A Nutrition Hack That Probably Doesn’t Matter After Thanksgiving
A Nutrition Hack That Probably Doesn’t Matter After Thanksgiving
I’ve stayed away from throwing shade at lay coverage of nutrition science since our articles, Nutrition Churnalism Part 1 and Part 2, that we published way back in August of 2023. It’s not that I don’t love a good observational study showing that eating chilies four times a week is more effective than any medication at preventing cardiovascular disease or that eating organic food cuts cancer risk by 25%. It is just that attacking these articles is like shooting the proverbial fish in a barrel. The underlying science is often flawed, and the lay coverage frequently features many of the seven deadly sins of churnalism.
Sin #1: Observational studies almost never prove causation
Sin #2: Extrapolation and generalization
Sin #3: Ignoring confounding, selection bias, and other epidemiological errors
Sin #4: Neglecting plausibility
Sin #5: The Disclaim and Pivot maneuver
Sin #6: Keep testing; report just once
Sin #7: Being incurious
But, as a special treat, it seems like the season called for a Sensible Medicine post thanking the New York Times, the paper of record, for an article arguing that we would be healthier if we ate our carbs the days after Thanksgiving, after they had spent some time in the refrigerator.
On November 8th, the Times published an article titled, Pasta and Rice May Be Healthier as Leftovers. Here’s Why. The gist of the article was that resistant starches, those that are harder to digest and are present in whole grains, beans, nuts, and other foods are better for us than regular starches, like those that are in rice, pasta and the lovely, mashed potatoes that I hope you recently enjoyed. Further, regular starches can be made more resistant by cooling them (and they actually remain resistant upon reheating).
Ok, this might seem too good to be true, and you know what they say about things that seem too good to be true, so let’s think about where ideas like this come from.
To support the “resistant starch” benefit, the author cites a (somewhat irritating) review article. The mechanistic arguments in favor of resistant starches involve lower postprandial glucose spikes and an “improved microbiome” among others. Diving into the literature, I found that the benefit of resistant starch has been studied mostly in small studies with surrogate endpoints. Fortunately some good meta-analyses have been published (here, here, and here). The current evidence suggests that resistant starch does lower post prandial sugars but has no effect on fasting plasma insulin, insulin sensitivity, acute insulin response and other measures. Probably not surprising, there is low to no evidence from controlled trials that a change to resistant starches improves clinical outcomes.
There is one very interesting study that the author of the Times article cites. The CAPP-2 RCT randomized people with Lynch Syndrome to 30 grams of resistant starch daily or placebo for up to 4 years. The researchers then looked at long-term cancer outcomes 10-years later. There were no differences in colon cancer incidence (the outcome that the authors expected to impact) but fewer participants randomized to resistant starch had non-colorectal lynch syndrome cancers (27 vs. 48; HR, 0.54; 0.33–0.86; P = 0.010).
The author of the NYT article makes the jump to equate resistant starch and fiber and discusses the benefits of fiber. Resistant starch and insoluble fiber are different. In the colon, insoluble fibers pass through unchanged while resistant starches are fermented by bacteria – hence the idea that resistant starch might have beneficial effects on the microbiome.
So, with that knowledge let’s get to how the NYT article racks up the churnalistic sins. The obvious issue here is that the benefit of resistant starches over simple sugars is based mainly on observational and surrogate outcomes. The leap from these to actual clinical benefits – decreased mortality and improved quality of life – is always a large one. The extrapolation from a 20-year study of people at high risk for malignancy to you sitting in your kitchen the days after thanksgiving is a huge one. Then, there is the idea that this imagined benefit would stand up to being extrapolated to simple starches, maybe made a bit hardier by a night in the fridge, is getting pretty extreme. It seems like our author has committed sin #1 -- observational studies almost never prove causation -- and sin #2 -- extrapolation and generalization.
The conclusion? I hope you enjoyed some simple and complex carbohydrates -- pecan pie, sweet potato pie, mashed potatoes, corn bread stuffing -- right off the oven. I hope you continue to enjoy cold leftovers out of the fridge or reheated in the coming days.
Have you recently read an article or story in the lay press that makes you crazy? Share it with me or write it up for us at Sensible Medicine.
Adam, the moral of the story seems to be eat your Thanksgiving feast with gratitude and grace and feel no guilt. I did my 5-mile run before the feast and I took two walks afterwards. I ate well and I felt no guilt. I get so tired of all the guilting of people for enjoying a good meal with their family. Some folks need to get a life and mind their own damn business.
Generally a good article debunking much of the nonsense that journalists propagate regarding food and diet. Early on the author refers to a two part article on Nutritional Churnalism published on this site in August of this year and I would object strongly to a few of the statements contained therein.
The worst: "The cholesterol lowering statins are the most effective medications ever developed for treating cardiovascular disease. These drugs, marketed under names like Lipitor, Zocor, and Crestor have revolutionized the care of patients with cardiovascular disease. The most impressive results we have for these drugs shows that, in high risk patients, statins decrease the risk of death from heart attack and stroke by 34%." He links to the article and it is the Scandanavian Simvastatin Survival Study (4S). In this trial the incidence of Cardiovascular Mortality was 5.0% and 8.5% in the drug and control groups respectively. The Total Mortality was 8.2% and 11.5%. So the risk of death figure of 34% is obviously a relative risk as the actual risk reductions were 3.3% and 3.5%. To cite only the relative risk is a tactic that I was surprised to see on this web site. Even more so since the same author cited in Part 2 a hypothetical study with incidences of 5% and 8% and dismissed that as random variation and characterized as "just noise".
I reviewed this 4S study along with several others in my book The Cholesterol Delusion and will quote a paragraph from it in order to illustrate a further weakness in the study:
"The most interesting feature of this trial (4S) was not even mentioned in the published report. There was a total disconnection between the improved outcomes and both the initial cholesterol level and the degree of cholesterol lowering attained. In other words, protection against CHD was the same whether the initial cholesterol was relatively high or low, and the degree of protection did not correlate with the degree of cholesterol lowering. Subjects whose cholesterol went down a relatively small amount benefited to the same degree as those whose cholesterol went down a lot. This disconnection is called "lack of normal exposure-response" and generally means that the factor under investigation is not the true cause of the disorder. In plain English, the small benefit conferred by the drug was not due to its effect on cholesterol but rather to some other cause,"
I apologize for the length of this comment but I feel compelled to counter statements like this that we see all too often regarding statin drugs.