172 The nutrition handbook for food processors in relation to colon cancer, and to be obtained with standardised procedures that can be applied to a wide enough range of foods for comparisons to be made Biomarkers are required because human death, disease and sub-optimal health are not permissible dependent variables, and many are the result of cumulative changes over long periods. Instead, intermediate biomarker 'end-points', markers of exposure to a food component, and food properties that research has already established as causal in disease and health must be used to assess health effects of food processing. Intermediate end-points must be either causal factors or correlated with changes that lead to end-points. For instance, hyperlipidaemia is an intermediate biomarker that is causally related to a true end-point atherosclerosis. However, as many factors are involved, evidence for the benefit of a product would be more convincing if several relevant biomarkers were measured. At present most biomarkers require clinical or laboratory mea- urement and are not widely used to monitor nutritional changes in the course of product development. A good deal of further work is required to develop tests hat are useful to industry 7.4.3 Validity that is balanced with practicality Validation is a crucial step in selecting variables that indicate effects of foods and food processes on biochemical precursors of health end-points. Because most foods are complex systems, ideal experimental trials in which one food factor is varied while all other variables are kept constant are not often possible, and there is a need to balance practical requirements of food processing with degree of nutritional validation. Given that final products should be comprehensively eval- uated, progress in food processing will often best be maintained by being pre- pared to sacrifice some degree of validity for expediency by appropriate choice of tests. as discussed in section 7.2.3 and illustrated in Table 7.1 7.4.4 Nutrition information that is up-to-date Nutrition s is constantly advancing, and as hard data throws new light on the relationship between a food property or component and a health end-point, indices of food effects on health are likely to change. For instance, heart disease is now considered to be infuenced less by intake of fat than by intake of specifi fatty acids such as saturated and trans-fatty acids. Such changes are not a sign that nutrition science cannot be relied on but that continuing research leads to A food company that had not kept abreast of nutritional knowledge recently formulated a new ' diabetic muesli bar, replacing all sucrose sources with dex- trins, in the belief that ' replacement would improve blood glucose control However, such wisdom was obsolete, because sucrose, being half fructose, induces a much lower blood glucose response than dextrins, which are rapidly digested glucose polymers. The new 'diabetic bar had a greater glycaemic impact than the unmodified versionin relation to colon cancer,40 and to be obtained with standardised procedures that can be applied to a wide enough range of foods for comparisons to be made. Biomarkers are required because human death, disease and sub-optimal health are not permissible dependent variables, and many are the result of cumulative changes over long periods. Instead, intermediate biomarker ‘end-points’, markers of exposure to a food component, and food properties that research has already established as causal in disease and health must be used to assess health effects of food processing. Intermediate end-points must be either causal factors or correlated with changes that lead to end-points. For instance, hyperlipidaemia is an intermediate biomarker that is causally related to a true end-point – atherosclerosis.39 However, as many factors are involved, evidence for the benefit of a product would be more convincing if several relevant biomarkers were measured. At present most biomarkers require clinical or laboratory mea￾surement and are not widely used to monitor nutritional changes in the course of product development. A good deal of further work is required to develop tests that are useful to industry. 7.4.3 Validity that is balanced with practicality Validation is a crucial step in selecting variables that indicate effects of foods and food processes on biochemical precursors of health end-points. Because most foods are complex systems, ideal experimental trials in which one food factor is varied while all other variables are kept constant are not often possible, and there is a need to balance practical requirements of food processing with degree of nutritional validation. Given that final products should be comprehensively eval￾uated, progress in food processing will often best be maintained by being pre￾pared to sacrifice some degree of validity for expediency by appropriate choice of tests, as discussed in section 7.2.3 and illustrated in Table 7.1. 7.4.4 Nutrition information that is up-to-date Nutrition science is constantly advancing, and as hard data throws new light on the relationship between a food property or component and a health end-point, indices of food effects on health are likely to change. For instance, heart disease is now considered to be influenced less by intake of fat than by intake of specific fatty acids such as saturated and trans-fatty acids.41 Such changes are not a sign that nutrition science cannot be relied on but that continuing research leads to clarification. A food company that had not kept abreast of nutritional knowledge recently formulated a new ‘diabetic muesli bar’, replacing all sucrose sources with dex￾trins, in the belief that ‘sugar’ replacement would improve blood glucose control. However, such wisdom was obsolete, because sucrose, being half fructose, induces a much lower blood glucose response than dextrins, which are rapidly digested glucose polymers. The new ‘diabetic’ bar had a greater glycaemic impact than the unmodified version. 172 The nutrition handbook for food processors