Protective Effect and Possible Mechanisms of Tripterygium Glycosides in Patients with Ankylosing Spondylitis: A Systematic Review and Meta-Analysis

Objective The safety and efficacy of Tripterygium glycosides (TG) were assessed for ankylosing spondylitis (AS) in accordance with the existing literatures. Materials and Methods Electronic literature was searched from Chinese VIP databases, Cochrane Library, Chinese Biomedical Literature Database, Wanfang Web of Science, EMBASE, Chinese National Knowledge Infrastructure, and the PubMed for the studies with the publication from the beginning to December 2021. Randomized controlled trials (RCTs) were included only. The major variables of result comprised erythrocyte sedimentation rate (ESR), C-reactive protein (CRP), Spinal Pain Visual Analog Score (SP-VAS), Bath Ankylosing Spondylitis Functional Index (BASFI), and Bath Ankylosing Spondylitis Disease Activity Index (BASDAI). Moreover, the secondary variables of result covered the overall clinical effective rate following the adverse drug reaction (ADR). We carried out the meta-analysis with the use of STATA 12.0 and RevMan 5.3. We used GRADE pro3.6.1 software to assess the quality of evidence. Results In general, we covered 15 randomized controlled trials with the focus of 1186 patients. As proven by our meta-analysis, TG as adjuvant therapy or monotherapy decreased the BASDAI, BASFI, SP-VAS, serum CRP, and ESR than control in patients suffering from AS. Additionally, TG treatment visibly improved the overall effective rate in AS. Nevertheless, TG was not found to significantly increase the rate of ADR in contrast to the control. Conclusion As indicated by our result, TG may be an option to treat AS. In this paper, we recommended strict trials with high quality and large samples sizes for confirming the finding here.


Introduction
Ankylosing spondylitis (AS) refers to a chronic disease that can cause inflammation. AS can result in damage of structures and inflammation, mostly occurring within the spine or the sacroiliac joint, which normally leads to patients' morning stiffness and chronic back pain and thus causes spinal immobility and ankylosis [1,2]. In some severe patients, it ultimately could lead to the rigid and completely fused spine. "Bamboo spine" is formed based on ossification within the fibrous ring's external fibers of intervertebral discs, forming adjoining vertebra's marginal syndesmophyte [3]. Besides, it frequently gives rise to extra-articular aggravation such as inflammatory bowel disease, psoriasis, and uveitis [4]. Normally, young men with 20 to 30 years old are more prone to AS [5]. The estimation of over 90% genetic role was achieved, the most significant correlation with human leukocyte antigen (HLA)-B27 [6]. Nevertheless, HLA-B27's specific pathogenic effect is still elusive though many hypotheses are proposed.
Currently, AS treatment strategy has been optimized remarkably in past decades by employing tumor necrosis factor-(TNF-) specific agents, disease-modifying antirheumatic drugs (DMARDs), nonsteroidal anti-inflammatory drugs (NSAIDs), and some biologics [7]. Nevertheless, under NSAIDs therapy in the long run, the possible risk covering risk increases in the cardiovascular system, the gastrointestinal tract, and kidney [8]. DMARDs, including sulfasalazine (SSZ) and leflunomide (LEF), have been rarely recommended for treating axial spondyloarthritis, since they have no effect on axial spondyloarthritis, while playing limited roles in treating peripheral manifestations when coexisting with the axial disease [9]. Using DMARDs in a long term would have negative effects, for instance, anaphylaxis, gastrointestinal reaction, liver injury, and leukopenia, thus, limiting their use in clinical practice. TNF inhibitor treatment discontinuation within patients usually leads to 84.6% of partial remission and a relapse within 79.2%, termed responders, in accordance with ASAS40 criterion [10]. Moreover, the huge cost is also a disadvantage of biologics, which should be taken into consideration. As a result, novel therapeutic options or agents are needed for AS treatment.
In recent years, the rising application of alternative and complementary medicine, covering Chinese herbal medicine for AS treatment, arouses wide concern [11,12]. Tripterygium wilfordii Hook, a woody vine pertaining to the Tripterygium genus, refers to a traditional Chinese medicine having anti-inflammation, antirheumatism, and immunomodulation impacts [13]. Tripterygium glycosides (TG), the main ingredients of Tripterygium wilfordii Hook, were employed in China to treat inflammatory diseases in the long run (e.g., AS, chronic nephritis, a wide variety of skin disorders, and rheumatoid arthritis) [14]. TG can achieve antiinflammatory effect, collateral dredging effect, swelling subsidence effect, detoxification, dampness elimination effect, wind dispelling effect, and inhibition effect on humoral and cell immunity [15,16]. TG has been demonstrated with the improvement effect of the clinical characteristics and the regulation effect of AS patients' serum biomarker [17]. Over the past few years, increasing randomized controlled trials with high quality proved TG safety and effectiveness in terms of AS treatment. Nevertheless, there has been rare meta-analysis or systematic review on TG safety and efficacy for AS treatment. For this reason, a systematic review and meta-analysis covering randomized controlled trials with high quality were conducted for assessing TG safety and effect for treating AS.

