Ankylosing spondylitis diagnosis: the delay, the HLA-B27 test and the imaging

Short answer

Ankylosing spondylitis diagnosis rests on a pattern, not a single test: back pain that behaves inflammatorily and started before 45, imaging of the sacroiliac joints, and supporting clues such as HLA-B27 and CRP. Nothing confirms it alone, which is why the pooled average wait from first symptom to diagnosis is 6.7 years (1, 5).

If you have spent years being sent to physical therapy, chiropractors and massage for a back that gets worse when you sit still, you're not imagining it and it isn't laziness or low pain tolerance. A pooled analysis of 64 studies put the average delay at 6.7 years, and it has not shortened over the decades studied (1). Women wait about 2.4 years longer than men (2). The delay is a property of the diagnostic pathway, not of you.

Note: this article explains how clinicians reach this diagnosis and is not medical advice or a treatment guide; a rheumatologist decides whether you have axial spondyloarthritis. Go to the emergency room or call 911 for loss of bladder or bowel control, numbness in the groin or inner thighs, new leg weakness, back pain after a fall or crash if you have a known fused or stiff spine, or back pain with fever, drenching night sweats or unexplained weight loss. A sudden painful red eye with blurred vision or light sensitivity needs same-day eye care (5, 7).

This page is about the route to a diagnosis. The symptoms themselves — where the pain sits, the fatigue, the stiffness, what it does over years — are covered in ankylosing spondylitis symptoms. Treatment is a separate subject and belongs with your rheumatologist.

Why does ankylosing spondylitis take years to diagnose?

Because nothing in the early years of the disease looks unusual, and the one test that was traditionally done first is the one that stays normal longest.

The numbers vary a lot depending on who is counted and how, and that variation is itself informative:

Treat the last row with care rather than quoting it as "the" number. It comes from a web survey recruited through a patient community, so people with long, difficult journeys are more likely to answer, and three-quarters of respondents were women — the group that waits longest anyway. The same study is still useful, because what it captures well is not the length of the wait but its shape: what people were told instead.

Four things stretch the pathway.

Back pain is extremely common and axial spondyloarthritis is not. Clinicians reason from a pretest probability of roughly 5% in an unselected person with chronic back pain, so the default assumption is mechanical (10). The job of the early workup is to raise or lower that 5%, not to confirm anything.

The X-ray lags the disease by years. Changes in the sacroiliac joints can take seven to 10 years to become visible on a plain film (7). A normal pelvic X-ray in a 26-year-old is close to meaningless, yet it is often filed as a negative result and the question is closed.

The first stop is rarely a rheumatologist. In the US survey, 87.2% had seen a general practitioner and 65.1% a rheumatologist; the people most likely to make this diagnosis were seen by only two-thirds of patients (14). NIAMS notes plainly that there is no single test for ankylosing spondylitis and that diagnosis may need a team (15).

Something plausible gets written down instead. In that survey, men were most often told they had a back problem (55.7%), a psychosomatic issue (23.0%) or sciatica (21.3%); women were most often told psychosomatic issue (40.8%), back problem (40.2%) or anxiety and depression (23.6%) (14). Each of those labels ends the search.

One more counterintuitive finding from the meta-analysis: delay was longer in high-income countries than in middle-income ones, and longer in people with no extra-articular involvement — no uveitis, no psoriasis, no inflammatory bowel disease (1). A back that is only a back is harder to place than a back that comes with an inflamed eye.

What makes back pain inflammatory rather than mechanical?

Its behaviour over time, not its location or its severity. This is the single filter that decides whether anyone orders the rest of the workup.

An international group, ASAS, tested which features best matched expert judgement of inflammatory back pain and published five, each with its odds ratio (4):

• Improvement with exercise — 23.1

• Pain at night (with improvement on getting up) — 20.4

• Insidious onset — it crept in, no injury — 12.7

• Age at onset under 40 — 9.9

• No improvement with rest — 7.7

Four of those five present counts as inflammatory back pain. In the validation cohort of 648 people this had 79.6% sensitivity and 72.4% specificity (4).

