This study compared titanium (Ti), palladium (Pd), platinum (Pt), and gold (Au) ion release following induced accelerated tribocorrosion from three Au alloy abutment groups coupled with Ti implants over time; investigated contacting surface structural changes; and explored the effect of Au plating. Three abutment groups, G (n = 8, GoldAdapt, Nobel Biocare), N (n = 8, cast UCLA, Biomet3i), and P (n = 8, cast UCLA, Biomet3i, Au plated), coupled with implants (Nobel Biocare), immersed in 1% lactic acid, were cyclically loaded. Ions released (ppb) at T1, T2, and T3, simulating 3, 5, and 12 months of function, respectively, were quantified by inductively coupled plasma mass spectrometry (ICP-MS) and compared. Surface degradation and fretted particle composition after T3 were evaluated with scanning electron microscopy and energy-dispersive X-ray spectroscopy (SEM/EDX). ICP-MS data were nonparametric, expressed as medians and interquartile ranges. SEM/EDX showed pitting, crevice corrosion, and fretted particles on the components. Released ion concentrations in all groups across time significantly decreased for Pd (P < .001, median range: 1.70–0.09), Pt (P = .021, 0.55–0.00), and Au (P < .001, 1.01–0.00) and increased for Ti (P = .018, 2.49–5.84). Total Ti release was greater than other ions combined for G (P = .012, 9.86–2.30) and N (P < .001, 13.59–5.70) but not for P (P = .141, 8.21–3.53). Total Ti release did not differ between groups (P = .36) but was less variable across group P. On average, total ion release was 13.77 ppb (interquartile range 8.91–26.03 ppb) across the 12-month simulation. Tribocorrosion of Ti implants coupled with Au abutments in a simulated environment was evidenced by fretted particles, pitting, and crevice corrosion of the coupling surfaces and release of ions. More Ti was released compared with Pd, Pt, and Au and continued to increase with time. Abutment composition influenced ion release. Au-plated abutments appeared to subdue variation in and minimize high-concentration spikes of titanium.

Biomedical implants placed in aqueous environments under cyclic loads are prone to corrosion, resulting in the release of metal ions and particles into the surrounding tissues.1,2  Different types of corrosion have been described and can coexist, including galvanic, pitting, crevice, fretting, and electrochemical.35  Simultaneous chemical, wear, and electrochemical interactions are defined as tribocorrosion.6,7  These factors can contribute to mechanical failure2,8  and possibly to biologic failure.

The osteolytic effect of corrosion products has long been highlighted in the orthopedic literature.1,9,10  This foreign-body reaction translates into the common clinical and radiographic orthopedic finding of focal aseptic osteolysis11  and is considered the main cause of aseptic inflammation and consequent prosthesis loosening.12 

Titanium (Ti) and its alloys are still the favored materials for dental implants because of their strength, biocompatibility, and corrosion resistance,7  but a recent systematic review documented many potential causes and time points of Ti particle and ion release in the adjacent soft tissues and bone.2  Biopsy studies have shown Ti particles surrounded by inflammatory cells such as lymphocytes and macrophages13,14  and areas of fibrosis and necrosis.15  The minimum concentration of Ti ions that causes toxic biologic effects has been variously reported and ranges from 9000 ppb to 9 ppb.1518 

In addition to Ti, there are numerous materials available for abutment fabrication, subject to corrosion, and also in contact with peri-implant tissues. Noble gold (Au)–based alloys may have comparatively reduced strength and hardness and are more resistant to tribocorrosion than nonnoble alloys are, such as cobalt–chrome (CoCr)–based alloys,19  but they result in higher differential galvanic corrosion of Ti because of lower electrochemical potentials.20 

The method of abutment fabrication and abutment geometry can also influence corrosion. Tribocorrosion can be accelerated when there is a misfit between components resulting in micro-movement and friction under cyclic load8  and screw loosening.2  It is commonly accepted that internal connections have superior fit and reduced micromovements under functional loading compared with external hex connections and are therefore potentially less subject to tribocorrosion.2  They also have reduced stress concentration in the surrounding bone at the neck of the implant at off vertical loading,21  which has been associated with marginal bone loss.22  However, internal stresses within the implants are higher, and recent studies indicate internal connections may have relatively higher fracture rates of components, including the implants, with increasing clinical times.23,24 

An up-to 30-year retrospective study of 3211 external hex implants resulted in a cumulative survival rate of 97.1%, indicating that the identified disadvantages may not translate to clinical failure.25 

External hex abutments with a prefabricated base and a burnout plastic sleeve will ensure a more predictable machined fit compared with fully cast plastic UCLA abutments26,27  but still maintain the advantages of form customization. However, contact between the dissimilar alloys could lead to galvanic corrosion.28 

An up-to-14-year clinical study observed 256 1-piece high-Au alloy cast plastic UCLA abutments with Au-plated implant coupling surfaces of the metal-ceramic implant-supported single crowns (ISCs Au).29  Marginal bone levels were relatively stable, and joint stability was excellent with a 5-year screw loosening rate of 1.73% (95% confidence interval [CI]: 0.22%–3.22%) compared with 8.5% after 5 years reported in a systematic review of biological, technical, and esthetic complications of ISCs.30  The effects of the Au-plated coupling surfaces have not been studied.

