Two devices, one misunderstanding, and it costs lives every year: chest compressions and a defibrillator do completely different jobs, and neither substitutes for the other.
Guidance changed at the end of 2025. The Resuscitation Council UK
published updated guidelines, and several points still in wide circulation — including in
training material and in charts written before then — are now superseded. Where this page
differs from something you learned, the changes are flagged.
1. What each one actually does
| Chest compressions | Defibrillation | |
|---|---|---|
| What it does | Moves blood. Keeps the brain and the heart muscle alive | Stops the heart's chaotic electrical activity so its own pacemaker can restart an organised rhythm |
| What it does not do | It does not restart the heart. It buys time | It does not circulate blood. A defibrillated heart with no compressions before it usually has nothing left to pump |
| When it helps | Always. Every cardiac arrest, every rhythm, every age | Only for a shockable rhythm — ventricular fibrillation or pulseless VT |
So the sequence is not "try CPR, then try the defibrillator". It is compressions continuously, with a shock inserted the moment one is available and advised. Every pause in compressions drops the pressure perfusing the heart, and it takes several compressions after a pause to get back to where you were.
"The AED said no shock advised, so they must be alright"
- This is the most dangerous misunderstanding in first aid, and it is common. An AED analyses the rhythm and shocks only if it is shockable. Asystole — a flat, silent heart — is not shockable. Neither is pulseless electrical activity.
- "No shock advised" therefore means one of two things: they do not need a shock because their heart is beating, or they do not need a shock because the rhythm is one a shock cannot fix. The device cannot tell you which, and it will not try.
- If they are unresponsive and not breathing normally, you carry on with compressions. The AED will re-analyse. Stopping because it did not shock is how people die next to a working defibrillator.
- The reverse is also worth knowing: a shockable rhythm degrades into a non-shockable one over minutes. Good compressions keep it shockable for longer, which is another reason the two are not alternatives.
2. Adults
- Unresponsive? Call 999 first. Changed in 2025 — call before assessing breathing, not after, with the call handler then guiding you. Speakerphone, so you never leave them.
- Not breathing normally? Occasional gasping is not breathing. Agonal gasps are common in the first minutes of arrest and are routinely mistaken for signs of life. If in doubt, start.
- 30 compressions to 2 breaths, centre of the chest, 5–6 cm deep, 100–120 per minute, full recoil between each.
- Compression-only is far better than nothing if you cannot or will not give breaths.
- Send someone for the nearest AED and keep going. If you are alone, call 999 and start compressions rather than leaving to find one.
3. Children and infants
Paediatric arrest is usually respiratory in origin rather than a sudden rhythm problem, which is why breaths matter more here than in adults.
| What to do | |
|---|---|
| If you are trained in paediatric life support | 5 rescue breaths first, then 15 compressions to 2 breaths. Paediatric first aid courses now teach this too, changed in 2025 |
| If you are only trained in adult CPR | Use 30:2 on a child. It is far better than hesitating — the fear of causing harm is unfounded and is why some children get nothing |
| Depth | At least one third of the chest depth — about 5 cm in a child, 4 cm in an infant |
| Hands | One or two hands for a child, whatever achieves the depth. For an infant, the two-thumb encircling technique — changed in 2025 from two fingers |
| Rate | 100–120 per minute, same as adults |
| Surface | Firm. Not a bed, not a mattress |
AEDs and children — simpler than it used to be
- An AED can be used at any age, including infants. Changed in 2025. Use paediatric pads or a paediatric mode if the device has them; if it does not, use the adult AED anyway. An adult AED is enormously better than no AED.
- Most infants in arrest are not in a shockable rhythm, so the device will usually not shock. The ones who are shockable benefit greatly, and you cannot tell which is which without attaching it.
- Child pad placement: one pad on the front just left of the midline, the other in the centre of the back. Tweaked in 2025. The aim is always to get the pads either side of the ventricles.
- Adolescents: adult pads, standard adult positions.
- Pads must not touch each other. Dry the chest first. Hairy chest: press firmly, and only shave if a razor is in the kit and it takes seconds.
4. Is this a cardiac arrest, a heart attack, a stroke, or something else?
Four things get confused with each other constantly, and the confusion matters because the response to each is different. The good news is that separating them at the scene takes two observations.
