Three-year occlusal wear of posterior composite restorations

Dent Mater 8:224-228, July, 1992

Three-year occlusal wear of posterior composite restorations M. A. Freilich 1, A.J. Goldberg1, R. O. Gilpatrick2, and R. J. Simonsen 2~ 1Departmentof Prosthodontics, Universityof Connecticut, School of Dental Medicine,Farmington, CT, USA 2Departmentof GeneralDentistry, Universityof Tennessee, Collegeof Dentistry, Memphis, TN, USA 3U.S. & International Dental ProductsDivision, 3M Health Care, St. Paul, MN, USA

Abstract. The specific aims of this study were to: 1) measure the occlusal wear of four different dental composite materials placed in the posterior teeth of adults; and 2) evaluate the effect of the clinical parameters, cavity class and tooth type on occlusal wear. Four differentvisible light-cured composite materials were usedto make the restorations in this study. The restorations placed for this randomized clinical trial were scored through the use of an indirect evaluation system (M-L scale). The total sample size per recall ranged from 90 to 142 restorations from baseline to 36 months, The mean wear at 36 months for Heliomolar, J&J Experimental (Adaptic II) and P-30 was 45- 54 pm, which is rather low compared to the recently reported wear of other composite materials, Marathon exhibited significantly greater wear with a mean of 174 pm at 36 months. The data also showed that cavity class and tooth type had no significant effect on the occlusal wear of the restorations made with the three low wear-rate materials, while restorations composed of the high wear-rate material exhibited more wear in molars than premolars; this effect was again not statistically significant. These data support the hypothesis that the overall wear of a composite restoration is more dependent on the material's properties than clinical parameters such as cavity class and tooth type.

Selection of subjects. As a subset ofalarger study, one hundred

INTRODUCTION Composite materials have been used to restore posterior teeth with varying degrees of success over the past two decades, Some of the earliest dental composite formulations proved unsatisfactory due to their poor resistance to occlusal wear (Phillips et al.,1973; Goldberg et al., 1984; Leinfelder et al., 1986a). Data from recent clinical trials have shown that the continued development of"posterior composites" has resulted in markedly improved resistance to wear (Boksman et al., 1986; Sturdevant et al., 1988; Lundin and Koch, 1989). While guidelines for the use of posterior composite materials have been proposed (Leinfelder and Lemons, 1988; Leinfelder, 1991), the effects of clinical parameters such as cavity class and tooth type on the wear resistance ofrestorations made with these materials remain unclear. In vivo studies with six- to sixty-month observations have found contrasting results when analyzing for these effects (Lutz et al., 1984; Boksman et al., 1986; Sturdevant et al., 1988; Rowe 1989; Bayne et al., 1991). It is possible that the effects of cavity class and tooth type on occlusal wear are largely dependent upon the particular composite material studied, and its overall

and forty six Class I and Class II restorations were placed in the posterior teeth of 47 healthy subjects (mean of 3.2 _+2.2 restorations per subject). The majority of the subjects who participated in this clinical trial were University of Connecticut Health Center employees and entering students. The age of this population ranged from 22 to 47 years (mean of 30 _+7.3 years). Potential subjects were screened by reviewing their dental and health histories and by performing clinical and radiographic examinations. Eligible subjects presented with a need for the restoration of posterior teeth as a result ofprimary or recurrent caries or defectiverestorations. Thoseindividuals who had a positive history of asthma, heart or liver disease, or those with signs or symptoms of TMJ disorders or malocclusions were excluded from the study. Furthermore, teeth selected for the study which upon preparation were not completely surrounded by enamel were excluded from the study. Study design. Four different visible light-cured composite materials were used to make the restorations in this study. They included three hybrid composites - "J&J Experimental" (commercially-available as Adaptic II - Johnson & Johnson, Skillman, NJ, USA; batch # 6459-82-1), "Marathon" (Den-Mat

