Failure Risk of Composite Resin and Amalgam Restorations: A Systematic Review and Meta-Analysis

Abstract

Objective

This systematic review and meta-analysis were conducted to compare the failure risk between amalgam and composite resin materials in permanent posterior teeth.

Material and methods

Study eligibility requirements included clinical trials and observational studies with at least 12 months of follow-up. English-language studies from 1990 onwards were the only studies included. This review follows the Cochrane Handbook for Systematic Reviews of Interventions and the guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA). Our search strategy included using the following databases: PubMed, Cochrane, and Google Scholar. The primary outcome was restoration failures, defined as restoration replacements, tooth and restoration fractures, secondary caries, postoperative sensitivity, and toothaches. We conducted a random-effects meta-analysis to determine the risk ratio (RR) of the included studies, and publication bias was assessed. The Cochrane Risk of Bias tool was employed to evaluate the quality of the clinical trials, while the Newcastle–Ottawa scale was used to assess the quality of other studies.

Results

The results were derived from 13 studies. The failure proportion for amalgam ranged from 0% to 50.0%, while that of composite resin restorations ranged from 0% to 62.7%. The meta-analysis did not find any statistically significant difference in failure risk between amalgam and composite resin restorations (RR: 0.96, 95% confidence intervals: 0.68-1.34). The Egger’s test results did not show any significant evidence of publication bias in the meta-analysis (P > .05).

Conclusion

This review did not reveal any statistically significant difference in the RR between composite resin and amalgam restorations. However, in their analyses, the 13 studies used varying definitions of failure and did not account for some important factors that might have influenced restoration failures. Future reviews need to account for other influential variables that contributed to restoration failures.

Introduction

The field of restorative dentistry is dominated by two dental materials: amalgam and composite resin.¹ Each material has distinctive advantages and disadvantages that influence a provider’s decision when selecting the appropriate treatment. Despite the benefits of both materials, providers have significantly shifted towards composite resin and away from amalgam. This shift is motivated by aesthetic preferences and concerns about mercury, although the evidence about mercury risks is limited.² The selection of restorative materials should be based on scientific evidence that prioritizes patients’ benefits.³

Amalgam, which is strong and durable, is particularly advantageous for posterior restorations due to its ability to withstand significant occlusal forces.⁴ In contrast, composite resin restorations offer superior aesthetic properties and support a more conservative approach that helps to preserve healthy tooth structure.⁵ However, the application of composite resin is highly technique-sensitive,⁶ which presents challenges during its use, especially for posterior teeth with proximal involvement. One important issue associated with composite resin fillings is polymerization shrinkage,⁷ a process in which resin-based dental materials contract during the curing or polymerization process.⁷ Although amalgam is praised for its strength and excellent marginal seal,⁴ it also has its drawbacks. Amalgam fillings require sufficient preparation depths to ensure effective mechanical retention.⁸ This requirement can pose challenges to the preservation of a healthy tooth structure, especially compared to the more conservative preparation required for composite resin fillings.⁵

The mercury content in amalgam has raised health and environmental concerns.^2,8 Although current research suggests that the risks associated with mercury may overstate,^2,8 the trend towards banning amalgam in some clinics continues, raising this question: Is this shift grounded in evidence-based decision-making?^9,10

Two meta-analyses conducted in 2015⁹ and 2021¹⁰ have shown a higher failure rate for composite resin restorations compared to amalgam restorations. Despite the increasing demand and preference for composite resin restorations, there is a lack of scientific evidence to fully support this trend. The most recent meta-analysis published in 2021 included eight studies from 1988 to 2016¹⁰ and highlighted the need for updated evidence.

Therefore, the aim of this systematic review and meta-analysis is to provide an updated and comprehensive evaluation by critically analysing the most recent data to compare the failure risk of amalgam and composite resin restorations in posterior teeth.

Materials and methods

The present systematic review and meta-analysis study adheres to the Cochrane Handbook for Systematic Reviews of Interventions¹¹ and the guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA).¹²

Focused question

What is the difference in the risk of failure between composite resin restorations and amalgam restorations in permanent posterior teeth?

