Thoughts on the Kempner Rice Diet

By Jason Fung, MD

Recently, Denise Minger wrote a thought provoking post “In Defense of Low Fat“. Denise is the author of a book called “Death by Food Pyramid” and came to fame writing a very thorough analysis of the China Study, which many people, (myself included) thought was far, far superior to the original analysis by the author T. Colin Campbell. I admire her open-mindedness and she has clearly put much work and thought into her latest post, too. Her humor also makes her writing quite special, and I encourage you to read it, even though it may be the longest blog post ever written. I quite agree with her main point that we cannot simply ignore evidence that contradicts us, but need to look closer. However, I do not agree with all her conclusions in her latest post, and the numerous comments made have not really addressed the key deficits of her arguments.

Denise describes the Rice Diet used by Kempner which consisted of rice, sugar and fruits and juice. This is extremely low in fat (<10%) as opposed to 20-30% of typical ‘low fat’ diets. Denise’s hypothesis is that at extremely low fat states, something ‘magic’ happens. Kempner was a clinician at Duke University in the 1930s who treated over 18,000 patients with this diet of refined carbs. And Denise states that it worked ‘ridiculously well’. There were 3 main groups of patients treated, as far as I can see – kidney disease, malignant hypertension and obesity/T2D.

So, this is taken as evidence that this extremely low fat diet works because it is extremely low in fat. That’s not accurate. We can say that the Rice Diet works, but we cannot infer that it works because it is low in fat. That is like saying that both antibiotics Amoxil and Azithromycin work for pneumonia because both start with the letter ‘A’. No, we can only say that it does work, but not why.

Further, we need to consider each disease separately. Why would low fat work for kidney disease at all? Why would the same mechanism cure kidney disease and high blood pressure and obesity all at the same time. Might not there be separate mechanisms? Let’s take a look.

Kidney Disease

Now, I’m a nephrologist (kidney specialist), so I’m not going to accept some hand-waving argument that low fat diets improve kidney disease. There are hundreds of different kinds of kidney disease. What kidney disease was Kempner treating? It was not diabetes or hypertension, but primary kidney disease or glomerulonephritis. Statistically speaking, we can assume that the majority of this was IgA nephritis – the most common type of inflammation in the kidney. Why would a low fat diet work here? There is no reason to think that it would.

However, a percentage of IgA nephritis is caused by dietary lectins (carbohydrate binding proteins) and reactions to other food antigens. Primary amongst these are gliaden and bovine serum albumin. Gliadin you probably know as the problem in gluten and celiac disease. There is a well described association between IgA nephritis and celiac disease. In other words, celiac disease and IgA nephritis often co-exist and are caused by a reaction to the dietary lectin, gliadin. The other major dietary problem in IgA nephritis is postulated to be bovine serum albumin, which you may have guessed is found in dairy products and cow’s milk.

In fact, there have been studies on and off using a so-called ‘elimination diet’ whereby all foods are eliminated with the gradual reintroduction of foods that are low in antigens (avoiding mostly wheat and dairy). In this study 11 of 12 patients improved. Not because it was low in fat (because it wasn’t), but because they removed dietary lectins and other antigens. The Kempner diet, with only rice and fruit and sugar was similarly a low antigen diet. So this may well have worked for IgA nephritis, but not because it was low in fat, but because it eliminated wheat (gliadin) and dairy (bovine serum albumin).

So, yes, the Kempner Rice Diet may have worked well in kidney disease – as shown in other elimination diet studies of this disease as well. But the mechanism is not the low fat.

Malignant Hypertension

The second major area that Kempner describes success is in the treatment of malignant hypertension. This is high blood pressure on Mark McGwire dose steroids. Remember that this was the 50s where there was just about no treatment for high blood pressure other than lying down in a dark room and thinking relaxing thoughts. Most of the drugs that we currently use for lowering blood pressure simply did not exist at that time. So high blood pressure was a really big problem, because there was essentially no treatment.

But the Rice Diet worked. But why? One of the things we have known for at least a half a century is that sodium restriction can lower blood pressure. To be sure, I don’t think there is any benefit to a low sodium diet in the person without high blood pressure. But that’s not the situation here. Here we have super super high blood pressure without any drugs to treat it. In the land of the blind, the one-eyed man is king. The Rice Diet has virtually zero sodium! If you kept on it, there’s a very good chance your blood pressure will go down – maybe even enough to save your life.

It doesn’t seem plausible to me to say that the Rice Diet worked because it was low in fat. It is much more likely that it worked because it was an extreme salt restriction diet. You could have also eaten an all olive oil diet and done well. Same goes for the claim that heart failure is treated. Yes, because it is a zero sodium diet.


