Modified Si-Miao Pill for Rheumatoid Arthritis: A Systematic Review and Meta-Analysis

Objective The aim of this review and meta-analysis was to assess the effects and safety of modified Si-Miao pill (mSMP) in treatment of rheumatoid arthritis. Design A systematic literature search was carried out in eight databases from their available dates of inception to April 2020. After screening, fifteen randomized, controlled trials (RCTs) comparing the effects and safety of mSMP in combination with western medicine (including disease-modifying antirheumatic drugs (DMARDs) and nonsteroidal anti-inflammatory drugs (NSAIDs)) in treating rheumatoid arthritis patients were included after screening. Results In comparison with DMARDs, or coadministration of DMARDs and NSAIDs, mSMP in combination with western medicine significantly lowered erythrocyte sedimentation rate (mean difference (MD) = -10.61, 95% confidence interval (CI) [−12.19, −9.03]), C-reactive protein (MD = −6.50, 95% CI [−8.43, −4.56]), rheumatoid factors (MD = −17.31, 95% CI [−24.34, −10.27]), swollen joint count (MD = −1.63, 95% CI [−2.29, −0.97]), tender joint count (MD = −1.98, 95% CI [−2.34, −1.62]), and morning stiffness time (MD = −24.37, 95% CI [−29.41, 19.33]) and ameliorated the condition of patients (odds ratio (OR) = 3.69, 95% CI [2.64, 5.14]). Additionally, mSMP in combination with western medicine seemed safer (OR = 0.49, 95% CI [0.30, 0.81]). Conclusion The results of the meta-analysis study have shown that mSMP in combination with western medicine therapies appears to be more effective and safer than western medicine alone in the treatment of rheumatoid arthritis including reducing inflammatory markers and adverse events and improving symptoms. Howbeit, more high-grade, large-scale RCTs of mSMP in various countries and regions are still needed.


Introduction
Rheumatoid arthritis (RA), one of the most common autoimmune diseases, is characterized by symmetrical inflammatory polyarthritis and progressive joint destruction of unknown etiology. It affects approximately 1% of the population at any age, burdening the social economy for its high disability [1,2]. At present, immunosuppression and anti-inflammatory effects are the main mechanisms of drugs to treat RA. Disease-modifying antirheumatic drugs (DMARDs), including conventional synthetic DMARDs, biological DMARDs, and targeted DMARDs, are thought to be first-line drugs internationally [3]. Nonsteroidal antiinflammatory drugs (NSAIDs) aim to relieve the pain and inflammation rapidly. However, the various side effects of DMARDs and NSAIDs, such as leukopenia and gastrointestinal, are common and cannot be ignored [4]. Epidemiological studies have found out that patients with RA treated with anti-TNF antibody therapy were at an increased risk of serious infections and a dose-dependent increased risk of malignancies [5]. In addition, huge medical care cost of biological and targeted DMARDs makes it infeasible in many countries, and only 8.3% of patients have received biological DMARDs in China [6]. Furthermore, quite a few people are not sensitive to the current DMARDs. In order to treat RA more effectively and safely with less cost, it is necessary to explore new pharmacologic treatment for RA.
Traditional Chinese medicine (TCM), as a multicomponent and multitarget approach, has been proven to be effective in the treatment of RA in terms of reducing toxicity and increasing the efficacy of mechanisms [7,8]. Although the use of TCM is common in Southeast Asia, it is rare in other regions. Si-Miao pill (SMP) is primarily composed of Phellodendri Chinensis cortex, atractylodes rhizome, achyranthis bidentatae radix, and Coicis Semen. It has been widely used in the treatment of various arthralgia diseases, especially rheumatoid arthritis in China. According to modern pharmacology, the Si-Miao pill and numerous monomers extracted from it are also invalidated efficaciously in treating RA in vitro and in vivo [9]. e modified Si-Miao pill (mSMP) is derived from the Si-Miao pill, adjusting the composition according to different syndrome types for more effectiveness [10][11][12][13][14][15]. While mSMP has been applied in RA for a long time, a systematically evidence-based study is vacant. We assessed the effects and safety of mSMP combined with western medicine (DMARDs and NSAIDs) in the treatment of RA systematically by the meta-analysis of randomized controlled trials (RCTs).

