New approaches to providing nutritional information J. A Monro, New Zealand Institute for Crop Food Research 7.1 Introduction Both food processors and consumers have a basic need for valid and relevant nutritional information; on the one hand to guide production and marketing of genuinely functional products, and on the other to allow selection of products according to efficacy. Data on product efficacy that are relevant, in the sense of being easily communicated, understood and appropriately applied are, however often unavailable. There is little to guide evidence-based food choice to meet such widespread health challenges as control of blood glucose levels, and maintenance of large bowel function, two examples that will be considered in detail in this o New approaches to nutritional information must, therefore, focus not only on ata that validly represent physiological changes linked to health, but also on transforming the data to show meaningfully the relative efficacy with which prod ucts can bring about change. This chapter focuses on new approaches to nutri- tional information that attempt to link food choice to health end-points through effective communication 7.2 Why food processors need new types of nutritional information Food processing has been shown to have large effects on a range of nutritional properties. Changes in macronutrients, such as starch and protein, 3and the destruction of vitamins during food processing, have been well documented and are updated in this book. Most studies have been constrained by
7 New approaches to providing nutritional information J. A. Monro, New Zealand Institute for Crop & Food Research 7.1 Introduction Both food processors and consumers have a basic need for valid and relevant nutritional information; on the one hand to guide production and marketing of genuinely functional products, and on the other to allow selection of products according to efficacy. Data on product efficacy that are relevant, in the sense of being easily communicated, understood and appropriately applied are, however, often unavailable. There is little to guide evidence-based food choice to meet such widespread health challenges as control of blood glucose levels, and maintenance of large bowel function, two examples that will be considered in detail in this chapter. New approaches to nutritional information must, therefore, focus not only on data that validly represent physiological changes linked to health, but also on transforming the data to show meaningfully the relative efficacy with which products can bring about change. This chapter focuses on new approaches to nutritional information that attempt to link food choice to health end-points through effective communication. 7.2 Why food processors need new types of nutritional information Food processing has been shown to have large effects on a range of nutritional properties.1 Changes in macronutrients, such as starch and protein,2,3 and the destruction of vitamins during food processing,4 have been well documented, and are updated in this book. Most studies have been constrained by
166 The nutrition handbook for food processors experimental design to a few foods and conditions, and apart from effects on nutrient levels measured in standard food analyses, nutritional information that reflects the effects of processing does not generally reach consumers in a form in which it can be widely used to choose foods. Similarly, food processors often do not have available means to apply, either accurately or practically, relevant nutritional criteria to select prototype products during development. Below are listed several reasons for the need for new nutritional information 7.2.1 Showing effects of food properties on nutritional qualities Much of the impact of processing on nutritional quality comes about through changes in physicochemical properties of food polymers, such as dietary fibre, hat cannot be represented by food composition values. Changes in structure asso- ciated with such processes as starch hydration and gelatinisation, milling, and extrusion,can have an large impact on the rate and extent of digestion, and onsequently on a range of physiological markers linked to disease end-points For example, the impact of cereals on blood glucose and insulin responses is increased markedly as particle size is reduced by milling, or as starch is 7.2.2 Tracking changes in nutritional quality Information from rapid, valid, but relevant tests is needed to guide processing to healthy products. Because of the expense, time, ethics, compliance and other issues involved in clinical trials, human subjects are not usually suitable for mon- oring effects of processing until potential products have been identified Food processing for improved nutrition may require pragmatic choice of dif- ferent tests at different stages in product development, to maintain momentum in product development. Active ingredients and formulations may be identified with screening tests, using indirect predictors of health effects, such as ingredient prop- erties. Responses in animal models may guide food processing further, and most promising products then taken into clinical trials, in which effects on biomarkers with established links to health end-points are measured, before a claim of efficacy is made. Increasing rigour in nutritional evaluation of functional prop. erties during the course of product development is illustrated in Table 7.1 As a general principle, the properties of foods that affect physiology should be measured under conditions as close as is reasonable to those in which the food property acts in vivo. Developing a soluble fibre-enriched product to lower blood cholesterol is an example. Ingredients of high soluble fibre content could be iden tified using soluble fibre analysis under simulated gastrointestinal conditions. 12 As hypocholesterolaemic effects of soluble dietary fibres result from increased intestinal viscosity, fibre viscosity should be measured, followed by in vitro digestion of products containing selections of viscous fibre sources, with mea surement of digesta viscosity. Promising products could be subjected to animal trials to establish that predicted gut viscosity and blood lipid changes occur
