Free Sugars Intake

WHO 2015. The headline surprise is which outcome the limit is protecting: not body weight, and not cardiometabolic disease. Dental caries.

The recommendations

“WHO recommends a reduced intake of free sugars throughout the lifecourse (strong recommendation).” “In both adults and children, WHO recommends reducing the intake of free sugars to less than 10% of total energy intake (strong recommendation).” “WHO suggests a further reduction of the intake of free sugars to below 5% of total energy intake (conditional recommendation).” [EXTRACTED (WHO - Sugars Intake 2015) Recommendations, p.4 and p.16]

And what each rests on, in WHO’s own words:

“The recommendation to limit free sugars intake to less than 10% of total energy intake is based on moderate quality evidence from observational studies of dental caries.” “…to less than 5% of total energy intake is based on very low quality evidence from ecological studies in which a positive dose-response relationship between free sugars intake and dental caries was observed at free sugars intake of less than 5% of total energy intake.” [EXTRACTED (WHO - Sugars Intake 2015) Remarks, p.16]

The general lifecourse recommendation is strong with no single GRADE rating attached to it — WHO grades the underlying outcomes rather than that recommendation (see the totality of evidence quote below, which names the ratings for body weight and caries).

The exposure is narrower than “sugar”

“Free sugars include monosaccharides and disaccharides added to foods and beverages by the manufacturer, cook or consumer, and sugars naturally present in honey, syrups, fruit juices and fruit juice concentrates.” [EXTRACTED (WHO - Sugars Intake 2015) Remarks, p.16]

Excluded: intrinsic sugars («incorporated within the structure of intact fruit and vegetables») and milk sugars (lactose, galactose) — on the stated ground that “there is no reported evidence of adverse effects” of either. Fruit juice is in; whole fruit is out. The 2015 definition also widened the 2002 one by adding “and beverages” and “and fruit juice concentrates”, while parts of the evidence base date from 1959 and measured total sugars.

Body weight — the effect is energy, not sugar

Contrast	Effect (95% CI)	Design	Certainty
Reduced free sugars, adults, ad libitum	0.80 kg lower (0.39 to 1.21)	5 RCTs	Moderate
Increased free sugars, adults	0.75 kg higher (0.30 to 1.19)	10 RCTs	Moderate
Isoenergetic exchange of free sugars for other carbohydrate	0.04 kg (-0.04 to 0.13) — null	11 trials	no GRADE profile
Reduced sugars, children	SMD 0.09 (-0.14 to 0.32) — null	5 RCTs	Moderate
Highest vs lowest SSB intake, children	OR 1.55 (1.32 to 1.82)	5 cohorts	Low

[EXTRACTED (WHO - Sugars Intake 2015) Summary of evidence, pp.12-13; Annex 1]

The third row is the load-bearing one, and it is the one that gets dropped when the 10% figure is quoted on its own. Swap free sugars for other carbohydrate at equal energy and body weight does not move: 0.04 kg, interval -0.04 to 0.13. So the weight effect of free sugars, on WHO’s own evidence, is an energy effect rather than a sugar-specific one — exactly the substitution logic of Energy Adjustment and What a Diet Coefficient Means: “reduce sugars” and “reduce sugars and replace the calories with other carbohydrate” are different interventions with different answers. WHO draws this conclusion itself, in the Remarks: “The excess body weight associated with free sugars intake results from excess energy intake.” So the reading is the guideline’s own, not an inference from its table. [EXTRACTED (WHO - Sugars Intake 2015) Remarks, p.16]

Note what this does and does not license. It bears on body weight only. Caries is sugar-specific by mechanism, and the isoenergetic swap says nothing about it.

And the primary review bounds the mechanism claim further than the guideline’s rendering does. Te Morenga - Dietary Sugars and Body Weight 2013 calls energy imbalance “a major determinant”, then immediately: “However, other less direct mechanisms independent of energy balance have been proposed” — naming fructose/uric-acid and ectopic-fat pathways that act “independently of an effect on body weight” and declaring them “beyond the scope of this review”. So the isoenergetic null licenses a claim about the scale, not about sugar’s health effects at large. The wiki’s own heading above is written in the narrower form for this reason. [EXTRACTED (Te Morenga - Dietary Sugars and Body Weight 2013) Potential mechanisms]

What the primary review adds to these rows

All five figures verified exactly against Te Morenga - Dietary Sugars and Body Weight 2013 — the transmission is numerically faithful. What the guideline’s compression drops changes how much weight they carry.

