Novel approach to assessing the primary stability of dental implants under functional cyclic loading in vitro: a biomechanical pilot study using synthetic bone

Abstract

Purpose

This pilot study was conducted to develop a novel test setup for the in vitro assessment of the primary stability of dental implants. This was achieved by characterising their long-term behaviour based on the continuous recording of micromotions resulting from dynamic and cyclic loading.

Methods

Twenty screw implants, each 11 mm in length and either 3.8 mm (for premolars) or 4.3 mm (for molars) in diameter, were inserted into the posterior region of 5 synthetic mandibular models. Physiological masticatory loads were simulated by superimposing cyclic buccal-lingual movement of the mandible with a vertically applied masticatory force. Using an optical 3-dimensional (3D) measuring system, the micromotions of the dental crowns relative to the alveolar bone resulting from alternating off-centre loads were concurrently determined over 10,000 test cycles.

Results

The buccal-lingual deflections of the dental crowns significantly increased from cycle 10 to cycle 10,000 (P<0.05). The deflections increased sharply during the first 500 cycles before approaching a plateau. Premolars exhibited greater maximum deflections than molars. The bone regions located mesially and distally adjacent to the loaded implants demonstrated deflections that occurred synchronously and in the same direction as the applied loads. The overall spatial movement of the implants over time followed an hourglass-shaped loosening pattern with a characteristic pivot point 5.5±1.1 mm from the apical end.

Conclusions

In synthetic mandibular models, the cyclic reciprocal loading of dental implants with an average masticatory force produces significant loosening. The evasive movements observed in the alveolar bone suggest that its anatomy and yielding could significantly influence the force distribution and, consequently, the mechanical behaviour of dental implants. The 3D visualisation of the overall implant movement under functional cyclic loading complements known methods and can contribute to the development of implant designs and surgical techniques by providing a more profound understanding of dynamic bone-implant interactions.

Graphical Abstract

INTRODUCTION

The long-term success of endosseous dental implants hinges on the complete osseointegration of the implant body into the alveolar bone [1]. Initially, a strictly 2-stage procedure was employed, as proposed by Brånemark et al. [2] in 1969, in which the implant body was left unloaded without the artificial dental crown for up to 6 months to ensure complete osseointegration. However, it is now widely recognised that mechanical loads are not per se the primary threat to osseointegration. Instead, the risk lies in excessive micromotions at the bone-implant interface due to insufficient mechanical anchorage. If the relative motion between the bone and implant exceeds a critical threshold of approximately 150 µm, it can disrupt the osseous healing process, trigger the formation of non-mineralised and fibrous tissue, and ultimately lead to implant loss [3]. In contrast, if a sufficiently strong initial mechanical bond exists between the implant and bone—termed primary stability [4]—extensive clinical studies have demonstrated high success rates, even for immediate and early loading of implant-supported single-tooth restorations [5].

In clinical practice, primary stability is typically assessed indirectly through the measurement of insertion torque during implant placement. The insertion torque reflects the forces and stresses between the implant and the bone and is therefore viewed as an indicator of bone quality and the likelihood of implant survival in vivo [3,6]. However, given the complex nature of mastication, primary stability is influenced by multiple factors, which cannot be fully quantified using insertion torque alone.

Various biomechanical studies have been conducted to further characterise primary stability in vitro or in silico using additional parameters, such as removal torque [7], pull-out force [8], load to failure [9] or implant stability quotient [10]. Despite their critical importance for osseointegration and provision of a direct measure of the relevant mechanical anchorage, few biomechanical studies have addressed implant deflection and the relative motion between the implant and the surrounding bone. Current methodologies employ static unidirectional loads to examine implant micromotions that arise from varying bone qualities [11], insertion torques [12] and implant orientations [13]. However, cyclic multidirectional loads, which are particularly physiologically common in the region of the posterior teeth [14,15], have not been incorporated.

Studies involving the use of cyclic multidirectional loads to explore the dynamic impacts of settling and fatigue have primarily been concentrated on specific components, such as the implant-abutment connection [16,17,18]. In these instances, excessive micromotions and fatigue are linked with other serious clinical complications, such as sudden abutment screw failure [19] or inflammatory reactions triggered by micro-leakage [20].

Given that dynamic effects can potentially lead to fatigue and sudden failure in the form of implant loosening, even well below the ultimate failure loads [21], the present pilot study proposes a novel approach for the in vitro characterisation of dental implant performance that considers the dynamic nature of mastication. In this technique, the primary stability of dental implants is directly examined through the relative movements between the implant and bone under functional cyclic loading. In long-term tests, implants in the posterior teeth region are subjected to alternating and multidirectional loading, while the resulting micromotions are concurrently recorded.

Our research was designed to enhance current methodologies by examining micromotions under functional cyclic loading over time, thereby providing a more profound understanding of the mechanisms, locations, and timing of implant loosening. These insights can be utilised to thoroughly explore the impact of varying implant designs and surgical techniques under controlled and reproducible conditions. This will ultimately contribute to the enhancement of dental implant performance, thereby improving future clinical outcomes and patient care.

The objective of this pilot study was to evaluate the functionality of the setup, the applied loading scenarios, and the optical 3-dimensional (3D) measurement by comparing 2 implant systems and to establish the presented methods as a foundation for further investigations.

The following hypotheses were investigated:

• 1) The experimental setup allows for the direct recording of micromotions between the implant and the bony environment resulting from applied loads.

• 2) In long-term tests, cyclic alternating loads can result in implant loosening, even when these loads are well below the maximum tolerable levels.

• 3) The bone surrounding the dental implants is crucial in distributing and dissipating masticatory forces, thereby playing an important role in the assessment of the in vivo performance of dental implants.

MATERIALS AND METHODS

Test setup

A custom-built 2-axis experimental setup was used to mimic the physiological loading of the lower posterior teeth during mastication. This setup simulated the alternating multidirectional loading of the dental crown by superimposing lateral movements of a mandibular component with a vertically applied chewing force (Figure 1).

Figure 1. Experimental principle for the superimposition of lateral movements of the mandibular component with a vertically applied chewing force. (A) Test setup concept. (B) Detailed view of the alternating off-centre loading of the dental crown.

The experimental setup incorporated a mandibular component, which consisted of an embedding tray that ran horizontally on linear bearings. This tray was driven by an eccentric mechanism designed to produce sinusoidal reciprocal movement. The mandibular bone was aligned and rigidly fixed in the embedding tray, enabling the implementation of cyclic mandibular movements along a buccal-lingual trajectory.

The vertical component simulated the maxillary antagonist, applying a vertically directed chewing force through a rounded stainless-steel load stylus. Due to the concurrent buccal-lingual movement of the mandible and the resulting relative motion between the 2 components, the load stylus reciprocally slid across the occlusal surface of the dental implant, alternating between the buccal and lingual cusp. The exclusive vertical linear bearing of the load stylus facilitated continuous contact with the dental crown, thereby ensuring uninterrupted force transmission. However, the point and direction of force application were permanently changing.

An optical 3-camera image correlation system (Q400; Limess Meßtechnik und Software GmbH, Krefeld, Germany) was employed to conduct non-contact measurements of the micromotions resulting from the loading process (Figure 2).

Figure 2. Experimental setup with optical 3-camera image correlation system.

Specimens

In this study, 5 synthetic bone models of the mandible (Model 8950; Synbone AG, Malans, Switzerland) were used to test dental implants in the region of the first premolars (region 34/44) and the first molars (region 36/46). The implant systems investigated were the CAMLOG SCREW-LINE and CONELOG SCREW-LINE (Camlog Vertriebs GmbH, Wimsheim, Germany) systems. These devices consist of slightly conical screw implants with a rounded apical end and a self-tapping thread, suitable for both immediate and delayed immediate implant placement. The study incorporated a total of 10 implants for each system, further divided into 5 implants for the premolar region and 5 for the molar region. Four experimental groups were created: Camlog premolar (CamPM, n=5), Camlog molar (CamM, n=5), Conelog premolar (ConePM, n=5) and Conelog molar (ConeM, n=5).

Both implant systems are externally identical in the threaded section, with a length of 11 mm and a diameter of 3.8 mm for premolars and 4.3 mm for molars. However, they differ in their internal configurations and the nature of their implant-abutment connections. The Camlog system features a distinctive tube-in-tube implant-abutment connection, characterised by a cylindrical interface. In contrast, the Conelog system employs a conical implant-abutment connection in which the interface is angled at 7.5°.

The superstructures were uniquely designed single crowns constructed from zirconia (GQ Quattro Disc Med OP; Goldquadrat GmbH, Hannover, Germany). The design of these crowns was stylised to mimic the geometry of natural teeth, with a height of 8.5 mm and an occlusal surface diameter of 6.0 mm for premolars and 7.5 mm for molars. These crowns featured a hemispherical fissure with an entry angle of 30°, replicating the geometric conditions of natural cusps [17,22]. The crowns were affixed to the abutments using Multilink Hybrid Abutment (Ivoclar Vivadent, Ellwangen, Germany).

Test preparation

An experienced oral surgeon from the Dental Competence Centre Leipzig (DENTALE) placed all implants in accordance with the manufacturer’s specifications. Premolars and molars from a single implant system were implanted on one side of the mandible, which was randomly assigned. The implants were placed at either positions 34 and 36 or positions 44 and 46.

Following implantation, the synthetic mandible bones were bisected in the sagittal plane. Each half of the mandible was then individually embedded in a cold and fast-curing polyurethane resin (RenCast® FC 52/53 Isocyanate/FC 53 Polyol; Huntsman Advanced Materials, The Woodlands, TX, USA), using a custom-designed, 3D-printed fixture to ensure precise positioning. The halves of the mandible were arranged such that the occlusal surfaces of the implants were horizontal, and the movement of the embedding tray led to a pure buccal-lingual movement of the implants. The embedding material was applied only up to the level of the mental foramen on the mandibular base, leaving the implant-bearing alveolar process unrestricted. This approach prevented any additional non-bony support of the implants and allowed for some flexibility in the synthetic bone structures.

Circular markers featuring stochastic speckle patterns were positioned at key locations to serve as reference points for the optical image correlation system (Figure 3). For each specimen, a single marker (M1) was centrally placed on the buccal surface of the dental crown. Three additional markers (M2-M4) were positioned on the synthetic bone, 1 mm beneath the alveolar crest. The bone marker M2 was situated directly beneath the specimen under observation (region 34, 36 or 44, 46). The remaining 2 bone markers (M3 and M4) were placed at the mesially and distally adjacent positions (regions 33, 35, 37 and 43, 45, 47, respectively). Another marker (MR), which was not attached to the bone, was used as a reference for the coordinate system.

Figure 3. Positions of the optical markers M1-M4 and MR, providing a representative example of absolute spatial motion and definition of the global coordinate system.

Before testing, the abutment screws were tightened to 20 N·cm using a torque ratchet (model J5320.1030; Camlog Vertriebs GmbH). After 5 minutes, the screws were retightened to the same torque, in accordance with the manufacturer’s specifications. Following this, the access holes were sealed using a thermoplastic resin. The final step involved polishing the sealed areas.

Test parameters

The implants underwent cyclical loading with a vertical test force of 20 N, regulated by a 6-axis force-torque sensor (model K6D40; ME-Messsysteme, Hennigsdorf, Germany). These tests, which were conducted under standard laboratory conditions, consisted of 10,000 sinusoidal cycles at a testing rate of 60 cycles per minute. Each load cycle involved the movement of the test stylus, beginning from the central position of the occlusal surface, moving to the lingual cusp, then to the buccal cusp, and returning to the central position. The range of the test stylus movement was determined by the driving eccentric mechanism, ensuring that it reached 60% of both the buccal and lingual cusp.

Data analysis

The optical image correlation system captured the motion of the markers (M1-M4) with an accuracy of less than 0.005 mm, at a sampling rate of 25 Hz. The data were collected at predetermined measurement points, each represented as cycle complexes composed of 10 consecutive cycles. Anticipating significant settling effects at the onset of the test, the intervals between measurement points were deliberately set to be shorter during the initial phase. The cycle complexes recorded were as follows: 1–10, 100–109, 200–209, 500–509, 1000–1009, 2000–2009, 3000–3009, 4000–4009, 5000–5009, 6000–6009, 7000–7009, 8000–8009, 9000–9009 and 9991–10000.

