Cancer also occurs on PUFA-free diets

Background 

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".

Conclusions

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."