INTRODUCTION

Despite the esthetic advantages of composite resins, restoration failure continues to be a problem. Studies have shown a significant increase in the size of cavity preparations when restorations are replaced.^1,2 Additionally, the esthetics of resins pose a problem of identifying margins at time of replacement and distinguishing resin from tooth structure. Complete removal of a defective composite restoration may result in unnecessary removal of additional tooth structure. Repairing an existing composite restoration may be a more conservative approach, especially if the defective restoration is very extensive or approximates the pulp. An alternative to replacing an entire restoration is to remove only the defective portion so that a repair can be accomplished. Minimally invasive repair of composites may extend the longevity of the restoration without further damaging teeth.³

The vast majority of composite restorations are placed in incremental layers and rely on the presence of the oxygen-inhibited layer of unpolymerized resin to bond the next increment.^4,5 The amount of unreacted methacrylate groups available for bonding on the surface of a composite decreases significantly after 24 hours.¹ Additionally, composites undergo hydrolytic degradation, which leads to swelling and crack propagation within the matrix and breakdown of the coupling agent, resulting in loss of filler particles.² As a result, bonding to aged or contaminated composite is unpredictable. Successful repairs depend on the development of an adequate interfacial bond between the old and new composite resin. Methods have been proposed to enhance the bonding through micromechanical and chemical means with repair strengths varying from 25% to 82% of the inherent strength of the composite substrate.^6-11

When performing composite repairs most North American dental schools teach the use of diamond finishing burs to create micromechanical retention followed by phosphoric acid etching and placement of a bonding agent before composite placement.³ Add & Bond (Parkell Inc, Edgewood, NY, USA) is a one-step primer marketed for composite repairs. The manufacturer claims repair strengths that surpass the inherent cohesive strength of the composite. Another chemical method is to silanate the resin before placing a bonding agent. This means placing Si-O groups on the exposed filler particles and matrix that can bond with the matrix of the repair resin.

The purpose of this in vitro study was to determine the optimal repair protocol by evaluating the flexural, diametral tensile, and shear bond strengths of composite repairs. The surface treatments were 1) phosphoric-acid etching and Optibond Solo Plus (OBSP; Kerr, Romulus, MI, USA); 2) Add & Bond, Parkell Inc, Edgewood, NY, USA ; 3) phosphoric-acid etching and Silane Bond Enhancer (Pulpdent Corp, Watertown, MA, USA) followed by OBSP; and 4) mechanical retention, phosphoric-acid etching and OBSP. Our null hypothesis was that there would not be any differences among the four repair protocols or any difference between the inherent strength of the resin substrate and the repair.

METHODS AND MATERIALS

The dental materials utilized and their compositions are listed in Table 1. The nano-composite Filtek Supreme Ultra (3M ESPE, St Paul, MN, USA) was used to make all specimens in this study by photo-curing the resin for 40 seconds with SmartLite IQ2 (Dentsply/Caulk, Milford, DE, USA) at an intensity of 500 mW/cm². When fabricating all specimens, a glass plate was placed against Mylar strips to compress the resin into the molds using finger pressure. Excess resin was removed before photo-curing. To delineate the repair joint during testing, two different shades of composite were used; B1B to make the original specimens and B5B as the repair composite. External surfaces of the specimens were visually examined, and a sharp knife was used to remove any burrs or irregularities that would prevent exact placement of the specimen into custom-made molds or testing jigs. Specimens were either repaired immediately or after being aged by placing them in 37°C deionized water for seven days. Twelve specimens were made for each control and experimental group.

Table 1: Materials Used in the Study

View Fullscreen

Flexural Strength

Composite resin (shade B1B) was placed in bulk in a stainless steel mold measuring 2-mm by 2-mm by 24-mm. The curing light tip measured 7 mm in diameter and was placed four times along the length of the beam, overlapping the previous photo-cured composite. These beams served as the positive control (n=12) for flexural strength testing. An additional 120 composite resin beams measuring 2-mm by 2-mm by 12-mm were made in a similar fashion. These were divided into 10 groups (eight experimental and two negative control groups) of 12 specimens. Five groups of 12 beams were repaired immediately and the other five groups were used for repairing aged specimens. One end of each beam was roughened with a coarse Sof-Lex (3M ESPE) disc for five seconds to simulate clinical finishing or partial removal of a composite resin restoration and then thoroughly rinsed for 10 seconds with air/water aerosol. The roughened end of the beam was treated with one of the experimental surface treatment protocols as explained in Table 2, placed back in the original mold, and composite resin in shade B5B was photo-cured against the treated end to make beams measuring 2-mm by 2-mm by 24-mm. Two negative controls consisted of immediate and delayed repairs without any surface modifications or treatments.

