Clinical Oral Investigations

https://doi.org/10.1007/s00784-021-03844-5

REVIEW

Adhesive strategies in cervical lesions: systematic review

and a network meta-analysis of randomized controlled trials
Fabiana Dias Simas Dreweck 1,2 & Adrieli Burey 1 & Marcelo de Oliveira Dreweck 3 & Alessandro D. Loguercio 1 &
Alessandra Reis 1

Received: 16 August 2020 / Accepted: 16 February 2021

# Springer-Verlag GmbH Germany, part of Springer Nature 2021

Abstract
Objectives A network meta-analysis (NMA) was performed to assess which adhesive strategy is most clinically effective in
treating non-carious cervical lesions (NCCLs).
Material and methods Studies were identified by a systematic search of electronic databases including MEDLINE via PubMed,
Brazilian Library in Dentistry (BBO), Cochrane Library, EMBASE, Latin American and Caribbean Health Sciences Literature
database (LILACS), Scopus, and Web of Science without restrictions on publication year or language. The grey literature was
also consulted. Only randomized clinical trials that compared different adhesive strategies in NCCLs in adult patients were
included. The risk of bias was evaluated by using the Cochrane Collaboration tool. A random-effects Bayesian mixed treatment
comparison model was used to compare adhesive strategies (3ER, 2ER, 2SE, and 1SE) at different follow-up times. The surface
under cumulative ranking curve (SUCRA) was estimated for each strategy. Heterogeneity was assessed by using the Cochran Q
test and I2 statistics. The quality of evidence was evaluated using the GRADE approach.
Results A total of 5058 studies were identified, 66 of which met the eligibility criteria and of these 5 were judged “low” risk of bias and
57 were meta-analyzed. We did not observe significant differences in the NMA analysis for any two pairs of adhesives, except for the
shortest follow-up for 2ER vs 3ER. The material 2SE ranked highest, although it differed only slightly from the other bonding strategies.
Conclusions No bonding strategy is better than the others.
Clinical relevance Adhesive efficacy cannot be characterized by its bonding strategy.

Keywords Dentin bonding agents . Dental bond . Non-carious cervical lesions . Clinical effectiveness . Network meta-analysis

* Alessandra Reis Introduction

[email protected]
Fabiana Dias Simas Dreweck Although numerous laboratory investigations have measured
[email protected] the bond strength values of different dental adhesives in an
Adrieli Burey attempt to predict their clinical outcomes [1, 2], randomized
[email protected] controlled trials (RCTs) remain the most appropriate research
Marcelo de Oliveira Dreweck design for assessing the clinical efficacy of any intervention
[email protected] [3, 4].
Alessandro D. Loguercio To evaluate the effectiveness and clinical performance of
[email protected] adhesive systems, the American Dental Association (ADA)
[5] recommends clinical trials on non-carious cervical lesions
1
Department of Restorative Dentistry, School of Dentistry, State (NCCLs), as composite resin restorations only remain bonded
University of Ponta Grossa, Rua Carlos Cavalcanti, 4748 Bloco M, to these lesions by the micromechanical interlocking produced
Sala 64-A, Uvaranas, Ponta Grossa, Paraná 84030-900, Brazil
by the adhesive systems [6]. The immediate, short-term, and
2
Campos Gerais Higher Education Center – CESCAGE, Ponta long-term bonding performance of adhesive systems is then
Grossa, PR, Brazil
3
evaluated by retention, marginal integrity, and marginal
Medicine Department, State University of Ponta Grossa, Uvaranas, discoloration.
Ponta Grossa, PR, Brazil
 Clin Oral Invest

Ideally, the clinical decision-making process should be CDR42018112672) and adhered to the recommendations of
based on the best available evidence encompassing RCTs; the Preferred Reporting Items for Systematic Reviews
however, the increasing number of RCTs published about (PRISMA-NMA) guidelines and the corresponding extension
adhesive systems prevents an overall conclusion about the for network meta-analysis [11].
most efficient adhesive strategy. Systematic reviews can facil-
itate this challenge through research synthesis of multiple
studies. Information sources and search strategy
Systematic reviews aim to collect, appraise, and synthetize
the amount of evidence in a specific domain. Although stan- A literature search was performed in MEDLINE/PubMed,
dard systematic reviews have reported on adhesive strategies Brazilian Library in Dentistry (BBO), Cochrane Central
for dental bonding [7–9], they present comparisons between Register of Controlled Trials (CENTRAL), EMBASE, Latin
pairs of different adhesive strategies and rarely involve multi- American and Caribbean Health Sciences Literature database
ple comparisons. When choosing adhesive systems for clini- (LILACS), and in the citation databases Scopus and Web of
cal purposes, several competing types of adhesive strategies Science, with no data or language restriction and using a
are available, and among them we can cite the 3-step and 2- predefined search strategy.
step etch-and-rinse adhesives (3ER and 2ER, respectively) Controlled vocabulary (MeSH and Entree terms) and free
and the 2-step and 1-step self-etch adhesives (2SE and 1SE, keywords, defined based on the concept of population (adult
respectively). Additionally, we can also have 1-step adhesives patients requiring NCCL restorations) and intervention (com-
being used with selective enamel etching and the universal posite restorations with adhesive systems), were combined
adhesives, which may be used as etch-and-rinse or self-etch within each concept using the Boolean operator “OR.” The
adhesives. Thus, one of the most important questions remain- concepts (population and intervention) were combined with
ing to be answered is which adhesive strategy offers the the Boolean operator “AND” and whenever possible with a
greatest benefits in terms of retention, marginal discoloration, validated filter from PubMed.
and marginal integrity. We hand-searched the reference lists of all primary
Clinicians who need to decide among these different strat- studies and the related article link of each primary study
egies would benefit from a single review that includes all in the PubMed database. Grey Literature in Europe
relevant adhesive strategies and presents their comparative (SIGLE), Google Scholar and abstracts of the
efficacy. Network meta-analysis is a technique that allows International Association for Dental Research (1998–
for the comparison of three or more interventions simulta- 2019), and ongoing and unpublished trials (Current
neously in a single analysis by combining both direct and Controlled Trials, International Clinical Trials Registry
indirect evidence across a network of studies, even when there Platform, Clinical Trials.gov, Brazilian Clinical Trials
are no head-to-head trials for some of the interventions. The Registry, and EU Clinical Trials Register) were also
authors are aware of a previous network review [10]; howev- searched.
er, limitations regarding date and database restrictions moti-
vated us to carry out another review on the topic.
Therefore, the aim of the present systematic review and Eligibility criteria
network meta-analysis was to establish a clinically meaningful
hierarchy of the different adhesive approaches to bond com- We included published and unpublished RCTs with paired or
posite resin restoration in NCCL cavities through the synthesis multiple restorations per participant that evaluated at least two
of available evidence from RCTs. To this end, we aimed to different adhesive strategies in NCCLs. Studies not adhering
answer the following PICO (population, intervention, com- to the inclusion criteria were excluded. The following exclu-
parison, and outcome) question: What is the comparative ef- sion criteria were applied to the studies:
fectiveness of adhesive strategies (I and C) for bonding com-
posite resin restorations in adult patients (P) in terms of reten- & Adhesive 2SE used in 3ER mode
tion (O) and what is the relative ranking of these strategies? & Compared two adhesives from the same strategy (without
control group)
& Conducted in other types of cavities (class I, II, III, or IV)
Materials and methods & With follow-up less than 12 months
& Performed dentin pretreatment
Protocol and registration & Cavities were lined with other materials
& Used 2nd and 3rd generation adhesives
This study protocol was registered with the International & Used phosphoric acid etching at low concentrations (e.g.,
Prospective Register of Systematic Reviews (PROSPERO- 10%)
Clin Oral Invest

Study selection and data collection process When trials compared more than two bonding strategies,
they were included in the meta-analysis separately to provide
Articles appearing in more than one database were considered more than one effect size. Data were extracted using intention-
only once. Subsequently, two reviewers (F.D.S.D. and A.B.) to-treat analysis by using the total number of failures of each
evaluated titles and abstracts to remove ineligible studies. treatment arm at each follow-up as the nominator and the total
Full-text articles were acquired from likely eligible studies, number of participants randomized at baseline as the denom-
and two reviewers classified those meeting the inclusion inator wherever trial reporting allowed.
criteria. Each eligible article received a study ID combining The primary outcome evaluated was restoration loss.
first author and year of publication. Three authors from this We calculated risk ratios for this binary outcome in a
study extracted relevant information about the study design, traditional pairwise meta-analysis, presenting their 95%
participants, type of adhesive, use of dental dam, enamel bev- confidence intervals at different follow-up periods (12 to
el, dentin preparation, number of operators, number of exam- 24 months, 36 to 48 months, and > 48 months). In case
iners, and evaluation criteria using customized extraction two or more adhesive systems from the same bonding
forms; in cases of disagreement, a decision was reached by strategy were investigated in the primary study, data were
consensus. Multiple reports of the same study (reports with merged to make a single entry. In case the study reported
different follow-up times) were extracted directly into a pre- data twice in the follow-up range described above, data
viously tested, single data collection form to avoid overlap- from the longer follow-up was taken.
ping data. The events were classified as dichotomous out- Transitivity was assumed to occur in all studies, meaning
comes at the end of each follow-up period. that the different sets of interventions were sufficiently similar
to provide valid indirect inferences. For this purpose, we ap-
plied narrow inclusion criteria to keep population and treat-
Risk of bias within individual studies
ments as similar as possible within and across treatment
comparisons.
The risk of bias (RoB) of the selected trials were carried out by
two independent reviewers using the Cochrane Collaboration
Statistical analysis
Risk of Bias Tool (version 1.0) for RCTs [12]. The RoB tool
contained six domains: sequence generation, allocation con-
We performed a network meta-analysis (NMA) by using the
cealment, blinding of the outcome assessors, incomplete out-
Bayesian model with the statistical package geMTC in R (ver-
come data, selective outcome reporting, and other sources of
sion 3.4.2). The mixed treatment comparison methodology,
bias. The last domain (other bias) was not used in the present
supported by the Markov Chain Monte Carlo hierarchy, was
study. During data selection and quality assessment, any dis-
chosen to carry out the NMA. This model allows for the si-
agreements between the reviewers were resolved through dis-
multaneous comparison of all four adhesive strategies and the
cussion and consultation with a third reviewer (A.R.). The
incorporation of trials with three or more arms. Random-
judgement for each entry consisted of recording low, high,
effects models with the DerSimonian and Laird variance esti-
or unclear risk of bias (either lack of information or uncertain-
mator and the inverse of the variance method were used. The
ty about the potential for bias).
convergence was based on the Brooks Gelman-Rubin criteria
with inspection of trace plots, and 20,000 interactions were
Summary measures and planned methods of analysis undertaken for 4 chains at a thinning interval of 10.
Heterogeneity was assessed using the Cochran Q test and I2
The adhesive strategies were classified as (A) 3-step ER, (B) statistics.
2-step ER, (C) 2-step SE, or (D) 1-step SE approaches. The The results of the network meta-analysis were
universal adhesives were classified into category B when used displayed in point estimates, 95% credible intervals
in the 2-step ER approach and into category D when used in (95% CrI). We also calculated the relative ranking for
the 1-step SE approach. Self-etch adhesives used in selective e a c h i n t e r v e n t i o n u s i n g t h e S u r f a c e U n d e r th e
enamel etching were considered either 1-step SE or 2-step SE. Cumulative Ranking curve (SUCRA), estimated within
The 4-step ER Syntac adhesive (Ivoclar Vivadent) was clas- the Bayesian framework. A SUCRA value of 100% indi-
sified as a 3-step ER adhesive. cates that the treatment is certain to be the most effective
In studies where 1-step SE adhesives were treated as a in the network, while a value of 0% indicates that it is
primer and coated with a hydrophobic adhesive layer, they certain to be the least effective. The larger the SUCRA
were considered a 2-step SE. The same was true for universal value, the better the rank of an intervention in the net-
adhesives when used in the 2-step ER mode; however, with an work. All analyses were implemented using the Meta
additional coat of a hydrophobic layer, the adhesive was clas- and geMTC packages of the R statistical software pro-
sified as 3-step ER and treated as such in this study. gram [13].
 Clin Oral Invest

Assessment of inconsistency population sample at different follow-up periods and received

the same study ID [16–30]. Therefore, 66 studies were included
A further assumption of NMA is consistency, the statistical in the qualitative synthesis and 57 in the quantitative analysis.
agreement between the direct and indirect comparisons. The Nine studies [31–39] were excluded from the quantitative studies
consistency assumption is the statistical manifestation of tran- because they were abstracts without data description as demon-
sitivity and depends on the statistical agreement between dif- strated in the flow diagram (Supplementary Fig. S1).
ferent sources of evidence. Statistical inconsistency was
checked using posterior plots and Bayesian p values produced Characteristics of included articles
by the node-splitting method of Dias et al. (2010) [14] by
testing the agreement between direct and indirect evidence. The characteristics of the 66 eligible studies are listed in
A p value equal to or greater than 0.008 was considered the Supplementary Table S3. Thirty-seven studies used the paired
threshold for significance after Bonferroni correction, as the design [31–33, 35, 37, 38, 40–70], twenty-eight studies per-
same data were used in six multiple comparisons. formed multiple restorations per participant [36, 39, 71–96],
and only one did not report this information [34]. The follow-
Small study effects and publication bias up periods ranged from 3 to 108 months, and great variability
was observed in the age range of participants in all the includ-
Publication bias was assessed by funnel plot asymmetry. The ed studies (18 to 88 years). A wide variation of commercial
presence of small study effects was evaluated by drawing a brands of each adhesive strategy and composite resin for res-
comparison-adjusted funnel plot that accounts for the fact that toration was used in the included studies.
different studies compare different sets of interventions. The Among the 3-step ER, the two most tested materials were
null hypothesis was tested by using the Egger test in all Optibond FL (Kerr) [31, 33, 34, 51, 60, 61, 75, 76, 85, 88, 95]
follow-up periods at a significance level of 5%. and Scotchbond Multi-Purpose (3M Oral Care) [41, 43, 52, 53,
57, 81] used in 11 and 6 comparisons respectively. The most
Assessment of the certainty of evidence using tested adhesives in the clinical trials were Adper Single Bond
grading of recommendations: assessment, (3M Oral Care) in the category of 2-step ER, Clearfil SE Bond
development, and evaluation (Kuraray) as 2-step SE, and Scotchbond Universal (3 M Oral
Care) as 1-step SE. In three studies, an extra layer of a hydro-
We followed the GRADE approach to appraise the confidence in phobic bonding resin (Scotch bond Multi-Purpose 3M Oral
estimates derived from the network meta-analysis of retention Care) was applied as in Perdigão 2019 [82] (Universal to 3ER,
rates following the Puhan et al. approach [15]. Direct evidence Universal to 2SE), Reis 2009 [63] (1SE to 2SE), and Sartori
from RCTs starts at high confidence and can be rated based on 2013 [65] (1SE to 2SE).
risk of bias, indirectness, imprecision, inconsistency (or hetero- The majority of the studies used cotton rolls, retraction cords,
geneity), and/or publication bias to levels of moderate, low, and and a saliva ejector to isolate the operative field [39, 41–43, 45,
very low confidence. The rating of indirect estimates starts at the 46, 49, 51, 52, 60, 62, 65–70, 72, 74, 76, 78–83, 86, 87, 89–95,
lowest rating of the pairwise estimates that contribute as first- 97], and three reported that the use of dental dam depended on the
order loops to the indirect estimate but can be rated further for location and access of the lesion [31, 84, 85]. Twenty-three stud-
imprecision or intransitivity (dissimilarity between studies in ies did not perform any preparation on either enamel or dentin
terms of clinical or methodological characteristics). If direct and [33, 44–46, 49, 54, 55, 58–60, 64–66, 70, 73, 75, 78, 80, 81, 87,
indirect estimates were similar (i.e., coherent), then the higher of 89, 90, 98], twelve studies prepared both enamel and dentin [51,
the ratings was assigned to the network meta-analysis estimates. 52, 69, 71, 76, 77, 79, 83, 84, 88, 95, 96], and eight did not report
this information [31, 32, 34–38, 82]. The number of operators
ranged from 1 to 6, the number of evaluators ranged from 1 to 3,
Results and in eight studies, this information was not reported [31, 32,
35–38, 50, 57]. The most used criteria for restoration evaluation
Study selection was the modified USPHS [34, 35, 38, 39, 42–47, 49, 50, 52, 54,
57, 58, 62, 65–67, 69, 72, 74, 75, 79, 81–83, 85, 86, 88, 89,
The search strategy was conducted initially on February 9, 2019, 91–94, 96, 97], but the FDI criteria [37, 51, 70, 77, 78] and the
and was updated on November 20, 2019 (Supplementary Vanherle method [76, 95] were also described.
Table S1). A total of 5058 studies were retrieved from electronic
databases. After removal of duplicates and title and abstract Assessment of the risk of bias
screening, 143 studies remained. Of these, 62 studies were ex-
cluded (Supplementary Table S2), which left 81 eligible random- The RoB of the eligible studies is presented in Fig. 1. With regard
ized controlled trials. Of these, fifteen reported the same to the specific items of the risk of bias assessment tool by the
Clin Oral Invest

Fig. 1 a Risk of bias graph according to the Cochrane Collaboration tool and b risk of bias summary
 Clin Oral Invest

Fig. 1 (continued)
Clin Oral Invest

Cochrane Collaboration, 65% of the included studies indicated an observed in the pair 1SE vs 3ER by split node analysis
unclear risk of bias for random-sequence generation, 85% for (p value = 0.006). Heterogeneity was analyzed in the
allocation concealment, 80% for blinding of participants and per- primary studies, and the pair 2ER vs 3ER showed
sonnel, 60% for blinding of outcome, 30% for incomplete data moderate heterogeneity (τ 2 = 0.236; I 2 = 28.8%).
outcome, and 40% for selective reporting. Overall, most of the Both Leave-One-Out and the Baujatplot method
RCTs had an unclear risk of bias, and only five studies were showed that the study of Van Dijken 2000 [91] was
classified as being at low risk of bias [56, 58, 59, 70, 75]. causing the heterogeneity (Supplementary Table S4 and
Fig. 2). The removal of this study corrected the het-
Evidence network erogeneity of the 2ER vs 3ER comparison and the
inconsistency of the 1SE vs 3ER comparison without
Figure 2 displays the network of the four adhesive strategies altering the NMA results for the 12- to 24-month fol-
(3ER, 2ER, 2SE, and 1SE) for each follow-up, where each node low-up.
represents an adhesive strategy. The strategies connected by a
line represent direct comparisons, with the number of pairs (from Loss of retention at 12 to 24 months Traditional pairwise
RCTs) reflected by the thickness of the edges and the number of meta-analysis for all possible pairs can be found in
restorations reflected by the size of the nodes. Supplementary Fig. S3. Figure 3 summarizes the direct,
Fifty-seven studies were included in three independent indirect, and pooled estimates for comparisons of bond-
meta-analyses for loss of retention at the different follow- ing strategies. We observed a significant difference in
up periods (12 to 24 months, 36 to 48 months, and > 48 the pair 2SE vs 3ER (RR = 0.72; 95% CrI 0.52 to 0.99)
months). In all follow-ups, the 1SE bonding strategy is the in the NMA analysis in favor of the 2SE bonding strat-
one with the highest number of placed restorations. egy. No other significant difference was found among
Although a high number of restorations were evaluated in any two pairs of bonding strategies.
the RCTs that reported short-term follow-ups, this number P values in Fig. 3 indicate the probability that the di-
was significantly lower in longer follow-up periods due to rect and indirect evidence is consistent. The smallest
participants dropouts and fewer RCTs. Bayesian p value found for inconsistency was equal to
In the 12- to 24-month and 36- to 48-month follow-ups, 0.05 (for 2ER vs 3ER) and therefore higher than the
there is evidence from all possible direct pairwise compari- threshold of 0.008 (after Bonferroni correction), showing
sons. At > 48 months, some pairwise comparisons show direct that we do not have evidence to reject the hypothesis of
evidence only (from head-to-head studies, e.g., 2ER vs 2SE), consistency. We did not observe heterogeneity (p > 0.38;
some show indirect evidence only (e.g., 3ER vs 2SE), and the I2 = 7%; Supplementary Table S5).
other pairs show both direct and indirect evidence.
Loss of retention at 36 to 48 months Traditional pairwise
Synthesis of network results meta-analysis for all possible pairs can be found in
Supplementary Fig. S4. No significant difference in the
In the first approach, we included all studies in the NMA was found among any two pairs (Fig. 4).
network analysis. An important inconsistency was Heterogeneity was observed (p < 0.01; I 2 = 72%;

Fig. 2 Networks of the comparisons of the adhesive strategies at 12 to 24 months (a), at 36 to 48 months (b), and at > 48 months (c). The size of the node
reflects the number of evaluated restorations and the thickness of the connecting lines the number of pairs being compared
 Clin Oral Invest

Fig. 3 Forest plot of direct, indirect, and network evidence for retention rates at 12 to 24 months produced by the split node method. QoE, quality of
evidence. 1Most studies are at unclear risk of bias. 2Imprecise estimates

Supplementary Table S6), but inconsistency between di- vs 3ER, as this was the single comparison with both
rect and indirect evidence was not detected. direct and indirect evidence. We found no significant
difference between pairs. Only four pairs had direct
Loss of retention at > 48 months Traditional pairwise meta- comparisons: 2ER vs 1SE, 2ER vs 2SE, 3ER vs 2ER,
analysis for all possible pairs can be found in and 3ER vs 1SE. The heterogeneity was high (p = 0.02;
Supplementary Fig. S5. Figure 5 summarizes the direct, I2 = 82%; Supplementary Table S7), but inconsistency
indirect, and pooled estimates for the comparison 2ER was not detected.
Clin Oral Invest

Fig. 4 Forest plot of direct, indirect, and network evidence for retention rates at 36 to 48 months produced by the split node method. QoE, quality of
evidence. 1Most studies are at unclear risk of bias. 2Imprecise estimates

SUCRA rankings for all study follow-ups for all study follow-ups. The probability of being the
best adhesive strategy varies in the study follow-ups,
In the primary probabilistic analysis, strategies were but the 2SE strategy is ranked as first in two study
ranked as having the higher probability of being the follow-ups (12 to 48 months and > 48 months; Fig. 6
first, second, third, and fourth at each study follow-up. a and c) and is ranked as second in the 36- to 48-month
Figure 6 illustrates the ranking and the SUCRA values follow-up (Fig. 6b).
 Clin Oral Invest

