Effects of Microecological Regulators on Rheumatoid Arthritis: A Systematic Review and Meta-Analysis of Randomized, Controlled Trials

In this study, the available data from published randomized, controlled trials (RCTs) of the use of intestinal microecological regulators as adjuvant therapies to relieve the disease activity of rheumatoid arthritis (RA) are systematically compared. An English literature search was performed using PubMed, Embase, Scopus, Web of Science and the Cochrane Central Registry of Controlled Trials and supplemented by hand searching reference lists. Three independent reviewers screened and assessed the quality of the studies. Among the 2355 citations identified, 12 RCTs were included. All data were pooled using a mean difference (MD) with a 95% CI. The disease activity score (DAS) showed a significant improvement following microecological regulators treatment (MD (95% CI) of −1.01 (−1.81, −0.2)). A borderline significant reduction in the health assessment questionnaire (HAQ) scores was observed (MD (95% CI) of −0.11 (−0.21, −0.02)). We also confirmed the known effects of probiotics on inflammatory parameters such as the C-reactive protein (CRP) (MD −1.78 (95% CI −2.90, −0.66)) and L-1β (MD −7.26 (95% CI −13.03, −1.50)). No significant impact on visual analogue scale (VAS) of pain and erythrocyte sedimentation rate (ESR) reduction was observed. Intestinal microecological regulators supplementation could decrease RA activity with a significant effect on DAS28, HAQ and inflammatory cytokines. Nevertheless, these findings need further confirmation in large clinical studies with greater consideration of the confounding variables of age, disease duration, and individual medication regimens.


Introduction
Rheumatoid arthritis (RA) is a systemic autoimmune disease with a chronic inflammatory process, which can cause symmetrical joint swelling, stiffening, arthralgia, and limited range of motion. Over time, progressive inflammation of the joints leads to cartilage damage, bone erosion, disability, and socioeconomic burdens [1,2].
The etiopathogenesis of RA is complex and involves the interaction between genetic and environmental factors. There are two major subtypes of RA (seropositive and seronegative) depending on the presence or absence of rheumatoid factor (RF) and anti-citrullinated protein antibodies (ACPAs). ACPAs can be found in sera many years before the onset of RA in approximately 67% of patients [2]. The most prevalent and strongly associated prebiotics by gut microbiota produces short-chain fatty acids (SCFAs). Several animal studies have implicated that SCFAs increased in collagen-induced arthritis (CIA) models fed prebiotics which presented alleviative inflammatory arthritis [26]. Synbiotics have both probiotic and prebiotic properties, and the proper combination of these two ingredients in a single product should be considered. Synbiotics were created to help probiotics survive in the intestines and they may improve colon implantation of probiotics, stimulate bacteria growth and modify the gut microbiota [27]. Based on the contributions to modulating gut microbial ecology mentioned above, microecological regulators could be a potential possibility for arthritis treatment.
The available data from randomized controlled trials (RCTs) are highly heterogeneous in terms of the study population, the characteristics of rheumatic diseases, the composition of supplements, and the results regarding activity scores and inflammatory markers. There are a few meta-analyses published to demonstrate the efficacy of probiotics or dietary supplementation in patients with inflammatory rheumatism; however, most of them included other inflammatory arthritis in addition to RA and mainly analyzed clinical variables. Furthermore, meta-analyses that discussed diet or dietary supplements conflated prebiotics with synbiotics, and thus the results are discordant and hard to interpret [25,[28][29][30]. The meta-analysis we performed pooled only RCTs with a detailed composition of microecological regulator supplementation in RA and therefore to provide more conclusive results. This meta-analysis and systematic review summarizes and analyzes the efficacy of probiotics, prebiotics, and synbiotics supplementation in RA, to provide a reference for the clinical application of microecological regulators in RA patients.

Materials and Methods
This study was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [31] (Supplementary Table S1). The study protocol was registered in PROSPERO International Prospective Register of Systematic Reviews (registration number CRD42022363172) [32].

