Potential effectiveness of Chinese herbal medicine Yu ping feng san for adult allergic rhinitis: a systematic review and meta-analysis of randomized controlled trials

Background Chinese herbal medicine formula Yu ping feng san (YPFS) is commonly used for allergic rhinitis (AR). Previous review had summarized the effectiveness and safety of YPFS, however without any subgroup analysis performed to provide detailed evidence for guiding clinical practice. YPFS was recommended for the management of AR by Chinese medicine clinical practice guideline, but the treatment duration of YPFS was also not specified. The aim of this study is to evaluate the effectiveness and safety of YPFS in treating adult AR with the most recent evidence, and attempt to specify the duration of utilisation through subgroup meta-analyses. Methods Seven databases were searched from their inceptions to September 2017. Randomized controlled trials (RCTs) evaluating YPFS for adult AR were included. Methodological quality of studies was assessed using the Cochrane risk of bias tool. Meta-analysis and subgroup meta-analyses were conducted for evaluating the effectiveness of YPFS. The Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach was used for rating the quality of evidence. Results Twenty-two RCTs involving 23 comparisons were included in this review. YPFS was compared to placebo, pharmacotherapy, and used as an add-on treatment compared to pharmacotherapy. Meta-analyses were feasible for the outcomes of four individual nasal symptom scores and “effective rate”. Four individual nasal symptom scores decreased after YPFS’ combination treatment: itchy nose (MD-0.46, 95% CI[−0.50, −0.42]), sneezing (MD-0.41, 95% CI[−0.47, −0.35]), blocked nose (MD-0.46, 95% CI[−0.54, −0.39]) and runny nose (MD-0.42, 95% CI[−0.58, −0.26]). Based on “effective rate”, meta-analysis showed that YPFS did not achieve better effect than pharmacotherapy (RR1.07, 95%CI [0.94, 1.22), but its combination with pharmacotherapy seemed more effective than pharmacotherapy alone (RR1.27, 95%CI [1.19, 1.34]) (low quality). Subgroup analysis suggested that YPFS was not superior to the second-generation antihistamine (RR1.04, 95%CI [0.90, 1.19]) (low quality). Further, YPFS’ combination treatment seemed more beneficial when it was used for more than three weeks (RR1.15, 95%CI [1.01, 1.32]). In addition, YPFS was well-tolerated for treating adult AR. Conclusion Chinese herbal medicine formula YPFS seems beneficial for adult AR. This potential benefit need to be further evaluated by more rigorous RCTs. Electronic supplementary material The online version of this article (10.1186/s12906-017-1988-5) contains supplementary material, which is available to authorized users.


(Continued from previous page)
Conclusion: Chinese herbal medicine formula YPFS seems beneficial for adult AR. This potential benefit need to be further evaluated by more rigorous RCTs.
Keywords: Yu ping feng san, Chinese herbal medicine, Allergic rhinitis, Treatment duration, Systematic review, Meta-analysis Background Allergic rhinitis (AR) is the most common type of noninfectious rhinitis; it is defined as a "symptomatic disorder of the nose induced by the IgE-mediated inflammation after allergen exposure" [1]. AR is characterized mainly by nasal symptoms including sneezing, itching, rhinorrhoea and nasal congestion, as well as a group of non-nasal symptoms involving eyes, ears, throat or chest. AR patients may also develop sinusitis and asthma, or be accompanied by malaise, weakness, and fatigue [2,3]. Globally, AR affects 10% to 20% of the population, and an increase of the prevalence has been observed in the past 40 years [1,4]. Clinically, AR is one of the most common diseases affecting adults [5], and it is considered a major chronic respiratory disease due to its high prevalence, impairments on patients' quality of life (QoL) and work/school performance, substantial economic impact and its co-morbidities [1,6].
Allergic rhinitis has traditionally been categorized as seasonal or perennial types according to the predilection time and triggering allergens, or may be classified as intermittent or persistent types based on its frequency and severity. However, both of these classifications have certain limitations [1,[5][6][7]. Regardless of the classification types, their clinical management are all same. Recommended by the most recent clinical guideline, effective treatments of AR are topical steroids, oral antihistamines, and immunotherapy [5]. This guideline also pointed out that Chinese herbal medicine (CHM) has often been utilized for managing AR in clinical practice, although the evidence supporting its efficacy and mechanism of action is uncertain [5]. As suggested by a systematic review published in 2012, CHM might be more effective than placebo for persistent allergic rhinitis, but a confirmed conclusion could not be drawn since all included studies suffered certain methodological limitations [8].One CHM formula, Yu ping feng san (YPFS), was recommended by Chinese medicine clinical practice guideline to manage AR despite its classification and severity, but the effective treatment duration is not clear [9]. YPFS formula contains three key herbs: Astragalus membranaceus (huang qi), Rhizoma Atractylodis Macrocephalae (bai zhu) and Radix Ledebouriellae Divaricatae (fang feng) [10]. In clinical practice, YPFS usually is used together with other herbs, e.g., Lilymagnolia (xin yi), Fructus xanthii (cang er zi) and Radix Angelicae Dahuricae (bai zhi) etc. The effectiveness and safety of YPFS for AR treatment has been investigated by a number of clinical studies, and been summarized by a review article published in Chinese [11]. Although the overall effectiveness of YPFS seems promising, the treatment details were not suggested due to the lack of subgroup analyses. Therefore, in depth meta-analyses of current evidence are needed for guiding the use of YPFS for adult AR treatment in clinical practice.

