Fructose Restriction Studies in Children

Table of Contents

1. 1. Why Lustig Designed a Fructose Restriction Study
2. 2. How the Nine-Day Study Was Designed
3. 3. The Children Who Took Part
4. 4. Blood Pressure, Blood Fats and Insulin
5. 5. Liver Fat and Fat-Making in the Liver
6. 6. Lipoprotein Particles and Apolipoprotein C-III
7. 7. D-Lactate and the Sugar-Damage Pathway
8. 8. Ceramides and Telomeres
9. 9. The Weight-Change Question in the Published Record
10. 10. How Lustig and Others Presented the Results
11. 11. Related Research on Fructose and the Liver
12. Key Research Papers
13. Connections

1. Why Lustig Designed a Fructose Restriction Study

By the time the study began, Lustig had already set out his fructose argument in his 2009 lecture “Sugar: The Bitter Truth” and in a series of review papers. In a 2010 review he described three parallels between fructose and alcohol (ethanol): both are turned into fat in the liver through de novo lipogenesis (the liver’s process of building new fat from sugar), which he links to insulin resistance, abnormal blood fats and fatty liver; fructose attaches to proteins (fructosylation) and generates superoxide, which he compares to the damage done by acetaldehyde, a breakdown product of alcohol; and fructose stimulates the brain’s “hedonic pathway”, creating habituation and “possibly dependence”.

In a 2013 review he summed the argument up in its title, “Fructose: it’s ‘alcohol without the buzz’”. There he writes that glucose raises insulin and is “not particularly sweet”, while fructose “does not generate an insulin response, and is very sweet”, and that in a well-fed state its breakdown products “overwhelm hepatic mitochondrial capacity”, pushing the liver toward fat-making and insulin resistance.

The obstacle, as the trial paper explains in its opening line, is that earlier human studies of fructose were hard to read. The authors write that intervention studies of dietary fructose “are confounded by positive caloric balance, changes in adiposity, or artifactually high amounts” — in other words, volunteers were often given extra calories, gained or lost body fat, or received far more fructose than people normally eat. The nine-day study was built to remove those three problems: the children kept their usual calories, were meant to keep their weight steady, and simply had sugar replaced by starch.

Back to Table of Contents

2. How the Nine-Day Study Was Designed

The design facts below come from the published papers. The study is registered on ClinicalTrials.gov as NCT01200043.

• Meals provided. For nine days the children received all their meals, prepared to match the share of protein, fat and carbohydrate in the diet each child reported eating at home.
• The one change. Dietary sugar was cut from 28% to 10% of calories and replaced with starch. The 2017 liver paper reports that this left fructose at 4% of total calories. The 2016 lipoprotein paper gives the fructose figures for its subgroup as 11.7% of daily calories before and 3.8% after, the missing calories being supplied as glucose in starch. The 2019 D-lactate paper describes the substitute as refined starch.
• Calories held steady (isocaloric). The children recorded their weight every day, and calories were adjusted with the aim of keeping weight the same.
• Before-and-after measurements. Tests were done on day 0 and on day 10: blood tests, an oral glucose tolerance test, a dual-energy X-ray absorptiometry (DEXA) body-composition scan and, in the liver analysis, magnetic resonance spectroscopy and imaging of liver and abdominal fat, with stable isotope tracers to measure fat-making in the liver.
• One group, compared with itself. There was no separate control group eating the usual diet. The 2017 paper states that “Paired analyses compared change from day 0 to day 10 within each child.”
• Allowing for weight change. The 2016 main paper states that its blood results were “controlled for weight change by repeated measures ANCOVA” (analysis of covariance, a statistical method that adjusts a result for another measurement).

PubMed classes the study as a clinical trial without randomisation: every child received the same diet, and each child’s day-10 results were compared with that child’s own day-0 results.

Back to Table of Contents

3. The Children Who Took Part

The 2016 main paper enrolled 43 children with obesity and metabolic syndrome (a cluster of high blood pressure, high blood fats, high blood sugar and abdominal fat): 27 Latino and 16 African-American children. The 2017 liver paper describes the children as 9 to 18 years old and as having “habitual high sugar consumption”, defined as fructose intake above 50 grams a day.

