Sunday, February 1, 2015

Mexican coke is unlikely to be superior to regular coke (HFCS vs. sugar debate)


There is a widely known passage in one of Peat's articles on the actual energy content of high fructose corn syrup-(HFCS)-containing beverages. This passage is often used by peatarians to explain the weight gain seen in populations with high HFCS-soda consumption and serves as a basis to prefer sucrose-sugared products compared to HFCS-products (such as mexican coke vs. US-american regular coke).
"Much of the current concern about the dangers of fructose is focussed on the cornstarch-derived high fructose corn syrup, HFCS. Many studies assume that its composition is nearly all fructose and glucose. However, Wahjudi, et al. (2010) analyzed samples of it before and after hydrolyzing it in acid, to break down other carbohydrates present in it. They found that the carbohydrate content was several times higher than the listed values. "The underestimation of carbohydrate content in beverages may be a contributing factor in the development of obesity in children," and it's especially interesting that so much of it is present in the form of starch-like materials." - Ray Peat in Sugar Issues

Relevant studies

The study referenced by Peat was a conference abstract, not a full article. Conference abstracts are in most cases not peer-reviewed and submitted at stages were your whole study is not finished yet (because you have to submit those often half a year earlier). This year, a full article was published by White et al. that tried to validate these findings by also analyzing the carbohydrate content in HFCS sodas. They concluded the following:
"Total sugars content of commercial beverages is consistent with common industry practices for canned and bottled products and met the US Federal requirements for nutritional labeling and nutrient claims. Prior concerns about composition were likely owing to use of improper and unverified methodology".
In their article, they also critcized the conference abstract by Wahjudi et al. I took this as a basis to personally contact the senior author of the Wahjudi study, Dr. Paul Lee. He wrote me the following in response:
"White is right about our abstract. We were not able to repeat the observations, and found out that the calibration samples were not properly prepared leading to systematic errors. "
On the question about whether mexican or US-american regular coke is better, the following things are interesting to know:
  • Mexican coke ingredients: Carbonated water, sugar, caramel color, phosphoric acid, natural flavors, caffeine (150 calories in 355ml)
  • Regular coke ingredients: Carbonated water, high fructose corn syrup, caramel color, phosphoric acid, natural flavors, caffeine (140 calories in 355ml)
  • In an interesting study, people consumed regular or mexicon coke in a blinded or non-blinded fashion (knowing what they drank or not and switching the container). It showed that when people don't know the content and just go by taste, people actually prefer the regular US-american and not the mexican coke. When there was a preference for mexican coke, it could better be explained by the preference of drinking out of a glass bottle vs. an aluminium bottle or by psychological perception, when the subjects were told they were drinking mexican coke (even though they actually got regular coke). This study emphasizes the decisive role of the mind when it comes to perception of foods.


The widely cited result by Wahjudi et al. is based on a systematic error. HFCS beverages don't contain more calories than listed on the labels. Given the wide range of studies showing that HFCS is practically identical in metabolic behaviour compared to sucrose, any adverse or beneficial effects seen in studies using HFCS as a calorie source would likely also be seen with the same amount of sucrose.

Mexican coke even contains slightly more calories than regular US-american coke. So all in all, you might just as well save the extra money and drink regular coke from a mexicoke glass container.

Higher metabolism, temperature and pulse and lower TSH associated with higher mortality


A central topic of peatarianism is to focus on increasing the metabolic rate using various methods. It is hypothesized that this leads to better health and longevity. Ray Peat also says that having a TSH at or below 0.4 would be optimal.
My recommendation is to eat to increase the metabolic rate (usually temperature and heart rate), rather than any particular foods. - Ray Peat (e-mail response)

I think it's best to keep the TSH around 0.4 - Ray Peat
(e-mail response)
I think it's good to have TSH below 0.4, and that probably contributes to loss of hair. - Ray Peat (e-mail response)

What does the evidence say? 

  • In a prospective study in men with around 40 years of follow-up, a higher basal metabolic rate was independently associated with a higher mortality compared to average metabolic rates. Low metabolic rates were not significantly associated with higher mortality (Figure: Mortality plotted according to quartiles of metabolic rate). enter image description here
  • In the same study, a higher temperature was associated with increased mortality compared to people with a lower temperature over 25-years of follow-up (Figure: X-axis: Survival time (yrs), Y-axis: Cumulative survival; plotted for higher or lower body temperature). enter image description here
  • Mice that have been genetically engineered to have a 0.3 to 0.5°C lower body temperature have an increased life span compared to normal mice despite having the same caloric intake.
  • In a prospective study in male and females with average 11 years of follow-up, a higher 24 hour energy expenditure was associated with higher mortality from diseases.
  • In a 16-year follow-up study, an elevated resting heart rate in men without cardiovascular disease is associated with an increased all-cause mortality, even when adjusting for cardiovascular risk factors. A lower resting heart rate wasnt associated with higher mortality (Figure: All-cause mortality according to groups of resting heart rate). enter image description here
  • In a prospective study of 52.674 participants with median 8.8 years follow-up, a TSH under 0.45 mIU/L is associated with increased all-cause mortality, cardiovascular mortality and atrial fibrillation compared to a TSH between 0.45 to 4.49 mIU/L
  • In two cross-sectional studies of long lived individuals, longevity was associated with lower fT3 and fT4 and higher TSH values
  • A genetic polymorphism that leads to higher TSH is associated with longevity
  • In a prospective study of 85-year olds with 4-years of follow-up, a TSH over 4.8mIU/L is associated with significantly lower mortality compared to subjects with average or suppressed TSH levels. Low serum T4 combined with high TSH was also associated with lower mortality.
  • Centenarians had a higher TSH compared to younger controls. No differences in fT4 values were found.
  • Italians "oldest-old (90-107 years) had a higher TSH, lower fT3/fT4 ratio and lower serum zinc and selenium than adult or elderly control subjects.


Having a higher metabolic rate, temperature or pulse or a TSH below 0.45 is associated with increased all-cause mortality. The underlying cause of these association is unclear. High metabolism or suppressed TSH can be a consequence of stress and elevated stress hormones, but also of high thyroid hormone action. The biomarker studies suggest that a slightly suppressed thyroid function may be protective, especially in elderly people. This effect may or may not be explained by the nutrient-sparing effects of hypothyroidism in populations prone to malnutrition.

Given the lack of trials that show that increasing temperature and pulse by dietary measures or suppressing TSH with thyroid medication actually improves health, any broad recommendation  to do should be stated in a careful manner. People need to be aware that they conduct a personal experiment when pursuing such hypotheses, especially when doing so for long periods of time.

All in all, it seems adequate to say that there is a narrow range between both hypo- and hypermetabolism that is associated with health in the long term.

The daily raw carrot is not antiseptic, it increases bacterial growth in the gut


Ray Peat's daily raw carrot is one of the most popular peatarian recommendations. Yet, it is among the recommendations that are the least questioned.
A daily raw carrot, shredded with a little vinegar and olive oil, can suppress bacteria. - Ray Peat
Sometimes having a daily carrot salad (grated, with salt, olive oil, and a few drops of vinegar) will stimulate (and disinfect) the small intestine enough to prevent fermentation. - Ray Peat

Where is the evidence for such claims?

When Peatarians are asked, they usually just quote 1-2 in vitro studies on the antimicrobial effect of carrots, such as this one. Although they sound convincing, caution should be applied when generalizing the evidence on much more complex systems such as the human gut with its microbiome and different compartments (i.e. stomach, small intestine, colon).

Interestingly, there is one older publication out there that tested the effects of raw carrots in humans. This one is also often quoted by Peatarians because it shows a cholesterol reducing effect of carrots, which is somewhat in line with Peat's observation of changes in estrogen/progesterone ratio (clear study data on that are lacking unfortunately).

Unfortunately, the second result of the study is only rarely discussed. The authors measured stool mass/fats and breath hydrogen, the latter being a marker for bacterial fermentation. This is what they found:

  1. The raw carrot more than doubled breath hydrogen, which is a marker for bacterial fermentation. This effect started to occur after around 10 days of starting the raw carrot.
  2. The excretion of fecal bile acids and fats was still increased 3 weeks after ending the carrot phase. 
Similar results have been obtained in another study. It compared the fermentation of different types of fibers in rats and man. Just like above, carrots seem to be fermented quite efficiently, similar to apples. Wheat bran in contrasts is one of the fibers that appears to be largely undigested. 


These studies show that the raw carrot is not "antiseptic" in the gut, but the opposite, that it increases overall bacterial mass when consumed regularly. The change takes several days to take place sugesting a slow adaption of the microbiome with a proliferation of bacteria that thrive on carrot-fiber. The change in microbiome may also explain the long lasting effect on fecal bile excretion.

By no means I am suggesting that eating the carrot is bad for you just because it favors microbial growth. The effect probably depends on how your microbiome reacts. I've heard from some people (2 of them with inflammatory bowel disease) that they got terrible stomach upset from eating the carrot. Here, the carrot seems to bring the microbial balance out of whack. In others, it might regulate things and lead to growth of beneficial bacteria while suppressing the "bad" ones.

In the light of these results, eating carrot fiber each day, but careful avoiding all other fibers doesn't make that much sense. Similar fibers such as from fruits or other vegetables could be equally beneficial (or deleterious, depending on the person). By the way, other fiber sources have also been shown to lower estrogens suggesting that the hormone regulating effects are not limited to the carrot but may apply to fiber in general.

Exercise at reasonable doses is beneficial to health


Even though you never specifically read him say "Exercise kills you", there is a distinct undertone in his writings against any kind of vigorous exercise, especially endurance exercise. He is only supportive of concentric weight training, low-intensity activities such as walking or very short high intensity exercise, but he also doesn't seem to mind much when one lives entirely sedentary. 
“The overlapping effects of estrogen, polyunsaturated fats, exercise, serotonin, histamine, lactic acid, nighttime, and hyperventilation, tend to be cumulative and self-stimulating. Degenerative changes in tissues are accelerated by all of these stress mediators.” - Ray Peat
"Exercise increases blood clotting, and so can increase the risk of strokes and heart attacks. [...] Walking is a better form of exercise." - Ray Peat
"Exercise, like aging, obesity, and diabetes, increases the levels of circulating free fatty acids and lactate."- Ray Peat
"Hypothyroid people, who are likely to produce lactic acid even at rest, are especially susceptible to the harmful effects of “aerobic” exercise. The good effect some people feel from exercise is probably the result of raising the body temperature; a warm bath will do the same for people with low body temperature." - Ray Peat
“Endotoxin (like intense physical activity) causes the estrogen concentration of the blood to rise.” - Ray Peat
Those are some pretty harsh statements. But are they actually true? Does endurance exercise really cause degenerative diseases such as heart disease? Is resistance exercise really better than endurance exercise? Does endurance exercise increase estrogen? Is it better to just sit on the sofa?

