Chinese herbal medicines for rhinosinusitis: a text-mining study with comparisons to contemporary research and clinical guidance

Abstract

Background

Rhinosinusitis is an inflammation of the paranasal sinuses and nasal cavity. It is managed with intra-nasal steroids, nasal saline irrigation, oral antibiotics and/or biologics. Chinese herbal medicines (CHMs) have long been used for nasal disorders, including rhinosinusitis, and feature in Chinese clinical guidelines for rhinosinusitis. Systematic reviews suggest some CHM formulations may be beneficial for the management of this condition.

Methods

This text mining study used an established methodology to search a database containing the full texts of more than 1,150 classical and pre-modern Chinese medicine books to identify references to disorders similar to rhinosinusitis, and the CHMs used as interventions. Ten search terms relevant to names of disorders in classical books and the major symptoms of rhinosinusitis were identified. Search results were downloaded, categorised, and analysed descriptively in SPSS®. Qualitative comparisons were made with the modern clinical Chinese medicine literature.

Results

Searches found 436 oral CHM formulae with 3,059 ingredients, 81 nasal or other topical CHMs with 142 ingredients, and 112 single natural products from classical pharmacopoeia used orally and/or topically. The earliest reference to a disorder similar to rhinosinusitis was in a Chinese medical book written approximately 2,000 years ago.

Three oral CHM formulae from the search have been tested in randomised controlled trials and one is included in a clinical guideline. A sample of 13 modern specialist textbooks on Chinese otorhinolaryngology still recommended nine of the oral classical formulae. Three of the seven herbs frequently included in the classical pharmacopoeia are still listed in the official Chinese pharmacopoeia for rhinosinusitis. Nasal formulae found in classical searches differed to those tested in randomised controlled trials, although the most frequent ingredient in the classical formulae was also frequent in modern formulae.

The pharmacological actions of the seven most frequently used herbs included anti-inflammatory, anti-allergic, antioxidant and/or anti-bacterial effects that may have contributed to their observed effects.

Conclusions

Results for classical interventions and the modern clinical literature overlapped for the higher frequency oral formulae and their ingredients, showing evidence of temporal continuity in their use for rhinosinusitis management. Gaps in the clinical and experimental evidence were identified, so there remains scope for further research into these CHMs to develop new interventions for rhinosinusitis.

Background

Rhinosinusitis (RS) is an inflammation of the paranasal sinuses and nasal cavity. Its symptoms include nasal blockage, obstruction and/or congestion; and nasal discharge (anterior/posterior nasal drip), with or without facial pain/pressure and/or reduction or loss of smell. RS can be verified by endoscopy or computed tomography (CT) [1,2,3].

When symptoms have continued for less than four weeks, it is acute RS (ARS). When two or more symptoms persist for 12 weeks or more, it is chronic RS (CRS) [3]. People with CRS can be broadly divided into those with and those without nasal polyps [2, 3]. Another category of RS is fungal [4].

In China, this disorder can be referred to by the traditional name bi yuan, the biomedical name, sinusitis (bi dou yan) and by the most recent biomedical term, rhinosinusitis (bi-bi dou yan). In traditional Chinese medicine (TCM), bi yuan is diagnosed according to the clinical symptoms of nasal blockage, turbid nasal discharge, facial pain and/or pressure and/or headache, and/or reduction in the sense of smell [5]. This clinical diagnosis does not require confirmation by endoscopy and/or computed tomography, although these tests are used in many TCM hospitals in China. In general, bi yuan encompasses disorders that would be biomedically diagnosed as ARS or CRS.

In conventional medicine, CRS management typically requires intra-nasal steroids, with nasal saline irrigation and/or oral antibiotic therapy during exacerbations, plus allergen avoidance in people with allergies [2, 3]. More recently, biologics have been added [6]. When these approaches don’t sufficiently control symptoms, surgery may be considered [1, 2]. Although these approaches are generally effective, antibiotic resistance remains an issue [7, 8], so there is a need to identify candidates for future product development.

Chinese herbal medicines (CHMs) have long been used for nasal disorders, including RS, and feature in Chinese clinical guidelines [5, 9,10,11]. Although termed ‘traditional’, modern TCM combines traditional practices with research results, uses modern diagnostic methods, and may combine TCM interventions with conventional biomedicine as ‘integrative Chinese–Western medicine.’ In China, TCM is part of the national healthcare system, with 4,630 Chinese medicine hospitals and 756 integrative Chinese–Western medicine hospitals operating in 2021 [12, 13].

From the perspective of evidence-based medicine, it has been proposed that traditional use that has been well documented could be considered a source of evidence [14], and such evidence could be considered along with the results of clinical and experimental studies as components of a ‘whole evidence’ approach to evaluating TCM interventions [15].

In the field of herbal medicine, systematic investigation of pre-modern texts has been proposed as a bioprospecting strategy in the search for new drugs [16, 17]. Subsequently, several text mining studies have been published based on European [18,19,20], Korean [21, 22] and Chinese sources [23,24,25,26,27,28], but none focussed on RS.

This text mining study identifies traditional formulations and their constituent natural product ingredients that have been used for conditions similar to RS during the classical and pre-modern period (until 1949). It then compares these with approaches recommended in contemporary CHM guidelines for RS management and examines the contemporary clinical research into these formulations. The study aimed to investigate the historical use of CHMs for RS and identify prospects for future clinical and experimental studies, which may lead to development of new RS treatments.

The research questions are:

• Which CHM interventions were used in ancient and pre-modern China for nasal disorders similar to RS?

• What are the similarities and differences between ancient CHM interventions, those included in contemporary clinical guidelines, and those tested in modern randomised controlled trials (RCTs)?

• To what extent does the information on CHM in modern specialist textbooks reflect the classical literature and clinical guidelines?

• What are the likely mechanisms of action of the frequently used herbs?

• Are there any gaps in the evidence that could be addressed in future research?

Materials and methods

Classical and pre-modern Chinese medical literature was searched electronically using the Encyclopaedia of Traditional Chinese Medicine (Zhong Hua Yi Dian –ZHYD) 5 th edition. This database contains the full texts of more than 1,150 classical and pre-modern Chinese medical books [29] written between the second century BCE and the mid twentieth century. It is one of the largest collections of such books [30]. Comparisons with other collections of classical books on Chinese medicine found it was representative of the classical and pre-modern medical literature in Chinese [31].

The text mining approach followed a procedure that has been described in detail elsewhere [32]. Terms reportedly used to refer to sinusitis and RS in the pre-modern and ancient Chinese literature were identified from medical dictionaries [33, 34], modern textbooks (Additional file 1) and clinical guidelines [5, 9,10,11]. Then test searches of ZHYD were conducted of all included books. Terms that did not locate any passages of text relevant to nasal disorders that could have been RS were discarded.

Books and guidelines identified the main names for sinusitis as bi yuan and nao lou. Additional terms related to nao lou included nao shen, nao beng, nao xie, kong nao sha and nao sha (all probable synonyms for ‘sinusitis’). In addition, we included two symptoms from the books and guidelines: zhuo ti (turbid nasal discharge) and bi chu chou qi (foul nasal odour).

We did not include the term bi se, since it could refer to many types of nasal blockage, or bi qiu, since it referred to conditions with clear rather than turbid nasal discharge and is a contemporary term for allergic rhinitis. We also did not include terms related to growths in the nose that could have been polyps (e.g. xi rou, bi zhi). Test searches showed these names alone did not locate good examples of RS. When a citation included one of these excluded terms in addition to the included search terms, it was still included in the data set, since treatments may have been used for multiple nasal disorders.

Search results from all books included in ZHYD were downloaded as semi-structured text. We identified distinct passages of text that referred to one or more of the search terms and included information relevant to symptoms, aetiology, definition of the disorder and/or medical management. These passages were defined as a ‘citation’ and were numbered and entered in a spreadsheet, along with the name of the book in which the citation was located, dynasty and year in which the book was written, description of the disorder, and intervention. Duplicate citations from the same book were identified and excluded. Books written after 1949 were excluded [32].

Citations were read in detail by at least two researchers (WL, JC, and/or BM) and scored yes (= 1) or no (= 0) based on pre-set criteria. Any issues were resolved through discussion between the two researchers, and where necessary with a third person (BM or LQ). If there was no consensus on the meaning of a citation, it was excluded. We did not calculate inter-rater reliability.