The image data sets captured from the 3 cameras were processed using ISTRA4D image correlation software (Dantec Dynamics A/S, Skovlunde, Denmark). This analysis quantified the absolute spatial motion of the markers (M1-M4) relative to the reference marker MR. Within the global coordinate system, the x-axis represented the mesial-distal component, the y-axis denoted the coronal-apical component, and the z-axis corresponded to the buccal-lingual component of spatial motion (Figure 3).

The relative motion between the dental crown (M1) and the directly underlying alveolar bone (M2) was used to characterise the buccal-lingual deflection of the dental crowns. The maximum buccal and lingual relative motions were averaged across all 10 consecutive cycles within each cycle complex. This produced the mean momentary deflection of the dental crown, calculated for all 14 measurement points.

The relative movements between the additional bone markers (M3 and M4), which were affixed to the mesially and distally adjacent tooth positions and M2 were utilised to examine the motions within the alveolar bone that extended beyond the immediate vicinity of the implant site.

Using geometric relationships, the 3D movements of 2 specific measurement points (P1 and P2) on the dental crown marker (M1) were linearly extrapolated to encompass the entire implant. Unique reference points located on the implant axis on the coronal plane (Cor), cervical plane (Cer), and apical plane (Api) were utilised to characterise the overall movement of the implant (Figure 4).

Figure 4. Geometric relationships necessary for extrapolating the 3-dimensional motion and shifting of the measurement points (P1 and P2), as well as the reference points on the coronal plane (Cor), cervical plane (Cer) and apical plane (Api) along the implant axis. The diagram compares the initial position (solid line) to a shifted position (dotted line).

Implant movements resulted in the displacement of the measuring points P1 and P2 to P1* and P2*, along with a shift of the reference points Cor, Cer, and Api to Cor*, Cer*, and Api*, respectively. The shifted measurement points were identified from the recorded 3D motion data. The displaced reference points were located on the implant axis, and their relative distances were defined geometrically. Consequently, the line between Cor* and Api* could be represented by a 3D linear function that was uniquely determined by P1* and P2*.

Statistical analysis

The raw data were processed and descriptive statistics were generated using Excel 2013 (Microsoft Corporation, Redmond, WA, USA). Graphical representation of the results was achieved using OriginPro 2016 (OriginLab Corporation, Northampton, MA, USA). Statistical analyses were performed using SPSS 24.0 (IBM Corp., Armonk, NY, USA). Given the sample size of n=5 for each experimental group, we employed distribution-free, non-parametric significance tests. The one-sided Wilcoxon signed-rank test was used to assess the micromotions of the implants within a single experimental group at the start and end of the test, thereby evaluating the loosening (2 dependent samples). The Kruskal-Wallis test was applied to examine the micromotions of the individual test groups at distinct time points for significant differences (multiple independent samples). Post-hoc Dunn-Bonferroni tests were conducted to perform pairwise comparisons of the individual experimental groups. The significance level was consistently set at P<0.05.

RESULTS

All samples were successfully tested. No failure was observed during the tests in terms of implant or abutment fracture.

Deflections of the dental crown

In all experimental groups, the mean buccal-lingual deflections followed a characteristic and qualitatively similar trajectory throughout the testing period. The deflections rose sharply during the initial 500 cycles, then slowly levelled off towards a plateau (Figure 5).

Figure 5. Mean buccal-lingual deflections of the dental crowns of the experimental groups CamPM, CamM, ConePM and ConeM over the 10,000-cycle testing period.

The mean buccal-lingual deflections were greater for premolars than for molars, irrespective of the implant system used. For both premolars and molars, the CAMLOG® implant system exhibited higher mean buccal-lingual deflections than the CONELOG® system. Within each experimental group, we observed a significant increase in the buccal-lingual deflection from cycle 10 to cycle 10,000 (P<0.05). The mean deflections increased by 0.241±0.179 mm for CamPM, 0.307±0.250 mm for ConePM, 0.224±0.146 mm for CamM and 0.283±0.398 mm for ConeM (Figure 6).

Figure 6. Buccal-lingual deflections in the experimental groups at the beginning (cycle 10) and the conclusion (cycle 10,000) of testing.

^*P<0.05, ^**P<0.01.

For the initial state at cycle 10, the mean deflections of the experimental group ConeM were significantly lower than those of the other experimental groups. However, after 10,000 test cycles, no significant difference was noted in the average deflections among the experimental groups (P=0.392).

