Fixed implant-supported restorations are extensively employed to rehabilitate partially or completely edentulous patients with predictable outcomes. However, despite their high percentages of success, dental implants are not free from possible complications with consequent failure, the causes of which are still the object of debate in the dental scientific community.
Moreover, even if all clinical parameters and changes in bone levels were combined, they may not be sufficient to predict the patient’s risk of developing peri-implantitis and its prognosis at the beginning of the inflammation process [
10]. For this reason, early diagnosis of pathologic BL and identification of early biomarkers for a peri-implant disease are essential. Diagnosis could be implemented by detecting immunological host-derived molecules, such as chemokines, cytokines, bone markers, and enzymes involved in peri-implant tissues turnover [
11]. Biomarkers such as pro-inflammatory cytokines (i.e., tumor necrosis factor-alpha (TNF-α), interleukin-1 beta (IL-1β), interleukin-6 (IL-6), and interleukin-17 (IL-17)) are classically associated with the initiation of the inflammatory cascade. They have also been proven to be stimulated by periodontal pathogens’ virulence. Other substances, including neutrophil elastase collagenase, alkaline phosphatase, and aspartate aminotransferase, have been weakly associated with peri-implantitis [
12]. Markers of bone tissues (i.e., osteoprotegerin (OPG) and soluble receptor activator of nuclear factor kappaB Ligand (sRANKL)), osteoclastogenic-related cytokines and chemokines (i.e., granulocyte colony-stimulating factor (G-CSF), matrix metalloproteinase-8 (MM-8), monocyte chemoattractant protein (MCP-1) are other important molecules which could be considered to understand better the immune-inflammatory profile of peri-implant disease [
13].
In the last decades, several studies have shown the presence of host-derived biochemical mediators in PICF, and levels of these inflammatory molecules have been proposed as a measure of active peri-implantitis. Biomarkers assessment in PICF may also be useful to identify specific markers responsible for the onset and development of peri-implantitis since its earliest stages when it is still clinically latent.
The initial search strategy provided a total of 158 articles: 101, 17, and 40 articles were found on PubMed/MEDLINE, Cochrane Library, and Google Scholar databases, respectively. After eliminating all duplicates, 127 possibly relevant studies were detected. In fact, a total of 31 duplicate articles were removed before the screening. After screening the articles’ titles and abstracts, 38 possibly relevant studies were detected for full-text examination. The final selection after full-text analysis included nine papers [
12,
14,
27,
28,
29,
30,
31,
32,
33]. A flowchart was drawn up describing the results of the study search and selection.
3. Included Studies
The nine included studies have been conducted from 2000 to 2022. One of the included studies was conducted in Turkey [
29], two in Finland [
27,
28], one in Sweden [
32], one in Iran [
30], one in Saudi Arabia [
33], one in Japan [
31], one in Switzerland [
12], and one [
14] in Italy. All the studies were written in English.
Detailed information for the nine included studies is reported in Table 1.
Table 1. Main characteristics of included studies.
Authors |
Study Design |
Number of Patients |
Number of Implants |
Assessed PICF Biomarkers |
Definition of BL * |
Type of Assay |
Main Outcomes |
Ma et al., 2000 [27] |
Cross- sectional study |
13 |
49 |
Collagenase-2 Collagenase-3 |
Group 1: BL < 1 mm; Group 2: BL from 1 to 3 mm; Group 3: BL > 3 mm |
Time-resolved immunofluorometric assay (collagenase-2) Quantitative immunoblot (collagenase-3) |
Collagenase-2 (p = 0.049) and collagenase-3 (p = 0.041) levels were significantly higher in Group 3 than in Groups 1 and 2. Collagenase-3 and collagenase-2 produced by adjacent bone osteoclast cells reflect irreversible peri-implant vertical BL around loosening dental implants. Measurements of collagenase-3 and collagenase-2 could be used as markers to indicate the degree of peri-implant vertical BL. |
Plagnat et al., 2002 [12] |
Cross- sectional study |
15 |
19 (healthy: 11; peri-implantitis: 8) |
ALP EA α2M |
Healthy implants: no radiographic evidence of BL. Implants with peri-implantitis: crestal BL greater than 20% in at least one site (mesial or distal) along the implant |
P-nitrophenyl-phosphate as substrate (ALP) low molecular weight Fluorogenic substrate (EA) ELISA (α2M) |
ALP and EA were correlated with the percentage of BL. ALP and EA could be promising markers of BL around dental implants. |
Ma et al., 2003 [28] |
Cross- sectional study |
13 |
46 |
Gelatinase b |
Group 1: BL < 1 mm; Group 2: BL from 1 to 3 mm; Group 3: BL > 3 mm. |
Modified urokinase assay |
The differences between activated (p = 0.044) and total gelatinase B (p = 0.026) levels were significant in the three BL groups. Furthermore, gelatinase B levels were increased in Group 3 compared to Groups 2 and 3. Activation of gelatinase B together with elevated mGI eventually reflects active phases of peri-implantitis and may prove to be diagnostically useful. |
Yamalik et al., 2012 [29] |
Cross- sectional study |
40 |
54 (Group P-1: 19 healthy implants; Group P-2: 27 implants with mucositis; Group P-3: 8 implants with peri-implantitis) |
Cathepsin-K |
The actual distance between two consecutive threads of the dental implant was used as a reference point |
Cathepsin-K activity assay kit |
Mean BL values for subgroups P-1, P-2, and P-3 were 1.242, 1.514, and 1.844 mm, respectively (p = 0.087). Mean total cathepsin-K activity levels of subgroups P-1, P-2, and P-3 were 3.637, 6.114, and 16.290 units, respectively. However, there is no positive correlation between the enzymatic profile of PICF and the BL measurements. Mean BL around dental implants did not significantly correlate with total cathepsin-K activity. |
Yaghobee et al., 2013 [30] |
Cross- sectional study |
32 |
41 |
IL-1β |
BL measured by intraoral periapical radiographs |
Enzyme-linked immunosorbent assay (ELISA) |
It seems that there Is a positive correlation between IL-1β level and BL (p < 0.0001) Mean BL: 1.66 mm. |
Sakamoto et al., 2018 [31] |
Cross- sectional study |
35 |
74 (healthy: 34; peri-implantitis: 40) |
Calprotectin and cross-linked N-telopeptide of Type I Collagen (NTx) |
BL of more than 2.5 or 3 mm evaluated around dental implants by intra-oral radiographs |
Enzyme-linked immunosorbent assay (ELISA) |
The mean BL rate of peri-implant disease sites was 42.7%, and that of healthy sites was 19.7%. The BL rate in healthy sites ranged between 6.9 and 41.8%, and that in diseased sites was between 7.7 and 80.0%. A positive correlation was observed between NTx amounts and the BL rate (ρ = 0.570, p < 0.001). Calprotectin and NTx in PICF are markers of inflammation and BL in peri-implant tissues and may be useful diagnostic markers for peri-implant diseases. |
Lira-Junior et al., 2019 [32] |
Cross- sectional study |
43 |
42 (mucositis: 20; peri-implantitis: 22) |
CSF-1; IL-34; IL-1β |
BL measured by intraoral periapical radiographs; mucositis: BL around the implant not reaching the first thread; peri-implantitis: BL involving at least two implant threads |
Commercial enzyme-linked immunosorbent assays |
There is no statistically significant correlation between CSF-1, IL-34, IL-1β, and BL. |
Algohar et al., 2020 [33] |
Cross- sectional study |
60 |
94 (healthy: 32; peri-implant mucositis; 27: peri-implantitis: 35 |
Procalcitonin |
BL: radiographic level of bone ≥3 mm apical of the most coronal portion of the intraosseous part of the implant after initial bone remodeling. BL is defined as the linear distance measured from the implant-abutment junction to the most coronal point of the alveolar crest. |
Enzyme-linked immunosorbent assay (ELISA) |
Mean BL in healthy, mucositis, and peri-implantitis groups were, respectively: 0.7 mm (0.5), 1.1 mm (0.6), and 2.5 mm (0.9). In the peri-implantitis group, a significant positive correlation was observed between crestal BL (p = 0.0013) and PICF procalcitonin levels. |
Menini et al., 2021 [20] |
Prospective cohort study |
7 |
14 |
MicroRNAs * |
BL was considered normal if it was ≤1 mm and increased if it was >1 mm |
Microarray technology |
MiRNAs may be used as biomarkers of peri-implant bone resorption. The following miRNAs were altered in the case of BL both in PICF and in soft peri-implant tissues: miR34 a; miR100; miR106 a; miR126; miR143; miR146 a; miR181; miR200; miR221; miR223; miR375; miR378; miR429; miR1248. |
Participants in all the included studies were in good general health and had not received any medication that could influence the peri-implant pathological process. All the studies specified the number of implants investigated, and some also specified the division into groups: healthy, affected by mucositis, or peri-implantitis. Healthy implants were defined as such when they showed no bone resorption and no signs of inflammation. However, the definition of pathological peri-implant BL was different between the articles. Five studies categorized “pathologic” BL in mm: BL ≥ 3 mm [
33], BL > 3 mm [
27,
28], and BL > 2.5 mm [
31]. Menini et al. considered “augmented” (versus “normal”) a BL > 1 mm [
14]. In one study, peri-implantitis was considered present if crestal BL was greater than 20% of the implant length in at least one site (mesial or distal) along the implant [
12]. Finally, three studies did not specify any millimetric cut-off value to define an augmented or pathologic BL [
29,
30,
33].
BL measurements were calculated from intraoral radiographs in all the included investigations.
Eight of the included studies were cross-sectional studies [
12,
27,
28,
29,
30,
31,
32,
33]. Only the study by Menini et al. was a prospective cohort study [
14].
The following biomarkers were examined: calprotectin, cross-linked N-telopeptides of type I collagen (NTx), cathepsin-K, alkaline phosphatase activity (ALP), elastase activity (EA), inhibitor α2-macroglobulin (α2M), procalcitonin, interleukins (IL-1β and IL-34), monocyte chemotactic protein-1 (MCP-1), collagenase-2, collagenase-3, gelatinase b, colony-stimulating factor (CSF-1), and miRNAs. MiRNAs were evaluated in one unique study using a microarray testing the expression of all 2700 human miRNAs [
14]. In this cohort study, miRNAs were extracted from PICF, analyzed, and they were considered predictive of BL.
Plagnat et al. evidenced the correlation between some specific biomarkers in PICF (elastase, α2-macroglobulin, and alkaline phosphatase) and BL around implants [
12].
Lira et al. considered the correlation between three other biomarkers (CSF-1, IL-34, IL-1β) and BL [
32].
The study by Yaghobee et al. assessed the correlation between IL-1β level and BL [
30], while in the article of Sakamoto et al., the association between calprotectin and cross-linked N-telopeptides of type I collagen and BL rate I was compared in sites affected by peri-implant disease and healthy sites [
30].
In contrast, in the comparative study of Yamalik et al., no correlation was found between the enzymatic profile of PICF (i.e., cathepsin-K) and BL measurements [
29]. In the study of Algohar et al., procalcitonin was collected from different sites to find a link with BL [
33].
Then, there are two studies Ma et al. [
27,
28] on the same sample of patients. They were both included since they reported the outcomes regarding different biomarkers: in the first study, the difference between activated and total gelatinase B levels was measured in three different BL groups (Group 1: BL < 1 mm; Group 2 BL from 1 to 3 mm; Group 3 BL > 3 mm), while in the other paper collagenase-2 and collagenase-3 levels were evaluated in the same BL groups.
In five articles, PICF collection was obtained using paper points left in the sulcus for 30 s [
14,
29,
30,
31,
32]. In two studies, a filter strip was placed into the sulcus for 4 min [
27,
28]. In one study, PICF was collected with paper strips left in the peri-implant sulcus for 15 s [
12]. Finally, in one study, paper cones were employed, but the insertion time was not specified [
33].
In most of the studies, PICF samples were collected at mesial and distal sites. In the study by Lira et al., it was specified that the sites were inflamed with gingival indexes of 1 or 2 [
32]. In the studies by Ma et al., the samples were collected from the site with maximum vertical BL and assessed by X-ray [
27,
28].
Biomarkers collected from PICF were examined through different types of assays: enzyme-linked immunosorbent assay (ELISA) [
12,
30,
31,
33], cathepsin-K activity assay kit [
29], P-nitrophenyl-phosphate as substrate [
12], low molecular weight fluorogenic substrate [
12], modified urokinase assay [
28], time-resolved immunofluorometric assay, quantitative immunoblot [
27], and commercial enzyme-linked immunosorbent assays [
32]. MiRNAs were evaluated with microarray technology [
14].
Because of the methodological heterogeneity of the included studies, and in particular, because different biomarkers were investigated in the different studies, a meta-analysis was not appropriate and was not conducted.
4. Results of Individual Studies
Most of the studies found a positive correlation between the biomarkers collected from PICF and BL. In the study by Sakamoto et al. [
31], the BL rate in healthy sites ranged between 6.9 and 41.8%, while that in diseased sites was between 7.7 and 80.0%; a positive correlation was observed between NTx amounts and BL rate (ρ = 0.570,
p < 0.001). In the study by Algohar et al. [
33], implants were divided into healthy and diseased (with mucositis or peri-implantitis), and it was found that mean BL in healthy, mucositis, and peri-implantitis groups were, respectively: 0.7 mm (0.5), 1.1 mm (0.6), and 2.5 mm (0.9). In the peri-implantitis group, a significant positive correlation was observed between crestal BL (
p = 0.0013) and PICF procalcitonin levels.
In the study by Menini et al. [
14], there was a positive correlation between miRNA expression profile and BL. Moreover, 14 miRNAs that were altered in PICF in the case of greater BL were also altered in the soft peri-implant tissue of the same implant sites.
In the study by Ma et al. [
27], it was found that collagenase-2 (
p = 0.049) and collagenase-3 (
p = 0.041) levels were significantly higher in Group 3 than in Group 1 and 2 (Group 1: BL < 1 mm; Group 2 BL from 1 to 3 mm; Group 3 BL > 3 mm). The authors concluded that collagenase-3 and collagenase-2 produced by adjacent bone osteoclast cells reflect irreversible peri-implant vertical BL around loosening dental implants. In the second study by Ma et al. [
28], the differences between activated (
p = 0.044) and total gelatinase B (
p = 0.026) levels were significant in the three BL groups. Furthermore, gelatinase B levels were increased in Group 3 compared to Groups 2 and 3. Activation of gelatinase B together with elevated mGI (modified Gingival Index) eventually reflects active phases of peri-implantitis and may prove to be diagnostically useful.
In the study of Plagnat et al. [
12], implants were divided into two groups: healthy implants that lost a mean maximum of 0.6 mm of crestal bone, whereas in the diseased group (that is, implants with a BL greater than 20% in at least one site-mesial or distal-along the implant), a correlation was found between EA and the percentage of BL.
Additionally, in the study by Algohar et al. [
33], in the peri-implantitis group, a significant positive correlation was observed between crestal BL (
p = 0.0013) and PICF procalcitonin levels. In fact, a significantly higher level of PICF procalcitonin was found in the implants with BL compared to the groups of healthy implants (
p = 0.039) and groups of implants with mucositis (
p = 0.042). Even the study by Yaghobee et al. found a positive correlation between IL-1β levels and BL (
p < 0.0001) [
30].
Only two studies did not find any statistically significant correlation between BL and PICF biomarkers. In Lira J et al., CSF-1, IL-34, and IL-1β were not related to BL [
33]. Moreover, in the article by Yamalik et al., the mean BL values for subgroups P-1, P-2, and P-3 were 1.242 mm, 1.514 mm, and 1.844 mm, respectively (
p = 0.087), and mean total cathepsin-K activity levels were 3.637, 6.114, and 16.290 units, respectively, with no positive correlation between the enzymatic profile of PICF and BL [
29].
5. Excluded Studies
Out of 38 papers for which the full text was analyzed, 29 were excluded from the systematic review [
13,
34,
35,
36,
37,
38,
39,
40,
41,
42,
43,
44,
45,
46,
47,
48,
49,
50,
51,
52,
53,
54,
55,
56,
57,
58,
59,
60,
61]. After full-text reading, the studies were excluded for two main reasons: (1) studies that did not measure BL; (2) studies that did not correlate PICF biomarkers with BL.
6. Quality Assessment
The risk of bias in the included studies was assessed using the Joanna Briggs Institute (JBI) tools. The purpose of this appraisal is to assess the methodological quality of a study and to determine the extent to which a study has addressed the possibility of bias. The questions for each checklist and the relative risk of bias are reported in Table 2 and in Table 3.
Table 2. Risk of bias for clinical studies included according to the critical appraisal tools of JBI Scale for analytical cross-sectional studies.
|
(1) |
(2) |
(3) |
(4) |
(5) |
(6) |
(7) |
(8) |
Ma et al., 2000 [27] |
No |
No |
Unclear |
Unclear |
No |
No |
Unclear |
Unclear |
Plagnat et al., 2002 [12] |
Yes |
Yes |
Yes |
Yes |
Unclear |
Unclear |
Yes |
Yes |
Ma et al., 2003 [28] |
Yes |
Unclear |
No |
Unclear |
Unclear |
Unclear |
Yes |
Yes |
Yamalik et al., 2012 [29] |
yes |
Unclear |
Yes |
Unclear |
Yes |
Yes |
Yes |
Yes |
Yaghobee et al., 2013 [30] |
Yes |
Yes |
Unclear |
Unclear |
Yes |
Unclear |
Yes |
Yes |
Sakamoto et al., 2018 [31] |
Yes |
Yes |
Yes |
Unclear |
Yes |
Yes |
Yes |
Yes |
Lira-Junior et al., 2020 [32] |
Yes |
Yes |
Yes |
Yes |
Yes |
Yes |
Yes |
Yes |
Algohar et al., 2020 [33] |
Yes |
Yes |
Yes |
Yes |
Yes |
Yes |
Yes |
Yes |
Table 3. Risk of bias for clinical studies included according to the critical appraisal tools of the JBI Scale for Cohort Studies.
|
(1) |
(2) |
(3) |
(4) |
(5) |
(6) |
(7) |
(8) |
(9) |
(10) |
(11) |
Menini et al., 2021 [20] |
Unclear |
Yes |
Yes |
Unclear |
Unclear |
Yes |
Yes |
Yes |
Yes |
Yes |
Yes |