Clinical Effectiveness of Restorative Materials for the Restoration of Carious Primary Teeth: An Umbrella Review

Since untreated dental caries remain a worldwide burden, this umbrella review aimed to assess the quality of evidence on the clinical effectiveness of different restorative materials for the treatment of carious primary teeth. A literature search in electronic bibliographic databases was performed to find systematic reviews with at least two-arm comparisons between restorative materials and a follow-up period ≥12 months. Reviews retrieved were screened; those eligible were selected, and the degree of overlap was calculated using the ‘corrected covered area’ (CCA). Data were extracted and the risk of bias was assessed using the ROBIS tool. Fourteen systematic reviews with a moderate overlap (6% CCA) were included. All materials studied performed similarly and were equally efficient for the restoration of carious primary teeth. Amalgam and resin composite had the lowest mean failure rate at 24 months while high-viscosity and metal-reinforced glass ionomer cements had the highest. At 36 months, high-viscosity glass ionomer cements showed the highest failure rate with compomer showing the lowest. Most reviews had an unclear risk of bias. Within the limitations of the review, all materials have acceptable mean failure rates and could be recommended for the restoration of carious primary teeth.


Introduction
Despite the notable decline in caries over recent years [1], millions of untreated carious primary teeth still constitute a global burden and a major health care challenge [2,3]. There are several treatment options for carious primary teeth with increasing invasiveness ranging from minimum intervention dentistry to restorative approaches [4,5]. Conservative caries management strategies aim to sustain hard dental tissue and to prolong the cycle of dental re-restoration, both of which increase the tooth's long-term preservation in the oral cavity [6]. However, it is recommended to restore carious lesions presenting with cavitated surfaces that cannot be cleaned properly or sealed in order to reduce the lesion's caries activity [6].
When it comes to the restoration of primary teeth, several factors may influence longevity, such as the operator's experience, child's age, type of tooth, cavity size, or

•
Comparing different dental materials for the restoration of carious lesions in primary teeth (the use of local anaesthesia and/or rubber dam isolation was not a limitation); • With a clear definition of lesion location, size, and depth; • With asymptomatic teeth (no history of pain, pulp exposure, infection, swelling, or evidence of periapical pathology); • With at least two-arm comparisons; • With a follow-up period of at least 12 months.

Exclusion Criteria
Reviews reporting: treatment of caries in permanent teeth (including those where results on primary dentition were not reported separately), with a single cohort of patients, management of superficial caries (i.e., radiographic lesions R1), and with a follow-up period <12 months were excluded.

PICO Format for the Included Studies
• Participants: Children up to the age of 12 years with deep carious lesions in their primary dentition, with no restrictions on participants' demographic characteristics. • Intervention: Any dental material placed as restoration in primary teeth: amalgam (A), resin composite (RC), compomer (CO), GIC, RMGIC, metal-reinforced GIC (MRGIC), HVGIC, and stainless steel crowns (SSC).

•
Comparator group: Any of the above restorative materials. • Outcomes: Primary outcomes were treatment failure/success (criteria used were not a limitation) and restoration quality (surface roughness, colour match, marginal integrity, tooth integrity, filling integrity, proximal contact, change of sensitivity, hypersensitivity, and radiographic assessment) [22].
Secondary outcomes were time until restoration failure occurs or re-treatment is needed, discomfort during restorative treatment or within 24 h after treatment, patient's and/or carer's perceptions of the restorative treatment, and the impact of the tooth or technique-related factors {e.g., surface(s) affected (single-or multi-surface lesions), the technique of carious removal (selective vs. complete), type of tooth (anterior/posterior tooth), isolation technique, and the type of adhesive (self-etch, etch-and-rinse, and universal)} on the clinical effectiveness of the restorative materials.

Search Strategy
A literature search was conducted (7 October 2021) in electronic databases (Embase, Web of Science, Dentistry & Oral Sciences Source (DOSS), Medline/PubMed, Scopus, LILACS, and The Cochrane Library (Cochrane Database of Systematic Reviews, Cochrane Methodology Register) considering the differences in controlled vocabulary and syntax rules. Reference lists of retrieved articles were hand-searched to identify additional studies not identified through electronic searches. No language or publication year restrictions were applied.

Data Collection and Analysis
The titles and/or abstracts of all studies retrieved from the search, and those from additional sources, were screened independently by two review authors (KS and SA) to identify those that potentially met the inclusion criteria. Full texts of the potentially included studies were reviewed and the list of included studies was finalised. Any discrepancies and disagreements were resolved through discussion with a third author (SG). All reviewers were calibrated prior to the initiation of the study, with intra-and inter-examiner reliability values being excellent, exceeding 0.85 for all cases.
Data were extracted independently and in duplicate by two review authors (KS and SA). For each review, the following data were recorded: a. publication details (authors and year of publication), b. review methodology (search strategy, objectives, number of included studies, and study design), c. review characteristics (interventions, control groups, and quantitative synthesis), and d. outcomes (main results and conclusions) including methods of assessment and quality assessment (risk of bias assessment tool used and method of grading the quality).

Overlap of Studies
Within the 14 reviews, 179 studies altogether were included without considering overlap (Supplementary Table S3). CCA was calculated at 0.06 (6%), with 'N' being equal to 179, 'r' equal to 101, and 'c' equal to 14. Overlap was moderate, with a moderate number of studies appearing several times within several reviews, slightly increasing the weighing of the results.

Study Characteristics
Literature searches in most reviews were performed within MEDLINE, Science-Direct, Web of Science, Google Scholar, Cochrane, and Embase databases, with two reviews [20,26] restricted to MEDLINE and EMBASE and another two [27,28] solely to MEDLINE. Two reviews also explored grey literature and dissertations [29,30] and four checked the National Institutes of Health and the World Health Organization International Clinical Trials Registry for ongoing trials [29][30][31][32]. A timeframe was defined from 1966 to 2021, with one review narrowing it between 1996 and 2017 [8], one between 2012 and 2016 [20], and one reporting a 10-year time frame [32]. Most studies placed no language restrictions, apart from three reviews [28,32,33] that included studies only published in English.

Records identified through database searching n=175
Screening Included Eligibility

Identification
Additional records identified through other sources/hand search n=54 Records after duplicates removed n=171 Records screened n=225 Records excluded n=117 Full-text articles assessed for eligibility Initially accepted n=11 Considered as maybe n=62 Hand-search n=35 TOTAL: n=108 Full-text articles excluded, with reasons Wrong outcome n=16 Insufficient follow-up n=6 In vitro studies n=1 Wrong publication type n=52 Permanent teeth n=10 Not retrieved n=9 TOTAL: n=94 Manuscripts included in qualitative synthesis n=14 Nine reviews included only randomized control trials (RCTs), three included both RCTs and non-RCTs [8,27,32], one included both RCTs and an observational study [34] and one included nine RCTs, one longitudinal study and one study with no information regarding its design [28] (Table 1). From the available data, 80 primary studies were of split-mouth design, and 70 of parallel-arm design. One review included initial studies with a partial split-mouth design [35], where the unit of randomisation was the tooth, not the patient. Interventions in the majority of the reviews involved conventional materials used for restoring single-or multi-surface carious lesions. Two reviews reported on stainless steel crowns [8,29], six reviews reported on atraumatic restorative treatment (ART) using mainly HVGIC [20,27,28,33,36], and one reported on different adhesive systems [32]. 17 RCTs (11 split-mouth and 5 parallel-arm) A CO, RC, and GIC Failure rate ART (n = 1); USPHS (n = 10); ART/USPHS (n = 1); DPDHS (n = 1)

24-96 Not performed
Mickenautsch et al., 2009 [38] 3 RCTs (2 split-mouth and 1 parallel-arm) ART using GIC A Longevity (dichotomous success/failure rates) ART (n = 3) 12-36 Two studies included Mickenautsch et al., 2011 [35] 6 RCTs (2 split-mouth, 1 parallel-arm, and 3 partial split-mouth)  14 RCTs (5 split-mouth and 9 parallel-arm); 1 observational study CRT, ART, and HVGIC Between each other Success rate and caries lesion arrestment ART (n = 4); ART and USPHS (n = 1); USPHS (n = 1); criteria by Innes  Outcomes reported were success rate, failure rate, annual failure rate, major failure, clinical performance, and caries arrest and/or progression at a follow-up period that varied from 12 months to 84 months. Most reviews used a combination of widely used criteria including FDI, modified USPHS, and ART. Four reviews included studies that used the author's own criteria [8,27,30,34], and two reviews did not report on the criteria used [29,37]. Meta-analysis was performed in most reviews except for four [26,27,32,36], with the number of initial studies included in each meta-analysis varying from 2 to 17.

Risk of Bias Assessment
Most reviews used the Cochrane collaboration tool for assessing the risk of bias, three reviews used criteria proposed by other investigators [26,35,36], and one did not include quality assessment for the included studies [28]. Overall, four reviews were at low risk of bias [8,29,32,33], three were at high risk [26,28,36], and the remaining had an unclear risk of bias ( Table 2). Regarding study eligibility, most reviews were at low risk except for four [26,27,30,36], which restricted search criteria, reducing the comprehensiveness and immediately increasing the bias by possibly leaving out eligible reviews.