Fig. 5 Forest plot of direct, indirect and network evidence for retention rates at > 48 months produced by the split node method

Small study effects and publication bias were found for the Egger test in all follow-up periods
(at 12 to 24 months [p = 0.76], at 36 to 48 months [p =
Publication bias was not observed in any of the study 0.08], and at > 48 months [p = 0.88]).
follow-ups (Fig. 7). Studies with high precision (high
sample sizes) are plotted near the point estimate, and Quality of evidence
studies with low precision (small sample sizes) are
spread evenly on both sides of the point estimate, cre- The ratings of the quality of the direct, indirect, and network
ating a roughly funnel-shaped distribution. We did not evidence can be seen in the Supplementary Table S8. In gen-
reject the null hypothesis because nonsignificant p values eral, the quality of evidence of the network meta-analysis
Clin Oral Invest

a b c

Fig. 6 Rankogram and SUCRA for loss of retention showing the cumulative rank order for each adhesive strategy at 12 to 24 months (a), at 36 to 48
months (b), and > 48 months (c)

(Figs. 3, 4, and 5) was graded as low because of unclear risk of Network meta-analyses (NMA) are capable of simulta-
bias and imprecision. neously addressing the comparative efficacy of multiple inter-
ventions by combining direct and indirect estimates of effect.
This statistical approach is adequate to evaluate multiple treat-
ment options. Differently from an earlier NMA [104], we did
Discussion not merge different follow-up periods, as failure rates seem to
increase over time and merging them may lead to misleading
This systematic review and network meta-analysis was con- conclusions.
ducted to allow a more comprehensive evaluation of all avail- In the traditional pairwise meta-analysis (Supplementary
able dental bonding strategies. Previous pairwise meta- Figs. S3, S4, and S5) from the present study, none of the
analyses have compared any two bonding strategies among bonding strategies was better than the other, with their point
all possible pairs [7, 99–101], some with flaws in their meth- estimates (risk ratio) always crossing the null value of one. By
odology. In two of them, no appraisal of the risk of bias of the adding indirect evidence to these direct pairwise meta-analy-
eligible studies was performed, and the authors did not choose ses, the resulting NMA did not increase precision enough to
the appropriate statistical analysis for data management [7, 8]. allow us to conclude that one bonding strategy is better than
Earlier systematic reviews attempting to include all adhe- any other, except for a single comparison (2ER vs 3ER) in the
sive strategies did not preserve the within-trial randomization shortest follow-up. Because of the imprecise estimate of this
[8, 102] as they pooled single arms across studies, ignoring the comparison, we cannot conclude that this finding did not oc-
comparator that was used in the primary studies and treating cur by chance.
data as if they came from a single large randomized trial. This Researchers wish to provide clinicians with the best choice
type of analysis discards the benefits of within-trial randomi- among available treatment options. This is why we deter-
zation (Li and Dickersin, 2013) [103] and should be avoided. mined ranks based on the Bayesian approach by calculating

a b c

Fig. 7 Funnel plot of included studies at 12 to 24 months (a), at 36 to 48 months (b), and at > 48 months (c)
 Clin Oral Invest

the Surface Under the Cumulative Ranking Curve (SUCRA). labeled by their bonding strategy, as it depends, among other
SUCRA value is a single numeric presentation of the overall things, much more on the balanced chemical composition of
ranking of the materials representing the probability of a treat- structural and functional monomers, solvents, polymerization
ment ranking best. An overview of the SUCRA values in the initiators, inhibitors, or stabilizers.
three study follow-ups highlights that the bonding strategy In all bonding strategies, there are efficient and inefficient
that ranked best (ranked as first in two follow-ups and as adhesives, and, when merged by the label of their bonding
second in the other) was the 2SE. Similar SUCRA values were strategy, results are similar. Future studies should focus more
also reported in Schwendicke et al. (2016) with the 2SE ad- on evaluating specific commercial brands both in short- and
hesive being ranked as first most often. long-term follow-up periods.
Although this result is easier to grasp compared with the Limitations of the present systematic review include that
other reported statistics, it should be interpreted with caution. the authors of RCTs of bonding studies have not reported the
SUCRA does not consider the magnitude of differences in study findings in a standardized way, and this may result in
effects between treatments. This means that the material misleading conclusions. In some clinical trials, events at the
ranked first may be only slightly better than the second ranked shortest follow-up period were not carried forward to the lon-
treatment. SUCRA evaluation will always provide a ranking gest follow-up periods, leading to misleading results.
of being the best even if the statistics that compare effect sizes Additionally, as the recall rate typically drops drastically in
do not demonstrate that any materials are better than the other. long-term follow-ups, review authors may calculate the reten-
Although 2SE was rated as being the one with the higher tion rates based on the number of recalled restorations and not
probability of being the best, this adhesive strategy was only on the total number of restorations placed at baseline.
significantly different from the 3ER adhesives in two study We performed data extraction from the primary studies by
follow-ups. Considering this fact along with the unclear risk of following the intention-to-treat analysis, always evaluating the
bias of most of the eligible studies, the conclusion that 2SE is worst-case scenario. However, there was no standardization in
the best approach seems unwise and precipitate. the primary studies regarding the reporting of events (cumu-
Even if not clinically or statistically relevant differences in lative or not) and dropout reporting over the follow-up pe-
the efficacy of treatments are found, the difference in their riods. The standardization of the collection of the data provid-
ranks will imply otherwise. This indeed was observed in the ing this information is recommended by the ADA guidelines
present study, as all comparisons except one (2ER vs 3ER; 12 [5], but this is not observed in most studies. All these concerns
to 24 months) were similar to one another. If, indeed, 2SE is to regarding data extraction indicate the urgent need to standard-
be considered the best bonding strategy as reported by ize the reporting of studies conducted in NCCLs. Instead of
SUCRA, one may conclude that the difference between this providing retention rates per follow-up, the use of survival
material and the other bonding strategies is not sufficient to analysis could provide better estimates of what occurs to the
justify changing the adhesive systems being used by clinicians adhesives over time.
or public health systems. It is important to highlight that there are some clinical
Reporting that different bonding strategies have similar variables that are usually described in the literature as
performance may make readers suspicious. For more than playing a role on the retention rates of composite restora-
20 years, there has been a widespread belief that 3ER systems tions in non-carious cervical lesion such as enamel bevel-
are the best bonding strategy for restorative and luting proce- ing, dentin roughening, and rubber dam isolation.
dures. Several laboratory studies with immediate and aged Although the present study did not evaluate these vari-
interfaces reported superior bond strength values for the 3ER ables, previous literature findings reported that we do not
systems. The authors of a recent meta-analytical review of have enough evidence to support the findings that neither
parameters on bond strength values, De Munck et al. 2012 the enamel beveling [107] nor dentin roughening [56] in-
[105], concluded that the 3ER Optibond FL performed best. creases retention rates. Similarly, the literature does not
However, we should bear in mind that laboratory and clin- have evidence of high quality concluding that rubber
ical findings often do not coincide. Although an earlier study dam usage improves retention rates of restorations [108].
by Van Meerbeek 2010 [4] reported a correlation between This discussion should encourage researchers in the devel-
laboratory and clinical data, this correlation could have been opment of well-designed, low risk of bias studies to allow
by chance because it was only found between “aged” bond more conclusive findings about these issues.
strength data and medium-term retention rates of adhesives. Another important consideration is that most of the RCTs
The authors of the present investigation are not concluding focused on short- and medium-term follow-ups, 12 to 48
that all dental adhesives have similar performance. There have months. In these short-term follow-ups, the number of events,
been short-term reports of high failure rates of some adhesives e.g., debonded restorations, is low, leading to imprecise esti-
when tested in some primary studies [37, 91, 106]. What is mates. Indeed, this was one of the reasons why the quality of
concluded is that the efficacy of adhesive systems cannot be the evidence was downgraded.
Clin Oral Invest

