Mar 14

Conventional CPR vs chest compression only

Journal reference: Ogawa T, Akahane M, Koike S, et al. Outcomes of chest compression only CPR versus conventional CPR conducted by lay people in patients with out of hospital cardiopulmonary arrest witnessed by bystanders: nationwide population based observational study. BMJ 2011; 342: c7106 [1]


Published: 27 January 2011

Evidence cookie says...

Conventional CPR may be associated with better rates of survival than chest compressions only when conducted by lay people in out of hospital arrests.

However, evidence from observation studies, though interesting, are methodologically weak and are not definitive.

Any CPR is better than none. The Australian Resuscitation Council CPR guideline remains relevant: [2]

  • chest compression ventilation ratio, 30:2
  • if unable or unwilling to provide rescue breaths, chest compression only at 100 per minute

See below for more details.

More details:

Article details

Study design:

retrospective population based observational study

Study aim:

to compare the effectiveness of chest compression only cardiopulmonary resuscitation (CPR) and conventional CPR on outcomes after cardiopulmonary arrest out of hospital

Methods summary:

  • retrospective review of a large national database of out of hospital cardiopulmonary arrests in Japan (a single emergency network covers the entire country)
  • study covered all consecutive arrests in Japan from 2005 to 2007
    • cardiopulmonary arrest outside medical facilities and were transferred to hospitals
    • were classified in the database as:
      • cardiopulmonary arrest confirmed by the emergency medical service on arrival at the incident, or
      • supposed cardiopulmonary arrest in which the patient had already been resuscitated when the emergency medical service arrived
    • baseline data obtained:
      • age and sex of the patient
      • whether the collapse was witnessed
      • whether bystander CPR was performed and the category of the bystander (layperson or emergency staff)
      • whether the dispatcher gave assistance by telephone
      • cause of cardiac arrest (cardiac vs non-cardiac origin)
      • initial identified cardiac rhythm
    • outcome data obtained:
      • survival
      • category of cerebral performance a month after hospital admission
      • cause of cardiac arrest as determined clinically

Inclusion criteria:

  • people with out of hospital cardiopulmonary arrests of cardiac and non-cardiac origin
  • witnessed by laypeople
  • CPR was provided by known layperson bystander(s)

Exclusion criteria:

  • unwitnessed arrest
  • arrest witnessed by emergency services (e.g., fire service staff, emergency specialists)
  • arrest witnessed by unknown bystander
  • lack of bystander CPR
  • rescue breathing only
  • CPR from unknown bystander

Primary outcome:

  • rates of one month survival
  • neurologically favourable one month survival
    • category one (good cerebral performance) and category two (moderate cerebral disability)


  • analyses by 20 year age categories (i.e., ages 0-19, 20-39, etc.)
  • analyses by duration between arrest witnessed and commencement of bystander CPR in 2 minute intervals (i.e., time 0, 1-2, 3-4, etc.)
  • origin of cardiac arrest (cardiac vs non-cardiac)
  • used χ2 test and calculated odds ratios with 95% confidence intervals
  • multiple logistic regression analysis to assess factors associated with better outcomes

Results summary:

  • Out of hospital arrests: 318 141 cases
    • exclusions:
      • not witnessed by bystander: 190 646
      • witnessed by emergency services: 25 521
      • witnessed by unknown bystander: 193
      • no bystander CPR: 56 851
      • CPR by unknown bystander: 3 225
      • rescue breathing only: 1 670
    • chest compression only CPR: 20 707
    • conventional CPR: 19 328

Primary outcome:

  • the results are adjusted for the following confounding factors:
    • age, sex, assistance from dispatcher, initial identified cardiac rhythm, cause of cardiac arrest, relation of the bystander to patient, use of public access automated external defibrillator, first shock from emergency medical staff, use of drug during CPR, and duration between bystander witnessing event to bystander starting CPR, duration to CPR by emergency staff, duration to patient’s arrival at hospital
  • one month survival rates:
    • adjusted OR (odds ratio): 1.17 (95% CI 1.06-1.29), P = 0.002
    • survival rates of patients receiving conventional CPR were higher than those receiving chest compression only CPR
  • neurologically favourable one month survival:
    • adjusted OR: 1.17 (95% CI 1.01 – 1.35), P = 0.037
    • results favour those patients receiving conventional CPR as opposed to chest compression only CPR

Other outcomes:

  • conventional CPR was associated with better outcomes than chest compression only CPR particularly in:
    • younger patients
    • non-cardiac cases
    • increasing delay before the start of CPR

Study conclusion:

Conventional CPR is associated with better outcomes than chest compression only CPR for selected patients with out of hospital cardiopulmonary arrest, such as those with arrests of non-cardiac origin and younger people, and people in who there was a delay in the start of CPR.


