Narrow‐diameter implants for treatment with fixed restorations in the posterior region: A systematic review and meta‐analysis

Abstract

Purpose

To evaluate the survival and complication rates of narrow‐diameter implants (NDIs) for treatment with fixed restorations in premolar and molar sites.

Methods

A systematic review and meta‐analysis were conducted following the preferred reporting items of systematic reviews and meta‐analyses (PRISMA) guidelines. Searches were performed according to a Population‐Intervention‐Outcome (PIO) format in MEDLINE, Scopus, and Cochrane Library, supplemented by manual cross‐referencing. Inclusion criteria encompassed clinical studies on NDIs (<3.75 mm diameter) restored with fixed dental prostheses (FDP) in the posterior region. Random‐effect models were employed to pool survival and complication rates, and heterogeneity was assessed using the Q test and I ² statistic. Subgroup analyses explored variations by jaw location, implant material, and follow‐up. The risk of bias was evaluated using the Cochrane tool for randomized controlled trials (RCTs) and the MINORS tool for non‐randomized studies.

Results

Thirty‐six trials involving 2741 NDIs were included in the systematic review and 29 in the meta‐analysis. Survival rates (SRs) ranged from 92.73% to 100% with follow‐up periods from 10.8 months to 12 years. Pooled SRs for maxillary (97.0%; 95%CI: 96.2%–97.8%) and mandibular (96.5%; 95%CI: 95.7%–97.3%) NDIs were not significantly different (p = 0.688). Comparing NDIs in the premolar and molar sites, jaw location demonstrated similar SRs of 97.5% (95% CI: 95.9%–99.1%) and 98.6% (95% CI: 96.5%–99.6%), respectively. Technical complication rates varied from 0% to 23.5%, with follow‐up periods ranging from 1 to 12 years. These complications primarily included screw loosening, fractures, and detachment of restorations. Data on biological complications were notably limited, and due to the heterogeneity in reporting, comparisons were not feasible.

Conclusion

NDIs demonstrate high survival and low complication rates for therapy of the partially edentulous patient with FDPs in posterior sites. The clinical performance is consistent across jaw location and implant materials.

The use of narrow‐diameter implants (NDIs), defined as implants with diameters less than 3.75 mm, has emerged as a promising alternative in implant dentistry, particularly in anatomical scenarios with limited bone volume or space restrictions. ¹ There is a clear trend in scientific databases that hits about NDIs are consistently increasing every year since 2010, indicating the increasing scientific interest in them.

Historically, NDIs were predominantly indicated for anterior regions, such as replacing incisors in the mandible as well as maxillary laterals with narrow mesiodistal spaces, due to concerns regarding their mechanical stability under high occlusal loads. ² An additional representative indication of NDIs is to improve the retention of mandibular overdentures in edentulous patients or in cases of geriatric or medically compromised patients in order to avoid invasive bone augmentation procedures. ³ , ⁴ However, recent advancements in implant materials and design, along with refined surgical protocols, have extended the use of NDIs to more challenging areas, including the posterior regions. ⁵ , ⁶ This expansion of the indication is driven by the need for minimally invasive procedures, particularly in patients who want to avoid complex augmentation techniques or have insufficient bone volume. ⁷

The posterior region of the jaw presents significant anatomical challenges for implant placement, such as reduced bone density, proximity to the maxillary sinus or inferior alveolar nerve, and a thinner alveolar crest. ⁶ These factors often necessitate alternative approaches to standard‐diameter implants (RDIs), which may not always be feasible without extensive bone grafting. ⁸ NDIs can potentially overcome these limitations and present a minimally invasive treatment plan that can eliminate surgical and economic burdens. ⁹ However, their reduced dimensions have traditionally raised concerns about their ability to withstand the occlusal forces typical in posterior regions. ¹⁰ The mechanical properties of NDIs, including their lower fracture resistance and the increased likelihood of screw loosening or abutment failure, remain key considerations that warrant detailed evaluation. ⁵ , ¹¹

Despite these concerns, emerging evidence indicates that NDIs may offer survival and success rates comparable to RDIs, ¹² even in the load‐bearing posterior regions, provided that careful case selection and precise surgical techniques are employed. ¹³ For instance, studies have reported cumulative survival rates (SRs) for NDIs ranging from 94% to 99% ² in posterior regions, suggesting that the mechanical limitations of NDIs may be less significant when proper implant materials, such as titanium‐zirconium (TiZr) alloys, and rigorous clinical protocols are applied. ⁶ , ⁷

While NDIs have traditionally been used in anterior regions due to concerns over mechanical stability, their application in posterior regions remains controversial due to higher occlusal forces and the anatomical challenges of these areas. However, with the introduction of new material combinations in implant dentistry, such as TiZr alloys, it is promising to evaluate whether NDIs can be utilized for the treatment of fixed dental prostheses (FDPs) in posterior sites for the replacement of premolars and molars. Clinically, some practitioners debate their efficacy in managing load‐bearing zones, where increased forces may potentially lead to complications such as screw loosening, abutment fractures, or prosthetic failure. ⁵ , ¹¹ Mechanically, NDIs are considered to possess lower fracture resistance compared to standard‐diameter implants, especially when subjected to continuous occlusal stresses. ² These issues cast doubt on their long‐term reliability in posterior regions. Furthermore, from an economic perspective, there is variability in cost‐benefit assessments.

