Update: 2014-11-01: Updated the rice protein section to use a much simpler calculation based off of the PDCAAS exclusively, since that does not require as many assumptions or as much data. Moreover, it turns out most of that information isn’t important.
I’ve been using my Soylent for about a week, and enjoying the taste and convenience. I’ve had friends express a couple concerns about the formula, and I have read some critical articles that range from analytical (“Good idea except for problems X, Y, and Z”) to sensational (“This will melt your insides and kill your chi”).
I wanted to share what research I found so anyone else interested in the same questions will have an easier time finding these answers. Some of these are already answered in these forums, but I figured it would be nice to have a lot of this in one place, instead of scattered across multiple threads. I also encourage anyone else with information on these topics to feel free to add.
Since it gets pretty lengthy, the topics in here so far are:
- Is rice protein a complete protein?
- Do you need to worry about soy and phytoestrogens in Soylent?
- The glycemic index and “sugar spike” of maltodextrin
- Is sucralose bad for you?
- GMO foods
- Absorbability of vitamins in soylent (not finished yet, may be added later)
Rice Protein Is/Isn’t a Complete Protein:
I have heard people say both things. Here’s what I found:
“Complete” proteins are proteins that have enough of all of the 9 essential amino acids (the amino acids the body can’t create). Brown rice is most significantly deficient in Lysine. It has all nine, but there’s not enough Lysine per serving to be considered complete. (Based on this: http://nutritiondata.self.com/facts/cereal-grains-and-pasta/5753/2)
The average adult is recommended to eat 0.8 g/kg of protein. For an 80 kg (180 lb) person, that’s about 65g a day. The protein source in soylent is specifically brown rice protein isolate (as of this blog post: http://blog.soylent.me/post/68180382810/soylent-1-0-macronutrient-overview).
Unfortunately, the calculation of the amount of lysine you get from rice protein is a bit complicated, because you need several numbers like the amount of lysine an adult must eat in a day, the amount of lysine in our protein, and the bioavailability (absorbability) of lysine in rice protein. It’s hard to find this last value, but what you CAN find something called the Protein Digestibility Corrected Amino Acid (PDCAA) value - a measurement of the “value” of a given protein based off of the amount and digestibility of the limiting amino acid. Rice Concentrate has a PDCAA about 50% (Less official sources found here: http://forums.truenutrition.com/showthread.php?38706-I-did-some-math-to-find-the-PDCAAS-optimal-blend-of-hemp-pea-rice-and-soy-protein and http://discourse.soylent.me/t/why-rice-protein/3680/36). I could not find the PDCAA of Rice Protein Isolate, but it feels safe to assume it is similar to Rice Protein Concentrate, since it should possess the same proteins in the same ratios and same forms, so I will assume Rice Protein Isolate has a PDCAAS of 0.50.
Here’s a formula for the PDCAAS:
PDCAAS = (Limiting Acid’s Bioavailability) * (Milligrams of Limiting Acid per Gram of Protein) * (Recommended Daily Protein Intake in Grams) / ( Milligrams RDA of Limiting Acid).
This is modified a bit from a formula found on Wikipedia:
PDCAAS = (mg of limiting amino acid in 1 g of test protein / mg of same amino acid in 1 g of reference protein) x fecal true digestibility percentage.
(Until someone corrects me), I have derived my formula based on the assumption that the “reference protein” is an “ideal amino acid ratio” protein where the following is true:
(Recommended Daily Protein Intake in Grams) / ( Milligrams RDA of Limiting Acid) = 1 / ( mg of amino acid in 1 g of reference protein). This assumption is supported by the description near table 10-24 in this article: http://www.nal.usda.gov/fnic/DRI//DRI_Energy/589-768.pdf.
Now, you could use the PDCAAS formula to calculate the bioavailability/digestability of lysine in this protein and then calculate exactly how much Lysine you absorb from this rice protein and compare that to your daily required intake (which is what I did originally), but there is an easier way. PDCAAS essentially represents the fraction of your daily required intake of the most limiting amino acid (lysine in this case) that you would get by eating a daily serving of the protein in question. This prevents us from having to find values (or make assumptions) about the amount of lysine you need in a day, the amount of lysine in rice protein, etc.
Since the PDCAAS of rice protein is 0.50, the daily intake for our 80 kg person is 65g, and we are eating 114g of rice protein in soylent, we are getting (114 g / 65 g) * (0.50) or 87% of our daily Lysine amount from rice protein.
So there’s not enough lysine in our rice protein for the average person, but luckily, we have another protein source in soylent: oat flour. Soylent has 110g oat flour as of this blog post: http://blog.soylent.me/post/68180382810/soylent-1-0-macronutrient-overview
I couldn’t find the PDCAAS of oat flour, but will assume that it is very close to the PDCAAS of Rolled Oats (0.57, based off a google search that returned several different unofficial websites citing the same number. Here is one: http://www.foodproductdesign.com/articles/2011/01/plant-based-proteins.aspx?pg=2), since theoretically, the forms and proportions of amino acids should be the same (or very similar) in both. I also am also assuming 7g of total protein per 40g of oat flour for the following calculation, based off of a company’s nutritional data for oat flour (http://www.bobsredmill.com/whole-grain-oat-flour.html)
So, the amount of our recommended intake of lysine that we get from oat flour is 0.57 * (110 g flour * 7g protein / 40g flour) / ( 65g protein a day) = 17%.
So, with 17% of our required intake from oat flour, and 87% from rice protein, that gives us 104% of our needed intake of lysine.
(For anyone looking for the original calculations and assumptions I made for this section, check the comments).
Addendum: one of the common trade names of brown rice protein, Oryzatein, claims to have a PDCAAS of 1.0, indicating that it is a fully complete protein (google returned many companies selling Oryzatein-made protein powders making this claim; here is one: http://instinctsportsnutrition.com.au/shop/brown-rice-protein/). In the blog post that explains that soylent uses brown rice isolate, it also specifically states that they are using a complete brown rice protein, suggesting that they may be using this product. I’m not sure how this higher PDCAAS is achieved (someone in the comments suggested that it might have something to do with the fact that cooking rice degrades the heat-sensitive Lysine, which lowers PDCAAS), but if it is true, then we have well more than the required amounts of all nine essential amino acids, and if that’s not the case, then we’re still in the clear even with a PDCAAS as low as 0.50.
Rice TLDR: even if brown rice protein is not complete, you are eating a higher than usual dose of protein and you are getting a second source of protein in the oat flour. Due to this, an 80 kg person will still get almost exactly their RDA of the most limiting protein - lysine.
Soy, Phytoestrogens, and Lecithin:
Despite the name, anyone who reads the ingredients of Soylent can tell you soy isn’t a big part of the formula, it’s only present as the emulsifier Lecithin. There’s 6g of it in a day of Soylent, which is a pretty small amount. Soy lecithen is also in a lot of other things you might eat (check anything with chocolate or peanut butter in it). I know that I am consuming way less soy by switching to Soylent. I didn’t delve into the processing procedure of lecithin yet, and will talk about GMOs later, but 6g is a very very small amount of soy, and probably pales in comparison to the amount of unusually processed or GMO food most of us eat every day ()
I did to delve into the phytoestrogens question, because some data suggested that lecithin was more estrogenic than regular soy (I’ll get to that later). This gets lengthy - there’s a TLDR at the bottom.
People claim that the isoflavones in soy (plant based estrogens) will cause gynecomastia and lower testosterone levels in men and contribute to some nasty cancer-related effects in women. As I did research, I focused on men (partially because gynecomastia is a scary word, but also because the data is much less conclusive for women).
I found loads of people both citing research saying that it definitively and scientifically causes testosterone problems in men, and people just as convinced that it was a myth. When two opposing sides pull out completely opposing scientific data, I tend to turn to meta-studies (i.e. studies that study multiple other studies - analyzing how good their data, analysis, and procedures were - and trying to come up with a consensus on the topic.) I found a meta-study that combined fifteen other studies and concluded that isoflavones (the soy phytoestrogens) do not cause any change in testosterone in men, nor does it affect sperm count, etc. (Source: http://www.fertstert.org/article/S0015-0282(09)00966-2/abstract). I am assuming that one or more bad studies suggested that isoflavones did affect testosterone, got a lot of press, and this meta-study looked at a wider range of studies and concluded those initial studies were bad or not reproducible.
