The present systematic review assessed the current existing evidence on the relationship between OLP and PIDs for the first time and attempted to determine whether OLP is a potential risk factor for PIDs. Of 139 studies, two, which contained 68 participants receiving 222 (OLP vs. non-OLP, 112 vs. 110) implants, with a 12- to 120-month follow-up time, were included and evaluated as having high quality. The proportions of implants with PIDs between the OLP and non-OLP groups were as follows: 19.6% (22/112) vs. 22.7% (25/110) for PIM and 17.0% (19/112) vs. 10.9% (12/110) for PI. Existing evidence seems to suggest that OLP is not a suspected risk factor for PIDs during the 10-year follow-up period.
Overall completeness and applicability of the evidence
The diagnostic criteria for OLP were clearly stated in the two enrolled studies, while those for PIDs were absent in the cross-sectional study. Moreover, in the two enrolled studies, PIDs were evaluated at only the implant level and not at the individual level, while OLP was diagnosed at the individual level. Hence, the evidence was insufficient to determine whether OLP is a risk factor for PIDs at the individual level.
Concerning secondary outcomes of PIDs, both studies had complete bleeding on probing data. After data combination, OLP was found to not be a risk factor for bleeding on probing at the implant level. For the secondary outcomes for PIDs, i.e., probing depth and bone loss, the data were complete but reported in different forms in the two studies. The prospective study set cutoff values for probing depth and bone loss to determine the percentages of involved implants, while the cross-sectional study directly presented the medians and ranges of the two parameters. Despite repeated contact with the authors, we obtained only the raw data of the prospective study. Hence, an attempt to combine the data of the two studies in terms of probing depth or bone loss could not be achieved. Ultimately, both studies found no significant difference in probing depth or bone loss between the OLP and non-OLP groups.
When the participant was used as the unit of analysis, the intragroup analysis of the OLP group failed to demonstrate a statistically significant relationship between the existence of desquamative gingivitis and PIM or desquamative gingivitis and PI, but 20 out of 25 implant patients with PIM were among those with desquamative gingivitis in the OLP group; hence, PIM with DG in the OLP group occurred more frequently than non-desquamative gingivitis in the OLP group in this subgroup of implant patients in the prospective study . Since the cross-sectional study presented insufficient information, although the corresponding authors were contacted in various ways and only the author of the prospective study provided the original data of the parameters of interest, no subgroup analysis was performed. Therefore, subgroup combinations in the meta-analysis were not possible. We considered implant failure as an outcome but get no eligible article. Five articles were excluded in the screening step due to case reporting, no control group, etc. Despite no parallel studies comparing implant failure between OLP patients and non-OLP controls, OLP patients have shown a similar implant failure rate (2.7%, 14/523) to populations without OLP (~2%) .
Overall quality, strength, and consistency of the evidence
The present systematic review included only two observational studies that met the inclusion criteria mainly because related research on implantation in OLP populations is scarce. However, the durations of the two enrolled studies were very similar (2003 to 2009 vs. 2005 to 2010), and the study populations were from the same country, which may be beneficial for data synthesis. Nevertheless, the results might be more applicable to specific ethnicities or countries and may not be easily generalized for people globally. The average age of the OLP patients in the two studies was quite different (64.5 vs. 53.7). This distinction may introduce uncertainty regarding the applicable age after pooling the data. Additionally, the mean follow-up time of the two studies was similar but did not exceed 5 years; therefore, it is unclear whether OLP is a risk factor for PIDs over an observation period longer than 5 years. The two studies contained detailed records of important confounders, such as smoking, but only the prospective study controlled for confounders. The response rate was not reported in either study; thus, the number of lost samples was unknown. In the prospective study, two implants in the non-OLP group were lost for unknown reasons, and the study did not state whether it was a PI-associated consequence. Even if that was the case, there was no difference in the results of the PI (RR=1.31, 95% CI: 0.70-2.45, P = 0.40) analysis between the two cases.
Case definitions for PIDs vary considerably in previous studies , and PID data were especially challenging to interpret. Only the prospective study, which reported clear definitions of PIM and PI, was evaluated according to the defined criteria. The cross-sectional study did not specify the diagnostic criteria of PIDs in detail, while the clinical indicators (e.g. BOP, PD, BL) were described in detail with table. the two studies had different definitions of diseases, and there may have been some bias in the data. The diagnostic criteria for PI in the study were not uniform, and limited studies were included in the meta-analysis, so no meta-analysis could be conducted to prove the effect of the differences on the meta-analysis outcomes. Clinical indicators such as bleeding on probing and bone loss are relatively easy to standardize, although no effect of OLP on these indicators has been observed. Implant position has been considered as a prognostic factor for implant success . However, we could not perform subgroup analysis with implant position as a grouping variable due to lack of raw data. Anyway, implants from the two studies showed equal distributions in jaws (maxilla/mandible: 52/60 vs. 63/49, c2=2.162，P=0.141) and tooth sites (20/92 vs. 23/89, c2 =0.259，P=0.611) between OLP and non-OLP groups, suggesting implant position might be not a source of heterogeneity to affect the conclusions.
