Breaths, and what CPR can and cannot do

There are two questions almost everyone brings to a CPR class, and they are the ones this lesson answers. The first is "do I have to do mouth-to-mouth?", often asked by someone who'd rather not. The second, usually asked later and more quietly, is "if I do it right, will they live?" The honest answers are "it depends who you are, and here is why" and "usually not, but far more often than if you do nothing". Both answers rest on evidence that is thinner than most people expect, so this is also the lesson that teaches you to read the labels in brackets that have followed every instruction since lesson 2.

What goes in the pause

For an adult, compressions come first in both countries. RCUK's 2025 guidelines say "Start chest compressions as soon as possible",¹ and the AHA's tell a lone rescuer with a phone to call and "immediately begin CPR, beginning with chest compressions" (AHA 2025, COR 1, LOE C-LD).² The American Red Cross, which teaches its own mnemonic for checking a casualty, agrees on this case: "For adult cardiac arrest and sudden pediatric arrests, the correct resuscitation sequence is compressions first, followed by breaths if the rescuer is trained and able."¹¹

What happens after the first thirty depends on you.

AHA 2025 Part 7, sections 6.2 and 8.4, and Part 4;²³ RCUK 2025 Adult basic life support.¹

The two columns say the same thing in different words. Everyone pushes; people who've been trained, and are willing, add two breaths after every thirty pushes; and in the UK the call handler assumes you haven't been trained until you say otherwise. ILCOR, whose evidence reviews both countries build on, puts it in one pair of sentences, which this lesson reads word by word further down.⁴

Notice two things that are not in the table. Neither country makes breaths a condition of doing CPR, and neither tells an untrained person to try breaths anyway. The AHA's summary is plain: "a chest compression–only approach is appropriate if lay rescuers are untrained or unwilling to provide breaths."²

How to give a breath

Know both countries' public versions word for word, because they're what you'll meet in a class.

The American Red Cross's steps: "Open the airway to a past-neutral position using the head-tilt/chin-lift technique", then "Pinch the nose shut, take a normal breath, and make complete seal over the person’s mouth with your mouth." Then "Ensure each breath lasts about 1 second and makes the chest rise; allow air to exit before giving the next breath".⁵ (Past-neutral means the head tilted back, beyond the position it rests in; St John's version says it plainly.) St John Ambulance's steps in the UK: "Tilt the casualty's head backwards, lift their chin, and pinch the soft part of their nose closed." Then "Blow into their mouth steadily and firmly for up to one second, until the chest rises," and "Remove your mouth and watch the chest fall."⁶ Two breaths, then straight back to thirty compressions.

Each step has a reason.

The head tilt. The British Red Cross explains it: "When a person is unresponsive, their muscles relax and their tongue can block their airway so they can no longer breathe. Tilting their head back opens the airway by pulling the tongue forward."⁸

The pinched nose and the seal. The AHA's description has you "pinch the person’s nose, create an airtight seal around the mouth, and provide a breath."² No guideline read here spells out why.

About one second, and only until the chest rises. RCUK 2025: "deliver just enough air to make the chest start to rise; avoid excessive ventilation."¹ The AHA grades "each breath over 1 s" as something that "may be reasonable" (COR 2b, LOE C-LD) and gives its reason for the timing: "The recommendation for using a 1-second inspiratory time is to keep the pauses in chest compressions as brief as possible during 30:2 CPR."² The American Red Cross asks you to keep any interruption to compressions under 10 seconds, and two one-second breaths fit inside that with room to spare.⁵

That's why one public page's wording needs a flag. The British Heart Foundation's CPR page, in its instructions for breaths, says "Seal your mouth over theirs and blow hard for a second."⁷ The guidelines do not say hard. They say just enough to make the chest start to rise. If you learned "blow hard", replace it with "until the chest rises".

If the chest does not rise. RCUK 2025: "If you are unable to ventilate the chest after two attempts, consider foreign body airway obstruction".¹ St John says the same for the moment itself: "If one or both don't work, return to chest compressions."⁶ You don't keep trying. Lesson 3 showed what a long pause costs, and lesson 8 covers the airway that is blocked.

Barriers. If you have a face shield or a pocket mask with you, the AHA's text compares them: "When using a barrier device, use of mouth-to-pocket mask provides more effective breaths than mouth-to-face shield."² Lesson 3 covered the rule for protective equipment: use it if it is in your hand, and do not delay compressions to find it.

Why compressions come first for an adult

The British Red Cross gives the reasoning in two sentences. First: "However, giving chest compressions is the most important thing to do because their blood already has some oxygen in it and the compressions will keep that blood pumping around their body, taking oxygen to their brain." Then: "Breathing into their mouth or nose tops up the oxygen in their lungs."⁸ Compressions move the oxygen that's already there; breaths add more.

The Red Cross's word is "some": no guideline read here says how much oxygen that blood carries or how long it lasts. The AHA's reason for encouraging breaths from the trained is a sentence about outcomes rather than mechanism: "Because CPR with breaths may lead to improved outcomes for adults in comparison with chest compression–only CPR, trained rescuers are encouraged to provide breaths along with chest compressions."²

And some arrests start without oxygenated blood. The AHA contrasts the two kinds: "Cardiac arrest following drowning is most often due to a hypoxic mechanism, whereas sudden cardiac arrest, particularly in adults, is more likely to occur with fully oxygenated blood."⁹ Hypoxic means short of oxygen. A person who drowns has run short of oxygen before the heart stops, so the reasoning that puts compressions first for a sudden collapse does not hold in the same way, and lesson 6 gives what each guideline makes of that. The AHA flags overdose too, as a kind of arrest where compressions alone may do less well, as you will see below.

The AHA's US estimate is that "1 in 6 to 1 in 3 OHCAs are related to an overdose, most commonly involving opioids."³ (OHCA is out-of-hospital cardiac arrest.) RCUK's UK figure is that "Approximately 80% of out-of-hospital cardiac arrests are due to a cardiac cause."²⁰ The two are counted differently, in different countries, and don't add up to one picture. Read together with the AHA's sentence above, they say, on this course's reading, that most adult arrests begin with oxygenated blood and a real minority don't. Keep that in mind. Lesson 6 gives the sequences for children and for drowning, where trained rescuers give breaths earlier, and lesson 7 covers the overdose.

How to read the label

Every instruction in this course has carried a label since lesson 2, and lesson 1 promised this lesson would explain it. Sleep lesson 6 gave you the idea already: a guideline grade has two halves, how strongly the panel recommends something and how sure it is about the evidence, and a weak grade is not the same as "it doesn't work". Here are the two versions you'll meet in first aid.

The AHA's version: a class and a level

The AHA gives every recommendation a class of recommendation (COR) and a level of evidence (LOE). Its guideline says what each is for. Of the class: "The specific wording is intended to reflect the risk-benefit ratio associated with the recommendation, with terms such as “is recommended” or “should” reserved for Class 1 recommendations." Of the level: it "reflects the quality of the evidence, with highest value (LOE A) placed on randomized controlled trials (RCTs) and meta-analyses involving RCTs."²

The AHA's Table 1 sets out both scales, and this is its content, shortened.²

From Table 1 of AHA 2025 Part 7, read from the publisher's image. The quoted cells are the AHA's words; the class names and the benefit column are its labels. The ‡ is the AHA's own footnote mark, on how evidence quality is judged.²

An RCT is a randomised controlled trial, where chance decides who gets which treatment, so the two groups differ only by that treatment and by luck. Registry and observational studies record what happened to people who did or did not get something, which is cheaper and bigger but leaves open why they got it.

Two sentences under the table matter more than the table. The first: "COR and LOE are determined independently (any COR may be paired with any LOE)." The second: "A recommendation with LOE C does not imply that the recommendation is weak. Many important clinical questions addressed in guidelines do not lend themselves to clinical trials. Although RCTs are unavailable, there may be a very clear clinical consensus that a particular test or therapy is useful or effective."² So a strong class can sit on expert opinion, and a weak class can sit on trials.

ILCOR's version: strength and certainty

ILCOR, the international committee that runs the evidence reviews, grades differently. Each recommendation is strong or weak, and the evidence behind it has a certainty of high, moderate, low or very low. Its reviewers rate the evidence "based on Grading of Recommendations Assessment, Development, and Evaluation criteria", the GRADE system you met in Sleep.⁴ In every ILCOR recommendation this course read, "We suggest" opens a weak recommendation, and "We recommend" opens either a strong one or a third kind of statement that carries no grade at all. That is this course's observation from the texts, not a rule ILCOR states. So the verb tells you a line isn't weak, and the brackets tell you which of the other two it is.

The one thing to take away

Most first aid recommendations are weak, on low or very low certainty evidence. That is not a reason to ignore them, and it doesn't mean the panel is unsure whether you should act. A weak grade usually describes a choice between two ways of acting, where the studies cannot tell them apart well, while the finding underneath is that acting beats not acting.

You met one of these in lesson 3. The review of injuries in people given compressions who were not in arrest was rated very low certainty, and it still told you enough to act: every study in it found injury uncommon, and nothing in it suggested holding back.

