Written by abjectly Simon Mills
This is a longer thought piece that reflects the many levels at which this question can be addressed. However it is still a summary review and we have not included detailed literature references. The foundations for these arguments were thoroughly researched in our seminal text. This has stood up well over the years and is recommended as a basis for any detailed personal updating.
http://cumbria-cottages.co.uk/catalog/Adminhtml_category In this review we conclude that herbs are not a public hazard, and that humans have always been very good at avoiding plants that upset them. However we have never been as sure of the effects of long-term exposure, and there are other reasons to be aware of what can go wrong. A few sensible precautions are warranted.
 Mills S & Bone K (2005) The Essential Guide to Herbal Safety. Elsevier/Churchill Livingstone, Edinburgh.
Theory versus practice
If we read all the safety threats in plants we might be too frightened to eat. Wheat, rye, and barley contain a protein called gluten that is hydrolysed in the digestive system to a peptide alpha-gliadin, a well-established intestinal irritant, that has caused many thousands of deaths around the world through coeliac disease and sprue. Apple seeds, and the kernels of apricots, plums and other stone fruits, as well as bitter almonds, contain glycosides that generate cyanide in the digestive system. Potatoes are members of the deadly nightshade family and produce the same poisonous alkaloids when they turn green under the influence of light. Many common household pulses, like soya bean, red kidney bean, and haricot bean (used in “baked beans”), contain various toxins, notably types of lectins called phytohaemagglutinins, as well as trypsin-inhibitors, that can only reliably be neutralised by boiling for at least 30 minutes.
Of course the foods listed above are generally safe to eat. The real point is to show how difficult it is to predict the toxicity of a plant only from the presence of toxic constituents. The action of the whole plant, and the way in which it is normally consumed, counts for more than any individual constituent list. We will see here that for the most widely used plant remedies there are similar reassurances.
Effects without side-effects?
A key reason why people turn to herbs appears to be the assumption that they are free from side-effects, “not like those drugs”.
On the other hand medical opinion is generally that any medicine that is effective must also be a risk, that it is impossible to have effects without side-effects, that if herbs are claimed to be free from side-effects they cannot then be effective at all. Paradoxically this aligns with the idea of the person being seen as a whole: the living body is so complex, so intricately interwoven that any intervention at one site may spark off reactions at other sites.
Yet we restate the principle that herbs are safer than conventional medicines. We refer to the uninterrupted use of the most established remedies by millions of people since prehistory. Even today the major herb formulations sold in health food shops sell in millions of units per year with remarkably few problems. There have also been extremely few cases of malpractice litigation brought against herbal practitioners (a fact reflected in the very low level of professional insurance premiums).
If we are right, how can we be both safe AND effective? These questions may be resolved if we look again at the nature of our remedies and the way they may work.
The remedies we use represent only a tiny proportion of available plant species around the world. It is certain that humans through history moved to using those plants that were quickly effective with minimum adverse effects. Short-term safety at least is built in to these millions of human experiences through the ages.
Further the benign qualities of these herbs arise from their complexity. The existence of tannins, mucilages, saponins, or other constituents in different ways buffer or modulate the effect of more active constituents, which are often in any case present in only low levels. It is after all the essential feature of herbal pharmacology that the whole package has potential that no isolated principle can mimic.
Most importantly the whole thrust of treatment with plants is different from that of conventional medicine. The herbs are used ideally to nudge healing responses in the body rather than directly attack the symptoms or pathology. A bitter stimulates digestive activity, a hot spice the digestion and circulation as well; laxatives provoke bile, bowel and urinary elimination; expectorants, diuretics, hepatics and lymphatics produce their own detoxifications, and so on. We trust the body to look after its affairs and seek only to help it on its way without disturbing it unduly.
If that is after all our claim, what is the evidence that our remedies are inherently safe to use?
How do we know?
There are two potential sources of safety information about any medicine:
- toxicological research, usually from animal studies;
- reported adverse effects from actual human use :
- published clinical reports
- post-marketing surveillance of herbal medicines.
Animal studies are still widely used to assess the safety of conventional medicines in spite of commitments to find reliable alternatives. In various parts of the world and over many decades herbal remedies have also been screened in this way and the evidence is available in the scientific literature. This generally shows that herbs are relatively safe, especially on oral consumption, with any adverse effects mostly tied to very high doses and injections.
