Integrative Medicine (Herbal Medicine Combined with Drug Therapy) for Behcet’s Disease: A Systematic Review and Meta-Analysis of Randomized Controlled Trials

This review aimed to investigate the efficacy of integrative medicine (herbal medicine combined with drug therapy) in the treatment of Behcet’s disease (BD). Eleven databases were searched from their inception to 7 December 2020, for randomized control trials (RCTs) that reported the effects of integrative medicine in treating BD. The risk of bias was assessed using seven domain criteria from the Cochrane Collaboration tool. Grading of Recommendations Assessment, Development, and Evaluation (GRADE) was used to assess the quality of evidence. The direction of the effect is also shown in the form of an albatross plot. Sixteen trials met the inclusion criteria and were analyzed. The overall risk of bias was determined to be uncertain. The meta-analysis showed a superior response rate with herbal medicine plus drug therapy (relative risk (RR) 1.19, 95% confidence interval (CI) 1.13 to 1.25, n = 1034, p < 0.00001, I2 = 0%, low certainty of evidence (CoE)) compared to drug therapy. Integrative medicine also lowered the recurrence rate after 2 months of follow-up (RR 0.27, 95% CI 0.09 to 0.76, n = 120, p = 0.01, I2 = 0%, low CoE). The erythrocyte sedimentation rate (ESR), C-reactive protein (CRP) level, and skin lesions were also significantly improved using integrative medicine, but equivalent effects were seen for oral ulcers, genital ulcers, and eye inflammation. Minor adverse events were reported in both groups. Our findings suggest that herbal medicine combined with drug therapy is more effective for the treatment of BD than drug therapy alone. Although the type of drug therapy used varied across the studies, integrative medicine was shown to improve the total response rate, skin lesions, the ESR, and the CRP level. However, the overall risk of bias of the studies was concerning, and the CoE was low. Information on adverse events (AEs) was also insufficient. In addition, the number of studies included for data synthesis for most outcomes was small. Future studies with rigorous RCTs may help establish the efficacy of integrative medicine in the treatment of BD.


Introduction
Behcet's disease (BD) is a multisystemic inflammatory disorder characterized by a range of manifestations, such as recurrent oral ulcers, genital ulcers, arthritis, vasculitis, and skin lesions. The age of onset of BD is usually 30-40 years [1,2]. The prevalence of BD is higher in the Middle East and Asia than in North America and northern Europe. In particular, its prevalence is highest in Turkey (80-370 cases per 100,000 persons) [3][4][5][6]. BD is commonly treated with steroids and immunomodulatory drugs such as corticosteroids, colchicine, and thalidomide, which often present side effects such as peripheral neuropathy, loss of appetite, nausea, diarrhea, and intestinal bleeding or perforation [7][8][9].
Integrative medicine refers to the combination of conventional drug therapies with complementary or alternative medicine (CAM) therapies such as acupuncture and herbal medicine [10]. Integrative medicine combines the benefits of both Western medicine and CAM [11]. It has, therefore, become increasingly prevalent and popular, not only in China but also worldwide [12]. Clinical trials have revealed that a combination of herbal medicine and drug therapy is more effective than herbal medicine alone and significantly reduces the symptoms and recurrence rate of BD. This improvement is accompanied by the expression of the cytokine LI-4 [13,14]. A recent publication also reported that herbal medicine was favorable for treating BD, showing the potential of herbal medicine in complementing conventional medication [15].
Two systematic reviews (SRs) have been published on the use of integrative medicine for the treatment of BD [16,17]. These SRs included seven randomized controlled trials (RCTs) that compared the effects of integrative medicine with drug therapy; however, the included studies did not have publication dates beyond 2014. Thus, this study aimed to update the evidence on the efficacy of integrative medicine in the treatment of BD.

Study Registration and Protocol Information
This review has been registered on PROSPERO CRD4201808496.

Data Sources
Eleven electronic databases, namely, PubMed, Embase, Cochrane Controlled Register of Trials (CENTRAL), China National Knowledge Infrastructure (CNKI), Wanfang Data, VIP Information, OASIS, DBpia, Research Information Service System (RISS), Korean Studies Information Service System (KISS), and KoreaMed, were searched from inception to 7 December 2020. Our search included studies in English, Chinese, and Korean languages. The search strategy included Medical Subject Headings (MeSH) and keywords that reflected the terms related to Behcet's syndrome, BD, Behçet's, integrative medicine, and integrated medicine. Detailed search terms are shown in Supplementary Materials.

Study Selection 2.3.1. Types of Studies
Eligible studies included RCTs or quasi-RCTs that compared integrative medicine with drug therapy. Postgraduate theses or dissertations were also eligible. Publications in the form of abstracts, conference proceedings, review articles, or other types of clinical studies, such as nonrandomized controlled studies, case-control studies, case reports and series, and animal studies, were not eligible.

Types of Participants
Eligible participants were patients of both sexes and all ages clinically diagnosed with BD. The studies had to meet the following diagnostic criteria for inclusion: We searched and screened eleven databases by checking the title and abstract. Furthermore, we independently extracted detailed data from the studies pertaining to the seven following domains: (1) first author, year of publication; (2) diagnosis, sample size, duration of treatment; (3) intervention group; (4) control group; (5) main outcomes; (6) results; and (7) AEs. In addition, we obtained the details on the prescription of herbal treatment.

Risk of Bias
As an assessment of the quality of each study, two authors (J.H.J and T.Y.C) used Cochrane Collaboration's tool to evaluate the risk of bias [22]. We evaluated seven domains, namely, random sequence generation, allocation concealment, blinding of participants and personnel, blinding of the outcome assessment, incomplete outcome of the data, selective reporting, and other biases, to assess the risk of bias. We categorized the risk of bias as low (L), high (H), or uncertain (U). Disagreements were resolved by another author (MSL).

Data Analysis
Data analyses were performed using Review Manager (Version 5.3.5) software provided by the Cochrane Collaboration. We quantified the effects of treatment as the RR with the 95% confidence interval (CI) for dichotomous data and as the mean difference (MD) with the 95% CI for continuous data. The chi-square test and Higgins I 2 test were used to assess heterogeneity. Additionally, we used Grading of Recommendations Assessment, Development, and Evaluation (GRADE) pro/GDT to assess the certainty of evidence (CoE) (21). Albatross plots were also generated using STATA/SE v.16.1 (StataCorp LLC, College Station, TX, USA) to visualize the effect of direction on the primary and secondary outcomes.  All studies were two-arm parallel-designed trials. The mean age of patients from the included studies ranged from 23.4 to 41.3 years. The average disease duration ranged from 2.5 to 12.6 years. The total sample size was 987, with the number of participants in each group ranging from 20 to 57. All included studies used integrative medicine (combination of herbal medicine and drug therapy) as the intervention and drug therapy alone as the comparator. The details of the included studies are shown in Table 1.
Herbal medicine interventions in the included studies were herbal prescriptions given in the form of oral decoctions. Thirteen studies used standard prescriptions or modified prescriptions, whereas the remaining three studies used pattern identification (PI)-based prescriptions. The standard and modified prescriptions included modified Gancao Xiexin decoction, modified Qingdai san, Shen's Shengdi Qinlian Tufuling decoction, modified Huanglian decoction, modified Yiqi Jiedu Quyu prescription, standard/modified Huiyan Zhuyu decoction, Baitouweng decoction, Yinshenhao decoction, Duanxia Shenshi decoction, Longdan Xiegan decoction, modified Chixiaodou Danggui San, modified Ziyin Yuyang decoction, and Leiling Jiedu decoction. The herb with the highest frequency of use across the prescriptions was Glycyrrhizae Radix et Rhizoma, followed by Scutellariae Radix and Coptidis Rhizoma. The herbal compositions of the included prescriptions are presented in Table 2. The conventional drugs included thalidomide, celecoxib, azathioprine, prednisone, levamisole, and cyclophosphamide (all administered orally), and dexamethasone (administered intravenously).

Risk of Bias
The overall risk of bias was judged to be uncertain using Cochrane Collaboration's tool for risk of bias assessment ( Figure 2).
Only four of the included studies reported the method used for random sequence generation, and none of the studies used allocation concealment methods [23][24][25]28]. As none of the studies reported the blinding of participants/personnel or outcome measurements, performance biases and detection biases for all studies were judged as unclear.
Most studies were judged as having a low risk of bias because follow-up information or outcome data were missing; however, one study was judged as having a high risk of bias due to the incomplete report of several outcomes [37]. One study (27) that reported a few dropouts was eventually judged as having a low risk of bias, as the number of dropouts was less than 10% of the randomized samples, and the remaining studies did not report the percentage of patients who were lost to follow-up or did not perform an intention-to-treat (ITT) analysis [28]. The judgment of unclear risk of bias was given to all studies for reporting bias, as none of the studies had study protocols and did not provide sufficient information for further assessment. Information such as the source of funding, sample size calculation, and trial registration was also insufficient to assess other potential biases in the included studies.

