Describe the findings of three studies on the use of antidysrhythmic agents in cardiac arrest
Discuss the function of randomized controlled trials
Identify how research impacts protocol changes and development
You arrive on the scene of a 45-year-old male in cardiac arrest. He’d just bought a treadmill to get in shape and was taking his first run on it when he complained of crushing chest pain, then collapsed. His 20-year-old son witnessed the collapse and called 9-1-1. The son was instructed to begin compression-only CPR.
You find the patient pulseless and apneic. Your second unit arrives at the same time and begins pit-crew CPR with immediate compressions. Your partner applies the monitor’s multifunction pads as compressions are ongoing. Seeing that you are nearing the completion of your first cycle of compressions, she charges your monitor. As you come off the chest, she quickly analyzes the rhythm, determines the patient is in coarse ventricular fibrillation (VF) and delivers a shock. Another medic immediately resumes compressions without a pulse check. Your total time off the chest was five seconds.
As an IO is established, you discuss the next steps. You remember a discussion your department had recently about the effectiveness of antidysrhythmic agents.
Medications have long been one of the mainstays of cardiac arrest management.1 Some medications have been deemphasized or eliminated based on good evidence,2,3 some because of a lack of availability,4 and the use of any drugs—even the most standard of all, epinephrine—is being questioned.5,6 So what does the literature say about the use of antidysrhythmic agents in cardiac arrest?
While there have been several different antidysrhythmic medications that have come and gone over the years (bretylium, for example), lidocaine and amiodarone are the most common used today. There are several studies that can help us understand the evidence base for these agents.
Paper #1: Amiodarone for Resuscitation After Out-of-Hospital Cardiac Arrest Due to Ventricular Fibrillation7
This is the first of the well-done early trials on this subject. This was the ARREST trial, a single-center (Seattle/King County, WA), double-blind, placebo-controlled, randomized controlled trial (RCT) comparing amiodarone to placebo in adult, nontraumatic out-of-hospital cardiac arrest with refractory VF. Data was collected between November 1994 and February 1997. This was a manufacturer-sponsored study, and the author was on the speaker’s bureau of Wyeth-Ayerst, the manufacturer of amiodarone.
This study was powered to detect only a difference in survival to hospital admission, not neurologically intact survival. To demonstrate why this is, the authors determined they would need to enroll 500 patients to have sufficient power to detect a 30% difference in survival to hospital admission between those receiving amiodarone and placebo. That’s a pretty big difference. If they assumed a smaller difference, that would require enrolling more patients.
The system worked 3,954 cardiac arrests before it enrolled the 500 needed. This demonstrates why it is so hard for single-center studies to look at neurologic survival.
The study protocol involved giving a single study drug one time. Paramedics administered a nonlabeled syringe containing either 300 mg of amiodarone or a similar volume of placebo after three defibrillations. Stacked shocks were the standard approach to the initial defibrillation at the time.
Survival to hospital admission was better with amiodarone than with placebo (44% vs. 34%, p=0.03). The adjusted odds ratio (OR) for admission with amiodarone was 1.6 (95% confidence interval 1.1–2.4, p=0.02). The authors calculated they would need to treat 10 patients with amiodarone to achieve one hospital admission. This number needed to treat (NNT) of 10 is a very good value; unfortunately, it is for an outcome—survival to hospital admission—that has no value to the patient. These authors also evaluated the groups in terms of survival to hospital discharge, but found no significant difference (13.4% with amiodarone vs. 13.2% with placebo). The study was not powered to detect this outcome, so we shouldn’t read too much into this finding. There was also no difference in discharge rate in any of the subgroups evaluated.
This study was well done and provided convincing evidence that amiodarone was superior to placebo for survival to admission. Unfortunately there was no difference in survival to discharge, the outcome patients most care about.
