Implications of Implant Framework Misfit: An Animal Study on an Ovine Model

According to numerous recommendations, clinicians always aim to provide implant prostheses that fit passively on implants.^1–3  An accurate fit of the implant prosthesis is believed to reduce the harmful stresses and strains within implants, screws, prosthesis, and the alveolar bone. An improperly fitting prosthesis is thought to initiate several biological and mechanical complications, such as screw loosening, joint instability, plaque accumulation, peri-implant bone loss, ceramic microcracks, and components fractures.^2,3 

While it is logical to aim for a passively fitting implant prosthesis, a certain level of framework distortion is unavoidable, and an absolutely accurate prosthesis fit is not achievable. Moreover, the inaccuracy of a prosthesis is increased by the numerous clinical and technical steps required for fabrication. Consequently, vertical, horizontal, and angular misfits in the range of several hundreds of microns can be introduced.^1,3,4  Currently, there is no clear guidelines on what is considered to be an acceptable implant prosthesis fit, and there is a lack of accurate clinical methods to assess implant prosthesis fit.^1,3  Therefore, it is worth evaluating the possible clinical consequences of prosthesis misfit.

Recently, with advances in digital dentistry, newer workflows have been introduced to shorten the duration of implant treatment, improve the patient experience, and increase the predictability of treatment. For example, implants can be digitally planned and placed by fully guided computer-aided placement protocols.^5–7  The favorable accuracy of the computer-aided implant placement has further simplified immediate implant restoration. Several authors discussed the production of immediate implant prosthesis in advance according to the virtually planned implants and before actual implant placement. Through such workflow, the implants are placed, and the prosthesis is immediately fitted on the implants.^8–10  Although this workflow has several obvious advantages, the fit of the prosthesis on the implants can be questioned, and the implications of this misfit should be investigated. This is even more important in the light of recent studies that indicated that the accuracy of implant placement according to the fully guided protocol was in the range of 1–2 mm.^5–7  Eventually, this may translate into a clinically misfitting prosthesis. Therefore, it is necessary to determine the effect of misfit of immediate prosthesis.

Few laboratory and animal studies have investigated the implications of implant framework misfit from biological and mechanical perspectives, and they demonstrated variable outcomes.^11–15  It has been hypothesized that the lack of a relationship between misfit and complications is related to the possibility of implant displacement and bone and prosthesis deformations.^12,13,16  This exploratory pilot investigation aims to evaluate the effect of deliberate framework vertical misfit on implant position and screw loosening. In addition, it evaluates the effect of timing of implant loading.

Ethics approval was obtained from the Florey Animal Ethics Committee, the Florey Institute of Neuroscience and Mental Health (17-080-UM). All animal handling was in accordance with the National Health and Medical Research Council, Australian Code for the Care and Use of Animals for Scientific Purposes. Three male merino sheep at the age of 3 years were included in the study. All mandibular premolars (total of 6 teeth per mandible) were extracted. The premolars were sectioned and surgically extracted. The extraction sockets were irrigated with saline to remove bone debris. The extraction sites were sutured with resorbable sutures (Polysorb, Covidien, Dublin, Ireland).

After 8 weeks, 4 implants (Southern Implants, Centurion, South Africa) were surgically inserted in each side of the mandible following the manufacturer's recommendations (Figure 1). All implants had external hex connections and were 3.75 mm × 13 mm. External hex implants were chosen because they allow direct measurements of the changes of the vertical gaps at the analysis phase of the study. Following implant placement, implant level impressions were taken with impression copings (Figure 2) and a custom-made sectional special tray (Vertex Dental). Tray adhesive (PVS tray adhesive, Kerr Corp, Orange, Calif) was applied on the trays, and heavy-body polyvinyl siloxane impression material (Kerr Extrude Extra type 1, Kerr Corp) was used for all the impressions. Healing caps were placed on the implants, and the wound was closed with resorbable sutures. Implant analogues were attached on the impression copings, and the impressions were poured with type IV dental stone (GC Fujirock EP, GC Corp).

