d e n t a l m a t e r i a l s 3 7 ( 2 0 2 1 ) 1273–1282

Available online at www.sciencedirect.com

ScienceDirect

journal homepage: www.intl.elsevierhealth.com/journals/dema

A prospective, multi-center, practice-based cohort

study on all-ceramic crowns

R.J. Wierichs a,b,∗ , E.J. Kramer c , B. Reiss d , F. Schwendicke e , J. Krois e ,

H. Meyer-Lueckel a , T.G. Wolf a,f
a Department of Restorative, Preventive and Pediatric Dentistry, School of Dental Medicine, University of Bern,
Switzerland
b Department of Biohybrid & Medical Textiles, Institute of Applied Medical Engineering, RWTH Aachen University,

Aachen, Germany
c Private Practice Norden, Germany
d German Society of Computerized Dentistry, Berlin, Germany
e Department of Oral Diagnosis, Digital Health and Health Services Research, Charité - Universitätsmedizin Berlin,

Berlin, Germany
f Department of Periodontology and Operative Dentistry, University Medical Center of the Johannes Gutenberg

University Mainz, Germany

a r t i c l e i n f o a b s t r a c t

Article history: Objectives. The aim of this prospective, multi-center, practice-based cohort study was to
Received 10 December 2020 analyze factors associated with the success of all-ceramic crowns.
Received in revised form Methods. All-ceramic crowns placed in a practice-based research network ([Ceramic Suc-
7 April 2021 cess Analysis, AG Keramik) were analyzed. Data from 1254 patients with (mostly in-office
Accepted 24 April 2021 CAD/CAM) all-ceramic crowns placed by 101 dentists being followed up for more than 5
years were evaluated. At the last follow-up visit crowns were considered as successful (not
failed) if they were sufficient, whereas crowns were considered as survived (not lost) if they
Keywords: were still in function. Multi-level Cox proportional hazards models were used to evaluate
(MESH): adhesives the association between a range of predictors and time of success or survival.
Ceramics Results. Within a mean follow-up period (SD) of 7.2(2)years [maximum:15years] 776 crowns
Clinical study were considered successful (annual failure rate[AFR]:8.4%) and 1041 crowns survived
Dental restoration failure (AFR:4.9%). The presence of a post in endodontically treated teeth resulted in a risk for
Longevity failure 2.7 times lower than that of restorations without a post (95%CI:1.4–5.0;p = 0.002).
Prospective studies Regarding the restorative material and adhesive technique, hybrid composite ceramics and
Risk factors single-step adhesives showed a 3.4 and 2.2 times higher failure rate than feldspathic porce-
Success analysis lain and multi-step adhesives, respectively (p < 0.001). Use of an oxygen-blocking gel as well
as an EVA instrument resulted in a 1.5–1.8 times higher failure rate than their non-use (p ≤
0.001).

∗
Corresponding author at: Department of Restorative, Preventive and Pediatric Dentistry, School of Dental Medicine, University of Bern,
Freiburgstrasse 7, CH-3010 Bern, Switzerland.
E-mail address: [email protected] (R.J. Wierichs).
https://doi.org/10.1016/j.dental.2021.04.005
0109-5641/© 2021 The Author(s). Published by Elsevier Inc. on behalf of The Academy of Dental Materials. This is an open access article
under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
1274 d e n t a l m a t e r i a l s 3 7 ( 2 0 2 1 ) 1273–1282

Significance. After up to 15years AFR were rather high for all-ceramic crowns. Operative
factors, but no patient- or tooth-level factors were significantly associated with failure.
The study was registered in the German Clinical Trials Register (DRKS-ID: DRKS00020271).
© 2021 The Author(s). Published by Elsevier Inc. on behalf of The Academy of Dental
Materials. This is an open access article under the CC BY-NC-ND license (http://
creativecommons.org/licenses/by-nc-nd/4.0/).

