Abstract
Objectives
This study compared the clinical performance of glass ionomer cement (GIC) compared to composite resin (CR) in Class II restorations in primary teeth.
Data
Literature search according to PRISMA guidelines including randomized controlled trials comparing Class II restorations performed with GIC, compared to CR, in primary teeth.
Sources: PubMeb, Scopus, Web of Science, VHL, Cochrane Library, Clinical Trials and OpenGrey, regardless of date or language.
Study
Ten studies were included in qualitative synthesis, and 9 in the meta-analyses (MA). Six studies were classified as low risk of bias, and 4 as “unclear”. Heterogeneity ranged from null to high (0% to 73%). GIC and CR presented similar failure patterns (risk difference −0.04 [−0.11, 0.03]; p = 0.25, I 2 = 51%), and the exclusion of studies with follow-up period <24 months, or grouping according to the type of GIC (conventional or resin-modified), or according to the type of isolation (cotton roll or rubber dam), or according to the evaluation criteria applied did not affect the pattern of the results obtained. GIC exhibited significantly lower values of secondary carious lesions (SCL) than CR (SCL: risk difference 0.06 [0.02, 0.10], p = 0.008, I 2 = 0%). The materials presented similar performance (p > 0.05) regarding the overall effect, as well as for marginal discoloration, marginal adaptation and anatomical form. The superiority of GIC was maintained when resin-modified GIC and rubber dam isolation were analyzed separately.
Conclusions
GIC and CR presented similar clinical performance for all criteria analyzed, except for secondary carious lesions, in which GIC presented superior performance, especially for the resin-modified GIC and with rubber dam isolation.
1
Introduction
Dental caries is one of the most prevalent diseases in the oral cavity, and its high prevalence is related to inadequate oral hygiene habits and ingestion of carbohydrate-rich foods [ ], as well as socioeconomic and behavioral factors. The proximal surfaces are the greatest contributors to the high prevalence of this disease [ ], especially in the primary dentition. Since the direct visual inspection of carious lesions in proximal surfaces is impaired by the presence of a contact surface between primary posterior teeth [ ], more invasive interventions are commonly performed, given that carious lesions in these surfaces are often detected in more advanced stages when compared with smooth surfaces. In these cases, restorative treatment is the most frequently performed.
Although amalgam restorations present high longevity [ ], their use has been increasingly discontinued, since they require more invasive operative techniques, demanding wear of intact tooth structure for adequate material retention [ ], in addition to concerns related to toxicity and environmental pollution [ ]. The aforementioned disadvantages, along with the poor esthetics of amalgam restorations, increased the attention to materials as composite resin (CR) and glass ionomer cement (GIC), due to the greater maintenance of intact tooth structure and their adhesion to the remaining tooth structure. These characteristics allow the use of more conservative restorative techniques, limiting the cavity preparation mainly to decayed tissue removal, thereby preserving the intact tooth structures.
Despite the favorable esthetic and mechanical properties of CRs, the restorative technique is more sensitive in relation to the use of GIC, as it involves a greater number of operative steps, combined to the higher sensitivity to moisture of this material. Consequently, in cases in which the use of rubber dam is inviable and/or patient compliance is limited, CR usually is not the first option of restorative material [ ], placing GIC as the material of choice. GICs are adhesive materials that release fluoride to the oral environment, and their insertion technique is faster compared to composite resins [ ], making this material an important resource for the treatment of children.
Isolated studies present conflicting evidence concerning the longevity of restorations in primary teeth, thus systematic literature reviews, especially involving meta-analysis, are a great tool to aid professionals in clinical decision-making. Within this context, two recent meta-analyses evaluated the performance of restorative materials (adhesive or not) in posterior primary teeth. One study [ ] concluded that adhesive materials with resin component (CR, resin-modified GIC and compomer) presented similar longevity rates to each other, with worst performance observed for silver-reinforced GIC. A subsequent study evaluated the effect of GIC restorations in the prevention of marginal carious lesions [ ]. The authors concluded that, while the rate of secondary caries was similar among the materials (amalgam, CR, polyacid-modified resin and compomer) for occlusal restorations, the clinical performance of GIC in occluso-proximal restorations was significantly better compared to the other groups.
