WO 2008/147226 PCT/NZ2008/000125 TITLE FOODS AND BEVERAGES LACKING GLYCATION PRODUCTS FIELD 5 This invention relates to dietary substances such as foods and beverages. for. human consumption, in particular to those having lowered content of glycation products,'and it relates to methods for creating such foods and beverages. DEFINITION Glycation is defined generally as the result of a sugar molecule such as glucose or fructose 10 bonding to a protein or lipid molecule without the controlling action of an enzyme. This specification refers to exogenous glycation, occurring outside the body. Glycation of proteins tends to make them resistant to enzymatic cleaving and significantly increases their half-life in the body. There is a class of compounds known as "early glycation products". "AGE" is an abbreviation for Advanced Glycation End-products and is used herein to refer to 15 glycation products in general. For example the first step in glycation of a peptide or protein is the formation of a Schiff base when the aldehyde group of a glucose (or like) molecule combines with the amino group of a lysine molecule in a peptide chain forming an mine or Schiff base which has a double bond between the carbon atom of the glucose and the nitrogen. atom of the lysine. The "Amadori 20 product" is the second stage in this version of formation of an AGE, made by rearrangement, wherein the hydrogen atom from the hydroxyl group adjacent the carbon-nitrogen double bond moves to bond to the nitrogen, leaving a ketone. The last step (an irreversible step) is oxidation of -the Amadori product. Well-characterized compounds of this type include NE carboxymethyl-lysine and methylglyoxal; also Amadori -phosphatidyl ethanolamines. 25 "Maillard reaction" or sequence refers to a chemical reaction between an amino acid and a reducing sugar; usually involving heat, and resulting in some extent of non-enzymatic browning as well as the development of flavours, being some of the range of "Maillard products" which are a type of AGE. "UHT" is an abbreviation for Ultra-High Temperature sterilisation treatment - a version of 30 pasteurisation typically run at about 140 deg C for several seconds only.
WO 2008/147226 PCT/NZ2008/000125 BACKGROUND This specification describes the present situation in relation to New Zealand commercially available milk-substitute products for babies and young children, then outlines an alternative method for manufacture of an infant formula. 35 The inventor offers the hypothesis that AGEs in the modern diet cause or promote the incidence of human diabetic disease, especially type I diabetes in children. Although direct substantive proof of such a hypothesis is desirable, it is of course quite hard and quite unethical to obtain. It is possible that use of bovine milk as a diet for infants is a genuine cause, but the wrong specific mechanism has been identified by other researchers. It is 40 prudent to consider all possibilities, because to know the specific mechanism lets one avoid its effects. This specification will consider one plausible mechanism - infant diets that include AGE-type components which are often found at significant concentrations in infant formulas based on cow's milk. Test results are described below. It is desirable to find the answer as quickly as possible, because the rising incidence of diabetes represents human suffering and a 45 load on any national health scheme. Financial advantages to a society. of reducing the incidence of diabetes could be estimated from the following: The cost of treatment of a child through life for type I diabetes is about NZD 700,000. The cost of treatment of an adult with advanced type II diabetes including the related diseases - very approximately NZD 300,000. The annual incidence of new cases in 50 New Zealand of type I diabetes ' 89 to 240 per million in ages 0-15 (reflecting location differences and date of study). The annual incidence of new cases of type II - about 2,000 per million in New Zealand adults. Further, the increasing incidence of juvenile (Type I) diabetes and of maturity-onset (Type II) diabetes in Western populations represents one of the most serious and economically damaging health problems today. 55 This section describes a number of mutually supporting observations: A plausible pathophysiology exists. For example, advanced glycation end products (AGEs) are neurotoxins and have been implicated in the aetiology of Alzheimer's disease (21, 22). The islets of Langerhans have a profuse neural component. Additionally, glycation alters the reactivity of peptides derived from intestinal protein digestion. Short peptide chains of dietary 60 origin are quickly destroyed in the gut wall or blood by dipeptidyl peptidase lV. Morphiceptin derived from cow milk beta-casein has different physiological and immunological properties when glycated (23) as compared with the unglycated peptide. Dipeptidyl peptidase 1V (DPP1V) is CD26, a T lymphocyte activation marker, important in immune responses (24). 2 WO 2008/147226 PCT/NZ2008/000125 Glycation of peptides which react with this immune marker (peptidase), but are not quickly 65 destroyed has implications for the development of diabetes related autoimmunity. Beta casomorphin-7; a peptide from the same area of the molecule of beta casein as morphiceptmi is released from the Al variant of cow casein, but not the A2 variant, and consumption of the Al variant, but not the A2 variant has been highly correlated with Type 1 diabetes incidence (11). The addition of modem processing producing glycation products provides the secular 70 trend link otherwise missing from this attractive hypothesis. Evidence that glycation renders peptides (in general) resistant to peptidases including serum and tissue peptidases is also available in reference 31. Advanced glycation end products are neurotoxins and have been implicated in the aetiology of Alzheimer's disease (18, 19). The islets of Langerhans have a profuse neural component. 