Finally, we would like to encourage researchers to conduct 7. Chee B, Rickman LJ, Satterthwaite JD (2012) Adhesives for the
restoration of non-carious cervical lesions: a systematic review. J
more studies of low risk of bias to eventually come up with
Dent 40:443–452. https://doi.org/10.1016/j.jdent.2012.02.007
more definitive conclusions about the research question 8. Peumans M, De Munck J, Mine A, Van Meerbeek B (2014)
evaluated. Clinical effectiveness of contemporary adhesives for the restora-
tion of non-carious cervical lesions. A systematic review. Dent
Mater 30:1089–1103. https://doi.org/10.1016/j.dental.2014.07.
007
Conclusion 9. Masarwa N, Mohamed A, Abou-Rabii I, Abu Zaghlan R, Steier L
(2016) Longevity of self-etch dentin bonding adhesives compared
to etch-and-rinse dentin bonding adhesives: a systematic review. J
Based on the results reported, we concluded that no strategy is Evid Based Dent Pract 16:96–106. https://doi.org/10.1016/j.
better than any other in terms of retention. The authors of this jebdp.2016.03.003
study discourage clinicians, researchers, and teachers from 10. Schwendicke F, Blunck U, Paris S, Gostemeyer G (2015) Choice
of comparator in restorative trials: a network analysis. Dent Mater
labeling the efficacy of the adhesives based on their bonding 31:1502–1509. https://doi.org/10.1016/j.dental.2015.09.021
strategy. However, we must bear in mind that these conclu- 11. Hutton B, Salanti G, Caldwell DM, Chaimani A, Schmid CH,
sions were based mostly on studies of unclear risk of bias. Cameron C, Ioannidis JP, Straus S, Thorlund K, Jansen JP,
Mulrow C, Catala-Lopez F, Gotzsche PC, Dickersin K, Boutron
Supplementary Information The online version contains supplementary I, Altman DG, Moher D (2015) The PRISMA extension statement
material available at https://doi.org/10.1007/s00784-021-03844-5. for reporting of systematic reviews incorporating network meta-
analyses of health care interventions: checklist and explanations.
Ann Intern Med 162:777–784. https://doi.org/10.7326/m14-2385
Funding This study was partially supported by the National Council for 12. Higgins JP, Altman DG, Gotzsche PC, Juni P, Moher D, Oxman
Scientific and Technological Development (CNPq) under grants 303332/ AD, Savovic J, Schulz KF, Weeks L, Sterne JA (2011) The
2017-4 and 308286/2019-7 and Coordenação de Aperfeiçoamento de Cochrane collaboration's tool for assessing risk of bias in
Pessoal de Nível Superior - Brasil (CAPES) - Finance Code 001. randomised trials. Bmj 343:d5928. https://doi.org/10.1136/bmj.
d5928
Declarations 13. Yavorska OO, Burgess S (2017) MendelianRandomization: an R
package for performing Mendelian randomization analyses using
Ethics approval This article does not contain any studies with human summarized data. Int J Epidemiol 46:1734–1739. https://doi.org/
participants performed by any of the authors. 10.1093/ije/dyx034
14. Dias S, Welton NJ, Caldwell DM, Ades AE (2010) Checking
consistency in mixed treatment comparison meta-analysis. Stat
Informed consent This article does not contain any studies with human Med 29:932–944. https://doi.org/10.1002/sim.3767
participants performed by any of the authors. 15. Puhan MA, Schunemann HJ, Murad MH, Li T, Brignardello-
Petersen R, Singh JA, Kessels AG, Guyatt GH (2014) A
Conflict of interest The authors declare no competing interests. GRADE Working Group approach for rating the quality of treat-
ment effect estimates from network meta-analysis. Bmj 349:
g5630. https://doi.org/10.1136/bmj.g5630
16. Aw TC, Lepe X, Johnson GH, Mancl LA (2005) A three-year
References clinical evaluation of two-bottle versus one-bottle dentin adhe-
sives. J Am Dent Assoc 136:311–322
17. Loguercio AD, Bittencourt DD, Baratieri LN, Reis A (2007) A
1. Heintze SD, Thunpithayakul C, Armstrong SR, Rousson V (2011)
36-month evaluation of self-etch and etch-and-rinse adhesives in
Correlation between microtensile bond strength data and clinical
noncarious cervical lesions. J Am Dent Assoc 138:507–514 quiz
outcome of class V restorations. Dent Mater 27:114–125. https://
535-7
doi.org/10.1016/j.dental.2010.09.005
18. D.G. Dall’orologio FF, Lorenzi R (2008) Restoration of cervical
2. Heintze SD, Ruffieux C, Rousson V (2010) Clinical performance lesions: 60-month results of a RCT. J Dent Res.
of cervical restorations–a meta-analysis. Dent Mater 26:993– 19. D.G.Dall’Orologio DL, Fazzi F, Lorenzi R (2009) Restoration of
1000. https://doi.org/10.1016/j.dental.2010.06.003 nccl: 6-year results of a RCT. J Dent Res.
3. Van Meerbeek B, De Munck J, Yoshida Y, Inoue S, Vargas M, 20. D.G. Dall’orologio FF, Lorenzi R (2010) Restoration of cervical
Vijay P, Van Landuyt K, Lambrechts P, Vanherle G (2003) lesions: 7-year resukts of a RCT. J Dent Res.
Buonocore memorial lecture. Adhesion to enamel and dentin: cur- 21. Robles NLA, Fu C, Givan D, Burgess J (2018) Clinical evaluation
rent status and future challenges. Oper Dent 28:215–235 of universal and two bottle total etch adhesives at 5 years. J Adhes
4. Van Meerbeek B, Peumans M, Poitevin A, Mine A, Van Ende A, Dent 97(A)
Neves A, De Munck J (2010) Relationship between bond-strength 22. Loguercio AD, Reis A (2008) Application of a dental adhesive
tests and clinical outcomes. Dent Mater 26:e100–e121. https://doi. using the self-etch and etch-and-rinse approaches: an 18-month
org/10.1016/j.dental.2009.11.148 clinical evaluation. J Am Dent Assoc 139:53–61
5. ADA (2001) American Dental Association Council on Scientific 23. Barceleiro LLM, Calazans F, Albuquerque E, Reis A, Loguercio
Affairs. Acceptance program guidelines, dentin and enamel adhe- A (2018) 18-month follow-up of composite restorations with non-
sive materials. MDP universal adhesive system. J Dent Res 97
6. Nakabayashi N, Kojima K, Masuhara E (1982) The promotion of 24. Perdigao J, Kose C, Mena-Serrano AP, De Paula EA, Tay LY,
adhesion by the infiltration of monomers into tooth substrates. J Reis A, Loguercio AD (2014) A new universal simplified adhe-
Biomed Mater Res 16:265–273. https://doi.org/10.1002/jbm. sive: 18-month clinical evaluation. Oper Dent 39:113–127. https://
820160307 doi.org/10.2341/13-045-c
 Clin Oral Invest