Risk of Bias Assessment
Most reviews used the Cochrane collaboration tool for assessing the risk of bias, three reviews used criteria proposed by other investigators [26,35,36], and one did not include quality assessment for the included studies [28]. Overall, four reviews were at low risk of bias [8,29,32,33], three were at high risk [26,28,36], and the remaining had an unclear risk of bias (Table 2). Regarding study eligibility, most reviews were at low risk except for four [26,27,30,36], which restricted search criteria, reducing the comprehensiveness and immediately increasing the bias by possibly leaving out eligible reviews.
In three reviews, concerns were raised regarding identification and selection, with the search strategy not being reproducible and the inclusion and exclusion criteria not being clearly defined [27,28,36]. Data collection and study appraisal were adequate in most reviews, except for four in which clinical and methodological heterogeneity of the included studies was either not performed adequately or was absent. Regarding synthesis of the findings, in most reviews, concerns were raised mainly due to the fact that metaanalysis was performed in studies of unclear or high risk of bias. Search strategy not reproducible; inclusion criteria not clearly mentioned; outcomes not predefined; no details given on data collection and appraisal; no clear data synthesis; and risk of bias tool not validated Innes et al., 2015 [29] Cochrane Collaboration Tool - Kilpatrick et al., 2007 [26] Criteria by Curzon and Toumba, 2006 Restricted data search; language restrictions; risk of bias tool not validated; no evaluation of clinical and methodological heterogeneity; and limited data on data collection and synthesis Mickenautsch et al., 2010 [33] Cochrane Collaboration tool Language restrictions Mickenautsch et al., 2011 [35] Based on the availability of evidence indicating successful prevention of selection and detection/performance bias from start to the end of each trial Risk of bias tool not validated; meta-analysis performed using studies of unclear or high risk of bias Pires et al., 2018 [31] Cochrane Collaboration Tool Meta-analysis performed using studies of unclear or high risk of bias Ruengrungsom et al., 2018 [27] Cochrane Collaboration Tool Restricted search strategy and data search; language restrictions; no details given on data collection and appraisal; primary studies with clinical and methodological heterogeneity; a majority of studies with an unclear risk of bias; and no meta-analysis performed

Risk of Bias Assessment
Most reviews used the Cochrane collaboration tool for assessing the risk of bias, three reviews used criteria proposed by other investigators [26,35,36], and one did not include quality assessment for the included studies [28]. Overall, four reviews were at low risk of bias [8,29,32,33], three were at high risk [26,28,36], and the remaining had an unclear risk of bias (Table 2). Regarding study eligibility, most reviews were at low risk except for four [26,27,30,36], which restricted search criteria, reducing the comprehensiveness and immediately increasing the bias by possibly leaving out eligible reviews.
In three reviews, concerns were raised regarding identification and selection, with the search strategy not being reproducible and the inclusion and exclusion criteria not being clearly defined [27,28,36]. Data collection and study appraisal were adequate in most reviews, except for four in which clinical and methodological heterogeneity of the included studies was either not performed adequately or was absent. Regarding synthesis of the findings, in most reviews, concerns were raised mainly due to the fact that metaanalysis was performed in studies of unclear or high risk of bias.

Risk of Bias Assessment
Most reviews used the Cochrane collaboration tool for assessing the risk of bias, thr reviews used criteria proposed by other investigators [26,35,36], and one did not includ quality assessment for the included studies [28]. Overall, four reviews were at low risk bias [8,29,32,33], three were at high risk [26,28,36], and the remaining had an unclear ri of bias (Table 2). Regarding study eligibility, most reviews were at low risk except for fou [26,27,30,36], which restricted search criteria, reducing the comprehensiveness and imm diately increasing the bias by possibly leaving out eligible reviews.
In three reviews, concerns were raised regarding identification and selection, wi the search strategy not being reproducible and the inclusion and exclusion criteria n being clearly defined [27,28,36]. Data collection and study appraisal were adequate most reviews, except for four in which clinical and methodological heterogeneity of th included studies was either not performed adequately or was absent. Regarding synthes of the findings, in most reviews, concerns were raised mainly due to the fact that met analysis was performed in studies of unclear or high risk of bias.  Table 3 summarises the main conclusions of the included systematic reviews and Table 4 summarises the results of the meta-analysis performed. Overall, all materials performed similarly and were equally efficient for the restoration of primary teeth with deep caries. Amalgam and resin composite had the lowest mean failure rate at 24 months and RMGIC and MRGIC were the highest ( Figure 2), respectively, at 36 months ( Figure 3). HVGIC showed the highest failure rate with compomer showing the lowest.   Table 3 summarises the main conclusions of the included systematic reviews and ble 4 summarises the results of the meta-analysis performed. Overall, all materials p formed similarly and were equally efficient for the restoration of primary teeth with d caries. Amalgam and resin composite had the lowest mean failure rate at 24 months RMGIC and MRGIC were the highest ( Figure 2), respectively, at 36 months ( Figure  HVGIC showed the highest failure rate with compomer showing the lowest.

Risk of Bias Assessment
Most reviews used the Cochrane collaboration tool for assessing the risk of bias, three reviews used criteria proposed by other investigators [26,35,36], and one did not include quality assessment for the included studies [28]. Overall, four reviews were at low risk of bias [8,29,32,33], three were at high risk [26,28,36], and the remaining had an unclear risk of bias (Table 2). Regarding study eligibility, most reviews were at low risk except for four [26,27,30,36], which restricted search criteria, reducing the comprehensiveness and immediately increasing the bias by possibly leaving out eligible reviews.
In three reviews, concerns were raised regarding identification and selection, with the search strategy not being reproducible and the inclusion and exclusion criteria not being clearly defined [27,28,36]. Data collection and study appraisal were adequate in most reviews, except for four in which clinical and methodological heterogeneity of the included studies was either not performed adequately or was absent. Regarding synthesis of the findings, in most reviews, concerns were raised mainly due to the fact that metaanalysis was performed in studies of unclear or high risk of bias.

Risk of Bias Assessment
Most reviews used the Cochrane collaboration tool for assessing the risk of bias, three reviews used criteria proposed by other investigators [26,35,36], and one did not include quality assessment for the included studies [28]. Overall, four reviews were at low risk of bias [8,29,32,33], three were at high risk [26,28,36], and the remaining had an unclear risk of bias (Table 2). Regarding study eligibility, most reviews were at low risk except for four [26,27,30,36], which restricted search criteria, reducing the comprehensiveness and immediately increasing the bias by possibly leaving out eligible reviews.
In three reviews, concerns were raised regarding identification and selection, with the search strategy not being reproducible and the inclusion and exclusion criteria not being clearly defined [27,28,36]. Data collection and study appraisal were adequate in most reviews, except for four in which clinical and methodological heterogeneity of the included studies was either not performed adequately or was absent. Regarding synthesis of the findings, in most reviews, concerns were raised mainly due to the fact that metaanalysis was performed in studies of unclear or high risk of bias.

Risk of Bias Assessment
Most reviews used the Cochrane collaboration tool for assessing the risk of bias, thr reviews used criteria proposed by other investigators [26,35,36], and one did not includ quality assessment for the included studies [28]. Overall, four reviews were at low risk bias [8,29,32,33], three were at high risk [26,28,36], and the remaining had an unclear ri of bias (Table 2). Regarding study eligibility, most reviews were at low risk except for fou [26,27,30,36], which restricted search criteria, reducing the comprehensiveness and imm diately increasing the bias by possibly leaving out eligible reviews.
In three reviews, concerns were raised regarding identification and selection, wi the search strategy not being reproducible and the inclusion and exclusion criteria n being clearly defined [27,28,36]. Data collection and study appraisal were adequate most reviews, except for four in which clinical and methodological heterogeneity of th included studies was either not performed adequately or was absent. Regarding synthes of the findings, in most reviews, concerns were raised mainly due to the fact that met analysis was performed in studies of unclear or high risk of bias.

Risk of Bias Assessment
Most reviews used the Cochrane collaboration tool for assessing the risk of reviews used criteria proposed by other investigators [26,35,36], and one did n quality assessment for the included studies [28]. Overall, four reviews were at l bias [8,29,32,33], three were at high risk [26,28,36], and the remaining had an u of bias (Table 2). Regarding study eligibility, most reviews were at low risk exce [26,27,30,36], which restricted search criteria, reducing the comprehensiveness a diately increasing the bias by possibly leaving out eligible reviews.
In three reviews, concerns were raised regarding identification and selec the search strategy not being reproducible and the inclusion and exclusion c being clearly defined [27,28,36]. Data collection and study appraisal were ad most reviews, except for four in which clinical and methodological heterogen included studies was either not performed adequately or was absent. Regarding of the findings, in most reviews, concerns were raised mainly due to the fact analysis was performed in studies of unclear or high risk of bias.

Risk of Bias Assessment
Most reviews used the Cochrane collaboration tool for assessing reviews used criteria proposed by other investigators [26,35,36], and quality assessment for the included studies [28]. Overall, four review bias [8,29,32,33], three were at high risk [26,28,36], and the remainin of bias (Table 2). Regarding study eligibility, most reviews were at lo [26,27,30,36], which restricted search criteria, reducing the comprehe diately increasing the bias by possibly leaving out eligible reviews.
In three reviews, concerns were raised regarding identification the search strategy not being reproducible and the inclusion and e being clearly defined [27,28,36]. Data collection and study apprais most reviews, except for four in which clinical and methodological included studies was either not performed adequately or was absent. of the findings, in most reviews, concerns were raised mainly due analysis was performed in studies of unclear or high risk of bias.

