Milk, protein and the rise of plants

I’m not entirely sure why, but the issues surrounding dairy milk and its substitutes seem to have featured prominently in my facebook feed lately. This has been partly in relation to nutrition, and partly about environmental impact. I regularly point people to a couple of resources on this, but the aim of this post is to cover both the nutrition and environmental impact in the same place.

Milk doesn’t agree with me…

This is a polite way of saying that milk gives you the shits. In this era of food fads, super-foods and insensitivities, we should be absolutely clear that lactose intolerance is a genuine physiological condition, not just something that afflicts middle class Guardian readers. As babies we have an enzyme called lactase that allows us to break down the lactose in milk. However, in the vast majority of world populations, the genes encoding this enzyme are not expressed in adults, which makes people lactose intolerant. The prevalence of this varies significantly as shown in the map below:

A gene mutation several thousand years ago allows some people to continue to digest lactose as adults. The map shows populations where this mutation did not occur. Taken from here . Whilst it’s straying off topic a bit, lactase persistence is a good example of ‘gene-culture coevolution’; the phenomenon of genes influencing culture and vice versa, as discussed in relation to lactase persistence here in a study of the teeth of ancient cultures.

Lactose is a disaccharide; a sugar consisting of a glucose molecule bound to a galactose molecule. Processing milk (e.g. into yoghurt and cheese) significantly reduces lactose content, so it is entirely reasonable for your annoying lactose intolerant friend who you think is completely faking it to chow down on a cheese pasty and even people who are lactose intolerant can generally tolerate a couple of hundred mililitres without significant ill-effects.

So it’s not just vegans who are interested in dairy milk alternatives, and this wider pool of consumers have driven a massive increase in variety and choice.

Nutritional profiles of dairy milk alternatives

When we’re substituting one type of milk for another, we need to be clear what it is that we are actually trying to substitute. In practice, most people approach milk substitution within the context of a relatively fixed eating pattern; you might be looking for something to have on cereal or in a cup of tea for example. There is nothing wrong with this, as long as the nutritional differences are understood.

And this is where we potentially end up with a lot of protein-deficiency. To most people the word ‘milk’ implicitly conveys some nutritional meaning that they assume will also apply to non-dairy milks. This is partly due to the fact that breast milk provides a complete food for babies and infants, and in the UK it is perhaps also a cultural and historical response stemming from nutritional programmes that led to milk being widely provided in schools. From this, many of us have a vague recollection that milk is high in protein and calcium, so we tend to assume that the same is true for milk substitutes. Unfortunately, it isn’t. I have compiled the table below to show these differences between cows milk and a range of non-dairy milks.

All values are per 100 ml. A normal ‘serving’ is 240ml. The RDI is taken from UK guidelines (NHS website) for adults. Where the RDI is different for men and women the average value is taken (the differences are slight). A more detailed breakdown of RDI’s for various age groups is available here.
* signifies fortification to achieve this value.

There’s lots of stuff in milk! In order to keep things simple, I’ve left out the data for 30 or so other principle components of milk, for which milk is not a major source in most diets and/or where deficiencies are uncommon (e.g. people don’t drink milk because of it’s vitamin C or Magnesium content). Contact me if you would like the complete data table.
Soya = Alpro soya unsweetened. Oat = Alpro oat unsweetened. Coconut = Alpro coconut unsweetened. Almond = Alpro almond unroasted unsweetened. Rice = Alpro rice original. Hazelnut = Alpro hazelnut original. Cashew =  Alpro cashew original. Pea = Querkee unsweetened.