Here’s the big one. The Rice Diet was CARBOCALYPSE, as Denise describes with all refined carbs, yet people lost weight and reversed their type 2 diabetes. This shoots an arrow directly into the eye of the Low Carb Crazies, right?  Well, not quite.

So here’s the diet:

Image from Duke University

A patient takes an average of 250 to 350 gm. of rice (dry weight) daily; any kind of rice may be used provided no sodium, chloride, milk, etc. has been added during its processing. … All fruit juices and fruits are allowed, with the exception of nuts, dates, avocados and any dried or canned fruit or fruit derivatives to which substances other than white sugar have been added. Not more than one banana a day should be taken. White sugar and dextrose may be used ad libitum; on an average a patient takes about 100 grams daily, but, if necessary, as much as 500 grams daily should be used. Tomato and vegetable juices are not allowed.

Wow. You can eat 2400 calories of refined sugar and lose weight! Dr. Atkins – take that! A graphical comparison of equal calorie diets looks like the above. But is this a fair comparison? Do both sides eat 2400 calories per day?

Well, what was the average intake of rice? 250 gm. That is a bit more than 1 cup, or 850 calories or so. On average, they also take 100 grams of sugar. That would be 387 calories. We don’t know about fruit, but this was the 1950s before white peaches, before golden pineapples, and well before jumbo California strawberries are available year round. So we’ll guess that an average person on the Rice Diet would be taking a total about 1400 calories a day, virtually all carbohydrates.

How does this compare to the standard diet? Well, an average person would take 2400 calories, for example, and we will estimate 60% carbohydrates. That make it about 1400 calories of carbohydrates. This Rice Diet was not Carbocalypse at all! It was simply the same number of carbs but without any of the protein or fat. Protein also stimulates insulin, and it may also contribute to weight gain. So what we are comparing is 1400 calories of carbohydrates alone, or 1400 calories of carbs with another 1000 calories of fat/ protein. No surprise here, the 1400 calories of carbs alone wins.

The truth, of course, is that the Rice Diet is only one more example of the unrealistic monotonous diets that occasionally gain faddish supporters. The potatoe diet, optifast, slimfast, the grapefruit diet etc. If you only eat one thing, it quickly becomes monotonous and all pleasure is lost in eating. So you will eat only if absolutely necessary to avoid starvation because, well, it’s gross. Kempner would, on occasion, whip his patients to help compliance. The problem is that once you start eating normally, all the weight comes back.

I could make a diet called the Fung PigBlood Diet. Eat as much as you like! Eat day or night! Eat anytime, anywhere! First three whippings for no extra charge! As long as you eat only Pig’s Blood, I guarantee you’ll lose weight. Sure – you’ll do fine, because you won’t eat most of the time. Why? Cause it’s super-gross. Oh, wait, does that mean you’ll be fasting a lot? Sure will. I’ll start booking a ticket for the Nobel ceremonies next year. Obesity – cured by pig’s blood!

This is also, of course, why medical low-fat supporters such as Drs. Fuhrman, MacDougall and Barnard all do not allow unlimited sugar and rice. Because it doesn’t work in a mixed diet! These fellows aren’t stupid, you know. Most of these low-fat advocates also avoiding refined grains and sugars, just like the Low Carb and Paleo groups. Can’t we all just get along?

Same goes for all the data on lowered usage of insulin on the Rice Diet. Yes, because insulin requirements depend on both carbohydrates and protein. That fits perfectly with our understanding of how things work. Eat same refined carbohydrates but zero protein – Need less insulin. Less insulin – less weight gain.

Roy Swank and Multiple Sclerosis

After finishing with the Rice Diet, Denise brings up Roy Swank’s low-fat diet and MS. This is bizarre for 3 main reasons. The man collected 150 patients, put them on a low fat diet and described how many of these patients did well. That’s great for them. But that’s not great science.

First, there’s no comparison group. How many of those patients would have done well without the low fat diet? This is also seen in the placebo effect, the no-cebo effect and why, until recently people were still eating mummies as medicine. The same goes for leaching, blood letting, exorcisms, and other such superstitious practices (tiger penises, bear claws).  You need to compare one group to the other. Just as the Terry Wahls story is great, it is also not science and cannot be used to prove anything. Placebo controlled studies, for example, have shown placebo to have almost as much effect on subjective symptoms of asthma as real drug. Sham surgeries have the same effect as real surgery on meniscectomy for knee pain.

