Adjunctive benefits of systemic metronidazole on non-surgical treatment of peri-implantitis. A randomized placebo-controlled clinical trial

ORIGINAL ARTICLE Adjunctive benefits of systemic metronidazole on non-surgical treatment of peri-implantitis. A randomized placebo-controlled clinical trial Carlota Blanco 1,2 | Alex Pico 1,2 | Jose Dopico 1,2 | Pilar Gándara 2,3 | Juan Blanco 1,2 | Antonio Liñares 1,2 1 Periodontology Unit, Faculty of Odontology, University of Santiago de Compostela, Santiago de Compostela, Spain 2 Medical-Surgical Dentistry Research Group, Health Research Institute of Santiago de Compostela, Santiago de Compostela, Spain 3 Oral Surgery Unit, Faculty of Odontology, University of Santiago de Compostela, Santiago de Compostela, Spain Correspondence Jose Dopico, Periodontology Unit, Faculty of Odontology, University of Santiago de Compostela, Rúa Entrerríos SN, 15782, Santiago de Compostela, Spain. Email: [email protected] Antonio Linares, Periodontology Unit, Faculty of Odontology, University of Santiago de Compostela, Rúa Entrerríos SN, 15782, Santiago de Compostela, Spain Email: [email protected]s Funding information This study was partially supported by grants from Osteology Foundation (16-070) and from the Spanish Society of Periodontology SEPA (research scholarship 2018). The funders had no role in the study design, data collection and analysis, decision to publish or preparation of the manuscript. SUNSTAR IBERIA, S.L.U. sponsored the perio-implant diagnostic test and analysis. Abstract Aim: To study clinical, radiographic, and microbiological outcomes after non-surgical therapy of peri-implantitis with or without adjunctive systemic metronidazole. Materials and Methods: A randomized placebo-controlled clinical trial was carried out in 32 subjects (62 implants) diagnosed with peri-implantitis. Implants received a mechanical non-surgical debridement session and systemic metronidazole or placebo. Clinical, radiographic, and microbiological outcomes were evaluated at baseline, 3, 6, and 12 months. Results: After 12 months, the test treatment resulted in significantly greater PPD reduction (2.53 vs. 1.02 mm) and CAL gain (2.14 vs. 0.53 mm) (pvalue <.05) in comparison with placebo. The test treatment also resulted in additional radiographic bone gain (2.33 vs. 1.13 mm) compared with placebo (pvalue <.05). There was a significantly greater decrease in Porphyromonas gingivalis,Tannerella forsythia,andCampylobacter rectus counts compared with the control group (pvalue <.05). At the end of follow-up, 56.3% of patients met the success criteria in the test group and 25% in the control group. Conclusions: The use of systemic metronidazole as an adjunct to non-surgical treatment of peri-implantitis resulted in significant additional improvements in clinical, radiographic, and microbiological parameters after 12 months of follow-up. This study is registered in ClinicalTrials.gov (NCT03564301). KEYWORDS antibiotic, metronidazole, non-surgical, peri-implantitis, treatment Clinical Relevance Scientific rationale for study: Recent case series studies have shown improvements in clinical and radiographic parameters with non-surgical treatment and systemic adjunctive antibiotics in the treatment of peri-implantitis. However, RCTs to confirm these findings are lacking. Principal findings: As initial therapy, systemic metronidazole as an adjunct to non-surgical therapy in peri-implantitis lesions achieved significant additional improvements in terms of clinical, Carlota Blanco, Alex Pico, and Jose Dopico contributed equally to this manuscript. Received: 29 July 2021 Revised: 21 September 2021 Accepted: 1 October 2021 DOI: 10.1111/jcpe.13564 This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes. © 2021 The Authors. Journal of Clinical Periodontology published by John Wiley & Sons Ltd. J Clin Periodontol. 2022;49:15–27. wileyonlinelibrary.com/journal/jcpe 15 radiographic, and microbiological outcomes compared with placebo. Despite these promising results, only 53.6% of the patients treated with adjunctive antibiotic met the success criteria. Practical implications: In light of the results of this RCT, modifications in treatment protocols for peri-implantitis could be implemented. 1|INTRODUCTION Peri-implantitis is defined as a “pathological condition associated with plaque that occurs in the tissues surrounding dental implants, characterized by inflammation of the peri-implant mucosa and the subsequent progressive loss of bone”(Berglundh et al., 2018). Although survival of dental implants is high (Howe et al., 2019), the prevalence of peri-implant diseases is also very significant. Using the case definition criteria proposed by Sanz and Chapple (Sanz & Chapple, 2012), a recent systemic review showed a prevalence of 18.5% at patient level and 12.8% at implant level (Dreyer et al., 2018). A study in a Spanish population, with data from 474 implants in 275 patients, showed an implant-level peri-implantitis prevalence of 20% (bone loss ≥2 mm) and 24% at patient level (Rodrigo et al., 2018). Considering the case definitions proposed at the 2017 World Workshop on the Classification of Periodontal and Peri-Implant Diseases and Conditions (Renvert et al., 2018) (bone loss ≥3 mm), 14% of patients and 11% of implants presented peri-implantitis in the aforementioned Spanish study (Rodrigo et al., 2018). It is widely accepted that poor oral hygiene, history of periodontitis, and tobacco smoking are well-documented risk factors for periimplantitis. Other factors such as diabetes, alcohol consumption, and genetic traits may also increase the risk of peri-implantitis (Schwarz et al., 2018). Peri-implantitis and periodontitis have common clinical features. However, marked differences exist. It has been shown that lesions of both pathologies show large amounts of inflammatory infiltrate, spreading to a more apical level in peri-implantitis (Berglundh et al., 2011; Carcuac et al., 2013). Moreover, the microbiota associated with peri-implantitis is characterized by a mixed anaerobic infection, with a composition similar to the subgingival microbiota of periodontitis lesions, but greater complexity, composed of predominantly non-cultivable gram-negative species (Lafaurie et al., 2017). There is no consensus on standards of care or clinical protocols for the management of peri-implant diseases (Renvert et al., 2013). In addition, different available therapeutic approaches have shown moderately successful outcomes (Renvert & Polyzois, 2018). It has been shown that non-surgical treatment is effective for peri-implant mucositis (Suárez-L opez Del Amo et al., 2016). In contrast, in periimplantitis, non-surgical therapy alone is not effective in a significant proportion of cases (Renvert et al., 2008). Despite the fact that the clinical endpoint for success may not be achieved in part of the cases, especially those with more advanced pathology, published case series studies have shown promising additional benefits when using systemic antibiotics as an adjunct to non-surgical therapy (Mombelli & Lang, 1992; Estefanía-Fresco et al., 2019; Liñares et al., 2019; Nart et al., 2020). Non-surgical treatment of advanced peri-implantitis lesions with intra-bony defects in combination with systemic metronidazole resulted in reduction in probing depth of 4.6 and 2.6 mm in radiographic bone gain after 54 months of follow-up (Liñares et al., 2019). To date, no randomized clinical trials have been conducted to assess the effects of systemic metronidazole as an adjunct to nonsurgical treatment of peri-implantitis. The main aim of this trial is to assess clinical, radiographic, and microbiological outcomes after nonsurgical treatment of peri-implant lesions in combination with adjunctive systemic metronidazole. 2|MATERIALS AND METHODS 2.1 |Experimental design This study was a triple-blind, randomized, placebo-controlled, clinical trial with 12 months follow-up. The study was carried out in accordance with the principles described in the Declaration of Helsinki on experimentation with human beings (Order SCO/256/2007), and was approved by the Ethical Committee of Galicia (2017/508). It is registered in ClinicalTrials.gov (NCT03564301). 2.2 |Patient screening Subjects eligible for the study were identified from patients referred to the Periodontology Unit of the University of Santiago de Compostela between January 2018 and July 2019. The study included subjects with (I) ≥18 years of age, (II) presence of at least one implant diagnosed with peri-implantitis defined as presence of bleeding on probing and/or suppuration, probing depth ≥6 mm and ≥3mm of detectable bone loss after initial re-modelling (Berglundh et al., 2018), and (III) absence of implant mobility. Subjects were excluded from the study if they (I) had history of allergy to metronidazole, (II) had received treatment with bisphosphonates, (III) were pregnant or breast-feeding, (IV) had received antibiotic treatment in the previous 3 months, or (V) had systemic conditions that contraindicated treatment. 