Not every analysis used every child. The numbers reported in each paper’s abstract are:

• Blood pressure, blood fats and insulin (2016, Obesity): 43 children.
• Liver fat, visceral fat and fat-making in the liver (2017, Gastroenterology): 41 children.
• Lipoprotein particles and apolipoproteins (2016, Atherosclerosis): 37 children.
• D-lactate (2019, Journal of Clinical Endocrinology and Metabolism): 20 children.
• Ceramides (2022, Nutrients): 43 participants.
• Telomere length (2025, Diabetology and Metabolic Syndrome): 43 children, described as Black and Latinx.

The study was approved by the UCSF Committee on Human Research and the Touro University Institutional Review Board, according to the 2025 paper’s ethics statement.

Back to Table of Contents

4. Blood Pressure, Blood Fats and Insulin

The first paper, published in Obesity in 2016 with Lustig as first author, reported the headline results for all 43 children after nine days:

• Diastolic blood pressure (the lower number) fell by 5 mmHg.
• Lactate in the blood fell by 0.3 mmol/L.
• Triglycerides (the main fat carried in the blood) fell by 46%.
• Low-density lipoprotein (LDL) cholesterol fell by 0.3 mmol/L.
• Glucose tolerance and hyperinsulinemia (too much insulin in the blood) both improved.
• Body weight fell by 0.9 kg (plus or minus 0.2 kg), and fat-free mass by 0.6 kg.

The authors ran a further, after-the-fact (post hoc) check on the 10 children who did not lose weight and report that their results “were directionally consistent” with those of the whole group. Their conclusion reads: “Isocaloric fructose restriction improved surrogate metabolic parameters in children with obesity and metabolic syndrome irrespective of weight change.”

The word “surrogate” is the paper’s own: blood pressure, blood fats and insulin are markers linked to later disease, rather than diseases themselves. The study did not run long enough to measure outcomes such as diabetes or heart disease.

Back to Table of Contents

5. Liver Fat and Fat-Making in the Liver

The second paper, published in Gastroenterology in 2017 with Jean-Marc Schwarz as first author and Lustig among the senior authors, looked inside the liver. Its stated aim was to measure the effect of the nine days on de novo lipogenesis — which the paper calls “the conversion of fructose to fat in liver” and describes as possibly “a modifiable pathogenetic pathway” — and on liver fat, visceral fat (fat around the organs), fat under the skin and insulin handling.

The paper reports these median changes from day 0 to day 10 in 41 children:

• Liver fat fell from 7.2% to 3.8%.
• Visceral fat fell from 123 cm³ to 110 cm³.
• De novo lipogenesis, measured as the area under the curve with stable isotope tracers, fell from 68% to 26%.
• Insulin kinetics (how insulin is released and cleared during a glucose test) improved.

The authors add that “These changes occurred irrespective of baseline liver fat” — children who started with little liver fat and those who started with a great deal both showed the change. Their conclusion: “Short-term (9 days) isocaloric fructose restriction decreased liver fat, VAT, and DNL, and improved insulin kinetics in children with obesity. These findings support efforts to reduce sugar consumption.” (VAT stands for visceral adipose tissue and DNL for de novo lipogenesis.)

The paper’s conflict-of-interest statement records that Lustig “wrote a book on obesity for the general public in 2012”.

Back to Table of Contents

6. Lipoprotein Particles and Apolipoprotein C-III

The third analysis, published in Atherosclerosis in 2016 with Alejandro Gugliucci as first author, examined the particles that carry fat and cholesterol through the blood. In 37 children it reported, from day 0 to day 10:

• Apolipoprotein C-III (apoC-III), a protein on fat-carrying particles that the paper links to atherogenic (artery-clogging) high triglycerides, fell from 8.7 to 6.5 mg/dl — the change the authors call “the most dramatic reduction”.
• Apolipoprotein B (apoB) fell from 78 to 66 mg/dl, and apolipoprotein E (apoE) from 4.6 to 3.6 mg/dl.
• Low-density lipoprotein particle size increased by 0.87 Ångström.
• Small dense low-density lipoprotein was present in 25% of the children and fell by 68% in them.
• High-density lipoprotein (HDL): small high-density lipoprotein fell by 2.7% and large high-density lipoprotein rose by 2.4%.
• The triglyceride-to-HDL-cholesterol ratio fell from 3.1 to 2.4.

