New NICE high sensitivity troponin guidance: 3 hours and done?

NICE high sensitivity troponin

The National Institute for Health and Care Excellence (NICE) from the UK has just published new recommendations for the use of high sensitivity troponin to rule out NSTEMI in the Emergency Department.  You can find it at this link.  I was honoured to be part of the team looking at this, as a specialist advisor.  Here are some of my thoughts on what NICE says and the implications for our practice.

[DDET The Bottom Line]

You may just want to know the bottom line.  If so, here it is..

You can rule out NSTEMI if you use a high sensitivity assay (Roche troponin T or Abbott ARCHITECT high sensitivity troponin I) providing that the levels are normal (below the 99th percentile) on arrival and 3 hours later.

NICE high sensitivity troponin summary

If that’s all you wanted to know, you can stop right there.  It’s that simple!  If you want to know a bit more about what that actually means, and if you want to understand the important caveats, keep reading and, of course, check out our podcast (which gives even more information) at the end!

[/DDET]

[DDET Does this guideline tell me what to do with patients who have chest pain?]

No.  This isn’t a NICE clinical guideline.  It’s simply a recommendation about how we can use certain diagnostic tests cost-effectively in the UK.  That means the scope’s fairly narrow.  The guidance only applies if your hospital uses one of two troponin assays:

  • The Roche high sensitivity troponin T (cut-off 14ng/L)
  • The Abbott high sensitivity troponin I (cut-off 26ng/L)

It’s also about ruling out (not ruling in) NSTEMI early on.  Because it’s a UK guideline and we have the 4-hour ED target, NICE considered that an ‘early rule out’ strategy would be one that gets patients home within 4 hours of arrival.

THIS IS

[/DDET]

[DDET Caveat 1: What if my hospital uses another troponin assay?]

There are lots of other troponin assays out there.  When NICE looked at the characteristics of each assay, only two could reasonably be classed as ‘high sensitivity’.  The others can still be used in practice but these recommendations don’t apply.  If you want to follow what NICE says and you’re using those assays, you still need to run tests at 10-12 hours from symptom onset because looking at the evidence for those assays just wasn’t part of this project.

You should also bear in mind that NICE didn’t look at anything other than troponin – so this guidance doesn’t cover things like copeptin or H-FABP.  It doesn’t cover the new decision rules like MACS, HEART, the ADAPT protocol or the ED-ACS score.

[/DDET]

[DDET Caveat 2: What about unstable angina?]

These recommendations are about ruling out NSTEMI.  That’s right – it means that unstable angina hasn’t been considered.  Troponin can’t help you to diagnose unstable angina because patients with that diagnosis have, by definition, no rise and/or fall of troponin.  This means that you still need to do some thinking.   Not everyone who has normal troponin levels at 3 hours can go home.  If the patients are high risk, you should still be admitting them to hospital.  You could leave this up to the clinical judgement of the treating doctor, but a better idea might be to use some kind of  formal risk stratification – whether that be the TIMI score, Goldman score, etc.

Troponin and unstable angina

[/DDET]

[DDET Caveat 3: How will this work with the 4-hour target in the UK?]

If you work in the UK, you’ll know that your patients need to have decisions made at an early stage about whether they can be discharged or need to be admitted.  If you fail on this mission, you may face a fate worse than death.  If you do a 3-hour troponin, that means the patient has the test before their 4 hours is up.  But there’s a big challenge for us here…

The first blood test is unlikely to be taken at exactly time zero (when the patient first sets foot in the ED).  It might be taken half an hour later, realistically.  The second test should then be taken 3 hours later – so that’s now three and a half hours after arrival.  After you’ve taken the blood, the lab has to test it for you.  That takes a minimum of an hour.  (Still longer at many places in the UK)  This means that your patient has been waiting for longer than 4 hours – and you’ll therefore still need to admit them to hospital.

So, ultimately, these recommendations might not stop us from having to admit patients to hospital – but they will at least reduce the length of stay.

[/DDET]

[DDET What’s the evidence for these recommendations?]

Ah, now you’re asking the best question of them all.  NICE ran a systematic review and an economic analysis for these recommendations – but it’s not in the public domain just yet.  When it’s published, I’ll let y’all know the full details including all the juicy gossip.  Keep posted!