Paper #2: Amiodarone as Compared With Lidocaine for Shock-Resistant Ventricular Fibrillation8
This was the ALIVE trial, a nicely done double-blinded RCT conducted in the Toronto EMS system between November 1995 and April 2001. It selected 347 consecutive adult patients with out-of-hospital, nontraumatic cardiac arrest found in VF/pulseless ventricular tachycardia (VT) who were still in a shockable rhythm after four shocks and administration of epinephrine. Patients were also included if they were initially in VF/VT, converted, but then had recurrent VF/VT. This trial was also sponsored by Wyeth-Ayerst.
Like the ARREST trial, the primary outcome was survival to hospital admission. The authors chose this outcome because they didn’t feel they could do a study large enough and with enough statistical power to detect the likely-small differences in more meaningful survival.
They found there was remarkably better survival to admission with amiodarone than with lidocaine: 22.8% vs. 12% with an OR of 2.17 (95% CI 1.21–3.83, p=0.009). Although not a primary outcome, they looked at survival to discharge with a modified Rankin score of 3 or below (0–3 is considered neurologically intact); they found no significant difference between amiodarone and lidocaine: 5% vs. 3%, p=0.34. Again, interpret this cautiously, since this study was not adequately powered to determine a difference this small. There were no differences in adverse events between the agents.
This was a well-designed study that showed amiodarone improved survival to hospital admission better than lidocaine. There were some significant limitations to the study, however. Most important, the outcome isn’t one we particularly care about. While it is nice to have more patients survive to admission, we want to see improved survival to a meaningful discharge. Unfortunately this study showed no benefit in this more important measure.
This study is also now more than 14 years old. Much has changed in how we address cardiac arrest. For example, we no longer recommend three stacked shocks. Currently we instruct medics to pay close attention to providing high-quality CPR, but this was not likely a focus when this study was done.
These two studies (ALIVE and ARREST) provided enough evidence of benefit to amiodarone that its use became prevalent in cardiac arrest. What was needed, though, was an RCT comparing both amiodarone and lidocaine to placebo in relation to neurologically intact survival. Fortunately that exact study was recently published.
Paper #3: Amiodarone, Lidocaine or Placebo in Out-of-Hospital Cardiac Arrest9
Dr. Peter Kudenchuk was the lead author on this large, federally funded study done through the Research Outcomes Consortium (ROC). This consortium includes 55 EMS agencies in the U.S. and Canada. The trial was conducted between 2012–2015 simultaneously with another trial that compared continuous vs. interrupted-compression CPR (and found no difference between the approaches).10 It was a randomized, placebo-controlled, double-blind trial comparing amiodarone, lidocaine and placebo. Since this trial was done very recently, providers were using current AHA guidelines. This trial also reported on compression metrics, including rate, depth, interruptions and compression fractions.
They enrolled adult patients with nontraumatic cardiac arrest who had an initial rhythm of VF/VT that was still present after one or more shocks. Their primary outcome was survival to hospital discharge. To achieve a power sufficient to detect a difference in this outcome, they needed to enroll 3,000 patients.
The main finding was that there was no overall advantage with either lidocaine or amiodarone compared with placebo or with lidocaine vs. amiodarone in terms of either survival to hospital discharge or neurologically intact survival.
There were some differences in secondary outcomes, though. Neither lidocaine nor amiodarone had higher admission rates than placebo. Lidocaine (but not amiodarone) had higher return of spontaneous circulation (ROSC) rates than placebo. There was a planned subgroup analysis based on whether the arrest was witnessed, and patients in this witnessed group had better survival to discharge with either amiodarone or lidocaine than placebo, but there was no significant differences between the two agents. Additionally there was more bradycardia that needed pacing in the amiodarone group than in either the lidocaine or placebo group.
This trial may, on first glance, appear to have different results than the first two trials, which both showed an advantage to using amiodarone. It turns out, though, that they don’t disagree. All three found an advantage to amiodarone in survival to hospital admission. They also agree there was no improvement in survival to hospital discharge. Although the first two trials weren’t powered to detect this more meaningful outcome, this larger trial was. Often, early, small trials will find impressive results, and subsequent larger trials see these improvements go away. This trial also was conducted in a time when there was a focus on high-quality, minimally interrupted compressions. The compression fraction in all arms of this study averaged 83%, which is likely much better than in the early studies, although it wasn’t possible to measure this in the earlier trials. Perhaps the higher-quality CPR reduced the potential effects of the drugs and poor-quality CPR might have produced a more noticeable effect with antidysrhythmics.
