1. Introduction
Dental implants are currently considered an effective treatment for functional and cosmetic rehabilitation of patients with partial or complete edentulousness [
1,
2]. The clinical success of dental implants is based on the principle of osseointegration, which involves bone growth in metal implants. Multiple factors, including biological [
3], may affect the success of the osseointegration. Peri-implantitis can lead to bone loss and finally implant failure [
4,
5]. Peri-implantitis, marginal bone loss, and implant failure are three outcomes associated with peri-implant diseases (PIDs) [
6,
7]. PID is a collective term for reversible peri-implant mucositis and irreversible peri-implantitis [
8]. Peri-implantitis could negatively affect the quality of life [
9].
In this view, meta-analyses [
6,
7,
8,
9,
10], reviews [
11], and original articles [
12,
13] demonstrated the role of several polymorphisms in PIDs. Proinflammatory cytokines, such as interleukins (ILs), are important biochemical mediators to control the host response to inflammation and to also stimulate the production and secretion of prostaglandins. Prostaglandins are associated with bone resorption and the metalloproteinases, which are involved in collagen degradation [
14]. As such, IL−1 may be a useful indicator and biomarker in diagnosing peri-implantitis, especially because it has an important role in the periodontitis pathogenesis, and because it interferes with immune and inflammation processes, tissue damage, and homeostasis [
15]. IL−1 is composed of 11 genes in the 430-kb fragment in the long arm DNA of chromosome 2, in the 2q12-q21 region. These genes produce the IL−1 alpha (IL−1A) and IL−1 beta (IL−1B) with genetic and biochemical differences but with the same biological functions [
10,
11,
12,
13,
14,
15,
16].
IL−1 receptor antagonist (
IL−1RN) gene regulates the synthesis of the IL−1ra antagonist protein, which can disrupt IL−1A and IL−1B function in competition for receptor binding [
17].
A thorough literature search identified five systematic reviews [
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15,
16,
17,
18,
19,
20] and three meta-analyses [
10,
11,
12,
13,
14,
15,
16,
17,
18,
19,
20,
21,
22] focusing on the associations between IL−1 polymorphisms and PIDs. Among these meta-analyses, one meta-analysis [
10] included the highest number of articles (13 articles) and reported an association between the occurrence of
IL−1A (−889),
IL−1B (−511), and
IL−1B (+3954) polymorphisms in patients with PIDs. In contrast, the meta-analysis [
10] reported no subgroup analysis, meta-regression, or trial sequential analysis (TSA); further,
IL−1B (+3954) and
IL−1B (+3954) polymorphisms were entered in the analyses without further distinctions, and last, the meta-analysis [
10] included studies with a deviation from the Hardy–Weinberg equilibrium (HWE) in their control groups, along with studies with sample sizes with less than 10 cases. To counter this, the present meta-analysis expanded upon previous meta-analyses in three ways. First, the number of included studies was higher. Second, the number of statistical procedures was higher, and the statistical procedures were more complex and sophisticated. More specifically, to counterbalance possible biases and heterogeneity in the results, we employed procedures such as meta-regression and trial sequential analysis (TSA). Third and relatedly, we deleted those studies, in the event that in their control conditions a deviation from the Hardy–Weinberg equilibrium (HWE) could be observed. Given this background, the aims of the present comprehensive meta-analyses were as follows: to evaluate the association of
1A (−889),
IL−1B (−511), and
IL−1B (+3954),
IL−1B (+3954), and
IL−1RN (VNTR) polymorphisms with PIDs; and to conduct subgroup analysis, meta-regression, and TSA. To this end, we removed those studies with a deviation from HWE, separately analyzed
IL−1B (+3954) and
IL−1B (+3954) polymorphisms, and we considered only studies with a minimum of 10 cases.
2. Materials and Methods
2.1. Study Design
The guidelines of PRISMA were followed while reporting this meta-analysis [
23]. The PECO (population, exposer, comparison, and outcomes) question [
24] was: are
IL−1 polymorphisms associated with PID risk among people with dental implants?