• The two adult arms fail differently — neither is simply firmer. The reduction estimate does not survive risk-of-bias exclusion (“the effect estimate was no longer significant… (-0.81 kg, -1.69 to 0.07)”, three of five trials dropped) but strengthens under a different exclusion: dropping the two trials that achieved almost no exposure contrast gives ”(-1.22 kg, -1.81 to -0.63)”, and dropping the completer-only trials leaves it significant at ”(-0.81 kg, -1.41 to -0.21)”. The increase estimate survives bias exclusion at “(0.96 kg, 0.06 to 1.85)” — a lower bound of 0.06 — but carries I2=82% and is the arm WHO downgraded for publication bias. So the arms are exposed on different domains, and ranking one above the other is not supported. Te Morenga’s own verdict is that the estimates hold: “sensitivity analyses showed that the trends were consistent and associations remained after these studies were excluded.” WHO’s profiles encode an asymmetry of their own — reduce arm “Serious risk of bias” with no publication-bias downgrade, increase arm the reverse — so the guideline did not flatten this; the symmetric-looking table above is this page’s summary, not WHO’s.
• Duration moves the effect five-fold. “(2.73 kg, 1.68 to 3.78)” for the two trials beyond eight weeks versus “(0.52 kg, 0.14 to 0.89)” for the shorter ones, P<0.001 — against the authors’ ceiling: «few data from the studies lasted longer than ten weeks.» The headline figures are short-term numbers.
• Publication bias is handled differently by arm, and the reduce arm is the exposed one. WHO downgraded the increase arm for it (“The funnel plot is suggestive of publication bias; therefore, the evidence was downgraded”) but not the reduce arm, where its profile reads “Publication bias was difficult to ascertain, given that only five studies were included in the analysis. The evidence was not downgraded for publication bias.” Te Morenga pooled all 15 ad-libitum trials precisely because «asymmetry cannot readily be assessed with 10 or fewer studies», and in that pool Egger’s was significant (P=0.001) with fixed-effects attenuating 0.78 kg to “(0.42 kg, 0.28 to 0.56)”. Note the quantity: the 0.78 -> 0.42 attenuation is the 15-trial pool, not the 5-trial 0.80 kg row. So the honest statement is that the reduce arm’s certainty was not downgraded for a bias its own review judged unassessable at that sample size.
• The child RCT null is an ADHERENCE null, not an efficacy null — and this changes what it means. Four of the five trials failed to achieve an exposure contrast (8-10 g/day, 0.1 glasses/day, 56 mL/day); the one provision trial achieved 75 g/day vs 12 g/day. “adherence to dietary advice… was poor.” The principle, stated symmetrically: an advice trial estimates the advice-plus- adherence package, so where the exposure contrast fails the null is uninformative about the exposure — and where it succeeds, the estimate is still an estimate of advice. That cuts both ways here: four of the five adult reduction trials were also advice-based, and Te Morenga frames the headline accordingly — “advice to reduce free sugars was associated with an average 0.80 kg reduction”. -> Framing a Decision Question Two post-census provision trials (de Ruyter; Ebbeling) then confirmed the effect where provision replaced advice — the positive counterpart of this same point. The durability caveat attaches to one of them: “after a further year’s follow-up with no further intervention, the difference between the groups was no longer significant.”
• The child cohort signal is a BEVERAGE signal. “Fourteen of these 15 studies reported the sugars exposure as a sugar sweetened beverage”, and outside the SSB pool “there was no evidence of association between increased sugars and adiposity.” Reading OR 1.55 as support for a free-sugars cap generalizes past what the review’s own vote-count will carry.
• The 10% figure is not located by the weight evidence — “compatible with a recommendation to restrict intake to below 10% total energy, currently available data did not allow formal dose-response analysis”, with a null metaregression (0.02 kg, -0.03 to 0.08; P=0.392). This confirms rather than contradicts what this page already held: the threshold is carried by the caries evidence. The authors attribute the failed dose-response to measurement error -> Measurement Error in Dietary Assessment.

[EXTRACTED (Te Morenga - Dietary Sugars and Body Weight 2013) Results; Discussion; Conclusions]

Scope limit worth carrying, which both sources state: weight-loss interventions were excluded by design — Te Morenga, “because the ultimate aim of the review was to facilitate the development of population based recommendations rather than nutritional recommendations for the management of obesity”; WHO, “Trials that were specifically designed to achieve weight loss were excluded.” So none of this speaks to sugar reduction as an obesity treatment, only as population guidance.

