed at the Oral Health and Sci-
ence University (OHSU) in Portland (USA) [3]. In principle,
enamel cusps are forced into contact with the specimens
through a layer of food like slurry (mixture of poppy seeds
and PMMA beads). The enamel cusps are drilled out of human
upper molars of similar shape giving them a spherical shape of
a diameter of 10 mm. The enamel stylus are polished with 600
grit and 1000 grit silicon carbide and polished with 5 ␮m alu-
minum oxide paste and then ultrasonically cleaned for 1 min.
Before mounting the specimens (n = 10) in the wear simulator,
they are kept in water for 24 h at 37 ◦ C. First the cusp is forced
with a load of 50 N on the surface, sliding across a linear path
of 8 mm producing abrasive wear. At the end of each path,
a static load of 80 N is applied to produce localized attrition
wear. For a whole test sequence 100,000 cycles at 1 Hz with
unidirectional movements are run. The mean vertical loss of
the abrasion and attrition wear facets is measured with a pro-
filometry device at 10 defined tracks. The values of tracks 4–6
correspond to the abrasion wear (OHUS ab) and the tracks 8–9
to the attrition wear facet (OHUS at).
2.8. Zurich-Method
The Zurich wear method has been developed at the Univer-
sity of Zurich (Switzerland) [2]. With a load of 49 N and a
frequency of 1.7 Hz the palatal cusps cut out of similar upper
molars push against the surface of the specimens (n = 6) that
are mounted on a rubber socket at 45◦ angle allowing the
antagonist to glide over the surface of the test specimen. The
test specimens are kept in water with exchanged tempera-
ture according to a thermocycling protocol (3000× 5 ◦ C/55 ◦ C).
After 120,000, 240,000, 640,000, and 1,200,000 cycles of load-
ing the specimens are submitted to a toothbrushing device
with a slurry of toothpaste for 30 min, 30 min, 100 min, and
140 min, resp. [31]. Additionally, at the end of the first phase
(120,000 cycles) the specimens are put into a solution of 75%
ethanol for 20 h to simulate the chemical degradation. After
each thermomechanical sequence, the maximal vertical loss
of both the specimens (occlusal contact area OCA) and the
antagonists as well as the vertical loss in the contact free area
(CFA) are calculated by using a computerized 3D-scanner [32].
The scanner is driven by step motors which scan the object
in 1 ␮m steps in the z-direction and in 100 ␮m steps in the
xy-direction [25,1–5,30,31].
The wear methods with their typical parameters are listed
in Table 1.
Four variables were used in conjunction with the Zurich
wear method: Phase 1 after 120,000 loading cycles, phase 2
after 240,000 cycles, phase 3 after 640,000 cycles and phase
4 after 1,200,000 cycles. These values were claimed to corre-
spond to six months, one year, 2.7 years and 5 years of clinical
wear, respectively [11]. For the Alabama method two variables
had been included in the analysis: generalized and localized
wear; for the OHSU method two different wear zones were
generated: abrasion and attrition. For the Ivoclar method the
vertical as well as the volumetric wear was taken into consid-
eration. A single variable was available for ACTA and Munich
(vertical loss after completion of all cycles).
A systematic search was carried out by using the databases
of MEDLINE (PUBMED, since 1966) and IADR (abstracts, starting
1984). The search month was March 2008. The search terms
were “wear” and/or “abrasion”, “in vitro” and the name or
acronym of the respective wear method. The wear data were
only used if the identical wear method parameters listed in
Table 1 were applied. As far as enamel wear is concerned, the
test parameters were different to those applied for compos-
ite or amalgam materials for the Alabama and OHSU method.
The studies included in the comparative analysis are listed in
Table 2.
Although the test parameters are identical, the wear data
may be influenced by the fact that different research institutes
carried out the tests. Therefore, the studies were grouped into
two groups: PUB 1 = from the same institute, namely the ones
who first published the method; PUB 2 = other institutes.
The materials were classified into 4 categories:
CAT 1: composites for intra-coronal restorations (direct and
indirect);
CAT 2: composites for crowns and bridges;
CAT 3: amalgam;
CAT 4: enamel.
2.9. Statistical analysis
In order to summarize the information of clinical measure-
ments, a single index was created for each material. Using
an average value of the measurements for each material,
however, would not provide a fair comparison between the
materials because of the unbalanced design and the increase
of wear over time. Instead, a linear mixed model with a fixed
material effect, a fixed time effect and a random patient effect
was considered as follows:
Yijk = mi + tj + pk + εijk
where Yijk denotes the logarithm of in vivo measurement of
material i at time j for patient k, mi is the effect of material i,
tj is the effect of time j, pk is the effect of patient k and εijk is
an error term; pk and εijk were assumed to be independent and
normally distributed. The age of the patients could not be con-
sidered, because this information was missing for about 43%
of the patients. A logarithm transformation was necessary to
reach normality. Calculating exp(effect) − 1 provided us with
an estimate of the percentage of median wear increase on the
original scale. Estimates of material effects mi were then used
as a summarizing in vivo index, to be correlated with the var-
ious in vitro variables. Further covariates such as gender of
the patients or tooth type were also included in the model.
Linear mixed models were calculated using the lme routine
implemented in the package nlme (version 3.1–83) from the
free statistical package R.
In order to be closer to a normal distribution, in vitro
variables were log-transformed, as suggested in a previous
publication where data of different laboratory methods were
compared [15]. To facilitate comparison, the log-values x of
each variable were further transformed into relative ranks, cal-
culated as 1 + (n − 1)(x − m)/(M − m), where n was the number of
materials considered (n = 28), m was the minimum (mean) and
Table 1 – Six laboratory wear methods and their wear generating parameters.
Wear method Force Force actuator Movement of stylus Stylus material Abrasive medium Number of cycles Outcome variables
ACTA 15 N Spring Wheel runs Metal wheel, slip rate Millet seed shells/rice 200,000 Vertical wear (␮m)

