A Randomised Controlled Trial Comparing a One‐Piece Zirconia Tissue‐Level Implant and Two‐Piece Titanium Bone‐Level Implant for Maxillary Single Tooth Replacement—3 year Results Authors: J. H. W. de Beus, M. S. Cune, L. Muijselaar, J. W. A. Slot, C. Jensen‐Louwerse, S. la Bastide‐van Gemert, H. J. A. Meijer, G. M. Raghoebar, U. Schepke Published: 2026 ABSTRACT ABSTRACTObjectivesThis RCT evaluated and compared the clinical performance of zirconia and titanium dental implants placed in the maxillary premolar region. Outcomes were assessed 3 years after prosthetic loading and included marginal bone level (MBL) changes, clinical parameters, aesthetic outcomes, and patient related outcome measures (PROMs).Materials and MethodsFifty patients were randomly allocated to receive either a zirconia one‐piece (ZrO2, n = 25) implant or a titanium bone‐level (Ti, n = 25) implant, restored with a lithium disilicate crown 3 months after implant placement. Primary outcome was change in MBL. Secondary outcomes included implant survival, peri‐implant tissue health, technical complications, aesthetic outcomes, and PROMs. Examinations were performed 3 years after crown placement.ResultsAfter 3 years, mean MBL change was −0.02 mm (SD = 0.78) for ZrO2 and −0.04 mm (SD = 0.36) for Ti; no statistically significant difference was found (p = 0.635). One ZrO2 implant failed to osseointegrate. Clinical parameters were favourable for both implant types, although slightly higher bleeding scores were observed for ZrO2. Differences in aesthetic outcomes and PROMs were found, favoring the ZrO2 implant group. None of these differences reached statistical significance.Conclusion and Clinical ImplicationsAfter 3 years of function, the ZrO2 and Ti implant as applied in this study showed comparable MBL changes. Survival, clinical and aesthetic parameters, and PROMs were generally favourable and largely comparable for both groups. These short‐term findings indicate that both implant types represent a viable option for clinical use in this application.Trial RegistrationCCMO registration: NL58957.042.16 Introduction Zirconia (ZrO2) implants are increasingly considered a relevant complement to titanium (Ti) implants in current implant dentistry, while titanium implants still remain the gold standard in clinical practice (Al‐Nawas et al. 2023; Sanz et al. 2019). Over the past decade, the clinical use of ZrO2 implants has expanded substantially, reflecting growing clinical and patient‐driven interests (Mohseni et al. 2024). Choosing ZrO2 implants may have several reasons, including patient preference for metal‐free restorations, reported or suspected hypersensitivity reactions to titanium, and aesthetic considerations, particularly in patients with a thin peri‐implant soft tissue phenotype (Sicilia et al. 2008; van Brakel, Cune, et al. 2011). From a biological perspective, concerns have been raised regarding the release of titanium particles and ions as a result of corrosion or wear, which have been shown to accumulate in peri‐implant tissues and might influence local immune reactions (Fretwurst et al. 2016; Ivanovski et al. 2022; Pettersson et al. 2019). In this context, an advantage of ZrO2 implants is the absence of corrosive degradation products. In addition, the white, tooth‐coloured appearance of zirconia may reduce the risk of mucosal discolouration compared with titanium, thereby offering potential aesthetic benefits, particularly in visually demanding regions (Thoma et al. 2016; van Brakel, Noordmans, et al. 2011). ZrO2 has been introduced as a metal‐free implant material combining favourable biomechanical characteristics with good biocompatibility (Kniha et al. 2021; Kohal et al. 2020). ZrO2 implants have demonstrated comparable osseointegration and mechanical stability to Ti implants (Bethke et al. 2020; Schepke, Meijer, Meijer, et al. 2017; Schepke, Meijer, Kerdijk, et al. 2017). Available clinical evidence indicates that ZrO2 implants used for fixed reconstructions—particularly for single tooth replacement—demonstrate good short term survival rates that are comparable to those of titanium implants during the first years of function (Padhye et al. 2023; Pieralli et al. 2017; Roehling et al. 2018). Furthermore, ZrO2 implants show favourable short‐ to medium outcomes with respect to marginal bone level (MBL) changes and peri implant tissue conditions, with results largely comparable to those reported for Ti implants (Kohal et al. 2020; Koller et al. 2020; Zuercher et al. 2024). Nevertheless, heterogeneity in study design, implant systems and follow‐up continues to limit the generalisability of these findings (Morena et al. 2024; Neugebauer et al. 2023; Padhye et al. 2023), and evidence beyond short‐term follow‐up remains limited. Randomised controlled trials comparing the material within the same study are still scarce (Duan et al. 2023; Mohseni et al. 2024; Morena et al. 2024). The 1‐year outcomes of the present randomised controlled trial comparing ZrO2 and Ti implants demonstrated largely comparable results for both implant materials (de Beus et al. 2024). While these findings support the short‐term clinical applicability of ZrO2 implants, extended follow‐up is required to evaluate the stability of these outcomes over time and to identify potential complications over time. The aim of the present study was to provide additional insights into the clinical application of ZrO2 implants by assessing the 3‐year clinical performance of ZrO2 and Ti implants used for single‐tooth replacement in the maxillary premolar region and to compare the treatment outcomes. Primary outcome was marginal bone level (MBL) change after 3 years of function. The null hypothesis was that there was no difference in MBL changes around ZrO2 and Ti implants after 3 years of clinical loading. Secondary outcomes included implant survival, peri‐implant soft tissue health, technical complications, aesthetic outcomes, and patient‐reported outcome measures (PROMs). Materials and Methods Study Design and Patients' Selection This prospective randomised controlled trial follows the same design and clinical protocol as previously described in the 1‐year outcomes of this study (de Beus et al. 2024). The current study describes the 3‐year observation period. Eligible participants (for in‐ and exclusion criteria see Table 1) requiring a single tooth implant‐borne restoration in the maxillary premolar region were recruited at the Department of Restorative Dentistry as well as the Department of Oral and Maxillofacial Surgery at the University Medical Center Groningen The Netherlands. A total of 50 participants were enrolled and randomly assigned, via computer‐generated randomisation (Sealed Envelope Ltd. 2022. Simple randomisation service. [Online] Available from: https://sealedenvelope.com/simple‐randomiser/v1/ [Accessed 11 Feb 2017]), to receive either the ZrO2 or the Ti implant supported restoration complex. TABLE 1Inclusion and exclusion criteria.Inclusion criteriaMissing first or second premolar in the maxillaWish to replace the missing premolar with an implantWilling to sign for informed consentBone height ≥ 10 mm beneath the maxillary sinusExclusion criteriaMissing teeth mesial or distal from implantation siteBone augmentation needed prior to implant placementOrthodontic treatment at the time of impression takingSevere bruxismAcute periodontitisHistory of implant lossDocumented extreme gagging reflexPoor medical condition (ASA a score 3 or higher)Previous therapeutic radiation of the head ‐neck regionChronic pain in orofacial systemYounger than 18 years at time of inclusionReduced mental capacity a American Society of Anaesthesiologists. Inclusion and exclusion criteria. American Society of Anaesthesiologists. Ethics Statement This study adhered to the ethical principles outlined in the Declaration of Helsinki and was approved by the Medical Ethics Committee of the University Medical Center Groningen (METc 2016.610; CCMO registration NL58957.042.16). Prior to inclusion, patients were informed about the study protocol and provided their written informed consent. Surgical and Prosthetic Procedures Implant placement was performed following a minimum post‐extraction healing period of 3 months. Patients received antibiotic prophylaxis 1 h prior to surgery (2 g amoxicillin, or 600 mg clindamycin alternatively) and were instructed to rinse twice daily with 0.12% chlorhexidine from 1 day before until 10 days after surgery. All surgical procedures were conducted under local anaesthesia by experienced surgeons. Participants in the zirconia group received a one‐piece ZrO2 implant (Straumann Pure, Straumann Group, Basel, Switzerland). Those in the titanium group were treated with a bone‐level Ti implant featuring anSLActive surface (Straumann BL RC, Straumann Group, Basel, Switzerland) (Figure 1). Both implant types share an identical intra‐osseous design (Figure 2). A healing abutment was placed after insertion of a Ti implant. Schematic representation of a section of the maxillary alveolar bone with the investigated implants in situ. Left: Pure ceramic implant; right: Titanium Bone Level RC implant fitted with a titanium RC VarioBase abutment (all components from Straumann Group, Basel, Switzerland). Image: Straumann (2023, https://www.straumann.com/content/dam/media‐center/straumann/de/documents/brochure/technical‐information/702824‐de_low.pdf, accessed October 20, 2023) (Guidance for Implant Removal Basic Information). Periapical radiographs of a ZrO2 implant and a Ti implant placed for replacement of a maxillary premolar, obtained 1 month (T1) and 3 years (T36) after crown placement, with a reference line used for the assessment of marginal bone level. A mesial or distal bone level apically of this reference line results in a negative value, the most severe value was used for analysis. Three months after implant placement, restorative treatment was initiated. Polyether impressions (Impregum, 3 M ESPE, Seefeld, Germany) were taken using semi‐custom trays (Border‐Lock, ClanDental, Maarheeze, The Netherlands). Closed‐tray copings were used for ZrO2 implants and open‐tray copings for Ti implants. All participants received a lithium disilicate crown (IPS e.max CAD, Ivoclar Vivadent, Schaan, Liechtenstein) 2 weeks after impression taking. The crowns for the ZrO2 implants were intraorally luted using a glass ionomer cement (Fuji One, GC Europe, Leuven, Belgium). All residual cement was carefully removed under microscopic inspection. The Ti‐implants received a screw‐retained crown. The crown was cemented extra‐orally onto a titanium‐base abutment (RC VarioBase, Ø 4.5 mm, gingival height 1.0 mm, Straumann Group), using an adhesive cement (Scotchbond Universal Adhesive and RelyX Ulimate Adhesive Resin Cement, 3 M ESPE, St. Paul, MN, USA), and fixed to the implant with a torque of 35 Ncm and subsequently sealed with glass‐ionomer material (Fuji II LC, GC Europe, Leuven, Belgium). The fixation screw was covered with polytetrafluoroethylene