Fatigue Reliability of 3 Single-Unit Implant-Abutment Designs

Abstract

Objectives: 

Because the mechanical behavior of the implant-abutment system is critical for the longevity of implant-supported reconstructions, this study evaluated the fatigue reliability of different implant-abutment systems used as single-unit crowns and their failure modes.

Methods and Materials: 

Sixty-three Ti-6Al-4V implants were divided in 3 groups: Replace Select (RS); IC-IMP Osseotite; and Unitite were restored with their respective abutments. Anatomically correct central incisor metal crowns were cemented and subjected to separate single load to failure tests and step-stress accelerated life testing (n = 18). A master Weibull curve and reliability for a mission of 50,000 cycles at 200 N were calculated. Polarized-light and scanning electron microscopes were used for failure analyses.

Results: 

The load at failure mean values during step-stress accelerated life testing were 348.14 N for RS, 324.07 N for Osseotite, and 321.29 N for the Unitite systems. No differences in reliability levels were detected between systems, and only the RS system mechanical failures were shown to be accelerated by damage accumulation. Failure modes differed between systems.

Conclusions: 

The 3 evaluated systems did not present significantly different reliability; however, failure modes were different.

Implant therapy has become a predictable treatment modality in oral rehabilitation with high success rates for restoring partial or complete edentulous arches and for single-unit replacements.^1,2 Despite these reported rates of the osseointegrated implant fixture,³ one commonly reported mechanical problem is screw joint instability^4–6; that is, loosening or fracture of the abutment or screws retaining single-crowns,^7,8 with a cumulative incidence of 12.7% after 5 years of clinical service in single-unit crowns.⁹

Several implant-abutment configurations have been designed to support prosthetic restorations. This joint may be classified as external or internal and incorporate features for rotational resistance, indexing, and lateral stabilization, described as hexagonal, octagonal, cone screw, cone hex, cylinder hex, spline, cam, cam tube, and pin/slot.¹⁰ The external hexagonal interface was originally intended to provide a method to engage the implant during surgical placement. In single-unit restorations, the external hexagon has been used to supply an anti-rotational mechanism, resulting in the exposure of the implant-abutment interface and abutment screw to greater external loads and bending moments, which can lead to screw joint opening and screw loosening.^11,12 The tapered integrated screwed-in abutments were introduced with an 8 degrees mating angle between implant and abutment that required a loosening torque 124% greater than the tightening torque of 25 N/cm.¹³ Considering the variety of connection designs of the implant-abutment, investigations concerning their probability of survival after fatigue are warranted.

The stability of the connection between different implant parts has been suggested as a prerequisite for the overall success of the reconstruction, especially in single-unit replacements that, unlike joined units, lack mutual or cross-arch stabilization.¹⁴ Screw joint instability or micromotion may be related to inadequate preload, inappropriate screw design, poor component fit,^15,16 settling of surface microroughness, excessive loading, and/or elasticity of bone. In addition, fatigue simulating occlusal function at loads below the yield strength of the connection may lead to steady loosening and/or abutment screw failure.¹⁷

Most commonly used in implant-abutment fatigue testing are unidirectional bend-release loads (carried out on asymmetric samples)^18–22 or rotational fatigue in symmetrical specimens.²³ The use of step-stress accelerated life testing (SSALT) has been recently reported for reliability evaluation of ceramic implants²⁴ and ceramic abutments.²⁵ However, comparisons of different implant-abutment connecting designs using SSALT have not been explored to date.

The purpose of this study was to evaluate the reliability and failure modes of 3 different implant-abutment designs subjected to SSALT. The tested null hypothesis was that there is no difference in reliability and failure modes between the implant systems.

Materials and Methods

Sample Preparation

Sixty-three Ti-6Al-4V implants 3.75 mm in diameter and 13 mm in length were divided into 3 groups (n = 21 each) according to implant system: (1) Replace Select (RS group, Lot no. 423497) (Nobel Biocare, Göteborg, Sweden,); IC IMP Osseotite (O group, Lot no. 2008030577) (Biomet 3i, Palm Beach Gardens, FL); and Unitite (UN group, Lot no. PRH066) (SIN – Sistema de Implantes, Sao Paulo, SP, Brazil). Each implant was assigned to receive its proprietary abutment: Snappy RP Abutment (Nobel Biocare; Lot no. 419350); Certain Conical Abutment (Biomet 3i; Lot no. 837178); and Abutment Cone Morse (SIN – Sistema de Implantes; Lot no. H10696).