Dental caries — a patient-important outcome with no trial behind it

First, the correction to a natural misreading: caries is what the threshold numbers are pinned to, but it is not the sole basis. WHO states the recommendations “were based on the totality of evidence reviewed regarding the relationship between free sugars intake and body weight (low and moderate quality evidence) and dental caries (very low and moderate quality evidence).” [EXTRACTED (WHO - Sugars Intake 2015) Remarks, p.16] Body weight and caries are both graded critical outcomes at moderate certainty in the same table — and the body-weight row has 5 RCTs and a pooled estimate, which the caries row does not.

Caries is a genuine patient-important outcome, not a surrogate. What carries the moderate rating is consistency, and it is striking: “42 out of 50 of those in children and 5 out of 5 in adults reported at least one positive association”; “7 out of 8 studies reported higher dental caries with higher sugars intake”; and the review names it the headline strength — “An important strength in this review is the consistency of the data, despite methodological weaknesses in many studies.” Against that, the evidence base is unusual:

• “No RCTs were identified that met the inclusion criteria” for children; “No RCTs or longitudinal cohort studies were identified” for adults.
• The guideline reports no pooled caries estimate at all. The review does report crude cross-design pools, which the guideline drops: «SMD for DMFT 0.82 (95% CI 0.67, 0.97)», and a risk ratio for caries prevalence. So a magnitude exists; it just never reaches the guideline.
• Certainty reached moderate only by an upgrade for large effect size from the observational floor of low (Upgrading Observational Evidence — one of the three narrow exits, exercised here).
• The <5% recommendation rests on 3 ecological studies (correlations r=0.6-0.8, intake falling from <15 to <10 kg/person/year), rated very low, plus an explicitly non-evidential argument: that caries effects are «cumulative, tracking from childhood to adulthood».

[EXTRACTED (WHO - Sugars Intake 2015) Summary of evidence, pp.13-15; Annex 1 Tables 1-6]

So a strong recommendation (<10%) rests on unpooled observational evidence upgraded by judgment, and a conditional one (<5%) on ecological correlation. Both are honestly labelled by WHO; neither label is visible where the numbers get quoted.

What the primary review adds — and it changes what the thresholds mean

Reading Moynihan - Effect of Sugars on Dental Caries 2014, the WHO-commissioned review these caries figures come from, recovers four bounds the guideline’s rendering does not carry.

• The two thresholds rest on evidence with different fluoride standing. For <10%: “All of the cohort studies used in the GRADE analysis comparing free sugars intake > 10% E and < 10% E considered fluoride exposure.” For <5%: “The national survey data used in the GRADE analysis comparing sugars availability < 5% E and > 5% E were obtained from non-fluoridated populations.” WHO states this too — “All three population studies were conducted in Japan on children with low fluoride exposure” — so it is not an omission, and both bodies then argue the transfer holds because “despite the protection offered by fluoride, the relationship between sugars and dental caries remains” (Moynihan, 11 studies cited).

• But on the GRADE indirectness domain they disagree, and this one IS joined.

  • Moynihan: the <5% studies “were undertaken in populations with low fluoride exposure, and thus show indirectness in extrapolation to populations with good fluoride exposure.”
  • WHO’s <5% profile footnote: “These studies were undertaken in populations with low fluoride exposure; however, there is no expectation of a difference in effect when extrapolating to populations with good fluoride exposure.”

  Same evidence, same question, opposite domain judgments — the commissioned review rates indirectness present, the guideline rates it absent. The not-joined checks do not fire: identical scope, unit and studies, and both positions are their holders’ own GRADE-domain statements rather than one chapter of a multi-position author. The bottom line nonetheless agrees — both rate the <5% evidence very low overall — so this is a disagreement about which domain carries the downgrade, not about how confident to be. Recorded as a distinction with a live disagreement inside it, not filed as a page-level [[tension]], because nothing downstream of it changes. And it is direct evidence against an over-strong reading of the section below: a commissioned review that contradicts its commissioner on a domain rating is not the same witness as it.

• Hitting the target does not eliminate the outcome. “even the groups with free sugars intake < 10% E had some caries (in 1/5 studies, children were free of cavities)”, and “it is not correct to say that it will prevent caries activity in all age groups across the life course.”