founders and endpoints, such data might not reflect the

1. Introduction effectiveness of daily dental care in general practice [8,9].
Contrastingly, in general practice dentist-, patient-, tooth-
In recent years several types of ceramics have been introduced
and material- level factors together influences the results. To
to indirectly restore teeth. In general, silica-based ceramics
resolve this, studies require a large number of restorations in
and oxide ceramics can be differentiated. Silica-ceramics offer
one dataset [10]. Practice-based studies offer the opportunity
high translucency and excellent aesthetic results associated
to (routinely) collect large datasets [11]. Furthermore, practice-
with reduced tensile strengths [1]. They can be subdivided
based studies allow to better understand the impact of daily
to feldspathic ceramics and (evolved) glass infiltrated ceram-
treatment decisions on restoration success in general dental
ics. Feldspathic ceramics were used for the first all-ceramic
practice [12].
crowns [2]. They are relatively brittle and show flexural
Thus, the aim of the present prospective, non-
strengths about 60−70 MPa [3]. Several attempts have been
interventional, multi-center, practice-based, clinical study
made to improve the low mechanical stability of feldspathic
was, firstly, to evaluate the longevity of different all-ceramic
ceramics and to expand the clinical use of all-ceramic crowns
crowns and, secondly, to analyze factors influencing the
[3]. For this, crystals of alumina, fluormica, leucite or lithium
success of these restoration after up to 15 years’ follow-up.
disilicate are added to feldspathic ceramics [2,3]. In leucite
glass-ceramics - the first developed pressable glass-ceramics
- the crystalline content is increased up to 50% [3], resulting
in flexural strengths about 160 MPa [3]. In lithium disilicate
2. Materials and methods
glass-ceramics the crystalline content is increased up to 70%
2.1. Study design
correlating with flexural strengths of about 400 MPa [3,4].
However, leucite glass-ceramics still seem to be insufficient
This study was a prospective, non-interventional, multi-
for posterior fixed dental prosthetics and lithium disilicate
center, practice-based, clinical study according to the
glass-ceramics can only be used for 3-unit bridges up to the
European guidelines for good clinical practice (Clinical trials
second premolar [3]. Contrastingly, oxide ceramics consist of
– Directive 2001/20/EC) [13]. This study is reported according
an opaque high-strength core onto which esthetic layering
to the STROBE guideline for cohort studies [14]. The study has
ceramic must be applied to accomplish a natural appearance
been registered in the German Clinical Trials Register (DRKS-
[1]. Their mechanical characteristics seem to be superior to
ID: DRKS00020271).
those of silica-ceramics. In vitro the zirconia- or alumina-
based ceramics show flexural strengths twice as high as the
ones for glass-ceramics [5,6]. Thus, oxide ceramics are recom- 2.2. Restoration selection
mended for posterior crowns and 3-unit posterior bridges and
offer the possibility of masking discolored teeth [3]. Dentists participating in restoratively-focused postgradual
Regarding all-ceramic single crowns acceptable failure education or training courses on CAD/CAM restorations were
rates after 5 years could be observed. In previous system- asked to insert consecutive cases of patients that had the
atic reviews on prospective as well as retrospective cohort need for at least one ceramic restorations in a standardized
studies 5-year success rates were estimated at 93% [4] or digital entry form For that, security and data protection condi-
92–96% [7]. Depending on the ceramic material and tooth tions of the practice-based research network ‘Ceramic Success
position success rates differed. For instance, glass-ceramic Analysis’ (CSA, AG Keramik e.V., Malsch, Germany) had to be
crowns showed a significant lower success rate than alumina accepted [13].
crowns or reinforced glass-ceramic crowns [4] and anterior all- We included all- participating in restoratively storations
ceramic crowns showed significantly lower failure rates than (85% of all cases), CAD-CAM labside fabricated restorations
posterior ones [4,7]. However, in one of the reviews compari- (10%) and restorations manufactured by dental laboratories
son between different ceramic materials and tooth positions (5%). No exclusion criteria applied, i.e. we did not specify the
were calculated indirectly [4] and direct comparisons could number of restorations per patient, the number of patients
only be found for tooth position [7]. Overall, both reviews con- and restorations per dentist, the materials, brands, and tech-
cluded that there is still a lack of information on the long term niques being used.
success of all-ceramic crowns. To minimize reporting bias of the involved dentists not
Most of the studies on all-ceramic crowns have been reporting on their own failures (more information on that can
university-based studies [4]. Although this may allow a more be found in the discussion) only crowns being followed up for
standardized data collection with regard to exposures, con- more than 5 years and all those restorations which had failed
 d e n t a l m a t e r i a l s 3 7 ( 2 0 2 1 ) 1273–1282 1275

during the first five years or afterwards were considered in the meaning whenever the crown was repaired (n = 27), rece-
present manuscript. mented (n = 98) or the tooth received an endodontic treatment
during the observation period (n = 111) the intervention was
2.3. Data extraction not considered as failed, but the observation period was then
censored. In contrast, the restoration was assessed as major
The following data were collected anonymously (without ref- failure (secondary endpoint: no survival) if the restoration was
erence to patient names): On dentist-level; country, sex, the renewed or the tooth was extracted.
dentist itself; patient-level; age, date of the first restorative
treatment, date of the second restorative (re-)intervention, 2.5. Statistical analysis and power analysis
date of the last visit, number of teeth/restorations (per patient)
being included in the study; on tooth-level; characteristics For descriptive purposes frequencies and percentages of mea-
of the involved tooth (Fédération Dentaire Internationale sured baseline characteristics as well as frequencies and
[FDI] notation system), clinical manifestation and sequel of percentages of different failure types were tabulated. Statis-
the caries if present (caries superficialis, caries profunda, tical analysis was performed using SPSS (SPSS 25.0; SPSS,
direct capping), mode of failure, the presence of an endodon- Munich, Germany). Time until any failure was the depen-
tic treatment, type of restoration (single-tooth vs. bridge dent variable. Kaplan–Meier statistics were used to calculate
anchor); on technique-level; cavity outline of the preparation significant differences between the groups (p < 0.05). For
(enamel vs. dentin), finishing line of the preparation (cham- Kaplan–Meier statistic the independent method was used to
fer, shoulder, partial shoulder), technique-related factors (use generate success curves up to 10 years [16]. The annual failure
of rubber dam, matrix, silane, oxygen-blocking gel, EVA oscil- rates (AFR) were calculated from life tables [17].
lating instrument, ultrasonic cementation); on material-level; For further analysis, only crowns being followed up for
materials being used (e.g. ceramic type, adhesive type, luting more than 5 years (n = 886) and all those crowns which had
material). failed during the first five years success: 490 crowns; survival:
The following data were not inserted in the electronic 272 crowns) or afterwards (109 crowns; 61 crowns) were con-
forms and therefore not collected: sidered.
Crude associations between baseline characteristics and
• Characteristics of study participants (e.g. demographic, time until failure were calculated by fitting separate models for
clinical, social) and information on exposures and potential each baseline characteristic as the independent variable. Fac-
confounders tors associated with time until failure (p < 0.25 [18,19]) in the
• Characteristics of the study dentists (e.g. demographic, separate models were entered in a non-clustered multivariate
experience, skills, ‘dentist profile’ [15]) and information on Cox regression model (independent model).
exposures and potential confounders For baseline characteristics with more than 100 variables
separate models and inclusion in multivariate models are not
2.4. Success, survival and failure of treatment appropriate. Nonetheless, dentists can significantly influence
success rates [11,20]. Therefore, for a second multivariate Cox
Assessment of the status of the all-ceramic crowns was done regression analysis the factors dentist and patient were used
in the same practice usually by the dentist who placed the as a cluster-specific random effect (dependent model) [21].
restoration when patients attended for routine care, recall or For the present study no prospective power or sample
when a problem occurred. The observation period started with size calculation was performed since this was a compre-
the restoration being inserted. When the restoration was still hensive data set from an ongoing practice-based research
in function at the last check-up visit and found to be clin- project. Regarding a retrospective power analysis for cate-
ically acceptable, the intervention was defined as success. gories being included in multivariate Cox regression analysis,
Whenever the restoration was replaced, or scheduled for this the analysis provided a power of ≥80% for the categories tooth
treatment at the last check-up, the intervention was consid- type, endodontic treatment, EVA instrument, liner or build-
ered as failed. Also, in cases where a tooth was extracted or up material. Nonetheless, due to the pragmatic design of the
where the replacement of the restoration was related to a present study the study is likely to be underpowered to detect
change in the prosthetic treatment plan, the restoration was moderate to clinically significant relative risks in some cate-
defined as failure. gories.
Success: If the crown was without clinical or radiographic
sings of failures (e.g.: loss of retention or chipping) at the
last follow-up visit the crown was judged as successful. 3. Results
Consequently, the crown was considered as failed (primary
endpoint: no success) if the crown was renewed, repaired, Between November 1996 and March 2019 6543 all-ceramic
recemented or the tooth was extracted / received an endodon- crowns in 4529 patients with at least one follow-up visit were
tic treatment. placed by a total of 140 dentists. However, only 1375 of these
Survival: If the crown was still in function at the last were followed up for more than 5 years or had failed dur-
follow-up visit without clinical or radiographic signs of tech- ing the first five years. The mean number of restorations
nical failure (e.g.: loss of retention or endodontic treatment) (standard deviation [SD]) per patient was 1.1 (0.5) and the num-
the restoration was judged as survived even though a re- ber of restorations per dentist was 14 (22). Characteristics of
treatment was needed in the meantime. In a statistical teeth/crownsare shown in Table 1 and appendix table A.1.
1276 d e n t a l m a t e r i a l s 3 7 ( 2 0 2 1 ) 1273–1282

