Peer-reviewed by Dr. Abdolghader Pakniyat
A practical, evidence-based approach for EMS, the Emergency Department, ICU, and anyone leading an adult resuscitation.
Clinical scope: Adult cardiac arrest. This article supports clinical reasoning; it is not a stand-alone termination protocol. Follow local law, institutional policy, and EMS medical-control requirements. Pediatric arrests require separate guidance. Termination during cardiac arrest is distinct from neurological prognostication after ROSC.
At some point during a prolonged cardiac arrest, somebody looks at the clock.
Twenty minutes.
Thirty minutes.
No ROSC.
Then comes the question:
Are we still doing something useful, or are we just continuing CPR because nobody wants to be the person who says stop?
This is one of the most difficult decisions in resuscitation.
And I think the first thing we need to get rid of is the idea that there is one number that gives us the answer.
Not 20 minutes.
Not an ETCO₂ of 10.
Not asystole.
Not cardiac standstill on ultrasound.
None of these, alone, means that the patient has no chance.
Termination of resuscitation should be a multimodal clinical decision based on the circumstances of the arrest, the physiology during CPR, the response to treatment, reversible causes, the patient’s wishes, and the system in which the arrest occurs.
That last part matters.
Because an unwitnessed OHCA in the field is not the same situation as a monitored VF arrest in the ICU.
Before asking when to stop, ask whether CPR should have started
In most adult cardiac arrests, the default is simple: start resuscitation.
There are exceptions. CPR may reasonably be withheld when there are objective signs of irreversible death, when resuscitation would create serious danger to the rescuer, or when there is a valid advance care plan or DNACPR order stating that the patient does not want CPR. The 2025 AHA ethics guidance continues to support this approach.
But once CPR has started, the decision becomes more difficult.
Now we are no longer asking whether cardiac arrest occurred.
We are asking whether the physiology is still potentially recoverable.
Before you prognosticate, make sure the resuscitation itself is good
This is something I think we sometimes forget.
A very low ETCO₂ during poor-quality chest compressions is not the same as a very low ETCO₂ despite excellent CPR.
Persistent asystole while severe hyperkalemia remains untreated is not the same as persistent asystole after reversible causes have been addressed.
Cardiac standstill after a prolonged interruption for ultrasound tells us very little about what good uninterrupted CPR might have achieved.
So before saying, “This patient is not responding,” I want to know:
What exactly have we given the patient the opportunity to respond to?
High-quality CPR. Appropriate defibrillation. Appropriate airway and ventilation. Guideline-directed ALS. Active search for reversible causes. Minimal interruptions. Appropriate treatment of the suspected cause.
Only then does the absence of response become meaningful.
I think about five questions before stopping CPR
- What was the context of the arrest? Was it witnessed? Was CPR started immediately? What was the no-flow time? Was the patient monitored? What was the initial rhythm?
- What has happened during the resuscitation? Has there been any ROSC, transient pulse, rhythm improvement, shockable rhythm, increasing ETCO₂, arterial pressure response, or other sign that circulation is improving?
- What does the physiology show? What is the ETCO₂ trend? Is there an arterial waveform? Is there cardiac activity on ultrasound? Importantly, none of these should be interpreted alone.
- Is there still something reversible or something different we should be doing? Hypoxia, hypovolemia, hyperkalemia, tamponade, pulmonary embolism, tension pneumothorax, toxicological arrest, hypothermia, pregnancy, or a patient who may be appropriate for an ECPR pathway can completely change the decision.
- What would the patient want? Do we know their goals of care? Is there a valid limitation of treatment? And are we continuing because there is a realistic physiological objective—or simply because stopping is uncomfortable?
That is the discussion I want my team to have.
Not simply: “We reached 20 minutes.”
EMS: this is where TOR rules have their strongest evidence
For adult nontraumatic OHCA, validated termination-of-resuscitation rules can be very useful.
The 2025 AHA guideline recommends using the appropriate rule according to the EMS system. [1]
For BLS or universal TOR, the classic combination is:
Arrest not witnessed by EMS + no shock delivered + no ROSC before transport.
For ALS systems, the rule is more restrictive:
Arrest not witnessed + no bystander CPR + no shock delivered + no ROSC after ALS care.
When all criteria are present, termination may be considered according to the EMS system’s protocol.
But there is an important word here:
Validated.
ILCOR’s 2025 recommendation is conditional and based on very low-certainty evidence. It specifically emphasizes that EMS systems should validate TOR rules locally and consider their own population, system, culture and values.
So a TOR rule is not something we copy from a paper and start using tomorrow.
It is a system tool.
And what about the famous 20 minutes?
Twenty minutes matters.
But not in the way we sometimes use it.
It is a prognostic marker, not an expiry time.
A good example is the modified Goto rule. In a very large Japanese OHCA population, the combination of initial asystole, unwitnessed arrest, more than 20 minutes of EMS CPR, and no prehospital ROSC predicted 1-month mortality with a positive predictive value of approximately 99.8%.
