Clinical performance of zirconium implants compared to titanium implants: a systematic review and meta-analysis of randomized controlled trials

PICOS question and eligibility criteria

Articles were included in this systematic review if they met the following inclusion criteria formulated using the PICOS question: participants (P), patients with one or more missing teeth who received one or more dental implants; intervention (I), veneered or monolithic Zr dental implants; comparison (C), traditional Ti implants. We did not restrict the type of abutment or prosthetic restoration. The primary outcomes (O) were survival rate, success rate, marginal bone loss (MBL), and probing pocket depth (PPD); secondary outcomes included plaque index (PI), bleeding on probing (BOP), pink aesthetic score (PES), and reasons for failure. The study design (S) of interest was randomized controlled trials (RCT).

Search strategy

Cochrane Central Register of Controlled Trials (CENTRAL), Medline via Ovid, EMBASE, and Web of Science were searched from inception to August 2022 by two authors (C. D. and L. Y.). The strings of the specific search strategies for these databases are presented in Table S1. The reference lists of the included studies were manually checked to identify potentially relevant studies. The articles obtained were imported into EndNote X9 (Thomson Reuters, Philadelphia, PA, USA) for screening.

Study selection

The titles and abstracts of all articles identified from the electronic database search were independently reviewed by two researchers (C. D. and L. Y). Full texts of potentially relevant articles were retrieved and further evaluated. Disagreements were resolved through discussion or by consulting an arbitrator (Y. C.). The reasons for exclusion were recorded for all the full texts. The study centers, ethical committee approval number, and full author lists of the selected studies were double-checked to detect replicated publications that reported on the same trial before being included.

Data extraction

Data extraction was performed independently by two researchers (C. D. and L. Y.) using a pre-designed form. When necessary, the corresponding authors of included studies were contacted to obtain additional information. The following data were extracted: authors’ names, publication date, country of origin, participants’ demographic characteristics, follow-up, intervention type (abutments, implants, or prostheses), and outcomes. Disagreements were settled through discussion with an arbitrator (Y. C.). Success was defined as the absence of biological or technical implant complications. A functional implant that still failed to meet the success criteria was considered to have survived. Failure occurred when the implant was removed for any reason.

Risk of bias assessment

The bias of the included RCTs was evaluated using the Cochrane Risk of Bias Assessment Tool via RevMan software (version 5.3, The Cochrane Collaboration, Copenhagen, Denmark) (Cochrane Training, 2023).

Two reviewers (C. D. and L. Y.) judged the risks as either “High”, “Low”, or “Unclear” based on seven domains: random sequence generation (selection bias), allocation concealment (selection bias), blinding of participants and personnel (performance bias), blinding of outcome assessment (detection bias), incomplete outcome data (attrition bias), selective reporting (reporting bias), and other bias. Disagreements were resolved through discussion between the two reviewers (C. D. and L. Y.). Any unresolved divergence was settled by an arbitrator (Y. C.).

Statistical analysis

The meta-analysis was performed using Review Manager 5.4.1 (The Cochrane Collaboration, London, UK) and STATA 15.0 (StataCorp LLC, College Station, TX, USA) software. Dichotomous variables were described as a percentage of the total number of events. The difference was calculated using the risk ratio (RR). Continuous variables were presented as mean and standard deviation and differences were evaluated using pooled mean difference (MD). The 95% confidence interval (CI) and P value were also calculated.

A Q-test based on the Cochran Chi-square test and Higgins I-squared statistic (I²) was used to detect statistical heterogeneity in this meta-analysis. A fixed-effects model was selected when I² was < 50%. When methodological heterogeneity was present, a sub-group analysis was performed. The subgroups were divided according to the follow-up duration or mucositis status. The stability of the results was evaluated using sensitivity analysis, which was performed by excluding one study from the meta-analysis and overserving whether the pooled results were significantly changed. If more than six studies were combined in the meta-analysis, funnel plots were used to assess publication bias. Qualitative descriptions were presented when data could not be pooled.

Certainty of evidence

The GRADE pro system (https://www.gradepro.org/) was used to evaluate the certainty of the evidence for all the outcomes (Schünemann et al., 2008). The included studies were RCTs with the highest pre-determined certainty of evidence. The certainty of evidence could be described as high, moderate, low, or ‘very low’ depending on the following five domains: risk of bias, inconsistency of effect, indirectness, imprecision, and publication bias (Guyatt et al., 2008).

Study selection

A total of 4,878 articles were identified through electronic and manual searches and 1,746 duplicates were excluded via EndNote X9. After initial screening by titles and abstracts, 3,107 records were excluded from the initial screening. We reviewed the full texts of the remaining 29 articles and finally included four RCTs with six records (Bienz et al., 2021; Koller et al., 2020; Osman et al., 2014b; Payer et al., 2015; Ruiz Henao et al., 2021; Siddiqi et al., 2015). A flow diagram of this study is presented in Fig. 1A.

