Current issues of ACP Journal Club are published in Annals of Internal Medicine


Review: In-hospital cardiopulmonary resuscitation rates over time are low

ACP J Club.1994 Jan-Feb;120:21. doi:10.7326/ACPJC-1994-120-1-021

Source Citation

Schneider AP 2d, Nelson DJ, Brown DD. In-hospital cardiopulmonary resuscitation: a 30-year review. J Am Board Fam Pract. 1993 Mar;6:91-101.



To study success rates and risks of in-hospital cardiopulmonary resuscitation (CPR) among elderly persons and other groups and over time (1960 to 1990).

Data sources

Studies published through July 1990 were identified by MEDLINE using the key words cardiopulmonary resuscitation, resuscitation, and heart arrest; bibliographies of relevant papers and reviews were also examined.

Study selection

English-language studies were selected if CPR was done in the hospital, sample size was > 5 adults, survival-to-discharge data were available, and the report was published between July 1960 and July 1990.

Data extraction

Data on number of patients, number of successful resuscitations (survival to discharge), long-term survival rates, age and sex of patients, status of patients (perioperative vs nonoperative), cardiac rhythm, primary diagnosis, length of CPR, complications, type of hospital (teaching vs community), and author optimism with CPR were extracted. Data were pooled for estimating rates.

Main results

19 955 patients in 98 studies had CPR after cardiac arrest; 2994 (15%) resuscitations were successful. The rate of CPR success did not change over time (P < 0.05). Shorter CPR was more successful (29% for CPR < 30 min vs 1% for CPR > 30 min). Younger patients had a higher CPR success rate (16% for < 70 y vs 12% for ≥ 70 y, P < 0.001), with a pooled odds ratio for CPR success for age < 70 years of 1.36 (95% CI 1.20 to 1.53). Sex did not affect success rates. The success rate of patients who had had surgery was higher than for patients who had not had surgery (31% vs 15%, P < 0.001). Cardiac rhythm affected CPR success rates (ventricular fibrillation or tachycardia 20%; electromechanical dissociation 7%, asystole 6%, and other 10%). Patients with several conditions had a low CPR success rate (< 5%): dissecting aneurysm (0%), sepsis (1.8%), central nervous system disorders (3.4%), trauma (3.6%), uremia (4.4%), and cancer (4.9%). Additional patients with pneumonia, congestive heart failure or pulmonary edema had success rates < 13%. Patients with chronic obstructive pulmonary disease, coronary artery disease, or myocardial infarction had moderate success rates (13% to 15%). Patients in shock had a high success rate (29%). Community hospitals had a higher success rate than teaching hospitals (19% vs 14%, P < 0.001). Complications, taken from 14 autopsy reports, were rib fracture (32%), marrow emboli (11%), hemopericardium (5%), and liver or spleen lacerations (5% each). 2% of patients with successful resuscitation had central nervous system impairment. For patients who died in the hospital, 46% died within 24 hours of CPR, 73% within 72 hours, 87% within 1 week; and 98% within 1 month. Among 2009 successful resuscitation patients in whom neurological status was described, < 1% were judged to have severe brain injury.


Cardiopulmonary resuscitation success rates have remained stable at about 15% for the previous 30 years. Age, primary diagnosis, type of hospital, operative status, and length of resuscitation affect success rates.

Source of funding: Not stated.

For article reprint: Dr. A.P. Schneider II, 1401 Harrodsburg Road, Suite B-375, Lexington, KY 40504, USA.


CPR is done on most hospitalized patients having cardiorespiratory arrest. Many studies have shown outcomes after in-hospital CPR and commented on the futility of this intervention for several subgroups of patients. By combining these results, Schneider and colleagues concluded that in-hospital CPR success rates have not improved over the last 3 decades. The authors have also identified other risk factors related to survival, such as age, initial rhythm, operative status, and primary diagnoses.

The authors did a comprehensive review of the literature and used sound methods in doing this meta-analysis. The variable quality and comparability of the original studies, diverse settings, and different population groups, however, must be taken into account. These factors as well as random error in outcome measurements may explain the inability to show a clinically important improvement in CPR success rates over the last 30 years. This study also found age to be inversely related to CPR success rates. It is unclear from the analysis provided whether age alone is a major determinant of survival or to what extent additional risk factors are also independently predictive.

Little doubt exists that CPR saves lives, but survival after cardiac arrest still remains unacceptably low. Success rates are particularly low after unwitnessed arrests, delays in the initiation of advanced life support, asystole or pulseless electrical activity, and in patients with some primary diagnoses. Only the last 2 of these prognostic factors were examined by the authors.

In summary, this overview confirms that CPR success rates are still low, suggesting a need for further research into CPR procedures and improved delivery of advanced life support. This report does not provide evidence supporting the futility of CPR in specific subgroups, although it identifies subgroups where CPR has < 5% chance of success.

Paul Hebert, MD, MHSc
Ottawa General HospitalOttawa, Ontario, Canada