The authors report that these changes in fasting blood fats tracked changes in insulin sensitivity, and conclude that nine days of isocaloric fructose restriction “improved lipoprotein markers of CVD risk” (cardiovascular disease risk) in these children.

Back to Table of Contents

7. D-Lactate and the Sugar-Damage Pathway

The fourth analysis, published in the Journal of Clinical Endocrinology and Metabolism in 2019 with Erkin-Cakmak as first author, followed a second line of Lustig’s argument: that sugar harms the body not only by being turned into fat but also by nonenzymatic glycation — sugar molecules sticking to proteins without an enzyme, the same family of reactions he had compared with alcohol damage.

The marker chosen was serum D-lactate, which the paper describes as “the detoxification end-product of methylglyoxal”, a reactive by-product of sugar breakdown. In 20 children, the paper reports:

• At the start, D-lactate was linked with liver fat and visceral fat, but not with fat under the skin.
• At the start, D-lactate rose with de novo lipogenesis, liver fat, triglycerides and the triglyceride-to-HDL ratio.
• After nine days, serum D-lactate fell by 50%, and the change in D-lactate tracked the changes in de novo lipogenesis and in insulin sensitivity.

The authors conclude that these findings “suggest that DNL and nonenzymatic glycation are functionally linked via intermediary glycolysis in the pathogenesis of metabolic syndrome and point to fructose as a key dietary substrate that drives both pathways.”

Back to Table of Contents

8. Ceramides and Telomeres

Ceramides (2022)

A 2022 analysis in Nutrients, led by Olson with Mietus-Snyder as senior author, measured ceramides — fat-based signalling molecules (sphingolipids) that the paper links to fat toxicity, insulin resistance and cardiometabolic risk. In data from 43 participants it reports that mean weight fell by 0.9 kg (plus or minus 1.1 kg) and that most total and individual ceramide levels fell significantly, including dihydroceramides, deoxyceramides and ceramide-1-phosphates. The change in each main ceramide type moved in the opposite direction to a composite insulin-sensitivity index, and the change in deoxyceramides moved with the change in de novo lipogenesis. The authors write that ceramides “may represent an intermediary link between hepatic DNL, insulin resistance, and CMR” (cardiometabolic risk).

Telomere length (2025)

A 2025 analysis in Diabetology and Metabolic Syndrome, led by Janet Wojcicki with Lustig as senior author, measured leukocyte telomere length — the protective caps on the chromosomes of white blood cells, which shorten with inflammation and oxidative damage. Across the nine days the children’s telomeres shortened on average by 0.05 T/S units, or 2.98% (with wide variation, plus or minus 8.74%). Higher starting levels of high-density lipoprotein cholesterol, apolipoprotein A-I and large high-density lipoprotein were associated with less telomere shortening by day 10, while rises in apolipoprotein A-I during the study were associated with more shortening. The authors report no link with other, non-HDL blood fats, and write that “Additional, larger studies are necessary” to understand how short-term fructose restriction, telomere change and high-density lipoprotein relate.

Back to Table of Contents

9. The Weight-Change Question in the Published Record

The study aimed to hold weight steady, but the 2016 main paper reports that the children’s weight fell by 0.9 kg over the nine days, including 0.6 kg of fat-free mass. Because weight loss on its own can improve blood fats and insulin, the authors addressed this in two ways described in the paper: their blood results were statistically adjusted for weight change, and a separate check of the 10 children who did not lose weight found results “directionally consistent” with the whole group.

The question was also raised in print. In March 2016, Khan and Sievenpiper of the University of Toronto published a letter in Obesity titled “Metabolic improvement with fructose restriction: Is it the fructose or the weight loss?” In the same issue Lustig published a reply, “Response to ‘Metabolic improvement with fructose restriction: Is it the fructose or the weight loss?’”. Both items appear in PubMed as comments on the trial paper; the letter and the reply are the authors’ published exchange on this point.

