To satisfy patient demand for metal-free fixed partial dentures that provide long-term stability, it is necessary to keep several issues in mind. It is well known that in addition to marginal adaption, high fracture strengths, and sufficient aesthetics, one of the most important aspects for long-term survival of luted fixed partial dentures (i.e., crowns) is the fixation of the restoration achieved as a quality factor for long-term success(14). Thus, in several studies the bond strength was used as a prognostic factor to evaluate clinical suitability, i.e. for veneer restorations on natural teeth (15). However, the question arises how it behaves with ceramic restorations on titanium abutments. Therefore the experimental designs of those studies, just like in this study, are (16, 17) to simulate the clinical situation, i.e. of crowns made of monolithic all-ceramic materials luted on implant abutments, and thus allows a realistic assessment of the retentive strength as a factor for long-term success that can be achieved when frequently used ceramic materials and luting agents are utilized. Regarding this the retentive strength depends on several factors: among other things the texture of the retentive surfaces respective to their pretreatment (18–21), the luting agent or their curing mechanism (22, 23) or the geometry of the retentive surfaces (24–28). So special attention was paid to allow a sufficient dwell time of cemented crowns in 100% humidity before they were placed in water. For some luting agents, placement in water immediately after removing cement excesses can lead to rapid hydrolytic degradation of the bond (for example if the luting agent has not completely cured) and a reduction in the bonding potential of the luting agent (17, 29). This does not absolutely correspond with the clinical situation, in which the cement is in contact with saliva rather than water. However, in this study long-term water storage was not conducted because Ernst et al. (17) has shown that it had no further significant effect on retentive strength, whereas thermocycling was carried out to simulate the thermal stress as a relevant factor for reducing bond strength (20, 30, 31). In this study it was valued that representative and commonly used ceramic materials and luting agents were used. Different classes of ceramic materials, feldspathic, lithium disilicate, a hybrid ceramic and ceramic materials with oxide ceramic components were utilized and regarding the luting agents also different types, which are frequently encountered in everyday clinical practice, such as adhesive composite materials, non- and self-etching, self and light curing and a reinforced glass ionomer cement. Although manufacturers of ceramic materials and professional societies make specific recommendations for the combination of the ceramic materials and luting agents the question arises to what extent other combinations influence the retentive strength. So, the first working hypothesis, that each ceramic material used would behave independently of the type of luting agent regarding the retentive strength, was only accepted for the Mark II and Empress CAD ceramics. For the other ceramic materials used this hypothesis had to be rejected. Regarding the second hypothesis, that all luting agents would result in equal retentive values independent of the type of ceramic material have to be rejected completely.
To explain the results, it is required to analyze the fixation, on the one hand the interface between the ceramic surface and the luting agent and on the other hand between the luting agent and the adhesive surface of the implant abutment. With this regard the location of the residue of the luting agent allows to judge the quality of the bonding and helps to understand the behavior of releasing of the bond during the withdrawal process. Historically dental ceramics first were bonded on natural teeth using and based on this there are already numerous publications and findings on the bonding interface between ceramic material, luting agent and the surface of natural teeth (31). Compared to the experimental setup of those studies further studies were conducted there was a titanium surface instead of human dentin.
The question is if there is a context between the location of the adhesive residue and the used luting agents and ceramic materials.
In this study regarding the feldspar ceramic and glass ceramic no significant differences between the luting agents were detected, the residues of the luting agents were primarily in the adhesive surfaces of the ceramic restoration and thus the prevailing adhesive failure was between the titanium surface and the luting agents and that show the basically good adhesive connection between all luting materials to the feldspathic (Mark II) and glass ceramic (Empress) materials tested. For the lithium disilicate ceramic (LiDS) the luting agent Rely X Unicem shows a significant higher retentive strength compared to Panavia (p < 0.001) whereas the residues were also overwhelmingly in the inner crown surface. For the hybrid dental ceramic PolyFSP more significant differences between the luting agents were detected. So, for the luting agents Multilink implant, Rely X Unicem and Fujicem a higher retentive strength was detected compared to Panavia 2.0, whereas it is noticeable that beside more significant differences between the luting agents for this hybrid ceramic materials more and more residue of the luting agents were located on the titanium abutment surface and less in the inner crown surface. For the presintered zirconia-reinforced lithium silicate ceramic (PsZirLS) and the fully crystalized zirconia-reinforced lithium silicate ceramic (FcZirLS) higher retention values, higher standard deviations for nearly all combinations with the different luting agents were detected and the residue of the luting agents were also increasingly on the abutment surfaces.