75 Additionally, glycation alters the reactivity of peptides derived from intestinal protein digestion. Short peptide chains of dietary origin are quickly destroyed in the gut wall or blood by dipeptidyl peptidase IV. Morphiceptin derived from cow milk beta-casein has different physiological and immunological properties when glycated (20) compared with the unglycated peptide. Dipeptidyl peptidase 1V (DPP1V) is CD26, a T lymphocyte activation 80 marker, important in immune responses (21). Glycation of peptides which react with this immune marker (peptidase), but are not quickly destroyed has implication for the development of diabetes related autoimmunity. Beta casomorphin -7 a peptide from the same area of the molecule of beta casein as morphiceptin is released from the Al variant of cow casein, but not the A2 variant, and consumption of the Al variant, but not the A2 variant has 85 been highly correlated with Type 1 diabetes incidence (8). Apart from AGE(f), some current food additives may contribute to the incidence of diabetes. For example, Vitamin C/ascorbic acid. The inventor has shown that ascorbic acid additions to food as an antioxidant have become increasingly widespread. (table 2) Vitamin C is oxidised itself by food processing, and oxidation products of ascorbate are diabetogenic (26). 90 It is not clear that it is dehydro ascorbic acid itself or further degradation products that is responsible for this beta cell toxicity (27). Lipoic acid. This acid is widespread in foods and is a potent lipid antioxidant. As with vitamin C the oxidised form of lipoic acid may be pro-oxidant and cause tissue damage (28). Prolonged heating in vitro in the presence of oxygen and subsequent storage increases the 95 proportion of this pro-oxidant. Although many environmental agents have been implicated.in the rising incidence of diabetes, only a handful of these satisfy the three criteria of increase in the environment of the past 3 WO 2008/147226 PCT/NZ2008/000125 three decades, causing diabetes in animal models, and having a plausible mechanism of action. Modem food processing and storage through glycation and oxidation may confer 100 adverse effects on some food components, and lead to the speculation that diabetes is a man made disease. The accelerating effect of elevated blood and tissue glucose found in overt diabetes on production of harmful glycation products has its mirror image in the production of harmful food contaminants resulting from many modem food preserving processes. Of particular concern in the New Zealand scene at least is the set of cow milk based infant 105 formulae. The processing involved yields high levels of Amadori products (17), and tables below, and consequently glycated peptides such as beta-casomorphin-7, which has been implicated in the aetiology of Type 1 diabetes. Also, ascorbic acid is frequently added to the formula at the time of manufacture, which will enhance the formation of Amadori products (AGE(f)), also allowing the formation of diabetogenic oxidised ascorbic acid products. It is 110 particularly likely that the introduction of dietary diabetogens in infancy may be harmful (29). Gliadomorphin derived from wheat gluten has also been proposed as a dietary diabetogen, but per capita hard wheat consumption has not increased in the past few decades. Nor is there an ecological association of wheat intake and Type 1 diabetes globally. However, modem food processing with resulting glycation could account for the secular trend. Blockade of the 115 receptors for advanced glycation products in general have been shown to inhibit autoimmune diabetes in the mouse (25). Ref 1 reviews more supporting evidence, as below. A number of publications have proposed that a higher Advanced Glycation Index ( AGE(t)) in a human tissue is correlated with various types of disease. A relationship between AGE(t) and the incidence of diabetes, in humans, is known. For instance ref 1 states that the rate of 120 endogenous formation of AGEs is markedly increased in diabetes mellitus, and refers to a correlation between AGE(f) and induced type I diabetes in the non-obese diabetic (NOD) mouse. Further, restriction of AGE(f) in the dietary intake of a group of diabetic sufferers resulted in significant reduction of two markers of inflammation, plasma C-reactive protein (CRP) and peripheral mononuclear cell TNF-, and also VCAM-1 a marker of endothelial 125 dysfunction. Even further, ref 1 discloses that when a group of diabetic patients were tested for flow-mediated vasodilatation (FMD) and were fed a single oral dose of a glucose-free, AGE-rich beverage, at 90 minutes the serum AGEs had increased while FMD was significantly impaired. This reference shows that dietary intake of AGE(f) does result in a compromise to health. Contrary to a widely held previous belief, intestinal absorbtion of 130 AGE(f) does occur and at sufficient levels to make problems. AGE(f) leads to protein cross linking and intracellular oxidant stress in tissue-culture models, like the effects of AGE(t). 4 WO 2008/147226 PCT/NZ2008/000125 AGEs are found in the composition of human diabetics. Ravandi et al (30) used normal phase HPLC with on-line electrospray mass spectrometry and found that glycated ethanolamine phospholipids made up 10-16% of the total phosphatidylethanolamine (PE) of red cells of 135 diabetic human subjects, but only 1-2% of PE of red cells of normal subjects. Such glycation products have been shown to be diabetogenic in a mouse model of Type 1 diabetes (16), and also a Type 2 diabetes rodent model (17). A variety of hypotheses are being considered in relation to possible causes of type I diabetes. For instance, some of the inventor's previous research has been concentrated on type Al 140 bovine beta-casein degradation products from cow's milk, namely beta-casomorphin 7. Those casomorphins are not found in type A2 bovine milk. An animal model using NOD mice as the test subjects is supportive of the.inventor's hypothesis, but experiments are incomplete. The inventor expects that it will be difficult to prove the