25. Loguercio AD, de Paula EA, Hass V, Luque-Martinez I, Reis A, 43. Aw TC, Lepe X, Johnson GH, Mancl L (2004) One-year clinical
Perdigao J (2015) A new universal simplified adhesive: 36-month evaluation of an ethanol-based and a solvent-free dentin adhesive.
randomized double-blind clinical trial. J Dent 43:1083–1092. Am J Dent 17:451–456
https://doi.org/10.1016/j.jdent.2015.07.005 44. Bittencourt DD, Ezecelevski IG, Reis A, Van Dijken JWV,
26. Dutra-Correa VKM, Ciaramicoli M, Pecorari V, Rodrigues F, Loguercio AD (2005) An 18-months evaluation of self-etch and
Saraceni C (2015) Randomized trial of four dentin adhesive sys- etch & rinse adhesive in non-carious cervical lesions. Acta
tems in NCCLs: 42-months. J Dent Res 94 Odontol Scand 63:173–178. https://doi.org/10.1080/
27. Dutra-Correa M, Kiyan VH, Ciaramicoli MT, Pecorari V, 00016350510019874
Rodrigues FP, Coury Saraceni CH (2019) Randomized clinical 45. Brackett MG, Dib A, Franco G, Estrada BE, Brackett WW (2010)
trial of four adhesion strategies: a 42 month study. Indian J Dent Two-year clinical performance of Clearfil SE and Clearfil S3 in
Res 30:487–495. https://doi.org/10.4103/ijdr.IJDR_466_16 restoration of unabraded non-carious class V lesions. Oper Dent
28. Van Landuyt KL, Peumans M, De Munck J, Cardoso MV, Ermis 35:273–278. https://doi.org/10.2341/09-266-c
B, Van Meerbeek B (2011) Three-year clinical performance of a 46. Brackett WW, Brackett MG, Dib A, Franco G, Estudillo H (2005)
HEMA-free one-step self-etch adhesive in non-carious cervical Eighteen-month clinical performance of a self-etching primer in
lesions. Eur J Oral Sci 119:511–516. https://doi.org/10.1111/j. unprepared class V resin restorations. Oper Dent 30:424–429
1600-0722.2011.00855.x 47. Burgess JO, Sadid-Zadeh R, Cakir D, Ramp LC (2013) Clinical
29. Van Landuyt KL, De Munck J, Ermis RB, Peumans M, Van evaluation of self-etch and total-etch adhesive systems in
Meerbeek B (2014) Five-year clinical performance of a HEMA- noncarious cervical lesions: a two-year report. Oper Dent 38:
free one-step self-etch adhesive in noncarious cervical lesions. 477–487. https://doi.org/10.2341/12-355-cr
Clin Oral Investig 18:1045–1052. https://doi.org/10.1007/ 48. Do FG (2013) Análise clínica e clínica/digital de restaurações de
s00784-013-1061-9 lesões cervicais não cariosas: efeitos dos sistemas adesivos e
30. Peumans M, Wouters L, De Munck J, Van Meerbeek B, Van tempos de análise. Universidade Estadual Paulista Júlio de
Landuyt K (2018) Nine-year clinical performance of a HEMA- Mesquita Filho UNESP
free one-step self-etch adhesive in noncarious cervical lesions. J 49. Dutra-Correa M, Saraceni CH, Ciaramicoli MT, Kiyan VH,
Adhes Dent 20:195-203. doi: 10.1016/j.pdpdt.2018.06.010. Queiroz CS (2013) Effect of chlorhexidine on the 18-month clin-
https://doi.org/10.3290/j.jad.a40630 ical performance of two adhesives. J Adhes Dent 15:287–292.
31. Armstrong SR, Cobb DS, Denehy G, Guzman-Armstrong S, https://doi.org/10.3290/j.jad.a29533
Hernandez M (2012) Clinical evaluation of one-step self-etching 50. Fu ARC, Givan D, Ramp L, Lawson N, Burgess J (2015) Three-
adhesive in non-carious cervical lesions. J Dent Res. year clinical evaluation of self-etching adhesives in class V resto-
32. Batalha-Silva S, GCL, Vieira LCC (2009) Clinical evaluation of rations. J Adhes Dent 94(B)
two adhesive systems in non-carious cervical lesions. J Dent Res. 51. Haak R, Hahnel M, Schneider H, Rosolowski M, Park KJ, Ziebolz
33. Blunck JSU, NS, Zaslansky P (2013) Two-year clinical perfor- D, Hafer M (2019) Clinical and OCT outcomes of a universal
mance of one-step self-etch adhesive in cervical restorations. J adhesive in a randomized clinical trial after 12 months. J Dent
Dent Res. 90:103200. https://doi.org/10.1016/j.jdent.2019.103200
34. G.D. Dall’Orologio RL (2006) Restoration of cervical lesions: 3- 52. Kim SY, Lee KW, Seong SR, Lee MA, Lee IB, Son HH, Kim HY,
year results of a RCT. J Dent Res. Oh MH, Cho BH (2009) Two-year clinical effectiveness of adhe-
35. Friedl K-AH K-H, Jung H, Schlittenbauer T, Lichtblau J, Schmalz sives and retention form on resin composite restorations of non-
G (2004) Clinical evaluation of composite restorations using dif- carious cervical lesions. Oper Dent 34:507–515. https://doi.org/
ferent adhesive systems. J Dent Res. 10.2341/08-006C
36. Helbig EB, Klimm HW, IES, Rietschel J, Richter G, Haufe E 53. Lawson NC, Robles A, Fu CC, Lin CP, Sawlani K, Burgess JO
(2004) A one-bottle and a self-etching adhesives in class V resto- (2015) Two-year clinical trial of a universal adhesive in total-etch
rations. J Dent Res. and self-etch mode in non-carious cervical lesions. J Dent 43:
37. Lopes LS, Calazans FS, Hidalgo R, Buitrago LL, Gutierrez F, 1229–1234. https://doi.org/10.1016/j.jdent.2015.07.009
Reis A, Loguercio AD, Barceleiro MO (2016) Six-month fol- 54. Loguercio AD, Costenaro A, Silveira AP, Ribeiro NR, Rossi TR,
low-up of cervical composite restorations placed with a new uni- Reis A (2006) A six-month clinical study of a self-etching and an
versal adhesive system: a randomized clinical trial. Oper Dent 41: etch-and-rinse adhesive applied as recommended and after dou-
465–480. https://doi.org/10.2341/15-309-c bling the number of adhesive coats. J Adhes Dent 8:255–261
38. Pavolucci MCG, Tiriduzzi P, Orsini G, Putignano A, Ferrari M, 55. Loguercio AD, Manica D, Ferneda F, Zander-Grande C, Amaral
Grandini S (2010) Class V resin composite restorations using two R, Stanislawczuk R, Carvalho RM, Manso A, Reis A (2010) A
different adhesives: prospective RCT. J Dent Res. randomized clinical evaluation of a one- and two-step self-etch
39. Walter EJR, Swift LWB Jr, Reymann H, Wilder AD Jr, Sturdevan adhesive over 24 months. Oper Dent 35:265–272
J, Ritter AV, Chung Y (2013) Clinical evaluation of dental adhe- 56. Loguercio AD, Luque-Martinez IV, Fuentes S, Reis A, Muñoz
sives of different bonding strategies. J Dent Res. MA (2018) Effect of dentin roughness on the adhesive perfor-
40. Abdalla, Garcia G (2006) Clinical evaluation of self-etch adhe- mance in non-carious cervical lesions: a double-blind randomized
sives in class V non-carious lesions (vol 19, pg 291, 2006). Am clinical trial. J Dent 69:60–69. https://doi.org/10.1016/j.jdent.
J Dent 19:392–392 2017.09.011
41. Araujo JF, Barros TA, Braga EM, Loretto SC, Silva e Souza Pde 57. Matis BA, Cochran MJ, Carlson TJ, Guba C, Eckert GJ (2004) A
A, Silva e Souza MH (2013) One-year evaluation of a simplified three-year clinical evaluation of two dentin bonding agents. J Am
ethanol-wet bonding technique: a randomized clinical trial. Braz Dent Assoc 135:451–457
Dent J 24:267–272. https://doi.org/10.1590/0103- 58. Matos TP, Gutierrez MF, Hanzen TA, Malaquias P, de Paula AM,
6440201302128 de Souza JJ, Hass V, Fernandez E, Reis A, Loguercio AD (2019)
42. Atalay C, Ozgunaltay G, Yazici AR (2019) Thirty-six-month clin- 18-month clinical evaluation of a copper-containing universal ad-
ical evaluation of different adhesive strategies of a universal ad- hesive in non-carious cervical lesions: a double-blind, randomized
hesive. Clin Oral Investig 24:1569–1578. https://doi.org/10.1007/ controlled trial. J Dent 90:103219. https://doi.org/10.1016/j.jdent.
s00784-019-03052-2 2019.103219
Clin Oral Invest

