Whole milk, not dairy overall, linked to prostate cancer deaths in men

Dairy is a staple across much of the world, and questions about whether it raises or lowers cancer risk have circulated for years, often with conflicting answers. Individual cohort studies are prone to chance findings and differ in the populations and dairy types they cover, which makes any single study hard to interpret. Cancer mortality is the outcome that matters most to people making everyday food choices. This analysis pooled the available population-based cohort evidence to ask whether dairy intake, overall and by type, is associated with the risk of dying from cancer.

The researchers searched PubMed and EMBASE and identified 11 population-based cohort studies covering 778,929 individuals. They extracted the data and calculated pooled relative risks for cancer mortality, with sensitivity analyses and subgroup analyses by region, sex and dairy type, including milk, yogurt, cheese and butter. They also explored dose-response relationships, testing whether risk changed with each additional daily serving. Because these were cohort studies, participants were followed over time rather than assigned to particular diets.

Total dairy intake showed no association with overall cancer mortality, with a pooled relative risk of 0.99. Milk at 0.97, yogurt at 0.88, cheese at 1.23 and butter at 1.13 were all non-significant across men and women combined. The exception was whole milk in men, associated with prostate cancer mortality at a pooled relative risk of 1.50. The dose-response analysis backed this up: each additional daily serving of whole milk carried a relative risk of 1.43 for prostate cancer death, a significant linear trend. A non-linear model also suggested that low total dairy intake was associated with reduced relative risk.

The overall null on dairy and cancer mortality is broadly reassuring and consistent with the mixed picture from individual cohorts. The whole milk and prostate cancer signal is the part that stands out, but it comes from a subgroup analysis of studies that followed people rather than testing diets on them, so milk drinkers may differ from non-drinkers in ways statistics cannot fully remove. Cause and effect cannot be established from this design. Whether lower-fat milk carries the same signal, and what mechanism might connect whole milk to prostate cancer, remain open questions.

For most people this is a reason not to worry about dairy as a category on cancer grounds: across roughly three quarters of a million people, total intake showed no link to cancer deaths. Men may find the whole milk finding worth knowing about, since it was specific, dose-dependent and statistically significant, though it stops well short of proof. The sensible reading is that this is information for weighing up habits, not a verdict on any single food. It also illustrates why dairy headlines so often conflict: the overall picture and specific subgroups can point in different directions.