Introduction

Osseointegration has revolutionized dental treatments, making implant-supported prostheses a popular choice for replacing missing teeth. These treatments boast impressive long-term implant survival rates, often exceeding 95% over a decade.^1–3

Traditionally, metal-based handcrafted restorations with feldspathic ceramics were the standard. However, advancements in CAD-CAM technology, combined with the increased demand for esthetic solutions, have expanded the possibilities for all-ceramic reconstructions, making them more affordable and efficient to produce.⁴

Yttria-stabilized tetragonal zirconia polycrystal (Y-TZP) ceramics are increasingly popular in implant dentistry due to their superior mechanical properties, low water solubility, reduced bacterial adherence, superior corrosion resistance, and biocompatibility.^5,6 Lithium disilicate glass-ceramic is another ceramic material that possesses excellent mechanical properties and translucency; therefore, it is increasingly used for making screw-retained implant crowns with various CAD-CAM systems. ⁷

The industrialization of dental prostheses with CAD-CAM technology has encouraged implant manufacturers to create components compatible with the fully digital process, like titanium base abutment (Ti-base).⁸ Ti-base abutment, which is a prefabricated abutment with an incorporated digital library, stands out for their uniqueness and distinction from customizable abutments due to their association with a digital library. Ti-bases, which are available in the digital library as open STL files, offer a range of geometries, heights, contours, and engaging and non-engaging connections, according to the restorative needs.⁹ After selecting a Ti-base and designing the full prosthesis or intermediate coping, the eventual STL file is delivered for milling. This milled restoration or coping fits over the Ti-base with minimal adjustment needed, ensuring a close fit before cementation.⁹ The literature presents various names for this prefabricated abutment, such as Ti-base abutment, titanium-bonding base, titanium insert, hybrid abutment, cementing cap, and titanium cylinder.⁸ The author used the “Ti-base” term consistently in this review.

Ti-bases present with several merits. They assist in transitioning to a digital workflow, either fully or partially, thus leading to a reduction in production costs and to a significant improvement in time efficiency.^4,10 In addition, zirconia abutments combined with Ti-bases address some of the issues related to one-piece zirconia abutments, which have shown a higher fracture rate compared to metal abutments, as well as increased wear at the implant connection.¹¹ Furthermore, combining a Ti-base with either a ceramic meso-structure (known as 2-piece hybrid abutment) or a fully contoured restoration (known as a 1-piece hybrid-abutment-restoration) transforms a standard component into a customized one that would support the peri-implant soft tissue and enhance esthetics.¹² Moreover, with this screwmentable hybrid-abutment-restoration, bonding process can be carried out in a controlled laboratory setting, with no risk of leftover cement.¹³

While short-term clinical outcomes for Ti-base implant-supported ceramic restorations were encouraging,^9,14 there are concerns about the long-term bonding stability. In vitro studies simulating five years of clinical use showed some crowns detaching from the titanium-base abutments,^12,15 and others displayed marginal gaps and slight movements between the components.¹⁵ Provided the adhesive bond is strategically placed beneath the peri-implant mucosa to hide the titanium in various clinical scenarios, this positioning may pose biological risks due to the potential for increased bacterial buildup if the adhesive bond degrades. Additionally, implant restorations with marginal misfits can result in greater crestal bone loss compared to those that fit accurately.¹⁶ Therefore, the debonding issue is the origin of most complications and, subsequently, failures.

Since concerns have been raised about the weak bond strength between the ceramic restoration and Ti-base,^12,15 several laboratory studies and a recent systematic review^{7,13,15,17,18} have investigated the bond strength and the factors that impacted the bond strength between ceramic restorations and Ti-bases; however, to the best of author’s knowledge, there is no systematic review that explored the clinical stability or durability of the resin bond between implant-supported fixed CAD-CAM ceramic restorations and Ti-bases.