Medicine manufacturers are generally required to pass on adverse effects reported to them and in some European countries doctors also have to report adverse drug reactions (ADRs) to the authorities. These post-marketing surveillance (pharmacovigilance) schemes also include herbs where they are categorised as medicines.
Most countries of the world pass their ADRs (as well as miscellaneous poisoning reports) to the WHO Uppsala Monitoring Centre in Sweden. In 2006 their director Ralph Edwards reported that they had received 3.6 million ADR reports from around the world. Of these over 41,000 (1.1%) had listed a herb and just over 17,000 (0.5%) had listed one as ‘suspected’. The reported ADRs most often associated with herbs were generally minor and transient symptoms: itching skin, urticaria and rash, nausea, vomiting, abdominal pain and diarrhoea and dyspnoea. “Very rare” serious symptoms included facial oedema, angioedema and anaphylactic shock, hepatitis, purpura, thrombocytopenia, Stevens Johnson syndrome, convulsions, erythema multiforme and circulatory failure. Most herbal ADRs per head of population arise from France and Germany where herbal medicine is part of the mandatory reporting system and where the true level of herbal ADRs relative to those of synthetic drugs might be clearer; yet even here the levels of reports are very low.
Where local research has been conducted, including in Asian countries where herb use is more widespread, a proportion of 1% of all adverse medicine reports keeps recurring. Even allowing for the relative inefficiency of ADR schemes this figure is reassuring. Another way to see the relative hazard involved with herbal and other natural supplements is to look at overall deaths involved compared to other causes. The following chart compiles mortality data from all 27 European member states (view the full size chart here).
When we look more closely at the adverse effects that have been published we see that they can fit into five main categories.
 Ralph Edwards, Director, WHO Collaborating Centre, Uppsala. Reporting to Pharmacovigilance of herbal medicines: current state and future direction London 24-26 April 2006.
Adverse effects I: Poisonous plants
Some plants medicines are dangerous when used incorrectly. A few were used as arrow poisons in days when much food was sourced by hunting. One of these, foxglove, is the origin of digoxin, a very widely prescribed medicine for a failing heart. Others like mandrake, bryony, aconite have been used in carefully prepared dosages in traditional medicine as well. Opium has been the most widely used pain relief since antiquity but is well known to cause problems in long term use and high doses.
These are plants whose effects are more like conventional prescriptions: with both positive effects, and negative side effects or adverse effects. These often contain strong phytochemicals called alkaloids which can have activity on the nerves, heart and other susceptible tissues. Other examples include the alkaloids of tobacco, lobelia, nux vomica, henbane, belladonna and stramonium.
Powerful toxic or near toxic activity from plants include blocking parasympathetic (or cholinergic) activity (eg the nightshades), damaging the liver (eg. pyrrolizidines alkaloids found in ragworts, comfrey, borage and coltsfoot – see also section on long-term safety), the kidney (eg aristolochia, amaranthus, colchicine), and increasing sensitivity to light (eg. excessive doses of St Johns’ wort).
There are a few plants restricted for sale to the public that herbal practitioners are permitted to use with careful dosing and application. These include poke root, lobelia, mistletoe, greater celandine, lily-of-the-valley, and ephedra. There are no reported adverse reports following such professional use.
The main safety concern with toxic plants is accidental consumption through foraging or by children, of plants such as hemlock, hogweed, deadly nightshade, arum lily, oleander, delphinium, bryony, laburnum, mandrake, mistletoe and columbine. National poisons units hold information about treatment and antidotes if such exposure has occurred.
Adverse effects II: Poor quality products
Inferior products are the most common causes of reported herbal adverse events.
Unfortunately there have always been perverse incentives for providers of medicinal products to doctor their wares. A classic and ancient example is ‘wetting’: a grower or gatherer paid by weight will be tempted not to dry the plant completely as even a couple of extra percentage points of water can substantially increase its weight. Unfortunately any dampness in stored ‘drugs’ (from the old European word drogen meaning ‘dried plant’) can encourage mould and associated toxins. Finding creative ways to make up orders or prices were to be expected. Harsh penalties for providing poor quality medicines were recorded as far back as ancient Egypt and throughout the mediaeval world, and as Jacob Bell, the founder the Royal Pharmaceutical Society of Great Britain, noted in 1841: “In reviewing the history of pharmacy in all ages we find that fraud has always prevailed to a remarkable extent”.
