A Fake Nutrition Controversy in Four Bullet Points
TL;DR No there haven’t been a series of recent studies that disagree about butter, only headlines.
NOTE: Keep in mind that often the press releases fuel the headlines as much as news articles, so news outlets and bloggers aren't the only ones complicit in creating confusion! This article isn't intended to single out a single publication as the cause of the issue.
EurekaAlert posted a press release titled “Saturated fat could be good for you” this week that made waves and eventually ended up with its own Daily Mail headline.
At the top of the Daily Mail story was this:
So either there is some HUGE discrepancy in the way nutrition scientists see the function of saturated fat…. Or press releases (and the subsequent headlines born from them) are just playing a giant 5th-grade-birthday-party game of telephone with the truth.
Firstly, did Tufts find that butter was harmless? Well, no. They just found that if you looked at the INDIVIDUAL FOOD BUTTER in people’s diets, the amount they reported eating (in diet recalls) wasn't correlated with heart disease outcomes (in fact it was slightly inversely correlated with heart disease and diabetes), but was statistically associated with overall mortality. This isn’t too surprising for reasons we’ll discuss below.
The concern over butter largely relates to its lower unsaturated fat content and greater saturated fat content; saturated fats can increase blood cholesterol levels whereas unsaturated fats have more favorable effects. To test the hypothesis that increasing saturated fats increases cardiovascular and/or total mortality in observational studies, like those cited in the daily mail, there are two important questions to answer:
1) Are total intakes of saturated fat associated with increased mortality?
2) if you replaced intakes of saturated fat with another nutrient, would you affect overall mortality?
"Are we a good fat? Or a bad fat?"
The butter publication from Tufts was a meta-analysis of previously published cohort studies and really only focused on whether butter intake alone, not total saturated fat intakes, was significantly associated with death. We might think of butter as being just a stick of solid fats, but in reality, food surveys have shown that it is only a relatively minor source of saturated fat in the diet (in this study, the effects of butter were considered in 14g servings (7g of saturated fat total)).
While butter can contribute to total saturated fat intakes and higher blood cholesterol, it’s certainly possible to eat very little butter and still consume a lot of other sources of saturated fats, such as cheese. Looking at high vs low butter consumption in these analyses doesn’t account for other sources of saturated fat intake, so it’s not super surprising that butter alone isn’t a major predictor of cardiovascular or total mortality.
In addition to considering total saturated fat intakes, you also have to ask yourself what else could people be eating instead. In observational studies looking at foods, you have to consider that people don’t eat in a vacuum.
When you look at butter intake as a percentage of total calorie intake (which observational studies often do), this means something has to fill the void that butter leaves if someone isn’t eating it!
Someone who doesn’t eat a lot of butter in these studies might be filling that caloric void with nuts and legumes, and in that case, you’ll likely see a benefit from removing butter. If it’s filled with trans-fat laden baked goods, you’ll make butter look relatively good! In large cohorts, where some people are doing the former, and some people are doing the latter, it’s not too shocking that you don’t find many significant associations between butter intake and mortality.
The unfortunate part of media presenting the Tufts butter study as though it says something entirely contrary to the Harvard study is that the two really aren’t incompatible! In fact, the Tuft findings - that a single source of saturated fat is weakly related to death - actually agrees with the Harvard study that followed.
Harvard's was another observational study that found that comparing the highest quintile of total saturated fat intake with the lowest was associated with an 8% increase in relative risk of mortality. Furthermore, those who ate the most polyunsaturated fats had an 11% decrease in total mortality.
So ultimately, these two studies don’t disagree as much as one might initially think. The totality of the evidence, from observational studies like these and from randomized controlled trials, supports current dietary recommendations to replace saturated fats with polyunsaturated fats.
The Daily Mail screengrab implies that the Harvard analysis was on the effect of butter - it was not. It was on the effect of total saturated fats. And this study was different in that it assessed the change in risk you see when you substitute one macronutrient in place of another.
Did Tufts then rescind its endorsement of butter? Also no. Consider that this sentence is in the original Tufts study listed:
“Our results suggest relatively small or neutral overall associations of butter with mortality, CVD, and diabetes. These findings should be considered against clear harmful effects of refined grains, starches, and sugars on CVD and diabetes; and corresponding benefits of fruits, nuts, legumes, n-6 rich vegetable oils, and possibly other foods such as fish on these endpoints. In sum, these results suggest that health effects of butter should be considered against the alternative choice.” [Emphasis mine]
So while they say that a little butter likely isn’t worse than other highly processed foods, it all depends on what you’re comparing with butter.
Exactly the same as the conclusion as the Harvard study.
And in Tuft’s follow up study purportedly saying butter should be avoided they conclude this:
“In comparison to carbohydrate, SFA, or MUFA, most consistent favourable effects were seen with PUFA, which was linked to improved glycaemia, insulin resistance, and insulin secretion capacity.”
Wow, it sounds really similar to the conclusion of the previous two studies.
So the Daily Mail headlines could have more accurately said something like this:
A study from Tufts said that the health effects of butter likely depend on what you replace butter with.
A study from Harvard said that the health effects of butter likely depend on what you replace butter with.
A later study from Tufts said that the health effects of butter likely depend on what you replace butter with.
Who wants to bet the last bullet point should actually look like this?
Now, a Norwegian study finds that the health effects of butter likely depend on what you replace butter with.
Want to know what the Norwegian "FatFunc" study found? Check in tomorrow for our take on that RCT!
Check out other times headlines and press releases have strayed from actual findings: