Opiorphin as a biomarker of orofacial conditions: a meta-analysis

The aim of this meta-analysis was to answer the following question: “Are there any differences in opiorphin biomarker concentrations between different orofacial conditions and controls?”. Two reviewers searched for observational studies that evaluated the levels of opiorphin in orofacial conditions, annotated in seven main databases and three that compile gray literature. Of the 443 articles obtained initially, 8 met the inclusion criteria for quantitative analyses. Relative percentages showed a mean 24.1% higher opiorphin concentration in chronic conditions (Burning Mouth Syndrome, Oral Potentially Malignant Diseases and Temporomandibular Disorder) compared to controls; 33.2% higher opiorphin in sustained pain (Symptomatic Irreversible Pulpitis, Symptomatic Apical Periodontitis, Painful Oral Soft-tissue conditions); and 21.7% higher opiorphin after stimuli (Corneal Foreign Body, Capsaicin). Meta-analysis revealed a standardized mean difference of 0.62 [0.02, 1.22] in the absolute concentration of opiorphin in saliva for the chronic group compared to the control. The analogous values for the sustained group and the stimulated group were 2.24 [0.34, 4.14] and 0.43 [0.00, 0.85], respectively. No differences in opiorphin levels were found for ‘after Local Anesthesia before Tooth Extraction’ or for apicoectomy. Based on the available evidence, in general, a statistically higher level of opiorphin is found in orofacial conditions. Salivary opiorphin levels are elevated in chronic, persisted and acute pain conditions, presumably reflecting a physiological homeostatic adaptative response to different conditions such as stress or pain. Salivary opiorphin might therefore be used as a valuable biomarker in several oral disorders.

Eligibility criteria.We included observational studies (case-control, cross-sectional and cohort studies) and case series (> 10 cases) that evaluated opiorphin in orofacial conditions, but included only adults (18-65 years old).The included studies evaluated opiorphin extracted from blood (including plasma), saliva (different preparations), urine, or tears, assessed by ELISA or chromatography.
Overall, the inclusion criteria were based on the PECOS question 41 : Population (P): Humans; Exposure (E): Opiorphin; Comparison (C): Controls; Outcome (O): Different concentrations; Study design (S): observational studies and case series.No data, sex or language restrictions were applied to the search strategy.
The exclusion criteria were as follows: (1) Studies in animals; (2) studies where no orofacial condition was evaluated; (3) studies where opiorphin was not evaluated through saliva, blood, urine or tears; and (4) literature reviews, intervention studies, books, letters, case reports (< 10 cases) and personal opinions.
Information sources and search strategy.Detailed individual search strategies were developed for each bibliographic electronic database: Cochrane, EMBASE, Latin American and Caribbean Health Sciences (LILACS), LIVIVO, PubMed (including Medline), Scopus and Web of Science.A gray literature search was performed on Google Scholar, Open Grey and ProQuest.All database searches were conducted from the starting Synthesis methods.Statistical pooling of data using meta-analysis was carried out where studies were considered combinable and relatively homogeneous in relation to design, interventions, and outcomes.Heterogeneity within studies was evaluated either by considering clinical (differences about participants, type of interventions and results), methodological (design, and risk of bias) and statistical characteristics (effect of studies) or by using the inconsistency index (I 2 ) statistical test 45 .
If quantitative synthesis was appropriate, analysis of the standardized mean difference was performed using RevMan 5.3, and heterogeneity was assessed using the Cochran Q test and I 2 statistics.For the analysis model, a fixed or random effect was based on an expectation of whether the intervention effects were truly identical, preferring the fixed-effect model if this was likely and a random-effects model if this was unlikely.Heterogeneity was calculated by I 2 , and a value greater than 50% was considered an indicator of substantial heterogeneity between studies.The significance level was set at 5%.The meta-analysis was performed with the aid of Review Manager software version 5.3.5 (Nordic Cochrane Center, Copenhagen, Denmark) for continuous data following the appropriate Cochrane Guidelines 45 .
We also considered generating a funnel plot as a graphic to address reporting biases, but in the end our sample size was too small (< 10 articles) for that method of analysis.
Risk of bias across studies and reporting bias assessment.The risk of bias across studies was considered in terms of an overall risk the study results may present, which could influence meta-analysis data.Methodological and statistical heterogeneity was evaluated by comparing the variability in study designs and the risk of bias.Furthermore, we also assessed the risk of bias due to missing results.

Certainty assessment.
A summary of the overall strength of evidence available was presented using "Grading of Recommendations Assessment, Development and Evaluation" (GRADE) Summary of Findings (SoF) tables, using GRADEpro software 45 .