Side by side with the ordinary mechanical pattern:

Two honest limits. First, inflammatory back pain screens; it does not diagnose. In referral studies only 17–33% of at-risk patients who met the inflammatory pattern turned out to have axial spondyloarthritis (5). Second, people with the condition do not always fit the pattern, and someone who has been avoiding exercise for years because of pain may genuinely not know whether movement helps.

This is also where other causes of lower back pain get separated out, and where a clinician starts looking for company: psoriasis, inflammatory bowel disease, a past episode of reactive arthritis, swollen fingers or toes as in psoriatic arthritis, heel pain, or an episode of acute anterior uveitis.

What does a positive HLA-B27 test actually mean?

That you carry a common gene variant. It shifts the odds; it does not make a diagnosis, and this is the point most often lost in a hurried appointment.

HLA-B27 is present in 6.1% of US adults aged 20–69 (124 of 2,320 tested in NHANES 2009), and it is not evenly distributed: 7.5% of non-Hispanic white adults, 4.6% of Mexican American adults, 1.1% of non-Hispanic Black adults (6). In raw numbers that is millions of Americans who carry it and will never develop spondylitis.

From the other direction, the gene is common in people who do have the disease: over 95% of white patients with ankylosing spondylitis test positive, around 80% of Mediterranean patients and around 50% of African American patients, by the Spondylitis Association of America's figures (7). A review of the diagnostic literature puts it at 70–90% of white AS patients (5).

Put those two together and you get the number that matters. Estimates of how many carriers go on to develop the disease run from "no more than 2%" (7) to about 5% of HLA-B27-positive white individuals (5). Either way, the overwhelming majority of positive results belong to people who are fine. The test is useful precisely because it is being ordered in someone who already has years of inflammatory back pain — that is, in a group where the pretest probability is nothing like the general population's.

Three practical consequences:

• A positive result is not a diagnosis. It is one SpA feature among several, and reviews state outright that "HLA-B27 positivity alone is not diagnostic" (5).

• A negative result does not rule it out. Roughly one in 10 to one in 20 white patients with AS are HLA-B27 negative, and the proportion is far higher in Black patients (5, 7). A negative HLA-B27 in someone with four of five inflammatory features and sacroiliitis on MRI changes very little.

• Ordering it as a screen is the wrong use. Tested at random, it generates far more false alarms than findings. Tested in someone with a years-long inflammatory pattern, it meaningfully moves the estimate.

If a result comes back positive and you are told to see a rheumatologist, that is the system working. If it comes back positive and you are discharged with the gene as the explanation for everything, ask what the imaging showed.

X-ray or MRI: which picture finds sacroiliitis?

Both look at the sacroiliac joints, where the base of the spine meets the pelvis, but they answer different questions. The X-ray shows damage that has already happened. The MRI shows inflammation happening now. In a disease that spends its first decade inflaming and its second decade damaging, the order you do them in decides how long you wait.

Plain X-ray of the pelvis. This is what the 1984 modified New York criteria were built on, and it is what the word "ankylosing spondylitis" historically meant: definite sacroiliitis visible on a radiograph, plus clinical findings (8). An X-ray is cheap and available everywhere, and it is the right test for showing erosion, sclerosis, narrowing and fusion. Its weaknesses are well documented: poor sensitivity in people with short symptom duration, and substantial disagreement between readers, especially when the only finding is sclerosis or a small erosion (5, 10). And the lag — up to seven to 10 years before changes show (7).

MRI of the sacroiliac joints. The protocol includes semicoronal and axial images with a T1 sequence and a water-sensitive sequence (STIR or fat-suppressed T2) (5). What radiologists look for is bone marrow edema — fluid in the bone next to the joint, the footprint of active inflammation — which appears before any structural damage. The ASAS MRI working group's 2016 definition of a positive scan is specific: two or more bone marrow edema lesions on the same slice, or one lesion in the same quadrant on two or more consecutive slices (9).

