https://www.sciencedirect.com/science/article/pii/S0022316626000222
Hmmm funded by the farmers. I am not myself a fan of full fat milk, but I don’t believe in demonising fat. However, perhaps it deserves a consideration.
chatGPT(5.5paid)
The paper is Anderson et al. (2026), “The Effect of Three Daily Servings of Full-Fat Dairy for 12 Weeks on Body Weight, Body Composition, Energy Metabolism, Blood Lipids, and Dietary Intake of Adults with Overweight and Obesity.” It is a randomized, parallel, multisite dietary intervention examining whether three daily servings of full-fat dairy can be incorporated into a diet broadly consistent with the 2019 Canada’s Food Guide without adverse metabolic consequences.
Summary
The study randomized adults with overweight or obesity into three 12-week groups:
- LD-ER: low dairy, with a prescribed 500 kcal/day energy restriction.
- 3D-EN: three servings/day of full-fat dairy, with other foods reduced by about 500 kcal to try to keep overall energy intake neutral.
- 3D-AL: three servings/day of full-fat dairy added ad libitum, with no instruction to reduce calories.
All groups simultaneously received dietary counselling based on Canada’s Food Guide. The stated hypothesis was that three servings of full-fat dairy would not worsen cardiometabolic biomarkers, while potentially improving intakes of nutrients that might otherwise be limiting.
The intervention dairy was quite specific: about 250 mL whole milk, 100 g Greek yogurt and 42 g cheese daily, rather than simply allowing participants to select whatever full-fat dairy they wished.
Main findings
The most important overall finding is a null adverse-effect result. After 12 weeks, adding three servings of full-fat dairy did not significantly worsen:
- body fat,
- waist circumference,
- fasting glucose,
- HbA1c,
- total cholesterol,
- LDL,
- HDL,
- resting metabolic rate,
- or kidney-related urinary measures.
The abstract reports that triglycerides temporarily increased at week 4 but had returned to baseline by week 12.
Weight differences were small. The low-dairy energy-restricted group lost about 0.69 kg, the energy-neutral dairy group gained about 0.35 kg, and the ad-libitum dairy group gained only about 0.14 kg. Thus, remarkably, simply adding three daily dairy servings without prescribing calorie reduction did not produce meaningful weight gain over 12 weeks.
The ad-libitum dairy group also showed within-group reductions in:
- hip circumference: about 0.25 cm;
- systolic BP: about 2.7 mmHg.
But these are much less convincing than the paper’s wording can make them sound, because the between-treatment differences in these changes were not significant. For example, the treatment-group comparison for systolic BP was P = 0.09.
Nutritionally, dairy consumption increased protein and calcium intake and in some analyses increased vitamin D, magnesium and potassium. Calcium intake in the dairy groups reached roughly 1,110–1,120 mg/day, versus about 760 mg/day in the low-dairy group.
Compliance was reasonably good: mean dairy adherence was about 80%, while adherence to the wider dietary recommendations reached about 77% by week 12. Participants’ reported understanding of Canada’s Food Guide also improved substantially.
The authors therefore conclude that three daily servings of full-fat dairy can be accommodated within the Canada Food Guide dietary pattern without evident short-term adverse effects on weight or cardiometabolic biomarkers.
What is novel?
The novelty is more modest than the paper’s topic might initially suggest. This is not the first randomized study showing that full-fat dairy does not necessarily worsen lipids, weight or metabolic risk.
The authors explicitly identify the novelty as the combination of a controlled full-fat-dairy intervention with simultaneous counselling to follow the 2019 Canada Food Guide.
I would separate the novelty into three elements.
1. Full-fat dairy tested within a whole-diet framework
Most dairy trials effectively ask:
What happens if dairy intake changes?
This study asks something closer to:
Can substantial full-fat dairy consumption be incorporated into an otherwise guideline-concordant diet?
That is arguably more relevant to real dietary recommendations. The authors themselves describe this as the principal strength and novelty.
2. The ad-libitum arm is particularly interesting
The 3D-AL group simply added three servings of dairy without explicit caloric restriction.
Despite adding a substantial amount of food energy, their average weight increased by only about 0.14 kg over 12 weeks.
That suggests some spontaneous compensation in the rest of the diet. The authors interpret this as evidence that appetite regulation adjusted to the dairy addition and discuss the possible importance of the dairy-food matrix.
This, to me, is probably the most scientifically interesting result in the paper.
It argues against the simplistic assumption:
additional dairy calories → equivalent positive energy balance → predictable weight gain.
It does not establish the mechanism, but it demonstrates substantial dietary compensation under free-living conditions.
3. Comparison of two different ways of incorporating dairy
The distinction between:
- adding dairy while attempting to remove equivalent calories, and
- simply adding dairy ad libitum,
is useful.
Interestingly, the investigators admit that their supposedly energy-neutral group did not actually achieve energy neutrality.
Paradoxically, that failure makes the ad-libitum result somewhat more interesting: participants apparently adjusted their intake naturally even without explicit calorie instructions.
Critique
There are several substantial weaknesses, some acknowledged by the authors and some insufficiently emphasized.
1. It was severely underpowered relative to the original design
This is the biggest problem.
The original planned sample was 153 participants, based on detecting a 2-kg difference in body weight with 80% power. COVID disruption meant that only 107 were recruited and 74 completed the study.
That leaves only about 24–26 participants per group for many analyses.
Consequently, the study provides reasonable evidence against large short-term effects, but much weaker evidence against small-to-moderate effects.
This distinction matters enormously when interpreting statements such as:
full-fat dairy “did not increase” LDL or HbA1c.
Statistically, failure to demonstrate a difference is not equivalent to demonstrating equivalence or non-inferiority.
A properly designed non-inferiority trial with prespecified margins would be needed to support a strong claim of “no adverse effect.”
The authors argue that the achieved sample was sufficient to detect changes in fasting glucose, lipids and HbA1c, but simultaneously acknowledge that the shorter study may have missed longer-term changes.
2. The intervention became 12 weeks rather than the intended 24 weeks
The study was originally intended to last 24 weeks, but funding/COVID-related constraints reduced it to 12 weeks.
This makes the repeated description of the intervention as assessing “long-term” dairy consumption questionable.
Twelve weeks is sufficient for:
- lipid changes,
- much of the HbA1c response,
- short-term weight effects,
but it says comparatively little about:
- sustained weight trajectory,
- cardiovascular events,
- diabetes development,
- atherosclerosis,
- or long-term dietary adaptation.
So the safe conclusion is:
three servings/day were metabolically well tolerated for 12 weeks, rather than that high full-fat-dairy intake is harmless over the long term.
3. The study’s control comparison is conceptually awkward
The control group was simultaneously told to:
- eat little dairy,
- choose low-fat dairy or plant alternatives,
- and reduce calories by 500 kcal/day.
The experimental groups differed not simply in dairy fat but also in calorie instructions.
That means this isn’t a clean test of:
full-fat dairy versus low-fat dairy.
Nor is it a clean test of:
dairy versus no dairy.
It tests dietary packages.
This is important because weight loss, blood pressure changes and food substitution patterns could arise from the caloric counselling rather than dairy itself.
A stronger design would have included, for example:
- three servings full-fat dairy;
- three servings low-fat dairy;
- isoenergetic non-dairy foods;
with otherwise identical counselling.
4. The intended caloric interventions did not work particularly well
The LD-ER subjects were told to cut 500 kcal/day, yet reported an average reduction of only about 213 kcal/day. And the 3D-EN group did not actually achieve energy neutrality.
This weakens the interpretability of comparisons between the three arms.
On the other hand, as noted above, it makes the behavioural compensation in the ad-libitum group an interesting observation.
5. Several “positive” findings are statistically fragile
The paper highlights reductions in hip circumference and systolic BP in the 3D-AL group.
But the crucial distinction is:
within-group significance ≠ evidence that the intervention caused an effect.
For example:
- SBP declined significantly within 3D-AL,
- but the difference in SBP change between groups was not significant (P = 0.09).
The same problem applies to some nutrient and anthropometric findings.
Because the study examines many outcomes, many time points and multiple subgroup/factor comparisons, some nominal P < 0.05 results would be expected by chance.
There does not appear to be a comprehensive multiplicity correction across the entire collection of endpoints. Tukey-Kramer correction was used for particular pairwise comparisons, but that is not the same as correcting the global false-positive rate across the many outcomes tested. Their statistical strategy involved multiple ANOVAs, paired tests and one-sample tests across many dependent variables.
I would therefore treat the BP and hip-circumference findings as exploratory rather than persuasive evidence of benefit.
6. The paper sometimes shifts from “no evidence of harm” toward “evidence of benefit”
This is probably the largest interpretive weakness.
