Insulin resistance calculation

Status
Not open for further replies.
Eddy Edson

Eddy Edson

Well-Known Member
Relationship to Diabetes
Type 2
For those interested, the "TyG Index" is a formula commonly used for estimating insulin resistance. In one version:

ln(fasting BG x fasting trigs)/2

with the quantities expressed in US-style mg/dl. You can easily find calculators on the Web, but here's one which conveniently allows you to enter values in UK-style mmol/l: https://www.optimaldx.com/calculators/tyg-index

along with a discussion and evidence summary: https://www.optimaldx.com/research-...ulation-biomarkers-triglyceride-glucose-index

Anything above about 4.5 is suggestive of insulin resistance.

(This is the most common form of the TyG Index used these days but actually it appears to be the result of an error transcribing the original formula from 2008, which somehow got embedded in the literature. Originally, it was

ln(fasting BG x fasting trigs/2)

With this original formula, the cut-off is around 8. Doesn't matter except for possible confusion if you see numbers quoted which use this original form. More dumbly, you sometimes see papers citing the original formula but reporting results derived from the other one.)
 
I like numbers so I got into Patient Access, looked up my last serum triglyceride result (1.5 mmol/l) and took my current 90 day waking average blood glucose (7.6 mmol/l), banged them in the equation and got a TyG of 4.9. This is above 4.5 so indicative of insulin resistance.

I then start asking questions...

.....like what would my waking blood glucose have to be to get below 4.5 with the same triglycerides. The answer is 3.5 mmol/l.

....and what would my triglycerides have to be to get below 4.5 with the same blood glucose. The answer is 0.7 mmol/l.

.... and what would trigs and BG would give me a 4.5 if both came down. The answer was trigs at 1 mmol/l and Bg at 5 mmol/l

Not sure what this means but I'm guessing nearly everybody will get a TyG of close to or above 4.5 since you get that with trigs and fasting blood glucose at what are considered normal levels.

On a wider level, the notion of insulin resistance is often talked about but I am not sure that there is a common understanding of what the term means. Don't think I could define it if asked.

Apart from those points it looks quite interesting and no doubt can be made into a good case for more research money!
 
Docb said:
I like numbers so I got into Patient Access, looked up my last serum triglyceride result (1.5 mmol/l) and took my current 90 day waking average blood glucose (7.6 mmol/l), banged them in the equation and got a TyG of 4.9. This is above 4.5 so indicative of insulin resistance.

I then start asking questions...

.....like what would my waking blood glucose have to be to get below 4.5 with the same triglycerides. The answer is 3.5 mmol/l.

....and what would my triglycerides have to be to get below 4.5 with the same blood glucose. The answer is 0.7 mmol/l.

.... and what would trigs and BG would give me a 4.5 if both came down. The answer was trigs at 1 mmol/l and Bg at 5 mmol/l

Not sure what this means but I'm guessing nearly everybody will get a TyG of close to or above 4.5 since you get that with trigs and fasting blood glucose at what are considered normal levels.

On a wider level, the notion of insulin resistance is often talked about but I am not sure that there is a common understanding of what the term means. Don't think I could define it if asked.

Apart from those points it looks quite interesting and no doubt can be made into a good case for more research money!
Click to expand...
I'm 4.0 - 4.2 in recent times.

If you actually wanted to find a definition for insulin resistance it's not very difficult, in terms of the gold standard glycemic clamp protocols. This is an expensive, risky research-only thing. The most common surrogate, validated in lots of studies, is HOMA-IR, which requires a fasting insulin measurement. This TyG Index apparently validates reasonably well against HOMA-IR and euglycemic clamp measures and is more convenient than either.

About 40% of adults in the US are insulin resistant to some extent, estimated using HOMA-IR. and I'm sure it wouldn't be hard to dig up a similar figure for the UK.

