Clinical relevance of tests on bond strength, microleakage and marginal adaptation


Dental adhesive systems should provide a variety of capabilities, such as bonding of artificial materials to dentin and enamel, sealing of dentinal tubules, reduction of post-operative sensitivity and marginal sealing to reduce marginal staining and caries. In the laboratory, numerous surrogate parameters that should predict the performance of different materials, material combinations and operative techniques are assessed. These surrogate parameters include bond strength tests of various kinds, evaluation of microleakage with tracer penetration between restorative and tooth, two-dimensional analysis of marginal quality with microscopes and mapping of the micromorphology of the bonding interface. Many of these tests are not systematically validated and show therefore different results between different research institutes. The correlation with clinical phenomena has only partly been established to date. There is some evidence, that macrotensile and microtensile bond strength tests correlate better with clinical retention of cervical restorations than macroshear and microshear bond tests but only if data from different test institutes are pooled. Also there is some evidence that marginal adaptation has a moderate correlation in cervical restorations with clinical retention and in Class II restorations (proximal enamel) with clinical marginal staining. There is moderate evidence that microleakage tests with dye penetration does not correlate with any of the clinical parameters (post-operative hypersensitivity, retention, marginal staining). A rationale which helps the researcher to select and apply clinically relevant test methods in the laboratory is presented in the paper.


Adhesive dentistry has generally revolutionized dentistry. With the possibility to bond artificial materials to enamel and dentin, there is no need for mechanical retention outlines to keep a restoration or a material in place, both for direct and indirect restorations . This enables dentists to accomplish defect-oriented preparations, which means that they can limit the removal of sound tooth substance to that area of the defect that needs to be restored, whether it is a carious defect, an erosive defect, a tooth fracture or an esthetical defect.

With the advent of adhesive dentistry, it has become possible to bond artificial materials to other artificial materials like composites, ceramics or metal alloys. This enables dentists to repair rather than replace restorations in the case of chippings, fractures, marginal caries or esthetic improvements required by the patient . A prospective clinical trial over 7 years even showed that the repair of defective direct restorations is more successful than the replacement of such restorations . In prosthodontics, the mechanical retention of crowns and fixed partial dentures is less dependent on the macromechanical retention design of the prepared tooth structure or indirect restoration when adhesive technologies are applied .

In orthodontics, the movement of teeth with brackets and other appliances would not be feasible without bonding them to the enamel.

Finally, in preventive dentistry the sealing of fissures or initial carious lesions has only become possible with the concept of adhesive bonding to enamel .

This short list of everyday dental applications illustrates the extent to which adhesive dentistry has penetrated different parts of dentistry and changed dental routines. The problem for dental professionals is to decide which adhesive technique or material/system is adequate for which type of bonding indication.

Dentistry has seen ongoing changes and the launch of innovative products. However, innovation was mostly driven by reducing the time required to fabricate or place restorations. This has become possible by reducing the number of steps or materials required for bonding to either natural or artificial materials as well as by reducing the activation times and by bundling several adhesive indications into one “universal” product. The general practitioner, however, is rather confused with the variety of different materials on the market and the way they are promoted. The most common method to promote and market these products is by presenting data gained in laboratory tests. Yet, neither dentists nor laboratory researchers have a clue as to what these tests say on the possible clinical outcome in terms of predictability and longevity. Since conducting clinical studies is complicated and expensive and the results tend to be available only several years later, the easiest thing to do is to use results from laboratory tests or in vitro simulations. However, this begs the question as to how relevant the information from such in vitro tests is when it comes to predict the clinical performance of dental materials.

Laboratory test results have consistently provided the basis for recommendations on how dentists should use composite materials in their daily clinical routine. Examples are special incremental and layering techniques for direct composite restorations , the use of a flowable composite under more viscous materials , soft-start polymerization and selective bonding instead of total bonding . But then, years later, clinical studies discovered that the restorations placed according to these recommendations did not perform better than those performed with other, usually simpler techniques. For instance, restorations placed with translucent wedges and transparent matrices do not better than those placed with wooden wedges and metal matrices . Likewise, restorations bonded with a “selective” bonding technique perform equally well as those bonded with a “total” bonding technique . To finish, cervical restorations polymerized with soft-start polymerization are comparable to those polymerized without soft-start polymerization . Two clinical studies even questioned the incremental technique in posterior composite restorations after the clinical results of fillings placed in bulk (one increment) in mid-sized cavities were comparable to the results of studies which used the incremental technique .

Thus, the purpose of this article is to critically analyze standard laboratory tests as a means of evaluating the adhesive properties of adhesive materials used for bonding to the natural tooth structure and artificial materials. Rather than looking at tests and then trying to correlate them to the results from clinical trials, this evaluation was conducted in the reverse order: First, we analyzed the evidence from clinical trials with respect to failures and their frequency and then looked at laboratory tests that may have simulated the phenomena observed in the clinic.

This article does not focus on the chemical or other interactions of adhesive systems with the dentin/enamel or other substrates nor does it cover the adhesion to restorative materials like zirconia, glass ceramics or alloys.

When do dentists need adhesion?

Adhesion to tooth substance is required in clinical situations if an inadequate amount of tooth substance has to be removed to produce enough mechanical retention. However, what does “sufficient retention” mean? Although not systematically analyzed, common sense in dentistry says that adhesion is required in the following clinical situations:

  • cervical restorations

  • build-up of fractured teeth (posterior/anterior)

  • short clinical die for full-coverage crowns or partial crowns

  • minimally invasive bridges like adhesive bridges with small lingual metal or ceramic retainers

For restorations in the posterior region (Class I, Class II), the preparation itself and/or the caries removal normally generates enough mechanical retention, thus making adhesion to tooth substance less important. The guidelines of cavity design preparations published by Black have been obsolete for several decades and were not even applicable for amalgam restorations .

