Main

The use of complementary alternative medicines (CAM) is well documented (Ernst and Cassileth, 1999). These are either used on their own (alternative) or in addition to conventional medicine (complementary) (Zimmerman and Thompson, 2002). This is particularly common in patients suffering from chronic disorders such as cancers and their associated physical and psychological problems. Depending on the definition and inclusion criteria chosen, estimates range from 7 to 64% in the reported prevalence of CAM use in cancer patients (Ernst and Cassileth, 1998). More recent studies have reported an even higher prevalence of between 70 and 80% (Richardson et al, 2000; Bernstein and Grasso, 2001; Ashikaga et al, 2002). The nature of CAMs used, for example, vitamins and other supplements, herbal remedies, physical and psychological treatments, also varies greatly (Risberg et al, 1998; Richardson et al, 2000; Sparber et al, 2000; Bernstein and Grasso, 2001; Ashikaga et al, 2002).

Patients with chronic illnesses who seek alternative therapies are likely to use conventional medicine regularly and simultaneously. However, they may not always inform their doctor of the concomitant use of alternative medicine. For instance, a study of Eisenberg and co-workers in the US showed that 96% of alternative-medicine users also sought a conventional medicine provider for at least one medical condition. In all, 28% used alternative medicine for the same medical condition, and 72% did not inform their physician (Eisenberg et al, 1993; Kessler et al, 2001). The reasons for CAM use have been widely investigated. Patients often wish to combine conventional and CAM approaches to improve their quality of life, to counter side effects, to achieve a sense of control and to match their life style with their world view (Austin, 1998; Sparber et al, 2000; Kessler et al, 2001).

However, the use of CAM and especially of herbal remedies and supplements is not without problems. Unconventional cancer therapies such as Laetrile, Essiac and coenzyme Q10 may not be effective (Ernst and Cassileth, 1999). Furthermore, CAMs have potentially dangerous side effects and interactions with conventional treatments. For instance, garlic and cod liver oil have anticoagulant effects (Fugh-Berman, 2000), and remedies acting on the cytochrome P450 system such as St John's wort, may interact with hormones, antibiotics and chemotherapeutic agents (Izzo and Ernst, 2001).

Many reviews of the potential dangers have been published, but clinical accounts are mostly confined to individual case reports of adverse events (Ernst, 1998). The purpose of this survey was to prevent potential health risks, which CAM users might encounter. We aimed to establish the type, frequency and pattern of herbal medicine and supplement use in a sample of cancer patients and to identify and quantify the potential for adverse side effects or drug interactions with conventional medicines.

Methods

We conducted a cross-sectional survey of patients attending the outpatient departments at the Royal Marsden Hospital, a specialist cancer centre using a multiple-choice questionnaire to estimate the presence, frequency and purpose of herbal medicines and supplement use. In addition, respondents were asked whether they had discussed their CAM therapy with their medical practitioners. The questionnaire was piloted on 5% of the sample, and amended as necessary. The completed questionnaires were returned to the Medicines Information Service at the Royal Marsden Hospital pharmacy. There they were scrutinised for potentially serious adverse effects or interactions with prescribed medicines using the web-based and library resources. If the potential for an adverse drug reaction or interaction was detected, the pharmacist (CS) issued a health warning to the patient and treating doctor or GP. The data were entered into a database and analysed descriptively using SPSS version 10. Patients gave written informed consent before participation in the study. The project had received ethical approval from the Royal Marsden Hospital Ethics Committee.

Results

Of the 500 patients invited to participate, 318 (63.6%) agreed to take part in the study, of whom 60.4% were female. As the study was conducted immediately after consent had been obtained, it was difficult to establish the reason for nonparticipation. However, 65.0% of the nonparticipants stated that the study did not apply to them as they were not taking any CAMs.