Methods
We conducted a systematic review and meta-analysis in accordance with the AMSTAR (assessing the methodological quality of systematic reviews) [18] and PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. There are no protocols preregistered for this review. This paper was registered in Research Registry (https://www.researchregistry.com/), with reviewregis-try1276 as the registration number. Major personal data was not acquired, so we did not need ethical approval. Some methods in this section referred to our previous study [19].  AND "qiang zhi xing ji zhu yan (i.e., ankylosing spondylitis in Chinese)" AND "sui ji dui zhao shi yan (i.e., randomized controlled trial in Chinese)". The detailed search strategy used for English databases and Chinese databases was provided as supplementary material (available here).

Types of Controls.
The control was given western conventional medication. We excluded studies involving controls of Chinese herbal medicine treatments.  (5) CRP (C-reactive protein). ADR (adverse drug reaction) and the clinical effective rate (ER) were the secondary outcomes.

Types of Studies.
We only involved randomized controlled trials investigating the safety and efficacy of Tripterygium glycosides (TG) for AS with no limit of publication states or languages. If this paper searched an article that contained 3 treatment arms, we merely extracted data for the control arm (s) as well as the arm (s) that entails TG. Quasi-randomized trials with researches allocating subjects were excluded according to the admission number order and the date of birth.

Exclusion Standards.
(1) Nonrandomized controlled trial; (2) randomized controlled trials in which the patients were not reported to suffer from AS; (3) the combination of TG treatment and other drugs; (4) consistent animal experiments, reviews, and abstracts; (5) the duplication of publications and randomized controlled trials containing not sufficient data.
2.4. Literature Selection. In this paper, we used the flow diagram of PRISMA to choose the involved articles. The result of literature was introduced to the Endnote X7 software. 2 2 Oxidative Medicine and Cellular Longevity authors evaluated the potential eligible studies independently after they screened the title and abstract for the removal of duplication and not relevant research or the randomized controlled trials that were not involved in the inclusion standards. Afterward, we obtained the full text of the rest potential researches and reviewed these researches. A third independent investigator would resolve the disagreement between the first 2 authors.
2.5. Data Abstraction. 2 reviewers searched the data independently. Next, the uniformity was examined by a third independent reviewer. We used a standard form containing the retrieved items, which covered the general information of studies: the author's (s) name(s), publication date, design of study, patient gender and age, number of sample, intervention approach covering TG alone, or supplement western conventional medication and course of treatment. For continuous outcomes, we extracted mean and standard deviation (SD) as well as participant number of each research. For dichotomous outcomes, the total number and the number of events of both groups were extracted. Under probable conditions, we recalculated the data according to other types of form, as an attempt to carry out pooled analysis. Discussions would be conducted to solve the disagreement of the mentioned 2 reviewers. If necessary, we established contact with the involved studies' authors to acquire the missing data or additional data.
2.6. Evaluation of the Quality of Included Studies. With the use of the Cochrane collaboration tool, 2 authors independently evaluated the method quality and the bias risk in the included randomized controlled trial studies [21]. The above Cochrane tool is capable of evaluating incomplete result data, selective result reporting, result evaluation blinding, blinding of subjects, allocation concealment, randomization, and other types of biases in terms of the respective item. It can also classify studies into high risk of bias, low risk of bias, or not clear level of bias.