When a CHM intervention was mentioned, the formula name (if any) and ingredients were added in separate columns. Formula names were according to the original book. When there was only a list of ingredients the citation was specified as ‘no name’ even when we could infer a likely formula name. In such cases the ingredients were still included in the data sets for herb frequency. When a citation included a formula that had additional ingredients these were regarded as modifications and it was grouped with other formulae with the same name, except when the main ingredients of the base formula were different. In such cases numbers were allocated to the formula name. When there were differences in the name of the same formula (based on the ingredients) such as Cang er san and Cang er zi san (Xanthium powder), these were grouped together as Cang er (zi) san for analysis. Similarly, when the preparation type was different, such as tang (decoction) or san (powder), but the ingredients were the same, these citations were grouped together for analysis. Individual herbs from classical pharmacopoeia were calculated separately.

Codes were allocated to indicate oral CHM, nasal CHM, or other therapies (not included in this paper). We could not reliably distinguish ARS from CRS, because duration of the disorder was not typically included in classical citations. When the description was not consistent with RS signs and symptoms (based on the pre-set criteria), the citation was excluded. Analyses were conducted in SPSS® using descriptive statistics to identify frequencies of search terms, herbal formulae, and their constituent herbs. Cross-tabulations included the source book, year and dynasty during which the book was written and the main disorders or symptoms.

Comparison with contemporary TCM practice was based on the 2012 clinical guideline for TCM, because it included syndrome differentiation plus recommended formulae [5], a convenience sample of 13 modern specialist textbooks on Chinese otorhinolaryngology sourced from libraries in China and Australia and internet searches (Additional file 1), and the official Chinese pharmacopoeia [35]. Comparisons with clinical research were based on two published systematic reviews of CHMs for RS [36, 37] and a recent book with chapters on CHM research [38]. Comparisons were qualitative to identify overlaps and discrepancies in the formula names and formula ingredients between pairs of data sets. All items were considered in comparisons although the results tables focus on the higher frequency items. Ranking based on frequency was used to facilitate comparisons between the formulae and their ingredients in the data sets derived from the classical books and the RCTs. No tests for statistical significance were conducted. Comparisons with the smaller data sets were based on the presence or absence of each item.

Results

ZHYD database searches yielded 1,901 results for the eight search terms. Following duplicate removal and exclusion of irrelevant and unclear passages of text, 678 citations were relevant to RS and included a Chinese medicine (CM) intervention (Fig. 1). Of these, 436 were of oral CHM formulae, 81 were of nasal or other topical CHMs, 112 were single items from classical pharmacopoeia and book sections on materia medica (ben cao) used orally and/or topically, and the remainder (n = 49) referred to acupuncture and related therapies.

Fig. 1

Flow diagram of the search and selection process for citations of interventions for disorders similar to rhinosinusitis

The search term bi yuan located 541 citations relevant to RS (79.8% of the total) followed by zhuo ti (148 citations, 21.8%) and nao lou (94 citations, 13.9%) (Table 1).

Table 1 Hit and citation frequency by search term

Oral formulae and their constituent ingredients

Of the 436 citations of oral herbal formulae, 105 were of named formulae and 168 formulae were not named. Most formulae (n= 277, 63.5%) were from the Qing dynasty (circa 1645–1911), and 128 formulae (29.4%) were from the Ming dynasty (circa 1369–1644) (Table 2).

Table 2 Dynastic distribution of citations relating to oral herbal formulae

Comparisons were made between the results of the searches of ZHYD and multiple samples from the modern literature on CHM for RS. The main comparisons are shown in Fig. 2.

Fig. 2

Diagram of comparisons between data sets. Abbreviations: CHM Chinese herbal medicine, RCT randomised controlled trial, ZHYD Encyclopaedia of Traditional Chinese Medicine (Zhong Hua Yi Dian)

The most frequent formula name in the ZHYD results was Cang er (zi) san (Xanthium powder) (n = 47), followed by Fang feng san (Saposhnikovia powder) (n = 25). These two formulae were first listed in books written during the Song dynasty (960–1279 CE) (Table 3).