Risk of Bias Assessment
Most reviews used the Cochrane collaboration tool for assessing the risk of bias, three reviews used criteria proposed by other investigators [26,35,36], and one did not include quality assessment for the included studies [28]. Overall, four reviews were at low risk of bias [8,29,32,33], three were at high risk [26,28,36], and the remaining had an unclear risk of bias (Table 2). Regarding study eligibility, most reviews were at low risk except for four [26,27,30,36], which restricted search criteria, reducing the comprehensiveness and immediately increasing the bias by possibly leaving out eligible reviews.
In three reviews, concerns were raised regarding identification and selection, with the search strategy not being reproducible and the inclusion and exclusion criteria not being clearly defined [27,28,36]. Data collection and study appraisal were adequate in most reviews, except for four in which clinical and methodological heterogeneity of the included studies was either not performed adequately or was absent. Regarding synthesis of the findings, in most reviews, concerns were raised mainly due to the fact that metaanalysis was performed in studies of unclear or high risk of bias.

Risk of Bias Assessment
Most reviews used the Cochrane collaboration tool for assessing the risk of bias, three reviews used criteria proposed by other investigators [26,35,36], and one did not include quality assessment for the included studies [28]. Overall, four reviews were at low risk of bias [8,29,32,33], three were at high risk [26,28,36], and the remaining had an unclear risk of bias (Table 2). Regarding study eligibility, most reviews were at low risk except for four [26,27,30,36], which restricted search criteria, reducing the comprehensiveness and immediately increasing the bias by possibly leaving out eligible reviews.
In three reviews, concerns were raised regarding identification and selection, with the search strategy not being reproducible and the inclusion and exclusion criteria not being clearly defined [27,28,36]. Data collection and study appraisal were adequate in most reviews, except for four in which clinical and methodological heterogeneity of the included studies was either not performed adequately or was absent. Regarding synthesis of the findings, in most reviews, concerns were raised mainly due to the fact that metaanalysis was performed in studies of unclear or high risk of bias.

Risk of Bias Assessment
Most reviews used the Cochrane collaboration tool for assessing the risk of bias, thr reviews used criteria proposed by other investigators [26,35,36], and one did not includ quality assessment for the included studies [28]. Overall, four reviews were at low risk bias [8,29,32,33], three were at high risk [26,28,36], and the remaining had an unclear ri of bias (Table 2). Regarding study eligibility, most reviews were at low risk except for fou [26,27,30,36], which restricted search criteria, reducing the comprehensiveness and imm diately increasing the bias by possibly leaving out eligible reviews.
In three reviews, concerns were raised regarding identification and selection, wi the search strategy not being reproducible and the inclusion and exclusion criteria n being clearly defined [27,28,36]. Data collection and study appraisal were adequate most reviews, except for four in which clinical and methodological heterogeneity of th included studies was either not performed adequately or was absent. Regarding synthes of the findings, in most reviews, concerns were raised mainly due to the fact that met analysis was performed in studies of unclear or high risk of bias.

Risk of Bias Assessment
Most reviews used the Cochrane collaboration tool for assessing the risk of reviews used criteria proposed by other investigators [26,35,36], and one did n quality assessment for the included studies [28]. Overall, four reviews were at l bias [8,29,32,33], three were at high risk [26,28,36], and the remaining had an u of bias (Table 2). Regarding study eligibility, most reviews were at low risk exce [26,27,30,36], which restricted search criteria, reducing the comprehensiveness a diately increasing the bias by possibly leaving out eligible reviews.
In three reviews, concerns were raised regarding identification and selec the search strategy not being reproducible and the inclusion and exclusion c being clearly defined [27,28,36]. Data collection and study appraisal were ad most reviews, except for four in which clinical and methodological heterogen included studies was either not performed adequately or was absent. Regarding of the findings, in most reviews, concerns were raised mainly due to the fact analysis was performed in studies of unclear or high risk of bias.

Risk of Bias Assessment
Most reviews used the Cochrane collaboration tool for assessing reviews used criteria proposed by other investigators [26,35,36], and quality assessment for the included studies [28]. Overall, four review bias [8,29,32,33], three were at high risk [26,28,36], and the remainin of bias (Table 2). Regarding study eligibility, most reviews were at lo [26,27,30,36], which restricted search criteria, reducing the comprehe diately increasing the bias by possibly leaving out eligible reviews.
In three reviews, concerns were raised regarding identification the search strategy not being reproducible and the inclusion and e being clearly defined [27,28,36]. Data collection and study apprais most reviews, except for four in which clinical and methodological included studies was either not performed adequately or was absent. of the findings, in most reviews, concerns were raised mainly due analysis was performed in studies of unclear or high risk of bias.  Table 3 summarises the main conclusions of the included systematic reviews and Table 4 summarises the results of the meta-analysis performed. Overall, all materials performed similarly and were equally efficient for the restoration of primary teeth with deep caries. Amalgam and resin composite had the lowest mean failure rate at 24 months and RMGIC and MRGIC were the highest ( Figure 2), respectively, at 36 months ( Figure 3). HVGIC showed the highest failure rate with compomer showing the lowest.

Risk of Bias Assessment
Most reviews used the Cochrane collaboration tool for assessing the risk of bias, three reviews used criteria proposed by other investigators [26,35,36], and one did not include quality assessment for the included studies [28]. Overall, four reviews were at low risk of bias [8,29,32,33], three were at high risk [26,28,36], and the remaining had an unclear risk of bias (Table 2). Regarding study eligibility, most reviews were at low risk except for four [26,27,30,36], which restricted search criteria, reducing the comprehensiveness and imme-diately increasing the bias by possibly leaving out eligible reviews.
In three reviews, concerns were raised regarding identification and selection, with the search strategy not being reproducible and the inclusion and exclusion criteria not being clearly defined [27,28,36]. Data collection and study appraisal were adequate in most reviews, except for four in which clinical and methodological heterogeneity of the included studies was either not performed adequately or was absent. Regarding synthesis of the findings, in most reviews, concerns were raised mainly due to the fact that meta-analysis was performed in studies of unclear or high risk of bias.  [35] Based on the availa-bility of evidence in-dicating successful prevention of selec-tion and detec-tion/performance bias from start to the end of each trial Risk of bias tool not validated; meta-analysis performed using studies of unclear or high risk of bias Pires et al., 2018 [31] Cochrane Collabora-tion Tool Meta-analysis performed using studies of un-clear or high risk of bias Ruengrungsom et al., 2018 [27] Cochrane Collabora-tion Tool Restricted search strategy and data search; lan-guage restrictions; no details given on data col-lection and appraisal; primary studies with clinical and methodological heterogeneity; a majority of studies with an unclear risk of bias; and no meta-analysis performed

Risk of Bias Assessment
Most reviews used the Cochrane collaboration tool for assessing the risk of bias, thr reviews used criteria proposed by other investigators [26,35,36], and one did not includ quality assessment for the included studies [28]. Overall, four reviews were at low risk bias [8,29,32,33], three were at high risk [26,28,36], and the remaining had an unclear ri of bias (Table 2). Regarding study eligibility, most reviews were at low risk except for fou [26,27,30,36], which restricted search criteria, reducing the comprehensiveness and imm diately increasing the bias by possibly leaving out eligible reviews.
In three reviews, concerns were raised regarding identification and selection, wi the search strategy not being reproducible and the inclusion and exclusion criteria n being clearly defined [27,28,36]. Data collection and study appraisal were adequate most reviews, except for four in which clinical and methodological heterogeneity of th included studies was either not performed adequately or was absent. Regarding synthes of the findings, in most reviews, concerns were raised mainly due to the fact that met analysis was performed in studies of unclear or high risk of bias.  [35] Based on the availa-bility of evidence in-dicating successful prevention of selec-tion and detec-tion/performance bias from start to the end of each trial Risk of bias tool not validated; meta-analysi performed using studies of unclear or high ri of bias Pires et al., 2018 [31] Cochrane Collabora-tion Tool Meta-analysis performed using studies of un clear or high risk of bias Ruengrungsom et al., 2018 [27] Cochrane Collabora-tion Tool Restricted search strategy and data search; la guage restrictions; no details given on data co lection and appraisal; primary studies with clinical and methodological heterogeneity; a majority of studies with an unclear risk of bia and no meta-analysis performed  Table 3 summarises the main conclusions of the included systematic reviews and Table 4 summarises the results of the meta-analysis performed. Overall, all materials performed similarly and were equally efficient for the restoration of primary teeth with deep caries. Amalgam and resin composite had the lowest mean failure rate at 24 months and RMGIC and MRGIC were the highest ( Figure 2), respectively, at 36 months ( Figure 3). HVGIC showed the highest failure rate with compomer showing the lowest.  Table 3 summarises the main conclusions of the included systematic reviews and ble 4 summarises the results of the meta-analysis performed. Overall, all materials p formed similarly and were equally efficient for the restoration of primary teeth with d caries. Amalgam and resin composite had the lowest mean failure rate at 24 months RMGIC and MRGIC were the highest ( Figure 2), respectively, at 36 months ( Figure  HVGIC Table 3 summarises the main conclusions of the included systematic reviews and Table 4 summarises the results of the meta-analysis performed. Overall, all materials performed similarly and were equally efficient for the restoration of primary teeth with deep caries. Amalgam and resin composite had the lowest mean failure rate at 24 months and RMGIC and MRGIC were the highest ( Figure 2), respectively, at 36 months ( Figure 3 Table 3 summarises the main conclusions of the included systematic reviews and Table 4 summarises the results of the meta-analysis performed. Overall, all materials performed similarly and were equally efficient for the restoration of primary teeth with deep caries. Amalgam and resin composite had the lowest mean failure rate at 24 months and RMGIC and MRGIC were the highest (Figure 2), respectively, at 36 months ( Figure 3) Table 3 summarises the main conclusions of the included systematic reviews and Table 4 summarises the results of the meta-analysis performed. Overall, all materials performed similarly and were equally efficient for the restoration of primary teeth with deep caries. Amalgam and resin composite had the lowest mean failure rate at 24 months and RMGIC and MRGIC were the highest (Figure 2), respectively, at 36 months ( Figure 3). HVGIC showed the highest failure rate with compomer showing the lowest.  Table 3 summarises the main conclusions of the included systematic reviews and Table 4 summarises the results of the meta-analysis performed. Overall, all materials performed similarly and were equally efficient for the restoration of primary teeth with deep caries. Amalgam and resin composite had the lowest mean failure rate at 24 months and RMGIC and MRGIC were the highest (Figure 2), respectively, at 36 months ( Figure 3) Table 3 summarises the main conclusions of the included systematic rev ble 4 summarises the results of the meta-analysis performed. Overall, all m formed similarly and were equally efficient for the restoration of primary tee caries. Amalgam and resin composite had the lowest mean failure rate at 24 RMGIC and MRGIC were the highest (Figure 2)

