Accuracy of dental implants inserted in clinical studies using static implantation guides: a systematic review and a meta-analysis

Background One of the main tasks of dental implantation is the precise insertion of the implant into the edentulous jaw. The purpose of a study was to determine the accuracy of implants inserted using static implantation guides in different conditions. Methods A search for clinical studies was performed in the PubMed databasis. The implantation procedure had to be performed on humans in vivo by using a static surgical guide. Pre and postsurgical CT scans were compared to evaluate the deviation of implants. A meta-analyses was performed to calculate the mean implant collar, apex, depth and axis deviations. A meta-regression analyses was performed by grouping study subgroups according to the method of surgery, implant insertion, jaw, guide support tissue, and the method of guide xation. Results 19 studies were included in the meta-analysis. Data on 1872 implants were analyzed. Mean deviation was 1.25 mm (95% CI: 1.00, 1.51) at the implant entry point and 1.52 mm (95% CI: 1.18, 1.86) at the apex. Mean axis deviation was 3.47o (95% CI: 3.01, 3.94). Depth deviation was 0.15 mm (95% CI: -0.36, 0.66). The following factors had a positive inuence on the implant precision: open surgical technique, fully guided implant insertion and xation of the surgical guide. Conclusions The surgical technique, implant insertion method, jaw, guide supporting tissue and xation has an effect on the deviation of inserted implants.

This systematic review was based on the PRISMA search guide for systematic reviews [21]. The review protocol is registered in the PROSPERO database for systematic reviews [22]. Registration number: CRD42017076584

Search strategy
The search was performed by one researcher using the MEDLINE (PubMed) data system. Systematic data search combination: dental AND (implant OR implants) AND (guide * OR computer) Filters used in search: • Studies published between 2012/09/01 and 2019/07/01 • In English Last date for systematic data search: 2019/07/01 During the rst stage of the search, titles of publications were read and works that did not t the topic were rejected. During the second stage, the abbreviations of the publications corresponding to the topic were ltered using the selection and exclusion criteria of research papers. An additional, manual search was carried out using the bibliographies of the selected publications with identical criteria for the selection and exclusion of research papers. In the nal stage, texts of the selected publications were read. Based on the criteria for inclusion and exclusion of works from the meta-analysis, selected publications were included in the meta-analysis.

Inclusion and exclusion criteria
The criteria for inclusion of research papers: (1) types of publications: clinical trials; (2) the subject is static implantation guides; (3) the accuracy of inserted implants is measured; (4) implants inserted in humans in vivo. Criteria for exclusion of research papers: (1) types of publications: literature reviews, meta-analyzes, clinical case studies, pilot studies that describe less than 10 patients; 2) implantation procedures on animals; (3) usage of cheekbones, the sphenoid bone (pterygoids), orthodontic mini-implants; (4) implant position deviations were not measured; (5) in vitro or ex vivo studies. Criteria for inclusion of research papers in meta-analysis: (1) implant osteotomy is prepared using a static surgical guide; (2) preoperative and postoperative computed tomography (CT) was performed on the same X-ray machine using identical parameters; (3) at least 10 patients participated in the study; (4) The position of inserted implants is measured using software and is based on postoperative CT scans; (5) inserted implant deviations are measured in a three-dimensional space; (6) The deviations of inserted implants are indicated in a standardized form (described below). Criteria for exclusion of metaanalysis: (1) static implantation guides used were modi ed by the author; (2) used dynamic computer assisted implantation; (3) less than 10 had implants inserted; (4) implant osteotomy preparation was completed by the freehand placement technique; (5) no preoperative or postoperative CT control; 6) implant deviations were measured without implant insertion; (7) The deviations of inserted implants are measured in a two-dimensional space.

