At the present time, functional foods, or nutraceuticals in their refined form, seem to exist largely in a kind of regulatory Utopia, unbound by the constraints and requirements of testing guidelines. Product developers and regulatory staff in companies that manufacture industrial chemicals, pesticides and in particular pharmaceuticals, may wonder how these half-food half-medicine products are able to arrive apparently unhindered on the shelves in our shops.
The size of the market last year approximated 930 billion in Japan, including FOSHU (food substances for specified health uses) and non-FOSHU products; $85 billion including dietary supplements in the U.S. and 830 million excluding dietary supplements in Europe. This, taken together with the minimal regulation of the standards of functional foods worldwide and only limited restrictions on health claims, makes the health food industry attractive to pharmaceutical and food manufacturers alike.
The demand for functional foods seems to be growing faster than the accompanying development of sound science. Consumers expect efficacy combined with safety. It seems evident that they assume food to be safe de facto. Food safety is a different paradigm from the safety of industrial chemicals, agrochemicals or pharmaceuticals. The public can understand that synthesized chemical materials may pose a health hazard. Any exposure, particularly ingestion, should be kept to a minimum and, in the case of pharmaceutical products, should be controlled by prescriptive dosing regimes. However, any substance labeled as a food is a different matter altogether.
Food, in addition to satisfying a basic physiological requirement, is also associated with pleasure and recreation and any notion that its nutritional composition may itself pose a health hazard is to be avoided. To a degree food additives such as colors, flavors or preservatives are accepted in minimal quantities and the safety testing of these is well established. Nevertheless, consumer pressure is against the introduction of novel "synthetic" additives and is for the reduction of existing chemical content, whether originating from crop pesticide residues or the food manufacturing process. An example of this is given by the public furor in the U.K. over the implications of the results of a feeding study conducted with rats and genetically-modified potatoes, which were released earlier this year. The study appeared to show that the potatoes, modified to express an insecticidal protein, could have adverse effects on certain organs and the immune system. Although the study has subsequently come in for criticism, the issues surrounding the safety of genetically modified organisms to human health and to the environment have since not been far from the media forefront in this country.
It is important to ensure that questions surrounding the safety and efficacy of functional foods are answered as far as possible before they are asked, but what are the pressures to do so?
Regulatory Requirements
In Europe, food law is intended to ensure a high level of protection of public health and safety to the consumer and to ensure that legislation is primarily based upon scientific evidence and risk assessment. The prime responsibility for safety is with industry, producers and suppliers. Functional foods are regulated under the EC 258/97 novel foods regulation. A novel food is defined as any food that has not been consumed to a significant degree within the European community. Novel foods are assessed and classed as:
(a) substantially equivalent to a traditional reference food;
(b) sufficiently similar to a traditional reference food;
(c) neither substantially equivalent nor sufficiently similar to a traditional reference food.
The degree of safety (toxicology) testing depends upon the degree of human exposure. More widespread exposure necessitates larger studies.
In the U.S., functional foods are regulated under the Federal Food, Drug and Cosmetic Act (FD&C Act) as amended by the Nutrition Labeling and Education Act of 1990 (NLEA) and the Dietary Supplement Health and Education Act of 1994 (DSHEA). Functional foods may be legally classified as conventional foods, dietary supplements or medical foods. Currently, providing the producer can confirm that all ingredients are Generally Regarded as Safe (GRAS), there is no requirement for toxicology testing. Otherwise, testing is needed in accordance with the food additive test guidelines.
In Japan, health foods for specified health uses are approved under the FOSHU license scheme. The approved health uses are limited and the licenses are difficult to obtain, requiring sound scientific data.
Efficacy
For functional food products, it is essential to substantiate the claimed health benefit. In addition to the above-mentioned regulatory pressures for safety data, the need to show efficacy of the functional food may also necessitate safety testing.
In the U.S., a claim for a health benefit must be submitted to the FDA within 30 days of the introduction of the product and in the U.K. all health claims must be substantiated in accordance with the Trades Descriptions Act (1968) and the Food Labeling Regulations (1996).