Methods
is study was designed according to the Cochrane Handbook for Systematic Reviews of Interventions and Preferred Reporting Items for Systemic Review and Metaanalyses (PRISMA) guidelines to ensure accuracy and reliability [16]. e review protocol was registered in the PROSPERO database before the start of the review process (CRD42019133738). e PRISMA checklist was presented in Appendix 1 in Supplementary Materials.

Search Strategy.
Only Chinese and English articles were concerned. We conducted a systematic literature search including eight Chinese and foreign databases to ascertain trials including CNKI Databases, Wan Fang Database, Chinese Biomedical Literature database (CBM), PubMed, EMBASE, the Cochrane Library, Clinical Trails, and Web of Science. All of the databases were searched from their available dates of inception to April 2020. Search strategies were combined as follows. For the English databases, free text terms ((rheumatoid arthritis) OR (rheumatism)) AND ((si miao) OR (four subtleties)) were applied. For the Chinese databases, the terms were (si miao AND (lei feng shi guan jie yan OR lei feng shi guan jie yan (RA in Chinese)) NOT si miao yong and NOT si miao xiao bi).

Selection
Criteria. We adopted the following criteria: (1) studies used mSMP in combination treatment with western medicine (including DMARDs and NSAIDs); (2) all partaken patients who were diagnosed with RA according to the authoritative diagnostic criterion of RA such as the 1987 or 2010 American College of Rheumatology (ACR)/European League Against Rheumatism Criteria; and (3) RCTs.

Exclusion Criteria.
Exclusion criteria were as follows: (1) studies using mSMP alone, si miao yong an, and si miao xiao bi decoction were excluded; (2) case reports, reviews, and animal experiments were excluded; (3) diagnosis criteria are unclear; (4) studies without regulatory outcomes or studies in which the evaluation of curative effect is not standard were excluded; (5) articles that have no available full text were excluded.

Types of Outcome Measures.
e primary outcomes were the main indicators correlating with disease activity including sedimentation rate (ESR), C-reactive protein (CRP), rheumatoid factors (RF), swollen joint count, and tender joint count. e secondary outcomes were some clinical symptoms reflecting disease activity consisting of effective rate, morning stiffness times, and adverse events (AEs). e decrease of ESR, CRP, RF, morning stiffness times, swollen joint count, and tender joint count and the increase of effective rate could reflect the therapeutic effects (TEs). AEs included abnormal liver function, hyperleukocytosis, acratia, erythema or itch of the skin, dental ulcer, and gastrointestinal discomfort.

Data Extraction and Management.
e data were extracted by two independent reviewers (HW and YAH). Any discrepancy was resolved by consensus or judged by the corresponding author (SHT and ZC). All relevant data including characteristics of trails, the first author, year of publication, baseline characteristics of patients, number of patients, duration of the study, intervention methods, and outcome measures were extracted and entered into a template table. We assessed the included studies' quality on the basis of Cochrane Collaboration's risk of bias tool. ere were three scores for each item, low risk, unclear, and high risk, according to following criteria: (1) random sequence generation, (2) allocation concealment, (3) blinding of participants and personnel, (4) incomplete outcome data, (5) selective reporting, and (6) other biases.
2.6. Statistical Synthesis and Analysis. Review Manager 5.3 and Stata 12.0 were used to calculate the differences of effects and safety of mSMP in treating RA between the mSMP/experimental groups (mSMP treatment in combination with DMARDs that were combined or were not combined with NSAIDs) and the control groups (DMARDs that were combined or were not combined with NSAIDs). We applied the odds ratio (OR) and 95% confidence interval (CI) to appraise dichotomous data and mean difference (MD) to continuous data. Heterogeneity was evaluated according to the chi-square test and the Higgins I 2 test. e fixed-effects model was applied when statistical heterogeneity was low that was I 2 ≤ 50% or Chi 2 test P < 0.10; otherwise, a random-effects model was employed. e subgroup analysis was performed to eliminate heterogeneity, and sensitivity analysis was adopted to probe the source of heterogeneity. If the data provided by the included studies were not appropriate for performing a meta-analysis, the study data were presented in narrative form. Publication bias was detected by Egger's regression asymmetry test.

