Medical Nutrition Therapy (MNT) for Diabetes Evidence based - PDF document

5/28/2013 Medical Nutrition Therapy (MNT) for Diabetes Evidence ‐ based Recommendations Sarah Kim, MD Assistant Clinical Professor, UCSF ‐ SFGH No financial disclosures 1

5/28/2013 Uses of MNT in diabetes To delay and manage To treat obesity and For metabolic control complications of prevent diabetes of diabetes diabetes ADA. Diabetes Care 2008; 31:s1 ‐ s78 Objectives 1. Review evidence supporting the following nutritional practices in relation to glycemic control in diabetes: – Carbohydrate counting & consistency – Carbohydrate restriction – Low glycemic index foods – Fiber 2

5/28/2013 Objectives 2. Review evidence supporting the following nutritional practices in relation to other health outcomes in diabetes: – Mediterranean diet – Low fat, calorie restricted diet (Look AHEAD) 1. Glycemic Control Definitely Beneficial • Carbohydrate Counting • Limiting sucrose • Carbohydrate Consistency • Fiber intake • Calorie restricted, low fat • Mediterranean diet • Low vs. High GI foods • Carbohydrate restriction 3

5/28/2013 I. CARBOHYDRATE COUNTING & CONSISTENCY Diabetes Control and Complications Trial • Although not a primary dietary intervention, the DCCT involved extensive nutritional training in the intensive arm • Techniques used: carbohydrate consistency, carbohydrate counting, the exchange system, and healthy food choices • Greater HbA1c reductions were associated with adherence to dietary advice – Following overall meal plan – Adjusting carbohydrate intake or insulin in response to hyperglycemia – Not over treating lows glucoses Delahanty L. Diabetes Care. 1993;16(11):1453 ‐ 8. 4

5/28/2013 Dose Adjustment for Normal Eating (DAFNE study) Multicenter randomized control trial examining the benefit of 5 ‐ day • outpatient intensive diabetes education with emphasis on carb counting and flexible eating 164 motivated subjects with T1DM and moderate to poor glycemic control • Outcomes: Glycemic control, hypoglycemia rate, QOL • Baseline N=164 Characteristics Age, yr 40 ± 9 Duration diabetes, yr 16.6 ± 9.6 HbA1c 9.4% Retinopathy 37% Neuropathy 17% DAFNE Study Group. BMJ. 2002;325:746 ‐ 751 Outcomes Control arm ‐ 0.5% vs baseline, p=0.001 ‐ 1.0% vs. baseline, p<0.0001 DAFNE Study Group. BMJ. 2002;325:746 ‐ 751 5

5/28/2013 Outcomes 6 months 11 * HbA1c, % 9 Intervention 7 Control Freedom to Eat Overall QOL 5 0 Pre Post Pre Post Pre Post ‐ 1 ‐ 2 * * ‐ 3 ‐ 4 ‐ 5 (less negative = better) ‐ 6 * p<0.05 DAFNE Study Group. BMJ. 2002;325:746 ‐ 751 DAFNE Expansion • 31 centers throughout UK & Ireland, 1163 participants • Retrospective database analysis of 639 participants 1 Year Follow Up 9 No changes in weight, * lipids or blood 8 HbA1c, % 8.51 8.24 pressure 7 6 5 Pre Post Hopkins et al. Diabetes Care. 2012; 35:1638 ‐ 1642 6

5/28/2013 DAFNE Expansion Severe Hypoglycemia 4 Episodes/subject/yr Modest improvements 3 in quality of life, 2 depression and anxiety * 1 * 0 Pre Post Pre Post Hypoglycemia aware Hypoglycemia unaware (sx’s with BG <54) Hopkins et al. Diabetes Care. 2012; 35:1638 ‐ 1642 Carb counting in T2DM 24 week randomized control trial comparing a fixed mealtime insulin • dosing vs. carb counting in patients with insulin ‐ requiring type 2 diabetes Simple Algorithm Carbohydrate Count N=136 N=137 Age 55 yr 55 yr BMI, kg/m2 38 37 HbA1c 8.1% 8.3% Diabetes duration 13 yr 13 yr >2 insulin injections/day 68% 58% Bergenstal et al. Diabetes Care. 2008;31:1305 ‐ 1310 7

5/28/2013 Intervention • Initial insulin doses: – Glargine: 50% of total daily dose before randomization – Glulisine: • 50% of total daily dose divided into 50%/33%/17% according to meal size ‐ OR ‐ • Carb:insulin ratio plus a simple correction factor (“± a few units”) Bergenstal et al. Diabetes Care. 2008;31:1305 ‐ 1310 Intervention • Targets: Fasting <95, pre ‐ prandial <100, bedtime <130 mg/dl • Metformin continued if used at baseline. • Weekly insulin titration by algorithm over the phone • Participants kept diaries (glucoses, insulin dose, carb intake, activity level, low glucoses) Bergenstal et al. Diabetes Care. 2008;31:1305 ‐ 1310 8

5/28/2013 Outcomes Bergenstal et al. Diabetes Care. 2008;31:1305 ‐ 1310 Outcomes Simple Algorithm Group Carbohydrate Count Group Severe 0.89 events/patient ‐ years 0.67 events/patient ‐ years P=0.58 Hypoglycemia (53 episodes in 19 subjects) (37 episodes in 19 subjects) (BG <36) Hypoglycemia 4.9 events/patient ‐ years 8.0 events/patient ‐ years P=0.02 (BG <50) Weight +3.6 kg (3.4%) +2.4 kg (2.3%) P=0.06 BMI +1.28 kg/m2 +0.83 kg/m2 P=0.037 Triglycerides ‐ 8.19 mg/dl (p=0.17) ‐ 13.19 (p=0.008) Bergenstal et al. Diabetes Care. 2008;31:1305 ‐ 1310 9

