Effects of probiotics on salivary cytokines and immunoglobulines: a systematic review and meta-analysis on clinical trials

Findings on the effects of probiotics on salivary cytokines and immunoglobulines have been conflicting. We aimed to perform a systematic review and meta-analysis on clinical trials that examined the effects of oral intake and local administration of probiotics on salivary cytokines and immunoglobulines in adults. We searched PubMed, MEDLINE, SCOPUS, EMBASE, and Google Scholar up to April 2020 for all relevant published papers assessing probiotic intakes and salivary cytokines and immunoglobulines. We included all randomized clinical trials that investigated the effect of oral probiotic supplementation or lozenges tablets on inflammatory biomarkers in adults. Studies that reported their effect sizes as mean ± SD or mean ± SEM were included. After excluding non-relevant papers, 8 studies remained in this review. Combining findings from 3 studies with 4 effect sizes, we found no significant reduction in salivary IgA concentrations after oral probiotic supplementation [weighted mean difference (WMD): −0.26; 95% CI: (−0.86, 0.35)]. A significant increase in salivary IL-1β concentrations reached after local probiotic supplementation (WMD: 28.21; 95% CI: 18.42, 38.01); however, no significant changes in salivary IL-6 concentrations after local probiotic supplementation was found (WMD: 0.36; 95% CI: −0.85, 1.56). We observed a significant increase in salivary IL-8 concentrations after local probiotic supplementation (WMD: 31.82; 95% CI: 27.56, 36.08). In case of salivary IL-10 concentrations after local probiotic administration, no significant reduction was seen (WMD: −0.02; 95% CI: −0.10, 0.06). we found that oral and local administrations of probiotics might influence some of salivary cytokines. However, additional clinical trials are required to examine these effects on further pro- and anti-inflammatory cytokines and immunoglobulines.

www.nature.com/scientificreports www.nature.com/scientificreports/ with regulated responses of immune system 12 . Probiotics have been reported to have local (direct) and systemic (indirect) effects on immune system 4 . For instance, they have been involved in maintaining of oral health through inhibiting the growth of pathogens 13,14 . Oral intake of probiotic drinks or supplements enhanced the secretory IgA in saliva 2,6,15 . In addition, local administration of probiotics in lozenges results in higher levels of salivary IgA and specific cytokines 13,14 . However, some other studies failed to find significant changes in salivary immunoglobulines or inflammatory cytokines by either oral intake or local administration of probiotics [2][3][4][5]7,[15][16][17] . Despite earlier investigations, there is no comprehensive systematic review or meta-analysis summarizing earlier findings in this regard. We conducted this systematic review and meta-analysis to summarize the available data about the effects of oral intake and local administration of probiotics on salivary cytokines and immunoglobulines in adults.

Methods
Search strategy. This systematic review and meta-analysis of clinical trials was conducted based on Cochrane library checklist. All articles published earlier than April 2020 were searched through PubMed, MEDLINE, SCOPUS, EMBASE, and Google Scholar, by two independent investigators to identify relevant articles. To obtain suitable MESH and non-MESH text words, an initial search on Medline was undertaken. The systematic search strategies through each database were provided in the supplementary material file. We had no restrictions of language or time of publication. To avoid missing any publication, a manual search was conducted on reference lists of all included studies as well as review articles. We didn't include unpublished data and grey literature, including dissertations, thesis, congress papers, and patentsin the current meta-analysis. In addition, duplicate citations were removed. Inclusion criteria. We included all randomized clinical trials that investigated the effect of oral probiotic supplementation or lozenges tablets on inflammatory biomarkers in adults. Studies that reported their effect sizes as mean ± SD or mean ± SEM were included. Publications were independently assessed by two reviewers considering the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) checklist. Any disagreements between the reviewers were resolved through discussion.In case of several publications with the same data set, we included only the most complete one 13,16 . If data for specific probiotics were reported separately, we considered them as a separate study in the analysis 1 .