Certainty of Evidence
The CoE for each outcome as assessed using GRADE was low or very low. The 'Summary of findings' table on the main outcomes is presented in Table 3. * The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; MD: mean difference; RCT: randomized controlled trial; RR: risk ratio; a . Overall risk of bias is uncertain. Only four studies reported the simple randomization method, and the remaining studies did not provide relevant information. All studies reported a lack of allocation concealment and blinding. Other risk of bias domains were also concerning due to poor reporting. Therefore, the studies included were judged to have serious methodological limitations. b . The sample size of each study is considered small, resulting in borderline imprecision. c . Heterogeneity across the studies is fairly high. GRADE Working Group grades of evidence. High certainty: We are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: We are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: Our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: We have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.
A meta-analysis of both studies on the recurrence rate after 2 months of follow-up showed that the group administered herbal medicine combined with drug therapy experienced a lower recurrence rate than the group administered drug therapy alone (RR 0.27, 95% CI 0.09 to 0.76, n = 120, p = 0.01, I2 = 0%, low CoE, Figure 3B).

Oral Ulcers
Only two studies reported the symptom score for oral ulcers [23,26]. The meta-analysis showed an equivalent effect between the herbal medicine combined with drug therapy group and the drug therapy group (MD −0.28, 95% CI −1.03 to 0.47, two studies, n = 120, p = 0.47, I2 = 72%, very low CoE, Figure 4A).

Genital Ulcers
Only two studies assessed the symptom score for genital ulcers [23,26]. The metaanalysis showed an equivalent effect between the herbal medicine combined with drug therapy group and the drug therapy group (MD −0.35, 95% CI −1.17 to 0.47, n = 120, p = 0.40, I2 = 77%, very low CoE, Figure 4B).

Eye Inflammation
Only two studies evaluated the symptom score for eye inflammation [23,26]. The meta-analysis showed an equivalent effect between the herbal medicine combined with drug therapy group and the drug therapy group (MD −0.32, 95% CI −0.84 to 0.21, n = 120, p = 0.24, I2 = 59%, very low CoE, Figure 4C).

Recurrence Rate
Only two studies that compared the effectiveness of herbal medicine combined with drug therapy with drug therapy assessed the recurrence rate [23,24]. One study [23] reported the recurrence rates between the two groups after 1 month, 2 months, and 3 months of follow-up. The recurrence rate was equivalent between the two groups after 1 month (RR 0.33, 95% CI 0.04 to 3.03, p = 0.33) and 2 months (RR 0.25, 95% CI 0.06 to 1.08, p = 0.06) of follow-up. After 3 months of follow-up, the recurrence rate was lower for the herbal medicine combined with drug therapy group than for the drug therapy group (RR 0.27, 95% CI 0.08 to 0.88, p = 0.03). Another study [24] reported the recurrence rate only at the 2-month follow-up (RR 0.25, 95% CI 0.06 to 1.08, p = 0.10).
A meta-analysis of both studies on the recurrence rate after 2 months of followup showed that the group administered herbal medicine combined with drug therapy experienced a lower recurrence rate than the group administered drug therapy alone (RR 0.27, 95% CI 0.09 to 0.76, n = 120, p = 0.01, I 2 = 0%, low CoE, Figure 3B).

Oral Ulcers
Only two studies reported the symptom score for oral ulcers [23,26]. The meta-analysis showed an equivalent effect between the herbal medicine combined with drug therapy group and the drug therapy group (MD −0.28, 95% CI −1.03 to 0.47, two studies, n = 120, p = 0.47, I 2 = 72%, very low CoE, Figure 4A).

Genital Ulcers
Only two studies assessed the symptom score for genital ulcers [23,26]. The metaanalysis showed an equivalent effect between the herbal medicine combined with drug therapy group and the drug therapy group (MD −0.35, 95% CI −1.17 to 0.47, n = 120, p = 0.40, I 2 = 77%, very low CoE, Figure 4B).

Eye Inflammation
Only two studies evaluated the symptom score for eye inflammation [23,26]. The meta-analysis showed an equivalent effect between the herbal medicine combined with drug therapy group and the drug therapy group (MD −0.32, 95% CI −0.84 to 0.21, n = 120, p = 0.24, I 2 = 59%, very low CoE, Figure 4C).