Table 2: Surface Treatment Protocols

View Fullscreen

RESULTS

The results for the three testing methods were not normally distributed. Therefore, geometric means, which are less affected by skewed data, were calculated and analyses (ANOVA and Tukey HSD multiple comparison) performed on the log scale of the geometric means.

Flexural Strength

The minimum, maximum, geometric means, and standard deviations for flexural strength are listed in Table 3 and failure modes in Table 4. The mean flexural strength of the intact resin beam was 102.9 MPa. The mean (geometric) repair strengths ranged from a low of 27.2 MPa (aged repairs using Add & Bond) to a high of 91.2 MPa (immediate repair using OBSP). All immediate and aged repairs were statistically similar to the intact resin beam except aged repairs using OBSP or Add & Bond. At best, 88% of the inherent flexural strength was achieved when the specimen was repaired immediately using OBSP. Aged specimens repaired using Silane Bond Enhancer with OBSP (68.6MPa) had more than twice the repair strength achieved using only OBSP (29.0 MPa). The immediate negative control (no surface treatment, primer, or dentin bonding agent) obtained 43% of the inherent flexural strength, and the aged negative control obtained 17% of the inherent substrate's strength. Most of the failures were classified as mixed or adhesive in nature.

Table 3: Flexural Strength (MPa)^a

View Fullscreen

Table 4: Failure Mode of Flexural Strength Repairs

View Fullscreen

Diametral Tensile Strength

The minimum, maximum, geometric means, and standard deviations are reported in Table 5 and failure modes in Table 6. The mean (geometric) diametral tensile strength of the inherent resin substrate was 59.0 MPa. Mean experimental diametral tensile repair bond strength ranged from a low of 18.0 MPa (immediate repairs using Add & Bond) to a high of 47.2 MPa (immediate repairs using OBSP). The diametral tensile bond strengths of all experimental groups were statistically less than the positive control except for immediate repairs using OBSP (47.2 MPa) or Silane Bond Enhancer with OBSP (44.6 MPa). Eighty percent of the inherent diametral tensile strength of the composite was obtained by an immediate repair using OBSP. The immediate negative control obtained 56% of the inherent diametral tensile strength and the aged negative control 19%. The failure mechanism was predominately mixed and adhesive.

Table 5: Diametral Tensile Strengths (MPa)^a

View Fullscreen

Table 6: Failure Mode of Diametral Strength Repairs

View Fullscreen

Shear Bond Strength

The minimum, maximum, geometric means, and standard deviations are illustrated in Table 7 and failure modes in Table 8. Mean (geometric) shear bond repair strengths ranged from a low of 9.4 MPa (aged/no treatment) to a high of 19.2 MPa (immediate/mechanical retention and OBSP). All experimental groups were statistically similar to the positive control (18.8 MPa) except for the group that was aged and treated with Silane Bond Enhancer and OBSP (10.6 MPa). Two experimental groups were numerically greater, although not statistically stronger, than the positive control. Both of these groups included repairs made using mechanical retention and OBSP repaired immediately (19.2 MPa) and repaired after aging (18.9 MPa). The immediate negative control achieved 60% of the inherent substrate shear bond strength; the negative aged control obtained 50%. Failure mechanisms in experimental groups demonstrated predominantly mixed or adhesive failures.

Table 7: Shear Bond Strengths (MPa ) ^a

View Fullscreen

Table 8: Failure Mode of Shear Bond Strength Repairs

View Fullscreen

DISCUSSION

Repair bond strength differences exist depending on the surface treatments. Additionally, differences exist between the original substrate and repairs; therefore, both null hypotheses are rejected. Repairs occur by three mechanisms: 1) micromechanical bonding due to surface irregularities, 2) chemical bonding between the two resin matrices, and 3) chemical bonding with the filler.¹² The purpose of using coarse Sof-lex discs with approximately 100 μm particle size is to simulate the immediate clinical finishing of the restoration. For aged specimens the Sof-lex disc served to remove the contaminated superficial layer of resin that would have been exposed to saliva. Additionally, their use creates micromechanical irregularities, resulting in increased surface area and free energy. North American dental schools that teach composite repairs recommend roughening the composite surface with a diamond bur.³ However, the degree of coarseness of the diamond bur was not identified. Depending on the manufacturer, a course diamond bur may have particle size ranging from 25 μm to 100 μm. We chose to use coarse Sof-lex discs because their particle size was similar to the range of commonly used coarse diamond burs. Additionally, many diamonds burs have a 4° to 7° tapper, which would result in a slanted repair junction.

