Wear of two denture teeth materials in vivo—2-year results

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/authorsrights

Author's personal copy

d e n t a l m a t e r i a l s 2 9 ( 2 0 1 3 ) e191–e204

Available online at www.sciencedirect.com

jo ur nal home p ag e: www.int l .e lsev ierhea l th .com/ journa ls /dema

Wear of two denture teeth materials in vivo—2-year results

Siegward D. Heintzea,∗, Gaby Zellwegera, Sandro Sbicegoa, Valentin Roussonb,Carlos Munoz-Viverosc, Thomas Stoberd

a Ivoclar Vivadent, Schaan, Liechtensteinb Statistical Unit, Institute for Social and Preventive Medicine, University of Lausanne, Switzerlandc Restorative Dentistry Department, University at Buffalo, USAd Department of Prosthodontics, Heidelberg University Hospital, Germany

a r t i c l e i n f o

Article history:

Received 3 October 2012

Received in revised form

20 December 2012

Accepted 25 April 2013

Keywords:

Wear

Denture teeth

In vivo

Laser

Study design

a b s t r a c t

Objective. (1) To quantify wear of two different denture tooth materials in vivo with two study

designs, (2) to relate tooth variables to vertical loss.

Methods. Two different denture tooth materials had been used (experimental material = test;

DCL = control). In study 1 (split-mouth, 6 test centers) 60 subjects received complete den-

tures, in study 2 (two-arm, 1 test center) 29 subjects. In study 1 the mandibular dentures

were supported by implants in 33% of the subjects, in study 2 only in 3% of the subjects.

Impressions of the dentures were taken and poured with improved stone at baseline and

after 6, 12, 18 and 24 months. Each operator evaluated the wear subjectively. Wear analysis

was carried out with a laser scanning device. Maximal vertical loss of the attrition zones

was calculated for each tooth cusp and tooth. A mixed linear model was used to statistically

analyse the logarithmically transformed wear data.

Results. Due to drop-outs and unmatchable casts, only 47 subjects of study 1 and 14 of study

2 completed the 2-year recall. Overall, 75% of all teeth present could be analysed. There

was no statistically difference in the overall wear between the test and control material for

either study 1 or study 2. The relative increase in wear over time was similar in both study

designs. However, a strong subject effect and center effect were observed. The fixed factors

included in the model (time, tooth, center, etc.) accounted for 43% of the variability, whereas

the random subject effect accounted for another 30% of the variability, leaving about 28% of

unexplained variability. More wear was consistently recorded in the maxillary teeth com-

pared to the mandibular teeth and in the first molar teeth compared to the premolar teeth

and the second molars. Likewise, the supporting cusps showed more wear than the non-

supporting cusps. The amount of wear did not depend on whether or not the lower dentures

were supported by implants. The subjective wear was correct in about 67% of the cases if it

is postulated that a wear difference of 100 �m should be subjectively detectable.

Significance. The clinical wear of denture teeth is highly variable with a strong patient effect.

More wear can be expected in maxillary denture teeth compared to mandibular teeth, first

molars compared to premolars and supported cusps compared to non-supported cusps.

Laboratory data on the wear of denture tooth materials may not be confirmed in well-

structured clinical trials probably due to the large inter-individual variability.

© 2013 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved.

∗ Corresponding author at: Ivoclar Vivadent AG, Bendererstrasse 2, FL-9494 Schaan, Liechtenstein. Tel.: +423 235 3570; fax: +423 233 1279.E-mail address: [email protected] (S.D. Heintze).

0109-5641/$ – see front matter © 2013 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.dental.2013.04.012

Author's personal copy

e192 d e n t a l m a t e r i a l s 2 9 ( 2 0 1 3 ) e191–e204

1. Introduction

Wear of denture teeth is still a clinical problem because exces-sive wear can be observed in some, yet not in all patients afterone year of service already [1]. The polymer material used indenture teeth is generally less wear resistant than the mate-rial used in polymer crowns. Mostly PMMA or variations ofit are used for denture teeth because these materials showexcellent gloss stability, limited plaque accumulation, reducedchipping and ease of handling and manipulation. However,one laboratory study has demonstrated that these materi-als are prone to higher wear compared to polymer materialsfilled with glass fillers [2]. Therefore, some manufacturers havelaunched denture teeth made of filled polymers (compositeresins). A clinical study has recently shown that the clinicalwear of denture teeth made of composite was significantlylower compared to the wear of denture teeth made of PMMA[3].

Biting forces of edentulous patients do not provide anexplanation for the increased wear of denture teeth. Althoughthe occlusal forces are mainly regulated by the masticatorymuscles in edentulous patients – the receptors of the peri-odontal ligament controlling the chewing forces are missing– both the maximum biting force and the biting force dur-ing mastication is lower in edentulous subjects compared todentate subjects [4].

Only a few clinical studies have been performed to inves-tigate the wear of denture teeth. This may be explainedby the technical difficulty in accurately measuring three-dimensional changes. For the quantification of material lossin dentistry, volumetric, mechanical and scanning electron-optical procedures have been used to date (reviews by Ref.[5]). In the 1970s a dial gauge and steel reference balls incor-porated in the denture base were used to measure the wearin dentures. In another study calipers were used in con-junction with reference markings [6]. This was followed bystereophotogrammetry [7] and by reflex microscopy [8] inthe 1980s and finally by a computerized coordinate mea-suring device coupled with an electronic ruby tipped probein the 1990s [9]. In the 1990s a triangulation laser sen-sor, computer-controlled specimen stage and sophisticatedmatching software with subpixel interpolation and a scan-ning speed of 5000–10,000 surface points per second wasdeveloped for wear measurements [10]. The accuracy of thelaser scanner and matching software to calculate verticaldifferences between unprepared teeth (n = 15) and the same

Table 1 – Composition of tested denture tooth materials.

Enamel and dentin Cervical

Experimental materialtest (weight%)

DCL control(weight%)

Both materials(weight%)

Polymethyl methacrylate 72 33–35 94–97Dimethacrylate 8 5–7 1–4Cross-linked polymethyl methacrylate 0 59 0UDMA/PMMA fillers 20 0 0Pigments <0.5 <0.5 <0.5Initiators and stabilizers <0.5 <0.5 <0.5

teeth after crown preparation was calculated to be 10.9 (±1.9)�m [10,11]. An in vitro study compared the analysis of ver-tical and volumetric loss of wear facets by a laser scanner(same as in the present study), by a mechanical sensor (pro-filometry) and by an optical sensor with each other usingmonochromatic aberration as measuring principle [12]. Inthis study a favorable agreement was found between allthree quantification methods for both volumetric and verticalloss. However, the mechanical sensor measured consistentlyhigher values compared to the optical sensor for the vol-umetric loss and consistently lower values for the verticalloss.

