In the aftermath of his commentary about butter in the New York Times, and my commentary about his commentary in my LinkedIn and The Huffington Post blogs, Mark Bittman and I -- along with several others, including Dr. Dariush Mozzafarian from Harvard, one of the authors of the study that set this all in motion -- were invited to discuss the health effects of butter on the NPR program, "On Point," this morning. Frankly, Mr. Bittman's on-air comments were very moderate, and very reasonable, far more so than his column, and in particular, the headline. Mr. Bittman let me know via email that the headline was not of his choosing, but that doesn't matter much, since the headline merely repeats ver batim his own words from the second line of his piece.
In any event, I would like the world to know that if Mr. Bittman's column had sounded just like his comments on NPR today, I would merely have appended: amen. But it did not. In fact, rather gratifyingly, his comments today sounded rather more like my column than his, at least to my ears. So amen to the on-air comments today, but not to the New York Times column which will doubtless reach many more people, and do some real damage as a result.
Here, in a dollop, is what I have to append to my prior commentary about the nature of that damage.
1) Regarding the study: the new study was very questionable in its significance and implications for many reasons, most salient of which is the complete failure to consider food substitutions. People eating less of A invariably eat more of B in its place. This study did not consider, address, or even mention B. We can, however, make a well-informed guess* about what 'B' is in our culture, and I'll return to that shortly.
2) Regarding Mr. Bittman, and with all due respect: all evidence suggests he opined on the study, in the New York Times no less, never having read it. If he had read it, by his own admission to me on a prior occasion, he lacks qualifications to interpret epidemiologic research as anything other than a guy with a decidedly non-expert opinion, and in this case a pretty clear bias into the bargain.
3) Regarding butter: the study was not about butter. It was about variation in fatty acids, both in dietary intake and in the bloodstream. So the only legitimate way to apply the findings to butter would be to compare the fatty acid profile of butter to the study findings. I think it can go without saying that Mr. Bittman did not do that. I did -- as, I presume, did most of my professional colleagues in nutrition. Here are the results of that comparison:
- The most abundant fatty acid in butter is palmitic acid, a 16-carbon saturated fat.
- Of 17 fatty acids included in the analysis, palmitic acid was associated with the three greatest increases in rates of heart disease; the greatest increase of all was associated with stearic acid, a 18-carbon saturated fat also abundant in butter.
- Therefore, while it's a leap of faith to apply this study to butter at all, if that is done -- the only conclusion possible is that butter, is indeed, a hazard to the heart, and on that basis best minimized or avoided -- not added back.
- As for the effects of butter on recipes, or taste buds, I defer to Mr. Bittman. But he should apply his opinions very cautiously to the public's coronary arteries before someone gets seriously hurt.
4) *Finally, that informed guess: Abundant evidence indicates clearly what we eat instead when we cut our saturated fat intake. Americans, alas, have not swapped out their salami for spinach, nor their Swiss cheese for Swiss chard. We have swapped out sources of saturated fat for sources of refined starch and added sugar, like Snackwell cookies.
- On the one hand, this readily explains why lower intake of saturated fat is not associated with lower rates of heart disease: we simply moved from one way of eating badly to another.
- But more interesting in the context of Mr. Bittman's argument that starch and sugar are the "real" culprits in our diets: If that were so, the lowest intake of saturated fat should be associated with the highest rates of heart disease, because less saturated fat means more sugar and starch. But that association was not seen; there was almost no variation in heart disease rates with variation in saturated fat intake. So this study also shows that the nutrients replacing saturated fat -- namely, sugar and starch -- are equally unassociated with variation in heart disease rates. This is a classic case of: what's good for the goose is good for the gander. There is no evidence here to support the assertion that something other than saturated fat is the "real" culprit.
- The reason for that, other than the important limitations of the study, is that there is no single "culprit" -- that effort to find a scapegoat is misguided and harmful. Rather, the overall pattern of the diet exerts an enormous influence on heart health -- both because of what the diet includes and what it excludes. The evidence is clear that overall variation in diet quality can dramatically influence heart health, and heart disease risk. And with suitable attention to foods and the pattern of the diet, the nutrients sort themselves out quite handily.
We can all benefit -- hedonistically, at least -- from Mr. Bittman's aptitudes in the kitchen. And when the dust clears, he and I agree -- in Michael Pollan's good company -- about a diet of minimally processed foods, mostly plants. But when it comes to epidemiology, he needs to be a lot more careful about impersonating an expert (whatever the associated headlines). He is otherwise at risk of being one cook too many, and doing nothing more productive for public health than spoiling the stew.
Dr. David L. Katz has authored three editions of a nutrition textbook for health care professionals. He is editor-in-chief of the peer-reviewed journal, Childhood Obesity, and President of the American College of Lifestyle Medicine. He was commissioned by Annual Review in Public Health to write the review article, Can We Say What Diet is Best for Health? He is the author, most recently, of Disease Proof.