The study’s design is stated in the papers: one group of children, nine days, starch in place of sugar, results compared within each child from day 0 to day 10, with no separate control group.

Back to Table of Contents

10. How Lustig and Others Presented the Results

On his own website’s fructose restriction page, Lustig headlines the study “Obese Children’s Health Rapidly Improves With Sugar Reduction Unrelated to Calories”. Of the 2019 D-lactate finding he writes there: “Short version: SUGAR IS TOXIC! And here’s the toxin.” These are his words, as published on robertlustig.com.

The study drew comment in the journals. Nature Reviews Endocrinology ran a research highlight on the 2016 paper titled “Nutrition: Breaking the fructose habit”. The 2017 Gastroenterology paper was accompanied by an editorial by Vos and Goran titled “Sugar, Sugar . . . Not So Sweet for the Liver”. The 2016 Atherosclerosis paper is listed in PubMed with three published comments.

The trial results also fit into Lustig’s wider writing on fructose and fatty liver disease. In a 2016 review, “Fructose and Nonalcoholic Fatty Liver Disease”, and in a 2013 review with Perito and Rodriguez on dietary treatment of nonalcoholic steatohepatitis (fatty liver with inflammation), his group set out the dietary side of the liver argument the trial was testing.

Back to Table of Contents

11. Related Research on Fructose and the Liver

Two other lines of research on the same subject are part of the background to the trial and the discussion that followed it.

Adults given fructose drinks (2009). In a 10-week study led by Stanhope, with Jean-Marc Schwarz among the authors, overweight and obese adults drank either glucose-sweetened or fructose-sweetened beverages supplying 25% of their energy needs. Both groups gained similar weight, but visceral fat increased significantly only in the fructose group. Fat-making in the liver and the 23-hour triglyceride curve after meals rose specifically with fructose, as did fasting apoB, low-density lipoprotein, small dense low-density lipoprotein and oxidised low-density lipoprotein, and insulin sensitivity fell with fructose but not with glucose. Fasting triglycerides rose by about 10% with glucose but not with fructose. The authors conclude that dietary fructose “specifically increases DNL, promotes dyslipidemia, decreases insulin sensitivity, and increases visceral adiposity in overweight/obese adults.”

Pooled feeding trials (2014). A systematic review and meta-analysis led by Chiu, with Sievenpiper among the authors, pooled 13 controlled feeding trials in 260 healthy participants. In the seven isocaloric trials, where fructose replaced other carbohydrates with no change in calories, it found no effect of fructose on liver fat or the liver enzyme alanine aminotransferase (ALT). In the six hypercaloric trials, where very high doses of fructose (104 to 220 grams a day) were added on top of the usual diet, liver fat and alanine aminotransferase rose — an effect the authors write “may be more attributable to excess energy than fructose”. They note that most of the trials were small, short and of poor quality, and that larger, longer and higher-quality trials are required.

The Lustig trial differs from both: its participants were children rather than adults, it removed sugar rather than adding it, and it ran for nine days.

