The effect of implant-abutment connection fit and proximal contact tightness on screw loosening

Abstract

PURPOSE

This study aims to evaluate how the fit between implant-abutment connection surfaces and the proximal contact tightness of crowns affect abutment screw loosening.

MATERIALS AND METHODS

Forty implants with internal hex connections were placed in stainless steel models in the maxillary right first molar region. Monolithic zirconia crowns with standard and 15 µm reduced proximal contacts were used with abutments with standard and 25 µm horizontally reduced hex connections. After cementation, four groups of specimens—standard abutment/standard proximal contact (control), reduced fit abutment/standard proximal contact (RS), standard abutment/reduced proximal contact (SR), and reduced fit abutment/reduced proximal contact (RR)—were subjected to simultaneous thermocycling (± 5 – 55℃) and chewing cycles (100 N, 1.6 Hz, 40 mm/sec lateral and vertical speed, lateral movement: 6 mm, vertical movement: 0.5 mm). Removal torque values (RTVs) were measured using a digital torque meter. Data were analyzed using two-way ANOVA and the Bonferroni test (α = .05).

RESULTS

All groups exhibited statistically significant differences compared to each other and the control group (P < .05). The RR group had the lowest removal torque values (RTVs) (21.70 ± 0.99 N·cm), while the control group had the highest (26.82 ± 0.70 N·cm). The RS group (24.78 ± 0.89 N·cm) had lower RTVs than the SR group (25.51 ± 0.52 N·cm).

CONCLUSION

The findings indicate that inadequate proximal contacts and horizontal misfit of the implant-abutment connection may lead to torque loss and screw loosening.

INTRODUCTION

Implant-supported restorations are the leading treatment for restoring aesthetics and function after tooth loss, but they may face mechanical complications, particularly abutment screw loosening.¹ This may result from insufficient preload, screw deformation, abrasion of screw roughness, overloading, and micro-mobility at the junction.²,³ Contributing factors include the implant system, casting defects, repetitive screwing cycles, settling effect, improper implant positioning, inadequate occlusal design, variations in hexagonal connections, stresses in the abutment and implant collar, increased abutment angulation from poorly fitting restorations, elevated collar height, inappropriate screw design, heavy occlusal loads, and incorrect insertion torque.³,⁴,⁵

Screw loosening is influenced by settling effect or embedment relaxation from the unavoidable micro-roughness of implant and screw surfaces, which reduces preload. Optimal torque is spent on smoothing rough surfaces instead of creating elongation stress for compression preload.⁶ Proper application in the implant-abutment connection is crucial for long-term success. To prevent screw loosening, it is recommended to use torque values above 30 N·cm and retorque after 10 minutes the initial torque.⁷

Previous studies highlight the need to minimize implant-abutment misfit to prevent mechanical complications.⁸,⁹ Irregularities in mating surfaces must be identified, as tightening the screw can flatten metallic contacts and reduce the distance between them.¹⁰ Misfits can occur both horizontally and vertically in internal-type implant-abutment connections. Barbosa et al. found that vertical discrepancies do not affect torque loss, while Kano et al. reported that horizontal misfits are more prevalent, indicating a need for further research on their potential effects on torque loss.¹¹,¹²

A harmonious occlusal relationship is crucial for the long-term success of implant-supported restorations, as it affects load distribution.¹³ Effective proximal contacts are crucial for load stability; non-ideal contacts can increase stresses around implants, which lack a periodontal ligament.¹⁴,¹⁵ Off-axis loading emphasizes the importance of interproximal contact points in stress distribution. Single implants without distal contacts in the posterior region are more susceptible to non-axial overload.¹⁵ This condition may alter stress distribution at the implant–abutment interface and potentially contribute to mechanical complications, including abutment screw loosening. However, evidence on the effect of proximal contacts on screw loosening in single implants is limited.

This study examines how fabrication-related misfit in implant-abutment connections and reduced proximal contact tightness affect screw loosening in cement-retained implant-supported single crown restorations. The null hypothesis (H₀) asserts that these factors do not affect screw loosening.

