Estimated reading time: 10 minutes
Introduction
Regular readers of St Emlyn’s will know we are big fans of targetted, evidence based testing to aid clinical decision-making and even have one of the world’s leading authorities on Troponin on the team (Rick Body).
Over the course of my career, I’ve seen the introduction of troponin testing (to replace the cardiac ‘enzymes’) at 12 hours after the onset of pain, to the advent of a single test, one hour rule out strategies, so it seems only inevitable that there are now papers looking at testing in the prehospital environment. If you need a reminder about troponin we have an excellent series of posts here, or you could listen to this podcast…
The Paper
The paper we will be discussing is from the European Heart Journal. As ever, we recommend that you read the whole paper yourself, and brilliantly the full text is free and available here
An early disclaimer
I have to start off by being completely honest – I found the primary outcome in this paper incredibly off-putting. It was focussed on ‘costs from a healthcare perspective at 30 days’. I realise that this is important in the wider scheme of things, but what I really want to know about is safety and accuracy. However, this was considered as a secondary outcome – looking a Major Adverse Cardiac Events (MACE) at 30 days, so for me, it was worth looking deeper.
What kind of trial is this?
This is an investigator-initiated, multicentre, open-label, randomized controlled trial in five ambulance regions in the Netherlands. In the region studied, there are 112 ambulances with 552 ambulance paramedics. All of the ambulance paramedics underwent mandatory training in various forms, with an online test on completion (although they didn’t say if they needed to pass!).
Tell me about the patients
Patients had either called their GP or ambulance for their pain. On assessment, they had to have ‘suspected NSTE-ACS’ only that the symptoms had to have been present for at least two hours prior to the arrival of the ambulance, although the authors don’t actually state what was used to decide on this suspicion. They then used the HEART score (the HEAR bit anyway – as T stands for Troponin) to decide if the patient was ‘low risk’ and therefore eligible for the study (if a HEAR of less than or equal to 3).
You may well be using the HEART score in your practice, but here is a reminder from a friend of St Emlyn’s Barbra Backus if not…
You’ll notice that only patients with a score of ≤ 3 for the whole of the HEART score gives a MACE of 1.9%. The assumption is then that their troponin will be under the normal limit and therefore the patient deemed ‘low risk’ and a cardiac cause for their pain ruled out. Clearly, if their troponin is raised at all this would put them into a higher-risk group needing further assessment. The ECG was interpreted by the paramedics and if there was any doubt it could then be transmitted to an independent cardiologist (although, again it is hard to know how much information they were given and whether this may have biased their interpretation- was it just ‘please look at this ECG’, or ‘please can you look at this ECG of a patient I think is having a heart attack?’?)
Once the initial screening had taken place, if eligible, the patient was invited to join the study and then shown an animated version of the trial information. Exclusions were all the things you might expect: ST segment elevation; suspected non-cardiac cause for the symptoms; decreased GCS (<8 – although if it was any less than 15 it would probably be hard to get informed consent); and plenty of others. Interestingly both the general practitioner and cardiologist could exclude the patient on the basis of needing to be seen in the ED.
Once the patient had given their consent, they were randomised to either the prehospital rule-out strategy or the ‘usual care’ (ie immediate transfer to the ED for assessment).
Tell me about the intervention
If randomised to the pre-hospital rule-out strategy the patients had on-site point-of-care troponin T measurement using the Roche Cobas h232. This test has a measuring range of 0.04 – 2ng/ml (40 – 2000ng/L). In comparison, the laboratory-based high-sensitivity troponin T assay manufactured by Roche Diagnostics has a limit of detection of 5ng/L. This means that the point of care test studied here cannot detect the small concentrations that we’re used to with laboratory-based high-sensitivity troponin assays. What’s more, the test does not have the same precision as the lab test. By that, we mean that the result may vary more often when we measure the same sample. We measure precision with the coefficient of variation, which is expressed as a percentage. Usually, we want the coefficient of variation to be less than 10% at the cutoff used to make clinical decisions. The manufacturer is unable to tell us the coefficient of variation at the upper reference limit of this point of care test, which suggests that the test does not meet the 10% criterion. It doesn’t meet the criteria for being labelled as a ‘high-sensitivity’ troponin test, which has become a standard of care in hospitals. The results are via venous blood sampling into a heparinised tube and are available in 12 minutes. For the ED rule-out strategy usual investigation pathways were followed.
What is MACE?
We often talk about MACE in cardiological trials so it is probably worth just reminding ourselves what this is:
- Major
- Adverse
- Cardiac
- Events
In this case, the researchers defined these at 30 days as one or more of the following events: acute coronary syndrome; unplanned revascularisation and all-cause death.
What about the outcomes?