assertion that the AGEs in infant formulas contribute to the onset and incidence of diabetes in a human model unless the 145 invention has been in widespread use for a period of time. Meanwhile, he intends to test that postulate by feeding groups of NOD mice on infant formulas or other milk products including measured amounts of AGEs to endeavour to ascertain a dose/response curve that relates AGE(f) to AGE(t) and onset of diabetes. The notion that diet-derived AGEs do contribute to the body's AGE pool, where they can 150 induce inflammation in healthy subjects, was referred to (such as in Ref 1; the contents of which are hereby deemed to be incorporated). The inventor has had a life-long interest in type I or type II diabetes and in identifying probable causes of diabetes so that the incidence of diabetes can be reduced. Symptomatic diabetes has been recognised for millennia, but the documented increase in 155 incidence of both Type 1 and Type 2 diabetes is a very recent phenomenon. Where reliable records are available, it appears that Type 1 diabetes incidence has doubled over the last two to three decades which is similar to the rise in Type 2 incidence (references 1,2). This is far too rapid to be caused by genetic changes, so environmental causes need to be examined. In the context of islet beta-cell damage, such environmental causes are likely to be of 160 microbiological or ingested agent nature. Any one or more of the major environmental candidates as summarised in table 1 are made more likely if they have become more prevalent over the last few decades. This candidacy is also enhanced if the agent is diabetogenic in either normal animals or diabetes prone strains, and if a plausible mechanism for this diabetogenicity exists. 5 WO 2008/147226 PCT/NZ2008/000125 165 Table 1: Previously suggested environmental diabetogenic agents. Microbiological agents 'Excessive' hygiene (ref 3) Specific viruses causing islet damage (4) Common vaccines (5, 6) 170 'Ingested' agents Overweight/obesity associated with excess food / inadequate exercise (7) Al beta casein containing cow milk (8) Gluten containing cereals (9) Coffee (10) 175 Bafilomycin (11) Arsenic (12) Nitrites and nitrates (13) The inventor will not review the evidence for a microbiological cause of the recent epidemic 180 of diabetes here. The excessive hygiene hypothesis meets all three criteria of secular trend congruity, animal model, plus a plausible diabetogenic mechanism, while the others listed in Table 1 lack one or more of these criteria, or have been disproved in other ways. Among the ingested agents, too much food, associated with obesity provides an attractive hypothesis for both Type 1 and Type 2 diabetes - with the accelerator hypothesis (7), whereas 185 the other specific agents listed do not fit with the secular trend - i.e. consumption of the agent has not increased 2-3 fold in the last three decades. Some do not fit with ecological considerations e.g. arsenic, to provide a holistic aetiology. Evidence that changes in modem food processing and storage techniques can provide potential diabetogens will be presented. This, coupled with the 'accelerator' hypothesis 190 provides a likely explanation for the recent epidemic of diabetes. Ascorbic acid (CAS no. 50 81-7) illustrates the time course argument. It is increasingly added to foods as an antioxidant. The global production of this preservative is shown in table 5. Most is used in modem food preservation processes, sometimes as salts or esters. Of course vitamin C, the L-enantiomer of ascorbic acid, has been recognised as a health-promoting agent in itself for some time. 195 Table 2. Global production of ascorbic acid in tonnes 6 WO 2008/147226 PCT/NZ2008/000125 Year Tonnes Notes 1933 0 First chemical synthesis 1934 <1992 <20,000 Estimated from amounts below and global market values for ascorbic and isoascorbic acid (Business Communications Company, Inc www. Bccresearch.com) 1997 97,000 Leo Hepner (Ihepner@probio.com) 2002 110,000 Viola R Development of a yeast based single step process for the manufacture of ascorbic acid (vitamin c) Scottish Enterprise Proof of Concept Fund, 2002 2004 130,000 Leo Hepner (lhepner@probio.com) It is estimated that 72% of global production is used in the food industry. The approximate annual per.capita consumption of ascorbic acid in North America and Europe is 100g (2005) 200 Oxidation of antioxidants as a result of food processing Ascorbic acid. From the above it can be seen that ascorbic acid additions to food as an antioxidant have become increasingly widespread. (table 2) Ascorbic acid is oxidised itself by food processing, and oxidation products of ascorbate are diabetogenic (23).. It is not clear whether it is dehydro ascorbic acid itself or further degradation products that is responsible 205 for this beta cell toxicity (24). Lipoic acid. This acid is widespread in foods and is a potent, lipid antioxidant. As with ascorbic acid the oxidised form of lipoic acid may be pro-oxidant and cause tissue damage (25). Prolonged heating in vitro in the presence of oxygen and subsequent storage increases the proportion of this pro-oxidant. 210 Discussion of the time-course evidence. Although many environmental agents have been implicated in the rising incidence of diabetes, only a handful of these satisfy the three criteria of increase in the environment of the past three decades, causing diabetes in animal models, and having a plausible mechanism of action. Modem food processing and storage through glycation and oxidation may confer 215 adverse effects on some food components, and lead to the speculation that diabetes is a man 7 WO 2008/147226 PCT/NZ2008/000125 made disease. The accelerating effect of elevated blood and tissue glucose found in overt diabetes on production of harmful glycation products has its mirror image in the production of harmful food contaminants resulting from many modem food preserving processes. Of particular concern in relation to the present invention is cow milk based infant formulae. 