The strongest result is:
no obvious deterioration in conventional metabolic markers.
That is useful.
The evidence that full-fat dairy improves those markers is substantially weaker.
Indeed, the authors’ final summary says that improvements in SBP, hip circumference and BMI were found in dairy-consuming groups.
That wording is arguably too favourable, given that several of those findings are within-group changes rather than statistically convincing treatment effects.
I would rewrite the conclusion as:
Three daily servings of the tested full-fat dairy foods did not produce detectable adverse changes in weight, body composition, glycaemia or conventional lipid measures over 12 weeks; they improved intake of several nutrients. Evidence for independent cardiometabolic benefit was inconclusive.
That is more faithful to the data.
7. Generalisability is restricted
The participants were relatively young — mean age about 36.6 — and although overweight or obese, they were metabolically fairly healthy at baseline.
The study excluded participants with substantial hypertension, diabetes, chronic disease and various other conditions. Baseline glucose, HbA1c, lipids and BP were generally in clinically normal ranges.
So this paper cannot readily establish safety in:
- older adults,
- people with diabetes,
- established cardiovascular disease,
- familial hypercholesterolaemia,
- advanced metabolic syndrome,
- or other high-risk populations.
8. “Full-fat dairy” is broader than the actual intervention
The intervention comprised particular quantities of:
- whole milk,
- Greek yogurt,
- cheese.
It does not necessarily generalize to:
- butter,
- cream,
- ice cream,
- very high-fat cheeses,
- sweetened dairy desserts,
even though these might colloquially all be described as full-fat dairy.
The appropriate unit of inference is therefore closer to a mixed milk/yogurt/cheese dietary pattern, rather than “dairy fat” in general.
9. No mechanistic measurements
The study offers explanations involving appetite regulation and the dairy-food matrix but does not really test them.
The authors themselves acknowledge one important omission: circulating fatty-acid profiles were not measured, which would have helped assess exposure and potentially distinguish effects of dairy fatty acids such as CLA, vaccenic acid and omega-3 fatty acids.
Other useful measurements could have included:
- ApoB,
- LDL particle number/size,
- postprandial lipemia,
- insulin and HOMA-IR,
- GLP-1/PYY/ghrelin,
- continuous glucose monitoring,
- fecal fat/calcium excretion,
- objective biomarkers of dairy-fat intake.
The absence of ApoB is particularly relevant because unchanged LDL-C does not necessarily establish an unchanged atherogenic particle burden.
10. Industry funding deserves consideration, although it does not invalidate the study
The paper reports no author conflicts, but funding came partly from Dairy Farmers of Canada, through Dairy Research Cluster 3, together with Agriculture and Agri-Food Canada and Mitacs. The authors state that the funders did not restrict publication.
That does not make the results unreliable, particularly because this was randomized and many of its main outcomes are objective.
Nevertheless, where conclusions are somewhat more positive than the underlying statistics warrant, funding context is reasonably considered when evaluating framing.
Overall assessment
I would regard this as a useful but underpowered pragmatic dietary trial, with a more important result than its modest size might suggest.
Its strongest contribution is not evidence that full-fat dairy causes weight loss or improves cardiovascular health.
Rather, it shows something narrower and quite interesting:
Adults with overweight or obesity could add roughly three daily servings of whole milk/yogurt/cheese for 12 weeks, within a broadly healthy dietary pattern, without appreciable weight gain or deterioration in conventional fasting lipid or glycaemic markers.
The 3D-AL arm is especially interesting because the participants apparently compensated for much of the additional energy without being instructed to do so.
I’d rank the conclusions approximately as follows:
Quite well supported:
three servings/day did not produce obvious short-term deterioration in weight, body composition, fasting lipids or glycaemia; dairy materially increased calcium/protein intake.
Suggestive:
humans compensate for much of the additional dairy energy through spontaneous changes in intake.
Weak evidence:
full-fat dairy lowers blood pressure or improves body composition.
Not established by this study:
full-fat dairy is cardioprotective, superior to low-fat dairy, or safe/beneficial over years.
The most valuable follow-up would be a 6–12 month, adequately powered trial comparing isoenergetic full-fat dairy, low-fat dairy and non-dairy controls, with ApoB and insulin-sensitivity measurements. That would turn the interesting negative finding here into a much stronger test of whether the fat content of dairy itself matters.