Some suggest that NDIs may lessen the need for intricate bone augmentation processes, presenting a more economical option for patients with insufficient bone volume. ⁷ Yet, others raise concerns that the potential requirement for more frequent maintenance or replacements may negate these initial savings. ⁶ Such complex issues demand further exploration to ascertain if advancements in implant materials, like TiZr alloys, alongside improved clinical protocols, have genuinely tackled these challenges. ¹⁴ , ¹⁵

Therefore, this systematic review and meta‐analysis aims to provide a comprehensive and updated analysis on the survival and complication rates of NDIs for treatment with FDPs in premolar and molar sites of the maxilla and mandible.

METHODS

Protocol and registration

This systematic review was conducted in accordance with the guidelines of Preferred Reporting Items of Systematic Reviews and Meta‐Analyses (PRISMA). ¹⁶ The review question was formulated according to the FINER criteria (Feasible, Interesting, Novel, Ethical, and Relevant); the protocol was registered in the International Prospective Register of Systematic Reviews (PROSPERO) (ID: CRD42023428880) (Table 1).

TABLE 1.

Quality assessment of the non‐randomized studies using the MINORS tool.

Study/Year	Study design	Aim	Inclusion of consecutive patients	Prospective data collection	Appropriate endpoints	Unbiased endpoint assessment	Adequate follow‐up	Loss to follow‐up <5%	Sample size calculation	Control group	Baseline equivalence	Adapted statistical analysis	Final score
Alrabiah et al. 26	Retrospective	2	2	1	2	1	2	2	0	1	1	2	18/24
Altinci et al. 27	Retrospective	2	2	1	2	1	2	2	0	1	1	2	18/24
Anitua et al. 28	Prospective	2	2	2	2	1	2	2	0	1	1	2	18/24
Anitua et al. 29	Prospective	2	2	2	2	1	2	2	0	1	1	2	18/24
Antiua et al. 30	Retrospective	2	2	1	2	1	2	2	0	1	1	2	18/24
Arisan et al. 31	Retrospective	2	2	1	2	1	2	2	0	1	1	2	17/24
Assaf et al. 32	Prospective	2	2	2	2	1	2	2	0	1	1	2	18/24
Chiapasco et al. 14	Retrospective	2	2	1	2	1	2	2	0	1	1	2	18/24
Comfort et al. 33	Prospective	2	2	1	2	1	2	2	0	1	1	1	17/24
de Souza Tolentino et al. 34	Retrospective	2	2	1	2	1	2	2	0	1	1	2	18/24
Degidi et al. 5	Retrospective	2	2	1	2	1	2	2	0	1	1	2	16/24
El‐Sheikh et al. 35	Prospective	2	2	1	2	1	2	2	0	1	1	2	18/24
Flanagan et al. 21	Retrospective	2	2	1	2	1	2	2	0	1	1	2	17/24
Geckili et al. 36	Retrospective	2	2	1	2	1	2	2	0	1	1	2	18/24
Grandi et al. 7	Retrospective	2	2	1	2	1	2	2	0	1	1	2	18/24
Kim et al. 22	Retrospective	2	2	1	2	1	2	2	0	1	1	2	18/24
Lambert et al. 37	Prospective	2	2	1	2	1	2	2	0	1	1	1	17/24
Lee et al. 25	Retrospective	2	2	1	2	1	2	2	0	1	1	2	17/24
Maló et al. 23	Retrospective	2	2	1	2	1	2	2	0	1	1	2	18/24
Maló et al. 38	Prospective	2	2	1	2	1	2	2	0	1	1	2	18/24
Marcantonio et al. 39	Retrospective	2	2	1	2	1	2	2	0	1	1	2	18/24
Mangano et al. 40	Retrospective	2	2	1	2	1	2	2	0	1	1	2	18/24
Mangano et al. 41	Retrospective	2	2	1	2	1	2	2	0	1	1	2	18/24
Pieri et al. 42	Retrospective	2	2	1	2	1	2	2	0	1	1	2	18/24
Romeo et al. 43	Retrospective	2	2	1	2	1	2	1	0	1	1	2	15/24
Saad et al. 44	Retrospective	2	2	1	2	1	2	2	0	1	1	2	18/24
Shi et al. 45	Prospective	2	2	1	2	1	2	2	0	1	1	2	18/24
Tolentino et al. 46	Retrospective	2	2	1	2	1	2	2	0	1	1	2	18/24
Vigolo et al. 47	Retrospective	2	1	1	2	1	2	2	0	0	1	2	14/24
Woo et al. 24	Retrospective	2	2	1	2	1	2	2	0	1	1	2	17/24
Xiao et al. 48	Prospective	2	2	1	2	1	2	2	0	1	1	2	18/24
Yaltirik et al. 49	Retrospective	2	2	1	2	1	2	2	0	1	1	2	17/24
Zinsli et al. 50	Retrospective	2	2	1	2	1	2	1	0	1	1	2	15/24

