- Email: [email protected]
Effect of restored and unrestored non-carious cervical lesions on the fracture resistance of previously restored maxillary premolar teeth
Journal of Dentistry Journal of Dentistry
26 (1998)
427-433
Effect of restored and unrestored non-carious cervical lesions on the fracture resistance of previously restored maxillary premolar teeth K.L. Osborne-Smith*, University
Dental
Hospital
F.J.T. Burke, T. MC Farlane, N.H.F. Wilson of Manchester,
Received
Higher
23 December
Cambridge
1996; accepted
Street, Manchester 29 May
Ml5
6FH,
UK
1997
Abstract Objectives: The effect of non-carious cervical lesions (NCCL) on tooth fracture resistance has not previously been investigated. The aims of this in vitro study were to examine the fracture resistance of a group of extracted maxillary premolar teeth with mesio-occlusal-distal (MOD) restorations of amalgam, and restored or unrestored simulated NCCL. Method: Forty sound maxillary premolar teeth were divided at random into four groups, each of 10 teeth, which were fixed crown uppermost and long axis vertical in stainless steel moulds. Groups 1,2,3 and 4 were prepared with standardized parallel-sided MOD cavities, then restored with amalgam. Groups 1,2 and 3 were further prepared with standardized NCCL. The NCCL in Group 1 were restored using a resin-modified polyalkenoate (glass-ionomer) cement, and the NCCL in Group 2 were restored with an adhesive composite resin system. The NCCL in Group 3 were left unrestored. The specimens were loaded compressively at 1 mm min-’ using a universal testing machine. Results: Mean fracture loads (KN) of 1.OS, 1.03, 0.98 and 1.14, respectively, were recorded for Groups 1, 2, 3 and 4. Two-way ANOVA and Scheffe’s Multiple Range Test showed no statistically significant difference between the groups. Conclusions: It is concluded that the presence of a standardized NCCL in an extracted maxillary premolar tooth does not reduce the fracture resistance of the tooth when loaded compressively at 1 mm min-‘. The restoration of NCCL with the materials tested did not result in an increase in the fracture resistance of the previously restored premolar teeth, when loaded compressively at 1 mm min-‘. 0 1998 Elsevier Science Ltd. All rights reserved. Keywords:
cervical; abfraction; erosion; abrasion; tooth flexure; tooth fracture
1. Introduction The pathogenesis of the non-carious cervical lesion (NCCL) has been a cause of discussion for almost a century, having been studied by G. V. Black in 1908[1] and subsequently by numerous workers. Traditionally, NCCL have been considered to be caused by a combination of erosion and abrasion[2,3]. The prevalence of cervical-erosion lesions is not well documented, although in 1949 Zipkin and McClure found that 27% of the 83 patients whom they studied had loss of tooth tissue attributable to erosion on the facial surfaces of teeth[4]. More recently, Sognnaes and co-workers found that 18% of the 10 000 extracted teeth which they examined were affected by erosive lesions, although their classification included lesions which would today be classified as lesions caused by abrasion[5]. * Correspondence should be addressed to: Mrs K. L. Osborne-Smith, University Dental Hospital of Manchester, Higher Cambridge Street, Manchester Ml5 6FH, UK. Tel: 0161 275 6718; Fax: 0161 275 6710. 0300.5712/98/$19.00 0 1998 Elsevier PZI SO300-57 12(97)00029-8
Science
Ltd. All rights
reserved
Bergstrom and Lavstedt reported an incidence of 31% of abrasive lesions in their cross-sectional study of 818 patients[6]. The role of acids in the aetiology of NCCL has been well documented, but it has been generally recognized that erosion and abrasion may act together in viva, and that there is some difficulty in distinguishing between erosion and abrasion lesions[7]. There is evidence that abrasion may produce NCCL in vitro, as demonstrated by Manly in 1944[8]. A number of authors have suggested that factors other than those traditionally considered in the pathogenesis of the NCCL may play a part in the development of such lesions. Among these authors are Lee and Eakle[9] who proposed that, in an ideal occlusion, functional loads are directed along the long axis of the tooth, while in eccentric movements the side towards which the tooth is bending is under compressive stress and the side opposite the force is under tension. Such occlusal loading causes deformation and flexure of the tooth resulting in disruption of the enamel
428
K.L. Osborne-Smith
Fig. 1. A non-carious cervical lesion prepared in the buccal premolar tooth, using a 110” tapered diamond bur.
et al./Journal
surface of a
crystals at the cervical region, thereby contributing to the formation of a NCCL. The damage caused by this process has been termed an “abfraction” by Grippo[lO]. Abfractions are typically wedge-shaped with sharp margins, and may have scratched or smooth surfaces according to whether there has been secondary abrasive or erosive action on the dentine. Secondary aetiological factors, such as toothbrush abrasion and acid erosion, may then cause further tooth substance loss with enlargement of the NCCL. Lee and Eakle proposed that the region of greatest stress on eccentric loading of a tooth is located at the fulcrum, corresponding to the shape and position of a NCCL close to the gingival margin[9]. Treatment of NCCL may be limited to palliative therapy if the causative factors have been eliminated and the clinician is able to monitor the NCCL on a regular basis. However, some clinical evidence suggests that restoration of these lesions may be necessary to prevent their enlargement[ 111. Primary indications for the treatment of NCCL include sensitivity, poor aesthetics and food stagnation. Consideration of the aetiology of the lesion and prevention should precede treatment and, where necessary, occlusal adjustment should be carried out to alter potentially harmful eccentric occlusal contacts[l2]. Restoration of a
of Dentistry
26 (1998)
427-433
NCCL is ideally by the use of adhesive materials, namely a glass-ionomer cement or a composite resin and a dentinebonding system[l3,14]. Studies on the deformation of teeth subjected to simulated occlusal loading have shown that removal of tooth structure causes increased cuspal flexure[l5]. The restoration of tooth structure using an adhesive acid-etch technique has been shown in strain-gauge studies to restore tooth stiffness significantlyll61. It has been suggested that when a NCCL is restored, the flexure of the restored tooth under load is reduced, thereby strengthening the teeth[ Ill. The method of compressive loading the premolar teeth in this study resulted in a point of application of the load to the cuspal inclines of the restoration, simulating eccentric occlusal loading. Such a method of loading exerts a compressive load to the restoration as well as to the tooth. A smaller diameter bar has previously been used to apply a compressive load to the central occlusal fossa simulating centric loading[ 171. Other methods of investigating stresses in teeth include finite element analysis, strain gauges and photo-elastic studies. Conflicting results from studies on tooth fracture resistance and stiffness of restored teeth may, in part, be explained by variations in test methods. The effect of NCCL on tooth fracture resistance has not previously been investigated. It is the aim of this study to examine the effect of restored and unrestored non-carious cervical lesions on the fracture resistance of maxillary premolar teeth containing moderately sized mesio-occlusaldistal (MOD) restorations of dental amalgam.