• The outcome is measured late, which biases toward understating low-intake harm. “in most studies, dental caries was diagnosed at the cavitation level - a late stage in the disease process… Pre-cavitation damage may occur at levels of sugars intake below that associated with low/no cavities.”

• Publication bias could not be assessed at all, and funding is mostly undisclosed: “In the absence of RCTs with which to conduct funnel plots and limited possibility to combine data, publication bias was difficult to assess. Of the 8 cohort studies included in the GRADE analysis, only 2 declared source of funding.” WHO’s four caries profiles carry no publication-bias footnote at all — the domain is simply not assessed there, which is a larger gap than an under-reported one.

Adults are near-unstudied — stated by both, though the recommendation’s phrasing does not repeat it. “The majority of studies identified were conducted in children, while only 4 studies were of adults” — and no adult cohort study exists. The adult conclusion is carried from child cohorts on a stated biological warrant (same etiology, disease tracking to adulthood), a reasoned indirectness non-downgrade rather than a skipped one. Same move as WHO’s LDL non-downgrade on Surrogate Outcomes — and NOT independent evidence of it, since both are the same body’s process.

[EXTRACTED (Moynihan - Effect of Sugars on Dental Caries 2014) Results; Discussion]

None of this is a second witness. See Why these are one source, not three below.

The second guidance family arrives — and the 5% agreement dissolves

SACN - Carbohydrates and Health 2015 is the independent reading gate 4 said was missing. It agrees with WHO’s <5% figure. That agreement is not corroboration, and the reason is unusually clean: the two bodies excluded each other’s evidence.

	WHO 2015	SACN 2015
<5% free sugars	conditional recommendation	primary, unqualified recommendation
<10% tier	strong	does not exist — one tier only
Evidence for the <5%	three ecological studies, very low quality	ecological studies excluded by design
Instrument	GRADE	own framework — GRADE evaluated and rejected
Warrant	dental caries	dental caries and total energy intake
Strength grades on recommendations	strong / conditional	none — no strength axis at all

The decisive pair of quotes. SACN’s inclusion rule: “Due to the wealth of data available and because of the concerns around their limitations, case-control, cross-sectional and ecological studies were not considered. Only prospective cohort studies and randomised controlled trials were considered for this report.” And on WHO’s caries review specifically: it “included different types of study designs such as non-randomised trials, population and cross-sectional studies which were not included in this report due to concerns of bias.”

So SACN reaches the threshold WHO could only reach at very low certainty, by a route that discards WHO’s evidence for it. Two arguments landing on one number.

• This is NOT a type-E convergence. Independent backing requires the same claim reached by compatible independent routes; here the routes are mutually exclusive by construction. Counting WHO and SACN as two witnesses for <5% would be the laundering the guard exists to catch — and it would be an easy mistake, because the surface facts (different country, different body, different instrument, same number) all point the wrong way.
• What it IS: a genuine tension-grade finding about the number’s warrant. The <5% figure is robust to method in a way neither body alone can show — but only if you accept that two incompatible evidence standards both point there. That is weaker than corroboration and stronger than a single very-low-quality result.
• The partial-independence caveat, checked at acquisition: SACN cites Te Morenga 7 times and Moynihan 5. It read WHO’s commissioned reviews — it distinguishes them on inclusion criteria rather than disputing their findings. So the reading is independent by rule, not by ignorance.

[EXTRACTED (SACN - Carbohydrates and Health 2015) 1.3, 6.67, S.18, A2.13]

And SACN adds a warrant WHO does not have. Its 5% rests jointly on caries and total energy intake, with the energy limb carried by randomised trials (“increasing or decreasing the percentage of total dietary energy as sugars when consuming an ad libitum diet leads to a corresponding increase or decrease in energy intake”) — where WHO’s threshold is caries-warranted. Note the energy argument is sugars-specific: in the same report total carbohydrate is energy-neutral. This is the strongest evidential position the <5% figure has anywhere in the corpus, and it is SACN’s, not WHO’s.

A third guidance family — and it sets no 5% tier at all

NNR - Nordic Nutrition Recommendations 2023 is the third Tier-A body the wiki holds on this question, and it lands on neither of the two positions above.