tape. All prostheses were fabricated by the same technician (Elysee Dental Labs, Groningen, The Netherlands). Static and dynamic occlusion were verified, and each patient received individualised oral hygiene instructions. All restorative procedures were performed by a single experienced clinician (United States). Blinding of investigators and participants was not feasible due to the visible differences between materials. In brief, patients allocated to the ZrO2 group were treated with a one‐piece, tissue‐level zirconia implant with a cemented, platform‐matched crown, whereas patients in the Ti group received a two‐piece, bone‐level titanium implant with a screw‐retained, platform‐switched crown. Data Collection and Follow‐Up All data were collected by two calibrated examiners (J.B. and L.M.) at three time points: prior to implant placement (T0), 1 month after crown placement (T1; baseline), and 3 years after crown placement (T36). The primary outcome was the change in MBL after 3 years of function. Secondary outcomes included implant survival, peri‐implant tissue health, technical complications, aesthetic outcomes, and PROMs. Marginal Bone Level Change MBL change was assessed using standardised intraoral radiographs taken at T1 and T36, employing individualised radiographic holders to ensure a consistent projection. Digital measurements were performed (DicomWorks software, University Medical Center Groningen) from a reference line from the implant shoulder to the first bone‐to‐implant contact on both mesial and distal sides (Figure 2). Changes were determined by subtracting the values of T1 from T36, the most severe value of either the mesial or distal side was taken for analysis (MBL = T36 − T1). Positive values indicated bone gain; negative values indicated bone loss. Measurements were performed by a single calibrated investigator (U. S.) (Schepke, Meijer, Kerdijk, et al. 2017). Clinical Parameters Clinical indices included implant survival, which was defined as being in situ and functional at the time of follow up. Others were plaque accumulation, presence of calculus, bleeding on probing, gingival condition, probing depth, and visual inflammation score. Plaque and bleeding were scored according to Mombelli (Mombelli et al. 1987) and gingival inflammation according to Loe and Silness (Loe and Silness 1963). Probing depth was measured at three buccal sites (mesio‐, mid‐, and disto‐buccal) using a standardised probe (Hawe Click‐Probe, Hawe Neos Dental, Switzerland) with a probing force of 0.25 N. The most severe probing depth was used for analysis. Plaque and bleeding indices were recorded as a single score per implant site, based on the overall clinical assessment of the buccal aspect. The restauration complex was evaluated using the FDI World Dental Federation functional criteria (Hickel et al. 2010) to monitor possible technical complications. Therefore, a modified version of the FDI criteria was used, in which only criteria 5, 7, and 8 were omitted. Technical complications were recorded when clinical intervention was required. Biological complications were assessed according to the 2017 World Workshop on the Classification of Periodontal and Peri‐implant Diseases and Conditions (Berglundh et al. 2018). Peri‐implant mucositis was defined as the presence of bleeding and/or suppuration on gentle probing, with or without increased probing depth compared with previous examinations, and in the absence of bone loss beyond crestal bone level changes associated with initial bone remodelling. Peri‐implantitis was defined as the presence of bleeding and/or suppuration on gentle probing, with an increased probing depth compared with previous examinations, and radiographic bone loss exceeding crestal bone level changes associated with initial bone remodelling (beyond the measurement error of 0.5 mm). Clinical measurements used for the diagnosis of peri‐implant diseases were obtained at the buccal aspect of the implant site only. Aesthetic Parameters Aesthetic evaluation at T36 was performed by a single, blinded, calibrated researcher (H.M.) using standardised digital photographs and assessed via the Pink and White Esthetic Scores (PES/WES) (Belser et al. 2009). Patient‐Reported Outcome Measures PROMs were obtained through a questionnaire consisting of 12 items addressing emotional, functional, and aesthetic satisfaction. The questionnaire, adapted from Guljé et al. (2014), used visual analogue scales (VAS) ranging from 0% to 100%, where higher scores indicated greater satisfaction. Differences in PROMs between the groups were assessed at T36, and additionally, it was examined whether there was a difference between the groups in the extent to which expectations were met. To evaluate this, expectation percentages at T0 were subtracted from those at T36 to determine any potential disparity in fulfilling expectations after 3 years of function. Sample Size Calculation Sample size estimation was based on published MBL change data for titanium and zirconia implants (Altuna et al. 2016; Gahlert et al. 2016). A mean MBL change of 0.36 mm (SD 0.06 mm) for titanium implants and 1.02 mm (SD 0.90 mm) for zirconia implants was used for the calculation. Based on these data, an effect size of 1.03 was determined, and with α = 0.05 and a power of 0.90, a required sample size of 21 participants per group was calculated using G*Power (v3.1.9.6) (Faul et al. 2007). To account for potential dropouts, 25 patients per group (n = 50) were enrolled. Statistical