4 Validating herbal therapeutics
Chapter contents
But do herbs actually work?
Among all the therapies that are called ‘complementary’, phytotherapy is the one that can draw on the most scientific support. This book is constructed largely on the foundation of published literature. It is clear that there is now a sufficient case to construct a rational therapeutic system from the older traditions. However, in spite of this and the presence in medicinal plants of many pharmacologically active constituents, the most persistent doubt expressed by the medical world is whether whole herbs actually work in practice.
The placebo response
The term ‘placebo effect’ itself is not wholly appropriate for this discussion. It derived originally from observations in double blind clinical studies where the treatment was compared with a dummy pill made to look as close as possible to the treatment, with neither the patient nor the practitioner aware of which was which. Early observations were that a significant proportion of subjects in such studies who were taking the dummy nevertheless got better. Initial impressions in the 1940s and 1950s, when such rigorous studies became more common, were that figures varied from trial to trial, but it appeared that about a third of subjects were likely to be ‘placebo responders’. This behaviour was put down, in the psychosomatic model emerging at the same time, as reflecting a particular suggestibility on the part of that proportion of the population. As a result the placebo response has been seen to be a non-serious event, not something to be confused with real medicine and at worst a confounding nuisance in establishing therapeutic efficacy for treatments.
• placebo benefits can occur in any proportion of a treatment group, from zero to almost 100%,1 depending on the condition and circumstances2
• there are no particular ‘placebo responders’ as such3
• placebos have time–effect curves and peaks, cumulative and carryover effects similar to those of active medications2; they can also generate significant levels of interactions with other medications4 as well as adverse effects5,6 (see also the discussion of the ‘nocebo effect’ on p. 109.)
• placebo responses can involve real cures over the long term7 – they are not, as often thought, transient, imaginary events8
• placebo response can lead to long-term benefit even in difficult conditions such as multiple sclerosis,9 ulcerative colitis,10 benign prostatic hyperplasia8 and schizophrenia.11
The shock also applies, however, to conventional medicine and especially surgery, where placebo responses, as evidenced by examining data from non-controlled clinical studies of surgical interventions which were later found to be valueless, are among the highest recorded.2 The simple instruction from a doctor carries enormous ‘placebo’ impact.12 In whole industries, such as those promoting antidepressant drugs, the impact of placebo relative to treatment response has probably been systematically understated.13
The way through this potential difficulty is to reconsider what the placebo response means.
Non-specific supports for self-repair: a different therapeutic strategy?
Indeed, some medical practitioners have got the point. There is a long unspoken tradition of non-specific prescribing, with vitamins, laxatives, aspirin and, unfortunately, antibiotics to keep the patient happy (placebo means ‘I please’), although actual prescription of placebos as such usually breaches ethical and legislative codes.
Many researchers prefer to use the term ‘non-specific effects’ to describe contributions to improvements in clinical studies that are not caused by the treatment in isolation. As well as the placebo effect itself, they include the natural course of the illness (‘getting better anyway’ is something ‘control groups’ of non-treated patients are supposed to quantify in the better organised clinical trials but even here confusion reigns about hidden ‘placebo effects’14), ‘spontaneous remission’ (generally used to describe recovery that cannot otherwise be explained), a trend for improvement (‘regression to the mean’) due to the fact that people tend to get recruited to studies (and come to obtain treatment) when they are at their worst. All these phenomena are aspects of self-repair. A shift in terminology so that the generally prejudicial ‘placebo effect’ becomes ‘non-specific effects’ will be welcome.
The difficulty of enquiry
There are, however, practical problems in pursuing good clinical research in herbal medicine:
1. To produce results carrying sufficient statistical weight is expensive and laborious (each trial has to be costed in research salaries plus logistical expenses). Herbal medicine in the West can boast no teaching hospitals or research institutes, nor funding by government or a wealthy industrial sector. The necessary infrastructure is lacking. Neither can the costs of undertaking research studies easily be justified commercially. The size of the market for any individual product is not comparable to that for any conventional drug. Commercial investment in clinical trials costing many millions of euros or dollars can only be justified if the manufacturer can recover the investment in the market. A crude herbal is free for anyone to copy and must therefore be transformed into something patentable and different from its natural origins. This leads phytomedicine manufacturers, therefore, to produce new extracts from plants that they can commercially protect. For example there are almost no clinical trial data for Ginkgo leaves as such: on the other hand for proprietary extracts of Ginkgo (EGb761 and other patented extracts15) there is a large evidence base.16 The same is largely true for black cohosh, saw palmetto, St John’s wort, horse chestnut and kava. It is difficult for the wider herbal community to claim efficacy for non-standardised products based on these clinical data. The very high and rapidly escalating costs of conducting clinical trials to modern manufacture, ethical standards and regulatory requirements puts off all but the most promising prospects.
2. The indications often claimed for herbal medicines include many without robust outcome measures. Most are destined for the self-medication or over-the-counter (OTC) market so are by definition directed at lesser degrees of morbidity where hard measures are elusive. By contrast most synthetic OTC medicines on the market have ‘switched’ from prescription status and have acquired their efficacy evidence on harder clinical indications in hospital or clinics. With more variable and lower grade symptoms among the patient population, and with a greater likelihood of self-limiting or other spontaneously changing conditions, clear treatment effects are harder to establish. The result is that it is usually necessary to recruit particularly large patient samples and to devise more artificial exclusion criteria to constrain sample variability. All this places extra logistical demands on those wishing to set up effective clinical trials for these products.
3. Herbs are complex medicines, occupying an unusual position in being medicines with many of the characteristics of foods. Being a complex of pharmacologically active chemicals, the whole package will have different properties from that of any single constituent acting alone. The action of the latter will not predict the effect of the former, particularly if the experimental evidence is based on work done on laboratory animals. It is therefore rare to find the satisfactory preclinical evidence often required by ethics committees for the approval of major clinical studies.
4. There are other unintended consequences of clinical research that can limit their benefits to investors in research. Even the best studies have failed to lead to medicine registrations outside central Europe for example, St John’s wort, Ginkgo and hawthorn (Crataegus spp.). In part this is linked to the indications established by such studies. In all three cases the evidence points to uses that are not appropriate for unsupervised self-medication: depression, cardiovascular disease and heart disease respectively. Therefore in countries where the public are more likely to use herbs in self-medication than through a practitioner, there is little incentive for doing such research. Looking ahead there are new reasons for concern. In Europe, where so much of the herbal clinical literature has been generated, there is a new regulatory regime. The 2004 Directive on Traditional Herbal Medicinal Products permits the registration (rather than licensing) as medicines of herbal products on the basis of their traditional use rather than on proven efficacy.17 Such products still have to comply with pharmaceutical good manufacturing practice and safety monitoring, but there is no longer any incentive to prove their efficacy in clinical trials. In fact, as the Directive does not apply ‘where the competent authorities judge that a traditional herbal medicinal product fulfils the criteria for authorisation’, any such data may bar a herbal product from the less expensive registration option and lead to a requirement that it be licensed as a full medicine. There is a concern that the new Directive will lead to a progressive devaluing of herbal medicines and a drying up of clinical research activity.18
5. The application of herbs and their effect on the body are not always the same as usually understood for conventional medicines. As has been suggested above, herbal medicines may be used more to evoke healing responses in the body than to attack symptoms; this generates a different research question. Clinical trial data will help with, but still not answer, the basic question ‘is this remedy going to be good for this patient?’ It is also likely that genuinely important benefits for a minority of the population will be overlooked.
6. There are some instances in herbal research where blinding will be difficult. For example, the impact of bitter remedies, potentially mediated by the effects on digestive functions of stimulation of the bitter taste buds, have played important parts in the claims of traditional herbal medicine: this is almost impossible to blind. There will as always be a role for the good single blind study, especially if other elements are rigorously controlled.