d e n t a l m a t e r i a l s 2 8 ( 2 0 1 2 ) 961–973
against 15%
specimens
Alabama 75 N Spring 30◦ rotation Polyacetal PMMA beads 400,000 Localized wear (␮m)
(Ø 2.5 mm)
Polyacetal Generalized wear (␮m)
(Ø 6.5 mm)
Ivoclar 5 kg Dead weight 0.7 mm Leucite reinforced Water 120,000 Vertical wear (␮m)
driven with step ceramic (Empress)
motor (Ø 2.4 mm) Volumetric wear
(107 ␮m3 )
Munich 5 kg Dead weight driven 8 mm Aluminum oxide Water 50,000 Vertical wear (␮m)
with pneumatic ceramic (Degussit)
cylinder (Ø 5 mm)
OHSU 20 N abrasion Electro-magnetic 6 mm Enamel Poppy seeds/PMMA 50,000 Mean vertical wear (␮m)
80 N attrition (Ø 10 mm) beads Abrasion
Attrition
Zurich 49 N Electro-magnetic Sliding Enamel Water 120,000 Vertical wear (␮m)
movement of (Ø 5 mm) 240,000
specimen 640,000
1,200,000

965
966 d e n t a l m a t e r i a l s 2 8 ( 2 0 1 2 ) 961–973

Table 2 – References from where the wear data were retrieved for the in vitro/in vivo comparison.
ACTA Alabama Ivoclar Munich OHSU Zurich
[10,15,26,42,43] [4,9,44–51] [15,52]. [53] [3,54–60] [15,61,62]

M was the maximum in vitro value (mean). The scale of each
variable was hence transformed into a “rank scale” from 1 to K
(with all variables being directly comparable), while the rela-
tive differences between the materials was taken into account.
To evaluate whether the combination of different labo-
ratory methods improves the correlation with in vivo data,
stepwise linear regression modeling was applied (SPSS release
16.0).
3. Results
The individual clinical wear data of 28 materials (including
enamel) for which also in vitro data of one or more labora-
tory wear methods were present were included in the analysis.
300
The mean age of the subjects for whom age data were avail-
able (n = 835) was 40 (±12) years and there were almost on
average twice as many females as males in each material
study (944 females, 540 males). The materials were mostly
applied in molars (1199 molar and 285 premolar restora-
tions). Almost the same number of restorations was present
in the maxilla and the mandible (766 versus 718). Furthermore,
96% of the intracoronal restorations were Class II restora-
tions and only about 4% were Class I restorations; 2-surface
restorations were more frequent than 3- or 4-surface restora-
tions (567 2-surface, 358 3-surface, 212 4-surface); replicas of
253 full-coverage crowns were available for analysis. The mean
number of individual wear data per material was 25 (range
14–42) per material group. The mean coefficient of variation
of wear data (1 year) was 53.7% (±14.8).

● Nulite F
250
200
average µm

●
● Targis ●
150

●
● ●
● ●
●
●
● ●
100

●
●
● ●
● ●
● ●
● ●
●
●
● ●
●
● ●
● ● Enamel
● ● Heliomolar
50

● ●
●
●
● ● SureFil
●
●
●
●
● Z250 ● ● Dispersalloy
●
● ●
●
● ●
● ●
●

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

time (years)

● Nulite F
5.5

●
5.0

●
●
● Targis ●
● ● ●
●
●
● ●
average log µm
4.5

●
●
● ●
● ●
●
● ●
● ●
● ●
●
● ● Enamel
4.0

●
●
●
●
●
● SureFil
●
●
● ● ● Heliomolar
● ●
3.5

●
● ●
● ● Z250 ●
●
● ● Dispersalloy
●
●
3.0

●
●
● ●
2.5

●
●

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

time (years)