One line, word by word

The recommendation at the centre of this lesson comes from ILCOR's 2025 review of adult basic life support:

"We suggest that bystanders who are trained, able, and willing, give chest compressions with rescue breaths for adults in cardiac arrest (weak recommendation, very low–certainty evidence)."⁴

Take it apart the way you'd take apart a sentence in a contract.

Who. "Bystanders who are trained, able, and willing". All three. Not trained means this line does not apply to you. Trained but unable (you can't get a seal, or you're alone and exhausted) or unwilling (a stranger, blood on their face, you just can't) also means it doesn't apply.

What. Compressions with breaths. Compressions are in both halves of every version of this advice.

For whom. Adults in cardiac arrest. Children and drowning are separate questions with their own lines, in lesson 6.

How strong. "Suggest", and the brackets confirm it: a weak recommendation.

How sure. Very low certainty, the bottom of ILCOR's scale. You'll see in a moment how little evidence sits under it.

What it doesn't say. It doesn't say breaths beat compression-only for everyone, it doesn't tell untrained people to try them, and it doesn't say the choice matters little. Read the sentence printed just above it in the same review: "We recommend that chest compressions be performed for all adults in cardiac arrest (good practice statement)."⁴ Read the two together.

Now look at what sits under it, because this is where reading a grade pays off. ILCOR's evidence for the question is thin: "No new studies that directly addressed this topic were found. The evidence remains 3 observational studies that compared bystander chest compression–only CPR with conventional CPR at a ratio of 15:2 and 30:2 in adults without DA-CPR instructions."⁴ (DA-CPR is dispatcher-assisted CPR, where the call handler talks you through it.) Two of those studies used a ratio of 15 compressions to 2 breaths, which is not taught any more. The one adult study that compared compression-only with today's 30:2 found higher survival with compression-only, an adjusted odds ratio of 1.60.⁴ (An odds ratio compares two groups' odds of surviving; the section on the Swedish study below takes the term apart.)

So the only direct adult comparison with 30:2 on the page, one observational study at very low certainty, points the other way, and ILCOR still suggests breaths for the trained. Its justification does not give one decisive reason. It lists considerations on both sides: that compression-only CPR "is preferred by the public and easier to learn and recall"; that untrained bystanders "are typically unable to deliver effective ventilations during simulated CPR"; and one of three older studies that found no overall difference between the two, a study from the 1980s in which 15:2 done correctly had higher unadjusted survival than compression-only, 31% against 20%. And it ends: "Both types of CPR are better than no CPR, and both should be taught in BLS/CPR training."⁴ That last sentence is the one the weak grade sits on top of.

Two readers, one argument

Imagine two people who've both read a little about this. The first says compression-only CPR is best, because the trials favour it. The second says breaths are best, because so many arrests now come from overdoses. This is a contested question, and both of them are partly right. Each case follows, quoted from the people who make it; where this course adds a reading of its own, it says so.

The case for compressions alone

It starts with randomised trials of what call handlers say. In each, patients were assigned at random to have the caller talked through either compression-only CPR or CPR with breaths.

• Svensson and colleagues, 2010, randomised 1,276 patients. Thirty-day survival was 8.7% with compression-only instructions and 7.0% with standard CPR, a difference that could easily have been chance (P=0.29).¹³
• Rea and colleagues, 2010, randomised 1,941 patients. Survival to hospital discharge was 12.5% against 11.0% (P=0.31), with trends towards compression-only among arrests of cardiac cause and those with a shockable heart rhythm. The authors' conclusion: "The results support a strategy for CPR performed by laypersons that emphasizes chest compression and minimizes the role of rescue breathing."¹⁴
• Hüpfl, Selig and Nagele, 2010, pooled three randomised trials of dispatcher instructions. Together they found "14% [211/1500] vs 12% [178/1531]; risk ratio 1·22, 95% CI 1·01-1·46" (a risk ratio is the relative risk Nutrition taught you), and their interpretation: "For adults with out-of-hospital cardiac arrest, instructions to bystanders from emergency medical services dispatch should focus on chest-compression-only CPR."¹²

Then the argument about people. The Japanese SOS-KANTO study, which followed 4,068 witnessed arrests, opened with the problem in one line: "Mouth-to-mouth ventilation is a barrier to bystanders doing cardiopulmonary resuscitation (CPR)". It found "no evidence for any benefit from the addition of mouth-to-mouth ventilation in any subgroup."¹⁵ And when Japan spread compression-only CPR nationally between 2005 and 2012, the share of people getting any bystander CPR rose from 34.6% to 47.3%, and survival with good brain function attributed to bystander CPR rose from 9.0 to 43.6 per 10 million people.¹⁶ That's an association over time, and the AHA reads it as "likely due to an increase in lay rescuers providing CPR".² ILCOR's dispatcher review puts the priority plainly: "We placed a higher emphasis on the importance of providing high-quality chest compressions and increasing the overall rate of bystander CPR over providing rescue breaths."⁴

