Is dairy healthy, or should you skip it?

Dairy is a package of protein, fat, calcium and milk sugar (lactose). People who avoid dairy over real or suspected lactose intolerance often take in too little calcium and vitamin D, which is the main nutritional cost of skipping it⁸.

Type matters more than the word "dairy." In the pooled cohorts, plain milk and total dairy were neutral for heart disease and death, while fermented dairy (yogurt, cheese, sour milk) showed a small advantage¹. Full-fat milk was linked to slightly more heart disease, and cheese to slightly less¹⁰. Yogurt, at about 80 g a day (a third of a cup), was linked to lower type 2 diabetes risk, with no extra benefit from eating more⁵. Butter showed no clear link either way in the 21-country PURE study². Sweetened dairy desserts are a different food from plain yogurt, and one French cohort tied them to more bowel cancer⁶.

Lactose. Lactose is the sugar in milk, and many adults make less of the enzyme (lactase) that digests it. That is not the same as being intolerant: most people who malabsorb lactose have no symptoms, and many who think they are intolerant do not malabsorb it⁸. A review of 36 randomized trials found that most people with presumed lactose intolerance tolerate 12 to 15 g of lactose, about a cup of milk⁷.

A1 and A2 milk. Ordinary milk contains two kinds of beta-casein protein, A1 and A2. A2-only milk is sold as gentler. A systematic review of 15 randomized trials plus other studies rated certainty as moderate that A1 milk causes more digestive discomfort, and low or very low for every other health outcome¹¹. A2 milk still contains lactose.

IGF-I. Milk and dairy protein raise IGF-I (a growth hormone signal) slightly, and higher IGF-I is linked to prostate cancer risk. The size of the effect is small⁹.

Grass-fed and organic. Organic milk, which comes from cows eating more pasture, has about 56% more omega-3 fats than conventional milk, but less iodine and selenium¹². It is a modest upgrade, not a different food.

Related reading: yogurt, kefir, raw milk and ghee.

Heart disease and death. Guo and colleagues pooled 29 prospective cohorts (938,465 people, 93,158 deaths) and found no association of total dairy or milk with mortality, coronary disease or cardiovascular disease. Fermented dairy was inversely linked, with a risk ratio of 0.98 per 20 g a day for both death and cardiovascular disease, and cheese 0.98 per 10 g a day for cardiovascular disease. Those small signals shrank when one large Swedish study was removed, and heterogeneity was very high. Several authors disclosed funding from dairy industry bodies¹. The PURE cohort followed 136,384 adults aged 35 to 70 in 21 countries for about 9 years. Compared with no dairy, more than 2 servings a day went with lower total mortality (hazard ratio 0.83), major cardiovascular disease (0.78) and stroke (0.66), but not heart attack (0.89, not significant). Milk and yogurt showed lower risk, cheese did not reach significance, and butter showed none². A 2021 meta-analysis found high-fat milk slightly worse (risk ratio 1.08 per 200 g a day for coronary disease) and cheese slightly better (0.96 per 20 g a day), again with heterogeneity¹⁰.

Diabetes. In 22 cohorts (579,832 people, 43,118 cases), total dairy carried a risk ratio of 0.97 per 200 g a day and yogurt 0.86 at 80 g a day, flattening beyond that⁵.

Bones, and where the data disagree. Two Swedish cohorts (61,433 women followed 20 years, 45,339 men followed 11) found that three or more glasses of milk a day in women went with a mortality hazard ratio of 1.93, and each glass with a 9% higher hip fracture risk. Markers of oxidative stress and inflammation were higher in milk drinkers. The authors warned about reverse causation and confounding³. A genetic design avoids much of that problem: among 97,811 Danes, lifelong lactase persistence, which raises milk intake, did not change hip fracture risk (hazard ratio 1.01), and pooling with five other Northern European studies found no fracture benefit, though femoral neck bone density was slightly higher⁴. Taken together, neither cohorts nor genetics show milk preventing fractures, while PURE found the opposite pattern for mortality.

Cancer. A systematic review of 172 studies found that milk raises IGF-I modestly (0.10 standard deviations per standard deviation of milk intake) and that each standard deviation of IGF-I carried an odds ratio of 1.09 for prostate cancer⁹. In NutriNet-Santé (101,279 adults, 2,503 cancers over about six years), total dairy was unrelated to overall, breast, prostate or colorectal cancer. Two subtypes, fromage blanc and sugary dairy desserts, were linked to colorectal cancer, which the authors said warrants further investigation⁶.

Lactose and A2. A review of 36 randomized trials found most adults tolerate 12 to 15 g of lactose, and evidence that near-zero-lactose milk beats that amount was insufficient⁷. An NIH panel concluded true prevalence of lactose intolerance is unknown and most people do not need to eliminate dairy completely⁸. For A1 versus A2, most trials measured intermediate markers and found no difference, yet pooled certainty was moderate for digestive symptoms¹¹.

Organic milk. A meta-analysis of 170 studies found organic milk higher in omega-3 fats (56%), conjugated linoleic acid (41%) and iron, and lower in iodine and selenium, with differences driven by grazing¹².

Conflicts. PURE and the Swedish cohorts disagree, and the fermented-dairy advantage depends on one study. No long trial has tested dairy against heart attacks, fractures or cancer.

Bottom line (confidence 3 of 5). Dairy is optional. Large studies find it neutral to mildly favorable, especially yogurt and cheese, and do not show that it builds stronger bones.