‘The body is not a one-note melody, but a symphony of many interactive components functioning synergistically … the active ingredient model does not stem from a strength of the scientific method, as often supposed; rather, it stems from a weakness – from the inability of the reductionist method to deal with complex systems.’²³

Synergy is an important hypothesis in herbal pharmacology in the context of the advantage of chemical complexity. It applies if the action of a chemical mixture is greater than the arithmetical sum of the actions of the mixture’s components: the whole is greater than the sum of the individual parts. In other words, there is a cooperative or facilitating effect between the components for a specific outcome. Another helpful definition of synergy is an effect seen by a combination of substances that is greater than would have been expected from a consideration of individual contributions.²⁴ A well-known example of synergy is exploited in the use of insecticidal pyrethrins. A synergist known as piperonyl butoxide, which has little insecticidal activity of its own, interferes with the insect’s ability to break down the pyrethrins, thereby substantially increasing their toxicity. This example emphasises one important possible mechanism behind synergy as applied to medicinal plant components: increased or prolonged levels of key components at the active site. In other words, components of plants that are not active themselves can act to improve the stability, solubility, bioavailability or half-life of the active components. Hence a particular chemical might in pure form have only a fraction of the pharmacological activity that it has in its plant matrix. This key mechanism for synergy, therefore, has a pharmacokinetic basis.

Methods for assessing pharmacological synergy were proposed in an extensive paper by Berenbaum in 1989.²⁵ Later Williamson²⁶ and Houghton²⁷ elaborated on these in the context of medicinal plants. Essentially, there are four mathematical techniques for demonstrating synergy proposed and discussed in these works, but only the isobole method is truly rigorous, although it is not always undertaken because of its complexity. Hence, several of the examples of herbal synergy discussed below need to be viewed in the context that they are only possible examples of synergy, since the isobole method was not performed.

The isobole method has the advantage that it is independent of the mechanism of action involved. Although some explanations of the isobole method can be complex, a simple way to describe it follows. For a series of two-component mixtures, the concentration or dose of component A in each mixture required to give a defined effect (such as an ED₅₀ or minimum inhibitory concentration, MIC) is plotted on the X axis against the concentration or dose of component B in the same mixture on the Y axis. The line joining all the points is then constructed. If there is no interaction between the effects of A and B, and their combination is merely additive, the line will be straight (Fig. 2.1). If there is a synergy between A and B, the line (the isobologram) will be concave. Alternatively, if antagonism exists the isobologram will be convex.

Figure 2.1 The isobologram method for demonstrating synergy.

There are some published examples in the literature of the isobole method demonstrating synergy for mixtures of phytochemicals from the same herb. Williamson provides the example of ginkgolides A and B from ginkgo in terms of an anti-PAF (platelet activating factor) effect on platelets in vitro.²⁶ In much earlier work in mice, a potentiating effect of different concentrations sennoside C on the purgative activity of sennoside A was observed.²⁸ In this study an isobologram of ED₅₀ values for laxative activity was not plotted, but one can be readily constructed from the data provided showing a concave line. Interestingly, a mixture of these compounds in the ratio 3:7 (which somewhat reflects the relative levels in senna leaf) gave the lowest ED₅₀ value.

In several published examples, only one mixture of two given phytochemicals from the same plant is tested, and hence an isobologram construct from just three points (A, B and A plus B) could be plotted. There is little point in doing this, however, since synergy is readily apparent from such a simple experimental design (if the activity of A plus B is greater than the expected activity calculated from the individual activities of A and B). Although this is a less robust means of demonstrating synergy than an isobologram constructed from four or more points, if there is a dramatic increase in activity from the mixture of A and B it is reasonable to assume that synergy has been demonstrated.

One example of where the combination of two components from the same plant has yielded a significantly greater effect than expected is a study on the upregulation of phase II detoxification enzymes in mice by the glucosinolate breakdown products from broccoli.²⁹ An orally administered mixture of crambene and indole-3-carbinol caused a significantly greater induction of glutathione S-transferase and quinone reductase that could be expected from the activity of oral doses of the individual treatments. Another example is the study where the three major curcuminoids from turmeric (Curcuma longa) were investigated for their in vitro nematocidal activity against second stage larvae of Toxocara canis.³⁰ Each of the three curcuminoids was ineffective on its own, whereas any combination of two, or all three together, was active. The combination of all three demonstrated the highest activity.

Another example identified in early research is the antibacterial activity of major components of lemongrass essential oil. While geranial and neral individually elicited antibacterial action, the third main component, myrcene, did not show any activity in vitro. However, when mixed with either of the other two main components, myrcene enhanced their activity.³¹ As pointed out in one review¹¹ this example is technically not one of synergy, but rather exemplifies potentiation, where an inactive compound enhances the potency of a bioactive one.

In an interesting study, the impact on serum lipids of four active components of Chinese hawthorn (Crataegus pinnatifida), and their combination, was investigated in mice fed a high fat and cholesterol diet.³² The four compounds (quercetin, hyperoside, rutin and chlorogenic acid), and their combination at percentages reflected in the berry, were each fed orally to the mice at a dose of 2.85 mg/kg. The greatest reduction in total cholesterol and LDL-cholesterol was found for the combination. For LDL-cholesterol, only the combination yielded a significant reduction versus the control group (p<0.01).

The above examples therefore demonstrate two types of synergistic interactions for phytochemicals within a plant. The first example is where the phytochemicals are all active in the test method used. The second example is where one or more of the phytochemicals are inactive, yet their inclusion in a mixture increases the observed effect.

Pharmacodynamic synergy has also been demonstrated for combinations of extracts or phytochemicals from different herbs. The in vitro antitumour activity for combinations of extracts of Corydalis and turmeric rhizomes (a traditional formulation) demonstrated a synergistic interplay, as determined using an isobologram.³³ Berberine from Coptis chinensis and evodiamine from Evodia rutaecarpa demonstrated synergistic antitumour activity in vitro against a human hepatocellular carcinoma cell line.³⁴ A pronounced convex curve was produced on an isobologram.

A recent review highlighted that two key developments are behind growing interest in synergy research in phytotherapy.³⁵ The first is the new methods of analytical chemistry and molecular biology that have become available in the 21st century. The second is an unexpected change of paradigm in chemotherapy that has appeared without drawing much attention as to its radical nature. This is the transition in chemotherapy towards a multi-drug approach. Multi-drug therapy is already widely practised in the treatment of AIDS and other infectious diseases, hypertension, cancer and many other diseases.

In particular, the review notes this multi-drug concept in cancer chemotherapy is often aimed at being biomodulatory, not just at direct destruction of the tumour.³⁵ Instead, suppression of different processes essential for the tumour’s survival are also targeted (such as angiogenesis, oncogene expression, anti-apoptotic mechanisms, immunological tolerance and inflammation). This concept of multi-targeted or polyvalent therapy is not new to phytotherapy and is discussed further below.

There are several examples in the herbal research literature where synergistic interactions are strongly implied, although they have not been absolutely demonstrated using the isobologram method. One example is the research cited by Williamson²⁶ and several others where cannabis extract demonstrated higher antispastic activity in mice than the equivalent dose of pure tetrahydrocannabinol (THC). As pointed out by Houghton,²⁷ this might not represent an example of true synergy. Rather other components in the cannabis extract might also have antispastic activity. In other words, there could be an additive effect between THC and other phytochemical components in the extract. However, the increased activity is so striking that synergy is a very plausible explanation. Testing a THC-free cannabis extract in the same model would have provided clear proof, one way or another.

Another example of a striking result that strongly implies a synergistic interaction is provided by Williamson.²⁶ This is the research where the antiulcer activity of a fraction of ginger extract was found to be 66 times that calculated from the individual components of that fraction. More recent similar examples include:

There are several examples of a pharmacokinetic basis for synergy, where chemical complexity results in enhanced solubility or bioavailability of key phytochemical components in plant extracts. The basic issues were discussed by Eder and Mehnert.³⁸ The isoflavone glycoside daidzin given in the crude extract of Pueraria lobata achieves a much greater concentration in plasma than an equivalent dose of pure daidzin.³⁹ Ascorbic acid in a citrus extract was more bioavailable than ascorbic acid alone.⁴⁰ More recently, consumption of a preparation of fresh kiwi fruit resulted in up to five times more effective ascorbate delivery to tissues than when ascorbate was administered via the drinking water to vitamin C-deficient mice.⁴¹ The kava lactones appear to increase each other’s bioavailability, especially when given as the herbal extract. (For more details, see the kava monograph under Pharmacokinetics.)

Improved pharmacokinetic parameters can also apply to combinations of herbs, although this is not necessarily always the case. The traditional Chinese formulation Huangqin-Tang decoction, together with decoctions of its individual herbal components, were studied in rats after oral dosing.⁴² The constituents/metabolites baicalin, wogonoside, oroxylin-A-glucuronide, risidulin I and liquiritin demonstrated higher overall bioavailability from the compound formulation than from the relevant single herb decoctions, due largely to a longer residence time.

Often synergy is invoked to explain experimental results that can be readily explained by additive effects. Synergy is not always necessary to justify the use of complex mixtures, be they single herbal extracts or herb combinations. Additive effects can be just as favourable to enhanced therapeutic activity and are probably more common in phytotherapy. A team of German scientists headed by the late Dr Hilke Winterhoff discovered that the procyanidins in St John’s wort extract (Hypericum perforatum) significantly increased the in vivo antidepressant activity of hypericin and pseudohypericin, probably by a solubilising effect (these compounds otherwise have poor solubility).⁴³ The same research team found that the flavonoids in St John’s wort have activity in an animal model of depression, with hyperoside and isoquercitrin showing significant activity.⁴⁴ The flavonoid perspective was backed up by another in vivo study from Italy that found a flavonoid-enriched St John’s wort extract exhibited a more significant influence on central neurotransmitters.⁴⁵ Later work from the German researchers demonstrated that a St John’s wort extract free of both hypericin and hyperforin (another phytochemical in the herb thought to play a role in the antidepressant activity) still exerted an antidepressant activity in rodent models.⁴⁶ Hence, it seems that hypericin, procyanidins (indirectly), hyperforin and the flavonoids can all contribute to the clinically verified antidepressant activity of St John’s wort.

Additive effects also apply to herb combinations. The reduction of amphetamine-induced hypermotility was studied in mice after a single oral dose of passionflower extract (Passiflora incarnata, 250 mg/kg), kava extract (100 mg/kg) or the two combined (250 mg/kg+100 mg/kg, respectively).⁴⁷ The combination yielded a much greater reduction in hypermotility over 2 hours than each individual herbal extract, which the authors attributed to a synergistic interaction. However, a careful analysis of the numbers suggests that the observed result is more likely to have been due to an additive effect (in terms of the differences observed from the control group). Additive benefits have also been observed at a clinical level. A combination of ginseng (Panax ginseng) and Ginkgo has been extensively investigated in volunteers for its acute and long-term impact on cognitive function. The effect of the combination appears to be significantly greater than either herb alone. This might be an example of synergy (which would be difficult to prove conclusively) but is at least an additive effect. (See the ginseng monograph for more details.)

As touched on above, the concept of multi-agent medicines is a developing theme in modern drug therapy. However, it represents nothing new for phytotherapy. Due to their chemical complexity, even a single herbal extract is a nature-designed multi-agent medicine that can simultaneously target a range of desirable pharmacological effects. There are many examples of this provided in the herbal monographs in Part Three. This helps to explain why identifying the ‘active constituent’ in many herbal extracts has proved to be so difficult. For most if not all herbal extracts the ‘active constituent’ is the whole extract itself, as illustrated by the above discussion of the antidepressant activity of St John’s wort. The potential for chemical complexity to confer polyvalent activity or polypharmacology can also explain the apparent therapeutic versatility of herbal extracts. As an example there is the rather large and novel range of clinical uses for Ginkgo, all somewhat supported by evidence from randomised, controlled trials.

It has also been argued that polyvalence may be behind some effects that have been attributed to synergy (although ultimately such a distinction is probably academic, as both explanations argue in favour of chemical complexity).²⁷ In whole organism or tissue studies, other compounds present in the mixture may be active against a range of targets, all acting to contribute to the observed outcome. Houghton goes on further to assert that:

As outlined by Houghton, polyvalence can occur at three key levels:

The third possibility overlaps with the concept of pharmacokinetic synergy underlining, as noted above, that the distinction between polyvalence and synergy might be quite arbitrary at times. There are many examples of these phenomena in the pharmacodynamic and pharmacokinetic sections of the monographs in Part Three of this book.

In a recent review, Gertsch observed that herbal extracts might be ‘intelligent mixtures’ of secondary plant metabolites that have been shaped by evolutionary pressures.⁴⁸ As such they could represent complex therapeutic mixtures possessing inherent synergy and polyvalence. Gertsch also notes that another important concept related to polyvalence is that of network pharmacology, as originally proposed by Hopkins.⁴⁹ In the context of plant extracts (which for commonly used herbs typically contain hundreds of potentially bioactive natural products with only mild activity) it is possible that different proteins within the same signalling network are only weakly targeted. However, this is sufficient to shut down or activate a given process by network pharmacology.⁴⁸ In other words, network pharmacology can explain how a number of weakly active plant secondary metabolites in an extract may be sufficient to exert a potent pharmacological effect without the presence of a highly bioactive compound.⁴⁸ In the context of herbal pharmacology, Paul Ehrlich’s concept of the magic bullet is supplanted by one of a ‘green shotgun’, to paraphrase Gertsch and James Duke. However, the herbal research that can verify such theoretical constructs is only in its infancy.

The ‘-omic’ technologies (genomics, transcriptomics, proteomics and metabolomics) are recent developments in molecular biology that have the potential to open new perspectives for our understanding of the multi-faceted activities of complex plant mixtures. However, their meaningful application to medicinal plant research poses several challenges.⁵⁰ The reviews of Wagner and colleagues and Efferth and Koch¹¹ provide a more complete discussion of the application of these techniques to natural product research.

Whether it is a manifestation of synergy or by other means, the hypothesis that herbal extracts might act at a pharmacological level as intelligent mixtures is intriguing. One such example could be the root of Pelargonium sidoides. Its phytochemical content is unremarkable, including oligomeric procyanidins (OPCs) and highly oxygenated simple coumarins.⁵¹ Yet a liquid preparation of the herb has been shown to be effective in several clinical trials. A meta-analysis published in 2008 reviewed its efficacy for acute bronchitis.⁵² Six randomised trials were included, providing results for a total of 1647 patients (treatment and control groups): one trial compared Pelargonium against a non-antibiotic treatment (N-acetylcysteine); the other five trials tested Pelargonium against placebo. In two trials the ages of the participants were 6 to 12 years, and 6 to 18 years. Adults were enrolled in the other trials. Key findings were:^52,53

Moreover, the herb had previously developed a traditional reputation as a treatment for tuberculosis.

Another example of an intelligent mixture could be willow bark. Its phytochemical content cannot be reconciled against its impressive clinical results (see the willow bark monograph for more details).

Essential oils (from the word ‘essence’) are mixtures of fragrant compounds that can be isolated from plants by the process of steam distillation. In this procedure, steam is driven through the plant material and then condensed, with the subsequent oil and water phases separating out. Since they are volatile in steam, and usually have pronounced aromas, essential oils are often referred to as volatile oils. However, this term is not accurate since they have boiling points well above 100°C. Often the oil is slightly modified by the steam distillation process, so that it does not exactly reflect what is found in the plant. In some cases, such as chamazulene produced from matricine in German chamomile, steam distillation actually produces substantial quantities of a new chemical.

Other processes are also used to produce essential oils and these include solvent extract using hexane or liquid or supercritical carbon dioxide, enfleurage (oil extraction of delicate essential oils in flowers) and expression (used to produce orange and lemon oils from the peel). Essential oils are important items of commerce, being used for perfumes, food flavourings and personal care and pharmaceutical products. They also comprise the medicines of the therapeutic system known as aromatherapy.

Essential oils are water-insoluble oily liquids that are usually colourless. Despite their being called oils, they are not related chemically to lipid oils (fixed oils) such as olive oil, corn oil and so on. Although often hydrocarbon in nature, they are unrelated to the hydrocarbon oils from the petrochemical industry. They will slowly evaporate if left in an open container, and placing a drop of oil on blotting paper can be used as a simple technique to test for adulteration with a fixed oil. If a fixed oil is present, an oily smear will remain on the paper a few days later.