Each pair of implants was restored by nonengaging implant level frameworks (Figure 3). Therefore, 4 frameworks were fabricated for each sheep. The right-side implants received immediate frameworks (2 days after implant placement), and the left-side implants received conventional loading frameworks (8 weeks after implant placement).¹⁷  For each side, one framework was well fitting (Figure 4), in which the framework cylinders were planned to fit accurately on the implants. The other framework was misfitting, in which a vertical gap of 0.5 mm was deliberately incorporated between the distal implant and framework cylinder (Figure 5). The casts were sent to a commercial milling center (Osteon Medical), and the frameworks were designed as a simple bar with a cross section of 3 × 3 mm, and they were out of occlusion to ensure they do not interfere with the function during chewing (Figure 6). The vertical gap was added virtually through the designing software.

After 2 days of implant placement, the immediate frameworks were attached to the right-side implants. The fitting procedure involved hand tightening one screw at a time, followed by tightening via a torque wrench to 35 Ncm. After 5 minutes, the screws were retightened by the torque wrench to compensate for any framework settling effect.^18,19  For the misfitting frameworks, the screw on the misfitting site was tightened first, followed by tightening the other screw. This ensured that the screws on the 2 implants shared the effect of framework misfit. After tightening, the screw access was sealed with light-body impression material.

After 8 weeks (first review), the immediate frameworks were removed, and implant level impressions were taken for all implants, as discussed previously. The generated casts were used to evaluate the alterations of vertical fit of the frameworks and implant orientation in relation to the framework. The immediate frameworks were reattached on the implants. In addition, the conventional frameworks were attached on the left-side implants. A similar screw-tightening protocol was followed for all frameworks.

At the second review, after 8 weeks, all animals were euthanized by administering an overdose of pentothal sodium. All frameworks were removed, and implant level impressions were taken for all implants. All steps of the interventions took place under general anaesthesia by induction with sodium thiopentone (12.0–20.0 mL 5%), which was maintained with a combination of isoflurane (2.0%) and oxygen administered via endotracheal tube. In addition, local anesthesia (bupivacaine hydrochloride 5.0 mg/mL) was administered at the surgical site to reduce postoperative discomfort. Before teeth extraction and implant placement, antibiotics (penicillin/streptomycin 3 mL/kg I.M) and preoperative anti-inflammatory (carprofen 4 mg/kg subcutaneously) were administered. For 3 days after surgery, postoperative anti-inflammatory and antibiotic agents were given. In addition, the mouth of each sheep was syringed with antiseptic mouthwash (chlorhexidine, 10 mL 0.2% aqueous) for 1 week after surgery. Through the whole duration of the experiment, sheep were kept in cages under daily monitoring.

The torque required to loosen each screw was measured on each implant as an indication of joint stability. This was done by setting a manual friction torque wrench (Astra Tech, Dentsply) to the minimal value and gradually increasing it in increments of 5 Ncm. The categorical torque values were 5, 10, 15, 20, 25, 30, and 35 Ncm. The loosening of each screw was attempted at the minimal value and increased incrementally until the screw was loosened.

The master casts and the casts produced from the first and second reviews were scanned by a laboratory surface scanner (Identica T300, Medit Identica, DT Technologies, Davenport, Ia). Implant level laboratory scanning bodies (ZFX Scan body, ZFX Dental, Zimmer Biomet) were attached on each implant analogue. Each implant pair was used separately for the evaluation. The generated STL files were imported to a 3-dimensional rendering software (Geomagic Studio, Raindrop, Geomagic Inc). The STL file was used and converted to a model with virtual implants. The implants of the first and second reviews were compared against the baseline implants. This was executed by superimposition of first and second review cast implants on the baseline master cast implants. The mesial implant was used as a reference, in which 3 widely distributed points were used for manual alignment. This was followed by automated best-fit alignment iteration to optimize the superimposition of the mesial implant. The correct alignment was confirmed by a heat map (Figure 7). Subsequently, the vertical gap on the distal implant was virtually measured from its center.

Because the number of included animals is small, only a descriptive evaluation of the results was used. The torque loss and vertical implant displacement were presented in graphs comparing the outcome of first and second reviews. For the immediate frameworks, the loosening torques of the first and second reviews were plotted, whereas for the conventional frameworks, only the results of the second review were available. For the vertical implant displacement, the results of the first and second reviews were included for the immediate and conventional frameworks. The outcome of the first review for the conventional frameworks was used as an indication on the accuracy of the impressions.

One implant failed before framework placement, and 3 implants failed after framework placement. A total of 3 frameworks (immediate fitting, conventional fitting, and conventional misfitting) were affected and excluded from the analysis of the study. Therefore, a total of 2 immediate fitting, 3 immediate misfitting, 2 conventional fitting, and 2 conventional misfitting frameworks were included in the study.

At the first review of the immediate frameworks, one of the screws retaining the fitting framework had a low torque of 15 Ncm, whereas the other screws of the fitting frameworks had a loosening torque in the range of 30 Ncm to 35 Ncm. At the second review, the torque loss values of the screws retaining the fitting frameworks were slightly greater and ranged from 25 Ncm to 35 Ncm (Figure 8). On the contrary, the screws retaining the misfitting frameworks had noticeably lower torque values at the first review (Figure 9). Two screws had 5 Ncm loosening torque, and one screw had a loosening torque of 10 Ncm. The rest of the screws had loosening torques in the range of 20 Ncm to 25 Ncm. Of note, at the second review, the loosening torque values increased to the range of 25 Ncm to 35 Ncm, which is comparable with the loosening torque of the screws retaining fitting frameworks.

The conventional fitting frameworks had loosening torques in the range of 25 Ncm to 35 Ncm in the second review (Figure 10). However, the conventional misfitting frameworks had generally less loosening torque (20–30 Ncm), which is more than the loosening torque for the immediate misfitting frameworks at the first review (Figure 11) but less than the loosening torque of the second review of the immediate misfitting frameworks.

The vertical displacement of the distal implants supporting the immediate fitting frameworks was 0.02 mm and 0.07 mm for the first review and 0.05 mm for the second review (Figure 12). This magnitude of displacement indicates minimal changes in the implant position. However, for the immediate misfitting frameworks, the vertical displacement of the distal implant was in the range of 0.21 mm to 0.54 mm for the first review and 0.65 mm to 0.83 mm for the second review (Figure 13). This is indicative of vertical displacement of the immediately restored implants if the framework has vertical misfit.

As anticipated, the conventionally restored implants showed minimal vertical movement at the first review prior to restoration. This was 0.04–0.06 mm for the fitting implants and 0.07–0.10 mm for the misfitting implants. Therefore, the impressions of the first review indicate that there is an error related to the impression, cast fabrication, and scanning that can be up to 0.1 mm. For the second review of the fitting frameworks, the results were minimal vertical displacement of the distal implant of the fitting frameworks (in the range of 0.04–0.09 mm; Figure 14), which was similar to the immediately restored implants with fitting frameworks. More vertical displacements were observed of the distal implants retaining the misfitting frameworks (0.12–0.48 mm; Figure 15). Therefore, it is clear that the magnitude of implant movement for the conventional frameworks is less than immediate frameworks by about 50%.

The present pilot study showed that the misfitting frameworks can be associated with permanent implant displacement and less loosening torque and joint stability than the fitting frameworks. This seemed to be more obvious for the immediately placed frameworks. However, at the second review, the loosening torque values were similar for all frameworks. This may be due to the permanent implant vertical displacement toward the framework, which may have reduced the magnitude of misfit after attaching the misfitting frameworks. Therefore, the fixation of frameworks with vertical misfit may result in a long-term reduction of framework misfit, as shown by several earlier studies.^{11,12,14–16}  Although some implants failed in the current study, none of the implant failures can be attributed to the misfit of the frameworks. Previous research supports that implants can withstand static loads that may occur due to fitting frameworks with misfit without affecting bone quantity and quality around implants.^{11,12,20–23}  On the other hand, this study indicated that implant movements due to misfit was primarily in the vertical direction. This may be due to the nature of the introduced gap and the nonengaging framework fitting surfaces.