Exclusion criteria. Studies were excluded if they were observational, editorial, letter to editor, comments, ecological or review papers. In addition, studies in which random allocation was not performed, had not control group or those conducted on animal models, pregnant or lactating women, children or elderlies were not included. Publications that examined the effect of another intervention along with probiotic supplementation, those that used symbiotics, examined only gene expression of inflammatory biomarkers or concentrations of inflammatory biomarkers in-vivo were not also considered eligible for the current study. Publications that examined gingival index, plaque index, bleeding, depth of pocket and etc. were excluded. The study by Garaiova et al. was excluded from systematic review and meta-analysis because its study population was children 18 . We also excluded the study of Dong et al. study form the meta-analysis due to not reporting any effect size 3 . In addition, the study of Jorgensen et al. 16 was excluded because the data were repeatedly reported in the study of Braathen et al. 13 . After these exclusions, 8 papers remained for the primary systematic review. We didn't consider two studies in the meta-analysis due not to reporting the data for control group 6 and in the end of trial for both groups 5 . Figure 1 illustrates the study selection process for systematic review and meta-analysis. Data extraction. The data were extracted independently and cross-checked by two reviewers (SE and AM).
Any disagreements between reviewers were consulted by principal investigator (AE). Quantitative data regarding effect-size measures such as mean and Standard Deviations (SDs) or mean and Standard Errors (SEs) or median and Interquartile Range (IQR) of inflammatory biomarkers before and after intervention in each groups; and mean (SD) changes in inflammatory markers after intervention in each group were extracted.In addition, information on first author's last name, publication year, subjects' heath condition, sample size, participants' sex, www.nature.com/scientificreports www.nature.com/scientificreports/ number of subjects in each group, participants' age, type of probiotics, study design (parallel/cross-over/other), type of control, duration of intervention and covariates were obtained. If data were reported as SEs or IQR, they were converted to SDs using appropriate formulas. When the concentration of an inflammatory biomarker was reported in different units, it was converted to the most frequently used one. Three studies had reported results in Figs. 1, 2, 6. We obtained the values from the figures by online "webplot digitizer" converting 2D Bar Plot to data. The values for SD changes were calculated using √S 1 2 + S 2 2 − 2 × r × S 1 × S 2 formula, in which r was computed for each individual study using SD 1 2 + S 2 2 − SD change 2 /2SD 1 SD 2 . The quality of studies and risk of bias of all eligible studies were assessed using the Cochrane Collaboration's tool for quality assessment of randomized controlled trials 19 . The quality assessment tool encompasses the following items: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting and other probable sources of biases.

Statistical analysis.
All effect sizes were calculated as mean ± SD of changes in the concentrations of inflammatory biomarkers between probiotic and control groups. The fixed-effects model was used to calculate the overall effect sizebecause random-effects model gives larger weights to small extreme studies 20 . We examined between-study heterogeneity by the Cochran's Q test and I 2 statistic. To find probable sources of between-study heterogeneity, subgroup analyses were conducted based on sex (Male/Female/Both genders), age (<40 year/>40 year), study design (Parallel/Cross-over), supplement dosage (=10 9 />10 9 CFU/day), duration of intervention (<3 /≥3 weeks) and probiotic type (Lactobacillus/Bifidobacter/Different types), using a fixed-effects model. The duration of 3 weeks and the dosage of 10 9 CFU/day were selected based on previous studies 21,22 . All statistical analyses were done using Stata software, version 11.2 (Stata Corp, College Station, TX). P < 0.05 was considered as statistically significant.

Results
Findings from the systematic review. The initial literature search yielded 407 unique studies. Based on titles and abstracts, 378 studies were excluded. Out of these, 21 studies were also excluded due to above-mentioned reasons. Finally, 8 articles that reported the effects oforal probiotic intake or probiotic containing lozenges tablets on salivary immunoglobulins or cytokines remained for the current study. Main characteristics of five studies that examined the effects of oral probiotic intake on salivary immunoglobulins are presented in Table 1. Five studies were done on healthy adults 1,2,5,6,15 . These studies were published between 2008 and 2016. Except for one study on men 15 , four other studies were performed on both genders. Total sample sizes in intervention and control groups were 231 and 129, respectively (54.92% female and 