However, adhesion is not only necessary for keeping the restoration in place but should also fulfill other tasks such as

  • sealing dentinal tubules to reduce post-operative sensitivity and

  • sealing restoration margins to reduce the risk of marginal staining and marginal caries.

When do dentists need adhesion?

Adhesion to tooth substance is required in clinical situations if an inadequate amount of tooth substance has to be removed to produce enough mechanical retention. However, what does “sufficient retention” mean? Although not systematically analyzed, common sense in dentistry says that adhesion is required in the following clinical situations:

  • cervical restorations

  • build-up of fractured teeth (posterior/anterior)

  • short clinical die for full-coverage crowns or partial crowns

  • minimally invasive bridges like adhesive bridges with small lingual metal or ceramic retainers

For restorations in the posterior region (Class I, Class II), the preparation itself and/or the caries removal normally generates enough mechanical retention, thus making adhesion to tooth substance less important. The guidelines of cavity design preparations published by Black have been obsolete for several decades and were not even applicable for amalgam restorations .

However, adhesion is not only necessary for keeping the restoration in place but should also fulfill other tasks such as

  • sealing dentinal tubules to reduce post-operative sensitivity and

  • sealing restoration margins to reduce the risk of marginal staining and marginal caries.

Which failures do occur with restorative materials in clinical studies?

Cervical restorations (Class V)

Class V non-carious non-retentive lesions are frequently used to clinically evaluate the effectiveness of adhesive systems. A meta-analysis on the clinical effectiveness of cervical restorations with glass ionomer derivates and composites revealed the following results :

  • On average, 10% of the cervical fillings were lost and 24% exhibited marginal discoloration after 3 years. The variability ranged from 0% to 50% for retention loss and from 0% to 74% for marginal discoloration. Hardly any marginal caries was detected.

  • The adhesive/restorative class had the most significant influence, with two-step self-etching adhesive systems performing best and one-step self-etching adhesive systems performing worst; three-step etch&rinse systems, glass ionomers/resin-modified glass ionomers, two-step etch&rinse systems and polyacid-modified resin composites were ranked in between.

  • Restorations placed in teeth with prepared dentin showed a statistically significant higher retention rate than those placed in teeth with unprepared dentin.

  • Beveling of the enamel and the type of isolation used (rubber dam/cotton rolls) had no significant influence.

Posterior direct restorations (Class I/II)

As early as in the 1970s of last century, composite resins were placed in posterior teeth. Those resins were peroxide initiated macrofilled composites that were placed in bulk. In those days, the enamel was not etched with phosphoric acid and the cavities were drilled in the same way as for amalgam fillings . Mostly calcium hydroxide or glass ionomer materials were placed as liners under the composite resin restorations. In the 1980s etching of the enamel became integrated into the operational procedure and it became common practice to use an unfilled, hydrophobic low-viscosity bonding material between resin and enamel. In the late 1980s the first dentin bonding agents were developed, but these materials still required separate etching of enamel. This method was later replaced by the so-called “total etch technique”, which involves the simultaneous etching of both enamel and dentin. In 1999 the first self-etching enamel–dentin adhesive systems were released on the market. In the meantime, these systems – either one-step or two-step – have gained popularity amongst dental practitioners (dental professionals) because they shorten and simplify the operational procedures.

A systematic review was carried out on composite posterior restorations . Fifty-nine studies met the inclusion criteria with more than 7100 composite restorations at baseline and more than 700 of which had been placed without enamel etching and/or adhesive system. The following conclusions relevant for the adhesive system can be drawn:

  • Restorations that were placed without enamel etching and bonding showed significantly more post-operative sensitivity than restorations that were placed with enamel etching or self-etching adhesives; there was no significant difference between the latter two groups. However, the difference between the enamel etching and the no etching/no bonding group was less than 2% of the restorations affected. This is in contrast to other studies that assume a frequency of post-operative sensitivity in posterior composite restorations that varies between 1% and 29% . Risk factors for the development of post-operative sensitivity include deep and large cavities, the removal of the smear layer and the type of tooth . For pulpal reactions and hypersensitivity, a multifactorial etiology must be assumed, involving – besides bacterial penetration due to gap formation – contraction stress of the composite at dentinal walls, residual caries, inadequate sealing of the opened tubules, etc. .

  • The frequency of marginal caries was low in most studies with a median prevalence of about 4% after 10 years for restorations with etch&rinse systems. Statistically, the restorations that were placed without enamel etching and bonding experienced significantly more caries compared to those that were placed with enamel etching or self-etching. However, the difference between both groups was again very small (3% versus 2% after 5 years) ( Fig. 1 ).

    Fig. 1
    Estimated median percentage of restorations across the studies and experiments without caries adjacent to the restoration in relation to the type of adhesive system .
  • The increase in restorations with marginal staining was dependent on the tooth conditioning technique. When the enamel was not acid-etched with phosphoric acid and when no adhesive system was applied, the decline was rapid and after 4 years already 60% of the restorations showed marginal staining. By contrast, marginal discoloration was found in about only 10% of the restorations if the enamel was etched and in about 20% if a self-etching system was used ( Fig. 2 ).

    Fig. 2
    Estimated median percentage of restorations across the studies and experiments without marginal staining in relation to the adhesive technique, adhesive system and observation time .
  • As for the outcome variable marginal integrity , the results were similar for restorations placed with etch&rinse and with self-etch techniques but not for restorations fabricated without enamel conditioning. For the latter group, the deterioration was not as rapid as in the case of marginal staining; after 4 years 35% of the restorations had detectable margins on average.

In cross-sectional investigations, marginal caries was usually cited as the most frequent reason for replacing a restoration, irrespective of the restorative material used and irrespective of the type of cavity and location in the mouth . There are several reasons for this discrepancy. Besides financial reasons due to the remuneration system, practitioners most often confound marginal staining with marginal caries. This holds true for both amalgam and composite restorations. Studies have shown that most of the dentists are not able to clearly identify caries that is associated with restorations. In one study, 9 dentists evaluated 228 teeth with different types of restorations . In only 37% of the cases, the diagnosis of the dentists corresponded with the histological analysis of the tooth with regard to caries and there was a considerable variability between the dentists. Special training and calibration can considerably increase the inter-examiner agreement .