Analysis Data were managed using the Research Electronic Data Capture system (REDCap; https://projectredcap.org). Statistical analyses were performed with SPSS (version 28.0 for Windows, SPSS inc., Chicago, IL, USA). Descriptive statistics were used to summarise data distributions. Normality was examined graphically; statistical significance was set at p < 0.05. Possible differences in MBL change between the groups after 3 years were analysed using the Mann–Whitney U test, the difference between the groups was expressed as a mean difference with 95% confidence interval. Differences in clinical parameters between implant types at each time point were evaluated using the Mann–Whitney U test and within‐group changes over time using Wilcoxon signed‐rank tests. Differences in the prevalence of peri‐implant mucositis and peri‐implantitis between the implant groups were analysed using the χ 2 test. When expected cell counts were < 5, Fisher's exact test was applied. The PROMs were also analysed with the Mann–Whitney U test to compare differences. Secondary outcomes were grouped into predefined domains (clinical parameters, aesthetic outcomes, and PROMs). Within each domain, Bonferroni correction was applied to account for multiple testing. Results Research Population A total of 50 patients were initially enrolled in the study. One participant withdrew before implant placement (n = 1) and one implant failed to osseointegrate (n = 1) before crown placement. The 3‐year evaluation included all patients whose implants remained in situ (n = 48). All participants were assessed at 1 month (T1‐baseline) and again 3 years after crown placement (T36). Consequently, 48 patients were available for per protocol analysis: 24 received a ZrO2 one‐piece implant with a cemented, platform‐matched crown and 24 received a Ti bone‐level implant with a screw‐retained, platform switched crown (Table 2). Figure 3 illustrates representative clinical cases at T36 of two patients—one treated with a ZrO2 implant and the other with a Ti implant. The study was conducted in accordance with the CONSORT 2010 guidelines (Schulz et al. 2011), a CONSORT flow diagram is shown in Figure 4. TABLE 2Case mix variables.ZrO2 (n = 25)Ti (n = 24)Mean age a 52.1 years (range: 23.1–72.5)52.4 years (range: 28.1–76.3)Gender (male/female)11/1413/11Tooth type (N)1473154624872568 a At the time of implant placement. Case mix variables. At the time of implant placement. Radiographic and clinical photographs of a patient from the ZrO2 group (a, b) and the Ti group (c, d), 1 month after crown placement, at baseline (a, c), and 3 years after crown placement (b, d). Consort flow diagram (Schulz et al. 2011). Marginal Bone Level Change MBL changes from T1 to T36 were evaluated on standardised intraoral radiographs (Figure 2). A Mann–Whitney U test was used to compare the MBL changes over time among the two implant types. No statistically significant difference was found (Z = −0.474, p = 0.635). The estimated mean difference in the outcomes of the MBL changes between the groups was 0.02 mm (95% CI −0.34 to 0.38). The outcomes of the MBL changes are presented in Table 3. TABLE 3Marginal bone level changes between 1 month after restoration placement (T1) and 3 years (T36) in function.T36–T1ZrO2 (n = 24)Ti (n = 24) p Most severe MBL change in mm (SD)−0.02 ± 0.78−0.04 ± 0.360.635 (Z = −0.474)Median (min–max)−0.04 (−1.73 to 2.30)−0.02 (−1.03 to 0.65) Note: A negative value indicates bone loss, a positive value indicates bone deposition. Differences between the groups were analysed using the Mann–Whitney U test. The difference was not statistically significant. Marginal bone level changes between 1 month after restoration placement (T1) and 3 years (T36) in function. Note: A negative value indicates bone loss, a positive value indicates bone deposition. Differences between the groups were analysed using the Mann–Whitney U test. The difference was not statistically significant. Clinical Parameters As mentioned before, at the 3‐year follow‐up (T36) the remaining 48 patients were evaluated. Implant survival rates after 3 years were 96.0% (95% CI 79.6%–99.9%) for the ZrO2‐group and 100.0% (95% CI 85.8%–100%) for the Ti‐group. Clinical outcome measures, summarised in Table 4, were generally favourable at both T1 and T36. At T36, median values for plaque, dental calculus, bleeding, and visual assessment of peri‐implant inflammation were consistently 0 in both groups, indicating healthy peri‐implant tissues regardless of implant material. For the clinical parameters, Bonferroni correction was applied (α = 0.05/10 adjusted significance threshold p < 0.005). After correction, none of the observed differences between the groups at T1 and T36 remained statistically significant. Nevertheless, at T36, the bleeding on probing score was slightly higher for the ZrO2 group (Z = −2.00, uncorrected p = 0.045). TABLE 4Clinical outcome measures around the implant site at baseline (T1) and 3 years (T36).T1T36ZrO2 (n = 24)Ti (n = 24)ZrO2 (n = 24)Ti (n = 24)Plaque‐indexScore 075.0%79.2%91.7%83.3%Score 120.8%37.5%8.3%12.5%Score 24.2%0.0%0.0%4.2% Calculus‐index Score 0100%100%100%91.7%Score 10.0%0.0%0.0%8.3%Gingiva‐indexScore 062.5%75.0%62.5%83.3%Score 133.3%20.8%33.3%16.7%Score 24.2%4.2%4.2%0.0%Bleeding‐indexScore 066.7%95.8%62.5%87.5%Score 133.3%4.2%20.8%8.3%Score 20.0%0.0%16.7%4.2%Score 30.0%0.0%0.0%0.0% Mean probing depth in mm (SD) 3.4 (1.1)2.8 (0.6)3.0 (1.1)2.8 (0.7) Note: Differences between the groups were analysed