2. Materials
and method
2.1. Selection of teeth Forty extracted maxillary premolar teeth were selected and examined visually and by transillumination to ensure that they were sound and free from defects, cracks and tooth substance loss. Any calculus deposits and soft tissue were removed from the teeth selected using a hand scaler. The teeth were divided into four groups, each of 10 teeth so that the mean measurement of the bucco-palatal width (BPWthe distance from the maximum convexity on the buccal and palatal surfaces) between and within each group varied by no more than 2.5%. Following post-extraction storage in buffered formal saline for 24 h, the teeth were stored in tap water at room temperature (20°C) except when aspects of the experimental procedure required isolation from moisture. Each tooth was fixed, crown uppermost and long axis vertical in a stainless steel mould (15 mm X 15 mm X 15 mm) which had a central cylindrical hole of I2-mm diameter, using chemically cured resin. The resin extended to within 2.5 mm of the cemento-enamel junction (CEJ).
K.L. Osborne-Smith
et al.Nournal
2.2. Experimental variables The teeth included restored as follows:
in the study were
prepared and
1. Standardized MOD amalgam restoration, NCCL restored with Vitremer (3M Healthcare, St. Paul, MN, USA). Group 2. Standardized MOD amalgam restoration, NCCL restored with Scotchbond Multipurpose Plus (SBMP + , 3M Healthcare) and 2100 (3M Healthcare). Group 3. Standardized MOD amalgam restoration, NCCL unrestored. Group 4. Standardized MOD amalgam restoration, no NCCL preparation. Group
2.3. Preparation and restoration of specimens 2.3.1. Amalgam restorations An amalgam cavity design was chosen which was considered to be typical for a moderately sized mesioocclusal-distal preparation[ 181 for amalgam in maxillary premolar teeth. The features of the preparation were as follows: width of proximal boxes and occlusal isthmus, l/2 BPW; depth of occlusal isthmus from buccal cusp tip, 5 mm; preparation walls, parallel; gingival margin, 1 mm above CEJ; gingival floor depth, 2 mm; pulpo-axial line angle, bevelled. The preparation was carried out using a parallel-sided, diamond bur (Hi-Di 648; Dentsply, Weybridge, UK) operating in a high-speed handpiece with water coolant, followed by a stainless-steelround bur (size 4, IS0 14, Dentsply, Weybridge, UK) in a slow handpiece (20000 rpm) without water coolant. The preparation was carried out under simulated surgery conditions, the measurementsbeing checked using RS Vernier calipers. The preparations were finished to exact dimensions using a parallel-sided round ended bur (DK Holdings, Staplehurst, UK), without water coolant, in a laboratory handpiece at a maximum speed of 8000 rpm. The handpiece was held in a Bachmann design parallelometer (Cendres et Metaux S.A., Bienne, Switzerland) which was used in conjunction with a Perthometer S8P (Feinpruf, Perthen, GmbH, Gottingen, Germany) to enable parallel cavity walls to be prepared. Finally, the internal line and point angles were rounded. The completed preparations were restored with a high copper amalgam (Dispersalloy; Dentsply, Weybridge, UK), packed using a G-packer with maximum compression against a Siqveland matrix. The restorationswere carved to the original dimensionsand shapeof the tooth by the useof a pre-preparation polyvinylsiloxane putty index.
of Dentistry
26 (1998)
427-433
429
2.3.2. Preparatiolz of cervical lesions V-shaped NCCL were preparedin the buccal surfacesof the teeth of Groups 1,2 and 3 in accordancewith the criteria defined by Bayne (personal communication, 1994) using a specially designed 110” tapered bur (DK Holdings, Staplehurst, UK) held in a laboratory handpiece, operating at 8000 rpm without water coolant. The specimenswere held rigidly in a parallelometer. A depth gauge was used to allow the cavities to be preparedto a depth of 2 mm from the CEJ, with the bur being run along the tooth in a mesialto distal direction to create a notch-shapeddefect (Fig. 1). 2.3.3. Restoration of NCCL The NCCL in the teeth of Group 1 were restored to original dimensionsusing Vitremer (3M Healthcare) which was dispensed,mixed and applied in accordance with the manufacturer’s instructions. The NCCL in the teeth in Group 2 were treated with the three constituents of the Scotchbond MultiPurpose Plus system (SBMPS) (3M Healthcare), and restored to original dimensions using ZlOO (Shade A2) restorative (3M Healthcare), with each material being used in strict accordancewith the manufacturer’s instructions. The NCCL in the teeth in Group 3 were left unrestored. 2.4. Fracture testing The restored teeth were stored in tap water at room temperature for 24 h prior to testing. They were then subjected to compressive loading at a cross-head speed of 1 mm min-’ in a Universal Testing Machine (RPC Howden Ltd., Leamington Spa, UK). Compressiveforce was applied by means of a 4-mm diameter steel bar placed along the midline fissure of the tooth. The force (KN) required to cause fracture was recorded. Articulating paper was used between the specimen and the steel bar to ascertain the exact point of contact and loading of the tooth/restoration occlusal surface. The fracture strength results were subjected to statistical analysisby two-way analysisof variance (ANOVA) and Scheffe’s Multiple Range Test at a significance level of 0.05. The fractured specimenswere examined to determinethe mode of fracture, using a classification designed for the investigation (Fig. 2). The shadedarea in Fig. 2 represents the fracture site. They were also examined for evidence of debonding or loss of the NCCL restorations. The results were recorded and the fracture mode analysed, considering the fractures in order of severity, using Kruskal-Wallis oneway analysis of variance by ranks (SPSS package).
3. Results Mean fracture loads (KN) of 1.08, 1.03, 0.98 and 1.14 were recorded for Groups 1, 2, 3 and 4, respectively (Table 1, Fig. 3). Two-way ANOVA and Scheffe’s Multiple
430
K.L. Osborne-Smith
MODE I la
et al./Jourtaal
Description Midline Occlusal fissure fracture
0
Amalgam and/or tooth fracture-- not midline
Group
Preparation/restorative materials used
n
Mean
s.d.
1
MOD/amalgam and NCCUVitremer (3M) MOD/amalgam and NCCL/ZlOO (3M) MOD/amalgam and NCCL unrestored MOD/amalgam
10
1.084
0.24
10
1.029
0.28
10
0.976
0.34
10
1.14
0.34
2 3
Buccal cusp fracture - sound tooth or amalgam remaining
4
Severe amalgam and/or tooth fracture involving more than one cusp
Vitremer (Group 1) or ZlOO (Group 2) resulted in fewer buccal cusp fractures but more palatal cusp fractures, compared with those teeth with unrestoredNCCL. No Vitremer (3M Healthcare) or ZlOO (3M Healthcare) restorations were lost during or after the compressive loading.
