Food and diet

Milk may be linked to bone fractures and early death

"Drinking more than three glasses of milk a day may not protect bones against breaking – and may even lead to higher rates of death," the Mail Online reports.

Do not be alarmed – your milkman is no Hallowe'en death-bringer. In fact, there are many reasons to treat this news – and the research behind it – with some caution.

The research comprised an analysis of two large Swedish cohort studies, in which a group of men and a group of women were given food questionnaires and then followed up for an average of 20 years. Researchers looked at whether how much milk they drank was linked to fractures or death during follow-up.

In women, drinking any more than 200g of milk daily (less than one glass) was linked to increased risk of dying during follow-up. This increased risk ranged from 21% for one to two glasses to an increased risk of 93% for three or more.

More than one glass a day was also linked to an increased risk of fractures in women. There wasn't such a clear link with either early death or fractures in men.

However, people involved in the study may not have been able to reliably estimate their volume of milk intake per day, and there may have been various unmeasured factors influencing the outcomes.

This was also a Swedish population, who may have distinct health and lifestyle factors or environmental influences, which means the results are not applicable to other populations.

For example, milk in Sweden is fortified with vitamin A (unlike the UK), and high levels of vitamin A intake have been linked to an increased risk of fracture.

These findings are undoubtedly worthy of further research, but people should not feel the need to drink less milk based on this study alone.

Where did the story come from?

The study was carried out by researchers from Uppsala University and the Karolinska Institutet in Sweden.

It was funded by the Swedish Research Council, and one of the researchers was reported to be an employee of the Swedish National Food Agency.

The study was published in the peer-reviewed British Medical Journal. This article is open access, meaning it can be accessed and read for free online.

The majority of the UK's media headlines are needlessly alarmist, although the actual reporting of the study tended to be more restrained. Many of the sources include quotes from independent experts, who discuss the limitations of the study and highlight the fact Swedish milk is fortified with vitamin A.

What kind of research was this?

This was an examination of the findings of two Swedish cohort studies – one in men and one in women – which aimed to investigate whether drinking more milk is linked to outcomes of fracture or mortality (death) from any cause.

The researchers say it is well known that a diet rich in dairy, containing high amounts of essential nutrients such as calcium and vitamin D, is considered to reduce the risk of osteoporotic fractures.

However, they say there may be undesirable effects because milk contains D-galactose, a type of sugar (although it tastes much less sweet than other types of sugar).

Experimental evidence in animals has suggested D-galactose is associated with ageing, with observations including oxidative stress (where damage occurs at the molecular level) to tissues, and changes to gene activity and the immune system.

The researchers say an injected dose of 100mg/kg of D-galactose has been shown to accelerate biological signs of ageing in mice, which is equivalent to 6 to 10g in humans, or the amount found in one to two glasses of milk.

The researchers therefore wanted to test their theory that a high consumption of milk may increase oxidative stress and inflammation in humans, and so increase the risk of mortality and fracture.

Cohort studies are a good way of looking at whether particular exposures are associated with disease outcomes. However, they cannot prove cause and effect.

In this study, important limitations include that the food questionnaire may not give a reliable indication of milk intake or of lifetime patterns.

Also, there may be a variety of other health and lifestyle factors (confounders) that are influencing any association between milk intake and fractures or mortality, which the study has not been able to take into account.

What did the research involve?

This study used data from two community-based Swedish cohorts:

  • The Swedish Mammography Cohort, which recruited more than 90,000 middle-aged to elderly women from two Swedish counties from 1987-90. Food frequency questionnaires were given to the women at enrolment and again in 1997. The current study included 61,433 women who completed both of these questionnaires.
  • The Cohort of Swedish Men, which recruited more than 100,000 middle-aged to elderly men from two counties in Sweden in 1997. The men were given a single food frequency questionnaire at enrolment, and this study is representative of 45,339 men who completed this questionnaire.

In both studies, the food frequency surveys questioned up to 96 foods and drinks consumed over the past year, including how many servings of the item per day or per week.

Dairy items included milk, fermented milk, yoghurt and cheese, with instructions that one serving of milk equalled one 200ml glass.

The researchers say milk intake was specified according to fat content, and they summed intake into a single measure representing total milk intake on a continuous scale.

Looking at outcomes, the researchers examined those recorded between enrolment for the two studies and the end of December 2010. All participants were linked to the Swedish Cause of Death Registry, so the researchers could identify any deaths related to all causes, cardiovascular diseases or cancers.

Fractures were identified by linking all participants to the Swedish National Patient Registry and by looking for any hospital admissions or outpatient visits with diagnostic codes related to fracture.

Regarding adjustment for confounders, the researchers adjusted their analyses for many factors, including age, body mass index (BMI), total energy intake, healthy dietary pattern, calcium and vitamin D supplementation, and physical activity levels.

