1. School of Biological and Molecular Sciences Glycaemic index and glycaemic load: current concepts and future challenges for the industry Jeya Henry Professor of Human Nutrition 22 June 2005

2. FDIN June 2005 School of Biological and Molecular Sciences Slow realisation of differences in CHO metabolism All foods are composed of protein, fat and carbohydrate By 1911 the importance of protein quality identified (Osborne & Mendel) By 1932 the effect of differences in fat composition were identified (Burr & Burr) It was not until 1981 that metabolic differences in carbohydrates were recognised (Jenkins et al, 1981) Why?

3. FDIN June 2005 School of Biological and Molecular Sciences Plasma glucose response to ingestion of 50 g CHO (Krezowski et al. 1987)

4. FDIN June 2005 School of Biological and Molecular Sciences Glycaemic index The glycaemic index (GI) is a method of ranking foods on a scale according to the extent to which foods raise blood glucose levels after ingestion Carbohydrate foods that break down quickly during digestion have the highest GI values - their blood glucose response is fast and high Carbohydrates which break down slowly, releasing glucose gradually into the bloodstream, have low GI factors

5. FDIN June 2005 School of Biological and Molecular Sciences GI calculation GI of a food: Incremental area under the blood glucose response curve for the test food containing 50 g available carbohydrate Corresponding area after equi-carbohydrate portion of a standard food X 100

6. FDIN June 2005 School of Biological and Molecular Sciences How is GI determined? Typical testing routine involves 12-15 subjects (age range 18-40 years) Subjects arrive at the laboratory after an overnight fast Finger-prick blood sample taken at time zero Subjects consume 50 g glucose in 200 ml water and blood samples taken at 15, 30, 45, 60,90 and 120 mins This procedure is done three times on each subject (on separate days with a minimum gap of 1 day between tests) Subjects then consume 50 g available carbohydrate from the test sample and blood samples are taken at time intervals as above

7. FDIN June 2005 School of Biological and Molecular Sciences GI testing of foods at Oxford Brookes FAO/WHO protocol (Carbohydrates in Human Nutrition, 1998) The GI rating of a food must be tested physiologically in recognised laboratories Oxford Brookes University is working with Tesco and several food companies globally to study the effects of low- and high-GI foods and to develop products that have lower GI ratings Tesco has commissioned Oxford Brookes University to develop QC for GI testing

8. FDIN June 2005 School of Biological and Molecular Sciences Classification Low GI 55 or less Medium GI56-69 inclusive High GI70 or more Does this classification have any biological significance?

9. FDIN June 2005 School of Biological and Molecular Sciences Continuous glucose monitoring Brynes et al. (Br J Nutr, 2005) showed that a low GI diet reduced glucose concentrations by 5% in the daytime and by 16% overnight Mean 24 h area under the curve for glucose at baseline (——) and in response to a diet with low glycaemic index (- - - - -) for eight subjects

10. FDIN June 2005 School of Biological and Molecular Sciences Are foods low in carbohydrate are also low in GI value – NO! Amount of carbohydrate does not predict GI At carbohydrate levels of 10-60% food products show a wide range of GI values At 25% carbohydrate, GI can be between 10 and 95

11. FDIN June 2005 School of Biological and Molecular Sciences Food factors that influence glycaemic response Nature and amount of carbohydrate Nature of the monosaccharide components-glucose -fructose -galactose Nature of the starch-amylose -amylopectin -resistant starch Cooking or food processing-degree of gelatinisation of starch -particle size -food form - cellular structure Other food components-fat and protein -dietary fibre -antinutrients -organic acids

12. FDIN June 2005 School of Biological and Molecular Sciences Food ingredient/macronutrient effect FoodGI (mean  sem) Classification Baked potato 93  8 High Baked potato with cheese 39  5 Low Baked potato with chilli con carne 75  7 High Baked potato with baked beans 62  6 Medium Baked potato with tuna 76  7 High Henry et al. (unpublished data)

