Abstract
Objectives:
To identify and analyze records of the treatment of dementia and memory disorders in the classical Chinese medical literature that were consistent with the signs and symptoms of Alzheimer's disease (AD), with the aim of determining which traditional medicines have histories of use for these disorders.
Methods:
Encyclopedia of Traditional Chinese Medicine (Zhong Hua Yi Dian), a database of more than 1000 classical and premodern Chinese medical books, was systematically searched. Search terms were identified from dictionaries, medical nomenclatures, guidelines, and specialist clinical manuals on aging, neurology, or brain disorders. Inclusion and exclusion criteria were used to identify citations of conditions whose signs and symptoms were consistent with the clinical features of AD. Passages of text identified by these terms were copied to Microsoft Excel spreadsheets, together with the identity of the source book and all relevant information on the disorder and the intervention. Each distinct passage of text was considered a citation. The frequencies of the traditional formulas used as interventions and their constituent ingredients were calculated.
Results:
The selection criteria identified 1498 citations of dementia and memory impairments derived from 277 different books written from circa 363 to 1945 AD. In 91 of these citations, memory impairment was associated with aging and was broadly consistent with the clinical features of AD. Although the interventions varied in name, Poria cocos, Polygala tenuifolia, Rehmannia glutinosa, Panax ginseng, and Acorus species consistently appeared as ingredients in multiple formulas for memory impairment in the context of aging.
Conclusions:
Memory impairment in older age was a recognized condition in the classical literature. Many of the traditional medicines frequently used as ingredients in classical formulas for memory impairment consistent with clinical features of AD remain in contemporary use, and experimental studies suggest biological activities relevant to AD.
Notes on Terminology
For Chinese terms, the Pin Yin system of transliteration is used. Chinese book names are translated into English, but formula names can be untranslatable so these are given in Pin Yin only. For names of traditional medicines in the text, the full Latin binomial and the Chinese name in Pin Yin are given in the first instance. In the text, Chinese disease or symptom names are in lower case italic (e.g., chi dai), herb names are given in lower case (e.g., yuan zhi), formula names have the first letter capitalized (e.g., Gui pi tang), and book names have each letter capitalized and appear in italics (e.g., Zhong Hua Yi Dian). A full glossary of the Chinese terms and names in Pin Yin used in the text appears in the supplementary material (Supplementary Table S1; Supplementary materials are available online at
Introduction
D
AD is a significant and increasing health issue in contemporary China and other Asian countries. 8,9 Traditional medicines (TMs) are commonly used in China for prevention and/or treatment of dementia, and research into TMs for dementia is a growing field. 10 Some conventional drugs used for AD have plant origins, including the acetylcholinesterase inhibitors galantamine, which is found in Galanthus, Lycoris, Narcissus, and Leucojum species, 11 and rivastigmine, which is a semi-synthetic derivative of physostigmine from Physostigma venenosum Balf. 12 Therapies of Asian botanical origin for which there is evidence of efficacy include huperzine A, derived from Huperzia serrata (Thunb. ex Murray) Trevis. 13,14 and EGb761, which is a standardized extract of Gingko biloba L. leaves. 15,16 Other plants used in traditional medicines show promise based on animal and human studies. 17 –23
While evaluation of plants and plant-derived compounds using experimental methods remains the mainstay of drug discovery from natural products, traditional use has been proposed as a valid source of evidence, 24 and text mining of the literature on traditional medicine has been used to research disease names 25 and identify plants and compounds for further study. 26 –31
In a previous study of Chinese TMs for dementia, Great Compendium of Chinese Medical Formulae (Zhong Yi Fang Ji Da Ci Dian [ZYFJDCD]), the largest printed book on Chinese formulas was searched; this includes 96,592 individual formula entries extracted from more than 685 books. 32 In recent times, voluminous collections of traditional Chinese medical literature have become available as digitized versions. 33 The current paper presents the results of a systematic analysis of Encyclopaedia of Traditional Chinese Medicine (Zhong Hua Yi Dian [ZHYD]), which is a database of more than 1000 classical and premodern medical books written or published before 1950. 33,34 About half (55.2%) of the books in ZYFJDCD also appear in ZHYD, 35 but these two sources differ in other important ways. ZYFJDCD is a curated book that does not include unnamed formulas; each named formula is listed once, except when the ingredients are substantially different; clinical applications tend to reflect the oldest citation, although there are some later additions; and the index is used to find formulas for a disease. 32 In contrast, ZHYD contains the full text of each book. 34 As a result, an electronic search can identify all citations of all formulas used as treatments for a particular disease or symptom in all the included books. Therefore, the appearance of a formula in multiple different books provides proxy measures of how many authors considered that formula worthy of inclusion and which formulas were more or less popular as interventions for the disease. Such an approach was not feasible with use of ZYFJDCD.