Some studies also used laser scanning for wear analy-sis in denture teeth [1,3,13,14]. The latter study discoveredincreased wear in partial dentures compared to completedentures. Another study revealed that wear analysis can berestricted to posterior teeth [1]. Small conventional trays canbe used in conjunction with a silicone material to facilitatethe removal of the impression from the posterior dentureteeth.

A clinical study of three different denture teeth mate-rials over 3 years did not reveal a significant differencein total wear by cuspal anatomy, tooth material, preferredchewing side or gender [9]. However, there was significantlymore wear for first molars and premolars compared to sec-ond molars and for maxillary teeth compared to mandibularteeth.

In the present study two different denture tooth materi-als were investigated for wear over 24 months. One materialwas a double-crossed linked polymer (DCL = control mate-rial) which was brought on the market in 1994 under thename Antaris/Postaris and the other an improved versionof the first material. The improved version comprised 20%of UDMA/PMMA fillers (experimental material = test material)(Table 1). In this material the polymer and the matrix werehomogeneously cross-linked, which was achieved by subject-ing both the pre-cross-linked polymer and the matrix to asecondary cross-linking process. In a laboratory study theexperimental material exhibited less wear during tooth brush-ing than the DCL material [15]. Thus, the main aim of thepresent multi-center study was to quantify the wear over 24months of the DCL and experimental denture tooth materialby using the replica method and a three-dimensional laserequipment.

The study was designed in two different ways: (1) split-mouth with both types of teeth in the same subject, and (2)two-arm study with both types of teeth in two different groups

Author's personal copy

d e n t a l m a t e r i a l s 2 9 ( 2 0 1 3 ) e191–e204 e193

of subjects. The reason for using two different methods wasto evaluate whether a split-mouth design is suitable for wearassessment as it can be argued that complete dentures are notfixed in the mouth and the wear of the left and right side ofcomplete dentures may be equalized even if the wear resis-tance of both materials is different.

The following questions were investigated:

1. Is there a difference in wear between the two materials? Isthe difference related to the study design (split-mouth ortwo-arms)?

2. How do the denture teeth resist wear over time?3. Is the wear related to tooth type, type of cusp, quadrant,

jaw or other regional intraoral factors?

The following hypotheses were formulated:

1. There is no difference between the two teeth materials,independent of the study design.

2. The wear increases over time but not linearly.3. There is more wear in molars than in premolars.4. The wear does not depend on the type of cusp, quadrant or

jaw.

The possible correlation of patient-related factors likegender, age, bruxism, and number of dentures before theinsertion of the last one will be presented in a separatepaper.

2. Materials and methods

2.1. Approval by ethical committee

The multi-center clinical trial had been approved by the ethi-cal committee of the Heidelberg University Hospital, Germany(No. 375/2006) as well as by the Institutional Review Board ofthe State University of New York, USA (SIS0380807E). Based onthe application the ethical committees were aware that thematerials were not only tested in subjects in Germany and theUSA but also in The Netherlands and Canada. As both mate-rials had a CE certificate and as two ethical committees hadapproved the multi-center clinical trial there was no need toapply additionally in the other two countries where the studywas carried out.

2.2. Selection of subjects

Subjects were recruited based on the following criteria:Inclusion criteria

• edentulous subjects with indication for full dentures

Exclusion criteria

• allergy against ingredients of the denture base material• subjects that wear the dentures for less than 6 h per day• subjects from which no compliance can be expected• subjects that received the first set of full dentures less than

12 months ago.

2.3. Study centers, subject population, randomization

2.3.1. Study 1 (split-mouth design)Ten test centers were selected and 10 subjects would receivecomplete dentures in each test center. Four denturist prac-tices were selected in Canada and five in the Netherlands.Furthermore, the Department of Prosthodontics of the Heidel-berg University Hospital was chosen as an academic referencecenter. In all ten test centers the involved denturists and inves-tigators were instructed on location by the monitor (S. Sbicego)about the purpose of the study, the study design, study planand study forms to be filled in. Furthermore, the involvedclinical operators were trained with regard to standardizedimpression making (see below). In all denturist centers andin the Heidelberg University Hospital the two materials underinvestigation (DCL and experimental material) were placedaccording to the split-mouth technique (design). The test teethand the control teeth were randomly assigned to the leftand the right side of the dentures. Randomization was per-formed at Ivoclar Vivadent using the minimization techniqueincluding the variables sex age, smoker/non-smoker and cen-ter, i.e. the test material was used in quadrants 1 and 4 andthe control material in quadrants 2 and 3. The teeth wereset in bilateral balanced occlusion using the Ivoclar Vivadenttemplate.

2.3.2. Study 2 (two-arm design)The Department of Restorative Dentistry at the Universityof Buffalo (USA) was chosen as test center for study 2. Twodifferent groups of subjects received the complete dentures.Each denture was made either by a trained prosthodon-tist or by residents of the Postgraduate ProsthodonticsProgram.

2.4. Clinical procedure

For each subject the clinical operator determined the sizeand color of the denture teeth and passed the informa-tion to the sponsor. The occlusal anatomy was equal forboth materials and for all subjects. The two different toothmaterials were coded with numbers and sent to the cen-ter. Neither the dental technicians nor the clinical operatorsknew which material was used for which quadrant orsubject.

For the posterior teeth SR Ortholingual molds wereused and for the anterior teeth SR Vivodent molds (IvoclarVivadent). The teeth were available in the following col-ors: A1, A2, A3 and A3.5. The dentures were fabricatedwith ProBase High Impact or SR Ivocap High Impact (IvoclarVivadent) denture base material according to the manufac-turer’s instructions for use (standard polymerization).