13. FDIN June 2005 School of Biological and Molecular Sciences Rate of starch breakdown will influence GI Three main factors influence glucose release 1. Access of enzymes to starch – cell wall as a barrier – protein matrix 2. Susceptibility of starch to hydrolysis – structural breakdown – molecular breakdown 3. Presence of other specific nutrients which affect stomach emptying and viscosity of gut contents

14. FDIN June 2005 School of Biological and Molecular Sciences Peanuts: very low GI (14). Why? Thick, resilient cell walls enclose limited amount of starch

15. FDIN June 2005 School of Biological and Molecular Sciences Three examples of newly developed foods with low-GI

16. FDIN June 2005 School of Biological and Molecular Sciences Glycemic load Glycemic load (GL) assesses the total glycaemic effect of a food/diet Calculated by multiplying the GI by the amount of available CHO in a serving/100 E.g. – GI of food item = 40 – 1 serving = 12 g available CHO – Therefore GL = 12 x 40 / 100 = 4.8 (rounded to 5)

17. FDIN June 2005 School of Biological and Molecular Sciences Glycemic load Glycaemic load represents both the quality and quantity of carbohydrate GL = (GI x amount of CHO in serving [g])/100  ‘QUALITY’ ‘QUANTITY’ Similar to: NDPE% = NPU x P:E%  ‘QUALITY’ ‘QUANTITY’

18. FDIN June 2005 School of Biological and Molecular Sciences GL classification Low GL <10 Medium GL10-20 High GL>20 We currently have no strong biological basis for this classification In contrast, we do have a strong biological basis for GI classification

19. FDIN June 2005 School of Biological and Molecular Sciences Glycaemic INDEX vs Glycaemic LOAD FoodPortion (g) CHO (g) GIGL Bread 1 slice30147010 2 slices60287020 3 slices90427029 Spaghetti Small portion150394216 Medium portion230604225 Large portion350914238 Pizza 1 slice11026369 2 slices220533619 3 slices330793628

20. FDIN June 2005 School of Biological and Molecular Sciences The role of GI and GL on health and well-being Three major diseases in the UK: – Diabetes – Obesity – Cardiovascular disease How does a low-GI/GL diet affect these?

21. FDIN June 2005 School of Biological and Molecular Sciences Global estimates of diabetes (1997) (Amos et al. 1997)

22. FDIN June 2005 School of Biological and Molecular Sciences % population with diabetes in Europe (1997) (Amos et al. 1997)

23. FDIN June 2005 School of Biological and Molecular Sciences Potential mechanism for the relationship between high-GI foods and insulin resistance (Augustin et al. Eur J Clin Nut 2002;56:1049-71)

24. FDIN June 2005 School of Biological and Molecular Sciences Low-GI diets in the management of diabetes Meta-analysis of randomised controlled trials [Brand-Miller et al. Diabetes Care 2003;26:2261- 7] 14 studies included in the final analysis (356 subjects) -203 type 1 diabetes; 153 type 2 diabetes After an average duration of 10 weeks, subjects with type 1 and type 2 diabetes who followed a low-GI diet had lower levels of HbA 1c (0.4%) and fructosamine (0.2 mmol/l) Clinically important reduction in HbA 1c

25. FDIN June 2005 School of Biological and Molecular Sciences High-GI vs low-GI - % change in HbA 1c or fructosamine Favours low-GIFavours high-GI

26. FDIN June 2005 School of Biological and Molecular Sciences Effect of low-GI diet on blood glucose control in type 2 diabetes (Augustin et al. Eur J Clin Nut 2002;56:1049-71)

27. Proposed model linking dietary GI and GL to body weight and obesity-related morbidities Macronutrient composition (carbohydrate amount) Glycaemic index (carbohydrate type) Glycaemic load Alterations in the endocrine profile Partitioning of metabolic fuels Regulation of food intake Body weight Risk for obesity-related morbidities Independent mechanism (insulin demand, insulin sensitivity) Source: Ebbeling & Ludwig (2001)