MATERIALS AND METHODS

To evaluate the effects of two distinct abutments, two distinct crowns, and the interaction between these abutments and crowns on removal torque value (RTV), a two-way ANOVA was planned. The minimum specimen size required for the analyses was determined through a power analysis conducted using the G*Power software program (G*Power 3.1.9.2; Heinrich Heine University, Düsseldorf, Germany). The Type I error rate (α) was established at 0.05, while the statistical power (1 - β) was set at 0.95. The effect size was defined as 0.848, consistent with findings from a similar study by Rutkunas et al.¹⁶ Consequently, the G*Power analysis indicated that a sample size of 21 would be required to detect an effect size of 0.848, suggesting that the minimum specimen size for each group should be 5. In the present study, 10 specimens were utilized for each group to ensure sufficient statistical power within the context of the research (n = 10).

Information regarding the groups is presented in Table 1. In the development of the master model, measurements were obtained from a typodont maxilla with a missing right first molar, which served as a reference (Frasaco GmbH, Tettnang, Germany). Three-dimensional design software was employed to create the design of the edentulous space and to prepare a partial crown for the second premolar and the second molar teeth, while ensuring that the contact surfaces facing the edentulous space (SolidWorks 25.0, Dassault Systèmes, Paris, France). In the study models, stainless steel, a rigid material, was selected due to its durability, as it is unlikely to wear in the proximal contact areas during dynamic loading, thereby minimizing the movement of the implants.

Table 1. Information on study groups.

Group	N	Specifications
Control	10	Standard abutments and crowns with standard proximal contacts
SR	10	Standard abutments and crowns with 15 µm reduced proximal contacts
RS	10	Reduced abutments with 25 µm reduced horizontal fit and crowns with standard proximal contacts
RR	10	Reduced abutments with 25 µm reduced horizontal fit and crowns with 15 µm reduced proximal contacts

Stainless steel study models were fabricated using a Computer Numerical Control (CNC) device. The cylindrical spaces, designed for the placement of implants and aligned with the external specifications provided by the manufacturer, were created using the wire erosion technique (Fig. 1). The fabrication precision of the models was evaluated by measuring nine distinct components in each model with a three-dimensional coordinate measuring device (Fig. 2). The observed dimensional deviations ranged from 0.2% to 0.6% across all regions, a range that falls within the sensitivity limits of the measuring device and is therefore considered acceptable.¹⁷

Fig. 1. Stainless-steel model.

Fig. 2. Dimensions of stainless-steel model.

Non-surface-treated internal hex implants (Oxy, Biomec SRL, Colico, Italy) with a diameter of 4.5 mm and a length of 10 mm were inserted into the prepared sites using dimethacrylate ester adhesive (Loctite 270, Henkel, Dusseldorf, Germany), which exhibits resistance to RTV ranging from 5 to 33 N·m. This adhesive was utilized to ensure a secure bond between the implants and the model, as well as to minimize any gaps between the model and the implants. The settlement depths were verified through measurements taken with a digital micrometer. Following a 24-hour curing period for the adhesive at room temperature, the abutments (ASAIN425, Oxy, Biomec SRL, Colico, Italy) were subsequently placed. In this study, the reduced abutments were fabricated by the implant supplier according to the specific dimensions requested for the study. These reduced abutments had the same properties as the standard abutments, except for the hexagonal diagonal diameter, which was 25 µm narrower (Oxy, Biomec SRL, Colico, Italy). The dimensions of the implant-abutment connection interfaces of the abutments used in this study are presented in Figure 3. After the abutments were placed, the stainless-steel models were scanned using a laboratory scanner (Imetric D104a, Imetric4D, Courgenay, Switzerland), and the design was created in the form of partial crowns for the second premolar and the second molar utilizing 3D design software (exocadDentalCAD, exocad GmbH, Darmstadt, Germany) (Fig. 4).

Fig. 3. The position, orientation, speed, and magnitude of the load within the chewing simulator.

Fig. 4. Three-dimensional designs of maxillary right second premolar and second molar.

This study involved the evaluation of two distinct single crown designs, characterized by standard proximal contacts and those reduced by 15 µm in the mesial and distal regions. The marginal gap was set to 0 µm, while the internal gap was set to 30 µm. To maintain design standardization, the same opposing occlusion and occlusal template were utilized (Fig. 5). Partial crowns were fabricated using modeling resin (Regodental Modeling Resin, Regodental, Istanbul, Türkiye) via a 3D printer (Anycubic Photon, Anycubic, Shenzhen, China) and were subsequently cemented with cyanoacrylate adhesive (Fig. 6). Single crowns were fabricated by milling (Roland DWX 50, Roland DGA Corp, Irvine, CA, USA) from 3Y-TZP monolithic zirconia discs (BruxZir Shaded, Glidewell Laboratories, Newport Beach, CA, USA) (LOT BZ0004039, BZ0001303). The tightness of the proximal contact was checked during the crown restoration placement phase and after cementation. The density of interproximal contacts in reduced crowns was evaluated using a 15 µm thick metal film coated on both sides with occlusal spray (Interproximal Space Ruler, Protector Dental, Guangzhou, China). The assessment involved determining whether the film could pass through the proximal contact area with minimal resistance (Bausch Arti-Spray, Dr. Jean Bausch GmbH & Co). Additionally, a 16 µm thick articulation paper (Bausch Gnatho Film, Dr. Jean Bausch GmbH & Co, Cologne, Germany), verified using a digital micrometer, was employed to ensure it could pass through with slight compression while simultaneously marking the contact points. Due to the highly operator-dependent nature of proximal contact assessment, all evaluations were performed by the same operator to reduce variability. Inter-rater reliability analysis was therefore not performed.

Fig. 5. Three-dimensional design with proximal contact of maxillary right first molar.

Fig. 6. Restorations fabricated using modelling resin and monolithic zirconia.

The distribution of abutments and crowns among the study models was determined through simple randomization, employing a random number generator. A single uncoated titanium alloy (Ti-6Al-4V) screw was used for each specimen, and the initial RTV was not measured. In accordance with established protocols, both insertion and removal torque were applied using a calibrated digital torque meter to ensure the reproducibility and accuracy of the results.¹⁸ An insertion torque was applied to the abutment screw using a digital torque meter which certified by the manufacturer according to Industrial Technology Research Institute (ITRI) Center for Measurement Standards to a trueness of ± 0.5% at 30 N·cm (SLQ-05Q1, Solude Digital Screwdriver, Solude, Taiwan). After a duration of 10 minutes, the same torque value was reapplied to maintain the preload. The head of the screwdriver was specifically milled to accommodate the digital torque meter for the application of torque. To minimize variations in torque application, a single experienced operator applied the insertion and re-torque procedures throughout the entire study. The digital torque meter was used according to the manufacturer’s standard protocol, and the torque was applied in a controlled and steady manner while avoiding excessive axial (vertical) force. Although no additional guiding apparatus was employed, the operator was instructed to hold the torque device perpendicular to the long axis of the abutment and to avoid pressing downward. This approach aimed to ensure consistency and repeatability across all specimens.

Monolithic zirconia crowns were cemented on the abutments using a urethane dimethacrylate-based cement (BJM-Cem Implant, BJM Lab, Tel Aviv, Israel) (LOT 4316CTTR), which was employed as a long-term temporary cement to allow the measurement of RTVs following dynamic loading. The cementation procedure was conducted while a constant force of 50 N was applied to the central fossa. Subsequently, a dynamic load of 100 N was applied vertically on the central fossa using a steel sphere with a diameter of 5 mm, which moved at a speed of 40 mm/s along both the vertical and horizontal axes, with respective sliding distances of 6 mm and 0.5 mm in the chewing simulator (Esetron Mechatronics, Ankara, Türkiye) (Fig. 7). During the application of 720,000 cycles at a frequency of 1.6 Hz, the specimens were subjected to thermal aging in distilled water at temperatures ranging from 5 to 55℃.

Fig. 7. The dimensions of the implant-abutment connection interfaces for standard and reduced abutments.

Following dynamic loading, the RTV was measured using a digital torque meter, and the resulting data were analyzed using statistical software (SPSS 15.0, SPSS Inc., Chicago, IL, USA). The normality of the data distribution was evaluated using the Shapiro-Wilk test. Levene’s test was employed to evaluate the homogeneity of variances among the groups. The data were analyzed concerning proximal contact, abutment fit, and the interactions between proximal contact and abutment fit using a two-way analysis of variance (ANOVA). The Bonferroni test was applied for multiple comparisons (α = 0.05).