Evidence from association studies

Looking in the study databases will quickly give you loads of studies that indicate that exercising more is associated with reduced mortality and morbidity compared to not exercising at all or only rarely:

  • Mortality of elite (!) endurance and mixed endurance-resistance athletes is reduced compared to the general population. For pure resistance athletes less consistent results were found, some found increased mortality.
  • In patients with coronary heart disease, regular strenuous physical activity between 5-10 hours per week (or 2-6x per week) is associated with reduced cardiovascular and all-cause mortality. At very high regular physical activity levels (daily strenuous activity; over 19 hours per week) there was an association with increased mortality, albeit still lower compared to people who did not exercise at all.
  • Regular joggers have significantly reduced mortality compared to non-joggers.
  • A high amount of sedentary activity is associated with higher all-cause, cardiovascular and cancer mortality, irrespective of physical activity.                      
  • People that newly start vigorous physical activity compared to people who don't change their activity level is associated with reduced all cause mortality, to a similar level of people who were continously active through the whole surveillance period.

Evidence from interventional studies

Now, a knowledgable peatarian will say that these are just population studies, and it is clear that people who are healthier in the first place move more, while people who are sicker move less (although the last study somewhat argues against that). Point taken, so lets look at the randomized controlled trials of exercise interventions which eliminate this problem:

  • A meta-analysis of randomized controlled trials of total 10,794 patients with coronary heart disease who got exercise-based rehabilitation vs. usual care showed that exercise reduced cardiovascular mortality and showed a trend of reduced all-cause mortality.
  • A meta-analysis of randomized controlled trials of total 2,188 patients with stroke comparing endurance, resistance, mixed and no training, showed reduced disability scores with exercise training. Not enough data was available for a conclusion on mixed or resistance training or on total mortality (although there were less deaths in the exercise groups).
  • A meta-analysis of randomized controlled trials of total 4740 patients with heart failure comparing exercise vs. no exercise showed a trend of reduced all-cause mortality in the exercise group and significantly improved quality of life and reduced hospital readmissions.
  • A notable long term study in heart failure patients compared 10 years of relatively boring aerobic cycle ergometer or treadmill exercise (moderate intensity; twice weekly for one hour) vs no exercise training and found significantly reduced cardiac mortality and hospital readmissions along with improved quality of life.
  • A randomized controlled trial in 262 sedentary patients with type 2 diabetics showed a reduction of insulin resistance by both endurance and resistance exercise (a slight trend of endurance being better). A combined training showed the best results.
  • A randomized controlled trial in 154 adults showed that endurance exercise with jogging, biking and brisk walking reduced duration and severity of respiratory infections compared to controls. Beneficial effects were also seen in an interesting third group of meditation.
  • A meta-analysis of randomized controlled trials in breast cancer patients showed that exercise interventions significantly improved quality of life and reduced fatigue compared to no exercise. The trials were not powered to show differences in mortality.
  • A randomized controlled trial of endurance exercise in overweight postmenopausal women showed that exercise significantly reduced serum estrogen levels compared to no exercise.
  • A randomized controlled trial in patients with mitochondrial myopathy showed improved symptom scores and higher exercise capacity in the exercise vs the control group


There is very good evidence from both observational as well as interventional studies that exercise is healthy. The argument that resistance exercise is to be preferred and endurance exercise to be avoided is not supported by the data. As with nearly everything in life (food, supplements, lifestyle), there is a J-curve like distribution, with both too much and too little exercise probably being harmful. But even acknowledging that, extreme exercise is still associated with lower mortality than no exercise at all. 

Another popular argument is that you should not exercise when you are hypothyroid or sick. This is also not supported by the data, as seen by the effect of exercise in people with heart failure and other disorders that are associated with severe metabolic and mitochondrial impairment. If exercise had made those metabolic conditions worse as Peat suggests, it would have also lead to worse outcomes. But as the opposite happened, it is likely that exercise improved both metabolism and mitochondrial health (e.g. 1, 2).

So in summary, the anti-exercise sentiments in Ray Peat's articles will likely deter more people from conducting reasonable-dosed exercise, than educating the small fraction of extreme-athletes from bringing their activity level down to physiologic levels.

Tocotrienols and red palm oil have more beneficial effects than expected


Tocotrienols are the unsaturated versions of vitamin E. 

Ray Peat is not very fond of them as he wrote:

"The unsaturated tocotrienols have hardly been tested for the spectrum of true vitamin E activity, and animal studies have suggested that it may be toxic, since it caused liver enlargement." - Ray Peat
I experimented with red palm oil and found it beneficial for my skin health. As red palm oil is very rich in tocotrienols, I really wondered whether this vitamin E form is that bad. 

(Picture Source:


While it is true that publications on tocotrienols only form around 1% of all publications on vitamin E, I couldn't find evidence that they cause liver damage.