At baseline (i.e. after an occlusal adjustment period of 1week), after 6 months (24 ± 2 weeks after baseline), 12 months(48 ± 2 weeks after baseline), 18 months (72 ± 4 weeks afterbaseline) and 24 months (96 ± 4 weeks after baseline) (of clin-ical service) extraoral impressions of the posterior teeth ineach quadrant were made with a polyvinyl siloxane impres-sion material (Virtual light/putty, Ivoclar Vivadent) using thedouble-mixing technique using one plastic tray for each quad-rant (Systemp.Impression Trays, Ivoclar Vivadent). Baseline

Author's personal copy

e194 d e n t a l m a t e r i a l s 2 9 ( 2 0 1 3 ) e191–e204

impressions were only taken after occlusal adjustment. Ifadjustment was necessary, the occlusal surface had beenpolished. Any other adjustment that has been made duringthe course of the study had to be reported by the clinicalinvestigators on a form that had to be sent to the sponsor.Before impression-making, the dentures were cleaned witha denture brush and submersed in an ultrasonic bath for atleast 15 min. If dental calculus was present in the fissuresand if this was not removed by the ultrasonic bath, the cal-culus was carefully removed with an ultrasonic handpiece.Before the study started, the denturists and dental instruc-tors of the universities were trained in impression-making.Red occlusion foil (Hanel 80 �m, Coltène/Whaledent, Lange-nau, Germany) was placed to mark the occlusal contact pointsin centric occlusion and the clinical status of the lower andupper jaw was photographed. The photographs were digital-ized to determine the reference points/areas during the wearanalysis with the computer program. Each center and eachsubject of each center was coded with numbers. Each clinicaloperator filled in a form recording the date of the impres-sion, subject number, sex and date of birth of the subject.At the recall sessions each clinical operator of study 1 sub-jectively assessed the wear by comparing the left and rightside and determining whether one of the sides showed morewear than the other or whether the wear was equal for bothsides.

2.5. Wear analysis

The impressions were sent to Ivoclar Vivadent inSchaan/Liechtenstein and poured out with white improvedstone (Fujirock Hard Rock, GC, Tokyo, Japan). The stone wasmixed under vacuum and filled into the impression using avibrator; the material was allowed to set in a pressure deviceat 2 bar. The casts were evaluated for imperfections (voids,pearls) with a stereo microscope (×6 magnification). Plasterpearls were carefully removed with a sharp scalpel bladeunder a microscope.

The wear quantification of all casts was performed by thesame operator (GZ) who was trained on using the method. Thebaseline and follow-up cast were scanned with a commerciallyavailable laser scanning device (etkon es1, formerly WillytecGmbH, now Straumann CADCAM, Gräfelfing, Germany). Theetkon es1 scanner scans the object from up to 10 differentangles thus providing true 3-D values of the whole tooth.With this technology the buccal and lingual surfaces canbe used as reference areas and wear analysis can be per-formed nearly independently of the amount of wear of theocclusal surface. Each denture quadrant was scanned sep-arately. Wear analysis was carried out with the Match-3Dsoftware (Willytec, Gräfelfing, Germany). Before the baselineand follow-up scan were superimposed, at least three refer-ence areas or points that were not abraded by wear accordingto the intraoral pictures were selected from at least two adja-cent teeth (e.g. 1st molar and 2nd premolar). The referenceareas were mostly located on the buccal and/or lingual sideof the teeth. The matching process was carried out with 1200minimal points and 8000 iterations and repeated several timesuntil the standard deviation of the last iterations reached a sat-isfying level (Fig. 1). Acceptance level was a standard deviation

below 30 �m. If these values were not reached the match-ing process was repeated with 2000 minimal points and/or10,000 iterations and/or the selection of other reference areas.However, the standard deviation was only one criterion fora satisfactory matching result. Other criteria were the dif-ferential picture and the distribution of the z-values. Thedifferential picture results of the superimposing and matchingprocess of baseline and follow-up model and includes nega-tive – which means material loss – as well as positive values.A symmetrical bell-shaped distribution of the z-values is anindicator for an adequate matching result. An additional cri-terion for an adequate matching result is the distribution ofthe z-values in the differential picture at areas that are notshowing much wear: the z-values should be in the range of±15 �m in these areas. If the standard deviation was abovethe acceptance level or the histogram showed an uneven dis-tribution of z-values the matching process was regarded asunsuccessful and the teeth were not included into the wearanalysis.

Wear analysis was carried out in the following way: Firstattrition areas were identified by means of the clinical pic-tures. In most cases, the wear zones on the differential picturecorresponded with the occlusal stops on the clinical picture.These areas were digitally cut and for each area the max-imal vertical loss was calculated by the software. The 99%quantile was used to eliminate outliers. For both premolars, 2wear measurements were taken (buccal and lingual). For bothmolars, 4 wear measurements were taken (mesiobuccal, disto-buccal, mesiolingual, distolingual). For each tooth, the meanand the maximum of these 2 or 4 measurements were con-sidered as summary measures, referred to in what follows asmean micrometer and max micrometer.

A Log-transformation (natural logarithms) was performedfor all these measurements to move closer to a normal distri-bution.

2.6. Statistical methods

To assess statistically the relationship between various fac-tors and the two outcome variables (Log mean, Log max), weused a linear mixed model including a random subject effectto account for the correlations within the measurements ofthe same subject. The center was introduced as a fixed effectin the model (with 7 possible values). Other factors/covariatesincluded in the model were the time of measurement (as a cat-egorical variable, with 4 possible values: 6, 12, 18 or 24 months),the material (DCL or experimental), the type of tooth (4, 5, 6or 7), as well as some characteristics of the subjects (age, sex,bruxing, smoking, time since last extraction and number ofprostheses used so far). The few missing values were imputedusing the median, respectively the most frequent value of thecorresponding factor. We repeated the analysis excluding thesubjects with missing values without changing our results.To take into account the factor quadrant, another two binaryvariables were included, one for the jaw (maxilla/mandible)and one for the side (left/right). In our model, the interceptrefers to the average Log wear measured after 6 months forthe tooth 4 from the left maxilla in center 1 with materialDCL.

Author's personal copy

d e n t a l m a t e r i a l s 2 9 ( 2 0 1 3 ) e191–e204 e195

Fig. 1 – Scans of baseline replica (left) and after 2 years of clinical service (right) and differential picture showing materialloss in red.

The significance of the slope coefficients in the modelwas assessed using Wald tests. To assess whether a set ofcoefficients was significant (which was useful to test whetherthe factors time, center or type of tooth were significant), alikelihood ratio test was carried out. We also used a likelihoodratio test to check whether a categorical time effect was sta-tistically superior to a linear time effect. The variances of therandom subject effect and of the random error were estimatedusing the method of restricted maximum likelihood. All cal-culations were performed using the lme routine in the nlmelibrary from the R free statistical package (version 2.5.1).

3. Results

3.1. Test centers – number of subjects

3.1.1. Study 1Unfortunately, 4 test centers (denturist practices in Canada)were not able to participate in the first study (split-mouth)due to illness of the denturist or impossibility of denturist toparticipate in the study thus reducing the number of test cen-ters to 6 (5 denturist practices and one university) and thenumber of subjects involved to 60 (Table 2). During the 2-yearperiod there were six documented subject drop-outs: two sub-jects with broken dentures, two subjects who did not wantto participate in the study any longer, one subject with failedimplants in the lower jaw, and one subject who died during thestudy (Table 2). In test center 10 two drop-outs had been com-pensated by the recruitment of two new subjects. At baseline

the complete dentures of the mandible were attached to den-tal implants in 23 of the 60 subjects (Table 2). In 42 of thebaseline 60 subjects the second molars had been mounted inall four quadrants (Table 2). In three subjects (two test centers)the lower denture had been relined, one at the 6-month recall,one at the 12-month recall and one at the 18-onth recall.

3.1.2. Study 2In the second study (named as test “center 2” as one of theother centers of study 1 discontinued with the study) only 14subjects of the 29 subjects at baseline completed the 2-yearrecall. The reasons for the high drop-out were concurrent ill-nesses, two subjects passed away and others could not attendthe required visits due to an unusual harsh weather that year.No replicas were made at the 18-month recall. For only onesubject the complete lower denture was supported by dentalimplants (Table 2). In 16 of the 29 baseline subjects the secondmolars were included in all four quadrants.

3.2. Data structure

Wear data were available for 764 teeth after 6 months, for 651teeth after 12 months, for 474 teeth after 18 months and for 575teeth after 24 months, whereas 973 teeth were measured atleast once (on average 11.7 teeth per subject) (Table 3). In study1 the largest amount of wear data and the highest number ofmatchable teeth derived from test centers 1 and 10 followedby test centers 8 and 9. In test centers 5 and 6, many replicascould not be evaluated due to poor quality of the impressions.

Author's personal copy

e196 d e n t a l m a t e r i a l s 2 9 ( 2 0 1 3 ) e191–e204

Table 2 – Number of impressions sent in relation to the individual test centers and recalls (months).

Test center 1 2 3 4 5 6 7 8 9 10

Implant 2 1 4 6 6 4 12nd molar 9 16 9 7 6 10 1

RecallBaseline 10 29 (15/14) 4 0 10 9 0 9* 10 12106 months 10 21* (10/11) 0 0 9 8 0 9 9 912 months 9* 15* (7/8) 0 0 9 5 0 9 8 1018 months 9 0 0 0 9* 3 0 9 8 7*

24 months 9 14* (6/8) 0 0 8* 3 0 9 8 8

Implant = number of subjects with lower dentures retained by implants.Center 2: in brackets: distribution of DCL versus experimental material.∗ Drop-out.

3.3. Wear analysis

Three denturists have reported that they have adjusted theocclusion on one subject each at the 6-month recall at twolower teeth each (second premolar and first molar). Theseteeth were excluded from further wear analysis. In two sub-jects the lower denture had broken after the 18 month recalland had to be replaced by a new denture. These subjects hadbeen excluded from the wear analysis of the 24 month recall.

Since both summary measurements (mean and max ver-tical loss) were highly correlated (Pearson correlation = 0.96,data not shown) and since the results of our statistical mod-els below were very similar when modeling the mean or themax vertical loss, only the results for the former are discussedbelow.

Tables 4a and 4b show the results of the linear mixed mod-els applied separately to the data of study 1 and study 2. Theresults were remarkably similar in both studies, even thoughthe intercept was somewhat lower in study 2 suggesting lesswear. In particular, the material effect was almost zero in bothstudies. Thus, we pooled the data from both studies, and recal-culated a linear mixed model, which is presented in Table 4c.Note that the “study effect” is represented in this model bythe factor “center 2”. Fig. 2 shows boxplots of the not trans-formed and Log-transformed raw measurements in functionof the factors material, tooth and quadrant, whereas Fig. 3illustrates some of the estimated effects in this model.

There was a consistent increase of the wear over time (asillustrated in Figs. 2 and 3), even though not quite linear, evennot on the log scale. The time effect was highly significant.The average Log wear increase was estimated at 0.41 between 6and 12 months, at 0.28 between 12 and 18 months, and at 0.16,

between 18 and 24 months, corresponding to estimated ratiosof raw measurements (on the original scale) of exp (0.41) = 1.51between 6 and 12 months, of exp (0.28) = 1.32 between 12 and18 months, and of exp (0.16) = 1.17 between 18 and 24 months,suggesting a slowing down of the wear increase with time.This was confirmed by the fact that the model with a categor-ical time effect was significantly superior to a model with onlya linear time effect.

Regarding the quadrants, the average difference of loga-rithms between the upper and the lower jaw was estimated at0.2, which was highly significant with more wear in the max-illa than the mandible. Significance was almost reached for thefactor side, with slightly more wear on the left side than on theright of the mouth (p = 0.07 for Log mean, and even p = 0.04 forLog max).

There was more wear for tooth = 5 and tooth = 6 than fortooth = 4 or tooth = 7. Overall, the tooth effect was highly sig-nificant, with an average increase of Log wear of about 0.10 forteeth 5 and 6 compared to tooth 4 (with respect to Log max,the average increase of Log wear for tooth 6 was estimatedat 0.28). The results were somewhat inconsistent for tooth 7,which had significantly less wear than tooth 4 with respect toLog mean, whilst it had significantly more wear than tooth 4with respect to Log max.

The center effect was, overall, highly significant (as illus-trated in Fig. 3). The ranking among the centers was thefollowing (from less wear to more wear): center 10 (−0.72),center 2 (−0.40), center 8 (−0.02), center 1 (0), center 9 (0.02),center 5 (0.17) and center 6 (0.22). The amount of wear didnot depend on whether or not the lower dentures were sup-ported by implants (p = 0.93). Overall, the fixed factors includedin the model (time, tooth, center, etc.) accounted for 43% of

Table 3 – Number of teeth with wear data available in relation to the individual centers (1/2/5/6/8/9/10), recall (6/12/18/24months) and type of tooth (4/5/6/7).