Buccal cusp and amalgam fracture
4. Discussion
Palatal cusp fracture - sound tooth or amalgam remaining
This in vitro investigation was designed to study the effects of restored and unrestored NCCL on the fracture resistanceof maxillary premolar teeth, restored with moderately sized MOD amalgamrestorations.It was anticipated that the amalgam restorations would simulate the clinical situation in which older patients typically present,with teeth with both NCCL and other restorations. Clinical conditions were simulatedwhere possiblethroughout the investigation and the specimenswere removed from the water-storage medium only when operative procedures were necessary. To allow comparisons to be made between and within groups, it was necessary to select teeth of standardized bucco-palatal width, to standardize the preparations and subsequentrestorative procedures and to standardize the application of force applied to the specimens. A 4-mm
Palatal cusp and amalgam fracture
Fig. 2. Classification when compressively
26 (1998) 427-433
Table 1 The mean fracture loads (KN) required to induce fracture in restored premolar teeth (Groups l-4), when compressively loaded at 1 mm min-’
lb Midline amalgam fracture
of Dentistry
of the mode of fracture loaded at 1 mm min-‘.
of restored
premolar
teeth
RangeTest showedno statistically significant differences (P > 0.05) in mean fracture load between the groups. The frequency of the mode of fracture of the teeth in Groups 1-4 is shown graphically in Fig. 4. The majority of fractures in Groups l-4 were ClassIa (n = 23), i.e. a midline amalgamfracture in a bucco/palatal direction. The restoration of the NCCL in Groups 1 and 2 with either 1.4 B z
q GPl 5jGP2 GP3 n GP4
1.2
5
1
E
0.8
z =) 0.6 b d 0.4 LL $ 0.2 Y
Fig. 3. The mean fracture
0
loads (KN)
GP GP GP GP
1 2 3 4
-
required
Standafdised Standardiid Standardised Standardised
MOD MOD MOD MOD
amalgam amalgam amalgam amalgam
to induce fracture
restoration, restoration, restoration, restoration
in restored
NCCL NCCL NCCL
premolar
restored with vitremer (3M Dental Products) restored with 2100 (3M Dental Products) unrestored
teeth (Groups
l-4),
when compressively
loaded at 1 mm min-’
K.L. Osborne-Smith
01’
et al./Journal
of Dentistry
iiwm-nn’
Noto Modes of fmcium in which a total of one or less specimens
of the mode of fracture
of restored
steel bar was used to apply a compressiveforce in an axial direction until fracture. Analysis of the resultsindicated that the majority of the forces applied were within the range of biting forces reported by Gibbs[l9]. Accordingly, the loads applied may be considered to be commensuratewith those applied during normal chewing. The teeth used were sound but of unknown age. As age may affect the fracture strength of teeth, 10 teeth were included in each group and the mean fracture strengths were used for comparison. While the NCCL preparations usedin the study were comparablewith thoseseenclinically (vi& in@), it is likely that in viva somereparative dentine may be laid down beneath areasaffected by NCCL. The potential weakening effect of the NCCL preparations in the present study may therefore have had a more profound effect upon the fracture resistance than in the in viva situation, where the teeth may have been strengthenedby reparative dentine formation. In a clinical study examining 457 NCCL in 101 patients, the internal angle was found to be 90-130” in 48% of cases, with Bayne reporting 110” to be the median internal angle[20]. Other workers]211 have reported the mean NCCL internal angle to be 105 t 27”. The standardized cervical preparation carried out in the present investigation may therefore be considered to be representative of the typical NCCL seen in the clinical situation. In addition, sinceBayne et aZ.[20] reported that 62% of NCCL occurred in the maxillary arch, the choice of maxillary premolar teeth may be consideredappropriate. The choice of a resin-modified glass-ionomercement and a compositeresin together with an enamel-dentine bonding system for the restoration of the NCCL was considered appropriate given the widespread use of such materials in Class V situations and their successin cervical cavities reported previously[ 13,141. Grippo has suggestedthat the restoration of NCCL would
431
427-433
GP4 GP3 ii -Amalgam/tooth fracture - not midline VII -Amalgam and paiatai cusp fracture
GPl GP2 I - Midllne amalgam fracture (buccopaiati) Vi - Amalgam and buccai cusp fracture
Fig. 4. The frequency
26 (1998)
fractured
premolar
from Group 1-4, have been excluded
teeth when compressively
for clarity.
loaded at 1 mm min-‘.
reduce flexure of teeth under occlusal load and thereby strengthenthe affected teeth[ 111.This strengthening effect was not noted in the present investigation, given that no statistically significant differences (P > 0.05) were found in the meanload required to causefracture amongor within the Groups l-4. However, the mean fracture loads for the groups of teeth with restoredNCCL (Groups 1 and 2) were higher than for the groups of teeth with unrestored NCCL (Group 3), although this difference was not statistically significant. Although the mean fracture strengths were not statistically significant, a trend towards greater fracture resistance is apparent in specimenswith restored NCCL when compared to specimenswith unrestored NCCL. A larger sample of teeth may have allowed the statistical tests to demonstrate a significant difference between the groups.The low power (23%) usedin the statistical analysis did not show a statistically significant difference in fracture strength of the four groups in this investigation. An increased number of specimensin each group (n = 87) would be necessaryto give 90% power, to detect a statistically significant difference in fracture strength at a significance level of 5%. However, the lengthy time required to produce eachspecimen(ca. 70 min) and the limited availability of suitable sound extracted teeth of appropriate dimension did not allow larger group sizes to be used in this study. The MOD restoration of amalgam may have been expected to predisposethe teeth to fracture, thereby emphasizing the effects of the NCCL preparation. However, from the resultsof a related investigation by the sameauthors(not published), it would appear that teeth with MOD amalgam restorations showed a greater resistance to fracture than teeth without MOD restorationsof amalgam.These findings are consistent with the work of others[17,22], with Watts and co-workers reporting that the greater force required to fracture teeth restored with amalgam may relate to the
432
K.L. Osborne-Smith
et al./Journal
plastic flow of directly loaded restorations of amalgam[ 171. These findings may be explained by the mode of application of the load to the occlusal surface of the restored teeth. Examination of the specimen teeth post-loading revealed that the teeth were loaded centrically as occlusal indicator marks were apparent on the central fossa of the amalgam restorations, while prior to loading the steel bar was found to rest on the cuspal inclines of the restored occlusal surfaces. The amalgam, being plastic, may have absorbed some of the applied load and, as a consequence, less load may have been applied to the tooth structure. Amalgam has a higher modulus of elasticity than dentine and, consequently, when loaded is able to withstand increased loads before fracture[23]. The present investigation involved the preparation of standardized NCCL in selected premolar teeth to enable comparisons to be made between groups of teeth in which the NCCL had been restored and not restored. It may be anticipated that application of an axial load would cause compression and flexure of the tooth specimens and lead to the debonding of the cervical restorations. However, none of the Vitremer or ZIOO restorations were lost when the fracturing force was applied. The restorations were placed under optimum conditions and on optimum substrate, namely, open dentinal tubules and intertubular dentine with no pulpal pressure or pulpal exudate. Nevertheless, the results indicate that the bonding strength of the restorations was sufficient to withstand the potential debonding forces resulting from tooth flexure. The results may have differed if a different method of loading had been applied, such as eccentric loading or fatigue testing, resulting in a different stress distribution within the restored tooth. Preservation of tooth structure is of primary importance to patient and clinician, and consideration of the mode of fracture of teeth following restoration is therefore necessary. It may be considered that fracture of a restoration is preferable to a catastrophic tooth fracture which may not readily permit restoration of the tooth. The non-parametric test used showed no difference in the distribution of fracture modes between the four groups. In the present investigation, the majority of fractures in Groups l-4 were of Mode Ia, i.e. the amalgam fractured in the midline in a bucco-palatal direction. This mode of fracture is seen in clinical practice and may be the result of centric loading, accepting that masticatory forces are applied at a greater speed than 1 mm min-‘. When the buccal cusp was sound (Group 4) or the NCCL had been restored (Groups 1 and 2), a greater number of palatal cusp and amalgam fractures (Mode VII) was seen than buccal cusp with amalgam fractures (Mode VI). Conversely, when the specimen teeth had an unrestored NCCL in addition to an MOD amalgam (Group 3), the buccal cusp and amalgam fractured (Mode VI) more frequently than the palatal cusp and amalgam (Mode VII). It may be considered that the presence of an unrestored NCCL weakens the buccal cusp, which may then fracture in preference to the more frequent mode of fracture involving the palatal cusp.
