A systematic review on the accuracy of zirconia crowns and fixed dental prostheses

Abstract Purpose The aim of this study was to review the fit and assess the accuracy of tooth-supported single and multi-unit zirconia fixed dental prostheses. Background The fit of zirconia restorations has been reported in several studies, but the accuracy of the manufacturing process is seldom discussed or used when drawing conclusions on the fit. Materials and methods A literature search of articles published in PubMed between 2 March 2013 and 1 February 2018 was performed using clearly defined inclusion and exclusion criteria. 841 articles were found and 767 were excluded after screening the title and abstract. After full-text analysis another 60 articles were excluded which left 14 articles to be included for data extraction. Fit was the mean of distances reported in the studies and accuracy was the fit minus the pre-set spacer Results For marginal gap of single crowns and multi-unit FDPs combined, the fit was 83 μm and the accuracy was 59 μm. The internal gap fit was 111 μm and the accuracy 61 μm. For the total gap, the fit was 101 μm, and the accuracy of the zirconia restorations was 53 μm. Conclusions Within the limitations of the present systematic review the fit of zirconia single crowns and multi-unit FDPs may be regarded as clinically acceptable, and the accuracy of the manufacturing of zirconia is ∼60 μm for marginal, internal, and total gap. Also, digital impressions seem to be associated with a smaller gap value.


Introduction
The fit of dental restorations is an important factor for the longevity of tooth-supported dental prostheses. A poor fit can affect the cement junction and result in dissolution, which may result in the loosening of the restoration or secondary caries [1]. Also, crowns with the poor marginal fit on subgingivally placed margins may increase bacterial retention and cause gingival inflammation [2].
There is no consensus on what is regarded as clinically acceptable fit, for marginal fit several authors suggest $100 lm [1,[3][4][5][6]. For internal fit, McLean and von Fraunhofer considered 120 lm clinically acceptable for dental restorations cemented with polycarboxylate cement [1,5]. Even though the internal discrepancies may be well over 200-300 lm most authors conclude that the results from their in vitro fit studies are clinically acceptable when the mean marginal gap is below or close to 120 lm [7][8][9][10][11]. The tooth-crown interface is divided into different areas; marginal, chamfer, axial, and occlusal. There are several areas or distances used to assess the marginal fit of the restoration, it can be measured as the marginal gap, the vertical marginal discrepancy, the horizontal marginal discrepancy, and the absolute marginal discrepancy [12]. The internal fit can be divided into discrepancy at the chamfer or cervical area, axial discrepancy and occlusal discrepancy, or as a mean of all the measuring areas/points [13,14].
The fit of a restoration can be measured using destructive techniques, where the crown or multi-unit fixed dental prosthesis (FDP) is cemented onto dies and embedded with, for example, epoxy resin, sectioned and analyzed microscopically [15,16]. Nondestructive techniques are also used, such as; clinical examination using an explorer, direct view of the crown margin using a microscope or scanning electron microscopy (SEM) [17,18], the silicone or impression replica method [5,19], micro-computed tomography (micro-CT) [20], and optical threedimensional (3D) scanning [21].
In order to evaluate the accuracy of a restoration, the settings for the marginal and internal spacer must be provided. Otherwise, the measurements only reflect the total deviation from the master model, however, since tooth-supported restorations seldom are manufactured with a 0 mm spacer setting, the results do not represent the accuracy. The results from the fit measurements should, therefore, be regarded as the fit, and the results minus the cement spacer setting, the accuracy.
Zirconia crowns and multi-unit FDPs are predominantly made using computer-aided design-computer aided manufacturing (CAD-CAM). A systematic review on the fit of CAD-CAM restorations of different materials found marginal gaps between 39-201 mm and internal gaps ranging from 23 to 230 mm [22].

Aim
The aim of this study was to review the fit and assess the accuracy of tooth-supported single and multi-unit zirconia fixed dental prostheses.

Exclusion criteria
Studies not meeting all inclusion criteria. Studies of implant-supported prostheses. Studies measuring fit after ceramic veneering. Studies measuring only the marginal gap. Studies where internal adjustments were made before a fit assessment.

Selection of studies
The titles were screened and abstracts from the studies found in the search described above, considering the inclusion criteria. After selection, the full texts of the studies were acquired. The full-text publications were screened according to the inclusion and exclusion criteria and 14 studies were included for data analysis ( Figure 1 and Table 1). The data collected from the studies were; Author, Year, in vivo/in vitro, Abutment teeth, Restoration type, Restoration material, Number of specimens per group, Preparation type, Cement spacer at margin, Cement spacer internal, Impression type, Scanner, CAD software, CAM machine, Fit assessment method, Number of measuring points per abutment, Die material, Restoration manufacturing method, Marginal gap, Cervical gap, Axial gap, Occlusal gap, Internal gap and Total gap. In studies where an internal or total gap was not reported but axial and occlusal values were, the internal and total gap values were calculated by the author. In this review, Holmes et al. definition of the marginal gap were used [12]. The internal gap was the mean of all the available internal measuring points (cervical, chamfer, axial, and/or occlusal) and the total gap was the mean of all the measuring points available in the studies (marginal, cervical, chamfer, axial, occlusal).
Two of the studies did not use an impression, the master model was scanned using a lab scanner and the restorations placed on the master model. The studies were included but the impression method used in those studies was grouped with the digital impression technique.

Statistical analysis
Descriptive data are presented as numbers and frequencies. Mean values were calculated as weighted values based on the individual group mean value and the number of test specimens per group. The fit was the mean of distances reported in the studies and accuracy was the fit minus the pre-set spacer.

