Ten tips for avoiding post-operative sensitivity with posterior composite restorations

Choosing the right material

While it almost goes without stating, taking time to make the correct diagnosis and then adopting a minimally invasive approach, is important to minimize post-operative problems. It may also be considered essential to choose, not only the correct material for a given clinical application, but also one that has been produced by a manufacturer with a pedigree in the field, ideally with research to back it up. This statement may be evidenced by research into branded versus non-branded (or ‘own label’) materials regarding resin-based²¹ and glass ionomer materials.²² Additionally, results of research by Shaw and colleagues²³ indicated that the batch-to-batch variability was greater for non-branded materials (ie their manufacturing might be in one facility one year and in a different one the following year), and research from workers in Norway and Poland²⁴ concluded that ‘at the present, the own brand label materials herein studied must be considered at the very least, a false bargain’ while expressing ‘legitimate concern’ over such materials. Most recently, Olegario and co-workers²⁵ examined the cost effectiveness of two low-cost glass ionomer cements (GIC) versus Fuji IX (GC) when used for 150 occlusal ART GIC restorations, which were then followed for 2 years. The results indicated that survival of restorations in the low-cost materials was less good, and, despite the fact that Fuji IX was more expensive, the low-cost materials were not cost-effective.

In these chastened financial times for dental practices, it may be tempting to purchase materials that are of low cost despite their less robust pedigree. However, the cost of one premature adhesive failure must be calculated against the saving in a 5-ml bottle of bonding resin.

Mechanisms for pulpal pain

In order to fully understand the potential mechanisms for POS, it may be considered relevant to examine the basic principles involved in bonding to dentine. The physiology of pulpal pain was explained by Brannstrom half a century ago in what has become known as Brannstrom's hydrodynamic theory of pulpal pain.²⁶ Essentially, this considered that pulpal pain was initiated by movement of tubular fluid as a result of an osmotic gradient from the pulpal aspect and the occlusal aspect of a dentinal tubule. Such movement of tubular fluid stimulates mechanoreceptors near the odontoblast processes and the A-delta fibres respond to this stimulation. The response of these fibres is proportional to the fluid flow, with factors that may be responsible for sensitivity being the remaining dentine thickness and the tubule diameter.²⁷ In that regard, Auschill et al¹³ considered that factors that can cause this fluid movement include dentine drying, heat resulting from cavity preparation, chemical agents and bacterial penetration. The contemporary interpretation of this theory is that, if a reliable (see previous section) bonding agent is employed, the tubules will be sealed and there should be little or no osmotic gradient, thereby minimizing the potential for POS.

Mechanism of dentine bonding

In the past, it was thought that the bond to dentine was dependent on the depth of penetration of resin into the dentinal tubules, but the work of Nakabayashi²⁸ informed us that bonding to dentine was facilitated by the application of an acid to the dentine surface in order to demineralize it and expose the collagen fibrils. Application of the bonding resin, followed by its polymerization, leads to the creation of a layer that is partly collagen and partly resin, termed the ‘hybrid layer’ (Figure 3). It therefore follows that overdrying/desiccation of the dentine leads to collapse of the collagen, and the resin does not penetrate it (Figure 4). In this regard, studies by Tay and colleagues²⁹ and Perdigao and co-workers³⁰ indicate that dry etched dentine resulted in low bond strengths. The concept of ‘wet bonding’, which was announced by John Kanca III in 1992 to some scepticism, has remained a central aspect of bonding to dentine, with Kanca adding that ‘dentine is inherently a wet tissue’.³¹

Contemporary bonding agents, such as the universal adhesives, may be used in self-etch, total-etch or selective enamel-etch modes, but the rule of not overdrying the dentine nevertheless applies.³²

What happens if the dentine is etched for a longer time than suggested by the manufacturers? The dentine surface is demineralized more deeply, more collagen fibrils are exposed and more dentinal tubules are opened. However, the bonding agent is generally designed to penetrate to a depth of, for example, circa 5 microns, so it will not form a hybrid layer to the complete depth of the demineralization, leaving an empty space (Figure 5). This then fills with fluid from the dentinal tubules, and when the tooth is subjected to thermal (cold or hot) and/or pressure stimuli during eating, this fluid-filled layer may expand or contract, resulting in tubular fluid movement, which the patient experiences as post-operative pain. Foxton,³³ in an excellent review, reinforced this statement, in stating that the phenomenon of over-etched dentine results in the resin not being able to fully penetrate to the full depth of demineralization, resulting in a zone of weakness beneath the hybrid layer. He added that this could influence the longevity of a resin composite restoration, particularly if it was placed in a tooth subjected to high occlusal stress.