An earlier study by the current authors investigated individual TMs for dementia based on searches of materia medica (ben cao) entries in ZHYD 36 and compared the results to the comprehensive contemporary book Great Compendium of Chinese Medicines (ZYDCD). 37 In the present study, the data derive from the formula and clinical literature and do not overlap with the earlier study.
In contemporary Chinese medicine, dementia may be described by using modern terms, including “senile dementia” (lao nian xing chi dai), “Alzheimer's disease” (a er ci hai mo ji bing), and “vascular dementia” (xue guan xing chi dai) or the traditional terms “chi dai” (dementia, mental feebleness), “jian wang” (memory impairment, forgetfulness), and related terms. 38 –40 These traditional terms were based on clinical presentation and had broader scopes of meaning than modern terms, so it cannot be assumed that a disorder described by using these traditional terms would have resulted from AD pathology and would now be diagnosed as AD. However, descriptions of disorders in traditional books can be identified whose signs and symptoms were consistent with the clinical features of AD or other disorders. This enables comparisons between the historical and contemporary literatures.
The present study identified and analyzed records in ZHYD of the treatment of dementia and memory disorders to determine the following: (1) whether the signs and symptoms in any of the records were consistent with AD; (2) which formulas and formula ingredients were most frequently used for age-related memory disorders; (3) whether the results were consistent with the studies of ZYFJDCD 41 and ZHYD materia medica books; 36 and (4) whether the results were consistent with contemporary Chinese materia medica books and guidelines. 38
Materials and Methods
As the sample of classical literature, ZHYD was chosen because it is representative and largely inclusive of other large collections of the classical and premodern Chinese medical literature. 33,35 The general approach to searching this database, data extraction, and coding has been described elsewhere. 42
Briefly, search terms were identified from dictionaries, medical nomenclatures, and guidelines 38 –40,43 –45 and traditional specialist clinical manuals on aging, neurology, or brain disorders. 46 –56 Trial searches were conducted by using each term. Terms for which the scope of meaning was nonspecific were excluded. The final search terms were grouped into jian wang and synonyms, chi dai and synonyms, and other terms relating to memory disorders (Supplement 1).
All passages of text identified by these terms were copied to Microsoft Excel (Microsoft Corp., Redmond, WA) spreadsheets (by I.W.Z., S.L., S.C., B.H.M.), together with the identity of the source book and all relevant information on the disorder and intervention. Each distinct passage of text was considered a single citation irrespective of how many times the search term was mentioned. Duplications were identified and removed. Inclusion and exclusion criteria were used to identify conditions whose signs and symptoms were consistent with the features of AD. Each citation was read and allocated codes (by S.C., I.W.Z.). Complex citations were read by two or more researchers (S.Z., B.H.M., I.W.Z.), and additional input was sought to resolve difficult issues.
Exclusion criteria were as follows: (1) conditions with sudden or recent onset (trauma, fever, epidemic, or seasonal disorder); (2) disorders specific to children or young people, or female obstetric disorders; (3) psychiatric disorders with symptoms unlike those of AD (e.g., dian kuang, madness); (4) conditions likely due to a cerebrovascular accident (e.g., zhong feng, hemiplegia, paralysis), Parkinson's disease (e.g., tremor), or epilepsy (e.g., seizures); (5) no TM intervention; (6) materia medica entry listing clinical applications of a single substance; and (7) book written after 1949. Citations that mentioned symptoms consistent with behavioral and psychological symptoms of dementia were not excluded.
Inclusion criteria were as follows: (1) specific TM intervention comprising one or more ingredients intended as a treatment for one or more of the search terms; (2) mention or description of memory impairment; and (3) mention of older people, aging, or promotion of longevity.
Citations satisfying all the exclusion criteria plus the first inclusion criterion were reported as level 1. Citations also satisfying the second and third inclusion criteria were reported as level 2. Level 1 is based on formulas for conditions that could not be eliminated as possible instances of AD. Level 2 is a subgroup in which memory impairment was a symptom and the formula was associated with aging. Data were entered into SPSS software (IBM, Armonk, NY) for analysis. Results were cross-referenced to the total data sets of the previous analyses of ZYFJCDC 41 and the ZHYD materia medica entries. 36
For the contemporary literature on materia medica, the comprehensive ZYDCD was selected; this contains 5757 entries with clinical applications. 36 For contemporary TM formulas, Chinese Guidelines for the Diagnosis and Treatment of Alzheimer's Disease and Other Dementias 38 was used. These two sources provided points of comparison between the classical and contemporary clinical applications of individual TMs and formulas. Scientific names of TMs are from Pharmacopoeia of the People's Republic of China and ZYDCD. 37,57
Results
After removal of duplicates, 1878 citations of TM interventions were analyzed. The most commonly used search term was jian wang (1059 citations) followed xi wang (274 citations), both of which refer to memory impairments. The chi dai group of terms were infrequent (63 citations). For level 1, 1498 citations were identified (Supplementary Table S2). These were derived from 277 different books written from circa (c.) 363 AD to 1945. Most of the books were written during the Ming (1369–1644) and Qing (1645–1911) dynasties (Table 1).‘Prescriptions for Universal Relief (Pu Ji Fang, c.1406), which is the largest book in ZHYD, provided 171 citations. The next most productive book was the Song dynasty encyclopedia Complete Record of Sacred Benevolence (Sheng Ji Zong Lu, c.1117) with 56 citations.
The dividing points between dynasties are open to interpretation, so the years have been adjusted to avoid overlap. When authors lived across two dynasties, the dynasty usually cited for the book was adopted. See May et al. 42 for how book years were determined.
The level 1 citations referred to 105 unnamed formulas and 527 different formula names. Gui pi tang was the most common formula name, followed by Tian wang bu xin dan, Ding zhi wan (aka Ding zhi xiao wan), and Tao ren cheng qi tang (Table 2). All formulas included 14,415 ingredients (mean, 9.6 ingredients/formula) comprising 398 different TMs. The most frequently used TMs were sclerotium of Poria cocos (Schw.) Wolf (fu ling or fu shen, 1143), root of Panax ginseng C.A. Mey. (ren shen, 867), root of Polygala tenuifolia Willd. (yuan zhi, 761), root and rhizome of Glycyrrhiza species (gan cao, 670); root of Angelica sinensis (Oliv.) Diels (dang gui, 522), and rhizome of Acorus species (chang pu, shi chang pu, 417) (Table 3). Rehmannia glutinosa Libosch. (di huang) root was used 538 times, with the steamed root (shu di, 261) being the most common form.
First book in group: the oldest book within the group of included citations, not the first book that included the formula. Dates are approximate. Formulas with the same name can vary in their ingredients, and the same combination of ingredients may have different names. In these data, formulas with the same core ingredients and the same name are grouped together, while those with different main ingredients are separated. Also, formulas with the same ingredients but different names have been grouped together. The frequency is for the name in the left column, and the number in parentheses includes modified versions of the formula. For scientific names of ingredients written in Pin Yin and Chinese characters for traditional medicines and book names, see list in Supplementary Table S1.
c, circa.
Scientific names based on Pharmacopoeia of the People's Republic of China 2010 and/or Great Compendium of Chinese Medicines.
Clinical applications based on Pharmacopoeia of the People's Republic of China.
Used in materia medica books in Encyclopedia of Traditional Chinese Medicine for memory impairment.
Used in materia medica books in Encyclopedia of Traditional Chinese Medicine for memory impairment and aging.
