Hey everyone, it's Mikki here. You're listening to Mini-Mikkipedia on a Monday. Today I want to chat about blood sugar, and more specifically, the utility of HbA1c, which is a marker of long-term blood glucose control, that is often run as a general health screen, if you like.

So if you get your annual health check, or you've got, I don't know, your doctor's checking other things, it's one of the more common markers.

So, haemoglobin A1c is, as the name suggests, haemoglobin that has a glucose molecule permanently attached to it.

Haemoglobin A is the main oxygen-carrying protein in red blood cells, and when glucose circulates in the blood, some of it spontaneously binds to that protein.

And this is glycation. Glycation is non-enzymatic, meaning that no enzyme controls it, and there's no regulatory break.

It just happens at a rate proportional to how much glucose is around, and how long the cell is exposed to it.

Now, red cells live for approximately 100 to 120 days, and you can't repair or reverse a glycation.

So a red cell is essentially a passive sort of recording device, if you like.

It accumulates this glucose exposure until it is cleared.

If you measure the whole population of circulating cells, and you get a weighted average of glucose exposure,

this is roughly over the preceding 8 to 12 weeks, as we understand it.

Now look, weighted is a key word.

The population contains cells of every age, these are red blood cells,

and the youngest ones are the most numerous contributors to the recent change,

which is why about 50% of the movement toward a new age.

which is why about 50% of the movement toward a new age.

which is why about 50% of the movement toward a new age.

a new value happens within the first month rather than being spread across the actual three months.

Now this is something to be mindful of because anything that affects the red cell lifespan

will alter HbA1c without glucose actually changing it at all. So if you have a shorter red cell

lifespan you can have a falsely low HbA1c which means those cells have cleared before they finish

glycating. And examples of this would be hemolysis of any cause so autoimmune, mechanical i.e.

athletes and running hemolysis is one example of this. Some hereditary conditions may also cause

this. Also hemoglobinopathies such as sickle trait, thalassemia trait which are both common

enough in a population like New Zealand with quite a high pacific and asian population.

Recipients for HbA1c.

Recipients for HbA1c.

Recipients for HbA1c.

Recent or ongoing blood loss such as heavy menstrual bleeding,

loss of blood through the GI tract, blood donation, a recent blood transfusion, advanced liver disease

and starting iron treatment in someone who is already iron deficient. So these are all examples

of of what could create these low HbA1c's where these cells are cleared before they finish

glycating. Of course on the flip side you can also have a falsely high HbA1c.

Of course on the flip side you can also have a falsely high HbA1c.

Where these cells hang around and they accumulate glucose. Examples of this are untreated iron

deficiency anemia, B12 and folate deficiency, splenectomy or functional aspenia or aplastic

anemia. So for athletes therefore you can have either of those actually going on. So as I

mentioned you've got that foot strike hemolysis and it's a classic one where you get that mechanical

destruction of red cells in the feet during running. With the additional contribution from

oxidative stress and catecholamine driven fragility. So you get this net effect as you get

a faster turnover of red blood cells and a lower A1c than what the true glucose exposure would

warrant. However iron deficiency is also very common particularly in female athletes and that

will push it the other way. So you can have a runner with mild foot strike hemolysis and low

ferritin whose two distortions may partly cancel each other out. And so you just

wouldn't know where that is coming from from the A1c number alone. So the practical consequence

of this is that for a high mileage athlete HbA1c may be one of your less reliable sort of indicators

whereas something like a fasting glucose, a fasting insulin and a cgm window will tell you

a lot more about their blood glucose levels over time. Of course if you're a cyclist or a swimmer

you have less of that mechanical hemolysis. Then

the effect isn't as uniform for these populations. And iron deficiency so untreated iron deficiency

will raise A1c. There was one case control comparison found that the mean HbA1c of 6.18%

versus 5.49% in iron deficient versus non-anemic adults despite comparable fasting glucose. So

in New Zealand terms that's roughly a difference of about eight millimoles from iron status alone.

With 6.1 being above 42 mmol per mole and 5.49 being below that sort of 39 mmol which is

getting quite close to that diabetes or pre-diabetes cut off if you believe in pre-diabetes.