Search Strategy
A comprehensive literature search was conducted to identify relevant articles from inception to November 2022 in the PubMed, EMBASE, Scopus, Web of Science, and Cochrane Central Registry of Controlled Trials databases. A manual search of the reference lists of all identified articles was carried out to find additional studies.
Search results were retrieved, and duplicates were removed using EndNote X9 software for Windows. Three independent reviewers assessed the relevance of selected retrieved articles. Screening of titles and abstracts was followed by full-text screening. Disagreements were resolved by discussion and consensus between reviewers and senior researchers.

Eligibility Criteria
We included any open-label or blinded randomized controlled studies that evaluated the efficacy of oral supplementation with intestinal microecological regulators (prebiotics, probiotics, and synbiotics) in adult patients with an established diagnosis of RA. The control therapy could be a placebo or another diet intervention. We excluded any uncontrolled studies, case reports, case series, letters, editorial comments, theses, literature reviews, book chapters, news, or only abstracts. We also excluded papers if their data could not be extracted or if they were not written in English.
Outcomes included RA clinical disease activity indices such as the disease activity score of 28 joints (DAS28), number of tender or swollen joints (TJC and SJC), Health Assessment Questionnaire, Disability Index (HAQ), visual analog scale (VAS) for disease activity provided by the patient, VAS for pain and global health score (GH score). Laboratory markers were the C-reactive protein (CRP) level, erythrocyte sedimentation rate (ESR), and levels of inflammatory cytokines. Disagreements in the determination of the eligibility of each study were resolved by discussion and consensus.

Data Extraction
Data of interest were extracted using a custom Microsoft Excel Office spreadsheet. The following data were extracted for each study: publication date, journal, study design, sample size, demographic characteristics (e.g., age, sex, disease duration, inclusion criteria, treatments such as DMARDs and symptomatic medications (GCs and NSAIDs)), microecological regulator (prebiotics, probiotics, and synbiotics) formulation details, outcome measures, side effects, and adherence.

Quality Assessment
Three independent reviewers assessed the risk of bias using the Outcomes Cochrane Collaboration tool for assessing risk of bias [33]. Records limited to abstracts were not assessed because of the lack of information about the study design. Any disagreement between them was resolved by discussion.

Statistical Analysis
All relevant quantitative data were, where possible, pooled in the statistical metaanalyses. The outcomes were the variation between the inclusion and evaluation endpoints between the two groups. A narrative synthesis was carried out to describe data extracted from articles that could not be included in the meta-analyses.
A meta-analysis was performed for all outcomes using the RevMan V 5.3 software package developed by Nordic Cochrane Centre (Review Manager (computer program), V 5.3. Copenhagen, Denmark: The Nordic Cochrane Centre, the Cochrane Collaboration, 2011). P values lower than 0.05 were considered significant. Statistical heterogeneity of the selected studies was tested using the Q-test (χ2) and reported with the I 2 statistic. Heterogeneity was considered to be significant when the χ2 test had a p value < 0.1 or I 2 test value > 50%. A fixed-effects model was used to calculate the pooled mean difference (MD) or standardized mean difference (SMD). In the case of significant statistical or clinical heterogeneity, a random-effects model was applied. Publication bias was checked with Egger's test.

Study Selection
The literature search of different databases revealed 4696 records, and one additional study was identified manually as shown in Figure 1. Of these, 2342 reports were duplicated and excluded by the Endnote software. The titles and abstracts of the remaining 2355 reports were screened and 2332 reports were excluded after the screening. Then, 23 studies were excluded after screening because of the wrong type of article (n = 4), or outcome (n = 6). One study was omitted for overlap with another study published by the same researchers at the same time [34]. Twelve articles were finally included in the qualitative synthesis. One study was excluded from the meta-analysis for having a high risk of bias [35]. 2355 reports were screened and 2332 reports were excluded after the screening. Then, 23 studies were excluded after screening because of the wrong type of article (n = 4), or outcome (n = 6). One study was omitted for overlap with another study published by the same researchers at the same time [34]. Twelve articles were finally included in the qualitative synthesis. One study was excluded from the meta-analysis for having a high risk of bias [35].
The characteristics of the individual studies are shown in Table 2. The total number of patients with RA in the included studies was 762; 219 patients were treated with prebiotics, 243 with probiotics and 143 with synbiotics. One study only evaluated Bacillus coagulants [41], three studies only evaluated Lactobacillus [40,42,44], and two studies assessed a mix of different probiotic types [39,43]. Lactobacillus is also the main strain in Preferred reporting items for systematic reviews and meta-analyses (PRISMA) flow diagram.