When the level 2 criteria were applied, the remaining 91 citations derived from 40 books (c.652–1839) with the most productive being Longevity and Life Preservation (Shou shi bao yuan, c.1406). Two citations referred to unnamed formulas, and the remainder cited 56 different named formulas. Of these, Shou xing wan was the most frequent (n = 12) followed by Xiao dan, modified Gu ben wan, and Ju sheng wan (Table 4). Remaining on the list were Gui pi tang (n = 2), Tian wang bu xin dan (n = 2), Er dan wan (n = 1), and modified Ding zhi wan (n = 1). A total of 970 ingredients representing 140 different TMs were contained in the formulas. The most commonly used ingredients were P. cocos (fu ling, fu shen, 78), P. tenuifolia (yuan zhi, 61), R. glutinosa (di huang, shu di, sheng di, 51), P. ginseng (ren shen, 47), and Acorus species (chang pu, shi chang pu, 34) (Table 5).
Formulas with the same name can vary in their ingredients, and the same combination of ingredients may have different names. In these data, formulas with the same core ingredients and the same name are grouped together, while those with different main ingredients are separated. Also, formulas with the same ingredients but different names have been grouped together. The frequency is for the name in the left column, and the number in parentheses includes modified versions of the formula. For scientific names of ingredients written in Pin Yin and Chinese characters for TMs and book names, see list in Supplementary Table S1.
First book in group refers to the oldest book within the group of included citations, not the first book that included the formula. Dates are approximate.
In Great Compendium of Chinese Medical Formulae search for memory impairment and aging.
Scientific names based on Pharmacopoeia of the People's Republic of China 2010 and/or Great Compendium of Chinese Medicines.
Clinical applications based on Pharmacopoeia of the People's Republic of China.
Used in materia medica books in Encyclopedia of Traditional Chinese Medicine for memory impairment.
Used in materia medica books in Encyclopedia of Traditional Chinese Medicine for memory impairment and aging.
Discussion
Of the candidate search terms, the most productive were those that contained the term wang which means “forget.” The terms associated with chi dai identified considerably fewer citations, and only one remained after the criteria for aging were applied (Supplemental Table 1). Overall, references to old people were few in the literature, so most of the associations with aging were for promoting longevity. One of the earliest citations was in Essential Prescriptions Worth a Thousand Gold for Emergencies (Bei Ji Qian Jin Yao Fang, c.652), in which the formula Wu bu wan was used for “Kidney qi deficiency, frequent forgetting (xi wang), dissociation (huang hu),” and a host of other symptoms, as well as for promoting longevity and slowing aging. However, this formula did not gain popularity. The same book listed Kai xin san, which included four of the most common ingredients (chang pu, yuan zhi, ren shen, fu ling) and was recommended for preventing memory loss in old age.
Descriptions of the nature of the memory loss tended to be brief, but some citations provided details suggestive of the profound memory deficits characteristic of AD. In Records of Pattern Discrimination (Bian Zheng Lu, c.1687), the formula Shen jiao tang was for jian wang with the following characteristics: “when the person talks to someone, they soon forget what they said; when someone speaks to the person, they cannot remember what they were told; it is as if nothing had been said.” However, it made no mention of age.
There was a marked change in the list of frequent formulas when the criteria for aging were added (Tables 2 and 4). Although several high-frequency formulas, such as Gui pi tang, were still included, it was at low frequencies. Instead, the most frequently cited formula was Shou xing wan, which is named after a god of longevity called Shou xing. This first appeared in the official formula manual of the Song dynasty, Prescriptions from the Great Peace Imperial Grace Pharmacy (Tai Ping Hui Min He Ji Ju Fang, c.1107) for jian wang and many symptoms atypical of dementia, so this citation was excluded. In later books, however, the applications of Shou xing wan were expanded. In Compendium of Materia Medica (Ben Cao Gang Mu, c.1578), it was used for “phlegm clouding the orifices (which means that the person's consciousness is dull or obscured), dissociation, and jian wang plus disordered speech,” which are all symptoms consistent with AD. In Wondrous Lantern for Peering into the Origin and Development of Miscellaneous Diseases (Za Bing Yuan Liu Xi Zhu, c.1773), Shou xing wan was combined with the formula Dao tan tang to treat “phlegm clouding the orifices, speaks as if demented, and memory loss.”
Although the lists of frequent formulas were different at level 1 and level 2, the individual ingredients of the formulas showed no marked differences (Tables 3 and 5). This indicates that while the formula names varied considerably, formulas for memory impairment tended to include similar core ingredients.
