While desire for a cure is less often stated as a reason, prevention of recurrence is a common reason for breast cancer patients to use CAM [19]. Furthermore, patients’ and physicians’ opinions diverge on what kind of evidence to use as a guide to CAM therapy. Patients place less emphasis on scientific evidence and rely on a much broader range of information, such as family, friends, and Internet sources [20]. Most oncologists place a much larger emphasis on scientific evidence [13], as they perceive complementary alternatives to be scientifically unproven [16]. However, when provided with online information, it has been shown that both patients and physicians have difficulty distinguishing high quality from low quality or biased information [21]. Therefore, in order for clinicians to be able to have a complete discussion with patients, they must first be aware of the existing clinical evidence.RISKS OF ALTERNATIVE CANCER CARE To fulfill the first caveat of clinical medicine, first do no harm, it is important to examine what evidence exists for risk associated with dietary supplement usage. Refusal of curative conventional treatment is often cited as a risk of CAM therapy by the conventional medical establishment [22]. Limited evidence does exist to support this concern. When alternative therapies were used as first-line treatment in a group of 33 women with breast cancer, early death and higher rates of recurrence were reported [23].

 Furthermore, when tested in rigorous clinical trials, all ‘‘natural cancer cures’’ examined so far, such as shark cartilage [24,25], either have shown no benefit or, in the case of others such as laetrile [26], were found to be both ineffective and toxic. Fortunately, even for patients with advanced disease, most (88%) receive CAM care concurrently with conventional treatment [27] and it is the minority (8% in one study) who choose solely alternative cancer care [28]. Concurrent use of CAM with conventional care has generally shown no change in survival time [29]; however, some exceptions of both increased and decreased survival will be discussed below. Risk from use of contaminated or adulterated dietary supplements has been suggested. When this occurs in products commonly used by cancer patients, patients can be exposed to unexpected ingredients that may themselves be toxic or may interfere with the action of pharmaceutic medications. In at least one case, substitution of one Chinese herb for another (Stephania for Aristolochia) caused acute nephrotoxicity and later development of genitourinary cancer [30]. Perhaps the most notorious example of a contaminated dietary supplement used by cancer patients is the herbal formula PC-SPES, which was found to contain warfarin, DES, and other substances [31,32]. These failures in quality control of botanical products have prompted concern in the public as well as the medical community. New dietary supplement manufacturing rules recently released by Food and Drug Administration are designed to address issues of dietary supplement quality [33]. In addition, relatively few herbs have toxic constituents that are not recommended for general use [34].

 One serious but infrequent side effect of concern to oncologists is liver toxicity, especially given the inherently toxic nature of chemotherapeutic agents. Hepatotoxictiy has been reported for some common herbs, such as chaparral (Larrea tridentate), comfrey (Symphytum officinale), and kava (Piper methysticum) [35]. However, not all herbs with reports of hepatotoxicity have equally compelling evidence. For example, black cohosh (Actea racemosa), often used to treat menopausal symptoms—including those in women with or at high risk for breast cancer—has been alleged in a number of case reports to cause liver damage by at least two separate mechanisms, and has been subject to regulation by a number of international regulatory agencies [36]. However, in contrast, an expert conference that convened at the NIH in 2004 reviewed all of the available data and concluded that there was no demonstrated mechanism of action of hepatic injury for black cohosh, as well as insufficient evidence of toxicity, to warrant stopping or modifying clinical trials currently using black cohosh [37].Herb-drug interactions, the form of adverse effect most often mentioned by oncologists, has been the subject of a number of reviews [38–42]. Despite these reviews and the concerns they raise, only two articles were identified that specifically tried to assess the degree of risk encountered by use of dietary supplements, including herbs, during cancer care [41,43]. Of the 76 chemotherapy patients surveyed to identify potentially negative chemotherapy-herb or vitamin interactions, only three of the patients were using herbs (St. John’s wort or Hypericum perforatumy and garlic or Allium sativum) that might have affected the metabolism of their chemotherapy.

 In 318 chemotherapy patients who were also using herbal remedies [43], 11% took supplements in higher than recommended doses and potential interactions were identified in 12% of the patients (n ¼ 20). Most of the warnings were given to lymphoma patients taking echinacea on the basis on potential adverse effects of immune stimulation. In neither study were confirmed interactions observed. Concern for interactions is based on the ability of herbal products to affect the cytochrome P450 enzyme system that is crucial in the metabolism of a number of chemotherapeutic agents.The investigators concluded that interactions with saw palmetto (Serenoa repens), cranberry (Vaccinium macrocarpon), black cohosh, milk thistle (Silybum marianum), and bilberry (Vaccinium myrtillus) were not expected, while specific cautions were made for garlic, ginkgo (Ginkgo biloba), soy (Glycine max), ginseng (Panax ginseng), valerian (Valeriana officinalis), and kava, largely on the basis of preclinical data. Despite the high level of concern expressed in these cited reviews, only a handful of human pharmacokinetic studies were identified in the literature that directly assessed the effect of any herb on a chemotherapeutic agent. Ten breast cancer subjects took 600 mg of a proprietary garlic extract containing 3,600 mcg of allicin twice a day for days 5 to 17 of their chemotherapy cycle [45]. Docetaxel pharmacokinetics assessed before and after the administration of garlic showed no change in peak concentration, area under the curve (AUC), or half-life. Milk thistle (Silybum marianum) was tested in six cancer subjects taking irinotecan [46]. 