2.2. Search Strategy
One author (M.S.) extracted the specific studies from the databases, and the same author removed duplicates and irrelevant studies.
The Web of Science, Cochrane Library, Scopus, and PubMed/Medline databases were searched for studies published until 9 September 2021, without any restrictions. The searched terms were:
(“oral implant*” or “dental implant*” or “peri-implant disease*” or “implant failure” or “implant loss” or “peri-implant” or “peri-implantitis” or “failing implant” or “implant bone loss”) and (“interleukin*” or “interleukin-1*” or “IL−1*” or “IL1*”) and (“variant*” or “polymorphism*” or “allele” or “genotype*”). In addition, we searched several sources (Google Scholar, Free Medical Journals, Library Genesis, and Science Direct) to retrieve relevant missed articles.
2.3. Inclusion and Exclusion Criteria
Inclusion criteria were (1) case-control studies; (2) dental PID was the outcome of interest; (3) studies reporting IL−1A (−889), IL−1B (−511); IL−1B (+3953); IL−1B (+3954); IL−1RN (VNTR), and composite genotype of IL−1A (−889)/IL−1B (+3953) and IL−1A (−889)/IL−1B (+3954) polymorphisms; (4) studies with the required data to calculate the odds ratios (ORs) with 95% confidence intervals (CIs) for genetic models; and (5) studies with no deviation of HWE in their control groups.
Exclusion criteria were (1) studies without the required data regarding genotype distributions, (2) animal studies, meta-analyses, review articles, book chapters, and letters to the editors; and (3) studies including less than 10 cases in each group (case and control groups).
The second author (H.M.) screened all the titles and abstracts based on the eligibility criteria and included/excluded studies for full-text review. Another author (D.S.B.) re-checked the relevant articles. In the event of low agreement, a third reviewer (S.B.) took a final solution.
2.4. Data Extraction
One author (M.S.) independently extracted the information or data from each study and another author (J.T.) rechecked them. If there was a disagreement between the authors, a third author (H.M) took the final decision.
2.5. Quality of Assessment
Two authors (M.S. and H.M.) independently evaluated the quality of each included article using the modified Newcastle-Ottawa Scale (NOS) questionnaire (a maximum total score of 9 was possible for each study) [
25].
2.6. Statistical Analyses
We used Review Manager 5.3 (RevMan 5.3) to calculate crude OR and 95% confidence intervals (CI) as an estimate of the association between IL−1 polymorphisms and PID risk in the five genetic models. To assess the pooled OR significance, the Z-test was applied with a p < 0.05. The I2 statistic showed the heterogeneity, we used the random-effect model, if there was a statistically significant heterogeneity (p < 0.1 or I2 > 50%); if there was no significant heterogeneity, the fixed-effect model was used. Ethnicity, PID outcome, and sample size were criteria for subgroup analyses.
We used Chi-square tests to calculate the p-value of the HWE in the control group of each study; in such cases, a p < 0.05 was considered as a deviation from the HWE.
We used Egger’s and Begg’s tests to plot and analyze the funnel plots; if a p < 0.05, then this was interpreted as publication bias. To evaluate the stability of pooled data, we used sensitivity analyses (“one study removed” and “cumulative analysis”). We used the Comprehensive Meta-Analysis version 2.0 (CMA 2.0) to calculate publication bias tests and sensitivity analyses.
We performed a meta-regression to survey the impact of publication year, ethnicity, PID outcome, and sample size on the pooled results. We used SPSS® version 22.0 (IBM Corporation, Armonk, NY; USA) to perform the meta-regression.
