Electrochemical Disinfection of Dental Implants Experimentally Contaminated with Microorganisms as a Model for Periimplantitis

Despite several methods having been described for disinfecting implants affected by periimplantitis, none of these are universally effective and may even alter surfaces and mechanical properties of implants. Boron-doped diamond (BDD) electrodes were fabricated from niobium wires and assembled as a single instrument for implant cleaning. Chemo-mechanical debridement and air abrasion were used as control methods. Different mono-species biofilms, formed by bacteria and yeasts, were allowed to develop in rich medium at 37 °C for three days. In addition, natural multi-species biofilms were treated. Implants were placed in silicone, polyurethane foam and bovine ribs for simulating different clinical conditions. Following treatment, the implants were rolled on blood agar plates, which were subsequently incubated at 37 °C and microbial growth was analyzed. Complete electrochemical disinfection of implant surfaces was achieved with a maximum treatment time of 20 min for Candida albicans, Candida dubliniensis, Enterococcus faecalis, Roseomonas mucosa, Staphylococcus epidermidis and Streptococcus sanguinis, while in case of spore-forming Bacillus pumilus and Bacillus subtilis, a number of colonies appeared after BDD electrode treatment indicating an incomplete disinfection. Independent of the species tested, complete disinfection was never achieved when conventional techniques were used. During treatment with BDD electrodes, only minor changes in temperature and pH value were observed. The instrument used here requires optimization so that higher charge quantities can be applied in shorter treatment times.


Introduction
While a consistent definition of periimplantitis [1] as well as diagnostic criteria defining this disease are still missing [2,3], periimplantitis has received tremendous attention during the past years [3]. In addition, considerable heterogeneity exists among reports on the prevalence of periimplantitis [3]. A systematic review and meta-analysis conducted by Dreyer and coworkers found prevalence on implant level between 1.1% and 85.0% [2], while Rakic and coworkers found prevalence of 12.8% at

Treatment of Experimentally Contaminated Implants as Periimplantitis Model
Forty-five commercially available dental implants with a medium rough surface (Straumann Bone Level Tapered 4.1 x 12 mm RC; REF: 021.5512; LOT: RP027) were exposed to different microbes for three days at 37 °C in rich medium [Brain Heart Infusion (BHI); Oxoid, Wesel, Germany] to allow biofilm formation on external and internal surfaces (Note: Implants were reused following sterilization in an autoclave in order to increase sample size; please cf. Table S1). The microorganisms applied included yeasts (Candida albicans, Candida dubliniensis), Gram-negative (Roseomonas mucosa) and Gram-positive bacteria (Enterococcus faecalis, Staphylococcus epidermidis, Streptococcus sanguinis), including spore-forming bacteria (Bacillus pumilus, Bacillus subtilis) ( Table 1). The microorganisms were chosen based on their robustness and occurrence in cases of infected root canals [31,32] as well as in cases of peri-implantitis [33,34]. Often found members of the genera Prevotella and Treponema

Treatment of Experimentally Contaminated Implants as Periimplantitis Model
Forty-five commercially available dental implants with a medium rough surface (Straumann Bone Level Tapered 4.1 x 12 mm RC; REF: 021.5512; LOT: RP027) were exposed to different microbes for three days at 37 °C in rich medium [Brain Heart Infusion (BHI); Oxoid, Wesel, Germany] to allow biofilm formation on external and internal surfaces (Note: Implants were reused following sterilization in an autoclave in order to increase sample size; please cf. Table S1). The microorganisms applied included yeasts (Candida albicans, Candida dubliniensis), Gram-negative (Roseomonas mucosa) and Gram-positive bacteria (Enterococcus faecalis, Staphylococcus epidermidis, Streptococcus sanguinis), including spore-forming bacteria (Bacillus pumilus, Bacillus subtilis) ( Table 1). The microorganisms were chosen based on their robustness and occurrence in cases of infected root canals [31,32] as well