«Intake of added and free sugars should be below 10 E%, and preferentially lower» [EXTRACTED (NNR - Nordic Nutrition Recommendations 2023) chunk 02, Box 7]

Parameter table (op-weave 2a) — built first, because the construct looked incommensurable:

Parameter	NNR 2023	WHO 2015	SACN 2015	Same quantity?
Regulated construct	«added and free sugars»	«free sugars»	«free sugars»	NNR=WHO yes; SACN NO — see the two definition rows
NNR’s definition	glossary: Free sugars = «Added sugars plus sugars naturally present in honey, syrups, fruit juices and fruit juice concentrates» (chunk 01)	«monosaccharides and disaccharides added to foods and beverages by the manufacturer, cook or consumer, and sugars naturally present in honey, syrups, fruit juices and fruit juice concentrates»	—	YES — NNR’s glossary is WHO’s 2015 definition
SACN’s definition [CORRECTED 2026-07-28]	—	(2015 form, above)	«all monosaccharides and disaccharides added to foods by the manufacturer, cook or consumer, plus sugars naturally present in honey, syrups and unsweetened fruit juices» (S.17)	NO — this is WHO’s 2002 wording; no fruit juice concentrates
Denominator	«E%: Energy percentage, i.e., percentage of total energy intake» (chunk 01)	”% of total energy intake”	%TE	YES
Upper threshold	<10 E%	<10%, strong	no 10% tier — one tier only	YES where both exist
Lower tier	none set — «preferentially lower», unquantified	<5%, conditional	<=5%, primary	comparable: all three address it
Primary warrant	nutrient density / adequacy	dental caries	caries + total energy intake	NO — three different warrants
Strength grade on the recommendation	none	strong / conditional	none	—

[CORRECTED 2026-07-28, SACN revisit] — “free sugars” names TWO constructs, not one

The 2026-07-27 version of this table filled the construct row with a flat YES for all three bodies. That was wrong, and wrong in the direction that made a comparison look cleaner than it is. NNR’s definition was checked against WHO’s and matched; SACN’s was never checked at all and was assumed to follow because it uses the same two words.

SACN adopts WHO’s 2002 wording, and it does so having quoted the 2015 form in the same report:

«The 2002 FAO/WHO Expert Consultation defined the term ‘free sugars’ as all monosaccharides and disaccharides added to foods by the manufacturer, cook or consumer, plus sugars naturally present in honey, syrups and fruit juices (WHO, 2003). In 2015, the term was elaborated to ‘free sugars include monosaccharides and disaccharides added to foods and beverages by the manufacturer, cook or consumer, and sugars naturally present in honey, syrups, fruit juices and fruit juice concentrates (WHO, 2015).» [EXTRACTED (SACN - Carbohydrates and Health 2015) chunk 01, 2.26]

And then recommends the earlier one for UK adoption:

«The definition for ‘free sugars’ be adopted in the UK. This comprises all monosaccharides and disaccharides added to foods by the manufacturer, cook or consumer, plus sugars naturally present in honey, syrups and unsweetened fruit juices.» [EXTRACTED (SACN - Carbohydrates and Health 2015) chunk 08, S.17]

The substantive gap is fruit juice concentrates. SACN’s adopted numerator omits them; WHO’s 2015 and NNR’s both include them. Fruit juice concentrate is the sweetener of choice in products marketed as no added sugar, so this is not a boundary case — it is a category that moves between counted and uncounted depending on whose definition is in force.

Two things this does NOT license, both of which the wording invites:

• Do not read «added to foods» as excluding beverages. SACN analyses sugar-sweetened beverages as a free-sugars exposure throughout the report, and WHO itself calls the 2015 change an «elaborated» form rather than a widening. On the beverage limb this reads as wording, not substance.
• Do not infer that SACN chose the older definition deliberately, or for a reason. SACN quotes both and comments on neither. Whether the choice was considered is simply not in the record, and the source page’s original note says so. An unexplained divergence is a G-gap, not a process defect.

What survives, and it is narrower than yesterday’s claim: NNR and WHO regulate the same construct on the same denominator; SACN regulates a narrower one. So the NNR-vs-WHO threshold comparison below stands as built, and any three-way comparison must carry this asymmetry explicitly.

This is a type-B disambiguation in its textbook form — one term, «free sugars», naming at least two distinct objects across bodies that all cite each other, with the difference nowhere flagged by any of them. -> Is the Food Category Doing Any Work

And it bears on the <5% agreement below. SACN’s <=5% and WHO’s <5% are numerically identical on different numerators — SACN’s excludes juice concentrates. The existing section argues that agreement is not corroboration because the two bodies excluded each other’s evidence; it is also not corroboration because they are not measuring the same quantity. That is a second, independent reason, and it was missed when this page had only two guidance families on it.