RESULTS

The data were normally distributed (P > .05), and variances were homogeneous across all groups (P > .05). A two-way ANOVA was performed to assess the effects of abutment fit, proximal contact, and their interaction on RTV. Results indicated that both abutment fit and proximal contact significantly influenced RTV (F(3, 36) = 74.507, P < .001, η² = .861). Abutment fit had a significant main effect on RTV (F(1, 36) = 135.083, P < .001, partial η² = .790), as did proximal contact (F(1, 36) = 76.071, P < .001, partial η² = .679). A significant interaction effect between abutment fit and proximal contact was also observed (F(1, 36) = 12.366, P = .001, partial η² = .256). Analysis of estimated marginal means with Bonferroni correction revealed that, among crowns with reduced proximal contact, reduced abutments with reduced fit had significantly lower RTV compared to standard abutments (M_difference = -3.81, SE = 0.36, P < .001). For crowns with standard proximal contact, reduced abutments also showed significantly lower RTV than standard abutments (M_difference = -2.04, SE = 0.36, P < .001). Among reduced abutments, crowns with reduced proximal contact had significantly lower RTV than those with standard proximal contact (M_difference = -3.08, SE = 0.36, P < .001). Additionally, within the standard abutment group, crowns with reduced proximal contact exhibited significantly lower RTV than those with standard proximal contact (M_difference = -1.31, SE = 0.36, P < .001). Mean and standard deviation values of RTVs are presented in Table 2, indicating significant differences among all groups (P < .05). The control group had the highest values (26.82 ± 0.70 N·cm), while the RR group had the lowest (21.70 ± 0.99 N·cm) (M_difference = -5.12, SE = 0.36, P < .001)(Fig. 8).

Table 2. Mean removal torque values and standard deviations (N·cm).

Abutment fit	Proximal contact	Mean	Std. Deviation	N
Reduced	Reduced	21.70	0.99	10
	Standard	24.78	0.89	10
	Total	23.24	1.83	20
Standard	Reduced	25.51	0.52	10
	Standard	26.82	0.70	10
	Total	26.17	0.90	20

Fig. 8. Bar charts showing the removal torque values (N·cm) in all groups.

DISCUSSION

The null hypothesis was rejected, indicating that horizontal misfit from abutment fabrication and proximal contact tightness affected screw loosening in cement-retained implant-supported single crowns. The control group with standard abutments and normal proximal contact crowns had the highest RTVs, while the RR group with reduced abutments and reduced proximal contact crowns had the lowest. Significant RTV differences between the SR and RS groups suggest that implant-abutment misfit may influence screw loosening more than proximal contact tightness. No screw loosening occurred in any group, despite decreased RTVs after dynamic loading, indicating maintained preload and connection stability. However, evidence on the impact of torque loss on connection stability in implant-supported restorations is limited.

The proximal contact area is crucial for stress transmission and minimizing implant mobility; however, research on its effect on implant-supported restorations is limited. Understanding this relationship could improve treatment planning. Specifically, when restoring a lost first molar with an implant-supported crown, the presence of a second molar has been shown to reduce stresses in the supporting simulated bone structure under various loading conditions.¹⁵ An effective proximal contact influences the stress pattern and load distribution of the implant.¹⁴ In this study, a higher loss of RTV was observed in implant-supported crowns with reduced proximal contact tightness compared to standard crowns, indicating that loss of proximal contact increases preload loss by elevating tension in the implant-abutment connection. Recent studies report a loss of proximal contact ranging from 53% to 65% between implant-supported fixed restorations and adjacent teeth.¹⁹,²⁰ Evaluating the stability of the abutment screw during clinical follow-ups is essential.

Implants do not adapt to musculoskeletal development or dentoalveolar changes, acting like ankylosed teeth.²¹ This can lead to loss of proximal contact with adjacent teeth due to tooth movement.²² While the mesial drift theory explains some contact loss, it can occur both mesially and distally.²⁰ Studies show mean proximal spaces of 0.33 mm and 0.167 mm between implant-supported restorations and adjacent teeth.¹⁹,²³ As people age, dental arch length decreases—by 1.0 mm in the maxilla and 0.8 mm in the mandible over 20 years, resulting in annual mesial migrations of about 50 µm and 40 µm, respectively.²⁴,²⁵ Another study found reductions of 0.67 mm in the maxilla and 0.71 mm in the mandible over 28 years, corresponding to annual migrations of 24 µm and 25 µm.²⁵ These changes affect the tightness of proximal contacts in implant-supported restorations. This study evaluated the impact of a 30 µm reduction in proximal contacts, with crowns having mesial and distal contacts each reduced by 15 µm.