using the Mann–Whitney U test. Bonferroni correction was applied (α = 0.05/10; adjusted threshold p < 0.005), and no statistically significant differences between groups were found. Clinical outcome measures around the implant site at baseline (T1) and 3 years (T36). Note: Differences between the groups were analysed using the Mann–Whitney U test. Bonferroni correction was applied (α = 0.05/10; adjusted threshold p < 0.005), and no statistically significant differences between groups were found. The technical performance of the restauration complexes was evaluated using the modified FDI World Dental Federation clinical criteria (Table 5). No technical complications were observed in either group during the 3‐year follow‐up. For the criteria related to approximal contour and wear, some restorations showed moderate deviations (scores 3–4). However, none of these findings required clinical intervention. TABLE 5Scores in the categories of the modified FDI criteria (Hickel et al. 2010)—functional criteria.B. Functional properties1. Clinically excellent/very good2. Clinically good3. Clinically sufficient/satisfactory4. Clinically unsatisfactory5. Clinically poor5. Fracture of material and retention ZrO2 Ti 100.0% (24) 95.8% (23) — 4.2% (1) — — — — — — ZrO2, Ti n = 24 7. Occlusal contour and wear ZrO2 Ti 58.3% (14) 91.7% (22) 25.0% (6) 8.3% (2) 16.7% (4) — — — — — ZrO2, Ti n = 24 8. Approximal contour and wear Mesial ZrO2 Ti 30.4% (7) 54.2% (13) 30.4% (7) 8.3% (2) 17.4% (4) 29.2% (7) 21.7% (5) 8.3% (2) — ZrO2 n = 23 Ti n = 24 Distal ZrO2 Ti 43.5% (10) 29.2% (7) 30.4% (7) 25.0% (6) 17.4% (4) 41.7% (10) 8.7% (2) 4.2% (1) — — Note: Five scoring categories can be distinguished in the modified FDI clinical criteria. Presented as percentages and absolute numbers per implant group after 3 years of function. No interventions were required. Scores in the categories of the modified FDI criteria (Hickel et al. 2010)—functional criteria. ZrO2 Ti 100.0% (24) 95.8% (23) — 4.2% (1) — — — — — — ZrO2, Ti n = 24 ZrO2 Ti 58.3% (14) 91.7% (22) 25.0% (6) 8.3% (2) 16.7% (4) — — — — — ZrO2, Ti n = 24 8. Approximal contour and wear Mesial ZrO2 Ti 30.4% (7) 54.2% (13) 30.4% (7) 8.3% (2) 17.4% (4) 29.2% (7) 21.7% (5) 8.3% (2) ZrO2 n = 23 Ti n = 24 ZrO2 Ti 43.5% (10) 29.2% (7) 30.4% (7) 25.0% (6) 17.4% (4) 41.7% (10) 8.7% (2) 4.2% (1) — — Note: Five scoring categories can be distinguished in the modified FDI clinical criteria. Presented as percentages and absolute numbers per implant group after 3 years of function. No interventions were required. At the 3‐year follow‐up, prevalence of peri‐implant mucositis was 33.3% (95% CI 15.6%– 55.3%) in the ZrO2 group (n = 8) and 12.5% (95% CI 2.7%– 32.4%) in the Ti group (n = 3). The difference between groups was not statistically significant (χ 2 = 2.95, p = 0.086). Prevalence of peri‐implantitis was 4.2% (95% CI 0.1%– 22.1%) in the ZrO2 group (n = 1), whereas 0% (95% CI 0.0%– 14.2%) of peri‐implantitis was observed in the Ti group (n = 0). Due to the very low number of events, no statistical comparison between groups was performed. Biological complications are presented in Table 6. TABLE 6Biological complications, prevalence of peri‐implant mucositis and implantitis, in the implant groups after 3 years of function.ZrO2 (n = 24)Ti (n = 24) p Peri‐implant mucositis33.3% (8)12.5% (3)0.086Peri‐implantitis4.2% (1)0% (0)N/A Note: Differences between groups were analysed using the Pearson chi‐square test. Biological complications, prevalence of peri‐implant mucositis and implantitis, in the implant groups after 3 years of function. Note: Differences between groups were analysed using the Pearson chi‐square test. Aesthetic Parameters In both groups, the PES/WES scores were generally acceptable (Table 7). Total aesthetic outcome for the ZrO2‐group was 16.42 ± 1.72 and 15.63 ± 2.45 for the Ti‐group. No statistically significant difference was observed between the groups (Z = −1.062, p = 0.288). After Bonferroni correction, none of the individual PES/WES parameters showed statistically significant differences between the groups. Nevertheless, slightly higher uncorrected scores were observed for the ZrO2‐group regarding the curvature of facial mucosa (Z = −2.110, uncorrected p = 0.035) and the root convexity and soft tissue colour and texture (Z = −2.113, uncorrected p = 0.035). TABLE 7Mean pink aesthetic scores and white aesthetic score (PES/WES) for the ZrO2 and Ti after 3 years (T36) in function (SD).T36ZrO2 (n = 24)Ti (n = 24) p Papilla volumeMesial1.62 (0.58)1.71 (0.55)0.550Distal1.50 (0.59)1.42 (0.65)0.693Curvature facial mucosa1.92 (0.28)1.67 (0.48)0.035Level of facial mucosa1.42 (0.58)1.67 (0.57)0.100Root convexity/soft tissue colour and texture1.29 (0.62)0.88 (0.68)0.035 Total PES 7.75 (1.29)7.33 (1.58)0.480Tooth form1.92 (0.28)1.75 (0.44)0.125Tooth volume/outline1.71 (0.46)1.62 (0.50)0.545Colour (hue/value)1.62 (0.50)1.54 (0.59)0.690Surface texture1.87 (0.34)1.87 (0.34)1.000Translucency1.54 (0.51)1.50 (0.51)0.775 Total WES 8.67 (1.20)8.29 (1.71)0.638Total PES/WES16.42 (1.72)15.63 (2.45)0.288 Note: Differences between the groups were analysed using the Mann–Whitney U test. Bonferroni correction was applied (α = 0.05/13; adjusted threshold p < 0.003). Mean pink aesthetic scores and white aesthetic score (PES/WES) for the ZrO2 and Ti after 3 years (T36) in function (SD). Note: Differences between the groups were analysed using the Mann–Whitney U test. Bonferroni correction was applied (α = 0.05/13; adjusted threshold p < 