Treatment of Experimentally Contaminated Implants as Periimplantitis Model
Forty-five commercially available dental implants with a medium rough surface (Straumann Bone Level Tapered 4.1 x 12 mm RC; REF: 021.5512; LOT: RP027) were exposed to different microbes for three days at 37 • C in rich medium [Brain Heart Infusion (BHI); Oxoid, Wesel, Germany] to allow biofilm formation on external and internal surfaces (Note: Implants were reused following sterilization in an autoclave in order to increase sample size; please cf. Table S1). The microorganisms applied included yeasts (Candida albicans, Candida dubliniensis), Gram-negative (Roseomonas mucosa) and Gram-positive bacteria (Enterococcus faecalis, Staphylococcus epidermidis, Streptococcus sanguinis), including spore-forming bacteria (Bacillus pumilus, Bacillus subtilis) ( Table 1). The microorganisms were chosen based on their robustness and occurrence in cases of infected root canals [31,32] as well as in cases of peri-implantitis [33,34]. Often found members of the genera Prevotella and Treponema were not included in the study, since these anaerobic and microaerophilic bacteria are highly oxygen-sensitive and less resistant against reactive oxygen species and even atmospheric oxygen concentrations. For control of biofilm formation, staining and quantitative analysis was carried out as described previously [48]. Isolate from infected root canal [50] The implants were either placed in (i) elastic silicone, stable against temperature, acid, base and oxidants and bacterial colonization (Bindulin, Fürth, Germany), mimicking periimplant soft tissue, polyurethane foam blocks (Cellular Rigid polyurethane foam 20pcf, Sawbones Europe AB, Malmö, Sweden) mimicking type IV alveolar bone ( Figure 3) according to the Lekholm and Zarb classification [51] or in bovine ribs. In the latter cases, osteotomies were created in preformed saucer-shaped defects applying the regular surgical protocol [39]. The defects simulated circular bone resorption under maintenance of the buccal and oral compacta resembling class Ie defects according to a clinical classification system [52]. These specimens were placed in containers filled with phosphate-buffered saline (137 mM NaCl, 2.7 mM KCl, 10 mM Na 2 HPO 4 , 2 mM KH 2 PO 4 , pH 7.4) and treated applying the following methods: (i) Mechanical debridement (stainless steel curettes (EXD11/12, HuFriedy, Chicago, IL, USA), polishing of accessible implant surfaces) and irrigation with chlorhexidine (Chlorhexamed FORTE ethanol-free 0.2%, GlaxoSmithKline Consumer Healthcare GmbH & Co. KG, Munich, Germany) for a total of 5 min. (ii) Air abrasion (AIRFLOW PLUS, EMS ElectroMedicalSystems GmbH, Munich, Germany) and irrigation with chlorhexidine (Chlorhexamed FORTE ethanol-free 0.2%, GlaxoSmithKline Consumer Healthcare GmbH & Co. KG) for a total of 5 min. (iii) Electrochemical disinfection using the electrode configuration described above applying different treatment times. For every species investigated, at least 3 biological replicates were tested for each treatment procedure.
After cleaning and disinfection, the implants were rolled five to seven times on Columbia Blood Agar plates (Oxoid, Wesel, Germany), which were subsequently incubated at 37 • C for one day. Bacterial growth was monitored and rated using an evaluation scheme adapted from a monitoring scheme of catheter infections [53]. To this end, each lane of the roll-out was rated from 0 (no growth) to 3 (strong growth) using a master sample as reference ( Figure 4). All experiments were carried out in independent replicates (n = 3 biological replicates).  . Evaluation scheme for microbial growth. Treated implants were rolled on Columbia Blood Agar plates and incubated at 37 °C. Growth of the individual roll-outs was scored from 0 = no bacterial growth, 1 = minor growth, 2 = major growth, 3 = strong growth.

