This week - in my ongoing series "Is That Thing You Eat Everyday Secretly Killing You?!" - #Erythritol!
I want to dig into a nice @NatureMedicine paper that suggests the sugar substitute might increase the risk of cardiovascular disease. (Thread/)
Erythritol is a non-nutritive sweetener used in all sort of products - toothpaste, gum, especially "keto friendly" stuff. Also monkfruit sweetener. It does NOT need to be labeled "artificial" since it can be found (in small quantities) in nature.
Data all comes from this paper @NatureMedicine - definitely a cut above your usual nutritional epidemiology fare - multiple lines of evidence here to tease out. nature.com/articles/s4159…
The researchers started with a metabolomic analysis of ~1000 people being screened for CV disease. Seems the focus was on sugar / sugar substitute metabolites but erythritol came out on top as most strongly associated with subsequent CV events.
Of course, this would need to be validated in an external dataset. Researchers did this too, finding in a US cohort, for example, that those in the highest quartile of erythritol levels had higher rates of CV disease over time.
But... correlation or causation, right? They adjusted for stuff like age, diabetes, BMI, etc - but still - people who use a lot of sugar substitutes are likely different in non-measurable ways from people who don't.
But what if there *is* causality here. How would that work? The researchers show that, in a test tube, higher levels of erythritol lead to increased aggregation of platelets (clotting, basically).
And in a model of carotid artery occlusion (in mice), erythritol speeds to the occlusion process... so there's at least some biologic plausibility here.
But if I've learned anything reviewing sugar substitute literature - it's to check the dosages. Just because you can give a rat cancer giving it 1000x the daily intake of some sugar substitute does not mean it is clinically relevant.
I'll use 45 uM as a "concentration of interest" here - since that appears to be the place where we see some of that platelet aggregation. Is that a reasonable concentration? The authors gave 8 healthy controls 30grams of erythritol and this is what the blood levels did:
After 30grams of intake - you can get super-high levels - persisting above 45 uM for almost 2 days. (This is likely based a bit on kidney function - 90% of erythritol is excreted unchanged in the urine).
But... is 30grams reasonable? The authors argue that 30grams is the average daily intake in the US based on this FDA document - but it looks like that's actually the 90th percentile...
And a european study (of people who exclusively use sugar-free foods / drinks) show intake levels way lower than that.
How much is 30 grams? Turns out I had a bag of erythritol in my closet.
I don't think most people get close to 30grams of erythritol - likely much much less - unless you are deep in the "keto" ecosystem. I therefore don't think we need to worry about the potential risk too much.
Remember, the risks of artificial sweeteners should not be evaluated in a vacuum - they need to be compared to the alternative - sugar - which is probably the single substance most responsible for the epidemic of overweight, obesity, metabolic syndrome, T2D, etc.
Still, there is a truism in the diet research space here. You probably can't have your (sugar-free) cake and eat it too.
I've definitely noticed this phenomenon in American diet culture of late.
But high-protein diets may not be all they are cracked up to be. Why? A new study in Nature Metabolism puts the blame squarely on a single amino acid: leucine.
🧵
OK there are three macronutrients. Your caloric intake will be comprised of some combination thereof, and basically all of the diet wars of the last 40 years can be cast in terms of macro content.
When I was growing up (it was a time called the 80s. Thing were... interesting) low-fat was the crazy. Remember "fat makes you fat"?
Does it seem like your hospital patients are getting more... complicated? They are!
🧵
Just went through this paper from @JAMAInternalMed which, like all good papers, comports with my prior beliefs ;-). No but seriously, I definitely feel like hospital complexity is rising.
Data comes from British Columbia - good for 2 reasons. 1) Universal EHR so you capture all hospital admissions 2) Universal healthcare - so you don't need to worry as much about access issues, etc.
I believe that RNA therapeutics will completely transform medicine. Imagine a future where, instead of taking medications daily, you take a shot maybe once a year. That future is, basically, here.
🧵
Currently, this is more or less how our drugs work. They do something to a protein. Inhibit it, cleave it, block its binding to some receptor, speed its degradation, etc.
And, if you really think about it, disease is mediated by proteins. High cholesterol? Protein problem. Sepsis? Proteins. Viral infection? Viruses hijack our cells to make their... proteins.
AI in Medicine is suffering from what I would like to call "The Cassandra Problem"
🧵(1/17)
I am thinking about this issues this week thanks to this study, appearing in @JAMANetworkOpen, which examines the utility of a model that predicts blood clots in hospitalized kids. (2/17)
The model takes in a bunch of data from the electronic health record, and makes what turns out to be a pretty accurate prediction about the risk of blood clot (AUC~0.90). (3/17)
Absolutely fascinating @JAMA_current paper this week suggesting that Cerebral Amyloid Angiopathy may actually be infectious. (Thread)
The background here is that CAA is the second most common cause of intracerebral hemorrhage (brain bleed) after hypertension. There are some genetic causes, but most are thought to be idiopathic. Just amyloid getting deposited for no clear reason.
But there have been some clues that there might be a transmissible agent involved. This @Nature paper describes cases attributed to contaminated cadaveric pituitary hormone. nature.com/articles/natur…
37 degrees Celsius, 98.6 Fahrenheit. That's normal body temperature, right?
Wrong.
New data suggests true "normal" is 98.0 degrees. (thread)
We get 98.6 degrees from this guy Karl Wunderlich, who measured 1,000,000 temperatures from 25,000 Germans in the mid 1800s. He was really the first to realize that "fever" was not itself a disease, but a symptom of a disease.
He took the average, got 98.6, and that was that. UNTIL NOW.
(Ok honestly, people have known this was wrong for a while, but the new data is cool).
From @JAMAInternalMed: