I do think there's some therapeutic value in there, somewhere. And she's right that the mainstream trial of low fat diets involves an awful lot of getting people to eat 30 percent instead of 35 percent, as if you could expect any sort of clinically significant therapeutic effect from that.

The low protein part of it is something that really needs to be tested, though. It's hard to add whole food proteins to a diet like that without changing all kinds of other stuff. What if instead of sugar, a bit of whey protein were added?

Another aspect of this diet--the fat level is so low, it verges on an essential fatty acid deficiency experiment. Or at least an essential fatty acid depletion one. In mice at least, if you get linoleic acid down below 1 percent, all sorts of diets that would otherwise be obesogenic/diabetogenic aren't. In a sane human diet, getting linoleic that low might not work, because we'd get some arachidonic acid from various meats. But in Kempner's crazy-ass diet, that's not an issue.

There are also studies where mice fed leucine-deficient diets rapidly lose their fat mass. A diet of rice and sugar sort of qualifies for leucine-deficient as well.