Data extraction:
A standardized form was used for data sampling. The data criteria of the publications is as follows: • Number of patients undergoing surgery The linear deviations of the implant position had to be measured between the centres of the apex and collar of the planned and inserted implants, calculating the global deviation -the distance between two points in a three-dimensional space. Global deviation is calculated in accordance with the distances in the buccolingual, mesiodistal, and apicocoronal planes. Linear depth deviations had to be measured in the apicocoronal plane by measuring the vertical distance between the centres of the planned and inserted implant collars. Depth deviations were given as positive values if the implants were drilled too deep. If the implants were not inserted deep enough, the depth deviation was given as a negative value. In publications, all linear deviations are presented in millimetres or micrometres, while in the metaanalysis linear data is presented in millimetres. The angular deviation had to be measured in a three-dimensional space by drawing out the centre line passing through the centres of the implant collar and apex (axes) and measuring the angle formed between the intersected axes of the planned and inserted implants (Fig. 1). Angular deviations were measured in degrees. All deviations had to be established using hardware comparing pre-and post-implantation CTs. If in the publications linear collar and apex deviations were measured separately in the three planes and did not include global deviations, the following mathematical formulas were used to standardize the results, and the global deviation and the standard deviation of said deviation were calculated manually. The calculations were performed twice. If the results did not match, they were calculated a third time.
Data that did not meet the standards used in this study was exluded from the meta-analysis. SD (x, y, z) = Standard deviation in x, y, z planes Formulas adapted from Tahmaseb and others [23].
Quality assessment: Article bias was assessed using the Cochrane method evaluating the: random sequence generation, allocation concealment, de ned inclusion / exclusion, blinding of participants and / or personnel, blinding of outcome assessment, incomplete outcome data, selective reporting. The evaluation was conducted by two researchers independently of each other and disagreements were resolved during the discussion. (Table 1) Statistical analysis: The statistical analysis was conducted using Comprehensive Meta Analysis ((CMA) [computer program]. Version 3.0. Englewood, USA, Biostat, 2017). Heterogeneity of the research was assessed using Cochrane's Q and I2 tests. The I2 test values were interpreted according to Higgins et al. [24], respectively: >25% = slight heterogeneity, > 50% = moderate heterogeneity, > 75% = signi cant heterogeneity. Separate meta-analyses were performed by subdividing the data by type of deviation (angular, linear deviations of depth, apex and collar). A metaanalysis was conducted using the Z test according to the Inverse Variance Weighted Random Effects Model. A meta-regression between different subgroups was conducted using the χ2 (Chi-Square) test. The data was divided into the following subgroups: • Jaws: upper, lower.
• Methods of guided implantation: open, closed.
• Implant insertion method: with guide, without guide.
• Type of guide support tissue: mucosa, teeth, bone.
The signi cance level of the statistical analysis was P < 0.05. The con dence intervals are 95% (95% C.I.) and the effect sizes are presented in the Forest plot graphs.
Vercruyssen and others studied the same population in two separate studies [35,38], however they measured different deviations: linear deviation of the collar and apex of the implant, angular deviation of the implant axis [35], and deviation of the implant placement depth [38]. Both articles were included in the meta-analysis.