Marginal caries does significantly occur more in patients who have a high caries activity. Unfortunately, only few clinical trials have evaluated the individual caries activity of the included subjects. In a prospective 5-year clinical study on posterior resin restorations that defined caries activity of a subject as having either more than 30 restored surfaces and/or high counts of mutans streptococci all the seven restored teeth with marginal caries out of the 51 available restorations were in the group of subjects with high caries activity .

Several in situ and clinical studies have shown that marginal staining is no indicator for marginal caries . The meta-analysis that was cited above showed that after 10 years the number of teeth with marginal staining was five times higher than the number of teeth with marginal caries on average (acid-etch), which supports the findings of in situ studies. Another fact that is of paramount importance is that practitioners most often judge the quality of a restoration by looking at its easily visible parts – most often the occlusal surfaces of posterior fillings or the labial surfaces of anterior fillings. These surfaces, however, are the areas where marginal caries rarely occurs. A clinical trial has shown that marginal caries is about 8 times more frequent at the gingival section of proximal surfaces in Class II composite fillings than at the occlusal site . This can be explained by different facts:

  • In the occlusal part of Class II restorations, the formation of a biofilm at the interface between the restoration and the enamel is restricted or impossible because of self-cleaning mechanisms in the mouth (saliva, tongue, etc.) and oral hygiene measures conducted at home (toothbrush, toothpaste). In contrast, the biofilm can develop relatively undisturbed, differentiate itself and offer a biological habitat for cariogenic bacteria at the proximal box of Class II restorations. Initially, after a Class II restoration has been placed, the concentration of interdental Streptococcus mutans decreases quite dramatically because of the manipulations that have been taking place. It takes about 6–9 months before the S. mutans concentration regains the baseline value and starts to increase again . Differences have been found among different restorative materials. In clinical and in situ studies, a significantly larger number of S. mutans was identified in the interdental plaque which had formed on the composite surfaces, compared to the interdental plaque adjacent to the unrestored enamel or amalgam restorations in the mouth of the same test subject.

  • The quality of the dentin and enamel bond in the gingival cervical area of Class II restorations is inferior to that in other areas of the cavity, even if a good adhesive system is used . If the filling’s margin lies just above the cemento–enamel junction, the operator has to deal with enamel which becomes ever thinner toward the edge and breaks easily. Furthermore, there are hardly any enamel prisms to etch at this site. Furthermore, a filling’s margin below the cemento–enamel junction lies in an area of dentin which contains no dentinal tubules in an area of approx. 100 μm, and further toward the pulp their number is still quite low; thus, the preconditions necessary for a good bond are not met .

  • A possible other important reason for the predilection site of marginal caries at the gingival margin of Class II restorations is the inadequate polymerization of the composite at this location, which will not be discussed further.

Anterior direct restorations

Another meta-analysis was carried out on anterior restorations (Class III, Class IV) and provided similar results to the analysis on posterior restorations, even if considerably fewer clinical studies could be retrieved in the literature and included in the review (21 studies, 2008 restorations at baseline) :

  • Marginal caries was also infrequent in anterior restorations and accounted for about 2.5% of replacements after a period of service of 10 years, independent of the tooth conditioning system used.

  • Anterior restorations placed with self-etching systems were about two times more likely to show marginal staining than those placed with the phosphoric acid enamel etch technique after 5 years (30% versus 15%).

Full-coverage indirect restorations

The luting agent should help to prevent the restorations from being dislodged. There is an inverse relationship between the adhesive properties of the luting agent and the macromechanical retention. The lower the macromechanical retention is the higher the adhesive strength has to be to compensate for the low degree of retention. Long-term clinical studies (up to 10 years) indicate that retention loss in single crowns occurred more often in crowns and FPDs that were cemented with zinc phosphate cement than with glass ionomer cements ; in both studies the retention loss frequency was 5% for zinc phosphate cement and 0% for glass ionomer cement. A meta-analysis of metal-ceramic FPDs ( n = 1307) revealed an estimated 10-year risk of 6.4% for retention loss (CI 3.9–10.4) ; for all-ceramic restorations the estimated 5-year risk for retention loss was calculated to be 2.3% (CI 1.2–4.6) . However, these studies did not differentiate between zinc phosphate and glass ionomer cements or other cements.

A systematic analysis in the longevity of metal-ceramic FPDs compared to zirconia-supported FPDs included also the evaluation of debonding/decementation . The frequency of debonding was higher for restorations luted with zinc phosphate cement than for restorations luted with any other type of luting agents, none of which revealed any difference in performance (RelyX Unicem, Panavia, glass ionomer cements, resin-modified glass ionomer cements). These data confirm the findings of the above mentioned studies on glass ionomer and zinc phosphate cements. Consequently, none of the luting agents, except for zinc phosphate cement, has a critical effect on debonding in standard routine preparations of single crowns and FPDs. In clinical situations with no or low mechanical retention, such as preparations for Maryland bridges or endo-crowns (endodontic crowns) or in situations with reduced stump height ≤3 mm and/or high angle of convergence (>30°), the adhesive properties of the luting agents are crucial for the prevention of debonding.

Conclusions from clinical trials

The most important conclusions to be drawn from the clinical trials are as follows:

  • Marginal caries most often occurs at the proximal gingival margins of restorations. It is, however, infrequent and does occur practically independently of the sealing properties of the adhesive system. Patient-related factors like high S. mutans counts and sucrose-rich diet are promoting factors for marginal caries.