And then of course the reverse happens on treatment so if you start someone on iron they mount this

increase in red blood cells and the circulating red cell population gets a whole lot younger and

A1c will fall. So a woman with heavy periods, low ferritin and an A1c of 42 millimoles per mole

who gets her iron sorted might see 37 millimoles about three months later without changing a single

thing about her diet. So if nobody knows about the iron then that may get attributed to a nutrition

intervention she might not even need. So it is worth when you're doing bloods have a look at

everything like this across the board. So I just wanted to bring up a couple of these points before

I dig into what I do want to chat about or I guess the meat of this episode is you know how useful or

what are some critiques around HbA1c that that we have to consider if you're using it as a marker

given how ubiquitous it is as a screening tool in doctor's office and I don't say I'm not saying

that's a bad thing at all but I just think this is super interesting. So you know you get your HbA1c

right and you get your A1c right and you get your A1c right and you get your A1c right and you get

approximately it might change every 12 weeks although as I said change occurs mostly in that

first month which is why you may see shifts in your A1c across a six-week period but in New Zealand

anywhere from about 30 to 39 is considered sort of normal if you like in New Zealand and then

at an A1c HbA1c of 41 this is when we would say pre-diabetes starts. So at

40 or 39 completely normal at 41 you've got pre-diabetes. Now in US units they equate to

sort of 5.7 percent at that 39 level and then it's 5.9 percent is at that HbA1c 41. Now pre-diabetes

kicks in from a WHO perspective at about 42 millimoles per mole or six percent and once you

hit an A1c you're going to get a lot of HbA1c. So you're going to get a lot of HbA1c and you're going to get a lot of HbA1c.

A1c of 50 that's when you are diagnosed with diabetes in New Zealand and that's 6.7 percent

on US units and in fact it actually kicks in earlier by the American Diabetes Association

and World Health Organization suggesting that an HbA1c at 48 millimoles or 6.5 percent is when

is enough to have you diagnosed with diabetes. So those are sort of the things that we're going to

talk about. So if you go to your doctor and they see 39 millimoles per mole they're just gonna they're

not really going to comment on that whereas if you go to your doctor at 42 millimoles per mole

they're gonna think okay you're in that pre-diabetes range we need to do something about this.

Hopefully you understand or know that I'm talking about type 2 diabetes here not type 1 diabetes.

So a couple of points about A1c which is why I'm chatting about this. The first thing is it's

actually one of the largest types of diabetes in the world. So if you go to your doctor and they see 39 millimoles per mole they're just going to they're not really going to comment on that.

So there is a huge study called Whitehall 2 and they've tracked people backwards from diagnosis

and found that insulin sensitivity declined steeply over the five years before a diagnosis

of type 2 diabetes and beta cell function rose between years four and three and then fell from

85 to 62 percent while fasting glucose stayed on the slow linear track until it had this complete

rise starting only in the late 90s. So if you go to your doctor and they see 39 millimoles per mole

23 years out so your glucose signal is the tail end of a process of insulin resistance that may have

been running for years and there's a Swedish cohort who found a similar thing. So fasting insulin

HOMA 2IR and HOMA 2B which are both markers of insulin resistance they differed from controls

up to 20 years before diagnosis. 20 years! By the time someone meets diagnostic criteria

beta cell function is already down 50 percent declining a like a further four percent per year

so a normal HbA1c is compatible with a decade of compensated insulin resistance and that's pretty

important particularly as you know through periods of our life we are more likely to be insulin

resistant such as perimenopause for example. The second reason is that your HbA1c may

miss a lot of dysglycemia that an oral glucose tolerance test might catch. So in a systematic

review of diagnostic accuracy HbA1c above 6.5 percent against an oral glucose tolerance test

as a reference had a sensitivity of just 50 percent and a specificity of 97.3 percent. So it's highly

specific but it's not that sensitive so it can miss a lot of people and screening study across

three data sets found that the HbA1c criteria missed 70 percent of people with diabetes and

