Without evidence of benefit, an intervention should not be presumed to be beneficial or safe.

- Rogue Medic

What is Evidence-Based Medicine?

First, Evidence-Based Medicine is a confusing term. Ironically, many people have different interpretations of what Evidence-Based means.

The whole idea of scientific evidence is to minimize the interference of bias, which is much more difficult than it seems. We all have biases. When our biases affect our treatment of our patients, there is a lot of potential for harm.

If I believe that a treatment is effective, but there is no good evidence that it improves outcomes –


Outcomes are what matter, not surrogate endpoints.

A surrogate endpoint is something that is measured, because it is easy to measure, not because it is important. In treating patients with antiarrhythmic medication, it is much easier to assess the frequency of PVCs (Premature Ventricular Contractions), than it is to assess survival. This was demonstrated very painfully in CAST (the Cardiac Arrhythmia Suppression Trial).[1],[2]


After a heart attack, if the patient has frequent PVCs, the patient is more likely to die.

Therefore, getting rid of the PVCs saves lives.

That is not an unreasonable conclusion.

The PVCs may progress to something worse, such as V Tach (Ventricular Tachycardia) or V Fib (Ventricular Fibrillation). Preventing these rhythms is important. We use antiarrhythmic drugs and/or implanted defibrillators for this reason.

But the conclusion assumes too much.

Getting rid of the PVCs does not saves lives.

Fiddling with the conduction system is doing something, but why should we assume that it is doing something good, unless we have good evidence that it improves outcomes?

PVCs are an indicator of an injured heart. Just an indicator. If the pulse oximeter alarm is going off because the patient’s oxygen saturation is low, turning the alarm off, or turning the pulse oximeter off is not going to make things better. If the patient is oxygenating adequately, regardless of the alarm settings, treating the false hypoxia would not be expected to produce a good outcome.

We used to automatically give lidocaine to chest pain patients to prevent V Fib.

We have to do something.

It’s in the algorithm.

What if there is a bad outcome?

We do not give lidocaine for chest pain any more.


Because almost every treatment is eventually discarded. Some are replaced by safer versions of the same treatment, but maybe even more are found to be harmful.

Lidocaine for chest pain was not based on any evidence of improved outcomes.

Lidocaine for chest pain, outside of a controlled trial, was a bad idea. This was just an uncontrolled, undocumented, experiment on unsuspecting patients without any explanation that the risks and benefits were unknown.

In CAST the patients receiving the study drugs encainide (Enkaid) and flecainide (Tambocor) were more than 3 times as likely to die as the patients receiving placebo.

Since these drugs were some of the most widely prescribed drugs at that time, probably tens of thousands of patients died because of these drugs.

Today, we use epinephrine in cardiac arrest because it improves outcome in dogs and rats who have cardiac arrest induced in the laboratory. Great news if you are a dog, or a rat, and EMS shows up to treat your cardiac arrest.

We also can show that epinephrine improves ROSC (the Return Of Spontaneous Circulation), but this improvement goes away before discharge from the hospital.

In other words, epinephrine only has good evidence of a surrogate endpoint – ROSC. With only evidence of improvement in producing a surrogate endpoint, epinephrine should never have been approved for use outside of controlled trials.

None of the evidence on epinephrine outcomes is well controlled.

There are two interpretations of the epinephrine evidence.

1. Epinephrine is harmful – in which case we never should have been giving epinephrine in cardiac arrest.

2. Epinephrine is neutral (it produces no improvement in outcomes) – in which case we never should have been giving epinephrine in cardiac arrest.


I added the two most recent studies to this, since they were not yet published when the review of vasopressors was published.[3] The quality and outcome of the added studies is my interpretation, but I think that others will rank them similarly.

The Hagihara study is more than 100 times larger than the combined numbers of all of the studies that came before it.

The Hagihara study is not positive.

The Hagihara study is not even neutral.

How much negative evidence do we need to make excuses for to keep using epinephrine in cardiac arrest?

If we ignore the evidence, then there is no limitation on what we can do for/to patients.

We can give cyanide for cardiac arrest, because maybe the homeopaths are on to something with their law of similars.

We can give medical marijuana, probably by positive pressure ventilation (being very careful not to inhale any ourselves).

We can give actual voodoo, which does have some scientific evidence, but not outcomes evidence and it is used to put people into a state that looks like cardiac arrest, but is not cardiac arrest.

There are no limits to what we can do to patients, if we do not base treatments on evidence.

We have to stop pretending that we First do no harm. Without evidence, we don’t know and we don’t want to know. Trust us, we don’t stop until we cannot ignore the evidence of harm.

If we are satisfied with surrogate endpoints and think that an expensive coma is better than whatever the outcome would have been without epinephrine, then we are just playing with the results of catecholamine toxicity and hoping we will find something that works as well as not using epinephrine.

What about evidence for using a long spine board and rigid EMS collars for spinal immobilization?

There is evidence of worse outcomes. There is no evidence of improved outcomes.

What about high flow oxygen for everything?

There is evidence of worse outcomes. There is no evidence of improved outcomes.

Too much of what we do can be described that way –

There is evidence of worse outcomes. There is no evidence of improved outcomes.

We need to grow up and start treating our patients with respect. They are not toys.

We are practicing witchcraft, not medicine.

We are pretending mythology is medicine.

Why is it ironic that many people have different interpretations of what Evidence-Based Medicine means?

Because reliance on scientific evidence is supposed to be about eliminating the influence of interpretation. If one person can read a paper and come to one conclusion, another person can come to the opposite conclusion, and a third person can come to a conclusion that differs with both of the others, and so on – then there is not enough well controlled research to justify subjecting patients to the risks of the proposed treatment.

Without clear evidence of benefit, we should only be using a treatment in controlled trials – otherwise we may never find out if the treatment is more harmful than beneficial.


[1] Mortality and morbidity in patients receiving encainide, flecainide, or placebo. The Cardiac Arrhythmia Suppression Trial.
Echt DS, Liebson PR, Mitchell LB, Peters RW, Obias-Manno D, Barker AH, Arensberg D, Baker A, Friedman L, Greene HL, et al.
N Engl J Med. 1991 Mar 21;324(12):781-8.
PMID: 1900101 [PubMed – indexed for MEDLINE]

Free Full Text Article from New England Journal of Medicine.

CONCLUSIONS. There was an excess of deaths due to arrhythmia and deaths due to shock after acute recurrent myocardial infarction in patients treated with encainide or flecainide. Nonlethal events, however, were equally distributed between the active-drug and placebo groups. The mechanisms underlying the excess mortality during treatment with encainide or flecainide remain unknown.

[2] C A S T and Narrative Fallacy
Rogue Medic
Mon, 20 Jul 2009

[3] Vasopressors in cardiac arrest: a systematic review.
Larabee TM, Liu KY, Campbell JA, Little CM.
Resuscitation. 2012 Aug;83(8):932-9. Epub 2012 Mar 15.
PMID: 22425731 [PubMed – in process]

CONCLUSION: There are few studies that compare vasopressors to placebo in resuscitation from cardiac arrest. Epinephrine is associated with improvement in short term survival outcomes as compared to placebo, but no long-term survival benefit has been demonstrated. Vasopressin is equivalent for use as an initial vasopressor when compared to epinephrine during resuscitation from cardiac arrest. There is a short-term, but no long-term, survival benefit when using high dose vs. standard dose epinephrine during resuscitation from cardiac arrest. There are no alternative vasopressors that provide a long-term survival benefit when compared to epinephrine. There is limited data on the use of vasopressors in the pediatric population.