Null hypothesis

There is no difference in the risk of failure between composite resin restorations and amalgam restorations in permanent posterior teeth.

Search methodology

We conducted an electronic search in PubMed/MEDLINE, Cochrane Central Register of Controlled Trials, and Google Scholar from 1990 onwards. The search methodology and the population, exposure, comparison, and outcome framework are detailed in Table 1. The references of the 13 included studies were examined to identify further relevant studies.

Table 1.

Systematic search strategy (PECOS strategy).

Search strategy
Population	#1 (“Dental Restoration Failure”[Mesh] OR “Dental Restoration, Permanent”[Mesh] OR “Dental Caries”[Mesh] OR “Molar”[Mesh] OR “Bicuspid”[Mesh])
Exposure and control	#2 (“Dental Amalgam”[Mesh] OR “Composite Resins”[Mesh])
Outcomes	#3 (Survival [tiab] OR success [tiab] OR failure [tiab] OR longevity [tiab] OR amalgam longevity [tiab] OR resin longevity [tiab] OR composite resin longevity [tiab] OR long-term [tiab] OR follow-up [tiab])
Search combination	#1 AND #2 AND #3
Database search language	English
Electronic database	PubMed/MEDLINE, COCHRANE Central Register of Controlled Trials, and Google Scholar

Selection criteria (inclusion)

• •Articles published in the English language.
• •Articles published from 1990 onwards.⁹
• •Articles including human participants.
• •Randomized controlled trials (RCTs), controlled clinical trials, and both prospective and retrospective cohort studies, and case-control studies with at least a 1-year follow-up.
• •Articles focused on amalgam or composite resin restorations placed on premolars, first molars, and second molars.
• •Articles focused on restoration failure as an outcome and defined it as fillings requiring restoration replacement, secondary caries, pain, sensitivity, and tooth/restoration fracture.

Selection criteria (exclusion)

• •Articles lacking a clear, definitive differentiation between composite resin and amalgam restorations.
• •Articles that include animal subjects.
• •In vitro studies.
• •Articles studying amalgam or composite resin restorations placed on anterior permanent teeth, third molars, root caries, and primary teeth.
• •Cross-sectional studies, case studies, case reports, and literature reviews.

Screening process

The search and screening process was conducted collaboratively by WA, AA, HA, and MA, who worked together at the same time and location. The search was performed from August 2023 to December 2023. Initially, titles and abstracts were reviewed, followed by a comprehensive assessment of full articles that adhered to predefined eligibility criteria (inclusion/exclusion) for subsequent data extraction.

After completing our comprehensive assessment, a manual search was conducted to ensure that the most relevant results were identified and included. Discrepancies between reviewers were resolved through detailed discussions with a fifth author, HQ, and a consensus was made and recorded. Authors of the studies were contacted via email for clarification when necessary.

Quality assessment

Three examiners (WA, HA, and MA) conducted their assessments independently. Any discrepancies in quality assessments were resolved through discussions. The quality assessments followed the Newcastle–Ottawa scale (NOS),¹³ designed for systematic reviews incorporating nonrandomized studies, including cohort studies across three main domains: selection, comparability, and outcome. For cohort studies, each relevant item in the selection and outcome/exposure categories earned one star. In the comparability category, studies received up to two stars. According to the NOS criteria, the highest possible score for a study is 9 stars/points, which is indicative of the highest scientific rigor.¹³ Studies achieving a score of 6 stars or more are considered to have high methodological quality. The Cochrane ‘Revised Risk of Bias 2 (RoB 2) Assessment’ was used to assess the quality of the RCTs.¹⁴

Data extraction

The data retrieved by three examiners (WA, HA, and MA) collaboratively from the studies included the following: author, study design, year of publication, follow-up duration, country of data collection, sample population, number of dropouts, type of teeth included (premolar/molar), number of restorations (amalgam and composite resin), total number of amalgam and composite resin restorations, number of teeth with failed amalgam and composite resin restorations, definition and measurements of failure, follow-up period, and patient age.