The case for breaths

The AHA makes it in its 2025 adult guideline: "Numerous observational studies and 1 large secondary analysis of an RCT found improved outcomes in patients with cardiac arrest who received both chest compressions and ventilations compared with those who received chest compressions only." Its conclusion: "Given the potential benefit of including both compressions and ventilations during CPR, if lay rescuers are appropriately trained, they should be encouraged to deliver breaths with compressions."²

It adds the point about who arrests now. Writing about those three trials, the AHA says: "It is noteworthy that the adult OHCA RCTs largely predate the rise in overdose-associated OHCA seen in the last decade." And in its list of open questions: "Compression-only CPR is easier to implement because of its simplicity and reduced contact but may be less efficacious than CPR with breaths in some etiologies of OHCA (eg, opioid-associated OHCA)."³ Add the one subgroup finding ILCOR reports for breaths, from a study that found no overall difference, and you have the case: when breaths are given well, by someone who knows how, they may add something, and the people most likely to need them are the ones the old trials had fewer of.

What each side says about the other's evidence

The compression-only side's best evidence is randomised; the breaths side's is mostly observational. The AHA itself names the weakness of the second kind: "While recent observational studies have reported improved neurologically intact survival outcomes with conventional CPR in adults, these studies are vulnerable to selection bias and confounding."³ In plain terms, and this gloss is the course's: people who give breaths may differ from people who do not, in training, confidence or speed, and those differences, not the breaths, could explain the result.

The nearest thing to a reply on the breaths side is in what the trials tested, and this reading is the course's. Look at what they compared: bystanders following a call handler's instructions, not trained rescuers choosing breaths for themselves. ILCOR's 2025 review also notes that the comparison in the dispatcher studies was often an old ratio: "As 15:2 CPR is no longer recommended, all outcomes were downgraded for indirectness."⁴ Indirectness is GRADE's word for evidence about a slightly different question from the one being asked. Mental Fitness taught you to ask what the comparison group got. ILCOR's table lists 15:2 as the comparator for the three adult trials' survival outcome, and marked every outcome down for it, so the comparison group got a call handler trying to teach mouth-to-mouth at 15:2 over the phone, during an emergency. That's a fair test of what a call handler should say, and not a test of what a trained rescuer should do.

Where it stands

The guidelines have split the answer by who the rescuer is, which is how they resolve it without deciding it. Call handlers talking to untrained callers give compression-only instructions, graded strongly in both systems (ILCOR 2025, strong recommendation, low certainty; AHA 2025 Part 4, COR 1, LOE A).⁴³ Trained, able, willing rescuers add breaths at 30:2 (ILCOR 2025, weak recommendation, very low certainty; AHA 2025, COR 2a, LOE B-R; RCUK 2025).⁴²¹ And both are far better than no CPR, which is the one part of this that is not contested.

Both bodies say what would settle it. ILCOR lists as a gap "The effect on outcomes of chest compression–only CPR compared with 30:2 CPR without dispatcher assistance". The AHA asks for "Research to identify age cutoffs and etiologies for which individuals in cardiac arrest benefit from full CPR with breaths, as opposed to compression-only CPR".⁴³ In other words, a comparison of trained bystanders doing 30:2 against trained bystanders doing compressions alone, in today's mix of arrests, with the causes recorded. Until then, neither reader in the argument can claim the evidence settles it.

Twice the survival: how that number is made

Lesson 1 quoted RCUK: "Bystander CPR and use of an AED increase the chances of survival by two to four-fold".¹ This section shows where a figure like that comes from, and how far it goes.

The largest study behind the idea is from Sweden. Hasselqvist-Ax and colleagues looked at 30,381 witnessed arrests between 1990 and 2011. Thirty-day survival was 10.5% when someone gave CPR before the ambulance arrived and 4.0% when nobody did. After adjusting for differences between the two groups, including age, sex, where the arrest happened, its cause, the first heart rhythm and how long the ambulance took, the odds ratio was 2.15.¹⁷ The authors' conclusion: survival with CPR "was more than twice as high as that associated with no CPR before EMS arrival."¹⁷

Two words in that need explaining.

