Antibody tests

[mandan]

Could Antibody tests (ICA, IA2, ZnT8A, GAD, IAA) falsely positive? I know they can predict LADA 5 years before the diabetes? How reliable are they?

 

[Inka]

@mandan My consultant took my results as very reliable. There was no mention of the possibility of the results being false at all. I had my antibodies test and the consultant told me the results and that it confirmed I was Type 1. She said that two or more antibodies usually mean a person is Type 1. What did your consultant or doctor say to you?

 

[Robin]

Could Antibody tests (ICA, IA2, ZnT8A, GAD, IAA) falsely positive? I know they can predict LADA 5 years before the diabetes? How reliable are they?

I believe the GAD test can throw up odd false positives, which is why the results are usually taken in the context of other pointers to Type 1 and not on their own. Usually the tests are ordered if there is already a strong suspicion of a type 1. I don’t know about other antibody tests.
I Googled ‘False positive GAD test' to check I was right, and a list of scientific articles popped up, of which this was one.

https://onlinelibrary.wiley.com/doi/pdf/10.1111/jdi.13307

 

[Inka]

Interesting @Robin Presumably that’s why they go with two or more for Type 1. I know GAD can be positive in other conditions not just diabetes. I always think of it as more of a general antibody. The other ones are more specific - or some of them, at least. My consultant only really talked about the ones I had. At least the OP has had them all tested rather than just GAD, as some people seem to have here.

 

[mandan]

@mandan My consultant took my results as very reliable. There was no mention of the possibility of the results being false at all. I had my antibodies test and the consultant told me the results and that it confirmed I was Type 1. She said that two or more antibodies usually mean a person is Type 1. What did your consultant or doctor say to you?

Thank you for your answer. My blood sugar levels were in march between 5.3-7.1. I had middle of march onetime hypo with 3.8, which can indicate that my insulin is not working well. But maybe I was eating just too much and was going up fast and after that coming down before dinner. The highest value after breakfast was 10.8. At last examination I had lower 3.7 fasting insulin level, which was much lower than one year ago with 6.5, normal range is 3-25. My doctor think that the 10 kg weight loss in some months is too much, the 1.42 c-peptide although in normal range, but on the lower end (normal between 0.48-5.05 ng/ml). My HbA1C is still in normal level with 5.9, under pre diabetes range. But maybe in normal people LADA could be diagnosed earlier. I am not sure, what to do. My doctor would prove his presumption with antibody tests, nothing more, not at all explaining what LADA is. Taking insulin were too early for me I think, but it can progress very fast and I could get ketoacidosis or go into coma.

 

[Inka]

So you haven’t had the antibodies tests yet @mandan ? If not, I’d definitely have them. LADA is a form of Type 1 in adults and progresses more slowly. You’re correct that sometimes things can go wrong quickly as the remaining beta cells fail.

I don’t think C Peptide can give a definite answer. I’d have the antibodies test too.

 

[mandan]

Thank you for your answer.

 

[mandan]

After 3 weeks of waiting, I got the results today. All of the antibodies are in normal interval.

ICA (islet cell autoantibodies) negative titer
GAD (glutamic acid decarboxylase) <5.00 reference 0-10
IAA (autoantibodies against insulin) <0.2 reference 0.0-0.4
IA2 (protein tyrosine phosfatase) <10.0 reference 0-10
ZnT8A (zinc transporter 8) <10 reference 0-15

"positive results in more than one of the marker antibodies (GAD, Islet cell, IA2 or insulin) can be associated with the onset of autoimmune diabetes."

IA2 Antibodies - Immunology Laboratory

www.ouh.nhs.uk

 

[Spathiphyllum]

Interesting @Robin Presumably that’s why they go with two or more for Type 1. I know GAD can be positive in other conditions not just diabetes. I always think of it as more of a general antibody. The other ones are more specific - or some of them, at least. My consultant only really talked about the ones I had. At least the OP has had them all tested rather than just GAD, as some people seem to have here.

There really aren't many disorders linked to GAD autoantibodies (GADA). There's T1D, and then there's a group of neurological disorders. (See for example https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8013924/ .)

The article for which Robin provided a link was discussing the fact that, where people tested only very weakly positive for GADA-- less than 6.5 U/ml-- many of those proved to be false positives, in that, when they were retested, the result was negative.

Where people are diabetic in HbA1c terms, and have very low C-peptide results, and have strong positive GADA results, that completely confirms T1D; it is not necessary to test positive for two or more autoantibodies. And, for adults, GADA is the most likely to be positive at the time of diagnosis.

 

[Spathiphyllum]

After 3 weeks of waiting, I got the results today. All of the antibodies are in normal interval.

ICA (islet cell autoantibodies) negative titer
GAD (glutamic acid decarboxylase) <5.00 reference 0-10
IAA (autoantibodies against insulin) <0.2 reference 0.0-0.4
IA2 (protein tyrosine phosfatase) <10.0 reference 0-10
ZnT8A (zinc transporter 8) <10 reference 0-15

"positive results in more than one of the marker antibodies (GAD, Islet cell, IA2 or insulin) can be associated with the onset of autoimmune diabetes."

IA2 Antibodies - Immunology Laboratory

www.ouh.nhs.uk

Mandan, you are very lucky to have a doctor who is careful enough to give you C-peptide and autoantibody tests, just to make sure you have the right diagnosis and treatment!

 

[Inka]

Where people are diabetic in HbA1c terms, and have very low C-peptide results, and have strong positive GADA results, that completely confirms T1D; it is not necessary to test positive for two or more autoantibodies. And, for adults, GADA is the most likely to be positive at the time of diagnosis.

I’m quoting my consultant. She said that two or more antibodies confirmed Type 1. It was part of a conversation about MODY.

I’m glad you’ve got your results back @mandan I found that actually having them on paper gave me more clarity.

 

[Spathiphyllum]

I’m quoting my consultant. She said that two or more antibodies confirmed Type 1. It was part of a conversation about MODY.

Ah! That makes sense. If the possibilities are MODY or Type 1, two or more antibodies would make it extremely unlikely that you had MODY. (See for example https://pubmed.ncbi.nlm.nih.gov/21395678/ )

In general, though, from everything I've read, there is more of a problem with false negatives than false positives. Through the course of development of T1, different antibodies tend to pop up at different times and then disappear. So even if a patient tests negative for all the relevant antibodies, a diagnosis of T1 may still be correct.

In my case, my consultant only had two antibody tests done, one I can't remember (ICA, I think?) and GADA; I gather it's not uncommon to do just two, because these tests are relatively expensive. The one I can't remember was negative. For GADA, the stated normal range was 0-5, and my result was over a thousand ... Add C-peptide well under 200 pmol/L, and that was good enough-- or rather bad enough!-- for him.

 

[Inka]

I had all mine done at the same time, along with C Peptide. This was years after my diagnosis yet interestingly I was still making antibodies. Not sure if this is good or bad! Anyway, I welcomed the information.

 

[mandan]

Thank you @Inka and @Spathiphyllum for your very detailled and kind answer. I am going back to the doctor end of May and he will decide what to do next, but according to his study from last year the doctor has to see other factors as well (nutrition, insulinproduction), not just the antibodies, so his decision could be insulin as well, although the tests are negative. If I had ketoacidosis it wouldn't be sure I have T1D I could have T2D as well. In Flashbush diabetes for example where antibodies are negative, but there was ketoacidosis, if keto is over insulin could be stopped as well.

There was no blood test for me between 2008 and 2020, maybe because I am too young and healthy and I don't need any. In this time interval I could develop diabetes. I am 42 years old now, so MODY is unlikely, because it is under 30 and with 28 everything was all right. I was diagnosed with glucose blood sugar test in the summer of 2020 with 40. But the test was done not the right way so I walked and drank a lot and this is why my blood sugar level came down. The doctor was so angry in 2022, when I visited him, that I waited two more years, without doing anything. The antibodies are for 5 years in my blood, but the doctor thought for one year it is T2D, so he lost time. Now I am taking 2*500 mg Metformin, going twice a week to swim and doing low-carb diet. He decreased from 2*1000 mg Metformin the dosis in november because I lost 12 kgs, so my BMI is now 19.9 and between 18.5-25 it is normal.

 

[everydayupsanddowns]

When we looked at antibody testing as part of the NICE Type 1 in Adults Guideline Development Group (I was a lay member) I was surprised how weak the evidence was, and how the tests were nothing like as clear cut as I had believed - in either direction.

You can have antibodies and not be T1. And you can have no antibodies and still be T1 (the antibodies don’t have around after all the beta cells have been splatted).

Checking multiple antibodies helps, as does checking for antibodies close to the time of diagnosis.

And viewing them in the context of cPeptide seems helpful too, particularly further on from diagnosis (after ‘honeymoon period’).

But the evidence was not sufficient to recommend their routine use.

———-—-——--

1.1.3 Take into consideration the possibility of other diabetes subtypes and revisit the diagnosis at subsequent clinical reviews. Carry out further investigations if there is uncertainty (see recommendations 1.1.7 and 1.1.8). [2022]
1.1.4 Measure diabetes-specific autoantibodies in adults with an initial diagnosis of type 1 diabetes, taking into account that:

• the false negative rate of diabetes-specific autoantibody tests is lowest at the time of diagnosis
• the false negative rate can be reduced by carrying out quantitative tests for 2 different diabetes-specific autoantibodies (with at least 1 being positive). [2022]

1.1.5 Do not routinely measure serum C‑peptide to confirm type 1 diabetes in adults. [2022]
1.1.6 In people with a negative diabetes-specific autoantibody result, and if diabetes classification remains uncertain, consider measuring non-fasting serum C‑peptide (with a paired blood glucose). [2022]

Revisiting initial diagnosis​

1.1.7 At subsequent clinical reviews, consider using serum C‑peptide to revisit the diabetes classification if there is doubt that type 1 diabetes is the correct diagnosis. [2022]
1.1.8 Take into account that the discriminative value of serum C‑peptide to diagnose type 1 diabetes increases the longer the test is done after initial diagnosis of diabetes. [2022]

Recommendations | Type 1 diabetes in adults: diagnosis and management | Guidance | NICE

www.nice.org.uk

 

[Spathiphyllum]

When we looked at antibody testing as part of the NICE Type 1 in Adults Guideline Development Group (I was a lay member) I was surprised how weak the evidence was, and how the tests were nothing like as clear cut as I had believed - in either direction.

You can have antibodies and not be T1. And you can have no antibodies and still be T1 (the antibodies don’t have around after all the beta cells have been splatted).

Checking multiple antibodies helps, as does checking for antibodies close to the time of diagnosis.

And viewing them in the context of cPeptide seems helpful too, particularly further on from diagnosis (after ‘honeymoon period’).

But the evidence was not sufficient to recommend their routine use.

———-—-——--

1.1.3 Take into consideration the possibility of other diabetes subtypes and revisit the diagnosis at subsequent clinical reviews. Carry out further investigations if there is uncertainty (see recommendations 1.1.7 and 1.1.8). [2022]
1.1.4 Measure diabetes-specific autoantibodies in adults with an initial diagnosis of type 1 diabetes, taking into account that:

• the false negative rate of diabetes-specific autoantibody tests is lowest at the time of diagnosis
• the false negative rate can be reduced by carrying out quantitative tests for 2 different diabetes-specific autoantibodies (with at least 1 being positive). [2022]

1.1.5 Do not routinely measure serum C‑peptide to confirm type 1 diabetes in adults. [2022]
1.1.6 In people with a negative diabetes-specific autoantibody result, and if diabetes classification remains uncertain, consider measuring non-fasting serum C‑peptide (with a paired blood glucose). [2022]

Revisiting initial diagnosis​

1.1.7 At subsequent clinical reviews, consider using serum C‑peptide to revisit the diabetes classification if there is doubt that type 1 diabetes is the correct diagnosis. [2022]
1.1.8 Take into account that the discriminative value of serum C‑peptide to diagnose type 1 diabetes increases the longer the test is done after initial diagnosis of diabetes. [2022]

Recommendations | Type 1 diabetes in adults: diagnosis and management | Guidance | NICE

www.nice.org.uk

Of course you're right that "You can have antibodies and not be T1. And you can have no antibodies and still be T1". But you go on to say, rightly, that testing for the relevant antibodies, and for C-peptide, is helpful. Most T1s at diagnosis will test positive for at least one of the relevant antibodies, and most T1s at diagnosis will be showing a lower than normal level of C-peptide.

Therefore, the reason that NICE did not recommend "routine use" of testing for antibodies and for C-peptide can only have been on grounds of cost. If these tests were cheap, they would be done routinely-- because they are helpful in getting the correct diagnosis.