Evidence Quality Assessment.
We evaluated the quality of the evidence according to the GRADE [22,23]. We classified the meta-analysis results into high evidence quality, moderate evidence quality, low evidence quality, or very low evidence quality. First, we classified the randomized controlled trial results into evidence with high quality. The quality of each result decreased due to indirectness, publication bias, inconsistency, imprecision, and risk of bias. We adopted GRADE pro3.6.1 for the investigation and synthesis of data.

Statistical
Analysis. The data were collected and input into the STATA software (V.12.0; StataCorp, College Station, TX) for meta-analysis. The bias risk assessment was evaluated by RevMan software version 5.3 (Cochrane Collaboration, Oxford, UK). For continuous outcomes, mean, standard deviation (SD), and sample number of each group were extracted. For dichotomous outcomes, the total number and the number of events of both groups were extracted. Standard mean difference (SMD) with 95% confidence interval (CI) was calculated to analyze continuous variables (ESR, CRP, SP-VAS, BASFI, and BASDAI). Relative risk (RR) with 95% CI was calculated for dichotomous variables (overall effective rate and ADR). Statistical heterogeneity among trials was assessed using the chi-squared test and I 2 statistic. The heterogeneity among studies was identified if the I 2 was greater than or equal to 50% or P value was less than or equal to 0.05. A random-effects model was applied when heterogeneity was detected or the statistical heterogeneity was high (P < 0:05 or I 2 > 50%), and then, further subgroup study and meta-regression analysis (the number of included studies was more than 8) were performed to detect the origin of heterogeneity. Otherwise, a fixed-effects model was used (P ≥ 0:05 or I 2 ≤ 50%). Publication bias was assessed by using Begg's and Egger's linear regression test in the metaanalysis, and we considered P < 0:05 to be significant. Stabilities of synthetic results were evaluated with a visual assessment of sensitivity analyses. The method of omitting each study in sequence was used for sensitivity analysis.

Subgroup Analysis.
We carried out subgroup investigation in accordance with the course of treatment, sample size, age, control medication, and TG dosage in Table 2 because of the significant heterogeneity of SP-VAS and BASDAI result in our paper. Nevertheless, according to the results of subgroup investigation, the mentioned factors were not the prominent heterogeneity source in terms of BASDAI and SP-VAS.
3.6. Publication Bias and Sensitivity Investigation. In this paper, we performed the Begg's test and Egger's test ( Figure 12) for examining ADR's possible publication bias within this meta-analysis. Consequently, according to the P For establishing the impact arising from each involved studies relating to BASDAI, BASFI, CRP, ESR, ADR, and ER pooled data to demonstrate the findings here to be robust, the sensitivity was evaluated through the exclusion of one study at a time and computing the pooled data for other randomized controlled trials. According to the sensitivity investigation results, no noticeable effect was exerted on pooled data when the respective study was eliminated, respectively, thereby indicating the finding here to be relatively robust (Figure 13).

GRADE Assessment.
We used the GRADE method for exploring the results' evidence quality, with moderate or low methodological issues and heterogeneity, as well as significantly low quality. According to Table 3, the evidence quality of 5 results was low, one result had very low evidence quality, and another result had moderate evidence quality. Thus, the overall evidence quality of this study was low.

Discussion
AS, the axial skeleton's chronic inflammatory disease, leads to the rigid and completely fused spine [39,40]. Medications (NSAIDs and DMARDs) used to treat AS have been used in clinical practice for many years [8]. However, the diverse adverse reaction such as anaphylaxis, gastrointestinal reaction, liver injury, and leukopenia limit their longterm use. Tumor necrosis factor blockers refer to a promising drug to treat AS. Nevertheless, tumor necrosis factor blockers are likely to lead to some significant side effects, covering neurological issues and reactivating latent tuberculosis [41]. Therefore, safe and effective anti-AS drugs are urgently required. With the progress of clinical practice and research, increasing evidence reveals that Chinese herbal medicine has prominent effects in preventing and treating AS [14]. TG, extracted from Traditional Chinese Medicinal plant Tripterygium, has been used in China and other Asian countries for the long-term treatment of inflammatory diseases covering AS, chronic nephritis, different skin disorders, and rheumatoid arthritis [15]. According to modern pharmacological research, TG can achieve anti-inflammatory effect, collateral dredging effect, swelling subsidence effect, detoxification, dampness elimination effect, wind dispelling effect, and inhibition effect on humoral and cell immunity [15,16]. Considerable randomized controlled trials have demonstrated that TG was advantageous in treating AS [27,31,34], whereas no high-quality systematic review and meta-analysis with large sample size on the safety and efficacy of TG to treat AS has been demonstrated. Thus, a systematic review and meta-analysis was carried out for the investigation of whether TG has safety and efficacy in patients suffering from AS to evidence clinical practice and scientific study.

