Eligibility Criteria

The following selection criteria were adopted:

1. Study design: randomized clinical trials (RCTs) or nonrandomized trials (CCT), prospective or retrospective;

2. Population: patients with fixed appliances needing TADs;

3. Intervention: patients in whom stainless steel (SS) MIs were used;

4. Comparison: patients in whom titanium MIs were used;

5. Outcome: success rate of MIs; and

6. Exclusion criteria: Patients with any bone disease or craniofacial syndrome; studies that used other types of TADs; opinion articles, animal or laboratorial studies, case reports, case-series, and literature reviews.

Information Sources

The following databases were searched: PubMed, Scopus, Web of Science, Cochrane, LILACS, OpenGray, and Google Scholar. A hand search was conducted by reading the references of the included articles for eventual additional relevant studies. The search was extended to the database clinicaltrials.gov. No restriction on language or date of publication was applied. The search was continued until July 2019.

Search Strategy and Study Selection

The databases were searched independently by two reviewers (P.M. and D.G.E.). Disagreements were settled through discussion and consensus and, when necessary, a third author's opinion (D.N.) was consulted. The search strategy was developed through a combination between MeSH, entry terms, and keywords related to the PICO (Population, Intervention, Comparison, Outcome) strategy using Boolean operators. The full search strategy for each database is illustrated in Table 1.

Table 1.  Search Strategy for Electronic Databases

View Fullscreen

The citations were saved in a reference manager (EndNote, x9 version; Clarivate Analytics, Philadelphia, Pa) and, at first, titles and abstract were analyzed according to the eligibility criteria. The selected articles were evaluated by full text, and a final selection was determined.

Risk of Bias Assessment

For the CCTs, the risk of bias (RoB) was performed following the ROBINS-I (Risk of Bias in Non-randomized Studies-of Interventions) tool.¹⁷ The checklist included the following three main domains of bias: preintervention, intervention, and postintervention. The RoB was judged for each domain and for overall evaluation as low, moderate, serious, critical, or no information (Table 2).

Table 2.  Bias and Domains Considered in Risk of Bias Evaluation According to the ROBINS-I (Risk of Bias in Non-randomized Studies-of Interventions) Tool

View Fullscreen

For the RCTs, the Cochrane RoB 2.0 tool¹⁸ was used. This tool assessed the following five main bias domains: bias arising from the randomization process, bias as a result of deviations from the intended interventions, bias as a result of missing outcome data, bias in the measurement of the outcome, and bias in the selection of the reported results. The RoB was judged for each domain and for overall evaluation as low, some concerns, or high. Each analysis of RoB was made by two researchers (P.M. and D.G.E.) and, in the case of disagreement, a third reviewer (D.N.) was consulted.

Evaluation of Quality of Evidence

The included studies were given a quality grade related to the MI success rate in accordance with the grading of recommendation, assessment, development, and evaluation (GRADEpro Guideline Development Tool, gradepro.org).¹⁹ This tool considered five aspects for rating the quality of evidence as high, moderate, low, or very low.

Data Extraction

Two reviewers collected the data independently (P.M. and D.G.E.), recording the following items: author, year and location, type of study, participants, type of MI and their characteristics, loading force, and healing time, MI location, follow-up period, purpose of installation, statistical analysis, outcome assessment, and outcomes. The quantitative data analysis was evaluated through risk ratio (RR). A meta-analysis was not performed because of the large methodological heterogeneity among the studies examined, principally because of the different sizes and location of MIs, as well as the amount of force used.

Study Selection

The electronic search revealed a total of 1680 articles: 327 from PubMed, 512 from SCOPUS, 217 from Web of Science, 20 from Cochrane, 52 from LILACS, 540 from Google Scholar, 12 from Clinical Trials, and 0 from OpenGray. After removing duplicates, 1261 studies remained. One article was added for screening after a hand search of the references in the included articles. After reading the titles and abstracts, 20 articles were evaluated by full text, and 14 were excluded. The reasons for exclusion were as follows: in vitro studies (n = 7), evaluation of only one type of MI (n = 3), animal studies (n = 2), and study not related to the research objective (n = 1; Table 3). As a result, six articles were included (Figure 1).

Table 3.  List of Excluded Studies With Reason

View Fullscreen

View Figure

• Download as Powerpoint
• Download Figure

Study Characteristics

The characteristics of the included studies are described in Table 4. Selected articles were published between 2009 and 2019. Five studies were CCTs, four^3,14,15,20 were prospective and one²¹ retrospective. All CCTs investigated interradicular MIs. Only one²² study was a RCT that investigated extraalveolar MIs.

Table 4.  Summary of the Data From the Studies Included in This Review^a

View Fullscreen

Table 4.  Extended

View Fullscreen

The mean age of participants ranged from 16.2¹⁵ to 29.6²¹ years. Great differences among the articles were observed regarding the number of MIs used that ranged from 10¹⁴ to 386²² per group. Only one study¹⁴ did not describe the number of male/female subjects and the mean age of patients.

In two articles,^14,15 the MIs were installed mesially to the molars, whereas one²² used extraalveolar MIs in the infrazygomatic crest. It was observed that the loading forces ranged according to the purpose of installation: for maxillary retraction, the highest load used was 227 to 397 g,²² whereas the canine retraction required the lowest forces, 90 to 100 g.¹⁵ Three studies^3,20,21 did not report the loading forces applied and did not standardize the installation sites and purposes of MI.

In relation to drill type, three studies^3,15,21 used different types of drills for the titanium and SS groups. One²¹ article used predrilled titanium MIs and self-drilling SS MIs. In another study, titanium MIs were self-drilling, whereas the SS MIs were predrilled¹⁵ In a third study,³ three groups were used, one composed of self-drilling SS MIs, another composed of self-drilling titanium MIs, and a third composed of predrilled titanium MIs. The greatest follow-up period was 12 months,^14,20 and the lowest was 160.8 days.¹⁵ Only one study did not report the follow-up period.²¹

RoB Within Studies

Regarding the CCTs, two^15,20 showed moderate RoB as a result of the differences in treatment between groups and the selection of results presented, whereas the other three articles presented a high RoB.^3,14,21 One²¹ of those was a retrospective study in which important information was lacking such as the force applied, follow-up period, installation purpose, and eligibility criteria. An error in the statistical analysis was found in one study.¹⁴ When the statistical analysis of this article was reevaluated, no differences were detected in any of the following statistical tests: chi-square, which was used in the article; Fisher exact test; G test; binomial; or RR (BioEstat, version 5.3; Mamirauá Institute, Belém, Pará, Brazil). For this reason, the study¹⁴ received a high RoB. Another study³ received a high RoB as a result of discrepancies between MI numbers by group, differences in MI locations, absence of clear eligibility criteria, and poor description of interventions (applied force, purpose of installation).

The RCT²² exhibited a low RoB as a result of an adequate sample along with standardization of the site of installation, the force used, the follow-up time, and the purpose of use for the two groups. The evaluations of RoB for all included studies are presented in Tables 5 and 6.

Table 5.  Risk of Bias in Nonrandomized Studies Selected

View Fullscreen

Table 6.  Risk of Bias in the Randomized Study

View Fullscreen

Results of Individual Studies

The lowest success rate found for SS MIs was 50%,¹⁴ and the highest was 100%.¹⁵ Regarding titanium MIs, the lowest success rate found was 84.46%,²⁰ and the highest was 100%.¹⁵

Only one study reported that titanium MIs showed higher success rates than SS MIs and that a higher implant failure was found in the upper jaw when compared with the lower jaw.¹⁴ The other five studies—one RCT²² and four CCTs^3,15,20,21—did not report statistical differences between the success rates of the two materials. One article did not report MI losses in either group.¹⁵

Age was a determinant factor for MI failures in two studies,^20,21 but divergent results were found. In one article,²⁰ older patients were more susceptible to MI failures, resulting in a 5% increase in failure risk for every 1-year increase in age among participants older than 30 years, whereas in another²¹ study, patients younger than 35 years presented a higher risk of MI failure than those older than 35 years. In two other studies,^3,22 age was not a determinant factor. A MI with a longer length²⁰ was cited as more prone to success, whereas another article³ found that diameter and length were not factors associated with MI failure. Installation in the attached gingiva²¹ was also cited as a major factor for MI stability.

Synthesis of Results

A meta-analysis was not performed because of the heterogeneity of methodologies, mainly because of the different sizes and locations of MIs, as well as the amount of force used. The RR for each study was analyzed, and no differences between the groups were found (Table 7).

Table 7.  RR of the Studies^a

View Fullscreen

Assessment of the Quality of Evidence

The evaluation of the evidence according to GRADE is described in Table 8. The quality of evidence was rated as high for the RCT as this work had excellent control of confounding factors, an adequate sample size, and few limitations. For the CCTs, the quality of evidence was rated as moderate because of the limitations of the study designs and differences between the intervention of groups.

