Introduction

Herbal medicine (HM) has been used as a treatment for allergic diseases for centuries [1] and can decrease nasal symptoms of allergic rhinitis (AR) [2]. Herbal usage has increased in the past three decades [3]. A self-report survey in Germany demonstrated that 26.5% of the participants used alternative medicine for allergy diseases [4]. To date, there are several potential HMs for treating AR worldwide. Yu ping feng san is a Chinese formula commonly used in Asian countries, especially in East Asia [5]. In Western countries, butterbur has become one of the most common herbs used as an adjunct treatment for AR since the first human trial in 2002 [6•]. The mechanisms of action of HM include anti-inflammatory, anti-allergic, and immunological effects [7]. Anti-inflammatory effects of HM interfere with type 1 hypersensitivity and inhibit the production of inflammatory cells such as mast cells, basophils, eosinophils, and monocytes. Anti-allergic effects of HM reduce the release of histamine, leukotriene, cytokine, and chemokine from inflammatory cells. The active compounds in HM modulate the immunological activities of mast cells [7, 8•]. As a result, these effects significantly relieve nasal symptoms of AR, including sneezing, itching, rhinorrhea, nasal obstruction [9].

The Allergic Rhinitis and its Impact on Asthma (ARIA) guidelines-2016 revision suggests conventional treatment options for treating AR including antihistamines, intranasal corticosteroids, and leukotriene receptor antagonists [10]. In addition to conventional therapies, HM is acknowledged as an alternative therapy [2, 11•]. Although the beneficial effects of HM have been extensively investigated, the findings are inconsistent [8•, 11•]. In some systematic reviews, the data consisted of different kinds of HM and various mechanisms of action. The extracted data were pooled without being categorized [9, 12•]. There is no systematic review that focuses on the mechanisms of action of HM [8•, 9, 12•]. The optimal duration of HM treatments to reach their maximum effects is unknown. It is unclear whether HM works only in a short period of duration or it has long-lasting benefit in controlling the symptoms of AR. This review aimed to assess the effects of HM therapy in patients with AR, in various durations and to analyze the effects of HM separately by mechanism of action.

Recent Findings

Study Selection, Data Extraction, and Analysis

The protocol of this systematic review was registered on PROSPERO (ID: CRD42020168367). Electronic searches were conducted on PubMed and EMBASE. Additional sources were manually searched for published and unpublished trials. The last search was performed on February 9, 2020. Combination of MeSH terms and keywords were “rhinitis, allergic, seasonal”, “rhinitis, allergic, perennial”, “rhinitis”, “*allergic rhinitis”, “hay fever”, “rhinoconjunctivitis”, “pollen allergy”, “herbal medicine”, “Chinese herb*”, “plant extract”, “phytomedicine*”, “herbaceous agent”, “eastern medicine”, “oriental medicine”, “alternative medicine”. Only clinical trials in humans and trials published in English were selected. The systematic review was performed under The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) format [13].

Randomized controlled trials (RCTs) studying AR patients of any age were included. Diagnostic criteria of AR followed the Allergic Rhinitis and its Impact on Asthma (ARIA) guidelines. Patients who had characteristic clinical symptoms of AR and allergies were confirmed by either skin prick test (SPT) or serum IgE test [14]. HM in any formulation (decoction, tablet, pill, powder, herbal patch, and nasal spray/drop). Duration of treatment was at least 1 week. There was no limit of the length of treatment. The comparison pairs of interest in this review were (1) HM versus placebo, (2) HM versus standard treatment, and (3) HM plus standard treatment versus standard treatment alone. Antihistamines and intranasal corticosteroids were acknowledged as standard treatments in this review. The outcomes were nasal symptoms, ocular symptoms, disease-specific quality of life (QOL), objective measurement for nasal patency, and adverse events. Exclusion criteria included trials related to homeopathy and immunotherapy, studies that were conducted with experimental extracts containing synthetic chemicals, conference abstracts, and crossover studies with the washout period less than one week (due to possible carry-over effects). Two authors (MPH and WC) independently screened the titles and the abstracts based on the pre-determined eligibility criteria and reviewed the selected articles comprehensively. The senior author (KS) resolved disagreements on study selection when necessary.

The extracted data were AR subtypes, number of patients who received HM and comparators, age, gender, formulation of HM, duration of treatment, and therapeutic outcomes. The mechanism of action of each HM was assessed. HMs were categorized into subgroups according to their actions. Data of HM with similar type of effects were pooled. One HM may have multiple effects and could be categorized into more than one subgroup. Anti-inflammatory effect was defined when the HM decreased the migration of inflammatory cells, including mast cells, eosinophils, basophils, and monocytes. Anti-allergic effect was defined when the HM reduced the release of cytokines, chemokines, or mast cell mediators, including histamine, leukotrienes, or prostaglandins [7, 15]. Anti-leukotriene effect was defined when the HM worked as leukotriene biosynthesis inhibitors or leukotriene receptors antagonists [16]. Anti-histaminic effect was defined when the HM diminished the skin wheal and flare responses in SPT [17]. In addition, anti-cholinergic and vasoconstrictor effects were extracted from the experimental trials where available.

Two authors independently reviewed the quality of the included studies following the Cochrane Handbook for Systematic Reviews of Interventions [18]. Five domains were assessed: random sequence generation, allocation concealment, blinding of outcome assessment, incomplete outcome data, and selective reporting. Each domain was determined as “low risk of bias” if the domain was well-described; “high risk of bias” if the method or data of the respective domain had not been mentioned; or “unclear risk of bias” if the domain data were only mentioned without a clear explanation.

The extracted data were pooled for meta-analysis. Risk ratio (RR) and 95% confidence interval (CI) were used for dichotomous data. Continuous outcomes were presented as mean difference (MD) or standardized mean difference (SMD) with standard deviation (SD) and 95%CI. Subgroup analysis by the AR subtype, quality of the included studies, and mechanism of the effects were conducted. If the change from baseline to endpoint was not available, the final scores were extracted. The standard error, interquartile range, range, and 95% CI were imputed if the SD was not reported. Discrepancies in treatment effects among different trials were assessed using a heterogeneity (I2) statistic. An I2 of <40%, 40-60% and >60% represented low, moderate and substantial heterogeneity, respectively. A fixed-effect model was used in low heterogeneity and a random-effects model was used if the heterogeneity was high for a more conservative estimate of the differences. All statistical assessments were conducted using Review Manager (The Nordic Cochrane Centre, The Cochrane Collaboration, Copenhagen, Denmark).

Results

A total of 2032 articles were retrieved for screening (2,030 from electronic search and two from manual search). Finally, 32 studies were included in qualitative synthesis [19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50] and 29 studies in quantitative synthesis [19, 20, 22,23,24,25, 27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43, 45,46,47,48,49,50] (Fig. 1). Data from three included studies [21, 26, 44] were not pooled for meta-analysis due to incomplete outcome data. The characteristics of the included studies are summarized in Table 1.

Fig. 1
figure 1

Flow diagram of the study selection

Full size image
Table 1 Characteristics of included studies investigating herbal medicine categorized by the mechanism of action
Full size table

There were 2697 participants with the mean age of 34.57 years. Fifty-two percent of the patients were female. Nineteen RCTs included patients with perennial allergic rhinitis (PAR) [20, 24, 28, 30, 31, 35,36,37,38, 40,41,42,43,44,45,46,47, 49, 50], and 12 RCTs included seasonal allergic rhinitis (SAR) [19, 21,22,23, 25,26,27, 29, 33, 34, 39, 48]. One RCT studied patients with both PAR and SAR [32]. Four RCTs recruited the patients under 18 years of age [28, 32, 41, 45].

  1. 1.

    Effects of Herbal Medicine

HMs by Mechanism of Action

Both anti-inflammatory effect and anti-allergic effect were in 18 RCTs [20, 22, 28, 30,31,32, 35, 36, 38,39,40, 42, 43, 45,46,47,48, 50], sole anti-inflammatory effect 4 RCTs [23, 26, 37, 41], sole anti-allergic effect 4 RCTs [19, 29, 33, 44, 49], anti-leukotriene effect 5 RCTs [21, 24, 25, 27, 34], and anti-histaminic effect 4 RCTs [32, 36, 46, 47]. There was insufficient data to assess anti-cholinergic and vasoconstriction effects.

HMs by Formulation

Oral HM were in 26 RCTs [19,20,21,22,23,24,25,26,27,28,29,30,31, 33,34,35,36,37,38,39,40, 43, 45, 47, 49, 50], intranasal spray or oil inhalation 3 RCTs [42, 44, 48], and external herbal patch or moxibustion 3 RCTs [32, 41, 46]. Data are shown in Table 1.

The duration of treatments ranged from 1 to 16 weeks.

The Comparison Pairs

HM versus placebo were in 27 RCTs, HM versus standard treatment 3 RCTs [21, 41, 50]. There were two RCTs investigating three arms of HM versus placebo and standard treatment [24, 27]. There was no study comparing HM plus standard treatment versus standard treatment.

Standard Treatments

Four RCTs used antihistamine [21, 24, 27, 41] and 1 RCT used a combination of intranasal corticosteroid spray and antihistamine [50].

