Autoantibody

What the research found.

• One survey covered 4,754 Americans aged 12 and over, tested at a single 1 to 80 dilution, and found antinuclear antibodies in 13.8 percent. The true value sits somewhere from 12.2 to 15.5. Females came out at 17.8 percent and males at 9.6 percent.

  Satoh and colleagues, Arthritis and Rheumatism, 2012

• One study screened 825 apparently healthy blood donors, of whom 130 tested positive on at least one test. Registry follow-up ran for five years. In that time only two of the 825 got a systemic autoimmune rheumatic diagnosis.

  Andraos and colleagues, Frontiers in Medicine, 2024

• In apparently healthy people, antinuclear antibody results fell steeply with the dilution used. At 1 to 40 it was 31.7 percent and at 1 to 80 it was 13.3 percent. At 1 to 160 it was 5.0 percent, and at 1 to 320 it was 3.3 percent.

  Tan and colleagues, Arthritis and Rheumatism, 1997

• One study assessed 263 patients sent to a rheumatologist for a positive antinuclear antibody, of whom 24 percent got an antibody-associated rheumatic diagnosis. Another 26.2 percent had no sign of any disease at all. A further 38.8 percent had a rheumatologic diagnosis of some other kind.

  Fitch-Rogalsky and colleagues, PLOS ONE, 2014

• One review pooled the studies of two rheumatoid arthritis antibodies. IgM rheumatoid factor caught 69 percent of the cases and cleared 85 percent of the people who didn't have it. Anti-CCP caught 67 percent and cleared 95 percent, which is why it counts for more.

  Nishimura and colleagues, Annals of Internal Medicine, 2007

What the word means

An antibody is something your immune system makes to pick out what doesn't belong, and a virus or a bacterium is the usual target. An autoantibody is one that has picked out part of you instead, so the prefix does all the work here. The finding it describes is a piece of information about your immune system rather than a diagnosis.

That difference is the whole subject of this page, and almost every misunderstanding about these tests comes from losing it. A result reporting an autoantibody says something is being recognized that shouldn't be. It doesn't say anything is being damaged, it doesn't say a disease is present, and it doesn't say one is coming.

The commonest of these tests is the antinuclear antibody, shortened to ANA, which looks for antibodies against parts of the cell nucleus. Every prevalence figure on this page describes that one test, because that's where the population data is. The general lesson applies more widely than the figures do.

How common they are in people with nothing wrong

The numbers here are why this page comes first, and they're worth reading before you read your own result. One survey covered 4,754 Americans aged 12 and over, tested at a single 1 to 80 dilution, and found antinuclear antibodies in 13.8 percent, with a true value from 12.2 to 15.5. That works out at more than 32 million people in the United States.

Females came out higher than males, at 17.8 percent against 9.6 percent, at a P value below 0.001. A separate study screened 825 consecutive apparently healthy blood donors and found 130 of them positive on at least one test, which is 15.8 percent. So two population samples gave two similar figures, and both sit far above the rate of any autoimmune disease.

So a positive result puts you in a group of tens of millions rather than in a small group of people with something wrong. That's the single most useful thing to know when a result like this comes back. It's also the thing least likely to be said in the appointment where you get it.

What happened to the healthy people who tested positive

The prevalence data tells you a positive result is common without telling you what happens next. Only one study in our sources followed that question. It's the most reassuring finding on this site and also one of the thinnest, so both halves belong here.

Of the 130 antibody-positive blood donors, 37 accepted a clinical assessment. Two women were diagnosed with Sjogren's disease, which the authors put at 1.5 percent of the antibody-positive group. Registry follow-up in that region ran for five years, and in that time only two of the 825 donors got a systemic autoimmune rheumatic diagnosis from a rheumatologist.

That's a small study, most positives turned down assessment, and five years isn't a lifetime. So it can't promise that nobody in that group develops something later on. What it argues strongly against is reading a positive result in a person with no symptoms as an early diagnosis, because living as though you're already ill is a real cost, paid for a finding that mostly goes nowhere.

Why the strength of the result changes what it means

A positive antinuclear antibody isn't one thing. The test gets run at a dilution, the result depends heavily on which dilution was used, and two people described as positive may be in quite different situations. The healthy population data makes that plain.

In apparently healthy people it was 31.7 percent at a 1 to 40 dilution, 13.3 percent at 1 to 80, 5.0 percent at 1 to 160, and 3.3 percent at 1 to 320. So nearly a third of healthy people are positive at the weakest dilution while about one in thirty are at the strongest. That study looked across four age groups from 20 to 60, and the rate didn't differ much between them.

The practical result is a question worth asking, which is at what dilution your own result was positive. A weak positive at a low dilution is common. A strong positive is much less common, and that's what gives it more weight in a clinical picture that fits it.

What happens when a positive result drives a referral

There's a study on this too, and it describes the situation many readers of this page are in. It covered 15,357 patients sent through one central rheumatology triage system, of whom 643 were sent because of a positive antinuclear antibody. Of those, 263 were assessed by a certified rheumatologist.

Of those 263 patients, 24 percent got a diagnosis of an antibody-associated rheumatic disease. That broke down into lupus at 9.1 percent, Sjogren's at another 9.1 percent, systemic sclerosis at 2.3 percent, and mixed connective tissue disease at 1.9 percent. A further 38.8 percent had a rheumatologic diagnosis of another kind, and 26.2 percent had no sign of any disease at all.

So the referral was informative for about a quarter of them in the way it was meant to be, and informative for another two fifths in a different way, and roughly one in four came out with nothing. That spread is what a positive result in an unselected person is worth. It's also why these tests belong in a clinical assessment rather than in a screening panel.

Why the tests are used at all

Nothing here makes autoantibody testing useless, and reading it that way would be the opposite error. In the right situation these results count for a great deal, and the situation is what changes their meaning. One antibody in somebody with swollen joints, a rash, or dry eyes is a different piece of information from the identical antibody in somebody who feels well.

The rheumatoid arthritis antibodies show how that plays out in numbers. One review pooled the studies, where IgM rheumatoid factor caught 69 percent of the cases and cleared 85 percent of the people who didn't have the disease, while anti-CCP caught 67 percent and cleared 95 percent. So anti-CCP counts for more, and not because it finds more cases, but because a positive result is far less likely to be a false alarm.

The other half of that arithmetic is the part people don't like. Both tests miss about a third of the people who have the disease, so a negative result rules out much less than it seems to. Seronegative disease is a recognized diagnosis, and symptoms that won't settle alongside negative antibodies are a reason to keep asking rather than a reason to stop.

What this page doesn't cover

If you've been told about a specific antibody, this page probably can't help with it. Anti-dsDNA, anti-Ro, anti-La, anti-Sm, anti-centromere, and the myositis panel all define real situations, and this site holds sources on almost none of them. Those absences are recorded in the review notes rather than filled in.

What the general lesson gives you is a way to hold whatever you've been told. Ask what the result was, ask at what strength it came back, and ask what it changes about the plan going forward. A result that changes nothing about the plan is worth understanding as just that.

If somebody has quoted you a figure for a specific antibody, asking for the reference is reasonable. These tests attract confident numbers from people who haven't checked them, and a plausible figure for a test this consequential does more damage than an admitted absence. A plausible number gets quoted onward.

Common misconceptions.

Related terms.

• ANA, the commonest autoantibody test and the one all the prevalence figures here describe.
• Rheumatoid factor, the clearest example of a test whose name oversells it.
• Anti-CCP, which carries more diagnostic weight without finding more cases.
• Lupus and blood tests, where the antinuclear antibody does its most useful work.

Questions patients ask.