Table 3 Frequent oral herbal formulae in classical citations by inclusion in guideline, modern textbooks, and clinical trials

One low frequency oral formula from the classical citations (Wen fei zhi liu dan, n = 1) was listed in the 2012 clinical guideline (Table 4), nine formulae from the classical citations (including modified versions) were listed in modern textbooks on Chinese otorhinolaryngology (Additional file 2), and three formulae found in the classical search were assessed in RCTs included in the systematic reviews (Table 3, Additional file 3).

Table 4 Summary of syndrome differentiation and oral herbal formulae from 2012 guideline by inclusion in classical search and clinical trials

In the 436 formulae for oral use, there were 3,059 ingredients (mean, 7 ingredients per formula). The most frequently used ingredients were Magnolia spp (xin yi (hua), n = 162), Angelica dahurica (bai zhi, n = 152), Glycyrrhiza spp (gan cao, n = 143), Xanthium sibiricum (cang er (zi), n = 135), and Mentha spp (bo he, n = 135). Most of these herbs were included in multiple books written from the Song dynasty (960–1279 CE) until the early twentieth century (until 1949) (Table 5).

Table 5 Frequently used ingredients in oral herbal formulae in classical citations by inclusion in clinical trials and modern Chinese pharmacopoeia

The comparison between the 22 most frequently used ingredients in classical formulae and ingredients of formulae tested in RCTs was based on one systematic review of oral CHMs for ARS and CRS that included 34 RCTs [36], and one systematic review of CRS following surgery that included 21 RCTs [37]. The first seven herbs from classical formulae, were in the top 10 ingredients in the formulae tested in the RCTs in each systematic review (Table 5). The Chinese pharmacopoeia [35] listed four of the 22 most frequently used ingredients for RS, each for bi yuan.

In the 112 classical pharmacopoeia citations, the most frequent herbs for oral use were Magnolia spp (xin yi (hua), n = 16), Xanthium sibiricum (cang er (zi), n = 14), Angelica dahurica (bai zhi, n = 13) and Piper longum (bi ba, n = 12). Of these, the three most frequent are still listed in the Chinese pharmacopoeia for bi yuan (Table 6).

Table 6 Frequently mentioned herbs for internal use in classical pharmacopoeia by inclusion in modern Chinese pharmacopoeia

Nasal formulae and their constituent ingredients

Fifty-two classical literature citations mentioned a CHM intervention for nasal application, mainly for bi yuan (n = 29). Most interventions had no name (n = 49). The exception was the powder Liu sheng san (n = 3). It could be sniffed into the nose or mixed with water and blown into the nose. This formula is not used in contemporary CHM [38]. The interventions used 142 ingredients. The most frequent were: Magnolia spp (xin yi (hua), n = 7) and moschus (she xiang, n = 7), followed by Allium fistulosum (cong (bai), n = 6) (Table 7).

Table 7 Frequently used ingredients in nasal herbal formulae in classical citations by inclusion in clinical trials and modern Chinese pharmacopoeia

The comparison with RCTs was based on a systematic review that included six RCTs: one of nasal plus oral CHM for ARS, four of nasal CHMs for CRS, and one of nasal plus oral CHM for ARS [38]. One of the high-frequency CHMs in the nasal formulae from the classical literature, Magnolia spp. (xin yi (hua)), was the highest frequency ingredient (n = 6) in the nasal formulae tested in the RCTs, but the other high-frequency ingredient, moschus (she xiang), was not used in any RCTs. Overall, of the 11 most frequently used ingredients from classical nasal formulae, four were used in one or more RCTs (Table 7). The Chinese pharmacopoeia listed two of the ingredients for RS.