The first point to note is that all of the non-dairy milks have been fortified with vitamins and minerals. This is interesting in itself; the manufacturers are acknowledging that these products do not substitute for dairy milk in nutrient terms. The next logical question is what are they fortified with and why? There seem to be a couple of underlying philosophies. Firstly, they are fortified with nutrients where a standard 240ml serving of cows milk provides a significant proportion of daily intakes; calcium, vitamin B12, and vitamin B2 (although rice milk is not). All are fortified to approximately the same level as that present in cows milk. Secondly, all of them are fortified with Vitamin D; this is not present in cows milk in the UK (although milk is fortified with vitamin D in some other countries). Vitamin D is only present in animal products (the NHS lists oily fish, red meat, liver, and egg yolks as being good sources), so fortifying vegan alternatives is a good way of increasing vitamin D intake in populations who would certainly be lacking in it otherwise. It seems entirely appropriate to fortify these non-dairy milks; broadly speaking it makes them nutritionally equivalent to cows milk in terms of the vitamins and minerals that one would expect to get from something that called itself ‘milk’. We do however have to buy into the idea that milk substitutes are processed products and that this is ‘good’ fortification rather than ‘bad’ additives.

The key thing in this nutritional comparison is the protein content. Dairy milk and other dairy products are a significant source of protein in many people’s diets. Soya and pea milks are the only non-dairy milks with a nutritionally useful amount of protein. Pea milk is not yet common in the UK, but it’s likely to become more so in 2019, as there are a couple of new brands available taking advantage of people’s concerns about soy. You don’t necessarily need to have protein in your milk substitute if you have enough elsewhere in your diet, but it is a good example of why we have to be careful when we substitute; you can’t just drink an equal volume of non-dairy milk and assume that it’s healthy.

Pea milk. To me it had the same taste that I associate with soya milk and don’t like very much. But it has a useful amount of protein. Claims to be produced in the UK, but I’m assuming that doesn’t mean the peas are grown here. Other brands are also available…

Environmental impacts

And what of the environmental impacts? Aficionados of dairy milk alternatives will be aware of the debate around soy bean production and the environmental damage it causes. International food trade is the major driver for rainforest destruction, and this is particularly acute in South America. In a study of deforestation drivers and international trade relating to Brazil, around a third of deforestation was directly linked to soy bean production, and the remaining two thirds to cattle farming directly. What’s more, around 67% of soya produced globally is fed to poultry and livestock, so the importance of livestock production as a driver for deforestation is difficult to over-state. It is far more effective to limit your meat consumption than to worry about soya milk.

The gold standard for measuring environmental impact is generally considered to be life cycle assessment (LCA). When we do LCA we always express the results in terms of a ‘functional unit’ which relates to the service provided. Choosing a functional unit for food is not always straightforward; what service is being provided? I’ve discussed this before in another blog post. The environmental impact of our milk substitute is going to be vastly different depending on our functional unit; are we talking about a drop in replacement (e.g. to put in tea or on cereal), or a replacement in nutritional terms? If our functional unit is based on protein requirements, our decision is pretty straightforward.

GHG emissions from various feedstuffs, expressed as kg CO2 equivalents per kg protein. Each dot represents a data point from an individual LCA study. The bands represent the range of results obtained, which vary according to the production system under study and the assumptions made. Taken from here. [this figure is reproduced from my blog post on low carbon diets here]

As we can see, beef and lamb are very poor environmentally, as is seafood. The differences between all of the other foodstuffs are relatively small. Consequently, if you eat beef/lamb/seafood even in small quantities, and you are interested in your environmental impact, you would do much better to reduce your consumption of these products and shift to any of the other foodstuffs with lower CO2 per kg of protein. Your milk consumption is pretty irrelevant.

But what if the functional unit is related to a ‘volume of liquid that tastes like milk and I can put on my cereal’? I’ve not found a comprehensive LCA dealing with all the non-dairy milk substitutes using similar assumptions, but as described here and here, they are relatively similar when considered on a volume basis (or at least the differences are small compared to the differences between non-dairy and dairy).