The MRI's weakness is the mirror image of the X-ray's. Bone marrow edema is not unique to spondyloarthritis. It turns up in mechanical back pain, in healthy volunteers, in athletes doing heavy training, and after childbirth (5, 10). An MRI report that mentions edema is not a diagnosis either — it is read together with everything else.

What changed in 2026. EULAR's updated imaging recommendations, presented in June 2026 and the first revision since 2015, move MRI of the sacroiliac joints ahead of radiography as the recommended first-line test in suspected axial spondyloarthritis. Where MRI is unavailable or contraindicated, radiography or, preferably, low-dose CT is used instead (12, 13). As of October 2026 this has been presented at a conference rather than published as a peer-reviewed guideline, and US practice and insurance rules will not change overnight. But it is a reasonable thing to raise if your workup stopped at a normal X-ray.

Radiographic and non-radiographic: the same disease, two pictures

This is the piece of vocabulary that confuses almost everyone, and it matters for insurance letters and for how seriously you are taken.

• Radiographic axial spondyloarthritis (r-axSpA) is what has long been called ankylosing spondylitis: definite sacroiliitis visible on X-ray (10).

• Non-radiographic axial spondyloarthritis (nr-axSpA) is the same clinical disease without those X-ray changes. The MRI may show sacroiliitis, or the diagnosis may rest on the clinical picture plus HLA-B27 (5, 10).

"Non-radiographic" means the X-ray is clean, not that the disease is mild. Disease activity, quality of life and response to treatment are broadly similar in the two groups (5). Somewhere between 5% and 25% of people with nr-axSpA develop radiographic changes within two to eight years, and up to 28% after 10 years or more (5). The current ACR, Spondylitis Association of America and SPARTAN recommendations cover both forms with near-identical advice (11).

What can blood tests say, and what can't they?

Less than most people expect. There is no blood test for ankylosing spondylitis.

CRP and ESR are general inflammation markers. They are raised in roughly 30–40% of people with axial spondyloarthritis, meaning up to 60% have active symptoms with entirely normal values (10). The Spondylitis Association of America puts the ESR figure at fewer than 70% of AS patients, and notes that anemia, infection and cancer raise it too (7). In the diagnostic weighting used by reviewers, a raised acute-phase reactant carries a likelihood ratio of only about 2.5 (5). A normal CRP does not close the question, and a raised one does not open it on its own.

Rheumatoid factor and antinuclear antibodies are not associated with this disease (7). A negative rheumatoid factor is sometimes handed back as reassurance; it is not relevant here. If rheumatoid arthritis is the alternative being considered, its own trajectory is described in the four stages of rheumatoid arthritis.

How do the criteria fit together — and why aren't they a diagnosis?

The ASAS classification criteria, published in 2009, are the framework most clinicians have in their head. The entry condition is back pain lasting three months or more with onset before age 45. From there, two routes (3):

• Imaging arm: sacroiliitis on X-ray or MRI, plus at least one SpA feature.

• Clinical arm: HLA-B27 positive, plus at least two other SpA features.

SpA features include inflammatory back pain, arthritis, enthesitis (notably heel), uveitis, dactylitis, psoriasis, Crohn's disease or ulcerative colitis, good response to NSAIDs, a family history of spondyloarthritis, and raised CRP. Overall the criteria performed at 82.9% sensitivity and 84.4% specificity; the imaging arm alone was 66.2% sensitive and 97.3% specific (3).

Here is the part worth knowing before you meet them in a report: these are classification criteria. They were built to make research cohorts comparable, and specialists warn explicitly against using them as diagnostic criteria, because doing so produces misdiagnosis in both directions (10). A real diagnosis is a rheumatologist weighing the whole picture, including how well everything else has been excluded.

That specificity problem is now being addressed. At EULAR 2026, ASAS and SPARTAN presented revised criteria built from the international CLASSIC study of 1,015 people referred with undiagnosed back pain, comparing each candidate item against the rheumatologist's own diagnosis. Sacroiliac MRI had the strongest association with diagnosis; the revised set keeps SI joint MRI, HLA-B27, raised CRP, and a tighter clinical list — inflammatory back pain, inflammatory bowel disease, acute anterior uveitis, heel enthesitis and psoriasis — and reached 79.5% sensitivity and 90.4% specificity in validation (12). Again: presented, not yet published, and still criteria for classification rather than a diagnostic checklist.