To conduct the trial sequential analysis (TSA) we used the TSA software (version 0.9.5.10 beta) (Copenhagen Trial Unit, Centre for Clinical Intervention Research, Rigshospitalet, Copenhagen, Denmark). Running TSAs reduce these statistical errors [
26], because each meta-analysis may create a false-positive or negative conclusion [
27]. Based on an alpha risk of 5%, a beta risk of 20%, and a two-sided boundary type we computed the required information size (RIS). Studies were considered to have adequate sample sizes and lead to valid results, if the analyses of the Z-curve reached the RIS line, or monitored the boundary line or futility area. Otherwise, the amount of information was considered not to be large enough, suggesting the need for more evidence. A threshold of futility area showed no effect before reaching the information size.
4. Discussion
The main findings of the present meta-analysis showed that there was no association between IL−1A (−889), IL−1B (−511), IL−1B (+3953), and IL−1RN (VNTR) polymorphisms and the risk of dental PIDs. In contrast, there was an increased risk of IL−1B (+3954) in the patients with PIDs. In addition, an association was observed between the composite genotype of IL−1A (−889)/IL−1B (+3953) and PIDs, but not between the composite genotype of IL−1A (−889)/IL−1B (+3954) and PIDs. Further, the subgroup analysis showed that ethnicity and PID outcomes influenced the association of IL−1A (−889) polymorphism and the risk of PID. Ethnicity, PID outcome, and sample sizes were significant factors for IL−1B (−511) polymorphism, while the sample size influenced the IL−1B (+3954) polymorphism. Further, based on meta-regression, the publication year was a significant predictor of the pooled results of IL−1B (+3954) polymorphism. Last, the TSA showed that there were inadequate sample sizes among the studies included in the analyses.
Three published meta-analyses [
10,
11,
12,
13,
14,
15,
16,
17,
18,
19,
20,
21,
22] investigated the association between the
IL−1 polymorphisms and the risk of PIDs. Junior et al. [
21] reported just two articles and reported that there was no association between
IL−1B (−511) polymorphism and the risk of implant failure based on the allelic model. Liao et al. [
10] included 13 articles reporting
IL−1A (−889),
IL−1B (−511),
IL−1B (+3954), and
IL−1RN (VNTR) polymorphisms and the risk of dental PIDs and also mixing the patients with peri-implantitis, implant loss, and marginal bone loss based on the allelic model. This study [
10] included both
IL−1B (+3953) and
IL−1B (+3954) in a similar analysis. The authors found that
IL−1B (−511) polymorphism and the composite genotype of
IL−1A (−889)/
IL−1B (+3954) on risk for implant failure and peri-implantitis. Third, meta-analysis [
22] included two articles to check the association of
IL−1 polymorphisms (
IL−1A (−889),
IL−1B (−511), and
IL−1B (+3954)) with early crestal bone loss around submerged dental implants that there was just an association between
IL−1B (−511) polymorphisms and early crestal bone loss. Our meta-analysis included 16 articles to investigate the association between the
IL−1 polymorphisms and the risk of PIDs. In addition, we included peri-implantitis, implant loss, and marginal bone loss as PIDs and mixed them in the first analysis such as the meta-analysis of Liao et al. [
10], but in subgroup analysis, we separately analyzed them for each polymorphism. Unlike the previously mentioned meta-analyzes, we used five genetic models, meta-regression, and TSA, as well as removed studies with a deviation from HWE in their control group to reduce bias and heterogeneity.
IL−1A (−889) polymorphism was found to be related to chronic periodontal disease in Brazilian cases [
43]. Further, for
IL−1A (−889) polymorphism, T allele compared to C allele induced a four times higher expression of IL−1 alpha [
44] and also TT genotypes compared to CC genotype [
45]. Similarly, Cosyn et al. [
16] presented the association between
IL−1A (−889) polymorphism and the risk of implant failure. In contrast and unlike previous studies [
28,
29,
30,
31,
32,
33,
34,
35,
36,
37,
38,
39,
40,
41], we were unable to identify an association between this polymorphism and the risk of PID. However, our subgroup analysis showed that the association between
IL−1A (−889) polymorphism and the risk of implant failure was statistically significant. Therefore, PID outcomes appeared to be important factors to explain the association between
IL−1A (−889) polymorphism and PID risk.