Discussion

Classical nomenclature

The term bi yuan indicates a disorder of the nose (bi), with the second character (yuan) suggesting a deep pool or well of water. It appears in the classical book Huang Di Nei Jing Su Wen ‘Huang Di’s Inner Classic – Basic Questions,’ which is considered a product of the Former Han dynasty (circa 206 BCE–24 CE), with additions from the later Han dynasty (circa 25–220 CE) [39, 40]. Chapter 37 (Qi jue lun) states: ‘heat from the Gall bladder [channel] enters the brain causing a burning feeling where the nose meets the forehead, and bi yuan. Bi yuan is continuous turbid discharge.’ This explanation is in accord with the literal meaning of the term bi yuan, and adds that the discharge is turbid, due to heat, and affects the brain. In reference to the external pathogen ‘heat qi’ (re qi), Chapter 74 (Zhe zhen yao da lun) states: ‘if it invades the lung, it [produces] cough and bi yuan’ [41]. This text again notes heat as a causative agent, but that it affects the lung and is associated with cough. The first of these two passages was quoted in multiple included citations. In terms of the aetiology, these two passages are reflected in the names of Chinese medicine syndromes – Stagnant Heat in the Gallbladder (dan fu yu re) and Wind–heat in the Lung Meridian (fei jing feng re) in a modern clinical guideline that includes five syndromes [5].

The term nao lou appeared much later than bi yuan in the literature. It literally means ‘brain leakage’ and appears to paraphrase Chapter 37 above. The earliest use of this term we identified was in the book Chi Shui Xuan Zhu ‘Black Pearl from the Red River’ (circa 1584), which described the symptoms of nao lou as ‘nasal discharge that was clear and/or turbid which lasted many years’ [38]. Notably, in the modern clinical guideline, the syndrome Wind–heat in the Lung Meridian includes white or yellow discharge plus cough among typical symptoms [5].

The book Wai Ke Zheng Zong ‘Orthodox Manual of External Diseases’ (circa 1617) states that nao lou was also called bi yuan [42], and subsequent classical books tended to agree [38]. Of the low-frequency search terms, nao shen was not the name of a disorder, just a term meaning ‘brain leaks’, and the terms nao beng, nao xie and (kong) nao sha (Table 1) appear to have been synonyms for nao lou [38].

From a modern perspective, we cannot be certain that any of the people diagnosed as having these disorders in premodern China would now be diagnosed as having RS. Based on the symptoms of the disorders described in the source books, these disorders were not inconsistent with RS. However, we do not know the prevalence of RS in premodern China and can only assume that it was not a rare disorder. It is also likely that a small proportion of disorders referred to as nao lou or its synonyms referred to leakage/discharge of fluid from the skull through the nose. It is notable that dizziness was a symptom of long-term nao lou [42]. In other cases, allergic rhinitis may have been the disorder, although this is usually included under the term bi qiu, and a wide range of other disorders in which the salient symptom was nasal discharge could have been included within the scope of bi yuan, nao lou and related terms. Nevertheless, a large proportion of the included cases would now likely receive a TCM diagnosis of bi yuan (sinusitis).

Comparisons between classical literature and modern practice

When compared, modern and classical TCM concepts of bi yuan and the interventions used for this disorder show similarities. When the 2012 clinical guideline was used as a reference point, only one of the five recommended oral formulae appeared in the classical search (Table 4). However, when the comparison was broadened to include 13 modern TCM specialist textbooks, that included about 15 different syndrome types and about 60 recommended formulae, nine classical formulae (15%) were still included (Additional file 2). Three of these formulae have been tested in RCTs (Additional file 3).

All the syndromes and oral formulae in the 2012 guideline (Table 4) were also included in one or more of the 13 specialist textbooks (Additional file 2), so the guideline appears to be a subset of modern literature. Overall, the textbook sample provided a broader scope of syndromes and formulae than the guideline and was probably representative of the scope of modern clinical practice.

In terms of evidence base, four of the five oral formulae in the 2012 guideline have been evaluated in one or two RCTs for people with CRS who also had the Chinese medicine syndrome appropriate to the formula (Table 4). Of the 33 named oral formulae tested in the RCTs, 14 (42%) were included in the specialist textbooks (Additional file 3).

When compared, oral formula ingredients that ranked highly in the ZHYD list (ranks 1–7) also ranked highly in the two systematic reviews (ranks 1–8) (Table 5). This suggests that despite the diversity in formula names, formulae tended to be based on similar herbal ingredients. This similarity was likely enhanced by the practice of modifying formulae to add herbs commonly used for RS.

In addition, when classical and modern pharmacopoeia were compared, four of the top seven herbs from the ZHYD list were still listed for bi yuan (sinusitis) in the Chinese pharmacopoeia (Table 6). Notably, three of these were high-frequency items in the RCTs. However, one of the herbs listed in the Chinese pharmacopoeia, xi xin (Asarum spp.), was much less frequently used in the classical formulae (rank 19) and RCTs (ranks 26 and 21) (Table 5). The use of this herb is restricted in many countries due to toxicity concerns.

