The progress of patients throughout the different phases of the study is illustrated in Fig. 1. Table 1 depicts the baseline demographic patient and implant characteristics. The included patients had a total of 128 implants of which 53 implants showed signs of peri-implantitis. Different implant brands and types with different implant surfaces were present, including Straumann (Straumann AG, Basel, Switserland; SLA® and SLActive® surface), Nobel Biocare (Nobel Biocare AB, Göteborg, Sweden; TiUnite® surface), Biomet 3i (Biomet Inc., Warsaw, Indiana, USA; OSSEOTITE® surface), Frialit-2, (Dentsply Friadent, Mannheim, Germany; FRIADENT® plus surface), and Pitt-Easy (Sybron Implant Solutions GmbH, Bremen, Germany; Puretex® surface). Three patients with each one implant with peri-implantitis were lost to follow-up (2 patients from the control group, 1 from the test group).

¹⁰Log-transformed mean bacterial anaerobic counts of the culture-positive implants for the control and test group before and after debridement and decontamination of the implant surface during the surgical procedure are depicted in Table 2. In both groups, the debridement and decontamination procedure resulted in a significant immediate reduction in counts of anaerobic bacteria on the implant surface. Although the reduction in total anaerobic load was greater in the test group, the difference did not reach the level of statistical significance (p = 0.108). However, in the test group, the total anaerobic load was significantly more often reduced below detection level than in the control group (20 out of 23 in the test group, 10 out of 17 in the control group, p = 0.042). No significant differences were observed in the ¹⁰Log-transformed mean bacterial anaerobic counts of the peri-implant sulcus, neither between control and test group nor between baseline and 3 months after surgery (Table 3).

Descriptive statistics of the clinical outcomes at baseline and follow-up are depicted in Table 4. At 3-month follow-up, 75% of the implants (66.7% of the patients) in the control group and 63.3% of the implants (53.8% of the patients) in the test group showed no clinical signs of inflammation (PPD ≤4 mm without bleeding and/or suppuration on probing) (Table 4). The results from the multilevel analyses regarding the effects of the intervention on BoP, SoP, and PPD are shown in Table 5. No significant differences in BoP, SoP, and mean PPD were detected between control and test group at 3 months after surgery, neither in the “crude” nor in the “adjusted” analysis.