28. FDIN June 2005 School of Biological and Molecular Sciences Energy balance Energy balance = Energy Intake - Energy Expenditure You will only be in positive energy balance when energy intake exceeds energy expenditure Positive energy balance leads to obesity

29. Male and female obesity levels in selected European countries Collated by the IOTF from recent surveys Yugoslavia Greece Romania Czech Rep. England Finland Germany Scotland Slovakia Portugal Spain Denmark Belgium Sweden France Italy Netherlands Norway Hungary Switzerland % BMI >30 3040300402010 20

30. FDIN June 2005 School of Biological and Molecular Sciences Are low-GI diets effective for weight loss? ReferenceDietSubjectsGreater weight loss? Clapp (1997, 2002) Low- vs high-GI, 24-28 (8 weeks gestation until term) Healthy pregnant women Yes (high-GI babies also overweight) Agus et al. (2000) Low- vs high-GI diet over 9 days 10 moderately overweight men No diff Greater reduction in leptin Speith et al. (2000) Ad lib low-GI vs hypo- energetic low-fat diet for 1 year 107 obese children Yes Bouché et al. (2002) Low- vs high-GI diet for 5 weeks + crossover 11 menNo, but:  in ab fat mass  lean mass trend Sloth et al. (2004) 10-week parallel study45 overweight women No (but appears to be NS tendency)

31. FDIN June 2005 School of Biological and Molecular Sciences Low-GI diet in the treatment of paediatric obesity (Spieth et al. 2000) 107 obese, but otherwise healthy children 43 subjects received hypo-energetic, reduced-fat diet (4.2 months) 64 subjects received low-GI diet (4.3 months) Low-fat group: energy restriction of approximately 250-500 kcal per day compared to their usual intake Low-GI group: emphasis on food selection not energy restriction

32. FDIN June 2005 School of Biological and Molecular Sciences Mean change in BMI by dietary treatment according to tertiles at baseline (Spieth et al. 2000)

33. FDIN June 2005 School of Biological and Molecular Sciences Sloth et al. AJCN 2004;80:337-47 Methods – 45 healthy overweight women – 10-wk parallel, randomised, intervention with 2 groups: high-GI or low-GI – High or low-GI foods (identical in energy, energy density, dietary fibre and macronutrient composition) replaced subject’s usual carbohydrate-rich foods Results – No significant between-group difference in body weight (P=0.31) – 10%  LDLc in low-GI group (P<0.05) – Larger decrease in TC in low-GI group

34. Mean (SEM) body weight changes during 10-wk ad libitum intake of high- or low-GI diet Body weight and body composition changes after 10- wk ad libitum intake of high- or low-GI diet

35. FDIN June 2005 School of Biological and Molecular Sciences Pawlak DB, Kushner JA & Ludwig DS. Lancet 2004;364:778-85 Male rats – High-GI diet (n = 11) – Low-GI diet (n = 10) Controlled diet to maintain bodyweight between the groups for 18 weeks Identical macronutrient composition (69% CHO, 20% protein, 11% fat as percentage of energy) Type of starch – High-GI100% amylopectin – Low-GI 60% amylose / 40% amylopectin

36. FDIN June 2005 School of Biological and Molecular Sciences Body composition changes (week 17) Mean (SE) in group High-GILow-GI Bodyweight (g)547.9 (13.4)549.2 (15.2) Body fat (g) 97.8 (13.6) 57.3 (7.2) Lean body mass (g)450.1 (9.6)491.9 (11.7) Adiposity (%) 17.5 (2.1) 10.3 (1.1) Consumption of a high-GI diet per se adversely affects body composition and risk factors for diabetes and CVD in animal models Adiposity was significantly greater in the high-GI group than in the low-GI group At 18 weeks, high-GI animals had more abdominal fat or “central adiposity” compared to the low-GI animals