Adulteration is an important issue in the trading of essential oils and many sophisticated techniques have been developed to imitate and extend essential oils. In some cases, such as oil of wintergreen, trade in the synthetic oil has completely supplanted the natural product. One survey found a large variability between the biological activities of different samples of oils and groups of oils under the same general name, for example lavender, eucalyptus or chamomile.¹⁰⁸ This reflected on the blending, rectification and adulteration that occurs with commercial oils. Of course, this issue does not apply for oils prescribed as part of the whole plant extract, as used by phytotherapists.

The synthesis and accumulation of essential oils are generally associated with the presence of specialised structures in the plant often located on or near the surface; for example, the delicate glandular trichomes (hairs) of the mint family (Lamiaceae or Labiatae). Essential oil composition can vary quite dramatically within a species and often distinct chemotypes are recognised. This means that the same plant species can produce quite different oils in terms of their chemistry, pharmacology and toxicology. From a biosynthetic perspective, components of essential oils can be classified into two major groups: the terpenoids and the phenylpropanoids. Any given essential oil might contain more than 100 of these components.

When the chemical structures of terpenoids are examined, they can be theoretically constructed from five-carbon isoprene units. Plants produce terpenoids using the mevalonic acid biosynthetic pathway. As mentioned earlier, the building block is actually isopentenylpyrophosphate, not isoprene, but the outcome is molecules based on the characteristic five-carbon units. Naming of terpenoids is based on multiples of 10 carbons (two isoprene units). Hence, molecules with 10 carbons are called monoterpenes; those with 15 are called sesquiterpenes (sesqui means one and a half) and so on. Only mono- and sesquiterpenes are found in essential oils. Higher terpenoids are too large and hence not volatile in steam. Diterpenes occur in resins as resin acids and triterpenes are found as saponins. Not even all mono- and sesquiterpenes are volatile in steam – the iridoids featured in the eyebright monograph are one example. Examples of monoterpenes in essential oils include limonene, geraniol, borneol and thujone. Bisabolol is a sesquiterpene.

Phenylpropanoids are far less common as components of essential oils. Their basic chemical skeleton is a three-carbon chain attached to a benzene ring. They are formed by the shikimic acid biosynthetic pathway and examples include anethole and eugenol.

Essential oil components can also be classified according to their functional groups. The most common compounds found in essential oils are hydrocarbons, alcohols, aldehydes, ketones, phenols, oxides and esters. These functional groups play a large part in determining the pharmacology and toxicology of the essential oil component; for example, ketones are more active and toxic than alcohols, and alcohols and phenols are more potent as antimicrobial agents, with phenols being more irritant. Essential oil components often exhibit optical isomerism (where the two isomers are mirror images of each other). For example, (+)-carvone isolated from caraway oil has a caraway-like odour and (−)-carvone isolated from spearmint oil has a spearmint-like odour.

Aromatherapy is a treatment system based on the use of essential oils. The oils may be inhaled, applied to the skin or orifices, added to baths or ingested. There is no doubt that ingested oils or those applied to the skin or added to baths are absorbed into the bloodstream in significant quantities.¹⁰⁹ However, the use of essential oils by inhalation, mainly to influence mental function, is more controversial. In fact, in the German-speaking world, aromatherapy is typically defined as the therapeutic use of fragrances only by means of inhalation. Evidence is accumulating that this form of aromatherapy also has a pharmacological basis and is not placebo, nor is it a manipulation of emotions via the sense of smell. A recent review of the therapeutic properties of lavender essential oil exemplifies this, with 14 clinical studies demonstrating possible beneficial effects from its inhalation, including enhanced sleep, decreased anxiety and improved cognition.¹¹⁰

Given the great chemical diversity of essential oils, it is not surprising that they exhibit a wide variety of pharmacological activities. However, some common themes do emerge, notably antimicrobial (including antiviral) and spasmolytic actions.

Most of the evidence for the antibacterial and antifungal activities of essential oils comes from in vitro tests, although case reports and clinical trials are scattered throughout the literature (see below). A review of the published in vitro work between 1976 and 1986 found that results were difficult to compare.¹¹¹ Test methods used differed widely and important factors influencing results were frequently neglected. One of these factors was the composition of the essential oil being tested (given the existence of chemotypes and adulteration). This was highlighted in another study of commercial essential oils that found a wide variation in the antimicrobial activities of commercial samples of thyme oil, eucalyptus oil and geranium oil, among others.¹⁰⁸

A review of more recent investigations summarised the antibacterial and antifungal activity data from around 50 studies.¹¹² Most tests assessed either MICs or the concentration required to kill 50% of the test micro-organism culture (EC₅₀). Many of the reviewed publications tested a range of oils and essential oil components. Wide variations in activity, depending on the test method, the essential oil tested and the test organism were tabulated.

Of 53 essential oils tested against four organisms, only a few oils exhibited remarkable activity, particularly thyme and origanum.¹¹³Pseudomonas aeruginosa was the least susceptible organism and Candida albicans the most susceptible. These oils stand out because they contain the highly active phenols thymol and carvacrol.¹¹⁴

Attempts have been made to identify the mechanisms behind the antifungal and antibacterial activities of essential oils and the key components responsible for this activity. Antimicrobial activity was said to parallel cytotoxic activity, suggesting a common mode of action, most probably exerted by membrane-associated reactions.¹¹⁵ In simple terms, it appears that the mobile and lipophilic nature of essential oil components, especially the monoterpenes, enables them to penetrate and disrupt cell membranes. Concentrations of tea tree oil that inhibit or decrease growth of Escherichia coli also inhibit glucose-dependent respiration and stimulate the leakage of intracellular potassium.¹¹⁶ According to a recent review, essential oils seem to have no specific cellular targets because of their great number of constituents.¹¹² This suggests a very low risk of the development of microbial resistance against essential oils. Being lipophilic (fat-soluble) they pass through the cell wall and membranes, disrupting their structures. In bacteria, permeabilisation of the membranes is associated with the loss of ions and a reduction of membrane potential, collapse of the proton pump and depletion of the ATP pool. Essential oils can also coagulate the cytoplasm and damage lipids and proteins. Damage to the cell wall and membrane can lead to the leakage of macromolecules and eventually to lysis.

In eukaryotic cells, essential oils can provoke depolarisation of mitochondrial membranes by decreasing the membrane potential, impact ionic calcium cycling and other ionic channels and reducing the pH gradient, affecting the proton pump and the ATP pool. Chain reactions from the cell wall or membrane invade the whole cell, leading to widespread oxidative damage.¹¹²

Of five components tested for antibacterial activity, cinnamic aldehyde was the most active, followed by citral, geraniol, eugenol and menthol.¹¹⁷ Essential oils with high concentrations of thymol and carvacrol usually inhibit Gram-positive bacteria better than Gram-negative bacteria,¹¹⁸ although they still possess a good broad-spectrum activity. In another study, linalool was the most active antibacterial agent and citral and geraniol were the most effective antifungal agents.¹¹⁹ Essential oils with a high monoterpene hydrocarbon level were very active against bacteria but not against fungi, with the exception of dill.¹⁰⁸ There was a negative correlation observed between cineole content and antifungal activity. In the case of tea tree oil, terpinen-4-ol was identified as the most important antimicrobial compound^120,121 and cineole detracted from its antifungal activity.¹²¹

Recent research has highlighted the significant in vitro antiviral activity of essential oils.¹²² Most of the studies have been conducted against herpes simplex virus (HSV)-1 and HSV-2, with quite potent activities (at the parts-per-million level) being observed. A viral envelope is necessary, as growth of non-enveloped (naked) viruses is not affected by essential oils. They appear to act on enveloped viruses by affecting the viability of the free virus (virion), probably by interfering with the viral envelope (which like the cell membrane is lipid in nature).¹²²

Clinical studies of the antimicrobial activity of essential oils have been published, with a focus on Australian tea tree oil used only topically. In an early double blind trial, 10% tea tree oil cream reduced symptoms of tinea pedis but was not effective in eradicating the fungus.¹²³ However, a 5% tea tree oil gel was as effective as a 5% benzoyl peroxide lotion in the treatment of acne and patients experienced few side effects.¹²⁴

A 2006 review of tea tree oil included published clinical trials.¹²⁵ In addition to the trials summarised above, tea tree oil demonstrated efficacy as an antibacterial mouthwash or gel in three trials (with a residual effect on oral bacteria), reduced methicillin-resistant Staphylococcus aureus (MRSA) carriage, and successfully treated fungal infections of the nails, mouth and skin. Since then, topical tea tree oil has demonstrated clinical efficacy in another randomised, double blind trial assessing its value in acne patients¹²⁶ and improved healing in recurrent herpes labialis after application of a 6% gel (although the results did not achieve statistical significance in this small trial).¹²⁷ Tea tree oil as a topical application is also quite effective for head lice infestation.^128,129

One review noted that other essential oils show bactericidal activity against oral and dental pathogenic micro-organisms and have been incorporated into mouth rinses.¹¹⁸

Information about topical use is important, but phytotherapists also rely on essential oils to achieve antimicrobial effects within the body. In particular, the excretion of essential oils via the lungs or urine might be expected to exert at least a mild antimicrobial activity at these sites. This is the basis of the use of juniper and buchu for urinary tract infections and one of the reasons for the use of thyme in respiratory infections. However, essential oil components are excreted into the urine in metabolised forms, mainly as glucuronide and sulphate conjugates. Hence, any antibacterial activity may not be reflected in the urine. Studies are necessary to understand further this traditional use.

Essential oils certainly have a promising role in the management of intestinal pathogens, bacterial or otherwise. In an open label trial, oil of oregano was orally administered to 14 adult patients whose stools tested positive for the enteric parasites Blastocystis hominis, Entamoeba hartmanni and Endolimax nana.¹³⁰ After 6 weeks of 600 mg/day of emulsified oil, there was a substantial reduction in detectable pathogens that correlated somewhat with an improvement in gastrointestinal symptoms. The clinical anthelmintic action of the (toxic) essential oil of wormseed (Chenopodium ambrosioides) is well documented, although this has been questioned.¹³¹

The spasmolytic activity of essential oils has been observed many times on isolated smooth muscle preparations and forms much of the basis of their use in functional gastrointestinal disorders.¹³² The effects of essential oils from 22 plants and some of their constituents on tracheal and ileal smooth muscle were investigated in one study.¹³³ All of the oils had relaxant effects on tracheal smooth muscle, the most potent being angelica root, clove and elecampane root. Sixteen oils inhibited the phasic contractions of the ileal muscle preparation, the most potent being elecampane root, clove, thyme and lemon balm. Two oils (anise and fennel) increased the phasic contractions. Spasmolytic activity has also been confirmed clinically. For example, peppermint oil added to barium sulphate suspension significantly relieved colonic muscle spasm (p<0.001) during barium enema examination in a double blind, placebo-controlled study involving 141 patients.¹³⁴ For more examples of the clinical spasmolytic activity of peppermint oil and its value in irritable bowel syndrome, see the peppermint monograph.

Carminatives relax sphincters and assist in the expulsion of intestinal gas. Their activity is somewhat related to spasmolytic activity. Certain essential oils or essential oil-containing herbs have been traditionally used as carminatives over many years. Oils of peppermint, sage and rosemary all relaxed Oddi’s sphincter, but peppermint was the most active.¹³⁵ The carminative activity of cardamom and dill was confirmed in human studies.¹³⁶ However, they were also shown to cause oesophageal reflux and should be used cautiously in susceptible patients.

Some essential oils are traditionally regarded as diuretics because they act as ‘kidney irritants’. The infusion and essential oil of juniper berries as well as terpinen-4-ol were tested for diuresis response in rats.¹³⁷ On initial dosing, all three test substances exhibited an antidiuretic effect, but a significant diuretic effect was established on repeated doses, with the infusion having the strongest effect. The ‘irritant’ effect of juniper oil on the kidneys was also investigated, since there are concerns in the literature about its long-term use. No nephrotoxic effects were observed in an animal model and the authors suggested that provided high-quality oil is used (distilled from the ripe berries), concerns about the kidney irritant effects of juniper are unfounded.¹³⁸ Essential oil terpenes used orally had shown some promise in dissolving small kidney stones in small, uncontrolled trials in adult patients, but these findings were not supported by controlled trials.¹³⁹ However, a positive pilot trial in children suggests further investigations are desirable.

Certain essential oils (or the herbs that contain them) are used as expectorants. A proprietary product containing myrtle oil was popularly prescribed by doctors in Germany as an expectorant and mucolytic agent for acute and chronic bronchitis and sinusitis. The oil contains limonene, cineole and alpha-pinene. An expectorant activity for this oil was confirmed in a clinical trial in patients with chronic obstructive airways disease.¹⁴⁰ Earlier animal experiments by Boyd suggested an expectorant activity for essential oils probably via influencing goblet cells to secrete more respiratory tract fluid (RTF) and mucus.¹⁴¹ In the 1940s, Boyd studied the effects of several essential oils in various experimental models.^142,143 The most pronounced increase of RTF was seen after ingestion of oil of anise. Interestingly ingestion of oil of eucalyptus had a moderate effect that was not eliminated by cutting efferent gastric nerves. This finding supports the premise that essential oils do not act as reflex expectorants (see later).

One intriguing aspect of Boyd’s research is that he was able to demonstrate that inhaled essential oils also acted as expectorants. These results are summarised in Table 2.1.

Table 2.1 Expectorant effect of inhaled essential oils or their components

RTF = respiratory tract fluid

Boyd found a biphasic effect in that high concentrations of essential oils suppressed RTF and mucin release, whereas lower doses (on the threshold of detection by the nose) had a pronounced stimulatory effect. A seasonal effect was also observed, with an increased activity in autumn (fall). These findings provide a rational basis for the hot lemon drink in respiratory infections, since active amounts of lemon oil would be inhaled while ingesting the hot drink.

The inhalation of essential oils into the lungs might exert a direct antimicrobial activity. A case report from Australia describes the successful use of inhaled eucalyptus oil to effect clinical improvement in a patient with tuberculosis.¹⁴⁴ A 28-year-old woman presented with cough, shortness of breath, fever, night sweats, weight loss and malaise. She had been unwell for 12 months. A sputum sample was positive for Mycobacterium tuberculosis. X-ray of the chest confirmed the diagnosis of pulmonary tuberculosis.

The patient initially refused conventional treatment and instead inhalation of a Eucalyptus globulus essential oil preparation (in an ethanol-water base) was initiated. A vapour inhaler was used to administer 3 mL of the preparation in 500 mL of boiling water 3 times a day for 3 weeks. After 10 days the patient reported reduced malaise, normal temperatures, improved appetite, absent cough and a gain of 4.5 kg in weight. Her sputum cultures were negative. The patient subsequently underwent conventional treatment.

A sedative activity following the ingestion of essential oils such as lavender is commonly recognised and was supported by early animal studies.¹⁴⁵ More recently, a proprietary essential oil of lavender at 80 mg/day was found to be clinically effective in sub-threshold anxiety disorder. Three randomised, controlled, double blind clinical trials were identified in a review, demonstrating significant improvements in anxiety scores after 10 weeks of treatment with the lavender oil.¹⁴⁶

Stimulant activity has also been attributed to some essential oils. Inhalation and oral doses of rosemary oil increased locomotor activity in mice.¹⁴⁷ Infusions of essential oil-containing herbs are often taken as diaphoretics, especially during acute respiratory infections. In this context, it is interesting to note that a case report describes a patient who exhibited pronounced diaphoresis attributed to up to 10 cups a day of sassafras tea (not recommended because of potential carcinogenicity).¹⁴⁸

Essential oils and essential oil components have been extensively studied in animal models for analgesic-like activity.¹⁴⁹ Overall, 43 bioactive essential oil components were identified in one review, mainly monoterpenes. However, these findings are yet to translate into clinical applications, apart from topical use of peppermint oil or menthol (for example) in headache and joint pain (see the peppermint monograph). In contrast, the local anaesthetic activity of oil of cloves is well described and the local anaesthetic activity of lavender oil has been demonstrated in an in vivo model.¹⁵⁰

Given their cytotoxic properties on bacterial and fungal cells, it is not surprising that essential oils have exhibited a wide range of antitumour activities against cancer cell lines.^112,122 Essential oils exert cytotoxicity at levels considerably lower than their antibacterial activity (in the parts-per-million range). As in bacterial cells, the cell membrane is one of the sites of cytotoxic activity.¹²² This antitumour research is particularly well developed (at least at the in vitro level) for thymoquinone, a component of the essential oil from the seeds of the black cumin (Nigella sativa).¹⁵¹

For more information about the pharmacology of essential oils and associated anti-inflammatory, antiulcer, spasmolytic, oestrogenic, expectorant, antimicrobial and analgesic activities, see the monographs on chamomile, fennel, peppermint and thyme in Part Three.