Other facts about pytoestrogens in soy lecithin:
I found three studies about soy lecithin and isoflavones:
Two could not find any known isoflavones in the lecithin (Source: http://www.encognitive.com/files/Dietary%20Phytoestrogens,%20Including%20Isoflavones,%20Lignans,%20and%20Coumestrol,%20in%20Nonvitamin,%20Nonmineral%20Supplements%20.pdf and http://www.sciencedirect.com/science/article/pii/S0308814603001213).
The third had interesting results (very little of which you can see from the abstract - I had to buy it: http://www.ncbi.nlm.nih.gov/pubmed/21801783). They could not find any genistein (the main isoflavone in soy) in the lecithin when they tested it. However, when testing foods with lecithen in it, and when testing lecithin itself, using a yeast-based estrogen-equivalency test, they found high estrogen-like activity in both the lecithin and the lecithin foods.
Someone with research experience or education will have a few serious problems with the conclusion that this is proof that lecithen is estrogenic.
- While they measured the “estroginity” of the lecithin, they never give a number. They just say it was “significant.” Without numbers, can’t be considered conclusive evidence of anything.
- They could not find isoflavones in the lecithin, and admit that the measurement of it’s estrogenity might be an unknown factor or contamination.
- Their conclusion that lecithin contributed greatly to the estrogenity of foods was based off small sample sizes and very variable data that (shown below) are indistinguishable from statistical noise. You can’t see the values from the abstract, but here’s a rundown:
They tested three chocolates. Numbers 1 and 2 had lecithin, 3 did not. (EEQ is some measurement of equivalent estradiol effect).
Chocolate 1: ~250 EEQ
Chocolate 2: ~90 EEQ
Chocolate 3: ~45 EEQ.
They also tested five baby formulas. Numbers 1-3 had lecithin, 4&5 did not. They measured:
Formula 1: ~21 EEQ
Formula 2: ~1 EEQ
Formula 3: ~20 EEQ
Formula 4: ~14 EEQ
Formula 5: ~4 EEQ.
They concluded that, except for formula 2, all lecithin foods had more EEQ than the non-lecithin versions, so lecithin must be the cause. While that is a tempting conclusion, their sample sizes are much to small to make such a conclusion: they have two lecithin chocolates, one non-lecithin chocolate, three lecithin formulas, and two non-lecithin formulas. The largest sample size is three. If you calculate the averages and standard error ranges of these, you get the following:
Lecithin Chocolate EEQ: 170 +/- 80
Non-Lecithin Chocolate EEQ: 45 +/- (infinity) (sample size = 1)
Lecithin Formula:14 +/- 7
Non-Lecithin Formula: 9 +/- 5
The measurements of the EEQ of Lecithin vs. Non-Lecithin foods fall within the error ranges of each other in both cases, and therefore no conclusions regarding the data can be made.
For people who don’t understand or feel comfortable with the idea of standard error, here is a more intuitive way to think of the same problem: For the chocolate set, what if they tested a second non-lecithin chocolate and it had a higher EEQ than chocolate 1? Their conclusion would be destroyed by adding only one new point. That is a very dangerous situation to be in: it’s like flipping a coin once, seeing heads, and assuming coin flips always land on heads. The formula set is even worse, considering the one with the lowest EEQ included lecithin. If they tested another lecithin formula and it was just as low as formula 2, the average EEQ of lecithin and non-lecithin formulas would be almost the exact same.
Phytoestrogen in lecithin TLDR: The only study suggesting that lecithin has estrogenic qualities has taken liberties applying conclusions to their data, and when accounting for standard error ranges is inconclusive. Three different studies (including the one mentioned above) conclude that they cannot find any isoflavones in soy lecithin. Moreover, while there are studies that both support and debunk the idea that isoflavones reduce testosterone in men, the prominent meta study sides with the idea that there is no effect on testosterone - my assumption is that one or two bad studies got very widely reported in the media, and the majority of these studies disagree with them.