0.003). Patient‐Reported Outcome Measures After 3 years of implant‐service, PROMs were generally high for both groups (Figure 5). Overall, the patients' expectations were exceeded. The total mean expected overall satisfaction before implant placement (T0) was 84.6 ± 11.6 and 82.2 ± 10.6 for the ZrO2‐ and Ti‐group respectively. Three years after treatment, the total mean measured satisfaction level was 96.2 ± 5.14 and 91.65 ± 7.80 respectively. Bonferroni correction was applied (α = 0.05/12; adjusted threshold p < 0.004). No significant differences in expectation levels compared to the observed satisfaction level after 3 years were observed between the groups. After correction for multiple testing, a statistically significant difference between the groups remained only regarding gum colour (Z = −2.991, p = 0.003). In addition, slight uncorrected differences were found in the statements regarding visibility (Z = −2.497, uncorrected p = 0.013) and gum shape (Z = −2.539, uncorrected p = 0.011), although these did not remain significant after Bonferroni correction. Calculated percentages of the median VAS scores of different statements with regards to patients' satisfaction for the implant supported prosthesis on either a ZrO2 or a Ti implant after 3 years in function (T36). Values ranged from 0 (very discontent or major concerns) to 100 (fully content, no concerns at all) percent. * Statistically significant difference Mann–Whitney U (α = 0.05/12; adjusted threshold p < 0.004) after Bonferroni correction for multiple testing. Discussion This randomised controlled trial aimed to evaluate the MBL changes, clinical parameters, aesthetic outcomes, and PROMs of ZrO2 and Ti implants for single tooth replacement in the maxillary premolar region after 3 years of clinical functioning. The aim was to provide additional insights into the clinical applications of ZrO2 as a ground material for dental implants with a follow‐up of 3 years. When interpreting these results, it is important to keep in mind that the ZrO2 and Ti implants used in this study differed not only in material but also in their transmucosal and restorative configurations (Figure 1). The ZrO2 implants are tissue‐level implants with a cemented crown (platform matched), whereas the Ti implants are bone‐level implants with a screw‐retained crown (platform switched). These design‐related differences should be considered when interpreting the results, as they may have influenced the observed outcomes beyond the effect of the implant material alone. The bone‐level Ti implant and the Pure ZrO2 implant were compared due to their identical intra‐osseous design. At the time of study initiation, no two‐piece ZrO2 implant was available yet. After 1 year of function the ZrO2 and Ti groups of the same population showed comparable MBL changes (Z = −0.774 p = 0.439), the same implant survival rates and no significant differences in clinical parameters and PROMs (de Beus et al. 2024). Data of the present study indicates that after 3 years of function MBL changes are still minimal, without any significant differences between the ZrO2 and the Ti groups. For the secondary outcomes, no statistically significant differences between the groups were observed after correction for multiple testing. However, the PES scores of the present study revealed differences between the groups, with the Ti group scoring higher for the level of facial mucosa, whereas the ZrO2 group scored higher for the root convexity and soft tissue colour and texture. With regard to peri‐implant tissue health, a difference in bleeding scores was found at 3 years to the disadvantage of the ZrO2 group accompanied by a trend towards slightly less positive gingival scores, while other clinical parameters remained largely comparable. Regarding the aesthetic scores, the previously observed difference in the level of facial mucosa was no longer present, whereas the advantage of the ZrO2 group for root convexity, soft tissue colour and root convexity persisted, and an additional difference in curvature of the facial mucosa became apparent. PROMs did not differ significantly after 1 year, whereas several appearance‐related PROMS differed in favour of ZrO2 after 3 years. The primary outcome measure, MBL change, showed minimal and clinically negligible alterations over the 3‐year observation period, with no statistically significant differences between the ZrO2 and Ti implant groups. These findings indicate that, with respect to the primary outcomes, both implant systems perform in a clinically comparable manner, supporting the null hypothesis that there are no differences in MBL change after 3 years between the groups. The MBL changes in the present study are smaller than those reported in several meta‐analyses (Morena et al. 2024; Padhye et al. 2023), which may be explained by the differences in implant site, implant design, loading protocols and even baseline definitions. Early bone remodelling around the same type of ZrO2 implants has been shown to occur predominantly during the first months after implant placement, followed by stabilisation (Kniha et al. 2018). Baseline measurements in the present study were obtained 1 month after crown placement, which makes the results not comparable. Future studies might consider including radiographic assessment from implant placement and directly after prosthetic rehabilitation, since early bone remodelling has been described during both phases and may influence subsequent measurements (Romanos 2023). In