Essential oils are highly concentrated from their original plant matrix and as a result can present specific safety issues when used as such. Their toxicology has already been comprehensively reviewed.¹⁵² Nevertheless, a few issues are worth noting here. Certain toxic essential oils, notably pennyroyal, tansy and parsley, have been used as abortifacients. However, as early as 1913 it was found that these oils have absolutely no specific or direct stimulating action on the uterine muscle.¹⁵³ In fact, consistent with the known pharmacodynamics of essential oils, they were found to inhibit uterine contractions. Their abortifacient action was concluded to be due to general poisoning or gastrointestinal irritation, which makes their use not only uncertain, but also extremely dangerous.

Thujone is a constituent of commonly used herbs such as wormwood, yarrow, thuja and sage. This compound is neurotoxic and its presence in liqueurs such as absinthe apparently caused widespread toxicity and abuse syndromes in the early 20th century.¹⁵⁴ The first sign of toxicity is a headache. High and prolonged doses of the above herbs are hence best avoided, unless they are low-thujone varieties. It has been suggested that thujone intoxication may have played a part in Van Gogh’s style of painting.¹⁵⁵ However, this has been recently challenged and it now appears that the syndrome known as absinthism is largely attributed to the toxic effects of ethanol.¹⁵⁶ Safrole, a major component of sassafras oil, is carcinogenic and its use should be avoided.

Topical use of essential oils can cause contact dermatitis, as one review of tea tree oil cases noted.¹⁵⁷ Oxidation, and hence the age of the oil, appears to increase the likelihood of a reaction.

Glucosinolates are sulphur- and nitrogen-containing glycosides responsible for the pungent properties of horseradish, nasturtium and mustard. The glucosinolate itself is not pungent. When it encounters the enzyme myrosinase, normally stored in another compartment of the cell, the aglycone is formed which then usually rearranges into a pungent and corrosive isothiocyanate. Isothiocyanates can also be formed from glucosinolates by steam distillation and so are called mustard oils. For this reason, they are sometimes classified as essential oils, but from a phytochemical perspective this is inappropriate. The structure of a typical glucosinolate is provided in the figure below. They are ionic in nature and occur in the plant as potassium salts.

Glucosinolates are also found in brassicas such as cabbage, broccoli and Brussels sprouts. As such, they are frequently consumed as a normal part of human diet.

In traditional herbal and folk medicine, strong skin irritants and inflammatory substances were empirically used to combat inflammatory processes in tissues and organs remote from the site where the irritant was applied. This is the principle of counterirritation. This poorly understood effect is recognised in pharmacology and can often provide misleading anti-inflammatory effects in test models for substances with no pharmacological activity other than being irritants. The mode of action of such skin irritants is characterised by an ability to influence deeper regions of the body, probably by reflex responses mediated by the nervous system. The stinging of arthritic joints with nettles and the subsequent reduction in pain and inflammation is one such example. Hyperaemic medicines can be used in the form of ointments, compresses, liniments or plasters. The mustard compress or plaster is still used in Europe today, particularly for bronchial infections and detoxification in chronic diseases. Mustard oil is highly corrosive and if applied for too long will cause blistering and may even permanently scar the skin.

The main component of nasturtium oil (Tropaeolum majus) is benzyl isothiocyanate. This has potent antibacterial and antifungal activity. In Europe, enteric-coated capsules of the oil are used to treat bronchial and urinary tract infections.¹⁵⁸ Horseradish preparations are used in the treatment of bronchial and sinus conditions. Presumably, the sulphur compounds confer a mucolytic effect. Recent research interest in benzyl isothiocyanate relates to its potential role in cancer.¹⁵⁹

Isothiocyanates and some of their products (such as goitrin) are goitrogenic and interfere with the function of the thyroid gland.¹⁶⁰ Goitrin inhibits iodine incorporation and the formation of thyroxin and its effects cannot be countered by iodine administration. Although this is a potentially toxic effect, it could possibly be exploited in cases of patients with hyperthyroidism.

The bulk of recent research attention on glucosinolates and their various transformation products (including sulforaphane, indole-3-carbinol and di-indoylmethane) has focused on their potential to prevent and modify cancer. This phenomenon has been known for more than 30 years. The anticarcinogenicity of these compounds, specifically in relation to brassica vegetables, has been recently reviewed for in vitro assays, animal experiments and various human studies.^161–163 Alterations in phase I, and particularly phase II, detoxification enzymes are suggested as mechanisms by which these plant constituents might inhibit chemical carcinogenesis or alter the level of cancer promoting hormones such as oestrogen. On the other hand, possible mutagenicity, tumour promotion and carcinogenicity are also discussed, indicating that caution should be exercised if recommending long-term intake of these compounds at doses well in excess of optimum dietary levels.

The following brief review of the data, especially for broccoli sprouts, is provided as a good example of the progress in this field of research. Epidemiological evidence suggests that diets containing vegetables from the Brassica genus (including broccoli, cabbage, Brussels sprouts and cauliflower) are associated with a reduced cancer risk. After reviewing the results of 7 cohort studies and 87 case-control studies, it was concluded that the association appears to be most consistent for lung, stomach, colon and rectal cancer.¹⁶⁴ A series of studies to find the active constituents and mechanisms of action began in the mid-1990s. As a result of this research, Brassica vegetables were found to be rich sources of inducers of phase II detoxification enzymes in vitro, with the isothiocyanate sulforaphane the main phase II inducer in broccoli.¹⁶⁵

As noted above, glucosinolates in Brassica vegetables are broken down to form isothiocyanates (the aglycones) by the action of the enzyme myrosinase. This occurs when the plant is crushed, but also during cooking or in the human intestine by intestinal flora.¹⁶⁶ Glucoraphanin is the glucosinolate precursor of sulforaphane. Broccoli sprouts contain about 15 times more glucoraphanin than the mature broccoli plant.¹⁶⁷

Sulforaphane was found to be a potent in vitro inducer of the phase II enzymes quinone reductase (QR) and glutathione S-transferase (GST).¹⁶⁸ Sulforaphane also induces phase II enzymes in vivo. In mice treated orally with sulforaphane, QR and GST were increased in many organs, including the liver.¹⁶⁸ It is a potent inducer of the multiorgan protector pathway and hence initiates the cytoplasmic transcription factor Nrf2 to migrate into the cell nucleus, where it activates the antioxidant response element, upregulating the production of antioxidant and phase II enzymes.

The induction of phase II enzymes is one of the main mechanisms by which sulforaphane exerts its chemopreventative activity. Orally administered sulforaphane has been shown to be protective against carcinogen-induced cancer at a variety of sites.¹⁶⁹ Taken together these findings suggest that sulforaphane can inhibit the development of cancer during the initiation and post-initiation stages.¹⁶⁹

The chemopreventative activity of sulforaphane probably involves many mechanisms. These include:¹⁶⁹

A randomised, placebo-controlled trial found that sulforaphane exerted a phase II inducing effect in participants who achieved a good bioavailability (that is, who were able to convert glucoraphanin to sulforaphane in their intestine). Two hundred healthy volunteers from a region in China with a high risk for development of hepatocellular carcinoma were enrolled. Urinary levels of the markers of two environmental carcinogens were assessed. A strong, highly significant inverse association was observed when the marker levels were compared with levels of dithiocarbamate excretion (a measure of the metabolism of glucoraphanin).¹⁷⁰ Oral administration of a broccoli sprout homogenate over 3 days induced mucosal phase II enzyme expression in the upper airway of human volunteers.¹⁷¹

Flavonoids are extremely common and widespread in the plant kingdom. They function as plant pigments and are responsible for the colours of many flowers and fruits. Being abundant in plants, flavonoids are commonly consumed in the human diet, especially if it is rich in fruits and vegetables.

The word ‘flavonoid’ is derived from the Latin word flavus meaning yellow and many flavonoids are indeed yellow in colour. However, many others are white and the special flavonoid-related anthocyanins are red, blue or purple. (For a discussion of the pharmacological properties of anthocyanins, see the bilberry monograph.) Flavonoids are also present in leaves, where they are said to protect the plant tissue against the damaging effects of ultraviolet radiation.

Flavonoids consist of a single benzene ring joined to a benzo-gamma-pyrone structure. They are formed from three acetate units and a phenylpropane unit (via the shikimic acid pathway). More than 2000 are known with nearly 500 occurring in the free (aglycone) state and the rest as O- or C-glycosides. Flavonoid glycosides are generally water-soluble. There are three main types, classified according to the state of oxygenation at carbon 3. These are flavones, flavonols and flavonones. The properties of isoflavonoids are discussed in the phyto-oestrogen section.

Unfortunately, the term flavonoid is used rather loosely in the pharmacological literature. It is often applied as a collective term to describe any plant polyphenolic compounds, including anthocyanins, green tea polyphenolics, the flavonolignans in St Mary’s (milk) thistle and oligomeric procyanidins and catechins, by way of example. None of these phytochemicals is actually flavonoids, and their bundling with flavonoids only serves to confuse any meaningful conclusions that can be drawn about the pharmacological properties of true flavonoids.

Most of the studies conducted on flavonoids have used in vitro models, often employing isolated enzyme systems. The findings of these studies need to be interpreted with caution, since it is uncertain that oral doses of flavonoid glycosides or even their aglycones can reach sufficient concentrations in living organisms to reproduce these effects. This reservation applies even more for oral doses of herbs containing flavonoids, since the flavonoid dose will be commensurately lower in the plant matrix. The issue was well illustrated by one early clinical pharmacology study that found, while quercetin and apigenin inhibited platelet aggregation in vitro, no significant effect was found in human volunteers.¹⁷² Fortunately, clinical studies involving flavonoids (albeit at doses typically much higher than can be achieved by phytotherapy) and flavonoid-containing herbs are on the increase. (For a further discussion on the bioavailability of flavonoids, see the Pharmacokinetics section.) There are now tens of thousands of published studies on flavonoids. Hence, the brief section in this chapter serves only as a primer to this vast body of literature.

The original pharmacological interest in flavonoids arose during vitamin C research in the 1930s. Studies by Hungarian workers indicated that a number of vegetables and fruits (notably citrus) contained substances capable of correcting certain abnormalities associated with scurvy. In particular, this new factor, designated as vitamin P, corrected the capillary fragility associated with ascorbic acid deficiency. Vitamin P was subsequently found to be a mixture of flavonoids. Additional research disputed that vitamin P was essential to maintain human life and the term was dropped in the 1950s. However, research did confirm the therapeutic value of flavonoids for fragile capillaries (actually fragile connective tissue surrounding capillaries) and as extenders of vitamin C activity, possibly through improved absorption and protection from oxidation, and by partially substituting for some of its biological functions.¹⁷³

Decreased capillary fragility denotes improved connective tissue tone and a reduced tendency for capillary contents to leak into surrounding tissue. This implies that flavonoids will prevent oedema associated with inflammation and stasis. Such effects from flavonoids are reasonably well established from clinical trials.

A proprietary mixture of micronised flavonoids consisting of 90% diosmin and 10% hesperidin, usually administered at a dose of 1000 mg/day, has been investigated in several clinical trials. Such double blind, placebo-controlled trials have shown that this flavonoid combination improves venous tone in normal volunteers,¹⁷⁴ enhances microcirculation in patients with venous insufficiency,¹⁷⁵ assists healing of venous ulcers¹⁷⁶ and relieves symptoms of acute haemorrhoids.¹⁷⁷ A recent critical review confirmed the benefits of this combination in patients with advanced venous disease.¹⁷⁸ A meta-analysis of clinical trials also supported its benefit in the treatment of haemorrhoids.¹⁷⁹ Of course, this is more a ‘nutraceutical’ therapy and it is arguable whether these doses have any relevance to phytotherapy, as they could not be realistically achieved by the administration of plant extracts.

Flavonoids are polyphenolic compounds (they contain several phenolic hydroxyl groups), some more so than others. The chemical properties that flow from this feature underlie many of the impressive in vitro pharmacological effects of these compounds. In particular, they are able to complex metal ions, act as antioxidants and bind to proteins such as enzymes and structural proteins (this last feature could also explain the ability of flavonoids to enhance the integrity of connective tissue).

The in vitro antioxidant properties of flavonoids were the focus of much early research.^180,181 The ability of flavonoids to complex pro-oxidant metallic ions such as iron probably augments their antioxidant effects in specific circumstances.¹⁸² Of particular interest is the capacity of flavonoids to inhibit macrophage-mediated oxidation of LDL and thereby attenuate atherogenesis.¹⁸³ The antioxidant properties of flavonoids could also contribute to their observed anti-inflammatory and antiplatelet effects,¹⁸⁴ and are related not only to their structural characteristics, but also to their ability to interact with and penetrate the lipid bilayers of the cell membrane.¹⁸⁵ Flavonoids scavenge the nitric oxide radical,¹⁸⁶ the superoxide anion¹⁸⁷ and singlet oxygen.¹⁸⁸ Like most other antioxidants, flavonoids can also act as pro-oxidants in particular circumstances.¹⁸⁹

The molecular promiscuity conferred by the phenolic groups of flavonoids is well demonstrated by the range of enzymes inhibited by them in vitro. Enzyme activities inhibited to varying degrees by flavonoids in vitro include cyclo-oxygenase,¹⁹⁰ lipoxygenase,¹⁹⁰ lens aldose reductase,¹⁹¹ xanthine oxidase,¹⁹² cGMP phosphodiesterase,¹⁹³ cAMP phosphodiesterase,¹⁹⁴ angiotensin-converting enzyme,¹⁹⁵ aromatase,¹⁹⁶ thyroid peroxidase,¹⁹⁷ hyaluronidase,¹⁸⁰ phospholipases¹⁸⁰ and protein phosphokinases.¹⁸⁰ However, more research is needed to determine which of these activities might realistically translate into clinical effects. Since enzyme-inhibiting activity depends on the structure of the flavonoid, some compounds are more likely than others to be clinically active.

Using isolated cells or organs in vitro, the activities demonstrated for flavonoids include antiviral activity, especially for the 3-methoxylated flavones,^181,195 anti-inflammatory activity,¹⁹⁸ antimicrobial activity,^195,199 inhibition of histamine release from mast cells,^200,201 antiplatelet activity,^195,202 tumour cell cytotoxicity, especially for highly methoxylated flavones,¹⁹⁵ and spasmolytic activity.^203,204 The flavonoid 8-isopentenylnaringenin (now 8-prenylnaringenin) isolated from the Thai herb Anaxagorea luzonensis was found to be an oestrogen agonist with an activity about 10 times greater than genistein.²⁰⁵ This flavonoid, which also occurs in hops (Humulus lupulus), has now been extensively studied as a phyto-oestrogen.²⁰⁶ It strongly activates oestrogen receptor alpha and at least partially accounts for the observed clinical value of hops in menopausal women experiencing mild vasomotor symptoms.²⁰⁷

Animal experiments (usually using high doses) have demonstrated a number of interesting pharmacological effects for flavonoids. One study found antioxidant effects in vivo for dietary flavonoids (quercetin and catechin), with reduced lipid peroxidation in rats.²⁰⁸ Flavonoids have also demonstrated hepatoprotective, antiulcer (gastroprotective)²⁰⁹ and analgesic activity in vivo.¹⁸⁰ Several flavonoids tested showed anti-inflammatory effects in acute and chronic animal models.^210,211 A favourable effect of quercitrin on experimentally induced diarrhoea was probably related to anti-inflammatory effects.²¹²

Flavonoids have exhibited preventative activity against chemical carcinogens and tumour promotion in animal models,^213,214 although most of the speculation over their cancer preventative role is based on in vitro studies.²¹⁵ Anxiolytic properties have been demonstrated for some flavonoids (including chrysin and apigenin), which selectively bind with a high affinity to the central benzodiazepine receptor.^216,217

The relationship between dietary flavonoid intake and cardiovascular disease was tested in several early epidemiological studies. The Zutphen Study in Holland found that flavonoid intake in elderly men (largely via tea, onions and apples) was significantly inversely associated with mortality and incidence of stroke.²¹⁸ A Finnish study also found that people with very low intakes of flavonoids had a higher risk of coronary disease,²¹⁹ but US studies found no significant association.²²⁰ These investigations are confounded by the difficulties in isolating effects due to just the flavonoids and the loose use of the term, as discussed above.²²¹ This was well illustrated by another US study that found the observed non-significant inverse associations for broccoli, apples and tea on cardiovascular disease incidence in women were not mediated by flavonoids.²²² The same researchers later detected no protection from dietary flavonoids on incidences of insulin resistance, type 2 diabetes²²³ and cancer,²²⁴ all from epidemiological investigations in women.