2.3 |Sample size calculation Sample size calculation was made for the primary outcome variable, probing pocket depth (PPD). Due to the absence of previous RCTs, 16 BLANCO ET AL. the 6 months visit data of the first 25 patients recruited for the study were used to estimate sample size. Standard deviation (SD) was 1.703 for the test group and 1.803 for the control group. The mean difference was 1.214 mm. A sample of 32 patients, 16 per group, was estimated to provide 80% power with an αof .05, accounting for 10% attrition. 2.4 |Randomization and allocation concealment Subject numbers were assigned in ascending order at the enrolment visit. Subjects were randomly assigned by a computer-generated table to receive one of the two treatments. A balanced random permuted block approach (4-unit block size) was used to prepare the randomization tables in order to avoid unequal balance between treatments. Restricted randomization (minimization) was performed, stratifying for smoking habits (yes or no) and history of periodontitis (yes or no). A randomization table was sent to the pharmacy, to prepare the medication for the first 25 subjects (12 tests and 13 controls). Twenty-five plastic bags containing metronidazole or placebo were sent back to the study coordinator (J.B.), who was not involved in the examination of patients, provision of the clinical intervention or data analysis, and was the only person who had access to them. Consequently, the study coordinator completed the treatment assignment and matched the code of the plastic bags with the subject number. The research nurse provided each subject with their corresponding plastic bag. Once 6 months data were available, sample size calculation was made, and medication or placebo was prepared in the same way as for the first 25 patients for the remaining seven subjects (four tests, three controls) and sent to the study coordinator (J.B.), who again was the only person who had access to them. The research nurse provided each remaining participant with their allocated plastic bag. Allocation concealment was only broken by the study coordinator (J.B.) for sample size calculation purposes and was not broken again until all data were collected and analysed. Thus, the treatment group was concealed to the patient, the clinical examiner, the treatment provider, and the statistician. 2.5 |Examiner calibration One examiner (J.D.) carried out all clinical measurements. The examiner was calibrated in measurement of PPD, CAL, and REC (Dopico et al., 2016). The measurement of radiographic bone levels was carried out by a different examiner (A.P.). The examiner was trained and calibrated in the measurement of radiographs using ImageJ software (ImageJ 1.47 V Wayne Rasband, National Institutes of Health) by another examiner (A.L.), who had been previously calibrated (Liñares et al., 2019) and represented the “gold standard”. Inter-examiner agreement was evaluated as the standard error of the mean difference of the measurements performed by the examiner (A.P.) and those performed by the “gold standard”examiner (A.L.) in 10 radiographs. The examiner showed 98.2% reproducibility (Liñares et al., 2019). 2.6 |Clinical parameters Clinical parameters were assessed using a UNC-15 periodontal probe by a calibrated examiner at six sites per implant. Data were collected at baseline, 3, 6, and 12 months. In case an implant-supported restoration did not allow adequate access to the implant, it was removed for clinical assessment at each appointment. Probing pocket depth (PPD) was calculated as the distance from the gingival peri-implant margin to the bottom of the peri-implant pocket (Mombelli & Lang, 1994). Recession (REC) was measured as the distance from the free marginal mucosa to the most apical portion of the crown. Clinical attachment level (CAL) was calculated as the sum of PPD and REC (distance from the most apical portion of the crown to the bottom of the periodontal pocket). Bleeding on probing (BoP) was assessed dichotomously in six sites per implant (Heitz-Mayfield, 2008). Full-mouth plaque score (FMPS) was calculated by assigning a binary score to each surface (1 for plaque present, 0 for absent) and calculating the percentage of total tooth/implant surfaces that revealed the presence of plaque detected by a periodontal probe (Tonetti & Claffey, 2005). Full-mouth bleeding score (FMBS) was recorded by assigning a binary score to each surface (1 for bleeding present, 0 for absent) and calculating the percentage of total tooth/implant surfaces that bled upon probing (Tonetti & Claffey, 2005). Mean values and confidence intervals for PPD, CAL, REC, deepest PPD per implant, BoP, FMPS, and FMBS were calculated. 