  • Marginal staining develops on the grounds of marginal irregularities (gaps, fractures). The frequency in relation to time of service depends on the type of adhesive system. Marginal staining per se is not linked to marginal caries. However, many practitioners confound marginal staining with marginal caries. Therefore, laboratory tests can help to identify products and mechanisms that show good in vitro performance with high external validity (correlation to marginal staining) and therefore also help to reduce the premature replacement of restorations because of suspected marginal caries.

  • Most post-operative sensitivities seem to be non-persistent and mostly resolve with time.

  • Retention of cervical restorations depends on dentin preparation and the adhesive system.

  • Retention of posterior restorations depends on the macromechanical preparation design of the teeth involved as well as the long-term bonding properties of the adhesive systems.

Which laboratory tests provide clinically relevant information?

Requirements for laboratory tests

Laboratory tests are useful for testing new operative techniques and materials before they are clinically implemented. The methods employed, however, should meet the following requirements :

  • 1.

    The results must be reproducible, i.e., when the same test is repeated under the same conditions and with the same materials, the same results should be obtained.

  • 2.

    The parameters which influence the test results must be known.

  • 3.

    The variability of the measured values must be low and within an acceptable range. The coefficient of variation, that is, the ratio of the standard deviation to the mean, should be under 20%. The coefficient of variation determines the number of specimens per group.

  • 4.

    If devices are employed for the test itself and/or to measure parameters and post-testing conditions of the specimens, then these devices must be suitable for the given purpose, that is, they must be qualified. This, in turn, must be proven and documented. A device may have to be calibrated before performing the test or measurement.

If all of these requirements are met, the test method is internally valid . Because the test method is used to provide information or prognoses about the clinical suitability of the material, the results must correlate with clinical findings. If this is the case, the test is also externally valid .

These requirements were described for medical devices and compiled under the title “ Good Laboratory Practice ” by regulatory authorities such as the Food & Drug Administration (FDA) in Washington or the European authorities in Brussels in the 1970s and 1990s, respectively . These specifications apply to medical devices in general and are not defined specifically for dental materials.

Bond strength tests


Two materials are normally bonded together by applying an adhesive between them. To test the strength of such a configuration, a straightforward method is to try to separate both materials with an adequate testing machine after having bonded them and to measure the force required to separate them. Various industrial sectors, such as the construction, aircraft or automotive industries, use established bond tests. For many of the different materials ASTM standard tests are published.


If PubMed is searched, the first article on bond strength tests for dental materials was published in 1965 by Bowen . In this publication, a tensile bond strength test with a restricted bonding area was elaborated and described. Since then, 4960 articles are listed in PubMed when searching for “bond strength” and “dental” and 2695 articles for “bond strength” and “dentin” and 1545 articles with the search terms “bond strength” and “enamel” (search period 1955–2012, search month April 2012). These articles advocate various test setups, such as the shear bond, microtensile, microshear, push-out and the fracture toughness test (see below). There is only a small degree of standardization, which explains the variation in results. Of the 4960 articles, there are only 12 relevant reviews which critically evaluate the different bond strength tests in terms of their strengths and weaknesses . All the other thousands of studies on bond strength testing in dentistry investigated various modifications of test setups and substrates. Additionally, most of them focused on testing specific materials. The first publications on the correlation between bond strength tests and clinical trials appeared only in 2010 and 2011 (see below).

Scientific evidence of bond strength tests

A systematic multi-center analysis on the factors influencing the various bond strength tests has not yet been performed. However, several factors that affect the results and are crucial for the implementation of laboratory bond strength tests can be identified on the basis of reviews on bond strength tests as well as on the basis of 40 years of bond strength testing published in numerous papers and presented at dental congresses:

  • Biological substrate : Studies indicate that bovine teeth is a good substitute for human teeth . This has been proven for the microtensile test , the tensile test and the shear bond test . However, only coronal bovine dentin should be used and not root dentin.

  • Cutting of teeth : Cutting should be performed with a qualified sawing equipment at low speed without producing cracks and fractures. Regular maintenance and calibration of the sawing equipment is a prerequisite.

  • Cutting level : The gingiva-coronal height (human molars) or labio-oral depth (bovine incisors) should be similar for all teeth. Bond strength values are higher in superficial dentin than in deep dentin .

  • Non-carious versus carious dentin : Bond strength values are higher in non-carious dentin than in carious dentin .

  • Eroded versus non-eroded dentin : Bond strength to artificially eroded dentin has been proven to be significantly lower than to sound dentin .

  • Regional differences within the same tooth : By means of microtensile tests, significantly lower bond strength values were found at the gingival floor of Class II restorations than at the axial or occlusal wall . This may be explained by the histological conditions which are less favorable to establish an adequate bond at this site (see above).

  • Number of specimens : As the coefficient of variation is between 20% and 40% – depending on the adhesive system to be tested and on the test method – it is necessary to use between 10 and 40 specimens per group provided that a difference of 25% in bond strength value is to be considered statistically significant. For Weibull statistics, at least 15 specimens are needed.

  • Preparation of substrate : It has to be ensured that the embedded substrate is polished virtually 100% parallel to the level of the testing device and 100% perpendicular to the composite cylinder.

  • Direction of force and size of bonded area : Bonded composite cylinders can be sheared (shear bond) or torn (tensile bond) away or pushed out of prepared cavities. The disadvantage of shear bond tests is that the composite cylinder is compressed while inadequate tensile forces are created in the tooth substrate due to the bending moment, as FEM calculations have shown . This explains why many cohesive failures occur with shear bond tests. In contrast, no such forces occur with tensile bond tests if performed correctly. In 1994 the tensile test with specimens of a bonded area of 3 mm was modified by cutting small sticks of 0.7 or 1.0 mm from a single tooth . Small bonded areas show higher values than large bonded areas . The advantage of these microtensile tests is that fewer extracted teeth are necessary and that regional differences in the tooth can be taken into account. However, the intra-tooth variability is greater than the inter-tooth variability . In 1998 another shear test with smaller specimens was proposed . In 1993 the so-called fracture toughness test was introduced to characterize the fracture resistance of the adhesive interface between tooth substance and resin by paying attention to crack initiation and propagation . Studies on the ranking of different adhesive systems in relation to the bond strength test carried out on them yielded conflicting results .

