The pendulum swings in nutrition are extreme. In Part 1 we critiqued the meta-analyses which generated momentum for the “saturated fats are fine” movement, but there was another suggested conclusion from certain of these meta-analyses: that carbohydrates are to blame, not saturated fats .
These conclusions were timely in the nutrition sphere, coming as they were at a time when a confused nutrition public was being exposed to the cult of “clean eating”, powered by the unprecedented access of totally unqualified “wellness gurus” to reach the masses via social media. And sugar was, and remains, Public Enemy No.1 for the “clean eating” paradigm.
So, what is the role of sugar in cardiometabolic disease? Why did these meta-analyses find “no association” between saturated fat and cardiovascular disease [CVD] when replaced by carbohydrates in the diet?
Comparing Saturated Fat to Carbohydrate in Meta-Analyses
Certain meta-analyses have suggested that replacing saturated fat with carbohydrates does not reduce risk of CVD . Reiterating the point made in the previous article, the quality of the primary included studies will define the results of any meta-analysis. In these meta-analyses, the primary studies included substitution trials which failed to specify the type of carbohydrate replacing saturated fat . This is a significant error. In population studies differentiating between unrefined, wholegrain carbohydrates and refined, simple sugar carbohydrates, substitution of saturated fat for unrefined carbohydrates decreases CVD risk, while substitution with refined carbohydrate either increases risk or at isocaloric levels risk remains unchanged .
Therefore, any pooled analysis that suggests a neutral effect of overall carbohydrate replacement of saturated fat arrives at a null association due to the fact that the studies showing increased risk and decreased risk cancel each other out . In addition, the de Souza et al. (2015) meta-analysis concluded that there was no association between saturated fat, CVD or type-2 diabetes [T2DM] when saturated fat replaces refined carbohydrates. That would be expected given that risk remains relatively similar at isocaloric replacement . Consequently, cohort studies that fail to address the quality of carbohydrate will generally result in a neutral effect of carbohydrate on CVD risk in replacing saturated fats.
These studies add to the confusion, because while they suggest carbohydrate to be a greater issue than saturated fat, they also obscure the true role of sugar in the development of CVD.
Lipoprotein Metabolism and Cardiometabolic Risk Factors
You’ll note I used the term “cardiometabolic” in relation to sugars. This term reflects the fact that CVD is a complex multifactorial disease, for which the influence of saturated fat on blood cholesterol levels is only one aspect. Other metabolic factors increase CVD risk by influencing the “atherogenic lipoprotein phenotype” defined by high blood LDL-cholesterol, low HDL-cholesterol levels, high circulating triglycerides [TGs], and a remodelling of LDL into small, dense lipoprotein subparticles . “Cardiometabolic” refers to these risk factors: central adiposity, insulin resistance, fatty liver, and the influence of these factors on other markers of CVD disease like triglycerides [TGs], HDL and lipoprotein subtypes .
Let’s go a bit deeper into the mechanisms by which CVD risk is increased through adverse effects on these biomarkers. Other than LDL, high blood TGs are an independent risk factor for heart disease even in otherwise healthy weight, normal individuals . LDL cholesterol has two patterns: A-type patterns are defined by large, less dense particles while B-type patterns are defined by remodelling of LDL into small, dense, atherogenic particles . LDL remodels into the B-type pattern when circulating TGs exceed a threshold around 1.5-1.6nM .
This is where we need to grasp some lipid metabolism. The density of a lipoprotein refers to its ratio of proteins to lipids: increasing density means less lipid storage. Chylomicrons [CMs] – which transport TGs absorbed from dietary intake – are 99% lipid, 85-90% TGs and 1-3% cholesterol, while very low-density lipoprotein [VLDL] is 90% lipid, 55% TGs and 25% cholesterol. CMs and VLDL are thus designed primarily for transporting TGs and other lipids. Conversely, HDL only consists of 44% lipids and 6% TGs, as its primary function is the transport of cholesterol for recycling. Neither LDL, and more specifically HDL, are designed to carry much TGs.
