Is milk dangerous? The recent Swedish study

[Jim Pickles]

A recent Swedish study has been in the news a lot, showing that increased milk consumption is associated not only with increased risk of bone fracture, but with increased death rates, cancer rates, heart disease, and general inflammation in the body. As someone who usually drinks a lot of milk (over 1 litre/day) I was concerned, so looked into it further. The original paper is freely available at http://www.bmj.com/c.../bmj.g6015.long.

The common advice for people who have had bone fractures or are at risk of osteoporosis is to increase their milk intake, so a clinic interested in bone fractures in the elderly looked at the data to see if milk indeed helped. They used extant Swedish medical records (possible in Sweden because records are so complete, and also linked to a unique patient identification number), and related them to self-reported food intake.

They found in women that those who had the maximum fresh milk intake (1.2 litres/day) had a 150% increase in all-cause death rate in any period (i.e. it went up to 250% of baseline rate), a 50% increase in rate of hip fractures, and a slightly higher (20% increase) in rate of other fractures. Overall cancer mortality went up by 50%, and CVD (cardiovascular disease) mortality by 200 %. The effects increased gradually with intake; i.e. there was no absolutely safe level.

Bad news!

The effects in men were much smaller, but men should not be complacent, because information was collected less adequately on the men, and they presented evidence that similar effects in men might have been lost in the noise in the data.

In addition higher milk intake was associated with increased levels of inflammatory markers in the bloodstream, showing that milk puts the body under inflammatory stress. Long-term low-level inflammation is known to be associated with the development of CVD, oxidative damage, more rapid aging, loss of muscle strength in ageing (sarcopenia) and possibly Alzheimer’s disease.

On the other hand, increased intake of fermented milk products (cheese, yoghurt) was associated either with none of these effects, or with increased protection. This points the finger at what is missing in fermented milk products, which generally (though not always- see below) contain lower levels of the milk sugar, lactose. The other components of milk stay in fermented products, so they are unlikely to be a factor.

In the gut, lactose is broken down into glucose and galactose (unless people have lactose intolerance, in which case it stays in the gut and causes bloating and diarrhoea). Galactose is known to be toxic – injected into the bloodstream of animals it leads to early death, more rapid ageing, neurodegeneration, and inflammatory responses (similar things happen in human beings with a deficiency in the ability to metabolise galactose). It does this in quantities (in relation to body weight) comparable to what human beings would get from milk. So galactose in anything other than very small quantities (it is an essential ingredient of many of the body’s biochemicals) looks like bad news.

I have therefore decided to cut my milk intake to about 0.5 litres or less/day.

Could there be anything wrong with the study? Studies like these are notoriously difficult, because you are monitoring people over a long time, the population monitored may not stay consistent, and there may be many confounding factors which have not been compensated for properly.

The biggest possible factor is what in medical terms is known as “reverse causation” – the people drinking more milk may have been told to do so because they are already known to be at higher risk of a fracture. But this does not explain (1) the reduced risk from eating fermented milk (which they are also told to do), and (2) the effects on all the other markers – CVD, cancer, inflammation, etc.

Secondly, self-reported diet – especially reported for a short period a long time ago – is notoriously poor as a predictor. They gave evidence that it went some way to being correct but this remains as a factor. But it still does not explain why they got the effects (errors would tend to dilute out any effects, not enhance them – and this is what they think happened with the data on men).

But the study was very good in many ways. It looked at the old and elderly over a long period. It had a very large number of participants, so was able to compensate for the many possible confounding factors (smoking, socio-economic status, other illnesses, etc). But there is always the possibility that the compensation was incorrect, as by its nature you are dealing with the unknown. It was certainly a good study by modern standard.

So why have previous studies not found the same? Probably because they did not separate out liquid milk from fermented milk products, which have opposite effects.

As for sugar in fermented milk products, the cheese in my fridge lists near-zero sugars (against about 5% in milk), but the yoghurt I normally get (labelled “no added sugar”) has the same sugar content as milk. It is quite possible that Swedish yoghurts are fermented out to a much greater extent than Australian ones – suggest caution is needed in choosing.

My good news for the day!

Jim.

[Rik]

As always, cohort studies do not give good insight in causative links since there can be confounding factors not accounted for. That is not to say this study is worthless, but it comes with a lot of ifs and buts and maybes.

[Frederik]

A problem with studies that show that X food is dangerous is that we do have to eat something. And that "dangerous" something has to be replaced with something else. And I would argue that replacing whole foods from the diet is an unhelpful path.

As far as milk goes. It does contain 5% sugar, and it is fairly rapidly digested and absorbed into the bloodstream. I will remain sceptical about the dangers of galactose. However, drinking 1L of milk has the same amount of sugar as a ½L of soft drinks. Also, calories that are drunk tend to be less satiating. So I would guess that the people that drink such a relatively high quantity of milk also has other sources of sugar in their diet.

Another problem is whether all milk is created equal? fat content, organic/non-organic, homogenized etc? Some studies have shown high fat dairy products to be beneficial in many regards, while not showing the same benefit from low-fat dairy. I have not heard about good studies done to test this, but I would argue that say very high quality grass fed whole milk is a food that should not be classified together with say conventional UHT skim milk. Though in the swedish case much of the dairy is likely to be of a decent quality.

I think that cohort studies is a very crude tool especially considering the many subtleties of our digestive system. Also, there is probably also a large degree of variation as to what is actually healthy for an individual. Genetic make up, body fat, activity level etc.

Also recall bias is notoriously a problem with diet studies.

I do think that fermented dairy products are more healthful for most people, most of the time. Also it is much more interesting from a culinary perspective.

just my ideas.

regards

Frederik

[thecolin]

Could vitamin K2 be a factor? It's found in fermented foods and it plays a very important role in putting calcium in the right places in your body.

[Adurst]

My wife and I had a long talk about this a couple of days ago.

She was saying that there is significant evidence that shows that the calcium in milk is in one of the most bio-available forms. But that lots of factors influence your body's ability to lay it down into bone.

The most important being weight bearing stimulatous. And that it is best done as children and youth.

As we age we all loose bone density. It we haven't built a good base earlier in life...it is too late.

Additionally, Generally as we age we become increasingly lactose intolerant.

I haven't read the study yet so will refrain from too much further comment till I do.

But given it was a bone clinic...I have questions of population bias....old frail women...ones drinking the most milk have the most significant osteo issues to start with as that is why they are prioritising milk/calcium consumption...but I have not read it.

We are the only animal I know of that drinks milk as adults.

[Jim Pickles]

Adurst: "But given it was a bone clinic...I have questions of population bias....old frail women...ones drinking the most milk have the most significant osteo issues to start with"

A bone clinic did the analysis, but the people studied were selected from a much wider population - the women were all those who were undergoing mammography in two counties of Sweden, and who were willing to fill in a food questionnaire (they ended up with 74% of all women in the target age group in those two counties).

Population studies have a number of difficulties, as we know. One is self-selection bias (the ones at greater risk of bone fracture may have been more willing to respond to the food questionnaire, even though it was in relation to mammography - difficult to see why though). Also there is the "reverse causation" effect I mentioned - that people most at risk of fractures may have been told to drink more milk. However, as I also said, there was a NEGATIVE correlation between cheese and other fermented milk consumption and fractures. Those at higher risk of fractures are also told to eat more of these. So why should reverse causation work with one type of food and not another?

Population studies especially in the early stages are never the last word. The proper picture only emerges over time as more studies are undertaken and possible confounding factors evaluated. But this study is as good as we have at the moment. As I also said, I find the relation with inflammatory markers in the blood convincing. Chronic inflammatory stress is something to be avoided, and it is for this reason that I am halving my millk consumption (just as I halved my sugar consumption - by reducing or omitting foods that have added sugar - a year ago); this, rather than the danger of osteoporosis in my case.

As I also said, but would like to emphasise, chronic even low-level inflamation is thought to be bad in a number of ways - predisposing to cardiovascular disease, osteoporosis, loss of muscle strength in ageing, senility, neurodegeneration, Alzhemer's disease, cancer.

The paper says "the findings merit independent replication before they can be used for dietary recommendations". Still, taking it at face value, the recommendation to women would be to limit milk consumption to 2 glasses equivalent (400 ml)/day (where effects are still small), and eat as much cheese as you like. Yogurt needs further investigation because as far as I can tell Australian yogurt has as much sugar in as fresh milk - the sugar does not seem fermented out as it may possibly be in Swedish yogurt.

Frederik: "Also, there is probably also a large degree of variation as to what is actually healthy for an individual. Genetic make up, body fat, activity level etc."

Of course people are different, and this is shown by a great spread in their data. We are talking about averages. You may be lucky, and be low-risk even with a high milk intake. On the other hand, you may be unlucky, and your risk may be much higher than the average. We all think we're healthy, until disaster strikes and we realise we're not. After all, some people live a long and healthy lives even though they are heavy smokers, even though on average smoking reduces life-span by 7 years.