Fig. 1 – Average in vivo wear measurements over time using the original scale (upper panel) or the log scale (lower panel).
Each line represents one material. Composites for intracoronal restorations, composite crowns, amalgam and enamel are
represented in red, blue, green and black, respectively.
 d e n t a l m a t e r i a l s 2 8 ( 2 0 1 2 ) 961–973
The number of materials for which in vitro data were avail-
able was 22 for Ivoclar, 20 for Zurich, 17/14 for Alabama
generalized/localized wear, 12/13 for OHSU abrasion/attrition,
12 for ACTA and 7 for Munich. In vivo wear data were avail-
able after six months (19 materials), one year (28 materials),
two years (27 materials) and three years (9 materials). We did
not consider in vivo wear data after more than three years
because they were missing for almost all materials. Among
the 28 materials, 21 belonged to CAT 1 (intracoronal compos-
ites), 5 to CAT 2 (composites for full coverage crowns: Artglass,
belleGlass, Sculpture, Solidex, Targis), 1 to CAT 3 (amalgam:
Dispersalloy) and 1 to CAT 4 (enamel).
3.1. Analysis of in vivo measurements
The average in vivo wear measurements along a time
axis are shown in the upper panel of Fig. 1, each line
representing one material. It is noticeable that both the level
of wear and the between-material variability were monoton-
ically increasing over time. The lower panel of Fig. 1 shows
the average in vivo wear measurements calculated on the
log-scale, where the between-material variability had been
stabilized, which facilitated data modeling. Note in particular
the nearly parallel increase of wear between the materials, an
indication that the effect of time was the same for each mate-
rial, and that a ranking of materials would be approximately
the same at each point in time.
A linear mixed model was fitted to the log wear mea-
surements, using the following factors: material (28 levels),
time (4 levels: 0.5, 1, 2 and 3 years), tooth (2 levels: premolar
and molar) and gender (2 levels) as fixed effects and patient
as random-effect. The effect of time was of course strongly
significant (p < 0.0001), with the percentage of median wear
increase being estimated to exp(0.62) − 1 = 86% between six
months and one year, to exp(0.49) − 1 = 63% between one year
and two years, and exp(0.37) − 1 = 45% between two and three
years. Overall, the factor ‘material’ was significant (p < 0.0001).
In line with Fig. 1, the best material with the lowest wear
was Dispersalloy, followed by Heliomolar, Z250 and SureFil,
whereas the worst material was Nulite F, followed by Targis,
Prodigy and Estilux. To summarize the in vivo performance
of the different materials, we calculated relative ranks from
the estimated coefficients of the factor ‘material’ in the lin-
ear mixed model, which are provided in the last column of
Table 1.
3.2. Correlation between in vitro and in vivo wear
Relative ranks for 9 wear variables are given in Table 3.
For the Zurich method, we considered an average of the
different Zurich time variables (corresponding to 0.5-year-
, 1-year-, 2.7-year- and 5-year equivalents). Scatterplots of
these relative ranks against the relative ranks of in vivo
performance are shown in Fig. 2. Spearman correlations
between in vitro and in vivo performance, together with asso-
ciated p-values, are provided. The correlation was negative
for the Alabama localized wear method, and was between
0.29 and 0.44 for the ACTA method, the Alabama generalized
wear method, the Ivoclar method (vertical and volumetric)
as well as for the Zurich and Munich method, with the
967
latter being calculated from only 7 materials. On the other
hand, the correlation was significantly positive for the OHSU
method, with the highest correlation being found for OHSU
abrasion (r = 0.86, p = 0.001). For the Zurich and Alabama gen-
eralized methods, the coefficient improved if only composites
for intracoronal Class II restorations were included versus
all materials (enamel, amalgam, and composited for full-
coverage crowns) (see Fig. 2). Also, the correlation slightly
increased for the Alabama localized and generalized wear
methods when only the results for materials published by the
same group were included (from −0.14 to 0.09 and from 0.40
to 0.49), while the correlation slightly decreased for the ACTA
method (from 0.43 to 0.36). For the Zurich variables, the cor-
relation slightly decreased over time (r = 0.49 at 0.5 year- and
at 1-year equivalent, r = 0.41 at 2.7 year- and r = 0.37 at 5-year
equivalent).
3.3. Ranking of materials with respect to in vitro
measurements
To summarize the in vitro performance, we calculated the aver-
age of the available relative ranks (AARR) calculated over six
in vitro variables, one for each method (ACTA, Alabama gener-
alized wear, Ivoclar vertical attrition, Munich, OHSU abrasion
and Zurich average). Values of AARR are provided in Table 1.
The best overall in vitro performance was achieved by enamel
(AARR of 7.4) followed by Dispersalloy (AARR of 8.5) and
SureFil (AARR of 8.5). The worst in vitro performance was
achieved by Nulite F and Prisma AP.H, although both materials
were measured by a single method only.
The AARR in vitro index was useful to investigate fur-
ther questions in an economical way. The global in vitro
performance of materials of CAT 1 and CAT 2 were, for exam-
ple, compared and this comparison showed that the average
AARR was 19.3 for the former and 14.0 for the latter. In a
Mann–Whitney test, this difference was almost significant
(p = 0.05).
When different laboratory methods were combined with
each other, the correlation between in vitro and in vivo relative
ranks of the materials was not significantly improved, which
was shown by the stepwise regression analysis (Table 4).
4. Discussion
This is the first study which systematically evaluated the rela-
tionship between laboratory data on wear and clinical wear
rates.
In the past, some efforts were made to correlate in vitro and
in vivo data, but those studies were marred by a number of
shortcomings related to an inadequate methodology of in vivo
wear measurements, inadequate statistical methods and/or a
limited number of materials [4,9–11]. From the early seventies
to the present date, the clinical wear of restorative materi-
als was subjectively evaluated directly on the patient together
with other clinical parameters using (modified) USPHS crite-
ria. USPHS criteria, however, are by far not an adequate tool
to measure wear [33,34]. In the eighties and early nineties,
clinical wear used to be quantified by comparing cast repli-
cas with a set of standards, known as scales. Two evaluators
 968
Table 3 – Relative ranks (RR) characterizing in vitro performance for 28 materials according to six methods (9 variables). Relative ranks characterizing in vivo
performance calculated via a linear mixed model. The first 21 materials are direct composite materials (CAT 1).
Material ACTA RR Alabama Alabama Ivoclar vert Ivoclar vol Munich RR OHSU ab RR OHSU at RR Zurich RR in vivo RR
loc RR gen RR RR RR
Adaptic 2 23.4 26.7 17.3
Alert 5.3 12.2 15.4 21.2 10.7 7.1 20.8 17.5
Brilliant 24.4 13.8 20 1 15.6
Charisma 17.4 20.9 28 14.5 19.8 14 15.6
Concept 1.1 14.7 14.8 6.8 1.6 17.8 14.1

d e n t a l m a t e r i a l s 2 8 ( 2 0 1 2 ) 961–973
Estilux 26.6 25.2 13.5 20.2 19.6
Heliomolar 28.0 17.2 1.0 1.0 1.0 6.6 10.9 12.7 7.8
Herculite 25.5 17.8 20.9 18.4 16.6
Nulite F 28.0 28.0
Occlusin 25.0 23.4 26.3 16.6 18.5
Pertac Hybrid 20 24.7 28 17.5 18.3
Prisma AP.H 28 18.8
Prisma TPH 16.4 20.4 25.8 18.2
Prodigy 10.9 16.3 16.5 20.5
Renamel Hybrid 15.4 18.4 17.0
Solitaire 17.2 20.4 20.2 24.7 4.1 19.7 20.1 15.4 16.6
SureFil 1.0 1.0 10.6 11.4 16.6 7.4 5.6 12.2 11.9
Tetric 13.4 27.4 28 24.2 14.9 14.1
Tetric Ceram 24.5 25.4 25.6 24.2 19.2
Z100 23.7 17.9 9.6 28.0 27.4 20.3 15.5 3.9 14.4 13.8
Z250 16.6 18.3 23.3 24 10.6