The implants and abutments were all evaluated and passed initial inspection under a polarized-light microscope (Leica MZ APO Stereomicroscope; Leica Systems, Wetzlar, Germany) for possible major superficial flaws. The implants were then embedded in polymethacrylate acrylic resin (Orthodontic Resin, Dentsply, Milford, DE) at a 30-degree angulation with respect to its long axis (ISO 14801:2007 standard). The top platform of each implant was positioned slightly above the potting surface.²⁴ After connection of the abutment to the bearing housing, the abutment screw was tightened with a torque gauge (Torque Wrench Ref. TMEC, SIN, Sao Paulo, SP, Brazil; Lot no. H60804) according to the manufacturers' instructions.

Central incisor crowns were waxed and cast in a cobalt-chrome partial denture alloy (Wirobond 280, BEGO, Bremen, Germany). The standardization of the thickness of the crowns was accomplished with a silicone index made from an impression of the waxed desired anatomy, which was used to guide waxing contour and anatomy. Then the crowns were cemented (RelyX Unicem, 3M ESPE, St. Paul, MN) on their respective abutments.

Single Load to Fracture

Three specimens of each group were assigned for single load to fracture (SLF) testing to determine the step-stress profile. A load was applied to the specimens at the incisal aspect using a universal testing machine (Model 5566, Instron, Canton, MA) equipped with a 10 kN load cell at 1 mm/min rate. The mean load to failure was calculated for the 3 different groups and used to determine 3 step-stress profiles (profile 1, n = 9 samples; profile 2, n = 6 samples; and profile 3, n = 3 samples) for fatigue accelerated life testing (Fig. 1).^24,26,27

Fig. 1: Step-stress profiles used for fatigue testing. Note that the loads started in the neighborhood of 100–200 N ending up to approximately 600–650 N. The specimens of each group were distributed across 3 profiles following the ratio 3:2:1 (n = 9 in profile 1, n = 6 in profile 2, and n = 3 in profile 3).

SSALT

The remaining 18 specimens of each group were subjected to SSALT according to the preestablished profiles and considering the distribution of 3:2:1 from mild to aggressive mode of test.^24,27 Each specimen was then fatigued (contact, load, and lift-off cycle) under water at 2 Hz in a servo-all-electric system (800L, TestResources, Shakopee, MN). A flat tungsten carbide indenter was attached to the simulator for the fatigue testing and the load was applied at the incisal edge of the incisor metal crown.

After fatigue testing, lever arm calculations were performed for each load at failure for the groups. Use level probability Weibull curves using a cumulative damage and power law relationship were calculated²⁸ (Alta Pro 7, Reliasoft, Tucson, AZ). Reliability (90% 2-sided confidence bounds) for completion of a mission of 50,000 cycles at 200 N load was determined for group comparisons. If the Weibull use level probability calculated beta was <1 for any group, meaning that failure was not influenced by fatigue,²⁹ then a Contour plot (Weibull modulus [m] vs. Eta [Characteristic Strength]) was calculated using final load at failure or survival of specimens.

Failure Analysis

Polarized light (Leica MZ APO Stereomicroscope, Leica Systems) and scanning electron micrographs (SEM) (Model S-3500N; Hitachi, Osaka, Japan) of failed samples were obtained for fracture analyses and comparisons between groups.³⁰ Survivor specimens were also inspected under polarized-light microscopy for possible crack initiation particularly on the side of the implant.

Results

SLF and Reliability

SLF mean values were 815 N for RS, 800 N for O, and 810 N for UN. The calculated reliability for a mission of 50,000 cycles at 200 N is presented in Table 1, and it shows that cumulative damage from loads reaching 200 N would lead to a survival of 81% of the RS, 79% of the O, and 72% of the UN implant-supported crowns. Given the overlap between upper and lower limits, no statistical difference was observed. For the O and UN groups, the β values of 0.46 and 0.36, respectively, indicated that load alone dictated the failure mechanism, and that fatigue damage did not seem to accumulate. In contrast, the β value of 2.1 for the RS indicated that failure rate increased over time due to damage accumulation (Fig. 2, a).

Table 1: Reliability for the Different Experimental Groups

Fig. 2: (a) Use level probability Weibull (probability of failure vs cycles). Note the beta values <1 for 3i Certain and SIN UN. (b) The Weibull 2-parameter contour plot (Weibull modulus vs characteristic strength) for groups comparison. Note the overlap between groups showing no significant differences in characteristic strength.

The Weibull 2-parameter contour plot showed a characteristic strength (Eta, which indicates the load at which 63.2% of the specimens of each group would fail) of 362.5 N for the RS, 348.1 N for O group, and 330.1 N for the UN group (not significantly different as per the overlap on plot presented in Fig. 2, b).