Results
Study selection.Our initial database searches up to June 2021, identified 443 studies.After eliminating duplicated hits, 133 studies remained of which 115 were excluded after title and abstract review, resulting in 18 articles.In addition, 71 studies were found with Google Scholar, 1 with OpenGrey, and 37 with ProQuest.Of these latter 109 studies, 3 from Google Scholar were selected for full-text reading.No additional study was selected following hand-searching of the reference lists of the included studies, although 1 further study was included based on suggestion by an expert.Thus, 22 studies became part of phase-2.The search was updated on October 28, 2022.We found a total of 103 more papers (9 in PubMed, 12 in Scopus, 1 in Cochrane, 43 in Web of Science, 0 in LILACS, 13 in EMBASE, 12 in LIVIVO, and 13 in Google Scholar); however, all 103 were excluded because "no orofacial condition was evaluated" (exclusion criterion #2).During phase-2, 14 of the 22 studies were excluded (reasons for exclusion are given in Appendix 2), leaving 8 studies for qualitative and quantitative synthesis.A flowchart of the process of identification, inclusion and exclusion of studies is shown in Fig. 1.

Study characteristics and results of individual studies.
In the 8 studies evaluated, mean sample size ranged from 22 38 to 144 46 , with a total of 338 subjects with one orofacial condition, and 118 healthy controls.
The proportion of women in the studies ranged from 40 31 to 90.4% 30 .Studies were conducted in Croatia 27,29 , France 30 , India 8,46 , Iraq 38 , and Turkey 31,34 .All studies were published in English.The study by Alajbeg et al. 27 was part of a clinical trials protocol, and the data were obtained by contacting the authors by e-mail.
Study designs included 5 case-control studies 8,27,29,30,34 , 1 randomized clinical trial 46 and 2 quasi-randomized studies 31,38 .Opiorphin levels were measured in saliva in 7 studies, except Boucher et al. 30 also tested blood and urine; Ozdogan et al. 34 measured opiorphin in tears.A human opiorphin ELISA kit was used in all but 2 studies; the exceptions were Alajbeg et al. 27 and Saláric et al. 29 who performed electrospray positive ionization-mass spectrometric multiple reaction monitoring (ESI+/MRM).Table 1 and Appendix 6 (descriptive methods for opiorphin collection) summarize the descriptive characteristics of the included studies.

Risk of bias in studies.
Risk of bias was heterogeneous among the 8 studies.Using JBI Critical Appraisal Tools, 2 studies were classified as having low risk of bias 29,30 , 3 as unclear 8,27,46 , and 3 as high risk of bias 31,34,38 .The higher risk of bias related to strategies to deal with confounding factors.The complete item list is presented in Fig. 2 and Appendix 3.

Results of syntheses.
Individually, TMD, SIP, SAP and CFB were associated with higher concentrations of opiorphin than the control, whereas BMS, POSC, OPMD and LA showed no difference.We further divided the conditions into 4 groups: chronic orofacial group (TMD, BMS and OPMD); sustained pain group (SIP, SAP and POSC); acute pain after local anesthesia group (tooth extraction and apicoectomy); and stimulated group (after CFB, and capsaicin), according to the physio-pathological processes underlying these heterogenous conditions.TMD, BMS and OPMD are chronic conditions (> 3 months) with peripheral and central involvement.SIP and SAP are similar conditions involving long-term alterations in nerve pathways (bacterial inflammation and sensitization) which peak in acute pain, but are not considered as chronic pain conditions.Regarding the others, we thought of merging the acute pain and stimulated groups but the studies are fundamentally different in nature.CFB and capsaicin provoke pain on a short-term basis (minutes or hours), whereas the anesthesia study aimed to suppress pain with an anesthetic.
In addition, we conducted a meta-analysis of the 8 selected studies (Fig. 3).To minimize bias, we used the standardized mean difference as a measure of effect size, because the studies all assess the same outcome but measure it in a variety of ways.The heterogeneity between the studies was high on this meta-analysis (I 2 : 70-90%) because the results were derived from different types of orofacial conditions, and a random effect was considered.Meta-analysis of the chronic group (TMD, BMS, OPMDs) showed a 0.62 [0.02, 1.22] standardized mean difference in the absolute concentration of opiorphin in saliva compared to controls.The sustained group (painful oral soft-tissue conditions vs. controls; and SAP and SIP, before vs. after treatment) showed a 2.24 [0.34, 4.14] standardized mean difference.The stimulated group (capsaicin, CFB) showed a 0.43 [0.00, 0.85] standardized mean difference in the absolute concentration of opiorphin 'after stimulus' when compared to 'before stimulus' .No meta-analysis was feasible for the acute pain after local anesthesia group, owing to a lack of SD data.In general, a statistically higher level of opiorphin was observed in orofacial conditions compared to controls.