Study Selection and the Basic Documents.
e flow chart of the study selection was given in Figure 1. In the primary 2 Evidence-Based Complementary and Alternative Medicine screening, we retrieved 176 articles. After removing duplicates, titles and abstracts of 104 studies were screened. 27 records were excluded because they were a summary of experience or reviews (n � 27), animals or cell experiments (n � 15), or irrelevant diseases or medicines (n � 24). After the full-text reading of the resulting 38 studies, 20 records were excluded, seven of which were due to self-controlled studies and thirteen of which were owning to inconsistent interventions and four of which were due to no available raw data. Eventually, 15 RCTs were enrolled in the meta-analysis, including 1349 participants in total [17][18][19][20][21][22][23][24][25][26][27][28][29][30][31]. e characteristics of the studies were exhibited in Table 1. ey were all carried out in China and published between 2007 and 2019 in Chinese. All of them were conducted as single-center trials.

Quality Assessment of Included Studies.
Most of the included RCTs were of poor quality according to the Cochrane Collaboration's risk of bias tool criteria shown in Figure 2.
Eight [17-19, 21, 22, 25, 28, 31] of the included studies indicated random sequence generation, two [19,21] of which are based on treatment order, one [18] of which was based on sortition randomization method, and the remaining five [17,22,25,28,31] were based on the table of random numbers. None of the articles mentioned allocation concealment or blind method, as well as intentional analysis. All patients completed the experiments, and no one lost the interview or dropped out. None of the trials reported other biases.

Publication Bias.
Egger's publication bias test showed that there were negligible publication biases for four outcomes in terms of RF (P � 0.309), morning stiffness time (P � 0.065), swollen joint count (P � 0.092), and tender joint count (P � 0.734) while there were significant publication biases for four outcomes in terms of ESR (P � 0.001), CRP (P � 0.014), effective rate (P � 0.04), and AEs (P � 0.032) for the evaluation rules that studies with P values over 0.05 in Egger's test were deemed low heterogeneity. e effects of lowering ESR and reducing need further exploration, as presented in Appendix 2 in Supplementary Materials.   Figure 3(a).
Morning stiffness time was determined in five trials (involving 428 patients) [21,[27][28][29][30]. e number of trial participants ranged from 40 to 146, with the trial duration varying from 42 days to 90 days. As illustrated in Figure 4(b), there was statistical heterogeneity between the studies based on the random-effects model (I 2 � 76%, P < 0.00001). e experimental groups were superior to the control groups regarding the reduction of the morning stiffness time (MD � −24.37, 95% CI [−29.41, 19.33]).

AEs.
AEs were reported in eight trials (involving 651 patients) [17, 23-25, 27, 28, 30, 31]. As illustrated in Table 3, the number of cases ranged from 0 to 18. Gastrointestinal discomfort was the most common adverse reaction, of which 18 cases were reported in the control groups and 10 cases in the experimental groups (mSMP + western medicine) according to Chen, Li, Qian, Zeng, Zhao, and Zhang's studies.
ere were six cases of erythema or itch of skin reported in control groups (western medicine) and only two cases in experimental groups in Zhao's study. ree trials reported hyperleukocytosis with a total of 3 cases in the control groups and no case in the experimental groups, the same as a dental ulcer. Two cases of abnormal liver function were reported in the control groups and only one case occurred in the experimental groups, as illustrated in Figure 4(c). ere was no statistical heterogeneity between the studies based on the fixed-effects model (I 2 � 0%, P � 0.77). Compared to the control groups, the mSMP groups could reduce the side effects apparently (OR � 0.49, 95% CI [0.30, 0.81]).

Random sequence generation (selection biases) Allocation concealment (selection biases)
Blinding of participants and personnel (performance biases)

Blinding of outcome assessment (detection biases)
Incomplete outcome data (attrition biases)   Evidence-Based Complementary and Alternative Medicine 3.6. Sensitivity Analysis. In the sensitivity analysis, no significant changes of heterogeneity related to ESR, CRP, and swollen joint count were observed. In terms of RF and morning stiffness time, the study by Li [27]was most likely to be the sources of heterogeneity study. When it was omitted, the change of heterogeneity was significant (RF: I 2 reducing from 91% to 55%, morning stiffness time: I 2 reducing from 76% to 44%), and the general conclusion remained the same, as presented in Appendix 3 in Supplementary Materials.