Methods
The review protocol was registered on PROSPERO (CRD42015024821), and the review was constructed following the PRISMA guidelines (Additional file 1).

Study selection criteria
This systematic review was designed to evaluate the effectiveness and safety of CHM formula YPFS for AR. No limits on language and publication type were placed on study selection. Studies were considered if they recruited adult AR participants (aged 18 years and above) without classification limited. Included studies were RCTs comparing oral YPFS in any preparation forms to placebo CHM or pharmacotherapy being recommended by clinical practice guidelines [1,7], or comparing the combination of YPFS and pharmacotherapy to the same pharmacotherapy. If YPFS was used together with other CHM herbs, only the studies using YPFS as the chief formula with other herbs as additional modification were included.
Studies were included if they reported one of the predefined outcomes, these are: the primary outcome measures being total nasal symptom score (TNSS) or individual nasal symptom scores; the secondary outcomes are effective rate (a composite outcome measure which calculates the change of nasal symptoms scores and nasal signs), QoL, recurrence rate in follow-up phase, serum specific immunoglobulin E (sIgE) level, serum interleukin 4 (IL-4), and adverse events (AEs). . Abstracts and full texts were screened for eligibility by two authors (QL and CSZ) independently according to the selection criteria. Disagreement was solved through discussion with the third author (LY).

Data extraction
Data were extracted and entered into a pre-defined Excel spread sheet, including study design, participants' demographic data, details of CHM and control treatments, outcome data and AEs.

Methodological assessment
Two authors (QL and LY) assessed the methodological quality of the included studies independently using the risk of bias tools according to the Cochrane Handbook version 5.1.0 [12]. Disagreements were resolved through discussion with another author (XG). Each item was assessed as low, high or unclear risk, with reasons recorded to support the judgments.

Statistical analysis
The RevMan 5.3 software was used for performing meta-analyses. Studies were grouped for analyses according to their comparisons and outcome measures. Dichotomous data were expressed as risk ratio (RR) and continuous data were presented as mean difference (MD), both with 95% confidence intervals (CI). A fixedeffect model was used when heterogeneity (I 2 ) was less than 50%; otherwise a random effects model was applied. Subgroup analysis was performed based on the type of pharmacotherapy medicine used in the control groups (e.g., second-generation antihistamine, intranasal glucocorticosteroids) and treatment duration (two weeks or more than three weeks). Publication bias was assessed using a funnel plot and Egger's test where more than ten trials in a meta-analysis.

GRADE
The Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach [13] was used to summarize and rate the quality of evidence. Summary of Findings (SOF) tables were prepared using the online software program "GRADE pro GDT" (https://gradepro.org/). The main comparisons being assessed using GRADE methods were identified through discussion.

Participants
In total, 2309 participants with AR were involved in the 22 RCTs, 309 participants were diagnosed with perennial allergic rhinitis, and 2000 participants were AR without specified subtype. All participants aged between 18 to 70 years. Of the reported gender information, there were a total of 1153 males and 1098 females. Participants were allocated to the YPFS group (n = 552), placebo group (n = 126), pharmacotherapy group (n = 1044) and the combination of CHM and pharmacotherapy group (n = 587). Five studies did not provide information on participants' duration of disease [17,21,26,29,31], the remaining 17 RCTs reported the duration of disease with a median of 4.5 years.

Dropouts
Three studies reported dropouts during treatment phase with reasons provided [14,28,30]. None of the dropouts was due to serious AE. The other 19 studies were considered to have no dropouts since they reported equal numbers of participants at randomization and study completion.

Risk of bias assessment
Eight studies were assessed as "low" risk for "sequence generation" since they used random number "tables" [14,16,20,22,23,29,33,35], others were "unclear". For "allocation concealment", one study was "low" risk because it used a sealed envelope method [14], others were "unclear" due to lack of information. One study employed placebo control and made effort to blind participants, personnel and outcome assessors, thus this study was assessed as "low" risk for these items [14]; the other 21 studies were "high" risk for "blinding of participants and personnel" and "unclear" risk for "blinding of outcome assessors". All studies were "low" risk for the "incomplete outcome data", since they were considered having no dropouts    Huangqi 10 g, baizhu 10 g, fangfeng 10 g, guizhi 6 g, cangerzi 10 g, xinyi 10 g, baizhi 6 g, wumei 10 g, gancao 6 g.

Effectiveness
The effectiveness of YPFS were firstly analysed according to the comparators as follow. All meta-analyses results are summarised in Table 3.

YPFS versus placebo
Two studies compared YPFS with oral placebo [14,15]. Meta-analysis was not possible because they reported data on different outcomes. One study reported data on RQLQ, it was found that the post-treatment total RQLQ score of the YPFS group was lower than that of the placebo group (n = 159, MD-8.57, 95%CI [−16.37, −0.77]) [14]. Another study reported post-treatment data on effective rate, serum sIgE and IL-4. Significant difference between the YPFS group and the placebo group was found for all these outcomes (n = 86, effective rate: RR13. 33 [15]. However, these results are only from single study, the small sample size and wide confidence interval makes the results unconvinced.

Safety
Ten studies did not report information on AEs [15, 17-21, 27, 32-34]. Eight studies reported there was no AE occurred in both groups [16, 22, 24-26, 28, 30, 35]. The remaining four studies reported AEs in treatment group [14,23,29,31], and control groups [23,29,31] during the treatment phase. According to these studies, the AEs possibly associated with the use of YPFS were acne and abdominal distension [14]; while the AEs possibly caused by pharmacotherapy included dryness in the nasal cavity [23], local mucosa irritation [31], sedation [31], skin rash [29,31], vomit and nausea [29]. All above AEs were transient and no medical intervention was required. Serious AE was not reported by any study. It was worth noting that the number of AEs occurred in the combination group was less than that of the pharmacotherapy group. For example, one study reported five cases of AEs in the combination group and 12 cases in the pharmacotherapy group [29], and another study reported three cases of AEs in the combination group and four cases in the pharmacotherapy group [31]. These studies concluded that adding YPFS to pharmacotherapy is safe and it may be beneficial for reducing the AEs caused by pharmacotherapy. All detailed information of AEs is provided in Table 1.

Publication bias
The publication bias could not be assessed for the comparison of YPFS versus pharmacotherapy since there were not enough studies included in one metaanalysis. For the comparison of YPFS plus pharmacotherapy vs. pharmacotherapy alone, there were 12 trials in the meta-analysis of effective rate. No publication bias was detected based on the funnel plot (Fig. 6) and the Egger's test (t = 1.30, 95% CI [−1.26, 4.79], p = 0.223).

GRADE for the main comparisons of YPFS
GRADE assessment was conducted for two main comparisons. The quality of evidence was low. Table 4 summarises the results.

Discussion
This systematic review evaluated RCTs of CHM YPFS including the most recently published articles. Studies compared CHM YPFS to placebo or pharmacotherapy, and those used YPFS as add-on therapy were all included in our evaluation. The evidence of YPFS effectiveness has been revealed through meta-analyses.