experimental design to a few foods and conditions, and apart from effects on nutrient levels measured in standard food analyses, nutritional information that reflects the effects of processing does not generally reach consumers in a form in which it can be widely used to choose foods. Similarly, food processors often do not have available means to apply, either accurately or practically, relevant nutritional criteria to select prototype products during development. Below are listed several reasons for the need for new nutritional information. 7.2.1 Showing effects of food properties on nutritional qualities Much of the impact of processing on nutritional quality comes about through changes in physicochemical properties of food polymers, such as dietary fibre,5 that cannot be represented by food composition values. Changes in structure associated with such processes as starch hydration and gelatinisation,6 milling,7 and extrusion,8 can have an large impact on the rate and extent of digestion, and consequently on a range of physiological markers linked to disease end-points. For example, the impact of cereals on blood glucose and insulin responses is increased markedly as particle size is reduced by milling,7 or as starch is gelatinised.6,9 7.2.2 Tracking changes in nutritional quality Information from rapid, valid, but relevant tests is needed to guide processing to healthy products. Because of the expense, time, ethics, compliance and other issues involved in clinical trials, human subjects are not usually suitable for monitoring effects of processing until potential products have been identified. Food processing for improved nutrition may require pragmatic choice of different tests at different stages in product development, to maintain momentum in product development. Active ingredients and formulations may be identified with screening tests, using indirect predictors of health effects, such as ingredient properties. Responses in animal models may guide food processing further, and most promising products then taken into clinical trials, in which effects on biomarkers with established links to health end-points are measured,10,11 before a claim of efficacy is made. Increasing rigour in nutritional evaluation of functional properties during the course of product development is illustrated in Table 7.1. As a general principle, the properties of foods that affect physiology should be measured under conditions as close as is reasonable to those in which the food property acts in vivo. Developing a soluble fibre-enriched product to lower blood cholesterol is an example. Ingredients of high soluble fibre content could be identified using soluble fibre analysis under simulated gastrointestinal conditions.12 As hypocholesterolaemic effects of soluble dietary fibres result from increased intestinal viscosity,13 fibre viscosity should be measured, followed by in vitro digestion of products containing selections of viscous fibre sources, with measurement of digesta viscosity. Promising products could be subjected to animal trials to establish that predicted gut viscosity and blood lipid changes occur 166 The nutrition handbook for food processors
New approaches to providing nutritional information 167 Table 7.1 Types of nutritional tests that may be used in food processing at different stages of product development Level of se in food processing Comments Ingredient Identifying ingredients May change in processing. Affected by with desired properties food matrix and gut milieu. Cheap Food De ent retention es physiochemical properties, composition Meeting labeling ailability, and bioactivity, but relatively cheap In vitro Predicting responses to gut May give an indication of effects of digestion on food properties, but not take other important food-host interaction Animal Predicting effects in the May be physiologically different to models whole body context humans Require validation. Compliance studie u le lly limited application Costly, slow, ethically difficult, study earing final product. population variable, compliance may be poor. Intermediate endpoints required Epidemiology Identifies possible health- Associations identified, but not cause relevant food factors effect relationship(uncertain) Ecologically valid. Not suitable for Case-control Identifying possible f Less controlled than experimental factors in health b methods leaving room for doubt about cause-effect Experimental Provides basis for a health Most rigorous, but findings apply to the claim experimental conditions -may lack external validity. in vivo. Final selections could then be clinically evaluated, to establish firmly their potential as functional foods for humans. 7. 2.3 Avoiding unjustified extrapolation between products The complexity of processing effects on food matrices makes nutritional proper ties susceptible to processing conditions, and extrapolation of function between different products uncertain. Nonetheless, it is a common cost-cutting measure for marketers to use a functional effect of a bioactive in one product, to promote other foods containing the bioactive, but with a different processing history. To reduce unjustified extrapolation, information is required from tests that are prac tical and inexpensive enough to use for detecting nutritional effects of different processing conditions in large numbers of samples