Maltodextrin and Sugar Spikes/Glycemic Index:
Many people seem specifically concerned that Maltodextrin is the main ingredient in soylent. Their concern is that, even though it’s technically an oligosaccharide (a medium-length chain of sugar - longer chains tend to digest more slowly) it is digested uncharacteristically quickly, resulting in a large spike of blood-sugar and insulin shortly after drinking it, much like drinking a very sugary beverage would (unlike, say, oatmeal, which is a longer chain that is digested slowly). Maltodextrin is not treated as a sugar on nutrition labels (so you won’t see much sugar on the nutrition info on Soylent), but it might be treated very similarly by the body. It’s worth noting that there IS a slow burning carbohydrate in soylent, as well: oat flour.
First off, there’s some smart people posting in the forums about this. A lot of my information is just reposted from this topic: http://discourse.soylent.me/t/dextrose-equivalent-glycemic-index-maltodextrin-sugar-and-soylent/16036
Maltodextrin can range from chains of sugars 2 to 20 units long. The longer chains are digested more slowly, avoiding this insulin spike, and the shorter chains are basically treated like table sugar. The Soylent blog claims that the Dextrose Equivalent (DE) of the maltodextrin in Soylent is 10 (maltodextrin’s DE can range from 3-20, 3 being the longest chains, 20 being the shortest chains). I’m honestly not sure how to convert DE to chain length, but I’d assume it’s somewhere in the middle of Maltodextrin’s chain range of 2 to 20 units (probably also about 10 units, not to be confused with the DE, which is also 10).
The Glycemic Index (GI) - essentially a measurement of how the food affects your blood glucose, which varies based off of the chain length, but according to an (unfortunately uncited) Wikipedia article on maltodextrin, the GI of maltodextrin ranges from 85 to 105 (source: http://en.wikipedia.org/wiki/Maltodextrin). I can’t find any information on how to convert from DE of maltodextrin to GI, but I will once again assume it’s somewhere in the middle based on the mid-range DE: I am going to take a Fermi-esque guess that we have a GI in the neighborhood of 95.
Note: I saw two ranges for the GI of Maltodextrin: one site (here: http://www.pcosupport.org/newsletter/articles/article121008-3.php) claimed it was 106-136. Wikipedia, and another site seemed to support the 85-105 range (here: http://www.sugar-and-sweetener-guide.com/maltodextrin.html) but NONE of these are official sites, and since the Wikipedia article is uncited, it they might simply be citing each other. I’m going to go with the lower number, because it turns out, either way, it’s too high.
The scale is built so that a GI of 100 corresponds to glucose. At an assumed GI of 95, even our 10 DE maltodextrin has a bit of a problem. Wikipedia says that a low GI food roughly 55 or less, a high GI food is 70 or more.
Evidently taking carbohydrates with fiber is evidently very beneficial for reducing the insulin spike of high carbohydrate (and high GI) meals:
Several studies have shown that the adverse metabolic effects of high-carbohydrate diets are neutralized when fiber and carbohydrate are increased simultaneously in the diet for diabetic patients. In particular, these studies demonstrated that a high-carbohydrate/high-fiber diet significantly improves blood glucose control and reduces plasma cholesterol levels in diabetic patients compared with a low-carbohydrate/low-fiber diet. In addition, a high-carbohydrate/high-fiber diet does not increase plasma insulin and triglyceride concentrations, despite the higher consumption of carbohydrates. Unfortunately, dietary fiber represents a heterogenous category, and there is still much to understand as to which foods should be preferred to maximize the metabolic effects of fiber. There are indications that only water-soluble fiber is active on plasma glucose and lipoprotein metabolism in humans.
Fortunately, Soylent 1.1 has fiber (27g/day; women are reccommended 25g a day, men 38g). But is it enough?
Here’s where I get stopped. I’ve exhausted my fund for buying expensive research articles today, but I would love to look inside that article and hope that they have a chart or some data that would tell me how much different quantities of fiber affects foods of different GI. Is anyone reading this a university student? Many universities offer access to pubmed or other journals through their libraries. I would love to see the inside of this article.