There is a scarcity of clinical research on the failure of ceramic prosthesis bonding to Ti-bases, since the Ti-base abutment concept is still a new treatment available in the market, and because the initial bonding failure usually presents as a micromovement that may not be detected except with microscopic evaluation after restoration removal, rending this difficult.¹⁸ In addition, clinical studies are costly and take a lot of time; however, they do answer on the real clinical behavior of different restorations being assessed.¹⁹ Therefore, this systematic review aimed to review all clinical studies reporting on the bond stability and durability between implant-supported fixed CAD-CAM ceramic restorations and Ti-bases from single crowns to full-arch prostheses; so as to deliver reliable clinical guidelines for a stable bonding protocol. The bond was considered stable if the restoration did not detach from Ti-base for the duration of the study.

Methods

Protocol

This systematic review was directed according to the guidelines delineated in the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA)²⁰ using the Population, Intervention, Comparison and Outcome (PICO) method.²¹ According to the PICO framework (population: implant-supported fixed CAD-CAM ceramic restorations, intervention: adhesion to Ti-bases, comparison: bonding protocol not applicable, and outcome: bond stability/ durability. As the current study is a systematic review, there was no need to gain approval from the ethics committee.

Focused Question

The focused question of the present review was: “How much is the stability and durability of the resin bond between implant-supported fixed CAD-CAM ceramic restorations and Ti-bases clinically?”

Eligibility Criteria and Exclusion Criteria

The inclusion and exclusion criteria were set based on PICO (Population, Intervention, Comparison, Outcome) guidelines as presented in Table 1.

Table 1 Inclusion and Exclusion Criteria for Clinical Studies

Information Sources and Search Strategy

The PubMed, Google Scholar, and Cochrane databases were investigated on September 21, 2024. Additional searching was executed in the references of included manuscripts, some related systematic reviews, and on the web sites of some journals: Clinical Implant Dentistry and Related Research; Clinical Oral Implant research; The International Journal of Oral & Maxillofacial Implants; The International Journal of Prosthodontics; Journal of Prosthetic Dentistry; Journal of Prosthodontics; Journal of Dentistry; Dental Materials; and Journal of Oral Implantology. Representative keywords are presented in Table 2.

Table 2 Systematic Review Search Strategy

Selection Process

The author executed the literature search by firstly screening through titles and abstracts; then, full-text articles were screened if title and abstract did not offer sufficient information. The author then read the full-texts of the included articles. Figure 1 displays the flowcharts of manuscripts identification and inclusion.

Data Collection Process

The author read the full-texts of the included articles and extracted all the relevant data of each one, using pre-determined fields in a uniform data extraction sheet.

Data Extraction

The gathered data for included studies comprised the authors’ name (year), study design, number of patients and restorations, restoration type, the evaluated ceramic materials, the used Ti-base system, the pre-treatment protocol and bonding system, follow-up time, restoration survival rate, number or rate of debonding incidence, and other prosthetic complications.

Risk of Bias Assessment

The risk of bias of the included Randomized controlled trials (RCTs) was assessed by the Cochrane risk of bias revised tool (RoB 2),²² the ROBINS-I tool was used to assess the non-randomized controlled clinical trials,²³ and the Joanna Briggs Institute’s (JBI) critical appraisal tool was used to evaluate the one-arm case series studies.²⁴

Synthesis Methods

Since heterogeneity was detected among the included studies, a meta-analysis was not conducted. Instead, summarization of the pertinent clinical studies was implemented.

Results

Study Selection

A total of 5,186 records were identified from electronic literature search through Sep. 21, 2024, whereas 4 records were identified after hand searching through references of included studies. After the duplicate removal, 3,184 records remained for screening based on titles and abstracts. Of these, 59 full-text articles were evaluated based on eligibility criteria. Ultimately, 40 studies^25–64 were included in this systematic review as shown in Figure 1 after exclusion of eighteen articles^65–82 for reasons outlined in Table 3 and one could not be retrieved.⁸³

Table 3 Studies Excluded from the Second Stage of Searching

Figure 1 Flowchart viewing the studies identification and inclusion method.