45.07% male). Participants in these studies were healthy people aged ≥18 years.Three studies were parallel 1,5,6 and 2 studies were cross-over trials 2,15 . Participants consumed the probiotic supplements or placebos as capsules 1,15 or milk-or fruit juice-based drinks 2,5,6 . Daily dose of supplementation ranged from 10 9 to 35 × 10 9 . All studies had control group, except for the study of Harbige et al. 6 . Administered probiotics were lactobacillus 1,5,6,15 , bifidobacter 1,2,5 and propionibacterium 5 . Three studies had used more than one type of probiotic 1,2,16 . Duration of trial ranged from 3 to 6 weeks. Measured outcomes were salivary IgA 1,2,5 , IgA1 6,12 , IgA2 6 , IgG 1 , IgM 1 and INF-γ 6 . The method of assessment of outcome in all studies was enzyme-linked immunosorbent assay (ELISA). Three studies had reported mean ± SE of salivary immunoglobuline concentrations before and after intervention 6 or their changes 1,2 . Table 2 presents   www.nature.com/scientificreports www.nature.com/scientificreports/ the results of quality assessment of eligible studies on oral probiotic intake. Two studies had poor quality 5,6 , two had good quality 1,2 and the remaining one study had fair quality 15 . The risk of bias was attributed to random sequence generation and blinding of outcome assessment in the included studies. Due to limited number of studies, we did not perform subgroup analysis by quality of primary studies. Table 3 provides characteristics of three studies 4,13,14 that examined the effects oflocal administration of probiotic tablets as lozengeson salivary cytokines and immunoglobulins. These studies were published between 2007 and 2017 and were conducted on both genders except for one study on females 14 . Total sample sizes in intervention and control groups were 93 and 66, respectively (68.79% female and 31.21% male). Participants were healthy people aged ≥18 years. Two studies were cross-over 13,14 and one study was parallel trial 4 . In these publications, participants were healthy participants 13,14 or periodontal patients 4 . Daily dose of supplementation ranged from 0.1 × 10 9 to 3 × 10 9 . The administered probiotics in these papers were various strains of lactobacillus. Duration of trials ranged from 3 weeks to 12 weeks. Measured outcomes were salivary IgA 13 , IL-1β 4,13,14 , IL-6 4,13,14 , IL-8 4,13,14 , IL-10 4,13,14 , IL-18 14 and TNF-α 4,13,14 . The method of assessment of all these variables was enzyme-linked immunosorbent assay (ELISA). All studies had reported mean ± SD of salivary cytokines and immunoglobuline concentrations before and after intervention. The quality assessment of included studies on local administration of probiotic tablets as lozenges revealed that two studies had fair quality 4,14 and the remaining one study 13 had good quality ( Table 4). Allocation concealment and blinding of outcome assessment were the major sources for risk of bias. Again, due to limited number of studies, we were not able to do subgroup analysis. www.nature.com/scientificreports www.nature.com/scientificreports/ Findings from meta-analysis. Combining findings from 3 studies 1,2,15 with 4 effect sizes, we found no significant reduction in salivary IgA concentrations after oral probiotic supplementation [weighted mean difference (WMD): −0.26; 95% CI: (−0.86, 0.35)] (Fig. 2). There were no significant between-study heterogeneity (I 2 = 0.0%, P = 0.427). No particular study had a significant influence on the summary effect in our sensitivity analysis. There was no proof of significant publication bias (Egger's test: 0.494).
There were 3 clinical trials examining local administration of probiotic tablets as lozenges on salivary IL-1β, IL-6, IL-8 and IL-10 4,13,14 . Combining three effect sizes from clinical trials, we found a significant increase in salivary IL-1β concentration after local probiotic supplementation (WMD: 28.21; 95% CI: 18.42, 38.01) (Fig. 3). There were no significant between-study heterogeneity (I 2 = 11.9%, P = 0.32). No particular study had a significant influence on the summary effect in our sensitivity analysis. There was no proof of significant publication bias (Egger's test: 0.89).
A significant increase in salivary IL-8 concentrations was observed after local probiotic supplementation (WMD: 31.82; 95% CI: 27.56, 36.08) (Fig. 5). However, a significant between-study heterogeneity was found (I 2 = 72.7%, P = 0.026). Due to limited number of studies we did not perform subgroup analysis to find possible source of this heterogeneity.

Disscusion
In the current meta-analysis, we found a significant increase in salivary IL-1β and IL-8 concentrations after local probiotic supplementation. However, no significant effects of oral probiotic supplementation on salivary IgA levels and also, no significant effects of local probiotic supplementation on salivary IL-6 and IL-10 concentrations were found in our meta-analysis. To the best of our knowledge, this is the first systematic review and meta-analysis summarizing the effects of oral and local probiotic supplementation on salivary immunoglobulines and cytokines.
