Milk, yogurt and cheese

Across large studies, people who eat dairy live about as long as people who don't, with a slight edge for yogurt and other fermented milks. Dairy is a convenient source of calcium and protein, and it may help frail older people avoid fractures, but nothing in the evidence says adults need it.

Dairy is optional: if you like it, up to about two servings a day is a sensible range, mostly plain yogurt or kefir, milk as you like it, and cheese a few times a week. Keep butter to a little. If you skip dairy, get your calcium elsewhere and choose an unsweetened, calcium-fortified soy drink if you want a milk substitute.

Total dairy is roughly neutral for how long you live

A serving in the studies below is about a glass of milk (200–250 mL), a pot of yogurt (125–175 g) or a slice of hard cheese (about 30 g).

Pooling 29 cohort studies of nearly 940,000 people, neither total dairy nor milk, full-fat or low-fat, was linked to death rates, heart disease or cardiovascular disease. Fermented dairy came out about 2% lower per 20 g a day, but that small signal disappeared when one large Swedish study was left out. Two of the authors had received funding from dairy-industry bodies.^[1]

The single studies point both ways, depending on what dairy replaces. In 136,384 adults across 21 countries, many of them low- and middle-income, eating more than two servings a day went with about 17% lower death rates and 22% less major cardiovascular disease than eating none.^[2] In about 218,000 US nurses and health professionals followed for up to 32 years, by contrast, total dairy made no difference until the top group, about four servings a day, which had 7% higher death rates. Swapping dairy for nuts, legumes or whole grains went with lower mortality, and swapping it for red or processed meat with higher.^[3]

An influential review by two Harvard nutrition researchers concluded that the US advice of three servings a day is not justified, that every nutrient in milk can come from other foods, and that guidelines should treat something like 0 to 2 servings a day as acceptable for adults.^[4] That is the stance this page takes.

Yogurt and kefir have the best record

The most consistent dairy result is for fermented milks, at about 6–7% lower mortality in pooled cohorts, a modest and observational signal that Fermented foods covers in full. Yogurt also tracks with less type 2 diabetes: about 80 g a day went with roughly 14% lower risk than none.^[5] In 2024 the US Food and Drug Administration allowed a yogurt health claim on diabetes while describing the evidence as limited.^[6] Kefir has its own page: Kefir.

Choose plain yogurt and add fruit yourself. Flavoured yogurts often carry a lot of added sugar.

Cheese raises cholesterol less than its fat predicts; butter doesn't get that pass

Dairy fat is about two-thirds saturated, the kind that raises low-density lipoprotein (LDL, "bad") cholesterol.^[7] Cheese seems to behave better than that. In five trials comparing hard cheese with butter carrying the same fat, around 145 g of cheese a day lowered LDL by about 0.22 mmol/L (8.5 mg/dL), 6.5%, compared with butter. That analysis was funded in part by dairy-industry bodies.^[8] An Irish trial of 164 adults aged 50 and over then gave the same 40 g of dairy fat a day as full-fat cheddar or as butter plus milk protein and calcium; LDL ended about 0.46 mmol/L lower with the cheese. The likely reason is the "food matrix", the physical structure the fat is locked inside.^[9]

Butter has no such matrix. In a blinded trial of 47 healthy adults, a moderate amount of butter raised LDL compared with both olive oil and their usual diet.^[10] The case for swapping butter for plant oils is on Dietary fats.

None of this makes cheese a health food: it is salty, and the trials used it in place of butter, not on top of a normal diet. A few servings a week fits a Mediterranean pattern; the MIND diet's stricter cheese limit was chosen from reviews, not measured in its own cohort, as Foods to avoid explains.

Full-fat or low-fat: no clear winner

Advice to choose low-fat dairy rests on its saturated fat. Outcome data don't clearly back it. Across 18 studies, people with the highest blood levels of two fatty acids that come mainly from dairy fat had about 12–14% less cardiovascular disease than those with the lowest; the few studies of death rates found no link.^[11] These markers show how much dairy fat people eat, not what it does, and they cannot tell cheese from butter. On the other side, whole milk was the one dairy food linked to higher mortality in the long US cohorts, about 11% per extra half-serving a day.^[12]

In practice: full-fat yogurt and cheese in modest amounts are fine. If your LDL cholesterol or apolipoprotein B (ApoB) is high, choose lower-fat milk, go easy on cream and keep butter to a minimum. The DASH diet for blood pressure was tested with low-fat dairy.

Bones: milk doesn't prevent fractures, but dairy may help frail older people

Milk is the classic bone food, yet the cohort data don't support it. Pooling six cohorts of about 195,000 women, each daily glass of milk made no difference to hip-fracture risk; in men the data were too few to call.^[13] Drinking more milk as a teenager did not lower hip fractures later in life either, and in men each extra daily glass went with about 9% more, partly because milk makes people taller and tall people fracture more.^[14] One Swedish study of 61,433 women and 45,339 men even found women drinking three or more glasses a day had higher death and fracture rates. The authors warned that confounding and reverse causation could explain it, and the pooled cohorts above did not show the same.^[15]

The one large trial points the other way, in a specific group. In 60 Australian care homes with 7,195 residents averaging 86 years, who were eating little calcium and protein, half the homes served extra milk, yogurt and cheese, adding about 560 mg of calcium and 12 g of protein a day. Over two years, those homes had 33% fewer fractures of any kind, 46% fewer hip fractures and 11% fewer falls; death rates did not change. The trial was funded largely by dairy organisations, which had no role in running it.^[16] It shows that food calcium and protein matter for frail older people, not that healthy adults need more milk.

Dairy is still the easiest way to reach the 1,000–1,200 mg of calcium a day recommended on Calcium: a glass of milk carries roughly 300 mg.^[17] What actually builds bone is covered on Bone density.

Protein for older muscles

Milk proteins are rich in leucine, the amino acid that switches on muscle building, which makes yogurt, skyr, cottage cheese and whey easy ways to reach the 25–30 g of protein per meal that older adults need. How much and when are on Protein and Leucine.

Cancer: likely protective for the bowel, a small signal for the prostate

Across 19 cohort studies, each extra 400 g of dairy a day (about two glasses of milk) went with about 17% lower colorectal cancer risk, mainly for colon cancer; cheese showed no clear link.^[18] The World Cancer Research Fund rates dairy as probably lowering bowel cancer risk.^[19]

For prostate cancer the direction is reversed. Across 32 studies, the same 400 g a day of dairy went with about 7% higher risk.^[20] The World Cancer Research Fund calls this evidence limited.^[21] Both effects are modest and observational.

Lactose intolerance rarely means no dairy

About two-thirds of adults worldwide digest lactose, the sugar in milk, poorly, from about 28% in northern and western Europe to about 70% in the Middle East.^[22] Most of them can still take 12–15 g of lactose, about a cup of milk, without trouble, especially with a meal.^[23] Yogurt has about half the lactose of milk, and hard and long-ripened cheeses have almost none; lactose-free milk is an easy swap.^[24]

Plant "milks" are not nutritional copies

Plant drinks vary widely. Unsweetened soy drink comes closest to cow's milk in protein; almond, oat and rice drinks carry far less.^[25] Check the label for added calcium and vitamin B12, since many brands are not fortified, organic ones especially. Soy foods have their own page: Soy.

Raw milk: real risk, no proven benefit

Pasteurisation kills the bacteria that cause the outbreaks. In the US, unpasteurised milk and cheese, eaten by only a few percent of people, caused 96% of the illnesses from contaminated dairy, about 840 times more illnesses than pasteurised products for the amount eaten.^[26] The claimed benefits haven't held up when tested: in a small randomised trial in adults with lactose intolerance, raw milk caused the same symptoms as pasteurised.^[27] Pregnant women, young children, older adults and anyone with a weak immune system should avoid it.

What to actually do

Treat dairy as optional. If you enjoy it, about 0–2 servings a day is a reasonable range; there is no evidence that more is better.
Favour plain yogurt and kefir, and sweeten them with fruit, not sugar.
Enjoy cheese a few times a week in modest portions, and use olive or other plant oils instead of butter.
Full-fat or low-fat is your call, unless your LDL or ApoB is high; then choose lower-fat milk and keep butter and cream minimal.
If you skip dairy, cover calcium and protein elsewhere: calcium-set tofu, fortified soy drink, sardines with bones, kale and bok choy, and enough protein at each meal.
Skip raw milk.