Although several laboratory and animal studies have evaluated the effect of framework misfit on implants, this study is unique in evaluating the effect of the time of fitting the frameworks. Vertical misfit was applied because it is simpler to produce in comparison with horizontal or angular misfit.^11,12,15  Further, the quantifications of its effect can be reliably monitored. A vertical gap magnitude of 0.5 mm was selected, as it challenges the implants and, at the same time, was shown by earlier studies to disappear as the retaining screws were tightened.^12,24,25  Further, this vertical gap magnitude was similar to the vertical errors introduced by computer-guided implant placement^5–7  and can subsequently be translated to prosthesis misfit, in which the prosthesis is produced prior to surgery.^8–10 

According to the present study, the retaining screws were influenced by a vertical misfit in the range of 0.5 mm. This agrees with several laboratory and mathematical studies^18,19,26,27  that confirmed the vulnerability of screws to loosening due to misfit. When the screw is tightened, the preload subjects the screw threads to tension and elastic deformation. This creates a clamping force between the screw and the implant, which keeps the prosthesis and the implant in contact. Screw stability is maintained as long as the preload preserves the elastic deformation of the screw surface. In misfit situations, part of the preload is used to approximate the framework-implant interface,²³  by deforming the framework and pulling the implant toward the framework fitting surface. This renders the screws more susceptible to loosening or deformation. In the present study, in addition to framework misfit, the vertical implant displacement may have accelerated the loss of preload within the retaining screws. This was obvious at the first review for the immediate misfitting frameworks, in which the torque loss values were considerably less than for the immediate fitting frameworks. Similarly, Jemt et al¹³  reported that the loosening torque average of the misfitting frameworks was 56%–66% of the loosening torque of the fitting frameworks.¹³  Retightening the screws seemed beneficial in retaining the torque at the second review and increasing the loosening torque to levels similar to the fitting frameworks. This may indicate that the retightening of the screws has compensated for implant displacement, and it can be a recommended practice.^18,19 

Several studies discussed the potential of implant displacement due to static load application in the form of fitting frameworks with deliberate misfit.^11,12,28  Specifically, this was observed in the vertical direction, which is consistent with the potential displacement of implant toward the framework fitting surface in a way that reduces or eliminates framework misfit. As the retaining screws were tightened, the gaps between the implants and the framework were reduced.^24,25,27  In the present study, the gap reduction was found to be in the range of 0.66–0.83 mm for immediate and 0.12–0.48 mm for conventional misfit frameworks, which was of similar magnitude to other implant studies that evaluated the effect of vertical misfit.^11,12  A study on rabbit tibia found the 1-mm vertical gap on the middle integrated implants was reduced by about 0.30 mm because of implant displacement (about 0.05–0.20 mm) and framework deformation (about 0.15 mm).¹²  Likewise, in another animal study, the gap reduction ranged from 40%–70% of the intended gap (0.50 mm on a middle implant).¹¹  The misfit reduction was 0.41 mm for immediate implants and 0.20 mm for integrated implants. Similarly, another study applied horizontal static orthodontic forces on implants and found that under certain force magnitudes, the implants can move in the direction of the applied forces.²⁸ 

Several mechanisms were proposed to explain implant displacement through the bone due to static load application. The preload-induced peri-implant strains are similar to Frost's minimum effective strain for remodeling, which eventually increase and conserve bone in response to increased usage.^29,30  Therefore, because of the application of constant static load, the bone is subjected to viscoelastic deformation and may remodel according to the new implant position by replacing the stressed bone with unstressed bone. As a result, the implant may permanently move and adapt to the existing framework fit.^14,15  In line with this observation, a histomorphometric evaluation found more bone deposition around the implants retaining misfitting frameworks than the implants retaining fitting frameworks.¹³ 

