In women, consumption of three glasses of milk or more a day is linked to higher death rate and risk of fractures, says study published in BMJ.
The findings may warrant questions about recommendations for milk consumption, although further research is needed, its authors said, as the association may be purely coincidental.
A Swedish team used data taken from 61,000 women aged 39-74 and monitored for about 20 years, and more than 45,000 men aged 45-79 followed for 11 years.
The volunteers gave details about diet and lifestyle, body weight, smoking habits, exercise frequency, education level and marital status.
By the end of this long study period, 25,500 of the group had died and 22,000 had suffered a fracture.
Higher milk intake was not accompanied by a lower risk of fractures but "may be associated with a higher rate of death", the study said.
Among the women, 180 per 1,000 in the group which drank three glasses of milk or more a day died during a 10-year period, compared to the group average, independent of milk consumption, of 126 per 1,000.
Among those who drank a glass or less per day, the rate was 110 per 1,000, co-author Karl Michaelsson of Uppsala University told AFP.
The figures for hip fracture was 42 out of 1,000 women who drank a lot of milk, 35 per 1,000 on average, and 31 per 1,000 of women who drank the least milk.
"Women who consumed three glasses or more per day had a 90 percent higher risk of death, 60 percent higher risk of hip fracture and 15 percent higher risk of any fracture compared to those who drank less than a glass," said Michaelsson.
For men, the difference in death rate was less pronounced: 207 per 1,000 among the three-glasses-a-day group over 10 years, 189 per 1,000 on average, and 182 per 1,000 among low consumers. There was no difference in fracture rates.
"The higher risk of mortality was evident with all types of milk: full-fat, half-fat and skimmed milk," Michaelsson added -- and started from a daily intake of about two glasses of milk.
At a lower consumption of half a glass to one glass per day, "there was a tendency of slightly reduced hip fracture risk" compared to zero intake, but the same was not true for mortality risk.
The team found that fermented milk products like cheese or yoghurt were associated with lower mortality and fracture rates, particularly in women.
One reason, the authors speculated, is that milk, but not cheese, is high in D-galactose, a type of sugar that in animal studies was shown to hasten ageing and shorten lifespan.
- Caution -
The researchers said it was impossible to draw any conclusions or make recommendations on milk consumption until further work is carried out.
The results may not apply to people of other ethnic origins with different levels of lactose tolerance, they said.
Milk also has different nutrient levels that depend on factors like food fortification and cow diet.
And the results could be skewed by a phenomenon called "reverse causation" -- osteoporosis sufferers at high risk of a bone break increase their milk intake, which then gets blamed when they suffer a fracture.
In a comment, Mary Schooling of the City University of New York School of Public Health said "the role of milk in mortality needs to be established definitively now" as consumption would rise with economic development.
Other experts noted shortcomings in the study, including that milk consumption was self-reported, often a flaw in dietary research.
Nor did the authors define the type of physical activity the men and women did -- whether it was weight-bearing and therefore bone-strengthening, or not.
The study "creates more questions than provides answers", said Catherine Collins, principal dietitian at St George's Hospital in London.