220 The processing involved yields high levels of Amadori products (14), and consequently glycated peptides such as betacasomorphine-7, which has been implicated in the aetiology of Type 1 diabetes. Ascorbic acid is frequently added to the formulae, which will enhance the formation of Amadori products, as well as allowing the formation of diabetogenic oxidised ascorbic acid products. It is particularly likely that the introduction of dietary diabetogens in 225 infancy may be harmful (26). Ascorbic acid is added to bread dough as an improver, meat to prevent unattractive colouration on storage and exposure to air, potato chips as a fat antioxidant, bottled fruit drink head space to prevent oxidation, infant formulae and many other foods. The often prolonged reaction time implicit in the use of ascorbic acid in .this way will ensure interaction with 230 proteins and lipids. Glycation products form, frequently in large amounts. This form of glycation fits with the secular trend of both forms of diabetes. Such glycation products have been shown to be diabetogenic in a mouse model of Type 1 diabetes (1 and 19), and also a Type 2 diabetes rodent model (20). The second criterion has thus been met. Other researchers are inclined to reach similar conclusions. See for example the Sebekovi K 235 et al review (27) of animal and human studies: Its abstract concludes: "It is concluded that an exaggerated intake of thermally processed foods may exert in vivo diabetogenic and nephrotoxic effects, induce low-grade inflammation, enhance oxidative stress, and promote atherosclerosis". The review did not consider infant foods per se. See also the current European Science Initiative, referred to as COST-927 which is entitled 240 "Thermally processed foods: possible health implications". The web site (available May 2007) is http://www.if.csic.es/proyectos/cost927/wg5.htm . That infant diets are likely to influence the incidence of type I diabetes is possible. Virtanen et al (28) studied 725 children with described feeding patterns and followed up the 33 siblings of children who developed diabetes. They conclude that the cow's milk hypothesis could apply 245 although the design of their study may emphasise a genetic connection. Akerblom et al (29) give evidence from a trial using 242 infants that infant diet may affect spontaneous beta-cell autoimmunity in infancy, suggesting it may be possible to manipulate spontaneous beta-cell autoimmunity by dietary intervention in infancy. 8 WO 2008/147226 PCT/NZ2008/000125 AGEs at significant levels are found in some infant foods, in forms consistent with process 250 induced changes. See Tables 3 to 9 below as illustrations of what the inventor considers are significantly high glycation product levels in the unimproved versions of some foods. ("Sleep Milk" is a product deliberately made to have high levels of AGE substances, by use of the JHT process plus addition of ascorbic acid plus extended storage). Heating in air of many foods in the presence of sugars (e.g. glucose, fructose, lactose etc.), 255 lipids or protein will produce a range of glycated products. Ascorbic acid behaves as a sugar in this context. An exemplar from the literature is shown in table 3. The more extensive the heating, the longer the time, and the greater the concentration of reactants, the more formed and the more advanced these non-enzymatic products e.g. UHT milk contains a thousand fold higher 260 amount of a glycation product than in-bottle sterilised milk (14) Modem food processing often involves the preservation of the product by heating, which then allows it to be stored long term before consumption, often with further reheating immediately before the meal. Antioxidants are often added. Table 3: Amadori -phosphatidyl ethanolamine (PE)results for milk products Glucose-PE nmol/g Lactose- PE nmol/g Human milk ND ND Cow milk ND 78 Infant Formulae A 35 78 B 112 88 C 124 49 Soy milk 268 128 265 Sourcedfrom (15) (ND = not detected). Measurements made by or for the inventor follow in tables 4-9 Are babies and children in New Zealand actually exposed to diet-derived AGEs? The practice of feeding infant formulas is widespread. They are the first choice alternative to breast feeding and could comprise 100% of an infant's intake. Tables 3 and 4 report results from the 270 use of the FAST test on various samples collected in New Zealand, on two different dates. Please note: many of the descriptors below are trade marks of their suppliers or manufacturers. 9 WO 2008/147226 PCT/NZ2008/000125 Table 4: FAST index measurements on a range of typical infant formulas. No. Sample FAST Index 1 Toddler Gold Plus Karicare 20.97 2 Toddler 3 Nurture Heinz 1 16.61 3 Infant Soy Formula Karicare 1 20.16, 22.65, 19.39 4 Toddler 3 Nurture Heinz 2 16.04 5 Sleep Milk (Al UHT + Ascorbic acid 22.54 6 Blue Anchor Fresh Milk (control) 8.97 275 The FAST index/test: (Fluorescence of Advanced Maillard products and: Soluble Tryptophan): (see Birlouez-Aragon et al, Int. Dairy J 12, (1), 2002, 59-67) Table 5: FAST index for milk samples and more infant formulas; no treatment. Sample FAST index Freeze-dried "Full Milk" 9.43 Autoclaved "Full Milk" 9.16 KariCare Gold 1 (from Birth) 11.26, 11.97 KariCare Gold Starter Formula 10.83 KariCare Starter Infant 1 9.91, 10.7 KariCare Gold Follow-on 2 7.34, 10 KariCare Toddler Gold (over 12 months of age) 21.33 Infant Starter Formula (from Birth) with omega LCPs 11.46 Follow-on Formula (over 6 month) 10 Follow-on Formula (over 6 month) with omega LCPs 8.71 Toddler (over 12 months) 22.14 Toddler Gold (over 12 months) 18.32 chor 9.69, 9.48 Fresh Full Milk (Blue) 8.84 ichor 12.55 anchor 6.86, 7.92 Anchor UHT Full Milk 23.59 Anchor UHT Light Milk 23.25 Pam's UHT. Full Milk 24.65 Bulk powder milk 11.37 Pam's full powder milk 19.52 Ridge A2 fresh milk 9.53 So Good soy milk 26.37 Meadow Fresh Farmhouse 10.58 10 WO 2008/147226 PCT/NZ2008/000125 Discussion: This standard test for the presence of Advanced Glycation End-products (AGEs) 280 shows, as expected, that