Four days before their second dose of irinotecan, subjects were given 200 mg of a commercially available milk thistle extract, standardized to 80% silymarin, three times a day. No significant effect on irinotecan clearance was noted despite a slight but borderline statistically significant decrease in the AUC. Serum concentrations of silybin, one of the constituents of milk thistle, were felt to be too low to be of concern for drug interactions. Most of the reviews cited above agree that the herb with the strongest risk of clinically significant interactions is St. John’s wort. Three human studies of the effect of St. John’s wort on imatinib confirm the need for caution. Ten subjects, given 400 mg of imatinib before and after a treatment with 300 mg of St. John’s wort three times a day [47] showed significant alterations in the pharmacokinetics of imatinib (32% reduction in AUC and a 29% reduction in maximal concentration). In a second study, 12 healthy subjects who were given 300 mg of a standardized proprietary St. John’s wort product three times a day also had a significant increase in imatinib clearance (43%), as well as a 30% reduction in AUC [48]. Significant reductions were also noted in half-life and maximum concentrations. More worrisome was a small study of five cancer subjects taking irinotecan concurrently with 900 mg per day of St. John’s wort extract. Plasma levels of SN-38, the active metabolite of irinotecan, were statistically significantly reduced (42%, P ¼ .033) and, more significantly, myelosuppression was less during cotreatment as well [49]. However, in none of the studies was the composition of the St. John’s wort extract described or independently confirmed by the investigators. 

In the case of St. John’s wort this is particularly important, as most of the induction of the CYP450 enzymes is felt to be because of hyperforin, one constituent of St. John’s wort extracts [50]. Interference with coagulation by herbs or other dietary supplements is of particular concern for oncology patients as they undergo surgery or other invasive procedures [51]. The investigators of this recent review postulate that antiplatelet actions, as well as interference with warfarin, could put cancer patients at risk. However, when commercially available extracts of ginkgo, garlic, Panax ginseng, St. John’s wort, and saw palmetto were given to 10 healthy volunteers for 2 weeks, no effect on platelet activity was demonstrated [52]. Likewise, evidence of interaction of warfarin with herbs is based largely on case reports of variable (mostly poor) quality, which is not confirmed by pharmacologic studies [53]. Herbal therapies with estrogenic, androgenic, or progesterone-like activity are a theoretic concern for patients with hormone-sensitive cancers, particularly breast, ovarian, endometrial, or prostate cancers [54]. However, the majority of the literature on hormonal effects of herbs focuses on the estrogen activity of herbs commonly used for treating menopause. In preclinical trials, soy isoflavones and red clover extracts have been shown to have estrogenic activity of uncertain clinical significance for estrogen receptor-positive breast cancer patients [55]. Black cohosh, although mistakenly referred to as a phytoestrogen, does not appear to have estrogenic activity as tested in a variety of in vitro, animal, and human studies [56–58]. Neither does it appear to increase breast density [59]. 

The only formal safety study of a proprietary black cohosh extract (Klimadynon) was done in normal menopausal women [58]. Four hundred women were given 20 mg of herbal drug for 12 months under close observation. No increased uterine hypertrophy or heptatoxicity was noted during the trial. Breast density in the subset of women who had mammograms before and after the trial was lower, suggesting no toxic effect on breast tissue. Activity via serrotonergic neurons in the hypothalamus is thought to account for the clinical effect of black cohosh on vasomotor symptoms [60]. Therefore, the conservative recommendation to avoid consumption of high amounts of herbs with in vitro estrogenic activity does not apply to the use of black cohosh extract (BCE). In summary, exposure has occurred from adulterated DS, such as PC-SPES, but few cases are noted. Evidence of hepatotoxicity of BCEs was not sufficient to warrant the interruption of ongoing clinical trials, nor is black cohosh a phytoestrogen. Despite widespread theoretic concern about herb-drug interactions, clinically significant interactions were only proved with St. John’s wort. Observations from in vitro screening and animal studies need to be tested in human beings to confirm the presence or absence of clinically relevant interactions. Evidence of interference with platelet function or warfarin activity was contradictory or absent for most herbs tested. Antioxidants Although antioxidant-rich foods are commonly associated with reduced risks of a variety of cancers, use of antioxidants, either singly or in formulas, as preventative agents for cancer has not been supported by large randomized trials [61]. 

In fact, at times antioxidant supplementation has been associated with harm when used preventatively [62]. Thus, use in conjunction with conventional therapy has remained a controversial area. Conventional practitioners usually have general prohibitions against use during chemotherapy or radiation [1]. However, a number of authorities have highlighted large amounts of preclinical and limited amounts of human clinical data in favor of use of at least some antioxidants [63–67]. Discussion of a limited selection of some of the key human clinical trials in this area will highlight key points in this debate. Conventional clinicians cite fear of decreasing the effectiveness of conventional therapy as their major concern with the use of antioxidants during chemotherapy or radiation [1]. Some evidence exists for this concern. During a large randomized, double blind, placebo controlled trial of 540 patients with head and neck cancer undergoing radiation [68], subjects were given either placebo or a combination of antioxidants (400-IU alpha-tocopherol, 30-mg beta-carotene) daily throughout radiation therapy and for 3 years afterwards. Although acute side effects of radiation were significantly less in the antioxidant group, quality of life was not improved significantly, and the rate of local recurrence was higher in the supplemented group (odds ratio or OR 1.37; confidence interval or CI 0.93–2.02). Long-term follow-up of these subjects showed that at a median follow-up of 6.5 years, all-cause mortality was significantly higher in the treated group (hazard ratio or HR 1.38; CI 1.03–1.85) [69].