Each year, US News & World Report publishes diet rankings based on the opinions of a panel of experts. This year, I served on the panel. In this post, I’ll discuss my general thoughts on the rankings, as well as the fact that the popular ketogenic diet tied for last place among the 40 diets.
Here’s how it works. US News staff summarizes information on each diet– including the diet’s recommendations and what current evidence says about its effectiveness– and provides it to a panel of expert reviewers. This is the same information that is publicly available on each diet’s US News webpage. It’s not perfect but they appear to put in a good faith effort, and they ask reviewers to recommend corrections if anything seems off. On the basis of this information, as well as the reviewers’ existing knowledge and optional literature searches, reviewers are asked to rate the diets in a number of different ways on a scale of 1-5. These scores are then compiled into four areas that are shared with the public:
- How effective is it for short-term weight loss?
- How effective is it for long-term weight loss?
- How easy is it to follow?
- How well does it support general health?
Note that these “diets” also often include lifestyle modification advice, such as physical activity, and these are considered in the scoring. US News also provides an overall score for each diet, which I suspect is an average of the four scores above.
The diet ranking system is necessarily an imperfect process. With 40 diets to review and rank, there isn’t enough time to do comprehensive literature reviews to get the best possible picture of effectiveness for each diet. This less-than-comprehensive approach leaves a certain amount of room for uninformed opinions to substitute for evidence, particularly in areas of evidence that reviewers may not be very familiar with. Still, the information it provides should be more helpful for selecting diets than the current alternatives, like reading Amazon reviews, Googling, or asking your aunt Martha.
The review panel includes people with a diversity of backgrounds, and they don’t agree on everything. This is probably by design. I often agree with the opinions of the rest of the panel, but there are also a number of places where I disagree. I’ll discuss a few of those below. I want to emphasize that the point of the exercise is not for everyone to agree.
For the first time, the Mediterranean diet secured the top spot for best diet overall, tied with the DASH diet. I think this is appropriate. The science of nutrition and chronic disease is complex and uncertain, but in my opinion the Mediterranean diet has more convincing evidence supporting its efficacy for general health than any other diet. It’s also relatively easy to follow, which supports long-term compliance.
However, it’s not really a weight loss diet and there is very little evidence that it causes weight loss as typically applied. Hence, I’m confused about why other reviewers rated it as an effective weight loss diet, especially in the long run. I suspect this has to do with two factors. First, a well-known randomized trial by Shai and colleagues found that a calorie-restricted Mediterranean diet caused weight loss that was similar to a low-carbohydrate diet, and superior to a moderate low-fat diet, over a two-year period (1). Four years after the end of the study (6 years total), people who had been assigned to the Mediterranean diet were maintaining their weight loss better than people who had been assigned the other two diets (2). Second, human nature. When we like something in general, that favorable opinion tends to color our view of more granular aspects of it, even if it isn’t deserved.
I think that if you adopt a standard Mediterranean eating pattern that doesn’t involve calorie restriction, you may lose some weight depending on what diet you’re coming from, but probably not much. I do think the diet shows promise for the long-term maintenance of lost weight, as suggested by the Shai trial.
This issue came up in a few other places as well. For example, the vegetarian diet was ranked fairly well for weight loss. This view seems to be based on observational studies showing that vegetarians tend to be leaner than their meat-eating counterparts. Yet there is little direct evidence that vegetarian diets promote meaningful weight loss, and if you expect to lose a substantial amount of weight just by cutting meat out of your diet, you’re probably going to be disappointed.
The ketogenic diet
The ketogenic diet is a very-low-carbohydrate, high-fat, moderate-protein eating style that forces the body to run almost entirely on fat for fuel. It’s extremely popular right now. Although the popularity of individual diets waxes and wanes, I tend to think that there is usually a nugget of truth underneath the hype.
The ketogenic diet limped in to last place in the rankings, tied with the Dukan diet (a high-protein, low-carb/fat rapid weight loss diet). It was rated by the panel as fairly effective for short-term weight loss, not very effective for long-term weight loss, very unhealthy, and very difficult to follow. Unsurprisingly given its popularity, the snub was not well received by ketogenic diet fans.
