Journal Watch: OHCA Rhythm Interpretation and Survival

Reviewed This Month

• Paramedic Rhythm Interpretation Misclassification is Associated With Poor Survival From Out-of-Hospital Cardiac Arrest
• Authors: Stoecklein HH, Pugh A, Johnson MA, Tonna JE, Stroud M, Drakos S, Youngquist ST
• Published in: Resuscitation, February 2022

It has been well established that early recognition of shockable rhythms followed by rapid defibrillation leads to better patient outcomes for prehospital cardiac arrest. Paramedics are often the first to interpret cardiac rhythms and administer treatments based on that interpretation. However, the relationship between paramedic rhythm interpretation and patient survival has not been well studied. 

The objective of the study we review this month was to estimate the sensitivity and specificity of classifications of shockable initial rhythms by paramedics using defibrillators equipped with ECG rhythm-filtering technology (ERF). ERF helps field providers identify underlying rhythms during chest compressions. The authors hypothesized that inaccurate initial rhythm interpretation would be associated with worse survival. 

Sensitivity and Specificity

Before we get too far into this study, let’s review sensitivity and specificity. In this study, sensitivity identifies the proportion of interpreted rhythms paramedics identified as shockable that were, indeed, shockable rhythms. Specificity identifies the proportion of rhythms paramedics identified as not shockable that were actually not shockable rhythms. 

To test their hypothesis the authors analyzed prospectively collected data from an ongoing quality improvement project for cardiac arrests. Since 2011 the Salt Lake City Fire Department has collected Utstein data elements and defibrillator files for out-of-hospital cardiac arrest (OHCA) cases attended by its EMS providers. To review, the Utstein data are reported uniformly to ensure apples-to-apples comparisons in resuscitation research. The analysis evaluated 9 years of data from the Salt Lake City Fire Department. The study period was from October 2011 to October 2020. 

Paramedics’ documented initial rhythms were compared to postevent physician interpretations of the same rhythms to estimate the sensitivity and specificity of paramedic identification of shockable rhythms and subsequent shock delivery to those patients. The authors also evaluated the association between misclassification of initial rhythms and neurologically intact survival to hospital discharge. 

Cardiac arrests were defined as either the administration of chest compressions by EMS or delivery of a shock by an AED. Arrests due to trauma, drowning, or strangulation were excluded. Cases where signs of irreversible death were present and those with do not resuscitate orders were also excluded. One interesting note about this study is that paramedics were trained to shock asystole. The authors say this was due to a sentinel event in this system in which paramedics misclassified ventricular fibrillation (VF) as asystole and withheld defibrillation. 

For the analysis the authors defined VF as a disorganized electrical activity without identifiable QRS complexes. Ventricular tachycardia (VT) was defined as regular, abnormally wide QRS complexes at a rate of at least 100 beats per minute. Asystole was defined as an isoelectric baseline without QRS complexes or with QRS complexes at a rate of less than 10 beats per minute. Pulseless electrical activity (PEA) was defined as QRS complexes occurring at a rate of 10–100 beats per minute with no pulse. Patient outcomes were obtained from review of hospital electronic medical records or direct contact with a medical provider involved in patient care. 

Results

The analysis included 863 resuscitation attempts. Thirty-nine percent were asystole, 22% were VF, 19% were PEA, 12% were unknown nonshockable rhythms, 6% were unknown shockable rhythms, and 1% were VT. There were 1756 shocks delivered during 542 of the 863 (63%) resuscitation attempts. The estimated sensitivity of paramedic-documented initial rhythms was 176/197 (0.89, 95% CI 0.84–0.93). The estimated specificity was 463/504 (0.92, 95% CI 0.89–0.94). 

There were 2 resuscitation attempts where VF was observed as the initial rhythm but no shock was delivered during the resuscitation. In both cases the paramedic interpretation of the initial rhythm was asystole. None of the 62 patients in either misclassified group survived to hospital discharge. However, survival with good neurologic function occurred in 36% of correctly classified shockable initial rhythms and 4% of correctly classified nonshockable rhythms. 

Misclassification of the initial rhythm was negatively associated with neurologically intact survival at hospital discharge (OR 0.05, 95% CI 0.003–0.82). Shockable initial rhythms (OR 8.07, 95% CI 4.28–15.20), witnessed arrests (OR 4.17, 95% CI 1.95–8.89), and cardiac arrests in public locations (OR 2.48, 95% CI 1.39–4.44) were positively associated with neurologically intact survival at hospital discharge. 

There were 60% of patients with correctly classified shockable initial rhythms who survived to hospital admission, but 16% of these patients had care withdrawn early. In patients with misclassified nonshockable initial rhythms, 29% survived to hospital admission, and half of these patients had early withdrawal of care. 

Conclusion

This was an important study that found that paramedics achieved high sensitivity for shock delivery to shockable rhythms. However, there was an 11% shock-delivery rate to PEA, and misclassification of initial rhythm was associated with poor survival. The authors noted that protocols to enhance accuracy of rhythm interpretation and speed of defibrillation attempts may improve outcomes. 

As with all studies, this one has some limitations. There may have been uncollected or unknown confounders that impacted paramedic rhythm interpretation accuracy and its association with outcomes. There were also likely differences in postarrest care that were not adjusted for. It is also important to note that the physicians who reviewed the rhythms did so in controlled environments, while paramedic interpretations occurred during the stress of cardiac arrest responses. Nevertheless, this is an important addition to the literature. 

Read this manuscript in its entirety for important information about intra-arrest shock accuracy and interrater agreement of physician rhythm interpretation we could not get to in this review.  

Antonio R. Fernandez, PhD, NRP, FAHA, is a research scientist at ESO and serves on the board of advisors of the Prehospital Care Research Forum at UCLA.