When looking at non-dairy milks with protein as a functional unit, the choice is between soya and pea milk (because the others are not useful protein sources. Soya and yellow peas (the source of pea milk) are both nitrogen fixing legumes, with similar fertiliser requirements. An LCA of pea and soya milk showed the two to be fairly similar in terms of environmental impacts, both on a volume basis and a protein basis (although it’s worth noting that the study was commissioned by a pea milk manufacturer). This is not particularly surprising given their agricultural similarities. However the study was based on US production systems. When looking at global agricultural production, it seems unlikely that yellow peas are associated with rainforest destruction in the same way that soya is, so in that regard one might regard pea milk to be better than soya on a protein basis.

But I’m already vegan, what milk is best?

If you’re vegan, my advice would be not to sweat the detail, you already have a very low impact diet. On one level this may seem like a bit of a cop out. But in my mind it is justified. Whilst LCA is the best way of determining environmental impacts of products, it is sensitive to the underlying assumptions made. This is why I like the figure above which shows the range of values from various studies for each foodstuff, reflecting the different assumptions made by each study team. Without detailed information on the precise source and method of manufacture of each of your protein choices, it’s just too difficult to know if the difference between oat milk and rice milk is bigger than the difference between two different brands of the same type of milk (for example). If you’re vegan, you are doing the right thing, just make sure you’re eating enough protein and then then invest some energy into looking at environmental impact in another aspect of your life.   

A word on cheese

The ‘what are we trying to substitute’ thinking applies just as much to cheese as it does to milk, and it’s a common vegan pitfall.  Having read the small print on vegan cheese packets I have yet to find one that has any real nutritional value. This is fine if you are trying to substitute the function of ‘something to put in between two slices of bread’ or ‘something that will melt on a pizza’. But they in no way substitute for cheese if the function you’re substituting has anything to do with nutrition.

Final thoughts

From an environmental perspective, if you eat beef or lamb, then the environmental impact of your milk choice is fairly irrelevant. From a nutritional standpoint, if you’re vegan or dairy-free, you should think about your protein consumption carefully. Soya milk and pea milk are the only high protein milk options, so if you prefer one of the other alternatives, make sure you get your protein elsewhere.

Postscript: Whenever I write about food I feel obliged to write a couple of sentences about my own diet. I don’t drink much milk (enough for a couple of cups of tea a day) but it is always cow’s milk. That’s mostly convenience as it’s what the rest of the household drinks. I eat meat once or twice a week. I have previously calculated that I probably don’t eat enough protein. My top protein tips are to eat wholemeal pasta and couscous instead of the refined versions; they are surprisingly high in protein. I also eat a hard-boiled egg most days and a couple of smoked mackerel fillets a week. Plus the usual advice of eating nuts and pulses.

3 thoughts on “Milk, protein and the rise of plants”

  1. Spot on, as usual. As a vegan leaning veggie, I have explored a lot of these substitutes. Cheapo soya milk from Tesco is my go-to, though only to throw on meusli or put in tea. I don’t like it neat, and I feel exactly the same about cow/goat milk – just don’t like the taste. My son made me a cup of tea a while ago, and I asked him to check whether the soya had gone off before adding it. He asked, cruelly, how he would know if it had. I see his point. If you want a stand alone drink, I suggest you try the Alpro roasted hazelnut milk stuff. Often on offer for a pound a pop. Sugar is second on the ingredients list, though, so bear that in mind as well. There are now some non-dairy cheeses out there that make quite a good fist of pretending … until you try to cook with them. Just as meat eaters seem unable to give up bacon, I seem unable to give up proper cheese. Cornwall is the home of Yarg, after all. If you have some space, and want to grow your own protein, I recommend Gigantes Plaki, giant climbing Greek beans that are very productive and grow well in the UK. They dry well, and could easily become a year round staple if you were really financially stretched.


    1. Thanks Dave. Yeah, non dairy cheeses taste ok, but if you read the ingredients there’s nothing nutrionally useful in them. Makes me glad I’m not vegan!
      You trying to convince me to eat butter beans by giving them a posh name?! Tssk. Thanks though, it hadn’t occur to me to start growing protein on the allotment, I will give it some thought.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: