JC: Are we fully LoDED?

Over the last few years many of us in the UK have started to incorporate high sensitivity troponin into the assessment of patients presenting with chest pain.

We have seen these samples taken at ever shorter intervals, aiming to discharge low risk patients safely, sooner from the Emergency Department (ED). This has been driven in part by the “Four Hour Emergency Access Target”^​1​ as well as increased crowding in overwhelmed EDs.

Before looking at the paper it is probably worth a quick bit of revision about some of the terminology used in “troponinology”. For a more in depth discussion a listen to the series of St Emlyn’s podcasts with Rick Body, or a read of his blogposts^​2​, is a good idea. As ever we suggest you read the paper for yourself (it is open access).

What is High Sensitivity Troponin (hs-cTn)?

A “high-sensitivity troponin” test refers to the analytical sensitivity and not the diagnostic sensitivity. To be highly sensitive the test needs to meet two criteria: that a level of troponin can be detected in 50% of “normal” people (ie it’s really good at finding incredibly some amounts of the troponin protein) and that the “co-efficient of variation” at the cut-off is less than 10% (if you measure a single sample repeatedly the result will not vary).

There are two other criteria that it is vital to understand before reading this paper: the limit of blank and the limit of detection.

The limit of blank is the highest value that you might reasonably detect when a troponin-free sample is tested. You might think that this would be zero, but this isn’t actually always the case – there’s always a little bit of noise. Ideally this should clearly be zero, but this isn’t actually the case – there’s always a little bit of noise.

The limit of detection is the smallest level of troponin you can reliably distinguish from the level of blank – it is this criteria that is used as one of the key tests in this study (along with a ECG)

The Paper

Limit of detection of troponin discharge strategy versus usual care: randomised controlled trial^​3​

In this paper the investigators wanted to see if using a negative single hs-cTn early in the clinical investigation (with an accompanying “non-ischaemic” ECG) of patients presenting to the ED with “low risk” chest pain would facilitate an earlier discharge than using established strategies.

Methods

This trial aimed to compare the “LoDED strategy” with “usual care” in adult patients attending the ED with chest pain that could be cardiac in origin. Crucially, the decision that the pain could be cardiac in origin was down to the treating clinician (whose level of experience wasn’t specified).

The LoDED strategy was a single hs-cTn below the level of detection with a non-ischaemic ECG and no ongoing clinical concern.

“Usual care” varied across the eight trial sites: three different hs-cTn assays were used; some using GRACE, HEART or TIMI scores as a risk stratification; and a variety of troponin assay results (some including the level of detection). Importantly, some sites already allowed a proportion of patients to go home after one troponin test, and some even used the limit of detection of the troponin assay to do that (although only in low risk patients using risk scores).

Patients were randomised 1:1 to either the LoDED strategy or the usual care and the primary outcome was discharge within 4 hours without a Major Adverse Cardiac Event (MACE) within 30 days. A MACE was defined as any of cardiac death, type 1 acute myocardial infarction, or emergency revascularisation.

The eight trial sites were all in the south of the UK, with only one (Royal London) in the capital.

The patients in both groups seem well matched, but it’s important to note that 88% of the entire sample declared their ethnic origin as “white”.

Results

A total of 632 patients were randomised between June 2018 and March 2019. The number of patients in each site ranged from 44 (University Hospital Southampton) to 172 (Royal Berkshire).

For the primary end point 46% were discharged within four hours (without MACE at 30 days) for the LoDED strategy and 37% using the usual care strategies, which wasn’t statistically significant. So, the trial was negative. Nobody who was discharged early in the intervention group developed a MACE. However, it’s worth noting that the study was’t powered to demonstrate a difference in the safety outcome (MACE) between groups.

Discussion

There is lots to consider within this paper and before starting the discussion I should declare that I work at one of the sites that was part of the study.

For many of us I think the patients we are seeking to use a strategy on are the very low risk (in some cases the “probably no risk, but I really want a test to back me up”). MACE are significant things to miss and the “do you remember that patient you sent home?” conversation is perhaps one we fear the most.

It’s interesting to see that almost 10% of the patients included in the study had a “highly suspicious” history (central chest discomfort with radiation to the jaw/arms precipitated by exertion and relieved by rest or nitrates) and a further 36% a “moderately suspicious history (some highly suspisious features, but some atypical features too). I think I would personally be very reluctant to include these patients in “low risk strategy”. and it would put far too much trust in the hs-cTn rather than the history. Although neither group who were sent home (ie had negative tests) had a MACE at 30 days the sample size is relatively small and we know that a single “missed” diagnosis could be disasterous. We should also acknowledge here that Emergency Physicians “gestalt” has limitations^​4​.

The heterogeneity between the usual testing strategies is rather surprising. Even those closely geographically related departments (often in the same Deanery) have varied approaches: some using LoD already, while others include sex specific cut offs and some the change (delta) between two tests.

This strategy places a lot of emphasis on the interpretation of the ECG and determination that it is “normal”. Although for many clinicians in the ED their ECG reading skills are good, we have no guarantee that this is true across the board and would need to have some quality assurance before allowing clinicians to use this strategy. With trainee doctors often only having four month placements this may be difficult to achieve.

Although this may tell us in this group that the patients do not have cardiac chest pain, we need to consider what their diagnosis may actually be. A key skill for the emergency physician is to “rule out the worst case”, but we do also have to attempt to tell the patient what may be causing their discomfort. The slight increase in time that the “usual” strategy often allows also does give us slightly longer to observe the patient and see if their symptoms develop. The four hour access standard is important, but many EDs have developed “Clinical Decision Units” or similar where a patient can be observed in relative comfort In a way the speeding up of their journey may be detrimental.

Conclusions

The authors conclude that “the clinical effectiveness of such a strategy is limited when compared with existing rule-out strategies…. and is likely to undermined by a range of system factors such as prolonged ED waiting times and crowding.

What do I take from this paper? Well, it confirms that patients with “low risk chest pain” and a normal ECG, a hs-cTn below the level of detection is hugely reassuring. We need to make sure the ECG skills of all our clinicians are of a high standard and our current rule out strategies are already very safe. There is huge variation between EDs regarding their current testing strategy and this difference is worth further investigation, and must be remembered when doctors in training rotate between departments.

Will this change my management? Not yet. To be absolutely honest I quite like the “excuse” to keep patients for three hours (don’t tell the Department of Health), and still think we have to try to make a positive diagnosis of what the patient has, not just tell them what they don’t have. I am really pleased that this high quality study was conducted and hope we can continue to support trials like this in our Emergency Departments.

The Podcast

Rick Body very kindly spent some time discussing this trial, and recent thoughts about the use of hs-cTn, in this St Emlyn’s podcast.

The Shownotes

The various organisations mentioned by Rick can be found here:

The Innovation Agency Webinar Series

The NHS Accelerated Access Collaborative

The CQUIN that will be implemented later this year (page 15 for the Troponin section)

The Draft NICE recommendations