Effectiveness and safety of YPFS
In this review, meta-analyses were feasible for the outcomes of four individual nasal symptom scores and "effective rate". Based on these outcomes, it was found that: 1) in terms of the nasal symptoms, adding YPFS to the second-generation antihistamines achieved better treatment effects (low quality evidence); 2) for "effective rate", although YPFS was not superior to  pharmacotherapy, adding YPFS to common pharmacotherapy was also more effective than using pharmacotherapy alone. Further, subgroup analysis showed that there was no significant difference between YPFS and the second-generation antihistamine in terms of "effective rate" (low quality evidence). In order to provide evidence for the duration of clinical administration, subgroup analysis regarding the treatment duration was conducted. Our results showed that adding YPFS to pharmacotherapy was beneficial when it was administered for more than three weeks, but this was not seen in the studies of only two weeks treatment. It is worth noting that, there has been experimental research evidence showing that six days' high and middle dosages of YPFS powder significantly increased the IFN-γ, the antibody of infectious laryngotracheitis and the immune organ indexes, and also reduced IL-4 in chickens [40]. One study reported that 14 days of oral YPFS could stimulate immune responses in allergic rhinitis mice [41]. Another study reported that after five to seven days intervention, different concentration of the glucosidic which extracted from YPFS produced anti-inflammatory and immune-regulatory effects [42]. These experimental studies indicated that YPFS alone may show immune-modulatory and anti-inflammatory effects after one to two weeks. In clinical management of AR, pharmacotherapies usually are fast-acting for symptom control but leaving 10% to 20% AR patients to endure bothersome symptoms [43,44]. Our results suggest that, although YPFS is not superior to pharmacotherapy, adding CHM YPFS to common pharmacotherapy for at least three weeks is helpful to improve the overall symptoms. Since YPFS has been recommended as AR management by current Chinese Medicine clinical practice guideline [9], but without recommendation on the treatment duration, our results may potentially have filled this gap and provide indication for further clinical practice. Nevertheless, these finding was only based on the outcome of "effective rate" since there were not enough data on nasal symptoms scores supporting these analyses. Further studies using globally well-accepted outcome measures to assess the treatment effects at different time points are needed to confirm such findings.
Reported AEs caused by YPFS were mild and transient. It was also reported that adding YPFS could reduce the number of AEs caused by cetirizine [29] or budesonide aerosol [31], though detailed explanation was not provided by these studies. Similar findings were reported by other clinical trials of YPFS for AR management [45,46] but without further exploration of the mechanisms. It is worth noting that some toxicity tests of YPFS also suggested that its toxicity is either minor or without any adverse effects [47,48].

Potential mechanisms of YPFS
Results from experimental studies showed that YPFS reduced the eosinophilic cells' infiltration and degranulation reaction, decreased the tissue edema [49,50], and reduced the immune factors including histamine, IgE, IL-4 and tumor necrosis factor-α (TNF-α) in serum [50]. Other researches indicated that YPFS was effective for AR through down-regulating the activity of IL-6 and TNF-α [51], activating interleukin 3 (IL-3), growth factor of mast cells, and inhibiting granulocyte colony stimulating factor (GM-SCF) and an inhibitory factor of mast cells proliferation [52]. IL-4 is one of the important Th2 cytokines related to AR, it can accelerate the synthesis of B cell and secrete a large number of IgE [7]. The result from five included studies in this review illustrated that the IL-4 level decreased after YPFS or its combination treatments [15,16,30,33,34], and one of these studies also reported the sIgE level declined after YPFS treatment [15]. Furthermore, a review article [53] suggested that YPFS may lead to immunoregulatory effects through the following pathways: 1), to promote the generation of IgA and IgG, the peripheral blood levels of CD3 + , CD4 + and CD4 + /CD8 + ; 2), to increase the amount of C3, C4 and the receptor of C3b, and activate the complement system involved in immune response; 3), to protect the phagocytosis of macrophage; and 4) to regulate the cyclic adenosine monophosphate (cAMP) in the cells of mice spleen bi-directionally.
There has been evidence showing that glucosides were the principal active ingredients of YPFS which are related to the anti-inflammatory and immunoregulatory effects [42], and claycosin and formononetin may be the effective compounds which could produce the antiinflammatory function by reducing thymic stromal lymphopoietin production via regulating Nuclear factor (NF)-κB activation [54].
In brief, there has been certain pre-clinical evidence supporting the use of YPFS for AR management due to its anti-inflammatory and immunoregulatory function. More details about its mechanism and the active compounds of YPFS are yet to be explored.

