Sugar: the newest and bitterest truth
Robert H. Lustig, M.D., M.S.L. Emeritus Professor Division of Endocrinology, Department of Pediatrics Institute for Health Policy Studies University of California, San Francisco
Adjunct Faculty, UC Hastings College of the Law
Chief Science Officer, EatREAL
Chief Medical Officer, Slendine
Plantrician Project, Anaheim, Sept. 25, 2017 Disclosures Decrease in U.S. Deaths from Heart Disease 1980–2000 offset by Type 2 Diabetes — and they’re not dying! Ford et al, NEJM 356:2388, 2007 T2DM increasing around the world
People with DM (in millions) for 2000, projection for 2010, and % increase
Zimmet et al. Nature 414: 782, 2001 T2DM increasing around the world
People with DM (in millions) for 2000, projection for 2010, and % increase
Zimmet et al. Nature 414: 782, 2001 285 MILLION IN 2010, up 8.9% per yr T2DM increasing around the world
People with DM (in millions) for 2000, projection for 2010, and % increase
Zimmet et al. Nature 414: 782, 2001 412 MILLION IN 2014, up 11.7% per yr T2DM increasing around the world
People with DM (in millions) for 2000, projection for 2010, and % increase Zimmet et al. Nature 414: 782, 2001 WTF?? The money is not going to hospitals, physicians, or Big Pharma
The money is not going to hospitals, physicians, or Big Pharma It’s going to chronic metabolic disease Two inconvenient truths Two inconvenient truths
• There is no medicalized prevention for chronic metabolic disease • There’s just long-term treatment Two inconvenient truths
• There is no medicalized prevention for chronic metabolic disease • There’s just long-term treatment
• You can’t fix healthcare until you fix health • You can’t fix health until you fix the diet • And you can’t fix the diet until you know what is wrong Reasons to eat a plant-based diet
1. Religious Reasons to eat a plant-based diet
1. Religious
2. Animal welfare Reasons to eat a plant-based diet
1. Religious
2. Animal welfare
3. Environmental/Sustainability Reasons to eat a plant-based diet
1. Religious
2. Animal welfare
3. Environmental/Sustainability
4. Cost Reasons to eat a plant-based diet
1. Religious
2. Animal welfare
3. Environmental/Sustainability
4. Cost
5. Metabolic?? “Diabetes is not and never was caused by eating a high carbohydrate diet, and it’s not caused by eating sugar.
The cause of diabetes is a diet that builds up the amount of fat into the blood.
I’m talking about a typical meat-based animal-based diet.
You can look into the muscle cells of the body and they’re building up tiny particles of fat, that’s causing insulin resistance.”
Dr. Neal Barnard, What the Health, 2017 Epidemiology Trends in macronutrient intake and obesity
Carbohydrate
Fat Obesity (BMI>30)
Protein
Based on NHANES data. Int J Obes 1998;22:39-47. JAMA 2002;288:1723. MMWR 2004;53:80-82. Trends in macronutrient intake and diabetes
Carbohydrate
Fat Diagnosed DM
Protein
Based on NHANES data. MMWR 2004;53:80-82. Diabetes Care 2004;27:2806. Prospective Urban and Rural Epidemiology (PURE)
Dehghan et al. Lancet epub Aug 29, 2017 Association of Carbohydrate, Glycemic Index and Load, and Food Groups With T2DM and Insulin Resistance in China
NS
NS
Villegas et al. Arch Intern Med 167:2310, 2007 Prospective Urban and Rural Epidemiology (PURE)
