The Derivative of Tripterygium wilfordii Hook F—Kunxian Capsule, Attenuated Rheumatoid Arthritis: A Systematic Review and Meta-Analysis

The study aimed to explore the efficacy and safety of Kunxian Capsule (KXC) in the treatment of rheumatoid arthritis (RA). The randomized controlled trials (RCTs) comparing the effects of KXC in patients with RA were included in this study. Weighted mean differences (MDs) were calculated for net changes by employing Review Manager meta-analysis software. Nine RCTs were included in the systematic review with a total of 747 patients. The overall effects showed that KXC alone or combined with disease-modifying antirheumatic and drugs decreased tender joint counts (P=0.02, MD = −1.07, 95% CI: −1.95 to −0.18), shortened duration of morning stiffness (P < 0.0001, MD = −9.01, 95% CI: −13.08 to −4.93), lowered erythrocyte sedimentation rate (P < 0.00001, MD = −5.27, 95% CI: −6.78 to −3.77), and reduced C-reactive protein (P < 0.0001, MD = −5.04, 95% CI: −7.28 to −2.80). The most common adverse events were gastrointestinal disturbances and abnormal liver function. These results suggest that KXC is likely to be a more effective and safe candidate for treating RA compared with conventional therapies.


Introduction
Rheumatoid arthritis (RA) is an autoimmune disease inducing synovial inflammation, which is characterized by tenderness, swelling, morning stiffness, joint destruction, and deformity [1,2]. e age-standardized prevalence and incidence rates are increasing, especially in countries such as Canada, Paraguay, and Guatemala. To reduce the ongoing burden of this condition, early diagnosis and treatment of RA are vital, especially among female patients [3]. e conventional therapies include nonsteroidal antiinflammatory drugs (NSAIDs), glucocorticoids, and diseasemodifying antirheumatic and drugs (DMARDs) [4,5]. With the exploration of molecular mechanisms, biological agents have been wildly applied in the treatment of RA [6,7]. However, owing to the adverse effects of conventional medications and high financial stress of biologicals [8], more effective and safe medications for RA are needed to be explored.
Tripterygium wilfordii Hook F (TWHF) has been applied in the treatment of RA in China for decades [9]. e previous research showed that TWHF could be as effective as synthetic DMARDs in the treatment of RA and the common adverse effects (AEs) were gastrointestinal discomfort, menstruation disorders, and amenorrhea [10]. Kunxian Capsule (KXC) is a derivative of TWHF and has little liver injury and reproductive toxics compared with TWHF [11]. Meanwhile, KXC has been employed in treating RA in clinical trials for a long time [12][13][14]. However, most of these clinical data origins from uncontrolled clinical trials or retrospective reports, and few multicenter clinical trials have been performed to verify the effects of KXC in the treatment of RA. erefore, the scientific evidence confirming the efficacy and safety of KXC in treating RA remains unclear. Given these uncertainties, it is essential to perform a quantitative meta-analysis of the efficacy and safety of KXC in patients with RA. Consequently, we carried out the metaanalysis and systematic review.

Materials and Methods
e Preferred Reporting Items for Systemic Review and Meta-Analyses (PRISMA) statement [15] was applied to design and report the study.

Search Strategy.
e following English databases were retrieved to confirm trials: PubMed, the Cochrane Library, and Clinical Trials.gov. In addition, the Chinese databases, such as the CNKI Database, Wanfang Database, VIP Database, CBM Database, and Chinese Clinical Trial Register were searched. All the databases were retrieved from their available dates of inception to the latest issue (June 2019).
Different search strategies were combined as follows. For the English databases, free text terms were employed, such as "kunxian capsule" and "rheumatoid arthritis" or "RA". For the Chinese databases, free text terms were applied, such as "kunxian jiaonang" (which means Kunxian Capsule in Chinese) and "lei feng shi guan jie yan" (which means rheumatoid arthritis in Chinese). A filter for clinical trials was selected for English databases. To include an adequate number of trials, the reference lists of pertinent publications were also retrieved to identify additional studies.

Selection Criteria.
Randomized controlled trials (RCTs) were included irrespective of blinding, publication status, or language. Studies were selected for reanalysis if they met the following criteria: (1) patients enrolled were diagnosed with RA, according to the 1987 or 2010 RA guidelines of the American Rheumatology Association; (2) in the intervention group, the participants took KXC alone or with other DMARDs; (3) in the control group, the participants took DMARDs; (4) duration was at least 12 weeks; and (5) outcomes were available.
Case reports, reviews, in vitro experiments, retrospective studies, or studies without control groups were excluded. For repeat studies, the authors were contacted to illuminate any ambiguities. e RCTs that lacked crucial data to allow for the calculation of net changes in outcomes and their variances from baseline to the endpoint were also excluded from the research. e literature was selected by two reviewers (YF Liu and Z Zhang) independently. e flowchart of the study selection has been generated in accordance with the PRISMA statement.