Study selection

The study selection process aimed to minimize bias and ensure relevance. Studies were included if they examined the use of NDIs with diameters less than 3.75 mm for fixed restorations in edentulous posterior regions. The inclusion criteria were limited to randomized controlled trials (RCTs), as well as prospective and retrospective clinical studies focusing on NDIs used for FDPs in the posterior regions of the maxilla or mandible. Only studies published in English were considered. The exclusion criteria included systematic reviews, commentaries, letters to the editor, in vitro studies, animal model studies, case reports, and clinical studies involving patients with severe medical conditions or those that did not adhere to the defined PIO (Population, Intervention, Outcome) format (Table 2).

TABLE 2.

Quality assessment of randomized controlled trials (RCTs).

Study	Random sequence generation	Allocation concealment	Blinding of participants and personnel	Blinding of outcome assessment	Incomplete outcome data	Selective reporting	Other sources of bias	Overall risk of bias
De Souza et al. 11	Low risk	Low risk	High risk	Low risk	Low risk	Low risk	Low risk	Moderate
Ioannidis et al. 51	Low risk	Unclear risk	High risk	Low risk	Low risk	Low risk	Low risk	Moderate
Tolentino et al. 52	Low risk	Low risk	High risk	Low risk	Low risk	Low risk	Low risk	Moderate

Abbreviation: RCTs, randomized controlled trials.

Information sources and search strategy

Based on the PIO criteria, a search strategy was developed and executed using an electronic search. The PIO question was formulated as follows: “What are the survival and complication rates of NDIs used for fixed restorations in the posterior region?”. The search strategy for this systematic review was designed to identify relevant literature on the use of NDIs for the restoration of edentulous posterior regions. A comprehensive search of electronic databases—MEDLINE (via PubMed), Scopus, and Cochrane Library, OVID—was performed, without any time restrictions, to capture all available studies until July 2024. The search terms included a combination of Medical Subject Headings (MeSH) and free‐text keywords to ensure thorough retrieval. The primary search terms used were “dental implants,” “NDIs,” “regular‐diameter implants (RDI),” and “posterior regions.” Boolean operators were employed to refine the search, combining terms such as “(‘dental implants’[MeSH Terms] OR ‘dental implant’[All Fields]) AND (regular[All Fields] AND diameter[All Fields])” to ensure specificity (Table S2).

In addition to the electronic searches, manual cross‐referencing of the bibliographies of relevant systematic reviews and included studies were conducted to identify any further potentially relevant publications that might have been missed.

Data extraction

The search was systematically documented, and all studies were screened for eligibility through a three‐stage process: title screening, abstract assessment, and full‐text review. Two reviewers independently performed all screening and data extraction processes, and any disagreements were resolved through discussion or by involving a third reviewer. The comprehensive approach ensured that the final pool of included studies was of high relevance and quality, forming the basis for the subsequent systematic review and meta‐analysis.

Risk of bias assessment

The quality of the included studies was evaluated using two different assessment tools. For randomized trials, the Cochrane Collaboration tool was employed to determine the risk of bias. For non‐randomized studies, a modified Methodological Index for Non‐Randomized Studies (MINORS) tool was used, consisting of 12 questions tailored explicitly for in vivo studies. Each item was rated from 0 to 2: a score of 0 was assigned if it was not reported, 1 if it was inadequately described, and 2 if it was reported sufficiently. The methodological quality assessment was conducted independently by two reviewers (A.P. and P.T.). In cases of disagreement, these were resolved by involving a third independent reviewer (N.S.). The potential risk of bias for each study was then classified as low, high, or unclear. Discrepancies were discussed until consensus was achieved among the reviewers. ¹⁷ , ¹⁸

Statistical analysis

Random‐effects models were employed to pool SRs across studies. ¹⁹ The Restricted Maximum Likelihood (REML) method was used for model estimation. The overall SR and its corresponding 95% Confidence Interval (CI) were calculated across all included studies. Heterogeneity among studies was assessed using the Q test and the I ² statistic, ²⁰ with I ² values representing the proportion of variability due to heterogeneity rather than chance.

Publication bias was evaluated using a funnel plot, and tests such as Kendall's rank correlation and Egger's regression were performed to measure asymmetry. Sensitivity analysis identified and excluded studies that disproportionately influenced the results. This improved the stability of the findings, evidenced by reductions in I ² and τ ² (between‐study variance).

Subgroup analyses were performed to explore heterogeneity and examined SRs by implant material, jaw location, and follow‐up duration. For robustness, influence diagnostics like Cook's Distance and DFFITS were applied to confirm the validity of the pooled estimates.

