All patient-related information and clinical recordings were recorded in a web-based clinical research form (VieDoc version 3.24, PCG solutions, Uppsala, Sweden).

Patients under 18 years of age; current smokers; patients who had undergone radiotherapy in the head and neck region, chemotherapy or systemic long-term corticosteroid treatment; patients who were pregnant or nursing; patients receiving medications known to induce gingival hyperplasia; patients with uncontrolled diabetes (HbA1c >6.5); patients who had taken systemic antibiotics less than 6 months prior to baseline; patients receiving bisphosphonate treatment; and patients with prosthetic factors that prevented clinical measurements were excluded from participation in the study. Implants with technical complications, such as loose supraconstructions, cement remnants, ill-fitted crowns with poor marginal contour or any type of prosthetic complication that according to the examiner, would be a local contributing factor to inflammation, were also excluded. Implants that were previously treated surgically for peri-implantitis and implants with overcontoured supraconstructions obstructing access for debridement and clinical measurements of more than three surfaces were also excluded. Before agreeing to participate in this study, all patients were provided with sufficient information via a patient information sheet and a consent form, which was signed prior to final inclusion.

After clinical recordings, the implant pockets were debrided with a chitosan brush (LBC, BioClean®, LABRIDA AS, Oslo, Norway) seated in an oscillating dental drill piece (ER10M, TEQ-Y, NSK Inc., Kanuma Tochigi, Japan) for 3 min and then irrigated with sterile saline (Fig. 1). The initial debridement was performed with local anaesthesia as needed. A second debridement with the chitosan brush was performed after 3 months.

The power calculation was based on data from a pilot clinical study evaluating the same test device performed at the Department of Periodontology, University of Oslo, in 2014. Descriptive statistics were presented for continuous variables (median and interquartile range), and proportions were presented for categorical variables.