Statistical Analysis
The ratings obtained for mechanical debridement, air abrasion and 5 min BDD electrode treatment of implants (n = 3) covered by E. faecalis, C. dubliniensis and multi-species biofilm were subject to comparative statistical analysis. To this end, the sums of the individual ratings were used. In view of the small sample sizes, the integer valued results cannot be assumed to be normally distributed. Therefore, the parameter-free Kruskal Wallis rank sum test (R, version 3.6.2, The R Foundation for Statistical Computing, Vienna, Austria; www.R-project.org) was applied to compare the different treatments. The level of significance was set at α = 0.05.

Biofilm Formation on Experimetally Contaminated Implants
It was the aim of this study, to investigate the elimination of biofilms on implants using BDD electrodes as new electrochemical treatment method for periimplantitis. As a prerequisite of our  . Evaluation scheme for microbial growth. Treated implants were rolled on Columbia Blood Agar plates and incubated at 37 °C. Growth of the individual roll-outs was scored from 0 = no bacterial growth, 1 = minor growth, 2 = major growth, 3 = strong growth.

Statistical Analysis
The ratings obtained for mechanical debridement, air abrasion and 5 min BDD electrode treatment of implants (n = 3) covered by E. faecalis, C. dubliniensis and multi-species biofilm were subject to comparative statistical analysis. To this end, the sums of the individual ratings were used. In view of the small sample sizes, the integer valued results cannot be assumed to be normally distributed. Therefore, the parameter-free Kruskal Wallis rank sum test (R, version 3.6.2, The R Foundation for Statistical Computing, Vienna, Austria; www.R-project.org) was applied to compare the different treatments. The level of significance was set at α = 0.05.

Biofilm Formation on Experimetally Contaminated Implants
It was the aim of this study, to investigate the elimination of biofilms on implants using BDD electrodes as new electrochemical treatment method for periimplantitis. As a prerequisite of our . Evaluation scheme for microbial growth. Treated implants were rolled on Columbia Blood Agar plates and incubated at 37 • C. Growth of the individual roll-outs was scored from 0 = no bacterial growth, 1 = minor growth, 2 = major growth, 3 = strong growth.

Statistical Analysis
The ratings obtained for mechanical debridement, air abrasion and 5 min BDD electrode treatment of implants (n = 3) covered by E. faecalis, C. dubliniensis and multi-species biofilm were subject to comparative statistical analysis. To this end, the sums of the individual ratings were used. In view of the small sample sizes, the integer valued results cannot be assumed to be normally distributed. Therefore, the parameter-free Kruskal Wallis rank sum test (R, version 3.6.2, The R Foundation for Statistical Computing, Vienna, Austria; www.R-project.org) was applied to compare the different treatments. The level of significance was set at α = 0.05.

Biofilm Formation on Experimetally Contaminated Implants
It was the aim of this study, to investigate the elimination of biofilms on implants using BDD electrodes as new electrochemical treatment method for periimplantitis. As a prerequisite of our experiments, implants were incubated in rich medium with distinct microbial species for several days. Subsequently, biofilm formation was tested by staining with crystal violet solution. In all cases, biofilm formation on implants was observed; however, in a species-specific amount. The strongest mono-species biofilm producers were B. subtilis and C. dubliniensis, moderate amounts of biofilm were produced by B. pumilus, R. mucosa and S. sanguinis, while the poorest colonization of implants tested here was observed for C. albicans, E. faecalis and S. epidermidis ( Figure 5). experiments, implants were incubated in rich medium with distinct microbial species for several days. Subsequently, biofilm formation was tested by staining with crystal violet solution. In all cases, biofilm formation on implants was observed; however, in a species-specific amount. The strongest mono-species biofilm producers were B. subtilis and C. dubliniensis, moderate amounts of biofilm were produced by B. pumilus, R. mucosa and S. sanguinis, while the poorest colonization of implants tested here was observed for C. albicans, E. faecalis and S. epidermidis ( Figure 5).