  • people in Japan who suffered an out of hospital cardiopulmonary arrest between 2005 and 2007
  • baseline demographics:
    • mean age: 72.9 years
    • sex: men 58.5%
  • differences between those who received conventional CPR and chest compression only CPR
    • those participants who received chest compression only CPR were more likely to have been:
      • male: 61.4% vs 55.3% (P < 0.001)
      • not younger than 20 years: 1.3% vs 2.5% (P < 0.001)
      • have been aged 60-79 years: 39.5% vs 35.4% (P < 0.001)
      • older in general: average age 73.1 vs 72.7 years (P = 0.03)
      • had a longer delay to bystander CPR: 4.1 vs 3.0 minutes (P < 0.001)
      • have had a family member as the bystander: 62.0% vs 46.7% (P < 0.001)
      • not have had an automated external defibrillator (AED) used by bystander: 0.6% vs 2.0% (P < 0.001)

Methodological weaknesses

  • the observational nature of this study naturally limits the statements of causality that can be made
  • the study used registry data and may have included patients without actual cardiopulmonary arrest
    • there was no validation of the registry data
    • this may have biases the results towards overestimating the effect of CPR
  • time to bystander CPR was based on the recall of the bystanders
    • there may be substantial inaccuracies in the judgement of time in a stressful situation
  • the quality of bystander CPR cannot be assessed by this study
  • there is a possibility of unmeasured and unadjusted confounding factors
    • there were substantial differences between the participants who received compression only CPR and conventional CPR
    • those who received compression only CPR would be expected to have a poorer outcomes due to these differences; more likely to be male, older, more likely to have had the arrest at home and attended by family members, less likely to have had access to an AED, more likely to have had a delay in CPR and had assistance from the dispatcher
    • it is conceivable that the bystanders who performed compression only CPR were less proficient at administering basic life support

Methodological strengths

  • large national-wide cohort of consecutive patients reduces the likelihood of a selection bias
  • the number of participants provides greater statistical power than previous studies examining compression only CPR
  • the study provides data on who might benefit most from conventional CPR as opposed to chest compression only CPR
  • cautious conclusion and frank discussion of the weaknesses and limitations of the study

Biases and conflicts of interests

  • nil declared and none seem obvious

Clinical relevance to primary health care

Recently, chest compression only CPR has started to become more popular. There is a prior observational study that reported chest compression only CPR by bystanders was associated with better outcomes [3].

This large nation wide observational study from Japan provides evidence to the contrary. Conventional CPR was associated with a better rate of one-month survival, and neurologically favourable one-month survival than chest compression only CPR.  The benefits were greatest for younger patients, those with a non-cardiac cause arrest, and those with a delay to CPR.  It should be noted that it is conceivable that there is an unadjusted confounding factor favouring conventional CPR that is reflected in the differences between the two groups.

Although the evidence is interesting, it is methodologically weak. It should not be considered definitive evidence favouring conventional CPR.

Nevertheless, a pragmatic approach is suggested by an accompanying editorial in the British Medical Journal: “definitive evidence is lacking, but either is better than no CPR.” [4] The current Cardiopulmonary Resuscitation Guideline (Section 8 ) remains relevant [2].  It recommends a chest compression to ventilation ratio of 30:2, with a compression rate of 100 per minute. Rescuers unable or unwilling to do rescue breathing should perform chest compression only at 100 compressions per minute.


  1. Ogawa T, Akahane M, Koike S, et al. Outcomes of chest compression only CPR versus conventional CPR conducted by lay people in patients with out of hospital cardiopulmonary arrest witnessed by bystanders: nationwide population based observational study. BMJ 2011; 342: c7106
  2. Australian Resuscitation Council and New Zealand Resuscitation Council. Guideline 8: cardiopulmonary resuscitation. Australian Resuscitation Council Online; article retreived 31 January 2011.
  3. SOS-KANTO. Cardiopulmonary resuscitation by bystanders with chest compression only (SOS-KANTO): an observational study. Lancet 2007; 369(9565): 920-26
  4. Jacobs IG. Chest compression over conventional CPR after out of hospital cardiac arrest? [editorial]. BMJ 2011; 2011; 342:d374
Print Friendly, PDF & Email

Leave a Reply

%d bloggers like this: