New long-term data adds weight to a question you face every breakfast.
A vast Norwegian study, stretching across three decades and tens of thousands of lives, now brings rare clarity to the full-fat versus low-fat milk debate. Its sheer scale, and a unique shift in national drinking habits, create a natural test case that few countries could replicate.
What the researchers did
Scientists in Norway drew on three cardiovascular screening waves carried out between 1974 and 1988. They followed 73,860 adults, average age 41 at entry, for 33 years. During follow-up they recorded 26,393 deaths, including 8,590 from cardiovascular disease.
The timing proved crucial. In the 1970s, whole milk dominated Norwegian fridges. By the 1980s, low-fat milk rose sharply. That switch allowed researchers to compare long-term outcomes for people who habitually drank different types of milk across their adult lives.
What they found
When the team looked at total milk intake, people who drank the most milk had a 22 per cent higher risk of death from any cause and a 12 per cent higher risk of cardiovascular death than those who drank the least. The signal did not rest on early deaths or pre-existing illness; excluding those cases barely changed the pattern.
The type of milk mattered: compared directly, low-fat milk was linked to an 11 per cent lower risk of death and a 7 per cent lower risk of cardiovascular disease than whole milk, after accounting for how much people drank.
In other words, the extra risk tied to heavy milk intake appeared to be driven mainly by full-fat milk. The 2025 paper in The American Journal of Clinical Nutrition reports that the association remained after adjustments for lifestyle and other factors.
| Comparison | Outcome | Relative difference |
|---|---|---|
| Highest milk intake vs lowest | All-cause mortality | +22% |
| Highest milk intake vs lowest | Cardiovascular mortality | +12% |
| Low-fat milk vs whole milk (adjusted for quantity) | All-cause mortality | -11% |
| Low-fat milk vs whole milk (adjusted for quantity) | Cardiovascular disease | -7% |
Who drank what
Patterns of who chose which milk varied. Low-fat milk drinkers were more often women, more likely to have higher education and less likely to smoke. Whole milk drinkers reported current smoking more often. The analysis adjusted for these differences, yet the contrast between full-fat and low-fat milk persisted.
Even after accounting for smoking, education and early deaths, low-fat milk kept showing a modest edge for heart health and longevity.
Why this matters for you
These findings fit with current NHS advice on saturated fat. Whole milk contains more saturated fat. Diets higher in saturated fat tend to raise LDL cholesterol, which can nudge up the risk of heart attack and stroke. Low-fat milk keeps core nutrients such as calcium and protein while trimming saturated fat and calories.
None of this says milk is “bad”. It points to a practical lever you can pull without abandoning dairy altogether: switch the type you pour.
What low-fat actually means
On British shelves, “whole” milk contains about 3.6% fat, “semi-skimmed” about 1.7%, and “skimmed” about 0.1%. That translates into meaningful differences in saturated fat per glass.
| Milk type | Typical fat per 100 ml | Typical saturates per 100 ml | Saturates per 200 ml glass |
|---|---|---|---|
| Whole | 3.6 g | ≈2.3 g | ≈4.6 g |
| Semi-skimmed | 1.7 g | ≈1.1 g | ≈2.2 g |
| Skimmed | 0.1 g | ≈0.06 g | ≈0.12 g |
How to act on this without fuss
- Keep your habits, change the milk: swap whole for semi-skimmed in tea, coffee and cereal.
- Choose plain milk over flavoured varieties that add sugar and calories.
- If you cook with cream often, try evaporated skimmed milk or a lighter crème fraîche in sauces.
- Check café drinks: a daily large latte with whole milk adds up fast; ask for semi-skimmed.
- Buy smaller bottles if you only use milk in hot drinks; less waste, easier portion control.
Where this study fits in the bigger picture
This research is observational. It shows associations, not proof of cause and effect. People who pick low-fat milk may also make other healthful choices. The authors controlled for many differences, yet unmeasured habits can still play a role. Even so, the consistency across analyses, the length of follow-up and the huge sample strengthen the signal.
Dairy itself is not one thing. Fermented products such as yoghurt and certain cheeses can behave differently metabolically. The balance of your whole diet matters: fruit, vegetables, whole grains, pulses, nuts, fish and exercise all shape risk more than any single item.
Practical nuances you might face
Family needs vary. Young children aged one to two typically use whole milk for energy and fat-soluble vitamins; many can move to semi-skimmed from age two if growing well, while skimmed is not advised before five. Athletes tracking calories may prefer skimmed for protein with minimal fat. People with lactose intolerance can choose lactose-free versions of low-fat milk. If you pick plant-based alternatives, aim for unsweetened products fortified with calcium and vitamin B12, and check protein levels, which vary widely.
For most adults, switching from whole to semi-skimmed or skimmed will lower saturated fat intake across the week without major sacrifice. Combine that swap with a few other heart-lean choices—more fibre, fewer ultra-processed snacks, regular movement—and your cholesterol and waistline may nudge in the right direction.
Key takeaways you can use today
A small change you control—what’s in your glass—linked to an 11% lower mortality risk compared with whole milk.
You do not need to abandon dairy. You do not need to count every gram. Start with the milk you buy next, keep an eye on flavoured drinks, and let the numbers do the quiet work over time.
Interesting, but it’s still observational. How robust are those 11–22% differences once you account for diet patterns and socioeconimic factors? Did the authors do sensitivity analyses for residual confounding or use negative-control outcomes? Also, any absolute risk deltas would be super helpful.