Failure Modes

All specimens fractured after SSALT. In general, the chief failure mode was screw fracture. Failures for the RS system was screw fracture in all specimens (n = 18), whereas in the O and in the UN systems screw fracture occurred in 11 specimens each. In the remaining samples, screw abutment failure (n = 5) and both screw and abutment fracture (n = 2) was observed. Polarized-light and SEM micrographs of the fractured surface of screws revealed compression curls, depicting the stress shift from tensile to compressive fields, which represent the end of the fracture event (Fig. 3).

Fig. 3: (a) Polarized-light and (b) SEM micrographs of a representative screw fracture after SSALT. White arrows show compression curl indicating the end of the fracture event and a strong bending component. Red arrows show the direction of crack propagation, while the pointers show the site of fracture origin. Images (c) and (d) are magnifications of (b) showing origin and the compression curl, respectively.

Discussion

This study evaluated 3 commercially available implant-abutment connection designs. According to the manufacturers, the RS system presents a tri-channel internal implant/abutment connection that offers the choice of 3 possible positions for connection, whereas the Certain Internal Connection system presents 6 of 12 internal connections with both a hex and a 12-point double-hex. The Cone Morse connection presents a hybrid system, showing a conical internal hex and a Morse taper connection.

The results of this study suggest that the geometric inherent differences of each design did play a role in the use level probability Weibull calculations, which showed that failure rate increased over time and were related to damage accumulation in the RS system, but not in the Osseotite and UN systems. Therefore, a fatigue-associated failure behavior was observed for the different tested systems.

Although subsequent probability Weibull calculation revealed no statistical difference in the Weibull modulus and characteristic strength, seen as an overlap between groups in the contour plot, higher variability in load at failure between systems is observed by the increasingly wider contour plot base observed for the UN, Osseotite, and RS groups, respectively. In agreement with this findings, Quek et al²¹ observed that load fatigue performance and failure location is system specific and related to the design characteristics of the implant-abutment connection.

The accelerated life testing method of step-stress was used in this investigation, where 3 different implant-abutment designs were tested under a chewing simulator in wet conditions. Through the interplay of variables such as at which step in a profile a specimen fails, in which profile the failures occur (mild, moderate, or aggressive), and how many fail at each step, the step-stress method allows for the prediction of specimen endurance at a given load with confidence intervals (based on calculations from a master Weibull distribution fit to the data). The rationale for using at least 3 profiles (mild, moderate, or aggressive) for this type of testing is based on the need to distribute failure across different step loads and allows better prediction statistics, narrowing confidence bounds.³¹

Considering the mastication forces of 206 N at anterior regions,³² the mean SLF values observed in this study suggests that all systems may withstand functional forces in this area. The main failure mode was screw fracture, observed mostly at the neck of the screw or between the screw threads. These fracture locations present a change in cross-sectional geometry along the length of the implant, likely related to an increase in stress concentration.²¹ According to Cehreli et al,³³ the screw is the primary responsible element that keeps the implant and the abutment assembled and the screw design may be a significant factor in joint stability.⁸ Thus, a contributing factor for screw failures is excessive bending of the screw-joint and settling effects arising from micromovements at the implant-abutment interface.^5,7,34 Specific to the RS, where screw fracture was observed in all samples, it can be assumed that the screw was effective in protecting the abutment from excessive functional load, serving as a safety mechanism in securing the implant and the surrounding structure from bending overload.^19,33 In agreement with this findings, Levine et al³⁵ did not observe implant fracture, whereas other studies^21,36 showed implant failures, such as fractures, mainly when narrow-diameter implants were used.²⁰

Conclusion

Although a restricted number of materials/processing conditions/specimen designs were investigated in this study, the methodology and the reliability established for implant-abutment designs provides the basis for comparison between different implant/abutment systems in future investigations. The specimen configuration (standardized implant-supported restoration with 3 different abutment designs and a metal crown chosen to isolate restoration materials variables) and loading condition (step-stress loading in water) used in this study showed that reliability was not significantly different between the 3 evaluated implant-abutment systems. On the other hand, failure modes differed between systems where screw fracture was chiefly observed in the RS system and mixed screw or screw and abutment fracture prevailed in the Osseotite and UN groups. Therefore, the postulated null hypothesis that stated that differences in reliability would not exist was accepted, and that differences in fracture modes would not exist was rejected.

Acknowledgments

Hitachi S3500N SEM imaging was made possible by the New York University College of Dentistry's cooperative agreement with the NIH/NIDCR. The authors also acknowledge CAPES, Brazil for Scholarship Process BEX 2434/09-1. Metal crown samples were provided by Marotta Dental Studio, NY, USA.