Study Characteristics

These comprised twenty retrospective investigations, twelve RCTs, and eight prospective. Of the 40 included studies, 29 studies were on zirconia framework bonded to Ti-bases, 7 studies were on lithium disilicate glass ceramic single copings bonded to Ti-bases, 3 studies were on both zirconia and lithium disilicate, and one was on non-specified CAD-CAM ceramic restorations. The ceramic material used for multi-unit fixed dental prostheses (FDPs) and full-arch prostheses was only zirconia, while zirconia and lithium disilicate glass ceramics were used for single copings.

Regarding the used bonding system, 17 studies did not specify the used bonding system, while the reported utilized resin cements were mainly dual-cure and only two studies utilized self-cure resin cement.

While 29 studies did not specify the type of bonding surfaces’ pre-treatment, the rest of the studies demonstrated that the pre-treatment ranged from no pre-treatment to sandblasting with 50μm alumina particles at 1- to 2.5-bar pressure for zirconia and Ti-bases, with or without MDP-containing primers application. Lithium disilicate bonding surfaces were usually etched with hydrofluoric acid and silanized prior to bonding.

In this systematic review, success of the bond was decided as the ceramic restoration remaining in place, luted to Ti-base, without any visible debonding in the study’s observation period. Survival was described as original restoration staying in situ at the time of follow-up with or without adjustment during the investigation period. If a restoration debonded but could be favorably rebonded, the survival rate was not affected. Prosthetic complication was designated as any unsatisfactory or unpredicted event happening during the investigation period but did not mandate remaking of restoration. These were documented but, for the aims of this review, were not considered in the success rates that focused on bond stability. Studies’ characteristics are presented in Table 4. An overview of the utilized resin cements is presented in Table 5.

Table 4 Study Characteristics of the Reviewed Studies

Table 5 Overview of the Utilized Resin Cements

Data Synthesis

Single Implant Restorations

Twenty-nine studies reported on prosthetic outcomes of single ceramic copings bonded to Ti-bases, either hybrid-abutment crowns (1-piece) or hybrid-abutments (2-piece). A total of 2,056 single ceramic copings were evaluated in this review, 711 were lithium disilicate copings and the rest were zirconia. Notably, 18 debonding events were reported (10 zirconia, 8 lithium disilicate) during follow-up ranging from 1 to 7.5 years. A recent RCT²⁵ reported on one occurrence of de-cementation of posterior zirconia crown from Ti-base out of 41 crowns in the first year of function, and it was successfully e-cemented during the 3-year observation period. Another recent RCT²⁶ reported 2 of 27 posterior zirconia crowns de-cemented from Ti-bases during the 4-year observation period. The authors did not report on the outcomes of these crowns. In a prospective study,²⁷ four events (5.3%) of debonding between posterior zirconia crowns and Ti-bases were reported during the 2-year observation period. These events were encountered only with crowns luted with dual-cure resin cement (RelyX Ultimate) and were successfully re-bonded with self-cure resin cement (Multilink Hybrid Abutment) for the duration of study.²⁷ In addition, a 3-year retrospective study²⁸ reported two of the 106 posterior zirconia abutments de-bonded from Ti-bases. They were successfully re-bonded for the duration of the study.²⁸ Another retrospective study²⁹ reported 1 out of 82 zirconia copings de-bonded from Ti-base during the 6-year observation period, and the coping was successfully re-bonded for the duration of the study. In a further recent retrospective study,³⁰ eight events (1.3%) of debonding between posterior milled lithium disilicate crowns and Ti-bases were recorded through a mean observation period of 6 years. The authors did not report on the outcomes of these crowns.³⁰

Multi-Unit FDPs

A total of 107 multi-unit zirconia FDPs besides approximately 52 multi-unit FDPs (supported by 105 implants³¹), were evaluated in this review. Nine events of debonding of zirconia restorations from Ti-bases were reported during a follow-up ranging from 1 to 6 years. Saponaro et al²⁹ in their retrospective study reported 2 debonds between zirconia abutments and Ti-bases supporting FDPs, where one patient was a bruxer with a three-unit cantilevered prosthesis, and the other had a four-unit anterior prosthesis. The zirconia abutments were successfully rebonded to Ti-bases for the duration of the study, but the patient with the four-unit anterior prosthesis subsequently experienced major porcelain fracture in his veneered zirconia prosthesis.²⁹ In addition, a 3-year RCT²⁵ reported seven events (16.3%) of Ti-bases’ debonding from two-implant supported zirconia FDPs in the first year of function. All Ti-bases except one were successfully re-bonded intraorally to zirconia FDPs for the duration of the study. The restoration that could not be re-bonded intraorally was due to mucosal overgrowth and it was sent to technician to re-bond it; however, it debonded again and considered as a failure.²⁵