Erythrocyte Sedimentation Rate and C-Reactive Protein Levels
Four studies showed a favorable effect of herbal medicine combined with drug therapy compared to drug therapy alone [25][26][27]29], while two studies showed an equivalent effect between the two groups [23,24]. Of the six studies [23][24][25][26][27]29], one was excluded from the meta-analysis due to incomplete reporting [37]. The meta-analysis showed that the effect of herbal medicine combined with drug therapy on the ESR was superior to that of drug therapy alone (MD −4.19, 95% CI −7.59 to −0.79, n = 338, p = 0.02, I 2 = 80%, very low CoE, Figure 4E).

Adverse events
Nine studies assessed AEs [23][24][25][26][27]29,[36][37][38]. Of the nine studies, two [26,36] reported no AEs for either integrative therapy or drug therapy. In the remaining seven studies, 30 AEs were reported in the herbal medicine combined with drug therapy group, and 67 AEs were reported in the drug therapy alone group. The details of the AEs are listed in Table 1.

Albatross Plot
The albatross plot showing the effects of direction and size range by p-value and the given sample sizes was generated for each included study ( Figure 5, different outcome groups are presented in different colors). For the dichotomous data, the points were scattered across the contour lines ( Figure 5A). All the points were clustered to the positive association side of the plot, implying that herbal medicine combined with drug therapy is favorable for the treatment of BD. For the continuous data, the points were scattered more towards the right side of the plot, with many points clustered around the null line, showing the equivalent effect of herbal medicine combined with drug therapy and drug therapy alone ( Figure 5B). groups are presented in different colors). For the dichotomous data, the points were scattered across the contour lines ( Figure 5A). All the points were clustered to the positive association side of the plot, implying that herbal medicine combined with drug therapy is favorable for the treatment of BD. For the continuous data, the points were scattered more towards the right side of the plot, with many points clustered around the null line, showing the equivalent effect of herbal medicine combined with drug therapy and drug therapy alone ( Figure 5B).

Summary of the Main Results
This systematic review revealed that herbal medicine combined with drug therapy appears to be more effective for the treatment of BD than drug therapy alone. Although the type of drug therapy used varied across the studies, herbal medicine combined with drug therapy was shown to improve the total response rate, skin lesions, ESR, and CRP level. However, the overall risk of bias of the studies was concerning, and the CoE was low. Information on AEs was also insufficient. In addition, the number of studies included for data synthesis for most outcomes was small. Future studies with rigorous RCTs may help establish the efficacy of herbal medicine combined with drug therapy in the treatment of BD.

Quality of the Evidence
The level of evidence as assessed by GRADE for the studies included in this metaanalysis was low. All the included studies had a high risk of bias. The evidence was downgraded by one level in the category of risk of bias due to a lack of information on the randomization, allocation concealment, and blinding methods in all studies. Further high heterogeneity of the symptom score outcome (oral ulcers, genital ulcers, and eye inflammation) downgraded the quality of evidence by one level in the category of inconsistency. In the category of indirectness, the included studies corresponded to the patient, intervention, comparison, outcomes (PICO) study, and it was ascertained that this effect was not serious. In the category of imprecision, the included studies had a small sample size, with 10 to 57 participants; thus, the quality of evidence was downgraded by one level. In general, the quality of reporting was poor in the included trials. Furthermore, the number of trials and total sample size included in our analysis were not sufficient to draw firm conclusions.

Potential Biases in the Review Process
There were several limitations to the present review. First, all the studies were published in China, limiting the overall generalization of the results. Despite searching eleven databases, all the studies that met the inclusion criteria were performed in China, where no studies with negative results have been reported [39]. Second, the methodological quality of the included studies was poor. There was a lack of information on randomization procedures and blinding methods that led to a concerning risk of bias. Third, the numbers of studies included in the meta-analyses of several outcomes, such as major BD symptoms and the recurrence rate, were small, resulting in a lack of significance and inconclusive results. In addition, the heterogeneity of herbal medicines used in combination with drug therapy highly limited the subgroup analysis and the comparison of therapeutic effects for herbal medicine combined with drug therapy to drug therapy alone. Such heterogeneity also makes it difficult to evaluate or identify herbal medicinal compounds effective for improving BD symptoms. Different types of integrative medicines and drug therapies used for different durations also made it difficult to compare the regimens. Fourth, herbal medicines used in the form of decoctions are not standardized and vary in their compositions and dosage, restricting the assessment of individual decoctions or single-herb effects. The diversity of herbal medicine administration within and across the included studies may have affected the overall results of our study, causing a certain degree of bias. Finally, we could not establish the safety of herbal medicine combined with drug therapy in treating BD, as almost half of the included studies did not report AEs.