224 Freilich et aL/Occlusal wear of composites

resistance to wear. The experimental composite materials used in the above studies, where the effects of these clinical parameters were analyzed, were found to exhibit greatly improved overall resistance to occlusal wear when compared to the earliest formulations. Materials which have been developed most recently, however, exhibit even greaterwear resistance. Since it is possible that the effects of these parameters are dependent on the overall wear characteristics of the material, it is logical to study the effects of cavity class and tooth type on these more wear-resistant, commercially-available matedals. The specific aims of this study were to: 1) measure the occlusa] wear of four different composite materials placed in the posterior teeth of adults; and 2) evaluate the effect of the clinical parameters, cavity class and tooth type on occlusal wear. In addition, pilot data were collected to study the effect of cavosurface margin configuration on the wear of these materials. MATERIALS AND METHODS

Inc., Santa Maria, CA, USA; batch # 256006 & 410283), and "P-30" (3M Dental Products Division, St. Paul, MN, USA; batch # 4RIP); and one microfil] composite, "Heliomolar Radiopaque" (Vivadent USA, Amherst, NY, USA; no batch # available at time ofplacement). The number and distribution of J&J Experimental and Marathon restorations were chosen to satisfythe guidelines ofADA acceptance in effect atthe time the study was begun (Council on Dental Materials, 1984). Accordingly, a minimum of two-thirds of the restorations was inserted into Class II cavity preparations, of which at least one-half involved molars. P-30 and Heliomolar Radiopaque were placed for comparison. The distributions of restorations by tooth and cavity type at placement for the four materials are shown in Table 1. The assignment of materials to study subject and tooth, and to clinical operator was made randomly. All restorations were placed by one of two experienced operators. For the restorations described in this paper, the cavity preparations were all of conventional design used for amalgam restorations and treatment was performed under optimal clinical conditions utilizing the rubber dam. The manufacturer's recommendations for each different material were followed throughout the various stages of tooth preparation, restoration placement, contouring, and finishing. With respect to cavity preparation, operator "one" utilized the butt joint configuration occlusal cavosurface finish line, while operator"two" placed an occlusal cavosurface bevel. Operator "one" placed 86 restorations while operator "two" placed 60 restorations. All exposed dentin was lined with a thin layer of calcium hydroxide (L. D. Caulk Div., Dentsply International Inc., Milford, DE, USA). Class II restorations were placed with the use ofa clear interproximal matrix (Contact Molar Bands, Vivadent USA) and anatomically formed wooden wedges, Each of the four composite materials was placed incrementally with a composite syringe (Centrix Inc. Stratford, CT, USA) after etching the enamel for 30 s with a gel etchant and placing a thin layer of the bonding agent supplied by the manufacturer. The composite placed in the proximal boxes of Class II preparations was polymer±zeal for the manufacturer's recommended time from the lingual, then from the buccal, and lastly from the occlusal direction. The occlusal surfaces of both Class I and Class II restorations and the buccal and lingual interproximal cavosurfaces were finished using 12fluted finishing burs and "Sof-lex" discs (3M Dental Products Division, St. Paul, MN, USA). Sof-Lexinterproximal finishing strips were also used to help establish interproximal contour. Collection of Data. The restorations were evaluated at baseline, 3, 6, 12, 24 and 36 months. This evaluation included the indirect assessment of occlusal wear through the use of stone replicas made from vinyl polysiloxane ("Express", 3M Dental Products Division) impressions of the experimental teeth. These replicas were independently scored by two examTABLE 1: SAMPLESIZEBYTOOTHAND CAVITYTYPEAT PLACEMENT

Material

Hetiomolar J&J Exp Marathon P-30 Total

Premolars Class l ClassII -1 2 -3

6 21 13 8 48

Molars Class l Classll 5 5 8 4 22

10 30 18 15 73

Total 21 57 41 27 146

TABLE 2: MEANOCCLUSALWEAR (pm) FORALL MATERIALS Recall Total Heliomolar J&J Exp Marathon P-30 Signif. (months) n p<