Filler particles in composite resin are coated with silane, which allows chemical bonding to the resin matrix. We anticipated that surface treatment with Silane Bond Enhancer followed by application of the dentin bonding agent would result in an additive effect compared with using only OBSP. This was the result in immediate repairs under shear testing (18.3 MPa compared with 13.6 MPa) and aged repairs under flexural testing (68.6 MPa compared with 29.0 MPa). When tested under flexural forces, Silane Bond Enhancer used with OBSP (immediate and aged specimens) provided statistically similar bond strengths to each other and to the intact substrate (Table 3). Silane Bond Enhancer had very little effect in diametral testing. The silane coupling agent on the filler particles must be replaced for adequate repairs to occur because surface preparation with Sof-lex discs removes the filler's silane layer.¹³ Improved repair strength tested under shear forces has been reported when the resin is first silicoated followed by silanization and application of a bonding agent.¹⁴ Our study produced comparable shear bond strengths without prior silicoating.

The manufacturer claims that Add & Bond produces repair strengths as strong as the composite's inherent strength but this study never produced values as numerically strong. However, the values were statistically similar for immediate repairs tested under flexural forces (102.9 MPa compared with 80.6 MPa) and in shear testing immediate repairs (13.3 MPa compared with18.8 MPa) and aged repairs (16.6 MPa compared with 18.8 MPa). In diametral tensile testing, Add & Bond was significantly weaker than the intact substrate for both the immediate repairs (18. 0MPa) and the aged repairs (23.6 MPa) compared with the intact substrate (59.0 MPa).

We anticipated an additive effect combining mechanical retention with OBSP but obtained mixed results. It was beneficial for aged repairs under flexural forces (44.9 MPa) compared with only OBSP (29.0 MPa) but not for immediate repairs, which had 40% reduction in repair strength. Mechanical retention combined with OBSP produced decreased diametral tensile repair strength of immediate repairs compared with only OBSP (29.1 MPa compared with 47.2 MPa). Mechanical retention with OBSP was most beneficial under shear forces, obtaining the highest shear bond strengths for immediate repairs (19.2 MPa) and aged repairs (18.9 MPa). The mixed results may be related to the size of the repair site area. Flexural specimens had a 4 mm² surface area, shear specimens 4.44 mm², and diametral specimens 18 mm². The diametral tensile specimens may benefit from additional retentive points instead of a single point. A quarter round bur is 0.5 mm in diameter and results in a retentive feature with a 90° shoulder and a rounded bottom. This is recommended by Vivas and others,¹⁵ who demonstrated improved flexural and shear bond strength compared with acid etching alone. Shen and others¹⁶ suggested that a quarter round bur may cause an area of stress concentration and should be eliminated by rounding the shoulder or widening the entrance. As an alternative, Yesilyurt and colleagues¹⁷ recommended the use of micromechanical retention produced by air abrasion or a diamond bur.

The use of bonding agents has been shown to significantly enhance the shear bond strength of composite repairs that were aged from one to 12 weeks before repair.¹⁸ The bonding agent is able to enter the microscopic mechanical features of the resin. Etching with phosphoric acid was used in our protocol because repairing a composite resin usually involves not only removing a portion of the restoration but also adjacent enamel and dentin. Surface imaging performed by Fawzy and others¹⁹ demonstrated no significant change in the composite's surface morphology after acid etching. The acid's action on resin may be limited to superficial cleaning.²⁰ Several studies have not shown an increase in bond strength when composite surfaces were etched with phosphoric acid.^6,7,10

In a study in which composite specimens 4 mm thick were aged, Rodrigues and others²¹ calculated 50% water saturation occurred within nine days. Additionally, Biradar and others²² determined that maximum water absorption occurred in the first week. Therefore, we chose seven days of water storage for our aging protocol. Water diffuses through the polymer chains and interface boundaries between the resin and filler particles. Hydrolytic deterioration of the resin results in elution of components and plasticization of the resin. Properties such as hardness, wear resistance, strength, and fracture toughness are affected by water absorption. Additionally, water absorption reduces the available unreacted methacrylate carbon double bonds necessary to chemically react with the repairing composite. The amount of water absorption is affected by the resin and the filler content, coupling agent, and curing time and distance.²³ Increasing the amount of filler content and improving its bond to the resin matrix decreases water absorption. As a result, hybrid composites have been shown to provide better repair strength than microfilled resins.²⁴ Although composites are hydrophobic, water absorption may increase their hydrophilicity allowing better penetration of the dentin bonding agent during the repair process. Resins containing bisphenol glycidyl methacrylate have lower conversion rates compared with other matrices.^25. Therefore, bis-GMA containing resins have more unreacted carbon double bonds to chemically react with the repairing composite resulting in higher repair strengths than resins containing different matrices.⁸ Ultimately, repair strengths are dependent on the microstructure and composition of the parent and repair materials as well as penetration of the adhesives into the retentive features of the aged resin.