Dehghan et al. Lancet epub Aug 29, 2017 The animal-based food effect on T2DM goes away after you control for iron/heme intake
Talaei et al. Am J Epidemiol DOI: 10.1093/aje/kwx156, 2017
Branched chain amino-acids or choline in red meat may contribute to insulin resistance and inflammation
Trimethylamine Oxide (TMAO), a bacterial metabolite of dietary choline, appears to be pro-inflammatory Newgard et al. Cell Metab 9:311, 2009 Zhu et al. Circulation 135:1671, 2017 Blood levels of dairy saturated fatty acids correlate with protection from T2DM
Yakoob et al. Circulation 133:1645, 2016 Surrogate markers LDL correlates with CVD, but not all that well
5/12/2014 0 (3000×3535)
http://www.sciencedirect.com/cache/MiamiImageURL/1-s2.0-S0168822713002672-gr2_lrg.jpg/0?wchp=dGLzVlV-zSkWA&pii=S0168822713002672 1/1 Ravnskov et al. BMJ Open 6:e010401, 2016 LDL levels and risk for CVD and T2DM
White et al. JAMA Cardiol 1:692, 2016 Hypertriglyceridemia is a better CVD risk factor
In Meta-Analysis (univariate risk) 46,413 Men (16 studies) 2445 Events 8.4 yr f/u
10,864 women (5 studies) 439 events 11.4 yr f/u
holds up in mutltivariate analysis
Austin et al, Am J Cardiol 81:7B, 1998
Atherogenic dyslipidemia— It’s not about LDL cholesterol, it’s about LDL particle number
Cromwell et al. J Clin Lipidol 1:583, 2007 The lipoprotein continuum
VLDL IDL LDLA large buoyant PATTERN “ A ” HDL
LDLA-B intermediate
“ ” PATTERN B LDLB small dense
“Total LDL” won’t tell you particle number -
There’s more LDLB than LDLA at the same total concentration TG and HDL change with LDL sizing
Rizzo and Berneis, Quart J Med 99:1, 2006 Insulin resistance determines liver lipid packaging
Adiposity, insulin resistance TG pool Low cytokines TG-rich sdLDL High FFA HSL LDL
LPL TG-rich HDL sdHDL
TGVLDL -rich VLDL Atherogenic Remnants trapped in the vascular wall while circulating
Krauss and Siri-Tarino, Endocrinol Metab Clin North Am 33:405, 2004 Randomized Clinical Trials Conclusions: A dietary intervention that reduced total fat intake and increased intakes of vegetables, fruits, and grains did not significantly reduce the risk of CHD or stroke.
Howard et al. JAMA 295:655, 2006 LDL-C lowering as primary or secondary prevention of CVD
Randomized controlled trials (RCTs) of drug (41) or dietary (3) interventions
• No overall benefit on mortality
• Most of these trials did not reduce CVD events
• Some of the drug studies reported harm
DuBroff, Evidence-Based Medicine.;22:15, 2017 No improvement in insulin sensitivity or intramyocellular lipid with LDL reduction due to simvastatin
Szendroedi al. Diab Care 32:209, 2009 LDL reduction with medications and incident diabetes
Wang et al. Sci Rep 7:39982, 2017
A high intake of omega-6 fats (vegetable oils) has not been proven as beneficial for our health and trans-fats have been shown to have negative health effects. The higher intake of vegetable oils and the increase in carbohydrate consumption in the last 30-40 years are the two leading factors behind the high rates of obesity and metabolic syndrome in the U.S. Saturated and monounsaturated fats are not. AM National Health Trust, UK RR Editor-in-Chief, JAMA Int Med PM Editor-in-Chief, BMJ Open Heart
Malhotra et al. Br J Sports Med April 26, 2017 doi: 10.1136/bjsports-2016-097285 Nature 487:27-29, Feb 1, 2012
New York Times, April 17, 2011 Hyperbole?