In the meantime, I will be trying to split out this insulin spike by drinking small amounts of Soylent throughout the day - hopefully “faking” the effect of eating a slow-digesting carb. I’m not sure if that is good science, but it makes me feel better.
It’s worth noting that the Soylent blog had this to say about the numbers:
Preliminary tests by beta testers and founders abiding by WHO glycemic index testing guidelines have found the GI to be rather low. More formal testing is planned for early 2014.
But they did not mention any specific numbers. One of the developers also mentioned in a forum topic that they originally used the more expensive tapioca derived maltodextrin - the slowest digesting maltodextrin, but the Soylent 1.1 recipe uses corn based maltodextrin - the fastest digesting maltodextrin. (Source: http://discourse.soylent.me/t/why-use-maltodextrin-at-all/5338/5)
Maltodextrin TLDR: Conservative guesses still put the glycemic index of the maltodextrin in our soylent uncomfortably close to straight glucose - very high. Eating high GI foods along with fiber, which is in Soylent, supposedly abates this problem, but without seeing inside the article linked above, we can’t know exactly to what quantity. I’m refraining from consuming large quantities of Soylent at once, and instead drinking it constantly throughout the day in an attempt to offset this, but I don’t know enough about nutrition to know if it works like that.
My research on sucralose was pretty limited to the Wikipedia page (here: http://en.wikipedia.org/wiki/Sucralose) because it started to look pretty conclusive, so most of this data is from that page:
Sucralose was discovered in the 70s, and waited through about 30 years of testing before it was declared fit for US citizens to eat. In that time, over 100 animal and clinical trials unanimously declared it safe. Note that safe and healthy aren’t quite the same thing, but it’s a chemical that is so sweet, you only end up eating trace amounts of it (it’s the least prevalent ingredient in Soylent), so theoretically it should have very little effect on your body.
The critiques of using Sucralose that I saw referenced a specific Duke University study. It found that feeding Sucralose to rats reduced their gut flora, caused some other intestinal problems, and promoted weight gain.
However, it turns out that this study was funded by the Sugar Association (who stands to lose a lot of money to Sucralose, and therefore cannot at ALL be considered objective). Moreover, a committee of experts from Rutgers, Harvard Public Health, New York Medical, and Duke itself convened and wrote a journal article concluding that the original study was “not scientifically rigorous and is deficient in several critical areas that preclude reliable interpretation of the study results.” (Suck it Dookies, go Heels). (Source: http://www.foodnavigator.com/Science-Nutrition/Sucralose-safety-scientifically-sound-Expert-panel).
The other two notable studies reporting health effects of eating sucralose were also not convincing: one found that it caused DNA damage in mice, but the dose was so large it would correspond to a human eating 11,450 packets of splenda (Source: http://www.sciencedirect.com/science/article/pii/S1383571802001286). Another article suggested that it affected the insulin response in obese people, but the sample size was fairly small (17), and previous testing suggests sucralose does not affect blood sugar in most adults (source: http://care.diabetesjournals.org/content/36/9/2530).
(Further research needed: I’ve heard about articles that supposedly prove that artificial sweeteners cause a of metabolism, resulting in weight gain. It involved something about making the body think that it was getting sugar but not able to digest it, so it holds onto the sugar it DOES get like some kind of ''starvation mode." I didn’t have time to look into this - you might want to, but I’m less worried, since the maltodextrin provides the body with carb calories to digest whenever we taste the sucralose).
Sucralose TLDR: About 100+ scientific studies find sucralose to be fairly safe (and have negligible effect on humans). Three prominent articles disagree, and have gained a lot of press for it, but two of them have serious conflicts of interests or procedural deficiencies, and the third doesn’t match previously reproduced data. Sucralose is probably very safe.
GMO Corn and Soy:
I’ll admit that I haven’t looked at hard studies for this one yet, but the concern is about the fact that the soy (lecithin) and corn (maltodextrin) used to make Soylent are GMO-sourced, and all GMO things everywhere are bad for you. Here’s my very opinion-based but hopefully elucidating two cents:
Don’t worry about the lecithin. Except for the most diligent informed consumer, we’re all getting more that 6g a day of GMO stuff. The poison is always in the dose - you don’t worry about the arsenic in the apples you eat or the radioactive metals in the bananas (both are naturally present), so don’t worry about this.