addition, it should be acknowledged that the sample size calculation was based on previously published data (Altuna et al. 2016; Gahlert et al. 2016) reporting relatively large differences in MBL change, especially for the ZrO2 implants, resulting in a large estimated effect size. Consequently, the present study was powered to detect relatively large differences between groups, and smaller differences may not have been identified. Implant survival after 3‐years of function was high in both groups, respectively 96.0% (95% CI 79.6%–99.9%) and 100.0% (95% CI 85.8%–100%) for the ZrO2 and Ti implants. These survival rates are consistent, or even slightly better, compared with previous randomised trials and systematic reviews reporting comparable survival rates for ZrO2 and Ti implants (Mohseni et al. 2024; Padhye et al. 2023; Roehling et al. 2023). Peri‐implant tissue conditions were generally favourable in both groups throughout follow‐up. Although higher bleeding scores were observed in the ZrO2 group at T36, this difference did not remain statistically significant after correction for multiple testing and should therefore be interpreted with caution. According to the current diagnostic criteria for peri‐implant health and disease as defined by the 2017 World Workshop (Berglundh et al. 2018), peri‐implant mucositis was observed in both groups, whereas peri‐implantitis occurred only incidentally and at a low frequency. The observed prevalence of peri‐implant mucositis in the present study (33.3% (95% CI 15.6%–55.3%) for ZrO2 and 12.5% (95% CI 2.7%–32.4%) for Ti) appears lower than the weighted mean implant‐based prevalence of approximately 29.48% (95% CI 22.65%–36.32%) reported in a systematic review (Lee et al. 2017), while the occurrence of peri‐implantitis (4.2% (95% CI 0.1%–22.1%) and 0% (95% CI 0.0%–14.2%), respectively) is also below the reported implant‐level prevalence of approximately 9.25% (95% CI 7.57–10.93). Differences in reported prevalence may be explained by the short follow‐up of this study and relatively small study population as well as by variations in diagnostic criteria as mentioned in the systematic review (Lee et al. 2017). In addition, it should be acknowledged that clinical parameters, including probing depth, were assessed only at the buccal site of the implants, which may have led to a possible underestimation of peri‐implant disease. Nevertheless, a trend was observed in which bleeding and gingival scores were slightly higher around ZrO2 implants, comparable to the results of Koller et al. (2020), whereas plaque accumulation and calculus were more frequently observed around Ti implants (Koller et al. 2020). This observation is in line with previous studies suggesting lower plaque accumulation on ZrO2 surfaces compared to Ti (van Brakel, Cune, et al. 2011). However, Kniha and colleagues concluded that current available data do not allow a clear preference to be stated for either ZrO2 or Ti with regard to surface biofilm formation (Kniha et al. 2021). The present findings do not support a potential advantage of ZrO2 with respect to peri implant soft tissue health. This assumption possibly originated from studies focusing on zirconia abutment materials rather than implant bodies (Kniha et al. 2021; Sanz‐Sánchez et al. 2018). Recent systematic reviews and meta‐analysis report on comparable peri‐implant clinical outcomes for both materials (Duan et al. 2023; Morena et al. 2024; Padhye et al. 2023). The observation in the present study that slightly higher bleeding scores coexist with low plaque scores and stable probing depths suggests that peri‐implant tissue health cannot be explained by plaque accumulation alone. Previous studies indicate that bleeding scores may also be influenced by tissue characteristics, implant design and soft tissue adaptation (Ivanovski et al. 2022; van Brakel, Cune, et al. 2011). Moreover, the observed differences cannot be attributed solely to the implant material, as the study design inherently combined material with prosthetic concepts. Previous evidence suggests that prosthetic design may influence peri‐implant tissue outcomes. Cemented restorations have been associated with a higher incidence of biological complications, potentially related to excess cement (Sailer et al. 2012), while platform switching has been reported to support marginal bone preservation and peri‐implant tissue stability compared to platform matching (Lin et al. 2025). After 3 years, PES/WES scores were favourable in both groups, with only minor differences and a slight overall preference for the ZrO2 group, which is in line with recent literature (Padhye et al. 2023). Notably, the absolute PES/WES scores in the present study were slightly higher than those reported by Padhye et al. (2023). Consistent with the one‐year outcomes of this trial, the criterion ‘root convexity/soft tissue colour and texture’ showed a difference in favour of the ZrO2 group. One possible explanation could be the metal‐grey colour of the implant and the titanium‐base used in the Ti‐group, which may have negatively influenced the clustered score, even though most of the transmucosal part of the implant‐crown complex consists of tooth coloured lithiumdisilicate. The use of the PES/WES index of Belser et al. (2009) allowed direct comparison with the one‐year results, which is the reason why this tool was used in this analysis, but the clustering of the soft tissue colour and texture into a single parameter still limits detailed