Utility of common investigations for suspected inflammatory arthritis in adults

[Music] Welcome to the Australian Prescriber Podcast. An independent, no-nonsense podcast for busy health professionals.

Investigations and particularly lab tests can be very important in rheumatology, but can definitely also be confusing. There's a lot of nuance to appreciate and wrong turns that can be made. I'm David Liew, your host for today and, as a rheumatologist myself, I know how difficult this can be. Thankfully, there's a new version of the rheumatology therapeutic guidelines in press and on top of that, 2 of the rheumatologists on the expert group advising the new version have written an article entitled Utility of common investigations for suspected inflammatory arthritis in adults. Walking us through some of the testing that's required is Nicola Cook, a rheumatologist from Perth. Welcome Nicola.

Thank you, David. Pleased to be here.

Talk me through your everyday diagnosing inflammatory arthritis when a new patient comes to see you. How do you take all the different clues that you have from the symptomatology and the rest of the history and all the other investigations and examination that might go along with that?

We start with the letter from the GP and often it does include a number of blood tests, but most importantly obviously we are talking to the patient and getting [a] history of suggestive symptoms of inflammatory arthritis or enlisting symptoms of other conditions that could be confused with this, whether it's nonspecific musculoskeletal pain, osteoarthritis, degenerative joint disease or other conditions such as fibromyalgia. So if symptoms are relatively acute or subacute, if there's prolonged morning stiffness, if the swelling in the joints, if there's symmetrical, small joint arthritis and so on. These are all things which are going to point more towards an inflammatory condition and then of course we examine the patients.

So by the time that we've taken a good history and done our examination, then generally that's 90% of what we need for making the diagnosis and then we turn to the investigations. So some of these will have been done by the GP and some we may order extra ones. And for some of the patients, the actual reason for referral may not be their symptoms but actually may an abnormal blood test which has turned up incidentally while looking for other things. Having done our history and examination, we then try and put the results of the blood tests that we have in front of us into the context as to what's relevant and what may be unhelpful or false positives.

And I guess that gets to the heart of this really and that feels like there are very few perfect laboratory investigations in rheumatology if there are any at all.

Yeah, well that's why really I think in rheumatology, more than any other specialty, we rely so much on the patient's history and the examination findings. A lot of so-called ‘rheumatological screens’ can be quite easily ordered as a very broad sweep to look for things that may be relatively nonspecific symptoms, and we have to always consider what the pretest probability is of the patient having a particular disease and therefore this will influence the likelihood of that blood test being positive, being of relevance because as you say, there is no such thing as a perfect blood test. We know that the majority of so-called normal ranges apply to approximately 95% of healthy people, so 5% of healthy people are going to have values outside of that range. The more blood tests you do, the greater the likelihood that you'll pick up some tests which are outside of range but normal for that individual. So that's where we need to look at our pretest probability for the condition and work out whether the positive result in that context is going to be useful in contributing to firming up a diagnosis or excluding it.

One of the problems is with the increasing use of Dr. Google that people often come to the doctor requesting investigations and knowing some of the things that they would like to have checked. And it's much quicker to just write a form than it is to explain what the implications of doing this test would be and what the implications of finding a positive result would be in the context of their lack of symptoms of that particular condition. I think that sometimes we're all guilty of setting up panels of investigations for certain symptoms or conditions and I suspect that there's a lot of GPs doing a so-called ‘rheumatological screen’ which would cover rheumatoid factor, anti-CCP [anti-cyclic citrullinated peptide antibodies], ANA [antinuclear antibodies], HLA-B27 [human leucocyte antigen B27], inflammatory indices and so on.

Now whilst these individual tests can be very helpful in certain scenarios, doing them all as a screen for a possible rheumatological condition is not generally helpful or logical. So, for example, if a patient presents with inflammatory sounding back pain, there really is no point in doing a rheumatoid factor or anti-CCP because it would be extremely unusual for patient with rheumatoid arthritis (for which those tests are useful) to present with back pain inflammatory or otherwise. And conversely, if somebody presents with a small joint inflammatory arthritis, it's probably not necessary to do an HLA-B27 as one of your first tests.

So it's appropriate to tailor the investigations to the nature of the presentation. And if a patient presents with nonspecific symptoms such as fatigue in the absence of anything more specific, doing a blood test such as an ANA can get you into a bit of difficulty. Because if you're doing it as a means to reassure the patient that there is nothing going on and [they are one of] the 1 in 20 people who have a positive ANA but are otherwise healthy, the patient is not going to be reassured when that blood test comes back positive. We just have to take a bit of time to explain the utility of the test and the pros and cons of them.

Let's talk through how you might use HLA-B27 in terms of the pretest probability and how it might be more or less useful in different situations. For example, like you said in someone who doesn't necessarily have back pain, HLA-B27 probably isn't particularly useful. Can you talk us through why that's a case then how it might be more useful in different situations?

Well, there are people who present with [peripheral] inflammatory arthritis for example, say psoriatic arthritis where there is an increased incidence of HLA-B27. It might be say 30 to 50% if they've just got a peripheral arthritis and no axial spondyloarthropathy, but I think that would not be the best first test. You probably need to look for other causes of peripheral arthritis in the first instance.

But if the patient comes with inflammatory sounding back pain, so again, pain which is worse in the morning perhaps through the night, low back into the buttocks, alternating buttock pain improved by exercise. Then if they've got investigations, for example they've already got plain X-ray changes of sacroiliitis. To be perfectly honest, the HLA-B27 is not going to contribute anything extra to your diagnosis because your pretest probability is so high that you've already essentially made your diagnosis.

However, if you have a patient who has clearly got noninflammatory back pain, they've got degenerative changes on their X-rays, they don't give any symptoms that sound inflammatory, then it [also] would be unhelpful to do the HLA-B27 because [in] 10% of the [healthy] population it will come back positive and then this is going to confound your diagnostic process and create greater anxiety in the patient and possibly the doctor as well. So those are the 2 extremes, a very low pretest probability in a very high pretest probability.

I think the most useful place for the HLA-B27 is where you've got the intermediate pretest probability where there's some inflammatory symptoms but they're not obvious. Maybe inflammatory indices are a little bit raised. Maybe they haven't got any changes on their X-ray because it's early on in the process, but they're young. So your diagnosis is not clean cut, but there I think the HLA-B27 test is going to be more useful to you because it may sway you one way or the other as to how likely this is to be an emerging inflammatory spondylitis.

That's a beautiful example. Let me ask you though, I think a lot of people would say, well, we have really busy clinical practice and we are trying to capture whatever we can. Is there any real harm in over investigation? Are we really going to see difficulties that come from it?

Well, I think yes you can because where the patient, or even the doctor, is looking for reassurance and so the pretest probability is low and then the test comes back positive, it's counterproductive. It doesn't provide reassurance, it does the opposite. It quite possibly can lead to further unwarranted tests which again have the risk of being positive or sending [you] down further rabbit holes. It may lead to referrals to specialists, which is an expense and contributes to the long waiting lists that we have in most states where it can take 6 months or so to get in [to see] a rheumatologist. A lot of these patients are referred just because of an abnormal blood test rather than suspicious symptoms. And then there's a cost not only to the patient but to Medicare [and] to the taxpayer.

Well that seems like a litany of good reasons to try and avoid over investigation. How would we best use rheumatoid factor in our patients?

Well, we have the rheumatoid factor and we have the anti-CCP antibody. And we know that rheumatoid factor is helpful in seropositive rheumatoid arthritis defined by positive rheumatoid factor, which is the case in perhaps 70% of patients with rheumatoid arthritis. Some 30% remain seronegative throughout. The higher the titre, the more likely it is to be significant and low titres are not diagnostic – under 30 [international units/L] is a nonspecific finding in most cases. Its specificity is relatively low in the order of 70% and it can be positive in a number of other conditions including interstitial lung disease and some chronic infections and so on. So, it's not a particularly specific test. It's still useful.

The anti-CCP is a more specific test. Its sensitivity is more than 90% and it may even be positive before the patient develops their symptoms of arthritis and a positive anti-CCP can be associated with relatively early erosive disease and therefore once a diagnosis of rheumatoid arthritis has been made in the presence of a positive anti-CCP, then it's prudent to treat them somewhat more aggressively as a prognosis is less good.

But even an anti-CCP is not 100% specific. If the anti-CCP is positive but the patient does not have symptoms or signs of inflammatory joint disease, we can't make a diagnosis of rheumatoid [arthritis] in that patient and it would not justify starting treatment. So if the anti-CCP is found to be positive in the absence of clinical features of rheumatoid arthritis, there is a possibility that that patient could go on to develop rheumatoid [arthritis]. So it's prudent to monitor them clinically, caution them about symptoms that they should report and that way you're going to pick it up early but not before it's actually there.

What about testing in monitoring? What about serial testing that sometimes you see happen?

These are really tests to help with diagnosis and once the patient has been diagnosed with rheumatoid [arthritis], they are not useful for monitoring. So serial testing is not helpful. There's no evidence that the rheumatoid factor or anti-CCP level goes up and down with disease activity. It's useful prognostically to have a rheumatoid factor and an anti-CCP documented. If they’re negative and you are quite sure your patient has rheumatoid arthritis, maybe you might like to check them down the track infrequently to see if they have seroconverted. However, it's probably not going to change your treatment, to be honest, but certainly if they are positive and you've made your diagnosis, there is no benefit whatsoever in repeating them serially.