Artglass 14.5 13 23.8 23.8 12.7 28.0 28.0 18.7

belleGlass HP 20.7 5.5 1.0 8.6 13.2 11.8 17.3
Sculpture 2.9 1.4 20.5 22.2 14.5
Solidex 6.9 12.2 10.7 14.7 16.2
Targis 24.5 12.5 16.5 20.7 21.6 10.0 20.6

Dispersalloy amalgam 14.9 16.8 5.1 2.2 5.9 1.0 1.0 19.1 1.0
Enamel 8.3 28.0 7.8 10.6 11.6 6.3 3.1 13.1
 d e n t a l m a t e r i a l s 2 8 ( 2 0 1 2 ) 961–973 969

r=0.43 p=0.17 (cat1: r=0.24 p=0.58) r=−0.14 p=0.64 (cat1: r=0.14 p=0.78) r=0.4 p=0.11 (cat1: r=0.57 p=0.09)

Heliomolar Enamel Prisma AP.H

25

25

25
Estilux
Herculite
Occlusin
Tetric
Brillant Targis
Ceram
Z100 Adaptic 2
Occlusin

relative rank alabama gen

relative rank alabama loc
20

20

20
belleGlass HP Solitaire
relative rank acta

Z100
Herculite Z250
Dispersalloy Z250 Solitaire Heliomolar
15

15

15
Targis
Dispersalloy Artglass
Tetric Artglass
Targis Alert
Solidex
10

10

10
SureFil
Z100
Enamel Enamel
Solidex
5

5
belleGlass
Alert HP Dispersalloy
Sculpture
Sculpture
0

0
SureFil Concept
SureFil belleGlass HP

0 5 10 15 20 25 0 5 10 15 20 25 0 5 10 15 20 25
relative rank in vivo relative rank in vivo relative rank in vivo
r=0.29 p=0.19 (cat1: r=0.06 p=0.83) r=0.31 p=0.16 (cat1: r=0.16 p=0.55) r=0.29 p=0.53 (cat1: r=0.61 p=0.2)

Z100 Tetric Pertac Hybrid

Charisma
Tetric Z100
25

25

25
Tetric Ceram Tetric Ceram
Pertac
SolitaireHybrid Tetric
Artglass Z250 Artglass
Z250
Sculpture Targis
relative rank ivoclar vert
20

20

20
relative rank ivoclar vol

Alert
Charisma

relative rank munich

Sculpture Targis Prisma TPH Z100
Solitaire
Pertac Hybrid Brillant
Renamel Hybrid
Charisma
15

15

15
Prisma TPH SureFil Prodigy
Renamel
Concept Hybrid
Alert Concept
Solidex
Brillant
belleGlass HP Artglass
10

10

10
SureFil Enamel
SolidexProdigy
Enamel
belleGlass HP
5

5
Dispersalloy
Solitaire
Dispersalloy
0

Heliomolar Heliomolar 0 Heliomolar

0 5 10 15 20 25 0 5 10 15 20 25 0 5 10 15 20 25
relative rank in vivo relative rank in vivo relative rank in vivo
r=0.86 p=0 (cat1: r=0.78 p=0.01) r=0.66 p=0.02 (cat1: r=0.44 p=0.2) r=0.44 p=0.06 (cat1: r=0.71 p=0)

Artglass Artglass Nulite

Adaptic 2
25

25

25

Occlusin Prisma TPH

Estilux
Tetric Ceram
20

20

20
relative rank ohsu ab

relative rank ohsu at

Herculite Alert
relative rank zurich

Solitaire Solitaire
Charisma Estilux
Dispersalloy
Herculite Concept
Pertac Hybrid
15

15

15

Occlusin Prodigy
Z100 Solitaire
Tetric
Charisma Z100
Charisma
Estilux
Heliomolar
SureFil
belleGlass HP
10

10

10

Alert Heliomolar
Targis

SureFil Alert
Heliomolar Concept
5

Enamel
SureFil
Z100
Enamel
Concept
0

Dispersalloy Dispersalloy Brillant

0 5 10 15 20 25 0 5 10 15 20 25 0 5 10 15 20 25
relative rank in vivo relative rank in vivo relative rank in vivo

Fig. 2 – Relative ranks characterizing in vitro performance according to 6 methods (9 variables) versus relative ranks
characterizing in vivo performance calculated via a linear mixed model, together with Spearman correlation r and associated
p-value (in parentheses, restricted to materials of CAT 1). Each dot represents one material. Composites for intracoronal
restorations, composite crowns, amalgam and enamel are represented in red, blue, green and black, respectively.