In another interesting Finnish study, the total dietary intakes of 10 054 men and women during the year preceding the baseline examination were determined with a dietary history method.²²⁵ Flavonoid intakes were estimated mainly on the basis of the flavonoid concentrations in Finnish foods. Participants with higher quercetin intakes exhibited lower mortality from ischaemic heart disease. The relative risk (RR) between the highest and lowest quartiles was 0.79 (95% CI: 0.63 to 0.99: p=0.02). The incidence of cerebrovascular disease was lower at higher kaempferol (0.70; p=0.003), naringenin (0.79; p=0.06) and hesperetin (0.80; p=0.008) intakes. Men with higher quercetin intakes had a lower lung cancer incidence (0.42; p=0.001), and men with higher myricetin intakes had a lower prostate cancer risk (0.43; p=0.002). Asthma incidence was lower at higher quercetin (0.76; p=0.005), naringenin (0.69; p=0.06) and hesperetin (0.64; p=0.03) intakes. A trend toward a reduction in risk of type 2 diabetes was associated with higher quercetin (0.81; p=0.07) and myricetin (0.79; p=0.07) ingestion.

Clinical studies of flavonoids (more as nutraceuticals than as herbal extracts) are yielding some interesting results. Citrus flavonoids have demonstrated promising results in patients with hypercholesterolaemia²²⁶ and in combination appear to exert significant anti-inflammatory activity in patients with osteoarthritis.²²⁷ Hesperidin, also from citrus, lowered diastolic blood pressure in a randomised, controlled trial.²²⁸ Flavopiridol, a flavonoid-derived drug, is being developed as a treatment for cancer and inflammation.²²⁹

The flavonoid most widely studied as a clinical nutraceutical is quercetin, and doses of around 500 to 1000 mg/day are typical. Generally, the clinical findings are modest, possibly reflecting on its low and variable bioavailability.²³⁰ Favourable effects were observed on endurance in a randomised, controlled trial,²³¹ although other trials have not been positive.^232–234 It did reduce blood pressure and oxidised LDL concentrations in overweight patients with a high cardiovascular disease risk phenotype at only 150 mg/day.²³⁵

For more information about the pharmacology of flavonoids, see the monographs on Astragalus, chamomile, Ginkgo, hawthorn, horsechestnut and licorice. In particular, the monographs on Ginkgo and hawthorn extensively review the cardiovascular effects of flavonoids, an area not emphasised in this chapter.

Flavonoid aglycones, but not their glycosides, are mutagenic in various assay systems. Quercetin is probably the most mutagenic (and most widespread) flavonoid. When glycosides are incubated with beta-glucosidase or bacteria possessing this enzyme, they acquire mutagenic properties.²³⁶ However, the presence of methoxy groups markedly decreases the mutagenicity of the flavonoid.²³⁷ Concerns about the mutagenicity and carcinogenicity of quercetin first arose among Japanese researchers who were searching for the carcinogenic compounds in bracken fern. However, those compounds were later identified to be ptaquilosides.²³⁸ Dietary quercetin also enhanced pretumorous lesions in a model of rat pancreatic carcinogenesis, indicating possible promoting and progressing effects,²³⁹ but lacked tumour-promoting effects in a different model.²⁴⁰ On the other hand flavonoids, including quercetin, have shown antimutagenic effects and are widely considered to be antimutagens.^241,242

With the initial discovery of the mutagenicity of quercetin and concerns about the carcinogenicity of bracken fern, tests were conducted to determine if quercetin was carcinogenic. While two studies were positive in rats, many other studies in rats, mice and golden hamsters failed to demonstrate carcinogenic activity.²⁴³ This is reassuring since quercetin is the most common flavonoid in the human diet. Proposed reasons for the lack of carcinogenicity included poor absorption and rapid microbial degradation, as already discussed.

However, the lack of carcinogenicity of a massive exposure to quercetin at 10% of diet suggests that more active mechanisms might be at work. This was confirmed in a study that demonstrated the rapid metabolic inactivation of quercetin by catechol-O-methyltransferase to form non-mutagenic methoxy groups on the flavonoid.²⁴³ This could be a major reason for the lack of carcinogenicity of other flavonoids in vivo. A more recent critical review of the data related to the safety of quercetin noted a lack of in vivo toxicity (including a lack of carcinogenic properties) and concluded that the weight of the available evidence supports the safety of this flavonoid.²⁴⁴

When tannins encounter mucous membranes, they react with and crosslink proteins in both the mucus and epithelial cells of the mucosa. The mucosa is consequently bound more tightly and rendered less permeable, a process referred to as astringency. If this phenomenon occurs in the mouth, such as when eating an unripe banana, a puckering and drying sensation is experienced. Astringency affords increased protection to the subadjacent layers of the mucosa from micro-organisms and irritant chemicals. It also leads to an antisecretory effect on the mucous membrane.

Since tannins are large polar molecules, they are generally poorly absorbed through the skin or gastrointestinal tract. Hence, the pharmacological effects of tannins can be largely explained in terms of their local effects on these organs (such as astringency) or effects within the gastrointestinal lumen. However, decomposition products of tannins (and the monomers of condensed tannins) can be absorbed and thereby do exert systemic effects (see the Pharmacokinetics section). The poor bioavailability of tannins is fortunate, since they can be quite toxic if absorbed in large amounts. Being smaller molecules, the green tea catechins do show some intact bioavailability in humans. However, their polyphenolic nature ensures that this is still relatively low.

One of the most notable effects of tannins in the gut is their dramatic effect on diarrhoea. It can be proposed that the effect of tannins is to produce a protective (if temporary) layer of coagulated protein on the mucosa along the upper levels of the gut wall, so numbing the sensory nerve endings and reducing provocative stimuli to additional peristaltic activity. Supporting this central astringent activity, tannins will also inhibit the viability of infecting micro-organisms, check fluid hypersecretion and neutralise inflammatory proteins. Because of their affinity for free protein, they will concentrate in damaged areas.

Condensed tannins were able to bind to and inactivate the hypersecretory activity of cholera toxin.²⁴⁶ At present, there are attempts to develop new therapies for cholera, and one approach showing promise is the inhibition of transport proteins involved in cAMP-activated chloride channels. One of these is the cystic fibrosis transmembrane conductance regulator protein (CFTR). (Mutations in CFTR cause cystic fibrosis, hence its name.) A hydrolysable tannin from the Chinese gallnut inhibited CFTR in vitro. Intraluminal placement of this tannin (0.6 mg/kg) in mice reduced cholera-toxin-induced intestinal fluid secretion by 75%, with no effect on intestinal fluid absorption.²⁴⁷

Roots from the genus Potentilla typically contain hydrolysable and condensed tannins.²⁴⁸ In a randomised, double blind, placebo-controlled trial, an extract of tormentil root (Potentilla erecta) effectively treated rotavirus infection in children.²⁴⁹ The duration of diarrhoea in the herbal group was 3 days, compared with 5 days in the control group (p<0.0001).

Tannins in herbs such as meadowsweet were traditionally regarded as beneficial in mild peptic ulceration and inflammation. This application is analogous to the use of the synthetic antiulcer drug known as sulcrafate. Sulcrafate is an astringent aluminium-based compound (aluminium is highly astringent). It is said to bind selectively to the exposed proteins at the ulcer base. The barrier thus created protects the ulcer crater from gastric contents. Antiulcer activity has been demonstrated for black tea extract²⁵⁰ and for condensed tannins.²⁵¹ Ellagic acid suppresses acid secretion²⁵² and pomegranate prevented experimentally induced gastric ulcers.²⁵³

Tannins and pseudotannins can also affect bowel flora composition. For example, tea fed to chicks significantly changed the levels of particular microflora.²⁵⁴ This may explain why rhubarb and other tannin-containing herbs reduced levels of uraemic toxins in rats with renal failure (perhaps at least in part by inhibition of the bowel flora production of some of these compounds).²⁵⁵ Whatever the mechanism, tannin-containing herbs such as rhubarb have been used in China to treat renal failure.

Local use of tannins on bleeding surfaces renders a styptic or haemostatic effect due to localised vasoconstriction and possibly an increased rate of coagulation. Since the presence of tannins is widespread in higher plants, this would explain the folk use of so many different plants as ‘wound herbs’. An aqueous extract of a tannin-containing herb demonstrated haemostatic activity due to vasoconstriction and the formation of an ‘artificial clot’ (presumably resulting from a tannin-protein reaction), that tended to produce a mechanical plug to arrest bleeding from small blood vessels.²⁵⁶ This styptic effect can also be useful for mild internal bleeding.

The topical application of tannins can exert favourable effects on burns, weeping eczema and viral infections. In the early 20th century, tannin sprays were applied to severe burns as preferred therapy. The tannin-protein complex thus formed acted as an artificial semi-permeable membrane known as an eschar.²⁵⁷ This procedure was abandoned because toxic levels of tannic acid (the hydrolysable tannin used) were sometimes absorbed through the damaged skin. Tannic acid also damaged epithelial stem cells and caused excessive scar formation. However, the procedure is still practised in China using condensed tannins, which are less toxic and kinder to the regenerating epidermis.

One of the notable properties of tannins that has emerged in recent research is their antioxidant effects. Given their polyphenolic nature, this is not surprising. Hamamelitannin (from witchhazel) and gallic acid were more active than ascorbic acid in scavenging reactive oxygen species.²⁵⁸ Oral administration of geraniin was found to lower the level of lipid peroxide in the serum and liver of rats suffering liver injury.²⁵⁹ Most of the antioxidant research on tannins and pseudotannins has focused on green tea catechins. As well as impressive in vitro effects, it appears that an antioxidant activity can also be achieved in the human body, presumably brought about by the absorption of decomposition products as well as the catechins (see the Pharmacokinetics section).

A recent review classified the key antioxidant mechanisms of tannins as comprising free radical scavenging, chelation of transition metals and inhibition of pro-oxidative enzymes and lipid peroxidants.²⁶⁰ Conflicting results have been reported for the antioxidant activity of green tea in clinical studies. The results may have been confounded by how the tea was prepared (e.g. brewing time, concentration) and other dietary and lifestyle factors.²⁶¹ A review of trials to 2000 found that antioxidant activity has been demonstrated, although a strong conclusion could not be drawn. In subsequent trials in healthy volunteers, green tea:

There are many other clinical benefits demonstrated for green tea catechins.

Good clinical evidence exists for the role of cranberry juice and extracts in the prevention of urinary tract infections.²⁶⁵ It is believed that consumption of cranberry results in urine that discourages bacterial adherence to the bladder wall. The A-type proanthocyanidins (condensed tannins) are thought to be key here, although they might be too large to be excreted as such in the urine.^266,267 Perhaps their decomposition products are active in urine?

As noted earlier, the pomegranate is a rich source of ellagitannins. Pomegranate extract and juice products standardised for their ellagic acid content are available (presumably the ellagic acid being measured is that contained in the ellagitannins). An initial uncontrolled clinical trial of pomegranate juice in patients with prostate cancer reported a significant prolongation of prostate specific antigen (PSA) doubling time. Anticancer effects in other types of cancers have been observed in vitro and in vivo.^268,269 Pomegranate peel extract also promoted wound healing in rats after topical application, an effect certainly predicted by its tannin content.²⁷⁰

As alluded to previously, the many and impressive in vitro effects of tannins may not have clinical relevance because of their poor bioavailability. This is highlighted by a study that postulated hydrolysable tannins were responsible for the antiprostatic activity of Epilobium species via inhibition of 5-alpha-reductase.²⁷¹ Sufficient quantities of almost any tannin will non-specifically inhibit almost any enzyme. Other effects of tannins unlikely to have clinical relevance include antiviral activity (except topically and in the gut lumen, see above),²⁷² inhibition of elastase,²⁷³ cytotoxic effects,²⁷⁴ reverse transcriptase inhibition,²⁵⁹ antimutagenic activity,²⁵⁹ host-mediated antitumour activity²⁵⁹ and inhibition of lipoxygenase.²⁵⁹

The lower than expected rate of coronary artery disease in France has been termed the French paradox. Several reasons have been proposed for this effect, notably the high consumption of red wine rich in OPCs. Independently of this development, a French scientist named Masquelier researched the antioxidant and connective-tissue-stabilising benefits of OPCs over many years. This led to the development of two OPC products (containing predominantly B-type OPCs and their monomers), one from grapeseeds and the other from the bark of the maritime pine (Pinus pinaster). These products are prescribed for applications similar to flavonoids and early clinical research supported their use for varicose veins, capillary fragility and chronic venous insufficiency.²⁷⁵ An impressive array of outcomes from randomised, controlled trials has now accumulated, demonstrating the remarkable clinical diversity of these OPC-rich extracts. For example, a recent review of pine bark extract noted benefits in cardiovascular disorders, type 2 diabetes, cognition, asthma, osteoarthritis and glaucoma, among others.²⁷⁶ Given the postulated beneficial effects of OPCs in heart disease, it should come as no surprise that hawthorn, the most important herb for the heart in modern phytotherapy, is rich in these substances (for more details on the cardiovascular pharmacology of OPCs, see the hawthorn monograph).

Important tannin-containing herbs not already mentioned include agrimony (Agrimonia eupatorium) and bistort (Polygonum bistorta). There is also a monograph on witchhazel in Part Three.

Tannins are found to some extent in many medicinal plants. The following comments about adverse reactions refer only to relatively high doses of herbs containing significant quantities of tannins. It is unlikely that incidental exposure to tannins at low levels has any significant negative impact on health. In fact, condensed tannins are found in several commonly consumed foods. In contrast, hydrolysable tannins are less common in foods (the pomegranate being one exception) and this suggests that the long-term therapeutic intake of this group of tannins should be avoided.

High doses of tannins lead to excessive astringency on mucous membranes, which produces an irritating effect. This probably led to the practice of adding milk to tea whereby the tannins preferentially bind to proteins in the milk rather than the gut wall. However, even adding milk does not prevent the constipation that can result from chronic intake of high levels of tannins. For these reasons, high doses of strongly astringent herbs should be used cautiously in highly inflamed or ulcerated conditions of the gastrointestinal tract and in patients complaining of constipation.

Chronic intake of tannins inhibits digestive enzymes, especially the membrane-bound enzymes on the small intestinal mucosa.²⁷⁷ Tannins complex metal ions and inhibit their absorption. One study found that as long as tea and iron are consumed separately, iron absorption is not affected.²⁷⁸ This iron-complexing property of tannins could be exploited in male patients with haemochromatosis, which is now recognised to be a relatively common disorder. (See Appendix C for more information on such potential interactions with tannin-containing herbs.) Tannins can also react with thiamine and decrease its absorption.²⁷⁹

Addition of tannic acid, a hydrolysable tannin, to the barium sulphate mixture used in barium enemas increases the yield and accuracy of the examination. The colonic mucosa stands out clearly and tumour visualisation is improved. However, the practice was banned in 1964 by the US FDA (Food and Drug Administration).²⁸⁰ Several deaths caused by acute hepatotoxicity, the majority in children, were attributed to this practice.²⁸¹ In these cases, quantities of tannic acid sufficient to cause massive liver damage were absorbed directly into the bloodstream from the colon. This effect is highly unlikely to follow from use of tannin-containing herbs. Nonetheless, some unexplained cases of herbal hepatotoxicity have been recorded. It is therefore prudent to avoid the use of high doses of highly astringent herbs in patients with very damaged gastrointestinal tracts, other than in the circumstances outlined above. Green tea extract consumption has been linked to rare cases of idiosyncratic hepatotoxicity.²⁸²

Tannins are carcinogenic when injected subcutaneously²⁸³ and herbal teas containing tannins have been implicated in the possible development of oesophageal cancers.^284,285 While these associations probably have little relevance to phytotherapy, they do suggest caution with the long-term oral and topical use (on damaged skin) of tannin-containing herbs.

Capsaicin has been the most commonly studied of the pungent compounds. As mentioned above, the C-fibre sensory neurons, which release inflammatory neuropeptides such as substance P, mediate a wide variety of responses including neurogenic inflammation, thermoregulation and chemical-initiated pain. Capsaicin functions to activate and then, at higher doses and over time, desensitise this class of neurons. This latter response, by a process known as tachyphylaxis, provides the basis for the therapeutic interest in capsaicin. Capsaicin stimulates C-fibres by interacting with TRPV1 receptors.³²⁸ The intense sensation of pain and heat which is experienced after eating a hot curry is testimony to this C-fibre activation. But, as experienced curry eaters will testify, they can tolerate hotter and hotter food over time due to tachyphylaxis.