Of the patients surveyed, 164 (51.6%) took herbal remedies and/or food supplements. In all, 133 different substances and combinations were recorded. Of these, 16 (9.8%) took CAM in the form of homeopathic preparations. Patients took on average 1.8 (±2.34) supplements; 40.9% took more than one substance and three patients took 10 or more preparations, and 17 (10.4%) only took herbal remedies, 69 (42.1%) only supplements and 78 (47.6%) a combination of both. Among the alternative remedies, Echinacea, evening primrose oil, ginkgo, milk thistle and essiac were most popular (Table 1a). Individual supplements included vitamin C, E and a combination of vitamin A, C and E (ACE), cod liver oil, selenium, beta-carotene, coenzyme Q10 and germanium. However, the majority took either multivitamins or other combinations, which were difficult to quantify in detail (Table 1b).

Table 1 (a) Alternative remedies taken (n=166a) (b) supplements and supplement combinations taken (n=324a)
Full size table

Half of all patients took CAMs for the nonspecific purpose of improving their health or in order to fight cancer, rather than for a specific indication such as boosting their immune system. Most patients took the remedies according to their purported indication, although many of the indications, particularly anticarcinogenic effects, are unproven. Patients with haematological cancer aimed to boost their immune system with echinacea. Patients with breast cancer used cod liver oil for joint pain and evening primrose oil for breast soreness or hormonal disturbances. Milk thistle was taken to detoxify the liver, presumably to counter some side effects of chemotherapy. One patient with lung cancer tried shark cartilage that is supposed to inhibit angiogenesis. In all, 41 (25.0%) patients took substances with psychoactive properties. However, 53 (32.3%) patients were not sure about the purpose of a remedy taken. For further reference, the suggested indications for all the listed remedies are listed in Appendix A.

The pharmacy issued health warnings for 20 (12.2%) patients taking herbal medicines or supplements (Table 2a). Most concerned the use of echinacea in patients with lymphoma. Owing to its immune system-stimulating activity, Echinacea could have interfered with corticosteroid and monoclonal antibody treatment (Natural Medicines Comprehensive Database, 2003). Further warnings were issued for cod liver/fish oil, evening primrose oil, ginkgo and garlic, all of which have coumarinic constituents, as an interaction with warfarin, aspirin and nonsteroidal anti-inflammatory drugs could lead to an increase in INR (Fugh-Berman, 2000; Natural Medicines Comprehensive Database, 2003). Patients were informed of a potential interference of Siberian Ginseng with antihypertensive therapy (Natural Medicines Comprehensive Database). Kava kava is potentially hepatotoxic (Escher et al, 2001; Russmann et al, 2001), which has led to voluntary withdrawal of all preparations from the UK market. We also issued a qualified warning to one patient taking beta-carotene, who was known to be an occasional smoker. Beta-carotene may increase the risk of prostate and lung cancer in smokers through enhanced production of beta-carotene oxidation metabolites if they are not neutralised by other antioxidants such as vitamin C and E (Heinonen et al, 1998; Patrick, 2000). In addition, 18 (11.0 %) patients reported taking supplements higher than the recommended doses. These included: vitamin C (5), vitamin E (4), multivitamins (3), zinc (3), calcium (2), cod liver oil (2) and one of each of the following: selenium, magnesium, glucosamine, germanium, folic acid, tomato tablets and beta-carotene.

Table 2 Warnings issued by (a) pharmacy: lymphoma (b) pharmacy: breast cancer (c) pharmacy: other cancers
Full size table

Only 46.3% using CAMs had discussed these with a health-care professional involved in their conventional treatment, and reported that 82.9% of the conventional practitioners gave a favourable or neutral response. Conversely, only 56 (34.1%) had consulted an alternative practitioner. Of these 78.6% had discussed their conventional medicines.

