• All-ceramic;
  • fixed partial denture;
  • inlay supported denture;
  • ceramic fixed partial denture


  1. Top of page
  2. Abstract
  3. Introduction
  4. Conclusions
  5. References

The effect of cavity design is a controversial and underrated factor in the clinical success of ceramic inlays and inlay supported prosthesis. Many articles and studies have been conducted into the advantages and disadvantages of isolated aspects of preparation design, but lacking is a review of the most relevant papers which bring together a consensus on all the critical features. Hence, a review and analysis of cavity depth, width, preparation taper and internal line angles is warranted in our attempts to formulate preparation guidelines that will lead to clinically successful, all-ceramic inlay restorations and ceramic inlay supported prosthesis.

Abbreviations and acronyms:

finite element analysis


intercuspal width


total occusal convergence


  1. Top of page
  2. Abstract
  3. Introduction
  4. Conclusions
  5. References

Successful restoration of the dentition has historically depended upon simultaneously respecting the three foundation principles of tooth preparation: mechanical preparation to achieve retention and resistance, hence ensuring longevity; aesthetic factors such as minimizing the appearance of margins and display of metal; and the biological consequences of achieving the first two factors which concerns the health and ultimate durability of the tooth and periodontium.

Growing patient demands for aesthetic “tooth-like” materials and concerns about the deleterious effects of metals has added a new consideration for the profession to address.

Magne,1 who advocated a new “biomimetic” approach for restorative and prosthetic dentistry via the use of ceramics and composite resins, heralded not only a change in techniques, but as stated by Roeters,2 a change in treatment philosophy. As a consequence of this major paradigm shift, the primary emphasis for dentistry is now not the restoration of the tooth, but rather its reinforcement and preservation.

Part 1 of this two-part investigation will review the literature with regards to the ideal inlay preparation design whilst Part 2 will present original research into the fixed partial denture (FPD) design. Three dimensional finite element analysis will be used to create an optimum all-ceramic bridge with regards to connector dimensions and embrasure geometry and validated with load-to-failure bench-top testing.

Inlay preparation design

It has become a basic tenet in dentistry that in order to replace a missing tooth (pontic) without resorting to the use of removable prosthetics or implants, it becomes necessary to attach the pontic to the adjacent teeth (abutments). Whether or not it is chosen to prepare the abutments in an effort to improve mechanical resistance and retention is a decision that the dentist must weigh up against the loss of tooth structure that accompanies such preparation and hence increased risk of tooth fracture.

Little or no preparation to the abutments means relying heavily upon adhesive technology with minimal mechanical assistance. The advantages are the conservation of tooth structure and thus the diminishment of the pulpal and periodontal consequences discussed above. Conversely, if the decision is made to prepare the abutments in order to impart mechanical resistance and retention to the FPD, then varying degrees of tooth reduction becomes necessary, with the associated complications arising with increasing tooth preparation. It is the challenge of replacing missing teeth and the restoration of aesthetics and function at minimal biological cost that is of concern to every practising dentist.3

Many articles and studies have investigated the advantages and disadvantages of the various aspects of preparation design and its effect on the clinical success of ceramic inlays. Milleding et al.4 stated that “the effect of cavity design on the strength of an inlay is a factor that is probably underrated”.

The main factors of preparation design that influence the longevity of the inlay/tooth complex are as follows: cavity depth; cavity/isthmus width; preparation taper, and the morphology of internal line angles. Figure 1 illustrates the idealized form of an MO inlay on the lower second molar.


Figure 1.  Ideal ceramic inlay preparation design.

Download figure to PowerPoint

Tooth preparation designs advocated for posterior ceramic restorations have been based upon recommendations made by GV Black (1836–1915) for cast metal and amalgam, resulting in considerable tooth structure removal, opposing walls that are too parallel and internal line angles too steep.

