There’s been an interesting edit in Marc Bittman’s sugar post, as he has now changed his tune on the PLoS one sugar study, now Bittman acknowledges obesity too is important. That was big of him, it is after all, the most important factor. Maybe my angry letter to the editor had an effect, but he’s grudgingly changed this statement:

In other words, according to this study, obesity doesn’t cause diabetes: sugar does.

To:

In other words, according to this study, it’s not just obesity that can cause diabetes: sugar can cause it, too, irrespective of obesity. And obesity does not always lead to diabetes.

The second sentence is totally unnecessary. Of course obesity doesn’t always cause diabetes, or heart attack or whatever. Nor do cigarettes always cause lung cancer. Nor does sugar intake always lead to obesity or diabetes. But obesity is the primary cause of type two diabetes, just as cigarettes are the primary cause of lung cancer, and who knows what sugar is doing.

Mother Jones, sadly, has decided to double down, calling the PLoS One study the “Best. Diet. Study. Ever.” It’s not, of course. It’s merely interesting and suggestive of an effect. It is not nearly proof of causation. They also laud the Mediterranean diet study (maybe it was supposed to be the Best. Study. Ever.?), however, they again show they’re not actually reading these papers because if you read our coverage of the study you’d know they didn’t actually study the Mediterranean diet! In a case of the blind leading the blind, they quote Bittman’s misinformed piece on the Mediterranean diet study

Let’s cut to the chase: The diet that seems so valuable is our old friend the “Mediterranean” diet (not that many Mediterraneans actually eat this way). It’s as straightforward as it is un-American: low in red meat, low in sugar and hyperprocessed carbs, low in junk. High in just about everything else — healthful fat (especially olive oil), vegetables, fruits, legumes and what the people who designed the diet determined to be beneficial, or at least less-harmful, animal products; in this case fish, eggs and low-fat dairy.

This is real food, delicious food, mostly easy-to-make food. You can eat this way without guilt and be happy and healthy. Unless you’re committed to a diet big on junk and red meat, or you don’t like to cook, there is little downside

Except for one critical fact. The subjects assigned to the Mediterranean diet did not have lower consumption of red meat, sugar and hyperprocessed carbs, or other junk! If you look at the supplementary data, you see that the subjects took the positive recommendations of the diet (olive oil, nuts, fish), and more or less ignored the negative recommendations (less meat, less spreadable fats/butter, less baked goods). If you look at figures like supplementary S6, the study groups did not change their diets in these categories relative to the controls, so the effects on their cardiovascular events relative to controls aren’t likely to be from the diet recommendations. When there were changes relative to baseline, even when statistically significant, the changes were tiny.

The participants in this study actually had a very high fat intake, about 35-40% of calories across all groups. And while there was a statistically-significant decrease in cardiovascular events like stroke and heart attack in both study groups (Med + olive oil, Med + nuts), only one arm of the so-called Mediterranean diet (Med + Olive oil) had a non-significant decrease in mortality, while the other arm (Med + Nuts) had a similar curve compared to the “do nothing” control. My interpretation of this, and it’s fine to be critical of it, is that this isn’t that meaningful. If anything, the only variable correlating with decrease in mortality was excess olive oil consumption (> 4 tbsp/day), not the Mediterranean diet. Either that, or eating nuts cancels out the beneficial effects of the diet on mortality.

This is why people always dump on nutrition science when it appears to change every 10 years. Results get overblown, and when the inevitable regression towards the mean occurs, we get blamed for it. The reality is, the press coverage of science is extremely poor, and there is not adequate critical analysis and presentation of results to their audience.

Comments

  1. #1 ralph
    March 2, 2013

    Instead of doom and gloom some hope on the horizon that people do not know about.

    There is a new Obesity pill called Belviq which is the first novel diet drug approved by the FDA in 13 years. It is due to hit the pharmacies in early March of this year. It has been proven safe with virtually no side effects and very effective for over 50% of those who take it and very effective on people with diabetic issues.

    In the 8000 people trials the top 40% of responders who stayed on the pill for 1 year they lost over 10% of their body weight. Twenty pounds for a 200 pound woman is excellent. 25% of the people who stayed on for 1 year lost over 15% of their body weight.

    In addition there was a significant reduction of HbA1c levels of -0.9 and fast glucose levels of -27. These effects were seen for patients even if the weight loss was much less than the top 40%. This new FDA approved pill promises to have a major impact of the diet market and the diabetic market.

  2. #2 Essante organics
    March 3, 2013

    Great post. I just located your blog and wished to let you know that I have certainly loved reading your blogs. At any rate I’m going to be subscribing to your feed and I really hope you are writing again soon. Essante organics

  3. #3 LH
    March 3, 2013

    “This is why people always dump on nutrition science when it appears to change every 10 years. Results get overblown, and when the inevitable regression towards the mean occurs, we get blamed for it. The reality is, the press coverage of science is extremely poor, and there is not adequate critical analysis and presentation of results to their audience.”

    Because, with diet probably more than anything else, people are looking for a magic solution, especially if it’s one that ascribes their currently unhealthy situation to factors beyond their control and so they are not to blame.

    “It’s not that you overeat junk food, undereat produce, smoke, drink, and don’t exercise; it’s that you had sugar!” “Eat this one thing everyday to lose 10 lbs in 10 days!” “[Fill in the blank] is evil and should always be avoided by everyone!” Etc.

    No one wants to hear “Your lifestyle choices, potentially in combination with genetics beyond your control, have resulted in your being overweight, diabetic, and/or at greater risk for significant health problems. Although you can’t change your genetics, you can control your lifestyle by minimizing known stressors such as obesity while increasing your consumption of foods known to have beneficial health properties and increasing your level of exercise. However, this needs to be maintained throughout the course of your life.”

    The media are just giving people what they want to hear. Not that the media shouldn’t be blamed too, because there is certainly supply side pushing going on through sensational headlines designed to sell papers/magazines. But it’s not like there’s a large market out there for people who want unbiased, nuanced reporting (on any issue).

  4. #4 Cvrai
    March 4, 2013

    “This is why people always dump on nutrition science when it appears to change every 10 years. Results get overblown, and when the inevitable regression towards the mean occurs, we get blamed for it. The reality is, the press coverage of science is extremely poor, and there is not adequate critical analysis and presentation of results to their audience.”

    Is this really the fault of the press or the institutions that deign to give dietary advice based on weak and shoddy studies? Why did the American Heart Association recommend their low-fat diet? Why did the FDA Food Pyramid have carbohydrates at the base of their nutrition guide? Why was the Atkins diet vilified? Why were eggs, dairy, and coconut oil smeared? Why are people given the impression vegetarianism is much healthier? Is salt really that bad? If the press is bad it’s because the science is bad and the press isn’t responsible for the science, the scientists are.

  5. #5 Cvrai
    March 4, 2013

    I can include red meat among the questionable assertions as well. A very large study says it is significant; another larger meta-analysis that includes the other study says it isn’t. Did they control or account for confounding factors? Do they even know what they might be? For all we know the use of preservatives and nitrosamines or maybe hepatitis C account for all the difference.

  6. #6 Matthew Herper
    March 4, 2013

    Re: The Mediterranean diet: Expecting a mortality benefit in a randomized controlled clinical trial of a diet in primary prevention is probably expecting too much. It’s very, very difficult to show a mortality benefit in preventing first heart attack and stroke.

    If your reading of that study is right (and I think you’re picking it apart way too much — we trust primary, not secondary endpoints) then that would be great news because we can all just eat lots of olive oil and have fewer heart attacks.

  7. #7 Mark
    March 4, 2013

    Matthew, I agree, that’s probably the way that this should have been presented, as that’s pretty much the only evidence-based statement you could make about this data. My main complaints are that the participants weren’t really on a Mediterranean diet, the controls weren’t on any kind of intervention at all, and while both of the Med diet groups had decreases in cardiovascular events, the two arms diverged on mortality. Very confusing.

    It does make me happy though because I use lots of olive oil in my cooking. At least it suggests that I won’t increase my risk of death with lots of olive oil use.

    Cvrai, your comments make no sense. What are these studies your talking about, and hep C? That’s out of left field.

  8. #8 Matthew Herper
    March 4, 2013

    But that kind of divergence on mortality happens all the time on CV studies. It’s the result of small numbers. It’s also why you’ll frequently see CV mortality, not mortality, as an endpoint — because a car accident can mess up the final result.

    That study may actually be the strongest clinical evidence we have that one diet can prevent heart disease over another. That it did it by adding the right kind of fat is a lesson. Remember that in a very well-done RCT (the WHI trial) a low fat diet failed to make a difference.

  9. #9 Mark
    March 4, 2013

    The WHI study is interesting, but two issues make that reading incomplete regarding obesity. The “low fat” diet group wasn’t counseled to start the diet to lose weight, weight loss and calorie reduction was not a goal. While they did lose a little bit, most of it was gained back by the end of the study. They just changed the proportions of fats in their diet. Further, while the low fat diet intervention did not result in a lower mortality compared to controls, subgroup analysis (always risky, but worth noting) showed those with less saturated fat intake, and higher fruit and vegetable intake trended towards lower CV mortality (again no statistical signficance). Further, the increased carbohydrate intake, at the very least, didn’t worsen cardiovascular risk, which had been a concern of other studies of fat replacement with carbohydrate might result in aggravated insulin resistance more atherogenic dyslipidemia.

    So while a “low fat diet” was prescribed it wasn’t a “weight loss” diet, so it doesn’t really speak to obesity as a risk factor. Also, the reduction in saturated fat intake wasn’t that great. In trials where the diet emphasized saturated fat reduction and weight loss if BMI elevated, the diet was effective, lowered serum cholesterol levels with positive effects on CVD events, and direct regression of atherosclerotic disease was even measured by angiography.

    The other fascinating thing is that the WHI study prospectively studied sugar intake. And guess what? They found no association between sugar intake and diabetes. A study of 36000 Iowan women showed no difference in diabetes related to carbohydrate intake. When broken down to individual sugars, interestingly, glucose and fructose at the highest quintiles of intake had a positive association, whereas sucrose had an inverse relationship to diabetes onset, and total starch, maltose and lactose had no relationship. The WHI study found the opposite with an inverse association between all of these sugars and diabetes onset in multiple regression analyses. Something might be happening here, but I kind of doubt it.

    The results on replacing saturated fats with polyunsaturated fats have been mixed, but the trials that failed to show a difference, like the Minnesota survey, still showed positive trends, albeit in a younger population. In general the data has shown a decrease in saturated fats to lower cardiovascular risk. And there are similar studies in culture and animals suggesting a negative effect of saturated fats on inflammation and insulin sensitivity. So, decreased saturated fats (about 10% of diet) is probably good for cardiovascular health and decreasing cholesterol and cardiac risk factors. However, replacing fat with carbohydrates is probably a bad idea, might increase your weight, and cause a more atherogenic dyslipidemia. Studies such as the Mediterranean diet one suggest, “hey maybe olive oil is alright”, but don’t speak to other aspects of the diet such as decreased red meat/chicken/dairy consumption, as the study participants didn’t adopt these behaviors. What hasn’t changed, and has been consistent through all of this, is that obesity increases morbidity and mortality. Since obesity is very hard to correct, prevention must be emphasized.

    I think we’re still pretty far away from knowing what is the ideal make-up (saturated, unsaturated, and polyunsaturated) of the fats we consume, but especially in patients with cardiac risk factors, the data are fair that avoidance of high levels of saturated fats may prevent disease progression and decrease cardiovascular events. I also agree with the sugar Nazis, eating a bunch of refined sugars is a terrible idea, I just think it’s more because it causes obesity than that there is a direct link between sugar consumption and pathogenesis of diabetes. Just because sugar aggravates diabetes doesn’t mean it causes it, similarly, just because salt aggravates hypertension, doesn’t mean we’ve been able to prove a link there either.

  10. [...] blogger Mark Hoofnagle is critical of Bittman for drawing the wrong conclusions from Lustig’s study and states what we [...]

  11. #11 Henry Miller
    March 5, 2013

    More on Bittman’s shortcomings concerning science, agriculture and journalistic ethics: http://dailycaller.com/2012/11/12/a-david-and-goliath-parable/

  12. #12 Mark
    March 5, 2013

    The only problem being it’s in the Daily Caller, a den of iniquity and lies if there ever was one. For Ann Coulter alone I’d suggest avoiding that rag.

  13. #13 Henry Miller
    March 5, 2013

    I’d suggest you judge the article on its merits, but if you really object to where it appeared, how about Forbes.com? See http://www.forbes.com/sites/henrymiller/2013/01/16/bittman-of-the-new-york-times-always-out-to-lunch/ and http://www.forbes.com/sites/henrymiller/2011/02/23/warning-labels-on-new-york-times-columns/.

  14. #14 Mark
    March 6, 2013

    Forbes is pretty hit or miss. I generally agree, Bittman on GMOs is a perfect storm of paranoia, Luddism, and ignorance.

  15. [...] Bittman changes his tune on Sugar Study, while Mother Jones Doubles Down Did We Doom the Mammoths? The end times of saber-toothed cats, giant wombats, and carnivorous koalas. So how do we fix the Ph.D/Postdoc glut? Sequester Starts, NASA to Lose $896 Million The Black Queen Hypothesis (actually sounds like niche construction more than anything else) [...]

  16. #16 S
    March 9, 2013

    OT, Mark – Do you have an email address so I send you something we discussed?

The site is undergoing maintenance presently. Commenting has been disabled. Please check back later!