I don’t remember exactly how I rated it myself, but here’s the gist of it. I think very-low-carbohydrate diets like the ketogenic diet are some of the more effective diets for short-term weight loss– although no diet is a silver bullet (3, 4, 5, 6). Ketogenic diets also cause weight loss in animal models under tightly controlled conditions (7, 8, 9). I don’t know how effective it is for long-term weight loss and the evidence in humans doesn’t seem very informative to me, but I seriously doubt it’s less effective than diets like Mediterranean and vegetarian that aren’t even weight loss diets (10). In general, my observation is that the more extreme you are on the carbohydrate:fat continuum, the more a diet tends to promote appetite control and weight loss. Very low-carbohdyrate and very low-fat diets both tend to be more effective for weight loss than their more moderate counterparts, and I suspect this also applies in the long run, at least to some extent. This is supported by animal research, which mostly seems to suggest that ketogenic diets can curtail long-term body fat gain (11, 12, 13).
I probably gave it a 2 or a 3 out of 5 in health, assuming that someone is coming from a typical unhealthy diet. The primary reason for this mediocre score is simply that its long-term effects on human health aren’t well known. I can’t really say that ketogenic diets are unhealthy in the long run, but I also can’t say they’re healthy, and that uncertainty makes me feel wary. This concern is compounded by the fact that (to my knowledge) no traditionally-living human population has ever been shown to exist in a state of chronic ketosis, which the diet promotes. Even the arctic Inuit, who traditionally eat a high-fat, low-carbohdyrate diet, were not in chronic ketosis, and the reason is rather troubling for ketogenic dieters: A genetic mutation preventing ketone production spread rapidly throughout the ancestral Inuit population, suggesting that avoiding ketone production conferred a powerful survival advantage (14). Furthermore, few of the people reading this are descended from Inuit and their genomes weren’t shaped by thousands of years of eating the Inuit diet. The vast majority of traditional diets that have been documented contain enough carbohydrate and/or protein to prevent ketosis, although the amount of carbohydrate varies greatly (15). Long-term nutritional ketosis doesn’t appear to be part of the ancestral human metabolic spectrum. Contrary to popular belief, it is probably not Paleo or ancestral. Shouldn’t that worry us a little?
The animal research on the health impacts of the ketogenic diet is a mixed bag. The diet usually prevents weight gain and its effects on lifespan are either neutral or beneficial in mice, which is reassuring (16, 17). However, it tends to cause fat accumulation and inflammation/injury in the liver and other tissues (18, 19, 20, 21, 22). Its effects on insulin sensitivity in animal models are inconsistent between studies. Rodents don’t naturally develop human-like cardiovascular disease, so they can’t help us understand the diet’s long-term impacts on the heart and blood vessels.
That said, we do have multiple randomized trials of very-low-carbohydrate diets, and they have generally found little evidence of harm, and some evidence of benefit, in people who have obesity and/or diabetes in the short-to-medium term (23, 24, 25). About the worst we can say is that they don’t lower “bad” LDL cholesterol as much as other diets. These findings make it seem less likely that the long-term effects are catastrophic. It’s also hard to deny that some people appear to benefit from the diet– particularly people with blood glucose control problems such as diabetes. If the alternative is having poorly controlled blood glucose, using blood glucose-suppressing drugs, and injecting insulin, I can see how a very-low-carbohydrate diet would start to make sense. We still need studies on whether the diet can prevent the complications of diabetes such as heart attacks. For people who are overweight, have blood glucose control issues, and who react well to it, very-low carbohydrate diets seem like a viable option and they don’t appear to cause obvious harm. We still have uncertainty about the long-term effects of such a diet, but given the alternatives I don’t think it’s unreasonable to try it.
Let’s be honest though– most people can’t stick with it. I agree with the other reviewers on this. The more extreme a diet, the harder it is to adhere to, and the ketogenic diet is extreme. “But wait”, you say, “I’ve been on the ketogenic diet for five years and it’s easy!” That may be true, but randomized controlled trials don’t lie. The average person can’t even stick with the diet for six months, as judged by urinary ketone levels (26, 27). The minority of people who find it easy, get good results, and stick with it are the ones who write about it on the Internet.
For the record, I’m not against the ketogenic diet and I doubt it’s a health disaster for people who tolerate it well. I do think people who adopt it should be aware of the uncertainty about its long-term effects. As with any extreme diet, it has strengths and weaknesses and it’s probably best viewed as a medical intervention akin to drugs and surgery, rather than a tool for supporting general health. I think it’s interesting enough that it deserves further research, particularly in the setting of diabetes.