Evaluating the effect of proximal contact insufficiency on the implant-abutment connection in clinical conditions is challenging due to factors like reduced dental arch length and migration. Nonetheless, the findings of this in vitro study are important despite its limitations. The evaluation of proximal contact and addressing contact loss through prosthetic or restorative techniques is essential for maintaining implant-abutment stability.

The passive fit of the implant-abutment complex can affect screw stability. Vertical misfit at the implant-abutment interface was measured in studies, revealing values of 3.2 to 4.33 µm for internal hex abutments and implants, with potential tolerances from the fabrication process.¹¹,²⁶,²⁷ Some research also assessed the horizontal fit by measuring the distance between the implant’s outer surface and the abutment’s contour.¹¹,²⁶ However, no standardized protocol for evaluating compatibility at this interface exists.

In a study, horizontal gaps of 22 to 100 µm have been reported as machining tolerances among implant components.²⁸ The effect of horizontal incompatibility within the implant-abutment complex on screw loosening is unclear. This study evaluated the impact of machining tolerance—considered a non-fabrication defect—on screw loosening by comparing torque loss in abutments that were 25 µm narrower than standard abutments. Minimizing machining tolerances is essential for a complete fit between abutment and implant surfaces, reducing the risk of mechanical and biological complications.⁹ However, manufacturers often do not disclose dimensional tolerance information. While evidence supporting the need for precision to ensure successful osseointegration is limited, inadequate fit at the implant-abutment interface has been linked to screw loosening and fracture.²⁹ One study reported a deviation of–18 µm in external hexagonal implants and 9 µm in abutments, while another found a 58 µm deviation between internal hexagonal implants and prefabricated gold abutments.⁹,³⁰ A horizontal gap of 40 µm was also identified between implants and cement-retained abutments over 16 years.³¹ Similarly, this study found that a dimensional difference of 25 µm in internal hex diagonal lengths influenced abutment screw loosening. Even when using original components, it is essential to consider manufacturing tolerances of implants and prosthetic parts; therefore, the use of the manufacturer’s original products is recommended in all instances.

This study was conducted using an in vitro mechanical model, and the findings should be interpreted within the inherent limitations of laboratory-based experimental conditions. The limitations of this study include the use of monolithic zirconia crowns. Future research on proximal contact tightness should consider metal-alloyed restorations, which do not require sintering, to minimize deviations from the three-dimensional design and reduce the need for additional laboratory procedures. Stainless steel models were used to minimize implant mobility; however, this approach may not adequately replicate conditions such as bone elasticity and osseointegration. Additionally, this study focused on cement-retained abutments with a single screw connection and an internal hex implant-abutment interface. Therefore, the findings may primarily apply to internal hex connection systems, while clinical behavior in conical connection designs may differ. The existing literature does not provide torque values that compromise screw connection stability, making it difficult to quantify the clinical tolerance limits of factors influencing preload loss. Consequently, additional studies are essential for a more precise evaluation of abutment screw loosening. While the compatibility of the implant-abutment interface is often highlighted as a significant factor affecting connection stability, there is a notable lack of studies and consensus on the definition and measurement of misfit. Screw loosening appears to be an inevitable issue, as fabrication deviations are evident even when using original prosthetic products from the manufacturer. Therefore, methodologies to define and measure implant-abutment interface misfits and establish fabrication tolerance limits are crucial.

CONCLUSION

The RTV is affected by the proximal contact tightness. Specimens with a 15 µm reduction in contact tightness are more prone to torque loss. Maintaining tight proximal contacts is essential, and during annual follow-ups, these contacts should be evaluated and if deemed necessary, the restorations should be replaced. The RTV is affected by the misfit at the implant-abutment interface. A 25 µm horizontal fit reduction increases the risk of screw loosening. Since this amount of discrepancy, which may also arise from fabrication processes, could be unavoidable, it is essential to assess the stability of the screw connection during clinical follow-ups. The most significant torque loss was observed in specimens with reduced proximal contact tightness and connection misfit.