Abstract
Objectives:
Assess whether an ancient text on herbal medicine accurately characterizes a class of herbal diuretics.
Design:
The Greek text of Dioscorides De materia medica was assessed for herbs stated to have diuretic activity, and then modern research was sought to determine how accurate the ancient assessment of these herbs was.
Results:
Of the 105 plants cited as having diuretic activity by Dioscorides, 56 (53.3%) genuses are confirmed as being diuretic in animal or human research. For another 38 (36.2%) genuses, no research related to diuresis could be identified. Six (5.7%) genuses had mixed results in modern research, whereas a mere 5 (4.8%) genuses were shown to not have diuretic activity. Considering the 67 genuses that were investigated, 56 (83.6%) were confirmed.
Conclusion:
This analysis confirms that Dioscorides was accurate in determining the diuretic nature of herbs, raising the possibility that he was right about other therapeutic suggestions concerning herbs he made. For the remaining herbs that have not been assessed for diuretic effect, it is not yet known if Dioscorides was accurate. Our findings suggest that the 38 herbs Dioscorides categorized as diuretics that have not been studied for diuretic function are candidates for research in this regard.
Introduction
Ancient texts on the therapeutic uses of plants and other natural substances have been increasingly consulted in recent years. Although scientists often search to identify leads for drug discovery, they also suspect that the information in these texts might be still valid for present-day medicine. 1 –3 Naturopathic doctors, herbalists, and other practitioners of natural medicine today base their practice in part on information in ancient texts. It is believed that the traditional uses of natural medicines are more likely to lead to renewed applications today than a random screening approach.
A fundamental question raised by this approach is the degree of exactness of the ancient knowledge of the action of medicinal plants. It is often stated that data in ancient treatises are not based on careful observation and do not result from repeated experiments the data of which have been accurately recorded and approached in an almost statistical way. On the contrary, information in ancient texts is often considered as randomly obtained, and in any case not based on solid evidence.
To offer a contribution to the analysis of this difficult question, in this essay we measure the degree of exactness of the information contained in the most important treatise on materia medica produced in Antiquity, which remained in use until late in the 19th century: De materia medica by the Greek Dioscorides (first century C.E.). Given the sheer quantity of data contained in the work (over 1,000 substances analyzed in detail), in this essay—the first proposing this approach—we focus on a specific property: diuresis. Additionally, many of the herbs Dioscorides describes as diuretic are used clinically this way today by naturopathic and practitioners of traditional medicine.
Such an approach raises the question of the method to be used to quantitatively evaluate the exactness of the data in the treatise. To do so, we have adopted an intuitive method, the validity of which we constantly verified while doing the research: we took Dioscorides' statements on diuresis at face value and we compared them with the data offered by contemporary publications reporting preclinical and clinical results. The vastness of contemporary literature to be potentially consulted required to impose limits to our research. We chose to consult standard databases of available medical literature that can be accessed through the Internet, selecting items offering evidence-based data.
On this basis, we checked the ancient information against current knowledge, something that made it possible to quantify in some way the degree of exactitude of Dioscorides' knowledge of diuretics.
Being a treatise of pharmacognosy and not a prescription manual, Dioscorides' De materia medica did not always specify the parts of the plants to be used, the quantity, or the way to administer them. However, in some cases, he provided this type of information. In 14 cases, the part to be used was the root, in 11 cases the fruit, in 5 the leaves, in 4 the seed, in 3 the whole plant, and in 3 others a decoction of it. As for the way of administering the plants, in 16 cases it was with a liquid which, in 9 cases, was wine. In two cases, the plant had to be taken as a foodstuff, which, in one case, could be either raw or boiled. There were no precise data about quantity, except in two cases in which of one the quantity of wine was specified (two drachms) and in the other, the length of the treatment (many days).
A further methodological issue consisted in correctly identifying the plants mentioned in Dioscorides' treatise. To do so, we apply the methods previously proposed by one of us to identify the plants mentioned in ancient botanico-medical literature and we submit ancient data to critical analysis. 4 –6
In this research, a philologist and historian of medicine who has frequented ancient medical literature for decades (Alain Touwaide) worked in collaboration with a naturopathic physician actively utilizing herbal medicines with patients (Eric Yarnell). We analyzed Dioscorides' text together, not only to propose plausible identifications of the plants, but also to properly understand the information on their diuretic uses and, further on, to check the validity of such information against contemporary literature. Through this crossdisciplinary collaboration and interaction, we do hope to have gained—and to promote—a synergistic understanding of the topic at hand.