The construct row is the one that had to be checked, and it reverses this page’s earlier reading. The NNR source page recorded that «NNR defines neither term here», treating the comparison as a same-quantity failure. That was wrong: NNR’s front-matter glossary defines free sugars in essentially WHO’s words, and since added sugars are a subset of free sugars under that definition, the phrase “added and free sugars” has the extension free sugars. The denominators match too. So the three thresholds are comparable, and the divergence between them is real rather than definitional. [EXTRACTED (NNR - Nordic Nutrition Recommendations 2023) chunk 01 — the glossary entries; the subset step is this page's] [searched: "monosaccharide"/"disaccharide"/"naturally present in honey"/"definition of sugar"/"intrinsic sugar" across all 10 NNR chunks]

The three-body picture, now that they can be compared:

	<10% tier	<5% tier	Warrant
WHO	strong	conditional	caries
SACN	absent	primary	caries + total energy
NNR	the only tier	absent	nutrient density

Read the direction and the threshold separately — they behave differently. All three bodies point the same way (less is better: WHO’s conditional tier, SACN’s primary one, NNR’s «preferentially lower»), so this is not a disagreement about direction. What diverges is where a number gets attached, and on what warrant. Per the telos, that is itself the finding: where guidance families place their thresholds differently on the same evidence, the evidence does not determine the number, and the wiki’s job is to show why rather than pick a side.

NNR’s warrant is the genuinely new one, and it is not a dose-response argument at all. Box 7 leads with adequacy, not disease:

«Restricting the intake of added and free sugars is important to ensure adequate intakes of micronutrients and dietary fibre (nutrient density) as well as to support a healthy dietary pattern. This is especially important for children and persons with a low energy intake.» [EXTRACTED (NNR - Nordic Nutrition Recommendations 2023) chunk 02, Box 7]

Caries appears second («Frequent consumption of foods with added and free sugars should be avoided to reduce the risk of dental caries») and environment third. So NNR’s 10% is a displacement threshold — sugar crowds out micronutrients and fibre — where WHO’s is a caries threshold and SACN’s is caries plus energy. NNR stratifies its own warrant — «especially important for children and persons with a low energy intake» — which follows from the displacement mechanism: a fixed proportion of energy from sugar crowds out proportionally more of a smaller nutrient budget. A caries threshold carries no such stratification, because caries risk tracks frequency and exposure rather than the surrounding energy budget. [EXTRACTED (NNR - Nordic Nutrition Recommendations 2023) chunk 02 — the stratification is NNR's; the mechanism reading is this page's]

And NNR names an evidence gap exactly where the other two set their second tier. Its carbohydrate Main data gaps block reads:

«There is a lack of studies on carbohydrates and health effects in pregnancy. There is also a lack of a standardized definition for dietary sugars (free and added sugars) and a lack of long-term studies measuring the impact of reducing intake of free and added sugars (especially below 10 E%)» [EXTRACTED (NNR - Nordic Nutrition Recommendations 2023) chunk 04, Main data gaps]

Two things follow, and only the first is NNR’s.

• NNR states the below-10 E% evidence is thin — the region in which WHO’s conditional <5% and SACN’s primary <=5% both sit. That is NNR’s own words about the evidence base.
• NNR does not say this is why it declined a lower tier. The adjacency is suggestive and the inference is tempting, but NNR never joins them, so the causal reading is the wiki’s, not the report’s, and is marked as such. [INFERRED (NNR - Nordic Nutrition Recommendations 2023) — the gap and the missing tier are both NNR's; the link between them is this page's]
• Note the second-order irony, which is real and worth keeping: the same data-gaps sentence flags «a lack of a standardized definition for dietary sugars (free and added sugars)» as a field-level gap, while NNR’s own glossary supplies a definition that matches WHO’s. Those are compatible — a body can use a definition the field has not standardized — but it means the comparability established in the table above is a property of these three documents, not of the literature they draw on. A fourth body could use the same words for a different construct. -> Is the Food Category Doing Any Work

No type-E is available here, and the reason is documented rather than assumed. NNR read the other two: it cites «WHO, 2015» and «SACN, 2015» by name in its carbohydrate chapter, summarising WHO’s two tiers accurately («it was recommended to limit free sugars intake to less than 10 E%. In addition, a conditional recommendation was set to limit the intake of free sugars to less than 5 E%»). A body that read both incumbents and chose a third position is a genuine third judgment on the same evidence — but it is not an independent witness to that evidence. The divergence is informative about the appraisal; it adds no new primary data. [EXTRACTED (NNR - Nordic Nutrition Recommendations 2023) chunk 04]

What this does NOT establish. NNR’s silence above 10 E% is not evidence that the <5% tier is wrong — NNR neither tested nor rejected it. Three bodies disagreeing is a statement about the state of the appraisal, not a verdict on the number.