Nature 487:27-29, Feb 1, 2012
New York Times, April 17, 2011 Toxicity:
The degree to which a substance can damage an organism
• Does not distinguish acute vs. chronic toxicity
Requisites:
• Must be an “independent risk factor”
• Exclusive of calories
• Exclusive of obesity
• Must establish causation Criticisms of Fructose Toxicity
• Animal models, not human studies
• Administration of excessive doses of fructose
Criticisms of Fructose Toxicity
• Animal models, not human studies
• Administration of excessive doses of fructose
WILL LIMIT DISCUSSION TO: HUMAN DATA, HUMAN CONSUMPTION, IN DOSES ROUTINELY INGESTED AND CONTROLLED FOR CALORIES AND OBESITY Sugar and Diabetes
— Plausibility
— Mechanisms
— Human Correlation
— Human Causation Sugar and Diabetes: Plausibility Histology of (N)AFLD
Normal (N)AFLD MRI Fat Fraction Maps
Obese Low Liver Fat = 2.6% MRI Fat Fraction Maps
Obese Low Liver Fat = 2.6% MRI Fat Fraction Maps
Obese Obese Low Liver Fat = 2.6% High Liver Fat = 24% MRI Fat Fraction Maps
Obese Obese Low Liver Fat = 2.6% High Liver Fat = 24% MRI Fat Fraction Maps
Obese Obese Thin Low Liver Fat = 2.6% High Liver Fat = 24% High Liver Fat = 23% MRI Fat Fraction Maps
Obese Obese Thin Low Liver Fat = 2.6% High Liver Fat = 24% High Liver Fat = 23% NAFLD and Metabolic Syndrome are congruent (if not the same)
Adults: Marchesini et al. Hepatology 37:917, 2003
Children: Schwimmer et al. Circulation 118:277, 2008 Epidemiology of NAFLD
Non-alcoholic fatty liver disease (NAFLD) has become epidemic
Steatosis: 45% Latinos 33% Caucasians 24% African Americans
NASH 5.5% of US Adults
Children: Steatosis in 13% of autopsy specimens ages 5-19 38% in obese autopsy specimens
Browning et al. Hepatology 40:1387, 2004; Schwimmer et al. Pediatrics 118:1388, 2006 NAFLD is a primary predictor of T2DM in Korean adults
Sung and Kim, J Clin Endocrinol Metab 96:1093, 2011 Intrahepatic fat explains insulin resistance better than visceral fat
Hepatic Insulin Sensitivity VLDL Index Secretion Rate
Insulin Stimulated Glucose Disposal Contribution Rate Of Free Fatty Acids Insulin To VLDL Stimulated Palmitate Suppression Rate
Fabbrini et al. Proc Natl Acad Sci 106:15430, 2009 Sugar and Diabetes: Mechanisms (20%)
(80%) Can you name an energy source that is:
Can you name an energy source that is:
Not necessary for life
Can you name an energy source that is:
Not necessary for life
There is no biochemical reaction in the body that requires it
Can you name an energy source that is:
Not necessary for life
There is no biochemical reaction in the body that requires it
Is not nutrition
Can you name an energy source that is:
Not necessary for life
There is no biochemical reaction in the body that requires it
Is not nutrition
When consumed in excess it is toxic
Can you name an energy source that is:
Not necessary for life
There is no biochemical reaction in the body that requires it
Is not nutrition
When consumed in excess it is toxic
We love anyway, and it’s addictive
Can you name an energy source that is:
Not necessary for life
There is no biochemical reaction in the body that requires it
Is not nutrition
When consumed in excess it is toxic
We love anyway, and it’s addictive
Answer: Ethanol (80%) (100%)
Variation of HDL and triglyceride levels based on consumption of added sugars in NHANES adults