[/DDET]

You can catch our podcast right here!

If you’re implementing these recommendations, we recommend that you audit what you’re doing.  There’s a tool to help with this on the NICE website.  You can find it here.  If you are running an audit, please get in touch.  We really should pool our efforts!

Until next time,

Rick

15 Comments

  1. Doc Xology

    We’ve been doing this at out centre for over 12 months now

    Reply
    1. richardbody (Post author)

      Thanks for commenting, Derek! It’s actually been in the ESC guideline for probably well over a year as well. This post and the podcast are about what NICE thought in their new recommendations and some of the really important caveats that people might not appreciate from reading the recommendations by themselves.

      Which assay are you using and what are your experiences of running the 3-hour rule out? Do you have the 4-hour target yet? How do you think it will work with that?

      Are you aware of the evidence for 3-hour rule outs? I’ll post again on that when the NICE systematic review is published. It might not be what you expect!

      All the best,

      Rick

      Reply
      1. DocXology

        1) Roche
        2) Main aim is to reduce cardiac admissions – ~ 2/day (75000 annual ED census)
        3) 4/hr (‘NEAT’) targets have not yet been strictly enforced and currently hospital ‘buy-in’ is largely absent
        4) Watch this space
        5) Yes and growing including an Oz study and our own local audit data. Currently cardiology will still admit abnormal ECG, abnormal first troponin (even if thought to be CCF, CRF)

        They are currently examining the role of ultra-high sensitivity troponin < 3ng/L as a rule out.

      2. richardbody (Post author)

        Hi Derek,

        That’s interesting. Do you have a reference for your Oz study?

        I’ll write more on the <3ng/L strategy to rule out AMI soon. My first thought when the high sensitivity assays came out was that being able to detect lower troponin concentrations will enable us to lower the diagnostic cut-off and hence have a strategy with higher diagnostic sensitivity and negative predictive value. The first evidence was very promising - http://www.sciencedirect.com/science/article/pii/S0735109711024247 (it’s FOAMed). Bandstein and her group in Sweden then ran a retrospective analysis of almost 15,000 patients and found that patients with an initial hs-cTnT <5ng/L (the limit of 'detection' of the assay rather than the limit of 'blank') who had no ECG changes had an extremely low probability of AMI. We need some prospective data and we have that, from two studies, and we'll hopefully publish them soon. Watch this space!...

        Rick

  2. Chris Krause

    Does Nice have an opinion regarding those patients with known HF, renal impairment, etc whose initial TnT (what we use in our shop) is already elevated – say, 24 when the 99%ile is 14 – but whose 3 hour repeat is unchanged?

    Reply
    1. richardbody (Post author)

      Hi Chris,

      Actually, while those are really important issues, NICE was really clear that this is outside the scope of the current project. They just wanted to look at how we might use hs-troponin to rule out NSTEMI in the Emergency Department. Anything else was outside of the scope for fear that the project would become too unwieldy. I know what they mean, to be fair. If the project got too large, it would become unfocused and there’s less chance of producing something robust and evidence-based.

      One really important issue is whether we should be evaluating the presence or absence of a delta at just 3 hours (and what constitutes a delta). There’s very little evidence on this. Because we’re not looking to discharge patients with a positive troponin so early, my recommendation would be to stick with a 6-hour delta (i.e. from two samples 6 hours apart) for now, when one of the troponin levels is positive.

      Reply
  3. Cyrille

    Hello, I found it surprising the hsTnI from Siemens not be recommended. It is one of the ”high-sensitive” Tn meeting the ”high sensitivity criteria” requirements edicted by the IFCC. Any reason for it no to be in those NICE recommendations?

    Reply
    1. richardbody (Post author)

      Dear Cyrille,

      Thanks a lot for your comment. Actually, the Siemens assay was used in one of the first two studies to report on the use of ‘high sensitivity troponins’. (See http://www.nejm.org/doi/full/10.1056/NEJMoa0903515). However, the criteria for ‘high sensitivity troponins’ were subsequently set as follows: (1) the precision of the assay must be such that the co-efficient of variation is <10% at the 99th percentile cut-off; and (2) the assay must be able to detect troponin levels in at least 50% of apparently healthy individuals.

