There are 30,000 out-of-hospital cardiac arrests (OHCA) in the UK each year for which resuscitation is attempted.
Ischaemic heart disease is present in 69 per cent and peak incidence is in the morning and evening on Mondays in winter.
Independent risk factors include male gender, previous MI, hypertension, drinking more than 10 cups of coffee per day and having a left ventricular ejection fraction below 40 per cent.
Survivors to discharge from OHCA in the UK number less than 5 per cent, although 73 per cent of these live for more than seven years.
Key predictors of outcome are effective bystander CPR (odds ratio 3.9, 95 per cent CI 1.1-14), having a shockable rhythm on arrival of a defibrillator, early defibrillation (odds ratio 0.92, 95 per cent CI 0.88-0.96), young age, being otherwise well and, surprisingly, having the cardiac arrest outside, rather than in, hospital.
All GPs should be able to perform effective CPR and, ideally, should have immediate access and the ability to use an automated external defibrillator (AED).
On encountering an apparently unconscious adult patient, confirm that it is safe to approach. Look for evidence of what could have harmed the patient - for example, a damaged electrical cable attached to a lawnmower.
Only when you have confirmed it is safe to approach, shout loudly 'Are you alright?' while gently shaking the patient's shoulder.
If there is no response and normal breathing is not clearly evident, position the patient on their back. This advice often creates great concern about worsening a spinal injury: in reality, unstable cervical spine fractures are very rare (incidence less than 0.003 per cent even in severe road accidents) and the most common cause of secondary damage to the cervical cord is hypoxia due to airway obstruction.
Open the airway by tipping the head back and lifting the chin upwards, perpendicular to the surface the patient is lying on.
Look for rise and fall of the chest, listen for breathing with your ear next to the patient's mouth, and feel for airflow against your cheek, all for 10 seconds. While you are doing this keep the head tilted back and the chin lifted, or the airway will obstruct and the patient will be unable to breathe.
The emphasis is on detecting normal breathing: agonal (noisy, irregular, gasping) breaths occur following 40 per cent of cardiac arrests and can continue for several minutes.
If the patient is not breathing normally, diagnose cardiac arrest at this point - it is no longer necessary to check for a carotid pulse as even healthcare practitioners are considered to lack sensitivity and specificity in detecting this sign.
As soon as cardiac arrest is diagnosed, obtain further help - ideally including an AED in the surgery or from a GP's car, and make sure an ambulance has been called.
While awaiting these, start chest compressions (not ventilation). Kneel by the patient's thorax (put them on the floor if they are in bed) and put the heel of one hand in the centre of their chest. Research shows that this instruction results in correct hand positioning far more often than the previously more complex advice to 'find the xiphisternum ...'
Place the other hand on top, keeping your fingers off the chest wall to avoid fracturing ribs at the costochondral margin. Keeping your arms straight, press vertically downwards smoothly to compress the chest by 4-5cm (1.5-2 inches); then allow the chest to recoil fully to its natural resting state (keeping your hands in position).
Ensure compression and recoil take an equal amount of time. Compressions should be delivered at a rate of 100 per minute in groups of 30.
After 30 compressions have been delivered, tip the patient's head back, lift their chin, pinch their nose and, ensuring a good seal with your lips around their mouth, blow in sufficient air to make the chest start to rise in one second.
Remove your mouth and repeat once. If the chest does not rise, check you have followed this sequence precisely - operator error is the most likely cause. Otherwise look in the airway and remove any visible foreign object such as food (a rare finding in practice).
Continue with cycles of 30 compressions to two rescue breaths. Do not expect the patient's heart to restart as this requires defibrillation or drug therapy, so don't give up.
Unfortunately, the quality of your chest compressions will deteriorate due to fatigue, and it is essential (if possible) that the person providing compressions is changed every two minutes.
The emphasis in the latest versions of resuscitation guidelines is very much on forceful uninterrupted chest compressions, since the evidence shows that application of this intervention is most likely to save lives.
Giving mouth-to-mouth is never pleasant, and indeed many would-be rescuers, including doctors, would rather do nothing than provide this treatment.
It is now widely accepted that failing to provide rescue breaths is not negligent, provided continuous chest compressions are given instead.
Indeed, the American Heart Association considers the evidence against giving rescue breaths is now sufficient for them to discontinue teaching mouth-to-mouth, and they have consequently adopted a compression-only CPR algorithm.
Professor Woollard is director of pre-hospital, emergency and cardiovascular care at Coventry University, West Midlands.