Table 1 – Frequency, number of failures of teeth included in study and bivariate Cox proportional hazard regression
analyses of time until failure by categories of each baseline characteristic for outcome success.
Category Teeth
Frequency Failures p-value HR 95% CI Estimated 95% CI AFR
Median
success time
[n (%)] [n (%)]
Overall 1375 (100%) 599 (44%) 103.9 98.6–109.2 8.3%
Patient-level
Age [years]
0−20 12 (1%) 3 (25%) 1.0 Reference 85.1 69.5–100.6 4.1%
21−40 334 (24%) 130 (39%) 0.389 1.7 0.5–5.2 115.5 106.7–124.4 6.9%
41−60 711 (52%) 299 (42%) 0.264 1.9 0.6–6 105.8 97.7–113.8 8.1%
>60 317 (23%) 166 (52%) 0.112 2.5 0.8–7.9 87.7 79.1–96.2 11.0%
n/a* 1 (0%) 1 (100%) 0.206 4.3 0.4–41.4 63.1 63.1–63.1 19.0%
Jaw
Upper 691 (50%) 308 (45%) 1.0 Reference 98.6 91.4–105.7 8.8%
Lower 684 (50%) 291 (43%) 0.589 1.0 0.8–1.1 108.7 101.9–115.5 8.5%
Number of restorations per patient
1 971 (71%) 373 (38%) 1.0 Reference 112.9 106.8–119 6.9%
2 144 (10%) 64 (44%) 0.024 1.4 1–1.8 97.2 83.8–110.6 9.7%
3 98 (7%) 63 (64%) <0.001 2.4 1.9–3.2 56.2 47–65.4 18.1%
4 38 (3%) 21 (55%) 0.001 2.2 1.4–3.3 66.5 48.3–84.6 16.2%
≥5 124 (9%) 78 (63%) <0.001 2.6 2.1–3.4 48.2 41.5–54.8 20.8%
Tooth-level
Tooth type
Incisive 56 (4%) 36 (64%) 1.0 Reference 40.8 33.7–48 19.3%
Canine 18 (1%) 15 (83%) 0.101 1.7 0.9–3 27.3 17.6–37.1 39.5%
Premolar 428 (31%) 180 (42%) <0.001 0.5 0.3–0.7 101.6 88–115.2 8.4%
Molar 858 (62%) 359 (42%) <0.001 0.4 0.3–0.6 107.1 100.8–113.4 8.0%
Wisdom 15 (1%) 9 (60%) 0.681 0.9 0.4–1.8 55.5 31.6–79.3 16.5%
Endodontic treatment
No 749 (54%) 340 (45%) 1.0 Reference 101.9 95–108.7 9.5%
Yes without post 397 (29%) 200 (50%) 0.265 1.1 0.9–1.3 89.1 81.5–96.7 10.6%
Yes with post 229 (17%) 59 (26%) <0.001 0.5 0.3–0.6 113.2 106.9–119.4 4.1%
Clinical manifestation of the caries
Caries superficialis 409 (30%) 164 (40%) 1.0 Reference 113.8 105.4–122.3 8.0%
Caries profunda 365 (27%) 169 (46%) 0.150 1.2 0.9–1.5 95.0 86.5–103.5 9.5%
Direct capping 11 (1%) 9 (82%) 0.014 2.3 1.2–4.5 55.4 30.9–79.9 20.1%
n/a* and no caries 590 (43%) 257 (44%) 0.629 1.0 0.9–1.3 98.8 92–105.6 8.4%
Technique-level
Cavity outline
Enamel 245 (18%) 94 (38%) 1.0 Reference 113.8 104.9–122.8 6.4%
Dentin 1130 (82%) 505 (45%) 0.002 1.4 1.1–1.8 99.8 93.4–106.3 9.2%
Rubber dam
Use 283 (21%) 100 (35%) 1.0 Reference 124.5 115.3–133.8 5.9%
Non-use 1092 (79%) 499 (46%) <0.001 1.6 1.3–2 95.1 90–100.2 9.6%
Matrix
Use 218 (16%) 87 (40%) 1.0 Reference 110.9 99.5–122.2 7.3%
Non-use 1157 (84%) 512 (44%) 0.075 1.2 1–1.5 103.3 97.6–108.9 8.9%
Silane
Use 1170 (85%) 484 (41%) 1.0 Reference 107.8 101.9–113.7 8.1%
Non-use 205 (15%) 115 (56%) 0.003 1.4 1.1–1.7 87.3 77.5–97.2 11.8%
Ultrasonic cementation
Use 262 (19%) 118 (45%) 1.0 Reference 103.5 93.4–113.7 9.1%
Non-use 1113 (81%) 481 (43%) 0.935 1.0 0.8–1.2 103.8 96.8–110.8 8.6%
Dental flossing
Use 1235 (90%) 545 (44%) 1.0 Reference 103.3 97.8–108.8 8.9%
Non-use 140 (10%) 54 (39%) 0.034 0.7 0.6–1 115.2 102.1–128.4 6.7%
Oxygen-blocking
Use 297 (22%) 160 (54%) 1.0 Reference 86.7 76–97.4 11.9%
Non-use 1078 (78%) 439 (41%) <0.001 0.7 0.6–0.8 108.0 102.1–113.8 7.9%
EVA instrument
Use 155 (11%) 84 (54%) 1.0 Reference 67.5 58.5–76.4 13.5%
Non-use 1220 (89%) 515 (42%) <0.001 0.6 0.5–0.8 106.8 101.3–112.4 8.2%
 d e n t a l m a t e r i a l s 3 7 ( 2 0 2 1 ) 1273–1282 1277