That sounds almost absolute.
But it was not.
In the validation cohort, 137 patients meeting the rule survived, and 31 had neurologically intact survival. [4]
That is exactly why I do not like saying:
“Twenty minutes means futile.”
The correct message is:
The longer high-quality resuscitation continues without any physiological response, particularly in an unwitnessed asystolic arrest without ROSC or a reversible cause, the worse the prognosis becomes.
That is very different.
The Emergency Department is a hybrid
This is where things become interesting.
Imagine two patients lying in adjacent resuscitation bays.
Patient one had an unwitnessed OHCA. No bystander CPR. Initial asystole. EMS has performed 30 minutes of ALS. No shock. No ROSC.
Patient two was talking to you five minutes ago, suddenly deteriorated, went into witnessed VF and received immediate CPR.
Both patients are now physically in the Emergency Department.
But physiologically and prognostically, these are completely different arrests.
For the first patient, the prehospital history is part of the resuscitation. The clock does not restart when the ambulance reaches the ED.
For the second patient, you are managing a witnessed, immediately treated in-hospital arrest.
So in the ED, I would not ask simply:
“How long have we been doing CPR here?”
I would ask:
What is the entire arrest story?
And then I would integrate that story with what I am seeing now: rhythm, response, ETCO₂, POCUS, reversible causes, and whether the patient has a realistic rescue pathway that has not yet been used.
ICU and IHCA: be very careful with termination rules
This is where we should resist borrowing OHCA rules.
In-hospital cardiac arrest is different.
Patients may be monitored. The arrest may be recognized immediately. CPR may start within seconds. We may already have vascular access, an arterial line, ventilatory support, laboratory results, and much more information about the underlying disease.
And, importantly, we still do not have a sufficiently reliable universal IHCA termination rule.
A Scandinavian study proposed an interesting combination:
Unwitnessed arrest + unmonitored patient + initial asystole + no ROSC within 10 minutes.
It performed well enough to deserve further investigation.
Then came an important reality check.
In 2026, investigators applied the rule to 359,686 adult IHCAs from 703 U.S. hospitals.
Among patients who met the proposed rule, 32.9% subsequently achieved ROSC and 5.1% survived to hospital discharge. [6]
Approximately one survivor for every 20 patients who would have met the termination criteria.
That is not an acceptable error rate for a stand-alone stop rule.
The ERC 2025 guideline also advises that TOR rules should not be used as the sole strategy for terminating IHCA resuscitation.
For the ICU code leader, context becomes even more important.
Why did this patient arrest?
What was happening immediately before the arrest?
Was there progressive refractory multiorgan failure, or did a relatively reversible event suddenly occur?
Did the patient have a monitored rhythm transition?
What does the arterial line show during CPR?
Is there a correctable mechanical or metabolic cause?
And what were the patient’s previously established goals of care?
That information is often much more useful than trying to force an OHCA prediction rule onto an ICU patient.
ETCO₂: use the physiology, not just the number
ETCO₂ is one of the most useful physiological signals we have during CPR.
Why?
During reasonably constant ventilation and CO₂ production, exhaled CO₂ depends substantially on how much blood is reaching the lungs.
In other words, during cardiac arrest, ETCO₂ becomes an imperfect window into pulmonary blood flow and therefore cardiac output generated by CPR.
So if compressions improve and ETCO₂ rises, that may tell us that forward blood flow has improved.
If ETCO₂ suddenly rises significantly, think about ROSC.
And if it remains extremely low despite optimized resuscitation, the prognosis becomes progressively worse.
But ETCO₂ is not pure cardiac output.
Ventilation, airway technique, epinephrine, bicarbonate, the cause of arrest and other factors can change the value.
The 2025 AHA guideline says that in an intubated adult, failure to achieve an ETCO₂ above 10 mmHg after 20 minutes of ALS may be considered as one component of a multimodal decision about termination.
It specifically does not recommend using ETCO₂ alone to stop resuscitation, and it advises against applying a specific ETCO₂ termination threshold to nonintubated patients.
So my approach is:
Follow the trend from the beginning. Do not suddenly look at the monitor at minute 20 and allow one number to decide whether the patient lives or dies.
POCUS: excellent tool, dangerous oracle
POCUS can be extremely useful during cardiac arrest.
It may identify tamponade.
It may support severe hypovolemia.
It may contribute to the assessment of pulmonary embolism.
It can show us whether apparent PEA on the monitor is accompanied by meaningful mechanical activity.
But ultrasound has two important dangers during resuscitation.
The first is interrupting chest compressions.
The second is believing the image is more definitive than it really is.
The 2025 AHA guideline states that POCUS may be considered for identifying reversible causes when performed by experienced clinicians without interfering with CPR. Its use for prognostication based on cardiac function remains uncertain.
Systematic-review evidence reinforces that caution.