Discussion
The studies analyzed in this literature review showed signi cant heterogeneity (I2 > 98%). Seeing as how the implantation protocols of individual studies differed, this heterogeneity was predicted prior to meta-analysis. Such heterogeneity results are observed in all metaanalyses published over the past 15 years that investigate implant accuracy using static implantation guides [3,[44][45][46][47][48]. According to Jung et al. [3], published in 2009, The International Team for Implantology (ITI), at the ITI Annual Conference, decided to ignore the heterogeneity of the research found in this review and to conduct a meta-analysis. During the selection of articles it was decided to not take into account the technique of computed tomography and the experience of the operating surgeons. Cone beam computed tomography (CT) was most commonly used in this study. Multi-layer and spiral computed tomography were used much less frequently. In a study conducted by Arisan and others, the accuracy of guided surgical implants was compared using cone bean and multi-layer computed tomography equipment. According to the results, there is no signi cant difference between these types of computed tomography. An identical conclusion was reached by Poeschl and others [49]. There is considerable inconsistency in the scienti c literature regarding the in uence of the clinical experience of the implanting surgeon on the accuracy of implant insertion. Some studies mention that the implanting surgeon's experience has a signi cant impact on implant insertion accuracy [50][51][52], while other studies mention that surgeon experience is not a signi cant criterion for examining implant accuracy [41,43,53] Wiele et al. [41], found that inexperienced surgeons achieved signi cantly lower errors in implant collar, apex, and depth. The angular deviation was higher in the inexperienced surgeon group, but this was not a statistically signi cant nding. In this study, a group of inexperienced surgeons consisted of periodontologists with no guided implantation experience and limited freehand placement experience. Cassetta et al. [43], discovered the opposite in their study -the implantation errors of the inexperienced surgeons were more signi cant in the collar and apex areas, but angular errors were smaller compared to experienced surgeons. These results were not statistically signi cant.
In implantology literature, it is accepted that a 2 mm deviation between the planned and actual position of the implant is clinically signi cant due to this distance being recommended between the implant and surrounding anatomical structures [20]. The mean implant position deviations obtained in this literature review are similar to the implant position deviations obtained in previous meta-analyses [3,[44][45][46][47][48]. In the literature, the mean implant collar deviation in ranges from 1.0 mm [45] [47], in this study it is 1.52 mm (95% CI: 1.18-1.86 mm), meaning that inserting multiple implants close to one another and leaving a recommended minimum distance of 3 mm of implants, their apex parts may intersect. The angular deviation in the literature ranges between 3.98o [44] and 5.7o [47], in this review it is 3.47o. This angular misalignment allows the prosthesis of inaccurately inserted adjacent implants without the use of hexagonal supports or special angular support [3,5]. Implant depth deviation was analysed by only two meta-analyses [44,45], yielding deviations of 0.74 mm and 0.6 mm deviations. The 0.15 mm deviation obtained in this meta-analysis is signi cantly smaller than that of previous meta-analyzes. This result can be explained by the fact that in both studies [44,45], the authors gave the vertical deviations of implant position only positive values, regardless of whether the implants were inserted too deep or not deep enough. In this study, the deviation of implants that are not deep enough is given a negative value, which results in a lower mean depth deviation when summarizing the results of several studies. The average implant depth deviation obtained in this meta-analysis is within the safe distance to adjacent anatomical structures, so implantation near important anatomical structures (maxillary sinus, inferior alveolar nerve) is safe. In most older meta-analyzes, implantation results obtained in in vivo, ex vivo, and in vitro studies were grouped together [44][45][46][47], which makes it di cult to compare subgroup analysis results from previous meta-analyses. The in uence of surgical methods on implant deviations was not analyzed in any of the reviews. The results of this review make it apparent that the open surgical implantation method result in smaller implant position deviations compared to the apless method, but this difference is not statistically signi cant. The open implantation method has drawbacks: due to the rupture of the cervix, resorption results in the loss of a small amount of the alveolar bone and can result in gum recession, postoperative swelling, discomfort and potential hematomas [54]. Dental implantation techniques are characterized by less postoperative discomfort, swelling and less frequent hematomas, while gum recessions do not develop as long as the bone is not damaged [53]. Currently, there is a lack of data in scienti c literature on the in uence of surgical methods on implant position deviations. The increased number of clinical trials conducted under standardized conditions would allow the formulation of speci c guidelines for clinical practice. Meta-analyzes by Ramos et al. [44], and Asche et al. [45], also suggest that implant insertion through a surgical guide results in smaller implant position deviations compared to freehand implant placement. According to Asche et al. [45], a xed surgical guide reduced implant position deviations more than a freely held surgical guide. The results of this meta-analysis regarding the type of surgical guide support do not coincide with those of previous reviews [45,46,48]. According to these reviews, surgical guides that are xed on the teeth have the smallest deviations of the implant position, while guides xed on the alveolar bone have the largest deviations. The results obtained in this meta-analysis are the opposite: the smallest deviations of the implant position are observed in the alveolar bone area, and the largest is recorded when the guide is xed on the adjacent teeth (Tables 2-4). This result may have been distorted by the extremely small number of scienti c articles and research subgroups that have used guides xed on the alveolar bone and teeth [27,33,35,38,42].
This literature review did not include studies using freehand implant placement methods. The main reason for this exclusion was the lack of scienti c literature and the heterogeneity of the methods used in articles. The systematic search has yielded 9 publications measuring the accuracy of freehand placement implants [35,38,[55][56][57][58][59][60][61]. The methodology of all publications varied greatly. In 3 studies, the accuracy of implants was tested ex vivo [55][56][57], one of them, that of Sherer and others [58], investigated the accuracy in porcine jaws. The other two studies investigated the accuracy of implants in human jaws ex vivo. 4 studies investigated the accuracy of freehand implant placement in vitro [58][59][60][61]. In all of the above studies, implantation accuracy was measured by different methods. Nickenig et al. [58], used cone beam computed tomography before and after implant insertion, Park SJ scanned jaw patterns after implant insertion using a threedimensional scanner, and Park C. et al measured postoperative implant positions using a coordinate measuring device. Two studies measured implant position deviations on humans, in vivo conditions, comparing implantation with surgical guides to non-surgical guides [35,38]. Both studies looked at the same population and measured different implant position deviations. When using the freehand placement method, the mean collar deviation of the implant was 2.7 mm (ranged 0.3-8.3 mm), whereas in identical conditions using a surgical guide, the implant collar deviation was 1.38 mm. The apex deviation was 2.9 mm (ranged 0.5-7.4 mm), using a guide − 1.58 mm.

Conclusion
The surgical technique, implant insertion method, jaw, guide supporting tissue and xation has an effect on the deviation of inserted implants. Due to deviations in the position of implants during guided implantation, it is recommended to use xed surgical guides and open implantation. These methods increase the accuracy of implantation and are safer than freehand placement. More clinical trials are needed to research the impact of implantation methods on implant deviation to obtain greater statistical signi cance.

Declarations
Ethics approval and consent to participate: Not applicable.

Consent for publication: Not applicable.
Availability of data and materials: All data generated and analyzed in this review are included within the article.
Competing interests: The authors declare that they have no competing interests with regards to authorship and/or publication of this paper.
Funding: Not applicable Author contributions: All authors have contributed equaly. All authors have read and approved the manuscript.   Linear deviation of implant depth categorized by jaw.

Figure 25
Linear deviation of implant depth, categorized by type of guide support tissue Figure 26 Linear deviation of implant depth categorized by guide xation method