When circulating TGs reach that threshold of 1.5-1.6nM, the usual equimolar exchange – where HDL offloads its cholesterol to VLDL in return for equal parts TGs, so VLDL can remove cholesterol to the liver – is impaired, and VLDL offloads too much TGs onto HDL and LDL . HDL is then burdened with TGs, and remodels into small dense subparticles, which are rapidly catabolized in the liver: thus, the low HDL aspect of the phenotype . With impaired TG clearance, LDL also remodels into smaller, denser subparticles, which remain in circulation for prolonged periods and are capable of penetrating the arterial endothelium .
The key point here is that there are two ways in which we end up with elevated circulating TGs: by impaired clearance – i.e. breakdown – of TGs from CMs and VLDL, and/or from increased TG synthesis in the liver. Both of these factors are strongly influenced by free sugars.
Mechanisms of Sugars’ Influence on CVD Risk Independent of Saturated Fat
Circulating TGs increase in response to a high intake of free sugars in both the fasting and post-prandial states . The increase in circulating TGs is driven by accumulation of liver fat, which precipitates insulin resistance in the liver, and causes an overproduction of VLDL .
In the post-prandial period, both hepatic and adipose tissue insulin resistance lead to increased concentrations of free fatty acids in circulation, which upregulates synthesis of new TGs and leads to impaired clearance of TGs . In addition to the increase in TGs synthesis from increased circulating free fatty acids, de novo lipogenesis of liver TGs is increased in response to carbohydrate overfeeding .
Therefore, carbohydrate-induced elevation of TGs in the post-prandial period can occur from a combination of increasing TGs synthesis in the liver, and impaired TGs clearance – which is strongly influenced by a fatty liver. This is central to the cardiometabolic risk of dietary sugars, as impaired TGs metabolism is a primary risk factor for CVD through driving the atherogenic lipoprotein phenotype discussed above . However, this is merely the mechanistic understanding: the relevant question is to what degree these effects occur through the habitual diet.
Controlled Studies on Sugar Intake and Cardiometabolic Risk
Fructose has been a big focus of certain anti-sugar advocates, in particular given the use of high-fructose corn syrup [HFCS] in the US as a sweetener, and the fact that HFCS-sweetened beverages account for 38% of the added sugars in the US diet . In a trial compared diets containing 25% energy from fructose with 25% from glucose, the fructose diet increased visceral fat deposition and dyslipidemia, effects that were not observed from the glucose diet despite similar weight gain between the groups . However, this trial used extremes of intake that are not representative of the typical diet and thus of questionable relevance beyond illustrating mechanistic pathways. Current data from both the National Health and Nutrition Examination Survey [NHANES] in the US and National Diet and Nutrition Survey [NDNS] in the UK shows average free sugar consumption in adults at 12.7% and 12.3%, respectively .
A further study by the same group comparing 145g fructose, 79g HFSC, and 145g glucose found significant increases in cardiometabolic risk factors from HFSC and fructose-sweetened beverages, with no effect from glucose-sweetened beverages . Again, the relevance to these overfeeding studies to CVD risk in the population must be questioned. Dyslipedimia induced by monosaccharide fructose intake shows a dose-response, with 50g/d increasing postprandial TGs and 100g/d increasing plasma circulating TGs . Achieving these thresholds would thus require consumption, respectively, of close to 100g and 200g of sucrose or HFCS. These levels of free sugar intake are not reflected in current average population consumption, although there are extreme quintiles of the population consuming >20% calories from added sugar .
This criticism was confirmed in a recent systematic review of fructose-feeding trials which found that supraphysiological doses of fructose used in trials where fructose was added to the diet averaged 187.3 g/day, and failed to reflect average population levels of intake . This is of particular relevance, as studies in which fructose is substituted isocalorically for other sugars show no adverse effect on cardiometabolic risk, where calories are controlled . A controlled trial comparing hypocaloric weight-loss diets matched for energy intake with 10% calories from either HFSC or sucrose, found improvements in plasma lipids and TGs, illustrating that the presence of 10% energy from free sugars did not negate the effects of weight loss on improving cardiometabolic risk . This is consistent with multiple systematic reviews and meta-analyses demonstrating that increased cardiometabolic risk from free sugars is a result of the contribution to energy excess .