Comparison with previous studies
The previous ZYFJDCD study used the index terms jian wang and chi dai for searching. These identified 127 formula citations for memory impairment and 19 formulas for memory impairment plus aging, 41 which are considerably fewer than in the present study. Of the 127 formulas for memory impairment, 98 (77%) were also found in the ZHYD search, including the most frequent formulas: Gui pi tang, Tian wang bu xin dan, and Ding zhi wan. For memory impairment plus aging, 14 (74%) of the formulas in the ZYFJDCD search were also on the ZHYD list, including Shou xing wan and modified Gu ben wan, but not Xiao dan or Ju sheng wan. The present study also found that Shou xing wan had the most impact among these formulas because it appeared in 12 books for memory impairment and aging whereas some other formulas, such as Qiang ji tang, appeared once and were not cited by later authors.
The ingredient frequency list at level 7 in May et al. (2012) 41 was similar to the level 2 list in Table 5. Although ZYFJDCD included 176 TMs, compared with 970 TMs in the present study, the lists are strikingly similar, with the five most frequent TMs being the same on both lists. There were many differences in the relative rankings of some TMs. For example, rhizome of Arisaema species (tian nan xing) had a relatively higher frequency in the present study, whereas rhizome of Atractylodes macrocephala Koidz. (bai zhu) had a relatively lower frequency. These differences were mainly due to the high frequencies of the formulas Shou xing wan (which contains Arisaema species) and Gui pi tang (which contains Atractylodes) in the ZHYD data.
In the study of ZHYD materia medica entries, 36 the same search terms were used. Four of the top five TMs in the present study were cited for both memory impairment and aging in the materia medica (Table 5). In general, the ingredients in the formulas for memory impairment and age were also listed in the classical materia medica books as used for memory impairment and disorders of aging, but not all the TMs listed in these books for these two indications appeared frequently in the formulas. For example, Dimocarpus longan (Lour.) Steud. fruit (long yan rou), and Gentiana scabra Bge. root (long dan cao) were frequently mentioned in the materia medica but were infrequent inclusions in the formulas. Conversely, some of the TMs that were frequent inclusions in the formulas, such as Zingiber officinale (Willd.) Rosc. rhizome (jiang), were not listed in premodern materia medica books as useful for memory disorders. This reflects a design feature of Chinese herbal formulas in which certain ingredients target the principal disorder, while others aim at secondary symptoms or play functional roles. Therefore, the presence of a TM on the frequency lists in Tables 3 and 5 does not necessarily mean it was intended for memory impairment.
Comparison with contemporary practice
In a recent clinical practice guideline for senile dementia, seven different formulas were suggested on the basis of syndrome differentiation, as is typical of contemporary Chinese medicine books 38 (Table 6). This syndrome differentiation approach was not evident in the classical books, which tended to provide a selection of formulas, each for different combinations of symptoms. However, in some citations the syndrome could be inferred from the stated actions of the formula, the cause, or the symptoms and signs.
“Sea of marrow” refers to the brain. “Orifices” refers to the sense organs of the head. The guideline is Tian (2012). 38
Based on level 1.
Of the seven formulas in the guideline, four were included in the ZHYD results: Gui pi tang, Qi fu yin, Huan shao dan, and Xi xin tang (Table 6). The other three formulas specified in the guideline were not identified for memory impairment in the ZHYD search results, but related formulas were. For example, the blood stagnation–removing formula Tong qiao huo xue tang was in the guideline, whereas other formulas with similar actions, such as Tao ren cheng qi tang and Di dang tang, were in ZHYD. In modern books, Shou xing wan modified to include most of the high-frequency TMs and appears occasionally for memory impairment, 58 but this formula was not mentioned in the guidelines. 38 When the TMs in Table 3 were compared with ZYDCD, six of the top 10 were listed for memory in this contemporary materia medica book. These two comparisons suggest continuity in the use of some formulas and many of the TMs. In general, TMs used historically for memory impairment associated with aging are still used in formulas for senile dementia and AD.
Limitations of the study
The sample of Chinese literature used for this search was large but not totally comprehensive; thus, it is possible that addition of more premodern books could change the results. However, on the basis of the comparison with ZYFJDCD, the identity of the high-frequency TMs and formulas at level 1 are unlikely to be affected. Because chi dai and related terms were relatively few, these results do not show how chi dai was treated in the premodern period. However, the citations for chi dai seldom mentioned memory or aging, and most were unlikely to have been AD.