Preserving tooth structure is beneficial to the overall health of the tooth and periodontal tissues. The use of the minimally invasive bonded restoration results in less trauma and superior prognosis.5–10 When designing a tooth preparation, be it for restorative or prosthetic reasons, it is imperative to balance the competing considerations of aesthetics; preservation of tooth structure and the periodontal complex, and maximizing the strength of the restoration.11 Cavity geometry and dimensions are dictated by traditions of cavity design, the properties of the restorative material, the techniques and technology applied and ultimately the inherent shape of the carious lesion.12

Preparation geometry for ceramic restorations in general, and inlays specifically, must be adapted to the specific properties of ceramics. Possessing a low tensile strength and high modulus of elasticity, the traditional retention/resistance principles for cast metal restorations must be relaxed and the simplest geometry employed.13 Low flexural strength is a limiting property of brittle materials such as ceramics because the failure mechanism most likely is that of tension or impact damage rather than compression, a property of which ceramics possess highly but is irrelevant in considerations of rupture and cyclic fatigue.14–17

The literature is conclusive in regards to the effects of tooth preparation; it further weakens teeth and increases the likelihood of fracture.18 Khera et al.19 examined the effects of cavity depth, isthmus width and remaining interaxial dentine on MOD cavity preparations via the use of 3D finite element analysis (FEA). A total of eight different cavity designs were prepared on human premolars, divided into three groups and compared with normal, unprepared teeth and to other cavity designs in the same group. It was demonstrated that the most crucial factor in the weakening of cusps was cavity depth, with the width of the isthmus alone being the least important.

Lin et al. studied the mechanical responses of MOD preparations on six human second premolars with the use of FEA. They concluded that pulpal wall depth was the most profound determinant in the likelihood of cuspal fracture and the deeper the pulpal depth, the greater the risk to the restored tooth.20

In a similar study, Lin et al.21 examined the biomechanics of 30 MOD cavity preparations on human maxillary second premolars via FEA. Stress levels were correlated to pulpal depth, isthmus width and interaxial thickness (width of the pulpal floor from axial wall-to-axial wall) with variations of the three design parameters being made and analysed. The results demonstrated that enlarging the volume of the MOD cavity significantly increased the stresses in enamel, and to a lesser extent dentine, with the stress intensity rising exponentially with cavity depth. For enamel, cavity depth is the most dominant factor influencing stress. However, for dentine, it appears that the length of the interaxial wall could be the most important factor.

Table 1 summarizes a number of studies evaluating the role of cavity depth and its relation to restoration and tooth strength. It demonstrates that a depth of 1.5 to 2 mm is ideal in minimizing tooth loss and providing sufficient thickness of material in order to ensure adequate functional life.

Table 1. Cavity depth recommendations for minimizing tooth fracture in Class II restorations
AuthorRecommendation for cavity depthComments
  1. NSR = no specific recommendation.

Banks (1990)151.5 to 2 mmUniformity of depth stressed.
Blaser et al. (1983)22Shallow floor considered is1.5 mmCavity depth most important factor. Width does not substantially weaken teeth if depth is shallow.
Donly et al. (1990)231.5 mm1.5 mm depth conservative Class II preparation has less marginal leakage than 2.0 mm conventional preparation.
Etemadi et al. (1999)111.5 to 2.0 mmStudy conducted on resin-bonded porcelain restorations. Rounded internal line angles recommended.
Goel et al. (1992)24NSRUnfavourable stresses increased with increasing cavity depth.
Homewood (1998)25NSRShallower cavity results in less cusp deflection.
Khera et al. (1991)19NSRCavity depth most significant factor in fracture of tooth, isthmus width the least.
Lin et al. (2001)20 Unfavourable stresses develop exponentially as cavity depth increases.
Malament and Grossman (1987)261.5 to 2.0 mmSmooth preparation with no sharp internal line angles recommended.
Malament (1998)271.5 to 2.0 mmCavity depth recommended for ceramic strength.
Milleding et al. (1995)41.5 to 2.0 mmRecommendation made specifically to minimize fracture of ceramic inlay but authors found that 1.5 mm cavity depth resulted in only 2% cusp fracture rate.
Nadal (1962)28NSRShallow floor and narrow occlusal outline recommended.
Rosenstiel et al. (2001)291.5 to 2.0 mmManufacturers recommendation.
Watts et al. (1995)30Cavity depth of ⅓ to ½ bucco-lingual width.Shallower restoration depth leads to decreased prevalence of tooth fracture.