Welsh et al. JAMA 303:1490, 2010 Meta-Analysis of Effects of Sugar on Triglycerides
Te Morenga et al. Am J Clin Nutr doi: 10.3945/ajcn.113.081521, May 7, 2014 Sugar and Diabetes
(Prospective Correlation) Fruit juice consumption increases risk for T2DM in women (Nurse’s Study)
RR 1.18 [1.10-1.26]
Bazzano et al. Diab Care 31:1311, 2008 SSB’s and BMI-adjusted risk of diabetes in EPIC-Interact (Europe)
Romaguera-Bosch et al. Diabetologia 56:1520, 2013 Associations between consumption of sugar sweetened beverages and fruit juice and incident type 2 diabetes: meta-analysis of prospective cohort studies
Sugar Not Adjusted I2 Adjusted for I2 Sweetened for Adiposity: Adiposity: Beverages Relative Risk Relative Risk (n=17)
Meta-analysis, crude: 1.25 (1.14 to 1.37) 89 — —
+multivariable adjusted 1.18 (1.09 to 1.28) 89 1.13 (1.06 to 1.21) 79
+calibration for 1.43 (1.20 to 1.70) 86 1.28 (1.12 to 1.46) 73 information bias +calibration for 1.42 (1.19 to 1.69) 85 1.27 (1.10 to 1.46) 73 publication bias
Fruit Juices Not Adjusted I2 Adjusted for I2 (n=13) for Adiposity: Adiposity: Relative Risk Relative Risk
Meta-analysis, crude: 0.97 (0.90 to 1.06) 79 —
+multivariable adjusted 1.05 (0.99 to 1.11) 58 1.07 (1.01 to 1.14) 51
+calibration for 1.06 (0.98 to 1.14) 49 1.10 (1.01 to 1.20) 29 information bias +calibration for Not detected — Not detected — publication bias
Imamura et al. BMJ 351:.h3576, 2015 *¥ *
* Prevalence Odds Ratio Prevalence
30g Added Sugar Quintiles - Mean sugar (g/day) 1 Adjusted for age, BMI z-score, energy intake, and physical activity * Statistically significant compared to 1st quintile (p < 0.05) ¥ Statistically significant compared to 2nd quintile (Wald test, p < 0.05 ) Rodriguez et al., Public Health Nutr 40:2424, 2016 Sugar and Diabetes
(Econometric Analysis) An international econometric analysis of diet and diabetes
Diabetes prevalence rose from 5.5% to 7.0% for 175 countries 2000-2010
Basu et al. PLoS One, e57873, 2013 An international econometric analysis of diet and diabetes
Diabetes prevalence rose from 5.5% to 7.0% for 175 countries 2000-2010
Model # countries Effect (95% CI)
Sugar
Sugar+controls
Sugar+controls+period
Overall
Basu et al. PLoS One, e57873, 2013 An international econometric analysis of diet and diabetes
Basu et al. PLoS One, e57873, 2013 An international econometric analysis of diet and diabetes Only changes in sugar availability predicted changes in diabetes prevalence
Every extra 150 calories increased diabetes prevalence by 0.1%
But if those 150 calories were a can of soda, diabetes prevalence increased 11-fold, by 1.1% (95% CI 0.03 — 1.71%, p <0.001)
This study meets the Bradford Hill criteria for Causal Medical Inference: —dose —duration —directionality —precedence
We estimate that 25% of diabetes worldwide is explained by sugar Basu et al. PLoS One, e57873, 2013 Sugar and Diabetes
(Causation) Sugar is toxic unrelated to calories
Lustig et al. Obesity 24:453, 2016
Gugliucci et al. Atherosclerosis 253:171, 2016
Schwarz et al. Gastroenterology 153:743, 2017 September 2017 cover article Strategy
• Isocaloric fructose restriction x 9 days in children who are habitual sugar consumers • No change in weight • Substitute complex carbs for sugar • Maintain baseline macronutrient composition of the the diet • Study in PCRC at Day 0 and Day 10 • Assess changes in organ fat, de novo lipogenesis, and metabolic health
