Dangers in the lore of the land

In the eyes of the consumer, 'natural' means 'safe' where herbal medicines are concerned. But this attitude is fraught with danger

Long before the pharmaceutical industry had even been imagined, herbal medicines were used to try and cure illnesses. In the absence of the clinically tested drugs we rely on today, folklore and old wives' tales were the basis of the treatments, with plants and their extracts featuring large. Recent years have seen a resurgence in interest in traditional herbal medicines: patients equate 'natural' with 'safe', all too often believing that formally tested medicines are artificial and therefore must be unsafe. This thinking is clearly erroneous - but also potentially dangerous. Far from being safe, many commonly used herbal medicines like St John's Wort and ephedra are actually very potent drugs in their own right, with all the implications this has for side-effects. And, most worryingly, they can have potentially serious interactions with prescribed medicines.

There are many culturally different reasons for the popularity of herbal medicines. The recent rapid rise in their use in the US owes a lot to the surge in 'new age' spiritualism. But in Europe - and in Germany in particular - herbal medicines are a part of life, and are even frequently prescribed by doctors; as many as 70% of German doctors routinely prescribe herbal medicines. And most of these are reimbursed by the German health system.

According to Shiew-Mei Huang of the FDA's office of clinical pharmacology and biopharmaceutics, nearly 20% of Americans are now taking at least one herbal product, a high dose vitamin, or both. Ginkgo, St John's Wort, ginseng, garlic and Echinacea each make up around 10% of the dietary supplement market in the US. And all can interact with prescribed drugs. 'Herbal-drug interactions can give altered responses to approved drugs, whether therapeutic failure or altered response,' she says.

regulatory framework

The US FDA does not regulate herbal products directly; rather, it treats them as foodstuffs with additional requirements. In 1994, the Dietary Supplement and Health Education Act set out a regulatory framework for safety and labelling, adding a requirement that label claims should not be misleading. The US National Center for Complementary and Alternative Medicine at the NIH supports research into clinical studies.

Australia's Therapeutic Goods Agency has an office of complementary medicine, which assesses the safety and quality, but not the efficacy, of herbal products. And while there is still no single system in Europe yet, one is currently being developed. The UK House of Lords Select Committee has expressed concerns that herbal products should be properly regulated and more research carried out. 'There is a clear need for international collaboration to avoid the duplication of research efforts,' says Iain McGilveray of the University of Ottawa in Canada

Indeed, while Germany is the third largest market for medicines in the world, it is the largest for herbal medicines, with a market worth around £2bn (€2.8bn) a year. And many of its mainstream pharma companies, such as Boehringer Ingelheim, have subsidiaries devoted to herbal products.

The European Commission's proposed new rules will control herbal medicines for the first time, much to the alarm of the health food stores such as Holland & Barrett in the UK, which is conducting a campaign against the legislation. The new EC rules are based on those introduced in Germany in the 1970s. Essentially, existing pharma legislation will be extended to cover traditional herbal remedies, with good manufacturing practice procedures analogous to those for pharma manufacturers being essential.

The big difference, however, is the need for formal clinical trials. These are clearly far too expensive for herbal medicine manufacturers, as their products are unlikely to be patentable, hence recouping the cost will be nigh on impossible. Instead, they will be able to use an existing rule that allows a product to be marketed without formal clinical trials as long as the company can show the product has a good safety record over a period of at least 30 years. This will have to include warnings about potential interactions with medicinal products. And practitioners are facing professional self-regulation controls, much in the same way as osteopaths and chiropractors.

Some people taking supplements are at much greater risk of developing adverse reactions than others. The ill are more susceptible to side-effects than healthy people, and environmental factors also have a great bearing. Perhaps the biggest factor is genetic make-up. As with conventional medicines, herbal products are metabolised by various pathways in the body, notably mediated by the cytochrome P450 enzymes, and the composition of these within the body is genetically determined.

The level of risk for most patients is low, but some, especially those taking toxic drugs at the same time, it is much higher. Reactions can include anaphylaxis and hypotension. But those most likely to take herbal remedies are actually those who are the most ill - yet, all too often, they do not tell their doctor what they are taking, vastly increasing the likelihood of untoward interactions. According to a White House study, as many as two-thirds of cancer patients take some form of herbal remedy. And the rates in AIDS patients are similar. powerful drugs

The big concern, according to McGilveray, is that these patient populations are exposed to many powerful conventional drugs. 'Most worrying is the use of Chinese herbal remedies,' he says. 'The problem is that no one knows precisely what's in them.'

According to Brian Foster of Health Canada's Therapeutic Products Directorate, there are around a quarter of a million species of flowering plants known, and somewhere around a quarter or a half of these have been used at some time for medical purposes. The World Health Organisation puts the figure closer to 80%. 'These products contain all sorts of compounds - carotenoids, monoterpenes, organosulphur compounds, phenolics, tocopherols, triterpenes and so on,' says Foster. 'We started to look at them because someone spotted that drinking tonic water while they were taking zidovudine made them feel unwell.'