In-vivo measurements of corrosion are difficult to obtain. Any assessment of the combined effects of chemical, wear, and electrochemical interactions (tribocorrosion) involves functional loading over time, but most in vitro models have used static immersion protocols.31,32  Dynamic immersion tests in a simulated oral environment have been described33  and standardized for testing oral implants.34  One model has combined immersion coupled with hydraulic cyclic loading machines to stimulate accelerated tribocorrosion.20,35  Ions released into the immersion fluid were analyzed with inductively coupled plasma–mass spectrometry (ICP-MS). Structural and elemental composition changes of the contacting surfaces were analyzed by scanning electron microscopy and energy-dispersive X-ray spectroscopy (SEM/EDX).

The aims of this study were to compare Ti, palladium (Pd), platinum (Pt), and Au ion release following dynamic immersion under cyclic loading (induced accelerated tribocorrosion) of 3 groups of Au alloy abutments coupled with Ti implants through ICP-MS of the immersion liquid; to investigate the influence of time on ion release; to evaluate structural and elemental composition changes to the contacting surfaces after loading with SEM/EDX; and explore the effect of Au plating the coupling surfaces.

The null hypotheses were (1) the structural integrity of the coupling surfaces of the abutments and implants would remain unchanged after loading, (2) the type and concentration of ions released would not be affected by the different abutment groups, (3) the type and concentration of the ions released would not be affected by time, and (4) Au plating the abutment coupling surface would have no impact on the concentration of ion release or the structural integrity.

The components used in this study and their manufacturers are presented in Table 1. A total of 24 Au alloy abutments across 3 groups defined by fabrication method and surface composition (G, N, P; each n = 8) were studied. Abutment groups were (1) group G, Au alloy fitting surfaces with cast-to-plastic sleeves that were cast and nonplated (GoldAdapt, Nobel Biocare); (2) group N, UCLA plastic copings that were cast and nonplated (Biomet 3i); and (3) group P, UCLA plastic copings that were cast and Au plated (Biomet 3i). No data were available in the literature to calculate a sample size for the study. Therefore, data were collected for Ti ion release for 3 samples from each group (G, μ = 14.34 ppb; N, μ = 32.72 ppb; P, μ = 8.17 ppb; σ = 16.07 ppb) and used to calculate the sample size based on α = .05, β = .8. Group G was compared with group P (resulting in a sample size of 7), and group N was compared with group P (resulting in a sample size of 1). To allow for increased heterogeneity across the study, a sample size of 8 per group was selected. This compares with a sample size of 6 used in previous studies on accelerated corrosion.20,34 

Table 1

Components of the implant-abutment couplings

Components of the implant-abutment couplings
Components of the implant-abutment couplings

The plastic sleeve of the 8 GoldAdapt cast-to abutments (group G) and the plastic abutment copings of the 16 UCLA (groups N and P) were cast according to the manufacturer's instructions with V-Classic high-Au content alloy (Cendres+Métaux SA) containing 94.06% of Au + Pt groups, 75% of Au, and 19% of Pd. After casting, abutments in only group P were Au plated with Aurogold C5 1.5 g/L solution (Alphabond Dental Pty Ltd) electrochemically deposited for 2 hours at 2.8 V. Abutments in groups G, N, and P were connected to external hex Ti implants (Brånemark System MkIII TiUnite RP, Nobel Biocare) with stainless-steel 24-carat Au-plated Gold-Tite Square UniScrew screws (Biomet 3i), torqued to 35 Ncm and prepared for cyclic loading. Figure 1 summarizes the experimental groups.

Figure 1.

The 3 types of abutments. G, GoldAdapt cast-to; N, nonplated UCLA; P, gold-plated UCLA.

Figure 1.

The 3 types of abutments. G, GoldAdapt cast-to; N, nonplated UCLA; P, gold-plated UCLA.

Close modal

The implants were embedded in Specifix-20 epoxy resin (Struers) poured into polystyrene containers, leaving the implant collar exposed. The abutment cylinders were coated with commercial resin (Kiko), leaving the collar exposed to limit the corrosion effect on the connection area.

Cyclic loading

The cyclic loading test protocol followed the methodology described in a recent study.20  Freshly prepared 1% lactic acid (0.1 mol/L lactic acid and 0.1 mol/L sodium chloride, pH 2.3, ISO 10271:201135 ) was added to the polystyrene container in a volume (2 mL) sufficient to produce a ratio of 1 mL of solution per 1 cm2 of exposed surface area (ISO 10271:201135 ). A universal hydraulic testing machine (Dartec HC10, Dartec Limited) was used to mechanically load each specimen, 30° off-axis with a peak force of 100 N varying sinusoidally (ISO 14801:200736 ). The total number of cycles was 240 000 at a frequency of 2.7 Hz, simulating 1 year of use.37 Figure 2 shows a schematic representation and an image of the specimen container mounted in the testing machine.

Figure 2.

Left: schematic representation of the loading apparatus. Right: specimen container was tightened in a 30° jig on the testing machine.

Figure 2.

Left: schematic representation of the loading apparatus. Right: specimen container was tightened in a 30° jig on the testing machine.

Close modal

Pre- and postloading SEM/EDX

As a convenience sample, 1 abutment from each group and its corresponding implant were selected for preloading SEM/EDX. After loading, the same specimens were once more subjected to SEM/EDX to assess if any visual structural changes had occurred on the contacting surfaces, such as the appearance of cracks, pits, and lose particles, and whether the elemental composition had changed. A XL30 FEG SEM (Philips) and an EDX detector with Inca software (Oxford Instruments Analytical) were used. Secondary electron images were taken at ×35, ×150, ×500, and ×1000 magnifications of various representative areas of the contacting surfaces.

Quantification of ionic release

Three time points for collecting the immersion liquid (T1, T2, T3) were calculated based on the set frequency (2.7 Hz) and total number of cycles (240 000 ≈ 12 months), so that they represented 3, 5, and 12 months in clinical function, respectively33  (Table 2). An equation was created to determine the collection times in hours, where 3600 is the number of seconds in 1 hour and 20 000 is the average number of cycles per month.
formula
Table 2

Collection times and their equivalent in months of clinical function.

Collection times and their equivalent in months of clinical function.
Collection times and their equivalent in months of clinical function.

At T1 and T2, loading was paused, the immersion solution was collected, fresh solution was added to the specimen container, and loading was resumed. At T3, loading was stopped, and there was a final collection of the immersion solution. Each time the immersion solution was collected, the jar was washed with 2 mL of 2% nitric acid to dissolve and collect any remaining particles (wash solution). The final solution for analysis for each time point consisted of a mixture of 1 mL of immersion solution, 2 mL of wash solution, and 7 mL of fresh 2% nitric acid, making a total of 10 mL of 1% lactic acid diluted by a factor of 10 in 2% nitric acid to comply with ICP-MS specifications.

The final solutions were analyzed for the presence of Ti from the implants and Pd, Pt, and Au from the abutments, quantified with ICP-MS (Varian/Bruker 800-MS Series, Analytical West) and expressed as parts per billion (ppb).

Statistical analysis

The SEM/EDX images and data were examined qualitatively for visual and elemental changes on the contacting surfaces. Ion release data were explored quantitatively. Ion release data were not normally distributed. The quantities of each ion (Ti, Pd, Pt, Au) released at each time point (T1, T2, T3), in relation to each abutment type (N, n = 8; P, n = 8; G, n = 8) measured by the ICP-MS were tallied and expressed as medians bounded by interquartile ranges (IQRs), in ppb. The following were investigated:

  • Differences between the individual types of ions released (Ti, Pd, Pt, Au) by each time point (T1, T2, T3), for each abutment group (G, n = 8; N, n = 8; P, n = 8) were visually explored with box plots, and the influence of the factors “group” and “time” on the ionic release was evaluated with a generalized linear mixed model (GLMM).

  • Differences in the total type of ions released (Ti_total, Pd_total, Pt_total, Au_total) by T3 for each abutment group (G, n = 8; N, n = 8; P, n = 8) were compared with the Kruskal-Wallis test (nonparametric test to compare continuous results from more than 2 independent groups, with Bonferroni correction for pairwise comparisons) and visually depicted on box plots.

  • Differences in the total ionic release (Ti + Pd + Pt + Au = grand_total) by T3 for all abutments (n = 24) were compared with the Kruskal-Wallis test and visually depicted on a box plot.

  • The total Ti release (from the implants) for all specimens in each group by T3 was compared with the total combined ion release of Pd, Pt, and Au (from the abutments) in each group by T3 using the Mann-Whitney test (nonparametric test to compare continuous results from 2 independent groups) and visually depicted in a box plot.

Data were analyzed with SPSS version 26.0 (IBM Corp). The significance level was set at 5%. The methodology was reviewed by an independent statistician.

Pre- and postloading SEM/EDX

One implant and abutment specimen from each group was assessed via SEM before and after loading. There was no qualitative difference between the specimens. When compared with preloading images, postloading SEM of the implant and abutment contacting the surfaces showed black spots and scratchlike marks, strongly indicating pitting and crevice corrosion (Figure 3). Given their relatively smooth machined surfaces, these changes were especially visible on the implants, in the vicinity of the hex. Postloading SEM also showed the presence of particles that were not released into the immersion solution.

Figure 3.

(a) Preloading scanning electron microscopy (SEM) of the internal surface of the hex of an implant showing a smooth, uniform machined surface (×150). (b) Postloading SEM of the same area showing the presence of pitting (×150). This implant was connected to an Au-plated abutment. (c) Preloading SEM of the external area of the coupling surface of an implant showing a smooth, uniform machined surface (×500). (d) Postloading SEM of the same area showing the presence of pitting and particles (×500). This implant was connected to a nonplated abutment.

Figure 3.

(a) Preloading scanning electron microscopy (SEM) of the internal surface of the hex of an implant showing a smooth, uniform machined surface (×150). (b) Postloading SEM of the same area showing the presence of pitting (×150). This implant was connected to an Au-plated abutment. (c) Preloading SEM of the external area of the coupling surface of an implant showing a smooth, uniform machined surface (×500). (d) Postloading SEM of the same area showing the presence of pitting and particles (×500). This implant was connected to a nonplated abutment.

Close modal

The EDX analysis identified the attached particles as Ti. External contaminants were not identified, and therefore, the particles present were likely the result of wear debris. Most Ti particles were seen close to the hex of the implants.

Quantification of ionic release

The results of the GLMM are summarized in Table 3. The summary data of Ti, Pd, Pt, and Au ion release over the study period are outlined in Table 4. The ion release at T1, T2, and T3 is visually depicted in Figure 4.

Table 3

Influence of time and group on the release of each element*

Influence of time and group on the release of each element*
Influence of time and group on the release of each element*
Table 4

Summary data of release of Ti, Pd, Pt, and Au ions across the T1, T2, and T3 time periods in ppb as median and interquartile range (Q1, Q3)*

Summary data of release of Ti, Pd, Pt, and Au ions across the T1, T2, and T3 time periods in ppb as median and interquartile range (Q1, Q3)*
Summary data of release of Ti, Pd, Pt, and Au ions across the T1, T2, and T3 time periods in ppb as median and interquartile range (Q1, Q3)*
Figures 4–7.

Figure 4. Box plots of ion release throughout the experiment time across the 3 groups (ppb). Note that each vertical axis has a different ppb scale. (a) Titanium. (b) Palladium. (c) Platinum. (d) Gold. P values for related tests are documented in Table 3. Figure 5. Box plots of the comparison of titanium (Ti) ion release from the implants compared with the combined release of platinum (Pt) + palladium (Pd) + gold (Au) ions from the abutments (ppb). There was a significant difference between ion release in groups G (P = .012) and N (P < .001) but not in group P (P = .141). Figure 6. (a) Box plots of the distribution of the total amounts of each element (titanium [Ti], platinum [Pt], palladium [Pd], gold [Au]) released across the 3 groups (ppb). Ti release was significantly greater than each of the other 3 elements in all 3 groups (P < .001). (b) Box plots of the distribution of the total amounts of Pt, Pd, and Au released from the abutments in the 3 groups. Figure 7. Box plots of the total ion release over the 3 time points (ppb). There was no significant difference between the groups (P = .32).

Figures 4–7.

Figure 4. Box plots of ion release throughout the experiment time across the 3 groups (ppb). Note that each vertical axis has a different ppb scale. (a) Titanium. (b) Palladium. (c) Platinum. (d) Gold. P values for related tests are documented in Table 3. Figure 5. Box plots of the comparison of titanium (Ti) ion release from the implants compared with the combined release of platinum (Pt) + palladium (Pd) + gold (Au) ions from the abutments (ppb). There was a significant difference between ion release in groups G (P = .012) and N (P < .001) but not in group P (P = .141). Figure 6. (a) Box plots of the distribution of the total amounts of each element (titanium [Ti], platinum [Pt], palladium [Pd], gold [Au]) released across the 3 groups (ppb). Ti release was significantly greater than each of the other 3 elements in all 3 groups (P < .001). (b) Box plots of the distribution of the total amounts of Pt, Pd, and Au released from the abutments in the 3 groups. Figure 7. Box plots of the total ion release over the 3 time points (ppb). There was no significant difference between the groups (P = .32).

Close modal

Across each group, the Ti release significantly increased over the simulated time periods (P = .018) and was significantly greater at T3 compared with T1 (P = .044; Figure 4a). In ppb, Ti release increased from 2.49 (IQR 0.16 to 6.68) to 5.62 (IQR 2.29 to 7.02) for group G, from 4.13 (IQR 2.20 to 8.59) to 5.84 (IQR 3.80 to 17.48)for group N, and from 2.50 (IQR 0.90 to 3.02) to 3.01 (IQR 0.83 to 3.82) for group P across T1 to T3. Across the entire time period, the total release (T1 + T2 + T3) of Ti ions did not significantly differ between groups (P = .36)

Group (P < .001), time (P < .001), and the interaction between both variables (P = .039) had a significant effect on Pd release (Figure 4b). Across each group, Pd release significantly decreased from T1 to T2 (P < .001) and from T1 to T3 (P < .001). At T1, group N released significantly more Pd (1.70 [IQR 1.13–2.89] ppb) than did group G (0.31 [IQR 0.18–0.64] ppb; P = .010) and group P (0.44 [IQR 0.25–1.08] ppb; P = .013). In ppb, Pd release decreased from 0.31 (IQR 0.18–0.64) to 0.09 (IQR 0.00–0.14) for group G, from 1.70 (IQR 1.13–2.89) to 0.34 (IQR 0.09–0.55) for group N, and from 0.44 (IQR 0.25–1.08) to 0.17 (IQR 0.14–0.23) for group P across T1 to T3. Across the entire time period, the total release (T1 + T2 + T3) of Pd ions differed between groups (P = .025) and was significantly higher in group N when compared with group G (P = .024).

Group (P < .001), time (P = .021), and the interaction between both variables (P = .010) had a significant effect on the release of Pt (Figure 4c). Across each group, Pt release significantly decreased from T1 to T2 (P = .034) and T1 to T3 (P = .017). At T1, group G released significantly more Pt (0.55 [IQR 0.35–0.88] ppb) than did group N (0.07 [IQR 0.03–0.13] ppb; P < .001) and group P (0.00 [IQR 0.00–0.02] ppb; P < .001). At T3, group G released significantly more Pt (0.19 [IQR 0.05–0.25] ppb) than did group P (0.00 [IQR 0.00–0.16] ppb; P = .016). In ppb, Pt release decreased from 0.55 (IQR 0.35–0.88) to 0.19 (IQR 0.05–0.25) for group G, from 0.07 (IQR 0.03–0.13) to 0.03 (IQR 0.00–0.13) for group N, and from 0.00 (IQR 0.00–0.02) to 0.00 (IQR 0.00–0.16) for group P across T1 to T3. Across the entire time period, the total release (T1 + T2 + T3) of Pt ions differed between groups (P = .002) and was significantly lower in group P (P = .003) and group N (P = .026), as compared with group G.

Group (P = .020) and time (P < .001) had a significant effect on the release of Au (Figure 4d). Across each group, Au release significantly decreased from T1 to T2 (P = .023) and T1 to T3 (P < .001). At T2, group P released significantly more Au (0.45 [IQR 0.16–1.14] ppb) than did group G (0.00 [IQR 0.00–0.11] ppb; P = .041). In ppb, Au release decreased from 0.12 (IQR 0.00–0.37) to 0.03 (IQR 0.00–0.14) for group G, from 0.70 (IQR 0.00–1.88) to 0.00 (IQR 0.00–0.24) for group N, and from 1.01 (IQR 0.37–1.56) to 0.24 (IQR 0.13–0.37) for group P across T1 to T3. Across the entire time period, the total release (T1 + T2 + T3) of Au ions did not differ between groups (P = .11).

The total release (T1 + T2 + T3) of Ti ions from the implants (Ti_total) was significantly greater than the release of ions from the abutments (Pd_total + Pt_total + Au_total) in group G (9.86 [IQR 7.43–18.96] ppb vs 2.30 [IQR 1.29–3.07] ppb; P = .012) and group N (13.59 [IQR 8.64–42.85] ppb vs 5.70 [IQR 1.77–6.88] ppb; P < .001) but not group P (8.21 [IQR 3.01–10.29] ppb vs 3.53 [IQR 1.63–4.51] ppb; P = .141; Figure 5). There was no significant difference in the total release of combined Pt, Pd, and Au over the 3 groups (Figure 6a,b).

The total release (T1 + T2 + T3) of all ions (Ti + Pd + Pt + Au) did not differ significantly between groups (P = .32; Figure 7). However, the total release appeared to be less variable across abutments that were plated (group P) when compared with abutments that were not plated (groups G and N; Figure 7). In ppb, (Ti + Pd + Pt + Au) release was 13.76 (IQR 8.85–19.99) for group G, 19.83 (IQR 10.27–50.37) for group N, and 11.10 (IQR 6.52–14.44) for group P.

On average, 13.77 (IQR 8.91–26.03) ppb of the total measured ions were released across the simulated 12-month period.

This study assessed induced accelerated tribocorrosion through measurement of ion release and structural changes associated with Ti implants and Au-based abutments commonly used in practice and explored the modulating effect of Au plating the abutment coupling surfaces. It allowed for an objective quantitative assessment of ion release through ICP-MS. The qualitative assessment of structural changes observed through SEM/EDX was subjective but nevertheless informative since features of corrosion could be visualized and compared with similar studies.20 

All 4 null hypotheses were rejected. The type and concentration of ions released by the different implant-abutment couplings was affected by both abutment composition and time, Au plating of the coupling surface influenced ion release, and the structural integrity of the contacting surfaces of the abutments and implants had degraded. Overall, the abutment composition and fabrication technique will influence corrosion of implants and abutments.

Au, Pt, and Pd were released by all abutments across the 12-month simulated time, and the amount released decreased significantly over time. Pd release was greater from the nonplated castable abutments (N) when compared with the cast-to (G) and Au-plated castable abutments (P). Pt release was greater from the cast-to abutments (G) when compared with the nonplated castable (N) and Au-plated castable abutments (P). Au release did not differ between groups. Across these groups, Au plating on the castable abutments appeared to reduce overall ion release.

Ti was released by all implants across the 12-month simulated time, and the amount released was both significantly greater than all other ions combined and continued to increase significantly over time. The amount of Ti release was not found to be related to abutment type, but substantial variation in release occurred across the specimens within 2 of the 3 groups. The variation was far less pronounced in the Au-plated castable group (P), as clearly delineated on the box plots (Figure 7). Although the sample size for this study was based on a power calculation from initial data observed in this experiment, the authors hypothesize that a difference is likely present between these groups but was not observed due to low numbers. The authors hypothesize that Au plating of the castable abutment surface will reduce Ti release from tribocorrosion, and further research is required to explore this potential outcome.

Immersion tests can be static or dynamic. Dynamic models such as the one used in this and other similar studies have the advantage of mimicking masticatory function to better represent the tribocorrosion processes taking place in the oral environment.20,33 

Implant-supported prostheses are expected to survive for long periods. An initial release of corrosion products, which then declines with time, may not affect tissue health or prosthesis longevity. However, if corrosion is sustained, or increases with time, there is a higher risk for accumulation of debris in the peri-implant tissues and toxic biologic effects. Time was a significant variable in this study. Overall, the release of Pd, Pt, and Au, albeit in relatively low concentrations, decreased over time, whereas that of Ti increased. This finding is in agreement with other studies.31,32 

There is conflicting information on what concentration of Ti ions is toxic. In vitro studies have shown that a concentration less than 9000 ppb has no significant effect on the viability of osteoblast or epithelial-like cells.16  On the other hand, an ex vivo study of human jaws with implants demonstrated that concentrations as low as 1940 ppb resulted in the presence of multinucleated cells and fibrotic and necrotic areas.15  When combined with other ions such as aluminum and vanadium, often used in Ti alloy implant systems, concentrations of just 750 ppb can be lethal to osteoblasts.17  However, in the presence of certain microorganisms, Ti ion concentrations as low as 9 ppb, affected the infiltration of monocytes and osteoclast differentiation by increasing the sensitivity of epithelial cells to microorganisms.18  Any lowering of the release of Ti ions would seem to be clinically desirable.

The maximum concentration of Ti ions found in this study—with the interquartile range observing over 40 ppb of Ti release for the nonplated group (N)—may or may not be below the toxic values variously described in the literature. However, the present study equated to only 12 months in function, and if Ti release is sustained or continues to increase with time, any toxic effects could be more enhanced.

In a previous study using a similar protocol with cast-to GoldAdapt abutments coupled with Ti implants, Au ions were released in comparable concentrations, but the Ti ion release was much greater, in the order of 120 ppb20  compared with 40 ppb (IQR 8.64–42.85) observed in this study for group N. Although most of the methodology was identical, following loading the implant/abutment couplings, the specimens in the previous study were immersed in 1% lactic acid medium for 6 days prior to ICP-MS. The fretted particles were thus subject to dissolution during this period, and this immersion may explain the disparity between the results. It is therefore likely that the Ti ion release measured in the present study underestimates the ion concentration that might be present if fretted particles were released into an alternative medium, such as peri-implant tissues.

Ti ion release from the implants that were coupled to the Au-plated abutments (group P) could be expected to be greater when compared with nonplated abutment groups, due to a greater difference in electrochemical potential. However, this was not the case. Although not statistically significant, group P appeared to release not only the least Ti but also the least overall concentration of ions. This may represent a clinical advantage.

Au plating of the coupling surfaces of metal abutments is not a common clinical practice, although it has been previously documented.29  In that study, cast UCLA abutments were used. The marginal bone levels were relatively stable, and joint stability was excellent. A recent study documented the use of an anaerobic sealing agent between implant/UCLA abutment couplings.38  The sealing agent (Loctite 2400, Henkel) reduced the misfit of the abutments and maintained higher screw preload compared with couplings without the sealing agent. However, the dimethacrylate ester composition of the sealant is subject to chemical and fretting decomposition and is potentially biologically toxic.

Au plating of abutment screws has improved coupling joint stability due to the ductility of the plate enhancing and maintaining screw preload.39,40  It is likely that the high ductility of the Au-plated abutment coupling surfaces used in this study reduces the microgap thereby improving the fit by filling in between microscopic disparities arising either through machining or casting processes and plastically deforming during screw tightening, similar to the sealing agent demonstrated in a previous study.38  The high ductility is also likely to assist maintenance of the screw preload, thus reducing micromotion in function and consequently lowering wear and fretting corrosion. This may explain the relatively low Ti and overall ionic release observed in group P, even given the greatest difference in electrochemical potential between groups. Varying Au plate thickness may have an impact and requires further research. Au is not biologically toxic in particulate form and has relatively low toxicity in ionic form. Au plating of coupling surfaces of nonnoble and harder abutment materials, such as CoCr alloys, may have a more significant reduction of Ti ion release. Internal connections have superior fit and reduced micromovements under functional loading compared with external hex connections2  but are still subject to tribocorrosion. This may be further reduced with Au plating the coupling surfaces. Further research is indicated.

The quantitative release of Ti particles still attached to the implants was not measured in this study. There may be a significant difference in particles generated between the groups, with the ductility of the elemental Au in the plated group resulting in less fretting corrosion.

The SEM/EDX analysis showed that pitting and crevice corrosion, as well as the presence of Ti particles, were particularly evident on the internal surface of the hex of the implants, in accordance with the findings of a similar study.20  This may be explained by the seclusion of the site, where lower oxygen availability accentuates corrosion.5  Pits that reach critical depths lead to high local stress accumulation and may act as crack-initiation sites.3  In a clinical situation, this can be accelerated by cyclic loading.8  Structural damage of the hex can compromise the success and longevity of the implant treatment.

The particles observed by posttreatment SEM/EDX had the same elemental composition as the underlining surface, indicating that they were not external contaminants but rather actual wear debris. Even though ICP-MS showed diminutive ionic amounts, the attached particles seen through SEM/EDX were not released into the liquid and therefore not accounted for by ICP-MS. Preparation of the specimen jars involving dilution of the 1% lactic acid immersion solution with 2% nitric acid, however, may have removed and dissolved particles that were then not seen with SEM/EDX, especially at the peripheries of the contacting surfaces, leading to an underestimation of the amount of particles produced. This could also explain the apparent concentration of Ti particles on the inner surface of the hex.

The identification of corrosion particles on the implant and abutment surfaces is important because, in a clinical situation, particles could be released into the peri-implant tissues and elicit an inflammatory response, as has been shown in recent oral biopsy studies.13,15  Furthermore, these particles can dissolve in the oral fluids and become a source of additive metal ions with potential toxic effects as previously discussed.

Techniques in implant dentistry, especially the use of printed alloys, are rapidly evolving. The use of noble metals is decreasing due to financial limitations. However, corrosion resistance, both electrochemical and tribocorrosion, of these new and cheaper materials is not likely to be as effective as the high noble alloys. A technique, such as Au plating the abutment coupling surfaces, which is relatively cheap and technically undemanding, may solve this problem as well as minimize microgaps and enhance implant/abutment joint stability.

This study observed tribocorrosion of Ti implants coupled with Au abutments in a simulated environment evidenced by fretting of particles, pitting, and crevice corrosion of the coupling surfaces and release of ions. More Ti was released compared with Pd, Pt, and Au and continued to increase with time. Abutment composition influenced ion release, with Au-plated abutments appearing to subdue variation in and minimize high-concentration spikes of Ti. Tribocorrosion of implant-abutment couplings and its impact on peri-implant tissues requires further investigation.

Abbreviations

Abbreviations
CI:

confidence interval

EDX:

energy-dispersive X-ray spectroscopy

GLMM:

generalized linear mixed model

ICP-MS:

inductively coupled plasma–mass spectrometry

ISCs:

implant-supported single crowns

ISO:

International Organization for Standardization

IQR:

interquartile range

SEM:

scanning electron microscopy

Special thanks to Mr David Boniface from the Eastman Dental Institute for the support with the statistical analysis, to Dr Nicky Morden and Dr Tom Gregory for the SEM/EDX analysis, to Wallis Franklin for the casting procedures, and to Dr Graham Palmer for the laboratory queries.

1. 
Archibeck
MJ,
Jacobs
JJ,
Roebuck
KA,
Glant
TT.
The basic science of periprosthetic osteolysis
.
J Bone Joint Surg Br
.
2000
;
82
:
1478
1478
.
2. 
Delgado-Ruiz
R,
Romanas
G.
Potential causes of titanium particle and ion relaese in implant dentistry
.
Int J Mol Sci
.
2018
;
19
:
3585
3620
.
3. 
Landolt
D.
The corrosion of metals
.
In:
Landolt
D,
ed.
Corrosion and Surface Chemistry of Metals. 1st ed
.
Lausanne, Switzerland
:
EPFL Press;
2007
:
11
13
.
4. 
Oldfield
JW.
Electrochemical theory of galvanic corrosion
.
In:
Hack
HP,
ed.
Galvanic Corrosion. 1st ed
.
Conshohocken, West Penn
:
ASTM International;
1988
:
5
22
.
5. 
Gittens
RA,
Olivares-Navarrete
R,
Tannenbaum
R,
Boyan
BD,
Schwartz
Z.
Electrical implications of corrosion for osseointegration of titanium implants
.
J Dent Res
.
2011
;
90
:
1389
1397
.
6. 
Souza
JC,
Henriques
M,
Teughels,
W,
Ponthiaux
P.
Wear and corrosion interactions on titanium in oral environmant: a literature review
.
J Bio Tribo Corr
.
2015
;
1
:
13
25
.
7. 
Apaza-Bedoya
K,
Tarce
M,
Benfatti
CA,
Henriques
B,
Teughels
W,
Souza
JC.
Synergistic interactions between corrosion and wear at dental implant connections: a scoping review
.
J Periodont Res
.
2017
;
52
:
946
954
.
8. 
Sridhar
T,
Vinodhini
S,
Kamachi Mudali
U,
Venkatachalapathy
B,
Ravichandran
K.
Load-bearing metallic implants: electrochemical characterisation of corrosion phenomena
.
Mater Technol
.
2016
;
31
:
705
718
.
9. 
Willert
HG,
Semlitsch
M.
Reactions of the articular capsule to wear products of artificial joint prostheses
.
J Biomed Mater Res
.
1977
;
11
:
157
164
.
10. 
Plummer
DR,
Berger
RA,
Paprosky
WG,
Sporer
SM,
Jacobs
JJ,
Della Valle
CJ.
Diagnosis and management of adverse local tissue reactions secondary to corrosion at the head-neck junction in patients with metal on polyethylene bearings
.
J Arthroplasty
.
2016
;
31
:
264
268
.
11. 
Zicat
B,
Engh
CA,
Gokcen
E.
Patterns of osteolysis around total hip components inserted with and without cement
.
J Bone Joint Surg Br
.
1995
;
77
:
432
439
.
12. 
Caicedo
MS,
Desai
R,
McAllister
K,
Reddy
A,
Jacobs
JJ,
Hallab
NJ.
Soluble and particulate Co-Cr-Mo alloy implant metals activate the inflammasome danger signaling pathway in human macrophages: a novel mechanism for implant debris reactivity
.
J Orthop Res
.
2009
;
27
:
847
854
.
13. 
Wilson
TG,
Valderrama
P,
Burbano
M,
et al.
Foreign bodies associated with peri-implantitis human biopsies
.
J Periodontol
.
2015
;
86
:
9
15
.
14. 
Fretwurst
T,
Buzanich
G,
Nahles
S,
Woelber
JP,
Riesemeier
H,
Nelson
K.
Metal elements in tissue with dental peri-implantitis: a pilot study
.
Clin Oral Implants Res
.
2016
;
27
:
1178
1186
.
15. 
He
X,
Reichl
F-X,
Wang
Y,
et al.
Analysis of titanium and other metals in human jawbones with dental implants—a case series study
.
Dent Mater
.
2016
;
32
:
1042
1051
.
16. 
Wachi
T.
Release of titanium ions from an implant surface and their effect on cytocline production related to alveloar bone resorption
.
Toxicology
.
2015
;
327
:
1
9
.
17. 
Alrabeah
GO,
Brett
P,
Knowles
JC,
Petridis
H.
The effect of metal ions released from different dental implant-abutment couples on osteoblast function and secretion of bone resorbing mediators
.
J Dent
.
2017
;
66
:
91
101
.
18. 
Mine
Y,
Makihira
S,
Nikawa
H,
et al.
Impact of titanium ions on osteoblast-, osteoclast-and gingival epithelial-like cells
.
J Prosthodont Res
.
2010
;
54
:
1
6
.
19. 
Wataha
JC.
Alloys for prosthodontic restorations
.
J Prosthet Dent
.
2002
;
87
:
351
363
.
20. 
Alrabeah
GO,
Knowles
JC,
Petridis
H.
Reduction of tribocorrosion products when using the platform-switching concept