Full-Arch Prostheses

Twelve clinical studies reported on prosthetic outcomes of 2861 zirconia full-arch screw-retained prostheses. While eight publications did not report on any incidence of debonding, four retrospective studies revealed 17 events of debonding from 2447 assessed prostheses.^32–35 All de-bonded titanium bases/cylinders were successfully re-bonded to zirconia prostheses either intraorally or in the laboratory for the duration of the studies.

Restorative Material

Regarding the effect of ceramic restorative material on the bond stability, Strasding et al³⁶ in their RCT showed that lithium disilicate and zirconia crowns bonded to Ti-bases performed comparably without any debonding event at 1-year follow-up.³⁶ De Angelis et al³⁷ in their cross-sectional retrospective study also revealed analogous clinical outcomes for both zirconia and lithium disilicate crowns bonded to Ti-bases without any debonding event after a 3-year follow-up. On the other hand, Graf et al³⁰ evaluated retrospectively the clinical performance of monolithic lithium disilicate crowns and veneered zirconia crowns luted to Ti-bases through a mean follow-up of 6 years. The study showed no significant differences in survival rate, significantly lower prosthetic complications for lithium disilicate crowns; however, more debonding events for lithium disilicate crowns over the observation period (8 debonds of 398 lithium disilicate crowns, 0 debond of 203 zirconia crowns).³⁰

Risk of Bias in Studies

Twelve RCTs^{25,26,36,38–46} were evaluated using the RoB 2 tool. For four RCTs,^38,42,44,46 the overall risk of bias was considered low, whereas the other 8 RCTs^{25,26,36,39–41,43,45} presented some concerns due to risk of bias in some domains as presented in Table 6. Two non-randomized controlled studies^47,48 were evaluated using the ROBINS-I tool and showed a serious risk of bias as presented in Table 7. All 26 one-arm non-controlled clinical studies^{27–35,37,49–64} were evaluated using the Joanna Briggs Institute’s critical appraisal tool and were found to have a low^{27–31,35,37,52,54–56,59–62,64} to moderate risk of bias^{32–34,49–51,53,57,58,63} as presented in Table 8.

Table 6 Risk of Bias Domains for Randomized Controlled Trials (RCTs)

Table 7 Risk of Bias Domains for Non-Randomized Clinical Trials

Table 8 Risk of Bias Check List for One-Arm Non-Controlled Clinical Studies

Discussion

The rapid advancement of digital technologies and the introduction of new ceramic materials that offer superior mechanical and esthetic qualities, have largely attributed to the notable transition from metal-ceramic to all-ceramic CAD-CAM implant restorations bonded to Ti-bases.^4,9

This systematic review aimed to review all clinical studies reporting on the bond stability and durability between implant-supported fixed CAD-CAM ceramic restorations and Ti-bases.

In this review, 18 debonding events from Ti-bases of 2,056 (0.88%) assessed single ceramic copings,^25–30 9 debondings of approximately 159 assessed multi-unit zirconia FDPs,^25,29,31 and 17 debondings of 2861 (0.59%) assessed full-arch zirconia prostheses,^32–35 were reported during observation periods ranging from 1 to 7.5. The reason why full-arch prostheses revealed the least reported debonding rate might be due to the delay in the visible clinical debonding of the prosthesis since it might remain stabilized on some bases even though some are detached. Another cause might be due to better distribution of forces on cross-arch splinted implants and copings compared with single crown.