Health Canada has now investigated almost 200 different natural products to establish levels of actives and the effects they have, including camomile tea, Ginkgo biloba, ginseng, echinacea and feverfew. One might think that all over the counter products contain the same (or at least very similar) levels of ingredients. But, the team found, this is not the case. Growing conditions can have a dramatic effect on the composition of the natural product. The climate makes a difference: humidity, rain, temperature or exposure to frost. And fresh products can be very different from processed ones, which can all too easily lose their volatile oils. Misidentification of plants and substitutions can also be an issue.

Several common foodstuffs are implicated in a number of drug interactions, top of the list being grapefruit and Seville oranges. While both these fruits contain flavonoids, particularly chrysin, naringenin and quercetin, the culprits for the drug interactions appear to be furanocoumarins, which are also found in celery leaves and lime juice.

These furanocoumarins increase serum levels of several drug products, notably antihistamines, calcium channel blockers and cyclosporin. Statins, buspirone and misoldipine have problems, too.

Mineral supplements can also be problematic. Calcium is essential for bone growth in children, and also in the prevention of osteoporosis, but it can interfere with drugs by intercalation, examples being the antibiotics ciprofloxacin, doxycycline and tetracyclines. The interaction between tetracycline and milk was first noted half a century ago. It has since been found that milk can cause a drop in serum levels of tetracycline of as much as 60 or 70%.

St John's Wort is commonly used as a 'safe' antidepressant, but it can cause a whole gamut of problems. Hypericum perforatum is a perennial with yellow star-shaped flowers native to Europe, where it has been used medicinally for at least 2,500 years, recommended by Hippocrates, Dioscorides and Pliny, for burns, snake bites and fever. It was rediscovered as a medicine in Germany in 1980, where it was approved as a 'safe' antidepressant in 1984, and several clinical trials have shown that it has antidepressant properties.

The product is standardised on the ingredient hypericin, despite the fact that it is now known that this is not actually the antidepressant component; hyperforin may act as an SSRI. Research at the FDA showed that in the short term St John's Wort had little effect on the major cytochrome P450 pathways, but chronic use selectively induced CYP3A, particularly in the small intestine.

As a result of this finding, drugs that are CYP3AZ substrates must now contain contraindications to St John's Wort on the label. These include Glivec, and mifepristine. And some St John's Wort products now have to have labelling indicating some of the drugs that can interact.

Kava kava has been implicated in causing liver damage. Kava originated in the Pacific islands, where it is taken in large quantities as a water extract, seemingly without problems. However, the pill product taken as a health supplement contains large amounts of kavalactone, which is essentially absent in the water extract. It is prepared from the root of the pepper plant Piper methyisticum, found in Polynesia, Micronesia and Melanesia. The water-soluble extract does not have the same pain-killing and sleep-inducing properties of the oil-soluble extract in the pills.

However, various reports have appeared in recent years of serious liver problems in patients taking kava products, including cirrhosis, hepatitis and liver failure, in some cases leading to the patient needing a liver transplant. Last year, FDA put out a warning letter, saying that great care should be taken, and no one should take kava products without first consulting their doctor. Several countries, including Canada, have removed products containing kava from the market altogether.

increased resistance

Another plant whose root is commonly used is echinacea. Echinacea angostifolia, which is native to the western US, has purple flowers whose shape led to its common name of the coneflower. The root extract is claimed to increase resistance to infection, and to have aphrodisiac properties. It has been used for centuries in native American medicine, and is now commonly used to ward off the common cold.

The Health Canada team looked at 14 echinacea products, and six biomarkers. Levels of cichorate ranged from 1 to 1,581 ppm in all the products looked at, and echinoside 16 to 604 ppm, but in only some of the products. Chlorogenate ranged from 2 to 3,230 ppm and trans caffeic acid 1 to 12 ppm. These vast differences in components make it very difficult to see precisely what active has what effect, if any.

Garlic, as well as being an important ingredient in cookery, is a common dietary supplement. Yet it too can interact with medicines. A serious adverse interaction was seen in two AIDS patients taking ritonavir.

Both of the hospitalised patients liked garlic, one taking four capsules a day - twice the recommended dose - and the other six or seven cloves of raw garlic. The Health Canada team investigated 10 readily available products, including different lot numbers of the same product, and found very different ingredient profiles.

Geoff Tucker of the University of Sheffield has carried out a large amount of work on drug-drug interactions, including with herbal products. He says that there are a number of special issues that have to be considered with herbal medicines. What are the relevant constituents of the remedy? There can be a very wide variability in their precise composition. The relevant enzymes they affect within the body must be identified. But it's not simple - there are a large number of products and chemicals, and we have no idea what systemic levels people will be exposed to. And there are some special risk groups who will be exposed to much greater amounts.

Tucker is using two approaches: in vitro screens and IVIV extrapolation, and in vivo 'cocktail' screens of the whole product. IVIV extrapolations are an attempt to simulate and predict interactions, using a Monte Carlo method to predict population distributions. The aim is to evaluate the risk to individuals, and spot patients who are at extreme risk.

They have had some success with drug-drug interactions, such as the interaction between ritonavir and methadone, and Tucker believes that there is no reason why predictions cannot be made for herbal medicines. The only problem is that the composition of the different products can vary so widely - and can vary enormously from batch to batch.

Perhaps Paracelsus had it right. He said that all substances are poisons, and there is none that is not a poison: it is the right dose that differentiates what is a poison and what is a remedy.