The meta‐analysis and related statistical analyses were conducted using STATA version 13.1 (Stata Corp., College Station, TX, USA) and RevMan software. RevMan was specifically used to calculate SRs as pooled risk differences (RD) to account for study‐level variability by finding the difference between the proportions of survival and non‐survival cases in each study. These RDs were then pooled using a random‐effects model to account for between‐study variability. The final pooled SR and its 95% CI comprehensively measure survival outcomes across the included studies.

RESULTS

Study selection

A comprehensive search was conducted, initially identifying 342 records through database searches. An additional 15 records were sourced through other methods, resulting in a total of 357 records after removing duplicates. During the screening phase, the titles and abstracts of these 357 records were reviewed, leading to the exclusion of 295 records that did not meet the inclusion criteria. Subsequently, 62 full‐text articles were assessed for eligibility, of which 26 were excluded due to ineligible population or intervention (Figure 1). Finally, 36 trials were included for data extraction, while 29 trials were used to perform a meta‐analysis.

FIGURE 1.

Flow chart outlining the systematic process of study selection, from identification and screening to eligibility assessment and inclusion, following PRISMA guidelines. PRISMA, preferred reporting items of systematic reviews and meta‐analyses.

Study characteristics

The systematic review encompassed 36 studies, with study designs predominantly retrospective (50%), followed by non‐randomized clinical trials (22.22%), case‐control studies (8.33%) and randomized controlled clinical trials (8.33%). The sample size of implants per study ranged from 10 up to 300, providing a robust dataset for analysis. The follow‐up durations varied considerably, from 10.8 months to 12 years, with a median follow‐up period of 4 years. This variability in follow‐up duration allowed for a comprehensive assessment of both short‐term and long‐term outcomes. Regarding the type of restorations, screw‐retained prostheses were the most frequently reported (50% of the trials included), while cemented restorations accounted for 30%. The remaining trials used a combination of both restoration types.

SRs

NDI SRs were consistently high across the trials analyzed, ranging from 92.73% to 100%. Notably, trials with follow‐up periods of less than 2 years frequently reported 100% SRs. However, a slight decline in SRs was observed in trials with follow‐up periods exceeding 4 years (92.73% to 97.73%). The highest SRs were associated with TiZr (98.4%; 95% CI: 97.6%–99.2%), and titanium (Ti) (98.4%; 95% CI: 97.6%–99.2%) implants.

Technical complications

Technical complications were reported in 52.78% of the studies, with incidence rates ranging from 0% to 23.5% with corresponding follow‐up periods ranging from 1–12 years. These complications primarily involved screw loosening, fractures, and detachment of restorations. The role of connection type was significant, with platform‐switching and CrossFit systems emerging as pivotal in reducing technical failures. Studies utilizing these advanced connection designs reported fewer mechanical issues, highlighting their importance in improving implant stability and longevity.

Biological complications

Data on biological complications were notably limited, with only 25.6% of the studies addressing this aspect. Among the studies that reported biological outcomes, bone loss, and peri‐implantitis were the most common concerns. However, considerable heterogeneity in the criteria used to define biological complications made comparisons across studies challenging or not possible.

Factors influencing outcomes

Several factors emerged as significant determinants of implant success and complication rates. Implant material played a crucial role, with Ti and TiZr implants achieving SRs exceeding 97%. There were no statistically significant differences in outcomes between these two materials, suggesting that both are equally suitable for clinical applications. Jaw location also influenced outcomes, with mandibular implants demonstrating slightly higher SRs and fewer complications than maxillary implants. This finding may be attributed to differences in bone density and load distribution between the two jaw regions.

Connection type further influenced implant success. Platform‐switching and CrossFit systems consistently showed lower rates of technical failures, underscoring their importance in mitigating mechanical complications. These connection designs are believed to enhance the distribution of occlusal forces, thereby reducing stress on the implant components. Lastly, the type of tooth restored did not significantly impact SRs, with both premolars and molars achieving SRs above 95%.

Longitudinal findings

The duration of follow‐up emerged as a critical factor influencing reported outcomes. Short‐term studies (<24 months) often reported SRs of 100%, reflecting the immediate success of NDIs. In contrast, studies with follow‐up durations exceeding 4 years showed a modest decline in SRs, ranging from 92.73% to 97.73%.

Manufacturers such as Straumann and BTI demonstrated consistently high SRs, ranging from 97.3% to 100%, regardless of follow‐up duration or implant diameter.

Meta‐analysis—Study inclusion and data extraction

A total of 29 studies were included in the final meta‐analysis, selected from an initial pool of 36 studies. These included three RCTs, 22 retrospective studies, and four non‐randomized clinical trials. Five studies were excluded during sensitivity analysis due to their disproportionate influence on heterogeneity and overall model robustness. ²¹ , ²² , ²³ , ²⁴ , ²⁵ Data extraction followed a predefined protocol, capturing study characteristics (authors, year, study design, sample size) and primary outcomes, including SRs and other key metrics (Table S1). Studies where SRs were not explicitly reported were calculated based on provided data, such as total cases versus events or effect sizes. Two independent reviewers conducted the data extraction, resolving discrepancies through discussion or consultation with a third reviewer.

