Eugene Murer at His Pastry Oven

Killer Croissants in the PURE Study?

SF Eater tells us that 13 bakeries in San Francisco have “killer croissants.” The Atlanta Journal Constitution tells us that croissants, along with white bread, tie us to an early death. Their source for this epidemiological wisdom is the BMJ. That treasure trove of epidemiology – the PURE study – has yielded another publication. This time the subject is refined grains. Researchers found that eating too much of them correlates with heart disease and an early death.

But whole grains didn’t correlate with better outcomes. White rice consumption was neutral, too. So what gives with this study of 148,858 people in 21 countries for 9.5 years? Shall we just take what we like from these conclusions and leave the rest?

A Big Jumble of Self-Reported Data

The PURE study is ambitious. PURE stands for Prospective Urban Rural Epidemiology. The goal was to collect data for at least 10 years from 140,000 individuals in more than 600 communities and 17 countries at various stages of economic development. For scale, the study has exceeded its goals.

But it attracts a fair amount of skepticism. Back in 2017, a sensational publication of data from the PURE study suggested that more fats in a person’s diet predicted lower risk of death. On the other hand, it suggested a higher risk of death with more carbs. This set off a wave of protests from nutritional epidemiologists complaining that you cannot trust this particular piece of nutritional epidemiology. Eugnia Gianos et al wrote:

“Oversimplification of macronutrients (fat, carbohydrates, protein) as good or bad sets the field back through misuse by the lay public and potentially, clinicians who are guiding the public.”

Oversimplification Bad, Food Good

The food that we eat is woven into our culture. Trying to find good and bad foods across 21 different countries in the PURE study might just be a fool’s errand. Too many croissants might not be a good thing, but on occasion, we can assure you that a croissant can be a very good thing. And white rice? Well, with more than a half of the study population coming from Asia, there’s a lot of rice consumed in this sample. So it’s not surprising that these number crunchers could not find a problem with rice.

The thing is that we have a strong impulse to label foods good or bad. Tufts Dean of Nutrition Science, Dariush Mozaffarian, sums up this thinking in a tweet: “Eat MORE good foods, ignore neutral foods, avoid bad.”

But too much of a good thing can be a bad thing. And what one person regards as bad food can easily be a delicacy that has a place in another person’s healthy diet. So we’ll pass on simplistic dietary advice about killer croissants.

Click here for the study in BMJ, here for reporting from the AJC, and here for more from Scimex.

Eugene Murer at His Pastry Oven, painting by Camille Pissarro / WikiArt

Subscribe by email to follow the accumulating evidence and observations that shape our view of health, obesity, and policy.


February 23, 2021

4 Responses to “Killer Croissants in the PURE Study?”

  1. February 23, 2021 at 7:46 am, Christine Rosenbloom said:

    Killer croissants? Why aren’t all the French dead if croissants are killing people?

  2. February 23, 2021 at 9:58 am, David Brown said:

    Good foods? Bad foods? As if foods could possess virtue. Better to think in terms of appropriate food choices.[1]

    Other important considerations are the impact of nutrient deficiency, imbalance, and excess on cell signaling[2] and mitochondrial function.[3]

    Web page references

    • February 23, 2021 at 10:32 am, Ted said:

      Good thoughts, David. I will say, though, that I don’t rely on YouTube for nutrition science.

  3. February 23, 2021 at 11:03 am, Stephen Phillips said:

    Food is not our moral compass and is neither good nor bad. All food contains nutrients and provides energy and cannot be labeled as healthy or unhealthy or good or bad. These words feed into the diet mentality and fuel the perfectionist, all or nothing forbidden food behaviors. Certainly some foods could be of advantage or disadvantage to specific medical conditions but this does not moralize them nor make them inherently good, bad, healthy or unhealthy.