Sugar Is the Word. Carbohydrate Is the Argument.

What you'll learn

Why "sugar" names a family, and which member behaves differently

Why honey is 100% added sugar by definition and a cranberry can look worse than a cookie

What the strongest trial on fructose actually found, and what it did not

Why the fat-versus-sugar argument is not settled, whatever you have been told

What a glucose monitor on a healthy arm can and cannot tell you

How SmartBar applies the deficiency framework

SmartBar assesses products using three things we can verify: ingredient composition, level of processing, and nutritional data. We use two organizing terms throughout our education. Deficiency means something the body needs and isn't getting. Toxicity means a burden the body must process and clear, not a poison.

Sugar is where that second definition matters most. Glucose is not a foreign substance: it is the main fuel your brain runs on and your body makes its own when it needs to. What creates the burden depends on quantity, source and structure, which sugar is involved, and whether it is adding energy or replacing something else. Treating sugar as a poison makes the topic simpler and the advice worse.

Start here: sugar is a family, not a substance

Four words do most of the work.

Glucose is the body's main circulating fuel. Every cell can use it, and your body makes its own if you eat none.

Fructose is the sugar in fruit and honey. Unlike glucose, much of an ingested dose is cleared on its first pass through the intestine and liver rather than circulating widely. Where that happens depends on dose: tracing work in mice shows the small intestine handles most of a low dose while higher doses spill over to the liver, though the human dose-response is not established.[27] That difference turns out to matter.

Sucrose is table sugar: one glucose bonded to one fructose, which split apart during digestion. So table sugar is half fructose by construction. Lactose is milk sugar, glucose bonded to galactose.

Then the wider category. Carbohydrate covers all of it, plus starch, which is simply glucose molecules chained together. Much of white flour is starch, long chains of glucose that digestion breaks apart rapidly. That is why a slice of white bread and a spoonful of sugar are closer metabolic relatives than the words suggest.

The aliases, and a label that does not mean what you think

Added sugars appear on ingredient lists under dozens of names: cane sugar, evaporated cane juice, corn syrup, high-fructose corn syrup, dextrose, maltose, invert sugar, molasses, fruit juice concentrate, brown rice syrup, agave nectar, honey. Our front-of-package article covers spotting them. The label line is more interesting, because it is a regulatory construct with edges.

FDA defines added sugars as sugars added during processing or packaged as such, including sugars from syrups and honey, and sugars from concentrated fruit or vegetable juice in excess of what the same volume of 100% juice would provide. The reference value is 50 g a day for anyone four and older, and 25 g for children aged one to three.[1]

Three consequences most readers never encounter.

Pure honey and pure maple syrup are, by that definition, entirely added sugar. Producers argued that a jar of honey labelled with 17 g of added sugars would suggest it had been cut with corn syrup; after more than 3,000 comments and an act of Congress, these single-ingredient products now show the percent Daily Value without the grams.[2]

Certain fruit juice concentrates are excluded, including those used toward a juice-percentage declaration or to formulate the fruit in jams.[1] A product can be sweetened with concentrate and, in defined circumstances, that sugar never reaches the added-sugars line.

And FDA's guidance records the cranberry problem: comparable products with the same total sugars declare zero added sugars because their fruit is naturally sweet, so the cranberry product looks worse for containing no more sugar.[2]

None of that is a scandal. It is what happens when a category built for a health message meets foods that do not fit it. The practical point: total sugars is the number that does not move.

The line is not natural against synthetic

Here is the division most writing gets wrong.

Table sugar is half glucose, half fructose. Honey is roughly the same split, and high-fructose corn syrup in its common form is close to it too. Maple syrup is mostly sucrose, so also about half fructose once digested. And agave nectar, the most natural-sounding of them, is the most fructose-heavy.

So a shopper choosing honey or agave over table sugar for metabolic reasons is not making the change they think they are making. On the axis that matters biologically, several of those products are the same thing or worse.

That does not make them equivalent in every respect. Honey and maple carry trace compounds table sugar does not, and whole fruit carries its fructose inside a structure that changes its delivery, which our juicing article covers in detail. But "natural" does not create a separate metabolic category.

Somebody tested exactly that. In a randomized crossover trial, 55 adults consumed 50 g of carbohydrate daily from honey, sucrose or high-fructose corn syrup for two weeks each. Glycemic, lipid and inflammatory responses were similar across all three, and all three raised triglycerides.[22] The study was funded by the National Honey Board and hypothesised that honey would come out better. It did not.

Source matters. Structure matters. Dose matters. "Natural" by itself tells you surprisingly little.

What fructose can do that glucose does not

The cleanest evidence is a randomized, double-blind trial in 94 healthy men. Each drank a beverage sweetened with 80 g/day of fructose, sucrose, or glucose in addition to their usual diet, or abstained, for seven weeks. Total calorie intake stayed stable, and liver fat synthesis was measured directly with tracer methodology.

Fructose and sucrose doubled basal hepatic fatty acid synthesis compared with control. Glucose did not.[3]

Sucrose behaved more like fructose because sucrose supplies both. Importantly, the groups did not report higher total energy intake or significant weight gain: the added drinks partly displaced other carbohydrate, fat or protein.[3] So this was not simply overeaters against non-overeaters. Neither was it a tightly controlled isocaloric substitution trial.

And mechanism is not outcome. Fructose substituted calorie-for-calorie for other carbohydrate has not consistently increased liver fat, while very high doses added on top of the diet have.[4] A meta-analysis of 169 controlled trials and 10,357 participants, covering 14 food sources at four levels of energy control, found the pattern depended on energy and source: body weight rose 0.28 kg in addition trials, fell 0.96 kg in subtraction trials, and did not change under energy-matched substitution. Sugar-sweetened beverages were the clearest adverse source in the excess-energy comparisons.[19]

The same research programme has applied that energy-control framework to liver fat, blood pressure and inflammatory markers, with many adverse associations attenuating or disappearing under energy-matched substitution.[19]

So fructose really is metabolised differently, and reviews agree it is a more potent inducer of liver fat synthesis than glucose while describing the link to metabolic syndrome itself as contentious.[5] That does not make energy balance irrelevant, and it does not make every source of fructose equivalent. Different metabolism does not automatically mean different outcome.

The outcome that needs less argument

Sugar's metabolism attracts the arguments. Some of its outcomes require less of one.

WHO recommends keeping free sugars below 10% of total energy, with a conditional recommendation below 5%.[20] "Free" matters: the category covers sugars added to food plus those naturally present in honey, syrups, fruit juice and juice concentrates. It does not cover sugars still inside intact fruit and vegetables.[20]

And one of the clearest harms on this page is one the sugar debate rarely mentions: teeth. The systematic review informing that guideline screened 5,990 papers and found 42 of 50 studies in children and 5 of 5 in adults reporting at least one positive association between sugars and caries, with moderate-quality evidence that caries is lower below 10% of energy.[21] Limiting free sugars to that level reduced but did not eliminate caries, and because caries is cumulative across a lifetime, even low levels in childhood matter — the reasoning behind the conditional 5% figure, where the evidence is weaker and ecological.[21]

Free sugars are a major dietary risk factor for dental caries. No insulin argument is required: the association is consistent enough that it independently supports limiting free sugars.[20][21]

Fat, sugar, insulin: the argument that is not over

You have probably met the claim that fat does not make you fat and sugar does, through insulin. Half of that is settled. The other half is an active scientific fight, and we are not going to pretend otherwise.

One part is settled: body-fat accumulation cannot be explained by saying dietary fat becomes body fat while carbohydrate does not. Our nutrition article covers that shift.

The harder question is whether carbohydrate has a special causal role through insulin, beyond the energy it supplies. That is the carbohydrate-insulin model: high-glycaemic carbohydrate raises insulin, which partitions fuel toward storage, which increases hunger and lowers energy expenditure.

Controlled inpatient feeding studies tested key predictions and failed to support them, leading one reviewer to conclude that important aspects had been experimentally falsified.[6] The same group found that calorie for calorie, restricting dietary fat produced more body fat loss than restricting carbohydrate in people with obesity.[7]

Proponents argue those short inpatient tests do not capture the model's longer-term predictions, and later published a revised formulation of it.[23] Their most prominent longer feeding study went the other way. 164 people held at stable weight after losing 10% of it, on 20%, 40% or 60% carbohydrate for 20 weeks. Total energy expenditure was roughly 200 to 250 kcal/day higher on the low-carbohydrate diet.[8]

Then the two camps fought over it in print. A reanalysis excluding participants for apparent non-adherence roughly halved the effect; the original authors replied that excluding on a post-randomisation variable linked to the outcome introduced severe confounding bias, and that the effect held once confounding was controlled.[9]

Where that leaves us. Insulin resistance is genuinely central to type 2 diabetes and metabolic syndrome. The exact role of carbohydrate in causing it is contested. What is not seriously contested is the practical core: rapidly digestible refined carbohydrate draws the most scrutiny, and it is where the evidence is strongest.

Anyone telling you this argument is finished is selling something, in either direction.

The glucose spike panic

Continuous glucose monitors are increasingly marketed to people without diabetes on the premise that flattening the post-meal curve is a foundation of health.

Start with the physiology. Glucose rising after you eat is not a malfunction. It is the system working, and in people without diabetes it comes back down.

Now the device. These monitors measure glucose in interstitial fluid rather than directly in blood, and they are not laboratory instruments. A randomized crossover trial found that a monitor overestimated glycemia, with the size of the bias varying by test food and by individual, and that a single fasting calibration was not enough to correct it.[10] That matters when a healthy user reads small differences between meals as precise biological signals.