Baseline 128

3

90

6

98

12

142

24

111

36

105

8+17 (n=19) 7 + 13 (n=12)

25 + 18 (n=12) 34 + 21 (n=14) 49 ± 28 (n=14)

4+ 7 (n=54) 5 ± 13 (n=41)

18+68 (n=38) *37 ± 25 (n=25)

1+ 4 (n=l7) 11 ± 13 (n=12)

25 ± 19 '61 ± 39 18 2 12 (n=42) (n=26) (n=18) 37 ± 26 *89 + 68 28 + 18 (n=44) (n=32) (n=21) 48 2 33 "150± 77 39 ± 22 (n=44) (n=32) (n=21)

NS

0.01 0.01 0.01 0.01

52 + 28 54 ± 40 "174± 87 45 + 25 0.01 (n=12) (n=44) (n=29t (n=20) * Denotesmaterialthat is significantlydifferentfrom the otherthreematerials iners who were trained in the use of the M-L scale and calibrated prior to evaluation of the restorations (Lugassy and Moffa, 1985). The examiners in this study, as well as exam±ners from other centers, have been able to differentiate intermediary values between the standard scores. Each examiner determined the overall average score for each restoration at the cavosurface margin. The two examiners' scores for each restoration were then averaged to obtain a mean score for each restoration at each recall (mean M-L measurements). Intraoral (USPHS) evaluations of these restorations were also made during the recall visits. Those results allow a comparison between direct and indirect evaluation systems and are reported elsewhere (Freilich et. al., 1992). Statistical Analysis. The mean M-L measurements were used for data analysis. All data were recorded, entered, verified, and analyzed using an SPSS software package on an IBM PS-2 computer. The M-L measurements of wear were converted to microns and treated as parametric data. One- and two-way analysis of variance tests (ANOVA) were used to evaluate the clinical factors and the materials at each recall. TheLeast-SignificantTestwasusedtodeterminewhichofthe experimental groups demonstrated statistically significant differences for a given factor.

RESULTS The total sample size ranged from 90 to 142 restorations at the various recall examinations. Table 2 shows the mean wear of each material at all recall exams. Marathon exhibited significantly greater wear than the other three composite materials (p < 0.01) at all evaluations after baseline. At 36 months, the mean occlusal wear for Heliomolar, J & J Experimental and P30 restorations ranged from 45 to 54 Bm while Marathon restorations exhibited 174 ~m of wear. The mean wear values are plotted against time for each of the four materials in Fig. 1. This figure indicates that wear increases with time, but at a decreasing rate. The most rapid wear occurs in the first 12 months. This pattern is typical of posterior composites. There is a large standard deviation for each material at each recall (Table 2). This is also typical of posterior composite restorations and is due to many factors including the patient, tooth position, etc. For purposes of evaluating the effects of clinical parameters

Dental Materials~July 1992 225

TABLE 3:36 MONTHCOMPARISONOF MEANOCCLUSALWEAR(l~m) FOR LOW-WEARMATERIALS(J&J,P-30AND HELIOMOLAR)ANDA HIGHWEARMATERIAL(MARATHON)ACCORDINGTO THREECLINICAL PARAMETERS Material CavityClass Tooth Type FinishLine Class I ClassII Premolar Molar Butt-jointBevel Low-wear 41+23 54+37 44+29 55_+37 "41+29 *63+37

(n=18) (n=58)

(n=26) (n=50)

(n=41) (n=35)