In response to such experiences, the supply of conventional medicines around the world is tightly controlled by ‘pharmaceutical standards’ of manufacture, involving detailed quality control and associated paper trails, with independent inspections, sometimes unannounced, as the ultimate sanction. In a few jurisdictions the supply of herbal remedies is as tightly controlled (‘herbal medicinal products’ in Europe and ‘therapeutic goods’ in Australia).
Medicinal standards are published in official national reference texts called ‘pharmacopoeia’. These set agreed definitions and analytical standards required to meet pharmaceutical standards. The early pharmacopoeia were largely devoted to plant drugs (Valeriana officinalis is one of many botanical plant names that reminds us that the plant was ‘official’ in the pharmacopoeia). Many herbs are still listed in pharmacopoeia (eg the European Pharmacopoeia, British Pharmacopoeia and those of China, India and elsewhere). There are also specialist references, the British Herbal Pharmacopoeia and the American Herbal Pharmacopoeia that provide rigorous monographs not covered in the official texts. Manufacturers may choose to use these standards to assure the quality of their products. There are some voluntary assurance schemes aim to apply similar standards, eg, the British Herbal Medicine Association’s HerbMark, that is rolling out for herbal practitioner suppliers.
Most herbal remedies on the market are supplied as food supplements. These are required to meet local food hygiene laws, have accurate labelling, recommended consumption levels, and appropriate allergen and information regarding safe use (eg in children and pregnancy). In the UK enforcement is provided by local trading standards offices and inaccurate information relating to allergens and other known risks renders the supplier liable to other legal measures. These are safeguards when purchasing herbal supplements from responsible manufacturers.
However the buyer needs to beware of less reliable sources.
Compromised quality in herbal products can come about through contamination, adulteration or substitution. Penugonda Contamination is the (often unintended) presence of other materials, including soil and related matter, soil-based heavy metals (some herbs, eg St John’s wort, bacopa, and gotu kola naturally accumulate these and need closer testing before entering the market), infective organisms or pesticides, all these most often due to loose quality controls. Testing for these is a key part of normal herb manufacturer quality control; of course sourcing herbs with organic certification is one way to reduce the pervasive contamination of all the food chain with pesticides. Recently specific controls have been widely instituted to reduce contamination of herb crops with ragworts containing high levels of hepatoxic pyrrolizidine alkaloids. Karor Adulteration is the intended addition of different materials to enhance appearance or intended effect: this can include pigments, isolated phytochemicals like polyphenols, or in modern times even synthetic drugs like steroids. In some early traditions heavy metals were added to plants to increase potency and this is still found occasionally. Substitution is when the plant is not what is labeled, either in error, due to linguistic confusion, or to reduce cost or when there is limited availability.
A notable example of the latter is the Chinese herb Aristolochia fangchi which has long been a commercial substitute for Stephania tetrandra (mandarin name: hang fang ji). Unfortunately its constituent aristolochic acid has been associated with serious kidney damage in Europe, notably when the substituted ‘stephania’ was administered by doctors as part of a conventional intensive dieting regime in Belgium, with many suffering cancer and terminal kidney failure. This single episode in the early 1990s led to a major review of herbal quality in the UK and other countries. A similar occurrence followed widespread substitution of skullcap by a species of germander (Teucrium chamaedrys) and associated cases of liver toxicity reported in France.
Another substitution example relates to black cohosh, which has been identified as linked to a cluster of severe liver complications; however suspect products were found to contain a different species entirely. Unfortunately adulteration of this valuable plant is common. In general the more expensive or in-demand the herb, the more likely it is that it will be substituted. Golden seal and ginseng are other classic examples.
Ginkgo leaf is often adulterated with added polyphenols and there are reports of adding pigments to sometimes substituted products marketed as elderberry and bilberry.
The best protection against consuming contaminated, adulterated or substituted herbs (and where there is no external quality controls) is to buy only from reliable and well-established producers, with transparent quality control measures, and reputations to protect. Avoid buying from suppliers you do not recognise and particularly so if these are marketed over the internet. In general if the product is surprisingly cheap it should be avoided.
Adverse effects III: Unpredictable reactions
There are three types of adverse reactions that cannot be predicted, but which can occasionally have serious consequences.