Pooled effect size

The pooled effect size across the 29 included studies was statistically significant: Effect Size (Intercept): 0.977 (SE = 0.004, z = 243.965, p < 0.001), indicating a high SR across studies.

Heterogeneity

Heterogeneity metrics showed negligible variability between studies after excluding the influential ones: Residual Variance (τ ²): 2.150 × 10⁻⁶, reflecting minimal between‐study variance. I ²: 0.403%, indicating nearly no residual heterogeneity. H ²: 1.004, suggesting that observed variability aligns with what would be expected by chance.

The test of residual heterogeneity was not significant: Q‐Test: Q = 31.795, df = 30, p = 0.377, confirming the absence of significant unexplained variability (Figure 2).

FIGURE 2.

Forest plot illustrating the pooled effect size analysis for the meta‐analysis. Each study's effect size is represented by a square (proportional to its weight in the analysis).

Sensitivity analysis

Sensitivity analyses identified five studies as disproportionately influencing the overall model: Kim et al. ²² , Flanagan et al. ²¹ , Maló et al. ²³ , Woo et al. ²⁴ , and Lee et al. ²⁵ These studies were flagged based on diagnostic metrics such as Cook's Distance, DFFITS, and leverage. Their exclusion significantly improved the model's heterogeneity metrics and reduced asymmetry in the funnel plot.

The sequential exclusion of these influential studies resulted in a reduction in I ² from 73.002% to 0.403%, indicating that nearly all variability was accounted for by chance. Additionally, it resulted in a reduction in τ ² to 2.150 × 10⁻⁶, reflecting minimal between‐study variance and a consistent pooled effect size (0.977), affirming the robustness of the findings.

Diagnostic plots post‐exclusion confirmed that individual studies no longer disproportionately influenced the model, and heterogeneity measures were within acceptable ranges.

Publication bias

The initial funnel plot exhibited moderate asymmetry, with several outliers contributing to the skew. After excluding the five influential studies, the funnel plot became more symmetrical, suggesting a reduced risk of publication bias. The improved distribution aligns with expectations for well‐conducted meta‐analyses.

Egger's Regression Test results were found to be before Exclusion: z = −2.549, p = 0.011, indicating significant asymmetry and potential publication bias, and after exclusion: z = −1.812, p = 0.070, indicating no statistically significant asymmetry.

Kendall's Rank Correlation Test results were Kendall's τ: −0.285, p = 0.027. While this suggests slight residual asymmetry, it is unlikely to meaningfully affect the robustness of the pooled results (Figure 3).

FIGURE 3.

Publication bias—Funnel plot.

Subgroup analysis

Jaw

Eleven studies reported implant SRs for implants placed in the maxilla. The pooled SR was 97.0% (95% CI: 96.2%–97.8%), with no observed heterogeneity (I ² = 0.0%, p = 0.909). This consistent SR highlights the predictability of implant outcomes in the maxilla, with minimal variability across studies.

Similarly, eleven studies provided data on implant SRs in the mandible. The pooled SR was slightly lower at 96.5% (95% CI: 95.7%–97.3%), with no heterogeneity (I ² = 0.0%, p = 0.909). These results confirm the reliability of implants in the mandible.

The test for subgroup differences (maxilla vs. mandible) indicates that the minor decrease in SR from 97.0% (maxilla) to 96.5% (mandible) is not statistically significant (Q = 0.161, df = 1, p = 0.688). This suggests no meaningful difference in implant SRs between the two jaw locations (Figures 4a,b).

FIGURE 4.

(a) Forest plot of SR in maxilla; (b) Forest plot of SR in mandible. SR, survival rate.

Tooth location

Twelve studies provided data on the SRs of implants in premolar sites. The pooled SR was 97.5% (95% CI: 95.9%–99.1%). Moderate heterogeneity was detected (I ² = 38.69%, p = 0.114), indicating some variability among the studies, potentially due to differences in implant design, placement techniques, or patient characteristics. This heterogeneity suggests a need for further subgroup analyses to explore contributing factors ​(premolars).

Twelve studies also examined implant SRs in molar sites. The pooled SR was 98.6% (95% CI: 96.5%–99.6%). Notably, no heterogeneity was detected (I ² = 0%, p = 0.789), indicating consistent findings across studies.

The test for subgroup differences between NDIs in premolar and molar sites confirmed that the slight difference in effect sizes is not statistically significant (Q = 1.071, p = 0.301). This indicates that both groups demonstrate similar effect sizes despite the observed variability for NDIs placed in premolar sites (Figures 5a, b).

FIGURE 5.

(a) Forest plot depicting the pooled SRs of implants in premolars; (b) Forest plot depicting the pooled SRs of implants in molars. SRs, survival rates.

Implant material

Twenty‐three studies with a total of 1740 NDIs reported on implants made of Ti. The pooled SR was 98.4% (95% CI: 97.6%–99.2%), with low heterogeneity (I ² = 9.02%, p = 0.323). These results underscore the well‐established biocompatibility and mechanical stability of Ti implants, making them a reliable choice for implant therapy across diverse patient populations and clinical settings.