Micromotions around the implant site

The mesially and distally adjacent measuring points (bone markers M3 and M4) exhibited cyclic deflections in the buccal-lingual direction. These deflections occurred synchronously and in the same direction as the load applied to the dental crown. The micromotions relative to the bone area directly below the loaded implant were 0.009±0.005 mm mesially and 0.010±0.006 mm distally, with no significant difference between them (P=0.458). In all experimental groups, the micromotions were more pronounced in the buccal direction. These micromotions experienced nonspecific fluctuations throughout the test period, without any preferential tendency for the buccal or lingual direction.

Overall implant movement

The spatial positions of the implant axis and defined reference points were determined for each data set of each cycle complex. By superimposing the projections of the respective shifted implant axes over a complete cycle complex, the momentary overall movement of the implant could be mapped. This is exemplified in Figure 7, which depicts a single test object (a molar) at cycles 10, 100, 1000 and 10,000. The red highlighted trace curves of the reference points (Cor, Cer, and Api) illustrate the extent of the momentary movement of the implant.

Figure 7. Overall spatial movement of the implant axis, using a single test object (a molar) as an example. The movements at cycles 10, 100, 1000, and 10,000 are shown, with the trace curves of the reference points Cor, Cer, and Api marked in red.

The overall spatial movement of all test objects exhibited an hourglass-shaped pattern, with a distinct intersection point representing the momentary pivot point of the dental implant. These pivot points were exclusively found in the region of the implant body and showed no significant variation between the different implant systems or positions considered (P=0.343). In relation to the cervical plane (Cer) at the coronal end of the implant body (0.0 mm) and the apical plane (Api) at the apical end of the implant body (11.0 mm), the pivot points were situated at an average of 5.5±1.1 mm across all test objects. The individual test objects displayed either coronal (n=11) or apical (n=8) shifts of the pivot point along the implant axis throughout the tests, with no discernible pattern or association with the implant systems or implant positions.

DISCUSSION

Numerous studies have underscored the importance of sufficiently high primary stability for the successful osseointegration of endosseous dental implants [2,3]. Consequently, major efforts are underway to enhance this primary stability. These efforts involve modifications to both the design of the implant and the surgical techniques used. The ultimate goal is to further decrease the period of edentulism for patients, without compromising the osseointegration process and, therefore, the long-term therapeutic success [21].

Previous biomechanical studies have primarily evaluated dental implants in vitro, applying uniaxial vertical [16,23], lateral [18] or worst-case oblique loading in accordance with ISO 14801 [7,24]. This approach stresses the implants to their maximum capacity in terms of material testing. However, to estimate the actual in vivo performance of various dental implant designs or the effects of surgical techniques, it is necessary to consider the interactions of the implants with the surrounding bone structures, as well as temporal and dynamic effects [11,19]. Dixon et al. [25] and Steinebrunner et al. [17] employed biaxial test setups, in which an inclined surface was cyclically loaded and unloaded along a linear path. This method simulates the dynamic eccentric loading of a single cusp, with a focus on the loosening of abutment screws, survival rates, and fracture loads in homogeneous resin test blocks.

The novel approach of the present pilot study was designed to enhance the complexity of dynamic in vitro modelling of masticatory loads on the posterior teeth, while concurrently recording the resulting micromotions in relation to the surrounding bone. The reciprocal buccal-lingual loading aligns with the typical loosening patterns for implants in the posterior region observed in practice [26] and represents a reasonable simplification, assuming mesial and distal support from proximal contacts with the adjacent teeth. The micromotions over time serve as an indicator of the present implant stability. They can also reveal direct correlations between loading and implant deflections, thereby providing a comprehensive characterisation of the location and progression of implant loosening, as well as settling effects.

The test setup was fundamentally influenced by the rocking-horse test, a method known from the field of shoulder arthroplasty. This test is utilised to cyclically apply off-centre loads to an artificial glenoid component, resulting in the gradual rocking of the implant out of the bone [27]. Analogous cyclic eccentric loads on the dental crown, along with constantly changing points and directions of force application, have been documented for the posterior teeth during mastication [14,28]. These conditions are deemed particularly critical as they can precipitate implant failure, even when the loads are well below the maximum tolerable levels [19,26].