Discussion
To our knowledge, this study is the first to systematically review the available evidence related to the concentration of opiorphin in patients with orofacial conditions compared to control subjects.Opiorphin levels were overall increased in OFP conditions.Relative percentages showed 24.1%, 33.2% and 21.7% higher opiorphin levels in chronic pain (TMD, BMS, and OPMDs), sustained pain (SAP, SIP, POSC), and after painful stimulus (CFB, capsaicin), respectively.Meta-analysis found significant standardized mean differences in the absolute concentration in all pain groups compared to controls.While interesting, these findings should be interpreted with caution since several factors may limit their value, as discussed below.
Technical issues.Opiorphin levels were not assessed by the same method in all 8 studies, potentially leading to differences in absolute values.Two studies used HPLC coupled to ionization 27,29 to measure opiorphin and 6 used ELISA 8,31,34,38,46,47 , with 4 different analysis kits.This might explain some discrepancy in the results.ELISA is easy to perform and relatively inexpensive but has a relatively high Limit of Detection (LOD), i.e., the smallest amount of the analyte that can be detected in the test sample.Electrospray positive ionization-mass spectrometric multiple reaction monitoring (ESI+/MRM) 29,48,49 provides much higher sensitivity and specificity but requires highly specialized equipment and software, which limits its usefullness.

OFP subtypes.
The studies included in this review encompass different orofacial conditions.Individually, TMD, SIP, SAP and CFB were reported to have higher concentrations of opiorphin than controls; however, no differences were noted for BMS, POSC, OPMD and LA.When considering the subcategories i.e. chronic pain, sustained pain, acute pain after local anesthesia, and stimulated acute pain, comparative analysis suggest that chronic pain conditions result in higher opiorphin levels, although not for all conditions since BMS data do not support this finding.Indeed two studies report contradictory results 29,30 .The study of Boucher et al. 29,30 reported non-significant differences in the concentration of opiorphin in controls compared to patients with BMS, whereas Salaric et al. 29,30 found a higher concentration in the BMS group.In addition, the results of these studies slightly differed when the saliva collected for analysis was stimulated saliva or non-stimulated saliva.Furthermore another study, not included in this meta-analysis 39 because only published in an abstract form, reported lower levels of salivary opiorphin in BMS patients.Overall, these discrepancies do not provide strong evidence for a link between opiorphin and BMS which might be related to the complex physiopathology of BMS.This condition is understood as a nociplastic condition including hormonal and neuropathic alterations, possibly related to stress 50 when the other OFP included in this review display stronger nociceptive/inflammatory components.Besides chronic conditions, sustained pain conditions also produced an increase of opiorphin, reinforcing the hypothesis that opiorphin is produced as a long-term adaptive response.However; it must be emphasized that some studies are characterized by a high risk of bias owing to lack of clinical information.For example, in the OPMD study 8 it is not clear whether BMS subjects were included.
Ozdogan et al. 31 measured opiorphin levels in pain-free patients 30 days after endodontic treatment, effectively a control group; the opiorphin levels returned to normal after a sustained rise elicited by pain of pulpitis or periapical periodontitis.This likely reflects a long term process even if the pain peaks for just one or a few years.Indeed, studies with local anesthesia, including subjects before tooth extraction 46 and apicoectomy 38 , showed no decrease of opiorphin after a few minutes or after one week, suggesting again sustained, long-term opiorphin production.For acute pain, measurements of opiorphin a few minutes after local anesthesia gave contradictory results: one study showed an increase in opiorphin levels of 143.1% 38 , whereas a second study reported a decrease of 4.1% 46 .