of Dentistry
26 (1998)
427-433
Finally, 3 of the 40 teeth demonstrated severe amalgam/ tooth fractures. No correlation was apparent between the load required to fracture these specimens and the severity of the fractures observed. It may be considered that such extreme results are due to normal biological variation, unseen damage during extraction, variations in pulp space or unseen fatigue stresses.
5. Conclusion It is concluded that the presence of a standardized NCCL of the dimensions used in this investigation in extracted maxillary premolar teeth with standardized MOD restorations does not reduce fracture resistance of the tooth under compressive loading at 1 mm min-‘. It is also concluded that the restoration of NCCL with Vitremer (3M) or ZlOO (3M) does not improve the fracture resistance of the specimen teeth when loaded compressively at 1 mm min-‘. Differences in mode of fracture were not noted between specimens with restored NCCL compared to specimens with unrestored NCCL. However, the teeth with restored NCCL suffered fewer fractures of the buccal cusp.
Acknowledgements The authors acknowledge for this study.
the support of 3M Healthcare
References [l] Black, G. V., A Work on Operative Dentistry. Pathology of Hard Tissues of Teeth, 1st edn. Medico-Dental Publishing, Chicago, IL, 1908, pp. 39-59. [2] Manly R. S. The abrasion of cementurn and dentine by modern dentrifices. Journal of Dental Research, 1941;20:583-595. [3] Eccles J. D. Dental erosion of non-industrial origin. A clinical survey and classification. Journal of Prosthetic Dentistry, 1979;42:649-653. [4] Zipkin I., McClure F. .I. Salivary citrate and dental erosion. Journal of Dental Research, 1949;28:613-626. [5] Sognnaes R. F., Wolcott R. B., Xhonga F. A. et al. Erosion-like patterns occurring in association with other dental conditions. Journal of the America1 Dental Association, 1972;84:571-576. [6] Bergstrom J., Lavstedt S. An epidemiological approach to toothbrushing and dental abrasion. Community Dentistry and Oral Epidemiology, 1979;7:57-64. [7] Sognnaes R. F., Dental hard tissue destruction with special reference to idiopatic erosions. In Mechanisms ofHard Tissue Destruction, ed. R. F. Sognnaes. American Association for the Advancement of Science, 1963, pp. 91-149. [8] Manly R. S. Factors influencing tests on the abrasion of dentin by brushing with dentrifices. Journal of Dental Research, 1944;23:59-72. [9] Lee W. C., Eakle W. S. Possible role of tensile stress in the etiology of cervical erosive lesions of teeth. Journal of Prosthetic Dentistry, 1984;52:374-380. [lo] Grippo J. 0. Abfractions; a new classification of hard tissue lesions of teeth. Journal of Esthetic Dentistry, 1991;3: 14-19. [l l] Grippo J. 0. Non-carious cervical lesions: the decision to ignore or
K.L.
Osborne-Smith
et al./Joumal
restore. Journal of Esthetic Dentistry, 1992;4(Suppl.):55-64. [12] Hand J. S., Hunt R. .I., Reinhardt .I. W. The prevalence and treatment implications of cervical abrasion in the elderly. Gerodontics, 1986;2:167-170. [13] Heymann H. O., Sturdevant J. R., Bayne S. et al. Examining tooth flexure effects on cervical restorations: a two-year clinical study. Journal of the American Dental Association, 1991;122:41-47. [14] Powell L. V., Johnson G. H., Gordon G. E. Factors associated with clinical success of cervical abrasion/erosion restorations. Operational Dentistry, 1995;20:7-13. [15] Douglas, W. H., Methods to improve fracture resistance of teeth. In International
Symposium
on
Posterior
Composite
Resin
of Dentistry
[18] [19]
[20]
[21]
Dental
Restorative Materials, ed. G. Vanherle and D. C. Smith. 3M, St Paul, 1985, pp. 433-441. [16] Morin D., DeLong R., Douglas W. H. Cusp reinforcement by the acid etch technique. Journal of Dental Research, 1984163: 1075-1078. [17] Watts D. C., EL-Mowafy 0. M., Grant A. A. Fracture resistance of
[22]
[23]
26 (1998)
427-433
433
lower molars with Class I composite and amalgam restorations. Dental Materials, 1987;3:261-264. Burke F. J. T., Investigation of aspects of optimum cavity design for composite inlays. M.Sc. thesis, University of Manchester, 1991. Gibbs C. H., Mahan P. E., Lundeen H. C. et al. Occlusal forces during chewing and swallowing as measured by sound transmission. Journal of Prosthetic Dentistry, 1981;46:443-449. Bayne S. C., Heymann H. O., Wilder A. D. et al. Substrate variables in dentin bonding systems for Class V restorations. Journal of Dental Research, 1994;73(1388):275. Yoon K. .I., Sturdevant J. R. Geometry of Class V cervical abrasion/ erosion lesions. Journal of Dental Research, 1994;73(1783):325. Stampalia L. L., Nicholls J. I., Brudvik J. S. et al. Fracture resistance of teeth with resin-bonded restorations. Journal of Prosthetic Dentistry, 1986;55:694-698. Goel V. K., Khera S. C., Senthil G. et al. Effect of cavity design on stresses in first molars. Journal of Dental Research, 1985;64:350.
26 (1998)
427-433
Effect of restored and unrestored non-carious cervical lesions on the fracture resistance of previously restored maxillary premolar teeth K.L. Osborne-Smith*, University
Dental
Hospital
F.J.T. Burke, T. MC Farlane, N.H.F. Wilson of Manchester,
Received
Higher
23 December
Cambridge
1996; accepted
Street, Manchester 29 May
Ml5
6FH,
UK
1997
Abstract Objectives: The effect of non-carious cervical lesions (NCCL) on tooth fracture resistance has not previously been investigated. The aims of this in vitro study were to examine the fracture resistance of a group of extracted maxillary premolar teeth with mesio-occlusal-distal (MOD) restorations of amalgam, and restored or unrestored simulated NCCL. Method: Forty sound maxillary premolar teeth were divided at random into four groups, each of 10 teeth, which were fixed crown uppermost and long axis vertical in stainless steel moulds. Groups 1,2,3 and 4 were prepared with standardized parallel-sided MOD cavities, then restored with amalgam. Groups 1,2 and 3 were further prepared with standardized NCCL. The NCCL in Group 1 were restored using a resin-modified polyalkenoate (glass-ionomer) cement, and the NCCL in Group 2 were restored with an adhesive composite resin system. The NCCL in Group 3 were left unrestored. The specimens were loaded compressively at 1 mm min-’ using a universal testing machine. Results: Mean fracture loads (KN) of 1.OS, 1.03, 0.98 and 1.14, respectively, were recorded for Groups 1, 2, 3 and 4. Two-way ANOVA and Scheffe’s Multiple Range Test showed no statistically significant difference between the groups. Conclusions: It is concluded that the presence of a standardized NCCL in an extracted maxillary premolar tooth does not reduce the fracture resistance of the tooth when loaded compressively at 1 mm min-‘. The restoration of NCCL with the materials tested did not result in an increase in the fracture resistance of the previously restored premolar teeth, when loaded compressively at 1 mm min-‘. 0 1998 Elsevier Science Ltd. All rights reserved. Keywords:
cervical; abfraction; erosion; abrasion; tooth flexure; tooth fracture
1. Introduction The pathogenesis of the non-carious cervical lesion (NCCL) has been a cause of discussion for almost a century, having been studied by G. V. Black in 1908[1] and subsequently by numerous workers. Traditionally, NCCL have been considered to be caused by a combination of erosion and abrasion[2,3]. The prevalence of cervical-erosion lesions is not well documented, although in 1949 Zipkin and McClure found that 27% of the 83 patients whom they studied had loss of tooth tissue attributable to erosion on the facial surfaces of teeth[4]. More recently, Sognnaes and co-workers found that 18% of the 10 000 extracted teeth which they examined were affected by erosive lesions, although their classification included lesions which would today be classified as lesions caused by abrasion[5]. * Correspondence should be addressed to: Mrs K. L. Osborne-Smith, University Dental Hospital of Manchester, Higher Cambridge Street, Manchester Ml5 6FH, UK. Tel: 0161 275 6718; Fax: 0161 275 6710. 0300.5712/98/$19.00 0 1998 Elsevier PZI SO300-57 12(97)00029-8
Science
Ltd. All rights
reserved
Bergstrom and Lavstedt reported an incidence of 31% of abrasive lesions in their cross-sectional study of 818 patients[6]. The role of acids in the aetiology of NCCL has been well documented, but it has been generally recognized that erosion and abrasion may act together in viva, and that there is some difficulty in distinguishing between erosion and abrasion lesions[7]. There is evidence that abrasion may produce NCCL in vitro, as demonstrated by Manly in 1944[8]. A number of authors have suggested that factors other than those traditionally considered in the pathogenesis of the NCCL may play a part in the development of such lesions. Among these authors are Lee and Eakle[9] who proposed that, in an ideal occlusion, functional loads are directed along the long axis of the tooth, while in eccentric movements the side towards which the tooth is bending is under compressive stress and the side opposite the force is under tension. Such occlusal loading causes deformation and flexure of the tooth resulting in disruption of the enamel
428
K.L. Osborne-Smith
Fig. 1. A non-carious cervical lesion prepared in the buccal premolar tooth, using a 110” tapered diamond bur.
et al./Journal
surface of a
crystals at the cervical region, thereby contributing to the formation of a NCCL. The damage caused by this process has been termed an “abfraction” by Grippo[lO]. Abfractions are typically wedge-shaped with sharp margins, and may have scratched or smooth surfaces according to whether there has been secondary abrasive or erosive action on the dentine. Secondary aetiological factors, such as toothbrush abrasion and acid erosion, may then cause further tooth substance loss with enlargement of the NCCL. Lee and Eakle proposed that the region of greatest stress on eccentric loading of a tooth is located at the fulcrum, corresponding to the shape and position of a NCCL close to the gingival margin[9]. Treatment of NCCL may be limited to palliative therapy if the causative factors have been eliminated and the clinician is able to monitor the NCCL on a regular basis. However, some clinical evidence suggests that restoration of these lesions may be necessary to prevent their enlargement[ 111. Primary indications for the treatment of NCCL include sensitivity, poor aesthetics and food stagnation. Consideration of the aetiology of the lesion and prevention should precede treatment and, where necessary, occlusal adjustment should be carried out to alter potentially harmful eccentric occlusal contacts[l2]. Restoration of a
of Dentistry
26 (1998)
427-433
NCCL is ideally by the use of adhesive materials, namely a glass-ionomer cement or a composite resin and a dentinebonding system[l3,14]. Studies on the deformation of teeth subjected to simulated occlusal loading have shown that removal of tooth structure causes increased cuspal flexure[l5]. The restoration of tooth structure using an adhesive acid-etch technique has been shown in strain-gauge studies to restore tooth stiffness significantlyll61. It has been suggested that when a NCCL is restored, the flexure of the restored tooth under load is reduced, thereby strengthening the teeth[ Ill. The method of compressive loading the premolar teeth in this study resulted in a point of application of the load to the cuspal inclines of the restoration, simulating eccentric occlusal loading. Such a method of loading exerts a compressive load to the restoration as well as to the tooth. A smaller diameter bar has previously been used to apply a compressive load to the central occlusal fossa simulating centric loading[ 171. Other methods of investigating stresses in teeth include finite element analysis, strain gauges and photo-elastic studies. Conflicting results from studies on tooth fracture resistance and stiffness of restored teeth may, in part, be explained by variations in test methods. The effect of NCCL on tooth fracture resistance has not previously been investigated. It is the aim of this study to examine the effect of restored and unrestored non-carious cervical lesions on the fracture resistance of maxillary premolar teeth containing moderately sized mesio-occlusaldistal (MOD) restorations of dental amalgam.