  • Restricted adhesion area : The adhesive area should be restricted to the area of the composite cylinder. Excess of adhesive creates higher bond strength than restricted areas and creates additional stress in the substrate .

  • Type of composite : In shear bond tests, the modulus of elasticity has an effect on the test result. The higher the modulus of elasticity, the higher is the bond value . Therefore, the same composite material should always be used, at least for shear bond tests.

  • Pre-test failures : If the bonded specimen falls (breaks) apart before testing, these specimens should receive a value “0”.

  • Test equipment : The universal testing machine has to be qualified for the intended use. The machine has to be calibrated on a regular basis by a qualified body .

  • Time interval between bonding and testing : Immediate bond tests do not give clinically meaningful results nor do they after 24 h. They should only be regarded as baseline values to interpret the decrease in bond strength after stressing and prolonged storage (see below).

  • Thermocycling : Short-term thermocycling (500×, 5 °C/55 °C) does not significantly alter the ultimate bond strength values . But also prolonged thermocycling (up to 10,000 cycles), has little effect on both the microtensile and macroshear bond tests . Systematic research on the influence of thermal stress on the bond strength has not been carried out to date. Therefore, it remains unclear whether the decrease in bond strength was caused only by exposure to water for a prolonged time or indeed by thermal stress or by both . There is no evidence that thermal stress alone leads to a clinically relevant deterioration or debonding of the adhesive interface.

  • Storage of bonded specimens in water : The storage of specimens in water (37 °C) results in a decrease in the bond strength compared with the baseline value – depending on the adhesive system and the test . Water can lead to a hydrophilic degradation of the dentin/enamel-composite interface.

  • Statistics : When microtensile or microshear tests are applied, sticks originating from the same tooth are not statistically independent of one another. If this dependency is not taken into account, it is entirely possible that an erroneous product ranking may result .

  • Variability : Study results from different study groups on bond strength tests are not comparable and even within the same study group great variability of the results have been reported. The intra-institute variability is about 20–40% (coefficient of variation) whereas the inter-institute variability is between 30% and 50%, with slight differences between the various test methods .

  • The microtensile bond strength of specimens from Class II restorations in vivo is lower than that of restorations in vitro .

Standards on bond strength tests

In 2012 an ISO standard on bond tests will be published (No. 29022: Dentistry – Adhesion – Notched-edge shear bond strength test). The test uses the Ultradent jig, which bears a notch which surrounds half of the specimen . The adhesive is applied unrestrictedly on either bovine or human dentin. The composite cylinder with a diameter of 2.38 mm is sheared off after 24 h storage of specimens in water. A round-robin test has been performed amongst several dental manufacturers with three adhesive systems. The results of the round-robin test, however, have not been published yet.

This standard will not replace the ISO Technical Specification (TS) with the title “Testing the adhesion to tooth structure” (No. 11405, first edition 1994, second edition 2003). The specification will, however, be revised, deleting e.g. the part of shear bond test. This specification represents still a useful approach to standardize some important variables for tensile bond strength . Unfortunately, few research workers follow these recommendations . Some important parameters listed in the ISO are:

  • Defined and limited bonding area

  • 6-month water storage before testing

  • Weibull statistics with a minimum of 15 specimens per group

  • Tensile bond strength: 90° angle alignment of the tensile forces acting on the specimen

  • If the coefficient of variation is above 50%, a thorough inspection of each process is recommended.

Because no internationally recognized standardized test protocol for the testing of adhesive systems is yet available, completely different bond strength values can be found and published for the same product, depending on the test institute.

A systematic review published in 2008 showed that the shear bond test was the most common type of test (used by 46% of the studies reviewed), human dentin was used in 77% of the investigations, and 24-h post-placement bond strength testing was the predominant time of specimen testing (67%) . The dimensions of the contact surface area or the specimen diameter was stated in only 65% of the papers reviewed.

Validation of bond strength tests

Few attempts have been made to validate the bond strength tests applied in dentistry. For the microtensile test several critical factors that influence the result have been identified, such as the diameter of the stick, type of jig, and trimming versus non trimming . For the shear bond test, no systematic analysis of influencing factors other than the universally applicable ones (thermocycling, restricted bond area, operator variability) have been published. One study indicated that the crosshead speed influences the test result .

As far as the comparability of the different test methods is concerned, Scherrer et al. published a review paper on the correlation between different bond strength test methods . The authors selected the following tests: shear, microshear, tensile and microtensile. The review revealed that a large variability for the same adhesive system evaluated with the same bond strength method was present not only at different test institutes (inter-institute variability) but also at the same test institute (intra-institute variability). The variability was similar for each test method. Then, the authors pooled the bond strength results for the individual adhesive system and the test method. The ranking between the six adhesive systems varied depending upon the test method chosen. However, this result must be interpreted with caution because the study pooled values which derived from the same test method but different test parameters. Only the data of adhesive systems that are tested with the same test parameters. This approach was applied in the following comparative analysis: In Fig. 3 , the mean Log-transformed bond strength results after 24 h of water storage of 12 different dentin adhesive systems tested with the microtensile, tensile and shear bond strength test are shown in a scatterplot together with an interpolation curve for each test method. The studies were retrieved from PubMed with a search for specific material combinations (search period 1946–2012, search month April 2012). Only studies that applied the same test protocol (human teeth, mid-coronal dentin, same diameter of test specimen according to test method, 24 h storage in water) were included; the included studies ( n = 24) comprised at least 4 different materials. It can be easily noticed that the absolute values differ widely for the same material and the same test method, highlighting the influence of the test institute. In most cases, microtensile tests yield higher values than the other two tests. If the test results are pooled in relation to the test method, none of the test methods is able to clearly differentiate between the different materials (see interpolation line in Fig. 3 ). It seems, however, that 1-step self-etch systems are associated with lower microtensile strength values than the other three types of adhesive groups (self-etch – 2 steps, etch&rinse 2 and 3 steps). No such clear differentiation can be observed with the tensile bond strength and shear bond strength test. In another approach, the studies were further reduced to those that evaluated the same adhesive systems. In order to end up with at least 3 studies using a microtensile test, it was possible to include only 3 adhesive systems. The absolute figures differ widely, as can be seen in Fig. 4 , but not the relative rankings between the materials except for one shear bond test (SB 2). The differences between the materials are more pronounced with the microtensile test method than with the macrotensile and macroshear bond strength test. These results confirm the results of a meta-analytic review of parameters involved in dentin bonding . Other important findings of the review were as follows:

  • The microtensile test discriminates more effectively between different adhesive systems than the macroshear bond test.

  • Thermocycling has a negligible influence on the decrease of bond strength – both for the microtensile and macroshear test.

  • Long-term storage in water significantly decreases the bond strength if tested with the microtensile method. However, if tested with the macroshear test, no significant decrease was detected.

  • The following parameters significantly affected the bond strength results (in decreasing order):

    • research group/institute

    • adhesive system

    • adhesive class (etch&rinse 3 step/2 step, self-etch 2 step/1 step)

    • substrate preparation (bur-cut or SiC-cut)

    • substrate origin

    • flexural modulus of composite.

Fig. 3
Log-transformed bond strength values (MPa) of 12 adhesive systems tested with 3 bond strength tests at 24 test institutes. At least 4 adhesive systems were tested at each institute ( references of studies are available on request ). Etch&rinse systems – 3 steps: 1 = OptiBond FL, 2 = Scotchbond Multipurpose. Etch&rinse systems – 2 steps: 3 = Excite, 4 = One Step, 5 = Prime&Bond NT, 6 = Single Bond. Self-etch – 2 steps: 7 = Clearfil SE, 8 = AdheSE. Self-etch – 1 step: 9 = iBond, 10 = Xeno III, 11 = Adper Prompt, 12 = G-Bond.

Fig. 4
(Top) Mean bond strength test results (MPa) of three adhesive systems tested after 24 h storage in water, using three test methods – 3 studies with microtensile bond strength test (MT), 2 with shear bond test (SB) and one with tensile bond test (TB). The data of the 3 systems were published in the same article and for MT and SB the test parameters were the same. (Bottom) The same test results presented with relative ranks (1–3). MT 1 , MT 2 , MT 3 , SB 1 , SB 2 , TB .

Correlation with clinical findings

In most cavity and abutment preparation designs, the largest area exposed after preparation is dentin. Therefore, bond strength on dentin should have its clinical correlate with retention. The sealing of the dentinal tubules is another important function of adhesive systems. The bonding to enamel is less important for the retention of the restoration but plays a more essential role in reducing gaps at the restoration margins and in preventing subsequent marginal staining. In posterior restorations about 90% or even 100% of the restoration margin is in enamel.

Only after 30 years of bond strength testing, efforts were made to relate the results of these tests to clinical findings. Even now, there are only a few publications which correlate the in vitro bond strength data with the clinical outcome of the tested adhesive systems. The clinical model most often used to test the effectiveness of adhesive systems is the restoration of non-carious cervical defects. Such defects are especially suitable to test adhesive systems due to: (1) practically no macro-mechanical retention is present, (2) straightforward clinical placement of the restoration and evaluation of debonding, reducing operator and evaluator variability, and (3) high prevalence, which makes patient selection simple and enables properly designed studies.

Only three publications on this topic have been found. In one of these studies, the microtensile bond strength data of 15 adhesive/restorative systems placed by the same operator were correlated with the clinical studies of non-carious cervical Class V restorations. No correlation was found between the retention rate of cervical restorations after 3 years and the microtensile test results after 8 h or 6 months of water storage ( Fig. 5 ) . There was, however, a very moderate correlation between marginal staining and bond strength values after 6 months of water storage. A comprehensive database of microtensile bond strength data and an equally comprehensive database on the retention rates of restorations placed in non-carious wedge-shaped defects at the same test institute (University of Leuven) found a moderate correlation for laboratory specimens that were submitted to artificial aging and the correlation was higher for 5-year data than for 2-year data . Another attempt was made to correlate the bond strength data with the retention of cervical restorations. In 2010 Scherrer et al. published data of laboratory studies on six dentin adhesive systems, available in the literature, and four laboratory methods (macroshear, microshear, macrotensile and microtensile bond strength test) . They pooled the data across the different studies in relation to the adhesive system and the bond strength test and calculated mean values and standard deviations. These data were correlated to estimated pooled 2-year retention rates of Class V restorations using the same adhesive systems and retrieved from the databank of the meta-analysis on cervical restorations . The results of the regression analysis for the pooled data revealed that only the macrotensile (adj. R 2 = 0.86) and microtensile tests (adj. R 2 = 0.64), but not the shear and the microshear tests, correlated more accurately with the clinical findings ( Fig. 6 ). This finding suggests two points:

  • Pooled data across different institutes may correlate more accurately with retention loss of cervical fillings than individual data from one test institute. The explanation may be that these pooled data characterize the variability and efficacy of a certain bonding system more appropriately.

  • Tensile tests correlate with the retention loss of cervical restorations, whereas the shear tests do not.

Fig. 5
Scatterplot of retention versus microtensile data for 15 adhesive systems, together with Spearman correlation rho and associated p -value . The numbers refer to different adhesive systems. Estimates of the adhesive effects in this model were used to summarize the clinical performance of each adhesive between 12 and 36 months. They were inverted and centered in such a manner that a positive value corresponds to a performance above average, and a negative value to a performance below average (a zero value represents average performance).