– Table 1 (Continued)
Category Teeth
Frequency Failures p-value HR 95% CI Estimated 95% CI AFR
Median
success time
[n (%)] [n (%)]
Finishing line of the preparation
Chamfer 522 (38%) 270 (52%) 1.0 Reference 87.0 80.1–93.8 11.5%
Partial shoulder 208 (15%) 102 (49%) 0.937 1.0 0.8–1.2 92.8 79.4–106.1 11.5%
Shoulder 593 (43%) 213 (36%) <0.001 0.6 0.5–0.7 111.8 105.5–118.1 6.5%
n/a* 52 (4%) 14 (27%) <0.001 0.3 0.2–0.5 146.7 131.1–162.3 3.4%
Material-level
Build-up
No 692 (50%) 317 (46%) 1.0 Reference 99.4 92.6–106.2 9.4%
Yes 683 (50%) 282 (41%) 0.100 0.9 0.7–1 111.3 104.6–118 8.0%
Adhesive type
Multi-step 765 (56%) 276 (36%) 1.0 Reference 115.5 108.7–122.2 6.7%
Single-step 192 (14%) 124 (65%) <0.001 2.2 1.8–2.8 77.0 65.6–88.4 15.8%
No adhesive 91 (7%) 32 (35%) 0.885 1.0 0.7–1.4 94.1 83.9–104.3 6.2%
n/a* 327 (24%) 167 (51%) <0.001 1.7 1.4–2 80.0 73.2–86.9 n/a
Ceramic type
Feldspathic porcelain (FP) 456 (33%) 169 (37%) 1.0 Reference 118.5 110.8–126.2 6.4%
Leucite glass-ceramic (LEU) 285 (21%) 98 (34%) 0.993 1.0 0.8–1.3 101.8 95.1–108.5 6.2%
Lithium dissilicate glass-ceramic (LD) 473 (34%) 205 (43%) <0.001 1.5 1.2–1.8 82.6 77.2–87.9 9.2%
Hybrid composite 79 (6%) 74 (94%) <0.001 6.1 4.6–8 24.0 19.9–28.1 50.1%
Zr und Al-oxide 82 (6%) 53 (65%) <0.001 2.6 1.9–3.6 64.3 51.1–77.4 17.6%
Luting material
Photoactivated luting agent 297 (22%) 107 (36%) 1.0 Reference 119.6 109.9–129.3 6.0%
Dual-cured luting agent 961 (70%) 428 (45%) <0.001 1.5 1.2–1.8 97.7 92.1–103.3 9.1%
Chemicalactivated luting agent 107 (8%) 61 (57%) <0.001 2.4 1.8–3.3 55.8 48.2–63.5 16.3%
Provisorial 10 (1%) 3 (30%) 0.934 1.0 0.3–3 72.0 55.4–88.5 3.9%
Type of fabrication
CAD-CAM chairside 1171 (85%) 487 (42%) 1.0 Reference 107.3 101.7–112.9 7.9%
CAD-CAM lab 136 (10%) 97 (71%) <0.001 2.9 2.3–3.6 40.0 33.8–46.3 26.0%
Individual lab 68 (5%) 15 (22%) 0.005 0.5 0.3–0.8 118.3 105.9–130.7 3.8%

Factors associated with time until failure (p < 0.25; bold) in the separate models were entered in the multivariate Cox regression model (Table 2).
* n/a: not available, for some crowns one or two (sub-)categories was/were not provided.