Studies documenting a liver protective effect of tocotrienols

Other benefits of tocotrienols

Red palm oil vs. other oils

Unrefined vs. refined palm oil


The European food safety authority concluded that tocotrienols are safe. In rats no adverse effects of tocotrienols were found for doses up to 120mg tocotrienol extract/kg bw/day (which would equal 8400mg/day in a 70kg person).
In regard to humans another review concluded: "Studies in humans at levels of 50–400 mg/day (equivalent to up to 6.7 mg/kg for a 60 kg human) for periods of 2 weeks to 18 months (56 days for the 400 mg/day study) have not been reported to cause adverse effects, other than occasional transient effects. Taken together, and given the pharmacokinetics of tocotrienols including their short half-life, consumption of 3–5 mg/kg/day, and possibly higher, would not be expected to cause adverse effects." One tablespoon (10g) of red palm oil only contains 5-7mg of tocotrienols, that is much less than any of the limits above.

In summary, I question RP's claim on the toxicity of tocotrienols. Given the repeated failures of alpha-tocopherol supplements in reducing heart disease or cancer in large trials (even the natural forms), more emphasis should be put on researching the less known vitamin E forms such as gamma-tocopherols or tocotrienols.

Eating more frequently not necessarily better than eating less frequently


According to Peat, people with metabolic problems should eat more frequently. 
Frequent meals are helpful during hypothyroidism, and help to prevent obesity, but when the thyroid and liver are working, 2, 3, or 4 meals are good. - Ray Peat (e-mail exchange)

Relevant studies

  • A recent randomized study (1; fulltext) compared two calorie-reduced diet patterns in people with diabetes. In one group people ate breakfast, lunch, dinner & snacks in between (6 meals), in the other just breakfast & lunch (2 meals), the total calorie intake was the same. The study showed better weight loss, glycemic control and less liver fat in the group that eats less frequently. Interestingly, the basal metabolic rate seemed to be worse in the group eating more frequently, albeit non-significant.
  • Another controlled study (2) compared the effects of meal size and frequency (2 vs 5 high-fat meals with the same calorie content) on endotoxinemia in lean and obese women. It found that endotoxin levels were significantly higher in the 5 meal group vs the 2 meal group. No significant differences in metabolic rate were noted.
  • An older study (3) compared the thermic effects of a single meal vs several small meals over a three hour period in women. The thermic effect of food was higher when only a single meal was consumed compared to the several small meals.
  • On the other hand, in a study in healthy young males (4), appetite control was improved when eating food at 5 different instances compared to a single meal, the latter leading to higher rises of insulin.
  • The metabolic effects of regular food timing (6x/day) compared to irregular eating times (3-9x) were studied in ten women (5). Women that ate regularly had a greater thermogenesis by food and lower insulin responses than the irregular eaters.


There is no clear-cut benefit of eating more or less frequently. Metabolically-sick people can in some instances even benefit from less frequent eating (e.g. above diabetes study). Regular eating at the same times each day seems to be beneficial for the metabolism.

High salt intake can increase catecholamines and blood pressure in certain contexts


The previous article talked about the calcium-excreting effects of high salt intake which can be problematic for people on low-calcium diets. Another potential problem are the context specific effects of salt on catecholamine and blood pressure.

Relevant studies

Study 1: 20 healthy men consumed three diets with varying salt content
  • Low salt: 0.6-6.1g salt/day
  • Normal salt: 6.1-10.5g salt/day
  • High salt: 10.6-15g salt/day (Normal salt diet + 9x0.5g salt capsules)
They looked at norepinephrine, epinephrine and renin. Renin decreased with increasing salt intake. In regard to norepinephrine and epinephrine, they found that the low salt diet caused the highest levels that was significantly lowered on the normal salt diet. However, the high salt diet lead to a signficant increase in norepinephrine and a trend of increased epinephrine levels.

Study 2: 10 healthy controls and 20 patients with hypertension consumed diets with varying salt intake
  • Low salt: 0.6g salt/day  
  • Medium salt: 5.9g salt/day
  • High salt: 11.7g salt/day
They then looked at effects on blood pressure an various stress hormones. They found that a higher salt had no blood pressure increasing effect in healthy individuals but lead to a significantly increased BP in previously hypertensive subjects. From these subjects, they identified 12 salt sensitive and 8 salt resistant people. In all subjects a lowering of aldosterone and renin was seen. In controls and salt-resistant hypertensives, norepinephrine and epinephrine decreased also with increasing salt intake. However in salt sensitive hypertensive patients, norepinephrine increased on the high salt diet.

 enter image description hereenter image description here



The results underline the general anti-stress effects of salt, especially on the renin-aldosterone system. However, they also indicate that at very high salt intakes and/or in salt-sensitive people, increasing salt could have detrimental effects. Besides causing calcium loss (1), salt can worsen hypertension and increased plasma catecholamines in susceptible people. These mechanisms could explain the increased rates of cardiovascular death in people with salt intakes higher than 15 grams/day compared to moderate salt intakes between 8 and 15 grams per day (2).

High salt intake increases urinary calcium loss


Peat supports adequate salt intake. He has plenty of reasons to do so, given that salt restriction increases aldosterone, which promotes potassium loss and has degenerative effects when chronically activated.

However, he does not talk much about the effects of too much salt. Some people may conclude from his work that the more they salt their food, the better it is. But there are some caveats that you should know.