Dietary sugars thus primarily contribute to cardiometabolic risk where they drive an energy excess, and consequent weight gain . This salient point has been conveniently misplaced by the LCHF crowd, who continue to advocate that sugars drive weight gain independent of calories, and that fat does not make you fat (see: Taubes, Gary). However, this is a moot point: isocaloric overfeeding of carbohydrate or fat in humans result in equal weight gain . Independent of macronutrient composition, energy excess drives adiposity .
However, in the first article we identified some of the limitations faced by nutrition science. We’re now faced with another: that controlled dietary intervention studies do not necessarily inform what is happening at a population level. People aren’t living in a randomized controlled trial, they are free-living in the modern obesogenic food environment. And under ad libitum conditions, where people are free to consume whatever they like, added sugar does drive weight gain . This isn’t independent of total calories: the reason is that people consume additional sugar in their diet, without making compensatory adjustments to reduce other dietary intake and maintain energy balance .
The Fallacy of the Single Nutrient Hypothesis, Part 2: Adverse Effects of Sugar
There is no evidence to support a unique cardiometabolic risk of sugar intake independent of total calories and adiposity . There are, however, risks that sugars do increase or add to. The most prominent of these is the effect of fructose on post-prandial lipid metabolism, where 50g per day increases post-prandial circulating TGs, an effect independent of bodyweight .
However, this effect is an interrelationship – between sugars, post-prandial TGs, and the duration of the post-prandial state. In fact, the extended fed state is becoming increasingly recognised as another problematic factor driving obesity and cardiometabolic disease. In the typical Western diet, people spend a majority of their day in post-prandial state . In the post-prandial state, elevations in circulating TGs occur in linear fashion relative to the fat composition of the meal . The process of atherosclerosis has been described as a “post-prandial phenomenon”, reflecting the adverse effects of elevated TGs on lipoprotein remodelling to small, dense subparticles, and the impaired clearance of these lipoproteins resulting in continued exposure to, and penetration of, the arterial space . That lower levels of fructose can increase post-prandial TGs may explain how there continue to be dose-responses observed between added sugars and CVD mortality: compared with consumption of 8-10% added sugars, consumption of 17-21% increased risk for CVD mortality by 38% . However, the linear dose-response of fat content in a meal on increasing post-prandial circulating TGs is arguably a greater issue, when we begin to move away from nutrients and start looking at the foods comprising the typical diet.
This brings us back to the fallacy of the single nutrient hypothesis, and the futility of attempting to separate one dietary component from the other in relation to cardiometabolic risk. High post prandial TGs are influenced both by the fat content of the meal, and the sugar content of the diet.
So, where does this leave us? Where we should have been looking in nutrition science all along: the composition of the diet pattern as a whole.
Diet Is the Sum of Its Parts
When you actually look at the foods in the typical adult Western diet – based on both US and UK adult data – you’ll see that the most significant contributors to calories in the diet are refined flour and potato-based products, cereals, animal fats and added vegetable oil fats . More specifically, we can look at actual food, beverage and environmental factors causative of increased calories from the 1970’s to the 2000’s: potato chips, pizza, sugar-sweetened beverages, French fries, processed animal meats, refined grain products, sweets and desserts, calories consumed away from the home, and increased calories coming from snacks .
That is an energy dense food diet. Where exactly do we start to identify one, single nutrient amidst that diet pattern to blame for cardiometabolic disease? Analyses that have looked at trends in macronutrient intake – the LCHF crowd love to single out increased carbohydrate consumption – are utterly redundant because the primary issue is increased dietary energy density. While technically, data supports that increased energy intake has primarily been attributable to increased carbohydrate consumption, it’s a misnomer: fat only decreased as a percentage of total calories, but fat intake in absolute terms as grams per day increased . In fact, population intake on a macronutrient level is not dissimilar to target levels, yet obesity has risen in the US and UK from 15 and 17% in 1993-1994 to 26% and 34%, respectively, in 2014 .