Statistical analysis

The primary outcome was a binary variable indicating restoration failure, which included secondary caries, restoration replacement, pain, sensitivity, and tooth/restoration fracture. We used risk ratios with 95% confidence intervals to determine whether composite resin or amalgam has a higher risk of failure in posterior permanent teeth. Heterogeneity was assessed using the I² statistic, with values up to 25% indicating low heterogeneity, values around 50% indicating moderate heterogeneity, and 70% or higher indicating high heterogeneity. In cases of significant heterogeneity (P < .10), results from the random effects model were considered, whereas the random effects restricted maximum likelihood model was used when heterogeneity was low. A random effects model was applied using the inverse variance method to account for variability among studies. Statistical significance was set at P < .05. Publication bias was assessed visually using a funnel plot, with asymmetry suggesting potential bias. Data analysis was performed using STATA 17.

Results

Literature search

The initial search identified 1336 articles in PubMed/MEDLINE, 89 articles in the Cochrane Library, and 5600 articles in Google Scholar. After an initial assessment and screening based on titles and abstracts, the selection was narrowed down to 292 studies. Upon further reading of the full texts of these 292 articles, studies were excluded for not meeting the inclusion/exclusion criteria, ultimately leaving 17 studies. Following a thorough review of these full texts, three additional studies were excluded due to missing or insufficient information, and one was conducted on cadavers, which did not meet our inclusion criteria. Consequently, 13 studies published from 1990 onwards were included in this review.15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27 The selection process is illustrated in Figure 1.

Fig. 1.

Flow diagram (PRISMA format) of the screening and selection process.

Study characteristics

These are summarized in Table 2. Our review includes five randomized clinical trials,^15,18,24, 25, 26 four prospective cohort studies,^17,19,20,22 and four retrospective cohort studies.^16,21,23,27 The number of participants in these studies ranged from 25²⁶ to 4856.¹⁷ The follow-up duration varied from 24 months²⁷ to 144 months.¹⁸ The type of material used in both amalgam (high copper content vs dispersed phase alloy) and composite resin (hybrid vs microhybrid) was reported in only two studies.^16,22 Five studies documented the techniques used to isolate the operative field,^{15,16,22,24,26} and only three studies included the adhesive system used for placing the restoration.^16,23,26

Table 2.

The main characteristics of selected studies.

Study	Year of publication	Study design	Follow-up period (mo)	No. of subjects	No. of subjects dropouts [%]	Age range (y) [mean age]	Number of restorations AM vs C
Jardim et al15	2020	RCT	60	233	130 [56%]	6-53 [17.2]	122 AM 177 C
Santos et al16	2023	Cohort (retrospective)	60	125	45 [NR]	26-86 [65.5]	69 AM 50 C
Vieira et al17	2017	Cohort (prospective)	12, 24, 60, 120	4856	NR [NR]	NR [42]	6266 AM 2010 C
Estay et al18	2017	Controlled clinical trial	12-144	34	3 [3.41%]	18-80 [26.5]	59 AM 30 C
Bencheikh et al19	2023	Cohort (prospective)	60	NR	NR [NR]	18-96 [NR]	5540 AM 17,753 C
da Fonseca Cumerlato et al20	2019	Cohort (prospective)	84	1303	303 [23.3%]	12 [NR]	17 AM 183 C
Opdam et al21	2007	Cohort (retrospective)	60	273	NR [NR]	23-77 [48]	1202 AM 747 C
Collins et al22	1998	Cohort (prospective)	96	72	26 [36%]	13-32.4 [16.8]	52 AM 161 C
Kim et al23	2013	Cohort (retrospective)	60	232	0 [0%]	NR [NR]	147 AM 676 C
Bernardo et al24	2007	RCT	84	472	35 [NR]	8-12 [NR]	856 AM 892 C
Letzel et al25	1989	RCT (multicentre)	12, 24, 36, 48, 60	447	109 [76%]	NR [NR]	232 AM 932 C
Kemaloglu et al26	2019	RCT	0.5, 6, 12, 36	25	5 [NR]	18-60 [NR]	20 AM 20 C
Birch et al27	2016	Cohort (population-based, retrospective)	24	NR	NR [NR]	Group 1: < 35 Group 2: 35-95 [NR]	791,723 AM 1870,123 C

Total No. of restorations: 20,700,059. Total No. of AM restorations: 806,305. Total No. of C restorations: 1893,754.

AM, amalgam restorations; C, composite restorations; NR, not reported.

Quality assessment

The quality assessment of the included studies was conducted using two methods: the NOS for observational studies and the Cochrane Collaboration’s tool for clinical trials (Table 3). According to the NOS, the quality assessment of the observational studies yielded scores ranging from 5 to 7 stars out of a possible 9, indicating low to intermediate levels of quality based on selection, comparability, and outcome measurements.

Table 3.

Quality assessment of the studies by Newcastle–Ottawa scale – Cohort studies.

Quality assessment Newcastle–Ottawa – Cohort studies
	Selection			Comparability		Outcome
Study	Representativeness of exposed cohort	Selection nonexposed cohort	Ascertainment of exposure	Demonstration that outcome of interest was not present at start of study	Comparability of cohorts on basis of design	Assessment of outcome	Follow-up long enough for outcomes	Adequacy of follow up	Total
Santos et al16	*	*	*	*	**	*	0	0	7/9
Bencheikh et al19	*	*	*	0	*	0	*	0	5/9
Vieira et al17	*	*	*	*	*	*	*	0	7/9
da Fonseca Cumerlato et al20	*	*	*	0	*	0	*	0	5/9
Opdam et al21	*	*	*	*	*	*	*	0	7/9
Collins et al22	*	*	*	0	**	0	*	*	7/9
Birch et al27	*	*	*	0	*	0	*	0	5/9
Kim et al23	*	*	*	0	*	*	*	0	6/9

A high risk of bias was observed across the RCTs included in this analysis, according to the Cochrane Collaboration’s tool. Each item in the risk of bias assessment was rated as ‘High risk of bias’, ‘Some concerns of bias’, or ‘Low risk of bias’, using criteria that included the randomization process, deviation from intended intervention, missing outcome data, measurement of the outcome, and selection of the reported results (Table 4).

Table 4.

Cochrane Revised Risk of Bias 2 (RoB 2) assessment.

Study	D1: Randomization	D2: Deviations from intervention	D3: Missing outcome data	D4: Measurement of outcome	D5: Selection of reported result	Overall bias
Jardim et al15	Low	Some concerns	Low	High	High	High
Estay et al18	Low	Low	Low	High	Low	High
Bernardo et al24	High	High	Some concerns	High	Low	High
Letzel et al25	High	High	Some concerns	High	Low	High
Kemaloglu et al26	Low	High	Low	High	Low	High

Risk of Bias 2 (RoB 2) assessment table. RoB 2 domain key: D1: Bias arising from the randomization process. D2: Bias due to deviations from intended interventions. D3: Bias due to missing outcome data. D4: Bias in the measurement of the outcome. D5: Bias in the selection of the reported result.

Meta-analysis

The failure proportion of amalgam ranged from 0%²⁶ to 50%,¹⁵ while the failure proportion of composite restorations ranged from 0%²⁶ to 62.7%.¹⁵ The meta-analysis revealed an risk ratios of 0.96 (95% confidence intervals: 0.68-1.34, P > .05), indicating no statistically significant difference between failure risk of amalgam and composite restorations (Figure 2). Due to the considerable heterogeneity found (I² = 97.02%; P < .05), a random effects model was used for the analysis of failures between the two types of restorations. The funnel plot demonstrated symmetry, indicating the absence of publication bias (Figure 3). The Egger’s test results showed a p-value of 0.8387, indicating no statistically significant evidence of publication bias in the meta-analysis.

Fig. 2.

A forest plot for the failure risk of composite and amalgam restorations.

Fig. 3.

Funnel plot for the studies reporting the failure risk of composite and amalgam restorations.