Summary of Evidence.
The current paper has been the initial systematic review and meta-analysis investigating whether TG monotherapy and TG supplied with western medicine have efficacy and safety to treat AS. On the whole, fifteen high-quality randomized controlled trials involving 1186 individuals with AS were included in the investigation. According to the major finding of existing systematic reviews and meta-analyses, using TG as adjuvant therapy or monotherapy for AS treatment evidently decreased BAS-DAI, BASFI, and SP-VAS, thereby demonstrating that TG 7 Oxidative Medicine and Cellular Longevity can effectively relieve the pain that is caused by AS. Next, the serum CRP and ESR were lowered by TG. In addition, TG treatment visibly elevated the overall effective rate in AS. Yet, there was no significant difference in the incidence of adverse reactions between the 2 groups, thereby demonstrating that TG has good safety and tolerability for AS patients. Thus, supporting evidence was provided in our paper, showing that TG may be highly recommended for planned use in patients suffering from AS.

Comparison with Existing
Studies. Some systematic reviews and meta-analyses demonstrated TG's efficacy and safety in AS treatment. As revealed by a meta-analysis [42] constituting 14 randomized controlled trials involving 996 patients, total effective rate, BASDAI, BASFI, and CRP showed no significant difference between the TG group and the western medicine group, inconsistent with the results of this paper. In their study, only 6 studies comparing TG with western medicine, and the rest 8 studies were compared with Chinese herbal compound. However, the studies with Chinese herbal compound control were excluded because numerous varieties of Chinese herbal compound of the control may result in high heterogeneity, and only the studies with western medicine control were involved. Thus, the sample size of this paper was significantly larger than theirs, which may be major cause of contradiction. Besides, part of the finding here was consistent with the data of Li et al. [3]. They performed a meta-analysis involving 11 trials that focused on 807 patients and exhibited that TG treatment obviously decreased the pain index, ESR in AS.
Nevertheless, the CRP was not reduced in their study. The data here have no consistency with the findings of existing researches in several fields, probably correlated with (1) the included studies in the existing meta-analyses varied considerably in quality. Nonetheless, herein, randomized controlled trials having a risk of bias score ≥ 4 on the basis of the Cochrane RoB tool were included, implying that only high-quality randomized controlled trials were included in this paper. (2) The diversity and complexity of Chinese herbal compound in the control in the existing studies are also major causes of the difference. Nevertheless, herein, we focused particularly on the single western medicine in the control, which can reduce the risk of heterogeneity. (3) AS is a chronic condition with diverse stages. The AS distinct stages can affect the progress of the disease along with the treatment response. However, many existing studies and this paper have not reported the stages of AS, which may also lead to contradictions. (4) The prevalence of AS in young men is 10 times higher than that in women. Therefore, the gender ratio difference between existing studies and this paper also affects the research results. (5) The course of treatment may affect the outcome. The duration of the current and current studies was 8-24 weeks. It is a paradox that too short a course of treatment may lead to poor efficacy, while too long a course of treatment may lead to serious adverse reactions.

Strengths.
The advantages of our meta-analysis studies include a clearly defined study issue, namely, reduced selection of randomized controlled trials with consistency,  Oxidative Medicine and Cellular Longevity fidelity, and bias. A precise study methodology was designed prior to the meta-analysis and an in-depth search of the literature was conducted. 2 researchers evaluated the protocols for the input data components, as well as the quality control of all data. All of the studies covered in herein constituted randomized controlled trials with an extraordinary number being high quality, which contributes to overcoming the drawbacks of the recall or selection bias in terms of nonrandomized researches. In addition, the number of trials and overall sample size was large (15 trials, 1186 participants). Subgroup and meta-regression assessments were carried out for determining the source of heterogeneity. Therefore, no publication bias was reported in this meta-analysis, and sensitivity estimates suggest that the results of the present meta-analysis have relative robustness.