Odds, not percentages. Odds compare the chance something happens with the chance it does not. At 10.5% survival, the odds are 10.5 to 89.5, about 0.117. At 4.0%, they're 4 to 96, about 0.042. The ratio of those odds is about 2.8 before any adjustment. The simpler ratio of the two percentages, 10.5 divided by 4.0, is about 2.6. When survival is uncommon, as it is here, the two ratios stay close, which is why "about twice" is a fair everyday reading of an odds ratio of 2.15. When the outcome is common, they drift apart, and a headline that turns an odds ratio into "twice as likely" can be well off. Statistics for Citizens, later on the Core, goes further into this.

Adjusted, and associated. A registry records what happened. Nobody decided at random who got CPR. People who get bystander CPR may differ in ways that matter: collapsing in front of someone, collapsing somewhere busy, being reached sooner. Adjustment accounts for the differences the researchers measured, and it can't account for the ones they didn't. That's why the authors say "associated", and why a Danish study that watched bystander CPR rise from 21.1% to 44.9% between 2001 and 2010, while 30-day survival rose from 3.5% to 10.8%, ended: "Because of the co-occurrence of other related initiatives, a causal relationship remains uncertain."¹⁸

Nutrition taught you the difference between an observational finding and an experimental one, and this is the same distinction. The observational evidence comes from large registries in several countries, all pointing the same way, and every guideline acts on it. ILCOR states it as "Bystander CPR more than doubles OHCA survival."⁴ What the registries cannot give you is the exact size of the effect for any one person.

What CPR can do, and what it cannot

On television, and in real life

Television shows a lot of CPR, and two studies have counted how it turns out.

In 1996, Diem, Lantos and Tulsky watched a season of ER and Chicago Hope and 50 episodes of Rescue 911: 60 resuscitations in 97 episodes. "Seventy-five percent of the patients survived the immediate arrest, and 67 percent appeared to have survived to hospital discharge." The patients were not typical either: "In the majority of cases, cardiac arrest was caused by trauma; only 28 percent were due to primary cardiac causes. Sixty-five percent of the cardiac arrests occurred in children, teenagers, or young adults."²¹ In 2015, Portanova and colleagues repeated the exercise with Grey's Anatomy and House: "CPR was depicted 46 times in the 91 episodes, with a survival rate of 69.6%."²²

Bars drawn to scale from zero. Drawn by this course from the figures in the two television studies' abstracts, the CARES 2025 national report and RCUK's 2025 epidemiology chapter.^21221920

The real figures are about a seventh of the television ones. In the US, the CARES registry recorded 140,144 cardiac arrests outside hospital in 2025, not caused by injury, where the emergency services attempted resuscitation. Of those, 10.5% survived to leave hospital, and 8.3% left with good brain function.¹⁹ In England, 9.5% of people were alive 30 days after a cardiac arrest outside hospital in 2024.²⁰

Don't set those two side by side as if they measured the same thing. The US figure is survival to discharge and the English one is survival at 30 days, and the two registries count their cases differently. They're both roughly one in ten, and that's as far as the comparison goes.

The British Red Cross says plainly what compressions alone do: "The chance of restarting the heart by chest compressions alone is very small. Usually, a heart needs an electric shock from an automated external defibrillator (AED) to restart." What compressions do instead: "Chest compressions pump a small amount of blood around the body to keep the organs alive, most importantly the brain."⁸ CPR keeps the chance alive until a defibrillator or a paramedic can use it. That is what it's for, and lesson 5 is the defibrillator.

Where the chance is higher

The same US report breaks the 10.5% down.¹⁹ The figures below are survival to hospital discharge in 2025. The bystander CPR and defibrillator rows leave out arrests in nursing homes and healthcare buildings, and arrests the ambulance crew saw happen, as the report's own table does.

Bars drawn to scale from zero. Drawn by this course from the CARES 2025 national report, dated April 20, 2026.¹⁹

Three things in that chart need a sentence each.

Arrests in public ended better than arrests at home, 20.9% against 8.9%. The report does not say why, and this course won't guess. Remember that the home is where most arrests happen, 71.5% of this registry's cases, so most of the people you might save are in the lower bar.¹⁹

Bystander CPR went with better survival, 12.9% against 7.7%. That's the same kind of gap as the Swedish study, with the same caution: it is an association in a registry.

A bystander used a defibrillator in only 2.5% of these arrests, and in those, 32.9% survived.¹⁹ That is the biggest number on the chart and the rarest event, and lesson 5 is about why.

Who gets bystander CPR

Lesson 1 said this lesson would report who actually receives bystander CPR. The registries show that it isn't evenly shared.

