This systematic review followed the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines [14] (Appendix 1) and was registered in the PROSPERO platform (CRD42021238063).
All this paper was conducted by two independent and duplicate reviewers and in case of doubt or disagreement, a third experienced reviewer performed the tiebreaker.
Search Strategy
The literature search was conducted using Medline (PubMed), Scopus, Web of Science, and Embase databases to identify relevant published papers. In addition to these, we performed a systematic search in OpenSigle (OpenGray) and in ProQuest for grey literature. There were no restrictions regarding language or year of publication.
We developed a search strategy to find relevant studies regarding our PICO question (Participants, Intervention, Comparison, Outcome), namely P: Children and adults presenting with restorative failure, I: Repair, C: Replacement, O: Restoration survival. Using a controlled vocabulary (Mesh terms), synonyms and related terms were selected and associated with Boolean operators, resulting in the search strategies for each database:
PubMed
|
((Success OR Survival OR Longevity) AND (Repair* OR refurbish* OR seal* OR Replacement) AND (amalgam OR "composite resin*" OR "resin composite*" OR compomer* OR "polyacid modified composite resin*" OR "polyacid-modified composite resin*" OR "dental restoration*" OR "glass ionomer cement*" OR "glass-ionomer cement*" OR GIC) AND (dental OR Dentistry OR teeth OR tooth OR molar))
|
Scopus
|
TITLE-ABS-KEY ( ( success OR survival OR longevity ) AND ( repair* OR refurbish* OR seal* OR replacement ) AND ( amalgam OR "composite resin*" OR "resin composite*" OR compomer* OR "polyacid modified composite resin*" OR "polyacid-modified composite resin*" OR "dental restoration*" OR "glass ionomer cement*" OR "glass-ionomer cement*" OR gic ) AND ( dental OR dentistry OR teeth OR tooth OR molar ) )
|
Web of Science
|
TS=((Success OR Survival OR Longevity) AND (Repair* OR refurbish* OR seal* OR Replacement) AND (amalgam OR "composite resin*" OR "resin composite*" OR compomer* OR "polyacid modified composite resin*" OR "polyacid-modified composite resin*" OR "dental restoration*" OR "glass ionomer cement*" OR "glass-ionomer cement*" OR GIC) AND (dental OR Dentistry OR teeth OR tooth OR molar) )
|
Embase
|
('success'/exp OR success OR 'survival'/exp OR survival OR 'longevity'/exp OR longevity) AND (repair* OR refurbish* OR seal* OR 'replacement'/exp OR replacement) AND ('amalgam'/exp OR amalgam OR 'composite resin*' OR 'resin composite*' OR compomer* OR 'polyacid modified composite resin*' OR 'polyacid-modified composite resin*' OR 'dental restoration*' OR 'glass ionomer cement*' OR 'glass-ionomer cement*' OR gic) AND ('dental'/exp OR dental OR 'dentistry'/exp OR dentistry OR 'teeth'/exp OR teeth OR 'tooth'/exp OR tooth OR 'molar'/exp OR molar)
|
OpenSigle
|
((Success OR Survival OR Longevity) AND (Repair* OR refurbish* OR seal* OR Replacement) AND (amalgam OR "composite resin*" OR "resin composite*" OR compomer* OR "polyacid modified composite resin*" OR "polyacid-modified composite resin*" OR "dental restoration*" OR "glass ionomer cement*" OR "glass-ionomer cement*" OR GIC) AND (dental OR Dentistry OR teeth OR tooth OR molar))
|
ProQuest
|
((Success OR Survival OR Longevity) AND (Repair* OR refurbish* OR seal* OR Replacement) AND (amalgam OR "composite resin*" OR "resin composite*" OR compomer* OR "polyacid modified composite resin*" OR "polyacid-modified composite resin*" OR "dental restoration*" OR "glass ionomer cement*" OR "glass-ionomer cement*" OR GIC) AND (dental OR Dentistry OR teeth OR tooth OR molar))
|
The results from the different databases were cross-checked using the EndNote reference manager program and eliminated manually to eliminate possible duplicates.
Study Selection and Eligibility Criteria
Using the search strategy, initially, potentially relevant publications were screened for inclusion based on title and abstract using the inclusion criteria: (1) Repair or replacement studies, (2) success, longevity, or survival data, (3) randomized clinical controlled trials. For the exclusion criteria (1) Loss to follow-up over 30%, (2) less than 12 months follow up, (3) anterior teeth.
Two independent reviewers participated in all phases of study screening and eligibility (Kappa = 0.87). All disagreements were resolved by discussion, and a third reviewer was consulted when necessary. When studies had the same sample, we considered the longer follow-up or the most complete data set.
Data Extraction
For each selected paper, the following information was systematically extracted: 1) Publication details such as authors, year and country of publication; 2) Study design; 3) Technique used (Repair or Replacement); 4) Sample characteristics such as number and age of participants, tooth type, number of faces involved and cavity type; 5) Material used; 6) Number and experience of operators; 7) Setting; 8) Outcome information such as evaluation criteria, survival percentage and follow-up time. In case of inconclusive or missing data, we contacted the authors through the e-mail address.
Risk of Bias Assessment
The RoB 2 tool assessed the risk of bias in randomized clinical trials [15], which approaches a fixed set of bias domains, focusing on different aspects of study conduct, reporting, and design. Each potential reason for bias can be rated as a 'high' or 'low' risk of bias or "having some concerns”.
Certainty of the evidence
Two reviewers assessed the certainty of the evidence using the GRADE tool (Grading of Recommendations, Assessment, Development, and Evaluation) in duplicate. The assessment was based on the risk of bias domains, inconsistency, indirectness, and publication bias, resulting in a high, moderate, low, or very low certainty of evidence. [16]
Data synthesis and statistical methods for the meta-analysis
A narrative synthesis was conducted to firstly describe the details of the studies, participant and intervention characteristics, and the results of the included articles. Then a meta-analysis was performed using fixed or random-effects models considering the failure of restoration as the outcome.
Heterogeneity between studies was statistically evaluated using the Chi-square test (Q test) and its inconsistency (I2). Values of I2 were classified as: not relevant heterogeneity (0–40%), moderate heterogeneity (30–60%), substantial heterogeneity (50–90%), and considered homogeneity (75–100%).
All analyses were performed using meta and metafor packages in the RStudio, version 1.3.95 (PBC, Boston, USA). The risk ratio and 95% confidence interval for pairwise comparison were calculated.