And the readings do not reproduce well. In a controlled inpatient feeding analysis, duplicate meals produced poorly reproducible individual responses on two monitor brands, with variability between repeats of the same meal roughly as large as variability between different meals.[11] If the same breakfast gives a different curve on different days, a personalised food ranking built from those curves is partly measuring noise.

That interpretation is contested. A 2026 editorial, drawing on a broader review of precision-nutrition work, argues that common analytical methods can themselves make post-meal responses appear less reproducible than they are.[12] The disagreement is partly about how the variability should be measured.

Some of it does work, and it is unremarkable. A systematic review of monitor-guided lifestyle intervention in adults without diabetes found that starting a walk before the post-meal peak significantly reduced glucose, insulin and C-peptide, and that the device worked as a motivational tool.[25] A walk after eating is good for you. It does not need a sensor to be good for you.

The more immediate concern may not be metabolic at all. A 2024 review of monitor use in people without diabetes found little consistent, high-quality evidence establishing clinical benchmarks or health utility in this group, and flagged gaps around device acceptability and effects on eating behaviour.[13] Turning normal biological variation into a score can create the problem the device was meant to solve.

If you have diabetes, none of this applies to you: your monitor is a clinical tool and your clinician sets how to use it. If you do not, the honest summary is that a normal body does this, the device is less precise than it appears, and watching the number is not the same as improving your health.

If you have diabetes, prediabetes, or take any medication affecting blood glucose, nothing on this page is a reason to change how you monitor, what you eat, or any medication. Those are decisions with your clinician.

Modern grain, and a myth that is wrong in a useful way

The popular story is that today's wheat has been engineered into something the human body no longer recognises. It is a good story, and the compositional evidence does not support it.

Analysis across 120 years of varieties found gluten content has remained constant, while overall protein declined slightly. Within the gluten fraction, gliadins fell around 18% and glutenins rose around 25%. And the scale matters: harvest year influenced protein composition more than cultivar did.[26]

Other comparisons point the same way. Older bread wheats grown alongside modern varieties have contained more protein, total gluten and gliadin,[14] while modern durum showed no increase in the major gliadin fractions and less omega-5 gliadin, an important allergen.[15]

What did rise is starch. Across a 150-variety diversity screen, protein declined with release date while starch increased, consistent with breeding for higher grain yield.[24]

So the claim that breeding loaded modern wheat with more gluten, making it uniquely harmful, is not supported by the compositional evidence. What rose was starch, with the protein diluted.

And the thing that changed most is not the plant at all. It is what happens after harvest: milling that removes the bran and germ, much of the fibre that leaves with them, and what gets added back in the package. That is the same food-structure argument our nutrition article makes, and it is where the evidence actually is.

One honest complication: when ancient and heritage grains have been tested in animal studies and clinical trials, more anti-inflammatory effects have been reported for the older varieties despite similar gluten content.[16] That signal exists, it is not explained by composition, and we are not going to bury it because it complicates a tidy paragraph.

Sweeteners, and three verdicts in one year

In 2023 three expert bodies assessed the same compound and produced three different-sounding answers.

The International Agency for Research on Cancer classified aspartame as possibly carcinogenic to humans, Group 2B, citing limited evidence.[17] The Joint FAO/WHO Expert Committee on Food Additives reaffirmed the acceptable daily intake of 40 mg per kg of body weight, concluding there was no convincing evidence of adverse effects.[17] And WHO issued a guideline advising against non-sugar sweeteners for weight control.[18]

They are not contradicting each other. IARC assesses hazard: could this cause harm under any conditions. JECFA assesses risk: does it, at realistic intakes. Group 2B also contains aloe vera and caffeic acid, which occurs in coffee, fruit and vegetables.

And WHO's objection is not safety. Its reasoning is that these sweeteners confer no long-term benefit in reducing body fat in adults or children. The evidence pattern is odd and worth knowing: short-term randomized trials show lower body weight, long-term observational studies show higher BMI and more incident obesity, and the trial benefit appears only when sweeteners are compared against free sugars.[18]

Two things follow. The guideline covers all non-sugar sweeteners, synthetic and naturally derived, so stevia and monk fruit are inside it rather than exempt. And it explicitly excludes people with pre-existing diabetes.

The strongest official criticism is not that these sweeteners are poisons. It is that using them as a long-term weight-control strategy has not delivered the benefit their marketing implies. WHO also grades its own recommendation as conditional, partly because the observational associations may be confounded by baseline characteristics and reverse causation.[18]

Is sugar addictive?

Our addiction article sets out why this question is harder than it looks. Gambling disorder is the only non-substance disorder in the DSM-5-TR addictions chapter. Gaming disorder is recognised in ICD-11 but appears in DSM-5 only as a condition needing further research. Sugar appears in neither.

That article's position is that the label is not the test. What matters is impaired control, increasing priority over other activities, and continuation despite harm. Those criteria can be applied to any behaviour without waiting for a committee, and no quantity establishes anything on its own.