4.4.
Limitations. However, this paper had some limitations. First, no protocol has been preregistered for our paper, probably leading to potential bias to our paper. Second, although randomized controlled trials were covered, the covered major studies had some intrinsic and methodological shortcomings: (1) only 10 trials reported information generated by randomization.
(2) Some studies did not report the blinded process or had unclear results, which might lead to unintentional or intentional biases to their results and reduce the credibility of their research conclusions. Three blinds are required in further trials. Third, AS is a chronic disease that requires lifelong treatment. The long-term safety and efficacy of drugs are the key to determine the clinical efficacy of therapeutic drugs. However, the treatment period of our paper was between 8 and 24 weeks. The long-term safety of TG was not found for AS because the duration of treatment of the covered studies was short, and no dropouts were revealed in a considerable number of the covered studies. Fourth, the dosage, administration methods, and course of TG treatments differed remarkably in the major randomized controlled trials. This clinical heterogeneousness might jeopardize the feasibility of the results of this paper. Fifth, we searched researches published in Chinese or English repositories only, so the potentially relevant randomized controlled trials published in other languages might be excluded. Besides, all randomized controlled trials covered here were made in China, a potential limitation to the generalizability of the results of this paper. Sixth, our paper had low quality of evidence due to the high risk of bias and inconsistency. Therefore, it is necessary to further carry out    Oxidative Medicine and Cellular Longevity controlled trials are urgently needed. In the future, we recommend guidelines, covering CONSORT 2010 statement [43], that are needed to establish and report on randomized controlled trials of TG. Second, although TG treatment was found to be safe for AS patients in the analyzed studies, the safety of TG on AS remains to be further studied. The

11
Oxidative Medicine and Cellular Longevity standard reporting format of ADR was developed [44], and it was suggested to pay close attention to ADR reports of TG. Third, to finally understand the long-term safety of tri-glycerides in patients suffering from AS, clinical trials and studies with long follow-up periods are recommended. As indicated by the results of this paper, TG can be used as an

12
Oxidative Medicine and Cellular Longevity

13
Oxidative Medicine and Cellular Longevity alternative treatment for patients suffering from AS, but further large-scale clinical studies are needed. In addition, the long-term safety, efficacy, and optimal dose of TG to treat AS should be explored. 4.6. Possible Mechanisms. Ji et al. [17] made a case-control investigation, reporting the ability of TG in improving the symptoms and signs of patients suffering from AS effectively on the basis of serum biomarker examination. The mechanism may be related to anti-inflammatory effects, inhibition of novel bone from being formed, and possible bone protective impact. Another study reported by them [45] showed that TG was efficient for the treatment of AS patients, and its mechanism of action may be correlated with the upregu-lation of cluster of differentiation (CD)4 + CD25 + CD127 low regulatory T cells and the downregulation of interleukin-(IL-) 17 levels in the peripheral blood. Furthermore, Zhang et al. [46] employed network pharmacology for the analysis of the active ingredients, the prediction of core TG targets, and pathways to treat AS, then preliminarily demonstrated the mentioned targets with the use of molecular docking. The enrichment investigation indicated that TG participates in various biological processes, covering acute inflammatory response, cell-matrix adhesion regulation, cell-cell adhesion, on the basis of TNF-α, nuclear factor-kappa B (NF-κB), and other signal pathways. However, the specific mechanism should be further clarified further in vivo and in vitro tests are required.  14 Oxidative Medicine and Cellular Longevity GRADE working group grades of evidence: high quality: further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: we are very uncertain about the estimate. 1 Some of the included studies did not report the implementation of blinding. 2 Heterogeneity (I 2 > 50%, P < 0:05) was found. 3 Some of the included studies did not report the details of random protocol. 4 Some of the included studies lack of allocation concealment.

Conclusion
In brief, we showed supporting evidence that at least to an extent TG as adjuvant therapy or monotherapy for AS reduces the BASDAI, BASFI, SP-VAS, serum CRP, and ESR without increasing the ADR. Furthermore, TG treatment visibly improved the overall effective rate in AS. Therefore, the findings of this paper proved TG as a possible candidate in terms of AS treatment. Nevertheless, given the heterogeneity along with small sample size, more randomized controlled trials with large multicenter and high quality should be performed to deepen the benefits of TG to treat AS.