      The Siemens assay fulfills criterion (1) but not criterion (2) - i.e. it doesn't detect troponin levels in >50% of apparently healthy individuals. It’s a shame because the limit of blank turned out to be 6ng/L but the assay can actually report levels lower than that. If the limit of blank turned out to be lower, it would have been a high sensitivity assay. We recently evaluated that assay (not yet published) and I can tell you that, had the limit of blank been lower, we might have achieved even more with that assay.

      The NICE project only looked at high sensitivity assays so the Siemens assay wasn’t evaluated at all. However, if we had looked at it, we would have noted that the sensitivity of the Siemens assay at 3 hours was 100% in Keller’s study (link above). That’s actually about as robust evidence as we have for any 3-hour rule out strategy. I’ll expand on that when I write a blog post about the systematic review later.

      Thanks again,

      Rick

      Reply
  4. will sargent

    This is a bit frustrating- there’s a lot of ‘this doesn’t show this, do this or that’ and there are no references on the actual NICE website…….
    We have the abbott assay- just need something to support the 3hr??!!

    Reply
    1. richardbody (Post author)

      Hi Will,

      The bottom line is that a test on arrival and then 3 hours after that is sufficient to ‘rule out’ NSTEMI if you use the Abbott ARCHITECT stat high sensitivity troponin I assay or the Roche high sensitivity troponin T assay. It’s worth noting that it’s 3 hours after arrival and not 3 hours after symptom onset, obviously! It’s also worth checking if the Abbott assay you’re using is their high sensitivity version. They are still selling their contemporary troponin assay as well.

      The NICE recommendation is long to read but the bottom line is in recommendation 1.2. You can also find the Diagnostics Assessment Report including a systematic review at this link… http://www.nice.org.uk/guidance/dg15/documents/highsensitivity-troponin-for-the-early-rule-out-or-diagnosis-of-acute-myocardial-infarction-in-people-with-acute-chest-pain-final-protocol2

      Hope that helps!

      All the best,

      Rick

      Reply
  5. will

    Cheers Rick,
    It kind of does and doesn’t. I suppose like everyone we are after the holy grail evidence based 3hr or less decision rule. Backus with bog standard trops supports this but as Amal Mattu suggests, it does imply discharging people home with a raised trop- and where did the criteria come from?- to me they are slightly arbitrary, but are clearly simple to use, and they give us the useful risk numbers…….
    I’m still going a bit in circles looking for the nugget of evidence supporting them (the hsTnI) in the NICE guideline. Dave Newman says we way over egg this (cardiac) pudding anyway, troponin shmoponin- or should that be hsmoponin?
    Anyway, we do indeed have the Abbott assay here in Darwin- our chief biochemist was somewhat hasty in implementing, but hey, what could go wrong?

    Reply
    1. richardbody

      Hi Will,

      Well, there’s a reason you don’t see anything about decision rules. NICE specifically restricted their scope to looking at troponin alone, and didn’t want to consider clinical information alongside that. That’s why you don’t see the limit of detection rule out strategies combined with ECGs in the systematic review, and it explains why the ADAPT, HEART and MACS protocols/decision rules weren’t evaluated.

      That ought to be within the scope of the NICE chest pain of recent onset guideline, which will be due for revision soon. Hopefully that will answer some of your questions!

      Rick

      Reply
      1. Will

        So essentially this was a technical review, yes? I just wonder why it mentions the 3hrs that’s all.

      2. richardbody

        Hi Will,

        It was a review for the Diagnostics Assessment Committee, and the scope was to evaluate high sensitivity troponin assays for rule out of NSTEMI within 4 hours of patients arriving in the ED. Any evidence relevant to that was considered but the committee was keen to only look at evidence for the test itself, rather than decision rules than also incorporate other information.

        All the best,

        Rick

  6. Pingback: Tasty Morsels of EM 106 – #FRCEM NICE Guidelines ACS - Emergency Medicine Ireland

Thanks so much for following. Viva la #FOAMed

Translate »