Discussion

This meta-analysis integrated data from 13 studies published between 1990 and 2023. The results did not reveal any statistically significant difference in failure risk between amalgam and composite resin restorations. This finding contradicts the results of two earlier meta-analyses conducted in 2015 by Moraschini et al⁹ and in 2021 by Worthington et al¹⁰ Both previous meta-analyses suggested that composite resin restorations in posterior permanent teeth have higher failure rates and reduced longevity compared to amalgam restorations.^9,10 Furthermore, the most recent of these meta-analyses included articles published up to 2016,¹⁰ underscoring the need for a more updated review of recent evidence. The meta-analysis of clinical trials by Worthington et al¹⁰ found mean failure rates of 7.5% for amalgam and 14.2% for composite resin over an average of 72 months of follow-up. Similarly, the meta-analysis of observational studies by Moraschini et al⁹ reported survival rates of 3.17% for amalgam and 1.71% for resin composites over at least 12 months of follow-up.

In our meta-analysis, 11 studies consistently identified secondary caries as a definitive criterion for restoration failure,15, 16, 17, 18^,20,22, 23, 24, 25, 26, 27 with seven studies reporting a higher number of failures in the composite resin restoration group^15,20,22, 23, 24, 25^,27 and four studies finding more failures in the amalgam restoration group.16, 17, 18^,26 Eight studies recognized tooth or restoration fracture as indicative of failure,^16,17,20, 21, 22^,24,26,27 with four studies reporting higher failure rates among resin composite restorations.^20,22,24,27

One study defined failure in terms of the necessity for subsequent endodontic treatment or tooth extraction, concluding a higher failure rate of 9.3% in the amalgam restoration group compared to 4.5% among composite resin restorations.¹⁹ Other indicators of failure cited in the studies included pain/sensitivity, issues with marginal adaptation or staining overhanging margins, and restoration displacement or loss.16, 17, 18^,20,21,25,26

Although our analysis did not demonstrate a statistically significant difference in failure risk between composite resin and amalgam restorations, a notable limitation of the included studies was the lack of detailed reporting on factors that can influence restoration failure: operative techniques, material brand/type, use of base/liners, isolation techniques employed, and data on adhesive systems used for restoration placement. The absence of data could obscure the assessment of restoration longevity and present challenges in evaluating the consistency and reliability of failure measurements across studies.

In our analysis, we demonstrated that studies published after 2013 reported higher success rates among the resin composite restoration group,^{17,19,23,26,27} suggesting potential advancements in the quality of composite resin restorations due to continuous improvements in their physical and mechanical properties. This may also explain the conflicting results between our meta-analysis and the two previously published meta-analyses.

The limitations of this study might be from the restriction of inclusion criteria to studies published only from 1990 onwards, which was done to align with previous meta-analyses and enable a comparison with their findings. Insufficient documentation in the included studies and a lack of consideration for confounding factors that may obscure the true relationship between exposure and outcomes. These limitations highlight the challenges in ensuring the reliability of the study’s findings, making it difficult to draw reliable conclusions. When further clarification on a study was requested from five authors, only one author responded.²⁷ Attrition bias was also evident due to undisclosed dropout rates in several studies.

The variability in how failure was defined across studies could have introduced potential performance and interpretation biases. To advance future research, it would be beneficial to expand on the current findings by including factors that may influence failure and longevity rates and to clearly define what constitutes restoration failure. Despite its limitations, this study has enhanced our understanding of the differences in failure rates and longevity between the two types of restorations.

Conclusion

This systematic review and meta-analysis indicate that there is not a statistically significant difference in failure risk between amalgam and composite resin restorations. However, this result must be interpreted with caution due to several limitations identified during the analysis. The insufficient reporting on other influential factors that could contribute to restoration failure highlights the gaps in the existing literature. Furthermore, the variations in how restoration failure was defined across the included studies introduced a level of heterogeneity that may have affected the results. To advance our understanding in this field, future research needs to focus on establishing standardized criteria for defining restoration failure and comprehensive reporting of all factors that may influence restoration outcomes. Such efforts will help in refining the assessment of restoration longevity and in guiding evidence-based decision-making in restorative dentistry.

Conflict of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this article.