Prove it: Drugs unnecessary during resuscitation attempts

By Kenny Navarro

Highlights:
The Scenario
The Review
What It Means for You

The Scenario

Medic 803 responds to a reported chest pain in a quiet residential neighborhood near the downtown area. The house is only about six blocks from one of the city’s trauma centers. About three minutes before the medic unit arrives on scene, the fire department first responders report that cardiopulmonary resuscitation is in progress.

Once in the house, the two medics find a male patient, age 68, lying on the floor. Two firefighters are performing CPR while the engine officer is comforting the patient’s wife. She says her husband has not felt well all day but did not want to go to the doctor’s office.

The firefighters report that the automated external defibrillator delivered a single shock just before the ambulance arrived. While CPR continues, the medic attaches the manual monitor/defibrillator and recognizes that the patient remains in ventricular fibrillation. While the ambulance EMT prepares the patient for transport, the medic quickly inserts a supraglottic airway.

There are no easily accessible intravenous sites, so the medic opts to defer IV access and medication administration until arrival at the emergency department.

Within five minutes of ambulance arrival, the crew loads the patient into the back of the ambulance. En route, the medic and the two firefighters continue to provide high-quality CPR and defibrillation attempts. After a three-minute transport interval, the patient arrives with CPR in progress and V-fib still present on the monitor. The ED staff establishes an IV, administers epinephrine and amiodarone, and continues the resuscitation attempt.

After the first ED countershock, the patient converts to asystole. The resuscitation attempt continues for another 30 minutes; however, the patient does not respond. At about one hour after the patient collapsed, the resuscitation attempt ends.

Later in the station, the EMT wonders if the patient could have survived if the medic had taken the time to establish an IV in the patient’s house and administered the advanced cardiac life support medications sooner.

The medic believes he made the right call because he knew establishing IV access would be difficult and the transport time was very short. The EMT is still not convinced and thinks that administering the medications in the house would produce better outcomes than deferring medication administration.

The Review

Intravenous Drug Administration During Out-of-Hospital Cardiac Arrest - Olasveengen, Sunde, Brunborg, Thowsen, Steen, & Wik, 2009

Researchers in Oslo, Norway, (Olasveengen, Sunde, Brunborg, Thowsen, Steen, & Wik, 2009) compared outcomes from a group of out-of-hospital cardiac arrest patients who received IV access and ACLS medications to a group who did not receive IV access or ACLS medications during the resuscitation attempt.

The EMS system in Oslo is a single-tiered system serving slightly more than 500,000 people. Two paramedics staff the ambulances 24 hours a day, although during the daytime hours of the workweek, an anesthesiologist serves as a third crewmember. Dispatchers, upon suspicion of a cardiac arrest, routinely dispatch two ambulances to the scene.

If the medics encounter V-fib, treatment guidelines require them to perform CPR for three minutes before attempting to defibrillate. Slightly different from American resuscitation standards, CPR cycles in Oslo consist of three-minute intervals resulting in a defibrillation attempt every three minutes for refractory V-fib.

In addition, every out-of-hospital cardiac arrest patient who achieves a return of spontaneous circulation (ROSC) receives therapeutic hypothermia regardless of the presenting arrhythmia or the arrest etiology.

In this study, researchers included all out-of-hospital cardiac arrest patients of non-traumatic etiology over the age of 18 years. Researchers excluded patients if:

• The EMS crew witnessed the cardiac arrest;
• Physicians who were not part of the ambulance team initiated or interrupted the resuscitation attempt;
• Asthma or anaphylactic shock was the presumed cause of the cardiac arrest.

Once the crew arrived and confirmed the absence of a pulse, one medic initiated CPR while the other medic opened a sealed envelope provided by the researchers. Based on the content of the envelope, medics would either establish IV access and administer standard ACLS medications (IV group) or withhold the IV and the medications (no IV group).

However, if a patient assigned to the no IV group regained a pulse, the medics could establish IV access five minutes after ROSC and administer post-resuscitation medications, as needed.

Researchers refer to this process of determining the patient group assignments as randomization. Randomization is important in clinical trials because it helps to reduce inadvertent bias or error from affecting the outcome.

For example, suppose instead of this sealed envelope randomization process, the researchers allowed the medics to determine which treatment group they wanted for their patient. Medics who have a bias toward ACLS drugs might assign younger V-fib patients to the IV group while assigning older asystolic patients to the no IV group, thinking that those patients would die anyway regardless of the ACLS drugs.

Group assignment made using this process would virtually assure that ROSC occurred more often in the IV group while death occurred more often in the no IV group. Instead, using a formal randomization protocol reduces assignment bias and provides greater confidence that any difference between the study groups is the result of the study variable and not error.

Variables for measurement
Researchers established many variables for measurement, including patient demographics, information about the arrest (i.e., location, bystander witnessed, etc.), EMS response intervals, CPR quality measurements (i.e., chest compression rate, pauses, and ventilation rate, etc.), and survival end-points including survival to hospital admission, one-year survival and neurological outcome.

The primary endpoint for the study was survival to hospital discharge. A primary endpoint represents the main measurement being studied (National Cancer Institute, n.d.).

Researchers monitored CPR quality using transthoracic impedance signals provided by the monitor/defibrillators. With this data, researchers could determine if one group of patients received better quality CPR, which obviously would affect the outcomes of each patient group and the study results.

During the five-year study period, the Oslo EMS system attempted to resuscitate 1183 patients. Of those, researchers excluded 267 patients from the analysis either because the patient did not meet the inclusion criteria or because the medics did not or could not properly randomize them.

Of the 916 patients randomized into the comparison groups, researchers excluded an additional 65 patients from the primary analysis because they met predefined exclusion criteria.