The concurrence of opinions regarding the recommended depth for cavities in order to minimize the incidence of tooth fracture is considerable. This needs to be balanced with the need to retain adequate bulk in the restorative material to ensure the long-term viability of the tooth/restoration complex. These competing issues influencing material strength can be successfully answered with current bonded restorations which rely significantly less on mechanical factors than traditional direct and cast restorations. In the case of ceramic inlay systems, the use of a resin cement to both retain the restoration and support the weakened tooth structure results in good long-term success.31 Zinc phosphates and glass-ionomer cements must be avoided. However, the former, because of its inability to bond, and the latter, due to its low modulus of elasticity, increases the flexure of the inlay and thus the rate of fracture.

Habekost et al.32 evaluated the in vitro fracture resistance of teeth restored with different designs of ceramic restorations. One hundred and twenty sound maxillary premolars were tested in three groups. Each group was prepared with three indirect restorations consisting of inlays, onlay with only lingual cuspal coverage and onlay with buccal and palatal cuspal coverage. Twenty intact teeth were selected as controls. Peak load-to-fracture was measured for each specimen. Results indicated that the fracture resistance of the teeth was related to the quantity of hard tissue removed and inlays showed a significantly higher fracture resistance than onlays. This suggests that unlike in the use of metallic materials and composite resins, where cusp capping is often viewed as being a preferred means of reinforcing a tooth, caution is needed for ceramic inlays.

Bonding of inlays to teeth increases the fracture resistance of the tooth.33–35 However, large MOD preparations severely undermine cusps to the degree that adhesive bonding of restorative materials does “not re-establish the fracture resistance of the tooth to its original levels.”16 Hence, minimizing the depth and overall width of any tooth preparation to the amount needed for adequate retention, resistance and convenience form must be of primary concern.

Table 2 summarizes a number of studies evaluating the relationship between enlarged cavity widths (specifically the intercuspal width defined as the distance between cusps) and tooth fracture strength. Universally, the consensus is to maintain as narrow cavity width as possible whilst maintaining acceptable strength in the restorative material; the recommendation is ⅓ intercuspal width (ICW) or less, with most recommendations suggesting ¼ or less.

Table 2. Cavity isthmus width recommendations for minimizing tooth fracture in Class II restorations
AuthorIsthmus recommendation (as a ratio of ICW)Comments
  1. NSR = no specific recommendation.

Bader et al. (2004)36NSRRelationship exists between fracture risk and dentinal support measured by intercuspal width proportion and restoration depth.
Blaser et al. (1983)22NSRWidth of MOD preparation does not substantially weaken the tooth if the pulpal depth is shallow.
Cavel et al. (1985)37≤⅓ ICWWider isthmus and/or more restored surfaces related to increased fracture susceptibility.
Christensen (1971)38≤⅓ ICWInlays with ICW > ⅓ have higher fracture risk.
Re et al. (1982)82NSRNo specific trend found between the fracture strength of restored teeth and preparations with various sizes of faciolingual width.
Homewood (1998)25NSRWider isthmus results in greater cusp deflection.
Joynt et al. (1987)40⅓ ICW⅓ ICW chosen for study, recognized that narrow ICW associated with reduced fractures.
Larson et al. (1981)41≤¼ ICWProportional isthmus width is possibly the most important measure of lost dentinal support associated with fracture resistance.
Lin et al. (2001)20NSRSmaller isthmus results in less stress, cavity depth most important factor.
Mondelli et al. (1980)42≤¼ ICWThe narrower the isthmus, the greater the load to cause fracture. A significant factor in preparation design.
Osborne and Gale (1980)43≤¼ ICW¼ ICW provides better resistance to fracture than ⅓ ICW.
Vale (1959)44≤¼ ICWIsthmus greater than ⅓ ICW significantly weakened.
Watts et al. (1995)30<⅓ ICWNarrower cavity width had statistically higher fracture strengths.

