NIH Peer-Reviewed Article: AMA Alone Not Diagnostic of PBC

[Kevin733]

"Only about 10% of patients who are AMA seropositive, but lack clinical features of PBC, subsequently develop PBC."

ncbi.nlm.nih.gov/pmc/articl...

[Karaliz]

Thanks for posting Kevin.....Many interesting articles!

My own state of Victoria, Australia makes an appearance in one of the studies related to the prevalence of PBC.

Karaliz

[Lizzy71 in reply to Karaliz]

I'm in Vic too and just read that there are under 100 of us in the State!!! I lived in London for a few years and also used Spatone when I was pregnant which comes from iron rich water in North Wales where PBC is more common. I wonder if either of those were the trigger? Have you traveled to the UK at all?

[Karaliz in reply to Lizzy71]

Hi Lizzy!

I have lived most of my life in Melbourne apart from travelling overseas a couple of times, including UK. As a ten year old I travelled through Europe for 12 months, including 6 months in the UK. One of the Drs I see for the OCA trial I'm on is investigating the possible link between Vit D and PBC - her PhD thesis. As PBC and many other autoimmune diseases are more common in the colder northern hemisphere countries, UK and Scandinavia included, this is a very interesting proposition I believe.

Good to hear from another Victorian!

Karaliz

[Lizzy71 in reply to Karaliz]

Hi Karaliz. Thanks for the reply and good to know someone else is out there! Yes Vit D makes sense and I'm taking it regularly. Good excuse to get a fab tan too!!

My maternal grandmother died at age 65 of liver failure from unknown origin so I think it was probably PBC. My Liver Specialist isn't that interested in causes :-). It interests me though as I have 3 little children and one is a girl so it would be good to discover a trigger. I work in the medical field so all those things interest me anyway :-).

[Karaliz in reply to Lizzy71]

Hi Lizzy

Yes, the trigger, causes, predisposition of individuals present a conundrum don't they ! I was diagnosed 13 years ago at age 38 when the youngest of my 4 children was 5 ( 2 boys and 2 girls). I was fatigued at the time but thought working as a nurse and looking after my brood were adequate explanations. No other symptoms at the time and in fact for the first 7 years almost forgot I had PBC. The wheels started to come off in 2010 with LFTs deteriorating significantly including dropping Albumin. Much imaging and another liver biopsy later revealed progression of disease - I now have cirrhosis - and decision that I was clearly no longer responding to Urso. My wonderful consultant recruited me into the OCA trial which has seen my LFTs improve over the last few years.

Long before PBC reared its head I developed Raynaud's syndrome but other than that there is very little evidence of autoimmune conditions in my family. Good to hear from you.

Karaliz

[sophottie in reply to Karaliz]

Hey Karaliz and Lizzy71,

I'm new on the scene here (in Melbourne too!) and am looking for info on a good specialist- hopefully a specialist in autoimmune liver diseases! I've got positive AMAs, elevated LFTs and just had a biopsy, so I'm waiting for a diagnosis... it's not much fun! I'm asymptomatic besides that... but I want to be informed and get a good team around me! Any suggestions?

Cheers

Sophie

[Chazzy]

Don't understand. Simple terms for me?

[Kevin733 in reply to Chazzy]

Chazzy,

It means that if you have AMA with no other symptoms, no other autoimmune conditions, and otherwise completely normal blood panels, the probability of going on to later develop PBC is only about 18% for women and 2% for men, for an overall average of 10%. This seemed relevant to your situation, so I thought I'd share.

[Chazzy in reply to Kevin733]

It is. Thank you. That's such positive news.

[Sarah4314]

Ama m2 would be different "odds" though right?

[Hidden in reply to Sarah4314]

I would say you are correct Sarah4314 - here is a link to the Mayo Clinic's 'AMA M2 serum' information - scroll down to the "Interpretation" paragraph where it states that AMA M2 is "highly specific" for PBC.

mayomedicallaboratories.com...

[Kevin733]

Chazzy, Sarah, and Dianne: This has been discussed and documented elsewhere on the site, but I'll review here. Most AMA-positive findings are for AMA-M2, and there have been papers published suggesting that a general AMA finding is as or more predictive of PBC than limiting only to the M2 subtype.

For example, this peer-reviewed paper, also from NIH, states that AMA is more accurate and more comprehensive in detecting PBC than AMA-M2:

ncbi.nlm.nih.gov/pubmed/245...

There is another important difference between the two articles: the NIH article is qualified with the condition that the patient "lacks clinical features of PBC." Because AMA isn't routinely tested for, almost everyone who gets an AMA test does so because a PBC clinical feature was detected. This, I believe, contributes to high specificity.

Also, any statistic that does not differentiate between the genders is inaccurate when applied to someone of a specific gender. The PBC gender ratio (female: male) is at least 9:1 (10:1 in the NIH article), whereas AMA is equally prevalent among men and women. From this it follows that the genders have different specificity.

Finally, note that the highest available estimate for PBC prevalence is 1 in 3000. For argument's sake, let's assume that PBC is even more prevalent, at 1 in 2000. The lowest available estimate for AMA positivity is 0.5%, which is 1 in 200. Finally, let's even assume away that there are AMA-negative cases of PBC, so that all PBC occurs in AMA-positive people. These statistics suggest only a 200 in 2000 (1 in 10, or 10%) probability that a given AMA-positive individual will go on to develop PBC. When combined with the 9:1 gender ratio above, that equates to 18% for women and 2% for men.

Regardless of the gender breakout, it's unclear why the prevalence of AMA is at least 10-fold greater than the prevalence of PBC, when it's so "highly specific" for the disease. The obvious explanation is testing bias--where only those with other indications of the disease get tested--which is why I suggested that the 10% figure applied only to those with no other PBC indicators.

Finally, let's suppose--contrary to the above statistics--that AMA is virtually diagnostic of the disease, so that (in women) it had a specificity of 99.9%. Based on the gender ratio and the fact that AMA occurs equally across gender, this would mean that its specificity in men would be only 11.1%. Assuming that equal numbers of men and women are tested (which would be necessary to avoid biased sampling), the overall specificity would then be the average of these numbers, which is 55.5%--only a little better than a coin flip.

My point is that while these high specificity statistics are out there, they do not occur in a vacuum. They do not occur when sampling the blood of gender-controlled samples who are otherwise healthy people. Therefore, they should be taken in that context. Meanwhile, we can use simple statistics to calculate more accurate PBC probabilities for otherwise healthy people. This approach is the only approach that makes sense, because healthy people are not routinely tested for AMA.