High-wear 167+77 178+93 128+50 189+92 159+82 188+92 (n=lO) (n=19) (n=7) (n=22) (n=14) (n=15) * Denotesspecificgroupsthat are significantlydifferent from each other within the sameclinical parameterand type of material, such as cavity class, tooth type, and finish line configuration, Marathon was classified as a high wear-rate material and was evaluated separately. For this same purpose, Heliomolar, J & J Experimental, and P-30 were classified as low wear-rate materials. The effects ofcavity class, tooth type and finish line configuration at 36 months can be seen in Table 3. Neither cavity class nor tooth type had a statistically significant effect on occlusalwear for either the highorlowwear-rate materials, It should be noted, however, that the difference in wear between the molar and premolar restorations made with the high wear-rate material was 189 _+92 pm and 128 _+50 ~m, respectively, compared to 44 _+29 pm and 55 _+37 pm for the low-wear materials. A two-way ANOVA examining these factors produced an interaction between material and tooth type which was almost statistically significant (p=0.06). Low wear-rate materials placed into cavity preparations with a "bevelled" finish line configuration had significantly more wear than those restorations placed into preparations with a "butt-joint" finish line at p < 0.01 for all recalls except the baseline examination. This difference between bevelled and butt-joint finish lines was not observed for the high wearrate material. When comparisons were made within cavity class, there was no statistical difference found between molar or premolar restorations.TheocclusalwearmeasuredforClassIIpremolars

and molars was 45 _+29 ~m versus 61 _+41 pm, respectively, for the low wear-rate materials and 129 _+ 50 ~m versus 207 _+102 I~m for the high wear-rate material.

DISCUSSION The earlier formulations of posterior composite materials have been shown to wear at a much higher rate under masticatory function than more recently developed materials. Three of the four materials tested in this study exhibited less than 60 pm of wear after 36 months of clinical use. The fourth material exhibited over 170 #m of wear. The pattern of wear for all four materials is comparable to results obtained for earlier products, as illustrated in Figure 2. The data for the earlier products is representative and was reported five to seven years ago (Go]dberget al., 1984; Leinfelder et al., 1986a). It is inappropriate to make absolute comparisons among materials studied in different clinical trials because of the differences in the study designs of these trials. Nevertheless, it is interesting to note the similar patterns of wear exhibited by the restorations reported here and in the references cited above, despite the important differences in study design characteristics such as eligibility criteria, distribution of tooth and cavity type, and experimental material composition and wear rate. The prediction of long-term wear from initial data was first approximated by Leinfelder et al. (1986b). More recently, Tayloretal. (1991) and Bayne etal. (1991) developed a mathematical model for predicting long-term wear which also accounted for numerous clinical parameters. In the present work, the data was fit to least squares (McLain, 1974), log and exponential algorithms (Systat, 1989). The latter was used to make the curves seen in the figures, although all models exhibited adequate curve fitting to the data. The lowest correlation coefficient between the data and any model was 0.85. Beyond its use in predicting long-term clinical wear, associating a physical model with an exponential or log algorithm could be extremely valuable for identifying the mechanism of wear. While various clinical guidelines have been proposed for the successful use of posterior composite materials, the effects of parameters such as cavity class and tooth type remain unclear. Based upon six-month in vivo observations of both

2OO

5OO

150

400

0

0

10

'

20 Time (months)

'

30

40

Fig. 1. Mean wear versus time for the four materials evaluated in the present study. J & J Experimenlale, P- 30 x, Heliomolar ~,and Marathon o.

226 Freilich et aL/Occlusal wear of composites

0

0

lo

20 Time (months)

30

40

Fig. 2. Mean wear versus time for the four materials evaluated in the present study compared to earlier posterior composite resins. J & J Experimental O, p - 30 x, Heliomolar and Marathon <>(present data); Fulfill z~ (Leinfelder et al., 1986a); Experimental Materials II, D, and 0 (Goldberg et al., 1984)