Our findings from the current meta-analysis were in line with previous clinical trials that showed no significant increase in salivary IgA levels after oral probiotic treatments compared to placebo 5,15 . In contrast, some studies indicated a significant increase in serum IgA concentrations by probiotic consumption 1,6 . Whereas Childs et al. reported a significant decrease in salivary IgA concentrations after probiotic intake 2 . Although some earlier studies have shown the effect of probiotic supplementation on systemic IgA antibody releasing and B cell stimulatory activity 23,24 , the salivary concentrations of IgA, as a marker of mucosal immunity, did not influence by probiotic supplementation. This might be explained by the age of participants. Most studies have enrolled elderly people, whom antibody responses might be different from healthy middle-age adults. Moreover, saliva volume and its contents might be affected by several environmental and neural factors. Therefore, salivary levels of IgA could also be influenced by psychological and physical stress 24 . Due to limited number of publications, we were unable to do subgroup analysis by sex, age group, design and duration of trials, dose and type of probiotics. These factors may also affect our findings. It must also be taken into account that exposure to probiotics in early life through diet might also contribute to immune responses and secretion of immune-globulins in body liquids 25 .
We found a significant increase in some salivary inflammatory cytokines including IL-1β and IL-8 concentrations by local probiotic administration. However, no significant changes in IL-6 and IL-10 were observed following probiotic supplementation. These findings were in agreement with several other reports from randomized clinical trials that showed a significant increase in salivary cytokines including IL-1β 4,14 . Against to this finding, some investigators failed to find any significant effects on salivary cytokines 4,13,17 . One should keep in mind that local administration of probiotics is different from oral supplementation. The effects of local ingestion of probiotics on immune system function basically depend on individual oral biofilm environment and oral hygiene and  Table 2. Study quality and risk of bias assessment of included studies on oral probiotic intake according to the Cochrane Collaboration's tool. U; unclear risk of bias, L; low risk of bias, H; high risk of bias. *Good quality: all criteria met; Fair quality: one criterion not met (i.e. high risk of bias for one domain or two criteria unclear); Poor quality: two or more criteria listed as high or unclear risk of bias.  26 . Individual oral biofilm and inflamed gums or healthy gums can differently respond to probiotic treatments. In addition, in case of gingivitis, in which we face with acute inflammation, local administration of probiotics for short-term cannot cool down inflammation due to elevated levels of inflammatory cytokines in these patients 27 . Moreover, in spite of immune-modulatory effects of local administration of probiotics and secretion to saliva, regular intake of probiotic products does not seem to be enough to initiate major alterations in oral biofilm 4 . It should also be kept in mind that the quality of primary studies can strongly influence the overall effect size. We assessed study quality in the current investigation and excluded studies with poor quality from the current analysis because of not reporting reliable effect sizes 5,6 . However, we could not perform subgroup-analysis based on quality of studies due to the limited number of publication in each area.
The possible mechanisms through which probiotic administration might affect salivary cytokines and immunoglobulines are not clearly understood. Among the possible suggested mechanisms are the effects of probiotics on increasing Treg function, through which they can induce the anti-inflammatory cytokine production, such as TGF-β, which can consequently lead to increased levels of IgA [28][29][30][31] . In addition, secretions of anti-inflammatory cytokines are up-regulated by probiotics through encouraging the anti-inflammatory M2 macrophages 32,33 .
Despite being the first meta-analysis on salivary cytokines and immunoglobulines, some limitations need to be considered. Due to limited number of publications, we were unable to do the meta-analysis on some other   Table 3. Effects of local administration of probiotic tablets as lozenges on salivary cytokines and immunoglobulins.
www.nature.com/scientificreports www.nature.com/scientificreports/ cytokines and immunoglobulines. The effects of probiotics are strongly dependent to age and primary exposure of host. This should be considered in the interpretation of the findings. We confined our meta-analysis to adult population and did not include studies that investigated children or adolescences. Moreover, despite the effects of salivary flow rate on the levels of salivary cytokines and immunoglobulins on one hand 34,35 and the effect of probiotic supplementation on salivary flow rate on the other hand 12 , none of the studies had considered normalized levels of cytokines for salivary flow rate. In addition, we did not register the protocol of the current study on PROSPERO registry system due to the delay in processing the submitted protocols for studies outside the UK. This lack of registration might be a source of bias for this review. However, this review and meta-analysis was designed and performed according to the Cochrane guidelines.
In conclusion, we found that oral and local administrations of probiotics were significantly associated with increased levels of IL-1β and IL-8 in adult population. However, additional clinical trials are required to examine these effects on further pro-and anti-inflammatory cytokines and immunoglobulines.   www.nature.com/scientificreports www.nature.com/scientificreports/