would compare the replica with the standards by means of time. (1) The Leinfelder scale [35], which used 6 calibrated die
loupes and assign a wear value. The scales were based on stone standards from clinical restorations, exhibiting approx-
the concept that material loss at the restoration margin was imately 100–500 ␮m of occlusal loss; (2) the Moffa–Lugassy
indicative of the loss of material over the entire restoration (M–L) scale [36], which used 18 standard dies with cylindri-
surface. Mainly three different scales were propagated at that cal incremental defects ranging from 25 ␮m to 1000 ␮m; (3)
970 d e n t a l m a t e r i a l s 2 8 ( 2 0 1 2 ) 961–973
Table 4 – Stepwise regression analysis of different combinations of wear methods with adjusted R2 and p-value.
Method Adjusted R2
OHSU abrasion 0.68
Alabama generalized 0.42
Ivoclar 0.31
Munich 0.26
ACTA 0.14
Alabama localized 0.01
OHSU attrition 0.36
Ivoclar + Alabama + OHSU 0.79
Ivoclar + OHSU ab 0.68
Ivoclar + Alabama generalized 0.42
Ivoclar + Zurich 0.43
OHSU abrasion + Alabama generalized 0.72
the Vivadent scale (modification of M–L scale developed by
V. Rheinberger, Ivoclar Vivadent, Schaan), which used tooth-
sized dies with restoration-like incremental defects. All scales
involved problems with internal and external validity. By
applying the standard scales, the agreement among differ-
ent evaluators can vary tremendously, especially amongst
inexperienced evaluators [37]; the Vivadent scale, however,
consistently achieved the highest level of agreement. Yet, even
the Vivadent scale was shown to lacking in accuracy, as it was
proven that the actual wear is systematically underestimated
when the results obtained with the Vivadent scale were com-
pared with those obtained with sophisticated laser equipment
[38].
By using the Leinfelder scale, the correlation coefficient
Spearman r between the clinical wear results and those gen-
erated with the Alabama method was 0.81 as calculated by the
author based on original data [4]. High correlation coefficients
(>0.8) were also reported for the OHSU method in conjunction
with clinical studies that used different methods to quantify
the wear in vivo [3]; however, the correlations were based on
only 2–3 resin materials.
For the ACTA method, in vitro/in vivo comparisons were car-
ried out using the clinical data of 7 different studies, which
used different methods to quantify wear at different time
intervals; the correlation coefficients were between 0.81 and
0.99 [10]. However, the number of materials was again only
between 2 and 4. Performing a comparative analysis based
on such a low number of materials or even on as few as
two is meaningless. A similar attempt using ACTA results
was made by another study, which reported correlation coeffi-
cients between 0.88 and 0.92 when the results of 7 composite
resins were compared [12]. In this study, the laboratory results
were related to pooled clinical results, which came from dif-
ferent studies.
In another comparative study, no correlation was found
between the results obtained with a laboratory method using
a slider-on-disc wear apparatus on 7 composite resins mate-
rials and an amalgam and the results obtained with the same
materials in 6 posterior restorations per material and follow-
ing them up for 6 months in vivo [39]. The in vivo replicas were
measured with a modified profilometer, which measured the
material loss in the central fossa by defining profiles relative
to reference points drilled in the cusps [40].
In the present study, the clinical wear was determined by
a single research institute only, following always the same
Number of materials p-Value
12 0.001
17 0.005
22 0.007
7 0.24
12 0.22
14 0.76
13 0.03
7 0.15
9 0.03
14 0.05
17 0.02
8 0.04
protocol for the placement of the restorations and the wear
quantification on replicas. It can therefore be assumed that
the wear rate of the different materials can be compared to
a certain degree. The wear was determined on replicas with
a method that uses a computerized microscope with narrow
focus adjustment. The accuracy of the method is adequate for
the research purpose.
It cannot be excluded that the wear rates might have been
different for some materials, if the material had been tested in
another population group. In the present study, generally one
material was placed per test person. A common method to
exclude the patient factor is to place restorations of different
materials in the same patient in what is known as the split-
mouth design. However, the number of materials that can be
inserted is limited to 2–4.
The variability of the clinical wear results was very high;
the mean coefficient of variation was 53%. This means that the
test person himself or herself plays a major role in the clinical
wear performance of a material. However, differences in the
anatomical shape of the restorations, which may vary from
one test person to another or from one operator to another,
can also render a restoration more prone to occlusal wear
than another restoration. For 17 of the 28 materials, the mean
clinical wear after two years was in a relatively narrow range
between 70 ␮m and 130 ␮m (see Fig. 2); due to the high vari-
ability of clinical results, these materials are statistically not
significantly different, if ANOVA with a post hoc test to adjust
for multiple comparisons is applied.
The high correlation coefficient published in the litera-
ture for certain methods was only confirmed for the OHSU
method, not, however for the Alabama and ACTA method. A
possible reason for this is that those studies included only
few materials, which may not be representative of the whole
diversity of the materials considered in the present study.
Some of the correlations increased and some decreased when
only composites for intracoronal Class II restorations versus
all materials (enamel, amalgam, and composites for indirect
use) were tested. The improvement, however, is not clinically
relevant. Furthermore, it has to be taken into account that the
selection of materials was not identical in all comparisons of
laboratory methods and clinical wear data for the comparison
of the laboratory methods with the clinical wear data. If one,
e.g. includes for the Ivoclar method only those materials that
were also included for the OHSU abrasion method, the correla-