  1. 2.

    Risk of Bias in the Included Studies

In general, the included studies had some selection bias: 25% had low risk of bias in allocation concealment and 59% had low risk in random sequence generation; 62% had low risk of bias in blinding of outcome assessment and incomplete outcome data while 78% had low risk of bias in selective reporting.

  1. 3.

    Total Nasal Symptom Score

Sixteen RCTs compared total nasal symptom score (TNSS) between HM and placebo [22, 24, 25, 27, 29, 33,34,35,36, 39, 40, 42, 45,46,47,48] and three RCTs compared TNSS between HM and standard treatment [24, 27, 50].

When the duration of treatment was ≤ 4 weeks, the effects favored HM over placebo (SMD −0.68; 95%CI −0.98, −0.38; p < 0.01; 11 RCTs) (Fig. 2). HM and standard treatment brought similar effects (MD −0.01; 95%CI −0.24, 0.21; p = 0.93; 3 RCTs). When the duration of treatment was 4–12 weeks, the effects still favored HM over placebo (SMD −0.22; 95%CI −0.4, −0.05; p = 0.01; 7 RCTs). Nevertheless, the effects were not statistically different from placebo when the duration of treatment was ≥12 weeks (SMD −0.49; 95%CI −1.13, 0.15; p = 0.13; 5 RCTs).

  1. 4.

    Individual Nasal Symptom Score

Fig. 2
figure 2

Improvement on total nasal symptom score: herbal medicine versus placebo at ≤ 4-week time point. CI confidence interval; df degrees of freedom; Std. mean difference standardized mean difference

Full size image

Sneezing, rhinorrhea, and nasal obstruction scores were assessed by fifteen RCTs [20, 25, 28, 30,31,32, 37, 38, 40, 42, 45,46,47,48,49] while itching score was assessed by eleven RCTs [20, 25, 31, 32, 37, 42, 45,46,47,48,49]. Data from two studies could not be pooled because the SDs could not be imputed [25, 31].

When the duration of treatment was ≤ 4 weeks, the effects favored HM over placebo in sneezing (SMD −0.23; 95%CI −0.44, −0.02; p = 0.03; 12 RCTs), rhinorrhea (SMD −0.32; 95%CI −0.58, −0.06; p = 0.02; 12 RCTs), nasal obstruction (SMD −0.36; 95%CI −0.57, −0.16; p < 0.01; 12 RCTs), and itching (SMD −0.36; 95%CI −0.62, −0.09; p < 0.01; 9 RCTs). When the duration of treatment was > 4 weeks, the effects favored HM over placebo only in nasal obstruction (SMD −0.34; 95%CI −0.66, −0.02; p = 0.04; 3 RCTs), the effects in other individual symptoms were not different from placebo. Two RCTs compared individual nasal symptom score between HM and standard treatment [41, 50]. HM brought similar effects with standard treatment in each symptom.

  1. 5.

    Total Ocular Symptom Score

Eight RCTs compared the total ocular symptom score (TOSS) between HM and placebo [22, 34, 39, 40, 42, 45, 46, 48]. When the duration of treatment was ≤ 4 weeks, the effects favored HM over placebo (SMD −0.32; 95% CI −0.58, −0.05; p = 0.02; 4 RCTs) (Fig. 3). When the duration of treatment was > 4 weeks, there was no statistically significant difference between the groups.

  1. 6.

    Disease-Specific QOL

Fig. 3
figure 3

Improvement on total ocular symptom score: herbal medicine versus placebo at ≤ 4-week time point. CI confidence interval; df degrees of freedom; Std. mean difference standardized mean difference

Full size image

Seventeen RCTs compared Rhinoconjunctivitis Quality of life Questionnaire (RQLQ) [19, 20, 22, 23, 30, 33, 36, 38,39,40,41,42,43, 46,47,48, 50]. Data in five RCTs were not pooled because the SDs could not be imputed [22, 38,39,40,41].

When the duration of treatment was ≤ 4 weeks, the effects favored HM over placebo (SMD −0.53; 95%CI −0.81, −0.25; p < 0.01; 9 RCTs) (Fig. 4) and over standard treatment (SMD −1.89; 95%CI −2.37, −1.41; p < 0.01; 1 RCT) [51]. When the duration of treatment was 4–12 weeks, the effects still favored HM over placebo (SMD −0.48; 95%CI −0.89, −0.06; p = 0.03; 7 RCTs). Nevertheless, the effects were not statistically different between the groups when the duration of treatment was ≥12 weeks (SMD −0.17; 95%CI −0.47, 0.12; p = 0.24; 3 RCTs).