This suggests that the primary driver is increased food energy supply and energy-density . Substantial evidence exists in support of a greater effect of energy-density on total energy intake than macronutrient variations . The strongest predictor of weight gain in the population, the real driver of obesity, has been – and remains – dietary energy density (Ibid.). Independent of macronutrient composition, the increase in energy in the food supply is sufficient to explain obesity . Attempting to separate one specific nutrient from another in relation to cardiometabolic risk is futile, given the abundant evidence that energy-dense nutrients – which includes both sugars and fats – drive cardiometabolic disease .
This is the crux of the issue: cardiovascular disease remains the leading cause of death in the Western world because it is inherently tied to obesity. Despite the success of reducing blood cholesterol levels in the population [primarily attributable to statin therapy], CVD rates remain where they are due to the adverse cardiometabolic effects of increased adiposity .
While sugar does have adverse cardiometabolic effects, the adverse impact of sugars is observed at high doses which are not typically replicated in the daily diet. Consequently, their primary adverse effect is in contributing to dietary energy density, driving weight gain and associated adverse cardiometabolic outcomes.
There is no longer any scientific validity to single-nutrient hypotheses in the context of the overall food pattern which is problematic for cardiometabolic disease. Singling out sugar is misconceived. Singling out fat is simplistic. The goal for public health is to address the outstanding problem: dietary energy density in the population.
One last point, in the context of clarifying that sugar is not a unique cardiometabolic risk independent of calories and overweight/obesity. I opened by saying the pendulum swings in nutrition are extreme. In the absence of the flexible dieting approach becoming a public health initiative anytime soon, the foods consumed at a population level are problematic. Unfortunately, the hardest challenge of combating “clean eating” and sugar [or fat] alarmism for evidence-based nutrition professionals is in clarifying these issues without becoming apologists for a food industry and environment that has so clearly precipitated diet-induced disease.
What we are trying to achieve is clarity so that a genuine shift can occur to reverse the tide. On a population level, that is a reduction in energy density in the food supply. On an individual level, that translates into a reduction in calories = weight loss = reduction in cardiometabolic disease.
Enough of single-nutrient demonization. Let’s talk about diet. In this respect, we can distill the following summation about sugar:
- There is no unique adverse cardiometabolic effect of sugar independent of total energy intake and adiposity.
- Sugar is problematic on a population level, because free-living subjects do not make compensatory adjustments in calorie intake, i.e. sugar contributes to energy excess.
- This effect does not exist in a vacuum: both sugar and fat are primary constituents of energy-dense foods which have been driving – amongst environmental factors – increased population calorie consumption and obesity.
- Dietary energy density is the primary driver of obesity, and associated cardiometabolic disease.
- Sugar as a reasonable proportion (8-10%) of calorie intake is not an issue.
- Siri-Tarino, P., Sun, Q., Hu, F. and Krauss, R. (2010). Saturated Fatty Acids and Risk of Coronary Heart Disease: Modulation by Replacement Nutrients. Current Atherosclerosis Reports, 12(6), pp.384-390.
- de Souza, R., Mente, A., Maroleanu, A., Cozma, A., Ha, V., Kishibe, T., Uleryk, E., Budylowski, P., Schünemann, H., Beyene, J. and Anand, S. (2015). Intake of saturated and trans unsaturated fatty acids and risk of all-cause mortality, cardiovascular disease, and type 2 diabetes: systematic review and meta-analysis of observational studies. BMJ, p.h3978.
- Li, Y., Hruby, A., Bernstein, A., Ley, S., Wang, D., Chiuve, S., Sampson, L., Rexrode, K., Rimm, E., Willett, W. and Hu, F. (2015). Saturated Fats Compared With Unsaturated Fats and Sources of Carbohydrates in Relation to Risk of Coronary Heart Disease. Journal of the American College of Cardiology, 66(14), pp.1538-1548.
