Prove it: Epinephrine administration improves outcome following cardiac arrest
During the 39-month study period, paramedics treated 4,130 out-of-hospital cardiac arrest patients
After a short response interval, Engine and Medic 57 arrive at a house in a quiet residential neighborhood near the downtown area. The caller reported a man having difficulty breathing. Inside the house, the rescuers find a male, age 57, lying on the floor. The patient is taking agonal breaths but does not have a palpable pulse. Two firefighters begin cardiopulmonary resuscitation (CPR) while the medics attach the monitor and begin preparing the advanced life support equipment.
The electrocardiogram (ECG) reveals ventricular fibrillation (VF). Medic Thompson immediately delivers a 150 joule countershock; the firefighters switch positions and resume CPR. Medic Davis quickly inserts a supraglottic airway and confirms placement and effective ventilation with waveform capnography. Medic Thompson is having difficulty establishing intravenous (IV) access.
At the two-minute mark, Medic Davis confirms VF and delivers a second countershock. During the ensuing two-minute period of CPR, both medics continue to attempt IV access but are unsuccessful. The Engine officer informs the medics that the two-minute CPR cycle is almost over. The patient remains in VF and receives a third countershock, after which, the firefighters switch positions and resume CPR. The crew spends the next two-minute period placing the patient onto a backboard and cot. Before moving the patient to the ambulance, Medic Davis delivers a fourth countershock.
At the next two-minute mark, Medic Davis notes that the patient is asystolic. Both medics continue to make unsuccessful attempts at gaining IV access during the short ride to the hospital. The emergency department (ED) team continues resuscitation attempts for ten additional minutes but the patient fails to respond.
Later in the station, both medics wonder if the patient could have survived if they could have administered intravenous epinephrine.
Researchers in Western Australia conducted the first randomized placebo-controlled clinical trial of epinephrine administration during resuscitation from cardiac arrest (Jacobs, Finn, Jelinek, Oxer, & Thompson, 2011). The research community considers randomized-controlled trials to be the gold standard for clinical research.
Paramedics at St. John’s Ambulance of Western Australia and other healthcare providers in that country follow cardiac arrest resuscitation guidelines established by the Australian Resuscitation Council (ARC) (Australian Resuscitation Council, 2011), who, along with the American Heart Association are member agencies of the International Liaison Committee on Resuscitation (ILCOR). Historically, ARC guidelines have not included medication administration for victims of cardiac arrest, owning to the lack of efficacy evidence for resuscitation drugs. Epinephrine administration was a new intervention for paramedics in this study.
Although researchers measure many different variables when collecting data, the investigation is designed to try to answer one specific question. In this case, researchers wanted to know if administering epinephrine results in more patients regaining a pulse and surviving long enough for physicians to discharge them from the hospital.
The research team will collect other data points such as gender of the patient, whether the arrest was witnessed, and the presenting EMS rhythm. However, the primary outcome measurement for this study was survival to hospital discharge. This variable determines the answer to the research question.
The researchers and statisticians choose all of the study methods and statistical analysis tests based on the one question. That one question even determines how many test subjects are required in order to find the difference between the two groups, if it even exists at all. If the researcher chooses a different question or a different variable, the study designers would need to reevaluate the data collection methods, employ new statistical tests, or alter the number of patients needed to detect a difference.
The researchers identified two secondary outcome measurements. Secondary outcome measurements are useful for determining the direction of future research projects, but readers must be careful about drawing definitive conclusions about those measures because the study was not designed to answer anything except the primary question.
One secondary outcome measure was prehospital ROSC, which researchers defined as return of a palpable pulse for greater than 30 seconds. The other was the patient’s Cerebral Performance Category (CPC) score at hospital discharge. The CPC divides patients into five categories based on their neurological function;
- Category 1 – normal function
- Category 2 – mild to moderate disability
- Category 3 – severe disability
- Category 4 – vegetative state
- Category 5 – dead
Researchers included all patients age 18 or older who suffered a cardiac arrest from any cause and received a resuscitation attempt by paramedics. Drug administration involved a vial that contained either epinephrine or saline. The label on the vial contained only a tracking number so that neither the medics nor subsequent healthcare providers would know if the patient received epinephrine or placebo.
If the patient presented in pulseless electrical activity (PEA) or asystole, the medics administered the study solution as soon as possible after establishing IV access. If the patient presented in ventricular fibrillation (VF) or pulseless ventricular tachycardia (pVT), the medics did not administer the solution until after delivering the third shock.
This strategy prevented study enrollment of patients who achieved early return of spontaneous circulation (ROSC), which could skew the results. Although the researchers measured the primary outcome variable using data from all the patients, the authors pre-planned to perform a subgroup analysis to differentiate primary and secondary outcome between shockable and non-shockable rhythms.
The medics administered a single dose of the study solution every three minutes until the patient achieved ROSC or received a maximum of ten doses. After administering the study solution, the medics flushed the IV line either by allowing the fluid to free flow or by administering a 30-ml saline bolus. The medics did not administer any other resuscitation drugs.