That leaves the corn maltodextrin. I’m not informed enough to tell you specifically about this, but maybe I can get you to think of it better:
First off, how much of the corn actually gets left in purified maltodextrin? Do you need to worry about that? There’s no/negligible protein in it, which would suggest that maltodextrin are purified to the point where any GMO protein, DNA, or chemical is no longer present, which would make GMO maltodextrin and non-GMO maltodextrin essentially the same. Maybe this can be an assignment to look into on your own, or perhaps it is something I can eventually come back around to.
Second, a point about throwing GMO foods out altogether:
GMO foods are a much broader beast than we are led to believe. Saying “GMO food is bad for your gut” is like saying “things that come from China are bad for your teeth.” It’s simply more complicated than that: everything depends on what genes/traits have been added to the plant. Most people quote the gut statistic because there is a variety of corn (Bt-Corn) that produces a pesticide naturally. Theoretically that pesticide doesn’t affect humans, but there’s some conflicting evidence that it leads to gut problems, which might make you want to stay away from it until we are sure. However, that’s not the only GMO crop, or even the only GMO corn. Here are some others I know about:
- Bt-Corn - mentioned above, creates a pesticide that might be bad for your gut. Ruh roh, might be worth avoiding until we figure out the science.
- Pesticide-Resistant (Corn, Everything) - potentially better than Bt-corn in the sense that the plant you’re eating doesn’t create pesticides, but potentially worse in that it allows us to use way more pesticides in farming, which could lead to environmental problems, and your food might be soaked in pesticides when you receive it. (Don’t fool yourself though: we use a ton of pesticides on non-GMO crops, too, and have been breeding them for increased resistance (so we can use even more) for years, even without using GMO techniques.)
- Bt-Cotton - cotton that produces it’s own pesticide. Better than Bt-corn because you aren’t eating it, so the gut worry goes away. There was some very sensationalist press about it causing suicides in India - from what I read, most people tend to agree that is essentially an invented statistic that caught a lot of press because it was exciting. All plants produce natural pesticides; Bt-cotton just produces a new one, which is kind of nice because you don’t have to create petroleum-based pesticides or spray them all over a field, avoiding some environmental problems.
- Golden Rice - Rice that has been engineered to produce vitamin A. Really not that bad, potentially saving lives in places where people can only eat rice and have vitamin A deficiencies.
- Drought Resistant Corn - Corn engineered to be more resistant to lack of rain. I think they use genes from other, hardier plants, so basically they have added genes and proteins that you are already eating (). (Evidently this strain doesn’t work so well, but hey, that doesn’t make it poison.)
See how there’s kind of a spectrum? Some of those are sort of suspicious, and some are much less frightening. My point is that you should be careful when you hear “all GMOs are bad because X.” They are a varied breed.
GMO TLDR: I don’t know. My opinion (not fact) is that there’s too little lecithin to care about the soy part, and I would suspect (but I don’t have any data either way) that extracting maltodextrin from GMO corn and non-GMO corn is essentially identical (i.e. any GMO proteins/DNA gets left behind with all of the other matter you are extracting it from). I would also say that it’s worth educating yourself about individual GMO plant breeds before throwing them all into the same group. There are multiple breeds for any plant. Some GMOs might be potentially scary, some are not at all, and they tend to get treated like one big thing. Vitamin producing? Not bad. Pesticide producing? Maybe I’ll stay away.
Soylent vitamin abosrbancy:
The concern here is that I have seen complaints that, while Soylent contains 100% RDA of most vitamins and minerals, it has poorly absorbed versions where you may not get the full 100%
Unfortunately I haven’t researched this one yet, but I saw it in one of the more sensationalist critiques. I feel like it’s worth some research, but I haven’t gotten there yet. I might not end up circling back around for this - that’s a lot of things to look up, but if someone in the comments gets excited about this topic I’d love to add it up here.