Results
The fourteen studies included in the analysis for this review presented fifteen results for single crowns, seven for three-unit FDPs, seven for four-unit FDPs, and one for six-unit FDPs. Four different fit measuring techniques were used; the silicone replica technique (18 results), the cement and section technique (eight results), the 3D scan technique (Four results), and the Micro-CT technique (Two results) ( Table 1). The scanners, CAM machines and zirconia materials used are described in Table 2.  For the marginal gap of single crowns and multiunit FDPs combined, the fit was 83 lm and the accuracy was 59 lm. The internal gap fit was 111 lm and the accuracy 61 lm. For the total gap, the fit was 101 lm, and the accuracy of the zirconia restorations was 53 lm ( Table 3). Eleven of the results were for restorations made from conventional impressions and 19 results were from digital impressions (Table 4).

Discussion
The purpose of this systematic review was to evaluate the fit and assess the accuracy of tooth-supported single and multi-unit zirconia fixed dental prostheses. The fit of the zirconia FDPs was within the range (max 120 lm) most researchers deem clinically acceptable [1,5,[7][8][9][10][11]. In an earlier review on the fit of CAD-CAM restorations, published in 2014, the marginal gaps ranged from 39 to 201 mm and the internal gaps from 23 to 230 mm [22]. These results are in accordance with the findings of the present study. In this review, based on studies published between 2013 and 2018, the marginal gaps ranged from 48 to 141 mm and the internal gaps from 59 to 238 mm. An improvement in the fit results could perhaps have been hypothesized due to the developments in CAD-CAM technology. However, this could not be seen in this comparison. It may be since, in the review by Boitelle, other materials such as glass-ceramics and alloys were included which could have affected the results. The choice of restorative material has been shown to influence the marginal fit, in a study by R€ odiger et al. with the same settings spacer for all materials, zirconia copings had significantly larger marginal gaps compared to titanium and cobalt-chromium copings [23]. Many of the included studies in this review aimed to compare the fit results of restorations from conventional and digital impressions [9,[24][25][26][27][28][29]. When the results were compared according to the impression technique, the fit and accuracy for all three fit assessment areas were slightly smaller for the digital impression technique. This supports the results from other studies on digital impressions, where single crowns and multiunit FDPs up to 8-units from digital impressions have shown comparable or lower fit values compared to conventional impressions [30][31][32].
When comparing the results from different studies one must be aware of the complexity due to the Table 2. Overview of the intraoral scanners (IOS), laboratory scanners, CAM systems, and zirconia materials used in the included studies.
In the systematic review by Boitelle, only about 50% of the 26 included studies reported the cement spacer settings [22]. In this review only 14 studies were included of the 74 that were originally analyzed in full-text, 25 of the 60 excluded studies did not report the settings. It is important to disclose as much information as possible about the production process and fit assessment technique in the materials section since the settings and parameters may affect the results and conclusions [35]. As an example, it would be a mistake to conclude that a technique or material with a spacer setting of 40 lm and a fit result of 70 lm is more accurate than a technique or material with a spacer setting of 60 lm and a fit result of 80 lm. The first technique is 30 lm from the aimed at spacer setting and the second technique 20 lm. Therefore; the accuracy is higher in technique 20 lm from the setting. In a study by Wettstein et al., the conclusion was that metal-ceramic FDPs had significantly smaller internal gaps compared to zirconia FDPs. However, when taking the spacer into consideration, the only significant difference found was that zirconia FDPs had a significantly smaller occlusal gap [35]. Other studies report the spacer settings but fail to use them when drawing conclusions [36,37].
The majority of the included studies used the silicone replica technique for fit assessment; the advantages of this technique are that it can be used both in vivo and in vitro, and it does not require expensive equipment. The disadvantages are that it is restricted to a two-dimensional analysis of the fit and only the specific points chosen are used. Also, there is a risk of rupture of the light-body material when removing the restoration and it is important to place the restoration and section the replica correctly [6,16]. Nevertheless, the method is considered reliable, although it may overestimate the gap with two to 11% [19,38].
The 3D scan technique can provide a 3D view of the fit, which can be used for both quantitative and qualitative assessments. It can also be used to isolate and measure the fit in specific areas or sections [32,39]. However, it is unclear if studies using the 3D scan technique presents results for the absolute marginal gap or marginal gap [40]. The 3D scan technique may not be the most suitable technique for measuring the absolute marginal gap due to uncertainty in if the outermost edge of the restoration margin is captured with the scanner [41]. Measuring the marginal gap could result in a smaller gap value than represented by the absolute marginal gap, earlier studies on milled CoCr and zirconia have found absolute marginal gaps of 185-260 lm [16,32] and 94-181 lm [42].
The accuracy of zirconia FDPs was $60 lm for MG, IntG and TotG. The dental laboratories could perhaps use this information when designing restorations, by changing the spacer settings to improve the fit. However, these results are based on a wide range of different scanners, design software, and CAM machines, all with several parameters that can affect the fit of a restoration. The dental laboratories should do their own tests and measure the accuracy of the restorations they manufacture and adapt the spacer settings accordingly. The easiest method would be the silicone replica technique.

Conclusions
Within the limitations of the present systematic review, the fit of zirconia single crowns and multiunit FDPs may be regarded as clinically acceptable, and the accuracy of the manufacturing of zirconia is $60 lm for marginal, internal, and total gap. Also, digital impressions seem to be associated with a smaller gap value.