Tooth 4 Tooth 5 Tooth 6 Tooth 7

Center/recall 6 12 18 24 6 12 18 24 6 12 18 24 6 12 18 24

1 33 24 33 31 32 25 31 31 33 21 29 29 29 20 24 282 72 47 0 38 77 51 0 38 78 47 0 35 51 30 0 205 1 7 3 4 9 20 11 18 5 10 9 13 6 7 13 136 4 5 3 0 21 14 11 12 23 18 16 8 24 14 11 118 12 11 20 11 26 27 29 17 26 31 33 20 13 21 22 169 9 14 15 11 22 24 24 24 26 23 28 24 27 24 28 2510 34 37 25 32 33 37 26 31 34 38 26 31 4 4 4 4

Author's personal copy

d e n t a l m a t e r i a l s 2 9 ( 2 0 1 3 ) e191–e204 e197

Table 4a – Results of linear mixed model of Log mean for study 1.

Value Standard error DF t-value p-value

(Intercept) 4.75 0.18 1814 26.56 0.000time12 0.37 0.03 1814 13.74 0.000time18 0.68 0.03 1814 25.20 0.000time24 0.85 0.03 1814 31.06 0.000mat2 0.02 0.02 1814 0.92 0.360right 0.03 0.02 1814 1.54 0.123mand 0.19 0.02 1814 10.39 0.000tooth5 0.09 0.03 1814 3.34 0.001tooth6 0.12 0.03 1814 4.34 0.000tooth7 0.21 0.03 1814 6.82 0.000

Table 4b – Results of linear mixed model of Log mean for study 2.

Value Standard error DF t-value p-value

(Intercept) 4.37 0.17 555 25.94 0.000time12 0.49 0.04 555 11.42 0.000time24 0.82 0.05 555 17.14 0.000mat2 0.01 0.23 20 0.06 0.952right 0.04 0.03 555 1.11 0.267mand 0.18 0.03 555 5.08 0.000tooth5 0.10 0.05 555 2.20 0.028tooth6 0.07 0.05 555 1.43 0.154tooth7 0.06 0.05 555 1.03 0.305

the variability, whereas the random subject effect accountedfor another 30% of the variability, leaving about 28% of unex-plained variability.

3.4. Modeling single measurements of wear data

As there were a few (22 out of 7027) zero values, they werereplaced by half of the minimum among the non-zero mea-surements (i.e. half of 7, or 3.5).

Boxplots of the logarithms of single measurementsare shown in Fig. 4, in function of the factors time(6/12/18/24 months), jaw (maxilla/mandible), tooth (4/5/6/7),lingual/buccal (l/b) and distal/mesial (d/m), the latter factorbeing only defined for teeth 6 and 7 (molars).

The most striking result was that the buccal measurementswere consistently lower than the lingual measurements in the

maxilla, whereas it was the other way round in the mandible. Astrong interaction between the factor jaw (maxilla/mandible)and the factor lingual/buccal was observed. For molars, themesial measurements were consistently higher than the dis-tal measurements in the maxilla, whereas these differenceswere not as clear in the mandible. This also suggests someinteraction between the factor jaw (maxilla/mandible) and thefactor distal/mesial.

Table 5 presents the results of a linear mixed model for thesingle Log measurements which included as fixed factors time,tooth, center, jaw, lingual/buccal and distal/mesial, as well asinteractions of the factor jaw with the factor lingual/buccaland with the factor distal/mesial. Both of these interactionswere highly significant. We received the following estimationsfor the average of the Log measurements (after 6 months,for tooth 4 and in center 1): 4.80 for lingual measurements

Table 4c – Results of linear mixed model of Log mean for pooled data of study 1 and study 2.

Value Standard error DF t-value p-value

(Intercept) 4.81 0.19 2376 25.40 0.000time12 0.41 0.02 2376 17.72 0.000time18 0.69 0.03 2376 26.75 0.000time24 0.85 0.02 2376 35.64 0.000mat2 −0.02 0.02 2376 −0.93 0.354right −0.03 0.02 2376 −1.84 0.066mand −0.19 0.02 2376 −11.68 0.000tooth5 0.10 0.02 2376 4.07 0.000tooth6 0.11 0.02 2376 4.65 0.000tooth7 −0.18 0.03 2376 −6.58 0.000center2 −0.40 0.17 66 −2.32 0.023center5 0.17 0.20 66 0.86 0.392center6 0.22 0.23 66 0.99 0.327center8 −0.03 0.20 66 −0.13 0.895center9 0.02 0.20 66 0.09 0.928center10 −0.72 0.20 66 −3.63 0.001

Author's personal copy

e198 d e n t a l m a t e r i a l s 2 9 ( 2 0 1 3 ) e191–e204

Fig. 2 – Boxplots with median vertical loss in relation to material, type of tooth, quadrant and recall interval, (above) nottransformed data, below Log transformed data.Material: 1 = DCL (red), 2 = experimental material (green). Tooth: 4 = first premolar (red), 5 = second premolar (green), 6 = firstmolar (dark blue), 7 = second molar (light blue). Quadrant: red = 1st quadrant, green = 2nd quadrant, dark blue = 3rd quadrant,light blue = 4th quadrant.

Author's personal copy

d e n t a l m a t e r i a l s 2 9 ( 2 0 1 3 ) e191–e204 e199

Fig. 3 – Main effects and variables of interest on mean Log wear along time.Material: 1 = DCL, 2 = experimental material. Tooth: 4 = first premolar, 5 = second premolar, 6 = first molar, 7 = second molar.

Table 5 – Results of a linear mixed model for the single Log measurements, including interactions of the factorsmaxilla/mandible with the factors lingual/buccal and distal/mesial.