 

[Inka]

…(the antibodies don’t have around after all the beta cells have been splatted)…

I know I’ve asked you this before, Mike, so I’m more asking everyone else for theories:

I still had antibodies more than 25 years after diagnosis. Does this mean I still have some beta cells popping up (that my immune system is merrily killing off in a kind of diabetic Whackamole 🙄 ) or are the antibodies just sitting there waiting and all primed in case a beta cell appears (eg through an islet transplant), or are they just left over from the initial immune attack all those years ago? I did ask my consultant briefly but our conversation was covering a lot in a short time, so I didn’t really get a response as she had moved on to something else. I’ve tried googling but they all say antibodies aren’t found too long after diagnosis.

 

[rebrascora]

…(the antibodies don’t have around after all the beta cells have been splatted)…

I know I’ve asked you this before, Mike, so I’m more asking everyone else for theories:

I still had antibodies more than 25 years after diagnosis. Does this mean I still have some beta cells popping up (that my immune system is merrily killing off in a kind of diabetic Whackamole 🙄 ) or are the antibodies just sitting there waiting and all primed in case a beta cell appears (eg through an islet transplant), or are they just left over from the initial immune attack all those years ago? I did ask my consultant briefly but our conversation was covering a lot in a short time, so I didn’t really get a response as she had moved on to something else. I’ve tried googling but they all say antibodies aren’t found too long after diagnosis.

I wonder if this is another individual thing. I was a blood donor from the age of 18 up until diagnosis at 56 and apparently my blood was specifically reserved for babies because it was very low in antibodies. I am therefore guessing that my body removes antibodies quite quickly after an episode has triggered their production or halts production once an episode has been resolved. I have a feeling this may play into my diabetes diagnosis in some way but I am not sure how.
I don't know how long antibodies survive in the body, so if they are still present 25 years after diagnosis for you @Inka, is that because your body is still producing them or because the original ones haven't been broken down.... which seems unlikely after 25 years that the same proteins would be floating around in your blood stream 😱!!
I think how much insulin you need may play into this as well considering that research mentioned above where many 50yr+ diabetics were found to still have some insulin production. It makes you wonder if we have more capability to regenerate beta cells than we think but that they keep getting hit when our immune system is triggered.

 

[Inka]

That’s very interesting about your blood being low in antibodies @rebrascora I didn’t realise that was something that could happen.

Your last paragraph is mainly what I mull over: are my beta cells trying to regenerate and getting killed off by the over-keen - and very persistent - antibodies? Interestingly, for a brief time after pregnancy, when the insulin resistance went and my immune system was still on ‘pregnancy time’, I only needed small amounts of insulin. I could eat more than double my normal carbs at a meal. I didn’t even have to count. I was shovelling in mashed potato, rice, etc, and staying in range or going low. During early pregnancy too, I could eat carbs without insulin sometimes. I always wonder if that’s the dialled down immune system in pregnancy allowing some beta cells to regenerate a bit. Obviously, that wouldn’t be a cure (having to trick the immune system into thinking you’re pregnant), but it would show that the beta cells could theoretically regenerate. I find it fascinating.

 

[Spathiphyllum]

…(the antibodies don’t have around after all the beta cells have been splatted)…

I know I’ve asked you this before, Mike, so I’m more asking everyone else for theories:

I still had antibodies more than 25 years after diagnosis. Does this mean I still have some beta cells popping up (that my immune system is merrily killing off in a kind of diabetic Whackamole 🙄 ) or are the antibodies just sitting there waiting and all primed in case a beta cell appears (eg through an islet transplant), or are they just left over from the initial immune attack all those years ago? I did ask my consultant briefly but our conversation was covering a lot in a short time, so I didn’t really get a response as she had moved on to something else. I’ve tried googling but they all say antibodies aren’t found too long after diagnosis.

Does this help? https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9750828/ ...

On a quick scan-- and this article was published in October 2022-- the short answer to your questions appears to be: 'We still don't know.' ; )

This article makes clear, though, that quite a lot of people do continue to test positive for the relevant antibodies even decades after diagnosis.

Put together with another article I posted on another thread the other day-- https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6668678/ -- a study of over a thousand people who had had T1D for 50 years or more: A lot of them still had some beta-cell function.

So, if you still are showing autoantibodies, this may indeed be because you still have some beta cells.

(This was just from a quick search of PubMed; if you want to burrow into it more, PubMed is the place to burrow!)