3.1. Success and survival

At the end of the observation time (mean (SD) follow-up

period: 7.2 (2) years (maximum: 15 years)) 56% of the crowns
(776 out of 1375) were considered as successful (Table 1), no
additional treatment was necessary. Moreover, 75% of the
crowns (776 out of 1158) survived (appendix table A.1). Regard-
ing the categories success and survival the AFRs were 8.4%
and 4.9%, respectively. However, AFR between dentists varied
widely (mean(95%CI): 12.6% (6.1%–19.1%)). The main failure
type was extraction of the tooth (n = 90) followed by ceramic
fracture (n = 45) and defective margin (n = 34). However, for
58 failures the reason for failure was not provided. The suc-
cess curves of restorations according to ceramic materials are
shown in Fig. 1. Success stratified according to adhesive types
are presented in Fig. 2.
Fig. 1 – Kaplan–Meier curve for all-ceramic crowns
according to the factor ceramic material.
3.2. Cox Regression Analysis

Crude bivariate associations between the different baseline liner or build-up material, adhesive type finishing line of the
characteristics and an increased failure rate are given in preparation clinical manifestation of the caries and type of
Table 1 and appendix table A.1. Endodontic treatment, cavity fabrication were possibly associated with increased failure
outline, oxygen-blocking gel, EVA instrument, ceramic type, rates (p < 0.25).
1278 d e n t a l m a t e r i a l s 3 7 ( 2 0 2 1 ) 1273–1282

Table 2 – Multivariate Cox proportional hazard regression analyses of time until failure as function of baseline
characteristics identified (for outcome success).
Category p-value HR 95% CI
Patient-level
Age [years] 0.292
0−20 1.0 Reference
21−40 0.819 1.144 0.4–3.6
41−60 0.694 1.259 0.4–4
>60 0.539 1.435 0.5–4.6
n/a 0.212 4.605 0.4–50.7
Number of restorations per patient <0.001
1 1.0 Reference
2 0.263 1.177 0.9–1.6
3 <0.001 1.767 1.3–2.4
4 0.113 1.467 0.9–2.4
≥5 <0.001 1.780 1.3–2.4
Tooth-level
Tooth type 0.007
Incisive 1.0 Reference
Canine 0.416 1.295 0.7–2.4
Premolar 0.046 0.675 0.5–1
Molar 0.014 0.623 0.4–0.9
Wisdom 0.470 1.317 0.6–2.8
Endodontic treatment <0.001
No 1.0 Reference
Yes without post 0.376 0.748 0.4–1.4
Yes with post 0.002 0.373 0.2–0.7
Clinical manifestation of the caries 0.272
Caries superficialis 1.0 Reference
Caries profunda 0.124 1.198 1–1.5
Direct capping 0.330 1.420 0.7–2.9
Na and no caries 0.198 1.514 0.8–2.8
Technique-level
Cavity outline
Enamel 1.0 Reference
Dentin 0.229 1.170 0.9–1.5
Rubber dam
Use 1.0 Reference
Non-use 0.630 0.940 0.7–1.2
Matrix
Use 1.0 Reference
Non-use 0.289 1.162 0.9–1.5
Silane
Use 1.0 Reference
Non-use 0.444 0.891 0.7–1.2
Dental flossing
Use 1.0 Reference
Non-use 0.319 0.843 0.6–1.2
Oxygen-blocking
Use 1.0 Reference
Non-use <0.001 0.559 0.5–0.7
EVA instrument
Use 1.0 Reference
Non-use 0.001 0.647 0.5–0.8
Finishing line of the preparation
Chamfer 1.0 Reference
Partial shoulder 0.579 1.1 0.8–1.4
Shoulder 0.046 0.8 0.7–1
Sonstiges <0.001 0.2 0.1–0.5
Material-level
Build-up
No 1.0 Reference
Yes 0.413 1.1 0.9–1.3
Adhesive type <0.001
Multi-step 1.0 Reference
 d e n t a l m a t e r i a l s 3 7 ( 2 0 2 1 ) 1273–1282 1279

– Table 2 (Continued)
Category p-value HR 95% CI
Single-step <0.001 2.2 1.7–2.8
No adhesive 0.037 0.7 0.4–1
Na 0.045 1.3 1–1.7
Ceramic type <0.001
Feldspathic porcelain (FP) 1.0 Reference
Leucite glass-ceramic (LEU) 0.997 0.999 0.8–1.3
Lithium dissilicate glass-ceramic (LD) 0.064 1.245 1–1.6
Hybrid composite <0.001 3.420 2.5–4.7
Zr und Al-Oxid 0.510 1.159 0.7–1.8
Luting material
Photoactivated luting agent 1.0 Reference
Dual-cured luting agent 0.059 1.3 1–1.6
Chemicalactivated luting agent 0.686 0.9 0.6–1.4
Provisorial 0.112 0.4 0.1–1.3
Type of fabrication
CAD-CAM chairside 1.0 Reference
CAD-CAM lab <0.001 2.214 1.6–3
Individual lab 0.003 0.426 0.2–0.7

Bold p-values (p < 0.05) indicate factors strongly associated with a de- or increased failure rate.

ces remained (non-)significant, with two exceptions. The

difference between endodontically treated teeth with and
without posts become non-significant.
In the clustered multivariate model, in which the factors
dentist and patient were used as a cluster-specific ran-
dom effect, and the model, in which all 6543 all-ceramic
crowns were included, the included (non-)significant pre-
dictors remained (non-)significant (appendix table A.3 and
appendix table A.4).