The application of a bonding agent for immediate repairs may not be necessary as Rinastiti and others²⁶ have demonstrated. The survival rate of free radicals in photo-cured resin is 14 days.^27,28 There probably exists sufficient unreacted carbon double bonds to react with the repair composite. Because the negative control groups were able to form a repair suggests that unreacted vinyl groups (C=C) in the original substrate remain available, but to a lesser degree. Negative control groups ranged from a low of 17% of the inherent substrate strength (aged flexural repairs without surface treatment) to a high of 60% (immediate shear bond strengths).

Shear testing had the largest percentage of mixed and cohesive failures. (Table 8). Della Bona and van Noort²⁹ suggested that shear testing measures the cohesive strength of the underlying composite rather the adhesive strength of the repair. The underlying composite will fracture only if the repair is stronger. However, it may also be an indication that the underlying composite was weakened because of water absorption and damage from using a Sof-lex disc. Flexural repairs had mainly adhesive and mixed failures at the repair site (Table 4). In shear testing there were 10 cohesive failures in the aged specimens using mechanical retention and OBSP. This may be the result of weakening of the parent substrate by the presence of mechanical retention in an already weakened matrix and stress concentration induced by the preparation design.

Repaired composites are subjected to complex intraoral forces during mastication and parafunctional habits. Of the three tests used in this study to replicate intraoral forces; diametral tensile is the most difficult to interpret. Failure must occur in the center of the specimen due to tensile forces if the diametral test is to yield useful results.³⁰ However, there is debate as to whether failure is actually caused by tensile forces acting at the center of the specimen.^30,31 When the specimen is loaded at a point, the fracture is due to shear and compressive forces at the loading point.³¹ If a specimen is loaded over a small flattened area (an arc less than 0.2× the diameter) or beneath soft pads, there are significant effects on the stress distribution patterns and diametral tensile values.³¹ All diametral tensile specimens in our study always failed along the diametral plane, and many exhibited cleft fractures, which we classified as mixed failures (see Figure 1). We believe that because point loading at the repair site continued, the specimens may have deformed resulting in a change from point loading to distributed loading. When this occurred stresses deviate from the recommended ideal tensile stresses at the center to more complex combined stresses.³¹ This may be the reason we observed mixed failures in specimens with lower diametral tensile strengths. Although our study did not exactly follow the testing methodology outlined in American National Standards Institute/American Dental Association Specification 27 for light-cured resins, the results we obtained were not drastically different from those obtained in studies that did.^32,33

View Figure

• Download as Powerpoint
• Download Figure

The minimally required composite repair strength that is necessary to survive clinically is not known. However, if the repair strength approaches the inherent strength of the nonrepaired composite or if a repair fails cohesively, this is an indication that the approach selected for repair may be clinically successful. In this study, several experimental groups were statistically equal to the flexural, diametral tensile, and shear strength of the inherent composite. In vitro studies have generally shown decreased repair strength of the composite repair. However, other studies have shown benefits of repairs on the longevity of the restoration.³⁴ Although the repair strength may be much lower than the inherent physical property being tested, there may not be a direct correlation to the clinical performance and longevity of the restoration.^35-38 Some clinical situations may involve immediate repair of a composite restoration. If a clinician determined that a suboptimal restoration was placed (open margin, inadequate proximal contact) and enough time remained in the appointment, then the clinician may decide to perform an immediate repair. If sufficient time is not available, it may be seven days before the patient could be reappointed to accomplish the repair. However, if the patient is new to the practice and the composite's composition is not known, there may not be a universally acceptable repair technique as composites of different compositions react differently when repaired.³⁹

CONCLUSIONS

1. Repairing composite is unpredictable, and surface treatments produce different effects depending on the mechanical property being tested.

2. Composite repair strengths are generally lower than the inherent strength of the composite.

3. The combination of three mechanical tests was not able to determine an optimal protocol for composite repair.