37. FDIN June 2005 School of Biological and Molecular Sciences Do low-GI foods reduce food intake at subsequent meals? ReferenceFoodSubjectsResult Barkeling et al. (1995) Pasta vs bread16 elderly menNo diff Benini et al. (1995)High vs low fibre meal8 healthy subjects No (but delayed return of hunger) Holt & Miller (1995)Rice (quick cook and varied amylose content) 9 adultsYes Stewart et al. (1997) Shredded wheat (SW) + glucose; SW + fructose 13 menNo diff Riguad et al. (1998)Psyllium fibre in bfast14 adultsYes, for rest of day Kong et al. (1999)75g fructose vs glucose8 healthy malesNo diff Ludwig et al. (1999)Oatmeal type12 obese teenage boys Yes, EI  81% Ball et al. (2003)Meal replacements16 adolescentsNo diff Warren et al. (2003)Breakfasts (cereal/bread) 37 preadolescents Yes

38. FDIN June 2005 School of Biological and Molecular Sciences Effect of altered GI breakfasts on subsequent food intake (Warren, Henry & Simonite, 2003) Low-GI – All-Bran – non-Swiss-style toasted muesli – traditional porridge – soya and linseed bread Low-GI and added sucrose – sucrose added to provide additional 10% energy High-GI – Corn Flakes – Coco Pops – Rice Krispies – White bread

39. FDIN June 2005 School of Biological and Molecular Sciences Effect of altered GI breakfasts on subsequent food intake (Warren, Henry & Simonite, 2003)

40. FDIN June 2005 School of Biological and Molecular Sciences Effect of altered GI breakfasts on subsequent food intake (Warren, Henry & Simonite, 2003) Lunch intake was significantly lower than after high-GI breakfast: **P<0.01

41. FDIN June 2005 School of Biological and Molecular Sciences The role of low- and high-glycaemic index breakfasts on food intake, appetite and satiety

42. FDIN June 2005 School of Biological and Molecular Sciences Test breakfasts Low-GI breakfast Soya & linseed bread (with low fat spread and diabetic jam) All-Bran ® Porridge – cereals served with either skimmed, semi-skimmed or whole milk All breakfasts served with 190ml unsweetened fruit juice High-GI breakfast Wholemeal bread (with low fat spread and jam) Weetabix ® Shreddies ® – cereals served with whole milk All breakfasts served with 190ml of low sugar squash + teaspoon of glucose powder (to increase the GI)

43. FDIN June 2005 School of Biological and Molecular Sciences Nutritional composition of test breakfasts Mean low-GI breakfast Mean high-GI breakfast Breakfast GI 46 77 Energy (kcal)300299 Protein (g) 10.7 7.6 Fat (g) 6.6 5.2 CHO (g) 49.4 55.5 NSP (g) 5.8 5.2

44. FDIN June 2005 School of Biological and Molecular Sciences Mean 24-hour energy intake for the low- and high- GI breakfasts 24-hour energy intake was significantly lower than after high-GI breakfast: *P<0.05

45. FDIN June 2005 School of Biological and Molecular Sciences Controversies Many of the studies to date have been short- term single meal effects Some studies have failed to match the macronutrient composition of test meals Measures of hunger/satiety are variable The effect of GI in the context of a mixed meal must be considered

46. FDIN June 2005 School of Biological and Molecular Sciences 8-week weight loss study on low-GI diets - foods provided BreadCrusty malted wheat52 Soya & linseed31 Wholemeal pitta56 All-in-One47 CerealsWholewheat muesli56 Hi-fibre bran43 Bran flakes50 BiscuitsSesame crackers50 YogurtLow-fat fruit49 SoupLentil44 Baked beans40