Applied here: much of what people describe as sugar addiction involves highly palatable combinations of sugar, fat and salt in engineered products rather than sugar alone, and the animal research most often cited uses intermittent access schedules that do not resemble ordinary eating. The honest answer is that the criteria are the right tool, the label is not established, and anyone confident in either direction is ahead of the evidence.

Key takeaways

Sugar names a family. Glucose and fructose are handled differently, and sucrose is half fructose by construction.

Natural versus refined is the wrong axis. Agave is the most fructose-heavy of the common sweeteners.

Fructose drove liver fat synthesis where glucose did not in a randomized trial, but across 169 controlled feeding trials the effect on body weight depended on energy and source, disappearing under energy-matched substitution.

One of the clearest harms is dental. Free sugars are a major dietary risk factor for caries, and that case needs no insulin argument at all.

The fat-versus-sugar-and-insulin argument is live, with falsification claimed on one side and a positive trial on the other.

A glucose monitor is not a laboratory instrument, and duplicate meals do not reproduce well enough to rank foods reliably.

Modern wheat has the same gluten content it had a century ago. What rose was starch, and what changed most was the milling.

The strongest official criticism of non-sugar sweeteners is about long-term weight control, not poisoning.

How to think about this in thirty seconds

Ask which sugar, not just how much. Fructose and glucose are not interchangeable.
Read total sugars, not only added sugars. One is a measurement, the other is a category with edges.
Treat a natural label as telling you about the source, not the metabolism.
Remember that a rise in blood glucose after eating is the system working.
Ask whether sugar is adding energy to the diet or replacing something else. That distinction changes the answer more reliably than any argument about insulin.

What SmartBar does with this

Added sugars are counted in every product we score, and processing level is assessed independently, which is why two products with identical sugar figures can score differently. That is deliberate. Sugar in a structure behaves differently from sugar without one, and the score is built to notice.

What we will not do is tell you sugar is poison. Our framework defines toxicity as a burden the body must process and clear. With sugar, that burden depends on how much, in what form, from what source, and whether it is adding energy or arriving inside an intact food. That description is less dramatic than a villain and much harder to break with an exception.

The most defensible advice is also the least dramatic: keep free and added sugars limited, prefer intact foods to isolated sweetness, and do not mistake a glucose curve for a health score.

This article is educational and is not medical advice. If you have diabetes or prediabetes, or take any medication that affects blood glucose, nothing here is a reason to change your monitoring, your diet or your medication; those are decisions with your clinician. If tracking food or blood glucose is causing you anxiety or affecting how you eat, that is worth raising with a qualified professional.

Keep exploring: Nutrition · Juicing and blending · Front-of-package claims · Obesity · Addiction