With respect to the luting agents the self-polymerizing luting agent Multilink showed significant higher retentive strength for the ceramic materials Enamic and Celtra Duo compared to the ceramics Mark II and Empress. For the luting agent Variolink as a dual-/light- polymerizing, resin-based dental luting material Enamic, Suprinity and Celtra offered higher retentive strength compared to Mark II and Suprinity further to Empress. Regarding the dual-polymerizing, self-adhesive resin cement Unicem the ceramic Empress, Enamic, Suprinity and Celtra offered a significant higher retentive strength compared to Mark II, and Enamic, Suprinity and Celtra are also higher in retentive strength compared to the material Empress. Fujicem as a self-polymerizing, resin-reinforced glass-ionomer luting cement offer a similar situation with a higher strength for Enamic, Suprinity and Celtra in comparison to Mark II and Empress and Suprinity and Celtra higher than e.max and the material Suprinity higher than Enamic. In contrast the dual-polymerizing, self-adhesive resin cement Panavia showed only for Suprinity a higher retentive strength compared to Mark II.
The interpretation of these results is certainly difficult because despite all standardization of the experimental setup various influencing factors must be considered, essentially the different ceramic materials and luting agents. Because of the standardized setup the reason for this must be in the factor ceramic material, luting agent and in the combination of them, the interface of titanium surface, luting agent and adhesive surface of the ceramic material. It is conspicuous that almost no significant difference between the luting agents in retentive strength for the feldspathic (FSC), the glass (LrGC) and the lithium disilicate (LiDS) material were noted. In summary the ceramic materials Enamic, Suprinity and Celtra showed in combination with the different luting agents tested relatively often higher values for retentive strength compared to the feldspathic Mark II and the glass ceramic Empress. Increasingly more differences between the luting agents were detected for the hybrid (PolyFSP), the presintered (PsZirLS) and fully crystalized (FcZirLS) zirconia-reinforced ceramic. Regarding these ceramic materials tested further studies stated also differences in fracture strength (32)with a higher fracture strength of hybrid (PolyFSP), the presintered (PsZirLS) and fully crystalized (FcZirLS) zirconia-reinforced ceramic compared to the feldspathic (FSC), the glass (LrGC) materials. It might indicate that the retentive strength could be related to the stability of the ceramic material. A hypothesis is that loosening of the crowns is caused by a failure of the adhesive connection initiated by microcracks that arises first in ceramic materials with lower stability. This could be an explanation for higher retention forces when ceramics with high a higher stability are used. However, the question is further to explain the observed significant differences of retentive strength between crowns within each ceramic material made of the high stable materials (PolyFSP), (PsZirLS) and (FcZirLS). Those differences were probably caused by the different luting agents which are for the crowns made of these materials mainly responsible for the formation of different adhesive connection patterns between the titanium surface, the layer of the luting agent and the ceramic surface. However, it must be clarified to what extent the results depend on individual components of the several luting agents used.
In summary, the benefit of this study is that in the clinical frequently used numerous ceramic material and luting agent classes were tested in a standardized laboratory setup with corundum abrasion of the titanium abutments and completely identical testing procedure, and an adjacent consistent pull-off setup without the presence of influencing factors like saliva or blood, that could affect the quality of the fixation of the ceramic crowns. Further it is to be expected that the monolithic ceramic crowns made of the ceramic materials und luting agents used will exceed physical masticatory forces, however certain combinations of ceramic materials and luting agents offer higher retentive strength.
Thus, the first hypothesis that all ceramic materials behave independently of the type of luting agent applied must be rejected to the materials LiDS, PolyFSP, PsZirLS and FcZirLS and the second working hypothesis, that the luting agent used would result in equal retentive strength independent of the type of ceramic material was completely rejected.