the UHT (ultra-high temperature) pasteurisation process does cause a rise in the FAST index above the value of about 10 that seems typical of ordinary pasteurised milk. The UHT process might have been used during the manufacture of other products although this is conjecture. We do know that the UHT process is becoming widely used nowadays, as in Europe. That AGEs result from heat treatment, according to the Maillard 285 reaction for example, is well-known. That dried milk powder will undergo a gradual browning (correlated with AGE content) during storage in standard conditions is also well known. Table 6: CML/Ne-(carboxymethyl) Lysine assay. Sample Notes ng/mL Human serum BE Fasting 3.706 Human serum BE After the meal -3.748 Human plasma BE Fasting 3.16 Human plasma BE After the meal 3.194 Milk Blue Anchor 2.86 Milk Light Anchor 2.606 Milk Green Anchor 2.204 Formula Toddler Heinz 3.506 Milk: A2, UHT,+ Ascorbic Stored 2 months 5.914 Sleep Milk, with honey 3.742, 3.394 Alginate (control) 0 Alginate QC# 1865/1 0.444 Alginate Mix 0.256 This assay shows that the N&-(carboxymethyl) Lysine content rises with storage. 290 Table 7: Furosine component of Raw Whole Milk and Formula Preparations. Sample Date or batch Corrected - mg/L Raw bovine milk 22/08/07 4.4 LCT WDM 13/08/07 5.1 LCT WDM 27/08/07 4.8 Soy infant formula 7180 96 Soy infant formula 7207 182 Soy infant formula 7141 91 11 WO 2008/147226 PCT/NZ2008/000125 Nurture 3 0237 107 Nurture 3 1255 166 Toddler Gold 7198 120 Furosine tends to be a better test for early glycation products. Table 7 shows that the freeze dried then reconstituted (LCD WDM samples) did not have significantly'increased furosine 295 content. Note that the furosine content of these sampled infant formulas was much higher than that of raw milk. ('Corrected' means made up, ready to consume). Table 8: Toddler (follow on) formula; made up then stored cold for a period. Time stored at 4 deg C FAST index Oh 22.14 24h 35.6 48 h 73.77 72h 71.19 One week 177.15 Table 8 illustrates effects of a common practice among mothers: to make up a batch of infant 300 formula and keep it in a refrigerator for up to a week. A surprisingly high level of AGEs, perhaps due to included ascorbic acid, is found. The particular infant formula tested here is widely used in New Zealand. Some infant formulas include added ascorbic acid. This compound is known to participate in reactions that form AGEs, even in the cold. Table 9 below demonstrates that ascorbic acid mixed with a cold sample results in about as much 305 AGE(f) as does heat treatment. Hence the inventor advises that ascorbic acid (Vitamin C) is not mixed into an infant formula before it is fed to an infant. Table 9: FAST index for milk samples: heating, ascorbic acid. Sample FAST index at 24 h Unheated group Blue 10.02 Blue + ascorbic acid 19.28 Heated group Blue 16.38 Blue + ascorbic acid 18.96 12 WO 2008/147226 PCT/NZ2008/000125 Standard pasteurised milk ("Anchor(TM) Blue") was used for the tests in Table 9; 4 x 20 ml 310 aliquots. 100 mg per 100 ml ascorbic acid was added to two. The heated group were kept at 60 deg C for one hour, then all were kept at 4 deg C for 24 hours. This shows that heating, and/or the addition of ascorbic acid to a cold sample, increases the FAST index. Gliadomorphin derived from wheat gluten has also been proposed as a dietary diabetogen, but per capita hard wheat consumption has not increased in the past few decades. Nor is there an 315 ecological association of wheat intake and Type 1 diabetes globally. However, modem food processing with glycation could account for the secular trend. Blockade of the receptors fori advanced glycation products in general have been shown to inhibit autoimmune diabetes in the mouse (22). Thus the third criterion of plausible mechanism (s) of action has been- met. References 320 1. Onkamo P, Vaananen S, Karvonen M, Tuomilehto J. Worldwide increase in incidence of Type I diabetes--the analysis of the data on published incidence trends. Diabetologia. 1999 Dec;42(12):1395-403. 2. Wild S, Roglic G, Green A, Sicree R, King H. Global prevalence of diabetes: estimates for the year 2000 and projections for 2030. Diabetes Care. 2004 May;27(5):1047-53. 325 3. Kolb H, Elliott RB. Increasing incidence of IDDM a consequence of improved hygiene? Diabetologia. 1994 Jul;37(7):729. 4. Filippi C, von Herrath M. How viral infections affect the autoimmune process leading to type 1 diabetes. Cell Immunol. 2005 Feb;233(2):125-32 5. Heijbel H, Chen RT, Dahlquist G. Cumulative incidence of childhood-onset IDDM is 330 unaffected by pertussis immunization. Diabetes Care. 1997 Feb;20(2):173-5. 6. Dahlquist G, Gothefors L. The cumulative incidence of childhood diabetes mellitus in Sweden unaffected by BCG-vaccination. Diabetologia. 1995 Jul;38(7):873-4. 7. Wilkin TJ. The accelerator hypothesis: weight gain as the missing link between Type I and Type II diabetes. Diabetologia. 2001 Jul;44(7):914-22. 335 8. Elliott RB, Harris DP, Hill JP, Bibby NJ, Wasmuth HE. Type I (insulin-dependent) diabetes mellitus and cow milk: casein variant consumption. Diabetologia. 1999 Mar;42(3):292-6. 9. Vaarala 0. Environmental causes: dietary causes. Endocrinol Metab Clin North Am. 2004 Mar;33(1):17-26, vii. 340 10. Virtanen SM, Rasanen L, Aro A, Ylonen K, Lounamaa R, Akerblom HK, Tuomilehto Is children's or parents' coffee or tea consumption associated with the risk for Type 1 diabetes 13 WO 2008/147226 PCT/NZ2008/000125 Eur J Clin Nutr. 1994 Apr;48(4):279-85. 11. Myers MA, Hettiarachchi KD, Ludeman JP, Wilson AJ, Wilson CR, Zimmet PZ. Dietary microbial toxins and type 1 diabetes. Ann N Y Acad Sci. 2003 Nov;1005:418-22. 345 12. Tseng CH. The potential biological mechanisms of arsenic-induced diabetes mellitus. Toxicol Appl Pharmacol. 2004 Jun 1;197(2):67-83. 13. Dahlquist GG, Blom LG, Persson LA, Sandstrom AI, Wall SG. Dietary factors and the risk of developing -insulin dependent diabetes in childhood. BMJ. 1990 May 19;300(6735):1302-6. 