This month I was honoured that a study I wrote with some terrific colleagues from Europe, Australia and the US has been published as the ‘Editor’s Choice’ in Academic Emergency Medicine. This was a secondary analysis from a large, multi-centre, international study I was lucky to be involved with: the TRAPID-AMI study. The primary goal of the study was to evaluate the accuracy of a 1-hour rule in and rule-out algorithm using high sensitivity troponin. This was recently published in Annals of Emergency Medicine with Christian Mueller as first author – and it’s open access. You can find it here. The sensitivity of that algorithm was 96.7% and the negative predictive value 99.1%.
In this secondary analysis, we wanted to evaluate whether patients who had an initial high sensitivity cardiac troponin T (hs-cTnT) concentration below the limit of detection (LoD) of the assay [that’s well below the usual cut-off, by the way] and a normal ECG could have an acute myocardial infarction ‘ruled out’. My interest in low cut-offs like this goes back some way. We first published a paper on this in the Journal of the American College of Cardiology back in 2011, showing a sensitivity of 100%. You can find it here – it’s also open access.
In a large retrospective study, Nadia Bandstein from Sweden showed that patients with a hs-cTnT concentration <5ng/L and no ECG ischaemia had a very low probability of acute MI. We validated that in another prospective study published fairly recently, again showing a sensitivity of 100%. Again – you’ll love this – it’s open access!
The only problem was that all the evidence seemed to be coming out of single centre studies. Now, we’ve got evidence from a study of 1,282 patients recruited at 12 centres in 9 countries and across 3 continents. We had study centres across Europe, in the US and in Australia. I think this is a real strength of the work – it means the findings are more likely to have external validity for people in many different environments. We also only included early presenters who arrived within 6 hours of symptom onset: just the sort of patient this early rule out strategy stands to benefit the most.
What did we find?
Our results were pretty promising. The sensitivity of this single test strategy was 99.1% and the negative predictive value was 99.6%. The confidence intervals are fairly tight as this is a multi-centre study (see below). The proportion of patients developing major adverse cardiac events (MACE) within 30 days was also fairly low at 1.3%, too. It’s worth noting that our definition of MACE didn’t include in-hospital revascularization and the incidence jumps to 3% with that included – but some people argue that is a subjective outcome.
How can I find out more?
You can read the full paper at Academic Emergency Medicine: please do! It’s open access and can be found at this link. You can also catch a most excellent blog and podcast by the SGEM, including an interview I did with Ken Milne, Justin Morgenstein and Corey Heitz – who are seriously brilliant at what they do. You can find the SGEM blog and podcast – called ‘Oh baby you’re too sensitive’ – at this link.
Once you’ve read their blog post and listened to their podcast, please also join in the online conversation. Search for and use the hashtag #SGEMHOP for your tweets: the best tweets in this conversation will be published in Academic Emergency Medicine – not a bad incentive for having a great online conversation!
What do the findings mean?
This strategy – to rule out AMI in patients with initial hs-cTnT <5ng/L and no ECG ischaemia – was recommended for clinical use by the European Society of Cardiology last year. Draft NICE guidance also includes this strategy. So, it looks like this is ready for prime time use: you can rule out acute MI with this strategy. However, there is reason for caution.
Many biochemists are a bit concerned that the precision of the hs-cTnT assay isn’t so great at such low concentrations. They say it could mean that one week we discharge a third of our patients with chest pain (as in this study) and another week it’ll be one in six patients (as in our study published in Clinical Chemistry). That might not be great for managing hospital beds.
The incidence of MACE isn’t so small either, especially when we consider in-hospital revascularisation – so we still need to think. We can’t discharge every patient with hs-cTnT <5ng/L and no ECG ischaemia as some could still have unstable coronary disease.
So, perhaps we can do better. I think we can – with prediction models that combine clinical information with hs-cTnT concentrations – but more on that another time!