And I suppose that some of the other tests that we're going to be talking about, there are a few of these, which if they're positive then it's a one and done. There's no reason to repeat. I guess ANA probably falls into that category as well.

If it's positive and you think the patient has an ANA-associated disease, such as lupus, then again it's not a useful monitoring test because the titre does not go up and down with disease activity. And even if the patient's in remission, it's not likely to go away. Anti-[ds]DNA is a bit different. That is more likely to change with disease activity and has some prognostic utility. If it rises in association with clinical activity, it lends support that you need to be doing something a bit more to control the disease. And if it rises and there isn't increased disease activity clinically, it's probably prudent to be monitoring the patient a bit more carefully. Not necessarily to be treating them just because the anti-[ds]DNA has gone up, but at least to be keeping a more careful eye on them.

So knowing that [anti-]dsDNA might be something that we use for monitoring but not ANA. Tell me about how you use ANA in practise to help make a diagnosis.

The antinuclear antibodies, it's a relatively nonspecific antibody and you do get quite a lot of false positives. Depending on the titre and the cutoff that you use, approximately 5% of healthy people have a positive test at a titre of 1 in 160. And this will tend to be higher in older people and in those who've got a history of autoimmune disease. It's also quite frequently positive in other autoimmune diseases such as say thyroid disease. So it's relatively nonspecific, but it is useful – certainly 95% at least of patients who do have lupus will have a positive ANA often at high titre.

And if they have the clinical scenario, clinical features consistent with a connective tissue disease, then with a positive ANA, it's prudent to go on and check their ENA (extractable nuclear antibodies) and their anti-[ds]DNA because these are more specific tests for specific connective tissue diseases. However, the likelihood of the ENA and anti-[ds]DNA being positive in the scenario of a negative antinuclear antibody is low. So unless there are very strong clinical indicators, in most cases if you have screened with an ANA and it's negative, then it's usually not necessary to go on to do the ENA and anti-[ds]DNA. So again, ordering them all upfront as a panel is probably not a good use of resources.

So I guess ANA, good rule out test, good negative predictive value, but it doesn't sound like it's the kind of thing that we want to put too much weight on just on the basis of a positive result. What about ordering the ANA in the first place? Should we be careful about how we order ANAs and whom we order ANAs in? There are other things that can give you a positive ANA. What about healthy populations? Do people have positive ANAs there as well?

Absolutely. Generally at lower titres. But when you are using your ANA for the so-called worried well with nonspecific complaints such as fatigue, really your pretest probability of a significant connective tissue disorder or lupus is really quite low. And yet you may end up with this person being [one of the] 1 in 20 people who are healthy but have a positive ANA and thereby probably reinforce their medical anxiety.

So let's talk a little bit about the ENA. You've mentioned these extractable nuclear antigens. How do we use them? How are they useful once you've got a positive ANA? How do they help to change our thinking?

In the appropriate clinical context, the ENA are more specifically related to connective tissue disorders. For example, the anti-centromere antibody is commonly seen in limited systemic sclerosis. The anti-Scl70 is frequently seen in association with [diffuse systemic sclerosis]. The anti-RNP is frequently seen in association with mixed connective tissue diseases. Their specificity is pretty high. Their sensitivity is not always that high. So the absence of these antibodies doesn't exclude these conditions, but it certainly firms up the diagnosis when you are clinically suspecting one of these specific connective tissue disorders.

So why don't we move on to ANCA testing, antineutrophil cytoplasmic antibodies. How would we use that and how good a test is that in diagnosis, in monitoring, where are its strengths and weaknesses?

Again, I think that we see quite a lot of ANCA tests being done as part of the rheumatological screen, which is probably not appropriate because it is a test which is very useful in the context of a small vessel vasculitis. But the features of a vasculitis are unlikely to be nonspecific or certainly not the only features. So if the patient has a suggestive rash, if they have glomerulonephritis, if they have interstitial lung problems, then these are obviously clinical contexts where vasculitis could be considered and the ANCA antibody is useful in these scenarios. But it should not form part of a general first tranche of tests in somebody with nonspecific symptoms or indeed somebody with inflammatory arthritis as their major presenting feature. I would not be requesting an ANCA [test] in those scenarios because ANCA-associated vasculitis is really quite rare and the features are usually quite specific.

Obviously there's a bit to navigate there and great lessons for us to take away. Nicola, thank you so much for joining us on the program today.

You're welcome. Thank you, David.

[Music]

Nicola Cook was a member of the Expert Group for Therapeutic Guidelines for Rheumatology Version 4. And I am on the Drug Utilisation Subcommittee of the Pharmaceutical Benefits Advisory Committee. I'm David Liew and thanks once again for joining us on the Australian Prescriber Podcast.