  1. 7.

    Objective Measurement for Nasal Patency

Fig. 4
figure 4

Improvement on Rhinoconjunctivitis Quality of Life Questionnaire (RQLQ): herbal medicine versus placebo at ≤ 4-week time point. CI confidence interval; df degrees of freedom; Std. mean difference standardized mean difference

Full size image

Nasal airway resistance (NAR) and peak nasal inspiratory flow (PNIF) were assessed. Four RCTs assessed anterior NAR in the inhalation phase [35, 44, 45, 50]. One RCT did not report the mean NAR [44]. There were no significant differences between the effects of HM and placebo (MD −0.07; 95%CI −0.19, 0.04; p = 0.22; 2 RCTs) [35, 45] and between the HM and standard treatment (MD −0.01; 95%CI −0.06, 0.04; p = 0.68; 1 RCT) [50]. Two trials assessed peak nasal inspiratory flow, however, neither of two trials had sufficient data for analysis [24, 49].

  1. 8.

    Subgroup Analysis by Mechanism of Action

The effects of HM on TNSS improvement were better than placebo in all subgroups of mechanism of action: anti-allergic effect (SMD −0.55; 95%CI −0.69, −0.4; p < 0.01, 12 RCTs), anti-inflammatory effect (SMD −0.61; 95%CI −0.88, −0.33; p < 0.01, 13 RCTs), anti-leukotriene effect (SMD −0.67; 95%CI −1.07, −0.27; p < 0.01, 4 RCTs), and anti-histaminic effect (SMD −0.5; 95%CI −0.91, −0.08; p < 0.01, 4 RCTs). The effects of HM on RQLQ improvement were better than placebo in anti-allergic effect (SMD −0.61; 95% CI −1, −0.21; p < 0.01, 9 RCTs) and anti-inflammatory effect (SMD −0.5; 95%CI −0.79, −0.21; p < 0.01, 9 RCTs). There was no difference on RQLQ improvement between HM with anti-histaminic effect and placebo (SMD −0.16; 95% CI −0.40, 0.09; p = 0.2, 3 RCTs).

  1. 9.

    Subgroup Analysis by AR Subtype

For the duration of treatment ≤ 4 weeks, the effects of HM on TNSS improvement were better than placebo in both the SAR (SMD −0.92; 95%CI −1.41, −0.43; p < 0.01; 5 RCTs) and PAR subgroups (SMD −0.47; 95%CI −0.77, −0.17; p < 0.01; 6 RCTs). For the duration of 4–12 weeks, the effects on TNSS favored HM only in the SAR subgroup (SMD −0.51; 95%CI −0.87, −0.16; p < 0.01; 3 RCTs), but not in the PAR subgroup (SMD −0.13; 95%CI −0.33, 0.06; p = 0.18; 4 RCTs). For the duration of ≥ 12 weeks, there were no differences between HM and placebo in both the SAR and PAR subgroups. The effects of HM on RQLQ improvement were better than placebo when the duration of treatment was ≤ 4 weeks in both the SAR (SMD −0.82; 95%CI −1.56, −0.08; p < 0.01; 3 RCTs) and PAR (SMD −0.38; 95%CI −0.61, −0.14; p < 0.01; 6 RCTs), but there were no differences after 4 weeks of treatments.

  1. 10.

    Subgroup Analysis by Quality of the Included Studies

The included studies that had at least one high risk of bias in one domain were defined as “Trials with high risk of bias” where others were defined as “Trials without high risk of bias.” In the trials without high risk of bias subgroup, HM significantly improved TNSS when the duration of treatment was ≤ 4 weeks (SMD −0.89; 95%CI −1.13, −0.65; p < 0.01; 8 RCTs) but there was no difference after 4 weeks. HM significantly improved RQLQ when the duration of treatment was < 12 weeks but there was no difference after this timepoint. In the trials with high risk of bias subgroup, there were no differences between the HM and placebo in both TNSS and RQLQ improvement regardless of the duration of treatment.

  1. 11.

    Adverse Events

Nine RCTs assessed headache, dry mouth/nose, dizziness, somnolence, and gastrointestinal pain/diarrhea events (Table 2). There were no significant differences in adverse events between the HM and other treatments.

