Covid/Coronavirus and diabetes - the numbers

[Lupf]

Many people with diabetes have reacted to yesterday's news that 25% of people who died from Covid-19 in UK hospitals also had diabetes. This is both understandable and scary, but it is not really unexpected.

Let me try to explain. I am T2, a scientist and I have been following what the epidemiologists and statisticians are learning about Covid. For example we are being told the daily death tolls. According to this 34,000 people have died in the UK from Covid by today 15 May. However, this number includes the deaths in hospitals, but not all deaths in care homes and at home. This number can be calculated by comparing the total number of deaths in the UK with the long-term average and sadly by 1 May there have already been 50,000 excess deaths in the UK

. These deaths are mostly due to Covid.

A few days ago I found a paper by the group of Ben Goldacre, https://www.medrxiv.org/content/10.1101/2020.05.06.20092999v1. It studies the in-hospital deaths by potential risk factors. For diabetes they find the following. For the group with "controlled Diabetes", defined as HbA1c<58 mmol/mol) the risk of death was between 1.40 and 1.60 times that of non-diabetics.
For the group with "uncontrolled diabetes" defined as HbA1c>=58 mmol/mol the risk of death was between 2.18 and 2.56 times that of non-diabetics. For the group with no recent Hb1Ac measure the risk was between 1.63 and 2.16 that of non-diabetics.
In summary, depending on the severity of our diabetes the risk of death is a factor 1.5 to 2.5 higher than for non-diabetics.

In the UK almost 4 million people or 6% of the population are diagnosed with diabetes. While diabetes is increasing in younger people, we know that the rate is higher for elderly people and is probably at least twice that high (12%) in the over 65 age group. We now need to multiply this number with the 1.5 to 2.5 and can estimate that between 18% and 30% of people who died from Covid also had diabetes. This is in agreement with yesterdays announcement of 25%.

If you have other questions on the science of Covid, please feel free to ask on this thread and I'll try to answer these.

update 20 May: In this thread in post #96 I discuss the latest paper by the group of Jonathan Valabhji, see the direct link Covid/Coronavirus and diabetes - the numbers

 

[bulkbiker]

Many people with diabetes have reacted to yesterday's news that 25% of people who died from Covid-19 in UK hospitals also had diabetes. This is both understandable and scary, but it is not really unexpected.

Let me try to explain. I am T2, a scientist and I have been following what the epidemiologists and statisticians are learning about Covid. For example we are being told the daily death tolls. According to this 34,000 people have died in the UK from Covid by today 15 May. However, this number includes the deaths in hospitals, but not all deaths in care homes and at home. This number can be calculated by comparing the total number of deaths in the UK with the long-term average and sadly by 1 May there have already been 50,000 excess deaths in the UK

. These deaths are mostly due to Covid.

A few days ago I found a paper by the group of Ben Goldacre, https://www.medrxiv.org/content/10.1101/2020.05.06.20092999v1. It studies the in-hospital deaths by potential risk factors. For diabetes they find the following. For the group with "controlled Diabetes", defined as HbA1c<58 mmol/mol) the risk of death was between 1.40 and 1.60 times that of non-diabetics.
For the group with "uncontrolled diabetes" defined as HbA1c>=58 mmol/mol the risk of death was between 2.18 and 2.56 times that of non-diabetics. For the group with no recent Hb1Ac measure the risk was between 1.63 and 2.16 that of non-diabetics.
In summary, depending on the severity of our diabetes the risk of death is a factor 1.5 to 2.5 higher than for non-diabetics.

In the UK almost 4 million people or 6% of the population are diagnosed with diabetes. While diabetes is increasing in younger people, we know that the rate is higher for elderly people and is probably at least twice that high (12%) in the over 65 age group. We now need to multiply this number with the 1.5 to 2.5 and can estimate that between 18% and 30% of people who died from Covid also had diabetes. This is in agreement with yesterdays announcement of 25%.

If you have other questions on the science of Covid, please feel free to ask on this thread and I'll try to answer these.

What about those with an HbA1c lower than the general population? I'm guessing zero data... but if you have seen anything?

 

[Tophat1900]

In the UK almost 4 million people or 6% of the population are diagnosed with diabetes. While diabetes is increasing in younger people, we know that the rate is higher for elderly people and is probably at least twice that high (12%) in the over 65 age group. We now need to multiply this number with the 1.5 to 2.5 and can estimate that between 18% and 30% of people who died from Covid also had diabetes. This is in agreement with yesterdays announcement of 25%.

This makes sense when the age group is applied. Context is everything.

 

[Bill_St]

https://www.cell.com/cell-metabolism/fulltext/S1550-4131(20)30238-2?
Looks at BG rather than HbA1c.

 

[copilost]

Worth pointing out that the HbA1c data comes from Primary care (i.e. local clinic not in hospital admission data). So this is an indication of diabetic control prior to hospital admission and most likely prior to infection with covid-19.

 

[urbanracer]

What about those with an HbA1c lower than the general population? I'm guessing zero data... but if you have seen anything?

It's a fascinating subject and in view of the research paper that @Mike d linked to in the Covid thread which discusses the prevalence of increased insulin resistance in the BAME community I have taken a quick look at the immune system and glucose requirements.

It would seem that the immune system is 'glucose hungry' and I presume that we may need glucose for energy to fight the invaders. But if you can't use insulin efficiently then where does that leave you?

Similarly, if one's a1c is very low, would a person have enough glucose to mount an efficient immune response? With our old friends the mice, glucose was “required for survival in models of viral inflammation, it was lethal in models of bacterial inflammation."

Some bedtime reading.

https://www.google.com/amp/s/amp.theatlantic.com/amp/article/498965/

https://www.scientificamerican.com/...e-response-in-the-flu-and-possibly-covid-191/

 

[kev-w]

https://abcd.care/relationship-between-diabetes-and-covid-19 gives a 'little info' (little as in not very much at all) so it's a waiting game as the data is still being 'looked at'.

 

[Lupf]

In the UK almost 4 million people or 6% of the population are diagnosed with diabetes. While diabetes is increasing in younger people, we know that the rate is higher for elderly people and is probably at least twice that high (12%) in the over 65 age group. We now need to multiply this number with the 1.5 to 2.5 and can estimate that between 18% and 30% of people who died from Covid also had diabetes. This is in agreement with yesterdays announcement of 25%.

Update:
@Brunneria posted the age distribution of diabetes at https://www.diabetes.co.uk/forum/th...ity-with-diabetes.171962/page-91#post-2262093. thanks.
Form this you can see that about a quarter or 1 million of diabetics are in each of the 60 to 69 and 70 to 79 age brackets.
From the population pyramid https://www.ons.gov.uk/peoplepopula...les/ukpopulationpyramidinteractive/2020-01-08
I find that 11% or over 7 million people in the UK are in the 60 to 69 age bracket. One in seven or 14% of these have diabetes.
Similarly almost 18% of the 66.4 million or close to 12 million people are over 65 and about 2 million that is one in six or 17% have diabetes.

Using these numbers and the elevated Covid risk factor for diabetics and elderly it is even less surprising that one quarter of people who have died also had diabetes.

 

[Lupf]

What about those with an HbA1c lower than the general population? I'm guessing zero data... but if you have seen anything?

I have not seen anything on lower Hb1Ac, so can't comment.

Important I am only explaining results and numbers, and each of us is an individual,
so please talk to your GP about your personal case.
For example. these studies mainly apply to T2. This is since about 9 out of 10 of us are T2.
However if you are T1, please be careful on how to interpret these results.

In the paper by the Ben Goldacre group you can find the elevated risk factors for many diseases and conditions. It confirms that age is the main factor, compared to the 50 to 60 age brackets, the risks for the 60 to 69 year olds is a factor of 2 higher. For the 70 to 79 year old the risk is a factor of 5 higher and for the over 80 year old it is more than a factor of 10 higher.

Relevant for diabetics is obesity. For a BMI > 30 the risk is 20% to 40% higher,
for a BMI over 35 it is 40% to 70% higher
and for a BMI > 40 the death risk increases by a factor to 2.6

 

[bulkbiker]

so please talk to your GP about your personal case.

My GP has less of a clue than the man in the street concerning these things I'm afraid..

I just wondered if anyone had seen anything.

Personally I'm very pleased with my low HbA1c and am pretty confident that it would be a major plus in fighting off any virus that hits.

 

[Honeyend]

Talks about diabetes and covid

 

[jjraak]

Many people with diabetes have reacted to yesterday's news that 25% of people who died from Covid-19 in UK hospitals also had diabetes. This is both understandable and scary, but it is not really unexpected.

Let me try to explain. I am T2, a scientist and I have been following what the epidemiologists and statisticians are learning about Covid. For example we are being told the daily death tolls. According to this 34,000 people have died in the UK from Covid by today 15 May. However, this number includes the deaths in hospitals, but not all deaths in care homes and at home. This number can be calculated by comparing the total number of deaths in the UK with the long-term average and sadly by 1 May there have already been 50,000 excess deaths in the UK

. These deaths are mostly due to Covid.

A few days ago I found a paper by the group of Ben Goldacre, https://www.medrxiv.org/content/10.1101/2020.05.06.20092999v1. It studies the in-hospital deaths by potential risk factors. For diabetes they find the following. For the group with "controlled Diabetes", defined as HbA1c<58 mmol/mol) the risk of death was between 1.40 and 1.60 times that of non-diabetics.
For the group with "uncontrolled diabetes" defined as HbA1c>=58 mmol/mol the risk of death was between 2.18 and 2.56 times that of non-diabetics. For the group with no recent Hb1Ac measure the risk was between 1.63 and 2.16 that of non-diabetics.
In summary, depending on the severity of our diabetes the risk of death is a factor 1.5 to 2.5 higher than for non-diabetics.

In the UK almost 4 million people or 6% of the population are diagnosed with diabetes. While diabetes is increasing in younger people, we know that the rate is higher for elderly people and is probably at least twice that high (12%) in the over 65 age group. We now need to multiply this number with the 1.5 to 2.5 and can estimate that between 18% and 30% of people who died from Covid also had diabetes. This is in agreement with yesterdays announcement of 25%.

If you have other questions on the science of Covid, please feel free to ask on this thread and I'll try to answer these.

Thank you @Lupf.

Most interesting.

Persinally think the better control does have a positive impact.
So worth aiming for, as most on here do.

So Cheers for the post.

 

[NicoleC1971]

I have not seen anything on lower Hb1Ac, so can't comment.

Important I am only explaining results and numbers, and each of us is an individual,
so please talk to your GP about your personal case.
For example. these studies mainly apply to T2. This is since about 9 out of 10 of us are T2.
However if you are T1, please be careful on how to interpret these results.

In the paper by the Ben Goldacre group you can find the elevated risk factors for many diseases and conditions. It confirms that age is the main factor, compared to the 50 to 60 age brackets, the risks for the 60 to 69 year olds is a factor of 2 higher. For the 70 to 79 year old the risk is a factor of 5 higher and for the over 80 year old it is more than a factor of 10 higher.

Relevant for diabetics is obesity. For a BMI > 30 the risk is 20% to 40% higher,
for a BMI over 35 it is 40% to 70% higher
and for a BMI > 40 the death risk increases by a factor to 2.6

I am a type 1 and assume I have elevated risk because my hba1c is teetering around 58 and hyperglecemia is a risk factor independent of age, hyperinsulinemia and age. Luckily I am under 50, female and otherwise healthy so other than taking vitamin D I am not taking more stringent measures to avoid getting this.
Btw your original post seemed to infer that diabetes is a marker of other CoVid risks e.g. age and obesity however I think more research will be needed to break down the causal mechanisms for each individual risk factor e.g. having high blood sugars is known to reduce the efficacy of our immune systems as does having excess insulin.
So if you are a well controlled diabetic this reduces one risk factor but if your good sugar levels are achieved by having excess insulin then I imagine you would still be more at risk.

It's a fascinating subject and in view of the research paper that @Mike d linked to in the Covid thread which discusses the prevalence of increased insulin resistance in the BAME community I have taken a quick look at the immune system and glucose requirements.

It would seem that the immune system is 'glucose hungry' and I presume that we may need glucose for energy to fight the invaders. But if you can't use insulin efficiently then where does that leave you?

Similarly, if one's a1c is very low, would a person have enough glucose to mount an efficient immune response? With our old friends the mice, glucose was “required for survival in models of viral inflammation, it was lethal in models of bacterial inflammation."

Some bedtime reading.

https://www.google.com/amp/s/amp.theatlantic.com/amp/article/498965/

https://www.scientificamerican.com/...e-response-in-the-flu-and-possibly-covid-191/

Of Mice and glucose needed for immune response, surely our saviour in times of glucose need is the liver i.e. gluconeogenesis therefore there's no such thing as too low unless you have a rare metabolic condition?

 

[jrussell88]

Well done. But surely the statisticians who have access to NHS data should be disaggregating the numbers by now?

Figures out of China were published on 24 January, and similar to this.

Can we not do better?

 

[Lupf]

Btw your original post seemed to infer that diabetes is a marker of other CoVid risks e.g. age and obesity however I think more research will be needed to break down the causal mechanisms for each individual risk factor e.g. having high blood sugars is known to reduce the efficacy of our immune systems as does having excess insulin.

Thanks for this comment. That was not intended, all I wanted to point out is that obesity, which affects many diabetics, has a higher risk. These are correlations, not markers. A proper evaluation of multiple risks needs to take these correlations into account. As far as I understand the causal mechanisms for age factors are not yet known.

 

[DavidGrahamJones]

Many people with diabetes have reacted to yesterday's news that 25% of people who died from Covid-19 in UK hospitals also had diabetes.

Does it say they only have diabetes or do they have other medical issues. I'm thinking aloud and not trying to make any real point. I'm assuming some people with diabetes are surviving Covid. Some might hardly have noticed anything other than a bad cold/flu type thing, others might have spent some time in hospital and others are sadly passing away. There must be something else going on with their health?

 

[urbanracer]

Does it say they only have diabetes or do they have other medical issues. I'm thinking aloud and not trying to make any real point. I'm assuming some people with diabetes are surviving Covid. Some might hardly have noticed anything other than a bad cold/flu type thing, others might have spent some time in hospital and others are sadly passing away. There must be something else going on with their health?

From what I have seen reported, it seems like many diabetics are dying from DKA due to C-19 causing massively elevated glucose levels.

And from reports in yesterday's press that doctors are being advised to monitor glucose levels closely, I am left "guessing" that they haven't been routinely doing it up to this point?

 

[copilost]

Does it say they only have diabetes or do they have other medical issues.

The paper reports adjusted hazard. That means if you compare two people who are same age, sex, BMI, etc etc for everything on the list in the table then the hazard reported for diabetes is the additional hazard regardless of all the other factors. Someone who looks exactly like you (for the listed factors) and doesn't have diabetes has less risk. Of course it may not be diabetes! The list doesn't include vitamin D status and it could be that diabetes is a proxy indicator for vitamin D status and that is the real risk (unlikely but it's an example of the uncertainty in these types of analysis). Also the analysis doesn't cover interactions i.e. what is the risk if you are both diabetic and have hypertension for example? You can't simply add the two figures together as this will probably overestimate the risk as it's likely they will share some of the risk. People will start to calculate these combined risks but you need a large sample size to get a useful result.

 

[Jamie H]

The paper reports adjusted hazard. That means if you compare two people who are same age, sex, BMI, etc etc for everything on the list in the table then the hazard reported for diabetes is the additional hazard regardless of all the other factors. Someone who looks exactly like you (for the listed factors) and doesn't have diabetes has less risk. Of course it may not be diabetes! The list doesn't include vitamin D status and it could be that diabetes is a proxy indicator for vitamin D status and that is the real risk (unlikely but it's an example of the uncertainty in these types of analysis). Also the analysis doesn't cover interactions i.e. what is the risk if you are both diabetic and have hypertension for example? You can't simply add the two figures together as this will probably overestimate the risk as it's likely they will share some of the risk. People will start to calculate these combined risks but you need a large sample size to get a useful result.

If you are referring to the NHS figures released on Friday, then it doesn't report adjusted hazard and its been quoted several times that people may have other comorbidities... This is what diabetes uk and partha kar are investigating so it can be broken down more... One area is also other existing comorbidities.

 

[Jamie H]

The paper reports adjusted hazard. That means if you compare two people who are same age, sex, BMI, etc etc for everything on the list in the table then the hazard reported for diabetes is the additional hazard regardless of all the other factors. Someone who looks exactly like you (for the listed factors) and doesn't have diabetes has less risk. Of course it may not be diabetes! The list doesn't include vitamin D status and it could be that diabetes is a proxy indicator for vitamin D status and that is the real risk (unlikely but it's an example of the uncertainty in these types of analysis). Also the analysis doesn't cover interactions i.e. what is the risk if you are both diabetic and have hypertension for example? You can't simply add the two figures together as this will probably overestimate the risk as it's likely they will share some of the risk. People will start to calculate these combined risks but you need a large sample size to get a useful result.

If you're referring to the NHS figures released on Friday there has been no hazard adjustment. That is what has caused the uproar and panic. Nothing about age, sex, BMI, control etc. All the areas required to perform hazard adjustment. They are just raw figures.