Why do women wait longer for a diagnosis?

Partly because the evidence that moves clinicians fastest is the evidence women are least likely to have early, and partly because of what gets written on the chart instead.

The size of the gap is well established: 8.5 years versus 6.1 years, a difference of 2.4 years, in 5,555 patients across 27 countries (2). Age at symptom onset is the same in men and women, or slightly younger in women — so the gap is in the pathway, not the disease (5).

Four contributors show up repeatedly:

• The X-ray finds less. Men show more severe radiographic progression; women are over-represented in the non-radiographic group, where the sex ratio is much closer to even than the 2–3:1 seen across spondyloarthritis as a whole (5, 10). If the pathway stops at a pelvic X-ray, it stops on women more often.

• The presentation is read as atypical. Women more often report enthesitis, pelvic pain, heel pain and widespread pain, and may start in the neck rather than the low back (5, 7).

• Widespread pain gets a different label. Women were significantly more likely than men to be told they had fibromyalgia — 20.7% versus 6.6% (14). The two can coexist, but one diagnosis given early tends to stop the search for the other; how fibromyalgia itself is assessed is covered in how to test for fibromyalgia.

• The symptoms get attributed to the mind. "Psychosomatic issues" was the single most common misdiagnosis reported by women in the US survey, at 40.8% (14).

None of this means a woman with back pain should assume spondyloarthritis. It means that if the inflammatory pattern is there and the X-ray was clean, "nothing on imaging" is an incomplete stopping point, and an MRI of the sacroiliac joints is a reasonable thing to ask about.

Living with an undiagnosed back: the questions people ask next

Which doctor actually makes this diagnosis?

A rheumatologist. The Spondylitis Association of America names the rheumatologist as the physician who commonly makes the diagnosis, and NIAMS adds that a primary care doctor usually coordinates between specialists (7, 15). A physical therapist, chiropractor or sports medicine doctor can be genuinely helpful for the pain and cannot diagnose this. If a referral has not been offered and you meet four of the five inflammatory features, ask for one directly.

Should I stop exercising before the MRI so it shows more?

No, and do not change anything about your care to influence a scan. It is worth knowing, though, that heavy training can produce bone marrow edema in healthy people, which is one reason radiologists read scans alongside the clinical picture rather than in isolation (5). Tell the clinician ordering the MRI what your training load actually is, so it can be factored in.

My X-ray was normal — is the question settled?

Not if the inflammatory pattern is there. Sacroiliac changes can take seven to 10 years to appear on a plain film, and a normal X-ray is exactly what early disease looks like (7). This is the whole reason non-radiographic axial spondyloarthritis exists as a category, and the reason 2026 imaging guidance puts MRI first (13).

Nobody in my family has it. Does that count against the diagnosis?

Only slightly. Family history is one SpA feature out of several, and plenty of people are the first in their family to be diagnosed — in part because older relatives with the same back went undiagnosed for decades (3).

How long does the workup itself take?

The tests are quick — a blood draw and one or two imaging appointments. The waiting is in referral queues and insurance authorisation for the MRI. Bringing a dated, written record of the pattern to the first appointment is the one part of the timeline you can shorten yourself.

How to bring this up with your doctor

What to bring. A written, dated record is worth more than a recalled summary, because the diagnosis turns on the pattern over time and that is precisely what blurs when you describe it in a 12-minute appointment. Note: the time you wake with pain, how long the morning stiffness lasts in minutes, what happens on a day with a long drive or a long flight versus a day with a long walk, and any episode of a red painful eye, a swollen finger or toe, heel pain, psoriasis or gut symptoms.

What to say. "My back pain started gradually when I was [age], it has lasted more than three months, it wakes me in the second half of the night, I'm stiff for [N] minutes every morning, it gets better when I move and worse when I sit still. That's the inflammatory pattern. Can we look at axial spondyloarthritis?"