Overall this trial is the best evidence to date on the effectiveness of antidysrhythmics in cardiac arrest. We aren’t likely to see a better study than this, considering the expense involved.
So What Do We Know?
So what do we know based on this research?
We have seen that amiodarone and lidocaine promote survival to hospital admission but not to discharge or neurologically intact survival. We have some evidence from a small trial that amiodarone does this better than lidocaine, but this was not confirmed in a much larger trial. In a subgroup of patients with witnessed arrest, both amiodarone and lidocaine improved survival to discharge, but neither was better than the other. There was also no difference in effectiveness in unwitnessed arrest. There is also a hint, merely a suggestion, that the longer the code has been going on, the more effective the agents may be.
Given what we know, what should we do differently?
We know there is not, to say the least, strong evidence that any antidysrhythmic agent improves meaningful survival. With that in mind, for an agency that is resource-limited and faced with the choice of buying more amiodarone or lidocaine instead of spending that money on, say, training on high-quality minimally interrupted CPR or AEDs, the choice is clear: Stop using these drugs and spend the money on things you know matter more.
If, however, your system has the budget to purchase drugs without compromising other needed resources, it probably makes sense to keep one of these agents available. There are several reasons for this. First, the largest prospective trial we have demonstrated an improvement in meaningful survival with either of these agents in patients with witnessed arrests. These are the patients most likely to be salvageable. Second, VF and pulseless VT are not the only conditions we use these agents to treat. Wide-complex tachycardias with a pulse and hemodynamic stability still benefit from antidysrhythmics, although there is no clear evidence that either one of these agents is superior to the other.11,12
Third, we have seen that although there is no statically significant meaningful survival benefit with either of these agents, there is a trend toward improvement with either of the antidysrhythmics compared to placebo. If (and this if is nothing but hypothetical) this trend represents a real but small difference, it is possible the sample size in the 2016 Kudenchuck paper was too small to detect it with statistical significance. Even a small 3% difference (what was observed in this study), if present, could represent an additional 1,800 lives saved a year in the U.S. and Canada, according to a recent calculation by Dr. Ken Milne.13
Since there is no overall harm from these agents (other than an increase in pacing in the amiodarone group), a trend toward overall benefit and a benefit in one subset of patients, it probably makes sense to continue using one of these agents, at least in witnessed arrests.
Randomized Controlled Trials
A randomized controlled trial (RCT) is one in which the interventions being studied (amiodarone and lidocaine in this case) are randomly assigned to eligible patients. They idea behind randomization is that all other potential factors affecting the outcome should be roughly equal and the decision to give one intervention rather than the other isn’t influenced by some unmeasured variable.
As an example, let’s say we’re evaluating whether insulin improves outcomes. We might look at all patients and break them into two groups, those given insulin and those not, and then look at the difference in, say, hospital length of stay. We would likely see that those who got insulin had longer hospital stays than those who didn’t. We might be tempted to conclude that insulin was responsible for the longer stays.
The problem with this, of course, is the distinct possibility that the group that got insulin was different than the group that didn’t—likely in a way the study didn’t measure. For example, it could be that the group getting insulin was sicker, perhaps having diabetic ketoacidosis. It would be incorrect to assume the insulin caused the worse outcome because there was likely another reason those patients were given insulin—they were sicker. The insulin, in this example, is a marker for a sicker patient population that required a longer stay and not the cause of the longer stay.
Randomization controls for this by taking all patients with a given condition—say hyperglycemia—and randomly assigning each of them to either get insulin or a placebo, something that is harmless but has no therapeutic benefit. The assumption is that the drug they get is the only thing different between the two groups. If, as is likely in this scenario, the patients getting insulin have shorter length of stays, we would then be able to say the insulin was responsible for this because everything else was the same between the groups.