The researchers set the criteria for statistical significance, or p value, at 0.05. When p values of a characteristic measured in both groups are less than or equal to 0.05, researchers consider the difference statistically significant and therefore not likely to have been caused by chance.

If the p value is greater than 0.05, the differences are not statistically significant and could be the result of chance alone.

The randomization process placed 433 patients into the IV group and 418 patients into the no IV group. The medics provided similar quality CPR to both groups. There were no statistical differences in the hands-off ratio (time without compressions during time without spontaneous circulation), number of chest compressions, or the preshock pause. Compression and ventilation rates for both groups fell within the recommendations of the 2005 European Resuscitation Council Guidelines for Resuscitation.

Table 1 presents some of the statistically significant and important study results. More patients in the IV group required defibrillation attempts, received a greater number of shocks per patient, and received CPR for a longer time interval compared to the no IV group.

More patients in the IV group achieved ROSC during the resuscitation, survived long enough for hospital admission, and survived long enough for ICU admission.

However, the researchers designed this trial in an attempt to answer one specific question. Based on the data and the goal of the study, the researchers concluded that there was no statistical significance in survival to hospital discharge between the cardiac arrest patients who received an IV and ACLS medication and those who did not.

What It Means for You

The American Heart Association (2005) 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 not perfect, this well-designed trial provides considerable insight into the role that ACLS medications play in resuscitation attempts for out-of-hospital cardiac arrest patients.

This study did not demonstrate any long-term survival improvements in cardiac arrest patients caused by out-of-hospital ACLS drug administration. Medics often argue that long-term survival cannot occur without short-term survival occurring first.

As this study did demonstrate an association between drug administration and short-term outcome improvements, many may interpret this finding as “proof” that ACLS medication administration does work. However, one must very careful when interpreting secondary outcome data.

One question
Researchers design most clinical trials to try to answer one question: what is the difference between two groups of test subjects if there is manipulation of one variable? 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, 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.

However, secondary outcome data does not represent wasted effort. Even though one must be careful not to infer a difference that does not truly exist, these secondary outcome statistics provide fertile ground for generating new hypotheses that can be tested in future clinical trials.

One limitation of the study is that the researchers could not prevent EMS personnel from knowing the treatment group assignments for each patient once randomization occurred, a technique called blinding. Without blinding to the treatment groups, there is always a possibility that medics could unintentionally perform differently with one group, thereby creating an uncontrolled variable that could affect the outcome of the study.

For example, suppose a medic believes that all cardiac arrest victims must have epinephrine administration in order to achieve ROSC. Upon learning that the randomization process assigns the patient to the no IV group, the medic may begin the resuscitation attempt with the preconceived notion that the patient cannot possibly survive.

This could affect the quality of the CPR or the timing of the defibrillation shocks. Although the data suggested the quality of CPR between the two groups was similar, that quality data was available for only three-fourths of the patients enrolled in the study.

Protocol compliance
Closely related is the issue of treatment protocol compliance. The authors did not provide details on whether the medics followed the recommended drug regimens for patients assigned to the IV group.

Other researchers have observed that medics may not follow ACLS medication administration recommendations in 86 percent of out-of-hospital cardiac arrests (Scliopou, Mader, Durkin, & Stevens, 2006). Similar studies observed a 35 percent non-compliance rate among emergency department physicians in actual resuscitation cases (Cline, Welch, Cline, & Brown, 1995), while 61 percent of anesthesiologists had major deviations away from ACLS recommendations in a simulated operating room resuscitation event (Kurrek, Devitt, & Cohen, 1998).

Finally, a mathematical formula called a power calculation can help researchers determine how many patients they need to enroll in each group in order to demonstrate statistically a difference between the two groups, if a difference truly exists. To make this calculation, researchers must provide their best educated guess as to what they think the difference will be. Based on the available literature and historical resuscitation rates in Oslo, the researchers thought the difference would be 100 percent (twice as many survivors in the IV group than in the no IV group).

Instead, the difference between the two groups was only 1.3 percent. In order to prove that a difference of that magnitude is statistically significant, the researchers would need to enroll 14,000 patients instead of the 900 they thought they would need. So, while they could not demonstrate a statistical differences between the two groups when measuring survival to hospital discharge, it could be because they only enrolled about 6 percent of the patients they would actually need for such a study.

While this study is one of the most important resuscitation studies in the past decade, it still does not provide a definitive answer to the question of whether ACLS medications improve long-term survival rates for out-of-hospital cardiac arrest patients. In the scenario, the medic’s deferment of medication administration seemed prudent, given the proximity to the hospital and the recognition of the difficulty in establishing IV access.

References

American Heart Association. (2005). Advanced cardiac life support guidelines, part 4: Advanced life support. Circulation, 112, III-25 - III-54.
Cline, D. M., Welch, K. J., Cline, L. S., & Brown, C. K. (1995). Physician compliance with advanced cardiac life support guidelines. Annals of Emergency Medicine, 25, 52-57.

Kurrek, M. M., Devitt, J. H., & Cohen, M. (1998). Cardiac arrest in the OR: How are our ACLS skills. Canadian Journal of Anaesthesia, 45, 130-132.
National Cancer Institute. (n.d.). Dictionary of cancer terms. Retrieved February 13, 2010 from www.cancer.gov/dictionary/?CdrID=44163.

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, 2222-2229.

Scliopou, J., Mader, T. J., Durkin, L., & Stevens, M. (2006). Paramedic compliance with ACLS epinephrine guidelines in out-of-hospital cardiac arrest. Prehospital Emergency Care, 10, 394-396.

The author has no financial interest, arrangement, or direct affiliation with any corporation that has a direct interest in the subject matter of this presentation, including manufacturer(s) of any products or provider(s) of services mentioned.