What to ask for. "Can we check HLA-B27 and CRP?" "Can I have imaging of the sacroiliac joints — and if the X-ray is normal, can we do an MRI?" "Can I have a rheumatology referral?" Asking for the sacroiliac joints by name matters: a lumbar spine X-ray is a different picture and will not answer this question.

If you're dismissed. Ask for the reasoning to be written in your notes: "Please document that we considered inflammatory back pain and the reason we are not referring." That request is usually enough to reopen the conversation, and it gives you something to take to a second opinion. The diagnostic delay data is a legitimate thing to cite.

When it is urgent. Go to the emergency room or call 911 for loss of bladder or bowel control, numbness in the groin or inner thighs, new weakness in a leg, back pain following a fall or crash if your spine is known to be stiff or fused, or back pain with fever, drenching night sweats or unexplained weight loss. A sudden painful red eye with blurred vision or sensitivity to light needs same-day care from an ophthalmologist — acute anterior uveitis is associated with this condition and is treated urgently (5, 7).

Start understanding your body

Learn what affects your energy, stress, sleep, and daily state. Get the app.

Explore your data

How Welltory helps — and what it cannot do

The limits, first and plainly. Welltory is a general wellness product, not a medical device. It does not diagnose, detect, predict, monitor, prevent or treat any condition. It cannot find ankylosing spondylitis or axial spondyloarthritis, it cannot see inflammation, it cannot read a sacroiliac joint, and nothing in the app substitutes for HLA-B27, an X-ray, an MRI or a rheumatologist. Welltory's own help centre also notes that measurements may not be accurate if you are under 18 or pregnant (16).

What it can do is hold the shape of the thing. What separates inflammatory from mechanical back pain is not how much it hurts but its pattern in time: worse after rest, better with movement, waking you at night, stiff for more than 30 minutes in the morning (4). That pattern is exactly what collapses into "it's been bad for a while" when you are retelling it in an appointment — and exactly what survives in a dated log.

You note what happened; the app records the rest the same way every day. Add a tag from the plus sign at the top of the feed — "woke 4am, back", "stiff 50 min", "desk all day", "walked 5km, better". When the Today screen marks a stress or rest stretch it asks "What happened?"; pick a suggested tag, type or talk, and correct the tags before saving. Alongside that, resting heart rate, heart rate variability, stress minutes, Battery and sleep are recorded identically each day, so a bad week is measured on the same scale as a good one rather than from memory (Battery and stress minutes need iOS with an Apple Watch or Oura; detailed sleep analysis needs an Apple Watch).

Over a few months, My Patterns shows what your tags have in common — which days cluster with poor nights or low Battery. It runs on iOS with an Apple Watch or Oura, and patterns start to appear at around seven tagged events in a month, so a couple of bad nights will not produce one. Anything it surfaces is an association to discuss, not a cause and not a finding.

A record you can hand over at an appointment. The Journal (Premium) keeps your measurements, tags and notes together, and the web app exports a CSV (Dashboard → choose a chart → Export). Six weeks of dated entries showing night waking, morning stiffness in minutes and the rest-versus-movement swing answers a rheumatologist's first three questions faster than any recollection can — and it is the kind of thing that gets a clean X-ray reopened.

The red flags do not change, whatever your data shows. Loss of bladder or bowel control, numbness in the groin or inner thighs, new leg weakness, back pain after a fall or crash with a stiff or fused spine, or back pain with fever, night sweats or unexplained weight loss means emergency care. A sudden painful red eye with blurred vision needs same-day eye care (5, 7).

How we made it

Made with AI tools, then edited and fact-checked by the Welltory team. See our Editorial & AI policy.

Jane Smorodnikova, co-founder of Welltory, built the methodology behind how we read physiological data.

Written by Tatsiana Yashyna.

Discounts for blog readers: up to 36% off

See what affects your energy, stress, sleep, and daily state with Welltory

Explore your data