.
J Dent Res
.
2018
;
97
:
995
1002
21. 
Balik
A,
Karatas
MO,
Keskin
H.
Effects of different abutment connection designs on the stress distribution around five different implants: a 3-dimensional finite element analysis
.
J Oral Implantol
.
2012
;
38(Spec No):491–496.
22. 
Koo
TE,
Lee
EJ,
Kim
JY,
et al.
The effect of internal verses external abutment comnection modes on crestal bome changes around dental implants: a radiographic naalysis
.
J Periodontol
.
2012
;
83
:
1104
1109
.
23. 
Lee
DW,
Kim
NH,
Lee
Y,
Oh
YA,
Lee
JH,
You
HK.
Implant fracture failure rate and potential associated risk indicators: an up-to 12-year retrospective study of implants in 5, 124 patients
.
Clin Oral Implants Res
.
2019
;
30
:
206
217
.
24. 
Yi
Y,
Koak
JY,
Kim
SK,
Lee
SJ,
Heo
SJ.
Comparison of implant component fractures in external and internal type: a 12-year retrospective study
.
J Adv Prosthodont
.
2018
;
10
:
155
162
.
25. 
Jemt
T.
Single implant survival: more than 30 years of clinical experience
.
Int J Prosthodont
.
2016
;
29
:
551
558
.
26. 
Barbosa
GA,
Simamoto
PC,
Júnior
Fernandes Neto
AJ,
de Mattos Mda
G,
Neves
FD.
Prosthetic laboratory influence on the vertical misfit at the implant/UCLA abutment interface
.
Braz Dent J
.
2007
;
18
:
139
143
.
27. 
Carr
AB,
Brunski
JB,
Hurley
E.
Effects of fabrication, finishing and polishing procedures on preload in prostheses using conventional ‘gold' and plastic cylinders
.
Int J Oral Maxillofac Implants
.
1966
;
11
:
589
598
.
28. 
Ucar
Y,
Brantley
WA,
Bhattiprolu
SN,
Johnston
WM,
McGlumphy
EA.
Characterization of cast-to implant components from five manufacturers
.
J Prosthet Dent
.
2009
;
102
:
216
223
.
29. 
Walton
TR.
The up-to-14-year survival and complication burden of 256 TiUnite implants supporting one-piece cast abutment/metal-ceramic implant-supported single crowns
.
Int J Oral Maxillofac Implants
.
2016
;
31
:
1349
1358
.
30. 
Jung
RE,
Zembic
A,
Pjetursson
BE,
Zwahlen
M,
Thoma
DS.
Systematic review of the survival rate and the incidence of biological, technical and aesthetic complications on singele crowns on implants reported in longitudinal studies with a mean follow-up of 5 years
.
Clin Oral Implants Res
.
2012
;
23
(suppl 6)
:
2
21
31. 
Cortada
M,
Giner
L,
Costa
S,
Gil
F,
Rodriguez
D,
Planell
J.
Metallic ion release in artificial saliva of titanium oral implants coupled with different metal superstructures
.
Biomed Mater Eng
.
1997
;
7
:
213
220
.
32. 
Cortada
M,
Giner
L,
Costa
S,
Gil
F,
Rodriguez
D,
Planell
J.
Galvanic corrosion behavior of titanium implants coupled to dental alloys
.
J Mater Sci Mater Med
.
2000
;
11
:
287
293
.
33. 
DeLong
R,
Douglas
W.
Development of an artificial oral environment for the testing of dental restoratives: bi-axial force and movement control
.
J Dent Res
.
1983
;
62
:
32
36
.
34. 
Alrabeah
GO,
Knowles
JC,
Petridis
H.
The effect of platform switching on the levels of metal ion release from different implant–abutment couples
.
Int J Oral Sci
.
2016
;
8
:
117
125
.
35. 
International Standards Organization.
ISO 10271:2011. Dentistry—Corrosion Test Methods for Metallic Materials. Geneva, Switzerland: International Standards Organization; 2011.
36. 
International.
Standards Oganisation. ISO 14801:2007(en). Dentistry—Implants—Dynamic Fatigue Test for Endosseous Dental Implants. Geneva, Switzerland: International Standards Organization; 2007.
37. 
Att
W,
Kurun
S,
Gerds
T,
Strub
JR.
Fracture resistance of single-tooth implant-supported all-ceramic restorations: an in vitro study
.
J Prosthet Dent
.
2006
;
95
:
111
116
.
38. 
Seloto
CB,
Strazzi-Sahyon
HB,
dos Santos
PH,
Assunçäo
WG.
Effectiveness of sealing gel on vertical misfit at the implant-abutment interface and preload maintenance of screw-retained implant-supported prostheses
.
Int J Oral Maxillofac Implants
.
2020
;
35
:
479
484
.
39. 
Park
C-I,
Choe
H-C,
Chung
C-H.
Effects of surface coating on the screw loosening of dental abutment screws
.
Metals Materials Int
.
2004
;
10
:
549
553
.
40. 
Byrne
D,
Jacobs
S,
O'Connell
B,
Houston
F,
Claffey
N.
Preloads generated with repeated tightening in three types of screws used in dental implant assemblies
.
J Prosthet Dent
.
2006
;
15
:
164
171
.