tion coefficient improves for the Ivoclar method. On the other
 d e n t a l m a t e r i a l s 2 8 ( 2 0 1 2 ) 961–973
hand, if one includes only those materials that were evaluated
with all laboratory methods, one ends up with only 2 materials
(Heliomolar and Z100), which were evaluated by all methods.
Although there are some similarities between the OHSU
and the Alabama method, such as the use of an abrasive slurry,
the relatively low force between the specimen and the stylus
and the big diameter of the stylus (OHSU 10 mm, Alabama
6.5 mm), only the OHSU method showed a strong correla-
tion with the clinical wear index. For the clinical wear the
entire occlusal surface is measured including attrition zones
(occlusal contact areas, OCA) and contact-free areas (CFA).
Therefore, the clinical wear rate represents the mean wear
across the entire occlusal restoration and it is conceivable that
the OHSU abrasion method comes close to what is measured
in vivo.
When the wear of individual composite materials is con-
sidered, the microfilled material Heliomolar (light-curing)
demonstrated the lowest clinical wear, followed by Z250 and
the packable material SureFil. The worst material was Nulite F,
which contains large fillers, followed by Prodigy, Targis; mate-
rials such as Tetric Ceram, Herculite, and Pertac Hybrid were
in between (see Fig. 1). None of the materials reached the low
wear rates of the amalgam Dispersalloy.
The laboratory methods showed quite different results
when they were compared to the clinical wear. Heliomo-
lar produced high wear rates in conjunction with the ACTA
and Alabama methods, while it showed low rates when sub-
jected to the Ivoclar, Munich and OHSU abrasion method
(see Table 3). When all laboratory methods were summed up,
the lowest available relative ranks (AARR) index was 7.4 for
enamel, followed by Dispersalloy (8.5) and SureFil (8.5), illus-
trating the fact that the various methods did not agree with
each other. On the opposite side of the ranking, the highest
AARR was obtained for Nulite F (28), Prisma AP.H (28), Adaptic
2 (25.1), Occlusin (24.9) and Tetric Ceram (24.7); the first three
materials were, however, only tested by 1 or 2 wear methods.
These materials contain fillers larger than 2 ␮m.
For both the clinical trials and some laboratory tests, there
was a tendency for microfilled composites to show lower wear
rates compared to composites with large fillers. Large fillers
can easily be removed and act as an abrasive medium, which
further accelerates the wear of the material. However, there
are materials such as Z100 and Tetric Ceram whose filler par-
ticles have a similar dimension but showed different wear
results in vivo and in vitro for the Zurich method but not for
the Ivoclar or ACTA method.
There are obviously many other material-related parame-
ters that come into play if a manufacturer intends to develop
a material with high wear resistance. First of all, the size of
the fillers reflects only the mean particle size. In most mate-
rials, however, the distribution of the filler particles does not
correspond to a normal distribution but rather to a distorted
one. Therefore, the mean particle size does not give valid
information about the amount of small, medium-sized and
large particles. This, however, may affect the resistance of the
resin to wear-induced attrition. Other important factors are
the shape of the filler particles, the distance between them,
the composition of the resin matrix, the chemical link between
inorganic fillers and the resin matrix and the conversion rate
after polymerization [25,41].
971
Other factors worth considering are related to the wear
methods themselves. Many of these methods involve devices
that are not qualified for the purpose to generate wear in
a standardized and reproducible way [1]. The force which
the actuators exert on the specimens is not regulated and
controlled; the same holds true for the distance at which
the stylus slides over the specimen. For the methods that
use an abrasive medium (ACTA, Alabama, OHSU), problems
arise with regard to the homogenous distribution of particles
between stylus specimen, thickening of the slurry because
of the evaporation of water and the impossibility to stan-
dardize organic components such as millet or poppy seeds.
A systematic review produced evidence that many of these
methods lack reproducibility, as the wear data of the same
composite material tested by the same authors using the same
wear method and the same wear parameters differed as much
as 72% from one publication to another [1]. Therefore, the
present study investigated the fact whether the correlation
with in vivo data differs if only data from the same research
group are taken or if all available data that used the same
wear method and parameters but are published by different
research groups are analyzed. There were, however, only slight
influences and for three methods (Ivoclar, Munich, Zurich) all
data came from the same research group or even the same
publication.
The combination of the results of different methods did
not significantly improve the correlation. Based on the lim-
ited amount of data available to us, we found no evidence
that it would be worth to install different wear methods in the
laboratory. Another question arises: How should the results of
a certain material be interpreted if they differ so largely from
one method to the other?
One of the conclusions of the present study is that only
one laboratory method showed a sufficient correlation with
the clinical wear data. The results of all the other methods
were only poorly correlated with the clinical data. Therefore,
one does not have to over-emphasize the results, and a high
correlation due to the heterogeneity of the population groups,
measurement errors, choice of criteria, wear measurement
methods, etc. could not be expected. Nevertheless, even if
laboratory wear methods do have limitations, they may be
able to roughly categorize a new material as to whether it
will probably exhibit a high or low wear rate or rank them
under same testing conditions. Furthermore, the between-
subject variability of wear rates is very high and limits also the
meaningfulness of clinical wear data, especially if the mean
differences between different materials are small. However,
there are materials with relatively low wear and those with