Medical Disclaimer: Read our full Medical Disclaimer

References

US Food and Drug Administration. Food Labeling: Revision of the Nutrition and Supplement Facts Labels, final rule, 27 May 2016; added sugars definition at 21 CFR 101.9(c)(6)(iii). (Added sugars defined as sugars either added during the processing of foods or packaged as such, including free mono- and disaccharides, sugars from syrups and honey, and sugars from concentrated fruit or vegetable juices in excess of what would be expected from the same volume of 100% juice of the same type. Excluded: fruit or vegetable juice concentrated from 100% juices sold to consumers; concentrates used toward the total juice percentage declaration under 21 CFR 101.30 or for Brix standardization under 21 CFR 102.33(g)(2); and fruit juice concentrates used to formulate the fruit component of jellies, jams or preserves under their standards of identity. Daily reference value no more than 10% of total calories: 50 g for ages 4 and older, 25 g for ages 1–3.)
US Food and Drug Administration (2019). "Guidance for Industry: Declaration of Added Sugars on Honey, Maple Syrup, Other Single-Ingredient Sugars and Syrups, and Certain Cranberry Products." (Following the Agriculture Improvement Act of 2018, single-ingredient pure honey, maple syrup and other pure sugars and syrups are not required to declare grams of added sugars but must still show the percent Daily Value, with an optional symbol leading to an explanatory footnote. The guidance records that comparable products containing equivalent total sugars declare zero added sugars because their fruit is inherently sweet, raising concern that certain cranberry products would appear less nutritious than products with the same sugar content. More than 3,000 comments were received on the draft.)
Geidl-Flueck, B., Hochuli, M., Németh, Á., et al. (2021). "Fructose- and sucrose- but not glucose-sweetened beverages promote hepatic de novo lipogenesis: A randomized controlled trial." Journal of Hepatology. PMID: 33684506. ClinicalTrials.gov NCT01733563. (94 healthy men, double-blind, randomized to 80 g/day of fructose, sucrose or glucose in sweetened beverages in addition to usual diet, or to abstention, for 7 weeks. De novo fatty acid and triglyceride synthesis, lipolysis and free fatty acid oxidation assessed by tracer methodology. Fructose- and sucrose-sweetened beverages produced an approximately 2-fold increase in basal hepatic fractional secretion rates of newly synthesised fatty acids compared with control; glucose did not. Effect observed with stable caloric intake. Authors declare no conflicts of interest.)
The Role of Dietary Sugars and De Novo Lipogenesis in Non-Alcoholic Fatty Liver Disease. PMCID: PMC4276992. (Records that two meta-analyses concluded fructose intervention trials examining liver outcomes have largely been conducted in healthy male participants receiving very high doses of 100–200 g/day and have often been confounded by excess energy intake. Cited so the body does not overstate reference 3.)
Fructose drives de novo lipogenesis affecting metabolic health. Journal of Endocrinology (2023) 257(2). DOI: 10.1530/JOE-22-0270. PMCID: PMC10083579. (Review of human intervention studies at physiological doses plus mechanistic animal work. States that despite numerous studies supporting a link between fructose consumption and metabolic syndrome, the link remains a contentious issue, while concluding that fructose is a more potent inducer of hepatic lipogenesis than glucose.)
Hall, K.D. (2017). "A review of the carbohydrate-insulin model of obesity." European Journal of Clinical Nutrition. (States that several logical consequences of the model were investigated in a pair of carefully controlled inpatient feeding studies whose results failed to support key model predictions, and that important aspects of the model have therefore been experimentally falsified, suggesting the model is too simplistic.)
Hall, K.D., Bemis, T., Brychta, R., et al. (2015). "Calorie for Calorie, Dietary Fat Restriction Results in More Body Fat Loss than Carbohydrate Restriction in People with Obesity." Cell Metabolism 22(3):427–436. (Controlled inpatient study; restriction of dietary fat led to greater body fat loss than restriction of carbohydrate, calorie for calorie, in adults with obesity.)
Ebbeling, C.B., Feldman, H.A., Klein, G.L., Wong, J.M.W., et al. (2018). "Effects of a low carbohydrate diet on energy expenditure during weight loss maintenance: randomized trial." BMJ 363:k4583. DOI: 10.1136/bmj.k4583; with secondary analyses. (164 participants with overweight or obesity who lost at least 10% of body weight, randomized to 20%, 40% or 60% carbohydrate test diets with protein controlled at 20%, over a 20-week test phase with energy adjusted to hold weight stable; total energy expenditure measured by doubly labelled water at four time points. Total energy expenditure significantly greater on the low- versus high-carbohydrate diet, reported at approximately 209–250 kcal/day.)
Hall, K.D., Guo, J., Speakman, J.R. (2019). "Do low-carbohydrate diets increase energy expenditure?" International Journal of Obesity 43:2350–2354; with the reply, Ludwig, D.S., Greco, K.F., Ma, C., et al. (2020). "Testing the carbohydrate-insulin model of obesity in a 5-month feeding study: the perils of post-hoc participant exclusions." European Journal of Clinical Nutrition 74:1109–1112. PMCID: PMC7340622. (The reanalysis excluded participants using a calculated "unaccounted energy" variable intended to reflect non-adherence, roughly halving the apparent diet effect. The original authors replied that an analysis based on a post-randomization variable linked to the outcome introduces severe confounding bias, and that with confounding controlled the effect on total energy expenditure remained strong, with sensitivity analyses showing robustness to plausible non-adherence.)
Hutchins, K.M., Betts, J.A., Thompson, D., Hengist, A., Gonzalez, J.T. (2025). "Continuous glucose monitor overestimates glycemia, with the magnitude of bias varying by postprandial test and individual — a randomized crossover trial." American Journal of Clinical Nutrition 121(5):1025–1034. DOI: 10.1016/j.ajcnut.2025.02.024. PMID: 40021059. PMCID: PMC12107490. (CGM overestimated glycemia, with bias magnitude varying by test food or drink and by individual; a single fasting capillary calibration was not sufficient to correct the bias. Limitation stated by the authors: one brand and model tested. Replaces the institutional commentary cited at v1 and v2.)