2. Materials
and method
2.1. Selection of teeth Forty extracted maxillary premolar teeth were selected and examined visually and by transillumination to ensure that they were sound and free from defects, cracks and tooth substance loss. Any calculus deposits and soft tissue were removed from the teeth selected using a hand scaler. The teeth were divided into four groups, each of 10 teeth so that the mean measurement of the bucco-palatal width (BPWthe distance from the maximum convexity on the buccal and palatal surfaces) between and within each group varied by no more than 2.5%. Following post-extraction storage in buffered formal saline for 24 h, the teeth were stored in tap water at room temperature (20°C) except when aspects of the experimental procedure required isolation from moisture. Each tooth was fixed, crown uppermost and long axis vertical in a stainless steel mould (15 mm X 15 mm X 15 mm) which had a central cylindrical hole of I2-mm diameter, using chemically cured resin. The resin extended to within 2.5 mm of the cemento-enamel junction (CEJ).
K.L. Osborne-Smith
et al.Nournal
2.2. Experimental variables The teeth included restored as follows:
in the study were
prepared and
1. Standardized MOD amalgam restoration, NCCL restored with Vitremer (3M Healthcare, St. Paul, MN, USA). Group 2. Standardized MOD amalgam restoration, NCCL restored with Scotchbond Multipurpose Plus (SBMP + , 3M Healthcare) and 2100 (3M Healthcare). Group 3. Standardized MOD amalgam restoration, NCCL unrestored. Group 4. Standardized MOD amalgam restoration, no NCCL preparation. Group
2.3. Preparation and restoration of specimens 2.3.1. Amalgam restorations An amalgam cavity design was chosen which was considered to be typical for a moderately sized mesioocclusal-distal preparation[ 181 for amalgam in maxillary premolar teeth. The features of the preparation were as follows: width of proximal boxes and occlusal isthmus, l/2 BPW; depth of occlusal isthmus from buccal cusp tip, 5 mm; preparation walls, parallel; gingival margin, 1 mm above CEJ; gingival floor depth, 2 mm; pulpo-axial line angle, bevelled. The preparation was carried out using a parallel-sided, diamond bur (Hi-Di 648; Dentsply, Weybridge, UK) operating in a high-speed handpiece with water coolant, followed by a stainless-steelround bur (size 4, IS0 14, Dentsply, Weybridge, UK) in a slow handpiece (20000 rpm) without water coolant. The preparation was carried out under simulated surgery conditions, the measurementsbeing checked using RS Vernier calipers. The preparations were finished to exact dimensions using a parallel-sided round ended bur (DK Holdings, Staplehurst, UK), without water coolant, in a laboratory handpiece at a maximum speed of 8000 rpm. The handpiece was held in a Bachmann design parallelometer (Cendres et Metaux S.A., Bienne, Switzerland) which was used in conjunction with a Perthometer S8P (Feinpruf, Perthen, GmbH, Gottingen, Germany) to enable parallel cavity walls to be prepared. Finally, the internal line and point angles were rounded. The completed preparations were restored with a high copper amalgam (Dispersalloy; Dentsply, Weybridge, UK), packed using a G-packer with maximum compression against a Siqveland matrix. The restorationswere carved to the original dimensionsand shapeof the tooth by the useof a pre-preparation polyvinylsiloxane putty index.
of Dentistry
26 (1998)
427-433
429
2.3.2. Preparatiolz of cervical lesions V-shaped NCCL were preparedin the buccal surfacesof the teeth of Groups 1,2 and 3 in accordancewith the criteria defined by Bayne (personal communication, 1994) using a specially designed 110” tapered bur (DK Holdings, Staplehurst, UK) held in a laboratory handpiece, operating at 8000 rpm without water coolant. The specimenswere held rigidly in a parallelometer. A depth gauge was used to allow the cavities to be preparedto a depth of 2 mm from the CEJ, with the bur being run along the tooth in a mesialto distal direction to create a notch-shapeddefect (Fig. 1). 2.3.3. Restoration of NCCL The NCCL in the teeth of Group 1 were restored to original dimensionsusing Vitremer (3M Healthcare) which was dispensed,mixed and applied in accordance with the manufacturer’s instructions. The NCCL in the teeth in Group 2 were treated with the three constituents of the Scotchbond MultiPurpose Plus system (SBMPS) (3M Healthcare), and restored to original dimensions using ZlOO (Shade A2) restorative (3M Healthcare), with each material being used in strict accordancewith the manufacturer’s instructions. The NCCL in the teeth in Group 3 were left unrestored. 2.4. Fracture testing The restored teeth were stored in tap water at room temperature for 24 h prior to testing. They were then subjected to compressive loading at a cross-head speed of 1 mm min-’ in a Universal Testing Machine (RPC Howden Ltd., Leamington Spa, UK). Compressiveforce was applied by means of a 4-mm diameter steel bar placed along the midline fissure of the tooth. The force (KN) required to cause fracture was recorded. Articulating paper was used between the specimen and the steel bar to ascertain the exact point of contact and loading of the tooth/restoration occlusal surface. The fracture strength results were subjected to statistical analysisby two-way analysisof variance (ANOVA) and Scheffe’s Multiple Range Test at a significance level of 0.05. The fractured specimenswere examined to determinethe mode of fracture, using a classification designed for the investigation (Fig. 2). The shadedarea in Fig. 2 represents the fracture site. They were also examined for evidence of debonding or loss of the NCCL restorations. The results were recorded and the fracture mode analysed, considering the fractures in order of severity, using Kruskal-Wallis oneway analysis of variance by ranks (SPSS package).
3. Results Mean fracture loads (KN) of 1.08, 1.03, 0.98 and 1.14 were recorded for Groups 1, 2, 3 and 4, respectively (Table 1, Fig. 3). Two-way ANOVA and Scheffe’s Multiple
430
K.L. Osborne-Smith
MODE I la
et al./Jourtaal
Description Midline Occlusal fissure fracture
0
Amalgam and/or tooth fracture-- not midline
Group
Preparation/restorative materials used
n
Mean
s.d.
1
MOD/amalgam and NCCUVitremer (3M) MOD/amalgam and NCCL/ZlOO (3M) MOD/amalgam and NCCL unrestored MOD/amalgam
10
1.084
0.24
10
1.029
0.28
10
0.976
0.34
10
1.14
0.34
2 3
Buccal cusp fracture - sound tooth or amalgam remaining
4
Severe amalgam and/or tooth fracture involving more than one cusp
Vitremer (Group 1) or ZlOO (Group 2) resulted in fewer buccal cusp fractures but more palatal cusp fractures, compared with those teeth with unrestoredNCCL. No Vitremer (3M Healthcare) or ZlOO (3M Healthcare) restorations were lost during or after the compressive loading.