- Mente, A., de Koning, L., Shannon, H. and Anand, S. (2009). A Systematic Review of the Evidence Supporting a Causal Link Between Dietary Factors and Coronary Heart Disease. Arch Intern Med, 169(7), p.659.
- Griffin B.A., & Cunnane S.C. (2009). Nutrition and metabolism of lipids. In M.J. Gibney, S.A. Lanham-New, A. Cassidy, H.H. Vorster (Eds). Introduction to Human Nutrition. West Sussex, UK: Wiley-Blackwell.
- Astrup, A., Dyerberg, J., Elwood, P., Hermansen, K., Hu, F., Jakobsen, M., Kok, F., Krauss, R., Lecerf, J., LeGrand, P., Nestel, P., Riserus, U., Sanders, T., Sinclair, A., Stender, S., Tholstrup, T. and Willett, W. (2011). The role of reducing intakes of saturated fat in the prevention of cardiovascular disease: where does the evidence stand in 2010?. American Journal of Clinical Nutrition, 93(4), pp.684-688.
- Griffin, B. (2015). Saturated fat: guidelines to reduce coronary heart disease risk are still valid. The Pharmaceutical Journal, 294(7858).
- Griffin, B., Freeman, D., Tait, G., Thomson, J., Caslake, M., Packard, C. and Shepherd, J. (1994). Role of plasma triglyceride in the regulation of plasma low density lipoprotein (LDL) subfractions: relative contribution of small, dense LDL to coronary heart disease risk. Atherosclerosis, 106(2), pp.241-253.
- Griffin, B. (2001). The effect of n−3 fatty acids on low density lipoprotein subfractions. Lipids, 36(S1), pp.S91-S97.
- Parks, E. (2002). Changes in fat synthesis influenced by dietary macronutrient content. Proceedings of the Nutrition Society, 61(02), pp.281-286.
- Kotronen, A. and Yki-Jarvinen, H. (2007). Fatty Liver: A Novel Component of the Metabolic Syndrome. Arteriosclerosis, Thrombosis, and Vascular Biology, 28(1), pp.27-38.
- Larsen, L. (2002). The role of de novo lipogenesis in development of obesity in man. British Journal of Nutrition, 88(03), p.331.
- Yang, Q., Zhang, Z., Gregg, E., Flanders, W., Merritt, R. and Hu, F. (2014). Added Sugar Intake and Cardiovascular Diseases Mortality Among US Adults. JAMA Internal Medicine, 174(4), p.516.
- Stanhope, K., Schwarz, J., Keim, N., Griffen, S., Bremer, A., Graham, J., Hatcher, B., Cox, C., Dyachenko, A., Zhang, W., McGahan, J., Seibert, A., Krauss, R., Chiu, S., Schaefer, E., Ai, M., Otokozawa, S., Nakajima, K., Nakano, T., Beysen, C., Hellerstein, M., Berglund, L. and Havel, P. (2009). Consuming fructose-sweetened, not glucose-sweetened, beverages increases visceral adiposity and lipids and decreases insulin sensitivity in overweight/obese humans. Journal of Clinical Investigation, 119(5), pp.1322-1334.
- Centers for Disease Control and Prevention: National Center for Health Statistics (2017). National Health and Nutrition Examination Survey Data. Hyattsville, MD: U.S. Department of Health and Human Services.
- Department of Health (2015). The Scientific Advisory Committee on Nutrition recommendations on carbohydrates, including sugars and fibre. London: Public Health England.
- Stanhope, K., Bremer, A., Medici, V., Nakajima, K., Ito, Y., Nakano, T., Chen, G., Fong, T., Lee, V., Menorca, R., Keim, N. and Havel, P. (2011). Consumption of Fructose and High Fructose Corn Syrup Increase Postprandial Triglycerides, LDL-Cholesterol, and Apolipoprotein-B in Young Men and Women. The Journal of Clinical Endocrinology & Metabolism, 96(10), pp.E1596-E1605.