Risk of Bias Assessment
Most reviews used the Cochrane collaboration tool for assessing the risk of bias, three reviews used criteria proposed by other investigators [26,35,36], and one did not include quality assessment for the included studies [28]. Overall, four reviews were at low risk of bias [8,29,32,33], three were at high risk [26,28,36], and the remaining had an unclear risk of bias (Table 2). Regarding study eligibility, most reviews were at low risk except for four [26,27,30,36], which restricted search criteria, reducing the comprehensiveness and imme-diately increasing the bias by possibly leaving out eligible reviews.
In three reviews, concerns were raised regarding identification and selection, with the search strategy not being reproducible and the inclusion and exclusion criteria not being clearly defined [27,28,36]. Data collection and study appraisal were adequate in most reviews, except for four in which clinical and methodological heterogeneity of the included studies was either not performed adequately or was absent. Regarding synthesis of the findings, in most reviews, concerns were raised mainly due to the fact that meta-analysis was performed in studies of unclear or high risk of bias.

Risk of Bias Assessment
Most reviews used the Cochrane collaboration tool for assessing the risk of bias, three reviews used criteria proposed by other investigators [26,35,36], and one did not include quality assessment for the included studies [28]. Overall, four reviews were at low risk of bias [8,29,32,33], three were at high risk [26,28,36], and the remaining had an unclear risk of bias (Table 2). Regarding study eligibility, most reviews were at low risk except for four [26,27,30,36], which restricted search criteria, reducing the comprehensiveness and imme-diately increasing the bias by possibly leaving out eligible reviews.
In three reviews, concerns were raised regarding identification and selection, with the search strategy not being reproducible and the inclusion and exclusion criteria not being clearly defined [27,28,36]. Data collection and study appraisal were adequate in most reviews, except for four in which clinical and methodological heterogeneity of the included studies was either not performed adequately or was absent. Regarding synthesis of the findings, in most reviews, concerns were raised mainly due to the fact that meta-analysis was performed in studies of unclear or high risk of bias.

Risk of Bias Assessment
Most reviews used the Cochrane collaboration tool for assessing the risk of bias, thr reviews used criteria proposed by other investigators [26,35,36], and one did not includ quality assessment for the included studies [28]. Overall, four reviews were at low risk bias [8,29,32,33], three were at high risk [26,28,36], and the remaining had an unclear ri of bias (Table 2). Regarding study eligibility, most reviews were at low risk except for fou [26,27,30,36], which restricted search criteria, reducing the comprehensiveness and imm diately increasing the bias by possibly leaving out eligible reviews.
In three reviews, concerns were raised regarding identification and selection, wi the search strategy not being reproducible and the inclusion and exclusion criteria n being clearly defined [27,28,36]. Data collection and study appraisal were adequate most reviews, except for four in which clinical and methodological heterogeneity of th included studies was either not performed adequately or was absent. Regarding synthes of the findings, in most reviews, concerns were raised mainly due to the fact that met analysis was performed in studies of unclear or high risk of bias.  [35] Based on the availa-bility of evidence in-dicating successful prevention of selec-tion and detec-tion/performance bias from start to the end of each trial Risk of bias tool not validated; meta-analysi performed using studies of unclear or high ri of bias Pires et al., 2018 [31] Cochrane Collabora-tion Tool Meta-analysis performed using studies of un clear or high risk of bias Ruengrungsom et al., 2018 [27] Cochrane Collabora-tion Tool Restricted search strategy and data search; la guage restrictions; no details given on data co lection and appraisal; primary studies with clinical and methodological heterogeneity; a majority of studies with an unclear risk of bia and no meta-analysis performed J. Clin. Med. 2022, 11, x FOR PEER REVIEW

Risk of Bias Assessment
Most reviews used the Cochrane collaboration tool for assessing the risk of reviews used criteria proposed by other investigators [26,35,36], and one did n quality assessment for the included studies [28]. Overall, four reviews were at l bias [8,29,32,33], three were at high risk [26,28,36], and the remaining had an u of bias (Table 2). Regarding study eligibility, most reviews were at low risk exce [26,27,30,36], which restricted search criteria, reducing the comprehensiveness a diately increasing the bias by possibly leaving out eligible reviews.
In three reviews, concerns were raised regarding identification and selec the search strategy not being reproducible and the inclusion and exclusion c being clearly defined [27,28,36]. Data collection and study appraisal were ad most reviews, except for four in which clinical and methodological heterogen included studies was either not performed adequately or was absent. Regarding of the findings, in most reviews, concerns were raised mainly due to the fact analysis was performed in studies of unclear or high risk of bias.

Risk of Bias Assessment
Most reviews used the Cochrane collaboration tool for assessing reviews used criteria proposed by other investigators [26,35,36], and quality assessment for the included studies [28]. Overall, four review bias [8,29,32,33], three were at high risk [26,28,36], and the remainin of bias (Table 2). Regarding study eligibility, most reviews were at lo [26,27,30,36], which restricted search criteria, reducing the comprehe diately increasing the bias by possibly leaving out eligible reviews.
In three reviews, concerns were raised regarding identification the search strategy not being reproducible and the inclusion and e being clearly defined [27,28,36]. Data collection and study apprais most reviews, except for four in which clinical and methodological included studies was either not performed adequately or was absent. of the findings, in most reviews, concerns were raised mainly due analysis was performed in studies of unclear or high risk of bias.  Table 3 summarises the main conclusions of the included systematic reviews and Table 4 summarises the results of the meta-analysis performed. Overall, all materials performed similarly and were equally efficient for the restoration of primary teeth with deep caries. Amalgam and resin composite had the lowest mean failure rate at 24 months and RMGIC and MRGIC were the highest (Figure 2), respectively, at 36 months (Figure 3). HVGIC showed the highest failure rate with compomer showing the lowest.

Risk of Bias Assessment
Most reviews used the Cochrane collaboration tool for assessing the risk of bias, three reviews used criteria proposed by other investigators [26,35,36], and one did not include quality assessment for the included studies [28]. Overall, four reviews were at low risk of bias [8,29,32,33], three were at high risk [26,28,36], and the remaining had an unclear risk of bias (Table 2). Regarding study eligibility, most reviews were at low risk except for four [26,27,30,36], which restricted search criteria, reducing the comprehensiveness and imme-diately increasing the bias by possibly leaving out eligible reviews.
In three reviews, concerns were raised regarding identification and selection, with the search strategy not being reproducible and the inclusion and exclusion criteria not being clearly defined [27,28,36]. Data collection and study appraisal were adequate in most reviews, except for four in which clinical and methodological heterogeneity of the included studies was either not performed adequately or was absent. Regarding synthesis of the findings, in most reviews, concerns were raised mainly due to the fact that meta-analysis was performed in studies of unclear or high risk of bias.  Table 3 summarises the main conclusions of the included systematic reviews and Table 4 summarises the results of the meta-analysis performed. Overall, all materials performed similarly and were equally efficient for the restoration of primary teeth with deep caries. Amalgam and resin composite had the lowest mean failure rate at 24 months and RMGIC and MRGIC were the highest (Figure 2), respectively, at 36 months (Figure 3) Table 3 summarises the main conclusions of the included systematic reviews and Table 4 summarises the results of the meta-analysis performed. Overall, all materials performed similarly and were equally efficient for the restoration of primary teeth with deep caries. Amalgam and resin composite had the lowest mean failure rate at 24 months and RMGIC and MRGIC were the highest (Figure 2), respectively, at 36 months (Figure 3) Table 3 summarises the main conclusions of the included systematic rev ble 4 summarises the results of the meta-analysis performed. Overall, all m formed similarly and were equally efficient for the restoration of primary tee caries. Amalgam and resin composite had the lowest mean failure rate at 24 RMGIC and MRGIC were the highest (Figure 2)

Risk of Bias Assessment
Most reviews used the Cochrane collaboration tool for assessing the risk of bias, three reviews used criteria proposed by other investigators [26,35,36], and one did not include quality assessment for the included studies [28]. Overall, four reviews were at low risk of bias [8,29,32,33], three were at high risk [26,28,36], and the remaining had an unclear risk of bias (Table 2). Regarding study eligibility, most reviews were at low risk except for four [26,27,30,36], which restricted search criteria, reducing the comprehensiveness and imme-diately increasing the bias by possibly leaving out eligible reviews.
In three reviews, concerns were raised regarding identification and selection, with the search strategy not being reproducible and the inclusion and exclusion criteria not being clearly defined [27,28,36]. Data collection and study appraisal were adequate in most reviews, except for four in which clinical and methodological heterogeneity of the included studies was either not performed adequately or was absent. Regarding synthesis of the findings, in most reviews, concerns were raised mainly due to the fact that meta-analysis was performed in studies of unclear or high risk of bias.