Carefully designed human trials are important for the conclusive evaluation of efficacy. There are no guidelines for the design and extent of these studies but they can follow a pattern that is well-established for the assessment of pharmaceutical products. Once the producer has decided upon a claim, which will legally allow the product to be sold as a food and not be considered a medicine, then it should be decided to what degree the health benefit should be established. Claim substantiation could vary from a single study in a few patients under controlled conditions in a clinical center to a large trial involving people consuming the product as part of their normal diet, followed by post-marketing surveillance.
Pharmaceuticals are given to healthy volunteers for initial safety assessment, prior to the first evaluation of efficacy in small groups of patients. The essential difference in assessing the efficacy of functional foods is that they are intended to maintain health, not to alleviate or cure a disease state. Therefore subject groups would usually be comprised of healthy people. Adverse and beneficial effects are therefore part of the same study. As the beneficial effect of the food is not intended to result in the recovery of the subject from a particular condition, it is important to select an appropriate set of biomarkers to monitor the results of exposure. Biomarkers should be relevant to the intended health claim and to the mechanism of action. For example, antioxidants might be evaluated by measuring paraoxonase activity or total antioxidant scavenging capacity; a sterol/stanol cholesterol-lowering agent could be assessed by measuring circulating cholesterols (HDL/LDL ratio) or a food containing ingredients for bone health may be monitored using serum alkaline phosphatase, vitamin K and calcium bioavailability. Whatever markers are used, they should be fully validated, easy to measure and reproducible in different centers. Care should be taken over their selection and the interpretation of results since they are likely to be influenced by the nutritional status of the individual and confounded by disease states or genetic polymorphisms.
In addition to giving an indication of efficacy, biomarkers can be used to quantify the kinetics of exposure to the active ingredient. Measuring the dynamic variation of the biomarker may be an attractive alternative to developing methods sensitive enough to detect small increases in functional food ingredients over endogenous circulating levels.
The efficacy trials themselves, in broad terms, may take three forms. A small pilot study measuring only the selected biomarker may be conducted. Following larger studies could include populations of volunteers on either a diet-controlled regime or on their habitual diet, using parallel or cross-over designs. These later studies establish the validity of the findings made in the pilot study, as well as demonstrating that under normal daily diet conditions the functional food still produces a detectable change.
The subjects must be carefully selected from age and gender groups who are to be targeted with the final product. For example, people at risk from cardiovascular disease or osteoporosis could be selected for studies with cholesterol-lowering agents or bone health supplements respectively. Health, genetic, cultural and climatic conditions may all contribute to the outcome of the study.
It is likely that there will be an increasing demand for comprehensive efficacy studies with functional foods. Single-dose studies only have so far been considered sufficient for at least one new product and this may change. Additionally, the World Health Organization issued a press release in January 1998 advising that, until further information becomes available on how beta-carotene and other carotenoids influence processes leading to cancer, these substances should not be promoted to the general population as tumor-preventive. Further to this point, if carotenoids are beneficial in this way, then perhaps the carotenoid-lowering effects of otherwise useful plant sterols should also be more carefully studied.
Safety
Prior to conducting efficacy trials, clinicians will need to be satisfied that the intended dose level of the material will not present a health hazard to the recipients. Although the functional food may have been already part of the human diet somewhere in the world, to a degree, the intended "dose level" of the food product may be much higher than previously encountered and the clinician may want to be reassured of the absence of potential adverse effects. Since we are considering food rather than pharmaceuticals, then in a worst case situation for certain products the human subject may be being given what amounts to unprescribed access to an unextracted, unrefined drug.
For a pharmaceutical active, typical pre-Phase 1 safety assessments would consist of single-dose studies in two rodent species, repeat dose studies (usually two or four weeks in duration) in one rodent and one non-rodent species (including toxicokinetics), some relevant safety pharmacology to evaluate effects on major and/or selected organ systems and at least two in vitro mutagenicity tests. For a food product with a history of human use, then the internal review body or ethics committee of a clinical center may be content with an abridged list consisting of single- and repeat-dose toxicity tests, the two mutagenicity tests and perhaps some safety pharmacology relating to intestinal function. If so, then these requirements would also approximate to the minimum indicated in the EU novel foods regulation (EC 258/97), for which two in vitro mutagenicity studies and one 90-day feeding study in a rodent species would be required for a food or ingredient considered not to be substantially equivalent to a traditional reference food so the needs for safety data can be satisfied with one study program.
When assessing a food product for any toxic potential, the concept of 'wholesomeness' has to be considered. This term is applied to the combined assessment of toxicology and nutritional balance. A potential problem exists in that the assessment of any toxic potential may be confounded by an effect of the functional food upon the nutritional status of the test system, whether in practice that means rodents in a 90-day feeding study or test cells or organisms in an in vitro mutagenicity study.
In a typical 90-day toxicology study, for example, usually a range of dose levels is chosen to give a high dose with some evidence of toxicity and a low dose with no observable adverse effects (a "NOAEL"). The low dose level would ideally be higher than human intake levels, in order to give a safety margin for risk assessment. It is generally considered that incorporation of a test material at levels of up to 5% of the animal feed is unlikely to cause a nutritional imbalance. In a conventional toxicology study, when the test material is a pharmaceutical or pesticidal active ingredient, evidence of toxicity, such as retarded growth, usually occurs at a much lower treatment level than this. However, for food materials, a much higher treatment level may be needed to achieve a similar effect. Invoking the "limit dose" concept for studies with food products where, for example, a dose level ceiling of 1000 or 2000 mg/kg body weight is considered high enough even without evidence of toxicity may be inappropriate because the intake by humans is likely to be relatively high for a functional food, which will erode the safety margin. Therefore, if high dose levels are to be used on animal feeding studies, then a thorough knowledge of the nutritional properties of the functional food (for example energy value, protein content and the bioavailability of micro nutrients) is essential, so that the test diets can be compensated for nutritional imbalance in order to achieve the normal healthy development of the animals.
Assuming the 90-day rodent study is conducted, then the first step would be a 14- or 28-day preliminary study using diet consisting of 50% functional food/50% compensated normal feed, in order to assess whether there would be any palatability problems or any adverse effects. A problem of unpalatability might be solved by using an alternative species, or in certain circumstances oral dosing, depending upon the nature of the test food substance.
If no adverse effects are predicted from preliminary assessments, then the 90-day rodent study can be conducted at the same administration level as used in the preliminary study. A group receiving a low dose level comparable to the predicted human intake should be included in order to provide reassurance of safety in case adverse effects materialize at the high dose. A control group would receive normal animal feed.
If the main feeding study is expected to show no adverse findings, the group size should be increased beyond the usual 10 of each sex typical of most 90-day rodent studies because the investigator, in attempting to demonstrate an absence of adverse effects, needs adequate statistical power in the design. The parameters investigated in an animal feeding study should include all usual and relevant toxicological parameters, including the biomarkers intended for evaluation in the clinical studies, where possible.
If the predicted usage levels of the functional food by consumers or clinical trial subjects is high, such that the safety margin generated by the animal feeding studies is low, then absorption and metabolism studies in animals followed eventually by similar studies in humans provide additional reassurance of the predictivity of the animal model.
It may be possible to avoid the complications of designating an animal feeding study with a functional macronutrient if the novel microconstituent or component can be isolated. Then a traditional toxicological approach can be adopted, where administration of relatively small amounts of test material would have only marginal nutritional implications for the test animals. This is one strategy that has been used for the safety testing of novel proteins expressed in genetically modified crops. In this case, for test purposes, large quantities of the novel proteins are synthesized in fermentation tanks, thus giving large quantities of unfamiliar or inappropriate foods to test species can be avoided. Perhaps one day all functional foods might originate from or come in the shape of genetically modified fruits and vegetables. This might also avoid the obvious "high tech" image perceived by consumers as a negative aspect of the industry.
Conclusion
In summary, it is to be expected that consumers, trading standards bodies and government regulators will want, and expect to see, evidence for the health benefits and safety of functional foods.
Tighter controls may result from this process in the future but, for the present, the opportunity exists for the entrepreneurial health food producer to seize the marketing advantage that is offered by a scientifically substantiated health claim.
NW