This blog post is based on a paper we’ve just had published in the Emergency Medicine Journal entitled: ‘The Manchester Acute Coronary Syndromes (MACS) decision rule: validation with a new automated assay for heart-type fatty acid binding protein’. You can find the full text at this link.
In May last year we (Simon and I, together with some great colleagues) published a paper in Heart reporting how we derived and externally validated the MACS (Manchester Acute Coronary Syndromes) rule for suspected cardiac chest pain. The MACS rule can be used to help you decide where to send a patient who presents to the ED with what could be cardiac chest pain. There are four options (and this can be customised to local practices): you can send patients home (immediately), to an ED observation ward (which means they’re ‘low risk’ and likely to go home after serial troponin tests), to an acute medical ward (which means they might need a bit more work-up even if serial troponins are normal) or to a cardiologist (or Cath Lab – if you’ve ‘ruled in’ an acute coronary syndrome). The MACS rule includes eight variables. To use it, you’d tell the computer whether each variable is present or absent and insert the biomarker results. The computer will calculate the estimated probability of a major adverse cardiac event within 30 days and will use that to tell you where it thinks you should send the patient.
In case reading whole journal articles isn’t your bag, or if you also like to know the story and interpretation behind the work, we also blogged about our initial paper, of course. We weren’t the only ones to write a blog post about it. We were really honoured that it was covered by the totally awesome The Bottom Line and by the great Ryan Radecki at EMLitofNote. If you follow those links, you’ll see that Ryan’s critical appraisal of the MACS rule wasn’t entirely positive. (Ed – understatement alert: he actually called it the ‘unusable Manchester chest pain instrument’). It might sound odd to you – but that kind of feedback is actually fantastic for us, as authors. That’s what #FOAMed can offer – open access peer review, free speech and an ongoing dialogue.
Why did Ryan Radecki call the MACS rule ‘unusable’ and was he right?
Ryan’s point was that, to use the MACS rule, you need to measure heart-type fatty acid binding protein (h-FABP). That’s not something we do every day, of course. Because h-FABP isn’t something we measure every day, there aren’t too many ways to measure it quickly using an automated analyser (like the tests we usually do in the ED). In the original study, we tested for h-FABP using an ELISA from Randox Diagnostics. If you happen to know anything about chemistry, running an ELISA is actually quite labour intensive. We ran ours in batches of around 40 samples each. Each batch took several hours to complete – and I’m indebted to Phil Pemberton for investing such a lot of time to do that. Obviously, if we were to run the tests ‘live’ for patients in the ED, we’d need a biochemist on hand 24/7 to run very labour intensive tests – and even then we wouldn’t get results back for several hours. So, yes, Ryan was right that (using an ELISA) the MACS rule is unusable.
So what’s changed?
Based at least in part on the very promising results that we’ve had regarding the MACS rule, Randox really helped to progress our research by developing a new h-FABP assay, which is fully automated just like any other lab test. It has a similar turnaround time to any other lab test (like troponin, for example) – so this really would make the MACS rule ‘useable’. However, it’s a different test to the one we ran originally. If you’ve listened to our troponin podcasts, you’ll know that not all assays are the same – so if this test really is to make the MACS rule ‘useable’ then we need to check that the rule would still perform well if we were to use it in practice.
That’s what we’ve done. We ran the new automated assay in stored samples from our external validation study. We also re-calibrated the MACS rule because the two h-FABP assays give slightly different results. For this, we used the correlation data that the manufacturer (Randox) provided. Then, we calculated the diagnostic performance.
What did we find?
We ran the new h-FABP test in 456 patients. If we use the MACS rule with the new test, we’d identify slightly fewer patients as being at ‘very low risk’ and thus eligible to be immediately discharged. In fact, just under 1 in 5 patients (18.9%) were classed as ‘very low risk’. Of those patients, none had an acute myocardial infarction – so sensitivity was 100% for that diagnosis. Two patients (2.3%) developed a major adverse cardiac event within 30 days. Now, 2.3% sounds like it’s quite a lot. However, there are major adverse cardiac events and there are major adverse cardiac events. Both of these patients underwent outpatient angiography and had coronary stenosis identified – but neither patient actually had a stent – and both did very well otherwise. So, my feeling (as the lead author) is that we actually defined the primary outcome badly. We said that patients with a new coronary stenosis had developed a major adverse cardiac event. However, it would be difficult to suggest that either of these patients ought to have been admitted to hospital in reality. They could have been discharged, continued to have their outpatient angiography and they’d have come to no harm.
Is the MACS rule just about ‘early rule out’?
No. We specifically designed the MACS rule to risk stratify all patients and to identify ‘high risk’ as well as ‘very low risk’ patients. In doing so, we hope that it’ll be able to make the most judicious use of resources. It’s all very well having a decision rule that sends a few patients home early – but if it increases admissions to Coronary Care Units (which are very expensive) then it’ll actually increase healthcare costs overall. In this study, we found that 1 in 9 (11.1%) patients were identified as being at ‘high risk’ using the MACS rule. Of those patients, 92.0% developed a major adverse cardiac event within 30 days.
What do the findings mean?
In summary, these findings show that the MACS rule will still work using the automated h-FABP assay that can be run by any hospital lab in real time. That ensures that the MACS rule is definitely ‘useable’ in practice. The next step is to find out what actually happens when we use it in practice. We’ve already started down that road – and we’ve just finished a new study looking at the feasibility of running an RCT of the MACS rule when compared with standard practice. We’re also about to start a new study that will compare the MACS rule against some of the best analysis – like the HEART score and the ED-ACS score. We’ll keep you posted about the results!
All the best,
Subscribe to the blog (look top right for the link)