(Later corrected the limit for women to 400 ml/day, from 800 ml/day)

Edited November 9, 2014 by Jim Pickles

[Adurst]

Thanks Jim for the clarification re: bone clinic

I'll make a couple comments...not related to the study design at all...but related to milk consumption.

1) I personally can't imagine drinking over a liter of milk a day...my body would not like it. When I was younger, this wasn't an issue...but now...anything more then 250mls in a sitting and 500mls a day will make me feel awful. I think this is a lactose issue, as cheese and yogurt is not issue...and in these cases the lactose is partially digested by the cultures. That said...I eat lots of yogurt and quark and cottage cheese and hard cheeses (personally hate soft Brie and similar cheeses...but that is another issue).

2) if you are consuming over 1ltr of milk a day that is 3000kj+ per day or almost half of your 8000kj a day energy requirement. I'd be worried about the rest of diversity in the diet. Especially if this is day in and day out. I get my clients to do a food diary...most eat the same 10 things over and over...just not enough diversity.

If you are consuming the low fat milk...I'm not sure about that as it removes the component that lowers the GI and hold critical vitamins and effectively increase the sugar content ratio/lactose.

Also you can get too much calcium. Which I've herd of leading to bone spurs (not sure though), and even solidifying in your veins.

As for Australian yogurt...much of the sugar is possibly added back in....we buy a bio-dynamic full fat yogurt...certain times of the year is very yellow due to cows diet.

[Jim Pickles]

In fact I think it must have been closer to the 500 ml/day I'm having now, and its low-fat (1.5%) milk, so thats 950 kJ/day. As for the diversity, yes, I do tend to eat the same things over and over - cant think what else to have, really! I'll count the number of ingredients and let you know if you know about this area (do different vegetables count as different ingredients?). Calcium deposits in arteries are known to be a predictor of cardiovascular disease, but a scan showed no deposits.

Thanks for your comments, Jim.

[Adurst]

yes each veggie is a different item....but you don't get to count red, green and yellow capsicum as different ingredients

families of veggies are a good way to think about it - (homework)

Umbellifers

Labiates

Solanaceae

Composites

Crucifers

Liliaceae

Rosaceae

Cucurbits

Chenopodiaceae

Fabaceae

Poaceae

• 
  

spices can play an important role in micro-nutrients and diversity...many have powerful health benefits

should be pretty easy to get a protein diversity

-chicken

-fishes (which fun fact is correct english when speaking about multiple fish species)

-beef

-lamb

-goat

-rabbit

-kangaroo

-emu

-deer

-beans/legumes/nuts

-eggs

-dairy/milk

I personally aim for 30 distinct items a week...but sometimes a big meal with leftovers like a roast and roast veggies gets us most of the way through the week

[thecolin]

Geeze, if I'd known there was going to be homework, I may have stayed away from this thread.

'Eating the rainbow' is a much easier way to get diversity into your diet without having to keep track of families of veggies. And colour is generally a good indicator of nutrient content. It's not a perfect system, but it balances positive outcomes with ease of execution.

A very important mechanism in calcium deposits and arteriosclerosis is K2 like I mentioned. But like I said, it's more a factor of calcium being where it shouldn't be due to vitamin deficiencies, not specifically a calcium excess. Check out the Rotterdam study if you're interested.

[Adurst]

A passion for my small community garden plot and crop rotation.

Yes lots of people like the "Eat the Rainbow"...it is an easy way to think and remember what to do.

The colours are produced by Phytochemicals (Plant Chemicals hahaha)...not quite the same as vitamin content by highly correlated.

http://www.exrx.net/Nutrition/ColorCodes.html

The only reason I don't like this is you can get varieties of the same vegetables in multiple colours...which actually have very similar chemical make ups: capsicums (red, Green, orange, purple, etc), same with cauliflower (green, purple, white etc)...this applies to many vegetables and the nutritional content is not wildly different between the cultivars.

Further, people often predominantly eat and/or have a sensitivity within a family. NightShades being the most common example. Tomatoes, Capsicum, eggplant, potatoes. Can easily eat the rainbow within the one family...and they contain alkaloids that some people are sensitive to.

Plus from a crop rotation point of view...if you don't plan these things...you'll have major problems.

[thecolin]

A passion for my small community garden plot and crop rotation.

That's great, where did you learn to garden? It's something I want to get into more, but I don't think it's something I'm particularly good at.

[Adurst]

That's great, where did you learn to garden? It's something I want to get into more, but I don't think it's something I'm particularly good at.