82 to 94 percent of those with pre-diabetes who would diagnose that by an oral glucose tolerance

test which is basically when you get a standard dose of glucose and your glucose response to that

measured over a two-hour period. In Chinese adolescents and young adults only 28 percent

of those with oral glucose tolerance tests defined pre-diabetes had an elevated HbA1c and

interestingly that two-hour value of the OGTT also carries this independent information right so

in close to 200,000 Chinese adults it remained predictive of cardiovascular disease, cancer and

diabetes in models already containing both fasting glucose and HbA1c. So it definitely is a lagging

measure and it's not particularly sensitive. Now I mentioned those in the US units like I said

HbA1c over 6.5 percent and just to remind you that equates to a New Zealand equivalent of that

48 millimoles which is still considered pre-diabetes in New Zealand. The third reason

HbA1c might not be the best in and of itself is that an average HbA1c will conceal what that

actually looks like over time. So two people at 40 millimoles per mole can have really different

glucose behavior and the shape appears to matter independently. So when they do analyses on

continuous glucose monitored data and this was a study of 8,000 people without diabetes they found

that three components the mean, the variation, the variance and all the other components of the

autocorrelation each independently associated with arterial thickness and hepatic steiosis and

stiffness. HbA1c captures only the first so it doesn't do a great job of assessing complete risk

and further in a coronary plaque study CGM derived indices reflecting both variability and

autocorrelation predicted necrotic core percentage better than fasting glucose.

HbA1c or two-hour oral glucose tolerance test for glucose because those three primarily reflect

just that main component alone and also as I mentioned even though we think about it such

as a three-month average modeling of red cell turnover only shows 50% of change toward a new

value. Also as previously mentioned we think about it as a three-month average but actually

50% of change is not a new value. So we think about it as a three-month average but actually

50% of the change toward a new value is reached in that first 30 days. So it's heavily weighted

the most recent month. So the fact that your A1c is sort of gives you an average across 12 weeks

tells you nothing about the control of that glucose over time and that does actually matter.

Now interestingly as well there are changes to HbA1c that aren't always driven by glucose

and of 60 HbA1c associated genetic

variants. 22 of these are erythrocytic rather than glycemic meaning that the erythrocytic genetic

score which is important doesn't predict diabetes at all. So between two people with identical HbA1c

can actually differ in true average glucose by more than about 3.3 millimoles per liter

and that is like a significant difference. So number five is about the measurement noise. So

under controlled conditions the HbA1c can actually differ in true average glucose by more than about

3.3 millimoles per liter and that is like a significant difference. So under controlled conditions

within person sort of changes is about 2.5%. This does mean that a change from 40 to 43 millimoles per

mole might actually be nothing. So in a real world situation it's worse right. So the largest study

ever done on this which was close to 600,000 people in the UK found that within individual

had a variation of 0.2 which is about three times higher than the controlled estimate.

It's rising with the mean HbA1c. So it means that modeling the combined effect of the different

assays used any imprecision and other non-glucose biological variation means that you could get an

HbA1c that is out by 1 to 19 percent depending on how much of that variation is non-glycemic

and that's that is really important actually. And that brings me to my next point really is that

it's much better to have a HbA1c than a HbA1c. So if you have a HbA1c and you have a HbA1c

for someone to evaluate their HbA1c according to their historic values themselves and not against

population averages per se. Particularly when we're talking about people who may be sitting

within this sort of normal reference range and it might be slowly creeping up. You want to evaluate

your own number according to what it has been like in the past because there is this individuality

of HbA1c. So population averages are not the same as the average. So if you have a HbA1c for someone

heart-offs can be considered a little bit like a blunt instrument. So your trend in this space

really matters. And if your HbA1c is rising along with other markers like something in your

cholesterol panel say your triglycerides are increasing or your waist circumference is

increasing or let's say your A1c is increasing alongside decreasing fitness or something like

that then that's something that you want to pay attention to. So if you have a HbA1c that is

all of your ability for figures and measurements including anti-f finale doubles and other aspects

of your body that other people have any ability for at least these data to be able to digamos and

attention to. So much like body mass index, actually, HbA1c isn't ethnically neutral.