Back to Table of Contents

Key Research Papers

1. Lustig RH, Mulligan K, Noworolski SM, Tai VW, Wen MJ, Erkin-Cakmak A, Gugliucci A, Schwarz JM. Isocaloric fructose restriction and metabolic improvement in children with obesity and metabolic syndrome. Obesity (Silver Spring). 2016;24(2):453-60. — PubMed PMID: 26499447
2. Schwarz JM, Noworolski SM, Erkin-Cakmak A, Korn NJ, Wen MJ, Tai VW, et al. Effects of Dietary Fructose Restriction on Liver Fat, De Novo Lipogenesis, and Insulin Kinetics in Children With Obesity. Gastroenterology. 2017;153(3):743-752. — PubMed PMID: 28579536
3. Gugliucci A, Lustig RH, Caccavello R, Erkin-Cakmak A, Noworolski SM, Tai VW, et al. Short-term isocaloric fructose restriction lowers apoC-III levels and yields less atherogenic lipoprotein profiles in children with obesity and metabolic syndrome. Atherosclerosis. 2016;253:171-177. — PubMed PMID: 27451002
4. Erkin-Cakmak A, Bains Y, Caccavello R, Noworolski SM, Schwarz JM, Mulligan K, Lustig RH, Gugliucci A. Isocaloric Fructose Restriction Reduces Serum d-Lactate Concentration in Children With Obesity and Metabolic Syndrome. J Clin Endocrinol Metab. 2019;104(7):3003-3011. — PubMed PMID: 30869790
5. Olson E, Suh JH, Schwarz JM, Noworolski SM, Jones GM, Barber JR, et al. Effects of Isocaloric Fructose Restriction on Ceramide Levels in Children with Obesity and Cardiometabolic Risk: Relation to Hepatic De Novo Lipogenesis and Insulin Sensitivity. Nutrients. 2022;14(7):1432. — PubMed PMID: 35406045
6. Wojcicki JM, Epel E, Lin J, Tai V, Schwarz JM, Noworolski SM, et al. Leukocyte telomere length change in children with obesity in the context of an isocaloric fructose restriction intervention. Diabetol Metab Syndr. 2025;17(1):94. — PubMed PMID: 40119470
7. Khan TA, Sievenpiper JL. Metabolic improvement with fructose restriction: Is it the fructose or the weight loss? Obesity (Silver Spring). 2016;24(3):549. — PubMed PMID: 26857209
8. Lustig RH. Response to “Metabolic improvement with fructose restriction: Is it the fructose or the weight loss?”. Obesity (Silver Spring). 2016;24(3):550. — PubMed PMID: 26853905
9. Phillips R. Nutrition: Breaking the fructose habit. Nat Rev Endocrinol. 2016;12(1):3. — PubMed PMID: 26585657
10. Vos MB, Goran MI. Sugar, Sugar . . . Not So Sweet for the Liver. Gastroenterology. 2017;153(3):642-645. — PubMed PMID: 28757270
11. Lustig RH. Fructose: metabolic, hedonic, and societal parallels with ethanol. J Am Diet Assoc. 2010;110(9):1307-21. — PubMed PMID: 20800122
12. Lustig RH. Fructose: it’s “alcohol without the buzz”. Adv Nutr. 2013;4(2):226-35. — PubMed PMID: 23493539
13. Lustig RH. Fructose and Nonalcoholic Fatty Liver Disease. J Calif Dent Assoc. 2016;44(10):613-17. — PubMed PMID: 29035477
14. Perito ER, Rodriguez LA, Lustig RH. Dietary treatment of nonalcoholic steatohepatitis. Curr Opin Gastroenterol. 2013;29(2):170-6. — PubMed PMID: 23283181
15. Stanhope KL, Schwarz JM, Keim NL, Griffen SC, Bremer AA, Graham JL, et al. Consuming fructose-sweetened, not glucose-sweetened, beverages increases visceral adiposity and lipids and decreases insulin sensitivity in overweight/obese humans. J Clin Invest. 2009;119(5):1322-34. — PubMed PMID: 19381015
16. Chiu S, Sievenpiper JL, de Souza RJ, Cozma AI, Mirrahimi A, Carleton AJ, et al. Effect of fructose on markers of non-alcoholic fatty liver disease (NAFLD): a systematic review and meta-analysis of controlled feeding trials. Eur J Clin Nutr. 2014;68(4):416-23. — PubMed PMID: 24569542

PubMed Topic Searches

1. Isocaloric fructose restriction in children
2. Papers from trial NCT01200043
3. Fructose, de novo lipogenesis and liver fat
4. Lustig RH and fructose

Further Reading

• Robert Lustig — Fructose Restriction (his website’s page on the study)
• ClinicalTrials.gov — NCT01200043

Back to Table of Contents

Connections

• Dr. Robert Lustig: Sugar, Fructose and Metabolic Health
• Robert Lustig: Life and Career
• Sugar: The Bitter Truth and the Fructose Argument
• Hypothalamic Obesity and Insulin
• Metabolical and the Case Against Processed Food
• The Hacking of the American Mind
• Contemporary Practitioners
• Non-Alcoholic Fatty Liver Disease
• Metabolic Syndrome
• Insulin Resistance
• Added Sugar and Weight
• Blood Sugar and Insulin (animation)