in vivo. Final selections could then be clinically evaluated, to establish firmly their potential as functional foods for humans. 7.2.3 Avoiding unjustified extrapolation between products The complexity of processing effects on food matrices makes nutritional properties susceptible to processing conditions, and extrapolation of function between different products uncertain.14 Nonetheless, it is a common cost-cutting measure for marketers to use a functional effect of a bioactive in one product, to promote other foods containing the bioactive, but with a different processing history. To reduce unjustified extrapolation, information is required from tests that are practical and inexpensive enough to use for detecting nutritional effects of different processing conditions in large numbers of samples. New approaches to providing nutritional information 167 Table 7.1 Types of nutritional tests that may be used in food processing at different stages of product development Level of Use in food processing Comments evidence Ingredient Identifying ingredients May change in processing. Affected by properties with desired properties. food matrix and gut milieu. Cheap and quick. Food Defining nutrient retention. Ignores physiochemical properties, composition Meeting labeling bioavailability, and bioactivity, but requirements. relatively cheap. In vitro Predicting responses to gut May give an indication of effects of digestion conditions. digestion on food properties, but not take other important food-host interactions into account. Animal Predicting effects in the May be physiologically different to models whole body context. humans. Require validation. Compliance good. Human Usually limited application Costly, slow, ethically difficult, study studies95 until nearing final product. population variable, compliance may be poor. Intermediate endpoints required. Epidemiology Identifies possible health- Associations identified, but not causerelevant food factors. effect relationship (uncertain). Ecologically valid. Not suitable for product testing. Case-control Identifying possible food Less controlled than experimental factors in health by methods leaving room for doubt about association. cause-effect. Experimental Provides basis for a health Most rigorous, but findings apply to the claim. experimental conditions – may lack external validity
168 The nutrition handbook for food processors 7. 2.4 Helping consumers choose foods for health Food properties can be used to select foods for health only if supported by useable efficacy data. Information from tests of the nutritional and functional propertie of foods needs to be easily used by both processors and consumers, to discrimi nate between products. Without scientific but communicable efficacy data, food processors cannot develop or ethically promote foods, and consumers cannot choose foods for real health effects 7.2.5 Gaining consumer confidence Consumer confidence in a product or company requires that claimed benefits are delivered Distrust is associated with perceptions of deliberate distortion of infor- mation, and having been proven wrong in the past. Therefore, the more that the benefits of a food product are exaggerated and overextrapolated, while scrutiny by food regulators and nutritionists continues, the greater the likelihood of gen- erating mistrust. Investment in tests to demonstrate real differences in efficacy of oducts is, therefore, a prudent strategy. 7.3 Limitations of food composition data in food processing A number of related reasons why simple direct relationships between constituent levels in foods and effects on health cannot be assumed are outlined below 7.3.1 People eating foods are complex systems consuming complex systems Food composition data are obtained from standardised analyses of discrete nutri ents, whereas the nutritional effects of food components are modulated by mul tiple interactions within the food matrix, within the gut, and within the body after absorption. Effects of nutrients in foods are, therefore, seldom the same when consumed in a food, as they would be if consumed as a pure nutrient. Yet, food composition is often used as a measure of ' nutritional quality', and nutrient formation panels have been the consumers' main guide to healthy food selec- tion. The ability to make informed food choices for health has therefore been quite restricted 7.3.2 Analytical data may not reflect bioactivit Bioactivity is determined by response of the body to food. Several metabolic steps may converge to contribute more to a response than might be expected from amounts of a single active component in a food. For instance, vitamin A acti- vity may be obtained not only from vitamin A as retinal, but also from a range of provitamin A carotenoid precursors. Therefore, analytical methods must
7.2.4 Helping consumers choose foods for health Food properties can be used to select foods for health only if supported by useable efficacy data.15 Information from tests of the nutritional and functional properties of foods needs to be easily used by both processors and consumers, to discriminate between products. Without scientific but communicable efficacy data, food processors cannot develop or ethically promote foods, and consumers cannot choose foods for real health effects. 7.2.5 Gaining consumer confidence Consumer confidence in a product or company requires that claimed benefits are delivered. Distrust is associated with perceptions of deliberate distortion of information, and having been proven wrong in the past.16 Therefore, the more that the benefits of a food product are exaggerated and overextrapolated, while scrutiny by food regulators and nutritionists continues, the greater the likelihood of generating mistrust. Investment in tests to demonstrate real differences in efficacy of products is, therefore, a prudent strategy. 7.3 Limitations of food composition data in food processing A number of related reasons why simple direct relationships between constituent levels in foods and effects on health cannot be assumed are outlined below. 7.3.1 People eating foods are complex systems consuming complex systems Food composition data are obtained from standardised analyses of discrete nutrients, whereas the nutritional effects of food components are modulated by multiple interactions within the food matrix,17 within the gut,18 and within the body after absorption.19 Effects of nutrients in foods are, therefore, seldom the same when consumed in a food, as they would be if consumed as a pure nutrient. Yet, food composition is often used as a measure of ‘nutritional quality’, and nutrient information panels have been the consumers’ main guide to healthy food selection. The ability to make informed food choices for health has therefore been quite restricted. 7.3.2 Analytical data may not reflect bioactivity20 Bioactivity is determined by response of the body to food. Several metabolic steps may converge to contribute more to a response than might be expected from amounts of a single active component in a food. For instance, vitamin A activity may be obtained not only from vitamin A as retinal, but also from a range of provitamin A carotenoid precursors.21 Therefore, analytical methods must 168 The nutrition handbook for food processors