Value Study error DF t-value p-value

(Intercept) 4.80 0.13 6937 36.47 0.000time12 0.40 0.02 6937 20.08 0.000time18 0.69 0.02 6937 30.97 0.000time24 0.86 0.02 6937 41.29 0.000tooth5 0.09 0.02 6937 3.78 0.000tooth6 0.02 0.02 6937 0.72 0.473tooth7 −0.32 0.03 6937 −12.30 0.000center2 −0.35 0.16 72 −2.22 0.030center5 0.21 0.19 72 1.12 0.265center6 0.24 0.20 72 1.17 0.247center8 −0.03 0.19 72 −0.15 0.883center9 0.05 0.18 72 0.27 0.791center10 −0.65 0.18 72 −3.65 0.000mand −0.47 0.02 6937 −21.27 0.000buccal −0.48 0.02 6937 −24.33 0.000mesial 0.18 0.02 6937 8.01 0.000mand:buccal 0.77 0.03 6937 27.28 0.000mand:mesial −0.12 0.03 6937 −4.12 0.000

Author's personal copy

e200 d e n t a l m a t e r i a l s 2 9 ( 2 0 1 3 ) e191–e204

Fig. 4 – Boxplots of Log mean vertical loss in relation to type of tooth, type of tooth cusp, quadrant and recall interval.l = lingual, b = buccal, m = mesial, d = distal. Tooth: 4 = first premolar, 5 = second premolar, 6 = first molar, 7 = second molar.

in the maxilla, 4.80–0.48 = 4.32 for buccal measurements inthe maxilla, 4.80 − 0.47 = 4.33 for lingual measurements in themandible, and 4.80 − 0.48 − 0.47 + 0.77 = 4.62 for buccal mea-surements in the mandible. For molars, the estimated averageincrease of Log for the mesial measurements (compared to thedistal measurements) was 0.18 in the maxilla, and only 0.06(0.18 − 0.12) in the mandible.

The pooled mean wear data for different tooth areas in rela-tion to type of tooth and time is presented in Fig. 5. Somediscrepancies compared to the general tendencies describedabove may be noted, suggesting further (more complicated)interactions. For example, the interaction between the fac-tor jaw and the factor buccal/lingual might not be exactly thesame for molars and for premolars.

3.5. Subjective assessment of wear

For subjects receiving both materials (study 1), a subjec-tive assessment was performed with three possible values,

indicating if one of the two materials (DCL or experimentalmaterial) produced less wear or if both materials were (approx-imately) equivalent after 6, 12, 18 and 24 months. A total of177 such assessments (dentures in maxilla and mandible andat different recall intervals) were available (on average 2.90assessments per subject). By retaining differences of 100 �mand more in wear, interpreted here as the “objective assess-ments”, we can summarize the results into a contingencytable (Table 6). This corresponds to an estimated proportionof agreement of 100/150 = 67% (95% CI: 59–74%).

4. Discussion

The strength of the present prospective clinical study isattributed to the relatively high number of subjects, the multi-center approach, the high level of standardization involvedand the randomization of test and control teeth.

Author's personal copy

d e n t a l m a t e r i a l s 2 9 ( 2 0 1 3 ) e191–e204 e201

Fig. 5 – Pooled mean Log vertical loss of all subjects of all test centers in relation to type of tooth, cusp and recall interval.l = lingual, b = buccal, m = mesial, d = distal. Tooth: 4 = first premolar, 5 = second premolar, 6 = first molar, 7 = second molar.

The technique of replica fabrication by means of impres-sion making with polyvinyl siloxane materials and improveddental stone has been proven to provide the accuracy requiredfor wear analysis [16]. The laser scanner device with the appro-priate software Match 3-D has been proven to be an accurate,reliable and efficient tool to measure wear on replicas [10].So far, this device has mainly been used for the wear analy-sis of direct and indirect composite fillings [5,17–19], ceramic

inlays and crowns and the antagonists opposing them [20–22].Three studies used the device for wear analysis of dentureteeth [1,13,14]. In the first two of these studies the mean ver-tical loss was calculated for the whole occlusal surface area,which explains the rather low wear values measured in thesestudies. As the accuracy of replicas made of impressions in vivoand scanned by the laser scanner device is only in the rangeof ±10–15 �m, negative and positive values are taken into the

Table 6 – Distribution of wear measurement by laser scanning in relation to subjective wear assessment – the thresholdvalue for subjectively detectable wear was 100 �m of objectively measured occlusal wear.

Subjective assessment Objective measurement

Experimental material less Same DCL less

Experimental material less 3 16 0Same 12 96 9DCL less 0 13 1

Author's personal copy

e202 d e n t a l m a t e r i a l s 2 9 ( 2 0 1 3 ) e191–e204

account when calculating the mean negate differences, thusunderestimating the real vertical loss. Therefore, the meanvertical loss of regions of interest (attrition zones) should beused for wear analysis.

Replicas made from intraoral impressions are subject to avariety of factors that influence the quality of the improvedstone model, namely distortions during removal, saliva con-tamination, plaque and calculus on the denture teeth. Incontrast to fillings and crowns, the advantage of complete den-tures is that impressions are made outside of the mouth undercontrolled conditions including meticulous cleaning of thedenture teeth. Thus, low standard deviations for the matchingprocedure may be expected. This was, however, not the casefor all test centers. Although the involved clinicians had beentrained on impression-making, the quality of the impressionswas quite different from one test center to the other.

Experience with in vivo wear measurements has shownthat the quality of matching and superimposing scans,expressed by the standard deviation of the superimposedpoints of the same region of the baseline and follow-up modelis inferior to the quality of replicas from in vitro models [12].While in vitro the standard deviation is between 5 �m and10 �m, it is between 15 and 20 �m under in vivo conditions.In the present study, a threshold value of 30 �m for posteriorteeth was chosen in order to increase the number of mea-surable units. This approach is justified because the verticalwear in the posterior region had a magnitude above 100 �m inmost cases. Thus, the suboptimal matching of the scans wasrelativized.

The occlusal adjustment before the taking of the baselineimpression and during the 2-year course of the study couldhave influenced the wear of the tested materials and couldalso explain the differences between the different test cen-ters. Even if one may suspect that not all involved investigatorshad kept to the rule that baseline impressions were only madeafter occlusal adjustments one should have seen those on thebaseline scans – but there weren‘t. The clinical investigatorsreported only in three cases, that they have carried out theocclusal adjustment after baseline. The adjustment before thebaseline impression-taking can be critical because of two rea-sons (1) the denture tooth material is composed by differentlayers of different materials with different physical properties,e.g. a material for enamel/incisal and dentin. However, thatwas not the case for the two DCL materials under investiga-tion. (2) The flattening of cusps can reduce the wear of thosecusps as sharp cusps wear more rapidly than flat cusps. Thisprobably has happened in the present study but reflects thenormal clinical procedure.

The results were very similar for both study designs. Inparticular, there was no difference between test and controlmaterial. However, this is not to say that different materialscannot be tested with a split-mouth design. The difference inwear between both materials was very small or insignificant.The outcome of the study could have been different if the wearof both materials had been of a greater and more significantmagnitude.