In fact, Benjamin Franklin, that famous fasting devotee, devised the first placebo controlled medical trial in order to debunk the myth of mesmerism or ‘animal magnetism’ whose proponents also claimed mystical healing powers. So, gathering a bunch of MS patients who did well is great, but it’s not great science and certainly proves nothing about low fat diets.

The second major point is that science means that this is a reproducible result. Has anybody else been able to replicate these findings? If not, then we generally would not accept them. If only one physician or researcher is able to show results, then those results are generally not accepted. The reason why?

Let’s review the despicable story of one Mr. Andy Wakefield, a former surgeon and medical researcher from Bath, England. He.wrote a 1998 paper published in the Lancet that alleged that there was a connection between MMR vaccines and autism. Brian Deere, a journalist dug into this story further and found a much deeper, more bile inducing story. Wakefield essentially made up the entire paper, because he had a vested interest ($$$) in proving this point.  This led to celebrities such as Jenny McCarthy taking up the no-vaccine cause and directly leading to several children’s deaths due to outbreaks of measles. Yes, kids died because the despicable Wakefield lied.

How was this found out? Well, this was in part because no other group could replicate the findings of the original paper. This is highly suspicious, and led to further investigations about the original paper which has since been retracted as well as its author.

Third, we mostly talk about diets in terms of affecting obesity and/or T2D. Whether it works or not for MS is mostly irrelevant to this discussion. Suppose I discover that a 100% fat diet is effective in a rare disorder affecting 1 person in the last 200 years worldwide. That’s great, but what relevance does it have to obesity of type 2 diabetes? Not very much. So maybe low fat diets are good for MS and maybe not. So what?

I’m going to end here with one more thought. I am always looking for studies that disprove my point of view of obesity/ T2D. That is the only way we learn. That is one of the reasons I read Denise Minger’s blog post with such interest. I don’t agree with its conclusions, but that’s OK. The point is to have open discussion and try to find our way to what matters in nutrition. For that, Denise, I would say thank you.

PS I will address the Pritikin Diet in another post, before this post threatens to be the world’s longest.

Update Oct 26, 2015

Several readers have contested the calorie count that I posted using different assumptions for amount of rice taken and fruit juices. I agree that it is certainly possible to get 2400 calories on the Rice Diet. However, the question is this. How many calories of carbohydrates were actually eaten for weight loss?

It appears that some of this depends upon why the Rice Diet was being used. For kidney disease and hypertension, calorie counts appear to average about 2200 calories/day. However, these were not patients that need weight loss.

In trying to figure out how many calories the weight loss people ate, a reader pointed out an interview with Dr. Don Rosati (Rice Diet Program Director) who picked up after Walter Kempner. For weight loss, it appears that unlimited rice and sugar was NOT the program, but instead was a severely calorie restricted diet of 800-1000 calories.

A patient, Jean Anspaugh wrote a book “Fat like us” in which she describes the Rice Diet as “spend(ing) all that money eating rice and fruit and starving to death”. Notice she did not say “I ate my fill of rice and fruit. So, perhaps the weight loss Rice Diet was not exactly the same as the kidney disease Rice Diet or the Malignant Hypertension Rice Diet.

Third, a piece written in JAMA 1974 by Dr. Eugene Stead, who became physician in chief at Duke University also mentions the Rice Diet as a great treatment of edema and hypertension. Notice there is no mention of obesity here.  However, he notes that “The rice diet was low in calories, protein and sodium chloride” (emphasis mine). A further interesting point was that many people who subsisted on this diet (and doing well) were eating less than 20g/day of protein – far below the recommended daily amount necessary and far, far, far below current intakes. I note this because I get a lot of grief about people who are worried about protein deficiency during fasting which is far overrated.

I realize that I have not reviewed all the studies, since, well, many of these papers are over 60 years old and I just can’t get them. However, there does seem to be a substantial amount written that indicated that the weight loss Rice Diet was not in fact, a 2400 calorie diet.

So, again, to compare the rice diet (800-1000 calories of refined carbs) to regular diet (2400 calories total with approx 1400 calories of refined carbs) you can see that there really is no mystery why the Rice Diet worked and why it died out.

By The Fasting Method

For many health reasons, losing weight is important. It can improve your blood sugars, blood pressure and metabolic health, lowering your risk of heart disease, stroke and cancer. But it’s not easy. That’s where we can help.

Jason Fung, MD

By Jason Fung, MD

Jason Fung, M.D., is a Toronto-based nephrologist (kidney specialist) and a world leading expert in intermittent fasting and low-carb diets.

Share this article with a friend

More articles you might enjoy…