4. Discussion

This multi-center, practice-based, clinical cohort study

prospectively analyzed the success of all-ceramic crowns. A
total of 6543 crowns with at least one follow-up visit were
Fig. 2 – Kaplan–Meier curve for all-ceramic crowns placed by a total of 140 dentists, of whom 1375 crowns have
according to the factor adhesive type. been followed-up for more than 5 years. The influence of sev-
eral baseline parameters on the success were analyzed. At
overall moderate annual failure rates the ceramic type, the
adhesive type, the build-up material, EVA oscillating instru-
The results of the non-clustered multivariate models ment and oxygen-blocking gel were significant predictors for
including factors possibly associated with an increased failure decreased time until failure for both success and survival.
rate in the bivariate models are shown in Table 2 (success) and A large variety of ceramic materials have been introduced.
appendix table A.2 (survival). After multivariate regression the Hybrid materials, including ceramic/polymer materials and
additional oxygen-blocking (HR (95% CI): 1.8 (1.4–2.0); p < 0.001) polymer infiltrated ceramic-network materials, are supposed
as well as the use of an EVA oscillating instrument (HR: 1.5 to combine the positive aspects of both ceramics and com-
(1.3–2.0); p = 0.001) shortened the time until any failure com- posites [22]. Compared to conventional ceramic materials
pared to treatment without them. In contrast, for all-ceramic hybrid materials are supposed to show good edge stability
crowns the presence of a post in endodontically treated teeth for minimal layer thickness, economical machinability, and
resulted in a significantly lower risk for failure than that of less brittleness [23–25]. However, until now only a few (con-
endodontically treated teeth without a post (HR: 2.7 (1.4–5.0); trolled) clinical studies [26] or case reports [27] have been
p < 0.001). The use of hybrid composite ceramics and single- published. In these studies, success rates of 97–100% were
step adhesives showed a 3.4 (95% CI:2.5−4.7) and 2.2 (95% reported. Contrastingly, in the present study hybrid compos-
CI:1.7−2.8) times higher failure rate than feldspathic porcelain ite ceramics showed a 2.7–3.0 times higher failure rate than
and multi-step adhesives (p < 0.001), respectively. CAD-CAM lithium disilicate glass-ceramic or Zr- or Al-oxide ceramics as
labside fabricated restorations showed a 2.2 (95% CI: 1.6–3.0) well as a 3.4 times higher failure rate than feldspathic porce-
higher risk for failure than chairside fabricated restorations. lain or leucite glass-ceramic. Thus, although hybrid composite
For survival, the included (non-)significant predictors for suc- ceramics seem to have superior material properties in vitro
1280 d e n t a l m a t e r i a l s 3 7 ( 2 0 2 1 ) 1273–1282

these factors seem to have limited relevance for the clinical facilitate treatment procedures within the daily clinical rou-
practice. tine, ‘gold-standard’ adhesives (3-step etch and rinse or 2-step
From a clinical perspective in restorative dentistry annual self-etch system) should be preferred, also for placing all-
failure rates at 5–10 years below 6% are considered as sat- ceramic crowns. Furthermore, in previous in vitro studies it
isfying [10]. Thus, the present study in private practice could be observed that due to the acidic nature of single-step
environments showed higher failure rates for success (AFR: adhesive the bonding of them to self/dual-cured composites
8.4%) and satisfying failure rates for survival (AFR:4.9%). Fur- might be compromised [30,31]. Thus, it might be assumed that
thermore, the present failure rates were higher than those this observation can also be seen in the present data. However,
of previous studies on all-ceramic restorations. In previews since several different (single-step) adhesives, luting materi-
reviews cumulative failure rates after 5 years varied between als, and ceramics have been used in combination, the present
20% and 0% [4,7]. However, for the present study data had study cannot satisfactorily address this assumption.
to be collected manually via an online platform. Thus, sev- In contact with activated monomer molecules, oxygen has
eral dentists might not succeed to upload follow-up data, e.g. a high affinity to form free radicals [32]. During the light cur-
due to a loss of motivation, the voluntary characteristic of ing, oxygen from the atmosphere inhibits the polymerization
the CSA network [13] or the factor that dentists do not want process [33]. In the delayed post-curing stage it could also be
to report one’s own failure. To minimize these reporting bias shown that the degree of monomer conversion and the surface
only crowns being followed up for more than 5 years and all micro-hardness were significantly higher in the absence than
those restorations which had failed during the first five years in the presence of oxygen [32]. Therefore, oxygen-blocking
or afterwards were considered. In contrast, this procedure may gels were developed by several manufactures to enhance the
result in failure rates being higher than they actually would be. polymerization process of adhesive materials. However, in the
This might also be seen in the increase of the AFR for survival. present study the additional use of an oxygen-blocking gel
AFR increases from 2.1% after a mean follow-up of 2.5 years resulted in a 1.8 times higher failure rate than the non-use.
to 4.9% after a mean follow-up of 7.2 years. The actual num- Interestingly, up to date in vitro the absence of oxygen was
ber of failed crowns after 5 years might, thus, in-between 2.1% mostly simulated with oxygen free conditions and not by using
and 4.9% and hence at the higher end of the range of previous oxygen-blocking gels. Furthermore, no in vivo study analyzed
university-based studies. the influence of oxygen-blocking gels for all-ceramic single
In the present study the assessment of the restorations crowns. Although there is a lack of evidence, three aspects
was done by the dentist who placed the restoration and not might explain the observed results: (1) The color and/or the
by a second blinded dentist. No intra- or inter-examiner cali- application form of the oxygen-blocking gels might reduce the
bration regarding treatment decisions was performed prior to intensity or change the spectrum of the curing light; thereby
the study. However, before joining this study, dentists were in reducing the strength of the tooth-restoration interface (2). For
specific continuing education and training courses. Nonethe- clinicians it might be easier to remove a slight excess of a lut-
less, due to the high number of included dentists, criteria for ing material than to subsequently (re-)fill a gap at the margin
failures might differentiate between dentists. This, of course, of a restoration. Consequently, when polishing the margins
causes difficulties to control evaluation bias and confounders. of the restoration the luting material with a lower degree of
However, predictors in the dependent model, in which the fac- monomer conversion might be removed (3). The use of an
tor dentist and patient were used as a cluster-specific random oxygen-blocking gel is an insufficiently trained working step;
effect and the independent model were (almost) the same. thereby, the clinicians’ focus is shifted from the correct inser-
Furthermore, a high number of more than 1000 all-ceramic tion of the restoration to the correct use of the blocking gel.
crowns could, thus, be included in one single study compared Nonetheless, further clinical studies are needed to improve
to recent systematic reviews in which 6006 crowns [4] or 2943 our knowledge about the use of oxygen-blocking gels.
crowns [7] had been included. Interestingly, the studies being For 155 crowns a sono-abrasion technique (EVA oscillating
included in the reviews also were (prospective or retrospective) instrument) was used to finish the interproximal or cervical
cohort studies. No randomized controlled trial was available. restoration areas. This technique is supposed to reduce risk
Although intra- or inter-examiner calibration has presumably of damage to adjacent surfaces when compared with rotary
been performed within a single study, it can be assumed that instruments [34]. Furthermore, it is supposed to not negatively
(1) for the 33 prospective cohort studies (being included in the affect the stability of the tooth-restoration interface [34,35].
reviews) criteria for failures differed between the studies and However, when used with different bonding systems sono-
that (2) for the 15 retrospective studies no calibration could abrasion seems to enhance or reduce the bonding strength
have been performed due to the study design. Thus, for both, to enamel and dentin [36]. With a two-step self-etch adhe-
the reviews and the present study, there are almost the same sive a significantly lower bonding strength to enamel and a
difficulties to control bias and confounders. significantly higher bonding strength to dentin was observed
In a recent study of the CSA network ceramic ceramic when compared with a three-step etch and rinse adhesive.
inlays/onlays showed a 2.4 times higher failure rate after Thus, it was indicated that in the clinical protocol the adhesive
the use of a single-step adhesive than after the use of a 3- materials and the sono-abrasion technique should be aligned.
step etch and rinse or 2-step self-etch system [13]. This is in However, in the present data set several different adhesives,
agreement with clinical studies on the longevity of posterior luting materials, and ceramics have been used in combination
ceramic coverages [28], non-carious cervical class V restora- with the sono-abrasion technique. This diversity can presum-
tions [29] and all-ceramic crowns (present study). Although ably also be seen in the present results: the use of an EVA
single-step adhesives should shorten treatment time and oscillating instrument resulted in a 1.6 times higher risk of
 d e n t a l m a t e r i a l s 3 7 ( 2 0 2 1 ) 1273–1282 1281