In terms of the TyG Index inputs, median non-diabetic population measures for trigs are less than 1.0 mmol/l and for fasting BG less than 5.0 mmol/l. No inconsistency there with a cut-point of around 4.5 for the TyG Index, given ~40% insulin resistance prevalence.
 
So insulin resistance is the condition where glucose is not absorbed into cells despite there being insulin present. Sounds simple but it raises all sorts of questions.

The biggest for me is why insulin levels are not measured along with blood glucose as a matter of routine. Is it because it can't be done easily or is it because the standard automated test kit is not configured to do it. Does anybody know?

Whatever the reason, it seems to me that it would be very useful to know whether a high blood glucose level was accompanied by a high or low insulin level. Would that not point to whether the high blood glucose (the diabetes) was due problems with the pancreas (high blood glucose, low insulin) or due to insulin resistance (high blood glucose, sufficient insulin)? Knowing that would lead to more accurate treatment choices more quickly would it not?
 
Docb said:
Whatever the reason, it seems to me that it would be very useful to know whether a high blood glucose level was accompanied by a high or low insulin level. Would that not point to whether the high blood glucose (the diabetes) was due problems with the pancreas (high blood glucose, low insulin) or due to insulin resistance (high blood glucose, sufficient insulin)? Knowing that would lead to more accurate treatment choices more quickly would it not?
Click to expand...
I agree. To this day I am uncertain to what extent it is low insulin or insulin resistance that is/was the cause of my high blood sugar. Using proxies like the formula above suggests I'm not insulin resistant but can't be sure.
 
Docb said:
The biggest for me is why insulin levels are not measured along with blood glucose as a matter of routine. Is it because it can't be done easily or is it because the standard automated test kit is not configured to do it. Does anybody know?

Whatever the reason, it seems to me that it would be very useful to know whether a high blood glucose level was accompanied by a high or low insulin level. Would that not point to whether the high blood glucose (the diabetes) was due problems with the pancreas (high blood glucose, low insulin) or due to insulin resistance (high blood glucose, sufficient insulin)? Knowing that would lead to more accurate treatment choices more quickly would it not?
Click to expand...
The euglycemic clamp gold-standard protocol involves infusing a fixed flow of insulin and varying flow of glucose into the subject's bod, searching for the equilibrium steady state where tissue absorption of the glucose matches the inflow, in the context of the insulin inflow. This provides a measure of insulin sensitivity/resistance.

Obviously it's not a useful method for mass screening etc.

Any mass method based on just a snapshot of insulin and BG, like HOMA-IR and its derivatives, has to incorporate a bunch of assumptions to achieve the same kind of thing, and the sensitivity and specificity delivered by these assumptions will depend on the characteristics of the population under examination.

I think that with HOMA-IR etc you can say that low/high results are good indications of absence/presence of insulin resistance, but what "low" and "high" cut points should be is an empirical, not an analytic question, and a bunch of people will always sit somewhere between them.

I think a commonly accepted general principal for any proposed screening test is that the link between the result and changes to care has to be clear. "How would this change current practice?" I think HOMA-IR etc are seen as not providing the required clarity.

On the other hand, the majority of insulin resistant people will have central adiposity. "Waist circumference < 0.5 height" provides both a screen and a target (as per the current NICE guidelines). Not just for insulin resistance, but a wide, wide range of other health conditions.
 
Study just published finds that (for the US population) the TyG Index is actually superior to HOMA-IR as a test for metabolic syndrome (which I think is the main reason you'd be interested in either of them):

​


Highlights​

  • Metabolic Syndrome (MetS), a major global problem, predisposes to both T2DM and ASCVD.
  • There is a paucity of data on the TG-glucose (TyG) index in the US population,despite it being a very reliable marker of insulin resistance in other countries .
  • We compared the TyG index to HOMA-IR as a discriminant of MetS in the US population (NHANES data)
  • It is a superior predictor of MetS than HOMA-IR in the US population and is more cost-effective.