Oral glucose tolerance test before and after isocaloric fructose restriction
Lustig et al. Obesity 24:453, 2015 Fasting Labs
Day 0 Day 10 β-coefficient p value (Adjusted Change) [95% CI] Heart rate (bpm) 83.1 ± 10.7 80.1 ± 11.3 -2.8 [-6.5, +0.9] 0.13
Systolic BP (mmHg) 122.6 ± 10.5 121.1 ± 9.9 - 1.39 [-4.9, +2.1] 0.43
Diastolic BP 68.8 ± 8.9 63.7 ± 7.5 - 4.9 [-8.1, -1.8] 0.003
Fasting lactate 1.2 ± 0.4 0.9 ± 0.3 -0.3 [-0.5, -0.2] <0.001 (mmol/L)
Lactate AUC 160.0 ± 34.5 129.0 ± 34.5 -31.2 [-41.9, -20.5] <0.001 (mM/120 min) HOMA-IR¥ 7.9 ± 4.8 5.2 ± 2.6 -2.7 [-3.8, -1.5] <0.001 AST (U/L) * 27.4 ± 14.1 23.8 ± 8.9 0.02
ALT (U/L) ¥ 28.9 ± 22.8 26.7 ± 19.6 -2.2 [-4.7, +0.3] 0.09
Fasting TG (mM) 1.4 ± 0.9 1.0 ± 0.5 -0.4 [-0.6, -0.2] 0.002
Fasting LDL-C (mM) 2.4 ± 0.6 2.1 ± 0.6 -0.3 [-0.4, -0.1] <0.001
Fasting HDL-C (mM) 1.2 ± 0.2 1.0 ± 0.2 -0.1 [-0.2, -0.1] <0.001
Fasting FFA (mM) 0.6 ± 0.2 0.7 ± 0.2 +0.1 [+0.1, +0.2] <0.001 DNL is the Conversion of Dietary Carbohydrates into Lipids
*) *) *) *) *) * *) *) *) *) *) * *) *) *) *) *) *) *) *) *) Sugar Into Fat (lipids)
Fructose Palmitate *) *) * * * Acetate
New Tracer Method using MIDA: Hellerstein and Neese, AJP 1999
DNL AUC Pre and Post Fructose Restriction
160 140 120 100 80 60 40 20 0 Day 0 Day 10
Schwarz et al. Gastroenterology 153:743,, 2017 Changes in liver, visceral, and subcutaneous fat (n = 37)
Schwarz et al. Gastroenterology 153:743,, 2017 Triglyceride-rich Lipoprotein DNL AUC (n=17)
14.00%
12.00%
10.00%
8.00%
DNL AUC – 6.00%
4.00%
Fractional TRL TRL Fractional 2.00%
0.00% 0 1 2 3 4 5 6 7 8 hour
D0 D10
Schwarz et al. Gastroenterology 153:743,, 2017 LDL subclasses A B C Day 0 Day 0 VLDL MID LDL1 LDL2 LDL3 (sd-LDL)VLDL MID LDL1 LDL2 LDL3 (sd-LDL) HDL subclasses Day 0 Large Intermediate Small
AUC % Day 10 Day 10 AUC % Day 10
AUC %
AUC %
Gugliucci et al. Atherosclerosis 253:171, 2016 Changes in Lipoprotein Subfractions
Change n Day 0* Day 10* (adj. for p value weight) LDL-1** (%) 37 30.1 ± 10.7 26.5 ± 9.3 - 26% 0.009
LDL-2† (%) 37 16.5 ± 13.5 11.8 ± 8.3 NS
sdLDL (3) (%) 10 1.9 ± 1.7 0.6 ± 0.7 -1.3 0.04 LDL Size (nm) 37 271.3 ± 3.1 272.2 ± 2.5 + 0.87 0.008
Small HDL (%) 37 14.6 ± 6.1 11.8 ± 5.5 - 2.73 0.001 Large HDL (%) 37 26.8 ± 7.9 29.3 ± 7.7 + 2.42 0.04
Apo-B** (mg/dl) 37 78 ± 24 66 ± 24 - 32% <0.001
ApoC-II** (mg/dl) 37 8.7 ± 3.7 8.3 ± 4.2 - 15% 0.19
ApoC-III** (mg/dl) 37 8.7 ± 3.5 6.5 ± 2.6 - 49% <0.001 TG / HDL ratio** 43 3.1 ± 2.5 2.4 ± 1.4 - 38% 0.02
** not normally distributed, log transformed, percent change, † non-parametric, Kruskal-Wallis
Gugliucci et al. Atherosclerosis 253:171, 2016 Changes in triglyceride and VLDL lipoproteins correlate with change in insulin sensitivity (Spearman R)
DHOMA-IR DC-peptide DFasting insulin DApo-B (mg/dl) 0.43 (0.01) 0.51 (0.001) 0.41 (0.01)
DApoC-II (mg/dl) 0.45 (0.005) 0.54 (<0.001) 0.48 (0.003)
DApoC-II! (mg/dl) 0.40 (0.01) 0.43 (0.01) 0.41 (0.01)
DTG / HDL ratio 0.31 (0.04) 0.50 (<0.001) 0.34 (0.03)
Gugliucci et al. Atherosclerosis 253:171, 2016 What the data say • Epidemiologic and randomized controlled trials do not support a positive association between animal-based fat, LDL, and diabetes
• Red meat may be promotive of diabetes (likely due to iron/oxidative stress, BCAA, or TMAO); yet dairy appears to be protective — does not suggest that insulin resistance is due to saturated fat.