59. Mena-Serrano A, Kose C, De Paula EA, Tay LY, Reis A, 75. de Paula EA, Tay LY, Kose C, Mena-Serrano A, Reis A, Perdigao
Loguercio AD, Perdigao J (2013) A new universal simplified J, Loguercio AD (2015) Randomized clinical trial of four adhesion
adhesive: 6-month clinical evaluation. J Esthet Restor Dent 25: strategies in cervical lesions: 12-month results. Int J Esthet Dent
55–69. https://doi.org/10.1111/jerd.12005 10:122–145
60. Moosavi H, Kimyai S, Forghani M, Khodadadi R (2013) The 76. Ermis RB, Van Landuyt KL, Cardoso MV, De Munck J, Van
clinical effectiveness of various adhesive systems: an 18-month Meerbeek B, Peumans M (2012) Clinical effectiveness of a one-
evaluation. Oper Dent 38:134–141. https://doi.org/10.2341/12- step self-etch adhesive in non-carious cervical lesions at 2 years.
110-cr Clin Oral Investig 16:889–897. https://doi.org/10.1007/s00784-
61. Mortazavi V, Samimi P, Rafizadeh M, Kazemi S (2012) A ran- 011-0565-4
domized clinical trial evaluating the success rate of ethanol wet 77. Hafer M, Jentsch H, Haak R, Schneider H (2015) A three-year
bonding technique and two adhesives. Dent Res J (Isfahan) 9:588– clinical evaluation of a one-step self-etch and a two-step etch-and-
594 rinse adhesive in non-carious cervical lesions. J Dent 43:350–361.
62. Pena CE, Rodrigues JA, Ely C, Giannini M, Reis AF (2016) Two- https://doi.org/10.1016/j.jdent.2014.12.009
year randomized clinical trial of self-etching adhesives and selec- 78. Jang JH, Kim HY, Shin SM, Lee CO, Kim DS, Choi KK, Kim SY
tive enamel etching. Oper Dent 41:249–257. https://doi.org/10. (2017) Clinical effectiveness of different polishing systems and
2341/15-130-c self-etch adhesives in class V composite resin restorations: two-
63. Reis A, Leite TM, Matte K, Michels R, Amaral RC, Geraldeli S, year randomized controlled clinical trial. Oper Dent 42:19–29.
Loguercio AD (2009) Improving clinical retention of one-step https://doi.org/10.2341/16-104-c
self-etching adhesive systems with an additional hydrophobic ad- 79. Kubo S, Kawasaki K, Yokota H, Hayashi Y (2006) Five-year
hesive layer. J Am Dent Assoc 140:877–885 clinical evaluation of two adhesive systems in non-carious cervical
64. Reis A, Manica D, Ferneda F, Amaral R, Stanislawczuk R, Manso lesions. J Dent 34:97–105. https://doi.org/10.1016/j.jdent.2005.
A, De Carvalho RM, Loguercio AD (2010) A 24-month random- 04.003
ized clinical trial of a two- and three-step etch-and-rinse technique. 80. Oz FD, Ergin E, Canatan S (2019) Twenty-four-month clinical
Am J Dent 23:231–236 performance of different universal adhesives in etch-and-rinse,
65. Sartori N, Peruchi LD, Guimaraes JC, Silva SB, Monteiro S Jr, selective etching and self-etch application modes in NCCL - a
Baratieri LN, Belli R (2013) Clinical effectiveness of a hydropho- randomized controlled clinical trial. J Appl Oral Sci 27:
bic coating used in conjunction with a one-step self-etch adhesive: e20180358. https://doi.org/10.1590/1678-7757-2018-0358
an 18-month evaluation. Oper Dent 38:249–257. https://doi.org/ 81. Perdigão J, Dutra-Corrêa M, Saraceni CH, Ciaramicoli MT, Kiyan
10.2341/12-014-c VH, Queiroz CS (2012) Randomized clinical trial of four adhesion
strategies: 18-month results. Oper Dent 37:3–11. https://doi.org/
66. Stojanac IL, Premovic MT, Ramic BD, Drobac MR, Stojsin IM,
10.2341/11-222-C
Petrovic LM (2013) Noncarious cervical lesions restored with
82. Perdigao J, Ceballos L, Giraldez I, Baracco B, Fuentes MV (2020)
three different tooth-colored materials: two-year results. Oper
Effect of a hydrophobic bonding resin on the 36-month perfor-
Dent 38:12–20. https://doi.org/10.2341/12-046-c
mance of a universal adhesive-a randomized clinical trial. Clin
67. Tian FC, Wang XY, Gao XJ (2014) Clinical evaluation of a two-
Oral Investig 24:765–776. https://doi.org/10.1007/s00784-019-
step etch-and-rinse adhesive and a one-step self-etch adhesive in
02940-x
non-carious cervical lesion. Beijing da xue xue bao [Journal of
83. Qin W, Song Z, Ye YY, Lin ZM (2013) Two-year clinical eval-
Peking University Health sciences] 46:58–61
uation of composite resins in non-carious cervical lesions. Clin
68. Türkün SL (2003) Clinical evaluation of a self-etching and a one- Oral Investig 17:799–804. https://doi.org/10.1007/s00784-012-
bottle adhesive system at two years. J Dent 31:527–534. https:// 0780-7
doi.org/10.1016/S0300-5712(03)00107-6 84. Ritter AV, Heymann HO, Swift EJ Jr, Sturdevant JR, Wilder AD
69. Yaman BC, Dogruer I, Gumustas B, Efes BG (2014) Three-year Jr (2008) Clinical evaluation of an all-in-one adhesive in non-
randomized clinical evaluation of a low-shrinkage silorane-based carious cervical lesions with different degrees of dentin sclerosis.
resin composite in non-carious cervical lesions. Clin Oral Investig Oper Dent 33:370–378. https://doi.org/10.2341/07-128
18:1071–1079. https://doi.org/10.1007/s00784-013-1079-z 85. Ritter HHA, Swift E, Sturdevant J, Boushell L, Lambert C (2015)
70. Zanatta RF, Silva TM, Esper M, Bresciani E, Goncalves S, Eighteen-month clinical evaluation of a two-component self-etch
Caneppele T (2019) Bonding performance of simplified adhesive adhesive. J Adhes Dent 94(A)
systems in noncarious cervical lesions at 2-year follow-up: a 86. Ruschel VC, Shibata S, Stolf SC, Chung Y, Baratieri LN,
double-blind randomized clinical trial. Oper Dent 44(5):476– Heymann HO, Walter R (2018) Eighteen-month clinical study
487. https://doi.org/10.2341/18-049- of universal adhesives in noncarious cervical lesions. Oper Dent
71. Abdalla AI, Garcia-Godoy F (2007) Clinical performance of a 43:241–249. https://doi.org/10.2341/16-320-c
self-etch adhesive in class V restorations made with and without 87. Sartori N, Pacheco E, Souza PTDR, Lopes GC, Peruchi LD
acid etching. J Dent 35:558–563. https://doi.org/10.1016/j.jdent. (2011) Performance clínica de dois sistemas adesivos em lesões
2007.02.006 cervicais não cariosas: 30 meses de acompanhamento. Arq Cent
72. Boushell LW, Heymann HO, Ritter AV, Sturdevant JR, Swift EJ Estud Curso Odontol Univ Fed Minas Gerais 47:119–126
Jr, Wilder AD Jr, Chung Y, Lambert CA, Walter R (2016) Six- 88. Scotti N, Comba A, Gambino A, Manzon E, Breschi L, Paolino D,
year clinical performance of etch-and-rinse and self-etch adhe- Pasqualini D, Berutti E (2016) Influence of operator experience on
sives. Dent Mater 32:1065–1072. https://doi.org/10.1016/j. non-carious cervical lesion restorations: clinical evaluation with
dental.2016.06.003 different adhesive systems. Am J Dent 29:33–38
73. Burrow MF, Tyas MJ (2007) Clinical evaluation of three adhesive 89. Tuncer D, Yazici AR, Ozgunaltay G, Dayangac B (2013) Clinical
systems for the restoration of non-carious cervical lesions. Oper evaluation of different adhesives used in the restoration of non-
Dent 32:11–15. https://doi.org/10.2341/06-50 carious cervical lesions: 24-month results. Aust Dent J 58:94–100.
74. Dalkilic EE, Omurlu H (2012) Two-year clinical evaluation of https://doi.org/10.1111/adj.12028
three adhesive systems in non-carious cervical lesions. J Appl 90. Turkun LS (2005) The clinical performance of one- and two-step
Oral Sci 20:192–199. https://doi.org/10.1590/s1678- self-etching adhesive systems at one year. J Am Dent Assoc 136:
77572012000200012 656–664; quiz 683. https://doi.org/10.2341/11-397-C
 Clin Oral Invest