Although the pain and burning from consumption of cayenne or capsaicin can be disturbing, no actual harm results from its consumption. In effect, the specific action on vanilloid receptors creates an illusion of pain and burning. Tissue damage is not concurrent with these sensations. This contrasts strongly with the mustard oils, which are highly corrosive and produce sensations of pain and burning in association with actual tissue damage. On the other hand, capsaicin is a pronounced irritant, as evidenced by the incapacitating effect of capsicum spray. Sometimes ingestion of cayenne does seem to produce lingering sensations of discomfort, and this probably highlights the role of substance P in neurogenic (nervous system-mediated) inflammation. Once the process of neurogenic inflammation has been triggered, it can become self-perpetuating. Neurogenic inflammation has been implicated in a number of chronic functional disorders of uncertain aetiology such as interstitial cystitis and irritable bowel syndrome.

The desensitisation of C-fibres has value for pain relief in a number of chronically painful disorders. Early controlled clinical trials of the topical use of capsaicin cream demonstrated symptom relief in osteoarthritis,³²⁹ neuropathy³³⁰ and postherpetic neuralgia.³³¹ Topical capsaicin was effective for painful skin disorders such as psoriasis and pruritus and possibly useful for neural dysfunction in the form of cluster headaches and phantom limb pain.^332,333 Vasomotor rhinitis could also be alleviated by topical capsaicin.³³⁴

However, Cochrane reviews have demonstrated mixed outcomes for the topical application of capsaicin, being negative for pruritus,³³⁵ mildly positive for neuropathic pain³³⁶ and stating insufficient evidence to support intranasal application in allergic rhinitis.³³⁷ Four weeks of a 0.0125% capsaicin gel was an effective treatment for mild to moderate knee osteoarthritis in a recent placebo-controlled trial involving 100 patients,³³⁸ but its role in this disorder is controversial,³³⁹ despite a positive meta-analysis (yielding moderate evidence of benefit).³⁴⁰

The higher fibrinolytic activity observed in Thai people has been attributed to daily intake of cayenne pepper.³⁴¹ Cayenne also increases gastric acid output.³⁴² Since gastric acid is a natural defence against gastrointestinal pathogens, it can be speculated that this could explain the preference for hot, spicy food in tropical countries. In the rat stomach there is clear evidence that capsaicin-sensitive sensory nerves are involved in a local defence mechanism against gastric ulcer. However, excessive capsaicin exposure caused tachyphylaxis and impaired this defensive mechanism.³⁴³ Perhaps the maxim of ‘too much of a good thing’ applies here.

Like capsaicin, piperine has attracted research interest, but for quite different reasons. Attention has focused on the capacity of piperine to enhance the bioavailability of other agents. These include aflatoxin B1 in rats³⁴⁴ and propanolol, theophylline,³⁴⁵ curcumin,³⁴⁶ vasicine and sparteine in humans.³⁴⁷ While there is good evidence that piperine inhibits drug metabolism by the intestine and liver,³⁴⁸ other possibilities include increased permeability of intestinal cells³⁴⁹ and even complexation with drugs.³⁵⁰ It strongly inhibits hepatic and intestinal aryl hydrocarbon hydroxylase and UDP-glucuronyl transferase.³⁵¹ Hence, piperine also has the potential to induce herb-drug interactions.

In traditional Chinese medicine, a mixture of radish and pepper is used to treat epilepsy. Piperine and some synthetic derivatives have been shown to be anticonvulsant agents that can antagonise convulsions induced by physical and chemical methods.³⁵² Antiepilepsirine, a derivative of piperine, is used as an antiepileptic drug in China.

Hot spices are used around the world for their general warming effects on the body, in modern terms stimulating circulatory activity. A mechanism for this universal experience is suggested. In a study that compared the effects of various pungent agents, an increase in catecholamine secretion, especially adrenaline, from the adrenal medulla was observed. Capsaicin was most active, but piperine and zingerone (from ginger) also showed activity, although allylisothiocyanate (from mustard) and diallyldisulphide (from garlic) did not have this effect. The effective principles were readily transported from the gut around the body.³⁵³ Capsaicin is known to increase energy expenditure in the body and boost basal metabolic rate, which has implications for its use in weight control.³⁵⁴

Epidemiological data reveal that the consumption of foods containing capsaicin is associated with a lower prevalence of obesity. Rural Thai people consume diets containing 0.014% capsaicin.³⁵⁵ Rodents fed a diet containing 0.014% capsaicin showed no change in caloric intake, but demonstrated a significant 24% reduction in visceral fat weight.

The above findings hark back to the use of cayenne promoted by Samuel Thomson, who regarded it as a life-promoting heating herb and a general metabolic stimulant. The physiomedicalists extended this concept and proposed that cayenne administered in conjunction with another herb would augment the particular stimulatory activity of that herb. Given what we now know about the bioavailability effects of piperine (and therefore possibly capsaicin), this apparently arcane dictum may have a rational basis.

A discussion of the pharmacology of the gingerols is included in the ginger monograph.

Investigations have been conducted to determine the potential mutagenic and carcinogenic activity of capsaicin and cayenne, but findings are contradictory.³⁵⁶ Results indicate that capsaicin also demonstrates chemoprotective activity against some chemical carcinogens and mutagens.³⁵⁷ Piperine appears to lack mutagenic activity.³⁵⁸ A review of the cayenne and capsaicin toxicity data did not find any reasons for concern.³⁵⁹

Saponins are phytochemicals that produce a foam when dissolved in water. Their name derives from the same root as the word soap (Latin sapo=soap). Like soaps or detergents, saponins are large molecules containing a water-loving (hydrophilic) part at one end separated from a fat-loving (lipophilic or hydrophobic) part at the other. In aqueous solution, saponin molecules align themselves vertically on the surface with their hydrophobic ends oriented away from the water. This has the effect of reducing the surface tension of the water, causing it to foam. For this reason, saponins are classified as surface-active agents. Similar to other surface-active agents, saponin molecules can align to form a spherical configuration within the water, creating a micelle. Micelles have a lipophilic centre, and this creation of a fat-loving compartment explains why detergents can dissolve grease and oils.

Saponins are glycosides (the sugar part comprises the hydrophilic end). Two classes are recognised based on the structure of their aglycone or sapogenin: steroidal saponins contain the characteristic four-ringed steroid nucleus, and triterpenoid saponins have a five-ringed structure. Both of these have a glycosidic linkage, usually at carbon 3, and share a common biosynthetic origin via the mevalonic acid pathway. Steroidal saponins are mainly found in the monocotyledons. Triterpenoid saponins are by far the most common. There are some unusual classifications; for example, the ginsenosides in ginseng are grouped with the triterpenoid saponins even though they exhibit a steroidal structure. Steroidal saponins typically contain extra furan and pyran heterocyclic rings, which is not a feature of the ginsenosides. (Furans and pyrans are respectively five- and six-membered rings containing oxygen.)

Saponins are consumed in many common foods and beverages including oats, spinach, asparagus, soya beans and other legumes, peanuts, tea and beer.

The pharmacological events that result from the ingestion of saponins can be broadly classified into two categories: the general effects from the detergent-like properties of intact saponins, and those specific actions that ensue after a saponin (or more usually the sapogenin) is absorbed into the bloodstream. This discussion will concentrate on the former, but will also provide a context for the specific activities of saponins. Also the anti-inflammatory, tonic, adaptogenic, aldosterone-like and mucoprotective properties (among others) of specific saponins are examined in detail in the monographs on Astragalus, black cohosh, Bupleurum, ginseng, horsechestnut, licorice, poke root, gotu kola and Withania.

Saponins are capable of destroying red blood cells (RBCs) by dissolving their membranes, a process known as haemolysis, releasing free haemoglobin into the bloodstream. RBCs are particularly susceptible to this form of chemical attack because they have no nucleus and therefore cannot effect membrane repair. Haemolysis explains why saponins are much more toxic when given by injection than when administered orally. The toxic dose of an injected saponin occurs if sufficient haemoglobin is released to cause renal failure (haemoglobin damages the delicate membranes of the glomerulus). After oral intake, much of the saponin is not absorbed intact. Typically, it is slowly and partially absorbed as the aglycone. The kidneys are thereby spared the sudden influx of haemoglobin. For a long time this feature of saponin toxicity was interpreted by many pharmacologists as an indication that saponins were largely inert after oral doses. While it is true that saponins and even their sapogenins are generally not well absorbed from the gut, there can be no doubt that they can exert significant pharmacological activity after ingestion (for example, the aldosterone-like licorice side effects).

A common misconception is that the haemolytic activity of saponins is related to their detergent-like characteristics. However, as early as the 1960s this was shown to be false. The mechanism of saponin-induced haemolysis was investigated by extracting the active haemolysing factor from ghost cells of saponin-haemolysed blood. The fact that only the corresponding aglycones could be extracted shows that hydrolysis of the glycosidic bond precedes haemolysis (RBC membranes possess a beta-glucosidase). The lack of haemolytic activity of a saponin was either due to the fact that it could not be hydrolysed by the beta-glucosidase or that it could not adsorb onto the RBC membrane.³⁶⁰ This is further supported by the observation that some sapogenins are also quite haemolytic (for example, glycyrrhetinic acid is actually more haemolytic than glycyrrhizin)³⁶⁰ and that haemolysis by saponins is inhibited by aldonolactones, which are glycosidase inhibitors.³⁶¹

Saponins are more or less irritant to gastrointestinal mucous membranes (whether this is related to their detergent or haemolytic properties is not understood). This irritant property creates an acrid sensation in the throat when a saponin-containing herb is chewed. One resultant effect, like the emetics, may be by upper gastrointestinal irritation to induce a reflex expectoration. Certainly many of the traditional expectorant herbs such as soapbark, senega, primrose root and ivy leaf are rich in acrid saponins. This reflex expectorant effect, and its relationship to emesis, has been demonstrated in animal models.³⁶² Presumably, it is mediated via the vagus nerve, as it was abolished when the afferent gastric nerves were cut.

The detergent effect of saponins helps to increase the solubility of lipophilic molecules via micelle formation. One example that illustrates this phenomenon is the kava lactones. These compounds are quite insoluble in water, yet water can readily extract a percentage of lactones from kava root, which also contains a significant amount of saponins. Moreover, the bioavailability of kavain from a kava matrix is much greater than for the pure compound (see the kava monograph), possibly due to the presence of saponins.

However, this phenomenon for kava does not necessarily apply to other herbs. The influence of saponins on the water solubility of compounds having poor aqueous solubility was investigated with the model compounds digitoxin, rutin and aesculin.³⁶³ The saponins that were tested represented the most common structural types. Only slight effects on aqueous solubility were obtained for all the model compounds. These effects include both enhancement and reduction in aqueous solubility, depending on the saponin, the concentration of the saponin solution and the model compound. Hence, saponins generally should not be regarded as solubilisers.

Early research suggested that the incorporation of saponins into the cell membrane probably forms a structure that is more permeable than the original membrane.³⁶⁴ Saponins readily increase the permeability of the mammalian small intestine in vitro, leading to the increased uptake of otherwise poorly permeable substances and a loss of normal function.³⁶⁵ The disruptive effect of saponins on the architecture of the enterocyte cell membrane can lead to the impaired absorption of smaller nutrient molecules, which are otherwise rapidly absorbed via specific transporters. This appears to be the case for glucose and ethanol, based on in vitro models.^366,367 Tablets containing an extract of the Ayurvedic herb Gymnema sylvestre are sold in Japan as a weight loss agent, which is attributed to reduced glucose absorption rates. Results from one study suggest that the saponin-rich herb Asparagus racemosus can act as a transdermal permeation enhancer for larger molecules.³⁶⁸

Sapogenins, both steroidal and triterpenoid, resemble cholesterol in their structure and may have a profound effect on cholesterol metabolism in the liver. Diosgenin is well studied in this regard. It interferes with the absorption of cholesterol of both dietary and endogenous origin; such interference is accompanied by increased rates of hepatic and intestinal cholesterol synthesis. Diosgenin also markedly enhanced cholesterol secretion into bile which, in conjunction with the unabsorbed cholesterol, resulted in increased faecal excretion of cholesterol (neutral sterols) without affecting excretion of bile acids.³⁶⁹ Higher levels of ingestion of saponins or sapogenins lead to cholestasis and jaundice associated with the presence of cholesterol-like crystals in hepatocytes. In grazing animals, this can in turn lead to photosensitivity due to the photosensitising effect of the high plasma levels of chlorophyll metabolites resulting from impaired biliary excretion (see the Tribulus monograph).

It is well established that saponin intake lowers plasma cholesterol levels in animal models.³⁷⁰ Various mechanisms have been put forward, such as increased faecal excretion of bile salts³⁷⁰ and loss of cholesterol via exfoliated mucosal cells,³⁶⁵ but the mechanism suggested by the above research on diosgenin seems most likely.³⁶⁹ In this context, it is pertinent to note that diosgenin (a sapogenin) inhibits cholesterol absorption, suggesting that the mechanism behind this phenomenon might have more to do with the haemolytic or steroid-like nature of the saponins rather than their detergent properties. As an interesting footnote, it has been suggested that the Masai, who consume 2000 mg of cholesterol per day yet maintain very low blood levels, might achieve this feat by their high consumption of saponins from both foods and medicines.³⁷¹ To date this potential for saponin-containing herbs has not really been exploited clinically.

Saponins may have beneficial effects on bowel flora; yucca saponins are added to commercial piggery feeds to suppress ammonia production. As testimony to the fact that a percentage of saponins pass through the digestive tract unabsorbed, the product brochures maintain that the faecal material is easier to hose down because of the detergent action of excreted saponins.

A recent review surveyed research of the effects of saponins on microbial populations and fermentation in ruminants.³⁷² The primary effect of saponins in the rumen appears to be inhibition of protozoa (defaunation), which might increase the efficiency of microbial protein synthesis and protein flow to the duodenum. Furthermore, saponins may decrease methane production via defaunation and/or directly by decreasing the activities and numbers of methanogens. Saponins may also selectively affect specific rumen bacteria and fungi, which may alter the rumen metabolism. The effects of saponins on rumen fermentation have not been found to be consistent. These discrepancies appear to be related to the chemical structure and dosage of saponins, diet composition, microbial community and adaptation of microbiota to saponins. The clinical effects of saponins on human bowel flora have not yet been adequately explored.

Saponins are very gentle detergents that can be used to wash the hair and skin in conditions such as acne without causing a rebound increase in sebum production. Decoctions of soapwort have been used to wash and restore ancient fabrics in stately homes; modern soaps are too harsh and disintegrate the fabrics.

One of the most interesting effects of saponins (or sapogenins) that follows from their ingestion is their capacity to interact with and influence steroid hormone metabolism. Much of this is covered in the monographs, but some of the principles behind this are worth highlighting here. The role of enzymes that metabolise steroid hormones in the regulation of the activity of these hormones has only been relatively recently appreciated. The most pertinent example for phytotherapy is 11-beta-hydroxysteroid dehydrogenase. This enzyme regulates glucocorticoid action by catalysing the interconversion of hydrocortisone (cortisol) and cortisone, an inactive steroid. In the kidney, hydrocortisone is oxidised to cortisone by the type 2 form of this enzyme, a reaction that prevents circulating hydrocortisone from occupying kidney mineralocorticoid receptors and stimulating a mineralocorticoid response. Aldosterone, being inert to 11-beta-hydroxy-steroid dehydrogenase (11beta-HSD), is then available to regulate mineralocorticoid responsive genes in the kidney on its own. Inhibition of 11beta-HSD in the kidney allows hydrocortisone to exert an additional aldosterone-like effect. This is exactly what licorice does.⁷

An important consequence of this mechanism for regulating hormone action is that compounds that inhibit these enzymes will appear to be acting as hormones (or antihormones). Thus the active sapogenin from licorice appears to behave like aldosterone, even though it has no affinity at all for mineralocorticoid receptors. Similar examples need to be better understood. For example, the anti-inflammatory effect of escin from horsechestnut is abolished in the absence of steroid hormones.

The influence of steroidal saponins on oestrogen metabolism may share similar aspects, but perhaps with some noteworthy differences. This is illustrated by research on Tribulus terrestris. Saponins from Tribulus appear to increase FSH (follicle-stimulating hormone) in premenopausal women, which in turn increases levels of oestradiol.³⁷³ They may do this by binding with, but only weakly stimulating, hypothalamic oestrogen receptors, which are part of the negative feedback mechanism of oestrogen control. The weak stimulus (as opposed to the strong stimulus of oestrogen) leads the body to interpret that oestrogen levels are lower than they really are. Consequently, the body increases oestrogen production via the negative feedback mechanism. In the postmenopausal woman, herbs such as Tribulus, wild yam (Dioscorea villosa) and false unicorn (Chamaelirium luteum) appear to alleviate symptoms of oestrogen withdrawal. It is possible that the binding of plant steroids to vacant receptors in the hypothalamus (in this low-oestrogen situation) is sufficient to convince the body that more oestrogen is present in the bloodstream than actually is, and calms the body’s response to oestrogen withdrawal.

