8 ms·

Association of insulin resistance marker with severity and mortality of Covid-19

95 points by sudoaza 6y ago

arkades 6y ago
Please note: The calculation of TyG is ln (fasting blood sugar) x triglycerides/2). Many studies make the error of calculating it as (ln (FBS x TG))/2. The only online calculator I've found, I think, falls into the latter category, or has worse errors - I didn't go through it too rigorously, but I put in values beyond what human life can sustain and didn't get close to the cut-off for this paper's bottom risk tier. If you look at studies/calculator using the latter calculation, it looks like this study looks at really, really severe diabetics. If you compare with the appropriate calculation, though, they're looking at more run-of-the-mill "not optimally treated and obese" diabetics. Hosseini 2017 did a paper analyzing a number of other TyG papers and calculating results under both calculation methodologies, for context. Please also note that this paper does not state when the results were collected. Insulin resistance/hyperglycemia is a symptom of sepsis - if these labs were drawn on already-severe patients, it would be entirely unclear whether they reflect a cause or an effect (or, as is almost certainly the case, both!).
- dreamcompiler 6y ago
  > The calculation of TyG is ln (fasting blood sugar) x triglycerides/2). You have a extra right parenthesis, which makes it ambiguous. Do you mean ln (FBS) x triglycerides/2 Or ln (FBS x triglycerides/2) ?
  - tw000001 6y ago
    Doesn't matter since both multiplication and division are...commutative? Don't remember the terminology but (10x2)/4 == 10x(2/4) == 5
    - rat9988 6y ago
      It matters here because he is using ln, and we don't know if we should stop at the first right parenthesis or the second. The word you are looking for is associative.
      - CydeWeys 6y ago
        Ooof, thanks for that. I didn't realize that was the natural log function because of the space. My brain was parsing it as the word "In" and essentially ignoring it.
        pbhjpbhj 6y ago
        There shouldn't be a space IMO, reading it out of context I couldn't see what the issue was either as I too was not parsing the ln as log_e. ln x or ln(x) but not ln (x)
    - indiv0 6y ago
      ln(5) * 10 / 2 ~= 8.047 ln(5 * 10 / 2) ~= 3.219 So it does matter.
  - arkades 6y ago
    Sorry, typo in my original post. It's ln(FBS x (Trig/2))
- graeme 6y ago
  So would it be accurate to say we could use that calculator if we: * Cut our glucose number in two * multiple their result by 2 Assuming you’re saying they are incorrectly dividing the ln by two, when they should be dividing blood sugar by 2. So we need to manually adjust the blood sugar, and multiply their final result by two to adjust for their ln. What’s an optimal result? Using the method above I got 7.5 https://www.mdapp.co/tyg-index-calculator-359/ https://www.mdapp.co/tyg-index-calculator-359/
hirundo 6y ago
"[triglyceride and glucose] index was closely associated with the severity and morbidity in COVID-19" So perhaps part of the reason why COVID-19 morbidity is lower in Japan/Korea/Taiwan compared to the U.S. is due to lower prevalence of metabolic syndrome. I wonder if that's also true for Europe.
- arkades 6y ago
  > It suggests a protective effect of a keto/paleo diet. No, if it can be taken at face value, it suggests a protective effect of a healthy lifestyle.
  - floatingatoll 6y ago
    If taken at face value, it suggests a protective effect of low insulin resistance as measured by the TyG marker. Everything else is interpretations and chains of logical reasoning. Neither of you are particularly wrong necessarily, but a third option is that someone could be sedentary and eating carbs every day and have low TyG. It's common to state that "activity XYZ will provide action-at-a-distance medical benefit ABC" because stating accurately what's going on takes more words that sound less certain: "Paleo and keto diets may weaken Covid-19 by lowering insulin resistance" "A healthy lifestyle may weaken Covid-19 by lowering insulin resistance" But it's really worth saying it like this, even though few do. (And yes, these aren't 'maybe' enough, but they're an improvement.)
    - arkades 6y ago
      You replaced the word "suggests" with "may". You're seeing a distinction between these two words that I apparently do not, since I take them both to indicate "this is a possibility, though not the only one, and a far from proven certainty." I would add that, even there, my statement was predicated on making an inference about dietary habits - something I cast doubt on with the opening, "if taken at face value." Something I further opined on in my other post, where I pointed out that the paper may be reversing causality. I feel like you're criticizing a strong inference having been made, that I think pretty clearly wasn't.
ashtonkem 6y ago
I've suspected for a long time that Insulin resistance is going to be one of the next big areas of focus for public health, but I thought that it was going to happen the moment Apple finally figured out blood glucose measurement through the skin. I did not see a pandemic being part of it.
- gumby 6y ago
  I love my Apple watch but I am astonished at your faith in their ability to do noninvasive glucose measurement. People have broken their picks on that particular coal face for decades. For that matter DM2 is an area of quite active research (and, like diagnosis, has been quite active for decades) as the financial payoff for any success is enormous. The only downside for transdermal diagnosis is the lack of consumables, which makes it a tough market to enter and to be funded for. That is an area that's good for Apple as they are already selling the platform, so this would be a feature that would add to sales. And one I'd use.
  - dreamcompiler 6y ago
    O2sat measurement has been noninvasive without consumables for a long time and many phones can do it now. Why do consumables matter?
    - gumby 6y ago
      Investors don’t typically like diagnostics as the margins and volume tend to be quite low. They also tend to be more “vitamin” than “aspirin”.* Consumables, a least, give you recurring revenue and even in some cases the opportunity for a razor-and-blades model. You’d be surprised how many med products are designed specifically to require consumables. * (funny analogy to use in a med context)
49para 6y ago
Insulin Resistance is the start of (all?) metabolic disease. So easy to resolve using fasting, intermittent fasting, keto, carnivore etc diets. Unfortuneately it slowly builds up over decades and only once disease has progressed do Drs move on to treat the resultant disease (and mainly with cholesterol lowering drugs). Instead of measuring fasting glucose levels (which indicate diabetes), insulin levels should be measured as they are the leading indicator.
- Trasmatta 6y ago
  > Insulin Resistance is the start of (all?) metabolic disease. So easy to resolve using fasting, intermittent fasting, keto, carnivore etc diets. Definitely not all. Not the least of which is Type 1 Diabetes, which is marked by the inability to produce insulin, not insulin resistance. And it's currently impossible to resolve, all Type 1 Diabetics have to be on insulin therapy to survive.
  - ashtonkem 6y ago
    Is Type 1 diabetes considered a metabolic disease? I've only heard of "metabolic disease" in reference to "metabolic syndrome", the precursor to Type 2 diabetes and heart disease.
    - asplake 6y ago
      There are many metabolic disorders - some due to specific genetic issues, and some severely life-limiting (eg children unlikely to reach adulthood). And as per the Wikipedia page for metabolic disorders, not to be confused with Metabolic syndrome or Diabetes.
    - Trasmatta 6y ago
      It's a metabolic disorder. I guess I don't really know the difference between metabolic disease / metabolic syndrome / metabolic disorder. All problems of the metabolism, but I'm sure there's some medical distinction between them.
  - dreamcompiler 6y ago
subsubzero 6y ago
Covid-19 is a very strange disease, it seems like it is amplified either way (severe vs. non-severe) depending on health. This seems different from the flu as the flu hits everyone very hard, the elderly/sickly especially hard. With covid some people who have it, do not and will not have any symptoms which cannot be said for the flu. I feel like the insulin marker data is an albatross, having high TyG means alot of systems in your body are not doing well, and the virus attacks weakness, it(covid-19) is also found to produce extreme clotting so that is probably why people with diabetes/heart disease and hypertension are all at high risk.
- sudoaza 6y ago
  Yes from what I heard having metabolic issues also is correlated with less development of antibodies after immunization.
- nradov 6y ago
  Asymptomatic infections are about as common with influenza as with SARS-CoV-2. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4586318/ https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4586318/ Some influenza strains hit young, healthy patients harder than the elderly by triggering a cytokine storm. This was particularly bad in the 1918 H1N1 pandemic. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4711683/ https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4711683/
- Synaesthesia 6y ago
  It certainly is weird. I read an article which said doctors around the world are often scratching their head at what they’re seeing. It can attack the rest of your organs, not just your lungs.
- sacred_numbers 6y ago
  Actually, this is not exactly true. There are actually a lot of people who are infected with the flu who are asymptomatic[1]. Based on my amateur research it does appear that the difference in severity between age groups is more pronounced for COVID-19 than for the flu, though. Part of this may be sampling bias, however, since we are probably catching far more asymptomatic COVID-19 cases than asymptomatic flu cases. [1]https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4586318/ https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4586318/
zxcvvcxz 6y ago
Fat and unhealthy people do poorly with virus. What a surprise. They should lock themselves in with a treadmill.
danans 6y ago
Given that insulin resistance is strongly correlated with obesity [1], I'm surprised that wasn't a factor they controlled for, especially since the respiratory difficulties associated with obesity seem to be a significant risk factor for death with Covid19 cases. 1. https://obesitymedicine.org/obesity-and-insulin-resistance/ https://obesitymedicine.org/obesity-and-insulin-resistance/
shakil 6y ago
Lets not confuse correlation with causation. All this study shows is people with insulin resistance are at significant risk of dying from Covid, it doesn't identify what actually kills them. However, if you look at the role Vitamin D plays [1] in suppressing cytokine storms, which is what actually pushes over an organism to the point beyond recovery from Covid, and then understand that Vitamin D deficiency is common [2] in Type 2 diabetes, you can begin to understand the fuller picture. 1. https://www.medrxiv.org/content/10.1101/2020.04.08.20058578v4 https://www.medrxiv.org/content/10.1101/2020.04.08.20058578v... 2. https://pubmed.ncbi.nlm.nih.gov/26375925/ https://pubmed.ncbi.nlm.nih.gov/26375925/

95 points by sudoaza 6y ago

Please note: The calculation of TyG is ln (fasting blood sugar) x triglycerides/2). Many studies make the error of calculating it as (ln (FBS x TG))/2. The only online calculator I've found, I think, falls into the latter category, or has worse errors - I didn't go through it too rigorously, but I put in values beyond what human life can sustain and didn't get close to the cut-off for this paper's bottom risk tier. If you look at studies/calculator using the latter calculation, it looks like this study looks at really, really severe diabetics. If you compare with the appropriate calculation, though, they're looking at more run-of-the-mill "not optimally treated and obese" diabetics. Hosseini 2017 did a paper analyzing a number of other TyG papers and calculating results under both calculation methodologies, for context. Please also note that this paper does not state when the results were collected. Insulin resistance/hyperglycemia is a symptom of sepsis - if these labs were drawn on already-severe patients, it would be entirely unclear whether they reflect a cause or an effect (or, as is almost certainly the case, both!).

dreamcompiler 6y ago
> The calculation of TyG is ln (fasting blood sugar) x triglycerides/2). You have a extra right parenthesis, which makes it ambiguous. Do you mean ln (FBS) x triglycerides/2 Or ln (FBS x triglycerides/2) ?
- tw000001 6y ago
  Doesn't matter since both multiplication and division are...commutative? Don't remember the terminology but (10x2)/4 == 10x(2/4) == 5
  - rat9988 6y ago
    It matters here because he is using ln, and we don't know if we should stop at the first right parenthesis or the second. The word you are looking for is associative.
    - CydeWeys 6y ago
      Ooof, thanks for that. I didn't realize that was the natural log function because of the space. My brain was parsing it as the word "In" and essentially ignoring it.
      - pbhjpbhj 6y ago
        There shouldn't be a space IMO, reading it out of context I couldn't see what the issue was either as I too was not parsing the ln as log_e. ln x or ln(x) but not ln (x)
  - indiv0 6y ago
    ln(5) * 10 / 2 ~= 8.047 ln(5 * 10 / 2) ~= 3.219 So it does matter.
- arkades 6y ago
  Sorry, typo in my original post. It's ln(FBS x (Trig/2))
graeme 6y ago
So would it be accurate to say we could use that calculator if we: * Cut our glucose number in two * multiple their result by 2 Assuming you’re saying they are incorrectly dividing the ln by two, when they should be dividing blood sugar by 2. So we need to manually adjust the blood sugar, and multiply their final result by two to adjust for their ln. What’s an optimal result? Using the method above I got 7.5 https://www.mdapp.co/tyg-index-calculator-359/ https://www.mdapp.co/tyg-index-calculator-359/

hirundo 6y ago

"[triglyceride and glucose] index was closely associated with the severity and morbidity in COVID-19" So perhaps part of the reason why COVID-19 morbidity is lower in Japan/Korea/Taiwan compared to the U.S. is due to lower prevalence of metabolic syndrome. I wonder if that's also true for Europe.

arkades 6y ago
> It suggests a protective effect of a keto/paleo diet. No, if it can be taken at face value, it suggests a protective effect of a healthy lifestyle.
- floatingatoll 6y ago
  If taken at face value, it suggests a protective effect of low insulin resistance as measured by the TyG marker. Everything else is interpretations and chains of logical reasoning. Neither of you are particularly wrong necessarily, but a third option is that someone could be sedentary and eating carbs every day and have low TyG. It's common to state that "activity XYZ will provide action-at-a-distance medical benefit ABC" because stating accurately what's going on takes more words that sound less certain: "Paleo and keto diets may weaken Covid-19 by lowering insulin resistance" "A healthy lifestyle may weaken Covid-19 by lowering insulin resistance" But it's really worth saying it like this, even though few do. (And yes, these aren't 'maybe' enough, but they're an improvement.)
  - arkades 6y ago
    You replaced the word "suggests" with "may". You're seeing a distinction between these two words that I apparently do not, since I take them both to indicate "this is a possibility, though not the only one, and a far from proven certainty." I would add that, even there, my statement was predicated on making an inference about dietary habits - something I cast doubt on with the opening, "if taken at face value." Something I further opined on in my other post, where I pointed out that the paper may be reversing causality. I feel like you're criticizing a strong inference having been made, that I think pretty clearly wasn't.

ashtonkem 6y ago

I've suspected for a long time that Insulin resistance is going to be one of the next big areas of focus for public health, but I thought that it was going to happen the moment Apple finally figured out blood glucose measurement through the skin. I did not see a pandemic being part of it.

gumby 6y ago
I love my Apple watch but I am astonished at your faith in their ability to do noninvasive glucose measurement. People have broken their picks on that particular coal face for decades. For that matter DM2 is an area of quite active research (and, like diagnosis, has been quite active for decades) as the financial payoff for any success is enormous. The only downside for transdermal diagnosis is the lack of consumables, which makes it a tough market to enter and to be funded for. That is an area that's good for Apple as they are already selling the platform, so this would be a feature that would add to sales. And one I'd use.
- dreamcompiler 6y ago
  O2sat measurement has been noninvasive without consumables for a long time and many phones can do it now. Why do consumables matter?
  - gumby 6y ago
    Investors don’t typically like diagnostics as the margins and volume tend to be quite low. They also tend to be more “vitamin” than “aspirin”.* Consumables, a least, give you recurring revenue and even in some cases the opportunity for a razor-and-blades model. You’d be surprised how many med products are designed specifically to require consumables. * (funny analogy to use in a med context)

49para 6y ago

Insulin Resistance is the start of (all?) metabolic disease. So easy to resolve using fasting, intermittent fasting, keto, carnivore etc diets. Unfortuneately it slowly builds up over decades and only once disease has progressed do Drs move on to treat the resultant disease (and mainly with cholesterol lowering drugs). Instead of measuring fasting glucose levels (which indicate diabetes), insulin levels should be measured as they are the leading indicator.