One caveat is that an increase of dietary salt increases the loss of calcium through the urine. This has been shown in both controlled animal and human studies. A possible explanation for this effect is that channels, that are important for reabsorbing calcium back into the body, are blocked by the excess sodium molecules secreted into the urine on higher-salt diets.

Studies where salt intake was increased

  1. Increasing dietary salt increases calcium loss in rats (1). In rats fed a low calcium diet, this results in activation of the parathyroid gland and osteoporosis (2) (3). In one study bone loss was also observed in rats on a normal calcium diet (4).
  2. In postmenopausal women, adding 3-6g of salt daily significantly increased calcium excretion (along with increased sodium and chloride excretion) and increased total urine volume (5).
  3. In healthy young women, NaCl but not KCl increased urinary calcium excretion (6).
  4. In a cross-over study on postmenopausal women, only a high-calcium/low-salt diet lead to a positive bone calcium balance, while on a high-calcium/high-salt diet bone calcium was unaffected. The low-calcium/low-salt and low-calcium/high-salt diets lead to a negative bone calcium balance (7).
  5. In 23-76-year old adults a high-salt diet increased urinary calcium and PTH compared to a medium-sodium diet (8).

Studies where salt intake was restricted

  1. In healthy subjects, a 4 day low sodium diet reduces urinary calcium excretion (13).
  2. In patients with urinary calcium stones and high calcium excretion, a 3 month long salt-restricted diet reduced urinary calcium excretion (9).
  3. In healthy 21–39 year-old adults, a salt-restricted diet didn't significantly decrease calcium absorption, while leading to a slight increase in PTH (10).
  4. In healthy postmenopausal women with a sufficent calcium and vitamin D intake through supplements, overrestricting sodium under 1500mg/day lead to unfavorable effects on bone density compared to women who ate ~3000mg/day (11).

Association studies

  1. In an association study, sodium excretion was the leading dietary factor associated with urinary calcium excretion (12).


  1. Overconsuming sodium will increase urinary calcium loss and this can be especially problematic for people that have a low calcium intake (i.e. the general public that doesn't consume much dairy but plenty of salted processed foods). Excessive urinary calcium loss may promote osteoporosis, hypertension and the development of kidney stones. This relationship supports the notion of eating salt in balance with the other minerals (calcium, magnesium and potassium).
  2. The mechanism may explain why some people tolerate high amounts of milk better with added salt (the increased urinary calcium excretion could regulate excessive calcium intake caused by the milk).
  3. On the other hand, over-restricting sodium in healthy people with an adequate calcium intake is also problematic because it can lead to an increase in stress hormones, most notably aldosterone, which could also impair bone health.
  4. Keeping salt in a normal range, by salting to personal taste, and balancing salt with adequate calcium, magnesium and potassium is probably the best option. Aldosterone has been shown to control salt intake, as high aldosterone levels correlate with higher salt pleasantness (13). There are no good reasons that we need to override this biological mechanism by overconsuming or over-restricting salt.

Neither coffee, milk or vitamin C affects heme-iron absorption from meats


Peatarians frequently recommend drinking coffee and milk with a meal of meat to inhibit iron absorption. It is also recommended to not combine vitamin C rich juices with meats as this increases iron absorption. 


The majority (70-90%) of iron in red meat is heme-iron (actually it is red because of it) (0). Heme-iron is a ferrous iron ion (Fe 2+) that's chelated in a large porphyrin ring. This form of iron is very efficiently absorbed by the body, more than non-heme iron (1). Unfortunately, uptake of this form of iron is also quite poorly regulated by the body. While there is a very efficient regulating mechanism for non-heme iron, that enhances iron uptake in deficiency states, while decreasing uptake in sufficient states, heme-iron absorption is only very marginally reduced in a sufficient iron state (2). Looking on a single meal, the maximum heme-iron absorption is saturable (3). So it might be wise to eat meat only on one occasion, instead of splitting it up throughout the day.

This difference might explain why higher consumption of (heme-iron rich) red meats (i.e. beef), but not (lower heme-iron) white meats (i.e. chicken), is associated with a higher risk for several diseases, such as type-2 diabetes, cancer and cardiovascular disease (4).

Given these data it would be great to have something that inhibits heme-iron intake. But there are bad news: Heme-iron absorption has been found to be highly resistant to dietary absorption inhibitors.
  • I couldn't find a single clear-cut study that showed that coffee inhibits heme-iron absorption. All studies I found only looked at non-heme iron. In one study they even added it to hamburgers but didn't measured heme-iron absorption (duh) (5). A new york times article based on an older study said that coffee did not affect heme-iron absorption (6). Unfortunately, I don't have access to the fulltext of that study (7). It is however plausible that it doesn't affect it, since heme-iron is absorbed differently than iron-salts. Experts in the field have a similar view, stating:
"Heme iron is relatively well absorbed under all circumstances. Moreover, heme iron absorption is relatively independent of meal composition and little affected by the enhancers and inhibitors that alter nonheme absorption." (8)
  • Calcium doesn't inhibit non-heme or heme-iron absorption in doses below 800-1000mg calcium per meal (9). So for milk to have an effect, you would have to drink roughly 800-900ml of it along with your meat. At these doses, calcium also hampers with zinc absorption (10).
  • Phytates inhibited non-heme-iron, but didn't affect heme-iron absorption (in contrast even seemed to slightly increase it) (11).
  • Meat itself appears to increase the absorption of heme-iron, compared to pure heme-iron (i.e. blood) (12).
Here are some studies that show that iron inhibiting/enhancing strategies have no or only very marginal effects on iron status in the long term. Maybe because the body itself is pretty good in increasing iron absorption when it is deprived of it and vice versa.
This all sounds depressing but there's also good news: 
  • In contrast to its effect on non-heme iron, Vitamin C doesn't increase the absorption of heme-iron (11) .


