Simple MIRACLE2 Score Could Help Triage Out-of-Hospital Cardiac Arrest

A new, simplified tool for predicting better neurological outcomes following out-of-hospital cardiac arrest (OHCA) could one day help in deciding whether a patient should be rushed for further cardiac workup and interventions or whether such resource use might be futile.

In the shorter term, lead author Nilesh Pareek, MBBS (King’s College Hospital NHS Foundation Trust, London, England), told TCTMD, a simple score could be used to better select patients for interventions being tested in this high-risk group, increasing the chance of success.

While a number of other tools already exist, said Pareek, the aim was to develop something as simple as possible. That ended up being a 10-point score that could easily be added up, or as a check-list that could be part of patient handover from emergency personnel.

“When you're served up a cardiac arrest patient by the ambulance service who's intubated, who may have had a lot of drugs, a lot of the details are unknown. Often you don't even know the patient's name,” Pareek said. “It's often very challenging to know which patients will survive and will have a good brain survival, and which patients have had a very long and unfavorable cardiac arrest situation and will go on to die of brain injury.”

As a result, he continued, patient management isn’t standardized. The choice of who should go on to echocardiography and angiography, or receive mechanical circulatory support and PCI, remains a judgement call on the part of treating physicians, said Pareek. “We felt that we needed a risk score to try and guide clinicians with provision of some of those invasive therapies.”

I think this is the first step in what could be quite an interesting journey where we understand what’s the best treatment for which patient group. Nilesh Pareek

The idea for MIRACLE₂ evolved out of his PhD thesis, he noted, inspired in part by gaps in current European and American guidelines that recommend that “favorable cardiac arrest circumstances” should signal considerations for angiography. “The problem is they never really clarified what those are,” Pareek observed.

Looking for a MIRACLE₂

For the study, prediction modelling with multivariable logistic regression was used to identify seven independent predictors of poor neurological outcome at 6 months using a retrospective sample of patients who’d had OHCA in the London area and were included in the King’s Out of Hospital Cardiac Arrest Registry (KOCAR).

It’s from these seven variables that the tools acronym, MIRACLE₂ is derived: missed (unwitnessed) arrest, initial nonshockable rhythm, reactivity of pupils (or rather, lack thereof), age, changing intra-arrest rhythms, pH < 7.20, and epinephrine administration. The subscript “2” serves as a reminder that use of epinephrine is worth two points. All of the others are worth one point with the exception of age: while 60 to 80 years carries one point, age over 80 entails three points. A score of 0-2 therefore represents low risk (5.6% risk of a poor neurological outcome at 6 months), 3-4 intermediate risk (55.4%), and 5-10 high risk (92.3%).

People working in resuscitation medicine would recognize all of these criteria as things that are commonly measured and biologically plausible, with the exception of “changing intra-arrest rhythms,” Pareek said. “That's a novel variable that, as a clinician, I've sort of picked up over the years as being a bad predictor—we collected that a priori, but it’s not [necessarily] a routinely collected variable.”

Validation of the score was done using two additional retrospective cohorts, one in another hospital system in London, England, and the other in Ljubljana, Slovenia. The MIRACLE₂ score performed “more favorably” in the London cohort (with an AUC of 0.91 and a calibration slope of 0.834) than in the Slovenian cohort (AUC of 0.84; calibration slope of 0.744). Compared with other risk prediction tools, MIRACLE₂ outperformed the OHCA score proposed by Adrie and colleagues in 2006 and the Cardiac Arrest Hospital Prognosis score, but it performed as well as the Target Temperature Management score.

The advantage of MIRACLE₂, Pareek argued, is that it can be done using a quick mental calculation, making it more likely to be used than tools that are nomograms or flowcharts, which are “challenging to use in an emergency situation,” whereas “simple things get better uptake.”

Avoiding Self-fulfilling Prophesies

Commenting on the study for TCTMD, Benjamin S. Abella, MD (University of Pennsylvania, Philadelphia), called it “well-executed.” MIRACLE₂ “looks reasonable,” he added, although it joins a somewhat crowded field of predictive scores.

“A reason why this is a such a tricky problem is that approaches to postarrest care have improved over the years, so any prediction score is imposed on a changing background,” Abella said. “Also, prediction scores for cardiac arrest run the risk of generating self-fulfilling prophecies: if a score suggests a poor outcome and the team decides to withdraw care, the patient will not survive. So prediction scores need to have a very high bar to not lead to withdrawal of care for patients who might have a small chance of recovery.”

Prediction scores need to have a very high bar to not lead to withdrawal of care for patients who might have a small chance of recovery. Benjamin Abella

For Pareek, the next step, will be large, prospective validation of the score in different healthcare settings, including nontertiary centers and all-comer populations—a study he’s trying to get off the ground now. From there, he said, he hopes MIRACLE₂could be helpful in research settings to better define and select the patient populations who could benefit from whatever intervention is being studied. In patients with a low MIRACLE₂score it might make sense to proceed with urgent angiography; in those with higher scores, however, prioritizing brain protection might be the better strategy.

He gave the example of a patient with a MIRACLE₂ score of 6: some practitioners will focus on the 5% chance of survival while others will point to the 95% chance of neurological injury.

“I’m not saying withdraw therapy and let those people die,” Pareek stressed, “but I think it's important that we think about stratifying this patient group, because up until now, we've not really had objective measures to guide us. And then I think if we can integrate those into appropriately designed clinical trials, we could then understand which patients benefit most from angiography, from Impella [Abiomed], from mechanical circulatory support, from hypothermia. I think this is the first step in what could be quite an interesting journey where we understand what’s the best treatment for which patient group.”