Interestingly, frameworks placed immediately in this study were associated with more noticeable permanent implant displacement than conventionally placed frameworks, which confirms the observations of an earlier study.¹¹  Duyck et al found the immediately loaded implants with misfitting frameworks had greater tendency to be displaced (about 70% displacement) than did conventionally loaded implants (about 40%). As considerable implant displacement was observed in the first review of this study, it is logical to assume additional mechanisms of implant displacement than bone remodeling. In addition to the mentioned earlier mechanisms of implant displacement, immediately loaded implants are distinguished with their greater susceptibility to displacement within the bone. One of the proposed mechanisms of preload-induced implant displacement is the microfracture of the bone around the implant threads¹¹  that occurs when the applied stress exceeds the fatigue strength of the peri-implant bone. Thus, as the retaining screws were tightened, it is easier for the implant to move vertically toward the framework fitting surface. Such a mechanism may be responsible for immediate reduction of the gap observed at the first review.^11,31–33  However, because the torque loss was considerable in the first review, most likely, this mechanism did not completely close the vertical gap at the time of insertion. The additional mechanism is the viscoelastic stress relaxation,³³  in which a recently prepared bone experiences reduction of stresses within a few weeks. These 2 displacement mechanisms may be responsible for the accelerated implant displacement under the immediate frameworks in comparison with the conventional frameworks. The microfracture and viscoelastic stress relaxation were also reported to stimulate adaptive bone remodeling and are anticipated to be followed by modeling of the bone at the implant interfaces.^32,34  The permanent replacement of bone is of longer duration and may explain the greater implant movement and more torque retention observed at the second review. Clinically, the primary stability is reduced over a few weeks and replaced with secondary stability when new bone is formed and matured.^35,36 

The outcome of this preliminary pilot study is relevant to recent digital workflows, in which computer-aided implant placement is executed and an immediate prosthesis is produced in advance before implant surgery.^8–10  Although computer-aided implant placement techniques are generally accurate, they still exhibit a certain degree of error that can range from 1 mm to 2 mm in the horizontal and vertical directions.^5–7  This is an accumulated error of digital planning, scanning, guide fabrication, guide fit discrepancy, tolerance of drills and sleeves, and implant deviation through the osteotomy.^5–7  Nevertheless, according to this study, the immediately loaded implants may displace in a way to improve the prosthesis adaptation. However, because of potential implant displacement, screw loosening may occur and may require retightening after a few weeks of service to ensure a tight connection with minimal leakage. On the other hand, although misfit-induced screw instability can be related to the deliberate misfit, it is difficult to assume screw loosening will occur due to the presence of reasonable and clinically undetectable misfit. One study that attempted to relate the measured misfit from the manufactured framework to screw loosening did not find a significant relationship between fit level and screw stability.³⁷  Similarly, a clinical study found no relationship between marginal bone loss and misfit of clinically acceptable frameworks.²⁰ 

The results of this study should be taken with caution. It is difficult to generalize the outcome of this study because of the limited number of animals and the difficulties in extrapolating the outcome of an animal study on humans. The limited number of animals may explain the considerable variation in the vertical displacement and torque loss values. It is important to note that for some implants, the magnitude of movement was larger than the intended gap. This could be due to errors associated with the methodology and complexity of implant displacement.^11,12  While this study did not reveal major consequences of a vertical framework misfit, it does not justify fitting inaccurate prosthesis. Further, framework inaccuracies can be more complicated and may not always manifest as vertical errors. As a result, a simple closure of misfit gap may not always be predictable, and premature binding and interferences may prevent gap closure. Moreover, the retaining screws can be subjected to excessive stresses,^15,27  and the remaining gap may cause biological complications. This animal study was limited in considering only static forces and excluded dynamic forces. Clinically, the prosthesis will be subjected to centric and eccentric occlusal loads, which may affect the implant response to misfit. There are indications from earlier studies that dynamic forces are associated with more marginal bone loss and complications than static forces.^22,38,39  Data from human clinical studies are needed to support the results of this study. This involves the mechanical and biological outcomes of immediate prostheses fabricated prior to implant placement.

Within the limitations of this exploratory pilot animal study, it can be concluded that the implant frameworks of the present study with vertical misfit of 0.5 mm seemed to be associated with less loosening torque of the retaining screws and more implant displacement. Immediate loading of the implants may increase the implant displacement. Retightening the retaining screws during the maturation of bone appears to increase the loosening torque. Given that the present study was conducted in an animal model, clinical studies are needed to confirm the outcome.

The authors acknowledge the guidance provided by Professor Warwick Duncan, the University of Otago, and Associate Professor Arun Chandu, the University of Melbourne, for their surgical input. The authors would like to thank Mr Alan McDonald, Vasculab, for conducting the anaesthesia and monitoring the sheep. The authors would like to thank Professor Rodrigo Marino, the University of Melbourne, for his recommendations on statistical analyses and data presentation. This study was funded by the Early Career Researcher Grant, the University of Melbourne.

The authors declare no conflict of interest in the products listed in the article.