As mentioned earlier I am, (rather naughtily) going to completely ignore the primary outcome stated by the researchers, but concentrate instead on the clinically relevant outcomes. It is important to note, however, that the trial was not adequately powered for these outcomes and so any results must be interpreted with caution (the authors estimate that to perform this trial would require over 17 thousand patients)
MACE in the study population
There were approximately 430 patients in each arm of the study, with 17 (3.9%) in the pre-hospital group and 16 (3.7%) in the ED group having MACE within 30 days of randomisation, so I think we can agree that these were indeed a ‘low risk’ group. Even if we did absolutely no testing at all we’d be right 24 out of 25 times. However, this group did include four with an ST elevation myocardial infarction (it’s hard to know when these diagnostic ECGs were obtained, but presumably it was not at the initial assessment and at some point in the following days).
MACE when patients were ‘ruled out’
Of course, what we are interested in is the number of patients in both groups who had been ‘ruled out’, but went on to have MACE. Of those that had a MACE in the prehospital group 15 of them had a raised troponin and so were referred to the ED. This does mean that 2 patients (0.5%) who had been ruled out in the prehospital group went on to have a MACE within 30 days. However, this also occurred with four patients (1%) in the ED strategy.
In this, albeit underpowered, study it does appear that the prehospital use of a HEART score to rule out major adverse cardiac events is at least as good as an in-hospital strategy
But surely all of this roadside assessment took time?
It did indeed – in fact the median on-scene time in the pre hospital group was over an hour (63 minutes) and in the ED group was statistically significantly less at 39 minutes. However, it seems that this extra time spent ‘in the field’ was more than worth it with the large majority of patients who were not transported having a safe outcome and no waiting in the ED. In fact, the back to availability time for the ambulances was very similar – the ED transportation time was about 15 minutes on average (and we can only imagine what the handover delay could be in the current UK system).
Healthcare resource use
Yes, yes, I know I said I wasn’t going to mention the cost-effectiveness analysis, but let’s just state the obvious: if patients were not taken into the ED there were significantly fewer investigations and thus much less cost. Not to mention the beneficial knock-on effect for the patients of not having long waits and reducing crowding in Emergency Departments.
The big question is could (and should) we do this in the UK?
There are several limitations:
- If a patient was ‘ruled out’ in the prehospital strategy their care was returned to their GP. Many of the patients assessed by our paramedic teams are direct 999 calls, without GP involvement. The crews would then have to contact the patient’s family doctor in order to make sure they were followed up.
- Paramedics in the Netherlands are all highly trained and all ambulances carry at least one trained member of staff. In the UK we can occasionally have crews without a paramedic.
- Ruling out a cardiac cause doesn’t mean there isn’t something that needs assessment – this should be part of the initial inclusion (or rather exclusion), but a negative troponin and low HEART score doesn’t mean nothing is wrong. The patient could still have another serious diagnosis that requires attention, and that may not have been picked up using the definition of MACE applied in this trial.
- Although we saw no difference in MACE between groups in this trial, it is still possible that patients in the prehospital group had missed myocardial infarction. Those patients may not have had a MACE within 30 days but their longer term prognosis is likely to be worse because they will miss out on the evidence-based treatments. To address that concern, we also need data from observational studies where all patients receive reference standard investigations for myocardial infarction. The anticipated results of the PRESTO study, led by Rick, will help us in that regard.
Isn’t it just all about the Bayes’?
Well yes and no. The truth is that a lot of these patients are really such low risk that we shouldn’t need to do a test on them at all. But that is not the world we live in. When there is even the slightest risk of an adverse (and in this case life-threatening) outcome clinicians want objective evidence to show that they did the right thing and to protect them from a perceived threat of litigation. In this case, doing a POC troponin seems to satisfy that desire.
Summary and final thoughts
I think point-of-care testing has real potential to help us both risk stratify and also rule out certain conditions before the patient even gets to hospital. We need to be bold and consider where and when we can use these strategies and then support and educate the staff who are using them. The early data from the ARTICA trial suggests that this method may be safe and would be incredibly helpful in a significant large group of patients who currently need to be seen in hospital.
References
- Cyril Camaro, Goaris W A Aarts, Eddy M M Adang, Roger van Hout, Gijs Brok, Anouk Hoare, Laura Rodwell, Frank de Pooter, Walter de Wit, Gilbert E Cramer, Roland R J van Kimmenade, Peter Damman, Eva Ouwendijk, Martijn Rutten, Erwin Zegers, Robert-Jan M van Geuns, Marc E R Gomes, Niels van Royen, on behalf of the ARTICA Investigators, Rule-out of non-ST-segment elevation acute coronary syndrome by a single, pre-hospital troponin measurement: a randomized trial, European Heart Journal, 2023;, ehad056, https://doi.org/10.1093/eurheartj/ehad056