350 14. Pellegrino L, Cattaneo S. Occurrence of galactosyl isomaltol and galactosyl beta-pyranone in commercial drinking milk. Nahrung. 2001 Jun;45(3):195-200. 15. Oak JH, Nakagawa K, Miyazawa T. UV analysis of Amadori-glycated phosphatidylethanolamine in foods and biological samples. J Lipid Res. 2002 Mar;43(3):523-9. 355 16. Peppa M, He C, Hattori M, McEvoy R, Zheng F, Vlassara H. Fetal or neonatal low glycotoxin environment prevents autoimmune diabetes in NOD mice. Diabetes. 2003 Jun;52(6):1441-8. 17. Sandu 0, Song K, Cai W, Zheng F, Uribarri J, Vlassara H. Insulin resistance and type 2 diabetes in high-fat-fed mice are linked to high glycotoxin intake. 360 Diabetes. 2005 Aug;54(8):2314-9. 18. Woltjer RL, Maezawa I, Ou JJ, Montine KS, Montine TJ. Advanced glycation endproduct precursor alters intracellular amyloid-beta/A beta PP carboxy-terminal fragment aggregation and cytotoxicity. J Alzheimers Dis. 2003 Dec;5(6):467-76. 19. Reddy VP, Obrenovich ME, Atwood CS, Perry G, Smith MA. Involvement of Maillard 365 reactions in Alzheimer disease. Neurotox Res. 2002 May;4(3):191-209. 20. Krowczuk S, Kostyra E, Kostyra H, Fraczek J Effect of non-enzymatic glycation on biological properties of morphiceptin. Pol.J.Food Nutr.Sci 2000; 50(1) 39-42 21. Reinhold D, Kahne T, Steinbrecher A, Wrenger S, Neubert K, Ansorge S, Brocke S. The role of dipeptidyl peptidase IV (DP IV) enzymatic activity in T cell activation and 370 autoimmunity. Biol Chem. 2002 Jul-Aug;383(7-8):1133-8. 22. Chen Y, Yan SS, Colgan J, Zhang HP, Luban J, Schmidt AM, Stem D, Herold KC. Blockade of late stages of autoimmune diabetes by inhibition of the receptor for advanced glycation end products. J Immunol. 2004 Jul 15;173(2):1399-405. 23. Patterson JW. Course of diabetes and development of cataracts after injecting 375 dehydroascorbic acid and related substances. Am J Physiol. 1951 Apr 1;165(1):61-5. 14 WO 2008/147226 PCT/NZ2008/000125 24. Domke I, Weis W. Reinvestigation of the diabetogenic effect of dehydroascorbic acid. Int J Vitam Nutr Res. 1983;53(l):51-60. 25. Scott BC, Aruoma 01, Evans PJ, O'Neill C, Van der Vliet A, Cross CE, Tritschler H, Halliwell B. Free Radic Res. 1994 Feb;20(2):119-33. 380 Lipoic and dihydrolipoic acids as antioxidants. A critical evaluation. 26. Leslie RD, Elliott RB. Early environmental events as a cause of IDDM. Evidence and implications. Diabetes. 1994 Jul;43(7):843-50. 27. Sebekovi K et al review "Dietary advanced glycation end-products (AGEs) and their health effects - PRO" Mol Nutr Food Res 2007; 51 1079-1084. 385 28. Virtanen et al Diabetes. 2000 Jun;49(6):912-7.) 29. Akerblom et al Diabetologia. 2005 May; 48(5):829-37. 30. Ravandi et al FEBS Lett 1996 Feb 26; 381 (1-2) 77-81 31. Lapolla A, et al "Enzymatic digestion and mass spectrometry in the study of advanced glycation end products/peptides" J Am Soc Mass Spectrom. 2004 Apr;15(4):496-509. 390 PROBLEM TO BE SOLVED The problem to be solved can be broken down into (a) identifying an at least probable cause for the raised incidence of diabetes in the modem world, and (b) devising at least one way to bypass the cause, in ordei to minimise the risk of disease or to minimise the possibility of a 395 risk of disease. OBJECT It is an object of this invention to provide at least one way to create a foodstuff with reduced or absent AGEs; ingestion of which is likely to enhance health, or at least to provide the public with a useful choice. 400 STATEMENT OF INVENTION In a first broad aspect this invention provides a method for making a foodstuff having a low proportion of AGEs; the method including the steps of: (1) obtaining fresh milk which has been kept chilled and not heat-treated since collection, (2) carrying out appropriate quality tests, (3) adding any desired additives, apart from those liable to undergo AGE-producing 405 reactions during storage, (4) sterilising the foodstuff by an ultrafiltration process in the cold, (5) freeze-drying the foodstuff, and (6) distributing the dried foodstuff to consumers in a cold, 15 WO 2008/147226 PCT/NZ2008/000125 dry state. Preferably the resulting foodstuff has a FAST index for AGE(f) of less than 10. Optionally the dried foodstuff is sold in packs along with separately packed additives. 410 Optionally the packs provide a convenient starting point for making up mixtures. Optionally, ultra-violet sterilisation in the cold may be used. In a related aspect the invention provides a foodstuff substantially unlikely to promote the incidence of diabetes. In a second broad aspect the invention provides a manufactured foodstuff based on a milk, 415 wherein' the foodstuff has been manufactured by a cold process including a step of sterilisation of the milk without heat, so that the foodstuff has a low concentration of glycated products, wherein the furosine content of the foodstuff is less than about 10 mg/i (when in a form that is ready for use), and is distributed for sale in a condition unlikely to form glycation products during storage. 420 Preferably the foodstuff has a FAST index of the foodstuff less than about 12 (the correspondence of the two tests FAST and furosine being a function of the degree of advancement of glycation) Preferably the foodstuff is substantially made from milk. Preferably the foodstuff is substantially made from bovine milk. 425 In a related aspect, the step of sterilisation without heat is an ultrafiltration step such that living organisms or reproducible parts thereof are excluded by size. Optionally the step of sterilisation without heat employs sterilisation using ultra-violet light such that the DNA of living organisms are disrupted. In a major related aspect, the foodstuff is manufactured so as to -exclude or delete chemical 430 materials liable to promote the formation of glycation products during manufacture or storage; said materials being selected from a range including ascorbic acid, lipoic acid, glucose and fructose, and oxygen within a pack holding the foodstuff. Optionally the made-up beverage may be mixed with known potentially AGE-promoting materials (for example: vitamin C) immediately prior to consumption. 