Removal of Biofilm from Implants Contaminated with C. dubliniensis
BDD application to implants placed in both, silicone and polyurethane foam for 10 min at constant 6 V was at least as effective as the control treatments i.e. mechanical debridement and air abrasion. Increasing the treatment times of BDD electrodes led to even better results ( Figure 6).

Removal of Biofilm from Implants Contaminated with C. dubliniensis
BDD application to implants placed in both, silicone and polyurethane foam for 10 min at constant 6 V was at least as effective as the control treatments i.e. mechanical debridement and air abrasion. Increasing the treatment times of BDD electrodes led to even better results ( Figure 6).

Removal of Biofilm from Implants Contaminated with E. faecalis
Mechanical debridement and air abrasion were not suitable for complete elimination of E. faecalis biofilm and disinfection of the implants (Figure 7). In contrast, using BDD electrode treatment, full removal of biofilm and complete disinfection of implants was achieved within 5 to 10 min depending on the model system used (Figure 7).

Removal of Biofilm from Implants Contaminated with E. faecalis
Mechanical debridement and air abrasion were not suitable for complete elimination of E. faecalis biofilm and disinfection of the implants (Figure 7). In contrast, using BDD electrode treatment, full removal of biofilm and complete disinfection of implants was achieved within 5 to 10 min depending on the model system used (Figure 7).

Removal of Biofilm from Implants Contaminated with E. faecalis
Mechanical debridement and air abrasion were not suitable for complete elimination of E. faecalis biofilm and disinfection of the implants (Figure 7). In contrast, using BDD electrode treatment, full removal of biofilm and complete disinfection of implants was achieved within 5 to 10 min depending on the model system used (Figure 7).

Removal of Multispecies Biofilms from Implants
When bovine ribs were used to insert implants, a multi-species colonization was observed, despite the fact that only E. faecalis was used for pre-incubation and colonization. Obviously, the tested bone material was already strongly contaminated with a number of different microorganisms and, consequently, a natural multi-species biofilm developed in the prepared bone. Treatment of these implants showed an inferior disinfection success. This result may be explained by the putative presence of spore-forming bacteria in the uncharacterized natural multi-species biofilm and the fact that bone debris, which was not removed prior to rolling the specimens on the blood agar plates, was sticking on the implant surface. However, compared to mechanical debridement and air abrasion, BDD treatment still performed best (Figure 8).
Agar plates after different treatment of implants infected with E. faecalis and placed in silicone. Lower panel: Quantitative comparison of growth on implants inserted in silicone (blue bars) and polyurethane foam (orange bars) depending on treatment protocols. (A) Mechanical debridement, (B) air abrasion, (C) BDD treatment for 0 min, (D) BDD treatment for 5 min, (E) BDD treatment for 10 min, (F) BDD treatment for 15 min (n = 3 for each methods and time point). Growth of implant rollouts was rated from 0 (no growth) to 3 (strong growth). Columns represent the mean of 3 independent biological replicated ± standard deviation (SD).

Removal of Multispecies Biofilms from Implants
When bovine ribs were used to insert implants, a multi-species colonization was observed, despite the fact that only E. faecalis was used for pre-incubation and colonization. Obviously, the tested bone material was already strongly contaminated with a number of different microorganisms and, consequently, a natural multi-species biofilm developed in the prepared bone. Treatment of these implants showed an inferior disinfection success. This result may be explained by the putative presence of spore-forming bacteria in the uncharacterized natural multi-species biofilm and the fact that bone debris, which was not removed prior to rolling the specimens on the blood agar plates, was sticking on the implant surface. However, compared to mechanical debridement and air abrasion, BDD treatment still performed best (Figure 8). . Growth of implant roll-outs was rated from 0 (no growth) to 3 (strong growth). Columns represent the mean of 3 independent biological replicated ± standard deviation (SD).