Why WHO and its own reviews are one source, not three

The wiki now holds WHO Sugars 2015 plus both reviews behind it. They are one witness for the appraisal and the recommendation — and NOT one witness for the primary evidence. That split is the whole of the finding, and an earlier draft of this section overstated it into a flat “one source, not three”.

Parameter table (op-weave 2a):

Parameter	Te Morenga 2013	Moynihan 2014	WHO Sugars 2015	Independent?
Who set the review question	WHO NUGAG subgroup	WHO NUGAG subgroup	itself	no — common origin
Who produced the GRADE ratings	authors, “then refined by the WHO… (NUGAG) Subgroup”	authors, «further refined with the guidance by the NUGAG Subgroup»	publishes them bylined to the review authors, with NUGAG input	no — co-produced
Cross-review GRADE lineage	assisted Moynihan’s tables	thanks “Prof. Jim Mann and Dr. Lisa Te Morenga… for their assistance with the GRADE summary tables”	—	no — the reviews are linked to each other, on this limb only
Funding	”WHO provided some funding"	"funded by Newcastle University’s Centre for Oral Health Research” — WHO commissioned but did not fund	n/a	mixed
Guideline text vs review text	not checked	near-verbatim reuse (indirectness passage)	reproduces it	no — textual reuse
Underlying studies	38 cohorts + 30 trials (weight)	55 studies (caries)	the whole graded base is these two reviews; the remarks cite further analyses	largely subset
Primary data collections	independent trials/cohorts	independent studies, disjoint from Te Morenga’s	collected none	YES — commissioning does not touch the trials

The last row is the one that bounds the conclusion, and the first draft of this table omitted it.

• No type-E is bankable across these three on the appraisal — question-setting, GRADE production and text are shared, so agreement on a certainty rating is one assessment reported three times.
• But the reviews are not one witness with each other on the evidence. They address disjoint outcomes (weight vs caries) over disjoint literatures with different confounding structures. The honest reason E is unavailable between them is disjointness — there is barely a shared claim to converge on — not lineage. Their one documented mutual dependency is narrow and runs through the GRADE tables (row 3), not the data.
• And the dependence is partial, which is why divergence is informative. The two disagree on the <5% indirectness domain (above): a commissioned review that contradicts its commissioner is demonstrably not the same witness as it. A flat “one witness” claim would have made that disagreement impossible, and it happened.
• What the pairing DID bank is F — the reviews bound and refine what the guideline compresses (the reduce arm’s bias exposure; the five-fold duration subgroup; the beverage-specific child signal; the energy-mediation carve-out; the crude caries pools). A refinement needs no independence.
• The usable rule, stated at the width the evidence supports: a guideline and its commissioned reviews are one witness for the appraisal, the certainty ratings and the recommendation — never for the primary data, and not necessarily for each other. The trap is real: every surface marker of independence is present (different authors, journals, years, outcomes) and the substantive ones on the appraisal limb are absent. -> Certainty of Evidence vs Strength of Recommendation

[INFERRED (Te Morenga - Dietary Sugars and Body Weight 2013; Moynihan - Effect of Sugars on Dental Caries 2014; WHO - Sugars Intake 2015) — the cells are quoted on the two source pages; the induction across them is this page’s.]

What this does NOT establish. It is not a criticism of WHO’s process — commissioning reviews and using their ratings is how guideline development is supposed to work, and WHO discloses it throughout. The finding is about how the wiki must count, not about whether the guideline is sound.

The companion guideline — do NSS help you hit this limit? [2026-07-29, WHO NSS 2023]

WHO’s free-sugars limit prompts an obvious follow-on: use non-sugar sweeteners to get there? WHO’s 2023 NSS guideline answers no — «WHO suggests that non-sugar sweeteners not be used as a means of achieving weight control or reducing the risk of noncommunicable diseases (conditional recommendation)» (-> Non-Sugar Sweeteners). The two guidelines chain into the substitution decision; they do not disagree.