Study Characteristics
The characteristics of the included studies are summarized in Table 1. Three studies were about prebiotics [36][37][38], six about probiotics [39][40][41][42][43][44], and three about synbiotics in RA patients [35,45,46]. The intervention in two studies was a combination of probiotics and a high-fiber diet [35,46]; thus, we placed these studies in the category of synbiotics.  12 (26) 24 (51) Normal diet DAS28-CRP 3.6 (0. The characteristics of the individual studies are shown in Table 2. The total number of patients with RA in the included studies was 762; 219 patients were treated with prebiotics, 243 with probiotics and 143 with synbiotics. One study only evaluated Bacillus coagulants [41], three studies only evaluated Lactobacillus [40,42,44], and two studies assessed a mix of different probiotic types [39,43]. Lactobacillus is also the main strain in synbiotics supplementation. Two studies assessed a high-fiber diet combined with probiotic supplementation [35,46]. Three different types of prebiotics were included in this review: standard mixed dietary plant-derived polysaccharides (dPPs) [38], polyphenols [36,37] and fructan [45]. The duration of intervention ranged from 8 weeks to 1 year. There were no studies included that provided an active control involving another probiotic strain. The comparator was placebo except in Nenonen et al. [35], which compared an uncooked vegan diet mixed with lactobacilli versus a normal diet, and in Vadell et al. [46], which compared an anti-inflammatory diet mixed with probiotics versus a typical Swedish diet. The main inclusion criteria were RA diagnosed according to the 1987 ACR/EULAR criteria [36][37][38][41][42][43]. Other inclusion criteria were a disease duration of more than 1 year [41,42,44], more than 6 months [39,45], and more than 2 years for Vadell et al. [46]. Treatments had to be consistently administered from 1 to 3 months prior to inclusion [37,38,40,42,46]. Eight studies referred to at least mild activity according to the DAS score [35,[37][38][39][40][41]45,46], while four studies did not specifically require the minimum DAS score [36,[42][43][44]. One study specified the requirement of no NSAIDs use [44] and two of no biological DMARDs use [39,44].

Risk of Bias Assessment
The risk of bias assessment is illustrated in Figures 2 and 3. The assessment of the results indicated that the quality of the included papers ranged from low to high. All studies were double-blinded random control studies, except two single-blinded studies [35,46] and one open-label study [36]. In addition, some studies did not adequately report key outcomes, such as the DAS28 and CRP level, which were also considered to have a moderate risk [41,42,44]. Nenonen et al. [35] was rated high risk because of the inappropriate measurement and data reporting of the outcome; thus, we excluded it from the meta-analysis. The main inclusion criteria were RA diagnosed according to the 1987 ACR/EULAR criteria [36][37][38][41][42][43]. Other inclusion criteria were a disease duration of more than 1 year [41,42,44], more than 6 months [39,45], and more than 2 years for Vadell et al. [46]. Treatments had to be consistently administered from 1 to 3 months prior to inclusion [37,38,40,42,46]. Eight studies referred to at least mild activity according to the DAS score [35,[37][38][39][40][41]45,46], while four studies did not specifically require the minimum DAS score [36,[42][43][44]. One study specified the requirement of no NSAIDs use [44] and two of no biological DMARDs use [39,44]

Risk of Bias Assessment
The risk of bias assessment is illustrated in Figures 2 and 3. The assessment of the results indicated that the quality of the included papers ranged from low to high. All studies were double-blinded random control studies, except two single-blinded studies [35,46] and one open-label study [36]. In addition, some studies did not adequately report key outcomes, such as the DAS28 and CRP level, which were also considered to have a moderate risk [41,42,44]. Nenonen et al. [35] was rated high risk because of the inappropriate measurement and data reporting of the outcome; thus, we excluded it from the meta-analysis.   The main inclusion criteria were RA diagnosed according to the 1987 ACR/EULAR criteria [36][37][38][41][42][43]. Other inclusion criteria were a disease duration of more than 1 year [41,42,44], more than 6 months [39,45], and more than 2 years for Vadell et al. [46]. Treatments had to be consistently administered from 1 to 3 months prior to inclusion [37,38,40,42,46]. Eight studies referred to at least mild activity according to the DAS score [35,[37][38][39][40][41]45,46], while four studies did not specifically require the minimum DAS score [36,[42][43][44]. One study specified the requirement of no NSAIDs use [44] and two of no biological DMARDs use [39,44]