Data sets were much smaller for nasal formulae than oral formulae. This meant we could not compare formula names because there were few formula names in the ZHYD data, and the 2012 clinical guideline did not include nasal formulae. Four of the 11 nasal formula ingredients (ranked 1–8) were also ingredients in formulae tested in RCTs, and two of these were listed in the Chinese pharmacopoeia (Table 7). A notable difference between classical and modern nasal formulae was the use of strongly aromatic ingredients in the classical formulae, such as moschus (she xiang), allium (cong bai), borneol (bing pian), pogostemon/agastache (huo xiang), frankincense (ru xiang) and myrrh (mo yao), whereas only borneol (bing pian) was used in RCTs. This may be because inhalations were common interventions in classical books, but the RCTs mainly used nasal irrigations.

When the herbs in the three searches of ZHYD were compared, of the top ten in the oral formulae, the herb with the highest ranking, Magnolia spp. (xin yi (hua)) was also the top-ranking herb in nasal formulae and in the classical pharmacopoeia (Additional file 4). Other herbs appearing in all three classical data sets were Angelica dahurica (bai zhi), Mentha haplocalyx (bo he), and Ligusticum chuanxiong (chuan xiong). Xanthium sibiricum (cang er (zi)) ranked highly in the oral formulae and the classical pharmacopoeia but was not present in the classical nasal interventions, although it was used in the RCTs of nasal interventions [38].

Possible mechanisms of action

The seven most frequently used herbs in classical oral formulae that were also frequently used in RCTs each show biological activity relevant to RS in experimental studies that tested their extracts and/or constituent compounds [38]. These activities are summarised below.

Xin yi (hua) is sourced from the flower buds of multiple magnolia species, with Magnolia biondii Pamp. (aka M. fargesii), M. denudata Desr. and M. sprengeri Pamp. being official. It has anti-inflammatory, antihistamine-like and antioxidant effects. Bai zhi, sourced from Angelica dahurica (Fisch ex Hoffm) Benth et Hook f; and A dahurica (Fisch ex Hoffm) Benth et Hook f var formosana (Boiss) Shan et Yuan, has anti-inflammatory, antihistamine-like, anti-microbial, and antioxidant effects. Gan cao, sourced from Glycyrrhiza uralensis Fisch, G. inflata Bat and G. glabra L roots, has anti-inflammatory properties and free radical scavenging activity [38].

Dried fruit from Xanthium sibiricum Patr (aka X. strumarium L) is the source of cang er zi, which has anti-inflammatory, anti-bacterial and antiviral effects, and inhibitory effects on histamine release. Bo he, sourced from the arial parts of Mentha haplocalyx Briq (aka M. canadensis L and M. arvensis L var haplocalyx Briq), has anti-inflammatory and free radical scavenging effects, and inhibitory effects on bacterial growth and histamine production. Chuan Xiong, sourced from Ligusticum chuanxiong Hort (LC) (aka L. wallichii Franch) roots and rhizomes, has anti-inflammatory and free radical scavenging activity. Huang qin, sourced from Scutellaria baicalensis Georgi roots, has anti-inflammatory, anti-allergic, antioxidant, anti-bacterial and anti-fungal effects [38].

In addition to the above seven most frequently used herbs, classical literature searches identified some herbs included frequently in classical pharmacopoeia as interventions for RS that did not appear in the official Chinese pharmacopoeia for this indication (Table 6). Of note is Piper longum L (bi ba), which is also used in foods. It was listed in the most comprehensive of the pre-modern pharmacopoeia, Ben Cao Gang Mu ‘Compendium of Materia Medica’ (circa 1578), for headache and bi yuan, and is still listed for these and other indications in some modern comprehensive pharmacopoeia [43, 44]. Therefore, we searched experimental literature in PubMed to investigate whether this herb has shown activity that may help explain its historical use for RS and related disorders.

Piper longum L fruit contains multiple compounds, including the alkaloids piperine and piperlonguminine, both of which were detected in rat plasma after oral administration, indicating these compounds were absorbed rapidly and cleared slowly over 24 h [45]. Piper longum has shown free radical scavenging activity [46]. In rats, an aqueous extract of its fruits showed low toxicity [47]. A study in rats that tested the dried leaf showed moderate antibacterial activity against Klebsiella pneumoniae, but not against other bacteria included in the screen. Significant anti-inflammatory activity was seen in the carrageenan-induced paw oedema model and other tests [48]. Oral administration of oil extracted from the dried fruit significantly reduced carrageenan-induced paw oedema [49]. Fruit extracts showed anti-inflammatory activity in the same model [47]. Powdered fruits of two varieties of Piper longum showed anti-inflammatory effects in carrageenan-induced paw oedema and formalin-induced paw oedema [50]. In models of Parkinson’s disease, alkaloid extracts (mainly piperine and piperlonguminine) reduced behavioural impairments, increased antioxidant activity, and reduced excessive proinflammatory cytokine release induced by injection of lipopolysaccharide [51, 52]. Overall, these studies showed anti-inflammatory activity in multiple animal models, but none were specific to RS.

Limitations

One limitation of this text mining study is that we used a single source for literature. While this was the largest collection available, it did not include all books written in pre-modern China, so we might have missed some less significant historical publications relevant to RS. Some reports of conditions similar to RS may not have used any of the search terms we identified, even though we used multiple search terms, so these may have been missed.

Although the citations from the classical literature referred to conditions similar to the modern conception of RS, the criteria for similarity were based on classical terminology for which the scope of meaning is likely to have been different to current usage. Therefore, it remains unclear whether a nasal disorder referred to in a classical citation would now be classified as RS. Moreover, we could not reliably distinguish between ARS and CRS. We assume that CRS was more likely in the classical citations since classical books tend to focus on more serious conditions, but this remains an assumption. There were few studies of ARS in the systematic reviews, so comparisons between classical results and clinical trial results were not feasible for ARS.

In the comparisons, we used a sample of modern TCM specialist textbooks based on availability. There was no deliberate selection of books, but as this was not a random sample of all possible books, bias is a possibility.

Analyses were descriptive based on frequency, rank and overlap between data sets. Higher frequency in a data set is an indication of usage or popularity and should not be misconstrued as indicating greater effectiveness of a formula or herbal ingredient for RS management. Similarly, presence or absence of an item in multiple data sets is not a measure of effectiveness.

Our selection of Piper longum for a mini review was based on its frequency alone, and other lower frequency CHMs may have shown stronger evidence.

Evidence gaps

Four of the five oral formulae in the 2012 guideline have been tested in RCTs (Table 4, Additional file 3), but the studies tended to have small sample sizes and be open label, so further, more rigorously designed studies are needed. Moreover, an RCT of the formula Yin Qiao San (Honeysuckle and forsythia powder) for RS was yet to be published and there have been few clinical studies of ARS.

The three most frequent formulae tested in the RCTs appeared in the sample of TCM textbooks but were not included as recommendations in the guideline (Table 4). This may be because these are modern commercially available preparations, whereas the guideline focussed on traditional formulae. Considering the advances in evidence based Chinese medicine over the past 20 years [53], there appears to be scope for TCM guidelines to be updated to include additional oral and nasal formulae based on the findings of clinical research and meta-analyses. This could assist practitioners to base their clinical practice on the best available evidence.

The summary of experimental studies of CHMs frequently used for RS did not find any studies in animal models specific for RS. While the CHMs’ reported benefits in RS could be due to their anti-inflammatory, anti-allergic, antioxidant and/or anti-bacterial effects, further experiments are needed to elucidate the effects of these herbs and their constituents in models directly relevant to RS. Moreover, since the herbs are used clinically as multi-ingredient formulae, their combinations could also be tested.

Conclusions

This text mining study identified references to disorders similar to RS in a Chinese medical book from approximately 2,000 years ago. Results for classical interventions overlapped with findings from two systematic reviews and a sample of 13 specialist textbooks. This showed evidence of temporal continuity in the use of certain orally administered CHMs in RS management. Nevertheless, there is scope for further clinical and experimental research into the frequently cited CHM formulations and their ingredients to inform clinical decision making and their mechanisms of action could be explored in animal models to identify future therapeutics.