One by One, Risk Factors Have Additive Impact on Cardiac Arrest Survival

Resuscitated patients with six of eight “unfavorable features” have a 10% or lower chance of surviving to hospital discharge.

One by One, Risk Factors Have Additive Impact on Cardiac Arrest Survival

For patients successfully resuscitated after cardiac arrest, it’s possible to predict their odds of survival by counting how many risk factors are present, data from the International Cardiac Arrest Registry (INTCAR) confirm.

Thanks to more-aggressive care, such as targeted temperature management (TTM) and coronary angiography/intervention, survival to hospital discharge for these patients has risen from around 25% to 50% in recent years, researchers say. Still, the burden of “multiple unfavorable features,” as worded in a 2015 algorithm from the American College of Cardiology (ACC) Interventional Council, can portend worse outcomes even in the face of this advanced treatment.

Exactly how many risk factors mattered, and to what extent, hadn’t previously been pinpointed.

Lead author Ahmed A. Harhash, MD (University of Arizona Sarver Heart Center, Tucson, and University of Vermont, Burlington), told TCTMD that, despite its contribution, the 2015 algorithm left clinicians with questions: “Would you treat someone who has one feature similar to someone who has 10? Would you treat someone who has a particular feature similar to the one next to it?”

What they learned in this study, published online today in the Journal of the American College of Cardiology, is that “the more of those features you have, the worse the outcome is, and those features have different weights—not all of them have the same significance.”

This is good news, said Harhash. “We’re definitely pleased we found the correlation that we were looking for.”

Looking for a ‘Tipping Point’

The ACC’s algorithm included 10 components: unwitnessed arrest, initial rhythm of nonventricular fibrillation, no bystander CPR, > 30 min from collapse to return of spontaneous circulation (time-to-ROSC), ongoing CPR, pH < 7.2, lactate > 7 mmol/L, age > 85 years, end-stage renal disease, and noncardiac etiology.

Harhash et al tweaked this list slightly to match up with INTCAR’s exclusion criteria and endpoint definitions, arriving at eight high-risk features that were analyzed in 2,508 comatose yet resuscitated patients treated between 2007 and 2017. Two-thirds were men, 73.7% had out-of-hospital cardiac arrest, 44.1% experienced ventricular fibrillation, and 19% had ST-segment elevation on ECG. Nearly all (94%) received TTM, 43% underwent angiography, and 23% were revascularized with PCI or CABG. Survival to hospital discharge was 39%.

Each of the eight risk factors apart from chronic kidney disease (CKD) was individually associated with the likelihood of survival to discharge. Patterns were largely similar for patients who did versus didn’t have ST-segment elevation.

High-risk Features for Resuscitated Cardiac Arrest Patients



Association With Survival

OR (95% CI)

Age > 85 Years


0.30 (0.15-0.61)

Time to ROSC > 30 Minutes


0.30 (0.23-0.39)

Nonshockable Rhythm


0.39 (0.29-0.54)

No Bystander CPR


0.49 (0.38-0.64)

Lactate > 7 mmol/L


0.50 (0.40-0.63)

Unwitnessed Arrest


0.58 (0.44-0.78)

Initial pH < 7.2


0.78 (0.63-0.98)



0.96 (0.70-1.33)


Patients who possessed three or more of these features had a predicted survival of less than 40%. Survival rates were 10% or lower when combining the three “strongest” risk factors (age > 85, time to ROSC > 30 min, nonshockable rhythm) or any six of the features.

The “tipping point” for poor prognosis seems to be at six or more features, the researchers conclude. “Such knowledge can assist physicians in identifying who are least likely to benefit from aggressive intervention, which can impact the discussion with families regarding realistic expectations. This may also improve resource allocation and direct access to invasive interventions for those more likely to survive.”

Harhash said that their prediction tool is just one part of decision-making. “Our point was to highlight what to look for in the ultra-early phase of triage” when patients are being seen in the emergency department, he explained, adding that their tool can’t help clinicians decide who should or shouldn’t proceed, for example, to angiography. Rather, it shows certain patients “have a good chance of survival and if you invest in them with invasive therapy . . . probably this will be another [factor] in improving their survival.”

For patients with six or more features, the information should be paired with the goals of care, with an eye toward resource allocation—particularly in times, like COVID-19, where such resources might be finite, he advised.

It’s important to remember, though, that these data predate the pandemic, stressed Harhash. The tool prediction tool probably can’t be applied when triaging cardiac arrest patients with confirmed COVID-19, since these events may stem from the infection and are thought to carry “dismal prognosis.” On the other hand, for clinicians facing strained healthcare systems in this era, the choice to prioritize patients more likely to survive must take into account the local situation, he said.

  • Harhash reports no relevant conflicts of interest.