Nor was I, but I wanted to be as well.

Got a plot at the community garden. Learned from trying, reading and the experience within the community.

[Kit_L]

I skimmed this, but want to make the point that increasing calcium (and the other minerals that harden the bone matrix) will have little effect on bone density if you are not also adding an activity that drives these (additional) nutrients into the bone. Availability ("bio"- or not) is only half the equation; driving uptake is the other.

From my reading, doing the former without the latter only increases the amount of excretion of these minerals, unless there was a chronic lack before the addition.

[Adurst]

Availability ("bio"- or not)

That made me laugh...and overall too true...you are too good at cutting to the chase/bone...as with all things we consume..there has to be the stimulus driving demand...otherwise it is expensive pee, or more fat and/or more expensive fat at that.

[Frederik]

Good point Kit.

And some of what I was driving at above, for some people in the context of their diet X amount of low-fat milk can be a "good" thing. In other cases it is decidedly a poor idea. Prescriptions that are not tailored and in context can only be appropriate some of the time. That is one of the major limitations with the food pyramid and other "helpful" prescriptions from the health authoratives.

The idea that we are machine like analogs (motors) that require X amounts of calories to run smoothly is outdated. For instance it also negates the fact that we have a gut filled bacteria. Any discussion of diet probably also should feature a discussion of said diets impact on our microbiota.

[Craig]

Also only skimmed, but there are three major things I see with most movement and diet related studies:

- The variety of movement a person is undertaking is rarely taken into account, and never taken into account properly. This clearly has an effect on absolutely everything (a person could be running 2 marathons a day and still be an unhealthy mover for example, with serious joint problems and lots of delayed immune responses happening in the body).

- Whether or not the participants are currently or throughout the process were experiencing delayed immune responses such as joint aching, inflammation, illnesses of any sort, sinus blockages, etc etc. The availability of the immune system resources to process food and drink has an enormous impact on whether or not things are processed properly or completely, if they are in use else where in the body, this could change the outcome significantly.

- The mental health of the person, particularly regarding stress levels. In general, it seems that over all stress levels are much higher in the population these days, and that also has an incredible impact on the bodies capacity to absorb nutrients and protect from problems.

"We are the only animal I know of that drinks milk as adults. "

​Yeah but we are also the only animals who cook our food so that's kind of a null point.

[Craig]

Also: the type of milk is super duper important. I know people who cannot in the slightest tolerate store bought milk (pasteurized and homogenized) but have no problem with raw milk straight from a cow or bought from those health stores that sell it as bath milk. Personally, i get immune responses from most milk (hayfever starts playing up, snotty nose etc), but the pasteurized and non-homogenized milk is fine for me, as is raw milk.

I highly agree with what adam said about variety. I've seen people who were previously intolerant to eggs, who then changed their diet (I think they shifted to a paleo type diet) and then all of a sudden eggs were fine with them. My instincts tell me that this would probably be the case for almost everyone in the study. There are soooooooooooo many factors that contribute to how the body works, I find it silly to try and blame single things for issues, as well as to apply it to the general population, because there isn't really such thing as a general human...we are all insanely different. What affects me strongly might have no affect on you.

[Jim Pickles]

Its possible to say "Its all very complicated, there are lots of individual variations, so many factors, no absolute rules", etc. That is correct, but where does this leave us, except doing what feels good for an individual? There are also general lessons that can be learned from studying the average effects in many people in large populations, which can then be used as a guidance for a baseline to start making the individual variations from. Denying the value of population studies means that we go back to the approach "My father smoked 20 a day and it didnt do him any harm, so I'll smoke too" (which as we know may be OK for some people, but on average will decrease your lifespan by 7 years).

[Adurst]

"We are the only animal I know of that drinks milk as adults. "

​Yeah but we are also the only animals who cook our food so that's kind of a null point.

???

No.

there are lots of things that only humans do...that doesn't make this specific point null.

all mammals start life consuming only milk...we then all ween...a number of biological processes happen...one of which is that we produce less or even stop producing lactase that digests lactose. Some genetic pools of humans are more or less susceptible based on historical diets. Northern Europe/Scandinavia people have the best lactase production.

When you consume milk beyond how much lactase your produce...you will cause issues that you'll feel.

And that's just what we know.

And even your point about cooking isn't null...we are the only animal that cooks milk...you yourself brought up pasteurized milk.

These are anomalies, unique to human milk consumption.

I wasn't agruing one way or another...just putting a relevant fact out.