2.7 |Radiographic parameters Intra-oral phosphor plate, long-cone parallel technique periapical radiographs were taken with the aid of custom-made holders silicone index (Griffiths et al., 1996) to evaluate peri-implant marginal radiographic bone levels at baseline, 3, 6, and 12 months. The same custom-made holders for each specific implant were used in all visits. Those radiographs were imported to a digital radiographic examination software (ImageJ 1.47 V Wayne Rasband, National Institutes of Health). All examinations were performed by an independent and calibrated examiner (A.P.) to the nearest 0.1 mm using IMAGE J software. The scale was set and calibrated by the height of the dental implant, which yielded a pixel/mm ratio. The implant shoulder was used as a fixed reference point. To assess linear changes at inter-proximal alveolar crestal bone height, the distance from the implant shoulder to the most coronal bone to implant contact (RBL) was determined both at the mesial and distal aspect of each implant and expressed as the mean of both measurements in millimetres. BLANCO ET AL.17 2.8 |Microbiological parameters Microbiological testing was performed in the deepest peri-implant pocket of each subject at baseline, 3, 6, and 12 months. A commercially available microbiological test was used (Guidor Perio-implant diagnostic test –Sunstar). Following elimination of the supragingival biofilm, samples were taken with two consecutive sterile medium paper points (Maillefer, Ballaigues, Switzerland). The paper points were kept in place for 10 s and transferred into one empty sterilized 1.5 ml Eppendorf tube. Subsequently, samples were sent for analysis by quantitative PCR (qPCR) for detecting and quantifying bacterial DNA. Standard curves for each bacterium were prepared by determining the Cq values of serial 10-fold dilutions of purified genomic DNA from each bacterium. Standard curves were constructed by plotting Cq values generated from qPCR against Porphyromonas gingivalis, Aggregatibacter actinomycetemcomitans,Tannerella forsythia,Fusobacterium nucleatum,andCampylobacter rectus (log CFU ml 1 ). The reduction of CFU counts of each group was calculated and the interaction of the microbiological status (presence or absence of studied bacterial species) with the treatment effect (test or control). 2.9 |Other parameters Age, gender, history of periodontitis, and smoking status were recorded. Subjects presenting with inter-dental CAL at ≥2 nonadjacent teeth were diagnosed with a history of periodontitis (Papapanou et al., 2018). Patients were classified as current smokers or former/never smokers. Data on the type of restoration (implant-supported single restoration, fixed partial denture, overdenture, or fixed complete denture), and retention (screw- or cement-retained) were also collected. Adverse events and compliance with the prescribed treatment were also recorded. 2.10 |Success criteria A case was considered successfully treated if PPD was reduced to 5 mm with no BoP or to <5 mm irrespective of BoP, and no further bone loss was detected between baseline and 1 year (Carcuac 40 subjects assessed for eligibility 32 RANDOMIZED 8 excluded Reasons: not meeting inclusion criteria (n = 7) refused to participate (n = 1) 16 test group: NST + metronidazole Follow up at: 16 control group: NST + placebo 3 months: 14 6 months: 16 12months:16 Follow up at: 3 months: 15 6 months: 16 12months:16 16 Analysed16 Analysed FIGURE 1 Flow chart for study patients 18 BLANCO ET AL. et al., 2016). This was evaluated both at patient and implant level. In cases where success criteria were not met upon completion of the study, further treatment was provided according to individual needs. 2.11 |Pre-treatment visits These visits were scheduled in order to help patients to achieve and maintain adequate self-performed plaque control. Supragingival debridement and elimination/reduction of plaque-retentive factors, including modification of existing restorations, were performed (de Tapia et al., 2019). Oral hygiene instructions were provided until an FMPS ≤20% was achieved (Gabay et al., 2015). 2.12 |Baseline examination Baseline examination took place on a separate appointment after pretreatment visits were completed. 