high wear. Therefore, laboratory methods should be able to
identify those materials and the method should be able to
show low wear of a standard material that also had low wear
in clinical trials, e.g. as is the case for amalgam.
5. Summary and conclusions
- Estimates of random effects in a linear mixed model pro-
vided a fair summary of in vivo performance of materials,
taking into account the unbalanced design, while using
the whole information available. The average of available
972 d e n t a l m a t e r i a l s 2 8 ( 2 0 1 2 ) 961–973
relative ranks over the various methods provided a useful
summary of laboratory wear. These indices could be used
to test hypotheses in an efficient way.
- The correlation between the various laboratory methods
and the in vivo results was different for the various methods.
As a whole, however, the correlation was significant only
for one method (OHSU abrasion) and weak or non-existent
for the other methods. Therefore, the first two hypotheses
were partially confirmed. The high correlations published
for certain methods were not confirmed in the present study.
- For the Zurich, Alabama generalized wear and Munich
methods, the correlation coefficient improved if only com-
posites for direct use were taken into consideration.
- The combination of different laboratory methods did not
significantly improve the correlation, thus rejecting the sec-
ond hypothesis.
- There is still a need for clinical wear studies as composites
with new monomers that reduce polymerization shrinkage
or provide biological effects (antibacterial and/or reminer-
alization enhancing effects) or composites with new fillers
may have wear patterns that completely differ from the
wear of the composites used thus far.
Acknowledgements
The authors are very grateful to Dr. Rella Christensen from
TRAC Research Institute (formerly CRA) who generously
provided the in vivo wear data. Without those data, the com-
parison of clinical wear and laboratory wear methods would
not have been feasible.
Furthermore the authors want to thank PD Dr. Till U.
Göhring, Zurich, for additional information about the results
of the Zurich wear testing.
references
[1] Heintze SD. How to qualify and validate wear simulation
devices and methods. Dental Materials 2006;22:712–34.
[2] Krejci I, Reich T, Lutz F, Albertoni M. In Vitro—Testverfahren
zur Evaluation dentaler Restaurationssysteme 1
Computergesteuerter Kausimulator. Schweizer
Monatsschrift für Zahnmedizin 1990;100:953–60.
[3] Condon JR, Ferracane JL. Evaluation of composite wear with
a new multi-mode oral wear simulator. Dental Materials
1996;12:218–26.
[4] Leinfelder KF, Suzuki S. In vitro wear device for determining
posterior composite wear. Journal of the American Dental
Association 1999;130:1347–53.
[5] De Gee AJ, Pallav P, Davidson CL. Effect of abrasion medium
on wear of stress-bearing composites and amalgam in vitro.
Journal of Dental Research 1986;65:654–8.
[6] ISO. Dental materials—guidance on testing of wear. Part 2:
wear by two-and/or three body contact. Technical
Specification. 2001; No. 14569-2.
[7] Pallesen U, Qvist V. Clinical evaluation of three posterior
composite resins: 20-year report. Journal of Dental Research
2005;84 (Special Issue B): Abstract No. 145.
[8] Ferracane JL. Is the wear of dental composites still a clinical
concern? Is there still a need for in vitro wear simulating
devices? Dental Materials 2006;22:689–92.
[9] Barkmeier WW, Latta MA, Erickson RL, Lambrechts P.
Comparison of laboratory and clinical wear rates of resin
composites. Quintessence International 2004;35:
269–74.
[10] Pallav P, Davidson CL, de Gee AJ. Wear rates of composites,
an amalgam, and enamel under stress-bearing conditions.
Journal of Prosthetic Dentistry 1988;59:426–9.
[11] Krejci I, Lutz F. In Vitro—Testverfahren zur Evaluation
dentaler Restaurationssysteme Korrelation mit In
vivo—Resultaten. Schweizer Monatsschrift für Zahnmedizin
1990;100:1445–9.
[12] Finger W, Thiemann J. Correlation between in vitro and
in vivo wear of posterior restorative materials. Dental
Materials 1987;3:280–6.
[13] Lambrechts P, Vuylsteke M, Vanherle G, Davidson CL.
Quantitative in vivo wear of posterior dental restorations:
four-year results. Journal of Dental Research 1985;62:370.
Abstract No. 1759.
[14] Söderholm KJ, Lambrechts P, Sarrett D, Abe Y, Yang MC,
Labella R, et al. Clinical wear performance of eight
experimental dental composites over three years
determined by two measuring methods. European Journal of
Oral Sciences 2001;109:273–81.
[15] Heintze SD, Zappini G, Rousson V. Wear of ten dental
restorative materials in five wear simulators—results of a
round robin test. Dental Materials 2005;21:304–17.
[16] Christensen RP, Christensen GJ, Bangerter V. 3-year in vivo
comparison of 7 posterior resins with silver amalgam.
Journal of Dental Research 1987;66:211. Abstract
No. 833.
[17] Christensen RP, Christensen GJ, Hein DK. 2 year clinical
performance sumary on 21 direct and indirect Class 2
materials. Journal of Dental Research 1994;73:180. Abstract
No. 628.
[18] Hein DK, Smith SL, Cordner CI, Christensen RP. 1-year
clinical performance of six direct Class 2 resins. Journal of
Dental Research 1995;74:165. Abstract No. 1225.
[19] Smith SL, Hein DK, Morrow TA, Cordner CI, Woods DR,
Christensen RP. 2-year clinical performance of eight direct
Class 2 restorative materials. Journal of Dental Research
1996;75:256. Abstract No. 1907.
[20] Smith SL, Hein DK, Aina T, Woolf SH, Christensen RP. 2-year
clinical performance of six direct Class 2 restorative
materials. Journal of Dental Research 1999;78:309. Abstract
No. 1632.
[21] Hein DK, Tolley JM, Wagstaff JS, Christensen RP. 2-year
clinical performance of 6 direct Class 2 resins. Journal of
Dental Research 2003;82 (Special Issue A): Abstract No. 35.
[22] Bangerter V, Christensen RP, Christensen GJ. Method for
determination of in vivo wear. Journal of Dental Research
1987;66:126. Abstract No. 158.
[23] Christensen R, Bangerter V. Apparatus for automated,
non-contact measurements of surface changes. Journal of
Dental Research 1990;69:126. Abstract No. 140.
[24] Russ JC, Hare TM, Christensen RP, Hare KT. SEM low
magnification stereoscopic technique for mapping surface
contours: application to measurement of volume
differences in human teeth due to polishing. Journal of
Microscopy 1986;144:329–38.