47. FDIN June 2005 School of Biological and Molecular Sciences 8-week weight loss study on low-GI diets - foods provided Basmati rice52 Pasta54 Lasagne47 Pasta bake23 Chicken fajitas42 Chicken chow mein47 Sweet & sour chicken41 Chilli beef noodles42 Fish pie40 Beef & ale casserole53 Mushroom stroganoff26 Dried fruit - Apricots32 - Peaches35 - Pears43 Nuts - Cashews25 - Mixed nuts24

48. FDIN June 2005 School of Biological and Molecular Sciences Measurements at baseline and week 4 Baseline (Mean  SD) Week 4 (Mean  SD) Body weight (kg) 86.7  10.5 84.7  10.3 BMI (kg/m 2 ) 32.4  4.131.7  4.0 Systolic BP (mm Hg) 139  27136  26 Diastolic BP (mm Hg) 83  1379  9 Fasting glucose (mmol/l) 5.5  0.7 5.4  0.9 Cholesterol (mmol/l) 5.0  1.5 4.4  0.8 Triglycerides (mmol/l) 1.3  0.7 1.4  0.5 HDL-cholesterol (mmol/l) 1.3  0.3 1.1  0.2 No. subjects: 12 Age: 46  11 yrs

49. FDIN June 2005 School of Biological and Molecular Sciences Weight profile for the study population SubjectBaseline (kg) Week 4 (kg) Change (kg) 191.689.3-2.3 281.278.6-2.6 3109.2106.2-3.0 486.282.5-3.7 578.176.2-1.9 681.580.2-1.3 788.384.7-3.6 898.798.4-0.3 979.476.6-2.8 1068.568.2-0.3 1189.086.2-2.8 1288.589.6+1.1

50. FDIN June 2005 School of Biological and Molecular Sciences Glycaemic load GL used widely in epidemiological studies (e.g. Salmeron et al. 1997; Schulze et al. 2004; Willet et al. 2002) Recent studies: – BMI positively associated with GI but not with daily carbohydrate intake or GL (Ma et al. 2005) – GL, fructose and sucrose related to an elevated colorectal cancer risk among men, but not women (Michaud et al. 2005) – GL an important independent predictor of HDL cholesterol, accounting for more than 20% of its variation in young people (Slyper et al. 2005)

51. FDIN June 2005 School of Biological and Molecular Sciences Implications of the GI/GL for the food industry Terms such as complex carbohydrates and sugars, which commonly appear on food labels, are now recognised as having little nutritional or physiological significance WHO/FAO recommend that these terms be removed and replaced with the total carbohydrate content of the food and its GI value However, the GI rating of a food must be tested physiologically in RECOGNISED LABORATORIES Whilst GL is the ideal way to express both the quality and quantity of carbohydrate, its use is best served in pre-packed food items of defined portion size Further work is required to understand the biological significance of the GL cut-off points

52. FDIN June 2005 School of Biological and Molecular Sciences Conclusions (I) Low-GI foods produce a smaller rise in blood sugar and can help control established diabetes Low-GI diets can improve the body's sensitivity to insulin Low-GI diets can help people lose weight and lower blood lipids Foods with a low-GI help people control their hunger, their appetite and their blood sugar levels Low-GI foods may be a useful dietary intervention for people suffering with Metabolic Syndrome

53. FDIN June 2005 School of Biological and Molecular Sciences Conclusions (II) Whilst the GI studies seem promising: – Further long-term research (1 year or more) on GI and energy regulation is required as almost all studies have been short-term – More studies need to be done in women and other ethnic groups – GI foods appear to have multiplicity roles: in energy regulation, insulin response and lipid metabolism

54. FDIN June 2005 School of Biological and Molecular Sciences Conclusions (III) The concept of GI is going to revolutionise our nutritional thinking Whilst GI/GL is not a nutritional panacea, mounting evidence suggests that it has a role in weight loss, diabetes and hyperlipidaemia Innovative ways to translate the concept of GI to the average consumer needs to be seriously explored The future challenge is to nurture collaborative activities between food industry, policy makers, FSA/EFSA and academics