New approaches to providing nutritional information 169 measure not only vitamin A in a food, but also carotenoids with the potential for conversion to active vitamin A, as vitamin A equivalents. Similarly, blood lipid is a biomarker that depends on more than dietary lipids as lipid type and carbo- hydrate intake, for instance, may also affect blood lipids 7.3.3 Availability or extraction in analytical systems may not equate with bioavailability Analyses are usually designed to measure all of a component of interest in a food, rather than the bioavailable fraction. Most samples are finely ground to facilitate complete extraction with an effective solvent. As a result, food components may be much more soluble during food analysis than during digestion in the gut Soluble dietary fibre is a good example; as prescribed in the American Associa- tion of Analytical Chemists method for soluble fibre analysis, food samples must be finely ground and extracted in hot buffer at 100C, whereas in the gut extrac tion is normally at 364C, from a mixture of particles. Soluble fibre extraction in gut conditions may therefore be much less than in fibre analysis 7.3.4 Physiological effects of food constituents depend on prior physiological state The effects of food constituents are emergent consequences of the interaction between food and body, and subject to the existing physiological state. A viscous polysaccharide may, for instance, lower blood cholesterol in subjects with hyper cholesterolemia, with much less effect in normals. Similarly, the impact of a food carbohydrate on blood glucose levels is affected by the capacity of the bodys cells to absorb glucose, which may be affected by insulin release, by sensitivity to insulin, and by the state of muscle glycogen reserves as a result of exercise The body may modulate uptake and utilisation of a nutrient in response to nutrient status; iron uptake is increased in states of iron deficiency,and vitamin C is excreted when uptake exceeds requirements. 0 7.3.5 A single analytical value may represent a group of compounds differing in nutritional properties Physiological effects often depend on physicochemical properties that vary within a food constituent class that is represented by a single analytical value. Total dietary fibre as a single value in a food table is a family of compounds of diverse form, physicochemical properties, and physiological effects. Dietary fibres exist as insoluble faecal bulking materials, such as wheat bran, that have little impact on blood cholesterol, or as viscous, fermentable, cholesterol-lowering polysaccharides, such as guar gum, that have relatively little impact on faecal bulk
measure not only vitamin A in a food, but also carotenoids with the potential for conversion to active vitamin A, as vitamin A equivalents. Similarly, blood lipid is a biomarker that depends on more than dietary lipids as lipid type and carbohydrate intake, for instance, may also affect blood lipids.22 7.3.3 Availability or extraction in analytical systems may not equate with bioavailability23 Analyses are usually designed to measure all of a component of interest in a food, rather than the bioavailable fraction. Most samples are finely ground to facilitate complete extraction with an effective solvent.24 As a result, food components may be much more soluble during food analysis than during digestion in the gut. Soluble dietary fibre is a good example; as prescribed in the American Association of Analytical Chemists method for soluble fibre analysis,25 food samples must be finely ground and extracted in hot buffer at 100 °C, whereas in the gut extraction is normally at 36.4 °C, from a mixture of particles. Soluble fibre extraction in gut conditions may therefore be much less than in fibre analysis.26 7.3.4 Physiological effects of food constituents depend on prior physiological state The effects of food constituents are emergent consequences of the interaction between food and body, and subject to the existing physiological state. A viscous polysaccharide may, for instance, lower blood cholesterol in subjects with hypercholesterolaemia, with much less effect in normals.27 Similarly, the impact of a food carbohydrate on blood glucose levels is affected by the capacity of the body’s cells to absorb glucose, which may be affected by insulin release, by sensitivity to insulin, and by the state of muscle glycogen reserves as a result of exercise.28 The body may modulate uptake and utilisation of a nutrient in response to nutrient status; iron uptake is increased in states of iron deficiency,29 and vitamin C is excreted when uptake exceeds requirements.30 7.3.5 A single analytical value may represent a group of compounds differing in nutritional properties Physiological effects often depend on physicochemical properties that vary within a food constituent class that is represented by a single analytical value. Total dietary fibre as a single value in a food table is a family of compounds of diverse form, physicochemical properties, and physiological effects.31 Dietary fibres exist as insoluble faecal bulking materials, such as wheat bran, that have little impact on blood cholesterol,32 or as viscous, fermentable, cholesterol-lowering polysaccharides, such as guar gum,33 that have relatively little impact on faecal bulk.34 New approaches to providing nutritional information 169