Although low debonding rates were recorded for different types of restorations, this result should be interpreted with caution since these reported ultimate debonding events were usually preceded by micromovements that could not be detected clinically. Some debonds also might not be visible for multi-unit prostheses except after most of the Ti-bases debonded. These undetected micromovements or debonds might lead to several technical and biological complications. Therefore, the number of these debonds should be taken seriously with addressing the probable underlying causes for each one.

With regard to the effect of Ti-base height and geometry on bond stability, Graf et al³⁰ who reported eight events (1.3%) of debonding between lithium disilicate crowns and Ti-bases through the 6-year observation period, blamed the short height of the used Ti-bases as a possible cause of this complication. Furthermore, Gonzalez and Triplett,³² who reported two events of debonding of the 44, investigated full-arch zirconia prostheses through a mean follow-up of 33 months, attributed the non-use of manufacturer’s recommended Ti-bases and the inadequate height of the excessively reduced bases as possible causes. Another 3-year RCT²⁵ that reported seven events (16.3%) of Ti-bases’ debonding from zirconia bridges in the first year of function, attributed the use of non-retentive, flat-cone Ti-bases as one of the possible causes. According to a recent systematic review of in vitro studies,¹⁸ the height of the Ti-base was cited as a significant factor for ceramic/Ti-base bonding stability. Although the critical height for Ti-bases was not specified in this review, a recent systematic review¹⁸ based on in vitro studies demonstrated that the critical height is approximately 3.5 mm, this can vary based on the Ti-base’s design and the inclusion of micro or macro-retentive elements. It seems from the limited clinical evidence that Ti-base heights and geometries could affect the bond stability over time since short and non-retentive, flat-cone Ti-bases lead to more debonding events. However, further clinical investigations are needed.

Considering the impact of ceramic restorative material on ceramic/Ti-base bond stability, Strasding et al³⁶ in their RCT and De Angelis et al³⁷ in their retrospective study showed that lithium disilicate and zirconia crowns bonded to Ti-bases performed comparably without any debonding event at 1-year and 3-year follow-up, respectively. On the other hand, Graf et al³⁰ in their retrospective study reported eight events (1.3%) of debonding for lithium disilicate crowns (8 of 398), while no debonding event was recorded for zirconia crowns (0 of 203) through 6-year mean observation period. Nonetheless, the authors attributed this complication to the short height of Ti-bases used in the study rather than to the type of restorative material.³⁰ However, an in vitro study showed that zirconia crowns had inferior bonding capability to Ti-bases compared with lithium disilicate after thermo-mechanical aging.¹⁵ This might be due to almost chemical inertness of the highly crystalline structure and to the lack of a glass phase of zirconia.⁸⁴ Due to controversy in the results, additional clinical investigations about the association between restorative material and the ceramic/Ti-base bonding stability might be conducted.

Regarding the micromechanical and/or chemical surface pre-treatments of ceramic and titanium, the reported surfaces’ pre-treatment of zirconia and Ti-bases in this review ranged from no pre-treatment to air abrasion with 50μm alumina particles at 1- to 2.5-bar pressure with or without a priori primer application. Lithium disilicate bonding surfaces were usually etched with hydrofluoric acid and silanized prior to bonding. A recent systematic review based on in vitro studies¹⁸ suggests that air-abrasion of the bonding surfaces of zirconia and Ti-bases using alumina particles of 45–50μm is the most effective method to enhance the retentive strength. Additionally, in vitro studies^13,17 demonstrated that the use of bonding systems comprising MDP-based primers and the application of universal primers before cementation appeared to be advantageous for sandblasted Ti-bases. Although most of the identified clinical studies showed favorable bond stability between ceramic restorations and Ti-bases with the reported surface pretreatment, it remains difficult to draw a conclusion on the most effective surface pre-treatment method. Given the limited clinical evidence on this topic, further research is needed.