The wear of the denture teeth in the present study after 1year was comparable to the wear measured in a pilot studywhich used the same wear quantifying protocol and system[1]. However, when the data were compared to other studies

using other methodologies, the wear rates were higher. Inthe present study, the median vertical loss of the first molarafter 2 years was about 250 �m while it was 225 �m (mean)after 3 years and about 150 �m (pooled data) after 2 yearsin another study [9]. In yet another clinical trial on the wearof denture teeth, the median maximum loss of material was146 �m after 1 year and 224 �m after 2 years [13]. However, thistrial included partial dentures or complete dentures in sin-gle jaws. Significantly more wear was observed if the dentureteeth were opposed by natural teeth, crowns or fixed dentalprostheses.

The arbitrarily chosen 100 �m wear difference of a subjec-tive comparison of 2 denture teeth, yielded a rate of agreementbetween of 67%. There were no “complete contradictions” (i.e.cases where the subjective assessment would rate one mate-rial better, whereas the objective assessment would show it tobe worse). However, as there was not much difference betweenthe two materials in this study, this might not be a samplewhich is really appropriate for such a comparison.

The clinical wear increased with time but not linearly,which is in line with clinical wear study on 3 denture toothmaterials over 3 years [9] and with a more recent studyover 2 years [13]. The primary factors affecting wear werethe subject and the type of tooth within the same subject.Vertical loss showed a large inter-individual variation – aresult which was also been found in a previous study usingthe same methodology [1]. In the present study, significantlymore wear was identified in the maxillary teeth than in themandibular teeth. This finding was consistent across all recallintervals and test centers and in line with some other studies[1,9] but not with all [13]. The reason for the higher wear ofmaxillary teeth compared to mandibular teeth could be thefact that maxillary dentures are more stable than mandibu-lar dentures, thus creating more force and attrition. Anotherpossible explanation is that the bilaterally balanced lingualcontact occlusion creates a larger cusp contacting area inmandibular teeth than in maxillary teeth, thus leading tohigher contact forces in the maxillary teeth during mastica-tion. On the other hand, the presence of implants retainingthe dentures in the lower jaw had no influence on wear.Higher wear in dentures retained by implants might have beenexpected because the chewing forces tend to be higher in thesepatients [4].

The denture teeth of the left quadrants tended to showmore wear than those of the right quadrants. However, thisfinding may be interpreted as a random finding, especially asthe difference was not statistically significant. Yet, the findingsmay also be interpreted as a reflection of the preferred chewingside, which was, however, not recorded in the study. One clin-ical study revealed higher masseter activity and biting forceon the preferred chewing side compared to the non-preferredchewing side in denture wearers [23].

As far as the wear of specific teeth is concerned, wear ofthe second premolar was similar to that of the first molarbut significantly more than that of the first premolar. Sucha difference was not found in a previous study [1]. Whethera special anatomical form or supposed higher biting forcesin the molar region may be the explanation for this findingremains to be proven. The wear results of the second molarsmust be regarded with caution, as these teeth had been set

Author's personal copy

d e n t a l m a t e r i a l s 2 9 ( 2 0 1 3 ) e191–e204 e203

only in 42 of the 89 subjects. Another finding is interestingand trivial at the same time. The lingual cusps of the maxil-lary teeth showed statistically significantly more vertical wearthan the buccal cusps, whilst it was the other way round inthe mandibular teeth. This finding can be explained by theocclusion concept (occusal scheme) that was followed dur-ing setting of the denture teeth. Invariably the lingual cuspsof the maxillary teeth had contact with the fossae of themandibular teeth and the buccal teeth of the mandibular teethwere in contact with the fossae of the maxillary teeth. Thesecusps can therefore be regarded as supporting cusps and aretherefore subjected to higher loading and consequently higherwear.

There was no significant difference in wear between bothmaterials, even though data from laboratory trials showedthat the experimental material produced significant less wearthan the DCL control material (see introduction). As the inter-individual variability was very high in the present clinicalstudy, slight differences in the wear resistance of a denturetooth material may not be clinically visible or provable. In clin-ical trials, the acrylic resin denture teeth showed more wearthan composite denture teeth [3]. This was confirmed in labo-ratory studies involving two-body [2,3,24] and three-body wearsimulation [25].

The relatively short duration of the study presents a lim-itation, as dentures should function for a longer period oftime. Very few data have been published on the longevity ofcomplete dentures. A retrospective study that was carried outin a general practice in Germany evaluating the records of332 subjects with complete dentures came to the conclusionthat after 5 years 83% of dentures were still in situ and after10 years about 50% were still in use [26]. If the wear after 2years were extrapolated, the average wear in molars wouldbe about 0.75 mm after 5 years and 1.5 mm after 10 years.Probably this extrapolation is inadequate as a 7-year clinicalstudy on 60 subjects with complete dentures and PMM dentureteeth revealed a sharp increase in wear after the second yearof clinical service [6]. In this study, a vertical loss of 1.1 mmwas measured for molar teeth after 7 years, but very simpleinstruments (calipers) were used. Much larger than the wearof the denture teeth, however, is the resorption of the sup-porting bone, with about 0.5 mm/year for the alveolar ridgeand 0.1 for the maxilla [27] which means that after 10 yearsthere is a loss of vertical height of about 6 mm due to boneresorption.

5. Summary

Using linear mixed models, we were able to describe the highlysignificant and non-linear increase of Log wear over time. Byincluding a random subject effect in the model, we were ableto consider data obtained from different designs (compari-son of two materials within or between patients) in the samemodel. No consistent significant difference could be identifiedbetween the two materials. Significant differences among thecenters were observed. The subject effect accounted for a largepart of the variability.

Teeth 5 and 6 showed significantly more wear than tooth4, which in turn had significantly more wear than tooth 7.

When modeling single measurements, there was more wearin the maxilla than in the mandible, and there was also morewear for the supporting cusps than for the non-supportingcusps. For maxillary molars, mesial measurements producedconsistently more wear than distal measurement.

6. Conclusions

The first three hypotheses were confirmed:

1. There was no difference between both materials, indepen-dent of the study design.

2. The wear increased over time but not linearly.3. There was more wear in the first molar than in the premo-

lars.

The fourth hypothesis was rejected because the jaws andcusps influenced the wear result. The inter-individual vari-ability of wear was very high and questions the usefulness oflaboratory tests to predict the clinical wear of denture toothmaterials.