Buccal cusp and amalgam fracture
4. Discussion
Palatal cusp fracture - sound tooth or amalgam remaining
This in vitro investigation was designed to study the effects of restored and unrestored NCCL on the fracture resistanceof maxillary premolar teeth, restored with moderately sized MOD amalgamrestorations.It was anticipated that the amalgam restorations would simulate the clinical situation in which older patients typically present,with teeth with both NCCL and other restorations. Clinical conditions were simulatedwhere possiblethroughout the investigation and the specimenswere removed from the water-storage medium only when operative procedures were necessary. To allow comparisons to be made between and within groups, it was necessary to select teeth of standardized bucco-palatal width, to standardize the preparations and subsequentrestorative procedures and to standardize the application of force applied to the specimens. A 4-mm
Palatal cusp and amalgam fracture
Fig. 2. Classification when compressively
26 (1998) 427-433
Table 1 The mean fracture loads (KN) required to induce fracture in restored premolar teeth (Groups l-4), when compressively loaded at 1 mm min-’
lb Midline amalgam fracture
of Dentistry
of the mode of fracture loaded at 1 mm min-‘.
of restored
premolar
teeth
RangeTest showedno statistically significant differences (P > 0.05) in mean fracture load between the groups. The frequency of the mode of fracture of the teeth in Groups 1-4 is shown graphically in Fig. 4. The majority of fractures in Groups l-4 were ClassIa (n = 23), i.e. a midline amalgamfracture in a bucco/palatal direction. The restoration of the NCCL in Groups 1 and 2 with either 1.4 B z
q GPl 5jGP2 GP3 n GP4
1.2
5
1
E
0.8
z =) 0.6 b d 0.4 LL $ 0.2 Y
Fig. 3. The mean fracture
0
loads (KN)
GP GP GP GP
1 2 3 4
-
required
Standafdised Standardiid Standardised Standardised
MOD MOD MOD MOD
amalgam amalgam amalgam amalgam
to induce fracture
restoration, restoration, restoration, restoration
in restored
NCCL NCCL NCCL
premolar
restored with vitremer (3M Dental Products) restored with 2100 (3M Dental Products) unrestored
teeth (Groups
l-4),
when compressively
loaded at 1 mm min-’
K.L. Osborne-Smith
01’
et al./Journal
of Dentistry
iiwm-nn’
Noto Modes of fmcium in which a total of one or less specimens
of the mode of fracture
of restored
steel bar was used to apply a compressiveforce in an axial direction until fracture. Analysis of the resultsindicated that the majority of the forces applied were within the range of biting forces reported by Gibbs[l9]. Accordingly, the loads applied may be considered to be commensuratewith those applied during normal chewing. The teeth used were sound but of unknown age. As age may affect the fracture strength of teeth, 10 teeth were included in each group and the mean fracture strengths were used for comparison. While the NCCL preparations usedin the study were comparablewith thoseseenclinically (vi& in@), it is likely that in viva somereparative dentine may be laid down beneath areasaffected by NCCL. The potential weakening effect of the NCCL preparations in the present study may therefore have had a more profound effect upon the fracture resistance than in the in viva situation, where the teeth may have been strengthenedby reparative dentine formation. In a clinical study examining 457 NCCL in 101 patients, the internal angle was found to be 90-130” in 48% of cases, with Bayne reporting 110” to be the median internal angle[20]. Other workers]211 have reported the mean NCCL internal angle to be 105 t 27”. The standardized cervical preparation carried out in the present investigation may therefore be considered to be representative of the typical NCCL seen in the clinical situation. In addition, sinceBayne et aZ.[20] reported that 62% of NCCL occurred in the maxillary arch, the choice of maxillary premolar teeth may be consideredappropriate. The choice of a resin-modified glass-ionomercement and a compositeresin together with an enamel-dentine bonding system for the restoration of the NCCL was considered appropriate given the widespread use of such materials in Class V situations and their successin cervical cavities reported previously[ 13,141. Grippo has suggestedthat the restoration of NCCL would
431
427-433
GP4 GP3 ii -Amalgam/tooth fracture - not midline VII -Amalgam and paiatai cusp fracture
GPl GP2 I - Midllne amalgam fracture (buccopaiati) Vi - Amalgam and buccai cusp fracture
Fig. 4. The frequency
26 (1998)
fractured
premolar
from Group 1-4, have been excluded
teeth when compressively
for clarity.
loaded at 1 mm min-‘.
reduce flexure of teeth under occlusal load and thereby strengthenthe affected teeth[ 111.This strengthening effect was not noted in the present investigation, given that no statistically significant differences (P > 0.05) were found in the meanload required to causefracture amongor within the Groups l-4. However, the mean fracture loads for the groups of teeth with restoredNCCL (Groups 1 and 2) were higher than for the groups of teeth with unrestored NCCL (Group 3), although this difference was not statistically significant. Although the mean fracture strengths were not statistically significant, a trend towards greater fracture resistance is apparent in specimenswith restored NCCL when compared to specimenswith unrestored NCCL. A larger sample of teeth may have allowed the statistical tests to demonstrate a significant difference between the groups.The low power (23%) usedin the statistical analysis did not show a statistically significant difference in fracture strength of the four groups in this investigation. An increased number of specimensin each group (n = 87) would be necessaryto give 90% power, to detect a statistically significant difference in fracture strength at a significance level of 5%. However, the lengthy time required to produce eachspecimen(ca. 70 min) and the limited availability of suitable sound extracted teeth of appropriate dimension did not allow larger group sizes to be used in this study. The MOD restoration of amalgam may have been expected to predisposethe teeth to fracture, thereby emphasizing the effects of the NCCL preparation. However, from the resultsof a related investigation by the sameauthors(not published), it would appear that teeth with MOD amalgam restorations showed a greater resistance to fracture than teeth without MOD restorationsof amalgam.These findings are consistent with the work of others[17,22], with Watts and co-workers reporting that the greater force required to fracture teeth restored with amalgam may relate to the
432
K.L. Osborne-Smith
et al./Journal
plastic flow of directly loaded restorations of amalgam[ 171. These findings may be explained by the mode of application of the load to the occlusal surface of the restored teeth. Examination of the specimen teeth post-loading revealed that the teeth were loaded centrically as occlusal indicator marks were apparent on the central fossa of the amalgam restorations, while prior to loading the steel bar was found to rest on the cuspal inclines of the restored occlusal surfaces. The amalgam, being plastic, may have absorbed some of the applied load and, as a consequence, less load may have been applied to the tooth structure. Amalgam has a higher modulus of elasticity than dentine and, consequently, when loaded is able to withstand increased loads before fracture[23]. The present investigation involved the preparation of standardized NCCL in selected premolar teeth to enable comparisons to be made between groups of teeth in which the NCCL had been restored and not restored. It may be anticipated that application of an axial load would cause compression and flexure of the tooth specimens and lead to the debonding of the cervical restorations. However, none of the Vitremer or ZIOO restorations were lost when the fracturing force was applied. The restorations were placed under optimum conditions and on optimum substrate, namely, open dentinal tubules and intertubular dentine with no pulpal pressure or pulpal exudate. Nevertheless, the results indicate that the bonding strength of the restorations was sufficient to withstand the potential debonding forces resulting from tooth flexure. The results may have differed if a different method of loading had been applied, such as eccentric loading or fatigue testing, resulting in a different stress distribution within the restored tooth. Preservation of tooth structure is of primary importance to patient and clinician, and consideration of the mode of fracture of teeth following restoration is therefore necessary. It may be considered that fracture of a restoration is preferable to a catastrophic tooth fracture which may not readily permit restoration of the tooth. The non-parametric test used showed no difference in the distribution of fracture modes between the four groups. In the present investigation, the majority of fractures in Groups l-4 were of Mode Ia, i.e. the amalgam fractured in the midline in a bucco-palatal direction. This mode of fracture is seen in clinical practice and may be the result of centric loading, accepting that masticatory forces are applied at a greater speed than 1 mm min-‘. When the buccal cusp was sound (Group 4) or the NCCL had been restored (Groups 1 and 2), a greater number of palatal cusp and amalgam fractures (Mode VII) was seen than buccal cusp with amalgam fractures (Mode VI). Conversely, when the specimen teeth had an unrestored NCCL in addition to an MOD amalgam (Group 3), the buccal cusp and amalgam fractured (Mode VI) more frequently than the palatal cusp and amalgam (Mode VII). It may be considered that the presence of an unrestored NCCL weakens the buccal cusp, which may then fracture in preference to the more frequent mode of fracture involving the palatal cusp.