The researchers assessed the risk of mortality or fracture according to categories of milk intake (less than 200g per day, 200-399g per day, 400-599g per day, and 600g per day or more) and for each additional 200g of milk per day corresponding to each additional glass of milk. They also looked at the effects of other dairy items, such as cheese and fermented milk products.

What were the basic results?

The women's cohort consumed, on average, 240g milk per day, and the men 290g – around one to two glasses day. The researchers observed a general trend that increased milk intake was associated with increased energy intake overall and increased intake of most other nutrients, while alcohol intake tended to decrease.


During an average 22 years of follow-up, 15,541 women died (25% of the cohort), with a third of these deaths as a result of cardiovascular disease and a fifth related to cancer.

The men were followed for an average of 13 years, during which time 10,112 died (22% of the cohort), with just under half of these deaths as a result of cardiovascular disease and just over a quarter caused by cancer.

In the women's cohort, compared with drinking less than one glass of milk a day (less than 200g/day), each increasing category of intake was associated with a 21% increased risk of death from any cause for one to two glasses, and 93% increased risk for three or more glasses.

Any intake above one glass a day was associated with an increased risk of cardiovascular death, but an increased risk of cancer death was only seen with intakes above two glasses a day.

In the men's cohort, the link with all-cause death was less strong. Intakes above two glasses of milk per day were associated with a 5-10% increased risk of death from any cause, but the links were only of borderline statistical significance, meaning these may be chance findings.

Looking at cause of death for men, there was a just significant increased risk of cardiovascular death above two glasses a day, but no significant link with cancer deaths.


In the women's cohort, 17,252 had a fracture (28%) during follow-up, while in the men's cohort, 5,379 had a fracture (12%).

In women, each increasing category of milk intake above less than one glass a day was associated with an increased risk of fracture of 7% for one to two glasses, and 16% for two or more. Risk of hip fracture specifically also increased with each intake above less than one glass.

In men, there was no significant link between milk intake and any fracture, or hip fracture specifically.

Other milk products

No increased risk was found with increased intake of other dairy products, such as cheese or fermented milk – in fact, the opposite was seen.

Higher intakes of other dairy products were associated with a lower risk of mortality and fractures in women. Risk reductions in men were more modest or were non-existent.

How did the researchers interpret the results?

The researchers concluded that, "High milk intake was associated with higher mortality in one cohort of women and in another cohort of men, and with higher fracture incidence in women."

However, they give due caution that, "Given the observational study design with the inherent possibility of residual confounding and reverse causation, a cautious interpretation of the results is recommended. The findings merit independent replication before they can be used for dietary recommendations."


This Swedish study found women who drink more than 200g (less than one glass) of milk a day have an increased risk of death and fractures. In men, the link between milk and risk of death or fractures was less strong.

There was no link between milk intake and fractures, and the risk increases for death above two glasses a day were small and of borderline statistical significance.

The study has various strengths, including its large population size of both men and women, and long-term follow-up. Also, the Swedish registries used to identify causes of death and hospital attendances for fracture are likely to be accurate and reliable.

However, there are important limitations to bear in mind when interpreting meaning from this study, as follows:

  • The study cannot prove direct cause and effect between milk and these outcomes. Although the researchers have tried to take account of various health and lifestyle factors, the study may not have been able to fully account for the influence of these factors (for example, former or current smoking categories were considered, but within these there is going to be a wide range of frequency and duration). There may also be other unmeasured factors influencing the association.
  • There may also be the possibility for reverse causation. For example, postmenopasual women who were at risk of, or had been diagnosed with, osteoporosis could have been at an increased risk of having a fracture and may have been increasing their milk intake to try to boost their calcium levels.
  • It may be hard to reliably estimate milk intake, particularly if you consider that people do not necessarily consume measured glasses of milk per day. Milk is added to drinks or cereal, or is used in cooking. Overall, this could make it difficult to give a reliable indication of milk intake. It is also hard to know whether these food frequency questionnaires represent a lifelong pattern.
  • Also, the study relates to a specific population of Swedish middle-aged to elderly men and women. This population may have particular health, lifestyle and environmental influences, meaning their results are not generalisable to all other populations. For example, Swedish milk is fortified with vitamin A, so results may not apply to the UK, where we do not have fortified milk as standard.

The reverse pattern – decreased risk of death and fractures in women with higher intakes of other dairy products such as cheese and yoghurt – further highlights the uncertain picture painted by these results.

The researchers clearly acknowledge the potential limitations of their research, saying that, "Given the inherent possibility of confounding and reverse causation, a cautious interpretation of the results is recommended."

The findings are undoubtedly worthy of further research, but people should not be overly concerned or feel the need to alter their milk intake as a result of this single study.

A balanced lifestyle is most important for health, including taking regular exercise, not drinking too much alcohol, avoiding smoking and having a healthy, balanced diet – milk contains many important nutrients and can be part of this. 

NHS Attribution