Risk of Bias Assessment
Most reviews used the Cochrane collaboration tool for assessing the risk of bias, three reviews used criteria proposed by other investigators [26,35,36], and one did not include quality assessment for the included studies [28]. Overall, four reviews were at low risk of bias [8,29,32,33], three were at high risk [26,28,36], and the remaining had an unclear risk of bias (Table 2). Regarding study eligibility, most reviews were at low risk except for four [26,27,30,36], which restricted search criteria, reducing the comprehensiveness and imme-diately increasing the bias by possibly leaving out eligible reviews.
In three reviews, concerns were raised regarding identification and selection, with the search strategy not being reproducible and the inclusion and exclusion criteria not being clearly defined [27,28,36]. Data collection and study appraisal were adequate in most reviews, except for four in which clinical and methodological heterogeneity of the included studies was either not performed adequately or was absent. Regarding synthesis of the findings, in most reviews, concerns were raised mainly due to the fact that meta-analysis was performed in studies of unclear or high risk of bias.

Risk of Bias Assessment
Most reviews used the Cochrane collaboration tool for assessing the risk of bias, thr reviews used criteria proposed by other investigators [26,35,36], and one did not includ quality assessment for the included studies [28]. Overall, four reviews were at low risk bias [8,29,32,33], three were at high risk [26,28,36], and the remaining had an unclear ri of bias (Table 2). Regarding study eligibility, most reviews were at low risk except for fou [26,27,30,36], which restricted search criteria, reducing the comprehensiveness and imm diately increasing the bias by possibly leaving out eligible reviews.
In three reviews, concerns were raised regarding identification and selection, wi the search strategy not being reproducible and the inclusion and exclusion criteria n being clearly defined [27,28,36]. Data collection and study appraisal were adequate most reviews, except for four in which clinical and methodological heterogeneity of th included studies was either not performed adequately or was absent. Regarding synthes of the findings, in most reviews, concerns were raised mainly due to the fact that met analysis was performed in studies of unclear or high risk of bias.

Risk of Bias Assessment
Most reviews used the Cochrane collaboration tool for assessing the risk of reviews used criteria proposed by other investigators [26,35,36], and one did n quality assessment for the included studies [28]. Overall, four reviews were at l bias [8,29,32,33], three were at high risk [26,28,36], and the remaining had an u of bias (Table 2). Regarding study eligibility, most reviews were at low risk exce [26,27,30,36], which restricted search criteria, reducing the comprehensiveness a diately increasing the bias by possibly leaving out eligible reviews.
In three reviews, concerns were raised regarding identification and selec the search strategy not being reproducible and the inclusion and exclusion c being clearly defined [27,28,36]. Data collection and study appraisal were ad most reviews, except for four in which clinical and methodological heterogen included studies was either not performed adequately or was absent. Regarding of the findings, in most reviews, concerns were raised mainly due to the fact analysis was performed in studies of unclear or high risk of bias.

Risk of Bias Assessment
Most reviews used the Cochrane collaboration tool for assessing reviews used criteria proposed by other investigators [26,35,36], and quality assessment for the included studies [28]. Overall, four review bias [8,29,32,33], three were at high risk [26,28,36], and the remainin of bias (Table 2). Regarding study eligibility, most reviews were at lo [26,27,30,36], which restricted search criteria, reducing the comprehe diately increasing the bias by possibly leaving out eligible reviews.
In three reviews, concerns were raised regarding identification the search strategy not being reproducible and the inclusion and e being clearly defined [27,28,36]. Data collection and study apprais most reviews, except for four in which clinical and methodological included studies was either not performed adequately or was absent. of the findings, in most reviews, concerns were raised mainly due analysis was performed in studies of unclear or high risk of bias.

Risk of Bias Assessment
Most reviews used the Cochrane collaboration tool for assessing the risk of bias, three reviews used criteria proposed by other investigators [26,35,36], and one did not include quality assessment for the included studies [28]. Overall, four reviews were at low risk of bias [8,29,32,33], three were at high risk [26,28,36], and the remaining had an unclear risk of bias (Table 2). Regarding study eligibility, most reviews were at low risk except for four [26,27,30,36], which restricted search criteria, reducing the comprehensiveness and imme-diately increasing the bias by possibly leaving out eligible reviews.
In three reviews, concerns were raised regarding identification and selection, with the search strategy not being reproducible and the inclusion and exclusion criteria not being clearly defined [27,28,36]. Data collection and study appraisal were adequate in most reviews, except for four in which clinical and methodological heterogeneity of the included studies was either not performed adequately or was absent. Regarding synthesis of the findings, in most reviews, concerns were raised mainly due to the fact that meta-analysis was performed in studies of unclear or high risk of bias.

Risk of Bias Assessment
Most reviews used the Cochrane collaboration tool for assessing the risk of bias, three reviews used criteria proposed by other investigators [26,35,36], and one did not include quality assessment for the included studies [28]. Overall, four reviews were at low risk of bias [8,29,32,33], three were at high risk [26,28,36], and the remaining had an unclear risk of bias (Table 2). Regarding study eligibility, most reviews were at low risk except for four [26,27,30,36], which restricted search criteria, reducing the comprehensiveness and imme-diately increasing the bias by possibly leaving out eligible reviews.
In three reviews, concerns were raised regarding identification and selection, with the search strategy not being reproducible and the inclusion and exclusion criteria not being clearly defined [27,28,36]. Data collection and study appraisal were adequate in most reviews, except for four in which clinical and methodological heterogeneity of the included studies was either not performed adequately or was absent. Regarding synthesis of the findings, in most reviews, concerns were raised mainly due to the fact that meta-analysis was performed in studies of unclear or high risk of bias.

Risk of Bias Assessment
Most reviews used the Cochrane collaboration tool for assessing the risk of bias, thr reviews used criteria proposed by other investigators [26,35,36], and one did not includ quality assessment for the included studies [28]. Overall, four reviews were at low risk bias [8,29,32,33], three were at high risk [26,28,36], and the remaining had an unclear ri of bias (Table 2). Regarding study eligibility, most reviews were at low risk except for fou [26,27,30,36], which restricted search criteria, reducing the comprehensiveness and imm diately increasing the bias by possibly leaving out eligible reviews.
In three reviews, concerns were raised regarding identification and selection, wi the search strategy not being reproducible and the inclusion and exclusion criteria n being clearly defined [27,28,36]. Data collection and study appraisal were adequate most reviews, except for four in which clinical and methodological heterogeneity of th included studies was either not performed adequately or was absent. Regarding synthes of the findings, in most reviews, concerns were raised mainly due to the fact that met analysis was performed in studies of unclear or high risk of bias.  Table 3 summarises the main conclusions of the included systematic reviews and Table 4 summarises the results of the meta-analysis performed. Overall, all materials performed similarly and were equally efficient for the restoration of primary teeth with deep caries. Amalgam and resin composite had the lowest mean failure rate at 24 months and RMGIC and MRGIC were the highest (Figure 2), respectively, at 36 months (Figure 3). HVGIC showed the highest failure rate with compomer showing the lowest.  Table 3 summarises the main conclusions of the included systematic reviews and ble 4 summarises the results of the meta-analysis performed. Overall, all materials p formed similarly and were equally efficient for the restoration of primary teeth with d caries. Amalgam and resin composite had the lowest mean failure rate at 24 months RMGIC and MRGIC were the highest (Figure 2), respectively, at 36 months ( Figure  HVGIC showed the highest failure rate with compomer showing the lowest.

Risk of Bias Assessment
Most reviews used the Cochrane collaboration tool for assessing the risk of bias, three reviews used criteria proposed by other investigators [26,35,36], and one did not include quality assessment for the included studies [28]. Overall, four reviews were at low risk of bias [8,29,32,33], three were at high risk [26,28,36], and the remaining had an unclear risk of bias (Table 2). Regarding study eligibility, most reviews were at low risk except for four [26,27,30,36], which restricted search criteria, reducing the comprehensiveness and imme-diately increasing the bias by possibly leaving out eligible reviews.
In three reviews, concerns were raised regarding identification and selection, with the search strategy not being reproducible and the inclusion and exclusion criteria not being clearly defined [27,28,36]. Data collection and study appraisal were adequate in most reviews, except for four in which clinical and methodological heterogeneity of the included studies was either not performed adequately or was absent. Regarding synthesis of the findings, in most reviews, concerns were raised mainly due to the fact that meta-analysis was performed in studies of unclear or high risk of bias.

Risk of Bias Assessment
Most reviews used the Cochrane collaboration tool for assessing the risk of bias, three reviews used criteria proposed by other investigators [26,35,36], and one did not include quality assessment for the included studies [28]. Overall, four reviews were at low risk of bias [8,29,32,33], three were at high risk [26,28,36], and the remaining had an unclear risk of bias (Table 2). Regarding study eligibility, most reviews were at low risk except for four [26,27,30,36], which restricted search criteria, reducing the comprehensiveness and imme-diately increasing the bias by possibly leaving out eligible reviews.
In three reviews, concerns were raised regarding identification and selection, with the search strategy not being reproducible and the inclusion and exclusion criteria not being clearly defined [27,28,36]. Data collection and study appraisal were adequate in most reviews, except for four in which clinical and methodological heterogeneity of the included studies was either not performed adequately or was absent. Regarding synthesis of the findings, in most reviews, concerns were raised mainly due to the fact that meta-analysis was performed in studies of unclear or high risk of bias.

Risk of Bias Assessment
Most reviews used the Cochrane collaboration tool for assessing the risk of bias, thr reviews used criteria proposed by other investigators [26,35,36], and one did not includ quality assessment for the included studies [28]. Overall, four reviews were at low risk bias [8,29,32,33], three were at high risk [26,28,36], and the remaining had an unclear ri of bias (Table 2). Regarding study eligibility, most reviews were at low risk except for fou [26,27,30,36], which restricted search criteria, reducing the comprehensiveness and imm diately increasing the bias by possibly leaving out eligible reviews.
In three reviews, concerns were raised regarding identification and selection, wi the search strategy not being reproducible and the inclusion and exclusion criteria n being clearly defined [27,28,36]. Data collection and study appraisal were adequate most reviews, except for four in which clinical and methodological heterogeneity of th included studies was either not performed adequately or was absent. Regarding synthes of the findings, in most reviews, concerns were raised mainly due to the fact that met analysis was performed in studies of unclear or high risk of bias.   Restricted search strategy and data search; language restrictions; no details given on data collection and appraisal;

Findings of the Reviews
primary studies with clinical and methodological heterogeneity; a majority of studies with an unclear risk of bias; and no meta-analysis performed Cochrane Collaboration Tool analysis was performed in studies of unclear or high risk of bias. Cochrane Collaboration Tool Restricted search strategy and data search; language restrictions; no details given on data collection and appraisal; primary studies with clinical and methodological heterogeneity; a majority of studies with an unclear risk of bias; and no meta-analysis performed analysis was performed in studies of unclear or high risk of bias. Cochrane Collaboration Tool Restricted search strategy and data search; language restrictions; no details given on data collection and appraisal; primary studies with clinical and methodological heterogeneity; a majority of studies with an unclear risk of bias; and no meta-analysis performed analysis was performed in studies of unclear or high risk of bias.  Table 3 summarises the main conclusions of the included systematic reviews and Table 4 summarises the results of the meta-analysis performed. Overall, all materials performed similarly and were equally efficient for the restoration of primary teeth with deep caries. Amalgam and resin composite had the lowest mean failure rate at 24 months and RMGIC and MRGIC were the highest (Figure 2), respectively, at 36 months ( Figure 3). HVGIC showed the highest failure rate with compomer showing the lowest.  Table 3 summarises the main conclusions of the included systematic reviews and ble 4 summarises the results of the meta-analysis performed. Overall, all materials p formed similarly and were equally efficient for the restoration of primary teeth with d caries. Amalgam and resin composite had the lowest mean failure rate at 24 months RMGIC and MRGIC were the highest (Figure 2), respectively, at 36 months ( Figure  HVGIC showed the highest failure rate with compomer showing the lowest. Cochrane Collaboration Tool included studies was either not performed adequately or was absent. Regarding synthesis of the findings, in most reviews, concerns were raised mainly due to the fact that metaanalysis was performed in studies of unclear or high risk of bias. Cochrane Collaboration Tool

Findings of the Reviews
Restricted search strategy and data search; language restrictions; no details given on data collection and appraisal; primary studies with clinical and methodological heterogeneity; a majority of studies with an unclear risk of bias; and no meta-analysis performed included studies was either not performed adequately or was absent. Regarding synthesis of the findings, in most reviews, concerns were raised mainly due to the fact that metaanalysis was performed in studies of unclear or high risk of bias. Cochrane Collaboration Tool Restricted search strategy and data search; language restrictions; no details given on data collection and appraisal; primary studies with clinical and methodological heterogeneity; a majority of studies with an unclear risk of bias; and no meta-analysis performed included studies was either not performed adequately or was absent. Regarding synthes of the findings, in most reviews, concerns were raised mainly due to the fact that met analysis was performed in studies of unclear or high risk of bias.  Table 3 summarises the main conclusions of the included systematic reviews and Table 4 summarises the results of the meta-analysis performed. Overall, all materials performed similarly and were equally efficient for the restoration of primary teeth with deep caries. Amalgam and resin composite had the lowest mean failure rate at 24 months and RMGIC and MRGIC were the highest (Figure 2), respectively, at 36 months (Figure 3). HVGIC showed the highest failure rate with compomer showing the lowest.  Table 3 summarises the main conclusions of the included systematic reviews and ble 4 summarises the results of the meta-analysis performed. Overall, all materials p formed similarly and were equally efficient for the restoration of primary teeth with d caries. Amalgam and resin composite had the lowest mean failure rate at 24 months RMGIC and MRGIC were the highest (Figure 2), respectively, at 36 months ( Figure  HVGIC showed the highest failure rate with compomer showing the lowest.  Table 3 summarises the main conclusions of the included systematic reviews and Table 4 summarises the results of the meta-analysis performed. Overall, all materials performed similarly and were equally efficient for the restoration of primary teeth with deep caries. Amalgam and resin composite had the lowest mean failure rate at 24 months and RMGIC and MRGIC were the highest (Figure 2), respectively, at 36 months ( Figure 3). HVGIC showed the highest failure rate with compomer showing the lowest.   Table 3 summarises the main conclusions of the included systematic reviews and Table 4 summarises the results of the meta-analysis performed. Overall, all materials performed similarly and were equally efficient for the restoration of primary teeth with deep caries. Amalgam and resin composite had the lowest mean failure rate at 24 months and RMGIC and MRGIC were the highest (Figure 2), respectively, at 36 months ( Figure 3). HVGIC showed the highest failure rate with compomer showing the lowest.   Table 3 summarises the main conclusions of the included systematic reviews and Table 4 summarises the results of the meta-analysis performed. Overall, all materials performed similarly and were equally efficient for the restoration of primary teeth with deep caries. Amalgam and resin composite had the lowest mean failure rate at 24 months and RMGIC and MRGIC were the highest (Figure 2), respectively, at 36 months ( Figure 3). HVGIC showed the highest failure rate with compomer showing the lowest.  Table 3 summarises the main conclusions of the included systematic reviews and Table 4 summarises the results of the meta-analysis performed. Overall, all materials performed similarly and were equally efficient for the restoration of primary teeth with deep caries. Amalgam and resin composite had the lowest mean failure rate at 24 months and RMGIC and MRGIC were the highest (Figure 2), respectively, at 36 months ( Figure 3). HVGIC showed the highest failure rate with compomer showing the lowest.   Table 3 summarises the main conclusions of the included systematic rev ble 4 summarises the results of the meta-analysis performed. Overall, all m formed similarly and were equally efficient for the restoration of primary tee caries. Amalgam and resin composite had the lowest mean failure rate at 24 RMGIC and MRGIC were the highest (Figure 2), respectively, at 36 month HVGIC showed the highest failure rate with compomer showing the lowest. [ 26,27,30,36], which restricted search criteria, reducing the comprehensiveness a diately increasing the bias by possibly leaving out eligible reviews. In three reviews, concerns were raised regarding identification and select the search strategy not being reproducible and the inclusion and exclusion cr being clearly defined [27,28,36]. Data collection and study appraisal were ad most reviews, except for four in which clinical and methodological heterogene included studies was either not performed adequately or was absent. Regarding of the findings, in most reviews, concerns were raised mainly due to the fact t analysis was performed in studies of unclear or high risk of bias.  Table 3 summarises the main conclusions of the included system ble 4 summarises the results of the meta-analysis performed. Overa formed similarly and were equally efficient for the restoration of prim caries. Amalgam and resin composite had the lowest mean failure ra RMGIC and MRGIC were the highest (Figure 2), respectively, at 36 HVGIC showed the highest failure rate with compomer showing the   Table 3 summarises the main conclusions of the inclu ble 4 summarises the results of the meta-analysis perform formed similarly and were equally efficient for the restora caries. Amalgam and resin composite had the lowest mea RMGIC and MRGIC were the highest (Figure 2), respect HVGIC showed the highest failure rate with compomer sh In three reviews, concerns were raised regarding identification and selection, with the search strategy not being reproducible and the inclusion and exclusion criteria not being clearly defined [27,28,36]. Data collection and study appraisal were adequate in most reviews, except for four in which clinical and methodological heterogeneity of the included studies was either not performed adequately or was absent. Regarding synthesis of the findings, in most reviews, concerns were raised mainly due to the fact that meta-analysis was performed in studies of unclear or high risk of bias. Table 3 summarises the main conclusions of the included systematic reviews and Table 4 summarises the results of the meta-analysis performed. Overall, all materials performed similarly and were equally efficient for the restoration of primary teeth with deep caries. Amalgam and resin composite had the lowest mean failure rate at 24 months and RMGIC and MRGIC were the highest (Figure 2), respectively, at 36 months ( Figure 3). HVGIC showed the highest failure rate with compomer showing the lowest.  Table 3 summarises the main conclusions of the included systematic reviews and Table 4 summarises the results of the meta-analysis performed. Overall, all materials performed similarly and were equally efficient for the restoration of primary teeth with deep caries. Amalgam and resin composite had the lowest mean failure rate at 24 months and RMGIC and MRGIC were the highest (Figure 2), respectively, at 36 months ( Figure 3). HVGIC showed the highest failure rate with compomer showing the lowest.   RC (n = 5); adhesive systems (n = 2); surface pre-treatments (n = 1); reducing etching time of an etch-and-rinse contemporary adhesive (n = 1); novel self-adhesive composites (n = 2); application modes of a universal adhesive (n = 1); bulk-fill RC (n = 1); sonic-resin placement system in bulk (n = 1); novel base RC (n = 1); contemporary adhesives (n = 1) NR 723 restorations Class-I (n = 494); Class-II (n = 229) RD (n = 6) CCR (n = 7); SCR (n = 1)

Findings of the Reviews
Novel approaches such as bulk-fill resin composites, self-adhesive restoratives, and adhesives have comparable performance to traditional materials. All materials were deemed clinically acceptable in children.

NR
The conventional technique showed a higher survival rate than ART for Class-I and multi-surface restorations with GIC.

NR
Hand

Amalgam (A)
Performance of amalgam was reported in nine reviews [8,20,26,27,31,[33][34][35][36] with the mean calculated failure rate at 24 months being 11% (range: 7%-17%) and at 36 months, 21% (range: 4%-36%). It was reported that amalgam, for multi-surface cavities in primary molars, can be expected to survive a minimum of 3.5 years but potentially more than 7 years and remains an appropriate treatment option. Meta-analysis was performed and found that amalgam, compared to GIC and RMGIC restorations, in primary molars exhibited a lower failure rate but not in a statistically significant manner. The quality of the evidence was of unclear risk of bias due to methodological heterogeneity.

Resin Composite (RC)
The mean failure rate for RC as calculated from the ten reviews reporting on its efficacy varied from 14% at 24 months to 20% at 36 months [8,20,26,27,[30][31][32]34,36,37]. Resin composite exhibited the lowest failure rates, with all resin-based restorative materials (resin composite and compomer) having no statistically significant differences based on a "moderate" level of evidence. In the review of Delgado et al. (2021) it was reported that innovative materials, such as bulk-fill resin composites, self-adhesive restoratives, and adhesives, perform comparably to conventional restorative materials and all are clinically acceptable [32].

Glass-ionomer Cements (GICs)
A total of 13 studies reported on GIC in various forms with the calculated mean failure rate at 24 months ranging between 16% and 21% and 16% and 35% at 36 months. At both time intervals, the lowest failure rates were reported with RMGIC and the highest with HVGIC for the restoration of primary teeth. Overall, results are inconclusive, with most reviews [20,26,27,29-

Amalgam (A)
Performance of amalgam was reported in nine reviews [8,20,26,27,31,[33][34][35][36] with the mean calculated failure rate at 24 months being 11% (range: 7%-17%) and at 36 months, 21% (range: 4%-36%). It was reported that amalgam, for multi-surface cavities in primary molars, can be expected to survive a minimum of 3.5 years but potentially more than 7 years and remains an appropriate treatment option. Meta-analysis was performed and found that amalgam, compared to GIC and RMGIC restorations, in primary molars exhibited a lower failure rate but not in a statistically significant manner. The quality of the evidence was of unclear risk of bias due to methodological heterogeneity.

Resin Composite (RC)
The mean failure rate for RC as calculated from the ten reviews reporting on its efficacy varied from 14% at 24 months to 20% at 36 months [8,20,26,27,[30][31][32]34,36,37]. Resin composite exhibited the lowest failure rates, with all resin-based restorative materials (resin composite and compomer) having no statistically significant differences based on a "moderate" level of evidence. In the review of Delgado et al. (2021) it was reported that innovative materials, such as bulk-fill resin composites, self-adhesive restoratives, and adhesives, perform comparably to conventional restorative materials and all are clinically acceptable [32].

Glass-ionomer Cements (GICs)
A total of 13 studies reported on GIC in various forms with the calculated mean failure rate at 24 months ranging between 16% and 21% and 16% and 35% at 36 months. At both time intervals, the lowest failure rates were reported with RMGIC and the highest with HVGIC for the restoration of primary teeth. Overall, results are inconclusive, with most reviews [20,26,27,[29][30][31][33][34][35][36][37] reporting a similar or even better performance of GICs when compared to conventional restorative materials. Characteristically, Dias et al. (2018) and Mickenautsch et al. (2010;2011) reported that when compared to RC and A, they presented a similar clinical performance for all criteria, except for secondary carious lesions, in which GIC presented superior performance [30,33,35]. Although, there are two reviews reporting worse results for GICs performance. In the review by van't Hof et al. (2006), it was reported that despite the high survival rates for single-surface ART restorations with HVGIC, survival rates for multi-surface ART restorations were low [28]. Similarly, Chisini et al. (2018) concluded that MRGIC exhibited the highest failure rate [8].

Compomer (CO)
The efficacy of compomers as restorative materials was reported in six reviews [8,26,27,31,34,37], with calculated failure rates differing slightly for 24 and 36-month follow-ups (19% and 13%, respectively). Based on a moderate level of evidence, there was no statistically significant difference calculated between compomer and both conventional restorative materials and novel approaches, underlying that there is no advantage among these materials. In the study by Siokis et al. (2021), where meta-analysis was performed comparing compomer with RC and RMGIC, there were no statistically significant differences reported between the materials based on the "moderate" quality of evidence (RR 1.12 [0.41, 3.02]; p = 0.83; I 2 = 57%, RR 1.04 [0.59, 1.84]; p = 0.88; I 2 = 1%, respectively) [37]. In the review by Tedesco et al. (2018), rank probability, calculated through network meta-analysis, showed that the best results for treatment of occlusal caries are expected using compomer [34]. In the same study, regarding the treatment of multi-surface caries, compomers were ranked third after the Hall technique and non-restorative caries treatment.

Stainless Steel Crowns (SSC)
Three reviews reported on the use performance of crowns for the restoration of carious primary teeth [8,29,34], with only one reporting calculable data. The failure rate was as low as 1% at 24 months and 4% at 36 months, underlying the excellent performance of the restorative material when compared to common conventional restorative materials. A meta-analysis reported that crowns have a reduced risk of a major failure at 24 months when compared to common filling materials (RR 0.18, 95% CI [0.06, 0.56], I 2 = 0%) based on high-quality evidence [29].
3.5.6. Secondary Outcomes Data on secondary outcomes was non-existent in most cases, as quantitative synthesis in almost all the reviews reported effectiveness in favour of specific restorative materials. In the review by Innes et al. (2015) [29], it was found that crowns were less likely to cause pain than conventional restorations at 12-24 months (RR 0.15, 95% CI [0.04, 0.67]; I 2 = 0%). Furthermore, participants reported more discomfort when a restoration was placed compared to crown placement (RR 0.56, 95% CI [0.36, 0.87]; I 2 = 0%). The discomfort was defined as 'moderate', 'intense', or 'very intense' pain reported by children and 'moderate' or 'significant' patient discomfort rated by the dentist during treatment [29].

Factors Affecting the Outcome
Different study designs and different handling of various materials regarding the necessity for application of local anaesthesia and/or isolation restricted the ability to assess the effect of the tooth and procedure-related factors on the outcome. The effect of singleor multi-surface restorations was demonstrated in four reviews [8,27,28,36], with most indicating that single-surface restorations exhibited lower failure rates as compared to multi-surface restorations (4.78%-7.6% vs. 9.46%-14.7%). Although, one review reported no difference in the weighted mean survival percentage of single-and multi-surface restorations [36]. Main reasons for the failure of both Class-I and Class-II restorations reported were secondary caries, restoration loss, and chipping of the marginal ridge with approximal contact loss.
Regarding the effect of tooth isolation on the outcome, this was only reported in two studies [8,30], indicating higher success rates in restorations placed with rubber dam isolation (93.6% vs. 77.5%). In the review by Dias et al. (2018), it was shown that in procedures performed using cotton roll isolation, there was no significant difference between materials for all parameters analysed [30].
The overlap of primary studies calculated by the CCA was 0.06 (6%) revealing a moderate risk of bias due to inadvertently including the results of primary studies more than once [23]. Results showed a wide range of failure rates with all materials performing similarly and being equally efficient for restoration of primary teeth with deep caries. Amalgam and resin composite had the lowest mean failure rate at 24 months, while RMGIC and MRGIC had the highest, respectively, at 36 months. HVGIC showed the highest failure with compomer showing the lowest. Main reasons for the failure of both Class-I and Class-II restorations reported were secondary caries, restoration loss, and chipping of the marginal ridge with approximal contact loss. Evidence from reviews reporting on factors affecting the outcome indicated that single-surface restorations and restorations placed using rubber dam isolation exhibited lower failure rates.
Conventional restorative materials, such as amalgam and resin composites, have a limited application in primary teeth. Despite their acceptable annual failure rate, their use in everyday practice is reduced. This can be attributed to patient-related factors that directly affect the execution of the restoration and therefore its longevity. Composite restorations are highly sensitive to the lack of complete moisture control, especially due to difficulties in cooperation, which can jeopardise the good performance of the material [8]. This underlines the increased prevalence of lost restorations reported in many studies. Similarly, amalgam, although having high durability with a survival range of 3.5 to 7 years, raises concern for its future use as a restorative material for Class-II cavities in primary teeth due to its toxicity and aesthetics [26]. Despite being considered an appropriate treatment option with high effectiveness and durability, its consideration as an option in future RCTs is questionable.
Less technique-sensitive materials are gaining interest and indicate high success rates in recent clinical studies. Their increased success rates are attributed to their biocompatibility and their easier and faster application, as compared to resin composites. It has been reported that the adhesion to the tooth structure is comparable to the micromechanical retention achieved by resin composite, giving the materials similar longevity rates [30]. Although, adhesion should not be the only criterion for retention of the restorative material as other factors directly related to the cavity (e.g., size, surfaces involved) may affect failure rates. At the same time, the less time-consuming application improves the procedure's acceptability by the patients and has a positive effect on behavioural shaping and overall management of even uncooperative patients [39]. Results from previous studies reported similar results regarding the annual failure rate of GIC combined with ART and conventional restoration with composite or amalgam [27].
Our study indicated an equally excellent performance of GIC with improved materials, such as RMGIC, with better physical properties, presenting better fracture and wear resistance [17]. Results from most studies reported better performance of the latter for the restoration of small to moderate-sized proximal cavities. It is notable that in a recent study, a more than 5-times higher risk of failure was found when GIC was used for the restoration of Class-II cavities as compared to RMGIC restorations [40]. There are also studies reporting that the RMGIC survival rate is better than RC but worse than compomer [27]. This was mainly attributed to the worse surface roughness, anatomic form, and marginal adaptation of RMGICs. These shortcomings, in combination with the deficiency of a good colour match, limit the overall use of the material.
GICs tend to form a stronger chemical bond, have a thermal expansion coefficient comparable to dentine, and release fluoride [17]. It has been suggested that fluoride released by these materials reduces tissue demineralisation and prevents caries recurrence, although the evidence is not strong as studies are not of high quality [41]. This is not in accordance with the results of other reviews that reported secondary caries as the main reason for the failure of GIC restorations, underlining that fluoride release does not affect the longevity of the material [8].
Evidence supports the use of simpler, less time-consuming techniques in a controlled environment for the treatment of deep caries in primary teeth. Simplification of adhesive systems and the use of flowable materials that are easily handled to achieve adaptation have been featured in RCTs with promising results. The limited lifespan of primary teeth favours their application, despite their questionable mechanical properties and shrinkage. However, the evidence to date is too limited to draw specific conclusions and further studies are required to investigate their efficacy.
In a similar way, bulk-fill composites are reliable materials for restoring primary teeth, which can decrease working time as they can fill the cavity in one step without layering. Studies comparing the material with other innovative techniques, such as reinforced GICs, confirmed their higher clinical success rate [32]. Compared to conventional restorative materials, they seem to have similar clinical outcomes, especially when referring to longevity.
Excellent performance with very low failure rates was also observed for SSC based on high-quality evidence. Despite their reliability and longevity, they present a much more invasive treatment option when compared to other restorative materials commonly used. Additionally, restorative options after failure are limited due to the increased hard tissue removal, while common restorative materials could be replaced even after further tissue loss due to secondary caries. This highlights that the success rate should not be the only factor to be considered for the ideal choice of material in each case.
Our review supports that the failure rate of restorations is affected by tooth and technique-related factors. It has been demonstrated in almost all the included reviews that failure rates drop when materials are used in single-surface restorations [8,27,28]. Similarly, the use of rubber dam isolation increases the longevity of the restorations as compared to cotton roll isolation [8,30]. Despite the fact that it restricts saliva contamination and improves moisture control, therefore improving adhesive properties, it is not always easily applicable. This underlines the fact that having ideal application conditions in children is uncertain and therefore the choice of restorative material should be made carefully.