Claims have arisen in the popular literature that the female body can manufacture progesterone from diosgenin, particularly if a wild yam cream is applied to the skin. This is despite the fact that wild yam contains dioscin and other steroidal saponins, not diosgenin, and there is no information about the dermal absorption of these compounds. Furthermore, no evidence exists for mammalian enzymes that are capable of effecting what is a difficult chemical conversion. The evidence that does exist strongly disputes the possibility of this conversion.³⁷⁴ Hopefully, the above discussion demonstrates that the interaction of saponins with steroid hormones is far more subtle than this. In fact, diosgenin appears to have oestrogenic properties in mice and lacks progesterogenic effects.³⁷⁵

In addition to this already impressive array of diverse activities for saponins, further effects have been demonstrated using in vitro and in vivo models. However, the results from the former types of studies need to be interpreted with caution, since in the majority of cases saponins will not be absorbed in sufficient quantities to manifest these effects systemically (and the sapogenins do not always share these properties). In vitro findings may have relevance to the topical activity of saponins and effects in the gut lumen, and where the saponin metabolite in the bloodstream has pharmacological properties similar to its parent saponin.

A review of the in vitro and in vivo effects of triterpenoid saponins has documented pharmacological properties such as antiallergic, antiatherosclerotic and antiplatelet, antibacterial, anticomplementary, antidiabetic, contraceptive, antifungal, anti-inflammatory, antileishmanial, antimalarial/antiplasmodial, anti-obesity, anti-proliferative, antipsoriatic, antispasmodic, antisweet, antiviral, cytotoxic/antitumor, gastroprotective, hepatoprotective, immunomodulatory, anti-osteoporotic and insulin-like.³⁷⁶

Saponins are gastrointestinal irritants. In milder examples, this can lead to oesophageal reflux in sensitive or overweight patients, which can be remedied by using enteric-coated preparations or by taking the saponin-containing herbs during a meal. In more severe examples, such as with poke root, this irritation can lead to acute vomiting and diarrhoea. If the irritation is sufficiently prolonged, erosion of the gastrointestinal mucosa can occur, resulting in substantial absorption of saponins into the bloodstream with the expected toxic sequelae. This is probably the mechanism via which acute oral doses of saponins cause death in humans.

Saponins are toxic to fish and other cold-blooded animals and have been used to kill snails that harbour the bilharzia parasite.³⁷⁷ However, normal intake of the majority of saponins is not toxic to humans, as evidenced by the fact that saponin intake by vegetarians is in the range of 100 to 200/day.³⁷⁰ Saponins are permitted as food additives in many countries (for example, to give a satisfactory head to beer), although there are inconsistencies in regulations.³⁷⁸ Long-term feeding of saponins to animals did not demonstrate any signs of toxicity.³⁷⁸

As noted previously, grazing animals that consume large amounts of saponins can develop cholestatic liver damage. In South Africa, consumption of Tribulus by sheep contributes to a bilirubin-induced disorder known as geeldikkop (see the Tribulus monograph for a more comprehensive discussion).³⁷⁹ While it is unlikely that normal human doses would cause cholestasis, this phenomenon should be considered in unexplained cases of this disorder in patients taking herbs. More importantly, saponin-containing herbs are best kept to a moderate level in patients with pre-existing cholestasis.

These compounds, also known as cardioactive glycosides, are steroidal glycosides. They are similar to, but essentially different from, steroidal saponins and constitute a well-defined and highly homogeneous group from both a structural and a pharmacological perspective. They (or the plants that contain them) have been used as conventional medical drugs for over 200 years and are still relatively widely used today, despite their potential toxicity and the controversy over their clinical value. There are two main types, which either have a steroidal aglycone with 23 carbons (the cardenolide glycosides) or 24 carbons (the bufadienolide glycosides). The sugar part usually consists of 2 to 4 sugars joined together (an oligosaccharide).

The properties of cardiac glycosides are very well documented. They inhibit the sodium-potassium cellular pump leading to a rise in intracellular calcium levels that increases the contractile force and speed of the heart muscle (positive inotropy). In patients with compromised cardiac function, this positive isotropic effect translates into increased cardiac output and associated events that flow from this central effect. Heart rate is decreased (negative chronotropy) via the autonomic nervous system. Digitalis glycosides in particular also cause electrophysiological changes in the heart: conduction velocity at the atrioventricular node is slowed, together with an increase in its refractory period (negative dromotropy). This last feature underlies the use of digitalis glycosides in supraventricular rhythm abnormalities such as atrial fibrillation.

It is beyond the scope of this chapter to discuss the pros and cons of digitalis therapy or the various adverse reactions and toxicity considerations associated with its use. Its medical use certainly appears to be in decline, although there is continued evidence of its frequent clinical utility.³⁸⁰ Phytotherapists find digitalis to be too powerful for comfort and prefer to see it used, if necessary, under appropriate specialist care (generally the legal restrictions on its use render it available only on a medical prescription).

In some countries, herbalists and physicians practising phytotherapy find value in the use of milder herbs containing cardiac glycosides, in particular lily of the valley (Convallaria majalis). Its properties are similar to those of digitalis, but much less cumulative. The principal glycoside is convallatoxin, but the plant contains many minor cardenolides. Convallatoxin is poorly absorbed as a pure compound but the other components in the herb are said to aid its absorption.³⁸¹ It is unsuitable for manifest congestive heart failure, but combines well with hawthorn in the management of milder forms of this condition, for digitalis hypersensitivity or for heart failure associated with a low pulse rate.³⁸¹ All the usual cautions governing the use of cardiac glycosides should be observed.

Recent reviews have highlighted the novel therapeutic applications of cardiac glycosides, in particular the accumulating in vitro and in vivo studies reporting antitumour activity.^382,383 The first generation of glycoside-based anticancer drugs are in clinical trials.

Anthraquinones, as the name implies, are phytochemicals based on anthracene (three benzene rings joined together). At each apex of the central ring is a carbonyl group (carbon double-bonded to oxygen), and these provide the quinone part. Not all anthraquinones are strictly quinones; for example, the sennosides are dianthrones consisting of two anthrone units, each bearing only one carbonyl group. Anthraquinones usually occur in plants as glycosides; for example, the sennosides from senna (Cassia species) are O-glycosides and the aloins from Aloe are C-glycosides. The cascarosides from cascara (Rhamnus purshiana) are unusual molecules in that they are C,O-glycosides, having one glucose linked to a central anthrone via a carbon atom and a second glucose linked via oxygen. If aglycones are present in dried herbs, they are always anthraquinones; anthrones are too unstable in the free state. Dianthrone glycosides such as the sennosides are not found in the living plant, being formed on harvesting and drying from monomeric anthrone glycosides.

Plants like rhubarb, senna and cascara have been used for their laxative effects since prehistory. Given their widespread contemporary use, it should not be surprising that their pharmacology is relatively well studied. Of particular interest is the pharmacokinetics of the anthraquinone glycosides. This topic will be more extensively discussed in the pharmacokinetics section, but, briefly put, these agents travel to the large bowel where bacterial action forms anthrone aglycones, the true active forms.

The laxative effect on the gut is largely a local one; systemic absorption is limited. Two distinct mechanisms are in force: a modification of intestinal motility and an accumulation of fluid in the intestinal lumen.³⁸⁴ Experiments in animals and humans have shown that the introduction of anthrones into the colon quickly induces vigorous peristaltic movements.³⁸⁴ Such a fast response is certainly not a secondary effect resulting from increased faecal water. This effect on motility is at least in part due to the release of prostaglandins, since its action is abolished by indomethacin and other cyclo-oxygenase inhibitors.³⁸⁵ However, other research shows that inhibition of intestinal tone and segmentation (and therefore reduced colonic transit time) could be the primary motility effect.³⁸⁴

Accumulation of fluid in the colonic lumen also leads to a laxative effect. While it has been suggested that this might be due to the inhibitory effect of anthrones on the sodium pump,³⁸⁴ these agents have instead been shown to stimulate active chloride secretion into the lumen, which is then balanced by an increase in sodium and water flow.³⁸⁶ Prostaglandins may be involved in this process³⁸⁷ but not PAF.³⁸⁸ Alteration of calcium transport may also play a role.³⁸⁴

The dual action of anthraquinones highlights an important aspect of their safe and effective usage. In lower laxative doses that produce a normal bowel motion, the effects on motility are dominant. In higher doses, electrolyte secretion and diarrhoea will predominate. Habituation and adverse effects are more likely to result from the excessive electrolyte loss associated with the use of high doses. Chronic abuse of laxatives raises aldosterone levels in response to the electrolyte loss, diminishing their effectiveness. Higher doses also empty a larger portion of the colon. The resulting natural absence of defaecation over the next day or so leads to anthraquinone reuse, perhaps at an even higher dose, and the cycle of laxative need is perpetuated.

Natural anthraquinones in the form of chrysarobin have also been used topically in the treatment of psoriasis. Chrysarobin, a mixture of substances including chrysophanol, is obtained from araroba or goa powder. Araroba is extracted from cavities in the trunk of the South American tree Andira araroba.³⁸⁹ Dithranol (or anthralin), a cheaper synthetic analogue of chrysarobin, has been the focus of the most studies.³⁹⁰ Chrysarobin is an effective topical agent for psoriasis, as was demonstrated in an open comparative study with coal tar and ultraviolet radiation.³⁹¹ However, it has a number of drawbacks: it is only stable in a greasy base and it irritates and stains the skin.

The antiproliferative action of dithranol, and presumably chrysarobin, is thought to be either through effects on DNA, probably mitochondrial DNA, which reduces cell turnover, or through activity on various vital enzyme systems.³⁹² A possible relationship with reduction-oxidation potential has also been observed.³⁹³ Other research on chrysarobin has been concerned with its tumour-promoting activity.³⁹⁴ Not surprisingly, this finding has further reduced its use in therapy.

Hypericin and pseudohypericin are dianthrones (structurally related to anthraquinones) with antiviral activity. Several anthraquinone aglycones including rhein, alizarin and emodin have also demonstrated antiviral activity against human cytomegalovirus.³⁹⁵ This may explain the traditional use of applying the leaves of Cassia species to viral skin conditions.³⁸⁹ It is unlikely that systemic antiviral effects would follow from the ingestion of anthraquinones, due to their low bioavailability.

Rhein is an anthraquinone aglycone found in rhubarb that inhibits the activity of cytokines in models of osteoarthritis.³⁹⁶ This observation led to the development of diacerhein (diacetylrhein), a synthetic derivative with better bioavailability. In early clinical studies, oral diacerhein at 100 mg/day improved symptoms in patients with osteoarthritis.³⁹⁶ Diarrhoea was a common side effect, as might be expected. A later Cochrane review of seven controlled clinical trials concluded that there is ‘gold’ level evidence that diacerhein (diacerein) has a small consistent benefit in the improvement of pain in osteoarthritis.³⁹⁷

Madder root (Rubia tinctorum) contains a characteristic spectrum of intensely coloured anthraquinone glycosides such as glycosides of lucidin and alizarin. As well as being used as a vegetable dye and natural food colouring, madder has traditionally been employed for the prevention and treatment of kidney stones. The anthraquinones in madder can function as chelating agents for some metal ions, such as calcium and magnesium. With oral doses of glycosides and aglycones, a pronounced calcium complexing effect and a significant reduction in the growth rate of kidney stones was observed in an animal model.³⁸⁹ A therapeutic oral dose of madder root will colour the urine slightly pink, which indicates that significant quantities of anthraquinones are excreted in the urine. Its regular use is said to slowly dissolve kidney stones. Given the concerns (below) over the carcinogenicity of madder, perhaps other anthraquinones sharing this property could be explored for potential value as oral chelation therapy in patients with atherosclerotic lesions.

Madder has been withdrawn from the German market due to concerns over its mutagenicity and potential carcinogenicity. In particular, rats metabolise alizarin glycosides to alizarin and 1-hydroxyanthraquinone, a known rodent carcinogen.³⁹⁸ One long-term feeding of madder root to mice failed to produce neoplastic lesions.³⁹⁹ However, studies that are more recent do tend to confirm carcinogenic activity.⁴⁰⁰

Anthraquinone laxatives may cause mild abdominal complaints such as cramps and abdominal pain and should be used cautiously in patients with these symptoms. Other side effects include discoloration of the urine and haemorrhoid congestion. Overdosage can result in diarrhoea and, coupled with prolonged use, may cause excessive loss of electrolytes, particularly potassium.³⁸⁴

Abuse of anthraquinone laxatives has been stated to cause damage to the myenteric plexus. However, results from animal studies and a controlled study in humans have challenged this finding.⁴⁰¹ Habituation can occur with laxative abuse, primarily due to hyperaldosteronism. The evidence is that anthraquinone laxatives used sensibly are unlikely to cause habituation. In fact, in several studies senna has been claimed to have a re-educative function and helped to restore normal bowel function.⁴⁰² For example, out of 210 patients in a psychiatric hospital, 44% were taking laxatives at the outset, but 3 months’ treatment with senna lowered this to 8%.⁴⁰³

Cathartic colon coupled with hypokalaemia (low serum potassium) is a characteristic finding associated with chronic laxative abuse (not just for anthraquinones). The cathartic colon is characterised by the existence of a segment of intestine (typically the ascending colon) that has become largely non-functioning. Cathartic colon presents as chronic diarrhoea resistant to therapy. This is probably a rare disorder and is often associated with psychological abnormalities.⁴⁰⁴ Cathartic colon can be reversed; a case was described of a 78-year-old woman with cathartic colon who, after treatment was switched to psyllium, demonstrated complete reversal of the condition after 4 months.⁴⁰⁵

Contraindications for anthraquinone laxatives include ileus from whatever cause. Use in pregnancy and lactation is controversial. In traditional Chinese medicine anthraquinone- containing herbs are contraindicated in pregnancy because they promote a downward movement of energy. Apart from this consideration, they are unlikely to cause adverse effects since their action in the gut is largely topical and systemic absorption is limited (see Pharmacokinetics). Results from several clinical studies on senna indicate that its normal use does not involve any increased risk for the pregnancy or the fetus.⁴⁰⁶ Moreover, clinical observations of infants and analytical studies of breast milk both lead to the conclusion that the treatment of lactating mothers with senna does not carry a risk of producing a laxative effect in the infant.⁴⁰⁷

Long-term use of anthraquinone laxatives leads to a condition known as melanosis (or pseudomelanosis) coli. This is a brown discoloration of the intestinal mucosa beginning at the ileocolonic junction and may extend to the rectum. This harmless pigmentation is not due to staining by the anthraquinones, but is instead a lipofuscin-like substance within the macrophages of the colonic mucosa.³⁸⁴ (Lipofuscin is a pigmented, peroxidised fatty acid residue found in many organs with advancing age.) However, the intrinsic colour of anthraquinones may play some part in the development of this pigment. It is generally accepted that melanosis is a benign, reversible condition.^408,409

The mutagenicity of isolated anthraquinones has been extensively studied and established in several models, particularly microbial assays such as the Ames test.⁴¹⁰ On the other hand, some studies involving both in vitro and in vivo models found an absence of mutagenicity for senna, sennosides and rhein.⁴¹¹ The same research group concluded that senna does not represent a genotoxic risk to humans when used periodically at normal therapeutic doses.⁴¹²

Nonetheless, their putative mutagenicity has led to the investigation of carcinogenic effects associated with the use of anthraquinone laxatives. One review examined the relationship between anthraquinone laxatives and colorectal cancer.⁴¹³ Danthrone (two studies) and 1-hydroxyanthraquinone (one study) were carcinogenic in rodent models. Three clinical studies did not show an association of colorectal cancer with laxative abuse, but these studies also included bulk laxatives. When melanosis coli was taken as an indicator of anthraquinone laxative use, a retrospective study of 3049 patients undergoing endoscopic diagnosis for suspected colorectal cancer found an association between anthraquinone use and colorectal adenomas.⁴¹³ A prospective study by the same research group found that 18.6% of patients with colorectal carcinoma had evidence of anthraquinone use.⁴¹³ From their data, they suggested that a relative risk of 3.04 could be calculated for colorectal cancer due to anthraquinone laxative abuse. However, these studies did not differentiate between natural and synthetic (danthrone) laxative use. This is significant, because relatively higher doses of danthrone are required for a laxative effect and, since it is not a glycoside, its systemic absorption is greater. It was later asserted that epidemiological studies do not give conclusive evidence for any association between anthraquinone use and colorectal cancer.³⁸⁴ Since then, a comprehensive review of the pharmacological and human data concluded that (1) there is no convincing evidence that the chronic use of senna can induce a structural and/or functional alteration of the enteric nerves or the smooth intestinal muscle, (2) there is no relation between long-term administration of a senna extract and the appearance of gastrointestinal tumours (or any other type) in rats, (3) senna is not carcinogenic in rats even after a 2-year daily dose of up to 300 mg/kg/day, and (4) the current evidence does not show that there is a genotoxic risk for patients who take laxatives containing senna extracts or sennosides.⁴¹⁴

Given the above concerns, whatever their basis in fact, most regulatory authorities now require that anthraquinone laxatives are labelled as a short-term treatment for constipation. Since the goal of the phytotherapist is to effect change (re-educate the bowel) rather than compensate for a deranged physiology, this approach is consistent with good herbal clinical practice.