Statistical Comparison Between Treatment Methods
The mean values of the ratings as well as the results of the Kruskal Wallis rank sum tests for comparing the three different treatment modalities are given in Table 2. In no instance, a statistically significant difference could be observed between the disinfection techniques applied. . Growth of implant roll-outs was rated from 0 (no growth) to 3 (strong growth). Columns represent the mean of 3 independent biological replicated ± standard deviation (SD).

Statistical Comparison Between Treatment Methods
The mean values of the ratings as well as the results of the Kruskal Wallis rank sum tests for comparing the three different treatment modalities are given in Table 2. In no instance, a statistically significant difference could be observed between the disinfection techniques applied. Table 2. Mean values and standard deviations of ratings recorded for implants contaminated with E. faecalis, C. dubliniensis (implants placed in silicone or polyurethane foam) and multispecies biofilm (implants placed in bovine rib) following 5 min of treatment with mechanical debridement, air abrasion and BDD electrode (n = 3 per microorganism and treatment method). Comparative statistics did not reveal any significant differences among the three modalities (p > 0.05).

Time-Dependent Removal of Biofilm from Implants Contaminated with Different Microorganisms
In addition to the microorganisms mentioned above and multi-species biofilm, the effect of BDD electrode treatment on mono-species biofilms formed by a number of different bacteria and yeasts was tested. The fastest elimination (within 10 min) was achieved for E. faecalis, R. mucosa, S. sanguinis and C. dubliniensis. A slightly longer treatment time of 20 min was necessary for the disinfection of implants colonized by C. albicans and S. epidermidis. In case of spore-forming B. pumilus and B. subtilis, a significant reduction of growth but no complete disinfection was reached under the experimental conditions applied due to the formation of highly resistant spores (6 V, 5-22 mA) ( Table 3). Table 3. Time for disinfection of implants placed in silicone depending on biofilm-forming species. All experiments were carried out in three independent biological replicates (n = 3) and the longest necessary treatment time is given. Implants were pre-incubated with the respective microorganisms to allow biofilm formation for three to five days.

Species
Maximum Time

B. pumilus
Incomplete disinfection within 60 min B. subtilis Incomplete disinfection within 60 min E. faecalis 10 min R. mucosa 10 min S. sanguinis 10 min S. epidermidis 20 min Yeasts C. albicans 20 min C. dubliniensis 10 min Mixture of microorganisms Multi-species natural biofilm developed on bovine ribs * Incomplete disinfection within 15 min * Implants were pre-incubated with E. faecalis, but bacteria were overgrown by a mixture of microorganisms already present on the cadaver bone.

Influence of BDD Electrode Treatment on Temperature and pH
Successful treatment of periimplantitis must not only rely on a reproducible disinfection protocol, but also on the absence of negative side effects. Therefore, changes in physico-chemical parameters depending on BDD treatment were tested. After treatment times of 25 min, changes in temperature in the range of 2 • C and changes in pH of 1 unit were recorded in a controlled setting with a reaction tube filled with 4 ml of phosphate-buffered saline and the respective probes mounted in the direct surrounding of the electrodes (Figure 9). Successful treatment of periimplantitis must not only rely on a reproducible disinfection protocol, but also on the absence of negative side effects. Therefore, changes in physico-chemical parameters depending on BDD treatment were tested. After treatment times of 25 min, changes in temperature in the range of 2 °C and changes in pH of 1 unit were recorded in a controlled setting with a reaction tube filled with 4 ml of phosphate-buffered saline and the respective probes mounted in the direct surrounding of the electrodes (Figure 9).