435 Preferably the foodstuff is provided in a dry form in a sealed sterile container ready to be made up into a beverage immediately prior to consumption. Preferably also, the foodstuff is or has been stored at a reduced temperature in the range of +5 to - 190 deg C in substantial absence of oxygen. 16 WO 2008/147226 PCT/NZ2008/000125 In one option, the foodstuff is formulated for use as an infant food or a food for children. 440 In another option, the foodstuff is formulated for use as a food for children or adults;. particularly but not exclusively for use by those at risk of developing diabetes. PREFERRED EMBODIMENT The description of the invention to be provided herein is given purely by way of example and is not to be taken in any way as limiting the scope or extent of the invention. '445 Figure 1: is a flow chart of a preferred process according to the invention. EXAMPLE 1 Principles and a rationale have been described within the background. This method for making a milk or a beverage or other food product based on milk avoids inherent glycation, as 450 would have arisen using the conventional heat pasteurisation or UHT pasteurisation followed by storage of the product as a liquid form. This process is used in the manufacture of a variety of products. The invention includes methods for providing foods and beverages with minimised glycation products, and the foods and beverages produced by such methods. METHOD: By way of non-limiting example, the following simplified process for 455 producing a foodstuff or an infant formula without included glycation products, so far as is possible, is described. Numbers refer to Fig 1. The method does not use pasteurisation or other heat treatment (e.g. UHT) to sterilise a mammalian milk product (or an equivalent such as soy or artificial milk), nor is gamma-ray irradiation recommended, and does not add ascorbic acid, lipoic acid, or a reducing sugar such as glucose. Preferred mammalian milks are 460 human, bovine, ovine, or caprine. Caprine milk is preferred under some allergies, although the allergen picture as currently understood may be altered if AGEs are considered as well. Preferably the finished foodstuff is supplied in a dry form which may be made up into a beverage immediately prior to consumption. Being dry - even anhydrous if that does not alter protein structure adversely - slows down any chemical reaction Anoxic- conditions help. 465 Further, the dry foodstuff is preferably stored at a reduced temperature in the range of +5 to 190 deg C as is convenient, for storage over a period of months. The made-up beverage would not be provided with reducing sugars or other potentially AGE-promoting materials. Alternative materials having anti-oxidant properties include plant flavonoid materials such as 17 WO 2008/147226 PCT/NZ2008/000125 luteolin. The user may mix these into the formula immediately (within at most an hour) prior 470 to consumption, because the AGE reaction is relatively slow. As a non-limiting example, the following process for producing a "finished foodstuff' substantially without included glycation products includes the following steps: 1. Establish a preferred supply chain of dairy farms 101, verified by testing 102 for applicable health matters (zoonoses, hygiene, contaminants, leucocyte count etc) and optionally also 103 475 testing to show that all bovine milk is of the A2 phenotype. 2. Collect (104) fresh, unpasteurised milk which has been kept chilled since collection, (but is not pasteurised, and if that requirement appears to present zoonosis risks or. spoilage difficulties, select an infectious agent-free source as substantiated by tests as above, or immunological tests of the animals or by culturing tests on milk samples). 480 3. Re-test 105 using appropriate quality tests on the milk (as are well-known in the dairy industry) in order to verify that the input material is safe and has low glycation content, at least to provide a baseline. Tests could include culture of samples, sensitive immunologically based tests or DNA/RNA type tests for specific organisms, and the like. 4. Separate (106) the cream, typically by centrifugation, and discard. Some additives, but not 485 those (such as ascorbic acid) liable to undergo AGE-producing reactions during processing or storage could be included before sterilisation. 5. Sterilise (see below) the fat-free milk. Example cold-sterilisation options include: 5A: 107 is an ultrafiltration process carried out in the cold, using a pore size sufficiently small to exclude all. known infectious agents. (Commercial processes are available) 490 5B: 108 is an "in the cold" ultra-violet light sterilisation process using for example a low pressure mercury-vapour lamp source, which particularly acts upon DNA. Commercial apparatus and processes are well known to those skilled in the art. Ultraviolet light irradiation may raise the level of AGE(f) more than cold filtration. 6. Some additives, flavours, etc if suitable (i.e. cold-prepared, sterile, low in glycation index) 495 may be mixed with the sterilised milk at this point (109). Ascorbic acid or lipoic acid, also glucose and other 6-carbon (reducing) sugars are examples of unsuitable additives. 7. Freeze-dry the foodstuff (at 110) without use of AGE-causing heat. Commercial processes are well known to those skilled in the art. Carry out further quality tests in order to validate the ultrafiltration process. Batches would be tested 111 at an appropriate sampling rate for 500 glycation index-related substances. For example, a manufacturer could test for furosine or use 18 WO 2008/147226 PCT/NZ2008/000125 the FAST index. , to ensure that the product after reconstitution has a furosine amount of preferably less than 10 mg/litre. 8. Package the material under nitrogen 112 and mark it with a short use-by date. 9. Store (113) the material while maintaining it cool (such as at or below zero deg C), and 505 distribute it through retailers or otherwise to consumers in a cold, dry state, under nitrogen, with a short "use-by" date. All other components are similarly cold-prepared. Preferably, periodically monitor (114) the storage conditions of the distributed foodstuff. 