2.13 |Intervention Both groups received one session of non-surgical debridement as described previously (Liñares et al., 2019) by an experienced periodontist (A.L.). Implant-supported restorations were, if possible, removed during treatment and relocated at the end of the treatment session. Before instrumentation, patients rinsed with 0.12% chlorhexidine digluconate (Perio-aid; Dentaid, Spain). Non-surgical treatment under local anaesthesia consisting of supra- and submucosal mechanical debridement using ultrasonic stainless steel scaling inserts (EMS ® , Electro Medical Systems) was performed in one treatment session with no time restrictions. Ultrasonic debridement was performed with simultaneous irrigation of 0.12% chlorhexidine digluconate (Perio-aid, Dentaid, Spain). After that, a stainless steel Columbia 4R/4L curette (LM Instruments Oy) was used to remove granulation tissue and minor mucosal curettage. Then, peri-implant pockets were irrigated with 0.12% chlorhexidine digluconate (Perioaid, Dentaid, Spain). Immediately after the treatment session, patients initiated a course of two tablets three times a day, for 7 days, consisting of metronidazole 250 mg (test) or placebo pills (control). 2.14 |Post-treatment visits Follow-up visits took place 1 week, 3, 6, and 12 months after the treatment session. In the post-treatment visits, oral hygiene advice and support were provided, early healing events were assessed and monitored, and any adverse events or additional medications were recorded. In the 1-week post-treatment visit, subjects were asked to return any medication not taken and/or the empty bottles. The number of pills not taken by the subject was recorded. 2.15 |Re-assessment visits Re-assessment visits were scheduled 3, 6, and 12 months after treatment. During these appointments, the examiner recorded any changes in medical history. Clinical measurements, radiographs, and microbiological samples were taken. At the end of the re-assessment appointment, a session of supra-gingival debridement was performed. Sites with residual PPD ≥5 mm also received ultrasonic submucosal debridement as described above. 2.16 |Statistical analysis IBM SPSS Statistics (version 24.0) was used to carry out statistical analyses. The Shapiro–Wilk test was applied to confirm if variables followed a normal distribution. A subject-level analysis was performed by computing a subject-level variable for each parameter. Numerical data were summarized as means and confidence intervals (CIs) for normally distributed continuous variables, categorical data were summarized as percentage. Means were calculated for PPD, REC, CAL, and RBL at patient, implant, and site level.MeansforPPD,CAL,REC,andBoPwerecalculatedatsite TABLE 1 Subject and clinical characteristics at baseline Parameter Test (n=16) Control (n=16) pvalue Smoking Never/former smoker 12/16 (75%) 11/16 (69%) .50 (χ2) Current smoker 4/16 (25%) 5/16 (31%) History of periodontitis Yes 9/16 (56%) 13/16 (81%) .12 (χ2) No 7/16 (44%) 3/16 (19%) Gender Male 9/16 (56%) 4/16 (25%) .74 (χ2) Female 7/16 (44%) 12/16 (75%) Age 58.31 (51.46–65.16) 60.74 (55.09–65.84) .71 (t-test) Systemic diseases Yes 3/16 (19%) 1/16 (6%) .1 (χ2) No 13/16 (81%) 15/16 (94%) Number of implants (included/total) 1.75 (1–4)/4.29 (1–16) 2.12 (1–5)/4.47 (1–12) .36 (t-test) BLANCO ET AL.19 level. Site-level data were divided into two subgroups according to PPD, in moderately deep (5–6mm) and deep (>6mm) sites to further explore the treatment effects within each of the subgroups. This later approach was also applied for another exploratory analysis on presence or absence of specific bacterial species in test and control groups. Independent samples t-test were performed for inter-group comparisons of numerical variables. Repeated measures ANOVA was used for intra-group comparisons of numerical variables. Chi-squared test was used for categorical variables. All tests were performed at a significance level of α=.05. 3|RESULTS 3.1 |Subject accountability Forty subjects were assessed for their eligibility before entering the study. Eight subjects were excluded: seven because they did not meet the inclusion criteria, while the other one was not willing to participate (Figure 1). Thus, 32 subjects (62 implants) were enrolled and randomly allocated to the test and control groups. All participants received the allocated intervention. The last treatment session took place in January 2020. Three patients, two from the test and one from the placebo group did not attend the 3-month visit because of reasons unrelated to the study (mobility restrictions related to COVID-19 pandemic). An intention-to-treat approach (ITT) was followed and all participants were included in the analysis (Figure 1). At the 3 months follow-up, the available data were analysed. No effort was made to input the missing data since the reason for not attending the follow-up appointment was unrelated to the study. Furthermore, data from later follow-up appointments was available. 3.2 |Subject characteristics at baseline Baseline characteristics of the 32 participants are displayed in Table 1. The mean age of the participants was 58.31 (51.46–65.16) and 60.74 (55.09–65.84) for the test and placebo groups, respectively. Twentyfive percentage of subjects in the test group and 31% in the placebo group were current smokers. Fifty-six percentage of patients had history of periodontitis in the test group and 81% in the control group. Nineteen percentage of patients in the test group and 6% in control group had systemic diseases (cardiovascular disease in all cases). None of these parameters showed statistically significant differences between groups (pvalue >.05). Regarding the type of restoration, in both groups, there were 10 implant-supported single restorations, 20 fixed partial dentures (2–4 units), three overdentures, and three fixed complete full-arch dentures. The majority of them were screw-retained (78%) versus cemented-retained (22%). All implants included had a moderately rough surface. TABLE 2 Mean clinical outcomes for baseline, 3, 6, and 12 months at subject level Test (n=16) Control (n=16) Parameter Baseline (n=16) 3 months (n=14) 6 months (n=16) 12 months (n=16) Baseline (n=16) 3 months (n=15) 6 months (n=16) 12 months (n=16) PPD (mm) 6.86 (5.78–7.94) 4.59 (3.99–5.19)* 4.42 (3.69–5.14)* 4.33 (3.64–5.01)* 5.89 (5.23–6.65) 5.04 (4.24–5.84)* 4.98 (4.18–5.77)* 4.87 (4.00–5.74) CAL (mm) 7.29 (6.20–8.40)** 5.36 (4.78–5.95)* 5.09 (4.32–5.86)* 5.16 (4.48–5.83)* 6.08 (5.42–6.74)** 5.79 (5.00–6.57) 5.56 (4.75–6.37) 5.54 (4.72–6.37) REC (mm) 0.44 (0.06–0.93) 0.78 (0.40–1.16) 0.67 (0.23–1.11) 0.83 (0.35–1.31) 0.19 (0.02–0.35) 0.75 (0.38–1.11)* 0.59 (0.28–0.90)* 0.68 (0.47–0.88)* ΔPPD (mm) 2.27 (1.17–3.37)** 2.44 (1.33–3.54)** 2.53 (1.37–3.69)** 0.85 (0.39–1.31)** 0.92 (0.46–1.37)** 1.02 (0.06–1.99)** ΔCAL (mm) 1.93 (0.92–2.94)** 2.21 (1.00–3.41)** 2.14 (0.97–3.30)** 0.29 (0.16–0.75)** 0.51 (0.09–0.94)** 0.53 (0.33–1.39)** FMPS (%) 20.92 (18.68–23.16) 28.31 (15.09–41.52) 27.38 (21.69–33.08) 30.77 (23.15 38,.39) 21.20 (19.41–22.99) 38.87 (27.18–50.55) 30.93 (19.87–41.99) 33.60 (23.89–43.31) FMBS (%) 39.85 (24.37–55.32) 17.77 (11.08–24.45)** 20.92 (14.47–27.38)** 20.77 (13.72–27.82) 50.47 (36.92–64.02) 34.40 (20.89–47.91)** 33.00 (23.20–42.80)** 29.47 (15.82–43.11) RBL (mm) 6.31 (5.39–7.22) 5.84 (4.93–5.76)* 4.66 (3.79–5.53)* 3.97 (3.16–4.79)* 5.54 (4.39–6.69) 5.22 (3.95–6.49) 4.72 (3.49–5.96)* 4.41 (3.36–5.46) ΔRBL (mm) 0.46 (0.18–0.75) 1.65 (0.98–2.31)** 2.33 (1.58–30.9)** 0.32 (0.01–0.65) 0.82 (0.34–1.29)** 1.13 (0.31–1.95)** Abbreviations: CAL, clinical attachment level; FMBS, full-mouth bleeding score; FMPS, full-mouth plaque score; PPD, probing pocket depth; RBL, radiographic bone level; REC, recession. *pvalue <.05 for intra-group comparisons; **pvalue <.05 for inter-group comparisons. 20 BLANCO ET AL. 3.3 |Adverse events and compliance At the 1-week follow-up visit, six subjects (38%) in the test group and five subjects (31%) in the control group reported adverse events. Two patients reported gastrointestinal disorder (one test/one control), four patients reported headache (two test/two control), two patients reported metallic taste (one test/one control), and two patients (two test/zero control) had oral tissue alterations (periimplant bleeding). All subjects returned the medication bottles. Fifteen subjects (94%) in the test group and 14 subjects (88%) in the placebo group completed the course of systemic medication as indicated. The reason for non-compliance in all cases was forgetting to take the medication. One subject in the control group missed 21 pills, and two subjects in the test group missed eight and six pills, respectively. An ITT approach was followed and no subjects were excluded from the analysis. No implants were lost during the study. 3.4 |Clinical outcomes 3.4.1 | Subject-level analysis Clinical outcomes at baseline, 3, 6, and 12 months are displayed in Table 2. At baseline, significant differences were observed in CAL between groups (7.29 mm test vs. 6.08 mm control) (pvalue <.05). In all follow-up visits, mean PPD (4.33 mm test vs. 4.87 mm control) and CAL (5.16 mm test vs. 5.54 mm control) were not significantly different between groups (pvalue >.05), but the amount of PPD reduction (2.53 mm test vs. 1.02 mm control) and CAL gain (2.14 mm test vs. 0.53 mm control) were significantly greater in the test group at the end of the study (pvalue <.05). Although a significant reduction can be observed in both groups, FMBS results were significantly different between groups at 3 (17.77% test vs. 34.40% control) and 6 months (20.92% test vs. 33.00% control) (pvalue <.05). Intra-group analysis showed that significant clinical improvements were present in both treatment groups. The test group achieved statistically significant improvements in PPD and CAL in all visits compared with baseline (pvalue <.05), while the clinical improvements in the control group only reached statistical significance at 3 and 6 months regarding PPD (pvalue <.05). REC in the test group increased, but the differences were not statistically significant (from 0.44 mm to 0.83 mm) (pvalue >.05). In contrast, the control group had a statistically significant increase in REC (from 0.19 mm to 0.68 mm) (pvalue <.05). FMPS remained stable throughout the duration of the study, and significant differences between groups were not found (pvalue >.05). 3.4.2 | Implant-level analysis Implant-level clinical outcomes at baseline, 3, 6, and 12 months are displayed in Table 3. At baseline, there were no significant TABLE 3 Mean clinical outcomes for baseline, 3, 6, and 12 months at implant level Test (n=28) Control (n=34) Parameter Baseline (n=28) 3 months (n=26) 6 months (n=28) 12 months (n=28) Baseline (n=34) 3 months (n=33) 6 months (n=34) 12 months (n=34) maxPPD (mm) 8.41 (7.62–9.21) 5.87 (5.25–6.50)* , ** 5.62 (4.97–6.28)* , ** 5.12 (4.52–5.73)* , ** 8.36 (7.62–9.11) 7.15 (6.41–7.89)* , ** 6.94 (6.10–7.78)* , ** 6.85 (6.01–7.68)* , ** PPD (mm) 6.78 (5.95–7.62) 4.65 (4.18–5.13)* , ** 4.54 (3.97–5.11)* , ** 4.34 (3.79–4.94)* , ** 6.32 (5.70–6.95) 5.74 (5.09–6.38)* , ** 5.56 (4.85–6.23)* , ** 5.43 (4.81–6.06)* , ** CAL (mm) 7.24 (6.43–8.05) 5.55 (5.07–6.04)* 5.48 (4.90–6.06)* 5.51 (4.92–6.11)* 6.59 (5.97–7.02) 6.47 (5.81–7.13) 6.19 (5.50–6.88) 6.09 (5.47–6.70) REC (mm) 0.46 (0.16–0.76) 0.90 (0.60–1.21)* 0.94 (0.51–1.36)* 1.15 (0.65–1.65)* , ** 0.26 (0.12–0.40) 0.74 (0.47–1.00)* 0.62 (0.40–0.84)* 0.65 (0.46–0.84)* , ** ΔmaxPPD (mm) 2.54 (1.75–3.33)** 2.79 (2.06–3.53)** 3.29 (2.30–4.28)** 1.21 (0.65–1.78)** 1.42 (0.73–2.12)** 1.52 (0.75–2.28)** ΔPPD (mm) 2.13 (1.42–2.85)** 2.24 (1.49–3.00)** 2.42 (1.53–3.30)** 0.59 (0.13–1.04) 0.76 (0.33–1.20)** 0.89 (0.14–1.64)** ΔCAL (mm) 1.69 (1.06–2.32)** 1.76 (1.00–2.52)** 1.73 (1.01–2.45)** 0.11 (0.36–0.58)** 0.40 (0.00–0.80)** 0.50 (0.18–1.18)** RBL (mm) 6.31 (5.55–7.06) 5.91 (5.17–6.66)* 4.99 (4.24–5.73)* 4.16 (3.39–4.94)* 6.00 (5.19–6.81) 5.77 (4.92–6.62) 5.20 (4.33–6.06)* 5.05 (4.19–5.92)* ΔRBL (mm) 0.40 (0.20–0.59) 1.32 (0.8–1.84) 2.15 (1.51–2.78)** 0.24 (0.05–0.42) 0.81 (0.39–1.22) 0.95 (0.39–1.51)** Abbreviations: CAL, clinical attachment level; maxPPD, deepest probing pocket depth; PPD, probing pocket depth; RBL: radiographic bone level; REC, recession. *pvalue <.05 for intra-group comparisons; **pvalue <.05 for inter-group comparisons. BLANCO ET AL.21 TABLE 4 Mean clinical outcomes for baseline, 3, 6, and 12 months at sites level Test Control Parameter Baseline (n=87) 3 months (n=83) 6 months (n=87) 12 months (n=87) Baseline (n=83) 3 months (n=83) 6 months (n=83) 12 months (n=83) Deep sites (>6 mm) PPD (mm) 8.65 (8.30–9.01) 5.36 (5.01–5.71)* , ** 5.38 (5.02–5.75)* , ** 5.04 (4.64–5.43)* , ** 8.54 (8.14–8.94) 6.86 (6.37–7.34)* , ** 6.75 (6.19–7.30)* , ** 5.95 (5.51–6.39)* , ** CAL (mm) 8.95 (8.60–9.30) 6.39 (6.75–6.35)* , ** 6.19 (5.79–6.59)* , ** 6.29 (5.85–6.72)* 8.65 (8.24–9.06) 7.30 (6.83–7.77)* , ** 7.23 (6.69–7.76)* , ** 6.63 (6.19–.06)* REC (mm) 0.30 (0.11–0.49) 1.02 (0.79–1.26)* , ** 0.81 (0.56–1.05)* , ** 1.25 (0.89–1.61)* , ** 0.11 (0.03–0.18) 0.45 (0.25–0.64)* , ** 0.48 (0.29–0.66)* , ** 0.67 (0.47–0.88)* , ** ΔPPD (mm) 3.29 (2.88–3.69)** 3.27 (2.79–3.74)** 3.61 (3.07–4.16)** 1.69 (1.33–2.05)** 1.80 (1.40–2.19)** 2.59 (2.02–3.16)** ΔCAL (mm) 2.57 (2.16–2.98)** 2.76 (2.29–3.23)** 2.66 (2.18–3.15) 1.35 (0.98–1.72)** 1.42 (1.02–1.82)** 2.02 (1.45–2.60) BoP (%) 87.4** 66.3** 63.2** 48.3** 96.4** 81.9** 85.5** 75.9** Parameter Baseline (n=51) 3 months (n=45) 6 months (n=51) 12 months (n=51) Baseline (n=70) 3 months (n=51) 6 months (n=65) 12 months (n=70) Moderate sites (5– 6 mm) PPD (mm) 5.76 (5.62–5.89) 4.07 (3.77–4.36)* , ** 3.78 (3.45–4.10)* , ** 3.93 (362–4.24)* , ** 5.55 (5.43–5.68) 5.00 (4.60–5.40)* , ** 4.58 (4.23–4.93)* , ** 4.85 (4.37–5.32)* , ** CAL (mm) 6.64 (6.29–7.00)** 5.07 (4.70–5.43)* , ** 5.22 (5.69–5.25)* 5.22 (4.83–5.62)* 5.89 (5.70–6.09)** 6.15 (5.65–6.66)** 5.54 (5.13–5.94)* 5.60 (5.05–6.08)* REC (mm) 0.89 (0.53–1.25)** 1.00 (0.64–1.35) 1.44 (0.99–1.90) 1.29 (0.92–1.65)** 0.34 (0.17–0.51)** 1.15 (0.86–1.45)* 0.95 (0.68–1.22)* 0.72 (0.46–0.98)* , ** ΔPPD (mm) 1.69 (1.40–1.97)** 1.98 (1.65–2.31)** 1.82 (1.53–2.12)** 0.55 (0.16–0.95)** 0.97 (0.61–1.33)** 0.71 (0.21–1.20)** ΔCAL (mm) 1.58 (1.16–2.00)** 1.42 (0.93–1.92)** 1.42 (0.95–1.89)** 0.26 (0.69–0.16)** 0.35 (0.1–0.72)** 0.32 (0.16–0.80)** BoP (%) 84.30 40** 49.00 31.4** 94.30 69.2** 62.90 57.1** Abbreviations: BoP, bleeding on probing; CAL, clinical attachment level; PPD, probing pocket depth; REC, recession. *pvalue <.05 for intra-group comparisons; **pvalue <.05 for inter-group comparisons. 