Coumarins owe their name to ‘coumarin’ which was the common name for the tonka bean (Dipteryx odorata), from which the simple compound coumarin was first isolated in 1820. Coumarins are benzo-alpha-pyrones (lactones of o-hydroxycinnamic acid) formed via the shikimic acid pathway. Except for a few rare cases, including coumarin itself which is unsubstituted, all plant coumarins contain hydroxy or methoxy groups in position 7. These substituted simple coumarins, such as scopoletin, aesculetin and umbelliferone, are common and widespread in higher plants and often occur as glycosides.

Simple coumarin has a pleasant vanilla-like odour. It is probably not present in the intact plant, but is rather formed by enzymatic activity from a glycoside of o-hydroxycinnamic acid (such as melilotoside) after harvesting and drying. This accounts for the odour of newly-mown hay, not present in the undamaged plant. Coumarin is used to perfume pipe tobacco and can be sometimes found as an adulterant in commercial vanilla flavouring.⁴¹⁵

The furanocoumarins are closely related furano derivatives of coumarin (furan is a five-membered heterocyclic ring containing oxygen) commonly found in the Rutaceae (rue family) and Umbelliferae (Apiaceae, celery family). Linear furanocoumarins are often called psoralens and are photosensitising agents (see below).

The furanochromones such as khellin from Ammi visnaga are structural derivatives of benzo-gamma-pyrone (furanobenzo-gamma-pyrones) and therefore are as much related to flavonoids as coumarins. In other words, the carbonyl group is opposite the oxygen rather than adjacent to it. However, they are usually classified as coumarins and will be considered in this section. Ammi visnaga also contains visnadin, a pyranocoumarin.

The widespread nature of coumarins means that they are consumed in the human diet, for example carrots, celery and parsnip.⁴¹⁶ Coumarins are fluorescent compounds and this property is widely utilised in a number of biochemical techniques. Simple substituted coumarins are also used as pigments in sunscreens.

Dicoumarol is a potent anticoagulant compound formed from coumarin by bacterial action in spoiled sweet clover hay (Melilotus species). Its discovery led to the development of modern anticoagulant drugs. Dicoumarol and related anticoagulants are hydroxylated in the 4 position. This is deemed an essential requirement (among others) for powerful anticoagulant activity. All of the common plant coumarins are not substituted at this position and therefore lack significant clinical anticoagulant activity, although many do possess very limited activity when given to animals in high doses.⁴¹⁷ Coumarin has anti-oedema, anti-inflammatory, immune-enhancing and anticancer activities which are more fully described in the Melilotus monograph.

Simple substituted coumarins such as scopoletin and umbelliferone have exhibited a diverse range of pharmacological activities. The spasmolytic activity of scopoletin is probably a major reason for the use of Viburnum species such as cramp bark and black haw for hypertension and dysmenorrhoea. Scopoletin and aesculetin were identified in black haw as having significant spasmolytic activity on guinea pig small intestine.⁴¹⁸ Another research team working at the same time also identified scopoletin as a component of the Viburnums that exhibited uterine spasmolytic activity.⁴¹⁹ It has been suggested that the spasmolytic activity of scopoletin might be due to a blockade of autonomic neurotransmitters.⁴²⁰

Further studies are required to establish if scopoletin, umbelliferone and related compounds possess clinically relevant anti-inflammatory and analgesic activities, but such activities have been established in animal studies.⁴²¹ Inhibition of cyclo-oxygenase and 5-lipoxygenase may play a role.⁴²² Aesculetin and umbelliferone are also strong xanthine oxidase inhibitors in vitro.⁴²³

Of eight simple coumarins tested, aesculetin was the most potent at protecting cells in vitro against oxidative damage.⁴²⁴ The coumarins all inhibited xanthine oxidase in vitro.⁴²⁴ Aesculetin possessed hypouricaemic activity in rats (100 mg/kg, ip) and mice (150 mg/kg, ip), but was devoid of xanthine oxidase inhibitory activity at these doses.⁴²⁵ Scopoletin (50 to 200 mg/kg, ip) was also hypouricaemic after induced elevation of uric acid in mice, but did not affect the serum uric acid level in normal mice.⁴²⁶ It exhibited both xanthine oxidase inhibition and uricosuric activity. However, the high doses of such phytochemicals given by injection probably have little relevance to phytotherapy.

Like coumarin, substituted coumarins may have a role to play in cancer prevention and treatment. Umbelliferone and scopoletin are antimutagenic^427,428 but were much less protective than coumarin in one model of chemical carcinogenesis.⁴²⁹ Aesculetin exhibited considerably higher cytotoxic activity than coumarin in vitro on two tumour cell lines, but scopoletin was inactive.⁴³⁰ Umbelliferone has similar cytotoxic activity to coumarin.⁴³¹ Two reviews have outlined the potential of simple natural coumarins in cancer therapy, but most of the data to date are in vitro.^432,433

Other in vivo properties of simple coumarins include gastroprotective activity for aesculin,⁴³⁴ and antiasthmatic activity for umbelliferone,⁴³⁵ all in mice, together with antithyroid activity for scopoletin in hyperthyroid rats⁴³⁶ and antidiabetic activity for umbelliferone in diabetic rats.⁴³⁷

Furanocoumarins have a long history of therapeutic use in humans. More than 3000 years ago, it was recorded in both Egyptian and Ayurvedic medicine that the ingestion of herbs containing psoralens followed by exposure to sunlight could assist in the treatment of vitiligo, a skin condition characterised by a loss of pigmentation.⁴³⁸ This traditional knowledge was developed into a modern therapy in the 1940s when xanthotoxin (8-methoxypsoralen, 8-MOP) plus sunlight exposure was introduced as a therapy for vitiligo.⁴³⁹ The treatment was not very successful, mainly due to phototoxic side effects, and it was only when ultraviolet A light sources (UVA) became available that significant advances were made. UVA radiation is less energetic and therefore less damaging than ultraviolet B (UVB) radiation. It was demonstrated that oral 8-MOP and UVA were highly effective in the control of psoriasis and the malignant skin condition mycosis fungoides.⁴³⁹ The treatment was called PUVA (psoralen plus UVA) and the new field of photochemotherapy was initiated.

Bergapten (5-methoxypsoralen, 5-MOP) is a psoralen with superior therapeutic characteristics in PUVA. Furanocoumarins in conjunction with ultraviolet radiation stimulate melanogenesis (tanning) and cause antiproliferative effects, but they also initiate phototoxic erythema (inflammation). For 8-MOP, the therapeutic dose is similar to the dose that causes erythema. With the use of oral 5-MOP it is possible to obtain melanogenic doses of drug-UVA or drug-sunlight combinations without the development of phototoxicity. This provides a wider margin of safety and permits the induction of a tan (known as a PUVA tan) that is more effective than a normal tan in attenuating the damaging effects of UV exposure. Psoralens are now the basis of yet another new field of therapy: photochemoprotection.⁴³⁹

The use of 5-MOP has also provided a more effective treatment for vitiligo and yields fewer side effects in PUVA therapy of psoriasis.⁴⁴⁰ As a further development in photochemoprotection, studies have shown that this furanocoumarin, in conjunction with UVB sunscreens, provides a faster tan (without burning) that is more protective against the harmful effects of ultraviolet radiation⁴⁴¹ and even chemical irritation.⁴⁴² It should be noted that certain citrus oils are rich sources of bergapten (5-MOP) and are used in these tanning products in preference to the synthetic chemical.

PUVA therapy is also used for psoriasis,⁴⁴³ although there are concerns over increased skin cancer risk (see below).⁴⁴⁴

Furanocoumarins in conjunction with UV light kill bacteria and inactivate viruses.^445,446 In addition, they may be responsible for the enhanced bioavailability that grapefruit juice affords to several pharmaceuticals (see the Pharmacokinetics section).

The decoction of the fruits of Ammi visnaga has been used since ancient times in Egypt as a spasmolytic for kidney stones and in the treatment of angina pectoris. The pyranocoumarin visnadin was isolated and exhibited positive inotropic and marked coronary vasodilatory activities. Visnadin was developed as a drug treatment for angina, possibly acting as a calcium channel blocker.⁴⁴⁷Ammi visnaga also contains the spasmolytic furanochromone khellin, adopted for the treatment of angina and asthma. Its use was discontinued because of side effects such as drowsiness, headache and nausea.

As further evidence that the archetypal plant constituents still provide much of the inspiration for the development of modern drugs, sodium cromoglycate was discovered in an attempt to find a derivative of khellin which was devoid of its side effects. Unlike khellin, the antiasthmatic activity of this drug is believed to result from increased stability of mast cells.⁴⁴⁸

The toxicology of coumarin is reviewed in the Melilotus monograph. It is hepatotoxic in the rat and also less commonly in humans, probably as an idiosyncratic reaction. Coumarins are structurally related to fungal aflatoxins, which are potent hepatotoxins, but it appears that the substituted coumarins do not share this property (see the Melilotus monograph).

Plant coumarins are incorrectly implicated as anticoagulant agents. This is a fundamental misunderstanding of the relevant pharmacology (see Chapter 5 and the Melilotus monograph). Although a case of increased haemorrhagic tendency was attributed to coumarin intake from tonka beans, sweet clover and woodruff, intake was excessive and there were a number of confounding factors.⁴⁴⁹

As noted above, the furanocoumarins are potent photosensitising agents. Chance exposure to these compounds in the field can lead to severe photodermatitis and blistering. The giant hogweed is one example of a plant that consistently causes this problem. Celery pickers can develop photodermatitis after handling celery infected with fungus, which induces celery to produce high levels of psoralens.⁴⁵⁰

Furanocoumarins are powerful mutagens in the presence of UVA and psoralens are light-activated carcinogens.⁴⁵⁰ In the presence of UV light, psoralens are capable of binding to and cross-linking DNA, which causes genetic damage and mutations. Only linear furanocoumarins (psoralens) are able to bind two DNA bases to cause cross-links. Non-linear furanocoumarins (isopsoralens) can only bind to one base to form monoadducts. Although concerns were expressed that the development of isopsoralens as photochemotherapeutic agents may increase carcinogenic risk because monoadducts are subject to error-prone DNA repair mechanisms,⁴⁵⁰ this does not appear to be the case.⁴⁵¹

The use of UV filters decreases the photomutagenic and photocarcinogenic effects of psoralens, but the employment of 5-MOP in sun tanning lotions remains controversial.^452,453 Given the advantages of a PUVA tan, it is likely that the benefits will outweigh risks, provided that UV exposure is carefully controlled. This may only occur under clinical supervision.

Use of PUVA therapy in psoriasis is postulated to be associated with an increased risk of skin cancer, particularly squamous cell carcinoma.⁴⁵⁴ Risk of developing skin cancer was claimed to be proportional to the number of treatments and the degree of exposure to UVA.⁴⁵⁴ However, these associations have been disputed by at least one study, which proposed that prior exposure to arsenic treatments or X-rays largely explained the higher incidence of skin cancer in PUVA-treated psoriasis patients.⁴⁵⁵ Significant exposure of the unprotected human eyes to PUVA may accelerate cataract formation, but this is readily alleviated by protective eyeglasses.⁴⁵⁴

Interest in the oestrogenic effects of plants first arose in the scientific world when clover disease was identified in Australia in the 1940s. It was observed that infertility developed in sheep after grazing on various species of clover (Trifolium species). Research interest at the time focused on understanding the factors that caused clover disease, which were identified as isoflavone glycosides. It was several decades later, when results from epidemiological studies suggested that the dietary consumption of soya products might have a protective role in breast cancer, that interest in the oestrogenic properties of isoflavones was rekindled.⁴⁵⁶

Early studies using various animal tissues demonstrated that, while isoflavones compete strongly with oestradiol at oestrogen receptors, their stimulation of these receptors is much weaker than oestradiol.⁴⁵⁷ In other words, they were thought to act as partial oestrogen agonists that could function as oestrogen agonists or antagonists depending on the hormonal milieu. The theory was that in a high-oestrogen environment (such as in the premenopausal woman), their displacement of endogenous oestrogens is postulated to have an antioestrogenic effect. In contrast, in a lower oestrogen environment, as in the postmenopausal woman, they were expected to provide a net oestrogenic effect. While this theory still has some relevance, the situation is now known to be more complex.

There are two isoforms of the oestrogen receptor, ERalpha and ERbeta. Their distribution and density varies, depending on the target tissue. The existence of an oestrogen receptor was first reported in the early 1960s. In 1996, an additional ER was discovered. This receptor was designated ERbeta and consequently the originally discovered receptor was renamed ERalpha.⁴⁵⁸ Both ERalpha and ERbeta may coexist in a tissue and relevant proportions often vary. ERalpha is the dominant receptor in the adult uterus. ERbeta is expressed in high levels in prostate, salivary glands, testis, ovary, vascular endothelium, bone, smooth muscle, certain neurons in the central and peripheral nervous system and the immune system.⁴⁵⁹ Both isoforms are present in the female breast and reproductive tissue.⁴⁶⁰

Each of the ERs may influence the function of the other. The effects of substances that interact with receptors when both forms are present are complex.⁴⁶⁰ It is hypothesised that ERbeta may modulate, or even antagonise, the actions of ERalpha.⁴⁶¹ For example, in the breast ERalpha and ERbeta exhibit opposing functions in cell proliferation: ERalpha promotes epithelial proliferation, whereas ERbeta has a restraining influence.⁴⁶² Other studies suggest several possible interactions for the two receptors: antagonistic, synergistic and sequential.⁴⁶³

The female sex hormones (the oestrogens) consist of 17beta-oestradiol, oestrone and oestriol. Oestradiol binds to ERalpha and ERbeta with equal affinity.⁴⁶⁴ The isoflavone phyto-oestrogen genistein has greater binding affinity for ERbeta than for ERalpha. This preferential binding of genistein to ERbeta indicates such isoflavones probably produce pharmacological and clinical effects quite distinct from oestrogens. In particular, they tend to exhibit antiproliferative activity.⁴⁶⁴ Other isoflavone phyto-oestrogens also bind more strongly to ERbeta.⁴⁶⁵

Furthermore, isoflavones bind to ERs with a much lower affinity than oestrogens and therefore produce less potent responses.^466,467 However, they initiate greater gene transcription of ERalpha compared to ERbeta.⁴⁶⁷

As noted above, isoflavones and other phyto-oestrogens are thought to compete with oestradiol for binding and activation of ERs,⁴⁶⁰ potentially decreasing the effect of oestradiol in vivo in some circumstances. However, there have been some in vitro results to the contrary.⁴⁶⁸ The effect depends on the doses (of both phyto-oestrogens and oestradiol). In a pilot study, a soy protein isolate containing isoflavones (120 mg/day) taken for 6 months did not prevent oestradiol-induced endometrial hyperplasia in postmenopausal women.⁴⁶⁹ (But interestingly, there was also no additive effect.)

Research on selective oestrogen receptor modulators (SERMs) such as tamoxifen has provided more insights into how non-steroidal molecules can interact with the oestrogen receptor. A summary of the knowledge thus far is that once a SERM binds to the ER it causes a change in its shape. This allows recruitment of co-activators if it is destined to elicit an oestrogenic response, or co-repressors if its response is anti-oestrogenic. The binding of the coregulatory molecules leads to the activation of the promoter sequence of the oestrogenic responsive gene. Isoflavones have been proposed as natural SERMs.⁴⁶⁵

Studies have found that isoflavones have both agonistic and antagonistic effects, though they are strong ERbeta agonists and weak ERalpha agonists. The presence of a correctly positioned phenolic ring and also the distance between the two opposing phenolic oxygens in the isoflavones structure is similar to that of 17-beta-oestradiol. This similarity allows the isoflavones to bind to the ER, effectively displacing 17-beta-oestradiol. This action may help explain how phyto-oestrogens help protect against breast cancer, because ERbeta inhibits mammary cell growth as well as the stimulatory effects of ERalpha.⁴⁶⁵ As with the SERMs, studies have shown that the recruitment of coregulatory molecules may be important in determining the function of phyto-oestrogens. In particular, isoflavones appear to trigger selectively ERbeta transcriptional pathways, especially transcriptional repression.⁴⁶⁵

The in vitro and in vivo studies on phyto-oestrogens can be divided into three general categories: chemoprevention, treatment effects and lifetime exposure studies (in vivo only).⁴⁶⁵ This literature is so vast that a thorough review is beyond the scope of this primer. However, one aspect worth emphasising is that isoflavones have been shown to prevent cancer or inhibit cancer cell lines in a variety of models.^470,471 Another is that isoflavones exert favourable effects in vivo and in vitro on parameters relevant to cardiovascular risk including insulin resistance, lipid peroxidation, haemostasis and endothelial function.⁴⁷² Probably of greater relevance are the many human studies of the impact of phyto-oestrogens on various health outcomes or parameters. Some of the key findings of these studies are featured below, in a brief outline of the reviews.

A meta-analysis concluded that soya isoflavone consumption was associated with a reduced risk of breast cancer, but this protective effect has only been observed in studies in Asian populations.⁴⁷³ Soya isoflavone intake was also inversely associated with the risk of breast cancer recurrence. Moderate consumption of isoflavones in diet does not increase the risk of breast cancer recurrence in Western women who have survived breast cancer, and Asian breast cancer survivors exhibit better prognosis if they continue consuming a soya diet.⁴⁷⁴

Results from Asian epidemiological studies suggest a beneficial impact of isoflavones on bone health.⁴⁷⁵ However, clinical trials have yielded conflicting results on bone mineral density and turnover markers that may reflect on differences in study parameters, type and dose of phyto-oestrogen used and the variable metabolism of isoflavones, especially in terms of equol production.⁴⁷⁵

A systematic review and meta-analysis of soya isoflavones versus placebo in the treatment of menopausal vasomotor symptoms concluded that there was a significant tendency in favour of soya.⁴⁷⁶ Another meta-analysis found that consumption of 30 mg/day of soya isoflavones (or at least 15 mg genistein) reduced menopausal hot flushes by up to 50%.⁴⁷⁷ Individual responses in menopause could be determined by bioavailability and metabolism, especially to equol.⁴⁷⁸

The published effects of isoflavones on the circulating hormones in pre and postmenopausal women were the subject of a systematic review and meta-analysis.⁴⁷⁹ In all, 47 studies were included. In premenopausal women, meta-analysis suggested that isoflavone consumption did not affect levels of oestradiol, oestrone or sex hormone binding globulin (SHBG), but did significantly reduce FSH and luteinising hormone (LH) (by about 20%). In postmenopausal women there was no impact of isoflavones on any of these hormones, although there was a 14% non-significant increase in oestradiol.

Epidemiological studies indicate that soya isoflavone consumption is possibly protective against prostate cancer.⁴⁸⁰ There is also some suggestion from clinical trials that isoflavones can slow the disease progression.⁴⁸¹

Isoflavone intake appears to confer cardiovascular benefits that may or may not be related to their phyto-oestrogenic properties. One systematic review identified improvement in arterial stiffness⁴⁸² and another used meta-analysis to establish a small reduction in blood pressure.⁴⁸³ Soya isoflavone intake improved flow-mediated dilation (an indicator of cardiovascular health) based on a meta-analysis of randomised, controlled clinical trials.⁴⁸⁴

Intestinal metabolism of isoflavones to their aglycone form is crucial for ensuring bioavailability and therapeutic activity. In other words, the intestinal microflora is pivotal in the metabolism of oestrogenic isoflavones. Individual differences in intestinal microflora have been proposed as a possible reason why there is some inconsistency in the clinical effects of isoflavones.⁴⁸⁵ In particular, daidzein is further metabolised by gut microflora to produce equol, a more active oestrogenic compound. There are large individual variances in the capacity to produce equol and hence possibly a therapeutic effect. Dietary fat consumption is known to reduce this capacity.⁴⁸⁶ On the other hand, dietary supplementation with fructo- oligosaccharides such as inulin is known to increase equol.⁴⁸⁷

In contrast to the isoflavones, research on the oestrogenic lignans enterolactone and enterodiol has not been as extensive. Initial interest in these compounds resulted from the observation in the early 1980s that their urinary levels in menstruating women exhibited a cyclic pattern during the menstrual cycle, with maximum excretion in the luteal phase.⁴⁸⁸ The relatively high concentrations of these new lignans in urine, their cyclic pattern of excretion and their increased excretion in early pregnancy suggested that they were a new class of human hormone. It transpired what had been found was a plant chemical modified by bacteria in the human digestive tract.⁴⁸⁸ Selective antibiotic administration to humans suppressed oestrogenic lignan formation.⁴⁸⁸

Enterolactone, like isoflavones, inhibits oestradiol-stimulated breast cancer cell growth in vitro.⁴⁸⁹ Lignan ingestion has been associated with increased concentrations of SHBG, but this was not confirmed in clinical studies.^489–491 Linseed intake by normal premenopausal women was consistently associated with longer luteal phase lengths and higher ratios of luteal phase progesterone to oestradiol.⁴⁸⁹ These findings may reflect favourably on breast cancer risk. An earlier animal study found that linseed supplementation reduced early risk markers for breast cancer.⁴⁹² Dietary studies and assays of urinary lignans in postmenopausal women have found that lignan excretion is significantly lower in the urine of women with breast cancer,⁴⁹³ although the situation is not as clear from recent prospective cohort studies.⁴⁹⁴

In more recent times, the health effects of linseed and SLDG supplementation have been investigated in a number of clinical trials. Results of trials involving whole linseed are confounded by its content of mucilage and fixed oil (rich in omega-3 fatty acids), especially the trials investigating its impact on blood lipids or glucose.

The following is obviously not a comprehensive review of the clinical data for linseed and its lignans and focuses on hormonal effects. Linseed (25 g/day) altered oestrogen metabolism more than soya (25 g/day), and significantly shifted oestrogen metabolites to less biologically active forms (2-hydroxyoestrone) in postmenopausal women.⁴⁹⁵ This was confirmed in other clinical trials at 10 g/d, in pre- and postmenopausal women.^496,497 At 10 g/day linseed decreased endogenous oestrogen levels and increased prolactin in postmenopausal women.⁴⁹¹ Linseed supplementation does not appear to benefit menopausal symptoms, although more studies are necessary.⁴⁹⁸

Epidemiological studies have not found adverse effects from the dietary consumption of isoflavones or lignans. However, when quantities well above dietary exposure are used in a therapeutic context, adverse events might well ensue. As noted above, the phyto-oestrogens are partial oestrogen agonists similar to the drug tamoxifen. Research has shown that while tamoxifen has potential for preventing breast cancer and cardiovascular disease, its use increases the risk of developing endometrial cancer. Moreover, concurrent intake of phyto-oestrogens and tamoxifen may reduce the therapeutic effects of the drug in breast cancer. For this reason, and because of the observation that phyto-oestrogens can stimulate the growth of oestrogen-dependent tumours in some circumstances, intake of these phytochemicals should be limited to dietary levels in women with oestrogen-sensitive breast cancers until clinical studies suggest otherwise. Such data are beginning to emerge (see the discussion on breast cancer risk below).

Controversy has arisen over the possible adverse effects of soya-based infant milk formulas. One earlier study found that infant exposure to isoflavones from these products was relatively much greater than levels shown to alter reproductive hormones in adults.⁴⁹⁹ It was suggested that further studies of possible developmental effects are highly desirable.

On the other hand, studies in males have found no indication of adverse effects. For example, a 2010 review found no evidence from rodent studies and nine clinical trials that isoflavone intake impacted oestrogen levels in males. Clinical evidence also indicates no impact on sperm and semen parameters in men.⁵⁰⁰ In addition, a meta-analysis that included 15 placebo-controlled clinical trials found that neither soya nor isoflavones altered testosterone parameters in men.⁵⁰¹

While the epidemiological evidence suggests that a lifetime of moderate to high isoflavone intake is protective against breast cancer incidence in women, the impact of such intake commenced later in life is uncertain. Concerns have been raised that a later intake of high levels of phyto-oestrogens might in fact increase the risk of oestrogen-associated cancers. However, these fears are not supported by the available data.

Mammographic density may be a predictor of breast cancer risk. Women with a high percentage of dense tissue are at 4-fold greater risk of breast cancer. Mammographic parenchymal patterns assess the variation between radiological appearances: fat appears dark and epithelium and stroma appear light. Percentage density is the ratio of dense area to breast area. A meta-analysis of randomised, controlled clinical trials on soya and red clover (Trifolium pratense) isoflavones found no impact on mammographic breast density in postmenopausal women.⁵⁰² However, isoflavones may cause a small increase in breast density in premenopausal women, which is of uncertain clinical relevance. An epidemiological study in 3315 Chinese women in Singapore found no significant impact of soya intake on breast density, with the suggestion of a possible reduction.⁵⁰³

The results of a case-control study involving nearly 800 participants conducted in New Jersey and published in 2009 suggest a reduction in endometrial cancer risk with intake of foods containing isoflavones (but not lignans) in lean women. Four case-control studies prior to this reported conflicting results for the role of dietary phyto-oestrogens in endometrial cancer risk, although there was a suggestion of a reduced risk for soya foods.⁵⁰⁴

A Cochrane meta-analysis of randomised, controlled, clinical trials published to March 2007 investigating relief of menopausal symptoms found no evidence that food or supplements containing phyto-oestrogens caused oestrogenic stimulation of the endometrium when used for up to 2 years. Trials of women who had breast cancer or a history of breast cancer were excluded.⁵⁰⁵ The dosage of isoflavones was 50 to 120 mg/day, with many of the soya products containing soya protein.

One concern is that isoflavones may adversely affect thyroid function and interfere with the absorption of synthetic thyroid hormone. A 2006 review evaluated the relevant literature describing the effects of soya on thyroid function.⁵⁰⁶ In total, 14 trials (thyroid function was not the primary health outcome in any trial) were identified where the effects of soya foods or isoflavones on at least one measure of thyroid function was assessed in presumably healthy people. Eight involved women only, four involved men, and two both men and women. With only one exception, no effects or only very modest changes, were noted in these trials. Collectively, the findings provide little evidence that soya foods or isoflavones adversely affect thyroid function in euthyroid, iodine-replete individuals. In contrast, some evidence suggests that soya, by inhibiting absorption, may increase the dose of thyroid hormone required by hypothyroid patients. In addition, there remains a theoretical concern based on in vitro and animal data that isoflavone intake by individuals with compromised thyroid function and/or marginal iodine intake may increase their risk of developing clinical hypothyroidism.

Because they are a vast and diverse group of archetypal plant constituents, it is consequently difficult to arrive at a consistent definition of alkaloids. The word derives from the term ‘vegetable alkali’, referring to the alkaline nature of these compounds, a property that results from the presence of nitrogen. The following definition is concise and descriptive: alkaloids are alkaline (basic) nitrogen-containing heterocyclic compounds derived from higher plants often exhibiting marked pharmacological activity. However, there are several exceptions to this definition. For example, the nitrogen in ephedrine is not part of a ring, so it is not heterocyclic. For this reason, it is sometimes referred to as a protoalkaloid. Berberine is not alkaline and some alkaloids are derived from animals and micro-organisms. Generally, alkaloids are white, but some are highly coloured; for example, berberine is yellow and sanguinarine is red. (By now it should be apparent that names of alkaloids end with the letters –ine.)

Alkaloids were the first chemical drugs to be derived from plants; a mixture of morphine and narcotine was isolated from opium in 1803. They have maintained an important role in conventional drug therapy since then. Names such as codeine, morphine, atropine, quinine, pilocarpine, theophylline, colchicine, pseudoephedrine and vincristine are a familiar part of modern drug medicine. These drugs are potent pharmacological agents and their properties are well described in conventional pharmacology texts. A high risk of adverse reactions ensues from their use. In contrast, while phytotherapists do rely on plants containing alkaloids, their activity is at the milder end of the pharmacological spectrum. Alkaloids as a group are not nearly as important to the phytotherapist as they are to the conventional doctor.

Accordingly, this section will not include a broad and detailed treatment of various classes and subclasses of alkaloids, but will instead focus on a few examples important to phytotherapy. Nevertheless, it is worthwhile to consider the basic structures of the important classes of alkaloids. These are provided in the following diagrams.

Examples of each class named are: pyrrolidine – hygrine; piperidine – lobeline; pyridine – nicotine; indole – strychnine; quinoline – quinine; isoquinoline – morphine; pyrrolizidine – seneciphylline; tropane – hyoscyamine; purine (xanthine) – caffeine; imidazole – pilocarpine; quinolizidine (norlupinane) – sparteine.

Most alkaloids are synthesised by the plant from amino acids. If they are not, they are called pseudoalkaloids. Most of the known examples of pseudoalkaloids are isoprenoids and are referred to as terpenoid alkaloids. Aconitine from various species of aconite is an example of a diterpene alkaloid and is one of the most toxic substances known. Steroidal alkaloids are also found in plants. Some are combined as glycosides, for example solanine from potato shoots.

Despite being the most important archetypal plant constituents from a conventional perspective, the function of alkaloids in plants is only beginning to be understood. They probably have a defensive or deterrent role, but other theories are still suggested in texts, including they are a by-product of primary metabolism.

Alkaloids have two key properties that determine much of their pharmacology: an ability to cross the blood-brain barrier and exert depressant or stimulant effects on the central nervous system (CNS), and an ability to interact with various neurotransmitter receptors. Examples include CNS depressants such as morphine and codeine, CNS stimulants such as caffeine and cocaine and sympathetic nervous system stimulants such as ephedrine.

Alkaloids are important to phytotherapy, but their role is less significant than other archetypal plant constituents, as demonstrated by the fact that only three alkaloid-containing herbs are covered by monographs in this text, namely Chelidonium, Berberis and Hydrastis. The research on berberine is now extensive and is comprehensively reviewed in the Berberis and Hydrastis monograph. Some additional examples of important alkaloid-containing herbs and their pharmacodynamic properties are discussed below.

Lobelia (Lobelia inflata) and ipecacuanha root (Cephaelis ipecacuanha) contain emetic alkaloids (lobeline and emetine, respectively) that act as reflex expectorants at sub-emetic doses (similar to the expectorant saponins). Oral preparations of lobelia are also used as an aid to stop smoking, since lobeline is very similar to nicotine in its pharmacodynamic actions.

Several alkaloid-containing herbs are used as mild analgesics and anxiolytics. California poppy (Eschscholtzia californica) is a traditional medicine used by the rural population of California for its analgesic and sedative properties. Animal studies have verified these actions and demonstrated low toxicity.⁵⁰⁷ A proprietary formula consisting of 80% California poppy and 20% Corydalis cava (both herbs are rich in isoquinoline alkaloids) has been the subject of clinical studies.⁵⁰⁸ Results from two clinical trials showed that disturbed sleep behaviour could be normalised by the combination, without evidence of carry-over effects or addiction.⁵⁰⁸ In vitro studies suggest that the two herbs co-operate in establishing an advantageous catecholamine status necessary for maintaining sedative and antidepressant effects.⁵⁰⁹

Various species of Ephedra contain the protoalkaloids ephedrine and pseudoephedrine.⁵¹⁰ Ephedra is a commonly used Chinese herb possessing diaphoretic, antipyretic, antiallergic and antiasthmatic properties. It is also favoured in phytotherapy, particularly for the latter two activities. Ephedrine has been used as a conventional treatment for asthma.

Broom tops (Cytisus scoparius) contains the alkaloid sparteine and is an important herb for the treatment of cardiac arrhythmias.⁵¹¹ Sparteine acts as a potassium channel antagonist⁵¹² and delays systolic depolarisation. Normal sinus frequency is slightly reduced, the refractory period is prolonged and the threshold is raised, reducing the risk of fibrillation and extrasystoles. Indications for broom tops include sinus tachycardia, ventricular extrasystoles and arrhythmias following a heart attack.⁵¹¹ High doses of the herb are necessary to achieve therapeutic levels of sparteine and related alkaloids. Sparteine was once widely used as an oxytocic drug, but it fell out of favour because of the uterine spasm that occurred in women who were unable to metabolise it effectively. About 5% of male and female volunteers studied were unable to metabolise sparteine by N-oxidation⁵¹³ and this defect appeared to have a genetic basis.⁵¹⁴