  • Coffee and milk at the usual dose don't have significant effects on heme-iron absorption. So don't expect both to protect you from iron toxicity of too much red meat. The only effective way is to limit heme-iron rich foods (i.e. by switching from red to white meat, or looking for other protein sources). At last, both coffee and milk have been found to hamper with zinc absorption. And red meat is a great source of zinc you don't want to miss out on.
  • It may still make sense to drink coffee/milk with meat when you know that your iron stores are high, as it could chelate the non-heme iron portion of it. For the average healthy person, the body already has a pretty decent system to regulate its non-heme iron intake.
  • Fruit or juice with meats will not enhance heme-iron absorption. So it can be combined occasionally.
At last, Ray Peat on this issue: 
  • "I think that's true, that coffee affects mainly non-heme iron absorption. The heme has toxic effects, forming carbon monoxide, apart from the iron." - Ray Peat (E-Mail advice)

Starch is not more fattening than sugar


Ray Peat repeatedly claims that starch is more fattening than sugar. He bases this argument on differences in insulin/glucose dynamics and animal feeding studies.

Eating “complex carbohydrates,” rather than sugars, is a reasonable way to promote obesity. Eating starch, by increasing insulin and lowering the blood sugar, stimulates the appetite, causing a person to eat more, so the effect on fat production becomes much larger than when equal amounts of sugar and starch are eaten.” -Ray Peat in "Glycemia, starch, and sugar in context"
"Fructose inhibits the stimulation of insulin by glucose, so this means that eating ordinary sugar, sucrose (a disaccharide, consisting of glucose and fructose), in place of starch, will reduce the tendency to store fat." -Ray Peat in "Glycemia, starch, and sugar in context"
“Starch is less harmful when eaten with saturated fat, but it’s still more fattening than sugars.” -Ray Peat (Quote found on FPS)
“When starch is well cooked, and eaten with some fat and the essential nutrients, it’s safe, except that it’s more likely than sugar to produce fat, and isn’t as effective for mineral balance.” -Ray Peat (E-Mail advice)
Paradoxically, the vast majority of people who switch from a high-starch diet to a low-starch high-sucrose peatarian diet gain weight. What does the experimental human evidence say?

Relevant studies

  • In a 14-day crossover study, 20 normal-weight healthy women ate a diet at their own pleasure (ad libitium), that was either rich in fat, starch or sucrose while keeping protein intake stable. After 14 days, the women lost weight on the starch diet (mean -0.7kg), while weight on the sucrose diet remained stable (mean +0.2kg). The difference in weight was likely mediated by differences in caloric intake, with women on the ad libitum high-sugar diet eating a higher amount of calories compared to the starch diet.
  • In a 6 week randomized crossover trial, 13 healthy males consumed either a higher-starch or a higher-sucrose diet while keeping total carbohydrates and energy intake stable. After 6 weeks, there was no significant difference in weight. Total and LDL cholesterol was higher in the higher-sucrose diet compared to the starch diet. Fasting glucose was unchanged but fasting insulin tended to be higher in the sucrose diet.
  • In a well designed 6-month randomized trial, 398 obese men and women ate ad libitum either a low-fat high-simple carbohydrate, a low-fat high complex-carbohydrate or a control diet. After 6 months, people in both low-fat diets lost weight compared to the control diet. The starch diet tended to be slightly superior (mean -1.8kg) than the sugar diet (mean -0.9kg) but the result was not significant and people seemed to eat more on the sugar diet. There was also a trend of an increase in insulin levels on the sucrose diet, while on the starch diet there was a trend of decreased levels compared to baseline. Fasting glucose tended to decrease more on the starch than on the sugar diet.
  • A meta-analysis of trials that tested isoenergetic exchanges of free sugars with other carbohydrates found no significant difference in body weight (a minimal trend of higher weight was seen in the higher sugar groups).


Ray Peats view of starch as being inherently more fattening than sugar is not in line with the experimental human evidence or the reports by peatarians. If you like trends then starch even seems to beat sugar in terms of weight loss. But if you really look objectively, then the only conclusion is that there is no difference whether you consume starches or sugar and the main determinant of body weight is total caloric intake. Peat's reasoning on starch's effects on insulin and hunger is also not in line with the human evidence: Starch was neither more insulinogenic nor did it lead to higher food-intake. The studies even suggest the opposite, that sugar is slightly more insulinogenic and leads to higher subsequent food intake than starches. 