failure than the non-use. Thus, the present results might indi- all-ceramic and metal-ceramic reconstructions after an
cate that, the used materials and the sono-abrasion technique observation period of at least 3 years. Part I: single crowns.
should be aligned when using the EVA oscillating instrument. Clin Oral Implants Res 2007;18(Suppl 3):73–85.
[5] Seghi RR, Denry IL, Rosenstiel SF. Relative fracture
However, it might also be speculated that EVA oscillating
toughness and hardness of new dental ceramics. J Prosthet
instruments were used often in those cases that deep sub- Dent 1995;74:145–50.
gingival outlines were present. Consequently, an indication [6] Lüthy H. Strength and toughness of dental ceramics. In:
bias would have led to the present finding. Since it was only Mörmann WH, editor. CAD/CIM in aesthetic dentistry.
recorded if the preparation outlines were in dentin or enamel Quintessence: CEREC 10-year Anniversary Symposium
but not if the outlines ware supra-,epi, or subgingival, the Chicago; 1996. p. 229–40.
present study cannot satisfactorily address this speculation. [7] Kassardjian V, Varma S, Andiappan M, Creugers NH, Bartlett
D. A systematic review and meta analysis of the longevity of
Within the limitation of this study, moderate failure rates
anterior and posterior all-ceramic crowns. J Dent
for all-ceramic crowns could be found in a private practice 2016;55:1–6.
environment after up to 15 years. Operative factors, but no [8] Kramer EJ, Meyer-Lueckel H, Wolf TG, Schwendicke F,
patient- or tooth-level factors were significantly associated Naumann M, Wierichs RJ. Success and survival of
with failure. The use of hybrid composite ceramics, the pres- post-restorations: six-year results of a prospective
ence of a liner or composite build-up, the use of a single-step observational practice-based clinical study. Int Endod J
2019;52:569–78.
adhesives, the use of an EVA oscillating instrument and the
[9] Perrin P, Meyer-Lueckel H, Wierichs RJ. Longevity of
use of an oxygen-blocking gel were significant predictors for
immediate rehabilitation with direct fiber reinforced
failure. However, for all-ceramic single crowns further studies composite fixed partial dentures after up to 9 years. J Dent
are needed to improve our knowledge about several of these 2020:103438.
factors and to analyze patient-related factors which have not [10] Opdam NJ, van de Sande FH, Bronkhorst E, Cenci MS,
been included in the present study. Bottenberg P, Pallesen U, et al. Longevity of posterior
composite restorations: a systematic review and
meta-analysis. J Dent Res 2014;93:943–9.
Funding [11] Wierichs RJ, Kramer EJ, Meyer-Lueckel H. Risk factors for
failure of direct restorations in general dental practices. J
Dent Res 2020, 22034520924390.
This study was funded by the authors and their institutions. [12] Mjor IA. Practice-based dental research. J Oral Rehabil
2007;34:913–20.
[13] Collares K, Correa MB, Laske M, Kramer E, Reiss B, Moraes
Informed consent RR, et al. A practice-based research network on the survival
of ceramic inlay/onlay restorations. Dent Mater
For this type of study, formal consent is not required. 2016;32:687–94.
[14] von Elm E, Altman DG, Egger M, Pocock SJ, Gotzsche PC,
Vandenbroucke JP, et al. The strengthening the Reporting of
Acknowledgement Observational Studies in Epidemiology (STROBE) statement:
guidelines for reporting observational studies. J Clin
Epidemiol 2008;61:344–9.
The authors are grateful to the dental practitioners who
[15] Collares K, Opdam NJM, Laske M, Bronkhorst EM, Demarco
treated the patients and provided their data and to the
FF, Correa MB, et al. Longevity of anterior composite
practice-based research network ‘Ceramic Success Analysis’ restorations in a general dental practice-based network. J
(CSA, AG Keramik e.V., Malsch, Germany) who concepted, Dent Res 2017;96:1092–9.
designed, and organized the data entries. [16] Chuang SK, Tian L, Wei LJ, Dodson TB. Kaplan–Meier analysis
of dental implant survival: a strategy for estimating survival
with clustered observations. J Dent Res 2001;80:2016–20.
Appendix A. Supplementary data [17] Linnemann T, Kramer EJ, Schwendicke F, Wolf TG,
Meyer-Lueckel H, Wierichs RJ. Longevity and Risk Factors of
Post Restorations after up to 15 Years: A Practice-based
Supplementary material related to this article can be
Study. J Endod 2020;47(4):577–84.
found, in the online version, at doi:https://doi.org/10.1016/j. [18] Wierichs RJ, Kramer EJ, Meyer-Lueckel H. Risk factors for
dental.2021.04.005. failure of class V restorations of carious cervical lesions in
general dental practices. J Dent 2018;77:87–92.
references [19] Wierichs RJ, Kramer EJ, Meyer-Lueckel H. Risk factors for
failure in the management of cervical caries lesions. Clin
Oral Investig 2017;21:2123–31.
[20] Collares K, Correa MB, Bronkhorst EM, Laske M, Huysmans
[1] Gehrt M, Wolfart S, Rafai N, Reich S, Edelhoff D. Clinical M, Opdam NJ. A practice based longevity study on
results of lithium-disilicate crowns after up to 9 years of single-unit crowns. J Dent 2018;74:43–8.
service. Clin Oral Investig 2013;17:275–84. [21] Chuang SK, Tian L, Wei LJ, Dodson TB. Kaplan-Meier analysis
[2] McLean JW, Hughes TH. The reinforcement of dental of dental implant survival: a strategy for estimating survival
porcelain with ceramic oxides. Br Dent J 1965;119:251–67. with clustered observations. J Dental Res 2001;80:2016–20.
[3] Saint-Jean SJ. Chapter 12 - Dental glasses and [22] Spitznagel FA, Vuck A, Gierthmuhlen PC, Blatz MB, Horvath
glass-ceramics. In: Shen JZ, Kosmac T, editors. Advanced SD. Adhesive bonding to hybrid materials: an overview of
ceramics for dentistry. Amsterdam: Elsevier; 2014. p. 255–78. materials and recommendations. Compend Contin Educ
[4] Pjetursson BE, Sailer I, Zwahlen M, Hammerle CH. A Dent 2016;37:630–7.
systematic review of the survival and complication rates of
1282 d e n t a l m a t e r i a l s 3 7 ( 2 0 2 1 ) 1273–1282

[23] Reich S. Tooth-colored CAD/CAM monolithic restorations. [31] Tay FR, Suh BI, Pashley DH, Prati C, Chuang SF, Li F. Factors
Int J Comput Dent 2015;18:131–46. contributing to the incompatibility between simplified-step
[24] Edelhoff D, Liebermann A, Beuer F, Stimmelmayr M, Guth JF. adhesives and self-cured or dual-cured composites. Part II.
Minimally invasive treatment options in fixed Single-bottle, total-etch adhesive. J Adhes Dent
prosthodontics. Quintessence Int (Berl) 2016;47:207–16. 2003;5:91–105.
[25] Coldea A, Swain MV, Thiel N. Mechanical properties of [32] Aromaa MK, Vallittu PK. Delayed post-curing stage and
polymer-infiltrated-ceramic-network materials. Dent Mater oxygen inhibition of free-radical polymerization of
2013;29:419–26. dimethacrylate resin. Dent Mater 2018;34:1247–52.
[26] Lu T, Peng L, Xiong F, Lin XY, Zhang P, Lin ZT, et al. A 3-year [33] Rueggeberg FA, Margeson DH. The effect of oxygen
clinical evaluation of endodontically treated posterior teeth inhibition on an unfilled/filled composite system. J Dent Res
restored with two different materials using the CEREC AC 1990;69:1652–8.
chair-side system. J Prosthet Dent 2018;119:363–8. [34] Decup F, Lasfargues JJ. Minimal intervention dentistry II: part
[27] Selz CF, Vuck A, Guess PC. Full-mouth rehabilitation with 4. Minimal intervention techniques of preparation and
monolithic CAD/CAM-fabricated hybrid and all-ceramic adhesive restorations. The contribution of the sono-abrasive
materials: a case report and 3-year follow up. Quintessence techniques. Br Dent J 2014;216:393–400.
Int 2016;47:115–21. [35] de Oliveira MT, de Freitas PM, de Paula Eduardo C,
[28] van Dijken JW, Hasselrot L. A prospective 15-year evaluation Ambrosano GM, Giannini M. Influence of diamond
of extensive dentin-enamel-bonded pressed ceramic sono-abrasion, air-abrasion and Er:YAG laser irradiation on
coverages. Dent Mater 2010;26:929–39. bonding of different adhesive systems to dentin. Eur J Dent
[29] Peumans M, Kanumilli P, De Munck J, Van Landuyt K, 2007;1:158–66.
Lambrechts P, Van Meerbeek B. Clinical effectiveness of [36] Van Meerbeek B, De Munck J, Mattar D, Van Landuyt K,
contemporary adhesives: a systematic review of current Lambrechts P. Microtensile bond strengths of an etch&rinse
clinical trials. Dent Mater 2005;21:864–81. and self-etch adhesive to enamel and dentin as a function of
[30] Cheong C, King NM, Pashley DH, Ferrari M, Toledano M, Tay surface treatment. Oper Dent 2003;28:647–60.
FR. Incompatibility of self-etch adhesives with
chemical/dual-cured composites: two-step vs one-step
systems. Oper Dent 2003;28:747–55.