In the present study—possibly the first of similar studies to be done in the future—we focus on the most important work compiled in Classical Antiquity on materia medica and the therapeutic uses of natural resources that has been transmitted to us, De materia medica by the first-century C.E. Greek author Dioscorides. We address the problems highlighted above by consulting the Greek text of De materia medica, identifying plants as methodically as possible, carefully assessing technical terms and their meaning in historical context, and by choosing an area of study (diuretic herbs), in which the lexicon does not seem to have been substantially transformed over time.
The ancient work traditionally identified as De materia medica by Dioscorides (1st century C.E.) was a major authority in Mediterranean drug therapy until the 19th century. It collected what was considered in its time definitive information about the plant, animal, and mineral substances, together with some manufactured products, used for the treatment of human diseases. 7,8 There is almost no information about Dioscorides. 9,10 He collected a vast amount of information and organized it by clinical relevance as opposed to alphabetically by name as was common practice in his time, or any other scheme. The work was continuously transmitted in Greek until the 16th century and even later, and also translated from the original Greek into Latin and Arabic in the Middle Ages. Only recently did an English translation become available. 11
In this study, we specifically assess Dioscorides' treatment of diuretic plants, and we determine if his assignment of this quality to plants is validated in modern scientific literature. This will help contemporary practitioners to understand the extent to which Dioscorides' work was scientifically valid, and therefore potentially applicable to current practice.
Materials and Methods
Approaching ancient texts is a complex scholarly and scientific endeavor. Using translations of ancient herbals into modern languages is a potential source of errors, misunderstandings, and anachronisms. Lexicology and exact meaning of single terms changed over time, particularly in science. Exact identification of the plants being discussed according to modern taxonomy is difficult in many cases, as the same names might have been used in ancient times for different species. Furthermore, the context in which statements about plants used medicinally were being made is important, among others because theories about diseases in ancient times often differed compared with the present. 12 We have therefore considered plants only at the genus level.
Based on a reading and thorough analysis of the full text of De materia medica in the Greek original language, a list of 105 medicines identified as diuretic was compiled. 5,13 In the present study, all references to Dioscorides' text are to the standard edition currently used. 14 Each plant and/or derivative is identified by a number made of two elements: the first refers to the book in which the substance appears in De materia medica (from 1 to 5), whereas the second refers to the chapter in any of the five books where the substance is studied in Dioscorides' work (for example, Iris is analyzed in the first chapter of book 1; reference thus is 1.1).
We created a computerized database in which we listed all the plants for which Dioscorides explicitly wrote that they were diuretic (diourêtikos * and the variant ourêtikos in Greek) or “provoked urination” (oura kinei, oura agei or ourêsin kinei). Plants that Dioscorides did not explicitly call diuretics, but mentioned as used for dropsy, dysuria, or similar conditions suggesting that they might have diuretic properties, were not included in this analysis.
Some items required special treatment. The plant akoron (Dioscorides, 1.2) ( = Iris pseudoacorus L.), and the wine made from it (Dioscorides, 5.63), were treated as one medicine. Although Pastinaca sativa L. is believed to be covered by two entries (sisaron, Dioscorides, 2.113, and elafoboskon, Dioscorides, 3.69), they are treated as separate items here, since Dioscorides handled them so distinctly.
Identification of plants was based on all available information: 15 the description of the plants in the Greek text of De materia medica and the ancient botanical literature analyzed in collaboration with Smithsonian Botany Curator Dan H. Nicholson and Smithsonian Scientific Program Specialist Emanuela Appetiti. The plant representations in the Greek articles of Dioscorides' text (over 30 articles from the early 6th century C.E. to the 16th), which had been previously searched for in all articles and databased by one of us (Alain Touwaide), were carefully analyzed to identify diagnostic elements and further compared with Dioscorides' text to check for correspondence between text and illustrations. Similarly, ancient mosaics and mural frescoes in Roman houses dating as far back as the first century C.E. (particularly of Pompeii and Ostia) were systematically searched to locate plant representations and check whether such representations did reproduce botanical illustrations from ancient books or were artistic creations by their authors.