Conflict of interest

The study has been designed and supervised by the spon-sor Ivoclar Vivadent. All clinical investigators in Germany, theNetherlands and Canada signed a contract and received afinancial compensation for the workload. The wear analysishad been done at the R&D headquarters of the sponsor bytrained and calibrated specialists.

Acknowledgements

The authors are grateful to all denturists that were involvedin the clinical study, namely Nancy Tomkins from Canada andPieter Smit, Ton Lameijer, Lodewijk Boersma and Rene Kam-stra from the Netherlands.

r e f e r e n c e s

[1] Schmid-Schwap M, Rousson V, Vornwagner K, Heintze SD.Wear of two artificial tooth materials in vivo: a 12 monthpilot study. Journal of Prosthetic Dentistry 2009;102:104–14.

[2] Ghazal M, Yang B, Ludwig K, Kern M. Two-body wear of resinand ceramic denture teeth in comparison to human enamel.Dental Materials 2008;24:502–7.

[3] Heintze SD, Zellweger G, Grunert I, Munoz-Viveros CA,Hagenbuch K. Laboratory methods for evaluating the wearof denture teeth and their correlation with clinical results.Dental Materials 2012;28:261–72.

[4] Fontijn Tekamp FA, Slagter AP, Van Der Bilt A, Van Hof MA,Witter DJ, Kalk W, et al. Biting and chewing in overdentures,full dentures, and natural dentitions. Journal of DentalResearch 2000;79:1519–24.

[5] Kunzelmann K-H. (German) Wear analysis andquantification of restoration materials in vivo and in vitro.Aachen: Shaker Verlag; 1998.

[6] Rarisch VB. (German) Results of the 7-year longitudinalstudy on the abrasion of polymer denture teeth in completedentures. Deutsch Zahnärztl Z 1982;37:776–7.

Author's personal copy

e204 d e n t a l m a t e r i a l s 2 9 ( 2 0 1 3 ) e191–e204

[7] Ogle RE, Ortman LF. Measuring wear of artificial teeth withstereophotography: part I. Journal of Prosthetic Dentistry1985;53:807–12.

[8] Adams LP, Jooste CH, Thomas CJ. An indirect in vivo methodfor quantification of wear of denture teeth. Dental Materials1989;5:31–4.

[9] Ogle RE, Davis EL. Clinical wear study of three commerciallyavailable artificial tooth materials: thirty-six month results.Journal of Prosthetic Dentistry 1998;79:145–51.

[10] Mehl A, Gloger W, Kunzelmann KH, Hickel R. A new optical3-D device for the detection of wear. Journal of DentalResearch 1997;76:1799–807.

[11] Mehl, A. 3D-Datenerfassung und CAD/CAM-Technologie in derrestaurativen Zahnmedizin – Entwicklung, Einsatz undPerspektiven eines neuen Verfahrens. Habilitation, University ofMunich; 1998.

[12] Heintze SD, Cavalleri A, Forjanic M, Zellweger G, Rousson V.A comparison of three different methods for thequantification of the in vitro wear of dental materials.Dental Materials 2006;22:1051–62.

[13] Stober T, Geiger A, Rues S, Dreyhaupt J, Rammelsberg P,Ohlmann B. Factors affecting wear of composite resindenture teeth - 24-month results from a clinical study.Clinical Oral Investigations 2012;16:413–20.

[14] Ohlmann B, Rohstock K, Kugler J, Gilde H, Nat R, Dreyhaupt J,et al. Influences on clinical wear of acrylic denture teeth: apilot study. International Journal of Prosthodontics2007;20:496–8.

[15] Shahdad SA, Wassell W, McCabe JF, Rusby S. In-vitrothree-body wear of two experimental denture teeth. Journalof Dental Research 2004, 83 Spec Issue A: Abstract No. 1389.

[16] Wise M. Fit of implant-supported fixed prostheses fabricatedon master casts made from a dental stone and a dentalplaster. Journal of Prosthetic Dentistry 2001;86:532–8.

[17] Söderholm KJ, Lambrechts P, Sarrett D, Abe Y, Yang MC,Labella R, et al. Clinical wear performance of eightexperimental dental composites over three yearsdetermined by two measuring methods. European Journal ofOral Sciences 2001;109:273–81.

[18] Perry R, Kugel G, Kunzelmann KH, Flessa HP, Estafan D.Composite restoration wear analysis: conventional methodsvs three-dimensional laser digitizer. Journal of the AmericanDental Association 2000;131:1472–7.

[19] Barkmeier WW, Latta MA, Erickson RL, Lambrechts P.Comparison of laboratory and clinical wear rates of resincomposites. Quintessence International 2004;35:269–74.

[20] Krämer N, Kunzelmann KH, Taschner M, Mehl A,Garcia-Godoy F, Frankenberger R. Antagonist enamel wearsmore than ceramic inlays. Journal of Dental Research2006;85:1097–100.

[21] Esquivel-Upshaw JF, Young H, Jones J, Yang M, Anusavice KJ.In vivo wear of enamel by a lithia disilicate-based coreceramic used for posterior fixed partial dentures: first-yearresults. International Journal of Prosthodontics2006;19:391–6.

[22] Esquivel-Upshaw JF, Rose Jr WF, Barrett AA, Oliveira ER, YangMC, Clark AE, Anusavice KJ. Three years in vivo wear:Core-ceramic, veneers, and enamel antagonists. DentalMaterials 2012;28:615–21.

[23] Garrett NR, Kaurich M, Perez P, Kapur KK. Masseter muscleactivity in denture wearers with superior and poormasticatory performance. Journal of Prosthetic Dentistry1995;74:628–36.

[24] Hahnel S, Behr M, Handel G, Rosentritt M. Two-body wear ofartificial acrylic and composite resin teeth in relation toantagonist material. Journal of Prosthetic Dentistry2009;101:269–78.

[25] Stober T, Henninger M, Schmitter M, Pritsch M,Rammelsberg P. Three-body wear of resin denture teethwith and without nanofillers. Journal of Prosthetic Dentistry2010;103:108–17.

[26] Kerschbaum T, Hirland K, Teeuwen R, Faber F. (German) Onthe survival rate of complete dentures. Deutsch Zahnärztl Z2007;62:459–64.

[27] Devlin H, Ferguson MW. Alveolar ridge resorption andmandibular atrophy: a review of the role of local andsystemic factors. British Dental Journal 1991;170:101–4.