Structured Abstract​

Background​

Metabolic Syndrome (MetS) is a cluster of cardio-metabolic features portending an increased risk for both type 2 diabetes mellitus (T2DM) and premature atherosclerotic cardiovascular disease (ASCVD). Homeostasis Model Assessment of Insulin Resistance (HOMA-IR) is a widely used surrogate measure of insulin resistance. The triglyceride-glucose (TyG) index is another validated measure of insulin resistance that predicts both diabetes and cardiovascular disease in low and medium-income countries, but only diabetes in high income countries.

Objective​

Due to the paucity of data on the TyG index in the US population, we compared the validity of the TyG index and HOMA-IR in predicting MetS.

Methods​

Data from the National Health and Nutrition Examination Survey (NHANES) between 2005 and 2018 on Non-Hispanic White(NHW), Hispanic American(HA), and African American(AA) individuals(n=5380) aged 20-80 years were used for analysis. Individuals were classified as having MetS based on three or more of its components. HOMA-IR and the TyG index were determined from fasting samples.

Results​

Both the TyG index and HOMA-IR were significantly increased in MetS and increased significantly with increasing severity of the syndrome. Also both indices correlated significantly with all 5 features of MetS, hsCRP and non-HDL-C. ROC-AUC analysis for TyG index was significantly greater than that of HOMA-IR in predicting MetS: 0.87(95% CI 0.85-0.88) versus 0.82(95% CI 0.81-0.83) respectively, p<0.0001. This was not evident for the small AA subgroup.

Conclusion​

The TyG index outperformed HOMA-IR in predicting MetS, a proxy for both T2DM and ASCVD, in a general US population and is a valuable biomarker.
 
Eddy Edson said:
Study just published finds that (for the US population) the TyG Index is actually superior to HOMA-IR as a test for metabolic syndrome (which I think is the main reason you'd be interested in either of them):

​


Highlights​

  • Metabolic Syndrome (MetS), a major global problem, predisposes to both T2DM and ASCVD.
  • There is a paucity of data on the TG-glucose (TyG) index in the US population,despite it being a very reliable marker of insulin resistance in other countries .
  • We compared the TyG index to HOMA-IR as a discriminant of MetS in the US population (NHANES data)
  • It is a superior predictor of MetS than HOMA-IR in the US population and is more cost-effective.

Structured Abstract​

Background​

Metabolic Syndrome (MetS) is a cluster of cardio-metabolic features portending an increased risk for both type 2 diabetes mellitus (T2DM) and premature atherosclerotic cardiovascular disease (ASCVD). Homeostasis Model Assessment of Insulin Resistance (HOMA-IR) is a widely used surrogate measure of insulin resistance. The triglyceride-glucose (TyG) index is another validated measure of insulin resistance that predicts both diabetes and cardiovascular disease in low and medium-income countries, but only diabetes in high income countries.

Objective​

Due to the paucity of data on the TyG index in the US population, we compared the validity of the TyG index and HOMA-IR in predicting MetS.

Methods​

Data from the National Health and Nutrition Examination Survey (NHANES) between 2005 and 2018 on Non-Hispanic White(NHW), Hispanic American(HA), and African American(AA) individuals(n=5380) aged 20-80 years were used for analysis. Individuals were classified as having MetS based on three or more of its components. HOMA-IR and the TyG index were determined from fasting samples.

Results​

Both the TyG index and HOMA-IR were significantly increased in MetS and increased significantly with increasing severity of the syndrome. Also both indices correlated significantly with all 5 features of MetS, hsCRP and non-HDL-C. ROC-AUC analysis for TyG index was significantly greater than that of HOMA-IR in predicting MetS: 0.87(95% CI 0.85-0.88) versus 0.82(95% CI 0.81-0.83) respectively, p<0.0001. This was not evident for the small AA subgroup.

Conclusion​

The TyG index outperformed HOMA-IR in predicting MetS, a proxy for both T2DM and ASCVD, in a general US population and is a valuable biomarker.
Click to expand...
1714175491836.png
 
Status
Not open for further replies.
Back
Top