Hengist, A., Ong, J.A., McNeel, K., Guo, J., Hall, K.D. (2025). "Imprecision nutrition? Intraindividual variability of glucose responses to duplicate presented meals in adults without diabetes." American Journal of Clinical Nutrition 121(1):74–82. DOI: 10.1016/j.ajcnut.2024.10.007. PMID: 39755436. PMCID: PMC11747189. NIH Laboratory of Biological Modeling. (30 adults without diabetes, duplicate meals presented roughly a week apart across four dietary patterns in an inpatient setting. Intraindividual variability of CGM responses was high, with intraclass correlations of 0.28 and 0.17 for the two monitor brands; the authors conclude that personalized dietary advice based on such responses requires more reliable approaches using aggregated repeated measurements. The journal issue carries an accompanying editorial on the finding. Status note: cited as a preprint at v2; the peer-reviewed publication supersedes it and the body no longer describes it as a preprint.)
Thomas, D.M., Fiore, C., Ludwig, D.S. (2026). "Inappropriate analyses make continuous glucose monitoring appear imprecise." American Journal of Clinical Nutrition 124(2):101388. DOI: 10.1016/j.ajcnut.2026.101388. PMID: 42547099. Published online 25 June 2026. (An editorial accompanying a broader narrative review of precision-nutrition studies, not a formal reanalysis of reference 11. Argues that commonly used analytical choices, including incremental area under the curve, can artificially reduce apparent reproducibility of postprandial glucose responses. Disclosed interests: two authors were associate editors of the journal, and one reported royalties from books advocating a low-glycemic-load diet. Cited so the body reports the challenge rather than presenting reference 11 as uncontested.)
Oganesova, Z., Pemberton, J., Brown, A. (2024). "Innovative solution or cause for concern? The use of continuous glucose monitors in people not living with diabetes: A narrative review." Diabetic Medicine 41(9):e15369. DOI: 10.1111/dme.15369. PMID: 38925143. (Found insufficient consistent high-quality evidence supporting CGM utility for detecting abnormal glucose, driving useful behaviour change or improving metabolic health in people without diabetes, and identified gaps around benchmarks, scoring, device acceptability and adverse effects on eating behaviour.)
Pronin, D., Börner, A., Weber, H., Scherf, K.A. (2020). "Wheat (Triticum aestivum L.) Breeding from 1891 to 2010 Contributed to Increasing Yield and Glutenin Contents but Decreasing Protein and Gliadin Contents." Journal of Agricultural and Food Chemistry 68(46):13247–13256. DOI: 10.1021/acs.jafc.0c02815. PMID: 32648759. (60 German winter wheat cultivars first registered between 1891 and 2010, grown over three harvest years. Protein and gliadin contents decreased, glutenin increased, and total gluten content did not change. Harvest year exerted a larger effect on protein composition than cultivar. The authors report no protein-level evidence for increased immunostimulatory potential of modern wheat. Replaces the institutional reporting cited at v1 and v2.)
Brouns, F., et al. (2022). "Do ancient wheats contain less gluten than modern bread wheat, in favour of better health?" Nutrition Bulletin. DOI: 10.1111/nbu.12551. (Comparisons of older heritage and modern bread wheats grown and analysed under identical conditions show older types contain more protein, including total gluten and gliadin, the major source of coeliac-active gluten. Cites Shewry et al. (2020) compositional analysis of 150 wheat varieties from seed banks, including 19th and early 20th century types, grown, harvested and milled under identical conditions, showing protein content declined from older to modern types in a linear fashion, associated with a parallel increase in starch content responsible for higher yields.)
Differences in gluten protein composition between old and modern durum wheat genotypes in relation to 20th century breeding in Italy. ScienceDirect S116103011730045X. (No significant differences between old and modern durum genotypes in α-type and γ-type gliadin expression, the major fractions triggering coeliac disease. A drastic decrease was observed in ω-type gliadin expression in modern genotypes, mainly ω-5 gliadin, a major allergen in wheat-dependent exercise-induced anaphylaxis.)
Differential Physiological Responses Elicited by Ancient and Heritage Wheat Cultivars Compared to Modern Ones. Nutrients (2019) 11(12):2879. PMCID: PMC6950659. (Notes that although ancient, heritage and modern wheats appear similar nutritionally with comparable gluten content, more anti-inflammatory effects have been associated with older varieties whenever tested in animal studies or clinical trials. Cited so the body does not present the compositional findings as the whole picture.)
World Health Organization / International Agency for Research on Cancer / Joint FAO/WHO Expert Committee on Food Additives (14 July 2023). "Aspartame hazard and risk assessment results released." (IARC classified aspartame as possibly carcinogenic to humans, Group 2B, citing limited evidence for carcinogenicity in humans. JECFA reaffirmed the acceptable daily intake of 40 mg/kg body weight, concluding there was no convincing evidence from experimental animal or human data that aspartame has adverse effects after ingestion. IARC assesses hazard; JECFA assesses risk. Both reports highlighted limitations in the evidence base and the need for further research.)
World Health Organization (2023). Use of Non-Sugar Sweeteners: WHO Guideline. (Advises against the use of non-sugar sweeteners to control body weight or reduce the risk of noncommunicable diseases, on the basis that they confer no long-term benefit in reducing body fat in adults or children. Records that higher NSS consumption led to lower body weight and BMI in short-term randomized trials but was associated with increased BMI and incident obesity in long-term prospective observational studies, and that trial effects are observed only when NSS intake is compared with free sugars. The recommendation covers all synthetic and naturally occurring or modified non-sugar sweeteners, and applies to all people except individuals with pre-existing diabetes. WHO states NSS are not essential dietary factors and have no nutritional value.)
Chiavaroli, L., Cheung, A., Ayoub-Charette, S., et al. (2023). "Important food sources of fructose-containing sugars and adiposity: A systematic review and meta-analysis of controlled feeding trials." American Journal of Clinical Nutrition. PMID: 36842451. (169 trials, 255 comparisons, n = 10,357, assessing 14 food sources at four prespecified levels of energy control over a median 12 weeks; GRADE assessed. Total fructose-containing sugars increased body weight by 0.28 kg, 95% CI 0.06 to 0.50, in addition trials; decreased it by 0.96 kg, 95% CI −1.78 to −0.14, in subtraction trials; and showed no effect in energy-matched substitution trials. Positive dose response for sugar-sweetened beverages in addition trials. See also the same group's parallel analyses using the identical design: liver fat, 51 trials, PMCID PMC9325155; blood pressure, 93 reports, PMCID PMC10427023; inflammatory markers, 64 trials, PMCID PMC9572084 — in each, effects were absent or minimal under energy-matched substitution. Note: this is one research group's prespecified programme rather than independent replications, and funding disclosures in this field are common; the article cites the consistency of the design rather than treating any single paper as decisive.)