• Prospective correlational data demonstrate associations between added sugar and and diabetes, exclusive of calories or obesity
• Econometric data show causal medical inference for added sugar and diabetes, exclusive of calories or obesity
• Interventional isocaloric glucose for fructose exchange shows improvements in fatty liver disease, insulin resistance, metabolic health, and lipids in children in just 10 days Foodstuffs and metabolic syndrome
• Transfats • Branched chain amino acids • Ethanol • Fructose
• Liver is the only site for energy metabolism • Not insulin regulated • No glycogen popoff, mitochondria are overwhelmed
Bremer et al., Pediatrics 129:557. 2012 Sugar and disease
• Causation – Diabetes – Heart Disease – Fatty Liver Disease – Tooth Decay
• Correlation – Cancer – Dementia Lustig, JAMA Pediatrics 171:212, 2017
PROCESSED food is high-sugar, low fiber
PROCESSED food is high-sugar, low fiber
REAL food is low-sugar, high-fiber
PROCESSED food is high-sugar, low fiber
REAL food is low-sugar, high-fiber
All diets that work are REAL food
PROCESSED food is high-sugar, low fiber
REAL food is low-sugar, high-fiber
All diets that work are REAL food
Oh, and by the way, sugar is plant-based
The Sugar Association Statement on Kearns JAMA Study
by The Sugar Association September 12, 2016
We acknowledge that the Sugar Research Foundation should have exercised greater transparency in all of its research activities, however, when the studies in question were published funding disclosures and transparency standards were not the norm they are today. The Sugar Association Statement on Kearns JAMA Study
by The Sugar Association September 12, 2016
Generally speaking, it is not only unfortunate but a disservice that industry-funded research is branded as tainted.
Most concerning is the growing use of headline-baiting articles to trump quality scientific research—we’re disappointed to see a journal of JAMA’s stature being drawn into this trend. And sugar is a primary cause of liver fat and insulin resistance
Collaborators
UCSF UCSF Clinical/Translational Ayca Erkin-Cakmak, M.D. M.P.H. Science Institute Andrea Garber, Ph.D., R.D. Laura Schmidt, Ph.D. Patrika Tsai, M.D., M.P.H. Claire Brindis, Dr.P.H. Emily Perito, M.D., M.P.H. Cristin Kearns, D.D.S. Stanton Glantz, M.D. Touro University Dept. of Biochemistry Jean-Marc Schwarz, Ph.D. UC Hastings College of the Law Alejandro Gugliucci, Ph.D. David Faigman, J.D. Marsha Cohen, J.D. SFGH Depts. of Medicine & Radiology Susan Noworolski, Ph.D. UC Berkeley Dept. of Nutr. Sciences Kathleen Mulligan, Ph.D. Pat Crawford, R.D., Ph.D. Kristine Madsen, M.D., M.P.H. Stanford Prevention Institute Lorrene Ritchie, Ph.D. Sanjay Basu, M.D., Ph.D. Paula Yoffe, B.A.
Union of Concerned Scientists Health Evolution Partners Environmental Working Group David Brailer, Ph.D. Dietitians for Professional Integrity