interpretation. Given the potential relevance of gingival colour differences when comparing ZrO2 and Ti implant supported restorative complexes, future studies may benefit from applying the PES score as described by Fürhauser et al. (2005) or from incorporating objective colour assessment methods, as previously suggested by Thoma et al. (2016). Future studies may also consider comparing clinician and patient based aesthetic outcomes, as described by Zuercher et al. (2024). In addition, a difference was observed for the criterion ‘curvature of facial mucosa’. However, such differences may more likely be attributable to variations of the transmucosal design (i.e., one‐piece implant with cemented crown, platform‐matched vs. bone level implant with screw retained crown, platform‐switched) of the prosthetic complexes, rather than to the implant material. Patient satisfaction remained high in both groups and exceeded patients' expectations prior to implant placement. Increasing attention has been given to PROMs in implant dentistry, with a previous study also demonstrating high PROMs for both ZrO2 and Ti implants, generally with only minor differences between materials (Delgado‐Ruiz and Romanos 2018). Direct comparison of absolute PROMs remains challenging due to methodological heterogeneity. Nevertheless, in the present study, a trend was identified for several appearance‐related PROM items—including implant visibility and perceived crown and gingival aesthetics—consistently in favour of the ZrO2 group. However, it should be noted that patients were not blinded to the type of implant received, which may have influenced subjective outcome reporting. A systematic review and meta‐analysis demonstrated that PROMs and clinician‐reported aesthetic outcomes at implant sites are not associated (Sadilina et al. 2025). This observation appears to be largely consistent with the findings of the present study, in which no evident association was observed between clinician‐based aesthetic scores and the majority of PROMs at 3 years, except for the specific appearance‐related parameters. Although no clear association between PROMs and clinician‐aesthetic outcomes has been established, it is noteworthy that the present 3‐year results showed more favourable scores for particularly implant visibility and perceived gingiva colour and shape. These findings coincided with differences between the groups in the PES criteria for the curvature of the facial mucosa and root convexity/soft tissue colour and texture, parameters that largely assess related aesthetic aspects of the peri‐implant tissues. In this context, the present findings suggest that both patients and objective measurement tools are able to identify differences for these specific aesthetic parameters between the two groups. Nevertheless, when considered alongside the PES/WES findings of this study, these results may provide a recommendation for future research focusing more specifically on the influence of implant material on the aesthetic outcomes and their relationship with PROMs, ideally using more specific assessment methods concerning the colour of the surrounding mucosa. Apart from that, the relevance of the differences found in this study appears to be limited. Overall, it is recommended that long‐term prospective clinical studies with standardised outcome measures are essential to obtain a more comprehensive understanding of the clinical performance of ZrO2 implants and to monitor potential biological and technical complications, ultimately allowing to establish well‐defined clinical indications for their routine use. Furthermore, the single‐centre design may limit the generalisability of the findings. Moreover, the 3‐year follow‐up should be considered relatively short for implant studies, and longer‐term data are needed to confirm the stability of these outcomes. Conclusion The ZrO2 one‐piece implant and the Ti bone level implant type for single tooth replacement in the maxillary premolar region, as applied in this study, perform comparably regarding MBL stability over a 3‐year period of clinical function. Survival, clinical and aesthetic parameters, as well as PROMs, showed favourable results for both groups. These short‐term findings indicate that both treatment modalities represent a viable option for clinical use. Author Contributions J. W. A. Slot: methodology, resources. C. Jensen‐Louwerse: methodology, resources. G. M. Raghoebar: methodology, resources, writing – review and editing, supervision. J. H. W. de Beus: data curation, investigation, formal analysis, validation, visualization, project administration, writing – original draft. U. Schepke: conceptualization, methodology, data curation, investigation, project administration, writing – review and editing, supervision. M. S. Cune: conceptualization, project administration, writing – review and editing, funding acquisition, supervision. L. Muijselaar: data curation, formal analysis, investigation. H. J. A. Meijer: writing – review and editing, supervision. S. la Bastide‐van Gemert: formal analysis, writing – review and editing. Funding This work was supported by ITI Research Grand, No. 1145‐2016. Conflicts of Interest The authors report an ITI Research Grant No. 1145‐2026 and support from the authors' institutions, during the conduct of the study, and declare to have no conflicts of interest related to this study. Supporting information Data S1: CONSORT 2025 checklist of information to include when reporting a randomised trial.