Light curing is not as simple as it seems

Light curing has been part of restorative dentistry for more than 50 years.³⁴ Do we take this for granted? It may seem elementary to state this, but it is essential that the exit ‘window’ of the light curing unit (LCU) is placed correctly (noting that type I units have a light guide and type II LED LCUs that have a lens in the head of the handpiece), directly above/below (depending upon which arch is being treated) the restoration being cured. It is also essential that this is held in place for sufficient time, the time being dependent upon the type of composite, its shade and increment thickness, and the type and condition of the LCU. In this regard, Al Omari and co-workers³⁵ reported POS in 30 patients who each received one Class II composite restoration, with the restorations being cured in either fast-curing mode with an LED LCU, or a step mode using the same light. The two-step curing mode resulted in a statistically significant reduced incidence and severity of POS compared with the fast-curing mode. The explanation for this could be that, while the fast mode will have had higher irradiance, the radiant exposure, and hence the extent of cure, may have been less with this mode because of the shorter cure time.

More recently, a study by Par and co-workers³⁶ reported on the effect of high-intensity rapid curing and conventional curing on stress in 1.5-mm composite specimens. The results indicated that the high intensity cycle did cure the composites faster, but at the potential expense of increased shrinkage force, the clinical relevance of that being potentially greater POS.

There is no strong evidence in the literature to relate placement method to clinical outcome. However, four papers are worth noting in this regard.

Light power output (or irradiance) has been noted to decrease with LCU age, although this effect is not so marked with contemporary LED LCUs as it was for older quartz-halogen lights. In this regard, results of a recent questionnaire-based evaluation of the ‘habits’ of UK GDPs indicated that 53% of respondents checked the intensity of their LCU regularly, with 33% stating that they checked their light every month and 27% every week.⁴¹ It has been considered that, for checking of modern LCUs, a 3-month frequency is probably adequate (Prof WM Palin, personal communication, November 2019), although others have suggested checking before each patient treatment, something considered by the present authors to be impractical.

The configuration factor

It is a fact of life that resin composite materials shrink on polymerization, with experts in the field, Davidson and Feilzer⁴² once calling this ‘unavoidable’. On the other hand, this may be seen as a good thing, since, if they expanded, the tooth could break. Hence, in a cavity, such as a Class I, where the cavity is surrounded by the walls of the tooth and the restoration is bonded to the walls and floor of the cavity, it therefore follows that there is only one surface at which the shrinkage can take place, namely, the occlusal surface, which is not bonded to any structure. In this regard, if we consider that an occlusal cavity may have four walls and a floor, which can be bonded to, then this presents a negative configuration factor, with this concept based upon the following formula by Feilzer et al in 1987:⁴³

For an occlusal cavity, the configuration factor (C-factor) is therefore 5/1 (Figure 6a). The authors who first suggested this concept, Feilzer et al, also suggested that a C-factor of >1 was necessary for the adhesives used around the cavity to survive the stresses of polymerization. Indeed, in their experimental model, all specimens failed cohesively at C-factors >2. Adhesives have improved since the time of that experimental work, but the concept still stands, namely, that the Class I and traditional box style V cavities (as opposed to a shallow saucerized Class V lesion as found in tooth wear) are a stressful situation for the bonding resin, and steps, such as incremental placement technique (with the composite material only touching one wall at a time) or the use of a low shrinkage stress (vide infra) material. It has been considered that an MOD preparation provides a more compliant structure than does a Class I preparation.⁶

It is also important that clinicians assess the compliance of the residual tooth tissue, given that weakened cusps may need reduction and capping. In the past, such cusp-coverage resin composite restorations were considered the domain of indirect techniques. However, the current authors consider that there is no reason why these should not be placed directly, given that this will present the patient with a reduced fee for the restoration, and a more convenient one-visit technique. The downside for the clinicians relates to an increased degree of difficulty in achieving the correct occlusion on the built-up cusps as compared with an indirect technique.

It therefore follows that, for a Class I, high C-factor cavity being restored with a resin composite material, the high potential stress means that incremental layering is necessary: not so with an extensive cusp replacement restoration, in which the resin composite material may be placed in horizontal increments (Figure 6b) (an example is presented in Figure 7), unless the resin composite thickness exceeds the manufacturer's cure depth.

The configuration factor also helps explain why Class IV restorations and large cusp replacement resin composite restorations in premolars do not generally debond (if correct placement procedures are used) – the bond is not stressed! In addition, as the direction of shrinkage is towards the most strongly bonded interface (enamel) de-bonding may occur at more weakly bonded interfaces for example, caries-affected dentine. Figure 8 presents a cusp replacement restoration in a first molar in which a crack in the cavity floor was identified.