IL−1B (−511) polymorphism is similar to
IL−1B (+3954) polymorphism, which was found to have a strong role in chronic periodontitis and inflammation [
46]. With regards to the association of
IL−1B (−511) polymorphism with the risk of PID, one study [
33] showed an elevated risk, while another study [
41] reported a protective role of this polymorphism. In this view, the present meta-analysis was unable to confirm the association between
IL−1B (−511) polymorphism and the risk of PID, and this zero association was already observed elsewhere [
16,
17,
18,
19,
20,
21,
22,
23,
24,
25,
26,
27,
28,
29,
30,
31,
32,
33,
34,
35,
36,
37]. However, the subgroup analysis showed that TC genotype has a protective role on marginal bone loss in Asian individuals. Therefore, the role of ethnicity should be considered when focusing on the association between
IL−1B polymorphisms and the risk of PID.
The
IL−1 gene polymorphism may have a negative effect on the results of peri-implantitis treatment in genotype-positive individuals, and the combination of
IL−1A (−889)/IL−1B (+3954) in peri-implant tissues may act as a risk factor that elevates tissue destruction [
36]. In this view, polymorphisms may be involved in osseointegration through the cumulative effect of multiple polymorphisms [
47]. In our meta-analysis, the pooled results showed that the combination of
IL−1A (−889)/IL−1B (+3953) could act as a risk factor for PID, while this was not the case for the combination of
IL−1A (−889)/IL−1B (+3954). Further, the prevalence of these combinations varied among ethnic groups [
36]. Therefore, in future studies, and due to the different results between the combination of
IL−1 polymorphisms and the risk of PID, the combination of
IL−1 polymorphisms with emphasis on ethnicity demand special attention. In addition, the functional genetic polymorphisms of
IL−1B (+3954) and
IL−1RN (VNTR) may diversify the production of IL−1b and IL−1ra proteins [
48,
49].
IL−1B and
IL−1RA may act as regulators of the inflammatory immune system [
50]; as a result, polymorphisms in these genes can affect inflammation and cause implant failure [
31,
32]. Given this background, two studies reported that
IL−1B (+3954) polymorphism could play a role in the pathogenesis of peri-implantitis and increase its risk [
16,
42]. In accordance, our meta-analysis confirmed the result that
IL−1B (+3954) polymorphism caused an increased risk of PID, but not for individuals with peri-implantitis. In the same vein, our meta-analysis and several other individual studies [
30,
31,
32,
33,
34,
35,
36,
37,
38,
39,
40] did not confirm the association of
IL−1RN (VNTR) polymorphism with the risk of PID.
The success of dental implants is determined by several factors such as clinical, biomechanical, and genetic risk [
51,
52]. Further, the synergistic effect of smoking and the positive
IL−1 genotype significantly increase the risk of implant failure [
17]. Regardless of the status of the
IL−1 genotype, smoking was associated with elevated peri-implant bone loss and implant failure [
53,
54]. In our meta-analysis, several studies did not report the smoking status, or data on smoking prevalence among case and control groups were not reported; as such, smoking status was not entered as a further factor in the present meta-analysis. However, future studies should consider the smoking status and its correlation with implant failure and the prevalence of
IL−1 polymorphisms.
Despite the new results, several limitations should be considered. First, based on TSA, there was a lack of sufficient sample sizes in the included studies. Second, only a very few studies were available for the two polymorphisms (IL−1B (+3953) and IL−1RN (VNTR)), given this, subgroup analyses and meta-regression analyses for these polymorphisms were not possible. Third, a high heterogeneity across the studies in several analyses was observed. Fourth, several confounding factors were observed in the pooled results.
In contrast, the strengths of the meta-analysis were first, the lack of publication bias across the studies in most analyses; second, the stability of the results; and third, studies with a deviation from HWE in their control group were removed.