Intolerance is the least severe and is an exaggeration of the activity of the substance in a few individuals. It is more likely to occur with foods than herbal remedies. With foods it is sometimes erroneously referred to as ‘food allergy’: however this is a much less common immunological reaction (see below)
Allergic/hypersensitivity reactions are dramatic over-responses of one part of the immune system associated with surface defences (perhaps originally targeted against parasites). They involve activation of a type of white blood cell (the mast cell) to release histamine and other inflammatory chemicals, and also the development of immunoglobulin (Ig) E antibodies. Hayfever (in response to pollen) is the most common, but true allergies can be to almost anything. In the case of food substances such as gluten or lactoglobulin from milk they are often ‘masked’ so that the direct cause-effect connection can be difficult to make. Dangerous frank hypersensitivity reactions are increasingly recorded to peanuts, shrimps and other crustaceans. Among conventional drugs such reactions are most likely with NSAIDs, angiotensin-converting enzyme (ACE) inhibitors and antibiotics. Very rarely these can include herbal remedies, with members of the daisy family (Compositae, including chamomile and echinacea) and mugworts among those most encountered. By definition the reaction once established is almost instantaneous.
Confusingly the term “allergic” is sometimes used to describe immune-related idiosyncratic reactions, when toxicity appears after several asymptomatic administrations of the compound (see below).
Idiosyncratic drug reactions (IDRs) are rare, unpredictable reactions to a substance that are not dose related and usually are delayed (although rapid on re-exposure). They include complex immune responses and are caused by the combination of a reactive metabolite of the substance with cellular proteins, and the production of DNA adducts. Symptoms can be severe and include anaphylaxis, fever, skin reactions, arthritis and lupus syndrome. Blood tests often show increased levels of eosinophil white blood cells. A notable type of IDR involves the liver and can closely mimic viral hepatitis.
IDRs account for most of the rare severe herbal adverse reactions. The observed liver problems that were associated with taking standardized kava-kava were probably idiosyncratic (Our safety book included a detailed case-by-case analysis of these reports and showed in forensic detail how misleading most were, concluding that the few clear cases were idiosyncratic rather than toxic, and incidentally giving a clean bill of health to traditional water extracts). Some conventional drugs, eg antiepileptics and antidepressants, are more likely to induce IDRs and there seem some chemical types more likely to do the same in plants: kava contains one such group, lactones, and they occur also in the germander responsible for the French cluster of liver disease reported earlier.
IDRs are rare and unpredictable and generally do not suggest inherent toxicity. However, this has not stopped prohibition of kava across UK and Europe and risks to the availability of black cohosh (even with adverse effects linked to an adulterant – see above). If there is a cluster of IDRs (as can happen for several reasons) then the fate of any herb involved can be threatened.
Adverse effects IV: Interactions
There is always the possibility that two ingested substances will interact in their metabolism within the body. This is particularly likely in the liver: here there are many complex mechanisms for processing any molecules that are not core food materials. Notable among these are cytochrome P-450 (CYP450) enzymes, a hierarchy of 14 enzyme families in the liver and most organs and tissues of the body. They are almost as adaptable as the body’s immune system and are designed to be able to metabolise an almost infinite range of molecules that might be taken in from the environment.
Such foreign molecules (sometimes referred to as xenobiotics) include medicines, and also most plant chemicals. Their metabolism includes engagement in complex dynamics with the hundreds of enzymes involved. Each CYP450 substrate can also inhibit the enzyme acting upon it, or by contrast can induce it, so that its effect on other molecules is changed.
One classic example is grapefruit. This is known to inhibit two enzymes: CYP2A6 and CYP3A4. The latter breaks down a wide range of medicines, including blood thinners, anti-cholesterol, heart and blood pressure medicines, anti-depressants, antihistamines, viagra and contraceptives, and notably anti-HIV and immunosuppressive drugs. In other words taking grapefruit at the same time as these medicines will reduce their breakdown rate and increase their activity in the body.
By contrast St John’s wort induces CYP3A4, so increasing the breakdown rate of the same medicines above. This means their effects may be reduced and St John’s wort remains the herb that doctors most worry about for its unintended potential for compromising the activity of sometimes critical medicines.
Low levels of interactions of these types are common between medicines and foods. Simple foods like garlic, brassicas, red wine and many common culinary spices can all interact with CYP450 enzymes and affect the availability of prescription medicines. However the key point is that most of these are 1) theoretical interactions (mostly observed in the test tube) and 2) with very little reported evidence of impact in the real world.