22 BLANCO ET AL. differences between test and control groups in any of the studied variables (pvalue >.05). The results favoured the test group in all visits. Mean PPD (4.34 mm test vs. 5.43 mm control), deepest PPD (5.12 mm test vs. 6.85 mm control), PPD reduction (2.42 mm test vs. 0.89 mm control), deepest PPD reduction (3.29 mm test vs. 1.52 mm control), and CAL gain (1.73 mm test vs. 0.50 mm control) showed statistically significant differences between groups at 12 month favouring the test group (pvalue <.05). Intra-group analysis showed that test group achieved statistically significant improvements in all studied clinical parameters in all visits compared with baseline (pvalue <.05). Statistically significant improvements in the control group were only present in mean PPD, maxPPD, and REC (pvalue <.05). 3.4.3 | Site-level analysis Clinical outcomes at moderately deep (5–6 mm) and deep (>6 mm) pocket categories at baseline, 3, 6, and 12 months are displayed in Table 4. At moderately deep sites, there were statistically significant differences favouring the test group in all variables except REC at 3 and 6 months (pvalue <.05), all favouring the test group. At deep sites, all the studied variables presented statistically significant differences favouring the test group (pvalue <.05). Intra-group analysis showed statistically significant clinical improvements compared with baseline in both groups at moderately deep and deep sites in all the studied variables (pvalue <.05). 3.5 |Radiographic outcomes At baseline, there were no significant differences between test and control groups, both at patient and implant level (pvalue >.05). At subject level, the amount of RBL gain was significantly greater in the test group at 6 (1.65 mm test vs. 0.82 mm control) and 12 months (2.33 mm test vs. 1.13 mm control) (pvalue <.05) (Table 2). At implant level, there was significantly greater radiographic bone gain at 12 months in the test group compared with the control group (2.15 mm test vs. 0.95 mm control) (pvalue <.05) (Table 3). 3.6 |Microbiological outcomes The number and percentage of subjects who tested positive to detectable levels of the studied species at baseline, 3, 6, and 12 months are displayed in Table 5. There was a statistically significant difference between groups in the detection of Porphyromonas gingivalis at baseline (15 test/9 control). At 12 months, Tannerella Forsythia and Campylobacter rectus was detected in significantly less patients in the test group compared with the control group (pvalue <.05). CFU counts at baseline, 3, 6, and 12 months are displayed in Table 6. Differences in mean total Tannerella forsythia counts between groups were found at 12 months (pvalue <.05). Regarding quantitative reduction of the studied bacterial species, statistically significant differences favouring the test group were found between baseline and 6 months in Porphyromonas gingivalis (pvalue <.05) and between baseline and TABLE 5 Mean microbiological outcomes for baseline, 3, 6, and 12 months Treatment group Baseline positive (%)/>10 6 (%) 3 months positive (%)/>10 6 (%) 6 months positive (%)/>10 6 (%) 12 months positive (%)/>10 6 (%) Aa Test 0/16 (0%) 0/16 (0%) 0/14 (0%) 0/15 (0%) Control 0/16 (0%) 0/16 (0%) 1/14 (7%) 1/16 (6%) pvalue .7 .7 .5 .3 Pg Test 15/16 (94%) 5/12 (42%)* 3/14 (21%)* 4/15 (27%)* Control 9/16 (56%) 6/15 (40%) 6/14 (43%) 7/16 (44%) pvalue .01 .6 .4 .3 Tf Test 14/16 (86%) 4/12 (33%)* 4/14 (29%)* 5/15 (33%)* Control 14/16 (86%) 6/15 (40%)* 8/14 (57%) 13/16 (81%) pvalue .7 .7 .3 .001 Fn Test 15/16 (94%) 9/12 (75%) 13/14 (93%) 11/15 (73%) Control 16/16 (100%) 13/15 (87%) 14/14 (100%) 16/16 (100%) pvalue .3 .2 .3 .06 Cr Test 13/16 (81%) 6/12 (46%)* 4/14 (28%)* 3/15 (2%)* Control 11/16 (69%) 9/15 (60%) 6/14 (43%) 16/16 (100%) pvalue .3 .3 .3 .04 Abbreviations: Aa, Aggregatibacter actinomycetemcomitans; Cr, Campylobacter rectus; Fn, Fusobacterium nucleatum; Pg, Porphyromonas gingivalis; Tf, Tannerella forsythia. *pvalue <.05 for intra-group comparisons. BLANCO ET AL.23