hybrid and micro fill composites placed in Class I and II cavities, Lutz et al. (1984) concluded that larger restorations exhibit more wear than smaller restorations. Sturdevantet al. (1988) found that restorations placed in molar teeth exhibited more wear than restorations placed in premolars for two materials studied for a five-year period, and that Class II restorations exhibited significantly greater wear than Class I restorations for one of these materials. This is in contrast to the findings ofBoksman et al. (1986) which showed that tooth type had no effect on occlusal wear, and Rowe (1989) who observed that while molar restorations exhibited more wear than premolars, the size of the restoration and the cavity class had no significant effect. For the above studies which measured occlusal wear by an indirect method, the mean wear ranged from 129 to 164 mm for 36-to 60-month observations, It is possible that the effects of these clinical parameters are largely dependent upon the particular composite material studied and its general wear characteristics, For this study, the effect of clinical parameters such as cavity class and tooth type was evaluated for the three materials which exhibited low rates of occlusal wear separately from Marathon, which exhibited a high rate of occlusal wear. The data have shown that these parameters have no statistically significant effect on the occlusal wear of restorations made with all of these materials. There is, however, a strong suggestion of a clinically interesting difference between premolars and molars (128 _+ 50 vs. 189 +_ 92 ~m, respectively) when analyzed for the high wear-rate materials, When a two-wayANOVA was used to examine the parameters of material and tooth type simultaneously, it was found that the p value for this interaction was 0.062. This indicates that the location of the restoration (molar vs. premolar) may be of concern only for composite materials that exhibit inherently greater wear. For these analyses, the lack of statistical differences (p<0.05) are likely due to the lack of statistical power to detect real differences between these subgroups. This results from both the high standard deviation and relatively modest sample sizes available for these subgroup comparisons. In contrast, cavity class and tooth type appear to have no clinically important effect on the wear observed for the restorations composed of the low wear-rate materials (41 _+23 vs. 54 +_37 um for Class I and II, respectively, and 44 _+29 vs. 55 _+37 pm for premolars and molars, respectively). These data support the aforementioned notion that the overall wear characteristics of composite materials may, in fact, have an impact on the effect oftooth type and/or cavity class on occlusal wear. Specifically, it appears that only the composite materials that exhibit a high rate of occlusal wear are affected by the location of the restoration. Although beyond the scope of the present clinical trial, it would be warranted to conduct laboratory and SEM studies of the materials tested. These studies may be able to determine the factors(s) responsible for the differences in overall wear rates and the reason(s) why the high wear material was more sensitive to tooth position. Such information might help to support or refute current theories of the wear mechanism in dental composites. The need to place an occlusal bevelin the cavity preparation of posterior teeth undergoing restoration with composite materials is unclear. Although this bevel may expose more enamel rods for bonding, thereby minimizing microleakage (Moore and Vann, 1988), it may also contribute to the early marginal fracture of the composite restoration where the

material is of a thin dimension. A comparison of the occlusal wear exhibited by restorations placed into cavities with bevelled versus butt-joint occlusal cavosurface margins was also made in this study. This comparison was made separately for both low-and high-wear materials. Low wear-rate materials placed into cavity preparations with a bevelled finish line configuration had significantly more wear than those restorations placed into preparations with a butt-joint finish line at p < 0.01 for all recalls except the baseline examination. This differencebetweenbevelledandbutt-jointfinishlineswasnot observed for the high wear-rate material. Therefore, the overall greater wear of the high wear-rate material appeared to completely disguise the effect of the cavosurface margin configuration. The difference between the bevelled and buttjoint finish line configurations observed for the low-wear materials may be the result of the geometry or angulation of the bevel and unrelated to the rate of wear of the restorative material. The bevelled preparation results in the exposure of a wider band of prepared enamel surface available for M-L assessment ofa given amount ofocclusal wear when compared to the band of enamel exposed for a butt-joint configuration. In addition, it must be recognized that it is possible that the differences in the occlusal wear observed between restorations of these two different subgroups may be due to differences between the clinical operators who restored the study teeth. As stated earlier, clinical trials have shown that more recently developed posterior composites exhibit increased resistance to occlusal wear. While the overall results of these studies appear to be quite favorable, it must be noted that the posterior composite restorations evaluated in these studies were placed by skilled clinicians operating under ideal conditions. One must be cautious when generalizing the findings observed during controlled clinical studies to the results that can be predictably attained in clinical practice, where the time required to create ideal conditions can be cost-prohibitive. ACKNOWLEDGMENTS The authorswishtothankDr.RalphV.Katzforhisassistance with the data management and analysis. This investigation was supported in part by a University of Connecticut Health Center Clinical Research Center Grant, Den-Mat, Inc. and Johnson & Johnson. Received August 30, 1991/AcceptedNovember26 1991