Parameter	WHO Sugars 2015 (this page)	WHO NSS 2023	Same quantity?
Target exposure	free sugars intake	non-sugar sweeteners (a route to cutting free sugars)	No — different exposures
Recommendation	reduce free sugars, <10%E (strong)	do not use NSS for weight/NCD control (conditional)	No — complementary, not competing
Preferred sugar substitute	(not addressed)	water, unsweetened foods, whole fruit — NOT NSS	—

The join is a distinction, not a tension (not-joined check (ii): different scope). WHO Sugars sets the goal; WHO NSS rules on one means to the goal, and prefers water/fruit over the NSS route because a sugar→NSS swap leaves «the overall quality of the diet … largely unaffected». No shared quantity is compared, so nothing here is a [[tension]].

NON-INDEPENDENT — same body. WHO NSS 2023 is the same organization as WHO Sugars 2015 (and shares the NUGAG Subgroup and reviewer pool), so its agreement is not type-E independent corroboration — the same counting rule as Why WHO and its own reviews are one source, not three above. It extends this page’s substitution frame; it does not add an independent backing.

Decision relevance

• The 10% limit is a dental recommendation. If someone’s decision is about adiposity or cardiometabolic risk, the caries evidence does not transfer, and the weight evidence says the effect runs through energy.
• Name the replacement. Cutting sugar and replacing the calories with other carbohydrate moved weight by 0.04 kg. The realistic alternative determines the effect (telos layer 3).
• Fruit juice sits inside the exposure and whole fruit outside it — a food-level distinction that a “reduce sugars” instruction loses entirely.
• Ecological studies are the weakest design for an individual inference (the ecological fallacy): the <5% threshold rests on three ecological studies rated very low by both the review and the guideline — the wiki’s clearest live example of a widely-cited number whose own sources claim little for it.

Limits

• 2015. Free-sugars/cardiometabolic outcomes beyond weight and caries are not covered here.
• No absolute effects: no risk differences, no baseline risk, no NNT anywhere. As with sodium, this cannot be ranked against other exposures without them.
• Two of eight caries cohorts measured total sugars, retained by inference from a correlation with added sugars — exposure misclassification inside the body of evidence carrying the main recommendation.
• The three-body comparison is an appraisal comparison, not an evidence comparison. NNR read WHO and SACN; SACN read WHO’s commissioned reviews. No two of the three are independent witnesses to the primary data, so agreement anywhere among them must not be counted as corroboration.

Self-critique of the NNR addition [run 2026-07-28, before commit]

• Over-claim: FOUND and fixed. The first draft read “three positions, three warrants — the guidance set does not converge.” Too strong: all three bodies agree on direction (less is better) and diverge only on where a number attaches and on the warrant. Rewritten to separate the two. The inflated version would have manufactured exactly the divergence the telos warns a divergence-scored fabric will produce.
• Mislabel: FOUND and fixed. NNR’s low-energy-intake stratification was tagged “route-(a)- adjacent”. It is not baseline-risk stratification; it is a property of the displacement mechanism. The route label was removed rather than swapped, because the correct route is not established here.
• Unsearched absence claim: FOUND and removed. The draft asserted the energy-intake stratification is one “the other two bodies do not make”. That absence was never searched in the WHO or SACN documents — only in what this page holds of them. Replaced with a mechanism statement about caries thresholds that does not depend on an unsearched absence.
• Same-quantity check: RUN, and it reversed the prior reading. The NNR source page had recorded the sugars comparison as a same-quantity failure on the ground that NNR defines neither term. The glossary search found NNR’s free-sugars definition and its E% definition, both matching WHO. The correction runs toward making a comparison possible, which is the direction that deserves more scrutiny, not less — so the definition row quotes both sides rather than asserting the match.
• Laundered independence: checked and denied. NNR cites WHO 2015 and SACN 2015 in the same chapter it sets its own threshold in. No [E-independent] claimed; the non-independence is stated in the body and in Limits.
• Residual risk: the strongest unverified step is the reading of «added and free sugars» as a union whose extension is free sugars. It follows from NNR’s own glossary, but NNR never states the identity, and a reader who takes the phrase as naming a deliberately broader construct would get a different comparison. Tagged in place.