An alternative treatment when using ‘conventional’ resin composite materials

If a ‘conventional’ resin composite material, which is not low shrinkage stress, is being employed, there is evidence from laboratory research that used a variety of placement techniques in extracted human teeth. The results of this indicated that the placement of a flowable composite base layer along with incremental curing of the composite, resulted in less leakage at the gingival margin of Class II cavities filled with conventional composite.⁴⁶ It may be of interest to note that accelerated curing with a plasma light led to increased leakage values. Today, the authors suggest that an appropriate material for this clinical situation might be one of the wide range of low shrinkage stress flowable composite resins, for example, SDR (Dentsply-Sirona, Weybridge, UK).

The deep Class II box and other margins

A deep interproximal box is a challenging clinical situation, given the difficulties in isolation subgingivally and the frequent lack of enamel in such situations. Regarding isolation, the authors suggest that rubber dam is of limited use for deep subgingival cavities and that isolation may generally best be obtained by the placement of a tightly wedged matrix band. In such a challenging isolation situation, the concept of proximal box elevation (also known as cervical margin relocation) comes into play, in which a material that is fast and simple to place is used to build up the interproximal box by 2–3 mm to a level at which isolation by rubber dam is possible (Figure 9). The authors suggest that chemically cured versions of a resin-modified glass ionomer (RMGI) are appropriate, given their fast and easy placement (with no intermediate bonding agent being required), lack of solubility (unlike some other glass ionomer variants) in the dilute organic acids found in plaque, reliable bond to dentine and some fluoride release.⁴⁷ The use of composite for this technique has also been suggested,^48,49 but has the disadvantage that the use of an intermediate bonding agent is required. Finally, it is stating the obvious that all margins, not only in the Class II box, must be adequately adapted, although the Class II box margin may be considered most important, insofar as it is generally difficult, or impossible to check on a regular basis. Importantly, it should also be added that, in cases where proximal box elevation is required (ie no enamel present at the base of the class II box margin), it is likely that the patient has high caries activity and therefore requires counselling regarding this.

Bonding technique

While it is important to follow manufacturers' instructions with regard to air thinning of a bonding agent in order to evaporate excess solvent, it is important to avoid using the three-in-one syringe at its full strength, as this could blow the bonding agent out of the cavity. This, in turn, will compromise the bond.

There is laboratory evidence^50,51 that rubbing in a bonding agent will improve bond strength, with recent work on bond strength to enamel, of seven universal adhesives used in self-etch mode, confirming this.⁵¹ There is also clinical evidence^52,53 that rubbing in a bonding agent will enhance the clinical performance of adhesives used in restoration of NCCL, with results of one study⁵² indicating an 82.5% retention rate after 2 years with passive application, and a 92.5% retention rate when the adhesive was scrubbed vigorously. Additionally, it may be of interest to note that the manufacturer's instructions for the clinical application of Scotchbond Universal (3M) indicate that this material should be rubbed in for 20 seconds.⁵⁴

Universal adhesives

Previous bonding agents were type specific, ie the self-etch bonding agents' bond strengths were compromised if the dentine was etched and the etch-and-rinse bonding agents performed suboptimally if the dentine surface was not etched. Universal adhesives can be used successfully in whichever etch mode is chosen by the clinician, but given that it has been considered that over-etching the dentine surface may lead to the formation of a fluid-filled layer beneath the hybrid layer, and pain on thermal stimulus and biting (see above), it is suggested that they are used in selective enamel-etch mode, ie do not etch the dentine. In this regard, there is a body of research that confirms the effectiveness of certain universal adhesives in selective enamel etch mode.⁵⁹ Reviews have also indicated that these materials have enhanced the performance of bonding to dentine,^59,60 while results of a randomized controlled trial have concluded that ‘this 2-year clinical evaluation showed that Scotchbond Universal performed similarly in restoring NCCLs compared with the etch&rinse (Single Bond, 3M, MN, USA) or self-etch (Clearfil SE, Kuraray) systems.⁶¹ It can therefore be expected that it will perform similarly in selective enamel etch mode. In addition, a review by Perdigao and colleagues has concluded that the clinical behaviour of Scotchbond Universal indicates that it is the current gold standard.⁶²

Bulk-fill composites

Bulk-fill composite materials, such as SDR (Dentsply/Sirona), were developed as a means of filling a cavity in a posterior tooth quickly. However, these required the placement of a layer of ‘conventional’ composite on the occlusal aspect of the restoration as their wear resistance was suboptimal. Such materials have been classified as bulk-fill base materials.⁴⁵ However, materials have been introduced (such as Filtek One, 3M, MN, USA, SonicFill, Kerr, Bioggio, Switzerland, and Tetric Evo-Ceram Bulk Fill, Ivoclar Vivadent, Leichtenstein) that do not require placement of a ‘capping’, and hence have been classified as bulk-fill restorative materials.⁴⁵ This group of materials have depths of cure of circa 5 mm owing to the presence, among other factors, of translucent fillers. There should, therefore, be no need for more than one light curing episode, although an additional interproximal light cure may be suggested, but this negatively impacts the potential time saving associated with these bulk-fill materials.