Limitations of the included studies
Proper randomization and allocation is essential to reduce selection bias for randomized controlled trials. However, in our systematic reviews, only 36.4% of the included studies (eight trials) applied appropriate methods for sequence generation, and only 4.5% (one trial) for allocation concealment. Blinding of participants, personnel and outcome measures was only achieved by one study, while other studies did not utilize proper placebo control to unsure blinding, and therefore they were not free of performance bias or detection bias.
In regards to the outcome measures, it was found the individual nasal symptom scores was reported by four studies but without TNSS being reported, and QoL was only reported by one study, while the "effective rate" was reported by the majority of the included studies. The "effective rate" assesses the proportion of participants who had achieved certain improvement of the nasal symptoms and signs. Considering that this outcome measure only reflects the overall improvement, it makes it indirect to compare the effectiveness results of these studies to other RCTs which reported data on nasal symptom scores and QoL, even though this outcome measure was recommended by several previous clinical guidelines [36][37][38][39]59] and had been widely used in China.
Considering AR is a chronic and recurrent condition, the long-term effects or the recurrence rate after treatment is of clinicians' interests. Two included studies reported data on the recurrence rate in follow-up phase [18,20]. However, there is no clear definition of "recurrence" stated by these studies. Further studies should consider using the internationally well-accepted outcome measures to evaluate the long-term treatment effects of YPFS, for example, the nasal symptom scores and QoL.
In terms of clinical trial reporting, none of the included study reported all items recommended by CON-SORT 2010 and its Extension for Herbal Intervention [60,61]. Also, in the absent of trial registration information or published protocol of the included studies, evaluation of selective reporting bias was not applicable in all of the included studies.
All these limitations reduced the certainty of our results. Future research should take all these aspects into consideration and improve the methodological quality of clinical trials' design and conduct.

Strengths and limitations of this review
The review synthesised the most recent clinical research following rigorous methodology recommended by the Cochrane Handbook, with GRADE approach being used to assess the quality of evidence. This review focused on YPFS treating adult AR, it not only included studies comparing YPFS to placebo or pharmacotherapy, but also evaluated the add-on effects of YPFS to pharmacotherapy. Outcome measures included symptoms scores, QoL, effective rate, recurrence rate, as well as objective outcomes such as serum sIgE and IL-4.
However, a few limitations downgraded the certainty of our results, they are: 1). The majority of included studies (21 out of 22) were all conducted in China and published in Chinese. 2).outcome measures in most included studies were not consistent with international guidelines. This leads to indirectness when interpreting the findings; 3). the overall methodological quality of included studies was low; 4). the long-term effects of YPFS were uncertain due to the lack of follow-up phase of most studies.

Implication for clinical practice and further research
The results of this review suggest that YPFS is effective for the management of adult AR as on add-on therapy. Clinical practice could consider adding oral YPFS to routine pharmacotherapy management and it may improve the effectiveness and safety of routine pharmacotherapy. A correlation of treatment duration and effects of YPFS was observed in this review, however, it requires further exploration. The mechanism of YPFS herbs and their active compounds are also worth more investigation.
Nevertheless, although YPFS in general is effective regarding individual nasal symptoms and the overall "effective rate", but there is insufficient data supporting the effectiveness towards TNSS and QoL. Further research should consider employing outcome measures recommended by international guidelines (nasal symptom scores and QoL) to evaluate the effectiveness of YPFS. A followup phase with well-accepted outcome measures is also important to confirm the long-term effects of YPFS. Lack of blinding is the major methodological flaw of most included studies in this review, further research should consider adapting placebo control and making efforts to the blinding of participants and outcome assessors.
In addition, although it was seen that adding YPFS to pharmacotherapy may reduce the common AEs caused by pharmacotherapy, the mechanism of such herb-drug