There are 3 important points that are rarely mentioned by Ray Peat but play a big role when comparing starches and sugar:

  1. Palatability and reward: There is good evidence that both palatability and reward are big determinants of total caloric intake. Easy palatable and rewarding foods, such as refined or sugared starch or non-starch products (i.e. cake, ice cream) increase overall caloric intake, contributing to weight gain, while shifting your diet to less palatable and rewarding food will promote weight loss. For more details I recommend the insightful review by Stephan Guyenet (1, 2).
  2. Liquid vs. solid foods: Likewise, there is good data that liquid foods are less satiating and suppress subsequent caloric intake less than solid foods (3). This is especially true for sugar-sweetened beverages which have been repeatedly in both observational and interventional studies shown to lead to weight gain and promote obesity (4). One explanation is that due to the rapid absorption of liquids in the gut, they cause less gastric stretch. Gastric stretch is an important trigger of satiety by activating stretch receptors (5).
  3. Type of starch: When people report they gain weight on starches, they should critically analyze their overall calorie intake and the type of starches they consume. Starches are often eaten together with fat or sugar (i.e. in cakes, cookies, pizza etc.) which make it easier to consume a larger amount of calories for the same satiety signal. This also makes comparison with fruits, with their inherently lower calorie content unfair, which are often not sufficient to satiate people alone. Sticking with more basic starches and combining them with less fat can already make a big difference.
At last, individual differences should not be discarded. It could be that some people are just born to be efficient starch-eaters while others run better on sugar-rich foods. For instance, the gene copy number of salivary amylase (which cuts down starch into sugar in the mouth) varies widely between people and could determine starch digestibility (6). People with a low copy number of salivary amylase were found to have a higher BMI, suggesting that efficient starch digestion leads to a quicker satiety feeling and thereby less overconsumption (7).  Differences in gut microbiome composition is another factor possibly involved in starch tolerability (8).

Yogurt does not deplete the liver's energy stores


"A spoonful or two of acidic yogurt isn't harmful, but a cupful of the
acidic type can be enough to deplete the liver's energy stores
because lactic acid is converted to glucose in the liver, requiring
energy. The "strained" type that isn't acidic is similar to cottage
cheese and is safe." – Ray Peat (Email advice)


For safety, I calculated the required energy to metabolize the lactic acid of 2 cups of yogurt (500g).
According to this source, yogurt contains 1.08% lactic acid.
500g yogurt would then contain 5.4g lactic acid.
5.4g lactic acid equals 5.34g lactate + 0.06 hydrogen (dissociates in gut).
Lactate is converted to glucose in the liver's Cori cycle.
6 mol of ATP is required to convert 2 mol of lactate into 1 mol of glucose.
5.34g lactate ~ 0,06 mol lactate
0,06 mol lactate requires 0.18 mol ATP to convert it into glucose.
To produce 0.18 mol ATP, 0.005 mol glucose is used up (1 mol glucose produces 36 mol ATP).
0.005 mol glucose = 0.9008g Glucose.


To completely metabolize the lactic acid of 500g yogurt, you only need 0.9g of glucose. Given that a human liver contains around 100-200g of stored glycogen and 500g of yogurt contains 23.5g lactose itself (Glucose+Galactose), this amount is totally negligible and does not deplete the liver's energy stores in a relevant way.

Neither milk, nor calcium increases the metabolism


One thing that you'll hear often in response, when people report that dairy caused acne, that "milk/calcium increases the metabolism and hence the need for nutrients such as vitamin A" (i.e. here). This motivates people to supplement with high amounts of vitamin A to antagonize the acne-forming properties of milk.

"The milk estrogen research isn't good. It also contains thyroid and progesterone and other protective substances. The high calcium content helps to increase the metabolic rate, and probably contributes to maintaining the anabolic balance." - Ray Peat

"Regarding milk and its tryptophan content, The calcium helps to keep the metabolic rate high, and the other nutrients help to steer tryptophan away from the serotonin path." - Ray Peat (E-Mail exchange)


I looked up studies on that subject and there are quite a few published. All I have found indicate that neither milk nor calcium increases one's metabolism. There are some other interesting findings.

  • In a 1-week trial with healthy participants (mostly male), an energy-balanced high-calcium diet did not increase the metabolic rate compared to a low-calcium diet. However, under energy deficient conditions, the high calcium diet lead to a higher fat oxidation.
  • In a 1-week trial with healthy normal-weight men, neither a high calcium carbonate or a high dairy calcium diet increased the metabolic rate. A non-significantly increased fat oxidation was found.
  • In a 3-week controlled trial with overweight children neither calcium carbonate nor milk products increased total energy expenditure.
  • In a 5-week controlled trial with overweight low-calcium consumers, an 800mg dairy calcium supplementation did not increase resting energy expenditure. However, a non-significant trend towards less carbohydrate oxidation and more fat oxidation was found
  • A 12-week RCT with overweight women, comparing control vs. 900mg calcium-carbonate vs. high dairy found no increase in total energy expenditure in either intervention group. In the calcium supplement group an increase in fat oxidation was found.
  • In a 9 month RCT with obese subjects, a recommended dairy diet did not lead to an increased resting metabolic rate compared to a reduced dairy diet.
  • In a 1-year controlled trial with healthy women, a chronic high-calcium intake increases fat oxidation to both low and high calcium meals compared to a chronic low-calcium intake. An increase in PTH during the trial correlated with a lower fat oxidation and vice versa.
  • In a meta-analysis of trials, both chronic and acute calcium intake was found to increase fat oxidation
  • In a randomized trial of adolescent girls with low dairy intake, girls randomized to increase their calcium intake did not loose more weight than girls on the continued low-calcium diet
  • In a 24-week trial with obese subjects on an energy-deficit diet, people randomized to 800mg calcium carbonate/day or a high-dairy intake lost more weight than people on an average calcium intake. Dairy intake lead to greater weight loss than supplemental calcium intake.
    Similar results have been found in a similar trial under energy-deficit conditions.
  • In a 1-week trial of healthy moderately overweight people (mostly female), a 1800mg calcium diet did not increase 24-hour energy expenditure compared to a 500mg calcium diet. A non-significantly increased fat oxidation and a decreased carbohydrate oxidation was found. However in the group with a high-calcium and normal-protein intake (%15 E of protein), a higher fecal fat and energy excretion was found. In the discussion, the authors said that calcium can form complexes with fat that hinder them from being absorbed, but on a high-protein diet, the calcium binds to the protein and not the fat, leading to a reduction in the fat excreting effect of calcium.
    These results have also been found in a meta-analysis of calcium trials.
    And in line with that study another study found that a high calcium meal reduces the lipid content in the blood compared to a low calcium meal, presumably by reducing fat absorption
  • In an observational study, high acute calcium intake was correlated with higher fat oxidation, which correlated with a lower respiratory quotient. Dairy intake was not a more important predictor of fat oxidation than total calcium intake.