Total occlusal convergence (TOC), defined as “that angle which is formed between opposing walls of a preparation”, is an important factor in cavity design and yet the aspect associated with the most contention.45 For complete crown preparations, TOC was one of the first preparation criteria given a specific quantitative value when Prothero in 1923 recommended a range of between 2° and 5°. This was later scientifically tested by Goodacre et al.44 with the recommendation increasing to between 6° and 7°.

The current practice of minimizing the axial wall convergence or the TOC to between 6° and 7° (or less) in the preparation of cast metal restorations47–50 is likely to lead to increased failure rates if used for ceramic restorations, and should be increased to approximately 15°.

Kaufman et al.49 examined the effects of varying the TOC angle (1°, 5°, 10°, 15°, 20°) on complete veneer crowns with controlled variations in height and diameter, and found that as the convergence approached parallelism (at least to within 5°), retention increased geometrically – this being related to the simple effects of geometry. No definitive recommendation was made as to a TOC angle as it was acknowledged that many factors influence the retention of a cast restoration (e.g., adaptation of the casting, texture of surfaces, elasticity of the casting to enable it to resist deformation and hence maintain the cement seal, etc.).

Livaditis,51 Shillingburg52 and Rosenstiel et al.29 have recommended a TOC of 5–7° for resin-bonded cast metal, intracoronal restorations due to the increased retention offered by the friction fit of the surfaces, whilst Jørgensen49 attributed the increased retention to the limiting of the “paths of insertion” and removal.

Mack examined the TOC angles of clinically prepared inlay and crown dies, and compared them to those prepared in a laboratory with the use of standard laboratory optical measurement equipment.53 It was concluded that the average TOC achieved in dental practice was about 16.5°– far removed from the textbook ideal of 5°. He showed that if a dentist looked over the preparation with a mirror and could sight all the walls, then the minimum taper achieved is 5°, 42′. An estimate of 12° was also made for the minimum convergence required in order to ensure an absence of undercutting clinically.

Ceramic restorations are fundamentally different to cast metal restorations in numerous ways. Chief amongst them is their very high modulus of elasticity and presence of numerous micro-flaws on the surface which renders them fragile in tension, hence highly brittle and likely to fracture during the luting procedure and under occlusal loading.11,54–56

Qualtrough and Wilson55 stressed the importance of the bonding procedure to the overall success of the ceramic restoration. Hypothesizing that unlike traditional cast metal inlays where the fit was critical to success, it is the bonding procedure in ceramic systems that may ultimately determine the longevity of the restoration, with smaller degrees of divergence between axial walls resulting in greater stress being imparted to the inlay and the increased likelihood for the need of adjustment. Unlike metals, resins and tooth structure, ceramics are unable to elastically deform to the same extent; hence the build-up of stresses is likely to occur from the cementation procedure if there are any discrepancies of fit, or if the fit is tight. The TOC angle must be relaxed in order to accommodate the inlay, minimizing straining and the build-up of stress.56

Table 3 demonstrates the current opinion with regards to increasing the TOC for ceramic inlays from the traditional 5° to 7° to approximately 20° when ceramic restorations are utilized. Values are also given for ceramic and metallic crowns as a comparison and a guide as to the fluctuating historical opinions.

Table 3. Total occlusal convergence angle recommendations
AuthorTOC recommendationComments
  1. NSR = no specific recommendation.