World Health Organization (2015). Guideline: Sugars Intake for Adults and Children. (Strong recommendation to reduce free sugars to under 10% of total energy; conditional suggestion of under 5%. Free sugars comprise monosaccharides and disaccharides added by manufacturer, cook or consumer plus sugars naturally present in honey, syrups, fruit juices and fruit juice concentrates; the guideline does not apply to sugars in fresh fruit and vegetables or in milk. Recommendations intended as a population benchmark for policy makers.)
Moynihan, P.J., Kelly, S.A.M. (2014). "Effect on caries of restricting sugars intake: systematic review to inform WHO guidelines." Journal of Dental Research 93(1):8–18. DOI: 10.1177/0022034513508954. PMID: 24323509. PMCID: PMC3872848; with Moynihan, P. (2016), "Sugars and Dental Caries: Evidence for Setting a Recommended Threshold for Intake," Advances in Nutrition 7(1):149–156. DOI: 10.3945/an.115.009365. PMID: 26773022. PMCID: PMC4717883. (5,990 papers screened, 55 studies eligible. 42 of 50 studies in children and 5 of 5 in adults reported at least one positive association between sugars and caries; moderate-quality evidence that caries is lower when free-sugars intake is below 10% of energy. At the under-5% threshold a significant relationship was observed but the evidence was judged very low quality and drawn from ecological studies; the review's own summary notes that conclusion may be overstated as it rested on a small subset. Limiting free sugars to 10% of energy reduced but did not eliminate caries, and caries is described as a lifelong progressive cumulative disease, which is the reasoning behind the lower conditional figure. Author declares no conflicts of interest.)
Raatz, S.K., Johnson, L.K., Picklo, M.J. (2015). "Consumption of Honey, Sucrose, and High-Fructose Corn Syrup Produces Similar Metabolic Effects in Glucose-Tolerant and -Intolerant Individuals." Journal of Nutrition. DOI: 10.3945/jn.115.218016. ClinicalTrials.gov NCT01371266. (Single-centre randomized crossover; 28 glucose-tolerant and 27 impaired-glucose-tolerant adults; 50 g carbohydrate daily from honey, sucrose or HFCS-55 for two weeks each with 2–4 week washout. Similar effects on glycemia, lipid metabolism and inflammation across all three; all three increased triglycerides. The authors' stated hypothesis was that honey would improve glycemia and insulin sensitivity relative to the others; it did not. Funded by a grant from the National Honey Board with USDA Agricultural Research Service support; the authors declared no conflicts and reported that the sponsor had no role in design, analysis, writing or the decision to publish. The prespecified hypothesis that honey would produce more favourable effects was not supported. A separate crossover trial in people with type 2 diabetes reported adverse effects of honey on LDL cholesterol and adiponectin, PMCID PMC7271300, so the finding is specific to these markers, doses and populations.)
Ludwig, D.S., Aronne, L.J., Astrup, A., et al. (2021). "The carbohydrate-insulin model: a physiological perspective on the obesity pandemic." American Journal of Clinical Nutrition 114(6):1873–1885. DOI: 10.1093/ajcn/nqab270. PMID: 34515299. (Presents the carbohydrate-insulin model as an alternative physiological paradigm, sets out testable predictions and addresses criticisms; the proponents' developed formulation. Cited only to establish that the model was refined after the earlier inpatient tests, which is what the body states.)
Shewry, P.R., Hassall, K.L., Grausgruber, H., Andersson, A.A.M., Lampi, A-M., Piironen, V., Rakszegi, M., Ward, J.L., Lovegrove, A. (2020). "Do modern types of wheat have lower quality for human health?" Nutrition Bulletin 45(4):362–373. DOI: 10.1111/nbu.12461. PMID: 33380903. PMCID: PMC7756780. (Analysis of the HEALTHGRAIN diversity screen of 150 wheat lines in relation to release date. Protein content declined with release date while starch increased, consistent with breeding for higher grain yield. Cited directly rather than through reference 14, because a quantitative claim rests on it.)
Use of Continuous Glucose Monitoring in Non-diabetic Individuals for Cardiovascular Prevention: A Systematic Review of Its Impact on Guiding Lifestyle Interventions. PMCID: PMC12612783. (Systematic search January 2020 to August 2025 across five databases; seven studies included; narrative synthesis due to heterogeneity. CGM used to personalise exercise timing, such as initiating walking before an individual's postprandial glucose peak, significantly reduced postprandial glucose, insulin and C-peptide; CGM also served as a motivational tool, increasing readiness for physical activity.)
Jang, C., Hui, S., Lu, W., Cowan, A.J., Morscher, R.J., Lee, G., Liu, W., Tesz, G.J., Birnbaum, M.J., Rabinowitz, J.D. (2018). "The Small Intestine Converts Dietary Fructose into Glucose and Organic Acids." Cell Metabolism. DOI: 10.1016/j.cmet.2017.12.016; with Jang, C., Wada, S., Yang, S., et al. (2020), "The small intestine shields the liver from fructose-induced steatosis," Nature Metabolism 2(7):586–593, DOI: 10.1038/s42255-020-0222-9, PMID: 32694791; and the review, Dong, Y., Li, W., Yin, J. (2024), "The intestinal-hepatic axis: a comprehensive review on fructose metabolism and its association with mortality and chronic metabolic diseases," Critical Reviews in Food Science and Nutrition 64(33):12473–12486, DOI: 10.1080/10408398.2023.2253468, PMID: 37671898. (Isotope tracing in mice: low fructose doses are approximately 90% cleared by the small intestine, while doses at or above 1 g/kg overwhelm intestinal absorption and clearance so that fructose reaches the liver and colonic microbiota. Intestinal clearance is increased by feeding and by prior exposure. The 2020 follow-up reports that the same amount of fructose is more lipogenic when drunk than eaten, and when given as a single bolus rather than split doses, indicating that lipogenesis follows when intake rate exceeds intestinal clearance capacity. Limitation stated by the authors: these are mouse data, and the dose-response in humans is described as critically important but not addressed experimentally. Published commentary notes that humans have a relatively small intestine and that fructose doses as low as 0.1 g/kg raise circulating fructose in people, so the murine intestinal shielding may not transfer directly; the body therefore states the dose dependence and says the human dose-response is not established.)