| Responsive? | Breathing normally? | What you do | |
|---|---|---|---|
| Cardiac arrest | No | No — or occasional gasping | CPR and an AED, now. Nothing else on this page matters until this is excluded |
| Heart attack | Yes — usually fully awake and frightened | Yes, though may be breathless | 999, sit them down, keep them still and calm. Aspirin 300 mg chewed if available and no allergy or bleeding, on advice from the call handler |
| Stroke | Usually yes, though may be drowsy | Yes | 999, FAST, note the time last known well, nothing by mouth and no aspirin — see the stroke chapter |
| Faint | Briefly no, then rapidly and fully yes | Yes throughout | Flat, legs up. See fainting |
| Seizure | No, then confused for many minutes | Often irregular during, then normal | Protect, time it, nothing in the mouth — the collapse chapter |
| Low blood sugar | Variable — confused, aggressive, drowsy | Yes | Glucose meter, then sugar if alert. Mimics all of the above |
A heart attack and a cardiac arrest are not the same thing, and the difference
is the single most common misunderstanding in this whole subject. A heart attack is
a plumbing problem: an artery blocks and heart muscle starts to die. The person is usually awake, in
pain, and their heart is still beating. A cardiac arrest is an electrical problem: the
heart stops pumping effectively and the person is unconscious and not breathing normally within
seconds. A heart attack can cause a cardiac arrest, which is why someone with chest pain needs
an ambulance rather than a lift — but most heart attacks do not, and most people having one are
entirely conscious throughout.
The two observations that sort all of it: are they responsive, and are they breathing normally? Everything else is refinement. If the answer to both is no, start compressions and stop diagnosing — you cannot make an arrest worse by treating it as one, and the commonest fatal error here is spending a minute deciding.
5. What “fibrillation” actually means, and why it matters
The words get used interchangeably and they describe genuinely different states of the heart, with different treatments.
| Rhythm | What the heart is doing | Shockable? |
|---|---|---|
| Ventricular fibrillation VF | Not stopped — quivering. The electrical signal that normally sweeps across the ventricles in an organised wave has broken into chaos, so millions of muscle cells fire independently. The muscle trembles rather than squeezes, and pumps no blood at all. On a trace it looks like an erratic scribble with no recognisable beats | Yes. A shock stops everything at once, so the heart's own pacemaker can restart something organised. This is the rhythm defibrillators exist for |
| Pulseless ventricular tachycardia pVT | Organised but far too fast to fill and empty. Regular, very rapid, and generating no useful output | Yes |
| Asystole the "stopped heart" | Genuinely stopped. No meaningful electrical activity at all — a flat line. There is nothing chaotic to interrupt | No. A shock has nothing to act on. Compressions, oxygen and treating the cause are the only options |
| Pulseless electrical activity PEA | The electrical signal looks reasonable but the muscle is not producing output — because it is empty, obstructed, compressed or poisoned | No. The answer is finding and fixing the cause, listed below |
Which is why “no shock advised” is not reassurance, as section 1 says: it means the rhythm is asystole or PEA, not that the heart is beating. And it is why good compressions keep a heart shockable for longer — VF degrades into asystole over minutes, and once it does, the defibrillator has nothing to offer.
Atrial fibrillation is a completely different thing and is not an emergency. It is chaos in the atria — the small upper chambers — while the ventricles carry on pumping, irregularly. People live for decades with it. Sharing the word “fibrillation” with the rhythm above is one of the more unfortunate coincidences in medical vocabulary. See single-lead ECG.
6. The reversible causes — what a team is looking for during a resuscitation
Traditionally taught as four Hs and four Ts. Worth knowing even as a bystander, because you frequently hold the information that identifies one of them and nobody at the hospital does.
| The four Hs | The four Ts |
|---|---|
| Hypoxia — not enough oxygen. Drowning, choking, asthma, smoke Hypovolaemia — not enough blood. Bleeding, internal or external Hyper- or hypokalaemia and other metabolic causes — kidney failure, missed dialysis, drugs Hypothermia — and remember nobody is dead until they are warm and dead |
Tension pneumothorax — air trapped, compressing the heart Tamponade — fluid around the heart, stopping it filling Toxins — opioids, tricyclics, cocaine, insulin, local anaesthetic Thrombosis — a clot in the lung or in a coronary artery |
What to tell the crew, because it changes the treatment
- What happened immediately before, and whether it was witnessed. Collapse mid-sentence suggests a rhythm problem; collapse after chest pain suggests a coronary cause; collapse in water suggests hypoxia.
- The medicines, and the packaging. Insulin, opioids, antidepressants and heart drugs all change what is given. Bring the box.
- Anything they had taken or been given, including recreational drugs. Nobody is in trouble.
- Whether they are on dialysis, and when they last went — a missed session is a classic potassium arrest.
- Recent surgery, a long flight, a swollen leg, or known cancer — all point at a clot on the lung.
- How long they were down before anyone started CPR. The single most useful number, and the one only you have.