Nine studies examined NDIs with a total of 308 implants made from TiZr alloys. The pooled SR was 98.4% (95% CI: 97.6%–99.2%). However, moderate heterogeneity was observed (I ² = 40.31%, p = 0.113), which may reflect differences in implant design, manufacturing processes, or clinical protocols. The test for subgroup differences between Ti and TiZr implants confirmed that the observed difference in heterogeneity and effect sizes is not statistically significant (Q = 1.497, df = 1, p = 0.221). This indicates that both groups demonstrate similar effect sizes despite the higher variability observed in the TiZr subgroup (Figures 6a,b).

FIGURE 6.

(a) Forest plot showing the pooled SRs for Ti implants; (b) Forest plot depicting pooled survival rates for TiZr implants across nine studies. SRs, survival rates. Ti, titanium; TiZr, titanium‐zirconium.

Follow‐up duration

Thirty‐one studies reported implant SRs for a follow‐up period of up to 5 years. The pooled SR was 98.2% (95% CI: 97.4%–99.0%), with low heterogeneity (I ² = 11.7%, p = 0.446). This finding highlights the implants’ short‐term reliability and predictability, with minimal variability across studies ​(< 5 years).

Eleven studies provided data on implant SRs for follow‐up periods exceeding 5 years. The pooled SR decreased slightly to 95.1% (95% CI: 91.7%–98.5%), with substantial heterogeneity (I ² = 72.3%, p < 0.001). This increased variability suggests that long‐term implant survival may be influenced by patient‐specific conditions (e.g., bone density, oral hygiene), clinical protocols, or complications that arise over time. The observed heterogeneity underscores the importance of standardized long‐term follow‐up studies to better assess implant therapy's durability ​(> 5 years).

The test for subgroup differences suggests that the slight decrease in SR from 98.2% (up to 5 years) to 95.1% (more than 5 years) is not statistically significant (Q = 1.147, df = 1, p = 0.284) (Figures 7a,b).

FIGURE 7.

(a) Forest plot showing pooled SRs for implants with follow‐up periods of up to 5 years; (b) Forest plot illustrating pooled survival rates for implants with follow‐up periods exceeding 5 years. SRs, survival rates.

Quality assessment

The quality of the included studies on NDIs was assessed using the MINORS tool. Scores ranged from 14 to 18 out of 24, averaging 17.5, indicating a moderate to high methodological quality across the 33 studies (Table 1). Most studies were retrospective, with only a few employing prospective designs, which inherently limit control over bias. Key strengths of the studies included well‐defined aims, appropriate clinical endpoints (e.g., implant SRs, marginal bone loss), adequate follow‐up periods, and low attrition rates. These aspects enhance the clinical relevance and reliability of the findings. However, significant limitations were identified: none of the studies included control groups, randomization, or sample size calculations, impacting the generalizability and internal validity of the results. Additionally, a few studies implemented blind or independent outcome assessments, posing a risk of assessment bias.

All three RCTs demonstrated a low risk of bias regarding randomization, outcome assessment, data completeness, and reporting transparency (Table 2). However, a high risk of bias was noted for blinding participants and personnel in each study, which may influence subjective outcomes. Allocation concealment was unclear in Ioannidis et al., ⁵¹ while Souza et al. ¹¹ and Tolentino et al. ⁵² maintained low risk in this domain.

DISCUSSION

This meta‐analysis comprehensively evaluated the SRs of NDIs across 29 studies, providing significant insights into their clinical efficacy. The pooled effect size was 0.977 (SE = 0.004, z = 243.965, p < 0.001), indicating a high SR for NDIs across diverse clinical settings. These findings are consistent with previous studies on NDIs, including those by Chiapasco et al. ¹⁴ , Geckili et al. ³⁶ , and Degidi et al. ⁵ , which reported SRs as high as 98.74% in clinical applications. These results underscore the reliability and predictability of NDIs in various implant placement scenarios, reinforcing their growing role in clinical implantology.

The analysis revealed that implant SRs were highly consistent across both the maxilla (97.0%, 95% CI: 96.2%–97.8%) and mandible (96.5%, 95% CI: 95.7%–97.3%), with no heterogeneity (I ² = 0.0%) in both regions. This result is in line with the findings of Geckili et al. ³⁶ , who reported high success rates for NDIs placed in both the maxilla and mandible. The consistent SRs observed across both jaw locations suggest that NDIs are reliable, irrespective of placement in the upper or lower jaw. ³⁶ Notably, the test for subgroup differences (maxilla vs. mandible) confirmed that the slight difference in SRs between the two regions (97.0% vs. 96.5%) was not statistically significant (Q = 0.161, p = 0.688), further supporting the uniformity of NDI performance across different jaw locations.

