Introduction

There is a plethora of ceramic materials available for clinicians to choose from, but limited scientific evidence to show which material is best in each specific case. There has been a steady increase in the number of clinical trials with ceramic materials, but there is still few randomized clinical trials, which are regarded as the highest level of evidence. Large differences in study population, methods used, success and survival criteria and follow-up time among the different prospective and retrospective clinical trials make comparison among the materials and treatment options difficult. This paper addresses the available recent clinical evidence for different treatment possibilities with ceramic materials. Furthermore, an analysis of the failures occurring and the reason for these are discussed.

When assessing clinical success, there are several criteria that need to be addressed. Most clinical trials report survival rates, but with no further disclosure of what is regarded as «survival» (1). Survival can thus mean everything from «no problems at all» to the restoration is «still in place» (Figure 1). According to Tan et al. (2), a survived restoration is still in use at the follow-up time, with or without reversible complications.

Success, however, is used as term for «a restoration with no complications during the observation period,» (2). Patient satisfaction is surprisingly seldom included in the evaluation. There are reasons to believe that every-day practices have lower success rates than what is generally reported in clinical trials where specialists with ample time perform the prosthetic treatment.

Survival of tooth-supported single unit crowns

Several systematic reviews have evaluated the success and survival of ceramic single unit restorations (3-5), but very few studies have more than 5 year's follow-up. When considering the success and survival of different materials and restoration types one must take into consideration that they are often made based on very different indications and conditions. Some reviews include meta-analyses where all the individual data from each of the included studies were similar enough to be put into one large pool for increased power of the statistical analyses. The meta-analysis has showed the survival rates of single unit ceramic crowns to be similar to metal ceramic ones with a 5 year survival rate of 94 % (5). Metal-ceramics is still considered as a «gold standard» with regard to clinical evidence. The posterior ceramic crowns seem more likely to fail than anterior crowns (5).

Veneers and adhesively cemented porcelain crowns

The use of shell-like adhesively cemented porcelain crowns and buccal veneers is well documented in multiple studies (6,7). The survival rates are high; around 90 % over 5-10 years and complications are few. Fractures are the most common complication (4 %), followed by debonding (2 %) and secondary caries (1 %). More complications occur when the preparation border extends beyond enamel or when cementation is performed without sufficient control of humidity. The restorations evaluated in these studies are mostly in the anterior region and on teeth with no or moderate previous damages which would significantly improve ease of maintenance and the potential for success and survival. Porcelains are used in veneered bi-layer ceramics and in metal-ceramics, as well as in aesthetic veneers and shell-like crowns, where glass-ceramics and zirconia cannot satisfy the need for optimal aesthetics.

Inlays-onlays

The survival and success of inlays and onlays have been studied in multiple trials (8). The overall 10-years survival rates are around 90 %. There are, however, large discrepancies in inclusion and exclusion criteria in the different studies, ranging from small MOD-inlays to large onlays which are closer to 3/4 crowns and so called «table-top» onlays for severely worn dentitions. Early trials of chair-side produced inlays show relatively high complication rates with marginal chipping, discoloration and secondary caries as common complications (9). The intra-oral scanning and CAD/CAM production has improved immensely since the introduction in the 1990's, however, and the success rates are naturally improving as a consequence. Fractures are the most common complication (4 %), followed by endodontic complications (3 %) and secondary caries (1 %) and debonding (1 %) (8).

Glass-ceramic crowns

Lithium- disilicate (LiO2) reinforced glass-ceramic crowns are among the best-documented treatment choices for restoring teeth in terms of number of published clinical trials (10). Based on a meta-analysis of a large number of restorations, the estimated 5-year survival rate of glass ceramic (leucite or lithium-disilicate reinforced) restorations was 96.6 %, whereas glass infiltrated alumina crowns had an estimated 5-year survival rate of 94.6 %. The long-term success is, however, less certain as few studies have more than 5 years follow-up. Again, one must consider that the information of the previous condition of the restored teeth in the different studies is limited, and it is thus difficult to compare the survival rates for each separate case.

Zirconia and Alumina (polycrystalline ceramics)

Alumina crowns were for many years the most used ceramic restorations in the Nordic countries. Studies of alumina crowns have shown a 5-year survival rate of approximately 93 % (3-5). Based on personal reports and feedback from dental technicians, the complication rates were significantly higher than reported in scientific publications. Both crown loosening and fractures have been experienced frequently. Alumina has been more or less totally replaced by zirconia over the last 10 years as the material of choice for ceramic restorations for teeth with moderate to substantial loss of substance and need for strong material. Relatively few clinical prospective studies on single unit zirconia crowns have been performed, but some recent reports indicate a 96 % 5-year survival rate, but with a significant drop in survival in the following two years to 75 %-93 % (11-14). Two retrospective studies based on dental laboratory data from up to 5 years of production show fracture rates of 3.35 % for bi-layered crowns and 2.0 per cent for monolithic restorations (15-16).

Few studies have exceeded mean five-year follow-up time in ceramic restorations. It should be kept in mind that the improvement of ceramic materials has developed quickly in recent years and the results of the long follow-up studies may include old materials. So far, there is no clinical documentation above one year of the clinical performance of translucent or high-translucent zirconia (cubic/anterior) materials that are recently introduced.

Survival of tooth-supported fixed dental prostheses (FDPs)

Two recent systematic reviews have addressed the survival rates of tooth-supported FDPs (17-18). Both found that the difference between ceramic and metal-ceramic restorations was not statistically significant after 5 years follow-up. The estimated survival rate of reinforced glass ceramics was 89.1 %, whereas glass-infiltrated alumina had 86.2 % and zirconia approximately 91 %. The survival rate of metal-ceramic FDPs was 94.4 % after five years. Regarding zirconia, long-term (7-10 years) follow-up studies have been published for multi-unit FDPs and the survival rate ranges from 75-100 % (12, 19-21). Veneer chipping was the predominant complication in the earlier studies.