7. After the heart restarts — what happens next and why
Getting a pulse back is the halfway point rather than the end. The hours afterwards determine whether the brain recovers, and the care is deliberate rather than a matter of waiting.
| What | Why |
|---|---|
| 12-lead ECG immediately | Looking for ST elevation. If it is there, emergency coronary angiography — a blocked artery is one of the commonest causes and re-arrest is likely until it is opened. Without ST elevation, the COACT and TOMAHAWK trials showed immediate angiography adds nothing for stable patients, so it is done selectively — promptly if there is shock, recurrent VF or ongoing ischaemia |
| Oxygen and carbon dioxide targeted, not maximised | Aim for a saturation of 94–98%. Too much oxygen after an arrest is harmful, and so is too little. Carbon dioxide is kept in the normal range rather than blown off |
| Blood pressure supported | A mean arterial pressure high enough to perfuse a brain that has just been starved — usually above 65 mmHg, and often a higher target. Fluids, and drugs to support the circulation if needed |
| Temperature controlled — and this changed | Actively prevent fever for at least 72 hours. ERC-ESICM guidance treats a temperature above 37.7 °C; a maintained target at or below 37.5 °C is the deliberately conservative band used in many protocols, which is why both figures appear in the literature. The older practice of deliberately cooling to 33 °C is no longer preferred: the TTM2 trial found it no better than strict normothermia, and the 2025 guidelines removed the legacy 32–36 °C option. Fever after an arrest consistently worsens brain injury; induced hypothermia does not consistently help |
| Glucose, potassium and seizures managed | Seizures are common after an arrest, often subtle, and worsen brain injury |
The investigations, and what each one is for
| Test | Looking for |
|---|---|
| Arterial blood gas and lactate | How badly the tissues were starved, and the trend. A falling lactate is one of the better early signs |
| Potassium, sodium, magnesium, calcium, kidney function | A potassium arrest is reversible and is missed if nobody looks. Also guides everything else |
| Troponin | Heart muscle damage — but interpret with care: CPR itself raises it, so a modest rise does not prove a heart attack |
| Full blood count, clotting | Bleeding as a cause, and the coagulopathy that follows a prolonged arrest |
| Glucose | Both as a cause and as something to control afterwards |
| Toxicology, and a paracetamol level | Where the history is unclear or overdose is possible |
| Echocardiogram | How well the heart is now pumping, tamponade, a valve problem, and signs of strain on the right side suggesting a clot on the lung |
| CT head, and often CT chest | A bleed in the brain as the cause of the arrest — which happens and is missed — and the extent of brain swelling. CT chest for a pneumothorax, and for the aorta |
| D-dimer and CT pulmonary angiogram | A clot on the lung. The D-dimer is of limited use here and is frequently over-interpreted — it rises after CPR, after any critical illness, and with age, so a raised result proves nothing. Where a clot is genuinely suspected, the CT pulmonary angiogram is the test, and a normal D-dimer in this setting is more informative than a raised one |
| Chest X-ray | Tube position, aspiration, rib fractures from compressions, pulmonary oedema |
And the timing rule that matters most to families. Judging how well the brain will recover is deliberately delayed until at least 72 hours after temperature and sedation are normalised, and uses several independent findings together rather than any single test. People who look hopeless at 24 hours sometimes recover well, which is precisely why nobody is asked to make an irreversible decision in the first day. If you are being told “it is too early to say”, that is a real answer rather than an evasion.
Then, before discharge
- Find and treat the cause — coronary disease, a rhythm disorder, a valve problem, an electrolyte cause, a drug.
- Consider an implantable defibrillator where the arrest was not from a reversible cause and the heart remains at risk.
- Screen the family where an inherited cause is possible, particularly in anyone young, and especially with a family history of sudden death — see inherited cardiac conditions.
- Expect cognitive and emotional consequences, in the survivor and in whoever did the CPR. Memory and concentration problems are common, and so is post-traumatic stress in bystanders and family. Both are treatable and both are routinely unmentioned.
8. The observations that actually change decisions
| Observation | Why it earns its place |
|---|---|
| Respiratory rate | The most sensitive single sign of deterioration in almost every acute illness, the earliest to move, and the one most often not counted. Count for a full minute in a child. Also the one no device does for you |
| Work of breathing | Nasal flaring, tracheal tug, grunting, recession, the ability to speak in full sentences. In children this precedes any change in saturation |
| Level of consciousness (AVPU) | A drop from Alert is significant regardless of every other number |
| Skin and capillary refill | Pale, mottled, clammy. Press a fingertip for five seconds; colour should return within two |
| Pulse | Rate, but also volume. A fast thready pulse says more than the number alone. A slowing pulse in a deteriorating child is pre-terminal |
| Blood pressure | Confirms shock but arrives late, especially in children, who compensate until they do not. Never delay treatment to obtain it |
| Temperature | Fever, and also the low temperature that can accompany sepsis. Both directions matter |
| Oxygen saturation | Useful, and trusted far beyond what it can support — it reads falsely high in carbon monoxide poisoning and in people with darker skin, fails in shock, and says nothing about whether carbon dioxide is being cleared. It has a chapter of its own because it is the most over-interpreted measurement in an emergency: pulse oximetry |
| Blood glucose | In anyone drowsy, confused, fitting or behaving oddly. Hypoglycaemia mimics almost everything and is reversible in two minutes |
The principle underneath all of it, which is the same one as elsewhere on this site. A single reading has error bars, and a reading that disagrees with how the person looks is more likely to be wrong than the impression is. The trend across three sets of observations is worth more than any one set, and the observation of an experienced person that someone is deteriorating outperforms most individual measurements. Devices exist to support that judgement, not to overrule it.