Three glasses of milk a day tracked with higher death rates, not stronger bones

Milk is widely encouraged, particularly for bone health, on the logic that its calcium should protect against fractures as people age. Whether heavy milk drinkers actually break fewer bones, and whether they live longer or shorter lives, had not been settled in large populations followed over time. Two big Swedish cohorts offered the chance to test both questions at scale, in women and men separately, across decades of follow-up. The answers matter because milk advice reaches almost everyone.

Researchers followed two cohorts from three counties in central Sweden: 61,433 women aged 39 to 74 at baseline between 1987 and 1990, and 45,339 men aged 45 to 79 at baseline in 1997. Participants completed food frequency questionnaires, with the women answering a second one in 1997. Survival models then linked milk consumption to time until death or fracture. Over a mean follow-up of 20.1 years, 15,541 women died and 17,252 had a fracture, 4259 of them hip fractures; over a mean of 11.2 years, 10,112 men died and 5066 had fractures, including 1166 hip fractures.

Women drinking three or more glasses of milk a day had an adjusted mortality hazard ratio of 1.93 compared with those drinking less than one glass, close to double. Each additional daily glass carried an adjusted hazard ratio for death of 1.15 in women and 1.03 in men. Fracture risk did not fall with higher intake in either sex, and in women the hazard ratio per glass for hip fracture was 1.09. In subsamples from two additional cohorts, higher milk intake was associated with a urinary marker of oxidative stress and with interleukin 6, an inflammatory marker in blood.

The result runs against the everyday assumption that more milk means stronger bones, and it complicates the case for recommending high intake as fracture prevention. As cohort research it cannot establish cause: people in declining health may change what they drink, and unmeasured differences between heavy and light milk drinkers could explain part of the pattern. The authors explicitly recommend cautious interpretation. What is needed next is evidence able to separate milk itself from the lives of the people who drink it.

If you drink milk because you enjoy it, nothing here says stop; if you drink several glasses a day purely as insurance for your bones, this study offers no support for that habit and some grounds for doubt. The mortality associations were larger in women than men, but in both cases they are associations rather than verdicts. It is a reason to treat milk as a food rather than a medicine.