Table 8.  GRADE Evidence Profile Table

View Fullscreen

Summary of Evidence

Most of the articles included^3,15,20–22 found no difference between the two materials. In these studies, the success rate for SS MIs ranged from 74.6%²¹ to 100%¹⁵ and for titanium ranged from 80.9%²¹ to 100%.¹⁵ Only one CCT reported a statistical difference between the groups.¹⁴ Nevertheless, statistical analysis was reevaluated, and no significant difference was detected. Because of this, this study showed a high RoB, and its conclusions should be analyzed with caution.

It is important to emphasize that some studies did not standardize the installation sites and purpose of MIs,^3,20,21 did not specify the time of follow-up,²¹ and did not measure the forces used.^3,20,21 If these factors might influence the stability of the MIs, it is essential that they be controlled and standardized between the two groups. Some studies that did not control the confounding factors obtained high rates of failure,^20,21 whereas those that controlled these factors had success rates above 90%,^15,22 and in one article there was no loss of MIs in any group.¹⁵ The high success rate found in that study¹⁵ may have been the result of the purpose of the MIs because canine distalization required the use of a lighter force. In addition, the installation occurred in the attached gingiva in patients with good oral health, characteristics that are predisposed to a higher success rate. Some included articles reported that the installation site,²¹ patient age,^20,21 jaw of insertion,¹⁴ and MI length²⁰ were factors associated with stability. A recent meta-analysis,³⁰ however, showed that MI diameter, length and design, patient age, and jaw of insertion had little effect on the failure rate. Insertion into attached gingiva was strongly related to higher success rates, whereas smoking was related to failure. Another meta-analysis¹² concluded that insertion in palatal sites and mainly no contact with roots were both associated with great success rates.

The literature reported that titanium MIs need prior drilling in higher density bones^26,28 to avoid fractures. However, a meta-analysis³¹ showed no difference in failure rates whether the MIs were self-drilling or predrilled. Three studies^3,15,21 included in this review did not use the same type of drilling for both groups. One article,¹⁵ which used self-drilling titanium MIs, obtained a success rate of 100%, the same rate found for predrilled SS MIs. In another article,²¹ titanium MIs were predrilled and SS MIs were self-drilling, and the success rates were the lowest for both materials among the articles included in this review; nonetheless, no differences were found between the groups. A third study used self-drilling SS MIs and both predrilled and self-drilling titanium MIs and, despite the differences among success rates (80% for SS MIs, 97.2% for self-drilling titanium MIs, 89.4% for predrilled titanium MIs), they were not statistically significant. Therefore, the type of drilling did not seem to be an important factor for the stability of the MIs.

The ROBINS-I tool evaluated the RoB in two studies as moderate^15,20 and as high in three others^3,14,21 as a result of limitations such as differences in treatment of groups, selection of reported results, and incorrect statistical analysis. The RCT²² presented a low RoB mainly because it used an adequate sample size and standardized the site of installation, the force used, the time of follow-up, and the purpose of use for both groups.

The quality of evidence of MI success rate was measured using the GRADE tool and presented results consistent with the assessment of RoB. The quality of evidence was high for the RCT once this article satisfactorily controlled confounding factors and possible bias. This was not seen in most CCTs, generating a moderate quality of evidence. Even with the need for better designed studies, the available evidence indicated that the material used to manufacture MIs appeared not to be an important variable for stability. The RR of included studies ranged from 0.85 to 5, but no statistical difference was found in any article (Table 7), demonstrating that the MI material was not a risk factor for failure.

Limitations

The CCTs^{3,14,15,20,21} included in this review had some limitations in their methodologies and study designs, which impacted their RoB assessment. Factors such as patient hygiene, installation site, and surgeon ability influenced the stability of the MI. Consequently, these factors should be controlled in all studies comparing SS and titanium MIs. Although not all studies did this, the success rates between groups were always close in almost all of the articles, with the exception of one,¹⁴ which did not present reliable evidence.

Another possible limitation was that the best available article, a RCT,²² evaluated extraalveolar MIs, which were used for different purposes and had different sizes than the interradicular MIs. Therefore, a RCT evaluating interradicular MI is necessary mainly because of the limited quality of the existing CCTs. Nonetheless, the aim of this review was to assess the influence of the material used to manufacturing the MI on its success rate, no matter if it was installed between the roots or in the infrazygomatic crest. In the RCT, there was control of possible confounding factors and randomization of the samples, so any possible difference between the groups would have been a result of the material rather than to other factors.

Clinical Considerations

Considering the available limited evidence, the results showed that the material used, titanium or SS, was not relevant to the stability of MIs. Accordingly, SS MIs are a good option because they have a lower cost and have shown clinical results similar to those made of titanium.