We could add to that list that we are the only animal that drinks milk from another species...unless humans are feeding it...like in cats or something.

Milk is a mother's food for a baby...we now do all sorts of crazy things to keep animals lactating for our consumption.

[Frederik]

This discussion has probably been derailed abit, but it is impossible to talk about foods and diet in isolation.

We are also the only animal that cooks its food. Rears it is own food. Spray its own food with poison and on on. Many of these things that are special to humans are probably also detrimental to our health or misguided - however, not all of them. Also, this points at some of the fallacies of the paleo diet. Just because a food or substance was not consumed prior does not mean that the human physiology cannot consume and thrive on it. And just because we have relied on a food in paleo times does not necessarily make it more appropriate.

There is a thing called the naturalistic fallacy in Philosophy - applied to this discussion, the fact that just because something is natural does not mean that is optimal or healthy. Botolinum Toxin is a perfectly natural substance it also happens to be incredibly toxic (I cannot readily come up with anything more toxic).

Milk and especially grass fed high fat dairy is one of the most nutrient dense foods available for human consumption. But it is context dependent. (I happen to come from an area were lactose persistence probably is somewhere between 95 and 100%.) But of course not all of these people should be eating copious amounts of dairy just because they happen to tolerate it.

[Craig]

I agree with all your points, specifically that we have a unique relationship with milk. I just find zero need to use "but other mammals don't do it" as an argument against drinking milk. Its much more useful to raise your other points about lactase decline and genetic history.

Jim: I think population studies are important but the results tend to be interpreted in a very black or white manner. Don't drink milk! Its bad! Don't eat grains, they are bad! I find it much more useful to present the results something like "70% of people show intolerance to milk. Try cutting it out for a few months and see if you feel. Better. If not you might be in the 30% in which case milk has all these great nutritional values." This is probably mostly the fault of the media hype train more than anything. I think we just need to reinforce what you said: use the population studies as a platform from which to start your individualisation, knowing full well that each person has potential to be the anomoly that is totally opposite to the statistical likelihoods.

[Jim Pickles]

Craig - I am not going by media hype. That is why I went back to the original paper, and always do if I can in these situations.

Unfortunately, in this case we are not going by what feels right for you in the short term. Somehow this issue may have got mixed up with lactose intolerance which can show itself immediately and which can be diagnosed in the short term. We are talking about long-term effects of high milk drinking, which may take decades to show themselves.

As an example, here is part of a figure from the paper:

showing inflammatory markers in the blood vs milk intake. While the mean levels go up with increasing intake, the range increases as well. Some people at the maximum milk intake have inflammatory markers at the same low level as some of those who have the lowest intake. But others have markers three or more times higher. If you know you are in the former group, fine. But for all you know, you may be in the latter one. How can you be sure? Inflammatory markers are a BAD THING, as they give rise to and show the body is undergoing the processes that lead to accelerated tissue senescence, cardiovascular disease, neurodegeration including Alzheimers disease, and some forms of cancer. It is this rather than the increased dangers of fractures that I bear in mind (though I should say that my fathers last decade of life was made very miserable by advanced osteoporosis of the spine, so I may have genetic tendency for that - or may not, of course).

What do we do, in an uncertain situation? We hedge our bets. In my case, this consists of keeping my milk intake within 500 ml/day (approx 2 glasses/day on the scale above). It is not an complete black and white situation as you say - rather, it is moderation in all things.

So while I agree with "I think we just need to reinforce what you said: use the population studies as a platform from which to start your individualisation, knowing full well that each person has potential to be the anomoly that is totally opposite to the statistical likelihoods.", we also need to recognise that sometimes we do not have enough information to allow us to make an individualisation, in which case our best bet is to go with the population mean.

[Craig]

Great response thanks for writing that up Jim!

I've been all aboard the inflammation train in regards to diet thanks to kit's friend bill Giles. He basically approaches food from the point of view of the immune system. Get to know what your personal delayed immune response are, test for foods that cause them. Cut out all foods that cause serious ones and only occasionally eat the ones that give you mild responses. I think following this method should keep your chance of immune markers really low!

Anyways more info here if you're interested:

http://billgiles-ecology.com.au

[Frederik]

Good points Jim. The main problem, for me, is that I very much doub that it is possible to control for confounders in population based studies.

Also, drinking 7 glasses of milk a day in and out that is a lot of milk.Out side some absurd bulking scheme in weight lifting or body building, I can think of few that actually have that high an intake - and it makes you question the confounders.

Sticking to two glasses of milk is probably not a bad way to go, all things considered.