Discussion
Despite using a simplistic prototype, it was shown that electrochemical disinfection of contaminated implant surfaces was possible when BDD electrodes based on niobium wires were used. In contrast to mechanical debridement damaging the implant surface and air powder abrasion leaving powder remnants on the implant surface, no alterations of the implant surfaces were identified with the use of BDD electrodes [54]. As expected, the different microbes tested in this study showed varying levels of sensitivity and hence required varying amounts of treatment time until complete disinfection was achieved. The worst performer in this respect were, besides an uncharacterized multi-species biofilm, B. pumilus and B. subtilis, both forming highly resistant spores under the experimental conditions applied, while E. faecalis, R. mucosa and S. sanguinis were safely

Discussion
Despite using a simplistic prototype, it was shown that electrochemical disinfection of contaminated implant surfaces was possible when BDD electrodes based on niobium wires were used. In contrast to mechanical debridement damaging the implant surface and air powder abrasion leaving powder remnants on the implant surface, no alterations of the implant surfaces were identified with the use of BDD electrodes [54]. As expected, the different microbes tested in this study showed varying levels of sensitivity and hence required varying amounts of treatment time until complete disinfection was achieved. The worst performer in this respect were, besides an uncharacterized multi-species biofilm, B. pumilus and B. subtilis, both forming highly resistant spores under the experimental conditions applied, while E. faecalis, R. mucosa and S. sanguinis were safely eliminated with a maximum treatment time of 10 min. With the non-optimized setup used here, these treatment times would be too long for clinical application. The critical aspect, however, is not the treatment time but the charge quantity applied. Future developments will hence include modified electrodes with increased surfaces.
Minor temperature and pH-value changes were observed after applying the electrodes for approximately 25 min and hence would not be expected in a clinical setting where clearly shorter treatment times are needed. Even if these changes would occur, negative effects on the patient cannot be expected.
It had been anticipated that disinfection would become more difficult when implants are placed in bone surrogate materials or cadaver bone as compared to having full access to all critical surfaces. While not expressible as a quantitative effect, this correlation was observed for the control treatments of chemo-mechanical debridement and air abrasion, respectively. In contrast, when electrochemical disinfection was applied, restricted access to implant surfaces obviously had less effect.
Recently, an electro-chemical approach for implant debridement and periimplantitis treatment was introduced utilizing the implant itself as an electrode (cathode) while a wire, not contacting the infected surface, was used as another electrode (anode) [44]. By this approach, not using BDD-coated electrodes, biofilm appears to be primarily removed by bubbles of electrochemically generated hydrogen at the cathodic implant surface. This theoretically bears the risk of spreading the biofilm into the surrounding tissue and may also generate corrosion problems on the implant surface by hindering the self-passivation of titanium after damaging the native oxide layer by a mechanical load. Furthermore, the hydrogen uptake of the titanium implant can lower the mechanical strength of the material by hydrogen embrittlement [55].
Given that the experiment at hand constituted the first application of BDD electrodes for biofilm removal from dental implant surfaces, a number of limitations have to be taken into account when interpreting the findings presented. A wide variety of bacteria and yeasts have been shown to be present in dental biofilms, while in the current experiment mostly mono-species biofilms were considered. Although different bacteria may form a synergistic biofilm, their resistance to disinfecting measures is not influenced by other species. However, with the spore-forming species used here, a worst-case scenario has already been tested. It has been shown that not only the clinical situation per se but also the defect morphology impacts the result of periimplantitis treatment [52]. As such, the model situations chosen here clearly present simplifications of reality, which were needed due to the prototypical stage of the BDD electrode setup. Due to the exploratory nature of this pilot investigation, comparative statistical analysis among the treatment modalities applied was limited to a subset of experiments where implants had been treated for exactly 5 min using all three disinfection methods. Future studies are under way with much greater sample size aimed at quantitatively comparing different treatment modalities.
Taking into account that numerous developmental steps including preclinical and clinical studies will be required prior to clinical application of an instrument based on a BDD electrode array, a probe-like instrument with a permeable cover can already be envisaged. Not requiring superstructure removal as well as the universal applicability also in periodontal and endodontic treatment would be major advantages of such an instrument.

Patents
Stefan Rosiwal, Andreas Burkovski, Maximilian Göltz and Matthias Karl have a filed a patent for the disinfection method described in this report.