Additional scope for interactions lie with the physical properties of some plant constituents. Plant tannins, mucilages, saponins, and resins can theoretically slow absorption from the gut; hot spices may increase it. However again the real impact is doubtful and in any case food is more likely to be concern than herbal remedies.
Serious herb-drug interactions are probably rare. This is the view of all those in the regulatory field who have watched the surveillance data closely over many years. However many modern medicines are very strong and are used in critical situations. The challenge is that powerful prescription medicines may also have a ‘narrow therapeutic window’: problems can occur if they are either over- or under dosed. If one is on a such a regime it is important to be careful to monitor the effects of taking other medicines, herbs and foods.
Adverse effects V: Transient and nocebo effects
Most reported reactions to herbs are not a threat to health, nor long term. Minor reactions to herbal treatment are quite common, especially at the beginning of a course of treatment.
There is an ancient precedent here. In the past herbal treatments were often robust and even unpleasant, to reflect the fact that they were often emergency medicine. Nowadays the emphasis is on supportive herbal treatments but in bringing about healthy changes in the body it is not surprising that there may be associated discomforts.
However the majority of such changes are transient and minor, including diuresis, changes in bowel performance, more productive coughing to clear an obstructed chest, or more sleepiness. It is also possible to see ‘healing crises’: short-term flare-ups in skin or headache symptoms, transitionary increases in menstrual symptoms. Practitioners usually include calming components into their initial herbal blends to reduce the tendency of some conditions to exacerbate but a few people’s constitutions are trigger-happy.
The placebo effect is noted in controlled clinical research trials among subjects taking the sham treatment. Surprisingly, placebo benefits can occur in any proportion of a treatment group, from almost none to most, depending on the condition and circumstances. Distilling much debate on this, placebo can best be seen as a testament to the powerful healing forces that the body has, and how little it sometimes takes to mobilise them. It also suggests that providing support and reassurance can exert real physiological effects on the body.
However placebos can also generate significant levels of adverse effects. This is referred to as the nocebo phenomenon. In one review of placebo-controlled trials involving healthy volunteers, adverse responses were reported in almost a fifth taking placebo, increasing up to a third on repeated dosing. Overall the most frequent adverse reports were headache, drowsiness, and energy loss. In another review of studies of five different medications in the areas of cardiology, neurology/psychiatry, metabolism, and gastroenterology, it was shown that the placebo side effect profile was largely similar to the side effect profile of the active treatment.
It appears that having expectations of adverse effects from taking a medicine is more likely to lead to their happening (doctors should never say “there is nothing more we can do”). It also seems that people with anxieties can more easily generate physical symptoms. If nocebo effects can happen to dummy pills then they will feature after any remedy. Most practitioners will know patients who react badly to almost anything they are given.
Long term effects?
Whereas we can be reasonably confident that short-term (‘acute’) side effects will have been well understood over thousands of years of human experience, we can not be so sure about the possibility of long-term (‘chronic’) consequences. Delayed effects of a substance can be hard to connect if they occur many years later, even with modern databases and record keeping. It is almost unimagineable that such connections would have been possible in the past.
The types of long term problems that might follow exposure to hidden ‘iceberg’ dangers include cancers, chronic inflammatory and auto-immune diseases, and chronic liver and kidney problems. However there is not much evidence that plant ingredients of any sort are linked to such conditions. Most evidence of connection comes from epidemiological observations of the effect of diet, lifestyle and environmental exposures. For example the International Agency for Research on Cancer, part of the World Health Organization lists about 200 known or probable carcinogens, including coal tar, asbestos and processed meat. The only plants that make the list are Aristolochia (see above), betel nut and tobacco. Some individual plant constituents have demonstrated mutagenic or carcinogenic activity in the laboratory or in animals (an example is the isolated compounds estragole and methyleugenol variously found in fennel, nutmeg, basil and tarragon, and coumarin found in Cassia cinnamon) but the case that these are dangerous to humans is contentious and there are only limited measures to restrict access to the plants that contain them. Exposure to some dietary ingredients (like gluten, animal products) has been linked to inflammatory and immunological problems. However the case against dietary plants is very limited.
There is one part of the body relatively vulnerable to untoward impacts of some plants: the liver. As we have seen this is central to the metabolism of non-dietary constituents, including from plants. The very things that can be valuable as medicines can pose extra work for liver metabolism, this compounded by excess alcohol consumption, fatty diets and prescription or other medication. Some of the most serious rare idiosyncratic adverse effects linked to plants, such as to germander, kava, and black cohosh (or substitute), also lead to chronic hepatic damage and consequent health problems in later years. This risk, even if rare, has led to such plants being widely restricted. However the rarity means that the risk of trouble to most plants is considered to be too small to present a public hazard.
However one case demonstrates the need for caution. Pyrrolizidine alkaloids (PAs) are a group of constituents in many plants, including ragworts, comfrey, borage and coltsfoot. Most of their toxicity data comes from animal studies and indicates that some PAs are hepatotoxic, genotoxic and carcinogenic. There are also human case reports of poisonings due to PA-containing plants, including deaths associated with grain crops contaminated with PA-containing weeds. PA poisoning in humans is characterised by acute hepatic veno-occlusive disease. The acute disease is associated with high mortality, and a sub-acute or chronic onset may lead to liver cirrhosis. There is more information on PAs in another of our Insight pieces here.
Because some of the plants involved have been valuable remedies there has been resistance among the herbal community to restrictions as these particular plants have not yet been linked to harm and there is reasonable doubt about the dangers of all PAs. However others have taken the view that the risks and uncertainties do not adequately reassure the public that taking these plants is safe and have voluntarily stopped supplying them. As far back as the early 1980’s the root of comfrey in internal remedies was withdrawn from trade catalogues for this reason (it remains an excellent external healing remedy). Most concluded that this was a case was not worth fighting, given a significant risk to the public, and a mode of action and delay that meant trouble could not be safely ruled out.
Caution I: Pregnancy and breastfeeding
There are always extra cautions in taking any remedies during pregnancy and a professional herbalist’s default position is to back off altogether during the first trimester (while the baby’s organs are forming) unless really necessary. Practitioners will not use strong herbs at all during pregnancy and any listed as risky elsewhere in this article should particularly be avoided. Many reference texts on herbs assume that if a herb has not been proved to be safe in pregnancy it should be avoided, without medical opinion at least. On the other hand many herbs have a good combination of widespread use by pregnant women with reassuring research evidence: our textbook rates many western herbs by tight criteria and finds many a pregnant woman can choose from.
There is evidence that regular consumption of large amounts of licorice through pregnancy might interfere with the birth and even embryonic development. This effect is linked to the hormonal properties of licorice (inhibition of an enzyme that converts cortisol to cortisone). The available evidence is that a safe to take up to 6 g of (or up to 250mg of glycyrrhizin) per week during pregnancy. If there is associated high blood pressure (eg pre-eclampsia) licorice should be avoided altogether.
Although precautionary statements have been made about the use of some herbal remedies during breastfeeding the evidence for risk is minimal. Plants with high levels of volatile oils will pass these easily through breastmilk so nursing mothers should be aware that these may affect their babies. This can actually be helpful, as in the case of reducing colic in infants by the mother consuming kitchen spices (although occasionally the reverse can happen). Some of these have additional reputations for enhancing breast milk production.
Caution II: Don’t overdo it!
In traditional medicine herbs were often taken in ‘heroic’ doses for maximum short-term effect and at much higher levels than are usually applied today. Nowadays doses may range from 0.25g-100g/day, with most in the lower range consistent with their use as supportive and gently nudging remedies. This provides a good margin of safety.
There are some cases where higher doses can present problems. Obviously all the risky herbs mentioned above are very dose sensitive. If you have high blood pressure you should restrict your licorice consumption to a few grams or less a day. You should avoid huge regular doses (more than 30g a day) of some of the spices like fennel with high estragole and methyleugenol content. Be particularly aware of dosing if you are also on prescription medicines with a narrow therapeutic window.
However common sense should prevail and the best advice is to observe recommended doses. Using herbs as part of our everyday diet, like chamomile, ginger and peppermint, or a sprinkle of a spice or herb on our food is, of course, completely safe. When using herbs more ‘medicinally’, in general one should always be looking for an exit plan in starting a herbal programme: the best herbal treatment is one that nudges the body to better performance. However there are some situations where long term use is needed, for example, adaptogenic herbs have incrementally benefit into the long term. But, as we have seen, we cannot be absolutely sure what long term effects might build up and focused use of herbs is always the best option. Do review what your plan is if you are taking the same herbs for more than a year. It is always a good idea to have a break within this time to check that the herbs are still being useful.
Be sensible and all will be well.