Doyle et al. (1990)5715°15° TOC significantly stronger than 5° on all-ceramic, complete crowns.
Eames et al. (1978)5820°20° TOC most likely to be seen clinically on complete crowns.
El-Ebrashi et al. (1969)472.5° to 6.5°Stress concentration increases slightly from 0° to 15°, increases sharply at 20°. Measured from models of complete crowns.
Esquivel-Upshaw et al. (2001)59Inlays with TOC of 5° significantly more fracture resistant than those at 20°.
Etemadi et al. (1999)1121° to 40°21° to 40° for internal tapers, 6° to 15° for external tapers, as measured from clinical models of porcelain inlays and onlays.
Gerami-Panah et al. (2005)6022°22° TOC results in less stress to the gingival connector area of an all-ceramic FPD than 12°.
Gilboe and Teteruck (2005)482° to 5°Recommendation for cast-metal complete crowns.
Goodacre et al. (2001)4610° to 20°Recommendation for complete crowns
Jørgensen (1955)61As parallel as possibleRecommendation for complete crowns. 5° TOC is twice as retentive as 10°; 20° is 62% the retention of 10° and 81% that of 5°.
Leempoel et al. (1987)6215.5° to 30.2°Review of working dies from dental laboratory. Crowns were in place 5 to 10 years and still functioning adequately.
Mack (1980)535° accepted consensusWhilst he accepts the consensus that the ideal TOC for inlays and crowns is 5°, 16.5° is more commonly seen clinically. 12° is the minimum required to avoid undercutting.
Milleding et al. (1995)4NSRInlay preparation designs must be relaxed from the traditional recommendations.
Malament and Grossman (1987)266° to 8°Recommendation for all-ceramic complete crown.
Nordlander et al. (1988)505° to 10°Theoretical ideal for complete crowns, but rarely seen clinically. Average seen clinically for premolars is 8° and for molars 12.5°.
Owen (1986)6312°Unless special jigs used, not possible to prepare teeth with TOC of less than 12° TOC. At this angle they still perform well.
Palacios et al. (2006)6420°Common journal finding for all-ceramic crowns.
Parker et al. (1993)658.4° for molars, 10° for premolarsCalculation of limiting average taper mathematically based on ½ arc sin (H/B). Less than this amount results in reduced resistance.
Qualtrough and Wilson (1996)55NSRFit must be relaxed for ceramic inlays.
Schwartz (1952)66Ideally 0°Pulpal and axial walls should be perpendicular.
Sobrinho et al. (1999)67No difference between 8° and 16°TOC of in-ceram crowns had no effect on their fracture strength. However, luting with zinc phosphate achieved significantly better results than GIC.
Wilson and Chan (1994)686° to 12°For extracoronal retainers, optimum thickness of cement occurs between 6° and 12°. Retention decreases significantly as TOC reduces from 9°. Larger than 25° TOC also results in significant decrease in retention.

Classic cavity design principles as advocated by GV Black recommended the use of flat walls and sharp internal line angles as the best way of maximizing the retention and especially the resistance form of restorations – this is especially true with regards to Class II restorations where traditionally even the axiopulpal line angle has been left deliberately angular.

Cavity design evaluation based upon the use of 2D photoelastic methods has revealed that any areas of angularity within tooth preparations and restorative materials give rise to significant stress concentrations. The pioneering work of Noonan,69 Mahler and Peyton70 and Haskins et al.,71 as well as others, concluded that “rounding of internal line angles [is] the most satisfactory modification of cavity preparation with respect to stress within the remaining tooth structure”.72

The stresses that accumulate within complex shapes is difficult mathematically to analyse. However, the use of photoelastic analyses of these complex and deleterious stresses has provided useful data to derive optimal design parameters for cavity preparations. Photoelasticity involves the construction of a model of the structure from a photoelastic material, i.e., a transparent material which exhibits birefringence. The photoelastic material exhibits birefringence upon the application of stress and the magnitude of the stress, and each point is displayed via the refractive indices.73 The evaluation of the stresses using this approach has immensely helped our understanding of the need for the rounding of all internal line angles, with special emphasis on the axiopulpal line angle. Interestingly, the stress distribution for bonded restorations is markedly different, with peak stress values occurring in the enamel at the site of contact with the opposing cusp.