  • Milk or calcium does not increase the metabolic rate according to the published evidence. The beneficial effects of dairy consumption on body weight appear to be partly mediated by a shift away from carbohydrate towards fat oxidation and a higher fecal excretion of fat. Milk products only favor weight loss in the context of a calorie-reduced diet, but in that context they are superior to low-calcium non-dairy diets.
  • In that sense, the acne-increasing properties of milk seem to be unrelated to a general metabolism increasing effect. Possibly, other mechanisms appear to be at play such as hormonal effects, gut irritation of the hard to digest casein/whey proteins or allergic/histamine/inflammatory reactions. The beneficial effect of vitamin A on acne may be unrelated to the metabolic rate and simply be a consequence of its anti-keratinization properties, leading to less clogging of sebum within the pores.

Cancer also occurs on PUFA-free diets


In several of his articles, Ray Peat suggests that cancer can't occur on PUFA-free diets. As proof, he repeatedly cites an old German publication by Bernstein and Elias from 1927 (source).
"Cancer can't occur, unless there are unsaturated oils in the diet." - Ray Peat in "Unsaturated Vegetable Oils: Toxic"
"In 1927, it was observed that a diet lacking fats prevented the development of spontaneous tumors." - Ray Peat in "Oils in Context"
"When I was studying the age pigment, lipofuscin, and its formation from polyunsaturated fatty acids, I saw the 1927 study in which a fat free diet practically eliminated the development of spontaneous cancers in rats (Bernstein and Elias)." - Ray Peat in "Unsaturated fatty acids: Nutritionally essential, or toxic?"
"In 1927, Bernstein and Elias found that rats eating a fat free diet had almost no spontaneous cancer [...]" - Ray Peat in"Cancer: Disorder and Energy"

Review of Bernstein and Elias study

I went through this study and found that it doesn't show what Peat cites it to show:
  • The study was done in mice, not in rats. There were several experiments with mice (2 per box) that got either a normal diet or a specially prepared lipidfree diet. Some mice got a lipidfree diet supplemented with cholesterol or lecithin.
  • It didn't study the effect of diet on spontaneous cancer but on cancer that has been injected into the mice.
  • In experiment B, tumors were found in 9/9 mice with the normal diet (mean tumor weight 0.96g) and in 5/7 mice with the lipidfree diet (mean tumor weight 0.4g). In percentage, tumor weight was roughly 60% smaller and tumor incidence was 30% less on the lipidfree compared to the normal diet. Mice on the lipidfree diet lost 18% of weight and gained 1% of weight on the normal diet.
  • In experiment R, tumors were found in 9/9 mice on the normal diet (mean tumor weight 0.18g) and in 7/9 mice in the lipidfree diet (mean tumor weight 0.1g). In percentage, tumor weight was roughly 60% lower and tumor incidence was 25% lower on the lipidfree diet. Mice on the lipidfree diet lost 19% of body weight compared to a 2.3% loss on the normal diet.
  • Interestingly, in experiment R mice that got a lipid-free diet supplemented with cholesterol had a higher tumor weight than both normal and lipid-free eaters (5/5 mice with tumors, mean tumor weight 0.74g). In experiment B, the cancer-increasing effect of cholesterol was less pronounced (10/15 mice with tumors, 0.6g mean weight). Any conclusions from that are limited as the level of cholesterol in the mice diet was very high (approx. 10%).
  • In the summary of all experiments (205 mice total), mean tumor weight on the normal diet was 6.1g and on the lipidfree diet it was 2g (~65% reduction). Unfortunately, they didnt report absolute tumor numbers in this summary. On average mice lost 13% body weight on the lipidfree diet and gained 1.3% on the normal diet.
  • At last, mice on the lipid-free diet (even without getting injected with cancer) had a increased mortality and died already after around 3 weeks. On the next morning, those mice were found without brains, as it was eaten by their "lipidhungry comrade".


Even though the details might have been lost in translation, Ray Peat misquotes the main findings of this study. In addition to the increased total mortality, mice on fat-free diets still grow cancer, so the claim that PUFA-free diets result in ("pratical"/"almost") freeness from cancer can't be uphold based solely on that reference. The more accurate description would have been: "Mice fed a lipid-free diet grow less injected tumors, but die much more often."