The results of the statistical analysis show that the homogeneity between the results was good. However, the limited number of enrolled studies made it unfeasible to draw a funnel plot to assess the variation in the studies. Although the sample size was small, the strength of the study evidence was high as analyzed by quality-control assessment, and the overall estimate of the meta-analysis represents the best available evidence.
Implications for practice and policy
Although there was no clear association between OLP and the occurrence of PIDs, implant-related issues in OLP patients cannot be easily ignored, and several issues are worth noting. First, in a clinical study , the gingival index of OLP group was calculated, and the mean value of periodontal index and rate of bleeding on probing were higher than those in control group, revealing that the periodontal condition in the OLP group was poor compared with that in the control group. Poor oral hygiene is a well-known risk factor for PIDs, therefore, OLP may have a certain impact on PIDs. Second, some scholars believed that the capacity of epithelium to adhere to the titanium surface of the implants in OLP patients may be affected because the adhesion of epithelial cells decreases, affecting the epithelial barrier around the implant surface . Third, desquamative gingivitis is a special type of OLP that causes gingival damage. An increased frequency of PIM was reported in cases of desquamative gingivitis compared with non-desquamative gingivitis OLP patients. Another study demonstrated that the presence of desquamative gingivitis was associated with relatively deep periodontal pockets  . The study inferred that desquamative gingivitis lesions might circuitously enhance the long-term risk for periodontal diseases via plaque accumulation by impeding proper oral hygiene around natural teeth and implants. It is important to clarify the constructive associations between desquamative gingivitis and PIDs in the future. Moreover, OLP is a T cell‑mediated chronic inflammatory autoimmune disease . PIDs in humans cause leukocytic infiltration through barrier epithelial migration and an increase in the proportion of T and B cells . In this context, these two conditions might aggravate each other via inflammatory pathways. Additionally, dysbiosis might be a potential link between OLP and PIDs, given that both conditions have been related to microbial alterations[34, 35] . Finally, systemic corticosteroid treatment in OLP patients was frequently associated with decreased bone mineral density, especially during the first 6 months of corticosteroid therapy . The osteoporotic effects of corticosteroids might cause more alveolar bone loss around dental implants during infection .
In this systematic review, the available evidence did not support OLP as a risk factor for PIDs. However, this does not mean that implant restoration can be performed indiscriminately in OLP patients. For patients with acute/erosive forms of OLP or desquamative gingivitis, immune system disorder and poor oral hygiene may be potential risk factors for PIDs. In clinical settings, whether OLP is well-controlled or not would be a concern during implant treatment planning. The two studies included in the current review had declared that implantation therapy was conducted during the remission stages of OLP. Some study has shown a very high implant failure rate (42/55) for OLP patients receiving implant placement during the acute stages . Reversely, controlled OLP patients and healthy condition have displayed an equal marginal bone loss around implant in four-year follow up . Obviously, selecting the remission stages are clinically appropriate and normative. OLP should not be considered as a contraindication for implant treatment in patients who do not have evident symptoms or mucosal erosive congestion and have good oral hygiene. Furthermore, the mental state of OLP patients might be improved after implant treatment, which in turn is better for OLP control, considering that those without implant treatment have been reported to have a poor quality of life, a weaker psychological profile and greater stress [15, 17]. Additionally, for OLP patients receiving glucocorticoid treatment, assessment of alveolar bone mineral density in the area to be implanted should be considered. Finally, it is worth noting the effect of implants in mouth on the recovery of OLP patients, although there were no significant differences in OLP signs and symptoms between the implant group (14 patients) and nonimplant group (15 patients) during the 12-24 month follow-up period . For OLP patients who have undergone implant restoration, OLP conditions may need to be monitored frequently, and implant maintenance planning may need to be personalized.
Implications for further research
This review systematically analyzed the existing evidence on relation between OLP and PIDs for the first time. Large-scale prospective trials are required to validate the findings.
The results suggest that OLP is not a potential risk factor for PIDs during the 1 to 10-year follow-up period. However, given the relatively small amount of evidence available, the final answer to this question depends on large, well-designed prospective and randomized clinical trials in the future. At present, the proportion of people undergoing implant treatment and the prevalence of OLP worldwide are not high, which may cause difficulty in researching the correlations between these two diseases; therefore, multicenter cooperative clinical research is needed. Additionally, OLP has various clinical classifications, and the prospective study included in this systematic review showed that OLP patients with desquamative gingivitis had a higher frequency of PIM. This suggests that in future studies, subgroup analysis of OLP may be helpful in exploring the associations between OLP and PIDs. Moreover, interventional studies on the effects of OLP treatment on peri-implant status are still lacking; these studies can further elucidate the associations between OLP and PIDs. Molecular biology techniques can also help to explore the microscopic effects of OLP on peri-implant pathophysiological processes at molecular level.