Fig. 6
Scatterplot and regression curves of mean pooled bond strength data of six adhesive systems in relation to the estimated pooled clinical data after 2 years for the same adhesive systems .

Therefore, bond strength tests should be carried out by different operators and/or research institutes to determine the reliability and technique sensitivity of the material under investigation.

The correlation of bond strength test values with the retention rate of cervical fillings might have been better if the tensile test had been performed with artificially eroded dentin rather than with non-eroded dentin. Eroded dentin reflects the clinical situation of non-carious cervical restorations more accurately, as in most clinical cases erosion is the cause for the cervical defect and most clinicians do not prepare the dentin, but rather only clean it with pumice or prophylaxis paste. Adhesive systems perform significantly worse in eroded dentin than in non-eroded dentin, as an in vitro study has shown . This finding is supported by clinical evidence as retention of fillings placed in unprepared cervical defects was significantly lower than those placed in prepared cervical defects as a meta-analysis has revealed . In this analysis the percentage of studies that included a preparation of dentin was about 50%.

Most of the bond-strength tests were carried out on dentin and – to a smaller extent – on enamel specimens. A good bond to enamel is important to reduce the risk of marginal gap formation and staining. However, studies investigating the correlation of bond strength values of cut enamel and the occurrence of marginal staining in posterior or anterior restorations have not been found in the literature.

If the bond strength tests are applied to fissure sealants, the results are misleading when compared to clinical results. For example, the bond strength tests showed an equal or even higher bond strength for the self-etching Primer Adper Prompt L-Pop in uncut enamel compared to the phosphoric acid etching method. This was true for both shear bond tests and microtensile tests . Light-polymerizable fissure sealants with phosphoric acid conditioning of the enamel produce a retention rate of about 80% after 2, 3 and 5 years, as a meta-analysis has proven . For Adper Prompt L-Pop self-etching primer/sealants, however, the clinical trials showed poor retention rates. In one clinical study, the retention rate of the self-etching primer Prompt L-Pop/Clinpro was only 44% after 1 year and in a school-based programme sealants placed with Prompt-L-Pop were 6 times more likely to show retention loss after 1 year than sealants placed with phosphoric acid conditioning of the enamel .

Assessment of bond strength tests

The value of bond strength tests to make a prognosis on the clinical performance of an adhesive system or dental material is limited. A bond strength test helps developers of new products to select between different material variants. The test should be standardized as much as possible and well documented. The universal testing machine should be qualified for the intended purpose. To assess the variability of the adhesive efficacy of adhesive systems, different operators with different levels of experience in adhesive testing should perform the test in the same institute or in different institutes. Pooling of results enhances significantly the correlation with the clinical performance.

Shear bond tests (macro and micro) are inadequate for the evaluation of the bond strength to dental hard tissues as clinically unrealistic stress is produced within the reaction zone. Tensile tests avoid such internal stress formation and their results correlate more accurately with the retention rate of cervical restorations. Microtensile tests are laborious and technique-sensitive and offer no advantage over macrotensile tests. For carrying out tensile tests, the ISO TS should be followed .

Testing the fracture toughness of the bonding interface can be regarded as a supplementary test to the macrotensile test, because it provides additional information. As it is technique-sensitive, it is not a routine test and the correlation to the retention loss of cervical restorations has not yet been established.

Simulation of retention loss in extracted teeth

As bond strength tests yield very variable results, it may be reasonable to simulate retention loss of cervical restorations in extracted teeth. A laboratory trial attempted to simulate restoration loss in non-retentive cervical cavities. For this purpose, premolars with wedge-shaped defects were selected ( n = 12) and restored with an adhesive system (Prompt L-Pop). Prompt L-Pop and his successor Adper Prompt L-Pop have a low pH and a strong etching effect on dentin which may be – together with other peculiarities – responsible for low bond strength values on dentin . Restorations placed with this adhesive system are characterized by a relatively high retention loss over a period of 2 years in clinical trials. In the laboratory, the restored teeth were centrically loaded 1.2 million times with simultaneous thermocycling, followed by another 1.2 million cycles of eccentric loading on the lingual cusp of the premolars to produce tensile stress on the buccal side of the tooth where the restoration was located . The result was surprising. None of the fillings was lost after 2.4 million thermo-mechanical cycles. In a second approach, 11 different adhesive systems (self-etch and etch&rinse) and one glass-ionomer cement were applied in cervical restorations of extracted premolars ( n = 12 per group) and submitted to a rigid programme of artificial aging including storing the teeth for 3 × 6 months in water, and subjecting them to various thermocycling and simulated mastication procedures . None of the adhesive systems – not even with the glass-ionomer cement – showed a loss of restoration during the simulation trial. Restoration loss only occurred when the conditioner (polyacrylic acid) was omitted for the fillings with glass-ionomer cement or the phosphoric acid etching of dentin and enamel for the fillings with an etch&rinse system. In sum, retention loss could not be simulated in fillings of extracted teeth when the materials were applied according to the recommendations of the manufacturer. However, self-adhering composite materials that are applied without separate adhesive system (e.g. Vertise Flow) may show retention loss in cervical fillings of extracted teeth. But this has to be proven.

Microscopic evaluation of interface

The interface between composite and dentin/enamel can be evaluated with a microscope either by cutting the specimen and evaluating the interface with a scanning electron microscope (SEM), a fluorescence microscope or a confocal laser scanning microscope (CLSM) after marking the adhesive with a fluorescent dye . The depth to which the adhesive system penetrates the dental hard tissue and the quality and thickness of the hybrid layer can be examined under the microscope. However, because the fluorescent dye does not chemically react with the adhesive system, the dye molecules can penetrate further into the dentin tubules than the adhesive system itself, thus distorting the results .