Addresscorrespondenceand reprint requeststo: M.A. Freilich

Department ofProsthodontics University ofConnecticutSchoolofDental Medicine

Farmington,CT 06032USA REFERENCES Bayne SC, Taylor DF, Wilder AD, Heymann HO, Tangen CM (1991). Clinical longevity often posterior composite materials based on wear. J Dent Res 70:344, Abstr. No. 630. Boksman L, Jordon RE, Suzuki M, Charles DH (1986). A visible light-cured posterior composite resin: results of a 3year clinical evaluation. J A m DentAssoc 112:627-631. Council on Dental Materials, Instruments and Equipment, American Dental Association, 1984. Revised Guidelines for Submission of Composite Resin Materials for Occlusal Class I and Class II Restorations.

Dental Materials/July 1992 227

FreilichMA, GoldbergAJ, GilpatrickRO, SimonsenRJ(1992). Direct and indirect evaluation of posterior composite restorations at three years. Dent Mater 8:60-64. Goldberg AJ, Rydinge E, Santucci EA, Racz WB (1984). Clinical evaluation methods for posterior composite restorations. JDent Res 63:1387-1391. Leinfelder KF, Taylor DF, Barkmeier WW, Goldberg AJ (1986a). Quantitative wear measurement ofposterior cornposite resins. Dent Mater 2:198-201. LeinfelderKF, Wilder AD, TeixeiraLC (1986b). Rates of wear of posterior composite resins. J Am Dent Assoc 112:829-833. Leinfelder KF, Lemons JE. (1988). Clinical Restorative Materials and Techniques. Philadelphia: Lea & Febiger, 87. Leinfelder KF (1991). Using composite resin as a posterior restorative material. J A m Dent Assoc 122:65-70. Lugassy AA, Moffa JP (1985). Laboratory model for the quantification of clinical occlusal wear. JDent Res 64: 1 8 1 , Abstr. No. 63. Lundin S, Koch G (1989). Class I and II composite resin restorations: A 4-year clinical follow up. Swed Dent J 13:217-227.

228 Freilich et al./Occlusal wear of composites

Lutz F, Phillips RW, Roulet JF, Setcos JC (1984). In vivo and in vitro wear of potential posterior composites. JDent Res 63:914-920. McLain DH (1974). Drawing contours from arbitrary data points. The Computer J 17:318-324. Moore DH, Vann WFJR (1988). The effect of a cavosurface bevel on microleakage in posterior composite restorations. J Prosthet Dent 59:21-24. Phillips RW, Avery DR, Mehra R, Swartz ML, McCune RJ (1973).Observations on a composite resin for Class II restorations: Three-year report. J Prosthet Dent 30:891897. Rowe AHR (1989). A five year study of the clinical performance of a posterior composite resin restorative material. JDent 17 Suppl:s6-s9. Sturdevant JR, Lundeen TF, Wilder AD, Taylor DF (1988). Five-year study of two light-cured posterior composite resins. Dent Mater 4:105-110. Systat (1989). The System for Graphics, Evanston, IL: Wilkinson, Leland. 255. Taylor DF, Bayne SC, Wilder AD, Sturdevant JR (1991). Tooth contact versus Leinfelder wear values of posterior composites. J Dent Res 70:344, Abstr. No. 634.