Risk of Bias Assessment
The risk of bias assessment is illustrated in Figures 2 and 3. The assessment of the results indicated that the quality of the included papers ranged from low to high. All studies were double-blinded random control studies, except two single-blinded studies [35,46] and one open-label study [36]. In addition, some studies did not adequately report key outcomes, such as the DAS28 and CRP level, which were also considered to have a moderate risk [41,42,44]. Nenonen et al. [35] was rated high risk because of the inappropriate measurement and data reporting of the outcome; thus, we excluded it from the meta-analysis.

• VAS of pain
Four studies provided direct data on pain in RA patients as measured with a 10 mm VAS [39][40][41]45]. We pooled all four studies in the meta-analysis, and no significant influence on the VAS of pain score was found, with an MD −0.49 (95% CI −1.40, 0.42) p = 0.29, I 2 = 86%, n patients = 182) ( Figure 6B).

• VAS of pain
Four studies provided direct data on pain in RA patients as measured with a 10 mm VAS [39][40][41]45]. We pooled all four studies in the meta-analysis, and no significant influence on the VAS of pain score was found, with an MD −0.49 (95% CI −1.40, 0.42) p = 0.29, I 2 = 86%, n patients = 182) ( Figure 6B).
The above studies and one study of prebiotics provided data on IL-6 and TNF-α [36,40,42,44]. We pooled the four studies and no significant IL-6 and TNF-α improvement were noticed in the meta-analysis (Figure 9).
The above studies and one study of prebiotics provided data on IL-6 and TNFα [36,40,42,44]. We pooled the four studies and no significant IL-6 and TNF-α improvement were noticed in the meta-analysis (Figure 9). three studies were pooled in the meta-analysis and demonstrated a significant improvement in the IL-1β level as a result of probiotic supplementation in 93 RA patients (MD −7.26 (95% CI −13.03, −1.50), p = 0.01, I 2 = 33%).
The above studies and one study of prebiotics provided data on IL-6 and TNF-α [36,40,42,44]. We pooled the four studies and no significant IL-6 and TNF-α improvement were noticed in the meta-analysis (Figure 9).

Tolerance Data
Seven studies provided information on side effects, and no side effects related to microecological regulators were reported in six of them [35][36][37][38][39][40][41][43][44][45]. Vadell et al. [46] reported that 29% of patients during the intervention periods experienced upset stomach symptoms, and most of them only existed at the start of the intervention period.

Discussion
The aim of this systematic review and meta-analysis was to identify the respective efficacy of intestinal microecological regulators in RA patients. Several statistically significant and possibly clinically meaningful effects were observed with microecological regulator intervention: (1) a significant decrease in the DAS28, (2) borderline benefits according to the HAQ, (3) a significant decrease in CRP levels in RA patients under probiotics, and (4) a significant decrease in proinflammatory cytokines in the probiotics group. This information may help inform clinical physicians and RA patients concerning the use of intestinal microecological regulators.
All except one study were rated with a low risk of bias arising from the randomization process for using appropriate random sequence generation and allocation concealment methods. Cannarella et al. [43] was considered to raise some concerns, because whiel they provided randomization methods, they did not report allocation concealment methods. Nine studies were rated as having a low risk of assignment to intervention bias, but Khojah et al. [36], Nenonen et al. [35], and Vadell et al. [46]. raised some concerns, because they were not double-blind RCTs. Alipour et al. [44], Mandel et al. [41], and Zamani et al. [39] were rated as raising some concerns regarding attrition bias due to outcome data loss at follow-up. Khojah et al. [36] and Mandel et al. [41] did not adequately report key outcomes, and were considered to raise some concerns regarding bias in the measurement of the outcomes, and Nenonen et al. [35] was rated as being of high risk because of the inappropriate measurement and data reporting of the outcomes. Due to the lack of a registered protocol or the insufficient data on several secondary outcomes, six studies were considered to raise some concerns regarding reporting bias. Overall, seven studies were rated as having a low risk of bias, four were rated as raising some concerns, and one was rated as having high risk. The two funnel plots of DAS28 and CRP were substantially symmetrical (Supplementary Figure S1). The results of Egger's test for DAS28 and CRP were p value =