of Dentistry
26 (1998)
427-433
Finally, 3 of the 40 teeth demonstrated severe amalgam/ tooth fractures. No correlation was apparent between the load required to fracture these specimens and the severity of the fractures observed. It may be considered that such extreme results are due to normal biological variation, unseen damage during extraction, variations in pulp space or unseen fatigue stresses.
5. Conclusion It is concluded that the presence of a standardized NCCL of the dimensions used in this investigation in extracted maxillary premolar teeth with standardized MOD restorations does not reduce fracture resistance of the tooth under compressive loading at 1 mm min-‘. It is also concluded that the restoration of NCCL with Vitremer (3M) or ZlOO (3M) does not improve the fracture resistance of the specimen teeth when loaded compressively at 1 mm min-‘. Differences in mode of fracture were not noted between specimens with restored NCCL compared to specimens with unrestored NCCL. However, the teeth with restored NCCL suffered fewer fractures of the buccal cusp.
Acknowledgements The authors acknowledge for this study.
the support of 3M Healthcare
References [l] Black, G. V., A Work on Operative Dentistry. Pathology of Hard Tissues of Teeth, 1st edn. Medico-Dental Publishing, Chicago, IL, 1908, pp. 39-59. [2] Manly R. S. The abrasion of cementurn and dentine by modern dentrifices. Journal of Dental Research, 1941;20:583-595. [3] Eccles J. D. Dental erosion of non-industrial origin. A clinical survey and classification. Journal of Prosthetic Dentistry, 1979;42:649-653. [4] Zipkin I., McClure F. .I. Salivary citrate and dental erosion. Journal of Dental Research, 1949;28:613-626. [5] Sognnaes R. F., Wolcott R. B., Xhonga F. A. et al. Erosion-like patterns occurring in association with other dental conditions. Journal of the America1 Dental Association, 1972;84:571-576. [6] Bergstrom J., Lavstedt S. An epidemiological approach to toothbrushing and dental abrasion. Community Dentistry and Oral Epidemiology, 1979;7:57-64. [7] Sognnaes R. F., Dental hard tissue destruction with special reference to idiopatic erosions. In Mechanisms ofHard Tissue Destruction, ed. R. F. Sognnaes. American Association for the Advancement of Science, 1963, pp. 91-149. [8] Manly R. S. Factors influencing tests on the abrasion of dentin by brushing with dentrifices. Journal of Dental Research, 1944;23:59-72. [9] Lee W. C., Eakle W. S. Possible role of tensile stress in the etiology of cervical erosive lesions of teeth. Journal of Prosthetic Dentistry, 1984;52:374-380. [lo] Grippo J. 0. Abfractions; a new classification of hard tissue lesions of teeth. Journal of Esthetic Dentistry, 1991;3: 14-19. [l l] Grippo J. 0. Non-carious cervical lesions: the decision to ignore or
K.L.
Osborne-Smith
et al./Joumal
restore. Journal of Esthetic Dentistry, 1992;4(Suppl.):55-64. [12] Hand J. S., Hunt R. .I., Reinhardt .I. W. The prevalence and treatment implications of cervical abrasion in the elderly. Gerodontics, 1986;2:167-170. [13] Heymann H. O., Sturdevant J. R., Bayne S. et al. Examining tooth flexure effects on cervical restorations: a two-year clinical study. Journal of the American Dental Association, 1991;122:41-47. [14] Powell L. V., Johnson G. H., Gordon G. E. Factors associated with clinical success of cervical abrasion/erosion restorations. Operational Dentistry, 1995;20:7-13. [15] Douglas, W. H., Methods to improve fracture resistance of teeth. In International
Symposium
on
Posterior
Composite
Resin
of Dentistry
[18] [19]
[20]
[21]
Dental
Restorative Materials, ed. G. Vanherle and D. C. Smith. 3M, St Paul, 1985, pp. 433-441. [16] Morin D., DeLong R., Douglas W. H. Cusp reinforcement by the acid etch technique. Journal of Dental Research, 1984163: 1075-1078. [17] Watts D. C., EL-Mowafy 0. M., Grant A. A. Fracture resistance of
[22]
[23]
26 (1998)
427-433
433
lower molars with Class I composite and amalgam restorations. Dental Materials, 1987;3:261-264. Burke F. J. T., Investigation of aspects of optimum cavity design for composite inlays. M.Sc. thesis, University of Manchester, 1991. Gibbs C. H., Mahan P. E., Lundeen H. C. et al. Occlusal forces during chewing and swallowing as measured by sound transmission. Journal of Prosthetic Dentistry, 1981;46:443-449. Bayne S. C., Heymann H. O., Wilder A. D. et al. Substrate variables in dentin bonding systems for Class V restorations. Journal of Dental Research, 1994;73(1388):275. Yoon K. .I., Sturdevant J. R. Geometry of Class V cervical abrasion/ erosion lesions. Journal of Dental Research, 1994;73(1783):325. Stampalia L. L., Nicholls J. I., Brudvik J. S. et al. Fracture resistance of teeth with resin-bonded restorations. Journal of Prosthetic Dentistry, 1986;55:694-698. Goel V. K., Khera S. C., Senthil G. et al. Effect of cavity design on stresses in first molars. Journal of Dental Research, 1985;64:350.