We compared all this material with dry specimens from the United States National Herbarium preserved at the Botany Department of the National Museum of Natural History at the Smithsonian Institution in Washington, DC, to assess the exactness of the diagnostic elements described in Dioscorides' text and in related plant illustrations. Also, using the database that one of us (Alain Touwaide) had created, we checked all the identifications previously made in scientific literature after Linnaeus, for the plants mentioned in the Greek and Latin medical treatises of classical antiquity, and also in any other ancient literary work. 16,17 As a result, we propose the best plausible identification for each plant. Their binomial designations and the name of their author(s) were confirmed by searching The Plant List 1.1, compiled in collaboration by the Royal Botanic Gardens at Kew and the Missouri Botanic Garden. Although species-level identification was attempted, we only considered plants at the genus level for our analysis. 16,17
Whether a plant has been documented as diuretic or aquaretic by modern scientific research was assessed by searching the databases Google Scholar, PubMed, Embase, ProQuest, and ScienceDirect. An herb that increased urine volume by a statistically significant margin compared with controls, regardless of the mechanism of action, was considered diuretic. In searching these databases, we introduced the Latin Linnaean identifications of the plants (genus and species), and their common English names in combination with the keyword diuretic. No limit was placed on date or language. Because all databases included English translations of the title, abstract, and keywords, it was unlikely that non-English articles about diuretics were missed.
To evaluate the accuracy of Dioscorides' knowledge of diuretics we focused on publications reporting clinical trials presenting primary data. We thus excluded secondary literature and review articles, with one exception: an English-language summary report of several clinical trials published only in Belarusian, Russian, and Ukrainian regarding Levisticum officinale W.D.J. Koch ( = ligustikon, 3.51). 18 The original trials discussed in this report could not be accessed.
Ethnobotanical reports required more careful consideration as they may contain the description of present-day practices resulting in some cases from the continuity of ancient knowledge. However useful and informative they might be to document possible traditional practice, they have not been taken into consideration in this study as we included in our research only modern scientific research studies to reduce risk of bias in assessments of whether or not plants are objectively diuretic.
Our focus on the databases above does not exclude that other relevant literature could be identified in a broader search. This might modify our conclusions on the unavailability of research for several of the plants identified as diuretics in Dioscorides' treatise. However, for feasibility reasons, we limited our search to the most commonly used databases with substantial coverage of contemporary literature, which are commonly believed among the medico-scientific community to offer the best sampling of currently available research.
The data obtained as described above were treated as follows.
Confirmed cases
Any genus for which there was at least one primary animal and/or human study that definitively showed a diuretic or aquaretic effect for the botanical genus under study, was considered fully confirmed.
One case required a particular treatment: while several positive reports confirmed the use of the root of petroselinon, 3.66 ( = Petroselinum spp.), one publication provided negative data about the leaf. This plant was regarded in this study as a confirmed diuretic since at least one of its parts is clearly active.
If there were two or more positive reports of diuretic activity based on human clinical trials yet negative animal studies, the plant was still considered confirmed (as in the case with akalêfê, 4.93 = Urtica dioica L. and ippouris 4.46 = Equisetum spp).
Mixed documentation
If there were any other type of mixed positive and negative reports on whether a plant genus was diuretic, it was considered to have mixed documentation.
Undocumented cases
Any Dioscorides' plant genus for which no animal or human study regarding its diuretic activity could be located in the databases consulted in this study with the search criteria mentioned above, was considered undocumented.
Discounted plants
Any plant genus for which the only primary animal or human study regarding its diuretic activity found it is not diuretic or aquaretic, was considered discounted.
Regardless of their validation as diuretics, plants with significant toxicity concerns based on primary animal and human research are denoted (Table 4).
Results
The list of diuretic herbs and research about them is summarized in Table 1. The list is sorted by categories as above and then, within each category, by Linnaean binomial designations listed in alphabetical order.