A 2024 systematic review and meta-analysis found that absent cardiac motion was associated with a higher likelihood of death in adult medical cardiac arrest, but accuracy was insufficient for use as a stand-alone termination tool. Study methods, scan timing, and definitions of cardiac activity varied substantially. [7]
That is useful prognostic information.
But association with poor outcome is not the same as certainty of futility.
This is why I like to think about POCUS asymmetrically:
Seeing cardiac activity may give me an important reason to continue.
Not seeing cardiac activity does not, by itself, give me permission to stop.
Some arrests should immediately make you cautious about routine TOR logic
Routine adult medical OHCA rules should not simply be transferred to every cardiac arrest.
The AHA specifically notes that standard TOR rules have not been validated for situations including overdose, trauma and IHCA.
And special circumstances such as severe hypothermia, drowning, pregnancy, major pulmonary embolism, toxicological arrest and selected patients who may be candidates for extracorporeal CPR require disease-specific thinking rather than a generic stopwatch approach.
Persistent or recurrent shockable rhythm should also make us think differently from prolonged refractory asystole.
The question becomes:
Is conventional CPR failing because recovery is impossible—or because this patient needs a rescue strategy that we have not yet provided?
Those are not the same thing.
How I would actually make the decision at the bedside
When I am reaching the point where termination is being considered, I do not want the code leader quietly deciding alone.
I would briefly summarize the arrest for the whole team:
“We have an unwitnessed medical cardiac arrest, initial asystole, no bystander CPR, no shockable rhythm and no ROSC despite prolonged high-quality ALS. ETCO₂ has remained very low despite optimizing CPR and ventilation. We have not identified a reversible cause. There is no special circumstance or additional rescue pathway available. Does anyone have information or a concern that should change our decision?”
That takes perhaps 20 seconds. Continue indicated CPR while this discussion occurs; it must not create an additional pause in compressions.
But it does several things.
It forces the leader to explain why stopping is reasonable.
It gives the team an opportunity to identify something missed.
And it changes termination from “the senior doctor decided” into a deliberate clinical decision based on the complete picture.
The 2025 ERC ethics guideline specifically recommends a team-based, planned termination decision in which the resuscitation is summarized and team members have an opportunity to contribute before CPR is stopped. It also recommends a debrief afterward. [2]
I think that is good practice regardless of whether you are standing beside an ambulance, in resus, or in the ICU.
So when should we stop CPR?
Not because 20 minutes have passed.
Not because the monitor shows asystole.
Not because ETCO₂ is 8.
Not because the ultrasound shows standstill.
And not simply because the resuscitation feels long.
We should consider stopping when multiple independent pieces of information begin to tell the same story:
The arrest had very poor prognostic circumstances.
High-quality guideline-directed resuscitation has been delivered.
There has been no meaningful physiological response.
No reversible cause remains untreated.
No appropriate rescue strategy remains available.
There is no special circumstance in which prolonged resuscitation may still offer benefit.
And continuing CPR is no longer offering a reasonable prospect of achieving an outcome consistent with the patient’s goals.
For EMS, validated and locally implemented TOR rules can help make that decision more objective.
For the ED, the entire prehospital course must follow the patient through the door.
For ICU and other in-hospital arrests, we currently need even greater caution because no universal IHCA termination rule is sufficiently reliable to make the decision for us.
ETCO₂ helps.
POCUS helps.
Time helps.
Rhythm helps.
Context helps.
But none of them should become the single judge.
That is probably the most important message.
Good resuscitation leadership is not only knowing when to continue fighting.
It is also knowing when the physiology, the evidence, the patient’s values, and the response to everything we have done are all telling us that continuing CPR is no longer helping.
And when we reach that point, stopping is not failure.
It is a clinical decision that deserves the same discipline, evidence, teamwork and humanity as everything we did before it.
References
- American Heart Association. (2025). Part 9: Adult Advanced Life Support.
- European Resuscitation Council. (2025). Ethics in resuscitation. Resuscitation, 215, 110734.
- ILCOR. Out-of-hospital cardiac arrest termination-of-resuscitation rules: systematic review and treatment recommendations.
- Goto, Y., Funada, A., Maeda, T., & Goto, Y. (2022). Termination-of-resuscitation rule in the emergency department for patients with refractory out-of-hospital cardiac arrest: a nationwide, population-based observational study. Critical Care, 26, 137.
- Holmberg, M. J., et al. (2025). Termination of resuscitation rules for in-hospital cardiac arrest. JAMA Internal Medicine.
- Moskowitz, A., et al. (2026). External validation of rules for termination of resuscitation in in-hospital cardiac arrest. Resuscitation, 220, 111019.
- Albaroudi, O., et al. (2024). Can absence of cardiac activity on point-of-care echocardiography predict death in out-of-hospital cardiac arrest? A systematic review and meta-analysis. The Ultrasound Journal, 16, 10.