10. Include user instructions to reconstitute (115) a pack of the freeze-dried components just before use, at which time other food materials may be added. This is the raw material to be 510 used as the starting point for building an appropriate infant formula using a variety of additives, flavours and the like, by adding water and possibly selected additives. Each pack is sufficient for a single feed, although resealable bulk packs may be provided. The freeze-dried foodstuff may be sold in packs along with separate additives. A list of materials believed to have inherently low glycation indexes yet appropriate for use with the milk-based component 515 would be made available. Note that the process avoids heat sterilisation, especially UHT (ultra-high temperature) pasteurisation and avoids the addition of "anti-oxidants" such as the currently popular ascorbic acid. Alternatively it is possible that plant flavonoid materials such as luteolin are substitute "anti-oxidants". 520 By way of quality control, the content of glycation products should be repeatedly assessed, such as by measuring representative AGEs or other glycation indicators using for example a test for furosine, and/or a test for Amadori -phosphatidyl ethanolamine (PE), a measurement of the FAST index, or an assay for CMLNs-(carboxymethyl) Lysine. The test is to be selected partly on the basis of the time over which the test materials are -presumed to have 525 accumulated. Furosine is an indicator of early glycation products, while FAST tests for advanced glycation products. Cold processed milk has low levels. Freshly harvested raw milk may serve as a control, and batches that have a high AGE content should be discarded. An infant formula made according to the invention should have a FAST index of preferably less than 10, or a furosine content of less than 10 mg/litre. The precision of that limit is not 530 yet established except by reference to test examples. A similar process would be followed for any beverage known for included AGEs, such as the popular carbonated colas. 19 WO 2008/147226 PCT/NZ2008/000125 VARIATIONS 535 The same principles may be applied to improve a wide variety of foods, since there is little current awareness of.the possible relationship between diabetes and glycation products. If that relationship is supported, the principles of the invention may be applied more widely than for foods based on dairy products. A wide range of foods or beverages designed for consumption by diabetic children or adult 540 persons, especially those on dialysis such as for kidney failure (whose dialysis fluids may have high AGE content), -may be made in a similar manner so that the AGE content is as low as possible (and stays as low as possible during storage) and so that the symptoms of those persons are not worsened as they will be (see ref 1) if the foods (or beverages) are high in AGEs. This may reduce the amount of insulin needed, or assist with the dietary restrictions 545 imposed on a Type II sufferer. The AGE(f) content should be included on the packaging. The postulate that low AGEs in a foodstuff is more healthy than high AGEs of say 10 units or more (mg/l by the Furosine test) can also be applied to foods for adults. Ref (1) includes an interesting list of the AGE content of selected foodstuffs, ranging from bananas at 0.1 kU/gram to a range of popular foods in the 22-80 kU/gram range. 550 As previously described in this specification, the foodstuff containing low amounts of glycation (end) products is believed to be significantly less likely to induce or promote the disease of diabetes than a foodstuff high in glycation (end) products. The scope of the invention extends to the manufacture and supply of low-AGE foods and beverages for adult consumption; in particular for consumption by persons already affected by 555 diabetes or diseases involving the -vascular system: such as heart disease, stroke, some auto immune diseases, and the like. INDUSTRIAL APPLICABILITY and ADVANTAGES Put simply, the invention creates foods or beverages having low or zero content of glycation products so that the consumers are not exposed to materials believed to be adverse to health in 560 many ways, especially since glycation products are believed capable of causing diabetes. The inventor believes that a substantial reduction in the incidence of diabetes will follow. 20 WO 2008/147226 PCT/NZ2008/000125 The inventor believes that infants, in particular, should not be exposed to diabetogenic substances in view of the very strong possibility that those substances within infant formulas are in part responsible for the rising incidence- an epidemic - of diabetes. 565 Animal studies, such as the NOD mouse studies cited in (1) show an increased onset of diabetes although it may be rather difficult to "prove" directly that, dietary AGEs in humans cause diabetes. However there does appear to be a high risk that they do. In many instances, if something makes a condition worse, it is likely to also initiate the condition. One might assume that there is a risk that an infant gut is more likely to absorb AGE(f) if present in a 570 diet, than an adult would. In ref (1), trials were carried out on adults only. In trials using normal rats, serum AGEs rose within 2 hours of ingestion. Finally, it will be understood that the scope of this invention as described by way of example and/or illustrated herein is not limited to the specified embodiments. Where in the foregoing description, reference has been made to specific components or integers of the invention 575 having known equivalents, such equivalents are included as if individually set forth. Those of skill will appreciate that various modifications, additions, known equivalents, and substitutions are possible within the scope of the invention as set forth in the following claims. 21