Sourcing note: this article cites federal regulations and agency guidance, WHO and IARC assessments, randomized controlled trials, controlled inpatient feeding studies, published methodological disputes, compositional analyses of crop varieties and systematic reviews, with DOI/PMID where available. Where the scientific literature is actively disputed, the article reports the dispute rather than selecting a side, and where a body of evidence rests on high doses or narrow populations the reference note says so. This article provides no individualized dose, treatment protocol or medical regimen. It provides no glucose-response optimisation protocol or personalised meal-sequencing instructions; where intervention studies are discussed, they are reported as evidence rather than as personalised instructions to the reader. Measurements are given in both metric and imperial. SmartBar's scoring criteria are described without individual point values. Cross-link targets (#) resolve once the site structure is live. Reading time is measured, not asserted: body words excluding the reference list and the metadata header, counted at 225 words per minute and rounded to the nearest minute.

PRODUCTION NOTE — DELETE BEFORE PUBLICATION

Zero bibliographic red flags. Every reference verified against a primary or authoritative source before the claim resting on it was written. Reference 12 is complete, including the disclosed editorial roles and book royalties of its authors.

One correction went beyond the review's instruction, deliberately. The review proposed replacing the fructose sentence with one attributing first-pass clearance to the intestine and liver. The underlying tracing work is in mice, its authors state that the human dose-response was not addressed, and published commentary notes that humans have a relatively small intestine and show circulating fructose at much lower doses. The body therefore states the dose dependence and says plainly that the human dose-response is not established, rather than swapping a liver-centric oversimplification for an intestine-centric one.

Reference 27 also supplies a mechanism the article had asserted without one: the same fructose dose was more lipogenic when drunk than eaten, and when given as a single bolus rather than split. That is liquid sugar and speed, in mice, and it is recorded in the reference rather than expanded in the body.

Two corrections are recorded inside the references rather than silently applied: reference 11's move from preprint to peer-reviewed publication, and reference 12's characterisation, which v3 described as a direct rebuttal when it is an editorial accompanying a broader review.

Numbering note. Reference 26 sits between 13 and 14 in the list, a consequence of resolving a mid-draft collision without renumbering the whole file. Renumbering happens once, manually, at production.

Targeted list for clinical review, so the gate stays a safety check rather than another editorial round: the fructose first-pass wording and its mouse-to-human boundary; the free-sugar and caries thresholds; the characterisation of the carbohydrate-insulin dispute; the CGM claims and the diabetes and prediabetes boundary; the WHO non-sugar-sweetener interpretation; and the final disclaimer. The reviewer is asked to identify errors, unsafe implications, missing clinically material qualifications or misleading wording, not to restyle a locked article.

Outstanding: clinical review; production cross-links; the single manual renumbering pass.