In the present study, the test force was set to 20 N. This corresponds to a mean load on a single posterior tooth during physiological mastication, according to studies involving in vivo measurements made using implanted 3D force sensors [14,29]. Our research was concentrated on physiological functional loads, which must be differentiated from maximum bite forces or pathological conditions such as bruxism, which can involve bite forces of up to 900 N [30].

The eccentricity of buccal-lingual loading was adjusted based on the diameter of the implant being tested (either premolar or molar). This adjustment allowed the test stylus to reach 60% of the buccal and lingual cusps, a loading path that aligns with those used in prior biomechanical studies [16,17,25]. The implants underwent cyclical loading for 10,000 sinusoidal cycles at a rate of 60 cycles per minute, a frequency that falls within the reported range of human masticatory frequencies [18,31]. The choice of 10,000 test cycles was intended to simulate the critical 2-week period for osseointegration following insertion [32]. This number of cycles was also sufficient to demonstrate loosening effects, as deflections tended to reach a saturation point after significantly changing during the initial ~1,000 cycles. In the existing literature, a higher number of test cycles has only been implemented in endurance tests of individual implant components [16,33,34].

The optical 3D measuring system employed in this study facilitated a direct, non-contact assessment of the micromotions induced by dynamic loading. The concurrent recording supplied insights into the instantaneous mechanical anchorage under load, thereby allowing for the detection of both gradual and sudden loosening effects. This method presents a substantial advantage over approaches in which the evaluation of the initial state (for instance, insertion torque), the subsequent loading scenario, and the evaluation of the final state (such as removal torque) are carried out separately in terms of temporality and causality [7,23].

To facilitate the concurrent recording of micromotions under load, the object being measured must consistently remain within the optical measuring system’s field of view. In the context of dental implants, only the superstructure is always visible externally and is used to record the micromotions through defined markers. Micromotions at the bone-implant interface can be determined only indirectly, through extrapolation. In biomechanics, sophisticated methods exist that allow the measurement of deflections and displacements directly at the bone-implant interface via special optical markers, such as the pin-sleeve technique [35]. However, these methods are invasive and unsuitable for dental implants due to their small dimensions. With such techniques, the bone-implant interface would be considerably disrupted, and the measurement process could affect the actual anchorage.

In this study, dental implants were inserted into a biomechanical synthetic bone model of the mandible, which was fixed only inferior to the mental foramen. In previous studies exploring the interaction with the osseous environment, dental implants were inserted into non-anatomical synthetic bone blocks [13,36].

In the anatomically shaped synthetic bone, micromotions were observed under dynamic cyclic loading, both directly at the implant loading site and in the mesially and distally adjacent areas of the alveolar bone. The micromotions of the alveolar bone occurred synchronously and in the same direction as the dental crown, as a consequence of interior motion of the implant.

The relative movements between 2 adjacent positions of the alveolar bone were approximately 0.01 mm, which fell within an order of magnitude of the total implant movement determined in other studies involving rigid fixation [16,25]. This finding underscores the critical importance of the yielding and mobility of the mandibular bone for the physiological distribution and dissipation of the applied masticatory forces [31]. These factors should be considered when assessing the in vivo performance of dental implants through in vitro experiments.

The deflections of the dental crowns increased significantly over time, especially during the first 500 cycles. This result emphasises the importance of sufficiently strong mechanical anchorage during the early postoperative phase [3]. Subsequently, the deflections began to stabilise at a certain saturation level as the external loads and the resistance at the bone-implant interface gradually achieved equilibrium [19].

In this study, the mean buccal-lingual deflections ranged from a minimum of 0.066 mm at the beginning of the tests to a maximum of 0.613 mm after loading. Initially, without the long-term influence of dynamic cyclic loading, the mean buccal-lingual deflections were of a similar magnitude to the deflections reported by Freitas et al. In their study, horizontal deflections resulting from comparable, but quasi-static, lateral loading ranged from 0.025 mm to 0.050 mm [12].

The increasing buccal-lingual deflections during testing represented a gradual loosening due to the increasingly unfavourable reciprocal and dynamic loading of the dental implants [19].

When comparing the experimental groups, the deflections tended to be higher in the premolars than in the molars. This observation aligns with the findings of previous studies, which have shown a direct correlation between a larger implant diameter and increased implant stability [37].