Agreements and Disagreements with Other Studies or Reviews
This review showed that herbal medicine combined with drug therapy is potentially effective for symptom relief in BD patients. Two previous reviews studied the effect of herbal medicine combined with drug therapy in the treatment of BD [16,17]. In our review, we found nine additional RCTs that adhered to standard clinical diagnostic criteria. The results of the analysis showed that the symptom improvement rate was higher with herbal medicine combined with drug therapy than with drug therapy alone. This result is similar to that of previous reviews, showing that the inclusion of additional studies did not change the significance of the results. Despite the addition of updated evidence, the small sample size and poor quality of the included studies remain the major limitations of our review.

Implications for Clinical Practice
Clinical manifestations vary greatly among BD patients, and conventional therapy highly depends on the severity of the disease, which often involves several body systems and organs. The main principle of BD treatment via drug therapy remains to be subduing inflammation during attacks and improving patients' quality of life by easing symptoms, increasing immunity, and reducing recurrence [40]. Although evidence supporting the use of azathioprine and cyclosporin A for ocular manifestations and interferon (IFN)α for mucocutaneous manifestations is available, evidence for vascular, gastrointestinal, and neurological involvement is still lacking [8]. To date, the effectiveness of drug therapy is based mostly on case reports/series and a limited number of RCTs.
Herbal medicine has been shown to have anti-inflammatory properties with favorable impacts on immune function and could play a critical role in complementing drug therapy [41]. The use of herbal medicine with anti-inflammatory properties also showed promising results in managing autoimmune diseases with oral manifestations, such as oral lichen planus, aphthous stomatitis, and Sjögren's syndrome, revealing the increasingly important role of herbal medicine in managing various inflammatory diseases [42][43][44].
Although evidence for proving the effectiveness of herbal medicine combined with drug therapy in the treatment of BD remains weak, accumulating evidence still shows encouraging effects. Medical professionals may still cautiously recommend the use of herbal medicine combined with drug therapy to patients who present side effects and resistance to drug therapy over a long duration.

Implications for Research
There are several concerns regarding the use of herbal medicine combined with drug therapy for the treatment of BD. First, the herbal prescriptions and drug therapies used in the included studies varied across studies. The included studies might not have been sufficient to demonstrate the complete effects for treating BD, and a subgroup analysis was unable to be conducted due to the small number of studies included. Second, all studies had a short duration of treatment. BD is an autoimmune disease with a high rate of recurrence. It is necessary to prove the effectiveness of herbal medicine combined with drug therapy on the recurrence rate by extending the period of treatment. Third, a number of AEs were reported in the herbal medicine combined with drug therapy group, and this has raised concerns about possible herb-drug interactions, such as pharmacokinetic and pharmacodynamic (PK-PD) interactions. The integration of herbal medicines with conventional drugs may cause changes in the movement, absorption, biochemical, and physiological effects of the herbs. Herb-drug interactions due to the concurrent use of herbs with conventional drugs may also magnify the effects of drugs, plausibly leading to the effectiveness of herbal medicine combined with drug therapy in treating BD. Fourth, the herbs with the highest frequency of use across the studies were Glycyrrhizae Radix et Rhizoma, Scutellariae Radix, and Coptidis Rhizoma. [45][46][47] These herbs have shown great potential in the treatment of inflammation-related diseases due to their anti-inflammatory active compounds. Future studies on these herbs, as single herbs or decoctions, should be considered to validate their usage and effectiveness in treating BD. In general, welldesigned RCTs in compliance with the CONSORT guidelines [48], long-term treatment periods, large sample sizes, and data on PK-PD parameters are warranted to guide the implementation of integrative medicine in clinical practice in the near future.

Conclusions
This review showed a significant improvement in symptoms upon the administration of herbal medicine combined with drug therapy in patients with BD. However, the included studies were performed on small sample sizes, had short treatment periods, and lacked detailed descriptions of the methodologies. To establish the effectiveness of herbal medicine combined with drug therapy in the treatment of BD, future RCTs designed in compliance with the CONSORT guidelines are required to ensure a larger sample size, longer treatment duration, and rigorous evidence-based treatment methodology.