Another method is to completely dissolve the tooth in hydrochloric acid and to measure thereafter the length of the resin tags and the thickness of the hybrid layer with SEM. Adhesive systems which condition the dentin and/or enamel with phosphoric acid demonstrate both better microretention than self-etching systems and thicker hybrid layers. The thickness of the hybrid layer, however, is not correlated with the bond strength, the marginal integrity or the clinical performance of the adhesive system. A qualitative microscopic analysis of the adhesive interface is a valuable tool to get an insight into the bonding mechanism of a dental adhesive. For instance, a study using silver nitrate penetration and TEM impressively showed where the paths of water infiltration occur at the adhesive interface . However, such investigations do not allow to make a prognosis on the clinical performance of a given material.

Evaluation of the marginal seal in vitro

Tracer penetration test


The hypothesis is that the penetration of different markers along the interface between the restoration and dental hard tissues in extracted teeth is a surrogate variable for the in vivo penetration of bacteria, fluids and other liquids, which may provoke hypersensitivity, pulpitis, marginal staining and/or marginal caries. Normally, restorations (Class V, Class II) are placed in extracted teeth, which are then subjected to a variety of different agents, such as organic dyes (fuchsine, methylene blue, rhodamine, erythrosine, eosin, etc.), silver nitrate, radioactive markers ( 45 CaCl 2 ) or bacteria. Alternatively, hydrodynamics may be used to measure the movement of liquids such as saline. The application of dyes represents the most commonly used method because of its simplicity. Penetration of dyes/tracers involves the cutting of teeth with sawing equipments.


According to PubMed, the first article on tracer penetration along the restorative hard tissue interface was published by Going et al. in 1960 . This study used radioactive tracers. Twelve years later, in 1972, the first author of this study summarized the knowledge available on tracer penetration/microleakage around dental restorations and concluded that more in vivo studies were necessary to validate the large amount of data collected in vitro. However, it was only 26 years later that efforts were made to correlate tracer penetration data to clinical findings (see below). In the meantime, several reviews on microleakage studies had been published ; many of them complaining on the non-standardization of the tests. The ISO standard on testing the adhesion to tooth structure describes a microleakage test in the cavities of third molars (mid-part of the buccal surface) with a diameter of 3 mm and a depth of at least 1 mm and a sample size of at least 10 . No specific dye tracer is recommended. An ordinal scale is used for the measurement of the tracer depth. No values for the acceptance of a given material had been postulated.

Scientific evidence on dye penetration studies

  • Type of substrate : Dye penetrations tests are carried out in both bovine and human teeth. However, a systematic analysis on the literature available was inconclusive on whether the study results achieved in bovine teeth can be compared to those achieved in human teeth .

  • Type of dye tracer : The type of dye plays a negligible role in dye penetration studies, except for methylene blue . The chemical solution of methylene blue is not stable at room temperature and under exposure to ambient light. Furthermore, hydroxyl ions can easily be reduced to leuco-methylene blue, which is colorless. In the presence of strong acids, the dye is not stable either.

  • Number of cuts : The more cuts are performed, the better is the level of agreement between test centers that have applied the same methodologies and materials .

  • Measurement device : The light microscope used for the evaluation of dye penetration should be qualified for this purpose as should be the software program designated to measure the length of the penetration.

  • Results according to substrate : Practically all the laboratory studies have shown that – within the same tooth and the same restoration – dye penetration is larger in sites where the margin is located in the dentin compared to sites where the margin is located in the enamel, independent of the adhesive system investigated .

  • Artificial aging : Unsystematic research on the influence of thermocycling and occlusal loading and the combination of both these stress factors has generated conflicting results . There is no evidence that thermal stress alone causes clinically relevant deterioration of the adhesive interface.

  • Comparability of results between test institutes : A systematic review on dye penetration studies for restorative dental materials concluded that a comparison of study results was not possible due to great variability in methodologies and parameters . The studies differ with regard to the substrate used (human/bovine), the dimensions of the prepared cavity, storage period until testing with dye tracer, type of stress and number of stress cycles (thermocycling, load cycling or a combination of both), the type and concentration of dye tracer, the length of immersion in dye liquid, the type of analysis of the dye penetration (cutting teeth), number of cuts and the evaluation method (metrically, score system).

  • Variability of results : The variability of test results is very high with a coefficient of variation up to 50% and more. In order to prove a statistical significant difference between test groups, many specimens per group are needed. One of the studies on dye penetration in Class II restorations revealed that a sample size of 12 would discriminate in the range a dye penetration of 1.0 mm at enamel margins and a dye penetration of 2.2 mm at dentinal margins .

  • Acceptance level : A validated acceptance level does not exist .

Correlation with other laboratory test methods

Dye penetration and bond strength

A systematic review on the correlation of bond strength tests and dye/tracer penetration did not reveal any significant correlation .

Dye penetration and quantitative marginal analysis with SEM

A correlation between dye penetration and the occurrence of marginal gaps is only partially proven, if at all .

Correlation with clinical findings

In vivo and in situ studies have shown that microleakage as such (alone) or the existence of marginal gaps does not correlate with the occurrence of hypersensitivity or the formation of marginal caries .

However, no systematic study has been conducted so far to correlate dye tracer or microleakage data to clinical findings on hypersensitivity, retention, marginal staining or marginal caries. It is not possible to extract data from the literature, because the methodologies of the dye penetration tests vary wildly from one study to another, thus making it impossible to correlate standardized laboratory data to clinical data. However, the author found microleakage data from the Clinical Research Associates test institute in Provo/USA (now: G Christensens Clinicians Report). In this study, cervical fillings were placed in extracted teeth using 18 different adhesive systems and the dye penetration was measured on an ordinal scale (published on an internet site in 2002, which is no longer available). The statistical analysis of these data in relation to the clinical results of Class V fillings revealed no correlation whatsoever for the variables of retention and marginal staining ( Fig. 7 ).

Nov 28, 2017 | Posted by in Dental Materials | Comments Off on Clinical relevance of tests on bond strength, microleakage and marginal adaptation
Premium Wordpress Themes by UFO Themes