Although all nonrandomized variables between the groups should be the same, they often aren’t, just by happenstance. If you take any two groups of people and compare them on enough variables, you’ll find a significant difference between the groups in terms of at least some of those variables. To control for these random differences in groups, researchers apply some statistical tests and then rerun their statistics. These are typically referred to as adjustments. If, for example, we’re looking at mortality, the results would be referred to as the adjusted mortality. It is a valid way of taking known (but unplanned) differences in two groups into account.
RCTs are often also double-blinded. This means neither the patient nor the provider knows if it’s the study drug or placebo being given. If either group knew, it could affect their perception of benefit in a biased way.
Your patient remains in VF despite two shocks, epinephrine and high-quality, minimally interrupted CPR. You find a consistently elevated EtCO2 of 45. Knowing that patients with witnessed arrests have a survival benefit from antidysrhythmic agents, you administer amiodarone in keeping with your protocol for witnessed arrests. You obtain ROSC and, while carefully transporting him with a mechanical CPR device in place should he rearrest, remain vigilant for possible bradycardias that might require pacing.
1. Standards for cardiopulmonary resuscitation (CPR) and emergency cardiac care (ECC). 3. Advanced life support. JAMA, 1974 Feb 18; 227(7): Suppl: 852–60.
2. Coon GA, Clinton JE, Ruiz E. Use of atropine for brady-asystolic prehospital cardiac arrest. Ann Emerg Med, 1981 Sep; 10(9): 462–7.
3. Papastylianou A, Mentzelopoulos S. Current pharmacological advances in the treatment of cardiac arrest. Emerg Med Int, 2012; 2012: 815857.
5. Joglar JA, Page RL. Out-of-Hospital Cardiac Arrest—Are Drugs Ever the Answer? N Engl J Med, 2016 May 5; 374(18): 1,781–2.
6. Perkins GD, Quinn T, Deakin CD, et al. Pre-hospital Assessment of the Role of Adrenaline: Measuring the Effectiveness of Drug administration In Cardiac arrest (PARAMEDIC-2): Trial protocol. Resuscitation, 2016 Nov; 108: 75–81.
7. Kudenchuk PJ, Cobb LA, Copass MK, et al. Amiodarone for resuscitation after out-of-hospital cardiac arrest due to ventricular fibrillation. N Engl J Med, 1999 Sep 16; 341(12): 871–8.
8. Dorian P, Cass D, Schwartz B, et al. Amiodarone as compared with lidocaine for shock-resistant ventricular fibrillation. N Engl J Med, 2002 Mar 21; 346(12): 884–90.
9. Kudenchuk PJ, Daya M, Dorian P; Resuscitation Outcomes Consortium Investigators. Amiodarone, lidocaine or placebo in out-of-hospital cardiac arrest. N Engl J Med, 2016 Aug 25; 375(8): 802–3.
10. Nichol G, Leroux B, Wang H, et al. Trial of Continuous or Interrupted Chest Compressions during CPR. N Engl J Med, 2015 Dec 3; 373(23): 2,203–14.
11. deSouza IS, Martindale JL, Sinert R. Antidysrhythmic drug therapy for the termination of stable, monomorphic ventricular tachycardia: a systematic review. Emerg Med J, 2015 Feb; 32(2): 161–7.
12. Long B, Koyfman A. Best Clinical Practice: Emergency Medicine Management of Stable Monomorphic Ventricular Tachycardia. J Emerg Med, 2016 Oct 14.
Jeffrey L. Jarvis, MD, MS, EMT-P, FACEP, FABEMS, is EMS medical director for the Williamson County EMS system and Marble Falls Area EMS and an emergency physician at Baylor Scott & White Hospital in Round Rock, TX. He is board-certified in emergency medicine and EMS. He began his career as a paramedic with Williamson County EMS in 1988 and continues to maintain his paramedic license.
Erin Lincoln, MS, NRP, has been in EMS since 2003, when she earned her EMT certification with Rice University EMS. She is currently pursuing her MD as a member of the Texas A&M College of Medicine Class of 2019. She can be reached at firstname.lastname@example.org or follow on Twitter at @ErinTWL.