Epinephrine has been part of cardiac arrest resuscitation since the 1960s. For most of that time, the rationale was straightforward: alpha-adrenergic stimulation increases coronary and cerebral perfusion pressure during CPR. More perfusion, better outcomes.
The 2025 AHA Guidelines for CPR and ECC left epinephrine itself almost entirely alone — and changed something next to it that matters just as much. Here's what moved and what didn't.
"For shockable rhythms, electricity comes first. The vasopressor waits its turn."
The key numbers
What did not change
The headline for anyone recertifying: the dose, the interval and the strength of the recommendation are all the same. One milligram, IV or IO, every three to five minutes, as a Class 1 recommendation. If you have been giving epinephrine correctly, you still are.
The goal has not changed either. Epinephrine's job during CPR is to raise coronary perfusion pressure and support cerebral perfusion. Absent a practical way to work out mid-arrest which patients will benefit, it remains standard therapy in cardiac arrest.
Timing still depends on the rhythm
For non-shockable rhythms — asystole and PEA — give epinephrine as early as access allows. Nothing about those rhythms is going to be corrected by electricity, so there is no competing intervention and delay buys you nothing.
For shockable rhythms — VF and pulseless VT — the 2025 Guidelines state that it is reasonable to administer epinephrine after initial defibrillation attempts have failed. Defibrillation is the definitive treatment; the vasopressor is adjunctive. In practice that usually means epinephrine goes in after the first couple of unsuccessful shocks, while high-quality compressions continue throughout.
This ordering is not theoretical. An observational study of in-hospital arrest found roughly 10% lower risk-adjusted survival at hospitals that most often gave epinephrine before the first defibrillation, compared with those that did so least often. It is observational data and cannot prove causation — but it points the same direction as the guideline, and the guideline is unambiguous about the priority order.
The real change: how you get access
This is the update most likely to alter what you actually do at the bedside, and it has nothing to do with the drug itself.
Previous practice treated IV and IO as broadly interchangeable in arrest — grab whichever is faster. The 2025 Guidelines are more specific. Establishing IV access for drug administration in adult cardiac arrest is now the Class 1 recommendation. Intraosseous access is reasonable — a weaker recommendation — if initial attempts at IV access are unsuccessful or not feasible.
The change follows several large randomised trials comparing the two routes, which did not show the improvement in outcomes that IO-first practice assumed. IO remains a legitimate and important option; it has simply moved from co-equal to second in line.
Read the recommendation carefully: it says attempt IV first, not persist with IV. If the initial attempts fail or peripheral access clearly isn't feasible in this patient, going to IO is explicitly endorsed. The error to avoid is the third and fourth blind stab at a vein while the drug goes ungiven.
What epinephrine does and does not buy you
The evidence is reasonably consistent that epinephrine improves the rate of return of spontaneous circulation. More patients get a pulse back. That is a real and meaningful effect.
The evidence for improved neurologically intact survival is far weaker. Getting a pulse back is not the same as getting the patient back, and the gap between those two outcomes is where most of the ongoing debate about vasopressors in arrest actually lives.
On alternatives: vasopressin, whether given alone or alongside epinephrine, offers no advantage as a substitute. Epinephrine remains the vasopressor.
Epinephrine is not the intervention that saves the brain. Compression quality, minimal interruptions and — in shockable rhythms — early defibrillation are. If attention shifts to the drug while compressions decay, the trade is a bad one.
Practical delivery
- Attempt IV first, then move to IO IV is the Class 1 route; IO is reasonable once initial IV attempts are unsuccessful or clearly not feasible. Decide early and commit — repeated blind attempts at a vein while no drug is being given is the failure mode to avoid.
- Flush every dose A drug sitting in a peripheral line is not circulating. Follow with a fluid flush and elevate the limb.
- Use the right concentration Arrest dosing uses the dilute cardiac-arrest preparation, not the concentrated ampoule used for anaphylaxis. Confirm the vial out loud — this is a classic high-stress substitution error.
- Announce and confirm "Epi one milligram IV" — "Epi one milligram IV, in." Without closed-loop confirmation, doses get given twice or not at all, and nobody can reconstruct which afterwards.
- Let the timekeeper own the interval Three-to-five minutes is easy to state and hard to feel. Someone should be tracking it explicitly rather than everyone estimating.
Do not give it during the pause
A small but common error: waiting for the rhythm check to push the drug. Every second of that pause is a second without perfusion, and the drug does not need a pause to be given.
Push epinephrine during compressions. Charge the defibrillator during compressions too. The only things that should stop compressions are the rhythm check itself and the shock delivery — and both should be measured in seconds.
Scenarios where the timing of the first dose actually matters. Give epinephrine before the first shock in a VF arrest and see how the case unfolds differently. The priority order stops being a sentence in a guideline and becomes something you have felt go wrong.
What every clinician should remember
- 1 mg IV/IO every 3–5 minutes, Class 1 — the dose, interval and strength are unchanged in 2025
- Non-shockable rhythms: give it as early as access allows
- Shockable rhythms: defibrillate first; epinephrine is reasonable after initial defibrillation attempts have failed
- New in 2025 — attempt IV access first (Class 1); IO is reasonable if initial IV attempts fail or aren't feasible
- Vasopressin, alone or with epinephrine, offers no advantage as a substitute
- Epinephrine reliably improves ROSC; its effect on neurologically intact survival is much less certain
- Give the drug and charge the defibrillator during compressions — never during the pause