The implant systems under investigation did not exhibit significant differences in the tests conducted on synthetic bone. Thus, the influence of the implant-abutment connection appears minor relative to the interactions between the implant body and the synthetic bone.

The ability to determine the spatial position of the implant axis facilitates the in vitro visualisation of implant movement within the bone. This has been previously characterised solely in numerical terms; thus, the present study complements existing in silico methodologies that utilise finite element analysis [13].

The specific patterns of 3D movement revealed an hourglass-shaped pattern with a characteristic pivot point. This hourglass shape has been identified as a common loosening pattern in the region of the posterior teeth [26,38], which can be attributed to the physiological reciprocal loads on the dental crown in vitro and in vivo.

Understanding the position and potential shifts of the pivot point under load is potentially useful in increasing the anchorage strength of dental implants. In the synthetic bone model of the mandible used in the present study, the pivot point was, on average, located in the centre of the implant body. The existing literature indicates that the pivot point for a natural tooth in human bone is in the apical third of the tooth root, while for dental implants, it is in the region of the implant shoulder [39]. Using the presented methods, future research on human donor bone could identify specific opportunities for improving the anchorage of dental implants, based on their actual 3D movement under cyclic dynamic loading.

In the test setup described, the physiological masticatory loads acting on the lower posterior teeth were represented by the coronal-apical and buccal-lingual force components for both premolars and molars. The mesial-distal force component was disregarded, based on the assumption that a single tooth replacement would be designed to have the implant supported by the teeth adjacent to it on the mesial and distal sides, thereby ideally restricting that degree of freedom.

In vitro biomechanical testing typically employs non-vital human or synthetic bone, which cannot reproduce the biological activity of living human bone. However, this study was focused on primary stability and, consequently, on purely mechanical anchorage as a prerequisite for subsequent osseointegration.

The synthetic mandibular bone models used in this study were anatomically accurate, were standardised in size and shape, and structurally mimicked the cortical and cancellous sections. Synthetic materials typically only approximate the actual mechanical properties of human bone, but they offer defined and consistent conditions, devoid of additional interference from anatomical variations in size, shape, bone density, or residual dentition. Consequently, synthetic bone models are particularly suitable for the development and testing of experimental setups and methods that can later be applied to human bone.

The decision to embed to the level of the mental foramen represented a compromise between the need for adequate fixation and the desire to allow a certain degree of flexibility. A further decrease in rigid fixation to enhance the physiological mobility of the mandible could potentially be beneficial.

A major limitation of this study was the limited number of samples. Nevertheless, this research served as a pilot study, pioneering a novel method to quantify the primary stability of dental implants. It also laid the groundwork for more extensive tests on human bone.

Furthermore, future studies should consider the insertion torque during implant placement. This additional parameter could be cross-referenced to assess primary stability in conjunction with implant movement.

The main conclusions of this study can be summarised as follows:

• 1) The proposed experimental setup models the dynamic loads acting on the posterior teeth during mastication. Through incorporation of an optical 3D motion measurement system, it enables the direct measurement of micromotions of the dental crown resulting from cyclic eccentric loading.

• 2) In long-term experiments conducted on synthetic bone, cyclic reciprocal loading, using an average masticatory force, results in significant loosening of dental implants. This could be more pertinent for assessing the primary stability and thus the osseointegration of single-tooth implants, compared to unidirectional worst-case loading scenarios.

• 3) The cyclic reciprocal loading of dental implants produces movements of the alveolar bone, which are also evident in the mesially and distally adjacent areas of the implant site. This observation highlights the area affected by masticatory loads and illustrates how these forces are distributed and dissipated through the mandibular bone. Consequently, when estimating the in vivo performance of dental implants, it appears crucial to consider the yielding and anatomy of the surrounding bone.

Our work complements known methods used for in vitro biomechanical investigation of dental implants, conducted to predict their in vivo performance. In general, a comprehensive biomechanical characterisation of various implant systems, surgical techniques, bone qualities, and diverse loading scenarios could potentially enhance dental implant procedures, thereby improving future patient care. Specifically, the 3D visualisation of overall implant movement, as presented in this study, offers novel opportunities to assess the complex and dynamic interactions at the bone-implant interface. It also allows the evaluation of the effects of specific modifications in implant designs or surgical techniques in an in vitro setting.