Discussion
Although mSMP has been applied in treating RA for several years and there were massive clinical reports and clinical trials in China, it is the first time to evaluate the efficacy and safety of mSMP in the treatment of RA by systematic reviews and meta-analyses. Biomarkers play an important role in guiding the clinical trials and therapies of RA. ESR and CRP are the most common experimental indicators reflecting the inflammatory activity of RA, which is useful to evaluate the condition and prognosis. RF is an antibody correlated with RA and predicts bone erosion and severe disease progression [32]. It generally does not change much for a short time, but its change can explain the disease to a certain extent. ACR20 and criteria for diagnosis and efficacy of TCM diseases and syndromes are the most common and authoritative standards to evaluate the efficacy of drug therapy. Swollen joint count, tender joint count, and morning stiffness time are the characteristic manifestations of RA which played a great role in assessing the disease activity. According to the 2018 Chinese guideline for diagnosis and treatment of rheumatoid arthritis, ESR, CRP, RF, anti-cyclic citrulline antibody, swollen joint count, and tender joint count should be taken into account in the treatment of RA. Since there were few studies on the measurement of anti-cyclic citrulline antibody, ESR, CRP, RF, swollen joint count, and tender joint count were regarded as primary outcomes. is statistical analysis revealed that mSMP in combination with western medicine seemed to be more effective and significant in reducing the levels of ESR, CRP, and RF, ameliorating morning stiffness time, swollen joint count, tender joint count, and the incidence of AEs incidence. Due to the limited studies included, we could not conduct subgroups' analysis in terms of dosage, treatment durations, the difference of control drugs to remove the high heterogeneity of CRP, and swollen joint count. We speculated on several possible reasons, such as patients with different disease activities, different regions, and not enough samples, which needs more and better quality clinical trials to explore. Furthermore, although the included article did not analyze whether the adverse events were caused by mSMP or western medicine in experimental groups, mSMP combined with western medicine had fewer adverse events than western medicine. RA belongs to "bi" in traditional Chinese medicine, while a large number of active RAs vest in "dampness-heat of bi" [7]. TCM syndrome of most of the patients included in the RCTs was dampness-heat. SMP was written by Zhang Bingcheng in    with precision medical treatment. On the other hand, the difference in composition and dosage of the prescriptions is also a disadvantage of standard evaluation. mSMP in the treatment of RA has also been confirmed according to modern pharmacology in vivo and in vitro. SMP was found to significantly inhibit the expression of IL-1β, IL-6, and TNF-α in adjuvant arthritis (AA) rats [34]. mSMP extract inhibited the release of inflammatory mediators, like NO and TNF-α, via the suppression of ERK and NF-κB-dependent pathways from lipopolysaccharide-stimulated mouse macrophages [14]. Wang found that SMP could downregulate the expression of VEGF in AA rats, thus inhibiting the formation of synovia pannus [35]. In addition, cortex Phellodendri, rhizoma atractylodes, and Semen Coicis had obvious anti-inflammatory and analgesic effects. Total saponins of Achyranthes bidentata could attenuate the acute inflammatory reaction and regulate immunity [36,37]. Moreover, SMP and one of the main components, berberine, were confirmed to regulate the blood lipids, increase the level of high-density lipoprotein, which is thought to improve rheumatism and benefit the cardio-cerebrovascular disease in terms of the common view that patients with rheumatoid arthritis are at increased risk of cardiovascular disease [38][39][40]. In condition, mSMP may lower CVD occurrences by lipid regulation and well work in hormonal-dependent RA [41]. One trial reported that mSMP could lower the recurrence rate during the one-year follow-up period.
ere were also a lot of shortcomings in this study. First, all the included RCTs were conducted in China and might cause selection bias. Most of the included RCTs were of poor quality according to the Cochrane Collaboration's risk of bias tool criteria for a missing message of allocation concealment, blind method, and intentional analysis. Second, the form, compositions, dosages, and treatment duration of mSMP were complex and changeable, which might undermine the credibility. Furthermore, the studies included in the systematic review were few in number and there was no unified standard for disease evaluation. Finally, the lack of relevant grey literature may lead to publication bias.

Conclusions
Overall, mSMP combined with western medicine was more effective and safer than western medicine in the treatment of RA. mSMP may play the pharmacological action by antiinflammation, regulating immunity, analgesia, and lipid regulation. In consideration of the low quality, single area of the given trials, and variable interventions, more large randomized controlled, double-blind, multicenter clinical trials with good methodological quality are needed to recommend mSMP as an alternative remedy for RA.

Conflicts of Interest
ere are no conflicts of interest.