Trasmatta 6y ago
> Insulin Resistance is the start of (all?) metabolic disease. So easy to resolve using fasting, intermittent fasting, keto, carnivore etc diets. Definitely not all. Not the least of which is Type 1 Diabetes, which is marked by the inability to produce insulin, not insulin resistance. And it's currently impossible to resolve, all Type 1 Diabetics have to be on insulin therapy to survive.
- ashtonkem 6y ago
  Is Type 1 diabetes considered a metabolic disease? I've only heard of "metabolic disease" in reference to "metabolic syndrome", the precursor to Type 2 diabetes and heart disease.
  - asplake 6y ago
    There are many metabolic disorders - some due to specific genetic issues, and some severely life-limiting (eg children unlikely to reach adulthood). And as per the Wikipedia page for metabolic disorders, not to be confused with Metabolic syndrome or Diabetes.
  - Trasmatta 6y ago
    It's a metabolic disorder. I guess I don't really know the difference between metabolic disease / metabolic syndrome / metabolic disorder. All problems of the metabolism, but I'm sure there's some medical distinction between them.
- dreamcompiler 6y ago

subsubzero 6y ago

Covid-19 is a very strange disease, it seems like it is amplified either way (severe vs. non-severe) depending on health. This seems different from the flu as the flu hits everyone very hard, the elderly/sickly especially hard. With covid some people who have it, do not and will not have any symptoms which cannot be said for the flu. I feel like the insulin marker data is an albatross, having high TyG means alot of systems in your body are not doing well, and the virus attacks weakness, it(covid-19) is also found to produce extreme clotting so that is probably why people with diabetes/heart disease and hypertension are all at high risk.

sudoaza 6y ago
Yes from what I heard having metabolic issues also is correlated with less development of antibodies after immunization.
nradov 6y ago
Asymptomatic infections are about as common with influenza as with SARS-CoV-2. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4586318/ https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4586318/ Some influenza strains hit young, healthy patients harder than the elderly by triggering a cytokine storm. This was particularly bad in the 1918 H1N1 pandemic. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4711683/ https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4711683/
Synaesthesia 6y ago
It certainly is weird. I read an article which said doctors around the world are often scratching their head at what they’re seeing. It can attack the rest of your organs, not just your lungs.
sacred_numbers 6y ago
Actually, this is not exactly true. There are actually a lot of people who are infected with the flu who are asymptomatic[1]. Based on my amateur research it does appear that the difference in severity between age groups is more pronounced for COVID-19 than for the flu, though. Part of this may be sampling bias, however, since we are probably catching far more asymptomatic COVID-19 cases than asymptomatic flu cases. [1]https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4586318/ https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4586318/

zxcvvcxz 6y ago

Fat and unhealthy people do poorly with virus. What a surprise. They should lock themselves in with a treadmill.

danans 6y ago

Given that insulin resistance is strongly correlated with obesity [1], I'm surprised that wasn't a factor they controlled for, especially since the respiratory difficulties associated with obesity seem to be a significant risk factor for death with Covid19 cases. 1. https://obesitymedicine.org/obesity-and-insulin-resistance/ https://obesitymedicine.org/obesity-and-insulin-resistance/

shakil 6y ago

Lets not confuse correlation with causation. All this study shows is people with insulin resistance are at significant risk of dying from Covid, it doesn't identify what actually kills them. However, if you look at the role Vitamin D plays [1] in suppressing cytokine storms, which is what actually pushes over an organism to the point beyond recovery from Covid, and then understand that Vitamin D deficiency is common [2] in Type 2 diabetes, you can begin to understand the fuller picture. 1. https://www.medrxiv.org/content/10.1101/2020.04.08.20058578v4 https://www.medrxiv.org/content/10.1101/2020.04.08.20058578v... 2. https://pubmed.ncbi.nlm.nih.gov/26375925/ https://pubmed.ncbi.nlm.nih.gov/26375925/