It is noteworthy, however, that the aforementioned reviews [ ] gathered data on Class I and II cavities simultaneously, and also included restorative materials that are rarely used or have been increasingly discontinued, which may have largely influenced the results. It is known that the longevity of Class II restorations is significantly reduced compared to Class I preparations [ ], since the greater number of surfaces involved leads to a greater interface area between the tooth structure and the restorative material, in addition to the loss of the marginal ridge, which poses an occlusal overload on the restoration [ ]. Additionally, considering that CR and GIC (conventional or resin-modified, hereafter abbreviated as C-GIC and RM-GIC, respectively) are the materials most widely used in clinical practice compared to other materials available, a direct comparison of the clinical performance of GIC and CR, especially in Class II restorations, might offer relevant information for the treatment of carious lesions in primary posterior teeth. Finally, analysis of the influence of type of GIC (C-CIG or RM-GIC) and isolation (rubber dam or cotton roll) on the clinical performance of restorations might also provide relevant information, both for pediatric dentists and public health services.
Thus, this systematic review and meta-analysis evaluated the clinical performance of GIC and CR in Class II restorations in primary molars. As secondary outcomes, the study also evaluated the influence of type of GIC and isolation (rubber dam or cotton roll) on the clinical performance of restorations.
2
Material and methods
2.1
Protocol and registry
This study was registered in database PROSPERO (registry CRD42015027751) and followed the PRISMA guidelines on the Preferred Reporting Items for Systematic reviews and Meta-Analyses [ ].
2.2
Search strategy
An electronic search was performed on the following databases: PubMeb, Scopus, Web of Science, Virtual Health Library (VHL), OpenGrey, Clinical Trials and Cochrane Library. A specialized librarian guided the entire electronic search strategy. Hand search was also performed to identify manuscripts that might not have been retrieved by the electronic search.
To find unpublished or ongoing studies, the registry of clinical trials was investigated on the website ClinicalTrials.gov ( www.clinicaltrials.gov ), without restriction as to date or language of publication. Additionally, the grey literature (produced at governmental, academic, entrepreneurial and industrial levels, in printed or electronic format, yet not controlled by commercial publishers) was searched using the grey literature database OpenGrey ( http://www.opengrey.eu/http://www.opengrey.eu/ ).
The search strategy, as well as the date of search for all databases, is presented in Box 1 . This search strategy was properly adapted to each database.
2.3
Eligibility criteria
The study included randomized controlled clinical trials comparing the clinical performance of Class II restorations performed with composite resin (CR) and conventional or resin-modified glass ionomer cement (C-GIC or RM-GIC, respectively) in primary teeth, in children of any age, according to the PICOS strategy described below:
(P) – Population: children with need of Class II restorations in primary teeth;
(I) – Intervention: use of conventional (chemically cured) or resin-modified (light cured) glass ionomer cement;
(C) – Comparison: use of composite resin;
(O) – Outcome: the primary outcome refers to the clinical performance of restorations, evaluated according to the presence of secondary carious lesions, marginal discoloration, marginal adaptation, longevity, retention and wear of restorative material, and anatomical form. The secondary outcomes included the influence of the type of GIC (C-GIC or RM-GIC) and isolation (rubber dam or cotton roll) on the primary outcome;
(S) – Study design: randomized controlled clinical trials.
Editorial letters, pilot studies, historical reviews, in vitro, cohort, observational and descriptive studies, case reports and case series were excluded. The study also excluded investigations evaluating other types of cavities (Class I, III, IV and V) and analyses on polyacid-modified composite resins (compomers) and silver-reinforced GIC as restorative materials.
2.4
Study selection and data extraction process
Papers appearing in more than one database were considered only once. Two reviewers (AGAD and MBM) independently analyzed the titles and abstracts of papers found on the databases. Potentially eligible papers were read in full text to clearly determine their eligibility. Data were extracted using a form based on other systematic reviews, in which the following data were recorded: study details (year of publication and authors); criteria for definition and evaluation; details of study methods (study design and follow-up period); details of participants (mean age and number of patients); details of restorative materials employed (restorative protocols: isolation methods and restorative material), and results.
2.5
Risk of bias in individual studies
The methodological quality and risk of bias of the included studies were analyzed by two independent reviewers (AGAD and MBM), using the Cochrane Collaboration tool for analysis of risk of bias ( ). The evaluation criteria comprised six items: random sequence generation, allocation concealment, blinded evaluations of results, blinding of participants and staff, results with incomplete data, selective report of outcome, and other possible sources of bias. The six domains were evaluated and the included studies were classified. During evaluation of the risk of bias, any divergences between reviewers were solved by discussion and consensus and, if necessary, with the aid of a third reviewer (LCM).
For each aspect of quality analysis, the risk of bias for each domain was identified following the recommendations of the Cochrane Handbook for Systematic Reviews of Interventions 5.1.0 ( ). Each criterion was scored as “yes”, indicating low risk of bias; “no”, indicating high risk of bias; and “unclear”, indicating lack of information or uncertainty about the potential of bias.
Only four among the six Cochrane domains were considered as key domains to evaluate the risk of bias. The studies were considered as “low” risk of bias if there was adequate randomization, allocation concealment, incomplete data and selective reporting. The two domains evaluating blinding (operators, participants and examiners) were not considered as key domain due to differences both in the operative technique and in the clinical aspect of materials. When the study was scored as “unclear” in the key domains, attempts to contact the authors (one weekly contact, for up to 4 weeks) were made to achieve additional information and allow definitive scoring as “yes” or “no”.
2.6
Meta-analyses
Data on included studies were obtained and analyzed using the software Revman 5.3 (Review Manager v. 5, The Cochrane Collaboration; Copenhagen, Denmark). Eleven meta-analyses (MAs) were performed to evaluate:
- (1)
The percentage of failure of restorations in all selected studies;
- (2)
The percentage of failure of restorations in studies with follow-up period equal to or greater than 24 months;
- (3)
The percentage of failure of restorations, subgrouping by type of GIC: Subgroup 1: C-GIC x CR; Subgroup 2: RM-GIC x CR;
- (4)
The percentage of failure of restorations, subgrouping by type of isolation: Subgroup 1: rubber dam isolation; Subgroup 2: cotton roll isolation;
- (5)
The clinical performance of the main parameters (marginal adaptation (MA), marginal discoloration (MD), anatomical form (AF) and secondary carious lesions (SCL)) considering all selected studies;
- (6)
The clinical performance of the main parameters (MA, MD, AF and SCL) considering studies with follow-up period equal to or greater than 24 months;
- (7)
The clinical performance of the main parameters (MA, MD, AF and SCL) including only studies using RM-GIC;
- (8)
The clinical performance of the main parameters (MA, MD, AF and SCL) including only studies using C-GIC;
- (9)
The clinical performance of the main parameters (MA, MD, AF and SCL) including only studies using rubber dam isolation;
- (10)
The clinical performance of the main parameters (MA, MD, AF and SCL) including only studies using cotton roll isolation.
- (11)
The percentage of failure of restorations, subgrouping by type of evaluation criteria applied: Subgroup 1: USPHS criteria; Subgroup 2: FDI criteria, subgroup 3: Serpa et al., 2017 criteria.
According to the criteria for the evaluation of restorations, data on the main parameters analyzed were dichotomized as “acceptable” (restorations without need of replacement or repair) or “unacceptable” (restorations presenting failures or requiring repair or replacement), as shown in Table 1 . The prevalence of unacceptable (events) and the total number of restorations per group were used to calculate the risk difference, at a confidence interval of 95%. The random effect was applied and the heterogeneity (intrinsic divergence among studies) was evaluated by the I 2 index. Sensitivity analysis was further conducted to estimate and verify the influence of studies, one by one, on the pooled results when the heterogeneity was moderate or considerable (30%–100%) [ ].
| Parameters | USPHS modified criteria [ ] | FDI World Federation criteria [ ] | Serpa et al. criteria [ ] | |||
|---|---|---|---|---|---|---|
| Acceptable | Unacceptable | Acceptable | Unacceptable | Acceptable | Unacceptable | |
| Failure | 1, 2, 3 | 4, 5 | 0 | 1, 2, 3, 4 | ||
| radiographic criteria | ||||||
| 0, 1 | 2, 3, 4 | |||||
| clinical criteria | ||||||
| Marginal adaptation | Alpha Bravo | Charlie Delta | 1, 2, 3 | 4, 5 | ||
| Marginal discoloration | Alpha Bravo | Charlie Delta | 1, 2, 3 | 4, 5 | ||
| Anatomical Form | Alpha Bravo | Charlie | 1, 2, 3 | 4,5 | ||
| Secondary caries | Alpha | Bravo Charlie Delta | 1, 2, 3 | 4, 5 | ||
2
Material and methods
2.1
Protocol and registry
This study was registered in database PROSPERO (registry CRD42015027751) and followed the PRISMA guidelines on the Preferred Reporting Items for Systematic reviews and Meta-Analyses [ ].
2.2
Search strategy
An electronic search was performed on the following databases: PubMeb, Scopus, Web of Science, Virtual Health Library (VHL), OpenGrey, Clinical Trials and Cochrane Library. A specialized librarian guided the entire electronic search strategy. Hand search was also performed to identify manuscripts that might not have been retrieved by the electronic search.
To find unpublished or ongoing studies, the registry of clinical trials was investigated on the website ClinicalTrials.gov ( www.clinicaltrials.gov ), without restriction as to date or language of publication. Additionally, the grey literature (produced at governmental, academic, entrepreneurial and industrial levels, in printed or electronic format, yet not controlled by commercial publishers) was searched using the grey literature database OpenGrey ( http://www.opengrey.eu/http://www.opengrey.eu/ ).
The search strategy, as well as the date of search for all databases, is presented in Box 1 . This search strategy was properly adapted to each database.
2.3
Eligibility criteria
The study included randomized controlled clinical trials comparing the clinical performance of Class II restorations performed with composite resin (CR) and conventional or resin-modified glass ionomer cement (C-GIC or RM-GIC, respectively) in primary teeth, in children of any age, according to the PICOS strategy described below:
(P) – Population: children with need of Class II restorations in primary teeth;
(I) – Intervention: use of conventional (chemically cured) or resin-modified (light cured) glass ionomer cement;
(C) – Comparison: use of composite resin;
(O) – Outcome: the primary outcome refers to the clinical performance of restorations, evaluated according to the presence of secondary carious lesions, marginal discoloration, marginal adaptation, longevity, retention and wear of restorative material, and anatomical form. The secondary outcomes included the influence of the type of GIC (C-GIC or RM-GIC) and isolation (rubber dam or cotton roll) on the primary outcome;
(S) – Study design: randomized controlled clinical trials.
Editorial letters, pilot studies, historical reviews, in vitro, cohort, observational and descriptive studies, case reports and case series were excluded. The study also excluded investigations evaluating other types of cavities (Class I, III, IV and V) and analyses on polyacid-modified composite resins (compomers) and silver-reinforced GIC as restorative materials.
2.4
Study selection and data extraction process
Papers appearing in more than one database were considered only once. Two reviewers (AGAD and MBM) independently analyzed the titles and abstracts of papers found on the databases. Potentially eligible papers were read in full text to clearly determine their eligibility. Data were extracted using a form based on other systematic reviews, in which the following data were recorded: study details (year of publication and authors); criteria for definition and evaluation; details of study methods (study design and follow-up period); details of participants (mean age and number of patients); details of restorative materials employed (restorative protocols: isolation methods and restorative material), and results.
2.5
Risk of bias in individual studies
The methodological quality and risk of bias of the included studies were analyzed by two independent reviewers (AGAD and MBM), using the Cochrane Collaboration tool for analysis of risk of bias ( ). The evaluation criteria comprised six items: random sequence generation, allocation concealment, blinded evaluations of results, blinding of participants and staff, results with incomplete data, selective report of outcome, and other possible sources of bias. The six domains were evaluated and the included studies were classified. During evaluation of the risk of bias, any divergences between reviewers were solved by discussion and consensus and, if necessary, with the aid of a third reviewer (LCM).
For each aspect of quality analysis, the risk of bias for each domain was identified following the recommendations of the Cochrane Handbook for Systematic Reviews of Interventions 5.1.0 ( ). Each criterion was scored as “yes”, indicating low risk of bias; “no”, indicating high risk of bias; and “unclear”, indicating lack of information or uncertainty about the potential of bias.
Only four among the six Cochrane domains were considered as key domains to evaluate the risk of bias. The studies were considered as “low” risk of bias if there was adequate randomization, allocation concealment, incomplete data and selective reporting. The two domains evaluating blinding (operators, participants and examiners) were not considered as key domain due to differences both in the operative technique and in the clinical aspect of materials. When the study was scored as “unclear” in the key domains, attempts to contact the authors (one weekly contact, for up to 4 weeks) were made to achieve additional information and allow definitive scoring as “yes” or “no”.