91. van Dijken JW (2000) Clinical evaluation of three adhesive sys- restorations: a systematic review and meta-analysis. J Dent 62:36–
tems in class V non-carious lesions. Dent Mater 16:285–291 53. https://doi.org/10.1016/j.jdent.2017.05.006
92. Van Dijken JWV (2004) Durability of three simplified adhesive 102. Peumans M, Kanumilli P, De Munck J, Van Landuyt K,
systems in class V non-carious cervical dentin lesions. Am J Dent Lambrechts P, Van Meerbeek B (2005) Clinical effectiveness of
17:27–32 contemporary adhesives: a systematic review of current clinical
93. Van Dijken JWV (2010) A prospective 8-year evaluation of a mild trials. Dent Mater 21:864–881. https://doi.org/10.1016/j.dental.
two-step self-etching adhesive and a heavily filled two-step etch- 2005.02.003
and-rinse system in non-carious cervical lesions. Dent Mater 26: 103. Li T, Dickersin K (2013) Citation of previous meta-analyses on
940–946. https://doi.org/10.1016/j.dental.2010.05.009 the same topic: a clue to perpetuation of incorrect methods?
94. van Dijken JW (2013) A randomized controlled 5-year prospec- Ophthalmology 120:1113–1119. https://doi.org/10.1016/j.
tive study of two HEMA-free adhesives, a 1-step self etching and a ophtha.2012.11.038
3-step etch-and-rinse, in non-carious cervical lesions. Dent Mater 104. Schwendicke F, Gostemeyer G, Blunck U, Paris S, Hsu LY, Tu
29:e271–e280. https://doi.org/10.1016/j.dental.2013.08.203 YK (2016) Directly placed restorative materials: review and net-
95. Van Landuyt KL, Peumans M, Fieuws S, De Munck J, Cardoso work meta-analysis. J Dent Res 95:613–622. https://doi.org/10.
MV, Ermis RB, Lambrechts P, Van Meerbeek B (2008) A ran- 1177/0022034516631285
domized controlled clinical trial of a HEMA-free all-in-one adhe- 105. De Munck J, Mine A, Poitevin A, Van Ende A, Cardoso MV, Van
sive in non-carious cervical lesions at 1 year. J Dent 36:847–855. Landuyt KL, Peumans M, Van Meerbeek B (2012) Meta-
https://doi.org/10.1016/j.jdent.2008.06.005 analytical review of parameters involved in dentin bonding. J
96. Zhou Z, Yu S, Jiang Y, Lin Y, Xiong Y, Ni L (2009) A random- Dent Res 91:351–357. https://doi.org/10.1177/
ized, controlled clinical trial of one-step self-etching adhesive sys- 0022034511431251
tems in non-carious cervical lesions. Am J Dent 22:235–240
106. Loguercio AD, Luque-Martinez I, Lisboa AH, Higashi C, Oliveira
97. de Oliveira FG, Machado LS, Sundfeld-Neto D, Giannini M,
Queiroz VA, Rego RO, Reis A (2015) Influence of isolation meth-
Briso AL, Dos Santos PH, Sundefeld ML, Sundfeld RH (2017)
od of the operative field on gingival damage, patients preference,
Two-year clinical evaluation of a nanofilled etch-and-rinse and a
and restoration retention in noncarious cervical lesions. Oper Dent
self-etch adhesive system containing MDPB and fluoride in non-
40:581–593. https://doi.org/10.2341/14-089-C
carious cervical lesions. Compend Contin Educ Dent 38:e1–e4
98. Reis A, Loguercio AD (2009) A 36-month clinical evaluation of 107. Schroeder M, Reis A, Luque-Martinez I, Loguercio AD,
ethanol/water and acetone-based etch-and-rinse adhesives in non- Masterson D, Maia LC (2015) Effect of enamel bevel on retention
carious cervical lesions. Oper Dent 34:384–391. https://doi.org/ of cervical composite resin restorations: a systematic review and
10.2341/08-117 meta-analysis. J Dent 43:777–788. https://doi.org/10.1016/j.jdent.
99. Krithikadatta J (2010) Clinical effectiveness of contemporary den- 2015.02.017
tin bonding agents. JCD 13:173–183. https://doi.org/10.4103/ 108. Wang Y, Li C, Fau-Yuan H, Yuan H, Fau-Wong MC, Wong MC,
0972-0707.73376 Fau-Zou J, Zou J, Fau-Shi Z, Shi Z, Fau-Zhou X, Zhou X Rubber
100. Santos MJMC, Ari N, Steele S, Costella J, Banting D (2014) dam isolation for restorative treatment in dental patients. 20(9):
Retention of tooth-colored restorations in non-carious cervical CD009858. https://doi.org/10.1002/14651858.CD009858
lesions-a systematic review. Clin Oral Investig 18:1369–1381.
https://doi.org/10.1007/s00784-014-1220-7 Publisher’s note Springer Nature remains neutral with regard to jurisdic-
101. Schroeder M, Correa IC, Bauer J, Loguercio AD, Reis A (2017) tional claims in published maps and institutional affiliations.
Influence of adhesive strategy on clinical parameters in cervical