- Livesey, G. and Taylor, R. (2008). Fructose consumption and consequences for glycation, plasma triacylglycerol, and body weight: meta-analyses and meta-regression models of intervention studies. The American Journal of Clinical Nutrition, 88, pp.1419-37.
- Choo, V. and Sievenpiper, J. (2015). The Ecologic Validity of Fructose Feeding Trials: Supraphysiological Feeding of Fructose in Human Trials Requires Careful Consideration When Drawing Conclusions on Cardiometabolic Risk. Frontiers in Nutrition, 2.
- Khan, T. and Sievenpiper, J. (2016). Controversies about sugars: results from systematic reviews and meta-analyses on obesity, cardiometabolic disease and diabetes. European Journal of Nutrition, 55(S2), pp.25-43.
- Swinburn, B., Sacks, G. and Ravussin, E. (2009). Increased food energy supply is more than sufficient to explain the US epidemic of obesity. American Journal of Clinical Nutrition, 90(6), pp.1453-1456.
- Te Morenga, L., Mallard, S. and Mann, J. (2012). Dietary sugars and body weight: systematic review and meta-analyses of randomised controlled trials and cohort studies. BMJ, 346(jan15 3), pp.e7492-e7492.
- Rippe, J. and Angelopoulos, T. (2016). Sugars, obesity, and cardiovascular disease: results from recent randomized control trials. European Journal of Nutrition, 55(S2), pp.45-53.
- Sies, H., Stahl, W. and Sevanian, A. (2005). Nutritional, dietary and postprandial oxidative stress. J Nutr., 2005(135), pp.969–72.
- Klop, B., Proctor, S., Mamo, J., Botham, K. and Castro Cabezas, M. (2012). Understanding Postprandial Inflammation and Its Relationship to Lifestyle Behaviour and Metabolic Diseases. International Journal of Vascular Medicine, 2012, pp.1-11.
- Mozaffarian, D., Hao, T., Rimm, E., Willett, W. and Hu, F. (2011). Changes in Diet and Lifestyle and Long-Term Weight Gain in Women and Men. New England Journal of Medicine, 364(25), pp.2392-2404.
- Nielsen, S., Siega-Riz, A. and Popkin, B. (2002). Trends in Energy Intake in U.S. between 1977 and 1996: Similar Shifts Seen across Age Groups. Obesity Research, 10(5), pp.370-378.
- Nielsen, S. and Popkin, B. (2003). Patterns and Trends in Food Portion Sizes, 1977-1998. JAMA, 289(4), p.450.
- Ernst, N., Obarzanek, E., Clark, M., Briefel, R., Brown, C. and Donato, K. (1997). Cardiovascular Health Risks Related to Overweight. Journal of the American Dietetic Association, 97(7), pp.S47-S51.
- Rolls, B. (2009). The relationship between dietary energy density and energy intake. Physiology & Behavior, 97(5), pp.609-615.
- Arnett, D., McGovern, P., Jacobs, D., Shahar, E., Duval, S., Blackburn, H. and Luepker, R. (2002). Fifteen-Year Trends in Cardiovascular Risk Factors (1980-1982 through 1995-1997): The Minnesota Heart Survey. American Journal of Epidemiology, 156(10), pp.929-935.
- Lammert, O., Grunnet, N., Faber, P., Bjørnsbo, K., Dich, J., Larsen, L., . . . Quistorff, B. (2000). Effects of isoenergetic overfeeding of either carbohydrate or fat in young men. British Journal of Nutrition, 84(2), 233-245.
- Lowndes, J., Kawiecki, D., Pardo, S., Nguyen, V., Melanson, K., Yu, Z. and Rippe, J. (2012). The effects of four hypocaloric diets containing different levels of sucrose or high fructose corn syrup on weight loss and related parameters. Nutrition Journal, 11(1).