So at equivalent glucose levels, like fasting glucose levels, A1c ran 5.5 millimoles per mole

higher in Pacific. They run two and a half times higher in Maori and 2.3 times higher in

Asian participants compared to European. So there is considerably different health risk

depending on your ethnic background. My eighth point around HbA1c, just to consider if you're

looking at it as a marker, is that it doesn't actually tell you what problem you have, which

is fine. You'll just need other tests, right? So two people at 45 millimoles per mole, which would

put them in that pre-diabetic category, can have complete different physiology. Analyses of at-risk

individuals identify groups with reduced beta cell function, high visceral and liver fat,

with a high blood pressure, and with a high blood pressure. So there's a lot of different

studies out there that show that people with high insulin resistance and high visceral fat with

high insulin secretion and their trajectories to diabetes does differ. So an intervention that

helps one isn't necessarily the intervention that may help another person. And just a single number

can't direct that. And that's why getting other markers of metabolic health is clearly important.

And this goes with point nine, I guess, as well, is that the Whitehall 2 study,

finding that I chatted about earlier, is that as a risk factor of itself, it may predict sort

of cardiovascular events. But if you remove other risk factors, such as dyslipidemia, such as waist

circumference, such as low physical fitness, then it actually does lose its significance as a risk

factor at that population level. So I just think that's pretty interesting. However, a couple of

things that I do need to mention that sort of speak to its usefulness is that in stable coronary

artery disease, HbA1c, rather than fasting glucose, improved cardiovascular outcome prediction

beyond traditional risk factors. And that optimal cutoff was about 5.85%, which by New Zealand

standards is about 40 millimoles per mole. And in another study,

in another meta-analysis, the relative risk for coronary heart disease was 1.2 per 1% higher HbA1c

versus 1.06 per 1 millimole per litre of higher fasting glucose. And in Whitehall 2,

prediabetes defined by HbA1c did carry a worse prognosis than prediabetes defined by fasting

or 2R glucose. So you can see that it's not all pointing towards HbA1c,

being a useful marker, clearly, because it is one of the most commonly run markers,

but in of itself may not give you a complete picture. I think one of the reasons why it's

so popular is that you don't have to be in a fasted state to do it. If you get it tested,

it's cheap, it's pretty reproducible at a population level, and it doesn't require

anyone to sit in the doctor's office for two hours and have 50 to 75 grams of glucose.

So it is a screening test that for an individual would be useful alongside fasting glucose,

maybe fasting insulin, that would be amazing actually, but also consider weight to height

ratio and triglyceride to HDL ratio. And so all of these markers will tell you a little bit more

about cardiometabolic health than just HbA1c alone. And what I would also say to this,

however, is that if you're at your doctor's office,

and you get an HbA1c of 40 or 41, and the doctor's not giving you advice to try to lower

your HbA1c, I think that's a red flag actually, if I'm honest. I think HbA1c, despite the critique,

is with other markers, a useful indicator, certainly a high HbA1c, unless of course,

you've got some of those situations that I mentioned at the top of this little mini

If you're getting your HbA1c measured, which you probably will be because it's cheap and it's

non-invasive and you can do it at any time, make sure you're also getting a lipid panel done in a

fasted state. You know your waist to height ratio, and you have an idea of your own sort of lifestyle

and habits, because that'll tell you how meaningful that number is. Anyway, this was what was on my

mind, so I thought I'd do a little mini on it. Hope you found it interesting, helpful, I don't know,

insightful. Let me know. I'm over on Instagram threads and X,

at Mikki Williden, Facebook, at Mikki Williden Nutrition, or head to my website,

mikkiwilliden.com, and book a one-on-one call with me there, or sign up to my recipe portal

access. I literally just put three recipes in today. It was ginger beef, delicious, a baked

date oats, and also chicken mince handhelds, these little wraps that I did, which are sort of like

burgers, but sort of not. They're delicious, and they are all in that recipe portal, along with

a thousand other recipes that are

macro-friendly, high-protein, appropriate carbohydrate. 99% of them have nutritional

information attached to them as well. All right, guys, you have the best week. See you later.

Bye-bye.