Tolerance Data
Seven studies provided information on side effects, and no side effects related to microecological regulators were reported in six of them [35][36][37][38][39][40][41][43][44][45]. Vadell et al. [46] reported that 29% of patients during the intervention periods experienced upset stomach symptoms, and most of them only existed at the start of the intervention period.

Discussion
The aim of this systematic review and meta-analysis was to identify the respective efficacy of intestinal microecological regulators in RA patients. Several statistically significant and possibly clinically meaningful effects were observed with microecological regulator intervention: (1) a significant decrease in the DAS28, (2) borderline benefits according to the HAQ, (3) a significant decrease in CRP levels in RA patients under probiotics, and (4) a significant decrease in proinflammatory cytokines in the probiotics group. This information may help inform clinical physicians and RA patients concerning the use of intestinal microecological regulators.
All except one study were rated with a low risk of bias arising from the randomization process for using appropriate random sequence generation and allocation concealment methods. Cannarella et al. [43] was considered to raise some concerns, because whiel they provided randomization methods, they did not report allocation concealment methods. Nine studies were rated as having a low risk of assignment to intervention bias, but Khojah et al. [36], Nenonen et al. [35], and Vadell et al. [46]. raised some concerns, because they were not double-blind RCTs. Alipour et al. [44], Mandel et al. [41], and Zamani et al. [39] were rated as raising some concerns regarding attrition bias due to outcome data loss at followup. Khojah et al. [36] and Mandel et al. [41] did not adequately report key outcomes, and were considered to raise some concerns regarding bias in the measurement of the outcomes, and Nenonen et al. [35] was rated as being of high risk because of the inappropriate measurement and data reporting of the outcomes. Due to the lack of a registered protocol or the insufficient data on several secondary outcomes, six studies were considered to raise some concerns regarding reporting bias. Overall, seven studies were rated as having a low risk of bias, four were rated as raising some concerns, and one was rated as having high risk. The two funnel plots of DAS28 and CRP were substantially symmetrical (Supplementary Figure S1). The results of Egger's test for DAS28 and CRP were p value = 0.537 and p value = 0.383, respectively, indicating that there was no publication bias.
We first analyzed the ability of microecological regulators to relieve the symptoms of RA patients. The present study found significant benefits of microecological regulators intervention on DAS28 reduction, which was consistent with the findings of Zamani et al. [39] and Alipour et al. [44]. In fact, we further identified a strong DAS28 response in RA patients to probiotics after removing Pineda et al. (MD −0.25 (95% CI −0.42, −0.08), p = 0.003). This might be because of the strict inclusion criteria, which require patients to have at least four swollen and four tender joints at enrollment. This led to a small sample size of 29 patients and perhaps a failure to demonstrate the efficacy of probiotics as an adjunctive therapy within three months. Prebiotics, more specifically, polyphenol supplementation, showed a better impact on disease activity in our subgroup analysis (Supplementary Figure S2). These results were consistent with our previous knowledge that probiotics and prebiotics can alleviate joint inflammation in CIA models [47][48][49][50].
Concerning quality of life, some researchers have proposed that the use of the HAQ may better reveal the functional status of RA patients in comparison to physical examinations and laboratory indicators [40]. The limited number of included RCTs may make it difficult to specifically evaluate the quality-of-life impact, and a borderline significant improvement in HAQ scores was observed. However, no significant effect was observed in subgroup analysis. This result was consistent with a previous meta-analysis conducted by Lowe J et al. [51].
Concerning the inflammatory markers, participants subjected to microecological regulator supplementation showed a borderline significant reduction in CRP levels. However, the reduction in CRP levels (MD −1.82 (95% CI −3.29, −0.35), p = 0.02) may not represent a clinically meaningful change. The pooled result might be influenced by two trials that employed a more sensitive test (hs-CRP instead of CRP) and had larger sample sizes. Regarding changes in the ESR, some studies reported normal baseline values of the ESR, but they did not provide the data, and this information could not be extracted a posteriori. Moreover, levels of IL-1β showed a significant decrease in the probiotics group, which was consistent with studies conducted by Alipour et al. [44] and Khojah et al. [36].