2.6
Meta-analyses
Data on included studies were obtained and analyzed using the software Revman 5.3 (Review Manager v. 5, The Cochrane Collaboration; Copenhagen, Denmark). Eleven meta-analyses (MAs) were performed to evaluate:
- (1)
The percentage of failure of restorations in all selected studies;
- (2)
The percentage of failure of restorations in studies with follow-up period equal to or greater than 24 months;
- (3)
The percentage of failure of restorations, subgrouping by type of GIC: Subgroup 1: C-GIC x CR; Subgroup 2: RM-GIC x CR;
- (4)
The percentage of failure of restorations, subgrouping by type of isolation: Subgroup 1: rubber dam isolation; Subgroup 2: cotton roll isolation;
- (5)
The clinical performance of the main parameters (marginal adaptation (MA), marginal discoloration (MD), anatomical form (AF) and secondary carious lesions (SCL)) considering all selected studies;
- (6)
The clinical performance of the main parameters (MA, MD, AF and SCL) considering studies with follow-up period equal to or greater than 24 months;
- (7)
The clinical performance of the main parameters (MA, MD, AF and SCL) including only studies using RM-GIC;
- (8)
The clinical performance of the main parameters (MA, MD, AF and SCL) including only studies using C-GIC;
- (9)
The clinical performance of the main parameters (MA, MD, AF and SCL) including only studies using rubber dam isolation;
- (10)
The clinical performance of the main parameters (MA, MD, AF and SCL) including only studies using cotton roll isolation.
- (11)
The percentage of failure of restorations, subgrouping by type of evaluation criteria applied: Subgroup 1: USPHS criteria; Subgroup 2: FDI criteria, subgroup 3: Serpa et al., 2017 criteria.
According to the criteria for the evaluation of restorations, data on the main parameters analyzed were dichotomized as “acceptable” (restorations without need of replacement or repair) or “unacceptable” (restorations presenting failures or requiring repair or replacement), as shown in Table 1 . The prevalence of unacceptable (events) and the total number of restorations per group were used to calculate the risk difference, at a confidence interval of 95%. The random effect was applied and the heterogeneity (intrinsic divergence among studies) was evaluated by the I 2 index. Sensitivity analysis was further conducted to estimate and verify the influence of studies, one by one, on the pooled results when the heterogeneity was moderate or considerable (30%–100%) [ ].
| Parameters | USPHS modified criteria [ ] | FDI World Federation criteria [ ] | Serpa et al. criteria [ ] | |||
|---|---|---|---|---|---|---|
| Acceptable | Unacceptable | Acceptable | Unacceptable | Acceptable | Unacceptable | |
| Failure | 1, 2, 3 | 4, 5 | 0 | 1, 2, 3, 4 | ||
| radiographic criteria | ||||||
| 0, 1 | 2, 3, 4 | |||||
| clinical criteria | ||||||
| Marginal adaptation | Alpha Bravo | Charlie Delta | 1, 2, 3 | 4, 5 | ||
| Marginal discoloration | Alpha Bravo | Charlie Delta | 1, 2, 3 | 4, 5 | ||
| Anatomical Form | Alpha Bravo | Charlie | 1, 2, 3 | 4,5 | ||
| Secondary caries | Alpha | Bravo Charlie Delta | 1, 2, 3 | 4, 5 | ||
3
Results
3.1
Selection of studies
After reading the titles and abstracts, 2722 duplicates were removed, with identification of 2937 studies. Among the 25 potentially eligible studies (read in full text), 15 were excluded due to the following aspects: lack of adequate control (studies without control group) (n = 2), retrospective study (n = 1), sample overlapping (n = 1), microorganisms count as the main response variable (n = 1), study protocols without results (n = 2), presentation of the results of class I and class II restorations without distinction between the two types of cavities (n = 2), presentation of the results of permanent and primary teeth without distinction between the two dentitions (n = 1) and restorations in permanent teeth (n = 5). Thus, ten papers remained for the qualitative synthesis, and nine for the quantitative analysis. The flowchart of study selection is described in Fig. 1 .
3.2
Characteristics of studies included
The characteristics of the 10 studies are listed in Table 2 . The follow-up period of Class II restorations in primary teeth in the included studies ranged from 6 to 48 months. The design of all 10 papers analyzed comprised randomized clinical trials. Six studies presented split-mouth design [ ] and five studies had parallel design [ ]. Seven studies applied USPHS criteria [ ], two applied the FDI criteria [ ] and one study applied its own criteria [ ]. The number of restorations ranged from 75 to 344, and the number of participants in each study varied from 31 to 180 children, aged 3 to 11 years. Seven papers reported restorations performed using rubber dam isolation [ ], while 3 studies reported the use of cotton roll isolation [ ].