If the utilized bonding system could affect the bond stability, Koenig et al 2019²⁷ reported that the four debonds (5.3%) that occurred between zirconia crowns and Ti-bases during the 2-year observation period, were encountered only with crowns luted with dual-cure resin cement (RelyX Ultimate). However, they were successfully re-bonded with self-cure resin cement (Multilink Hybrid Abutment) for the duration of the study.²⁷ In addition, Saponaro et al,²⁹ who reported 2 occurrences of debonding between zirconia abutments and Ti-bases for multi-unit FDPs and one debond for single restoration, advocated the use of self-cure resin cement rather than the initially used dual-cure resin cement (Panavia F2). Furthermore, Gonzalez & Triplett,³² who reported two events of debonding of the 44 investigated full-arch zirconia prostheses through the 33-month observation period, considered that one of the factors that might contribute to debonding was the use of dual-cure resin cement that was inadequately polymerized under the zirconia prostheses. It appears from the limited clinical evidence that self-cure resin cement offers a more stable bond between zirconia restorations and Ti-bases over time; however, no recommendation can be given for the best luting agent, and further clinical research is needed.

In addition, biomechanical prosthesis- and patient-related factors might contribute to prosthetic complications like debonding of restorations from Ti-bases. Saponaro et al²⁹ reported 2 occurrences of debonding between zirconia abutments and Ti-bases in multi-unit FDPs and another debonding was for single crown. One patient was a bruxer with a three-unit cantilevered prosthesis, another had a four-unit anterior prosthesis, and the other was a bruxer. The authors attributed these complications partly to biomechanical patient- and prosthesis-related factors.²⁹ Inadequacy of passive fit is another factor that could affect the bond stability, as reported in a 3-year RCT,²⁵ where seven events (16.3%) of Ti-bases’ debonding from zirconia FDPs were recorded in the first year of function. The authors attributed these debonds to the probable inadequacy of passive fit of these bridges and to the use of non-retentive, flat-cone Ti-bases.²⁵ Impassive fit is usually manifested as abutment screw loosening; however, the non-retentive design of Ti-bases makes them the weakest link in the system and thus de-cemented earlier.²⁵ It appears from the limited evidence that inadequacy of passive fit and biomechanical patient- and prosthesis-related factors could affect bond stability. However, further clinical investigations are required.

Although the current systematic review is a valuable addition to the current literature since it is the first review based on clinical studies that shed the light on this important topic, this review presents with some limitations. One of the limitations is that most of the included studies reported 1–5-year follow-up data; as such, the clinical data are still regarded short-term. In addition, the risk of bias in the included studies ranged from low to moderate, with two presented with serious risk of bias. A further limitation is the issue that a meta-analysis could not be implemented due to the relative heterogeneity of the data. Moreover, due to scarcity of clinical research reporting on bond stability, this review included studies even though the utilized bonding protocol was not specified. More well-designed clinical research assessing the impact of different “genuine” Ti-bases geometries, heights, cementation protocols, surface pre-treatments, restorative materials, and the interaction between these factors on bond stability, with long-term observation periods beyond 5 years, is highly advocated.

Conclusions

From the limited evidence available, this systematic review concluded that CAD-CAM implant-supported all-ceramic restorations bonded to Ti-bases demonstrated relatively high bond stability during observation periods ranging from 1 to 7.5 years; whereas full-arch zirconia prostheses revealed the least reported debonding rate (0.88%) compared with single ceramic copings and multi-unit zirconia FDPs, respectively. However, these results should be interpreted with caution since they represent the ultimate visible debonding events in the short-term evaluation. Although most of the identified studies showed favorable bond stability between ceramic restorations and Ti-bases with the reported surface pretreatment, it remains difficult to draw a conclusion on the most effective surface pre-treatment due to limited clinical evidence. It appears that ceramic restorative materials may affect bond stability; however, it remains difficult to draw conclusions on the behavior of different ceramic restorative materials due to controversy in the results. In addition, it seems that Ti-base heights and geometries could affect the ceramic restoration/Ti-base bond stability over time since short and non-retentive, flat-cone Ti-bases lead to more debonding events. It also appears that self-cure resin cement offers a more stable bond between zirconia restorations and Ti-bases over time; however, no recommendation can be given for the best luting agent, and further clinical studies are required. Lastly, inadequacy of passive fit and biomechanical patient- and prosthesis-related factors could affect the bond stability between implant ceramic restorations and Ti-bases.