By sex. Blewer and colleagues looked at 19,331 arrests in the Resuscitation Outcomes Consortium registry from 2011 to 2015. "In public locations, 39% (272/694) of females and 45% (1170/2600) of males received BCPR ( P<0.01), whereas in private settings, 35% (2198/6328) of females and 36% (3364/9449) of males received BCPR ( P=NS)."²³ (BCPR is bystander CPR; NS means not significant.) So in public, women were less likely to get CPR, and at home there was no real difference. The CARES 2025 figures for public arrests show a smaller gap in a different registry and a different year: 46.2% of women and 48.8% of men.¹⁹

By race and ethnicity. Garcia and colleagues studied 110,054 witnessed arrests in the CARES registry from 2013 to 2019. Black and Hispanic people received bystander CPR at home in 38.5% of cases against 47.4% for White people, and in public in 45.6% against 60.0%. Their conclusion: "Black and Hispanic persons were less likely than White persons to receive potentially lifesaving bystander CPR at home and in public locations, regardless of the racial or ethnic makeup or income level of the neighborhood where the cardiac arrest occurred."²⁴

By place, in the UK. RCUK's 2025 epidemiology chapter describes English "hot spots", areas with "incidence above 60.2 per 100,000 per year and bystander CPR rates below 78.7% (the national medians)", which "are more likely to be urban, densely populated, and socioeconomically deprived."²⁰

These are established associations, measured in large registries. What they are not is a measured cause. The studies take some things into account (Garcia's team looked within neighbourhoods of each make-up and income), but none of them recorded why a bystander did or didn't start CPR, and this course does not know what sits behind the gaps. Lesson 1 reported what members of the public guess the reasons are for the sex gap, and those were guesses, labelled as such. Lesson 5 comes back to the one explanation the AHA does discuss, for women and defibrillator pads.

The pandemic changes, and a page that has not caught up

The ERC's 2025 adult guideline says its COVID-19 changes are over: "COVID-19 patients should be treated as any other patients. Modifications to CPR are no longer required."¹⁰ RCUK 2025 recommends 30:2 for anyone trained, as the table at the top of this lesson shows.¹

The British Heart Foundation's CPR page has not changed. In answer to "Should I give rescue breaths?", it says: "It’s your choice if you would like to give rescue breaths. Hands-only CPR is very effective and is the recommended way to give CPR since the Covid-19 pandemic. However, you may wish to give rescue breaths to a family member."⁷ The middle sentence dates itself: it describes CPR as changed by the pandemic, and RCUK's 2025 guidelines give 30:2 to anyone trained. The October 2025 guidelines are recent. If you're trained, follow RCUK.

There is also a difference inside the UK. The British Red Cross's step-by-step page for someone unresponsive and not breathing gives compressions only in its numbered steps, and offers breaths in its questions below: "If you feel able to, combine chest compressions with breathing into their mouth or nose."⁸ St John Ambulance's page puts 30:2 in its main steps "If you are trained to do so".⁶ Both fit RCUK 2025, which gives breaths to the trained and compressions to everyone. They just lead with different halves of it.

What people get wrong

"CPR usually works, like on TV." Television showed survival of around 70% in two studies. Real survival to discharge in the US was 10.5% in 2025, and 30-day survival in England was 9.5% in 2024.^21221920

"It isn't real CPR without mouth-to-mouth." Both countries tell untrained rescuers to give continuous compressions, and call handlers give compression-only instructions to adults (RCUK 2025; AHA 2025, COR 1).¹²³

"Hands-only is the recommended way since COVID." RCUK 2025 and the AHA 2025 give 30:2 for trained rescuers, and the ERC says its COVID-19 modifications are no longer required.¹²¹⁰

"Blow hard." Just enough to make the chest start to rise, over about a second (RCUK 2025; AHA 2025, COR 2b). Too much air goes into the stomach.¹²

"Always tilt the head and blow first." Not for an adult. Compressions come first in both countries; breaths first belong to children in the UK, who get five breaths before compressions, and to trained rescuers in some drowning cases, in lesson 6.¹²

"A weak recommendation means it doesn't matter." It usually means the evidence can't separate two ways of acting. ILCOR says both kinds of CPR beat none.⁴

"LOE A means the class is 1, and LOE C means it's weak." The AHA scores them separately, and says a C-level recommendation "does not imply that the recommendation is weak."²

"Bystander CPR has been proven in trials to double survival." The doubling is from registries, adjusted and consistent, and its authors call it an association.¹⁷¹⁸

Practice

Where to learn with your hands

A breath is harder to learn from a page than a compression, because the seal, the tilt and "just enough" are all things you feel. A class with a manikin that shows the chest rising is where this comes together. For practice at home, the BHF's free RevivR course and RCUK's Lifesaver film are the two free tools this course points to, and neither replaces a class.²⁵ Lesson 16 has the full list.