[25] Roulet J-F. Degradation of dental polymers. Basel: Karger;
1987.
[26] De Gee AJ, Pallav P. Occlusal wear simulation with the ACTA
wear machine. Journal of Dentistry 1994;22:S21–7.
[27] Leinfelder KF, Beaudreau RW, Mazer RB. An in vitro device
for predicting clinical wear. Quintessence International
1989;20:755–61.
[28] Mehl A, Gloger W, Kunzelmann KH, Hickel R. A new optical
3-D device for the detection of wear. Journal of Dental
Research 1997;76:1799–807.
 d e n t a l m a t e r i a l s 2 8 ( 2 0 1 2 ) 961–973
[29] Kunzelmann K-H. Verschleissanalyse und—quantifizierung
von Füllungsmaterialien in vivo und in vitro. Aachen:
Shaker Verlag; 1998.
[30] Zantner C, Kielbassa AM, Martus P, Kunzelmann KH. Sliding
wear of 19 commercially available composites and
compomers. Dental Materials 2004;20:277–85.
[31] Krejci I, Albertoni M, Lutz F. In vitro—Testverfahren
zur Evaluation dentaler Restaurationssysteme. 2.
Zahnbürsten-/Zahnpastaabrasion und chemische
Degradation. Schweizer Monatsschrift für Zahnmedizin
1990;100:1164–8.
[32] Krejci I, Reich T, Bucher W, Lutz F. Eine neue Methode für die
dreidimensionale Verschleissmessung. Schweizer
Monatsschrift für Zahnmedizin 1994;104:160–9.
[33] Hickel R, Roulet JF, Bayne S, Heintze SD, Mjör IA, Peters M,
et al. Recommendations for conducting controlled clinical
studies of dental restorative materials (Science Committee
Project 2/98—FDI World Dental Federation). Journal of
Adhesive Dentistry 2007;9(Suppl 1):121–47.
[34] Taylor DF, Bayne SC, Sturdevant JR, Wilder AD. Comparison
of direct and indirect methods for analyzing wear of
posterior composite restorations. Dental Materials
1989;5:157–60.
[35] Leinfelder KF, Taylor DF, Barkmeier WW, Goldberg AJ.
Quantitative wear measurement of posterior composite
resins. Dental Materials 1986;2:198–201.
[36] Moffa JP, Lugassy AA. Calibration of evaluators utilizing the
M–L occlusal loss scale. Journal of Dental Research
1986;65:302. Abstract No. 1197.
[37] Bryant RW. Comparison of three standards for quantifying
occlusal loss of composite restorations. Dental Materials
1990;6:60–2.
[38] Perry R, Kugel G, Kunzelmann KH, Flessa HP, Estafan D.
Composite restoration wear analysis: conventional methods
vs. three-dimensional laser digitizer. Journal of the
American Dental Association 2000;131:1472–7.
[39] Lutz F, Phillips RW, Roulet JF, Setcos JC. In vivo and in vitro
wear of potential posterior composites. Journal of Dental
Research 1984;63:914–20.
[40] Meier C, Lutz F. Komposits kontra Amalgam: Vergleichende
Verschleissfestigkeitsmessungen in vivo: 1-Jahres-Resultate.
SSO: Schweizerische Monatsschrift für Zahnheilkunde
1979;89:203–12.
[41] Turssi CP, Ferracane JL, Vogel K. Filler features and their
effects on wear and degree of conversion of particulate
dental resin composites. Biomaterials 2005;26:4932–7.
[42] Pallav P, de Gee AJ, Werner A, Davidson CL. Influence of
shearing action of food on contact stress and subsequent
wear of stress-bearing composites. Journal of Dental
Research 1993;72:56–61.
[43] Hu X, Shortall AC, Marquis PM. Wear of three dental
composites under different testing conditions. Journal of
Oral Rehabilitation 2002;29:756–64.
[44] Suzuki S. Does the wear resistance of packable composite
equal that of dental amalgam? Journal of Esthetic and
Restorative Dentistry 2004;16:355–65, discussion; 365–7.
[45] Suzuki S, Nagai E, Taira Y, Minesaki Y. In vitro wear of
indirect composite restoratives. Journal of Prosthetic
Dentistry 2002;88:431–6.
973
[46] Barkmeier WW, Latta MA, Erickson RL, Wilwerding TM,
Simister BG. Evaluation of a generalized wear model for
composite. Journal of Dental Research 2002;81 (Special Issue
A): Abstract No. 3844.
[47] Barkmeier WW, Latta MA, Erickson RL, Wilwerding TM.
Simulated wear rates of resin composites and clinical
relationships. Journal of Dental Research 2006;85(Special
Issue A). Abstract No. 393.
[48] Suzuki S, Leinfelder KF. Effect of organic peroxide
additives on wear resistance of composite resin.
Compendium of Continuing Education in Dentistry
1997;18:1006–11.
[49] Suzuki S, Kobashigawa A, Leinfelder KF. In vitro wear of
direct and indirect nanhybrid composites. Journal of Dental
Research 2005;84(Special Issue A). Abstract No. 3089.
[50] Suzuki S. Simulated enamel wear during occlusal contact.
American Journal of Dentistry 2004;17:373–7.
[51] Latta MA, Barkmeier WW, Murdock CM. Localized wear of
newer-generation composite resins. Journal of Dental
Research 2003;82(Special Issue B). Abstract No. 1291.
[52] Heintze SD, Cavalleri A, Forjanic M, Zellweger G, Rousson V.
A comparison of three different methods for the
quantification of the in vitro wear of dental materials.
Dental Materials 2006;22:1051–62.
[53] Zantner C, Kielbassa AM, Martus P, Kunzelmann KH. Wear of
aluminum oxide antagonists by commercially available
composites and compomers. American Journal of Dentistry
2004;17:285–90.
[54] Condon JR, Ferracane JL. Factors effecting dental composite
wear in vitro. Journal of Biomedical Materials Research
1997;38:303–13.
[55] Ferracane JL, Choi KK, Condon JR. In vitro wear of packable
dental composites. Compendium of Continuing Education
in Dentistry Supplement 1999:S60–6.
[56] Freiberg RS, Ferracane JL. Evaluation of cure, properties and
wear resistance of Artglass dental composite. American
Journal of Dentistry 1998;11:214–8.
[57] Clelland NL, Pagnotto MP, Kerby RE, Seghi RR. Relative wear
of flowable and highly filled composite. Journal of Prosthetic
Dentistry 2005;93:153–7.
[58] Clelland NL, Villarroel SC, Knobloch LA, Seghi RR. Simulated
oral wear of packable composites. Operative Dentistry
2003;28:830–7.
[59] Correr GM, Alonso RC, Consani S, Puppin-Rontani RM,
Ferracane JL. In vitro wear of primary and permanent
enamel. Simultaneous erosion and abrasion. American
Journal of Dentistry 2007;20:394–9.
[60] Heintze SD, Cavalleri A, Zellweger G, Ferracane JL. Influence
of the antagonist material on the wear of different
composites in two different wear simulators. Dental
Materials 2006;22:166–75.
[61] Besek M, Windeler T, Schmidlin PR, Schug J, Göhring NT.
Verschleiss. In: Zahnfarbene adhäsive Füllungen im
Seitenzahnbereich. Zurich: Eigenverlag PPK; 2004.
p. 221–30.
[62] Göhring TN, Besek MJ, Schmidlin PR. Attritional wear and
abrasive surface alterations of composite resin materials in
vitro. Journal of Dentistry 2002;30:119–27.