Intensity of pain and opiorphin levels.
A correlation between pain intensity on a visual analog scale (VAS) and opiorphin level was supported in only one study.Ozdogan et al. 31 observed a positive correlation in the painful SIP and SAP pre-treatment group.Other studies did not find or did not report this parameter.However, it must be emphasized that these data are at high risk of bias.For instance, Al-Saffar et al. 38 claimed an inverse correlation between opiorphin levels and VAS post-LA but provided no numeric pain evaluation before anesthesia, reporting only "painful patients" in need of apicoectomy.The same flaw in study design was found in the study of Parida et al. 46 where no pain scores were measured in patients needing tooth extraction.Based on this literature, we could not find an association between local anesthesia and changes in opiorphin levels, and we could not extrapolate the results for direct association between VAS scores and opiorphin level.Further studies are necessary to document this association.
Time course of opiorphin release.Alajbeg et al. 27 stimulated the oral mucosa of subjects with capsaicin, and detected no opiorphin release in control subjects but in TMD patients.
Although not yet published, this is the first study to document acute release of opiorphin in response to a painful stimulus in humans which seems to occur only in certain conditions.The study of Ozdogan et al. 34 also supports the release of opiorphin after a CFB painful stimulus, although with a different time course (hours vs. minutes).It must also be mentioned that local anesthesia, before silencing peripheral nerves, is often accompanied by a pricking pain due to the needle insertion which may also generate stress, and could explain contradictory results.Therefore, more studies related to the time course of opiorphin release after nociceptive stimulation are needed.
Taken together, as the conditions reviewed here include mainly painful and stressful conditions, the data suggest that opiorphin is released in response to pain and/or stressful situations; interestingly, the only study to report a pathological non-painful condition, i.e. oral potentially/malignant conditions 8 , reported no significant increase of salivary opiorphin levels, thus supporting this assertion.
As a consequence, administration of opiorphin or its analogs might be useful in therapeutics.Indeed, studies suggest an analgesic effect of administration of a dual enkephalinase inhibitors, in animals' models of ocular pain 51 and migraine 52 .
Influence of different factors on opiorphin release.Various factors have been described that can influence opiorphin levels, such as age and sex 30,53 , systemic health, use of medications, the most stimulated salivary gland 29 , the body fluid from which opiorphin is collected 30 and psychosocial profile 16 .
Evidence is already available for higher concentrations of opiorphin in males compared to females 53 ; in younger healthy adults (mean age 26 ± 6 years) 53 compared to older ones with BMS (59 ± 12 years) 30 ; in nonpregnant volunteers compared to sixth-month pregnant 53 ; in unstimulated saliva secreted mainly by the submandibular glands compared to stimulated saliva, which is secreted mainly by the parotid glands 29 ; and in serum compared to saliva 30 .In addition, one study found no correlation between the levels of opiorphin and systemic conditions or drug consumption 29 .Another important point concerns whether and how the psychosocial status of the patient may alter the salivary opiorphin levels.Patients with anxiety may experience a more negative emotional response to pain and increased susceptibility to stress 54 .Furthermore, sialorphin increases under acute stress conditions in rats, suggesting that psychosocial status may influence opiorphin levels in human subjects; therefore, studies related to this topic, i.e., different stressful conditions, should be encouraged 16,55 .
Future directions.The present review emphasizes the need for better study designs and improved clinical information.Multicentric designs should be favored to control for cultural differences.Confounding factors such as age, sex, systemic health, use of medications, the body fluid sampled and the psychosocial profile of patients should all be analyzed, as well as pain levels in control groups, and tests conducted before and after intervention.Initiative on Methods, Measurement, and Pain Assessment in Clinical Trials (IMMPACT) recommendations can provide supportive guidelines 56 .
This review highlights the lack of knowledge related to physiologic conditions of opiorphin release.Experimental studies in both acute and chronic conditions should be encouraged, as well as dissociating pain and stress effects on opiorphin release.Finally, trials with different types of nociceptive stimulations such as capsaicin can be considered in future studies in order to decipher the mechanisms of opiorphin release in acute and chronic pain.

Conclusions
The results of the present review may not be generalized due to the aforementioned limitations of the included studies, the higher risk of bias in some studies regarding strategies to deal with confounding factors, and very low GRADE level of evidence.Based on the available evidence, this meta-analysis suggests that salivary opiorphin levels are elevated in chronic, sustained and acute pain conditions, reflecting a physiological homeostatic adaptative response to different conditions such as pain and psychic stress.Salivary opiorphin might therefore be used as a valuable biomarker in oral inflammation.

Figure 2 .
Figure 2. JBI Critical Appraisal Tools to assess risk bias summary in (A) studies; (B) quasiexperimental studies; and (C) randomized clinical trials.
This article does not contain any studies with human participants or animals performed by any of the authors.

Table 1 .
Flow diagram of the literature search and selection criteria.Adapted from PRISMA.Summary of descriptive characteristics of the included articles (n = 8).ELISA Enzyme Linked ImmunoSorbent Assay, F female, HADS hospital anxiety and depression scale, iBMS Idiopathc Burning Mouth Syndrome, NR not reported, OPMDs oral potentially malignant disorders, SAP symptomatic apical periodontitis, SIP symptomatic irreversible pulpitis, SWS stimulated whole saliva, TMD temporomandibular disorders, UWS unstimulated whole saliva, VAS Visual Analogue Scale.