Dioscorides' Diuretic Plants and Modern Research
Table 2 shows the tabulation of the categories, and in effect, the degree to which Dioscorides was correct in his assignation of diuretic properties to these plants. Of the 105 species considered, 56 genuses are confirmed in some way (53.3%), and another 39 (37.1%) have not been subjected to any published research on their diuretic activity. Only 5 (4.8%) of the plant genuses have been shown to not have diuretic effects, whereas 6 (5.7%) had shown mixed effects in research.
Tabulation of Categories of Herbs Labeled Diuretics by Dioscorides
For definitions of categories, see the methods section. Percentages may not add to 100% due to rounding errors.
Of the plants which have been subjected to research on whether they are diuretic (67 items), a total of 56 are confirmed diuretics (83.6%) and 5 (7.5%) were discounted as diuretics. The remaining herbs (6, 9.0%) showed mixed results.
Out of those diuretic plants listed by Dioscorides that have been the object of modern research on their diuretic effects, only a handful have been assessed in clinical trials (Table 3). Conservatively, a total of nine of Dioscorides' diuretic plant genuses have been assessed in this fashion. Of the nine genuses identified by Dioscorides as diuretic that have been studied in human trials regarding this property, 8 of 9 (88.9%) supported that they are indeed diuretic.
Human Trials Regarding Diuretics in De materia medica
Discussion and Conclusions
Based on the results presented above, it appears that Dioscorides' rate of accurately considering plants as diuretic, when looking at those that have been researched for this property in the modern scientific literature compiled in the database that we have consulted, is very high (83.6% of those herbs that have been studied). Such accuracy rates allow us to argue that Dioscorides may have accurately observed the diuretic action of plants in practice, instead of having gained knowledge of their action through random luck.
In some cases, we had to make judgment calls on whether to consider herbs confirmed. In the case of elenion, 1.28 ( = Inula helenium L.), there is no clear research on whether it is a diuretic. One study was located concerning a compound in a related species. It is about the sesquiterpene lactone inulicin, isolated from the related species I. japonica Thunb., which is defined as diuretic. 113 Given the remoteness of this report from the whole plant under study, it is not considered relevant. Consequently, this plant genus is categorized as having no modern research on its diuretic nature.
Three additional genuses were assessed in clinical trials for diuretic activity, but none was included in our analysis. Levisticum officinale WDJ Koch (lovage), was confirmed as a diuretic but only as part of a formula that also contains leaves of Centaurium erythraea Rafn. (European centaury) and of Rosmarinus officinalis L. (rosemary). Although lovage makes up a major portion of this formula (∼33%), European centaury and rosemary are themselves diuretic, making it difficult to determine how much lovage contributed to the clinical trial results. 114 One human clinical trial was located that showed a decrease in size of kidney stones in people treated with Smilax kêraia, 2.146 ( = Vigna unguiculata (L.) Walp.). However, a definitive assessment of a diuretic mechanism of action was not shown. 115 Kassia 1.13 (Cinnamomum cassia Blume) was an effective diuretic as part of the formula wu ling san, but its individual contribution to the outcome could not be determined. 116 It was confirmed to be active on its own, however.
Several of the plants in Dioscorides list are presently known to raise potential toxicity concerns (Table 4). Some of them may not be appropriate for routine internal use based on these concerns, although, as will be shown below, most of the exact species believed to be involved have not been shown to cause a problem, whereas they do so only in overdose or with topical exposure. The only one of the plants for which Dioscorides mentioned as problematic is skilla, 2.171 ( = Drimia maritima (L.) Stearn). According to Dioscorides, it is abortifacient and requires caution in the dosage in patients suffering from “internal wounds” (possibly referring to peptic ulcers).
Toxicity Concerns with Dioscorides' Diuretic Plants
The Apiaceae is the family most represented in Dioscorides' diuretic list with 18 species (Table 5). Most members of this family contain furanocoumarins, which can cause photosensitivity. 128 –130 One analysis of the levels of furanocoumarins in petroselinon, 3.66 ( = Petroselinum spp.) found that they were almost never high enough to cause photosensitivity in humans after oral consumption, but were high enough to cause contact photodermatitis in susceptible individuals. 131 In fact, many of the cited cases involve contact photosensitivity and not problems after oral ingestion. However, this problem is still possible after oral ingestion of members of the Apiaceae family.
Members of Apiaceae on Dioscorides' Diuretic Plants List
It should be noted that a member of the Rutaceae family, pêganon, 3.45 ( = Ruta graveolens L.) is also reportedly photosensitizing, including after oral ingestion by humans. 117,132
The asaron, 1.10 ( = Asarum europaeum L.) contains aristolochic acid, a well-established cause of nephrotoxicity and urothelial cancer. 133 –135 Reports of toxicity in humans from the specific species A. europaeum were not encountered. Caution is warranted with intake of this plant until more information is available; it appears particularly ill advised to use it for longer than a few days.
The ampelos leukê, 4.182 ( = Bryonia cretica L., bryony) contains ribosome-inactivating lectins and proteins. 136 In clinical reports cited in textbooks, bryony is considered fairly toxic, with the potential to cause severe gastroenteritis and even death in overdose. 137
Various species of Drimia and Scilla (skilla, 2.171, and uakinthos, 4.62) contain cardiac glycosides that, in overdose, can cause arrhythmias and can even be fatal. 138,139
The lithospermon, 3.141 ( = Lithospermum officinale L.) may contain hepato- and nephrotoxic unsaturated pyrrolizidine alkaloids that are found in other members of the genus. 140,141 It can suppress thyroid-stimulating hormone, gonadotropins, and prolactin secretion, which is useful therapeutically in some patients with hyperthyroidism and hyperprolactinemia, but which make it contraindicated in hypothyroidism, pregnancy, and lactation. 142,143
Various Teucrium spp. have been shown to contain neoclerodane diterpenoids. Some of these species, although not T. scordium L. (skordion, 3.111) specifically, have apparently caused various forms of hepatotoxicity in humans. 144,145
This research is limited by several factors. The work has been made on the assumption that the plants mentioned in Dioscorides, De materia medica, can be identified with some degree of plausibility according to present-day taxonomy. This is a sine-qua-non condition for this kind of research because incorrect identifications will not make it possible to locate relevant publications on whether a species is a diuretic. Also, misidentifications could lead to misattributing studies to the wrong genus. A great deal of scholarly study has been devoted to identifying which plants Dioscorides was describing and analyzing in De materia medica to reduce this problem as much as possible. Nevertheless, the identifications we have used in this essay still need to be considered as plausible hypotheses.
The method by which we categorized plants as confirmed based on animal research only could be questioned for supporting their efficacy in humans. It is unknown in general how well animal models predict efficacy of diuretic plants in humans. Five of the seven plants (all but amugdalê 1.123 = Prunus dulcis (Mill.) D. A. Webb, and druos balanoi 1.106 = Quercus spp.) confirmed to be diuretic in human trials reviewed in this study had prior animal studies supporting their diuretic nature. However, the one plant not confirmed in a human clinical trial to be diuretic, Petroselinum, also had animal research suggesting it is diuretic. This still seems to indicate that animal research is more likely than not to specifically predict whether an herb will be diuretic in humans. Clearly more research is necessary to answer this question.
Although the human clinical trials conducted on the plants said by Dioscorides to be diuretic are of small size and although not all meet the highest standards of methodological rigor, they allow for a preliminary measure of degree of the accuracy of Dioscorides' statements about which herbs are diuretic. Of the clinical trials reported in the literature compiled in the databases under consideration, 8 (88.9%) support that these plants are diuretic. Absence of clinical trials on most of Dioscorides' plants does not equate to proof of their lack of efficacy.
Another limitation is the lack of discussion of the plant part that is diuretic in Dioscorides, and therefore a difficulty of being certain that modern research was using the same material that he recommended.
Our findings suggest that the 38 herbs Dioscorides categorized as diuretics that have not been studied in any capacity as diuretics are excellent candidates for research in this area. Further research, particularly human trials, is also warranted on those herbs that are already confirmed solely by animal research.
Footnotes
Acknowledgments
E.Y. is pleased to acknowledge the School of Naturopathic Medicine at Bastyr University for the financial support it generously provided for this research. A.T. is thankful to the late Dan H. Nicholson, Smithsonian Botany Curator, who devoted time and energy to review together the identification of plants. Also, he thanks Emanuela Appetiti, then a Scientific Program Specialist at the Smithsonian, for the assistance provided during this research.
Author Disclosure Statement
E.Y. is president and part owner of Heron Botanicals. A.T. has no competing financial interests.