Connections

Lesson 3 told you to keep pauses short; this lesson filled the pause, and told you who should. The labels in brackets since lesson 2 now have their key, and every lesson from here uses them without explaining them again. Lesson 5 is the defibrillator, the thing behind the 32.9% in the chart, and the reason CPR is a bridge rather than a cure. Lesson 6 covers children and drowning, where the arrest usually starts from lack of oxygen and breaths come earlier for children in the UK and for trained rescuers, and lesson 7 covers the overdose.

From earlier on the Core: Sleep gave you the two halves of a guideline grade, and here they are in two more systems. Mental Fitness taught you to ask what the comparison group got; in the compression-only trials, it got callers talked through 15:2 over the phone. Nutrition taught you what an observational study can and cannot show, which is the whole distance between "associated with twice the survival" and "doubles your chances". Statistics for Citizens, later on the Core, goes further into odds, risks and confidence intervals.

Go deeper

• ILCOR, "Basic Life Support: 2025 International Liaison Committee on Resuscitation Consensus on Science With Treatment Recommendations", Circulation, 2025, free at ahajournals.org. This course read its treatment recommendations and both compression-only sections in full and skimmed the rest, checking the passages quoted here against the published text. The two sections on compression-only CPR, with and without the dispatcher, are a few pages each and show a panel reasoning in public about a question it cannot settle.
• AHA, Part 7: Adult Basic Life Support, 2025, section 8, "Ventilation and Compression-to-Ventilation Ratio", read in full here from an Internet Archive capture, and its Table 1 on the publisher's site. Keep Table 1 open while you read any AHA guideline.
• Hüpfl, Selig and Nagele, "Chest-compression-only versus standard cardiopulmonary resuscitation: a meta-analysis", The Lancet, 2010, free through PubMed Central. This course read the abstract only. It's the pooled analysis behind the AHA's risk ratio of 1.22.
• CARES, 2025 national survival report, five pages, read in full. The whole US picture on one page of numbers, with the definitions at the bottom.

Sources

1. Resuscitation Council UK, 2025 Resuscitation Guidelines: Adult basic life support, published 27 October 2025 . Read level: full, direct, re-read on 2026-09-24 for this lesson. Supports compressions first, the rescue breaths section (30:2 for the trained, just enough air, two attempts, continuous compressions for the untrained), the call handler's compression-only instructions and 30:2 for callers who know how, and "two to four-fold".
2. M. E. Kleinman, J. E. Buick, N. Huber and colleagues, "Part 7: Adult Basic Life Support: 2025 American Heart Association Guidelines for CPR and Emergency Cardiovascular Care", Circulation 152 (suppl 2), 2025 . Read level: full, from the Internet Archive capture 20251026225858, sections 4.4, 6.2, 7.2 and 8 re-read on 2026-09-24; Table 1 read from the publisher's image on ahajournals.org the same day. Supports the class and level definitions, the lay rescuer recommendations with their COR and LOE, 30:2, the breath recommendations and their reasons, the barrier-device sentence, and the case for breaths in the AHA's words.
3. C. Dezfulian, J. G. Cabañas, J. R. Buckley and colleagues, "Part 4: Systems of Care: 2025 AHA Guidelines", Circulation 152 (suppl 2), 2025 . Read level: partial, from the Internet Archive capture 20260208194135: the telephone CPR section, its knowledge gaps and the public access naloxone synopsis, re-read on 2026-09-24. Supports the compression-only instruction for call handlers (COR 1, LOE A), the trials and the pooled risk ratio as the AHA reports them, "largely predate", the observational caution, the research gaps, and the overdose estimate.
4. J. E. Bray, M. A. Smyth, G. D. Perkins and colleagues for the ILCOR Basic Life Support Task Force, "Basic Life Support: 2025 International Liaison Committee on Resuscitation Consensus on Science With Treatment Recommendations", Circulation, 2025 . Read level: the treatment recommendations and both compression-only sections in full, the rest skimmed, from the accepted manuscript on ilcor.org; every quoted sentence except the 1980s-study figures checked against the typeset text on ahajournals.org on 2026-09-24. Supports both compression-only sections (evidence, recommendations, justification, knowledge gaps), the good practice statement definition and the GRADE basis. The count of three adult trials of dispatcher instructions is from the manuscript's table.
5. American Red Cross, CPR Steps, undated . Read level: full, direct, re-read on 2026-09-24. Supports the US public steps for a breath and the 10-second limit on pauses.
6. St John Ambulance, How to do CPR, clinically reviewed 28 April 2025 . Read level: full, direct, re-read on 2026-09-24. Supports the UK public steps for a breath and returning to compressions after two attempts.
7. British Heart Foundation, How to do CPR, undated . Read level: full, direct, re-read on 2026-09-24. Supports "blow hard for a second" and the pandemic-era line on hands-only CPR.
8. British Red Cross, Unresponsive and not breathing, undated . Read level: full, direct, re-read on 2026-09-24. Supports why the head tilt opens the airway, why compressions come first, what compressions alone can and can't do, and the page's offer of breaths.
9. D. Cao, A. M. Arens, S. L. Chow and colleagues, "Part 10: Adult and Pediatric Special Circumstances of Resuscitation: 2025 AHA Guidelines", Circulation 152 (suppl 2), 2025. **Read level: partial, from the Internet Archive capture 20251027153841** (the drowning section among others). Supports the contrast between hypoxic and sudden cardiac arrest.
10. M. A. Smyth, S. van Goor, C. M. Hansen and colleagues, "European Resuscitation Council Guidelines 2025 Adult Basic Life Support", Resuscitation 215 suppl 1, 2025, 110771. **Read level: full, direct, at the publisher**, checked on 2026-09-24. Supports the withdrawal of the COVID-19 modifications.
11. American Red Cross, A-B-C vs. C-A-B, published July 15, 2024 . Read level: full, direct. Supports compressions first for adult cardiac arrest.
12. M. Hüpfl, H. F. Selig and P. Nagele, "Chest-compression-only versus standard cardiopulmonary resuscitation: a meta-analysis", The Lancet 376, 2010, pp. 1552 to 1557. **Read level: abstract only.**
13. L. Svensson and colleagues, "Compression-only CPR or standard CPR in out-of-hospital cardiac arrest", New England Journal of Medicine 363, 2010, pp. 434 to 442 . Read level: abstract only.
14. T. D. Rea and colleagues, "CPR with chest compression alone or with rescue breathing", New England Journal of Medicine 363, 2010, pp. 423 to 433 . Read level: abstract only.
15. SOS-KANTO study group, "Cardiopulmonary resuscitation by bystanders with chest compression only (SOS-KANTO): an observational study", The Lancet 369, 2007, pp. 920 to 926 . Read level: abstract only.
16. T. Iwami and colleagues, "Dissemination of chest compression-only cardiopulmonary resuscitation and survival after out-of-hospital cardiac arrest", Circulation 132, 2015, pp. 415 to 422. **Read level: abstract only.**
17. I. Hasselqvist-Ax and colleagues, "Early cardiopulmonary resuscitation in out-of-hospital cardiac arrest", New England Journal of Medicine 372, 2015, pp. 2307 to 2315 . Read level: abstract only.
18. M. Wissenberg and colleagues, "Association of national initiatives to improve cardiac arrest management with rates of bystander intervention and patient survival after out-of-hospital cardiac arrest", JAMA 310, 2013, pp. 1377 to 1384 . Read level: abstract only. Its study excluded arrests of presumed non-cardiac cause.
19. CARES (Cardiac Arrest Registry to Enhance Survival), CARES Survival Report, National Data, 2025, Non-Traumatic Etiology, dated April 20, 2026, and the CARES 2025 Metrics Summary . Read level: both full, direct, re-read on 2026-09-24. Supports every US registry figure in this lesson and both charts' US bars.
20. Resuscitation Council UK, 2025 Resuscitation Guidelines: Epidemiology of cardiac arrest, 27 October 2025 . Read level: full, direct (the out-of-hospital section), re-read on 2026-09-24. Supports England's 30-day survival in 2024, the share of arrests with a cardiac cause, and the hot spots.
21. S. J. Diem, J. D. Lantos and J. A. Tulsky, "Cardiopulmonary resuscitation on television: miracles and misinformation", New England Journal of Medicine 334, 1996, pp. 1578 to 1582. **Read level: abstract only.**
22. J. Portanova and colleagues, "It isn't like this on TV: revisiting CPR survival rates depicted on popular TV shows", Resuscitation 96, 2015, pp. 148 to 150 . Read level: abstract only.
23. A. L. Blewer and colleagues, "Gender disparities among adult recipients of bystander cardiopulmonary resuscitation in the public", Circulation: Cardiovascular Quality and Outcomes 11, 2018, e004710 . Read level: abstract only.
24. R. A. Garcia and colleagues, "Racial and ethnic differences in bystander CPR for witnessed cardiac arrest", New England Journal of Medicine 387, 2022, pp. 1569 to 1578 . Read level: abstract only.
25. British Heart Foundation, RevivR, and Resuscitation Council UK, Lifesaver. **Read level: both pages full, direct**; neither course was taken.