5/28/2013 Carbohydrate counting/estimating • Carbohydrate counting has glycemic benefits in type 1 diabetes, with additional modest benefits in quality of life • Carbohydrate counting improves glycemic control in type 2 diabetes who need mealtime insulin and may be equally effective as fixed ‐ dose insulin coupled with consistent carbohydrate meals CARBOHYDRATE AMOUNT 10

5/28/2013 Carbohydrate intake and glycemic control • Observational data from large studies involving T1DM 1 and T2DM 2 show an association between higher carb/lower fat intake with lower HbA1c • Interventional studies show conflicting results 1. Delahanty et al. Am J Clin Nutr 2009; 89:518 ‐ 524 2. Xu et al. Am J Clin Nutr 2008; 86:480 ‐ 487 Carbohydrate restricted diets in T2DM Study N Intervention Control Length Samaha 51 37% CHO+41% Fat 51% CHO+33% Fat 6 mo 2003 Samaha et al (6 month study): Stern 109 120g CHO+93g Fat+74g Pro 230g CHO+69gFat+74g Pro 1 yr 2004 Severely obese American subjects, 51 of whom had diabetes • Randomized to Low Carb (30g CHO per day) vs. Low Fat Diet (<30% fat) Westman • 50 19% CHO+59% Fat 44% CHO+36% Fat 6 mo High attrition: 33% in Low Carb, 47% in Low Fat 2008 • Weight loss and TG levels were better in Low Carb group • Wolever 156 40% CHO+40% Fat 50% CHO+25% Fat 1 yr HbA1c dropped in subjects with diabetes: 7.8 ‐ > 7.2% (p=0.06) pp j (p ) 2008 • (low GI) Wolever et al (1 yr study): Haimoto 127 45% CHO+33% Fat 57% CHO+26% Fat 1 yr 2008 156 Canadian subjects with diabetes • Randomized to Low Carb (40% CHO per day) vs. Low GI or Low Fat • Davis 105 20 ‐ 25g CHO 25% Fat 1 yr Haimoto et al (1 year study): No significant differences in HbA1c, lipids, weight 2009 • 127 Japanese subjects with T2DM • Lower diastolic blood pressure ( ‐ 3 mm/Hg) in low carb group • Elhayany 124 35% CHO+45% Fat 55%CHO+30%Fat 1 yr Randomized to simple carb reducing instructions vs. unrestricted diet • 2010 (low GI, high MUFA) (low GI, high MUFA) In Low Carb, HbA1c drop : 7.5 ‐ > 6.7% (p<0.001) and BMI, LDL lowered significantly • Iqbal 144 60g CHO+60g Fat+55g Pro 190g CHO+40g Fat+80g Pro 1 yr 2010 (+Orlistat) Ajala et al. Am J Clin Nutr. 2013:97:505 ‐ 16 11

5/28/2013 Carbohydrate restricted diets in T2DM Ajala et al. Am J Clin Nutr. 2013:97:505 ‐ 16 Carbohydrate amount • In studies conducted thus far, reducing percent of dietary carbohydrates may have a small benefit in glycemic control and lipid profile • Studies are heterogeneous in terms of degree of carbohydrate restriction and difficult to compare 12

5/28/2013 GLYCEMIC INDEX High GI (>70) Low GI ( <55) White rice Brown rice White bread Whole grain bread Potato Tortilla Cornflakes Milk Pizza Beans Carrots Spaghetti www.glycemicindex.com 13

5/28/2013 Glycemic Glycemic Index Load 71 4 41 16 Impact of low GI diet on HbA1c Children with T1DM Elliott et al. Cochrane Database Syst Rev. 2009; 21 14

5/28/2013 Low GI vs High ‐ Cereal Fiber Diet 155 subjects with T2DM randomized to Low GI (69.6) or High GI (83.5) diet Jenkins et al . JAMA 2008; 300(23):2742 ‐ 53. Low GI diet in T2DM Ajala et al. Am J Clin Nutr. 2013:97:505 ‐ 16 15

5/28/2013 Glycemic Index • Choosing low GI foods may have a small benefit in glycemic control in diabetes FIBER 16

5/28/2013 Fiber and glycemic control Most studies are small and of short duration (not shown) • Study Population Duration Intervention Control Findings Milne Type 2 18 month 30g Fiber 21 g Fiber No change in A1c, 1994 Diabetes (21g achieved) (17g achieved) lipids, or weight N=64 Giacco Type 1 6 month 50g Fiber 15g Fiber In hi fiber compliant, 2000 Diabetes (40g achieved) (15g achieved) ‐↓ 0.2% A1c N=54 ‐↓ hypos by 0.8 x/mo ‐ no change in lipids Franz et al. J Am Diet Assoc. 2010;110:1852 ‐ 1889 Fiber and T1DM Lo Fib: HbA1c 8.6  9.1% (p=NS) • Hi Fib: HbA1c 8.8  8.6%, p<0.05 • Giacco et al. Diabetes Care. 2000; 23: 1461 ‐ 1466 17

5/28/2013 Fiber and glycemic control • There is insufficient evidence to recommend high fiber diets to improve glycemic control • Recommend daily intake for all = 14g/1000kcal MEDITERRANEAN DIET 18