If the patient failed to respond to treatment and remained asystolic for at least 20 minutes, the EMS medical protocols allowed the paramedics to terminate further resuscitation efforts in the field. If the medics transported, the hospital staff provided subsequent care in accordance with the ARC guidelines however, they did not know the contents of the study vial.
All paramedics underwent a training program before beginning the trial. The training program included familiarization with the pharmacology of epinephrine, the study protocol, refresher training on IV cannulation, and cardiac arrest megacode simulation.
Before the trial began, the research team performed a statistical analysis called a power calculation to determine how many patients they would need to enroll in order to show a difference in treatment if a difference really existed. Based on outcome history in that area, researchers estimated that 5% of the cardiac arrest patients would survive long enough to be discharged from the hospital.
They also estimated that epinephrine administration would result in a 2% survival improvement. Using those two estimates, the calculation determined that researchers would need 2213 patients per group to be able to show this 2% improvement. In order to account for missing data that often appears in clinical trials, the researchers planned to include about 5000 patients.
The researchers initially intended to include several EMS agencies in the trial. Unfortunately, all but one agency dropped out of the study. Researchers collected patient care data from the paper documents used by the EMS agency. The State maintains a patient care database that researchers used to obtain outcome data. Hospital records provided the CPC status at discharge. Researchers obtained the dispatch and response time points from the ambulance computer-aided dispatch (CAD) system.
During the 39-month study period, paramedics encountered 4130 out of hospital cardiac arrest patients. Researchers excluded the overwhelming majority of these patients from analysis for a variety of reasons, including field pronouncement without a resuscitation attempt.
The medics randomized 601 patients to the two study groups; however, the researchers had to exclude an additional 67 of these because the medics failed to record the randomization number (identification number on the medication vial). This left 534 patients for the study sample; 262 patients randomized to the placebo group and 272 patients randomized to the epinephrine group.
There was no difference between the groups with respect to age, gender, location of arrest, cardiac etiology, whether the arrest was witnessed, bystander CPR, initial arrest rhythm, ambulance response interval, or airway management technique. Medics transported more of the patients in the epinephrine group than the placebo group and this difference met statistical significance (88.6% vs. 82.1%; p = 0.03).
The authors reported the outcome data as odds ratios, a numerical representation of the odds of achieving ROSC for patients receiving epinephrine divided by the odds of achieving ROSC in the patients who receive placebo. This is similar, but not the same as probability. When the ratio is one, the odds that one group will achieve ROSC is the same as in the other group. A ratio greater than one means the odds of achieving ROSC is greater for patients who receive epinephrine (Goldin, 2007).
Before reporting the actual results, one must understand that although often appearing as precise numbers, all measurement outcomes involve some degree of error. This could include factors such as mathematical rounding error and measurement error. Therefore, researchers often report a 95% confidence interval for their measurements.
This means that if it the research team could eliminate all sources of error, the true odds of achieving ROSC after receiving epinephrine would be within this interval 95% of the time. If the interval includes the whole number one statisticians generally consider the odds equivalent between the two groups, even if the odds for that group are greater than one.
The odds of achieving ROSC in the field were 3.4 times greater for patients who received epinephrine than patients who received the placebo. The researchers are 95% confident that the true odds of achieving ROSC in the field for patients who receive epinephrine are between 2.0 and 5.6 times greater. In addition, the odds of surviving long enough or admission to the hospital were almost two and a-half times greater if the patients received epinephrine in the field.
However, both of these data points are secondary outcome measures, which are useful for informational purposes only. Remember, the researchers designed this investigation to answer a different question – Does epinephrine administration in the field improve survival to hospital discharge rates?
At first glance, it appears the odds of surviving long enough for hospital discharge doubles for patients who receive epinephrine in the field. However, because of uncontrollable measurement error, the actual odds could be anywhere between 0.7 and 6.3. Because the first number in the interval falls below one, we cannot state with any certainty that the odds are better for patients who receive epinephrine compared to those who receive placebo. In a subgroup analysis of shockable versus non-shockable rhythms, epinephrine administration did not significantly improve the odds of survival to hospital discharge for either group.
Clinicians have long debated the value of drug administration during resuscitation attempts for patents who suffer cardiac arrest. Adding epinephrine to an EMS system that did not previously administer IV resuscitation medication did not produce an increase in survival to hospital discharge (Ong et al., 2007). A Norwegian study of adult cardiac arrest published in 2009 could not demonstrate a statistically significant improvement in long-term survival for patients who received an IV and ACLS medication compared to those who did not (Olasveengen et al., 2009).
The American Heart Association (Neumar et al., 2010) acknowledges that there is insufficient evidence to support or refute the administration of any sequence of medications to human cardiac arrest victims with the intent of increasing survival to hospital discharge. Although this study did show that prehospital epinephrine produced significant short term-benefits, it could not demonstrate any long-term survival benefits.
One limitation of this study is that the researchers included all patients who suffered cardiac arrest, regardless of their presenting rhythm. Survival for patients presenting in asystole and PEA is dismal (Deasy et al., 2011; Survey of Survivors After Out-of-hospital Cardiac Arrest in KANTO Area, Japan (SOS-KANTO) Study Group, 2011). It is reasonable to assume that any study that includes subsets of patients who researchers expect to die will limit their ability to show value in any intervention.
Although a Swedish study of cardiac arrest that limited the analysis to patients who presented in ventricular fibrillation could not find survival benefits from epinephrine administration, medics in that study administered epinephrine in doses far lower than recommended by the ILCOR (Herlitz et al., 1995). Perhaps a resuscitation study designed to include only patients with a reasonable chance of survival (shockable rhythms) who receive the recommended dose of epinephrine would find different results.
Although researchers performed a power analysis to determine how many patients they would need in order to detect a survival difference, four out of the five agencies originally involved in the study dropped out because of the pressures they faced in using an unproven therapy. In addition, the authors report that press reports during the investigation questioned the ethics of the study, which hampered the researcher’s ability to convince the agencies to return to the study.
What was a reasonable enrollment goal for five agencies became unattainable for one agency, given time and financial constraints. As a result, the authors could not enroll enough patients to detect long-term survival advantage if they truly existed.
Another limitation (one the authors fully admit) is that they did not measure CPR quality in the two groups. Without objective measurement, there is no way to know if medics performed differently for one group compared to the other, thereby creating an uncontrolled variable that could affect the outcome of the study. One could make the argument that blinding the medics as to the vial contents reduces the chances of introducing bias into one particular group; however there is no way to know for sure without measuring the variable.
The results of this first double blind, randomized, and controlled trial of epinephrine in a human population adds to the growing body of evidence that suggests that epinephrine administration does not improve long-term survival rates for patients who suffer an out-of-hospital cardiac arrest.
Secondary outcome measurements suggest the drug may be beneficial in improving short-term outcome, but ultimately, the goal is to return patients to a normal or near-normal quality of life. The results did not demonstrate harm resulting from epinephrine administration; therefore, one could argue that medics should continue to administer the drug. Epinephrine MAY have value for patients in cardiac arrest. At this point, the medical community is still unclear on what that value might be.
Australian Resuscitation Council. (2011). Advanced life support guidelines. retrieved from www.resus.org.au
Deasy, C., Bray, J. E., Smith, K., Wolfe, R., Harriss, L. R., Bernard, S. A., & Cameron, P. (2011). Cardiac arrest outcomes before and after the 2005 resuscitation guidelines implementation: evidence of improvement? Resuscitation, 82(8), 984-988. doi:10.1016/j.resuscitation.2011.04.005
Goldin, R. (2007). Odds ratio. Retrieved from http://stats.org/stories/2008/odds_ratios_april4_2008.html
Herlitz, J., Ekstrom, L., Wennerblom, B., Axelsson, A., Bang, A., & Holmberg, S. (1995). Adrenaline in out-of-hospital ventricular fibrillation. Does it make any difference? Resuscitation, 293, 195–201.
Jacobs, I. G., Finn, J. C., Jelinek, G. A., Oxer, H. F., & Thompson, P. L. (2011). Effect of adrenaline on survival in out-of-hospital cardiac arrest: A randomized double-blind placebo-controlled trial. Resuscitation, 82, 1138–1143. doi:10.1016/j.resuscitation.2011.06.029
Neumar, R. W., Otto, C. W., Link, M. S., Kronick, S. L., Shuster, M., Callaway, C. W., Kudenchuk, P. J., Ornato, J. P., McNally, B., Silvers, S. M., Passman, R. S., White, R. D., Hess, E. P., Tang, W., Davis, D., Sinz, E., & Morrison, L. J. (2010). Part 8: Adult advanced cardiovascular life support: 2010 American Heart Association guidelines for cardiopulmonary resuscitation and emergency cardiovascular care. Circulation, 122[suppl 3], S729–S767. doi:10.1161/CIRCULATIONAHA.110.970988
Olasveengen, T. M., Sunde, K., Brunborg, C., Thowsen, J., Steen, P. A., & Wik, L. (2009). Intravenous drug administration during out-of-hospital cardiac arrest: A randomized trial. Journal of the American Medical Association, 302(20), 2222-2229. doi:10.1001/jama.2009.1729
Ong, M. E., Tan, E. H., Ng, F. S., Panchalingham, A., Lim, S. H., Manning, P. G., Ong, V. Y. K., Lim, S. H. C., Yap, S., Tham, L P., Ng, K S., and Venkataraman, A. (2007). Survival outcomes with the introduction of intravenous epinephrine in the management of out-of-hospital cardiac arrest. Annals of Emergency Medicine, 50(6), 635–642. doi:10.1016/j.annemergmed.2007.03.028
Survey of Survivors After Out-of-hospital Cardiac Arrest in KANTO Area, Japan (SOS-KANTO) Study Group. (2011). Atropine sulfate for patients with out-of-hospital cardiac arrest due to asystole and pulseless electrical activity. Circulation Journal, 75(3), 580-588. doi: 10.1253/circj.CJ-10-0485