Discussion

Our survey confirms that there is a high prevalence of herbal medicine and supplement use in cancer patients. A substantial proportion of patients used remedies that have the potential to cause serious adverse reactions or drug interactions. To our knowledge, this survey is the first attempt to identify these potential risks for an actual sample of cancer patients before adverse events have emerged. However, we do not know how these potential risks translate into actual events, and research is required to establish the frequency and seriousness of such side effects and drug interactions. As this study was based on voluntary participation and CAM users seemed to be more likely to participate, we may have overestimated CAM use. However, even if all nonparticipants did not use any form of alternative remedy, the proportion of CAM users would still be 33%. Nonparticipation did not affect the risk estimates, that is, the main area of interest in this study. It was also difficult to draw a clear line between remedies and supplements as these overlap and many patients took combinations.

Although most patients had discussed their use with a health-care professional, there remained a considerable potential for harmful effects. There may be different reasons for this. Medical practitioners may not have the expert knowledge required to deal with the large number of potential risks or may not have the time to do so in routine outpatient clinics. Also, patients may not accept their doctors' opinion and may argue that conventional cancer treatment can be equally toxic. Thus patients may require more education on the benefits of CAMs and their risk management. For instance, patients need to know that for some vitamins, effectiveness is only established when taken in fruit and vegetables but not as supplements (Moertel et al, 1985) or that effectiveness of supplements may be confined to specifically selected populations (Blot et al, 1993; Russell, 2000). They also need to know that supplements may be associated with adverse events including bleeding and liver failure (Palmer et al, 2003) or fail to work, for example, high dose vitamin C (Creagan et al, 1979). Only recently, the UK Food Standards Agency has reduced the safe upper limit for many supplements (Food Standards Agency, 2003). Also, the potential for CAM to interact with drugs given during diagnostic procedures or radiotherapy needs to be recognised. For instance, kelp can interact with contrast agents containing iodine, as used in bone and thyroid scanning (Eliason, 1998). Antioxidants binding free radicals or remedies increasing photosensitivity may interfere with radiotherapy (Ernst, 1998).

Our survey highlights the importance for conventional health-care professionals to discuss CAM use with their patients. Clinicians need to be aware of CAM-induced side effects or interactions and identify hazards, advising patients accordingly and avoiding uncritical encouragement of potentially harmful use. Otherwise, prescribers may expose themselves to criticism and possibly litigation (Cohen and Eisenberg, 2002). Equally patients should be encouraged to disclose information about CAMs to health-care professionals. Such discussions need to be conducted sensitively in order to avoid alienating patients who may feel that they have not been taken seriously or have been criticised for using CAM. Also, given that about one-third of the remedies used had psychotropic effects, the question of whether CAM users have special psychological needs should be explored.

Also, research on CAMs and their interactions with conventional medicines needs to keep pace with the development of new cancer therapies. Although in randomised controlled trials the proportion of CAM users should be equal in each trial arm, the trial outcome could theoretically be influenced if a CAM specifically interacts with the trial agent but not with the control medication/placebo.

Doctors will need to devote time to discussing CAM use in outpatient clinics, although the complexities of side effects and interactions may require clinics that are run jointly with a local medicines information and toxicology services that provide access to and interpretation of herbal formularies, reference texts and web-based resources such as Natural Medicines Comprehensive Database (2003) (naturaldatabase.com) and Longwood Herbal Task Force (www.mcp.edu/herbal). Also, pharmacists have a key role in updating physicians and sharing important information gathered from patients with other health-care professionals (Klepser and Klepser, 1999). Service models need to be designed and tested to meet this challenge.

Contributors: Ursula Werneke and Judith Earl initiated the research participated in protocol development and coordinated the survey. Ursula Werneke coordinated the statistical analysis and wrote the first draft. Judith Earl and Cordula Seydel screened the questionnaires and issued the health warnings. Oded Horn conducted the statistical analysis. Paul Crichton assisted protocol development and facilitated the study. Dominic Fannon participated in the conduct of the survey, and the literature review. All authors contributed to interpretation of the results and the final manuscript. Ursula Werneke and Paul Crichton are guarantors of the paper.

Funding: Department of Psychological Medicine, Royal Marsden Hospital.