When evaluating implant survival by tooth location, molar implants exhibited a slightly higher SR (98.6%, 95% CI: 96.5%–99.6%) than premolar implants (97.5%, 95% CI: 95.9%–99.1%). While the premolar group displayed moderate heterogeneity (I ² = 38.69%, p = 0.114), suggesting some variability across studies, the difference in SRs between premolars and molars was not statistically significant (Q = 1.071, p = 0.301). These findings are consistent with other studies, which also reported high SRs for NDIs in both anterior and posterior locations, confirming the versatility of NDIs in supporting a range of prosthetic rehabilitations, including single‐tooth and multi‐unit restorations. ²⁴ , ³² , ⁵³ The observed heterogeneity in the premolar subgroup may be attributed to factors such as differences in implant design, surgical protocols, and patient characteristics, as highlighted in previous studies by Mangano et al. ⁴¹ and Olate et al. ⁵⁴

The analysis of SRs by implant material revealed no significant differences between Ti and TiZr implants, with both materials showing a pooled SR of 98.4% (95% CI: 97.6%–99.2%). These findings are consistent with the studies of Chiapasco et al. ¹⁴ and El‐Sheikh et al., ³⁵ where TiZr implants (e.g., Roxolid) demonstrated similar clinical outcomes to traditional Ti implants. TiZr alloys are increasingly used in NDI fabrication due to their superior mechanical strength and biocompatibility. ² , ¹⁵ The observed moderate heterogeneity in the TiZr subgroup (I ² = 40.31%, p = 0.113) suggests that variability in the implant designs or clinical protocols may have influenced the outcomes. However, the test for subgroup differences between Ti and TiZr implants confirmed that the difference in SRs was not statistically significant (Q = 1.497, p = 0.221), indicating that both materials provide comparable long‐term performance.

Implant SRs were also analyzed based on follow‐up duration. For studies with a follow‐up period of up to 5 years, the pooled SR was 98.2% (95% CI: 97.4%–99.0%) with low heterogeneity (I ² = 11.7%, p = 0.446), highlighting the short‐term reliability and predictability of NDIs. These results are consistent with the findings of Geckili et al., ³⁶ where NDIs placed over a short‐term follow‐up period exhibited minimal marginal bone loss and high SRs. However, for studies with follow‐up periods exceeding 5 years, the pooled SR decreased slightly to 95.1% (95% CI: 91.7%–98.5%), with significant heterogeneity (I ² = 72.3%, p < 0.001). This increase in variability is in line with the observations made by Mangano et al., ⁴¹ where long‐term implant survival was influenced by factors such as patient‐specific conditions (e.g., bone density, oral hygiene) and the clinical protocols employed. This reduction in SRs over longer follow‐up periods may reflect the cumulative impact of mechanical wear and biological factors over time. The observed heterogeneity in long‐term follow‐up studies underscores the importance of consistent patient management and standardized follow‐up protocols to better assess NDIs' durability.

The quality of the studies included in this meta‐analysis was moderate to high, with an average MINORS score of 17.5 out of 24. Most of the studies were retrospective, with a few employing prospective designs, limiting the control over biases such as selection bias and confounding factors. Despite these limitations, the studies generally had well‐defined aims, appropriate clinical endpoints, adequate follow‐up periods, and low attrition rates, which enhanced the reliability of the findings. However, none of the studies included control groups, randomization, or sample size calculations, limiting the results' internal validity and generalizability. The assessment of risk of bias in RCTs showed that all studies demonstrated low risk in terms of randomization, outcome assessment, data completeness, and reporting transparency, but high risk was noted for blinding of participants and personnel.

The overall results from this analysis suggest that NDIs are a reliable treatment option across different jaw and tooth locations and implant materials. Despite minor variations in SRs based on follow‐up duration and tooth location, the overall findings indicate that NDIs perform similarly across these subgroups. The absence of significant differences between jaw locations (maxilla vs. mandible), tooth locations (premolars vs. molars), and implant materials (Ti vs. TiZr) further supports the conclusion that NDIs are versatile and effective for a wide range of clinical indications.

The slight decrease in SRs observed for long‐term follow‐up studies and the variability in premolar SRs highlight areas for further research. Future studies with more consistent methodologies and extended follow‐up periods are needed to better understand the factors influencing the long‐term success of NDIs, particularly in patient‐specific conditions, surgical protocols, and implant design variations.

While this review provides valuable insights into the performance of NDIs, several limitations must be acknowledged. Most included studies were retrospective, which introduces the risk of selection bias and limits the ability to draw causal conclusions. Additionally, the lack of RCTs and control groups in many studies impedes the ability to directly compare NDI outcomes with other implant types under similar conditions. The variability in follow‐up durations across studies further complicates the interpretation of long‐term outcomes, highlighting the need for more uniform and extended follow‐up periods in future research. Also, the lack of standardized reporting underscores a critical gap in the current literature, emphasizing the need for more rigorous and uniform documentation of peri‐implant health and technical complications in future research.

The limited reporting of biological complications, such as peri‐implantitis and bone loss, also constitutes a significant gap in the literature. Future studies should prioritize standardized and comprehensive reporting of technical and biological complications to facilitate better comparisons and more robust conclusions.