Table 2 Adverse events of herbal medicine and risk ratio
Full size table

Discussion

This systematic review and meta-analysis demonstrated the beneficial effects of HM for treating AR. HM improved total nasal symptoms, individual nasal symptoms, total ocular symptoms, and disease-specific quality of life. These beneficial effects persisted in the high quality RCTs subgroup analysis. In contrast, there were no differences between the HM and placebo in the trials with high risks of bias subgroup. In addition, HMs brought beneficial effects like standard treatments, including antihistamines and intranasal corticosteroids. HMs with anti-inflammatory activities contains plant steroids, of which the structure is close to corticosteroids [50]. Subgroup analyses showed that HMs with anti-allergic effect and anti-inflammatory effect were effective. These effects controlled the early-phase and late-phase symptoms. In addition, anti-histamine and anti-leukotriene effects were also revealed. Jung et al. [36] demonstrated the ability of fermented red ginseng to suppress the wheel and flare response in SPT as a part of the anti-histamine effect. Butterbur and Pycnogenol showed the ability to inhibit leukotriene biosynthesis, like zileuton [21, 24, 25, 27, 34]. These effects decrease mucus hypersecretion in the airways and enhance mucociliary functions [51]. Choosing an appropriate HM should be based on the mechanisms of action of the HM that could improve the prominent symptoms of AR.

The results of this study showed the benefits of HMs up to twelve weeks duration then the benefits decreased. Tachyphylaxis has been known for a long time in other medicines such as antihistamines, intranasal decongestants, and opioids. To the best of our knowledge, this is the first systematic review showing evidence of tachyphylaxis of herbal medicine. There is no evidence regarding whether increasing the dose of HM can restore the original response. Physicians should be aware that the HM response decreases after three months of treatment. The subgroup analyses by AR subtype showed that both the patients with SAR and PAR benefited from the HMs. However, the patients with PAR experienced tachyphylaxis after 4 weeks.

To date, the evidence supporting the HM treatments for AR is unclear. The recommendation of HM is controversial [8•]. A systematic review utilized a modified Delphi method by Wu et al. [6•] showed that butterbur extract was one of the potential alternative treatments for sinusitis and rhinitis. Unlike Western HMs, Eastern HMs have composite ingredients containing different herbs. Therefore, it is difficult to identify the original or individual component that provides the primary beneficial effects. Lenon et al. [39] studied a new formula that was developed from an existing HM formula, by selecting 7 out of the 18 individual herbal ingredients and found no differences between the HM and placebo. There are four meta-analyses evaluating the effects of Chinese HMs. These meta-analyses included several studies that were published in Chinese. However, those studies were not included in our review. A meta-analysis by Wang et al. [9] showed the benefits of Chinese HM over placebo or inactive comparator in the assessment of TNSS. In contrast, Zhang et al. [12•] reported no differences in TNSS or individual nasal symptom scores between the HM and placebo or inactive comparator. Although they found beneficial effects on RQLQ favoring the HM, the heterogeneity was substantial. Another meta-analysis by Zheng et al. [52] assessed pediatric AR from 19 RCTs and showed benefits of Chinese HM over antihistamines. Luo et al. [3] assessed adult patients with AR from 23 RCTs and showed that the Chinese HM formula, Yu ping feng san, was effective for managing adult AR.

Based on the results of this meta-analysis, HM decreased nasal and ocular symptoms related to allergic rhinitis and improved quality of life with no difference from standard treatments. Nevertheless, beneficial effects did not persist after 12 weeks. In addition to the benefits of HM as a sole therapy, its role as an addition to standard treatment also had favorable therapeutic outcomes [17]. Arpornchayanon et al. [45] assessed the effects of cetirizine and HM combination and showed that the combination was superior to cetirizine and placebo. The findings of this study showed that HM was safe and tolerable. This is in agreement with previously published articles which reported no differences in adverse events between the HM and control groups. However, diarrhea or liver toxicity was reported in some cases [8•, 11, 49]. In clinical practice, the authors suggest that HM should be considered as a primary treatment only for a short-term treatment. Standard treatments such as antihistamines and intranasal corticosteroids are the first-line drugs for the long-term treatment while HM can be used as an option or as an adjunct to standard treatment to boost up the treatment effect [6•].

This study had several limitations. The systematic search did not search for articles published in languages other than English. Therefore, our meta-analysis could not cover all current studies. In addition, the included studies had high heterogeneity for outcomes assessment. Subgroup analyses, by mechanism of action of the HM, the AR subtype, and quality of the included studies, were conducted to investigate the heterogeneity. The heterogeneity persisted because different kinds of HM were investigated together.

Conclusion

Evidence from this meta-analysis showed the benefits of HM for treating AR patients. HMs improved nasal symptoms, ocular symptoms, and disease-specific QOL when compared to placebo. Beneficial effects of HMs were similar to standard treatments but only revealed in a short-term treatment, less than 12 weeks. In general, HM is considered safe. In practice, standard treatments such as antihistamines and intranasal corticosteroids should be considered for a long-term treatment.