Regretfully, due to the fairly limited number and high heterogeneity of eligible RCTs in respective subgroups, we were not able to select the most effective type of probiotics or prebiotics by comparing our present data. From the above analysis, however, we can propose that probiotics could be more effective than prebiotics for RA patients in adjunct with disease-centered treatment. In addition, in prebiotics, polyphenol supplementation is more likely to be used as adjuvant therapy. Synbiotics has no obvious advantage over the two in our meta-analysis.
Many studies have proposed a "gut-joint axis" and suggested that inflammation in the gut mucosa can precede joint manifestations [52]. The gut microbiota is related to RA etiology through several autoimmune pathways, such as the regulation of T helper and T reg cell functions and the induction of immune tolerance. Besides probiotics' local effect on gut health, such as diminishing harmful bacteria, data from animal and human studies revealed that probiotics modulate locally and systemically the immune system [11,[53][54][55]. Evidence from clinical and animal studies suggested SCFAs as possible mediators of these functions. SCFAs are also the main fermentation products of prebiotics by gut microbiota. It can influence the B lymphocyte's cellular proliferation, inhibit germinal center B cell, and plasmablast differentiation, as well as innate natural killer T (NKT) cell cytokine production [56,57]. Furthermore, animal and human studies have shown that prebiotics can improve immunity functions by increasing the population of protective microorganisms and decreasing the population of harmful bacteria by Lactobacilli and Bifidobacteria [58,59]. Beyond SCFAs, probiotics modulate the immune response by directly affecting the immune system. Concerning the innate immune system, probiotics can blind specific TLRs, affect downstream signaling, promote the expression of proteins that negatively regulate TLRs activity, and thus reduce inflammation induced by different pathogens [23,60]. T cells are essential to the adaptive immune response. Several animal studies of RA have shown that probiotics tend to generate a Treg immune response, promote the conversion of T cells into Tregs expressing the forkhead box transcription factor (FoxP3), and enhance the suppressive function of pre-existing Tregs [61,62]. The increase in anti-inflammatory and the decrease in pro-inflammatory cytokines are both associated with the upregulation of FoxP3-positive Treg cells.
We analyzed the variation in the outcome values before and after supplementation, as Mohammed et al. [30] and Sanchez et al. [63] did, thus eliminating differences in baseline data between the intervention and control groups. The strengths of our meta-analysis are that we compared variations in outcome measures between the two groups and included only RCTs which were rated as having a low or moderate risk of bias. Another strength of our meta-analysis is that we pooled only studies that were human RCTs with a detailed composition of microecological regulators. This makes our analysis results more concordant and easier to interpret. We extensively and comprehensively analyzed the effects of different types of prebiotics, probiotics, and synbiotics supplementation in RA patients, provided the effect of pro/pre/synbiotics clearly and separately based on the stringent definition and discrimination of all these three supplementations, and showed a great variety of microecological regulator treatment options in use.
The most significant limitation of the present study was the insufficient number of RCTs that were eligible for analysis and the high heterogeneity, which affects the conclusion of our study. On the other hand, the variation in data and the presence of incomplete data can also affect the reliability and validity of the results. The sample sizes across the included studies were generally small, and therefore, the clinical significance of outcome changes was insufficient. This systematic review of the literature provided very different baseline characteristics and inclusion criteria for RA patients. As the average disease duration was approximately 9 years in all studies, the current analysis did not provide information on patients with newly diagnosed RA.

Conclusions
All of the above findings suggest that intestinal microecological regulators have great potential to improve the outcome of established therapies in RA patients, especially prebiotics in improving symptom severity and probiotics may have a promising role in upregulating inflammatory markers such as CRP and IL-1β. Nevertheless, further studies are required to consolidate these effects and further investigate the efficacy and safety of microecological regulators in newly diagnosed RA patients, particularly preclinical RA patients.