Data Extraction and Management.
e data were extracted by two independent reviewers (Zhang and Chen), and contradictions were resolved by consensus or were judged by another reviewer. e studies' quality was evaluated according to the Cochrane handbook 5.3. e risk of bias of the included trials was evaluated by the following items: (A) random sequence generation (selection bias); (B) allocation concealment (selection bias); (C) blinding of participants and personnel (performance bias); (D) blinding of outcome assessment (detection bias); (E) incomplete outcome data (attrition bias); (F) selective reporting (reporting bias); and (G) other bias. e primary outcomes were tender joint count (TJC), swollen joint count (SJC), and duration of morning stiffness (DMS). e secondary outcomes included erythrocyte sedimentation rate (ESR), C-reactive protein (CRP), rheumatoid factor (RF), and anticyclic citrullinated peptide (CCP). Adverse events were also collected from the studies. For the trials that applied a three-armed group design, the outcomes of the groups were extracted if they satisfied the inclusion criteria. In terms of vagueness or absence in the publications of the outcomes, the authors were connected and related data has been extracted by consensus if the authors were unavailable.

Data Synthesis and Analysis.
e effects of KXC on patients with RA were calculated as differences between the KXC group and the control group, utilizing the Review Manager meta-analysis software, version 5.3. To guarantee the credibility of the results, the net changes in all the outcomes were analyzed as the mean differences (KXC minus control) in changes (endpoint minus baseline) for parallel trials. e weighted mean differences and 95% confidence intervals (CIs) were calculated for continuous data. Heterogeneity was estimated through the chi-square test and Higgins I 2 test. A fixed-effect model was applied when the studies were sufficiently alike (P > 0.10); otherwise, a random-effects model was applied. For the subgroup with high heterogeneity, a random-effects model was performed. A Z score was calculated to determine the overall effect, with a significance set at P < 0.05. Publication bias was analyzed by funnel plot if the number of the included studies ≥10.
Two subgroup analyses were performed to diminish the clinical heterogeneity: KXC compared with DMARDs and KXC plus DMARDs compared with DMARDs.

Study Selection.
e process of study selection was shown in Figure 1. After filtering, 5 animal experiments, 6 reviews, 4 studies without sufficient data, 2 studies with duplicated data, and 4 studies whose control intervention did not meet the inclusion criteria were excluded. According to the selection criteria defined in the Materials and Methods, 9 RCTs were included in the meta-analysis. e three-armed group design was applied in two of the studies. To prevent sample duplication, the data of the two groups were included [16,17]. In one study, there were three groups: leflunomide alone, leflunomide plus KXC, and leflunomide plus methotrexate [16]. e data of leflunomide plus KXC versus leflunomide alone was extracted [16]. One study reported that KXC monotherapy was compared with methotrexate monotherapy or KXC plus methotrexate [17]. e data of KXC 2 Evidence-Based Complementary and Alternative Medicine monotherapy versus methotrexate monotherapy was extracted [17]. In the KXC monotherapy subgroup, three studies were included [17][18][19]. In the combined therapy subgroup, six studies were included [16,[20][21][22][23][24]. e characteristics of the studies were indicated in Table 1. Together, 747 RA patients were included in the study.

Study Descriptions.
e included studies were published as full texts between 2010 and 2019. All the RCTs were performed in China. All the studies were published in Chinese. Eight studies were conducted as single-center trials, and only one study was a multicenter trial [17].
Except for one study with a duration of six months [19], all the other studies were 12 or 24 weeks. e outcomes were measured at 4 weeks and 12 weeks in two studies, and the outcomes of 12 weeks were extracted [17,18]. e outcomes were determined at 4 weeks, 8 weeks, and 12 weeks in one study, and the results of 12 weeks were extracted [22]. e outcomes were detected at 4 weeks, 8 weeks, 12 weeks, and 24 weeks in one study, and the results of 24 weeks were extracted [16]. Figure 2, all the included trials were of low quality due to unclear randomization, deficient allocation concealment, inadequate blinding, and undescribed withdrawals and dropouts.

Publication
Bias. As the number of included studies <10, the funnel plot was not conducted to estimate the publication bias.