Moreover, although platform‐switching and CrossFit systems were associated with fewer technical complications, the influence of other design factors (e.g., thread design, surface treatment) on long‐term implant success remains underexplored. Further research should focus on these design factors' impact on biological and technical outcomes, especially in longer‐term follow‐up studies.

This meta‐analysis provides strong evidence for the high SRs of NDIs and their reliability as an alternative to standard‐diameter implants in clinical practice. Despite some observed heterogeneity, the pooled results highlight the predictability and effectiveness of NDIs in various clinical settings. Future studies should address the gaps in knowledge regarding the long‐term outcomes and factors influencing implant survival, particularly in more challenging clinical scenarios.

The findings of this systematic review demonstrate that NDIs are a reliable treatment option for posterior rehabilitation, with high SRs and low complication profiles. These results are clinically significant as NDIs provide a viable solution for patients with anatomical limitations, such as reduced alveolar ridge width, without requiring extensive bone grafting procedures. Moreover, the observed consistency in outcomes across various implant materials and jaw locations reinforces the adaptability of NDIs in diverse clinical scenarios.

This review also highlights the high SRs of NDIs, with pooled estimates exceeding 95% across various jaw and tooth locations. However, the literature offers limited data on whether the restorations were splinted or non‐splinted. This distinction is vital since splinted restorations may help distribute occlusal forces more evenly, thus reducing the risk of technical complications such as screw loosening or implant fractures.

While several studies included in this review reported on technical complications, few explicitly mentioned the type of restoration connection (splinted vs. non‐splinted). Neglecting to account for this factor could lead to misleading conclusions regarding the success of NDIs in high‐stress areas. Future studies should prioritize this variable to clarify its impact on long‐term implant stability. For now, clinicians should remain cautious and consider splinting as a potential strategy to mitigate complications, especially in patients with elevated mechanical risk factors.

The omission of significant biological complications, particularly peri‐implantitis and marginal bone loss, may lead to an overestimation of implant SRs. Although peri‐implantitis has a considerable long‐term influence on implant stability, it was underreported in the included studies. This creates a potential reporting bias, as complications that may ultimately lead to implant failure are not consistently documented. Consequently, the SRs presented in this meta‐analysis may reflect immediate or short‐term outcomes rather than the cumulative impact of both mechanical and biological factors over time. Future studies should incorporate standardized definitions and reporting protocols for complications to ensure a more accurate evaluation of implant performance.

Limitations

Despite the strengths of the present pooled synthesis, several limitations should be taken into account. First, studies with varying methodological rigor were included, and some studies needed long‐term follow‐up data, which may limit the ability to assess the durability of NDIs over extended periods. Secondly, the heterogeneity in reporting outcome measures, such as definitions of implant success and complication types, made direct comparisons challenging. Additionally, while subgroup analyses were performed, the small sample sizes in specific categories may reduce the robustness of these findings. Finally, the review does not incorporate a direct comparison with RDIs, which limits the ability to conclude whether NDIs are superior or equivalent to RDIs in similar clinical conditions.

Recommendations for future research

To address the limitations identified in this review, including variability in methodologies and follow‐up durations, the following directions for future research can be proposed:

Multicenter RCTs: Such trials would enhance the external validity of findings by encompassing diverse patient populations, clinical protocols, and implant systems.

Standardized outcome measures: Adoption of frameworks like the Implant Dentistry Core Outcome Set and Measurement (ID‐COSM) would improve comparability across studies by ensuring consistent definitions for survival, success, and complications.

Extended follow‐up periods: Studies with at least five to ten years of follow‐up should be prioritized to capture long‐term outcomes and complications. Regular assessment intervals and comprehensive data on both technical and biological failures will provide a clearer understanding of implant longevity.

Patient‐reported outcomes: Incorporating measures such as quality of life and patient satisfaction will provide a holistic perspective on the success of NDIs, further guiding clinical decision‐making.

CONCLUSIONS

Based on the results of the screened literature, it can be concluded that NDIs demonstrate high survival and low complication rates for treatment of the partially edentulous patient with FDPs in posterior sites. In addition, the clinical performance is consistent across jaw location and implant materials.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest related to this research.

PATIENT CONSENT STATEMENT

This study does not involve individual patient data; therefore, patient consent is not applicable.

ADHERENCE TO GUIDELINES

This systematic review and meta‐analysis adhered to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta‐Analyses) guidelines. The PRISMA checklist is provided in the supplementary materials.

Supporting information

Supporting Information

Supporting Information

Pachiou A, Tsirogiannis P, Ioannidis A, Joda T, Sykaras N, Naka O. Narrow‐diameter implants for treatment with fixed restorations in the posterior region: A systematic review and meta‐analysis. J Prosthodont. 2025;34:670–685. 10.1111/jopr.14057

DATA AVAILABILITY STATEMENT

Data supporting the findings of this study are available from the corresponding author upon reasonable request. Due to the nature of this research, participant‐level data are not publicly available to maintain confidentiality. Any additional study‐level data, including extracted information and supplementary analyses, can be provided upon request.