The use of the criteria for aging made the results more consistent with AD, but aging was infrequently mentioned overall. As a result, this approach greatly limited the size of the included data. Hence, it is debateable whether level 2 was more specific for AD than level 1. There is also the question of whether the age-related memory impairments cited in the historical literature were due to AD pathology. Although citations that included references to cerebrovascular factors, tremor, and seizures were excluded, it remains probable that such symptoms were present but unreported in at least some of the included citations. In the absence of autopsy evidence, whether AD was present is uncertain; thus, it is possible that there was no AD in premodern China and all citations referred to other abnormalities. What the analyses did show was that the TMs used most frequently in formulas for memory impairment were also used frequently for memory impairment in older people or with the aim of slowing the effects of aging. In addition, these TMs remain in use for AD and other dementias.
Another challenge in the interpretation of these data is whether the TMs referred to in the premodern literature were derived from the same species as considered official in contemporary books. In the case of the TM chang pu, this may be sourced from several Acorus species, and the source species for other TMs have varied over time and according to region. This issue needs to be considered when candidates for clinical or experimental research are selected. 36
Experimental studies on the five common TMs
Historical use alone does not provide evidence that any of the TMs were effective for memory impairment or AD, but experimental studies of the five most common TMs and/or their constituents suggest benefits are plausible. P. ginseng has improved cognitive outcomes in some clinical trials. 59 In rats with experimentally induced memory impairment, the ginsenosides Rg1 and Rg5 improved memory and task performance 60,61 and the ginsenosides Rg1, Rg3, and Rg5 have all demonstrated downregulatory effects on β-amyloid. 60 –62 The compound catalpol from R. glutinosa has shown neuroprotective effects in cells treated with β-amyloid, 63,64 and it may be able to increase the synthesis of neurotransmitter acetylcholine. 65
An extract of P. tenuifolia was reported to protect neurones from damage by β-amyloid, 66 and the compound tenuigenin inhibited secretion of β-amyloid. 67 In cognitively impaired mice, a triterpenoid saponin from Polygala roots improved learning and memory. 68 A water extract of P. cocos was reported to protect cells via suppression of oxidative stress and apoptosis induced by β-amyloid. 69 An extract of Acorus tatarinowii protected cells from β-amyloid–induced neurotoxicity. 70 In rats injected with β-amyloid, oral administration of the compound β-asarone ameliorated β-amyloid–induced cognitive impairment; 71 in a mouse model, β-asarone decreased β-amyloid deposition and downregulated β-amyloid levels in the cortex and hippocampus. 72 The combination of Polygala, Acorus, and Poria species showed antioxidant functions and improved learning in memory-impaired mice. 73 These and other TMs identified in these analyses may provide fertile fields for research into interventions for the management of memory impairments and AD.
Conclusion
This study identified several citations in the classical literature that described conditions characterized by profound memory impairment, but whether these were due to AD pathology or other causes remains uncertain. What is evident is that memory impairment associated with aging was a recognized condition. The formulas recommended for this condition varied considerably in name, but certain ingredients, including P. cocos, P. tenuifolia, R. glutinosa, P. ginseng, and Acorus species, appeared in multiple formulas and in multiple books for memory impairment in the context of aging. This same group of TMs also appeared in a systematic search of the largest printed compendium of Chinese formulas. 41 These five TMs remain in clinical use for memory impairment, 37,38 and experimental studies suggest biological activities relevant to AD. 59,68 –70,72,74
Footnotes
Acknowledgments
This project was funded by the China-Australia International Research Centre for Chinese Medicine and partially supported by an International Research Grant from the Guangdong Provincial Academy of Chinese Medical Sciences, Guangdong Province, China. The authors thank the Preventive Health National Flagship Program, Commonwealth Scientific and Industrial Research Organization for providing partial funding support in the early stages of this project. They also thank Dr. Angela Yang, Dr. Claire Zhang, Dr. Takako Tomoda, and Michael Owens for their help during the project.
Author Disclosure Statement
No competing financial interests exist.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