Effects of Interventions
3.6.1. e Outcome of TJC. As illustrated in Figure 3, only one study compared KXC with DMARDs in terms of TJC (involving 159 subjects), and there was no significant   Figure 3, only one study compared KXC with DMARDs in terms of SJC (including 159 subjects), and there was no significant difference between the two groups (P � 0.78, MD � −0.14, 95% CI: −1.11 to 0.83). ree studies compared KXC plus DMARDs with DMARDs regarding SJC (including 160 subjects) and there was a significant difference between the two groups (P � 0.02, MD � −1.42, 95% CI: −2.57 to −0.27). Moreover, the overall effect (involving 319 subjects) demonstrated that there was no significant difference between the two groups (P � 0.08, MD � −0.67, 95% CI: −1.42 to 0.07). Figure 3, two studies compared KXC with DMARDs with regard to DMS (involving 219 subjects) and there was a significant difference between the two groups (P � 0.0003, MD � −8. 28 Table 2, the adverse events were reported in six trials (involving 97 patients), 44 patients in KXC group, and 53 patients in the control group, respectively. e frequent adverse events were gastrointestinal discomfort and abnormal liver function.

Discussion
Complementary and alternative medicine has been extensively employed in the treatment of inflammatory arthritis       Evidence-Based Complementary and Alternative Medicine [25]. KXC consists of four individual herbs which adhere to the compatibility principle of traditional Chinese medicine: Tripterygium hypoglaucum, Herba Epimedii, Fructus Lycii, and Semen Cuscuta. e main component of the KXC is TWHF, which derives from Tripterygium hypoglaucum [26,27]. Herba Epimedii was widely used in the clinic to enhance bone fracture healing [28]. Fructus Lycii had many pharmacological and biological activities, such as anti-inflammation, antioxidation, antiapoptosis, immune regulation [29]. Semen Cuscuta had a protective effect on the reproductive system [30]. In a word, KXC, as an upgrade of TWHF, served as strengthening the therapeutic effect and weakening the side effects of TWHF [31][32][33][34].
ere were two meta-analyses of KXC in the treatment of RA [35,36]. Nonetheless, no review protocol was registered. However, one of the meta-analysis lacked data and forest plots [35]. e other meta-analysis did not have strict inclusion criteria [36]. Our analysis included more samples, performed rigorously screening, and had two subgroup analyses.
e results demonstrated that KXC had significant effects in alleviating DMS, ESR, and CRP, whether in two different subgroups or as a whole. ese results were consistent with previous meta-analyses [35,36]. Moreover, TJC was reduced in KXC group, which was similar to the previous meta-analysis [35].
In addition, the study also demonstrated that KXC plus DMARDs group could reduce SJC and RF, but KXC alone or the overall effect of KXC did not lower SJC and RF. However, the previous analyses suggested that KXC could reduce SJC and RF [36]. We considered that the number of RCTs treated with KXC alone is small, which could lead to the difference.
Our study indicated that KXC had no effect on CCP, and previous meta-analyses did not analyze the outcome CCP. e CCP is a diagnostic parameter that cannot reflect disease activity; thereby KXC could not reduce CCP.   Lin et al. [17] Lu et al. [23] Quan and Wang [24] Tian [16] Lu [19] Total events

Experiment Control Experiment Control Experiment Control Experiment Control Experiment Control Experiment Control
Number of patients showed that KXC had immunosuppressive, anti-inflammatory, and analgesic effects [37]. Animal experiments demonstrated that KXC downregulated the expression of IL-8 and cIP-10 mRNA, reduced synovial inflammation, alleviated the formation of pannus, decreased joint damage and fibrous tissue proliferation, and promoted tissue repair in the synovium of rats with collagen induced arthritis [38][39][40]. Additionally, triptolide influenced CD4+ and CD8+ cells distribution in Peyer's patch of DA rats with collagen induced arthritis [41].
However, some limitations of this systematic review should be mentioned. First, all the included trials were conducted in China, which inferred a high risk of selection bias. e results could not be applied to other regions in the world. Second, the majority of the studies published in Chinese were of poor quality. Only one multicenter RCT was confirmed. ird, the limited number of trials (from two to six) and participants included in each subgroup obscured the supportive evidence of KXC for RA. Fourth, the heterogeneity included in each subgroup was also obvious, especially in the outcomes CCP and RF. We consider that the differences in the quality of reports, intervention methods, doses, and duration of treatment lead to heterogeneity. Finally, the most important outcomes (ACR 20, ACR 50) were not reported except for two studies [17,22]. Considering this, all the results should be carefully interpreted.

Conclusion
In conclusion, KXC is effective in the treatment of RA through lowering TJC, DMS, ESR, and CRP. KXC may function as anti-inflammation as well as immunosuppression. In a word, KXC, as an upgrade of TWHF, serves as enhancing the efficacy and weakening the side effects of TWHF. KXC, as a novel "phytoimmunosuppressant," is promising in the treatment of RA and other autoimmune diseases. Considering the low quality of the included trials, more well-designed RCTs are needed before we can recommend KXC to replace or combine with conventional therapies.

Data Availability
e data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
All authors state that there are no conflicts of interest.