Study characteristics

The basic characteristics of the included studies are presented in Table 1. These studies were published between 2014 and 2021, with a mean follow-up period of 23 months. A total of 109 participants with 248 implants were evaluated for comparison purposes. All implants in the RCTs were placed using the delayed approach. Payer et al. (2015) used two-piece implants (Koller et al., 2020), while the rest of the studies used monotype or one-piece implants. In addition, Osman et al. (2014b) rehabilitated completely edentulous patients (Siddiqi et al., 2015), whereas other researchers have only partially restored edentulous dentition. Moreover, only Bienz et al. (2021) compared Zr and Ti implants under healthy and experimental mucositis conditions.

Risk of bias

The quality assessment and risk of bias of the included RCTs are shown in Fig. 1B. Two studies were assessed as having an elevated risk of bias due to excessive loss during follow-up (Osman et al., 2014b) or baseline incomparability between the control and experimental groups (Payer et al., 2015). Moreover, we determined that the risk of bias was unclear in this study. Other studies were assessed to have a minimal risk of bias.

Data synthesis

(1) Survival rate: Four articles provided survival data and were included in this meta-analysis (Koller et al., 2020; Osman et al., 2014b; Payer et al., 2015; Ruiz Henao et al., 2021). No significant difference was found between Zr and Ti in the survival rate (RR = 0.91, CI [0.82–1.02], P = 0.100, I² = 0%) (Fig. 2A). No significant heterogeneity was observed in the subgroups with different follow-up durations.

A total of 24 Zr implants failed for several reasons, including osseointegration failure (n = 18), fracture (n = 3), adequate torque verification (n = 2), and excessive occlusal loading (n = 1). However, only 11 Ti implants failed because of osseointegration failure (n = 10) and peri-implantitis (n = 1).

(2) Success rate: Four articles provided survival data and were included in the meta-analysis (Koller et al., 2020; Osman et al., 2014b; Payer et al., 2015; Ruiz Henao et al., 2021). A statistically significant difference was found in the survival rate of different implant materials (RR = 0.87, CI [0.78–0.98], P = 0.030, I² = 0%) (Fig. 2B). There was no significant heterogeneity between the subgroups at the different follow-up times.

The reasons for the lack of success in the Ti group are described above, while there are three more implants in the Zr group owing to their deep placement.

(3) MBL: Three studies with four reports were included in the meta-analysis (Koller et al., 2020; Osman et al., 2014b; Payer et al., 2015; Ruiz Henao et al., 2021). One study reported pre- and post-changes in MBL, while others only reported MBL at the end of the follow-up period. Osman et al. (2014b) favored Ti rather than Zr implants because of decreased MBL (Ti: 0.18 ± 0.47 mm, Zr: 0.42 ± 0.40 mm, P = 0.024). A meta-analysis of other studies showed a significantly higher difference in the MBL of Zr implant materials compared to Ti implant materials (MD = 0.25, CI [0.02–0.49], P = 0.033, I² = 0%) (Fig. 2C). There was no heterogeneity in the subgroup follow-up durations.

(4) PPD: Three articles provided survival data and were included in this meta-analysis (Bienz et al., 2021; Ruiz Henao et al., 2021; Siddiqi et al., 2015). The two subgroups were divided according to mucosal conditions (normal versus experimental mucositis). The results showed no significant difference between Zr and Ti in terms of PPD (MD = −0.07, CI [−0.19–0.05], P = 0.250, I² = 31.3%) (Fig. 2D). Mucosal conditions may explain the source of heterogeneity.

(5) BOP: Bienz et al. (2021) and Koller et al. (2020) reported no significant differences in BOP between Zr and Ti implants under healthy conditions, but the results were more favorable for Zr implants under experimental mucositis conditions, with a significantly lower BOP (Zr: 21.7 ± 23.6%, Ti: 32.5 ± 27.8%, P < 0.001).

(6) PI: Bienz et al. (2021) noted that the results were more favorable for Zr implants due to a lower PI, with a significant difference under mucositis conditions (Zr: 68.3 ± 31.9%, Ti: 75.0 ± 29.4%, P < 0.001), while Siddiqi et al. (2015) and Koller et al. (2020) concluded that PI was not significantly different between Zr and Ti implants.

(7) PES: Two studies with three reports that assessed PES suggested a positive performance of Zr implants to improve aesthetic outcomes (Koller et al., 2020; Payer et al., 2015; Ruiz Henao et al., 2021). In the gingival margin position, the contour of the labial surface and the color and texture of this peri-implant tissue, the results were slightly better when a Zr implant was placed (Ruiz Henao et al., 2021). Zr implants showed a significantly higher difference (P = 0.004) compared to Ti implants at baseline and the 6-, 12-, 18-, and 24-months follow-up, but this difference was not observed at the 80-month follow-up (P = 0.428) (Koller et al., 2020; Payer et al., 2015).

Sensitivity analysis and publication bias

All sensitivity analyses showed the robustness of the meta-analysis (Figs. S1–S4). Publication bias was not assessed for all outcomes due to the limited number of included studies.

Certainty of evidence

The GRADE certainty of evidence ranged from moderate to very low (Table S2). The success rates were downgraded to moderate, the survival rate, PPD, BOP, and PES were low, and the MBL and PI were very low. The grade of the evidence was lowered by limitations in study quality, inconsistency of results, or imprecision of the study design.