Strengths and Limitations
This umbrella review critically appraises the accessible evidence and presents a comprehensive overview of current restorative treatment approaches for carious lesions in primary teeth, which is its principal strength. The 14 included systematic reviews presented a broad spectrum of restorative treatment approaches for carious primary teeth, which have not been previously covered by one systematic review exclusively.
However, there are limitations of this umbrella review that need to be addressed. First and foremost, several review authors mentioned the heterogeneity among the included primary studies [20,29,31,33,36]. As far as the study design is concerned, five systematic reviews included only RCTs [8,27,28,32,34] and the included RCTs were further subdivided into split-mouth and parallel-arm studies. Pires et al. (2018) reported on a heterogeneity with regard to the sample size, the type of cavity and isolation, the restorative materials and techniques under investigation, as well as the length of the follow-up [31]. The comparison of studies and the evaluation of outcomes were complicated by the fact that a broad spectrum of restorative materials was included, with a few being used in a limited number of initial studies. It was further hampered since, within the same restorative material class, material compositions or adhesive concepts may have been modified by manufacturers over time to improve the material properties. The influence of the operator on the results of the studies also did not play a role in the included reviews [7]. By using different evaluation criteria with various cut-off points to assess the longevity of restorations, the data become even more heterogeneous [8,32]. The heterogeneity observed among the different primary studies often impedes the direct comparison of results. The quality of evidence is further restricted given the finding that several systematic reviews performed meta-analyses based on primary studies with an overall unclear or high risk of bias [20,30,31,34,35,37].
Secondly, a limit in the dropout rate for primary studies included in the systematic reviews was not chosen as an eligibility criterion for this umbrella review because of the wide variation of numbers presented within the included systematic reviews. Kilpatrick and Neumann (2007) observed a high dropout rate among primary studies on the use of amalgam in Class-II cavities in primary molars. In addition to this, inconclusive reporting of the results and different handlings when dealing with censored data, in case of restorations lost to follow-up or exfoliated teeth, hampered the calculation of failure rates [26].
Thirdly, the follow-up time among included primary studies was variable, ranging from 6 to 60 months. On the one hand, the lifespan of primary teeth in the oral cavity is limited because of the physiological exfoliation [8,26,32,34] which limits the maximum follow-up period, especially when older children are included. On the other hand, the occurrence of secondary caries at the restoration margins takes time, which is why its observation requires studies with a long-term follow-up [42,43].
Finally, the blinding of outcome assessors is often problematic, if not impossible, when restorative materials of different clinical appearances are compared with each other or with preformed crowns [20,29,31,36]. Therefore, the potential risk of bias in the measurement of the outcome has to be taken into consideration when interpreting the outcome of these studies.
The heterogeneity among included systematic reviews presented different study designs, several comparisons, and various outcome measures that precluded the quantitative synthesis of results. In five systematic reviews, meta-analyses were not performed and another two of the systematic reviews conducted network meta-analyses which did not allow for a comparison with conventional meta-analyses. Non-independence of samples was rarely assessed and causes of variation among included studies were not always discussed. For included studies with split-mouth design, cluster-level analysis and analysis of clustering effects were not reported. Ultimately, the unclear or high risk of bias among the majority of included systematic reviews ruled out meta-analyses.

Recommendations for Future Research
There is a need for further well-designed RCTs to overcome the limitations of studies on the restorative treatment of carious primary teeth and to increase the internal validity of future ones as well as the systematic and umbrella reviews [8,[30][31][32][33]35,37]. This requirement was also observed in studies on (non-)vital pulp therapy in primary teeth [44]. In general, it is recommended to register the trial protocol from the start [45] and to follow the Consolidated Standards of Reporting Trials (CONSORT) to increase the transparency and completeness of reporting [33,35,45,46]. The trials should be conducted in parallelarm design to avoid the shortcomings of split-mouth studies, with adequate random sequence generation and allocation sequence concealment to avoid bias arising from the randomisation process [35].
Future sample size calculations based on power analyses should take the high dropout rate observed among primary studies into account to obtain meaningful results after longer follow-up periods [47]. The inclusion of younger participants with a narrower age range would support the longer follow-up periods that are needed, given the shorter lifespan of primary teeth in the oral cavity leading to their physiological exfoliation [29,47]. In this respect, the handling of exfoliated teeth, censored data, and dropouts needs to be clarified, as the number of restorations at the different follow-ups affects the calculation of failure rates [47].
The caries risk of participants should be reported, as it was shown that a higher caries risk is associated with increased susceptibility to restoration failure [43,48]. The fact that a secondary caries was found to be the main reason for failure in primary [8] and permanent teeth [43,48] shows that the participants are at risk of caries even after the restorative treatment and illustrates the need for behavioural changes by employing additional caries-preventive strategies to achieve a long-term success [8]. Innes et al. (2015) recommended extending the inclusion criteria of future studies to children with special needs or developmental defects of teeth and to include general dental practitioners without specialisation in an attempt to increase the generalisability of the outcome [29].
As far as the teeth are concerned, the type of included teeth and the cavity class should be mentioned. This is connected to the extension of the carious lesions, which were frequently found to be stated insufficiently [29,36]. Therefore, the carious lesion depth and its impact on the outcome should be reported [29].
The experience of the operators performing the restorative treatment should be mentioned because it may influence the clinical performance of restorations [27,48]. The reporting of treatment-related factors, such as the choice of anaesthesia, the isolation technique, and detailed descriptions of restorative materials and techniques facilitates the interpretation of results, which is desirable for future studies.
For the outcome assessment, validated and internationally accepted criteria should be used to determine the restorative treatment success. The clinical criteria introduced by the FDI World Dental Federation in 2007 have been used in an increasing number of clinical trials and may be a viable option for further studies [49,50]. In addition to the classical outcome measures, patient-related and -reported factors should be further documented, such as discomfort, pain, and the impact on the oral health-related quality of life [29,47].

Conclusions
Within the limitations of the current review, the conclusions drawn were: • All restorative materials have acceptable mean failure rates and their use for the treatment of carious primary teeth is supported. • Among common conventional restorative materials, amalgam and resin composite had the lowest mean failure rates at 24 months and compomer at 36 months. • Stainless steel crowns had the lowest failure rate at 24 months and 36 months compared to all other restorative materials for primary teeth. • Limited reviews indicated that single-surface restorations and restorations placed using rubber dam isolation exhibited lower failure rates.

•
The main reasons reported for the failure of both Class-I and Class-II restorations were secondary caries, restoration loss, and chipping of the marginal ridge with approximal contact loss.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/jcm11123490/s1; Table S1: Search strategy; Table S2: Reason for exclusion of reviews retrieved for the current umbrella review; Table S3: Citation matrix for the primary studies included in the reviews of the current umbrella review. Informed Consent Statement: Not applicable.

Data Availability Statement:
The data presented in this umbrella review are available in the manuscript and the supplementary material. Further information is available on request from the corresponding author.