Resin bonded ceramic fixed dental prostheses (RBFDPs)

Ceramic resin retained FDPs have previously been rather unsuccessful. Resent development in bonding to zirconia shows, however, promising results. One study of 108 zirconia restorations restoring missing incisors showed a survival rate of 98 % after ten years (22). Of these 108 restorations, six had debonded, but all could be rebonded successfully. It is important to note that only one wing was used for retainer in both these trials. Furthermore, a special attention was made to follow a strict protocol in order to achieve sufficient bonding. Trials with ceramic RBFPDs with two wings have significantly lower success rates, since both fractures and loss of retention is experienced more frequently for two-winged restorations (23). There are however very few publications and further results should be awaited.

Survival of implant-supported single crowns

A survival of 97.1 %-100 % has been reported for implant-supported ceramic crowns after three to five years (14,24-25). Both monolithic lithium disilicate implant-supported single crowns, and monolithic zirconia crowns have been studied. Even though the survival was high in all studies, the success rates was significantly lower (87.5 %-91.7 %), indicating that there was or had been complications with the restorations within the follow-up.

Survival of implant-supported FDPs

Five-year survival rate of implant-supported zirconia-based FDPs has been found to be 100 % in two studies with zirconia-based restorations, bi-layered (18) and monolithic (24). The number of patients were, however, low and the results must be regarded with caution.

Biological complications

Technical complications

As discussed above, most published clinical trials have relatively low rates of technical failures. Fractures remain the most common technical complication, and are responsible for between 0.5 to nearly 6 % annual complication rates in different studies, depending on fracture mode and the type of the restoration (single or FDPs, supported on teeth or implants) (3, 17). There are two main types of fracture modes, fractures in the veneering material (chipping) and core fracture (total fractures or catastrophic fractures). These can be further divided into four fracture modes based on localization and severity (Figure 2). It is, however, uncertain whether the success rates are equally good in normal everyday practices. Based on communication with practicing dentist and dental technicians, fractures remain a clinical problem. A large British registry of revision rates and survival of dental restorations from regular dental practices showed that ceramic crowns have a shorter time in clinical function than metal-ceramic ones (36). The exact reason is not evident based on the registry, but fractures are considered the major contributing factor. Crown loosening is the other major technical complication with annual failure rates between 0.11 % and 1 % in different studies (3, 17). The lack of periodontal receptors surrounding the implants can probably explain the high technical complication rates (most commonly chip-off ceramic fractures) in implant-based restorations due poorer chewing precision and thus to higher mastication loads compared to natural teeth. Furthermore, implant-based restorations often have long coronal height compared to tooth-based restorations creating unfavorable force vectors.

Most papers involving systematic fractographic analyses of the fractures of retrieved dental crowns that have failed during clinical practices are case reports or case series. Retrieval of crowns failed by chipping is difficult without destroying the remaining restoration. Furthermore, chipped veneer is often adjusted in situ instead of being replaced as long as function or aesthetics is not severely reduced. Core failures are easier to collect and a relatively large number of cases has been identified by several authors (37-45)

The prevalent assumption has been that ceramic crowns fail due to contact damage at the point of occlusion. However, analyses of retrieved crowns reveal that most clinical crowns produced of modern ceramics with core failures originate from cracks starting in the cervical margin, and usually in the approximal area or in the palatinal region (Figure 3). Some crowns fracture from inner radial cracks, sometimes referred to as cone cracks, in the area opposing the occlusal load (Figure 4). Chipping failures, on the other hand (Figure 5), seem to originate from poor veneer support, traumatic occlusion, improper occlusal adjustments, defects or pores in the veneering material or accidental biting on hard objects (46-48). Chipping failures are more commonly observed on implant-based than in tooth-supported restorations. Additionally, fractures of zirconia abutments have been observed, but the occurrence rates are uncertain (Figure 6). The number of restorations analyzed by standardized fractographic analyses is, however, not yet large enough to draw any definitive conclusions concerning the cause-effect relationship of clinical fractures overall. The restorations analyzed vary greatly in size, shape and materials used. The fracture may be caused by many factors, such as; wear, grinding, material flaws, residual stress, thin margins, machining cracks or a combination of two or more factors (49). Multi-unit restorations, on the other hand, mostly break in the connector area. The few publications addressing this, reveal that the connector areas in fractures restorations are very often poorly designed, with sharp angles or asymmetric shape, and smaller than recommended. Furthermore, grinding marks causing cracks are often evident (Figure 7). Based on the existing publications, it can be concluded that margin initiated fracture is a fairly common failure mode for core fracture of crowns. FDPs fracture due to poorly dimensioned connectors. Traumatic occlusion or poor occlusal adjustment are common causes of chipping.

The cause for crown loosening is, unfortunately, seldom included in scientific publications. There can be many reasons for crown loosening: poor retention form of the abutment, too thick cement space, improper use of cement, contamination of surfaces or poor adhesion between cement and the surfaces in restoration or tooth. All of these factors can be at play for ceramic crowns. The need for rounded edges and finish lines reduces the area available for mechanical retention. Furthermore, there is still debate regarding the efficiency of adhesive bonding to alumina and zirconia, although Kern et al. have shown very promising clinical results with their method for bonding to zirconia (50).

Conclusion

Despite the fact that the ceramic materials available today are stronger and more versatile than ever, failures do occur in clinical use. Although biological complications are rare compared to metal-ceramic restorations, technical complications occur at similar or on higher level than for metal-ceramic restorations. Choosing the right material for each case and proper handling of the materials is crucial and can reduce the number of complications.