We did not classify peri-implant diseases into peri-implant mucositis and peri-implantitis in this pilot study. Peri-implant mucositis does not show BL, whereas peri-implantitis shows BL of more than 2.5 or 3 mm on intra-oral radiographs [39, 40]. Figuero et al. [2] introduced plural diagnostic criteria for peri-implant mucositis and peri-implantitis. Rakic et al. [5] defined peri-implantitis as a PD of more than 5 mm, BOP positive, and BL of at least two threads of implant. Furthermore, Sanz et al. [41] proposed their opinion for the radiographic assessment of alveolar bone in peri-implant treatment. However, difficulties are associated with accurately measuring 2–3 mm of alveolar BL on a radiograph taken by a regular method and assessing BL levels by implant threads when implant species differ. We evaluated BL around dental implants using Schei et al.’s method [30], which has been used to evaluate BL rate in periodontal diseases. The mean BL rate was significantly higher at peri-implant disease site than at healthy sites without inflammation and deep PD. Therefore, we did not distinguish peri-implant mucositis and peri-implantitis that were diagnosed by measuring bone level on radiograph in the present pilot study. Biomarkers for BL may be more accurate than clinical BL indicators because PICF NTx amounts were found to correlate with BL rates determined by Schei et al.’s method (ρ = 0.570, P < 0.001). Biomarkers for bone metabolism in PICF and clinical, radiological assessment of bone level may accurately diagnose peri-implant mucositis and peri-implantitis.

Bone-related proteins including ICTP, osteocalcin (OCN), and RANKL have been studied as BL biomarkers in peri-implantitis. ICTP, a cross-linked C-telopeptide of type I collagen, is a marker for bone degradation, and its levels in PICF were significantly higher from peri-implantitis sites than from healthy sites [9, 42]. However, Tümer et al. [13] did not detect a significant difference in PICF ICTP levels between peri-implantitis and healthy sites. RANKL is a main mediator of osteoclast formation and associated with bone resorption [43]. Soluble RANKL (sRANKL) concentrations in PICF were significantly higher from peri-implantitis sites than from healthy implant sites (P < 0.01), and its levels correlated with clinical indicators such as PD (ρ = 0.309, P = 0.034) and BOP (ρ = 0.327, P = 0.024) [44]. In the present study, NTx amounts and concentrations showed similar significant differences to sRANKL between the peri-implant disease and healthy groups (amount: P < 0.01, concentration: P < 0.05), and a stronger correlation was observed between NTx amounts and PD (ρ = 0.434, P < 0.001). In contrast, Arikan et al. [9] showed that sRANKL concentrations in PICF were significantly higher in healthy groups, while Sarlati et al. [45] reported no significant difference in PICF sRANKL concentrations among healthy, peri-implant mucositis, and peri-implantitis groups. OCN is a major non-collagenous protein in bone and is associated with bone metabolism [46]. The mean OCN concentration in PICF from peri-implantitis sites was approximately 1.5-fold that of healthy groups [13], and this finding was similar to the result for NTx in PICF. Although OCN levels in PICF samples were significantly higher from peri-implant mucositis sites without BL than from healthy sites, OCN levels in PICF from peri-implantitis with BL was not significantly different from those in PICF from healthy and peri-implant mucositis sites [47]. These conflicting findings do not necessarily suggest that ICTP, sRANKL, and OCN are reliable biomarkers for alveolar BL. Few studies showed a relationship between the PICF levels of bone-related markers and those of clinical indicators for alveolar BL. NTx levels in GCF samples were significantly higher from periodontitis sites than from healthy sites [28]; however, the relationship between NTx levels in PICF or GCF and BL levels has not yet been investigated. NTx in PICF may be a reliable biomarker for evaluating BL in peri-implantitis because PICF NTx levels correlated with the BL rate as well as PD and had high sensitivity and specificity for predicting peri-implant diseases.