Couegnat et al.74 utilized structural shape optimization procedures based on FEA to derive optimized designs for the second upper premolar. This relatively new technique allows adaptations to be made to cavity designs involving the build-up of material at overloaded zones and the reduction or no build-up of material at underloaded zones to be analysed mathematically and displayed as a scalar function (similar to a photoelastic image). Their results indicated that the “notches” which are created at internal line angles are a principle source of stress concentration in non-bonded internal restorations, whereas the principle source of stress for onlays and other external restorations existed in the restorative material itself. Rounding of all line angles and the orientation of prepared cusps tips perpendicular to the occlusal load is recommended for the reduction of stresses.

Arola et al. utilized FEA to analyse the stress distribution and potential for cyclic fatigue crack growth within Class II amalgam cavities.75 From their results it was concluded that subsurface cracks developing in the dentine along the buccal and lingual margins during cavity preparation can significantly reduce fatigue life and may be the principle source for premature restoration failure. The authors opine that it is the instruments and techniques used in tooth preparation that must be examined closely as cracks as small as 25 μm can lead to fracture in 25 years.

Table 4 demonstrates the findings from a number of studies with regards to the stresses caused by sharp internal line angles to both tooth and restoration.

Table 4. Stress analysis in dental materials and cavity preparations
Arola et al. (1999)75Subsurface cracks introduced during cavity preparation with conventional burs may serve as a principle source for premature restoration failure.
Arnetzl and Arnetzl (2006)13Geometry of cavities for ceramics must be refined and relaxed, with the simplest of forms to increase their fracture resistance.
Banks (1990)15The transfer of distribution of stresses in an efficient manner is of equal importance to the strength and toughness of the restorative system. Preparations for ceramics must have smooth surfaces and rounded, smooth flowing internal and point angles.
Bell et al. (1982)76Cuspal failure is often related to fatigue failure of the cusp, initiated from small cracks propagating under repeated loading. Evidenced from clinical observations and mathematical modelling.
Braly and Maxwell (1981)77Recognition that any inlay restoration, in particular MOD inlays weakens the remaining tooth. Preserving the natural tooth structure, rounding of sharp line angles and designing castings that don’t extend onto uninvolved parts of the tooth is essential.
Cameron (1964)78First mention of “cracked tooth syndrome” and its correlation to restoration size and postulated mechanisms for crack propagation.
Couegnat et al. (2006)74Restorations not bonded to the tooth structure are most likely to fracture at the internal line angles. Rounding of internal line angles results in reduced von Mises stress values.
Etemadi et al. (1999)11Advocates rounding of all internal line angles.
Haskins et al. (1954)71Pioneering work advocating the rounding of internal line angles.
Kahler et al. (2006)79Fatigue is considered to be the principle mechanism of tooth fracture. The axiopulpal line angle in the dentine is the site of high stress concentration, with cracks as small as 25 μm leading to failure.
Malament and Grossman (1987)26Preparations for ceramics must be smooth and not have sharp line angles. Gingival margins must be either a chamfer with a rounded gingivoaxial line angle or a rounded shoulder.
McDonald (2001)80Emphasis on rounding of internal line angles and a chamfer or rounded shoulder finish-line for posterior ceramic restorations.
Milleding et al. (1995)4Smooth supporting surfaces and softly rounded contours reduce the degree of tensile and bending forces.
Snyder (1976)81Cavities should be prepared as conservatively as possible.
Soares et al. (2006)56Sharp angles and knife edge prepared cusps tend to concentrate stress.
Vale (1959)44Sharp internal line angles result in increased incidence of tooth fracture.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Conclusions
  5. References

As a result of the above analysis, it may be concluded that the idealized inlay preparation design should have the following dimensions in order to best achieve a balance between the preservation of tooth structure and strength of material: cavity depth of between 1.5 and 2 mm; isthmus cavity width of ≈⅓ the intercuspal width; TOC of ≈20°, and rounding of all internal line angles. However, it should be borne in mind that clinically, preparations tend to be wider and often deeper than recommended9 and that the presence of existing restorations and caries will often dictate preparations much larger than ideal. Hence, the above dimensions are ideal recommendations that may require changes in clinical settings.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Conclusions
  5. References